UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Functional analysis of Pannexin 3 during mammalian and avian osteogenesis, and the physiology of Pannexin… Bond, Stephen Raymond 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2013_spring_bond_stephen.pdf [ 6.12MB ]
JSON: 24-1.0073472.json
JSON-LD: 24-1.0073472-ld.json
RDF/XML (Pretty): 24-1.0073472-rdf.xml
RDF/JSON: 24-1.0073472-rdf.json
Turtle: 24-1.0073472-turtle.txt
N-Triples: 24-1.0073472-rdf-ntriples.txt
Original Record: 24-1.0073472-source.json
Full Text

Full Text

Functional analysis of Pannexin 3 during mammalian and avian osteogenesis, and the physiology of Pannexin 1 ohnologs in teleost fish by Stephen Raymond Bond B.Sc., The University of Waterloo, 2005 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Cell and Developmental Biology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2012  Stephen Raymond Bond 2012 Abstract The pannexins (Panx) are a family of chordate channel proteins with ancient homology to the invertebrate gap junction proteins called innexins. Only three distinct pannexin paralogs have currently been described in the Verte- brata, although all three are represented by orthologs in each of the major vertebrate clades. As public sequence databases have expanded however, a fourth pannexin sequence has revealed itself in teleost fish, apparently repre- senting a duplication of Panx1. The first set of objectives for this thesis was to assess the likely genetic source of this duplication event, determine if and where the two Panx1 paralogs (denoted panx1a and panx1b) are expressed in a model teleost (zebrafish), and characterize some of the physiological properties that differ between the two gene products. Data will be pre- sented indicating that the whole genome duplication event punctuating the teleost radiation ∼ 350 million years ago is the likely source of this duplica- tion, panx1a is widely distributed while panx1b is expressed primarily in the brain and eye, and neo- and/or subfunctionalization is discernible through differences in subcellular trafficking and distinct electrophysiological charac- teristics. The second set of objectives was the generation and dissemination of a public web based graphical user interface for the automatic design of restriction-free cloning primers, which is a method heavily used for the gen- eration of custom DNA plasmids in this thesis. The third set of objectives focused on characterizing Panx3 expression in osteogenic tissues. In par- ticular, determining how Panx3 expression is regulated in osteoblasts and hypertrophic chondrocytes, and describing the physiological consequences of this expression. Using in vitro techniques, it was determined that the osteogenic transcription factor Runx2 is necessary, but not necessarily suf- ficient, to promote expression of Panx3. To identify the functional impor- tance of the protein in vivo, a replication competent avian retrovirus was used to force over-expression and to deliver shRNA into embryonic chicken forelimbs. Increased expression had no overt impact on bone development, while knockdown resulted in a bone dysplasia, manifested as a reduction in total mineralized bone volume, and correlated with reduced expression of the hypertrophic chondrocyte marker COL10A1. ii Preface Versions of chapters 2–4 have been published in peer reviewed journals:  Chapter 2: Pannexin 1 ohnologs in the teleost lineage. S. R. Bond, N. Wang, L. Leybaert, and C. C. Naus. J Membr Biol, 245(8):483–93, Aug 2012.  Chapter 3: Rf-cloning.org: an online tool for the design of restriction- free cloning projects. S. R. Bond and C. C. Naus. Nucleic Acids Res, 40(Web Server issue):W209W213, Jul 2012.  Chapter 4: Pannexin 3 is a novel target for runx2, expressed by os- teoblasts and mature growth plate chondrocytes. S. R. Bond, A. Lau, S. Penuela, A. V. Sampaio, T. M. Underhill, D. W. Laird, and C. C. Naus. J Bone Miner Res, 26(12):291122, Dec 2011. Several co-authors have contributed directly in performing experiments and/or collecting data, and assisting in interpretation:  Alice Lau manually acquired thousands of images for figure 4.1, which I later merged together into a montage.  Dr. Arthur Sampaio provided me with Runx2 expression plasmids and generated the microarray data included in figure 4.5C.  Dr. Silvia Penuela and Dr. Dale Laird furnished me with the antibod- ies that made chapter 4 possible.  Nan Wang performed the electrophysiological recordings behind fig- ures 2.8, 2.9, and 2.10, and was instrumental, along with Dr. Luc Leybaert, in helping me understand the physiological relevance of the data.  Kathy Fu handled most of the radiation for the situ experiments in chapter 5, and regularly assisted with the sample collection and pro- cessing for that chapter. All other experiments were performed by me, and I was responsible for writing each of the manuscripts. Figures were produced by me, unless oth- erwise credited. iii Table of Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1 Gap junctions . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.1 Connexins . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.2 Innexin . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2 Pannexins . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.2.1 Discovery . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.2.2 Biochemical properties . . . . . . . . . . . . . . . . . 14 1.2.3 The pannexin channel . . . . . . . . . . . . . . . . . . 23 1.2.4 Expression and distribution . . . . . . . . . . . . . . 28 1.2.5 Physiological and pathological relevance of pannexins 34 1.3 Bone development . . . . . . . . . . . . . . . . . . . . . . . . 39 1.3.1 Intramembranous ossification . . . . . . . . . . . . . . 41 1.3.2 Endochondral ossification . . . . . . . . . . . . . . . . 43 1.3.3 Connexins in bone . . . . . . . . . . . . . . . . . . . . 46 1.3.4 Pannexins in bone . . . . . . . . . . . . . . . . . . . . 47 1.4 Motivation, objectives, and highlights . . . . . . . . . . . . . 48 iv 2 Pannexin 1 ohnologs in the teleost lineage∗ . . . . . . . . . 50 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . 52 2.2.1 Phylogenetic analysis . . . . . . . . . . . . . . . . . . 52 2.2.2 Real-time qPCR . . . . . . . . . . . . . . . . . . . . . 53 2.2.3 Cloning zebrafish pannexins . . . . . . . . . . . . . . 54 2.2.4 Cell culture . . . . . . . . . . . . . . . . . . . . . . . 54 2.2.5 Western blot . . . . . . . . . . . . . . . . . . . . . . . 54 2.2.6 Visualizing pannexin-EGFP . . . . . . . . . . . . . . 57 2.2.7 Electrophysiological recording . . . . . . . . . . . . . 57 2.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . 57 2.3.1 At least four pannexin genes are present in the teleost lineage . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.3.2 The two teleost panx1 genes likely originate from the R3 WGD event . . . . . . . . . . . . . . . . . . . . . 61 2.3.3 Distinct expression profiles of panx1a and panx1b . . 64 2.3.4 Subcellular dynamics and localization of the zebrafish pannexins . . . . . . . . . . . . . . . . . . . . . . . . 66 2.3.5 Physiological properties of zebrafish Panx1 channels . 67 2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 3 RF-Cloning.org: An online tool for the design of restriction- free cloning projects∗ . . . . . . . . . . . . . . . . . . . . . . . 74 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 3.2 Project workflow . . . . . . . . . . . . . . . . . . . . . . . . . 75 3.3 Recommended rf-cloning protocol . . . . . . . . . . . . . . . 79 3.4 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . 81 3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4 Pannexin3 is a novel target for Runx2, expressed by os- teoblasts and mature growth plate chondrocytes∗ . . . . . 82 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . 84 4.2.1 Animal care . . . . . . . . . . . . . . . . . . . . . . . 84 4.2.2 Antibodies . . . . . . . . . . . . . . . . . . . . . . . . 84 4.2.3 Cell culture . . . . . . . . . . . . . . . . . . . . . . . 84 4.2.4 Immunofluorescence . . . . . . . . . . . . . . . . . . . 85 4.2.5 Microarray . . . . . . . . . . . . . . . . . . . . . . . . 85 4.2.6 Western blot . . . . . . . . . . . . . . . . . . . . . . . 86 4.2.7 Promoter alignment . . . . . . . . . . . . . . . . . . . 86 v 4.2.8 Plasmids . . . . . . . . . . . . . . . . . . . . . . . . . 86 4.2.9 Transcription reporter assay . . . . . . . . . . . . . . 87 4.2.10 Real-time qPCR . . . . . . . . . . . . . . . . . . . . . 88 4.2.11 Chromatin immunoprecipitation . . . . . . . . . . . . 88 4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.3.1 Panx3 is expressed in both endochondral and intramem- branous bone during embryonic development . . . . . 89 4.3.2 Panx3 is expressed by pre-hypertrophic chondrocytes, hypertrophic chondrocytes, and mature osteoblasts . 89 4.3.3 The Panx3 promoter is responsive to Runx2 . . . . . 95 4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5 In vivo role of Panx3 during endochondral ossification . . 108 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 5.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . 109 5.2.1 Plasmid construction . . . . . . . . . . . . . . . . . . 109 5.2.2 Cell culture . . . . . . . . . . . . . . . . . . . . . . . 111 5.2.3 Western blot . . . . . . . . . . . . . . . . . . . . . . . 111 5.2.4 Virus preparation . . . . . . . . . . . . . . . . . . . . 112 5.2.5 Infection of developing forelimb . . . . . . . . . . . . 112 5.2.6 Real-time qPCR . . . . . . . . . . . . . . . . . . . . . 112 5.2.7 Whole mount in situ hybridization . . . . . . . . . . . 114 5.2.8 Radioactive in situ hybridization on tissue sections . 114 5.2.9 Histological analysis of mineralized forelimb bones . . 114 5.2.10 Cell division assay . . . . . . . . . . . . . . . . . . . . 115 5.2.11 Optical projection tomography . . . . . . . . . . . . . 115 5.2.12 Statistical analysis . . . . . . . . . . . . . . . . . . . . 115 5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.3.1 Panx3 is expressed in intramembranous and endochon- dral bone in chick . . . . . . . . . . . . . . . . . . . . 116 5.3.2 In vivo knock-down of PANX3, but not over-expression, alters long bone morphology during development . . . 118 5.3.3 Panx3 knockdown has no effect on the expression of genes involved in chondrocyte proliferation or early differentiation . . . . . . . . . . . . . . . . . . . . . . 121 5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 5.4.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 131 5.5 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 131 vi 6 Extended discussion and concluding remarks . . . . . . . . 132 6.1 Panx1a and Panx1b are probably ohnologs . . . . . . . . . . 132 6.2 Greater functional diversity through gene duplication . . . . 133 6.3 Effects of epitope tags . . . . . . . . . . . . . . . . . . . . . . 135 6.4 Regulation of the Panx3 promoter requires, but also tran- scends Runx2 . . . . . . . . . . . . . . . . . . . . . . . . . . 136 6.5 RCASBP delivery of shRNA in vivo . . . . . . . . . . . . . . 138 6.6 The effects of Panx3 on endochondral bone development . . 139 6.7 Panx3 C-terminal leucine zipper . . . . . . . . . . . . . . . . 143 6.8 A place for RF-Cloning.org in the molecular biology commu- nity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 6.9 The future of RF-Cloning.org . . . . . . . . . . . . . . . . . . 147 6.10 Personal appeal for greater computational literacy in the life sciences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 6.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Appendix A Journal club publication — Passing potassium with and without gap junctions . . . . . . . . . . . . . . . . . . . . . . . 201 vii List of Tables 1.1 Connexin nomenclature . . . . . . . . . . . . . . . . . . . . . 5 1.2 Amino acid conservation of pannexin orthologs between species 17 1.3 Amino acid conservation between pannexin sequences within a given species . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.1 Pannexin protein sequences . . . . . . . . . . . . . . . . . . . 53 2.2 Zebrafish qPCR primers . . . . . . . . . . . . . . . . . . . . . 55 2.3 Cloning primers . . . . . . . . . . . . . . . . . . . . . . . . . . 56 5.1 Chicken qPCR primers . . . . . . . . . . . . . . . . . . . . . . 113 viii List of Figures 1.1 Connexin and pannexin schematics . . . . . . . . . . . . . . . 7 1.2 Phylogenetic relationship between pannexins . . . . . . . . . 16 1.3 Primary structure of the three chicken pannexin sequences . . 19 1.4 Identification of 70 kDa Panx3 Western blot band in FBS . . 21 1.5 Osteoblast and chondrocyte differentiation . . . . . . . . . . . 40 1.6 Intramembranous ossification . . . . . . . . . . . . . . . . . . 42 1.7 Endochondral ossification . . . . . . . . . . . . . . . . . . . . 44 2.1 Multiple pairwise alignment of Panx1 sequences . . . . . . . . 59 2.2 Phylogram illustrating the phylogenetic relationship between pannexins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 2.3 Genomic architecture of the two zebrafish panx1 genes . . . . 61 2.4 Syntenic relationship between the zebrafish and mouse chro- mosomal regions containing panx1 genes . . . . . . . . . . . . 63 2.5 Relative mRNA levels of pannexins expressed by zebrafish tissues, as measured by qPCR . . . . . . . . . . . . . . . . . . 65 2.6 Relative mRNA levels of pannexins expressed by zebrafish tissues, as measured by qPCR . . . . . . . . . . . . . . . . . . 67 2.7 Exogenous expression of EGFP-tagged zebrafish pannexins in HeLa cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 2.8 Whole-cell voltage clamp of EGFP, Panx1a-EGFP, and Panx1b- EGFP transfected HeLa cells . . . . . . . . . . . . . . . . . . 71 2.9 Single channel recordings of Panx1b-EGFP demonstrating unitary event activity and single channel conductance . . . . 72 2.10 Panx1a-EGFP is sensitive to carbenoxolone . . . . . . . . . . 73 3.1 Schematic of a typical restriction free cloning protocol . . . . 76 3.2 RF-Cloning.org input page . . . . . . . . . . . . . . . . . . . 78 3.3 RF-Cloning.org output page . . . . . . . . . . . . . . . . . . . 80 4.1 Expression of Panx3 in stages E13-E15.5 mouse embryos . . . 90 4.2 Expression of Panx3 in a stage E17-17.5 mouse embryo . . . 91 4.3 The Panx3 antibody is blocked by pre-absorption with anti- genic peptide . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 ix 4.4 Localization of Panx3 and markers of long bone development in a representative long bone . . . . . . . . . . . . . . . . . . 94 4.5 Panx3 expression is induced during osteoblast differentiation 96 4.6 The Panx3 promoter contains putative binding sites for tran- scription factors associated with bone development . . . . . . 98 4.7 Promoter truncations . . . . . . . . . . . . . . . . . . . . . . . 99 4.8 Runx2 induces reporter expression from the Panx3 promoter in HEK-293 cells . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.9 Runx2 induces reporter expression from the Panx3 promoter in primary osteoblasts . . . . . . . . . . . . . . . . . . . . . . 101 4.10 Chromatin-IP indicates Runx2 association with Panx3 pro- moter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.11 Runx2 is insufficient to induce Panx3 expression in fibroblasts 103 4.12 Panx3 is expressed in a limited number of non-osteogenic tissues104 5.1 Panx3 in situ hybridization during chick development . . . . 117 5.2 Panx3 over-expression in chicken embryo . . . . . . . . . . . . 119 5.3 Knockdown of PANX3 using RCASBP::mir-30a constructs . 120 5.4 OPT scans of stage 41 forelimb bones . . . . . . . . . . . . . 122 5.5 Histomorphometric analysis of RCASBP::PANX3 -14 infected forelimbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 5.6 Stage 35 forelimb radio in situs: PANX3, PTCH, BMP7 . . . 126 5.7 Stage 35 forelimb radio in situs: LEF1, RUNX2, IBSP . . . . 127 5.8 Stage 35 forelimb radio in situs: WNT5A, WNT5B, MMP13 128 5.9 Real-time qPCR for bone genes in forelimb . . . . . . . . . . 129 6.1 Panx3 promoter binding elements . . . . . . . . . . . . . . . . 137 6.2 pENTR3C-miR-30a construct . . . . . . . . . . . . . . . . . . 140 6.3 Panx3 leucine zipper . . . . . . . . . . . . . . . . . . . . . . . 144 6.4 Plasmid map output comparison between original Savvy out- put and new RF-Cloning.org . . . . . . . . . . . . . . . . . . 146 6.5 RF-Cloning.org plasmid management dashboard . . . . . . . 148 6.6 RF-Cloning.org site traffic . . . . . . . . . . . . . . . . . . . . 150 6.7 Human knowledge vs. PhD . . . . . . . . . . . . . . . . . . . 152 x Acronyms Kvβ3 voltage-dependent potassium channel β subunit. ΔCT delta-CT. ΔΔCT delta-delta-CT. [Ca2+]i intracellular calcium concentration. [Ca2+]o extracellular calcium concentration. [K+]o extracellular potassium concentration. I-V non-linear current-to-voltage relationship. Acan aggrecan. AJAX asynchronous JavaScript and XML. ALP alkaline phosphatase. Alpl alkaline phosphatase. ANOVA analysis of variance. ATP adenosine-5’-triphosphate. BCA bicinchoninic acid. Bglap osteocalcin. BMP bone morphogenic protein. bps basepairs. BrdU Bromodeoxyuridine. bzATP 3’-benzoylbenzoyl adenosine 5-triphosphate. xi CaMKII calcium/calmodulin-dependent protein kinase II. cAMP cyclic adenosine monophosphate. CBX carbenoxolone. cDNA complementary DNA. CDS coding sequence. cKO conditional knockout. co-IP co-immunoprecipitation. Col1α1 collagen type I alpha 1. Col10α1 collagen type X alpha 1. Col2α1 collagen type II alpha 1. CREB cAMP-responsive element binding protein. CT carboxy terminus. Cx connexin. DIG Digoxigenin. DM1 Dermo1/Twist2. DNA deoxyribonucleic acid. Dre15 Danio chromosome 15. Dre5 Danio chromosome 5. ECM extracellular matrix. EO endochondral ossification. ER endoplasmic reticulum. FACS fluorescence-activated cell sorting. FBS foetal bovine serum. FGF fibroblast growth factor. xii GEO Gene Expression Omnibus. GFAP glial fibrillary acidic protein. GFP green fluorescent protein. HEK-293 human embryonic kidney. HP hydrostatic pressure. HRPO horseradish peroxidase. Hspg2 Perlecan. Ibsp integrin binding sialoprotein. IC50 50% inhibitory concentration. Ihh Indian hedgehog. IL-1α interleukin-1α. IL-1β interleukin-1β. Inx innexin. IO intramembranous ossification. IP immunoprecipitation. IP3 inositol 1,4,5-trisphosphate. LPS lipopolysaccharide. MC3T3-E1 mouse calvarial pre-osteoblast cells. Mmu9 mouse chromosome 9. mRNA messenger RNA. MYA million years ago. N-CAM neural cell adhesion molecule. NCBI National Center for Biotechnology Information. xiii NeuN neuronal nuclear antigen. NFM-T 5% non-fat milk + 0.1% Tween20. NIH National Institute of Health. NMDA N -methyl-d-aspartate. NPPB 5-nitro-2-(3-phenylpropylamino)benzoic acid. nt nucleotide. ODDD oculodentodigital dysplasia. OPT optical projection tomography. P2X7R purinergic receptor P2X ligand-gated ion channel, 7. PAMP pathogen-associated molecular pattern. Panx pannexin. PAR-1 protease-activated receptor-1. PCR polymerase chain reaction. PFA paraformaldehyde. PM plasma membrane. Pth1r parathyroid hormone 1 receptor. Pthlh parathyroid hormone-like hormone. qPCR quantitative real-time PCR. RA retinoic acid. RCASBP Replication Competent ALV LTR with a Splice acceptor. RIPA radioimmunoprecipitation assay. rRNA ribosomal RNA. RT-PCR reverse-transcription PCR. xiv Runx2 Runt-related transcription factor 2. SCAM substituted cysteine accessibility method. SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis. SEM standard error of the mean. shRNA small-hairpin RNA. SOAP Simple Object Access Protocol. Sox Sry-related HMG box. Sp7 osterix. Spp1 osteopontin. SVG scalable vector graphics. Tm melting temperatures. TESPA (3-Aminopropyl)triethoxysilane. TGF-β transforming growth factor β. UPGMA unweighted-pair-group method with arithmetic mean. URL uniform resource locator. Vj transjunctional voltages. Vm membrane potential. VDR/RXR vitamin D3 receptor - retinoid-X receptor heterodimer. VEGFR vascular endothelial growth factor receptor. WGD whole genome duplication. WISH whole mount in situ hybridization. Wnt wingless-type MMTV integration site family. WT wild type. XML extensible markup language. xv Acknowledgements The relationships and friendships forged over the past 7 years will stay with me for as long as I draw breath, and perhaps longer if I can get myself implanted into a computer. . . To begin, I thank Dr. Christian Naus for bringing me into his lab. Neither of us knew what we were getting into back in the winter of ‘04, when we sealed the deal with a couple emails and a phone call. Thank you for taking the risk on me, and giving me the chance to spread my wings. It’s been a wild ride at times, but you’ve always been unwaveringly supportive. I would also like to express my deepest gratitude to Dr. Joy Richman, for her tremendous contributions to the latter half of my degree. Working with you and your lab has been one of the most gratifying experiences of my PhD, and I thank you for taking on an unofficial supervisorial role. My only hope is that the work we have done together continues to bear fruit for you in the future. I must also acknowledge Suresh Nimmagadda and Kathy Fu from Joy’s lab, for all their technical support and mentorship. You guys are the best. I would like to thank my ‘lab parents’, John and Lynne Bechberger for all the little things that keep a family together, be it birthday cakes or afternoon chats about whatever in TC. To keep the family metaphor going a bit longer, Dr. Sin is the bossy yet ardently supportive Aunt that you always want to do cool things with. We’ve had some great discussions Sin (occasionally heated!) and I know you’ll make a wonderful faculty member one day, with droves of athletic underlings. To Dr. Vince Chen, what a treat it’s been to work with you on our various little projects, and to be able to put my programming experience to use! I’ll always appreciate your willingness to help anyone and everyone, along with your enthusiasm for science and life. Many thanks to the graduate students of the Naus lab, past and present (Dave Bates, Charles Lai, Cima Cina, Mike Kozoriz, Annie Aftab, and Max Le Vasseur), for fostering a dynamic and engaging atmosphere. Our Friday afternoon wine and science sessions were always enlightening, and if I ever have graduate students of my very own, I’ll be sure to encourage the activity. The sundry of undergraduate students who passed through the lab will also always have a special place in my heart. Of particular note are Alice Lau and Esther Park, who I was privileged enough to supervise personally. I hope I was able to pass on some knowledge, and to fan the spark you both possess that just might blossom xvi into a true love for scientific inquiry. To our collaborators at the University of Western Ontario, Dr. Dale Laird and Dr. Silvia Penuela, thank you for the antibodies and the feedback on our manuscript, and also to Dr. Michael Underhill and Dr. Arthur Sampaio here at UBC, for the expertise and constructs. To our more recent collaborators at the University of Ghent, Nan Wang and Dr. Luc Leybaert, it was such a pleasure working with you both. I would also like to acknowledge my advisory committee (Dr. Doug Allan, Dr. Michael Underhill, Dr. Joy Richman, and Dr. Robert Nabi) for trying to keep me on focus, as well as The Michael Smith Foundation and NSERC, the administrators of which felt I was worthy of personal funding. And finally, my most passionate thanks go out to my lovely wife Kristen. We’ve stood beside one another throughout our respective degrees, giving each other strength in the darker times, and celebrating the brighter ones. We may not know where our careers will lead us now that we have these letters behind our names, but at least we’ll be going there together. xvii Dedication I dedicate this work to my parents Robert and Barbara Bond, who en- couraged me from the very beginning to pursue an education and then do something inspired; perhaps even inspiring. You have always been support- ive of my dreams, and never prodded me with the threat of disappointment. Somehow, I think that made me work all that much harder to never disap- point you. I love you both. xviii Forward “You never know what life is like, until you have lived it.” -Marilyn Monroe I would like to take a moment to discuss the structure of this thesis, and to place you, the reader, into a similar mindset as myself from the very beginning. The project(s) I set out to work on during my degree have morphed and wavered insistently, and for better or worse, my direction has often been at the mercy of my interest at any given time. While this has definitely caused me some grief, and perhaps taken an extra year or two of my life, the education I will leave with has probably been more rounded as a result. What you are about to read is an amalgamation of three distinct topics, whose details I will defer to the introduction in chapter 1, but which generally have pannexins as their unifying thread. Chapters 2, 4, and 5 in particular focus on pannexin, while chapter 3 is a bioinformatics project that played a role in a lot of the molecular cloning I performed over the years. The pannexin literature is a little strange, in that it is almost always intimately tied to the connexin (i.e., gap junction) literature. This thesis will be no different, so I want to impress upon you up front that pannexins are not connexins. Maybe, in some tropical sea that existed in deep history, the two genes could be found to share a common ancestral origin that predisposed them towards forming low resistance aqueous pores, but this is truly a case of convergent evolution; any contemporary similarities are the result of independent innovation. Sometimes however, it may feel like the pannexin community is paying ritualistic lip service to the connexin field by referring back to it almost ubiquitously in publications, perhaps as a way of showing gratitude for the significant role it played in giving us our start. I think the actual reason is far more pragmatic though, and is based in the fact that we can draw from the rich history of connexin research as a scaffold over which new pannexin work can be stretched. As an analogy, de novo genomic assembly from shotgun sequencing is quite difficult for anything larger than a bacterial chromosome, both in terms of 1 algorithmic efficiency and brute force computing requirements, but the task becomes orders of magnitude easier if you have a reference genome already in hand. The new sequencing data can be aligned to the reference sequence, and thus neatly indexed, instead of begin held fragmented in memory as it awaits integration into a contig. So too can new pannexin studies be neatly indexed against the already mature connexin field, instead of trying to independently identify interesting target pathways and re-design every tool/technique de novo. For this reason, chapter 1 begins with a historical overview of the gap junction field, which is used as a primer and a segue for a comprehensive review of how the pannexin field has developed since its birth a little over twelve years ago. As the title of the thesis suggests, the ultimate focus (or at least the ‘headlining’ focus) is how Panx3 is involved in bone development, so the final section of the introduction will describe bone development, along with what is currently known about how connexins and pannexins effect the process. Chapters 2–4 are formatted as manuscripts, each with its own indepen- dent introduction, materials/methods, results, and discussion section (al- though the results and discussion are merged in chapters 2 and 3). Because the results are discussed in each chapter, the final chapter has been ded- icated to expanding upon these discussions; exploring weaknesses in each study and proposing future directions. Also, I have taken some stylistic lib- erty in this concluding chapter, often switching away from the traditional scientific passive voice in favour of the first person. Hopefully my change in tone can be pardoned — This is the largest document I have ever produced, and I would really like to be a part of it. In closing I would like to thank you dear reader, for opening this thesis and peering inside. Most likely you are a member of my examining com- mittee, and to you I give a special thanks for taking so much time out of your busy schedule to critically review my work. To anyone else perusing these words, whatever it is that has brought you to them, I sincerely hope you find a smidgen of data or a sliver of insight that helps you in your own studies, because then I will know I have helped push science forward. —Stephen R. Bond 2 Chapter 1 Introduction “An education is learning what you didn’t even know you didn’t know.” -Daniel J. Boorstin 1.1 Gap junctions In 1959, Edwin Furshpan and David Potter observed action potentials prop- agating through the crayfish giant motor synapse with a delay of 0.1 ms, or about an order of magnitude faster than synaptic delays previously recorded. They also demonstrated that hyperpolarization of the post-synaptic fibre could ‘leak’ backwards into the pre-synaptic fibre [1]. These observations did not adhere to the canonical view of the synapse, whereby action poten- tials were understood to propagate unidirectionally via chemical intermedi- aries [2, 3]. The authors postulated a ‘synaptic rectifier’ directly linking the cells, leading to the eventual discovery of intercellular gap junctions. These large aqueous pores span the plasma membranes of adjacent cells, creating a cytoplasmic syncytium that ions and small molecules can rapidly diffuse through. There are two large families of unrelated (yet structurally con- vergent) transmembrane proteins within the Metazoa that oligomerize into gap junctions, and these have been named the connexins (Cxs) and innexins (Inxs) [4]. 1.1.1 Connexins The Cxs are a large family of channel forming proteins found solely in Chor- dates. The primary role of Cxs is the creation of gap junctions, linking the intracellular compartments of adjacent cells to facilitate rapid commu- nication and to maintain balance in the concentration of many ions and small molecules. However, the cytoplasmic components of Cx are also a rich source of protein-protein interactions, and the physiological relevance of Cxs 3 is extensive, reaching well beyond that of a simple passive pipeline between cells. Nomenclature Two Cxs nomenclature systems have been historically used in the literature. One is based on the molecular weight of monomeric subunits (e.g., Cx43 for the 43 kDa isoform), while the other groups genes into paralogous clades and numbers each according to its order of discovery (e.g., Gja1 for Cx43, and Gjb1 for Cx32) [5]. The molecular weight system has been more widely adopted, and will be the convention used herein, but it should be noted that confusion can arise when comparing orthologous genes because of changes in respective protein mass over evolutionary time. In these cases, the ’Gj’ nomenclature formalized by the Human Genome Organization Gene Nomen- clature Committee will be referred to. The respective gene name mappings for the two systems have been summarized for convenience in Table 1.1. Molecular structure Each monomeric Cx is characterized by four transmembrane domains with cytoplasmic C- and N-termini. Following translation, and during transit towards the plasma membrane (PM) through the endoplasmic reticulum (ER) and Golgi apparatus, six Cx subunits oligomerize into a tightly reg- ulated channel structure called a ‘connexon’ [6]. Once integrated into the plasma membrane, connexons from each of two adjacent cells are able to form highly stable electrostatic interactions, creating a mature gap junction channel (Figure 1.1). Once formed, gap junctions aggregate into distinct foci called ‘plaques’, forcing the membranes of the linked cells into such close proximity, that in 1962 Maynard Dewey and Lloyd Barr mistakenly concluded that the outer lamellae of these juxtaposed membranes were fused into a single uniform layer [7]. Electron microscopy was used to observe these areas of contact, and they were given the name ‘nexus’. A year later (1963), David Robertson published a collection of higher resolution electron micro- graphs illustrating the nexus at Mauthner cell synapses in goldfish, which he called the ‘synaptic disc’ [8]. The outer membrane lamellae were shown to remain intact, but were brought to within 30-50Å of one another instead of the usual 100-200Å inter-membrane gap of the synaptic cleft. By rotating the angle of view of the synaptic disc 90, a regular hexagonal subunit pat- tern emerged, repeating with a period of about 95Å — clearly illustrating the hexameric nature of gap junctions within these plaques. The link be- 4 Human Mouse Chicken GJA1 CX43 Gja1 Cx43 GJA1 CX43 GJA3 CX46 Gja3 Cx46 GJA3 CX46 GJA4 CX37 Gja4 Cx37 GJA4 CX39 GJA5 CX40 Gja5 Cx40 GJA5 CX40 Gja6 Cx33 GJA8 CX50 Gja8 Cx50 GJA8 Cx50 GJA9 CX59, CX58 GJA10 CX62 Gja10 Cx57 GJA10 CX62 GJB1 CX32 Gjb1 Cx32 GJB1 CX32 GJB2 CX26 Gjb2 Cx26 GJB3 CX31 Gjb3 Cx31 GJB4 CX30.3 Gjb4 Cx30.3 GJB5 CX31.1 Gjb5 Cx31.1 GJB5 CX31.1 GJB6 CX30 Gjb6 Cx30 GJB6 CX31 GJB7 CX25 GJC1 CX45 Gjc1 Cx45 GJC1 CX45 GJC2 CX47, CX46.6 Gjc2 Cx47 GJC2 CX47 GJC3 CX30.2 Gjc3 Cx29 GJD2 CX36 Gjd2 Cx36 GJD2 CX36 GJD3 CX31.9 Gjd3 Cx30.2 GJD3 CX31.9 GJD4 CX40.1 Gjd4 Cx39 GJD4 CX40.1 GJE1 CX23 Gje1 Cx23 Table 1.1: Two nomenclature conventions are present in modern connexin literature, using either molecular weight or evolutionary relationships, and the link between each is summarized here for the human, mouse, and chicken genes. This data was retrieved from the NIH GenBank, June 2012 5 tween regions of close membrane apposition and ‘electrotonic transmission’ was reported by Michael Bennett soon after [9], and in 1967 Jean-Paul Revel coined the term ‘gap junction’ [10]. When a connexon is composed entirely from a single Cx isoform it is clas- sified as ‘homomeric’, but with more than 20 Cx isoforms in mammals (many with overlapping expression patterns), ‘heteromeric’ connexons are a com- mon occurrence [11–15]. The composition of apposing connexons within a given gap junction can also be either the same (homotypic) or different (het- erotypic) [16], allowing for a lot of combinatorial diversity in gap junction architecture. Among the various Cx homologs, the N-terminus, transmem- brane domains, and extracellular loops share the greatest level of sequence conservation [17]. Of particular note are three cysteine residues in each of the two extracellular loops (spaced CX6CX3C in the first and CX4CX5C in the second), which link the loops of a monomer together through disulphide bonds [18–20] to facilitate gap junction formation/function [21]. Conversely, there is considerable diversity within the intracellular loop and C-terminus of Cxs. These are the regions where most posttranslational modifications occur, such as phosphorylation [22] and ubiquitination [23], as well as many protein-protein interactions [24]. The first crystal structure of a gap junc- tion was solved to 7.5Å in 1999 by Vinzenz Unger using a carboxy-terminal truncation of Cx43 [25], and Cx26 was more recently solved in its entirety to a resolution of 3.5Å by Shoji Maeda [26]. The Cx26 model informs us that the intracellular entrance of the pore is constructed from residues of transmembrane domains 2 and 3, with the cytoplasmic N-terminus fold- ing inward to line the entrance funnel with positively charged side chains. Deeper inside the channel, the luminal wall is primarily constructed from the first transmembrane domain and first extracellular loop. 6 CH H H S S C H H H SC H H H S C H H H C H H HS SC H H H Connexins Pannexins C H H HS S C H H H SC H H H C H H H S → N-Linked Glycosylation S C H H H → Cysteine Side Chain Figure 1.1: Cxs and Panxs are both understood to reside primarily in the plasma membrane as tetra-spanning monomers, which oligomerize into hexameric transmembrane pores. Cxs ‘hemichannels’ at the surface of adjacent cells readily associate with one another to form gap junctions to create a cytoplasmic syncytium, but it is generally accepted that glycosylation of Panxs prevents similar coupling through steric interference 7 Biophysical properties of the intercellular channel Historically, gap junctions have been reported to allow passage of atomic ions and small molecules of up to about 1.2 kDa in size [27], but this view is largely derived from work with Cx43 (the most commonly expressed iso- type), and is both overly simplistic, and not representative of Cx diversity. Each isotype has its own unique conductive properties when incorporated into homomeric gap junctions, ranging from highly selective to broadly per- missive with unitary conductances from as little as 10 pS to more than 350 pS [28], and the picture is only complicated further when heteromeric and heterotypic gap junctions are considered [11, 12, 14, 15]. However, once this diversity is recognized as a caveat, it is reasonable to state in general terms that gap junctions are able to facilitate the passage of nearly all physio- logically relevant atomic ions and soluble cytoplasmic molecules under 1.2 kDa in size; including K+, Ca+, Cl−, inositol trisphosphate, glucose (and its metabolites), nucleotides, small interfering RNAs, amino acids, and even small peptides [29, 30]. It is tempting to assume that channels with the highest unitary conduc- tance will also permit the passage of the largest solutes, but only a weak correlation actually exists between these two properties. For example, Cx37 has a unitary conductance (using KCl as the internal pipette solution) that is an order of magnitude greater than that of Cx45 [31], yet both are com- parably permeable to a number of dyes and tracers ranging in mass from 279 Da to 443 Da, with ionic charges ranging from -2 to +2 [31, 32]. In fact, Cx45 is considerably more permeable to the fluorescein derivative diCl-F than Cx37 [31]. This discrepancy is a product of the amino acid side chains lining the entrance and lumen of the pore, which create a complex internal topology (both polar and steric). All else being equal, the conductance of a channel (i.e. the amount of current at a given voltage) is directly pro- portional to its length and the square of the channel’s average radius. An interesting consequence of this physical relationship is that the conductive potential of a channel to very small particles is only marginally effected by a localized constriction within the pore, and yet, such a feature will com- pletely block the transit of larger species that can still enter the channel and occupy space on either side of the narrowing [28]. Gating A number of chemical gating mechanisms exist that regulate the gap junc- tion pore [33]. For example, a drop in intracellular pH facilitates dimeriza- 8 tion of adjacent connexin carboxy-terminal tails within a connexon, which subsequently interact with their own intracellular loop domains to block channel conductance [34, 35]. Intracellular calcium concentration ([Ca2+]i) rises have a more indirect inhibitory effect on mature gap junctions, requiring calmodulin as an intermediary [36], but as we will see in section 1.1.1, extra- cellular calcium concentration ([Ca2+]o) has a critical roll in modulating the open probability of unpaired connexons. A wide range of pharmacological agents have also been used over the years to block gap junction channels, including glycyrrhetinic acid derivatives (carbenoxolone (CBX) in particu- lar), mefloquine, volatile anaesthetics (e.g., halothane and ethrane), certain straight chain fatty alcohols (e.g., heptanol and octanol), and cyclodextrins [37]. Gap junctions are also strongly regulated by a number of voltage gating mechanisms [38]. Some (but not all) isotypes are weakly sensitive to ab- solute membrane potential (Vm) between the cytoplasm and extracellular space [39, 40], while all isotypes have the ability to sense transjunctional voltages (Vj) between coupled cells [41]. There are actually two Vj gating mechanisms (commonly referred to as ‘fast’ and ‘slow’ [42]) that each result from separate structural features within Cx [43]. Fast-gating is characterised by rapid (≤1 millisecond) transitions to various subconductive states of be- tween ∼5–40% maximal channel conductance, depending on the isotype [44]. The voltage sensitivity of fast-gating is also isotype dependent, with some responding to positive Vj (e.g., Cx26 and Cx40), while others respond to negative potentials (e.g., Cx43 and Cx32) [41]. In the case of homotypic gap junctions this polarity responsiveness makes little difference; the gap junc- tion has a voltage sensor in each of two coupled cells (one in each connexon), so a depolarization in one cell is electrically equivalent to a hyperpolarization in the other. If a gap junction is heterotypic however, with one connexon of opposite polarity sensitivity from the other, then the coupled cells will have an asymmetric response to Vj [45]. Unlike fast-gating, the slow-gating mech- anism requires major conformational changes in the channel [46] that are characterized electrophysiologically by a stepwise reduction in conductance spanning 10s of milliseconds [47], eventually leading to complete occlusion of the pore. All Cx isotypes create slow-gating channels sensitive to negative potentials [41], which is critically important — not so much as a means of regulating mature gap junction conductance, but to keep unpaired connex- ons closed while they are translocated through the excretory pathway and the plasma-membrane before coupling. 9 Hemichannels Within the gap junction community, the term ‘hemichannel’ is often under- stood to be broadly synonymous to the term ‘connexon’, but various defini- tions have evolved following the discovery of functional unapposed connexin channels. For the sake of clarity, a connexon will be described here as any hexameric structure formed by connexins, as part of a gap junction or not, and whether functional or not; while ’hemichannel’ will be reserved for that subset of connexons that display channel activity independent of a partner connexon. Given the highly permeable nature of gap junctions, any significant level of connexon hemichannel activity would quickly dissipate the electrochemi- cal gradient across the plasma membrane, leading to cell death. This lethal effect was first demonstrated in 1991, when over-expression of Cx46 in Xeno- pus oocytes resulted in a depolarizing positive feedback loop that led to osmotic cell lysis [48]. A number of mechanisms exist to prevent this from occurring under normal physiological conditions, including chemical gating by 2mM [Ca2+]o [49, 50], as well as a high closed state probability at negative membrane potentials [48]. Conversely, hemichannels can open in response to membrane depolarizations and a drop in [Ca2+]o (below 1 mM), and are also sensitive to mechanical stimuli, pH, and redox potential (reviewed in [51] and [52]). In general, the electrophysiological properties of hemichannels are very similar to their respective gap junctions in terms of selectivity, with twice the conductivity [53]; as would be expected by reducing the overall length of the channel by half. Hemichannels have been described in an extensive array of both physiological and pathological processes, including (but not limited to) inflammation [54], cell cycle control [55], and bone homoeostasis [56], as well as renal [57], cardiac [58], and brain [59] function. There is some controversy over the interpretations of many earlier hemichannel stud- ies however [60, 61], largely due to the discovery of pannexin channels which possess many similar biophysical and pharmacological characteristics to Cx hemichannels (discussed further in section 1.2). 1.1.2 Innexin The first biophysical observations of a gap junction were recorded between arthropod neurons [1], and then subsequent studies found electrically equiv- alent features in other diverse Metazoans, including the earthworm [62], leech [63], fish [9], and mammals [64]. Once the Cxs were identified, they were reasonably assumed to be the sole gap junction forming proteins; con- 10 served throughout the animal kingdom. It turned out this assumption was incorrect however, and Cxs are in fact an innovation exclusive to the chor- date lineage [4]. The innexins are a completely unrelated family of proteins responsible for forming gap junctions within the myriad invertebrata, with surprisingly similar functional properties to the Cxs given their ancestral independence. Discovery The first Cx complementary DNAs (cDNAs) were cloned in 1986 from mam- malian liver samples [65, 66] , and yet, invertebrate Cxs sequences remained strangely elusive throughout the subsequent decade. In 1994 Thomas Barns proposed that a collection of genes present in the worm and fly might code for a new class of gap junction proteins [67], and he named the group ‘OPUS’ after the genes ogre, Pas, unc-7 and shakB. The shakB cDNA was later ex- pressed exogenously in paired oocytes, where it was indeed shown to code for a gap junction protein [68], while at the same time the OPUS nomencla- ture was discarded for being confusing (Pas and shakB were different alleles of the same gene, and ‘OPUS’ failed to impart any sense of function to the group). The name of these genes was instead replaced with ‘invertebrate analogues of the connexins’ — innexin [69]. In the course of proposing this new name, Pauline Phelan commented on the fact that over 90% of the C. elegans genome had been analysed with no sign of a Cx sequence, and pos- tulated that they are probably not present at all [69], so the hunt for Cxs in worms and flies quietly came to an end. In the intervening years the volume of genome sequencing data has expanded exponentially, and there is still no sign of connexins outside the phylum Chordata. Furthermore, it is now well accepted that if the Cx and Inx families are derived from either half of an ancient gene duplication, their sequences have since diverged beyond our ability to infer that homology [70]. There are 8 Inx genes present in the arthropod Drosophila melanogaster [71, 72], 25 in the nematode Caenorhabditis elegans [73], and 21 in the annelid Hirudo verbana [74], with many more identified throughout the var- ious Metazoan clades [70]. The protein products from all of these genes are clearly homologous, and appear to originate from a single gene that would have been present in the ancestral species at the base of the Metazoa some 900 million years ago (MYA) [70]. 11 Structure and function Inxs are predicted to have the same gross structural topology as Cxs, with four transmembrane domains, one intra- and two extracellular loops, and cytoplasmic C- and N-terminal tails [67]. The Inxs also contain highly con- served cysteine residues in their extracellular loops necessary for channel formation/function [75], but instead of three per loop like Cx, Inx has only two [70]. While the ability of invertebrate Inxs to oligomerize into hex- amers has not been demonstrated directly, they are unquestionably able to form intercellular gap junctions [69], and cluster together into large plaques at regions of very close cell-to-cell contact [76, 77]. Also like the Cxs, these junctions can be homomeric or heteromeric [78], as well as homotypic or het- erotypic [79]. In terms of pore activity, many Inxs form the canonical highly conductive, non-specific, Vj sensitive intercellular channels typical of a gap junction [69, 78, 80–82], but a number of other interesting properties have also been observed amongst this group. For example, gap junctions created from Hminx1 (leech) are completely insensitive to Vj [83], Drosophila inx2 is dependent upon inx3 as a carrier for proper localization to the plasma membrane [77], and although not yet shown conclusively, it is thought that a number of Inxs in the C. elegans reproductive system do not create gap junctions at all, perhaps in favour of forming hemichannels [73]. Not much is known at the molecular level, but tryptophan scanning of Drosophila inx8 has recently shown residues H27, T31, L35, and S39 (first transmembrane domain) to be positioned along one face of an α-helix that is likely involved in a helix-helix interaction necessary for channel function [82]. In mammals, the only cells commonly known to not express Cxs, are ter- minally differentiated skeletal muscle, spermatozoa, and erythrocytes [84]. Similarly, broad surveys of Inxs in Drosophila and C. elegans show complex expression patterns throughout embryogenesis and in adult tissues, often with isotype overlap [71, 73], and disruption of normal expression can have harmful consequences. For instance, inx7 is necessary for proper axon guid- ance in the peripheral nervous system of the Drosophila embryo [85], and in conjunction with Inx6, is also required for long term memory formation [79]. In the worm, phenotypes associated with Inx disruption include the eat-5, and unc-7/unc-9 mutants (three of the earliest observed, and the only C. elegans Inx genes to not conform to the standard naming scheme), which respectively result in desynchronization of pharyngeal muscle contractions [86], and severe impairment of forward movement along with egg-retention [87, 88]. 12 1.2 Pannexins As discussed in section 1.1.2, the Cxs are a chordate specific family of pro- teins not represented in the genomes of any Invertebrata, which instead utilize Inxs to form intercellular gap junctions. It turns out that Inxs are expressed alongside Cxs in chordates, but instead of competing with Cxs as a redundant class of gap junction, these proteins (named pannexin) primar- ily function as distinct aqueous pores between the intra- and extracellular space. A considerable amount of attention has been focused on the biochem- ical and functional properties of Panxs during the intervening years since their discovery in 2000, and much of the literature pertaining to this work will be reviewed in the following sections. 1.2.1 Discovery Inxs from nematode worms and flies have been the most extensively stud- ied, but by best estimates these two lineages diverged over 900 MYA [89]. This far pre-dates the split between the chordate lineage from that of the In- secta, so from a phylogenetic perspective it is reasonable to predict that Inxs may also be present in chordates. Yuri Panchin was the first to show how extensively the Inx genes have radiated throughout the Metazoa, aligning sequences from example species within Platyhelminthes, Nematoda, Arthro- poda, Mollusca, and even Chordata. In light of this diversity, and especially because of their presence in non-invertebrates, Panchin argued that the ex- isting naming convention was not appropriate; suggesting that this family of genes instead be re-branded as ‘Pannexin’, because the Latin prefix ‘pan’, meaning ‘all’, would better represent reality [90]. Unlike the switch from OPUS however, the community was already becoming accustomed to the ‘Innexin’ nomenclature, and it has remained largely intact within the inver- tebrates. Even so, the new name has been adopted in the case of chordate Inx homologs, and from lancelet to man these genes are known as ‘pannex- ins’. The initial discovery of mammalian Panxs was achieved by BLAST searching GenBank against known Inx sequences [90], and several subse- quent studies have gone about confirming homology between the Inxs and Panxs by more rigorous statistical means [70, 91–93]. Interestingly, de- spite the structural and functional features common among the Inxs and Cxs, there is no discernible conservation of primary amino acid sequence, and as such, they are no more related to one another than they are to other tetra-membrane spanning proteins like claudins and occludin [93]. Recently, 13 another protein has been identified that may well share ancestry with Panxs, named leucine-rich repeat containing 8 (LRRC8). Bioinformatic analysis in- dicates that a fusion between the entire transmembrane region of an ancient Panx gene fused with a leucine-rich repeat-containing domain, which has taken on the role of C-terminal tail [94]. To date no biophysical analysis has been performed on this unique protein, but if it represents yet another large plasma membrane channel it may help explain some of the conflict- ing reports regarding permeability within the Panx literature, which will be discussed in subsequent sections. 1.2.2 Biochemical properties While there are only three primary Panx paralogs in chordates, there is com- pelling evidence that they can intermix to form heteromeric pores, and that post-translational modifications are critical for proper intracellular localiza- tion and function. Furthermore, the rapidly growing list of Panx interacting proteins implies that these molecules could play a complex and multifunc- tional role in cellular processes. Isotypes Three Panx genes are present in most model organisms, and the products of these genes have been designated Panx1, Panx2, and Panx3. These genes originate from duplication events that occurred prior to the radiation of all higher vertebrates, and as a result the three Panxs are orthologous across taxa (i.e., Panx2 in Xenopus is equivalent to Panx2 in human, etc. See Fig. 1.2), which is not the case for Inx genes [70, 92]. Sequence alignments indi- cate that the Panxs are well conserved between species, sharing on average >70% identity and >80% similarity at the amino acid level (Table. 1.2). Panx1 and 3 are also relatively well conserved, sharing ∼60% identity and ∼75% similarity between paralogs, but Panx2 is significantly more divergent (Table. 1.3). Splice variants for Panx1 and 2 have also been predicted, with evidence for three different Panx1 isoforms generated in rat (full length = Panx1a; total loss of exon 3 = Panx1c; and partial loss of exon 4 = Panx1d) [91, 95, 96], and at least two Panx2 isoforms are actively transcribed in fish [97]. The teleost fishes also have a fourth Panx paralog, retained as two ‘ohnologs’ of Panx1 following a whole genome duplication event approxi- mately 350 MYA. This duplication presents an interesting opportunity to assess the plasticity of Panx1, the most highly expressed of the Panxs, when evolutionary pressures are relaxed. As we will see in chapter 2 the two 14 paralogs have indeed diverged functionally, both in terms of transcriptional regulation as well as at the channel level. 15 bfPanx drPanx3 xtPanx3 lcPanx3 ggPanx3 mmPanx3 drPanx1a drPanx1b lcPanx1 mmPanx1 ggPanx1 xtPanx1 drPanx2 mmPanx2 lcPanx2 ggPanx2 xtPanx20.2 ciPanxA ciPanxB frogxt → mousemm → coelacanthlc → zebrafishdr → sea squirtci → lanceletbf → chickengg → Figure 1.2: Phylogenetic relationship between pannexin sequences from M. musculus, G. gallus, X. tropicalis, L. chalumnae, D. rerio, B. floridae, and C. intestinalis. Multiple pairwise global alignments were performed with a gap open penalty of 11 and gap extension penalty of 1, using the Blosum45 cost matrix 16 Pannexin 1 Pannexin 2 Pannexin 3 Identity 73.5% ± 3.0 71.3% ± 4.9 75.5% ± 5.9 Similarity 82.6% ± 3.0 80.0% ± 4.4 85.2% ± 4.4 Table 1.2: Amino acid conservation of pannexin orthologs between species. Values represent the average conservation among the sequences from M. musculus, G. gallus, X. tropicalis, L. chalumnae, and D. rerio, based on multiple pairwise global alignment used to generate the phylogenetic tree in Fig. 1.2. ‘Identity’ was called for exact match residues, and ‘similarity’ was called for residue substitutions with Blosum45 scores greater than 0 Panx1/Panx2 Panx1/Panx3 Panx2/Panx3 Identity 24.9% ± 2.5 62.4% ± 3.2 25.9% ± 1.9 Similarity 37.0% ± 2.0 74.2% ± 3.0 37.0% ± 2.6 Table 1.3: Amino acid conservation between pannexin sequences within a given species (i.e., paralogs). Values represent the average conservation among the sequences from M. musculus, G. gallus, X. tropicalis, L. chalum- nae, and D. rerio, based on multiple pairwise global alignment used to gen- erate the phylogenetic tree in Fig. 1.2. ‘Identity’ was called for exact match residues, and ‘similarity’ was called for residue substitutions with Blosum45 scores greater than 0 Structure As discussed in section 1.1.2, the gross structure of the Inx superfamily is very similar to that of Cx, with four transmembrane domains and cytoplas- mic C- & N-termini (Fig. 1.1). To date, there have been no reported crystal structures of a Panx protein, but a number of indirect methods have been used to query the relative importances and/or 3D position of many specific residues. The most common method has been site directed mutagenesis, targeting specific residues within the sequence that are expected to play a role in tertiary/quaternary structure. From this work we now know that the four conserved extracellular cysteines are required for the formation of functional channels [75, 98, 99] (Fig. 1.3). Of the other cysteines found in Panx1, converting the intracellular residue C346 or the transmembrane residue C40 to serine results in a constitutively open channel that rapidly degrades plasma membrane potentials, leading to cell death [75, 98, 100]. Feng Qiu took the mutagenesis approach to a whole other level by system- 17 atically modifying 81 of the 84 extracellular Panx1 residues one at a time, followed by careful electrophysiological characterization of the mutant pro- teins’ abilities to form channels [101]. Of the mutants generated, 24 ablated voltage gated channel activity (although the cause, be it changes in fold- ing, trafficking, or direct channel blockage, was not assessed in this report). Substituted cysteine accessibility method (SCAM) and electron microscopy have been used to gain a better understanding of the actual pore structure of Panx1, and it appears that residues 3-7, 10, and 12 in the N-terminus, residues 58-62 from the first transmembrane domain, and residues 414-426 from the extreme C-terminus contribute the hydrophilic pore lining moieties [98], and the outer entrance of the pore is around 17-22 Å (29.-30.5 Å in the case of Panx2) [99]. 18 Panx1 Panx2 Panx3 † † † † → N-Linked Glycosylation Figure 1.3: The primary amino acid sequence (colour coded using the ClustalX colour scheme) of the three chicken Panx proteins is superimposed over the predicted association each protein has with the plasma membrane. Of note, are the glycosylation sites found on one of the two extracellular loops in each isotype, the relatively long C-terminal tail of Panx2, the four conserved extracellular cysteine residues (∗), and the leucine residues found in the C-terminal tail of Panx3 († these will be discussed in section 6.7) 19 Oligomerization Given that Panxs were shown to form channels quite soon after they were first discovery, it was assumed that the monomeric protein oligomerized into an analogous structure to the connexon [102]. Electron microscopy of cell membranes expressing Panx1 or Panx2 supports this notion of a ‘pannexon’ [99], as do cross-linking experiments where sodium dodecyl sulfate polyacry- lamide gel electrophoresis (SDS-PAGE) and Western blot revealed Panx1 bands at sizes expected for dimeric and hexameric interactions [103]. In- triguingly, similar cross-linking experiments for Panx2 suggest that they ac- tually form octameric structures, which has never been observed in Cxs [99]. A number of studies have identified an approximately 70 kDa Panx3 band in Western blots, with speculation that it may represent a dimer that is insen- sitive to the reducing/denaturing conditions of SDS-PAGE [96, 104, 105]. All of these studies use the same polyclonal antibody however [104], and while there has been mention of this species being identified using other antibodies, the blots have not been published [105]. During my own expe- rience using this antibody (see chapter 4), I noticed that the 70 kDa band (but not the predicted 42 kDa monomeric band) was also present in condi- tioned growth media taken from cultures of differentiating mouse calvarial pre-osteoblast cells (MC3T3-E1) and in lysates of cells not known to express Panx3. Further investigation indicated that foetal bovine serum (FBS) ac- tually contains a large amount of the 70 kDa species (Fig. 1.4), so the extent that cell cultures are washed prior to lysis can potentially have a sig- nificant effect on the intensity of the band. Perhaps there is a large soluble Panx3 species that is being released into the plasma, or this band could sim- ply represent a non-specific antigen that is nonetheless sensitive to cognate blocking peptide [104]. In either case, if the antigen is also detectable in rodent serum, it could explain why the 70 kDa band is so intense in organs with high blood/lymph content like the spleen and thymus [104], despite the conspicuous absence of detectable Panx3 mRNA in these tissues [91, 102]. In terms of heteromeric pannexons, immunoprecipitation (IP) experi- ments have shown interactions between Panx1 and 2, as well as between Panx1 and 3, but not between Panx2 and 3 [106, 107]. Heteromeric inter- mixing does not appear to be conducive to channel activity however, with Panx1/2 co-expression reducing the size of voltage activated currents rela- tive to Panx1 expression alone [106]. Nor is the configuration overly stable, since Panx1/2 pannexons degrade much more rapidly than their monomeric counterparts [99]. 20 75 kDa M C3 T3 -E 1  co nd iti on ed  m ed ia He La  ly sa te NI H3 T3  ly sa te NR K lys at e M C3 T3 -E 1 lys at e 50 kDa 37 kDa 75 kDa anti-γ-tubulin anti-Panx3 anti-Panx3 50 kDa 37 kDa Ho rs e s er um Si gm a F BS Hi Cl on e F BS Go at  se ru m No n- fa t m ilk Bo vin e s er um  al bu m en A B Figure 1.4: (A) Protein was isolated from conditioned growth media off of an MC3T3-E1 culture, as well as lysates from a number of cell lines. The 70 kDa band was prominent in all samples, but only differentiated MC3T3-E1 cells expressed the 42 kDa band. (B) Protein was isolated from a number of commercial serum sources, as well as from milk and purified bovine serum albumin. In no case was the 42 kDa Panx3 band observed, but FBS from two separate vendors contained the 70 kDa species 21 Posttranslational modification Unlike the Cxs, where phosphorylation plays key roles in function and life cycle [108], the Panxs do not seem to be phosphorylated at all [103, 104]. They do however, and again this is in contrast to Cxs, undergo glycosyla- tion on residue N254 on Panx1 (N246 in fish), N86 on Panx2, and N71 on Panx3 [103, 104, 107, 109, 110]. Initially, glycosylation in the ER results in a high mannose form, and following movement into the Golgi, this is further processed into a complex form [107, 111]. Blocking glycosylation through site directed mutagenesis, or by pharmacological agents, reduces the abil- ity of Panxs to traffic to the plasma membrane [103, 104]. Palmitoylation is also thought to play a role in Panx2 dynamics in hippocampal neural progenitor cells, possibly at residue 246. It appears that palmitoylation prevents trafficking of Panx2 to the cell surface in immature neurons, but the depalmitoylated form is free to move to the plasma membrane following neuronal differentiation [112]. Interacting proteins The primary method used to identify the growing list of Panx interacting proteins has been co-immunoprecipitation (co-IP) following some variant of rational target selection. As we will see in sections 1.2.3 and 1.2.5, there is a strong relationship between Panx1 and P2X receptors (and with the purinergic receptor P2X ligand-gated ion channel, 7 (P2X7R) subtype in particular), which are adenosine-5’-triphosphate (ATP) gated ion channels. Of the seven P2X receptor subtypes, P2X2R, P2X3R, P2X4R, and P2X7R have all been shown to co-IP with Panx1 [95, 113–115]. Panx1 also co-IPs with many components of the inflammasome [114], and P2XRs are only one possible component of these structures, so it still remains to be determined whether Panx1 and P2X receptors directly interact, or require other adaptor proteins. The G protein-coupled P2Y receptors are a second major class of ATP binding proteins, and there is some evidence that Panx1 associates with these as well [116]. Of particular relevance to the study in chapter 2 (where we will see significant localization of Panx1 to areas of membrane ruffling), is the interaction between Panx1 and the actin cytoskeleton. The C-tail of Panx1 co-IPs and co-sediments with actin, and pharmacological disruption of the microfilaments (but not of tubulin microtubules) reduces cell surface stability and motility of both Panx1 and 3 [111]. Various other cell sur- face receptor and channel proteins have also been shown to interact with Panxs, including voltage-dependent potassium channel β subunit (Kvβ3), 22 which may play a role in regulating the gating effects of redox potential on Panx1 [117, 118]; the voltage sensor protein dihydropyridine receptor, which could be mediating an interaction between Panx1 and P2Y2 [116]; the α1-adrenergic receptor, which may be interacting with Panx1 as part of a signalling microdomain that regulates vascular smooth muscle tone [119]; and stomatin, which through an interaction with the C-terminal domain of Panx1, inhibits channel activity [120]. 1.2.3 The pannexin channel Given their relationship to the Inxs it was initially thought that Panxs were gap junction forming proteins, but that notion has now been largely dis- carded. Instead, Panxs are understood to form large gated pores between the intra- and extracellular space, reminiscent of Cx hemichannels. Intercellular gap junctions A small number of studies have reported the ability of Panxs to form inter- cellular channels. For example, over-expression in Xenopus paired oocytes revealed appreciable levels of transjunctional current attributable to Panx1, as well as co-expressed Panx1/2 (but not Panx2 or Panx3 on their own) [102]. These channels are surprisingly insensitive to Vj, with recorded cur- rents varying linearly with membrane potential even at large driving forces (> +60 mV or < -60 mV). These channels take a very long time to form however (upwards of 24 hours), and the macroscopic currents are relatively weak when compared against Cx based gap junctions [103]. Intriguingly, enzymatic removal of bulky N-glycans from the cell surface dramatically en- hances the ability of Panx1 to couple [121], implying that even though the Panxs may have retained their ancestral ability to dock with one another, this configuration is largely inhibited in vivo. Panx gap junctions have also been inferred following over-expression in studies using cultured cell lines. Rat C6 glioma cells appeared to acquire the ability (albeit a weak one) to pass sulphur rhodamine 101 when Panx1 was introduced [122], and a human prostate adenocarcinoma line (LNCaP) was shown to increase the movement of calcium between neighbouring cells following Panx1 transfection [123]. In a more recent study, ∼ 2.5% of paired Neuro2A cells over-expressing Panx1 became electrically coupled, but the observed current was very weak at ∼32 pS [124]. Panx3 has also been reported to form calcium permeable gap junctions in a study by Ishikawa et al, using cultured osteoblast [125]. On the other hand, different groups have completely failed to measure Panx 23 coupling, using both dye transfer as well as electrophysiological techniques [104, 126], and criticisms have been published in response to the studies that do support coupling. These include the potential for an un-physiologically relevant response to massive over-expression; a lack of proper control for induction of endogenous Cxs; and a more general inability within the com- munity to observe Panx behaving in a fashion one would expect from a gap junction, such as sequestration to areas of cell-to-cell contact [99, 127– 129]. Since about 2006, the general attitude towards Panx gap junctions has largely degraded to that of ‘academic peculiarity’, instead of ‘common cellular feature’ [130] — and while a complete disavowal of physiological relevance is probably not prudent, the burden of proof currently rests heavy on the shoulders of anyone making such a claim. Non-junctional channels Since the very first biophysical study performed by Roberto Bruzzone in 2003, Panx1 has been known to form non-junctional channels that medi- ate robust macroscopic currents across the plasma membrane of individual cells [102]. These channels are characterized by a unitary conductance from ∼350 pS to >500 pS depending on the charge carrier [110, 131–134], with non-selective permeability to solutes as large as 1.5 kDa [61]. Opening of the channel tends to occur at positive membrane potentials and requires 30-70 ms to transition from fully closed to fully open [102, 110, 131]. There are also at least 5 open substates (corresponding to 5%, 25%, 30%, and 90% of the fully open unitary conductance) [131], and following activation, Panx1 channels slowly deactivate again over several seconds to a lower sub- conductance state [102, 131]. Another group now claims that the unitary conductance of Panx1 does not exceed 68 pS, and is in fact anion selective [135], but further evidence will be required to resolve this conflict. The initial oocyte studies that revealed such robust Panx1 channel activity in response to membrane potential, completely failed to note any significant currents attributable to either Panx2 or 3 [102]. Subsequent work however, using sulforhodamine B uptake as a readout, suggests that both of these proteins can indeed form cell surface channels following mechanical stimula- tion (i.e., from the fluid shear force caused by dripping the dye directly onto monolayer cell culture) [104, 105, 107]. Another study has gone a step fur- ther in terms of Panx2, where purified baculovirus liposomes were shown to be permeable to ascorbate if Panx2 was present, and channel activity could be electrophysiologically measured in oocytes by implementing a protracted voltage ramp protocol (-100 mV to +100 mV over 70 s) [99]. Heteromeric 24 Panx1/2 pannexons show channel activity, but the currents reported are significantly lower than Panx1 currents alone, and no channel activity has been reported for Panx1/3 or Panx2/3 heteromers [102]. Pharmacological blockers To help tease apart the details of Panx function, efficacious and highly se- lective drugs that block channel activity are desirable. Several compounds have been identified that suppress Panx1 currents, with varying potency and specificity, and some have been used to great effect in the collective effort to understand the physiological relevance of this channel. CBX is a synthetic derivative of glycyrrhetinic acid, found in licorice, that has been used to block Cx based channels for over 25 years [136], and was one of the first agents shown to directly inhibit Panx1 currents [106]. Panx channels are more sensitive to the drug than are Cx channels [106], so low dose CBX (10- 30 μM) is commonly used to differentiate between pannexon activity and connexin hemichannels [126, 137–139]. Mefloquine is another compound that reversibly blocks both Panx and Cx channels, but again Panx1 is far more sensitive with a 50% inhibitory concentration (IC50) of ∼50 nM [134], versus >10 μM for many Cx isotypes [140]. The drug probenecid (com- monly used to treat the symptoms of gout) has proven popular [119, 141– 144] after it was shown to block Panx1 channels with an IC50 of ∼150 μM without effecting Cx channels, even at doses as high as 5 mM [145]. Interestingly, the effects of probenecid and CBX on Panx1 are severely mit- igated when co-expressed with Kvβ3, so studies that fail to see an effect from these blockers cannot immediately rule out Panx1 activity [118]. An- other popular reagent is a mimetic peptide against the first extracellular loop of Panx1, called 10Panx1, that impedes passage of small ion currents, dyes, and ATP [61, 113, 146–148]. This effect can be partially replicated by other peptides of equivalent size, or polyethylene-glycol with a molecu- lar weight of 1500 Da, so the blockage may result from steric interference within the pore itself rather than from specific interactions with the ex- tracellular loop [61, 149]. Two other compounds that show promise are glyburide, with an IC50 of 45 μM [150], and the chloride channel blocker 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB), with an IC50 of ∼50 μM [145]. 25 Gating Panx channels are generally gated shut to prevent degradation of the elec- trochemical gradient across the plasma membrane, and a number of stimuli have now been identified that either enhance this closed state or induce channel opening. Voltage gating is perhaps the key process regulating chan- nel activity in the absence of other stimuli, because normal (i.e., negative) membrane potentials keeps the channel closed [102]. At positive voltage the characteristic high conductance pore does becomes evident, but aside from excitable cell types like neurons and myocytes, membrane resting potentials above 0 mV are uncommon. Panx1 channel activity is also suppressed by acidification of the cytoplasm [151], increasing intracellular redox potential [75, 100, 118], and (at least partially) through an interaction with a erythro- cyte integral membrane protein, stomatin [120]. In terms of activating the channel, at least four distinct mechanisms have been identified which can overcome gating at negative membrane potentials: 1) Mechanical stimulation, like increased fluid shear force [104, 105, 107, 152] or the application of suction to a patch pipette after excising membrane from a Panx expressing cell [131], increases permeability for the duration of the stimulus. Cell swelling has also been proposed as a semi-mechanical stimu- lus for Panx1 activation [95, 132, 138, 153, 154], but others have argued that some channel other than Panx1 is responsible [147, 155]. 2) Increasing ex- tracellular potassium concentration ([K+]o) to ≥ 20 mM causes Panx1 (but not Panx2) channels to open even at hyperpolarized potentials (-100 mV) [99, 114, 131, 156, 157]. 3) Proteolytic cleavage of the Panx1 C-terminal tail at residues 376-379 results in a constitutively open channel, generally in association with apoptosis, and caspase 3 and 7 are the primary pro- teases involved [158, 159]. The mechanism of action in this case appears to be non-covalent binding between the C-tail and the interior of the pore, which is lost to dissociation/diffusion once the C-tail has been cleaved [160]. Interestingly, adding an exogenous peptide into the cell with the same se- quence as the cleaved C-tail fragment is sufficient to block the processed channel, but only at negative membrane potentials. At positive potentials the peptide has no effect, so the cleaved channel behaves ‘normally’ as long as the peptide is present in high enough quantity to match the dissocia- tion constant. This implies that the extreme C-tail is also involved in the overall voltage sensitivity of the Panx1 channel [160]. 4) Unlike the Cxs, Panxs are insensitive to [Ca2+]o [106, 137], but they open readily at nor- mal resting potentials if [Ca2+]i is raised to mM levels [110, 151]. Several processes known to influence Panx1 activity are able to generate this level 26 of [Ca2+]i, including caffeine stimulation of the ryanodine receptor [161], thrombin activation of protease-activated receptor-1 (PAR-1) or histamine activation of the histamine receptor [162], adrenergic receptor alpha1 and muscarinic acetylcholine receptor activation of q-subtype G-protein [163], and perhaps most importantly, stimulation of various purinergic receptors [113, 151]. Members from the P2Y and P2X receptor classes have both been implicated in ATP mediated Panx1 activation, but do so through different mechanisms. P2Y1R and P2Y2R are G protein-coupled receptors that me- diate calcium release from internal stores through phospholipase C depen- dant inositol 1,4,5-trisphosphate (IP3) generation [164], and have both been shown to cause Panx1 currents following ATP stimulation [139, 151]. The P2X receptors are themselves channel proteins, and upon activation can increase [Ca2+]i by facilitating transit across the plasma membrane from the extracellular space. Direct interaction between P2XRs and Panx1 [95] ensures that local increases in [Ca2+]i are detected by Panx channels, cre- ating biphasic macroscopic currents in the presence of ligand [113]. P2X7R is a particularly prominent Panx1-interacting isotype that is dependent on phosphorylation by src-tyrosine kinase [165]. With a pore large enough to pass solutes up to 1.5 kDa, an open Panx1 channel should have no trou- ble allowing ATP to flow along its concentration gradient from inside to outside of the cell. This property was first demonstrated by Li Bao with some clever reversal potential measurements from a patch pipette using an ATP gradient between the pipette and bath solution [131]. The flow of ATP into the extracellular space stimulates more purinergic receptors, setting up a positive feedback loop on the affected cell as well as diffusing outward to stimulate other cells nearby [124]. This positive feedback is eventually derailed because ATP binds to the extracellular loops of Panx1, forcing the channel into a lower conductance state that is no longer permeable to ATP [166]. The concentration of ATP needed to stimulate P2X7R is 50-100 fold lower than is required to deactivate Panx1 however [101] (UTP and GTP are also inhibitory [137], but are not overly relevant because do not reach the same physiological concentrations as ATP), so the system pro- vides an elegant mechanism for amplifying an ATP stimulus. Care must be taken here though, in light of a recent report that casts some doubt on the ‘ATP induced ATP release’ mechanism; Panx1/2 double knockout as- trocyte cultures were still able to release ATP in response to the P2X7R agonist 3’-benzoylbenzoyl adenosine 5-triphosphate (bzATP), and CBX was equally as effective at blocking this release as it was in wild type cultures [167]. Also, there are other non-ATP induced ATP release mechanisms that must be considered, for instance cytokine stimulated wild type (WT) and 27 Panx1−/− macrophages release equivalent levels of ATP [168], and erythro- cytes release Panx1-independent ATP following stimulation with iloprost (a prostaglandin metabolite) [150]. 1.2.4 Expression and distribution Northern blots provided the first Panx expression data, using commercially available mRNA samples from mouse [102] and human [91]. From this ini- tial work Panx1 was reported to be expressed by a wide range of tissues, Panx2 was essentially restricted to the brain, and Panx3 was not expressed at appreciable levels in adult tissues except in the skin. A number of no- table discrepancies existed between these two studies however, such as very high levels of Panx1 in the heart, skeletal muscle, and testis from the hu- man samples verses nearly undetectable levels in the mouse, and there was surprisingly weak Panx2 expression in human spinal cord when compared to the extremely high level in the mouse. As we will see in the following sections, these conflicts have been readdressed over the 8 years since the ini- tial observations were published, along with a great number of other details, especially as antibodies have become available. Furthermore, evidence will be reviewed that reveals Panx expression within tissues is often cell type specific, and in many cases developmentally regulated. Trafficking As integral membrane proteins, Panxs are first expressed in the ER [111, 121, 123, 169]. There is some evidence that Panx1 and Panx3 can be retained in the ER as functional calcium leak channels [123, 125], and depending on the cell type, the fraction of Panx retained inside the cell can be significant [170]. In cells where targeting to the plasma membrane is evident however, Panx first traffics to the Golgi via Sar1 dependant COPII vesicles [111], and glycosylation plays a critical role in regulating these movements (at least for Panx1) [103, 104]. While in the ER, Panx1 is initially N-glycosylated to a high mannose form (GLY1) that can be separated from the un-glycosylated form (GLY0) using SDS-PAGE; the high mannose polysaccharides are fur- ther processed into a complex form (GLY2) in the Golgi [121]. Using drugs to prevent glycosylation, or transfecting cells with glycosylation defective Panx mutants, severely disrupts the amount of the protein that translocates to the PM [103, 104, 107, 170]. Once in the PM, Panx1 and Panx3 are sequestered in Triton X-100 insoluble lipid rafts [171], and can remain there for hours or days [104, 121]. When they are eventually removed they appear 28 to be shuttled off to lysosomes for degradation [103, 172], but the mecha- nism controlling their internalization remains a mystery — it is not mediated by any of the standard clathrin, caveolin, or dynamin dependent endocytic pathways [172]. Panx2 continues to be an enigma, as our lab and others have most often observed it localizing to small intracellular puncta (See Fig. 2.7) [173, 174], the identification of which is still under investigation. Panx2 does nevertheless still traffic to the PM in the right circumstances [99], perhaps in response to de-palmitoylation [112]. Bone Osteogenic cell types are of particular interest because they are among only a very limited group of cells that express Panx3. Cultured osteoblasts up- regulate Panx3 expression significantly when they are stimulated to differ- entiate and mineralize [104, 125, 170], as do the ATDC5 and N1511 chon- drocytic cell lines [169]. At the tissue level, Panx3 antibodies reveal strong expression in bones derived from both endochondral ossification (EO) as well as intramembranous ossification (IO) [125, 169, 175], and closer exam- ination of EO derived long bones indicates that induction in growth plate chondrocytes only occurs at the pre-hypertrophic to hypertrophic stage of differentiation [169]. Over-expression and knockdown studies indicate that Panx3 is important for normal differentiation of these cell types [125, 169], but this will be discussed in much more detail in chapters 4 and 5. Panx1 is also expressed in osteoblasts [170]. Brain The expression of all three Panxs have been reported in the brain [122, 175]. In rodents, Panx1 is highly produced by undifferentiated neuroblasts, with down-regulation occurring as these cells commit to a specific lineage. As a result, relative expression within the whole brain peaks in late embry- onic/early neonatal animals, and is dramatically reduced in adults [176, 177]. Not all neuroblast derived cells lose Panx1 however, with expression continu- ing in many mature neurons, such as excitatory principle cells, cortical and hippocampal interneurons, GABAergic Purkinje cells, dopaminergic neu- rons, and cholinergic motoneurons [176, 178, 179]. A number of neuron rich regions in particular retain significant Panx1 expression into adulthood, in- cluding the hippocampus, inferior olive, substantia nigra, dentate gyrus, layer V of the prefrontal cortex, and several layers in the olfactory bulb [178–180]. At the subcellular level Panx1 can be observed localizing to 29 neuronal synapses, although distribution is heavily skewed to the postsy- naptic side of the cleft [179]. While most of the Panx1 expressing cells identified in the brain are neuronal nuclear antigen (NeuN) positive (neuron specific marker), co-expression with the glial marker glial fibrillary acidic protein (GFAP) has also been observed in tissue sections [156, 178], and primary astrocyte cultures tend to have significant levels of Panx1 expres- sion [122, 126, 155, 181, 182]. Cerebellar white matter oligodendrocytes are also a possible source of Panx1 in the brain, but the reports are conflicted [102, 176–178, 183]. In contrast to Panx1, Panx2 expression is much lower early in brain de- velopment; becoming more pronounced in the adult [177]. The subcellular distribution of Panx2 also changes considerably as neurons mature. It stays predominantly intracellular in immature neuronal progenitor cells, then ex- pression is turned off for a short period as the progenitor cells transition towards terminal differentiation, followed by re-expression in mature neu- rons, but this time Panx2 traffics to the plasma membrane [112]. Many different neural subtypes throughout the brain express the protein, with no apparent correlation to specific neurotransmitter molecules, degree of elec- trical connectivity, or cellular origin [173]. Panx3 has been observed at the transcript level in hippocampal mRNA and as actual protein in whole brain samples, but in both cases the overall expression appears to be very low [91, 175]. Circulatory system One of the most striking indications that Panx and Cx are truly different entities, came with the discovery that Panx1 is present in the membrane of erythrocytes [132]. These cells have no nucleus, so the proteins they carry around must survive the lifetime of the cell. Cx are well known to be short lived with half lives on the order of a few hours [184], but as we will see in section 1.2.4, Panxs are stable for much longer. Functionally, Panx1 appears to facilitate the response to low oxygen. As erythrocytes encounter an area with reduced O2 ATP is released, and this release is sensitive to CBX, probenecid, and the 10Panx1 mimetic peptide [185]. Interaction between the released ATP and P2Y receptors and/or P2X7R on the surface of the erythrocytes can stimulate even more ATP release, as well as release of epoxyeicosatrienoic acids [132, 186], which are then free to interact with receptors on the blood vessel walls. ATP is a potent vasodilator, so the final result will be widening of the blood vessels and greater perfusion of the tissue to increase oxygen tension [132]. Panx1 also seems to have vasoconstrictive 30 properties, at least in the thoracodorsal artery, where phenylephrine induced narrowing of the artery is inhibited by mefloquine, probenecid, and Panx1 siRNA [119]. This is again probably ATP mediated, because blocking P2Y receptors had a similar effect on smooth muscle tone as did blocking Panx. A series of exceptionally elegant experiments published by Donald Vessey over the past two years have uncovered a role for Panx1 in ischemic pre- /post-conditioning of the heart. Using the Langendorf ex vivo perfused heart model, he has shown that CBX, mefloquine, and Brilliant Blue G (P2X7R inhibitor) are all able to completely block the cardio-protection gained from ischemic pre-/post-conditioning in a 40 minute ischemic insult/40 minute reperfusion protocol. These blockers have no effect if exogenous sphingosine 1-phosphate or adenosine are present however, indicating that their release through membrane channels is critical [187, 188]. Furthermore, the short ischemic conditioning events can be replaced by treating the heart with ATP before or after the ischemic insult, but the protective effects are lost if Panx or P2X7R inhibitors are present during the challenge [189]. These effects are likely mediated through the Akt/PI3K pathway, which is primed during pre-conditioning for rapid response to subsequent G-protein coupled receptor activation by sphingosine 1-phosphate [188, 189]. Ear Panx1 is expressed in many of the cells found throughout the spiral limbus, organ of Corti, spiral prominence at the cochlear lateral wall, and Reissner’s membrane [175]. In the organ of Corti in particular, expression is restricted to the epithelial cells at the top of the organ early in development (E16.5 in mouse), and then as the animal matures, expression is induced in both the inner and outer sulcus, Claudius cells, and neurons in the Scarpa’s and spiral ganglia [190]. Panx2 expression has very little overlap with Panx1 except in the Scarpa’s and spiral ganglia neurons in the organ of Corti [190], and instead is found in the stria vasularis side of the boundary between the stria vasularis and spiral ligament in the cochlear lateral wall [175]. Panx3 is restricted to the bones of the cochlear lateral wall and modiolus [175]. Eye In the retina, Panx1 is expressed in the ganglion cell layer (along with Panx2), inner nuclear layer, and at the periphery of the outer nuclear layer, primarily in ganglion, amacrine, and horizontal cells, with very little ex- pression in the inner plexiform layer. Expression is highest in the juvenile 31 retina, although expression remains in the retinal ganglion cells into adult- hood [109, 171, 176]. At the subcellular level, Panx1 can be observed in the dendrites of horizontal cells [110]. Hydrostatic pressure (HP) is an important parameter in eye physiol- ogy, and must be tightly controlled to prevent conditions such as glaucoma. Neuronal death in the retina can be exacerbated in the presence of elevated extracellular ATP, which increases following a rise in HP presumably by passing through Panx1 channels [191]. HP induced ATP release from tra- becular meshwork cells however, outside the retina, is critical to maintaining homoeostasis. Purinergic signalling initiates MMP-2 mediated degradation of the extracellular matrix between trabecular meshwork cells, thus reduc- ing aqueous humor outflow resistance. Ang Li used a substantive collection of pharmacological agents to show that Panx1 accounts for about one-third of total ATP release in this system, with Cx hemichannels and P2X7R ac- counting for nearly all the remaining release [138]. Immune system As we will see in section 1.2.5, a lot of effort has been expended in under- standing how Panxs affect immune-response. At least a dozen manuscripts have reported on the presence of Panx1 in macrophages [113, 142, 146, 159, 167, 168, 192–197], and a smattering of others have demonstrated Panx1 expression in other components of the immune system, including T-cells [198–200], polymorphonuclear neutrophils [201], and microglia [202]. In neu- trophils, Panx1 appears to be part of a much larger protein complex that includes formyl-peptide receptors, human tweety homolog 3 (a maxi-anion channel), P2Y2R, and actin, and this complex is enriched at the leading edge when neutrophils are mobilized in response to formyl-peptides (which are released from ruptured host cells and/or invading bacteria) [201]. Sim- ilarly, Panx1 co-localizes with P2X1R and P2X4R at the immune synapse following T-cell activation, and pharmacologically blocking Panx1, or using siRNA to reduce expression of the P2X receptors, decreases T-cell response to hypertonic stress [199]. To my knowledge, Panx2 and Panx3 have not been reported to be expressed by any type of immune cell, and they have been explicitly shown to not be expressed in macrophages [197]. Lung Panx1 and a small amount of Panx2 can be found in primary airway epithe- lial cells, and Panx1 has a really striking subcellular localization pattern at 32 the apical pole (equivalent to the luminal side, associated with cilia) when these cells are grown in an air-liquid interface culture [141]. Using a hypo- tonic bath solution to induce swelling of these cells caused an increase in ATP release and propidium iodide uptake, which was sensitive to the Panx inhibitors CBX, probenecid, 10Panx1 mimetic peptide, and Panx1 siRNA. Furthermore, similar swelling experiments performed on trachea explant cul- tures demonstrated a near complete loss of ATP release when the tissue was taken from Panx1−/− mice [203]. Muscle Exceptionally high levels of Panx1 were observed in human skeletal muscle and heart samples during the initial Northern blots performed by Ancha Baranova in 2004 [91]. Later work with rat atrial myocyte cultures confirmed this expression in freshly isolated cells, although once out of the animal the myocytes down-regulate Panx1 expression to undetectable levels within two days. Large (300 pS) probenecid sensitive ion channels were well correlated with Panx1 expression, and activation of these channels was sufficient to induce an action potential (via voltage-gated sodium channels) [161]. Panx1 may also play an ATP dependant role in maintaining [Ca2+]i in skeletal muscle after a strong activating stimulus, possibly through an interaction with the dihydropyridine receptor at the sarcoplasmic reticulum [116, 204]. Reproductive tissue Both Panx1 and Panx3 are expressed by subsets of cells within the testis, efferent ducts, and epididymis of the male reproductive tract. Orchidec- tomy had little effect on the expression of Panx1 throughout the epididymis, but the surgery caused Panx3 to be dramatically up-regulated unless post- operative testosterone was supplied, implying that androgens are able to modulate its expression [96]. Skin Panx1 and Panx3 have been observed in the epidermis and underlying cell layer of E13.5 mouse skin, as well as thin skin from 21 day old mice. Panx1 could be seen in newborn skin, but not in adult thick skin (e.g., foot pad), and the opposite was reported for Panx3. Over-expressing Panx1, but not Panx3, in rat epidermal keratinocyte organotypic 3D cultures resulted in disorganization of the epidermis and reduced vital layer thickness, associated with increased CK14 expression, and reduced involucrin expression [105]. 33 Tongue In taste buds, the receptor cells do not form synapses with the afferent nerves responsible for transmitting information to the brain, but instead are associated with serotonergic presynaptic cells. Upon activation, the receptor cell releases ATP, and activation of purinergic receptors on the presynaptic cell induces neurotransmitter release into the excitatory synapse. There is currently discord in the literature as to how ATP is released from the receptor cells, with one group claiming Panx1 channels are responsible [126, 143, 205], and another believes Cx hemichannels are the primary route [206, 207]. Both groups however, rely exclusively on pharmacological methods to support their positions, and while the evidence appears to favour Panx1, knock-down/knock-out studies are still conspicuously absent. 1.2.5 Physiological and pathological relevance of pannexins Pharmacological agents, over-expression, RNAi, and mouse knockout lines have all been used to modulate the function of Panx channels both in vitro and in vivo, to try and decipher their role in physiological processes. Four separate Panx1 knockout mice have now been generated, and all of them are fertile with no outwardly identifiable phenotype [156, 157, 159, 167, 168, 197, 203, 208, 209]. A Panx2 knockout mouse has also been recently described, and it is viable as well [167]. This normality has actually caused some dis- cord in the community, given the large body of pharmacological data that has been used to support claims that Panx1 is critical to specific processes; inflammation in particular. What this ultimately means remains to be seen, but could indicate that our current crop of ‘specific’ Panx inhibitors have off-target effects that still need to be identified, or perhaps there are com- pensatory mechanisms that are up-regulated when Panx is removed from the system. Irrespective, there are strongly supported cases, particularly in cells or tissues under stress, where Panx plays a role; sometimes by reducing the severity of a particular challenge, and sometimes exacerbating it. Roles in disease The first indication that Panxs might be involved in a specific pathology came in 2006, when Roger Thompson proposed that the massive disrup- tion in electrochemical gradient that occurs across the plasma membrane of hippocampal neurons challenged by oxygen/glucose deprivation, could be the result of Panx1 channel activity [210]. The depolarization in ques- tion did not occur until after about ten minutes of continuous insult, and 34 was reversible if normoxic conditions were reasserted soon after the drop in membrane resistance, but was cytotoxic if the depolarization was allowed to persist for more than ∼5 minutes. CBX inhibition and dye transfer ex- periments supported the hypothesis that Panx1 was involved, and others have now gone on to show that significantly less damage is suffered by oxy- gen/glucose deprived cells/tissues when Panx1 is either blocked or knocked out [167, 209, 211]. Given that Panx1 can induce depolarization and over-excitability follow- ing ischemia, it was a natural extension to ask if it could also be responsi- ble for other cases of dysregulated neuronal activity, such as during seizure events. Epileptic-like seizure can be induced by stimulating brain slices with N -methyl-d-aspartate (NMDA), or treating live animals with the muscarinic acetylcholine 1 receptor agonist pilocarpine or the kainate receptor agonist kainic acid [156, 212, 213]. Intriguingly, interfering with Panx1 activity (us- ing drugs, RNAi, or knockout) significantly reduced the negative effects of NMDA and kainic acid induced seizures [156, 212], while the opposite was true following pilocarpine treatment [213]. The reason for this discrepancy is probably rooted directly in the activating receptors; the muscarinic acetyl- choline 1 receptor leads to intracellular accumulation of IP3 [214], and the accompanying rise in [Ca2+]i has a significant impact on hyper-excitability (by depolarizing the cell) as well as inducing Panx1 opening. IP3 can then exit the cell through the Panx channels, thus reducing its intracellular con- centration. ATP is of course released at the same time, and activation of P2X7R desensitizes the muscarinic acetylcholine 1 receptor [215], so Panx1 and P2X7R ultimately have a quieting effect if pilocarpine is the agonist used to induce seizure. On the other hand, Panx1 opening provides no neg- ative feedback on the system when NMDA receptors or kainate receptors are the source of epileptic discharge. Sustained status epilepticus elevates [K+]o [216], which is the likely source of Panx1 currents in these cases, and their opening enhances excitability by depolarizing the cell and releasing in- tracellular ATP stores. While P2X7R activation may have a quieting effect on the muscarinic acetylcholine 1 receptor, it is otherwise believed to po- tentiate the progression of seizures, and depleting intracellular ATP stores interferes with the phosphorylation state of GABA receptors to decrease the hyperpolarizing effects of these channels [217]. This is a fascinating insight into just how important context can be when interpreting results relating to Panx activity, and indeed, biological processes in general. Analysis of public gene expression databases reveals a correlation be- tween Panx2 and post diagnostic survival time in glioma patients, and a weaker, although suggestive, link between Panx1 and cancer in general [218]. 35 This laid the foundation for a series of studies performed by Charles Lai, a graduate student in our lab, that illustrated how over-expression of either Panx1 or Panx2 can dramatically reduce the tumorigenicity of the rat glioma C6 cell line [122, 174]. The underlying mechanism for this could reside in a general suppression of cell cycle progression, which has been shown to occur in otherwise normal epidermal keratinocytes over-expressing Panx1 [105], or from enhanced actin/actomyosin facilitated cellular compaction when Panx1 is present [219] (of course these two processes are not necessarily mutually exclusive). Whether modulation of Panx has a place in the clinic however, still remains to be seen. As we will see in the next section, a lot of effort has been expended on describing how Panxs are involved in the inflammatory response, but a small amount of work has also revealed how Panx can be hijacked by invading pathogens. For example, the negative impact α-toxins from haemolytic E. coli or S. aureus have on erythrocytes is reduced by CBX, mefloquine, or probenecid [220, 221]. The reason Panx1 (the only Panx expressed in red blood cells) is harmful during haemolytic infection once again relates back to ATP release and activation of a purinergic signalling cascade; without P2X receptor activation, the α-toxin is unable to integrate into the plasma membrane. HIV is also heavily dependant on purinergic signalling early on during the infection process, and as contacts are made between infected and uninfected lymphoblasts, Panx1 is translocated to the contact site where it can increase the local concentration of extracellular ATP [222]. Between Panx and the purinergic receptors, the receptors appear to be the more critical component during infection, not to mention the existence of other ATP release mechanisms, so it seems unlikely Panx will be the next big drug target in the fight against AIDS. Inflammation and immunity The role that Panx1 plays in the innate immune response has been of con- siderable interest over the years, although conflicting results abound. In particular is the process leading up to interleukin-1β (IL-1β) release from pro-inflammatory cells like macrophages. Two separate activation signals are required in relatively quick succession to initiate an acute inflammatory response. The first signal is often an interaction with a foreign pathogen- associated molecular pattern (PAMP) (e.g., lipopolysaccharide (LPS) bind- ing to Toll-like receptors), which induces synthesis of the inactive ‘pro’ form of IL-1β via NFκB signalling. The second stimulus can then come from a cytotoxic T-cell or the activation of P2X7R if extracellular ATP rises 36 [223]. Following the second signal, inflammasome assembly/activation en- sues; caspase-1 becomes active, which then processes IL-1β into its bioactive form before it is released to stimulate inflammation in the surrounding tis- sue. Activation of this cascade via P2X7R was initially though to require Panx1 channel activity, because treating macrophages with Panx1 siRNA, 10Panx1, probenecid, or CBX prevented ATP mediated dye uptake and interleukin-1α (IL-1α), IL-1β, and IL-18 release due to an impairment of caspase-1 processing [113, 142, 192–194]. Further work has even gone so far as to show that simply activating Panx1, by elevating [K+]o for example, is sufficient to activate the inflammasome [114], but the sudden availability of Panx1 knockout mice in the last two years has thrown everything back into question. Macrophages harvested from these animals showed no deficiency in inflammasome activation in response to many different stimuli, leading to caspase-1 dependent cleavage and release of IL-1β [159, 167]. P2X7R still appears to be important however, so one is left to wonder whether the pharmacological blockers so heavily used in the original studies are maybe interfering with the purinergic receptor itself, or with some other as yet unidentified component in the inflammasome activation pathway. Only time will tell. Despite the confusion surrounding Panx and inflammasome activation following the knockout studies, macrophages from these mice still revealed deficiencies in other respects. Cell fusion, for example, occurs between cy- tokine activated macrophages, but the process was severely impaired if the leukocytes were harvested from P2X7R −/− or Panx1−/− mice [168]. Also, the ‘find-me’ signal release by apoptotic cells (as reported by Chekeni et. al., in what may become a landmark paper for the field given how rapidly it is attracting citations [158]) is suppressed in Panx1−/− macrophages [159]. It should be noted that chemotaxis itself is not significantly impacted by Panx1, because a gradient of complement component 5a is still equally at- tractive to knockout cells [196]. A number of studies have also looked at the role Panx1 has to play in T-cell function; concluding that it regulates immune priming following systemic hypertonic stress [200], as well as IL-2 synthesis, calcium intake from the extracellular space, and proliferation in response to T-cell receptor mediated MAPK activation [198, 199]. Like the earlier macrophage studies though, these were all the result of pharmaco- logical inhibition, so we must reserve final judgement until the experiments have been repeated with Panx1−/− cells. 37 Cell cycle Given the potential tumour suppressive effects of Panx [122, 174], it is nat- ural to question how these proteins influence cell division. This is one of the rare cases where Panx3, instead of Panx1 or Panx2, has been the primary subject of study, and indeed a number of over-expression experiments have all indicated that this channel protein is anti-mitotic [105, 125, 169]. At the biochemical level, two mechanisms have been proposed for these observa- tions: 1) Panx3 expression reduces intracellular cyclic adenosine monophos- phate (cAMP) synthesis in response to growth hormone stimulation, leading to reduced cAMP-responsive element binding protein (CREB) associated (pro-mitotic) transcription [169]. 2) Panx3 expression increases [Ca2+]i, possibly as an ER leak channel, thus increasing calmodulin activity and calcineurin induced NFATc1 de-phosphorylation; NFATc1 then promotes differentiation [125]. Cellular morphology Although not studied in close detail, there is evidence that Panx1, and to a lesser extent Panx2, are able to alter cellular morphology. In cultured C6 glioma cells over-expression has a flattening effect that increases the surface area each cell occupies on the substrate [122, 174]. In the case of Panx1 at least, this could relate back to the interaction between the C-terminal tail of the channel and filamentous actin [111] — perhaps the presence of Panx is sufficient to nucleate actin polymerization in the absence of other strong polarizing stimuli. There is even some evidence to support this, in that these same Panx1 over-expressing C6 cells tend to have a more developed actin cytoskeleton, and they also tend to aggregate much more densely in 3D culture [219]. Calcium waves Calcium waves are generally characterized by a focal rise in [Ca2+]i within a population of cells; potentially starting from a single point, the intracellular rise is then communicated to neighbouring cells, which in turn experience their own rise in [Ca2+]i. This process builds a wave front that can prop- agate many dozens of cell lengths from its origin, and is at least partly mediated by an extracellular diffusible agent because it can spread between cells not in physical contact [224]. ATP is this diffusible agent [225]. As was discussed in section 1.2.3 rises in [Ca2+]i induces Panx1 channels to open, even at normal resting potentials; ATP is then free to diffuse outwards and 38 interact with G-protein coupled P2Y receptors on adjacent cells, leading to IP3 synthesis and its associated calcium release which can open the Panx1 channels in these cells as well. Thus the actions of Panx1 could, in theory, support a propagating calcium wave [151]. The real story is more compli- cated than this model would suggest though, because gap junctions are an important mediator of calcium wave progression when cells are in contact [226]. Perhaps more importantly however, when Panx1−/− cochlear organ- otypic cultures were tested for their susceptibility to calcium waves they did not appear to be hindered at all [208]. If anything, calcium wave spread was increased in knockout astrocytes, possibly as a result of increased Cx43 expression [157]. At this time, it is still unclear if Panxs have a significant part to play in calcium wave propagation in vivo. 1.3 Bone development In the early- to mid-Cambrian period some 513-542 MYA, a number of jawless Metazoans began strengthening select tissues with calcium phos- phates; providing the first foundational innovation that would allow natural selection to develop the modern day endoskeletons of all contemporary verte- brates [227]. Today’s bones are a highly dynamic and multifunctional tissue, providing the mechanical support necessary for our large bodies, protecting delicate internal organs, storing minerals that are needed to maintain ho- moeostasis (e.g., calcium) as well as removing others that are toxic (e.g., heavy metals), production of hormones and growth factors, and acting as a nursery for nascent blood cells during hematopoiesis. During develop- ment two separate processes are responsible for bone formation depending on where those bones are located in the body. In the head and face IO pre- dominates, with osteoblasts depositing layers of ossifying matrix to create the mature bone in an ‘inside-to-outside’ pattern, while EO is responsible for bone formation throughout the remainder of the body, whereby an interme- diate cartilaginous anlage is formed prior to mineralization. In the following sections these two processes will be described in greater detail, and the role gap junctions play in bone development, maintenance, and repair will be introduced. Chapters 4 and 5 investigate how Panx3 is regulated during bone development, as well as how it impacts this development in vivo, so the rational for these studies will be laid out in section 1.3.4. 39 Runx2 Runx2 Runx2 Runx2 Runx2 Runx2 Runx2 Col1a1 Col2a1 Col11a2 ACAN Col10a1 Ibsp Mmp13 Spp1 Spp1 Ibsp Alp1 Sox9 Sox9 Sox9 Immature osteoblast Preosteoblast Pluripotent mesenchymal cell    Immature chondrocyte Prehypertrophic    chondrocyte Chondrocyte in permanent cartilage Induction Inhibition    Terminal  hypertrophic   chondrocyte   Mature osteoblast osteocyte adapted from Komori. Cell Tissue Res. 339(1):189–195, 2010 Figure 1.5: Osteoblast and chondrocyte differentiation. Osteoblasts and chondrocytes are derived from a common ‘osteochondral’ progenitor mes- enchymal cell type, and progression towards one lineage or the other is heavily influenced by the transcription factors Runx2 and Sox9. Runx2 drives expression of collagen type I (Col1a1), osteopontin (Spp1), and alka- line phosphatase (Alp1) in progenitor cells, and is necessary for commitment to the osteoblast lineage, but its effects are superseded by expression of chon- drogenic Sox9. Later stages of chondrocyte maturation are however heavily influenced by Runx2, which is required for terminal differentiation of pre- hypertrophic chondrocytes, driving expression of genes like collagen type X (Col10a1), osteopontin, and matrix metalloprotease 13 (Mmp13). 40 1.3.1 Intramembranous ossification Most craniofacial bones are the product of IO. To begin, neural crest de- rived mesenchymal cells are directed to migrate to locations of future skele- tal elements by patterning factors like wingless-type MMTV integration site family (Wnt), bone morphogenic protein (BMP), and fibroblast growth fac- tor (FGF), where they aggregate into densely packed condensations and arrest mitosis. When the condensation reaches critical mass, there is a down-regulation of extracellular matrix (ECM)/adhesion factors like neural cell adhesion molecule (N-CAM) and syndecan, and the osteogenic tran- scription factor Runt-related transcription factor 2 (Runx2) up-regulates the genes necessary to facilitate differentiation and osteoid synthesis. The maturating osteoblasts deposit layers of collagenous matrix and hydroxya- patite to progressively build the bone from the inside out [228]. This first round of mineralization is rapid, but the layers of collagen and calcifying os- teoid are deposited haphazardly, which leads to a mechanically weak ‘woven’ bone. Later, osteoclasts infiltrate the woven bone to begin remodelling, and subsequent osteoblast activity is more controlled; producing well organized alternating concentric layers (i.e., lamellae) of collagen and mineralized ma- trix. This secondary ‘lamellar’ bone is much more resistant to torsional force, although it requires significantly more time to form. A subset of os- teoblasts become embedded in their own osteoid and differentiate into long lived osteocytes, sending projections throughout the bone via canaliculi that link up with one another via gap junctions to form a large syncytium. Os- teocytes are important sensors of mechanical stress and injury, as well as modulators of osteoclast driven remodelling [229]. 41 Osteoblast Center of Ossification Mineralized Bone Osteoblast Osteocyte Blood vessels infiltrate Mesenchymal Cell Blood Capillary Mesenchyme Condenses Adapted from Yue-Ling Wong, IMEJ multimedia Figure 1.6: The process of intramembranous ossification begins with condensation of cranial neural crest derived mesenchymal cells, which differentiate into osteoblasts. Osteoid is secreted by these cells to form an ossification centre, leading to fully mineralized bone matrix. In the final phase of development, blood vessels are recruited into the new bone and remodelling begins. 42 1.3.2 Endochondral ossification EO is responsible for the formation of nearly all bones in the apical and appendicular skeleton. Similar to IO, EO begins with local condensation of mesenchymal cells, although the osteochondral progenitor cells originate from somitic mesoderm as opposed to cranial neural crest. Furthermore, instead of differentiating into osteoblasts, the aggregated mesenchymal cells create a cartilaginous anlage of the future bone. Elongation occurs lon- gitudinally due to chondrocyte proliferation near the terminal ends of the bone, as well as from hypertrophy of terminally differentiated chondrocytes in physeal plates. Associated with final chondrocyte maturation is a tran- scriptional switch heavily influenced by Runx2, resulting in up-regulation of many genes associated with osteoblast function, causing the chondrocytes to deposit a calcifying osteoid before succumbing to apoptosis. 43 Uncalcified Cartilage Calcified Cartilage Bone Blood Vessels Adapted from Alberts, Molecular Biology of the Cell. 4th Ed. Hypertrophic Resting Prehypertrophic Apoptotic Adapted from Kronenberg H.M., Nature 423(6937):332-336, 2003 Proliferating Sox9 Pth1r Pth1r Pthlh Sox9 Col2a Col2a Col10a Col2a Sox9 Runx2 Ihh Ihh Runx2 Hspg2 Figure 1.7: The process of endochondral ossification is characterized by the formation of a cartilaginous scaffold, where the innermost chondrocytes become hypertrophic and ossify, then finally undergo apoptosis as angiogen- esis and remodelling begin. Endochondral bone forms a characteristic ‘growth plate’ structure, with distinct populations of chondrocytes of increasing maturity and distinguishing gene expression as indicated. 44 Anlage patterning Transforming growth factor β (TGF-β)/BMP mediated expression of adhe- sion molecules like N-CAM, N-cadherin, and fibronectin choreographs the recruitment of mesenchymal cells into dense aggregates [230–233], and mem- bers of the forkhead family of transcription factors interact with Smads to stimulate proliferation [234]. Syndecan in the surrounding ectoderm binds to tenascin and fibronectin in the chondroprogenitors, leading to down- regulation of N-CAM and inducing differentiation into chondrocytes, thus helping delineate the size and shape of the future bone [234]. Perhaps the most characteristic regulator of chondrogenesis is Sox9, which is induced very early in condensing mesenchyme and is responsible for orchestrating ECM genes like Col2a1 [235], Col11a2 [236], and aggrecan [237]. Condition- ally knocking out Sox9 from the lateral plate completely blocks condensa- tion, while deletion of the gene during or soon after condensation prevents differentiation into chondrocytes [238]. Although not involved in early con- densation, Sox5 and Sox6 work in conjunction with Sox9 to induce expres- sion of select ECM genes and facilitate chondrocyte differentiation [239]. Chondrocyte maturation Following chondrification of mesenchymal condensations, mitosis continues around the periphery while the innermost cells exit the cell cycle into a state of terminal differentiation, and this is regulated by the Indian hedge- hog (Ihh)/parathyroid hormone-like hormone (Pthlh) loop as first described by Beate Lanske in 1996 [240]. Soluble Pthlh diffuses inwards from the peri- chondrium and extreme ends of the cartilage to interact with parathyroid hormone 1 receptor (Pth1r) on proliferating and prehypertrophic chondro- cytes; maintaining mitotic activity and suppressing further differentiation. When the interstitial space between the innermost cells and the periphery extends beyond the diffusivity of Pthlh however, the chondrocytes begin to enter their final state of differentiation. Ihh expression is induced, which diffuses back outwards to further enhance proliferation in those cells still in range of Pthlh, as well as signalling back to the peripheral chondrocytes to stimulate continuing Pthlh production [241]. Meanwhile, Sox9 delays transition to the pre-hypertrophic fate, but once the cell becomes commit- ted, it interacts with Mef2c to complete terminal differentiation [242], and is subsequently down-regulated. The cell swells considerably (i.e., becomes hy- pertrophic), accounting for much of the longitudinal growth of the bone, and the prevailing collagen expressed switches from Col2a1 to Col10a1. Wnt/β- 45 catenin signalling facilitates the down-regulation of Sox9 and subsequent activity of Runx2 [243], which induces expression of many osteogenic fac- tors (e.g., osteopontin (Spp1) integrin binding sialoprotein (Ibsp)) leading to mineralization of the cartilage [244]. Ultimately, the hypertrophic chon- drocytes begin to recruit blood vessels by stimulating endothelial vascular endothelial growth factor receptors (VEGFRs) with Perlecan (Hspg2) [245], and then undergo apoptosis. Osteoclasts and osteoblasts eventually infil- trate mineralized cartilage to replace it with mature bone [241]. 1.3.3 Connexins in bone The presence of Cx40, Cx43, Cx45, and Cx46 have all been documented in bone [246]. Cx43 has been especially well studied thanks to oculodentodig- ital dysplasia (ODDD); an autosomal dominant congenital disorder that is caused by channel inactivating mutations. ODDD is characterized by a number of dysmorphic bone features, including syndactyly of the hands and feet, craniofacial hyperostosis, and various dental abnormalities [247]. Dur- ing very early limb patterning, Cx43 is expressed primarily in the apical ectodermal ridge, but is later found in the condensing mesenchyme prior to chondrogenesis [248]. Following differentiation however, the nascent chon- drocytes begin to down-regulate Cx43 expression, and eventually it is only found in the perichondral region of the anlage [249]. Knocking down Cx43 in the chick embryo results in limb outgrowth and fusion deficiencies (of- ten leading to truncations or divisions) [250] and facial abnormalities [251]. Surprisingly, the Cx43−/− mouse has only minor skeletal patterning defects (there are a few in the craniofacial bones, thought to result from altered neural crest cell migration), although both EO and IO derived bones have significantly retarded mineralization rates [252]. Cx43−/− mice are not vi- able after birth, due to severe heart defects [253], so a Cx43fl/fl line was crossed with a Dermo1/Twist2 (DM1)-Cre mouse to create a conditional knockout (cKO) from early chondro-osteogenic precursor cells [254], as well as an osteocalcin-Cre mouse to delete Cx43 from only osteoblasts and os- teocytes [255]. These mice are viable, and the mineralization delay during development is not permanent; the mature bones tend to be larger however, and display a distinct cortical thinning due to higher than normal osteo- clast activity (particularly in the medullary cavity) [256]. The mechanism behind the mineralization phenotype involves Cx43 gap junctional coupling between osteoprogenitor cells, the loss of which (e.g., treating with gap junc- tion blockers) is sufficient to suppress osteoblast differentiation [257, 258]. As one might predict, this also influences fracture repair. Following a break, 46 Cx43-cKO mice take significantly longer to re-mineralize the trauma site, and even after more then a month the strength of the repaired bones (in terms of torsional rigidity) is only about half of that in WT animals [259]. The syncytium of osteocytes that responds to less severe forms of mechani- cal perturbations than a fracture (such as loading or unloading) also relies on Cx43, sending signals out to effector cells like osteoblasts and osteo- clasts lining the bone [255]. What is surprising however, is that knocking out Cx43 actually results in a greater relative anabolic response to load- ing (i.e., repetitive strain increases bone mass by a greater percent) [256]. The reason for this is that the decreased rate of osteoid deposition and mineralization suffered by the Cx43-cko animals from slower osteoblast dif- ferentiation is more than counterbalanced by a reduction in the osteocyte dependent RANKL/OPG ratio — leading to suppressed osteoclast activity [256, 260]. The opposite effect is seen during mechanical unloading (such as during limb paralysis or suspension), where Cx43-cKO suffers less relative bone loss than WT [261, 262]. Cx45 has been shown to interact with Cx43 in osteoblasts using co-IP [263], forming heteromeric channels with conductance properties different from those formed by either isotype independently [264], and while Cx46 is also present in osteoblasts, it does not appear to assemble into channels at all [265]. In terms of regulating osteogenesis, it is still unclear if Cx45 or Cx46 play a role, but Cx40 has been clearly implicated in proper patterning of EO derived skeletal elements. The Cx40−/− mouse suffers multiple mal- formations, including elongation of the digits, fusion events between bones of the wrist, and ectopic growth of an extra cervical pair of ribs [266]. 1.3.4 Pannexins in bone A microarray dataset was uploaded into the Gene Expression Omnibus (GEO) database by Dr. Eric Hoffman in 2003, comparing the expression of genes between a fractured bone that had begun to heal normally and one that was non-union (i.e., the broken ends where unable to rejoin). Of all the genes analysed, Panx3 showed the highest expression increase in the nor- mal verses non-union samples (GEO accession GSE494). Further analysis of the public UniGene database circa 2007 revealed a significant enrichment of mouse Panx3 (accession Mm.217159) in bone and teeth; a fact that had been briefly commented on by Ancha Baranova [91], but it had not been investigate further. At the very outset of the work described in chapters 4, this was essentially the extent of what was known in terms of Panxs and osteogenic tissues. The handful of publications that have since confirmed 47 this expression profile and/or explored its significance have already been discussed in section 1.2.4, and will be further addressed in chapters 4 and 5, so will not be exhaustively restated here. Two publications coming out of Yoshihiko Yamada’s lab at the NIH however do much to set the stage for the coming discussion [125, 169], even though they were published after much of the work in chapter 4 was already complete. In these manuscripts Panx3 expression was unequivocally shown to be expressed by osteoblasts as well as prehypertrophic and hypertrophic chondrocytes, and data was presented suggesting that this expression influences osteogenesis. Given the relative dearth of Panx3 expression outside osteogenic cell types, and especially its absence in chondrocytes until they approach terminal differentiation, the transcriptional regime controlling it is very likely the same (or at least rec- ognizably similar) to that controlling genes involved in mineralization like Spp1 or Ibsp. Indeed, multiple pairwise alignments of the Panx3 promoter from various mammalian sources clearly revealed canonical binding sites for osteogenic transcription factors, providing targets for study (see chapter 4). The Yamada lab manuscripts claim that over-expressing Panx3 in vitro enhances differentiation/osteogenic potential of osteoblasts an chondrocytes, while shRNA knock-down has the opposite effect. A more critical appraisal of these results will be provided in chapter 6, but they nonetheless pro- vide motivation to pursue the effects of Panx3 misexpression on bone de- velopment in vivo. As chapter 5 will show, considerable care has gone into streamlining a robust viral delivery system for shRNA in avian embryos, and endogenous levels of Panx3 were successfully reduced by ∼75% in the developing chicken forelimb. These in vivo results support the in vitro ob- servation by Ishikawa et. al. that knockdown retards long bone elongation [125], but we will also see that in vivo over-expression has no obvious effects. 1.4 Motivation, objectives, and highlights Chapters 2–4 can be separated into three general topics: Panx1 ohnologs Chapter 2 is an exercise in evolutionary analysis, describing the functional divergence of two separate panx1 genes in teleost fish. The goal of this study was to determine the origin of the genetic duplication, and to assess some of the physiological changes each gene product has acquired since their split. Genomic architecture strongly supports my hypothesis that the genes are in fact ‘ohnologs’, originating from a whole genome duplication event 48 that punctuated the radiation of the ray-finned fishes some 350 MYA. The expression profiles of these genes are clearly shown to have diverged, and EGFP-tagged versions of the proteins exogenously expressed in cell culture suggests that they have different trafficking patterns and channel properties. RF-Cloning.org Chapter 3 describes the restriction-free cloning method of molecular cloning (used extensively to create the various plasmid in the remaining chapters), and the development of a web service that automates the design process for the hybrid DNA primers central to the technique. The primary personal motivation driving this project was a desire to gain experience in computer science and bioinformatics, while the end product is a freely available re- source for the molecular biology community. The effort has resulted in a steadily increasing user base since the server’s official publication in June 2012. Panx3 in bone Chapters 4 and 5 explore the role or Panx3 in bone cells — providing a basis for its expression and describing the effects of altered expression in vivo. In chapter 4, the osteogenic transcription factor Runx2 is shown to transac- tivate the Panx3 promoter, leading to expression in mineralizing cell types like osteoblasts and hypertrophic chondrocytes. Given this distribution it is hypothesized that Panx3 is involved in the development of bone, so an avian retrovirus is used in chapter 5 to deliver exogenous Panx3, as well as Panx3 shRNA, into the developing forelimb of chicken embryos. Panx3 over-expression had no obvious impact of long bone morphology, although knockdown resulted in significant shorting of the limb, and was associated with changes in gene expression of late chondrocyte differentiation markers. 49 Chapter 2 Pannexin 1 ohnologs in the teleost lineage∗ “There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.” -Charles Darwin 2.1 Introduction Early chordate evolution was punctuated by two whole genome duplication (WGD) events (named ‘R1’ and ‘R2’) approximately 400–600 million years ago (MYA) [268, 269], with a third major WGD occurring in the ances- tral teleost lineage (R3) between 320 and 350 MYA [270]. These events were followed by significant reshuffling of the polyploid chromosomes via interchromosomal exchange, accompanied by massive loss of replicate genes through inactivation or deletion, and may have contributed to episodes of rapid speciation and radiation [270, 271]. While most duplicate genes de- rived from a WGD (i.e., ohnologs) are expected to be lost over time [272], at least 3–4% of the ohnologs created during the R3 WGD have been re- tained [273]. Neofunctionalization and subfunctionalization are believed to be the primary drivers of duplicate gene retention because random muta- tion will eventually inactivate at least one copy, unless a selective advantage is gained from retaining both [274]. Gap junction genes (connexins) have been notably well preserved in the teleosts following R3, with ∼37 func- tional members present in most species versus ∼21 mammalian members ∗A version of this chapter has been previously published [267] 50 [97]. Wagner [275] has argued that overall functional complexity in a pro- tein strongly correlates with the probability of both paralogs being retained if duplication occurs, or more specifically, that greater complexity increases the opportunity for subfunctionalization; mutations that ablate a different functional property from each copy can render both indispensable. Gap junction proteins are expressed by almost every vertebrate cell type, they interact with a diverse group of binding partners, and are involved in many physiological processes [24, 276], so it is perhaps unsurprising that so much connexin diversity has been retained by the ray-finned fishes. A second family of ‘gap junction-like’ proteins also exists within verte- brates, named ‘pannexin’ [90]. This small group of channel proteins (Panx1, -2 and -3) is homologous to the much larger innexin family, which are the invertebrate analogs of the vertebrate connexins. Innexins and connexins have a very similar structural topology and share many functional char- acteristics [92], yet it is unlikely these protein families are derived from a common ancestral gene that would have been classified as encoding a gap junction-forming monomer. Instead, they are believed to be the products of convergent evolution [70, 93]. While it appears that pannexins are able to form intercellular gap junctions to a limited extent in over-expression systems [102, 122], under normal physiological conditions they are now un- derstood to function as large unitary pores between the intra- and extracel- lular compartments [103]. As such, the ‘hemichannel’ nomenclature often used to describe unpaired connexin or innexin channels has been deemed inaccurate when referring to pannexin channels [128] and, thus, will not be used in this report. Regardless, the number of physiological processes in which pannexins have been implicated has rapidly grown in recent years [129], with numerous studies having assessed the diversity of innexins and pannexins across many phyla [4, 70, 91–93, 277]. Pannexin expression has even been studied specifically in fish [97, 110, 133, 179], but to date no one has determined if extra pannexins have been functionally preserved following R3. Here, we report that panx1 has in fact been retained as two independent ohnologs (panx1a and panx1b) and describe several features of the genes and EGFP-tagged versions of the gene products. 51 2.2 Materials and methods 2.2.1 Phylogenetic analysis Pannexin coding sequences were downloaded from the National Center for Biotechnology Information (NCBI) and Ensembl databases (Table 2.1) and analyzed within the Geneious 4.8 bioinformatics platform [278]. Global alignment of protein sequences was executed using the Blosum45 cost ma- trix, with an open gap penalty of 11 and an extension penalty of 1. Con- sensus phylograms were generated with the Geneious Tree Builder, using unweighted-pair-group method with arithmetic mean (UPGMA) in conjunc- tion with the Jukes-Cantor genetic distance model. Subsequently, boot- strapping with 1,000 replicates was used to estimate clade confidence [279]. Syntenic gene blocks associated with panx1 and shared between zebrafish and mouse were identified using the online synteny database [280]. Species Protein Accession D. rerio (zebrafish) Panx1a ENSDARP00000034149 D. rerio (zebrafish) Panx1b ENSDARP00000088509 D. rerio (zebrafish) Panx2 ENSDARP00000101872 D. rerio (zebrafish) Panx3 ENSDARP00000120825 G. aculeatus (stickleback) Panx1a ENSGACP00000017487 G. aculeatus (stickleback) Panx1b ENSGACP00000025338 G. aculeatus (stickleback) Panx2 ENSGACP00000017793 G. aculeatus (stickleback) Panx3 ENSGACP00000026588 G. gallus (chicken) Panx1 XP 001235339.2 G. gallus (chicken) Panx2 ENSGALP00000000002 G. gallus (chicken) Panx3 ENSGALP00000001381 G. morhua (Atlantic cod) Panx1a ENSGMOP00000002876 G. morhua (Atlantic cod) Panx1b ENSGMOP00000005970 G. morhua (Atlantic cod) Panx2 ENSGMOP00000020015 G. morhua (Atlantic cod) Panx3 ENSGMOP00000001486 L. chalumnae (Coelacanth) Panx1 ENSLACP00000005461 L. chalumnae (Coelacanth) Panx2 ENSLACP00000019572 L. chalumnae (Coelacanth) Panx3 ENSLACP00000010289 M. musculus (mouse) Panx1 ENSMUSP00000126405 M. musculus (mouse) Panx2 ENSMUSP00000124354 M. musculus (mouse) Panx3 ENSMUSP00000011262 O. latipes (medaka) Panx1a ENSORLP00000002658 Continued on next page... 52 Species Protein Accession O. latipes (medaka) Panx1b ENSORLP00000017032 O. latipes (medaka) Panx2 ENSORLP00000020445 O. latipes (medaka) Panx3 ENSORLP00000023811 O. niloticus (Nile tilapia) Panx1a XP 003456453.1 O. niloticus (Nile tilapia) Panx1b XP 003453849.1 O. niloticus (Nile tilapia) Panx2 XP 003447117.1 O. niloticus (Nile tilapia) Panx3 XP 003449860.1 T. nigroviridis (green spotted puffer) Panx1a1 ENSTNIP00000008776 T. nigroviridis (green spotted puffer) Panx1a2 ENSTNIP00000004506 T. nigroviridis (green spotted puffer) Panx1b ENSTNIP00000008995 T. nigroviridis (green spotted puffer) Panx2 ENSTNIP00000011195 T. nigroviridis (green spotted puffer) Panx3 ENSTNIP00000011448 T. rubripes (fugu) Panx1a ENSTRUP00000016615 T. rubripes (fugu) Panx1b ENSTRUP00000026699 T. rubripes (fugu) Panx2 ENSTRUP00000011590 T. rubripes (fugu) Panx3 ENSTRUP00000006083 X. tropicalis (Western clawed frog) Panx1 ENSXETP00000014090 X. tropicalis (Western clawed frog) Panx2 ENSXETP00000061696 X. tropicalis (Western clawed frog) Panx3 ENSXETP00000032094 Table 2.1: Pannexin protein sequences used for phylogenetic analysis 2.2.2 Real-time qPCR Twelve separate tissues were harvested from three individual zebrafish ac- cording to an approved University of British Columbia Animal Care proto- col (# A07-0288). Total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA), according to the manufacturer’s directions, and treated with DNAseI to remove genomic contamination. Relative pannexin expression was measured between tissues using quantitative real-time PCR (qPCR) on a CFX96 real-time qPCR machine (Bio-Rad, Hercules, CA). Primers are listed in 2.2, and 10 ng of total RNA was analyzed in duplicate 10 μl reactions using the iScriptTM One-Step RT-PCR kit with SYBR green (Bio-Rad). Expression levels were standardized against 18S ribosomal RNA (rRNA) (delta-CT (ΔCT)), and delta-delta-CT (ΔΔCT) was dynamically 53 based on the tissue with the lowest ΔCT in a given experiment. The tissues from each animal were analysed separately and then averaged. 2.2.3 Cloning zebrafish pannexins Total zebrafish messenger RNA (mRNA) was reverse-transcribed with Su- perScript III (Invitrogen), and each of the four pannexin cDNAs was PCR- amplified using primers containing 5’ EcoRI or EcoRV sites 2.3. PCR products were digested and ligated into pBlueScript downstream of the T7 promoter. The four genes were then subcloned into pEGFP-N1 using an appropriate double digest and ligation (panx1a, HindIII/BamHI; panx1b, KpnI/BamHI; panx2, HindIII/PstI; panx3, HindIII/BamHI), followed by restriction-free cloning (see chapter 3 for details) to remove the stop codons and to add a small linker sequence (GAAQSK) between the pannexins and EGFP. 2.2.4 Cell culture Human cervical carcinoma HeLa cells (American Type Culture Collection, Manassas, VA) were maintained in DMEM + 10% FBS in a humidified 37C incubator with 5% CO2 and transfected with the pEGFP-N1 constructs us- ing FuGENE6 (Roche, Indianapolis, IN) per the manufacturer’s directions. To produce stable over-expression, the growth medium was supplemented with 500 μg/ml G418 and cells were subjected to fluorescence-activated cell sorting once per week for 4 weeks to enrich for EGFP expression. 2.2.5 Western blot Transfected HeLa cell lysates were collected with radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM Tris-HCl [pH 8.0], 0.5% Sarko- syl, 1% IGEPAL, 0.1% SDS) and separated on 10% Tris-glycine SDS-PAGE gels. The protein was transferred to Immuno-Blot PVDF (Bio-Rad) and then blocked in 5% nonfat milk +0.1% Tween20 (NFM-T). Membranes were probed with an horseradish peroxidase (HRPO)-linked α-GFP mouse mon- oclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) diluted in 3% NFM-T for 2 h at room temperature. HRPO activity was visualized by treating the membrane with SuperSignal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL) and exposing/developing Bioflex Econo Film (Clonex, Markham, Canada). 54 Target Primer sequence Anneal. Size panx1a F: 5’-TTCGCTCAGGAAGTTTCTGTCGGT-3’ 60.3C 179 bp R: 5’-ACTGCCACCAGCAGCAGGATATAA-3’ 60.3C panx1b F: 5’-TAAGTATAAAGGCGTGCGGCTGGA-3’ 60.3C 169 bp R: 5’-ATACGCAGCCTGTCTCATCGTGAA-3’ 60.2C panx2 F: 5’-AGCCCTCGGTAATCAAAAGAC-3’ 54.8C 139 bp R: 5’-CGGACTGGTTTAGGTTTCTCTG-3’ 55.2C panx3 F: 5’-AGGGAAAACTCAATCTGGTGG-3’ 54.9C 146 bp R: 5’-GGTCTCCTTCAACTCATGGTAC-3’ 54.9C Table 2.2: qPCR primer sequences used to measure zebrafish pannexin ex- pression 55 Target Primer sequence Anneal. Size panx1a F: 5’-TAGAATTCGTCATGGCTATAGCGCACGCG-3’ 61.0C 1273 bp R: 5’-TAGAATTCTTAGATGACCCTCTGGCGGACAG-3’ 60.3C panx1b F: 5’-TAGAATTCACAATGGCTATAGCGCGGGTAGC-3’ 61.4C 1288 bp R: 5’-TAGAATTCTTAGACAACCCTTTGTCTTACATCCTTCACACAG-3’ 60.9C panx2 F: 5’-TAGAATTCGCCGGAATGCAGAATATCCTCGAGCAG-3’ 62.8C 1978 bp R: 5’-TAGAATTCTTAACACTCTCCAGCAGCAATCAGCTCG-3’ 62.1C panx3 F: 5’-TAGATATCAGCATGTCTATCGCCAACACGGC-3’ 61.9C 1354 bp R: 5’-TAGATATCTTAGGCATAACAGCATGTTGTGGTCTCCTTC-3’ 61.6C Table 2.3: Primer sequences used to clone the zebrafish pannexin cDNAs 56 2.2.6 Visualizing pannexin-EGFP Pannexin-EGFP transfected HeLa cells were grown in eight-well ibiTreat μ-Slides (Ibidi, Munich, Germany) for 12 h and supplemented with 20 mM HEPES buffer (pH 7.4) just prior to imaging. Confocal microscopy was performed on a Leica (Nussloch, Germany) TCS SP5II Basic VIS system, using the special photomultiplier R 9624 with low dark current. Time-lapse and z-stack images were analyzed with ImageJ∗. 2.2.7 Electrophysiological recording HeLa cells were bathed in a recording chamber filled with a modified Krebs- Ringer solution consisting of (in mM) 150 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2, 2 CsCl, 1 BaCl2, 2 pyruvate, 5 glucose and 5 HEPES (pH 7.4). The standard whole-cell recording pipette solution was composed of (in mM) 130 CsCl, 10 Na-aspartate, 0.26 CaCl2, 1 MgCl2, 2 EGTA, 5 tetraethylammonium (TEA)-Cl, and 5 HEPES (pH 7.2). Pipette resistance was 3–4 MΩ. Whole-cell recording was performed as described previously [281] with an EPC7 PLUS patch-clamp amplifier (HEKA Elektronik, Lambrecht, Ger- many). All currents in whole-cell configuration were filtered at 1 kHz (7-pole Besselfilter). Data were acquired at 4 kHz using an NI USB-6221 data- acquisition device from National Instruments (Austin, TX) and software written by J. Dempster (University of Strathclyde, Glasglow, UK). Steady- state currents for non-linear current-to-voltage relationship (I-V ) relations were measured between the ninth and tenth seconds of the 10 s membrane potential steps and are expressed as mean ± standard error of the mean (SEM). For outside-out patch recording, the recording electrode was lifted and pulled away from the cell after establishment of whole-cell configuration, thus excising the membrane patch attached to the electrode. 2.3 Results and discussion 2.3.1 At least four pannexin genes are present in the teleost lineage As of Ensembl release 66 (February 2012), seven species of ray-finned fishes have been annotated: zebrafish, stickleback, fugu, green spotted puffer, At- lantic cod, Nile tilapia, and medaka. Each of these species has a single ∗http://rsbweb.nih.gov/ij/ 57 record for Panx2 and Panx3, and two listings for Panx1. The green spotted puffer is an exception, with three listings for Panx1. Within each species analyzed, the two Panx1 proteins share an average 55.0% sequence identity (±6.9%) and 67.6% similarity (±7.3%). They retain the classic innexin- specific P-X-X-X-W motif in the second transmembrane domain and two cysteine residues in each extracellular loop [92] (Fig. 2.1). They also con- tain a charged K or R residue relative to position 75 in the mouse, which is thought to be involved in ATP-mediated channel regulation [166]. In- terestingly, we do not observe any conservation of the cysteine residue at position 282, previously reported to regulate channel activity of zebrafish Panx1 [133]. The authors of this study chose to use a bulky tryptophan residue to replace the native cysteine, which is a common practice when attempting to identify transmembrane residues with side chains that inter- act with the main body of the protein; it does not, however, reveal much about the actual function of the specific residue being replaced [282]. In this light, it seems that the impacts on channel activity are more likely to be a product of steric interference than ablation of a novel functional innovation associated with this particular cysteine. The only exception to the aforementioned conserved features is the ex- tra Panx1 sequence in the green spotted puffer, which appears to be the product of a partial duplication event that truncated the coding sequence immediately after the first cysteine residue in the second extracellular do- main (C234). Even if this truncated protein is still actively translated, it is very unlikely it can participate in normal channel activity because all four extracellular loop cysteines are needed for formation of active channels [75]. Multiple pairwise alignment splits the teleost Panx1 sequences roughly into two orthologous clades, and these groups combine into a single sister clade relative to Panx1 sequences from more distantly related vertebrates (Fig. 2.2). The zebrafish sequences complicate this phylogeny to some degree because both paralogs group into a single clade within the teleost Panx1 branch. This is probably due to the fact that the zebrafish belongs to the taxonomic group Otocephala, as opposed to the rest of the fish species in this study which belong to the Euteleostei, and these two lineages diverged approximately 250–300 MYA [89]. One of the sequences does, however, group more tightly with its respective clade, so this property along with genomic positioning will be utilized below for naming purposes. 58 {P-X-X-X-WCC CC * + Figure 2.1: Multiple pairwise alignment of Panx1 sequences. Classic pan- nexin/innexin specific residues are identified, including four cysteine residues in the extracellular loops (C), a P-X-X-X-W motif in the second transmem- brane domain, and a positively charged residue thought to facilitate ATP gating of the channel (⊕). A cysteine residue in Panx1a (∗) has also been reported to facilitate channel activity in zebrafish, but is not present in any other sequence included in this study. Predicted transmembrane domains are denoted by filled boxes 59 mmPanx2 xtPanx2 ggPanx2 lcPanx2 olPanx2 drPanx2 gmPanx2 onPanx2 gaPanx2 tnPanx2 trPanx2 xtPanx1 ggPanx1 mmPanx1 lcPanx1 olPanx1a onPanx1a gaPanx1a gmPanx1a trPanx1a tnPanx1a1 tnPanx1a2 drPanx1a drPanx1b tnPanx1b trPanx1b gmPanx1b olPanx1b gaPanx1b onPanx1b xtPanx3 lcPanx3 ggPanx3 mmPanx3 drPanx3 olPanx3 gmPanx3 tnPanx3 trPanx3 gaPanx3 onPanx3 frogxt → mousemm → coelacanthlc → zebrafishdr → green spotted puffertn → fugutr → sticklebackga → Atlantic codgm → medakaol → Nile tilapiaon → chickengg → Panx2 100 100 75 54 100 68 76 92 100 41 92 Panx1 100 75 78 100 94 81 93 82 53 100 97 42 64 80 10049 75 83 100 Panx3 100 98 49 99 100 59 100 100 96 63 0.1 Figure 2.2: Phylogram illustrating the phylogenetic relationship between pannexins. The tree was generated from aligned protein sequences using UPGMA and Jukes-Cantor genetic distance, with branch length propor- tional to genetic distance. Bootstrap values are indicated at branch points, and were calculated through 1000 replicates 60 2.3.2 The two teleost panx1 genes likely originate from the R3 WGD event Clupeocephala is the lowest taxonomic group to include all of the teleosts present in this study, so the most parsimonious time frame for the duplica- tion of panx1 precedes the Clupeocephala/Elopomorpha split 300–350 MYA [89]. Gene duplications occur through various processes, including retro- transposition, errors during homologous recombination, and whole chromo- some or whole genome duplications [283]. It is highly unlikely that the extra copy of panx1 resulted from reintegration of a processed mRNA into the genome by a retrotransposon, because the exon architecture is nearly identical between the two genes (Fig. 2.3). Although not impossible, it is also unlikely that the duplication was the product of an unequal crossover event. These are usually characterized by a tandem repeat pattern [284], and the two panx1 genes are located on separate chromosomes. For this process to have generated the current localization of the two panx1 genes, a homologous recombination error would need to have been followed in rela- tively quick succession by an interchromosomal exchange and, depending on when this occurred relative to the WGD, either one or two extra panx1 par- alogs must have then been deleted from the genome. A more parsimonious explanation is that the extra gene was produced during the R3 WGD. 1kb panx1a panx1b Figure 2.3: Genomic architecture of the two zebrafish panx1 genes. Exons are indicated by red boxes, with intronic regions represented by thin lines. Feature lengths are to scale The initial period following an instance of tolerated polyploidy (i.e., when the extrachromosomal load does not kill or sterilize the organism) is gener- ally associated with genomic instability, and recombination shuffles alleles 61 between all homologous chromosomes during meiosis [285]. Homologous recombination is also thought to facilitate rapid genomic downsizing of a neopolyploid species [286], with deletions presumably remaining innocuous so long as the affected regions are present on the replicate chromosome. Eventually, sufficient divergence causes the karyotype to once again assume a diploid state, but two key signatures of the duplication will remain for a considerable time afterwards. First, large syntenic regions can be expected to exist between the duplicated chromosomes, preserving gene order and ori- entation. Second, comparison against the genome of a related species that did not undergo WGD should reveal a pattern of gene interleaving between the duplicated chromosomes relative to the homologous outgroup chromo- some [270]. The chromosomal neighbourhoods of the two zebrafish panx1 genes contain signs of both synteny and gene interleaving. For example, the panx1 gene on Danio chromosome 15 (Dre15) is flanked by the same set of genes observed on mouse chromosome 9 (Mmu9), while a block of genes adjacent to the panx1 gene on Danio chromosome 5 (Dre5) is many megabases away on both Mmu9 and Dre15 (Fig. 2.4). The genes adjacent to panx1 on Dre5 have been previously annotated as ‘b’ ohnologs (e.g., tm- prss13b, tmprss4b, and scn4bb), while the genes on Dre15 are annotated ‘a’. As such, the zebrafish panx1 genes should now be referred to as panx1a and panx1b on Dre15 and Dre5, respectively. Furthermore, despite both of the zebrafish panx1 gene products clustering into a single clade of the remaining teleost panx1 sequences, panx1b shares greater similarity with that clade, so the genes for the other species included in Fig. 2.2 have been annotated accordingly. 62 Dre5 pa nx 1b  (3 9.4 Mb ) tm prs s13 b (39 .5M b) tm prs s4b  (3 9.5 Mb ) mp zl2  (39 .5M b) scn 4b b ( 39 .5M b) Mmu9 Dre15 Pa nx 1 ( 14 .8M b) Am ot l1 (1 4.3 M b) An kr d4 9 (1 4.6 M b) M re 11 a ( 14 .6M b) panx1a (2.5Mb) MED17 (127Kb) C11orf90 (133Kb) KIAA1731  (119Kb) mre11a (2.8Mb) ankrd49 (2.8Mb) AMOTL1 (2.8Mb) zgc:165649 (2.4Mb) He ph l1 (1 4.9 M b) 22 00 00 2K 05 Ri k ( 15 .0M b) M ed 17  (15 .1M b) 58 30 41 8K 08 Ri k ( 15 .1M b) tmprss4a (13.0Mb) tmprss13a (13.1Mb) scn4ba (13.0Mb) MPZL2 (46.9Mb) M pz l2 (4 4.9 M b) Sc n4 b (4 4.9 M b) Tm pr ss 4 ( 45 .0M b) Tm pr ss 13  (45 .1M b) Figure 2.4: Syntenic relationship between the zebrafish and mouse chromosomal regions containing panx1 genes. Individual genes are represented by solid boxes (red for panx1, blue for all others), and orthologous pairs are indicated with connector lines 63 2.3.3 Distinct expression profiles of panx1a and panx1b To compare expression levels of each pannexin throughout the adult ze- brafish, mRNA was prepared from 12 separate tissues and analyzed by real- time qPCR (Fig. 2.5). In line with previous studies, expression of panx2 was primarily restricted to the eye and central nervous system [97, 102, 109], and panx3 was highest in skin [105]. The distribution of zebrafish panx1a has previously been reported in the central nervous system, muscle, heart, liver, kidney, and retina [97, 110], similar to our observations here showing the near ubiquitous expression pattern characteristic of mammalian panx1 [102]. Previous attempts to isolate or measure panx1b transcript were unsuc- cessful [110], but we were able to observe robust expression in complemen- tary DNA (cDNA) prepared from brain and eye with more modest relative levels of expression in heart, kidney and spleen. 64 020 40 60 80 100 R e l a t i v e  m R N A  l e v e l s panx1a 0 20 40 60 80 100 R e l a t i v e  m R N A  l e v e l s panx2 0 20 40 60 80 100 R e l a t i v e  m R N A  l e v e l s panx3 0 20 40 60 80 100 R e l a t i v e  m R N A  l e v e l s panx1b Figure 2.5: Relative mRNA levels of pannexins expressed by zebrafish tissues, as measured by qPCR. Each graph is the average of three independent experiments, where ΔΔCT was based on the tissue with highest expression in a given experiment. Error bars represent standard error 65 2.3.4 Subcellular dynamics and localization of the zebrafish pannexins The coding sequences of all four zebrafish pannexin genes were amplified from a multi-tissue preparation of total cDNA and cloned into the expres- sion vector pEGFP-N1, so trafficking of the proteins could be monitored live. HeLa cells were chosen for this study because they express very little endogenous connexin [32], and while reports on the expression of panx1 in HeLa are mixed [170, 178, 287], we were unable to observe the protein by Western blot (data not shown). Western blot analysis of lysates taken from stably transfected HeLa cultures confirmed the presence of EGFP-tagged products of expected size for all constructs (Fig. 2.6). Time-lapse imaging revealed distinct cellular distributions for each of the four pannexins (Fig. 2.7). A fraction of Panx1a-EGFP localizes to the plasma membrane, with concentrations in areas of membrane ruffling (Fig. 2.7a). This is consistent with previous reports of Panx1 localizing to the leading edge of motile cells [288], probably through direct interaction with filamentous actin [111]. Panx1b-EGFP also localizes to the plasma mem- brane with recruitment to dynamic membrane ruffles, but most of the cells analysed also had a fraction of the protein associated with mobile intra- cellular vesicles with diameters of about 200–500 nm (Fig. 2.7b). Many vesicles within the endocytic pathway are approximately this size [289], but we observed no co-localization between Panx1b-EGFP and the early endo- some marker EEA1 (data not shown); thus, the identity of these vesicles remains undetermined at this time. Mammalian Panx2 has been reported to localize primarily to small intracellular vesicles [173, 174], similar to what we observed with the zebrafish ortholog (Fig. 2.7c). Multiple splice variants exist for this protein however [97], of which we isolated only the ‘C’ vari- ant, while context-dependent depalmitoylation has been shown to facilitate trafficking of Panx2 to the plasma membrane [112]. Finally, Panx3-EGFP expression appeared to be primarily intracellular (Fig. 2.7d). ‘Normal’ lo- calization of Panx3 is variable in the literature, and it is also probably cell type dependent and can be disrupted when tagged with GFP [111, 169, 170]. Given all these potentially confounding factors with regard to pannexin traf- ficking, we do not assume our results are necessarily equivalent to normal in vivo dynamics but instead hope to highlight the differences among the four proteins when expressed in a common environment, with particular emphasis on the two Panx1 proteins. According to the classic duplication- degeneration-complementation model [272], we expect Panx1a and Panx1b to have undergone some degree of neofunctionalization at the transcriptional 66 level and/or the physiological level. While the differences we have observed in relative mRNA expression and intracellular distribution of the EGFP- tagged proteins appear to support this position, the clearest evidence of neofunctionalization would be a measurable difference in channel proper- ties. Pa nx 1a -E GF P Pa nx 2- EG FP Pa nx 3- EG FP Pa nx 1b -E GF P EG FP 37 50 75 100 kDa Figure 2.6: Immuno-blot confirming exogenous expression of EGFP tagged zebrafish pannexins in HeLa cells, using an anti-GFP antibody. Several of the lanes were rearranged with image processing software for easier com- parison (as indicated by thin black lines), but all lanes are from the same Western blot. 2.3.5 Physiological properties of zebrafish Panx1 channels Individual Panx1a-EGFP, Panx1b-EGFP or EGFP control transfected HeLa cells were voltage-clamped at a holding potential of -30 mV for 1 s, fol- lowed by voltage steps of 10 s duration to potentials in the range of -60 to 67 20 μm Panx3-EGFP 20 μm20 μm Panx1a-EGFP Panx1b-EGFP 20 μm Panx2-EGFP Figure 2.7: Exogenous expression of EGFP-tagged zebrafish pannexins in HeLa cells. Panx1a localized to the cell membrane, and is recruited to areas of membrane ruffling (arrow). Panx1b localizes to the cell membrane as well, but is also present in intracellular vesicles (arrowhead). Panx2 was exclusively observed in small intracellular vesicles, while Panx3 was more diffuse 68 +60 mV in 10 mV increments (Fig. 2.8). Similar to previous reports for Panx1a [133], the macroscopic currents from cells expressing our constructs were characterized by an outwardly rectifying, I-V at positive potentials. Panx1b-EGFP appears to have a voltage threshold for activation (i.e., the point where I-V breaks from linearity) between +20 and +30 mV, which is higher than the 0 to +20 mV necessary to activate Panx1a-EGFP currents. Panx1b-EGFP also exhibits a longer activation time than Panx1a-EGFP, re- quiring upward of 8 s to reach steady state upon membrane depolarization versus < 250 ms for Panx1a-EGFP. These results match reasonably well with those reported previously for Panx1a properties [110, 133]. Record- ings from outside-out patches of Panx1b-EGFP transfected cells displayed unitary events upon stepping the membrane potential to +30 mV or above (Fig. 2.9a), and analysis of an all-event histogram indicates a Panx1b-EGFP single-channel conductance of ∼123 pS (Fig. 2.9b). This is nearly fourfold lower than has previously been reported as the unitary conductance of fish or mammalian Panx1 [131, 133]. Unexpectedly, unitary opening events could not be resolved from any excised outside-out patches taken from over 60 Panx1a-EGFP transfected cells, and yet the increase in whole-cell currents at positive membrane potentials does still appear to be the result of pannexin channels because the current was completely inhibited by 25 μM carbenox- olone, returning the I-V plot to a linear relationship without affecting the steady-state currents at negative membrane potentials (Fig. 2.10). The ad- dition of bulky tags like GFP to the carboxy terminus (CT) of connexin can alter channel gating properties and unitary conductance [281, 290, 291], and while CT tagging Panx1a with EYFP has been shown to have no untoward effect on unitary recordings [133], the addition of an EGFP tag to the CT of human Panx1 completely blocked channel activity [137]. As such, it is not unreasonable to speculate that the tag in our system could be causing a partial blockade and/or altering open probability, and indeed, closer ex- amination of whole-cell currents taken from Panx1a-EGFP transfected cells revealed infrequent but well-resolved unitary events of ∼276 pS at potentials ≥ +50 mV (data not shown). These events could represent the occasional transition of a partially open channel to a more fully open state, but further work with the untagged proteins will be required to properly resolve this issue. 69 2.4 Conclusion In the current study we have demonstrated that a fourth pannexin gene is present and actively expressed in the ray-finned fishes. This gene is probably a holdover from the teleost R3 WGD event, representing a split of panx1. As such, the two panx1 genes should now be referred to as panx1a and panx1b. These genes display distinct differences in tissue distribution, with the panx1a expression pattern mimicking the near ubiquity of mammalian panx1, while panx1b is heavily enriched in the brain and eye. Exogenous over-expression of the zebrafish ohnologs reveals potential differences in the intracellular vesicles with which each protein associates, but both clearly traffic to the plasma membrane, particularly to areas of cell ruffling. At the channel level the two proteins appear to have distinct physiological proper- ties, in terms of both gating and conductance, although future electrophys- iological characterization of the untagged versions of these channels will be needed to fully assess the extent of their differences. Taken together, our results indicate that the two panx1 genes and gene products have undergone some degree of neofunctionalization or subfunctionalization, as would be ex- pected according to conventional evolutionary theory. To our knowledge, this is the first time functional properties of panx1b have been reported. 70 EGFP Panx1b-EGFP Panx1a-EGFP 2 s -30 mV 60 mV -60 mV A B -20 -50 50 100 150 200 -60 -40 20 40 60 EGFP (n=15) Panx1a-EGFP (n=18) Panx1b-EGFP (n=10) V  (mV)m I   (pA)m 1 s 50 pA Figure 2.8: Whole-cell voltage clamp of EGFP, Panx1a-EGFP and Panx1b- EGFP transfected HeLa cells. A) The protocol included a brief holding potential of -30 mV, followed by 13 consecutive 10 s holding steps starting at -60 mV and increasing depolarization by 10 mV per step. Depolariza- tion to positive membrane potentials evoked progressively increasing mem- brane currents from Panx1a-EGFP and Panx1b-EGFP compared to EGFP controls, but the activation time was much longer for Panx1b-EGFP than Panx1a-EGFP. B) I-V plot demonstrating the voltage-gated pannexin cur- rents, in contrast to the linear (background) I-V relationship recorded in HeLa cells expressing EGFP only 71 + 60 mV 1 s 25 pA + 50 mV + 40 mV + 30 mV + 20 mV + 10 mV 0 0 500 1000 1500 2000 2500 C O1 2 3O O 123 γ (pS) N um be r o f e ve nt s 246 369 Panx1b-EGFPA B Figure 2.9: Single-channel recordings of Panx1b-EGFP demonstrating uni- tary event activity and single-channel conductance. A) Representative traces from excised outside-out patches reveal single-channel activity for Panx1b-EGFP channels at membrane potentials of +30 mV and above. B) An all-point histogram representing all six example traces illustrates a uni- tary conductance of ∼123 pS between the closed state (C) and fully open state (O1). The peaks at O2 and O3 are both multiples of 123 pS and, thus, are most likely the result of multiple channels in the excised patch. The histogram shows some background activity that may be caused by sub- conductance states 72 604020 A B 50 100 -40-60 -50 -100 0 Panx1a-EGFP + CBX (25 μM) -20 0 Panx1a-EGFP n = 3 (paired) Panx1a-EGFP Panx1a-EGFP + CBX (25 μM) V  (mV)m 4 s 50 pA I   (pA)m Figure 2.10: Panx1a-EGFP is sensitive to CBX. A) Representative traces from a single Panx1a-EGFP expressing cell recorded in the whole-cell con- figuration and subjected to incremental 10 mV steps from -60 to +60 mV before (left) and after (right) treatment with CBX. B) I-V plot illustrating the reduction in voltage-activated currents from Panx1a-EGFP expressing cells following CBX treatment as well as the lack of effect on steady-state currents at negative membrane potentials 73 Chapter 3 RF-Cloning.org: An online tool for the design of restriction-free cloning projects∗ “An identifying characteristic of the Khorana approach to the production of synthetic polynucleotides for the solution of bio- logical problems has been a willingness to use both chemical and enzymatic tools as appropriate, and this catholic approach is the foundation of all the molecular technology employed in modern molecular genetics.” -Michael Smith 3.1 Introduction Classical DNA cloning generally involves cleaving a destination plasmid and a target insert sequence with restriction enzymes, and then stitching them together with DNA ligase. This approach is enormously convenient and straight forward when the appropriate restriction sites are well posi- tioned in the sequences being manipulated, but becomes problematic when these restriction sites are not present. A number of ligation-independent cloning technologies have been developed that utilize recombinase enzymes, such as the Gateway system by Invitrogen and In-Fusion by Clontech. Gateway requires the presence of recombinase-specific attR and attL se- quences [293], so there is no gain in overall flexibility. In-Fusion allows for recombination to occur at user-specified sites, but the per-reaction cost can ∗A version of this chapter has been previously published [292] 74 be high due to the proprietary recombinase required. polymerase chain reac- tion (PCR) based restriction site-free cloning (RF-cloning) overcomes these limitations, and has been described in the literature numerous times in the past decade [294–298]. RF-cloning is based on the overlap extension site- directed mutagenesis technique first described by Steffan Ho in 1989 [299], and commercialized by Stratagene under the name QuikChange [300]. The technique is initiated with a pair of primers each designed with complemen- tary sequence to both the desired insert and the destination plasmid. High- fidelity PCR is used to first amplify the insert sequence, and then the result- ing product is purified for use as a ‘mega-primer’ in a secondary PCR reac- tion, using the destination plasmid as template. During this second reaction the destination plasmid is amplified in its entirety. PCR customarily results in geometric expansion of product, which can amplify point mutations that occur during synthesis. The mega-primer on the other hand only initiates amplification from the parental destination plasmid, resulting in arithmetic accumulation of daughter molecules; this is sufficient for cloning purposes while minimizing mutations. When these newly synthesized strands anneal, the mega-primers act as long complimentary overhangs that circularize the plasmid, forming a nicked hybrid molecule. DpnI is used to degrade the methylated parental plasmid while leaving the unmethlyated in vitro syn- thesized hybrid plasmids intact and ready to transform competent bacteria (Fig. 3.1). Currently, designing the hybrid RF-cloning primers is a manual task. Four separate primer sequences must be created (two for the insert and two for the plasmid) with compatible melting temperatures (Tm), requiring the investigator to arbitrarily test sequences of varying length until the proper conditions are achieved . To automate primer design, we have written an algorithm that accepts a user specified insert sequence, destination plasmid sequence, and the desired insert sites as input. The algorithm then returns hybrid primers with the correct orientation and compatible Tm. The service can be accessed through our web based user interface, or through direct extensible markup language (XML) requests. 3.2 Project workflow From the RF-Cloning.org home page (Fig. 3.2), the user supplies the se- quence of their insert and destination plasmid in plain text or FASTA for- mat. The desired insertion sites in the destination plasmid can be specified numerically, or by placing exclamation marks (!) directly into the plas- 75 DpnI DpnI DpnI DpnI Synthesized destination Insert Parental destination Figure 3.1: Hybrid primers are designed with complementarity to the de- sired insert (green) and the destination plasmid (blue). A first round of PCR is performed to create a ‘mega-primer’ comprising the insert sequence flanked by sequences complementary to the destination plasmid. During a second round of PCR, the mega-primer initiates replication of the destina- tion plasmid (pink). Since the entire destination plasmid is replicated in the reaction, mega-primer binding to a daughter molecule fails to expose a free 3’ end for polymerase elongation, so accumulation of new product is linear. The destination plasmid must be purified from a DAM+ bacterial stain, since DpnI is used to selectively degrade parental DNA after the sec- ond PCR reaction, leaving the unmethylated daughter products intact. The reaction can then be used to transform competent bacteria 76 mid sequence. A list of common plasmids is supplied as a pull-down menu, and the sequence and associated plasmid map are loaded automatically if one is selected. If the user inputs a custom destination plasmid sequence, common features and restriction sites can be automatically identified by clicking the ‘draw plasmid’ button; this will then dynamically generate a plasmid map. The database of common plasmid features has been adapted from a listing at AddGene∗, which is an expanded list of the features used by PlasMapper [301]. A number of useful sequence manipulation tools, some of which have been adapted from those found at Sequence Massager†, are also included to perform common tasks, such as removal of numbers and whitespace or converting a sequence to its reverse complement. When the user runs the project, they are redirected to the output page, where they will find the newly designed hybrid primers, recommended PCR conditions, and new construct sequence (Fig. 3.3). The hybrid primers are necessarily a direct function of the insert sequence and the insertion sites within the destination plasmid, and the design algorithm constrains the final primers to a minimum length and annealing temperature of 18 bases and 55◦C on the insert side, and 22 bases and 60◦C on the plasmid side. These values are based on general PCR primer design best practices, as well as rf-cloning spe- cific recommendations [294, 302]. Annealing temperature is defined by the Wallace-Itakura rule for sequences less than 14 bases long (C = 4[G + C] + 2[A + T]), while nearest-neighbour thermodynamics are used for sequences 14 bases long or greater [303, 304]. The nearest neighbour calculations as- sume a primer concentration of 500 nM, monovalent cation concentration of 50 mM, and divalent cation concentration of 0 mM. Of course magnesium has a stabilizing effect on DNA hybridization [303], and is also a necessary component of the PCR reaction buffer, so the calculated annealing tem- perature is 5-10◦C below the actual primer Tm expected in the final PCR reaction. The default minimums for length and annealing temperature can be over-ridden by the user from the main page using the ‘advanced settings’ form. All new projects are assigned a unique 32 byte hash code and immedi- ately saved to the database at runtime, and the hash code is appended to the output page uniform resource locator (URL) so the user can bookmark their projects. The length of the hybrid primers and insertion sites can be manually adjusted on the output page by clicking the associated arrow but- tons, and a save button is provided to update the project in the database ∗http://www.addgene.org/plfeatures.html †http://www.attotron.com/cybertory/analysis/seqMassager.htm 77 Figure 3.2: 1) Optional information about the project. 2) ‘Orientation’ refers to complementarity of the insert sequence relative to the destination plasmid, and ‘arrow’ refers to how the insert will be drawn in the Savvy map. 3) Collapsible form fields where the default primer length and annealing temperature parameters can be modified. 4) Destination plasmid sequence can be entered manually (numbers and FASTA comments are ignored) or by selecting popular plasmids from the dropdown menu. 5) Insert sites within the plasmid sequence should be designated numerically in the fields provided (with the very front of the plasmid is position 0), or by placing exclamation points (!) directly into the sequence. 6) Insert sequences must be entered manually (numbers and FASTA comments are ignored). 7) Plasmid maps are dynamically generated based on destination plasmid sequence. 8) When ready, the project can be submitted to the server 78 should any changes be made. If users would prefer to not keep track of their project hash keys manually, they are able to register a free account to access a simple project/plasmid management system where they can save their work. Throughout the site, users are able to graphically visualize plasmids using an integrated version of ‘Savvy’. Savvy is a scalable vector graphics (SVG) plasmid map drawing software adapted from the original version 0.1 source code kindly provided by Dr. Malay Basu‡. These plasmid maps are generated dynamically, and if the user wishes to retain a copy for their own records, the SVG file can be downloaded by clicking the associated ‘print’ symbol in the lower right hand corner of the image. The SVG format can be opened by vector graphics manipulation software from various commercial vendors (e.g., Adobe Illustrator and CorelDRAW) as well as the open source SVG editor, Inkscape§. Plasmids can also be exported in genebank format, which is importable by most plasmid management platforms. 3.3 Recommended rf-cloning protocol By default, the starting hybrid primers will be at least 40 basepairs (bps) long with a Tm of at least 55 ◦C for the primary PCR (amplification of insert) and at least 60◦C for the secondary PCR (extension around the plas- mid). High fidelity deoxyribonucleic acid (DNA) polymerase should be used for all PCR reactions, and it has been our experience that iProof (BioRad, Hercules, CA) and Phusion (New England Biolabs, Ipswich, MA) produce consistent results. To generate the mega-primer, use a standard 50 μl PCR reaction (1x PCR buffer, 200 μM dNTP, 500 nM of each primer, 1 U poly- merase, user defined amount of starting template), and cycle 30-35 times with the RF-cloning.org recommended annealing temperature, followed by product purification (e.g., by gel extraction). A recent study has identi- fied the optimal conditions for the secondary PCR reaction [294]: a molar insert:plasmid ratio of ≥20, using 20-50 ng of parental plasmid as starting material, 18 to 20 amplification cycles (note that cycling more than 20 times has significant detrimental effects), and setting the annealing temperature to 5-10◦C below Tm. A 20 μl reaction is sufficient for the secondary PCR, and upon completion it should be treated with 20 U of DpnI for two hours at 37◦C (DpnI is active in the iProof and Phusion PCR buffers), followed by a 20 minute inactivation at 80◦C. RF-cloning reactions are inherently low ‡http://www.bioinformatics.org/savvy/ §http://inkscape.org/ 79 Figure 3.3: 1) A unique 32 byte hash code is generated for all new projects, and is present in the URL for bookmarking purposes. 2) The hybrid primers are color coded, blue for sequence complementary to the plasmid, and green for the insert. The length of the primers can be adjusted by clicking on the arrow buttons if the user wishes to alter the annealing temperature. 3) If the insert site needs to be adjusted, the user can use the provided arrow buttons. 4) The secondary PCR conditions are optimized for iProof or Phusion as the polymerase, so the user should follow manufacturer’s instructions if using another high fidelity enzyme. “Insert” refers to the mega-primer purified from the primary PCR reaction. 5) The entire sequence of the new plasmid is output, with insert in green and parental plasmid in blue. 6) The plasmid map can be drawn by specifying the positions of markers manually, or by auto-finding common features. Commercially available restriction enzyme cut sites can also be specified or automatically identified. If desired, the plasmid can also be exported as a genbank file. 7) All projects are automat- ically saved, but making changes to the output page will activate the save button so those changes can be uploaded to the database. If the user has registered an account to access the plasmid management system, the save button will attach the project to their profile. 80 efficiency, so using super-competent bacteria (>108 CFU/μg pUC18) for the subsequent transformation can be beneficial, although we usually find sub-cloning grade competent cells (106 CFU/μg pUC18) to be sufficient. The design algorithm at RF-Cloning.org can also be used to create hybrid primers compatible with other overlap-extension based methods, such as fusion PCR[305] or the recently described transfer-PCR[306], although the parameters associated with the downstream protocol will need to be opti- mized by the user. 3.4 Implementation RF-Cloning.org runs on a standard LAMP configuration (Debian web server, Apache2, MySQL, and PHP), and the user interface has been successfully tested on Firefox 3+, Opera 9+, Safari 3+, Internet Explorer 9+, and Chrome web browsers, on Windows XP/Vista/7, Debian based distributions of Linux, and Mac OS-X. All critical functionality of the site is also present when using Internet Explorer 7 and 8, but due to incomplete support of the SVG web standard by these older browsers, they are unable to download the Savvy plasmid maps. Many features of RF-Cloning.org communicate with the web server with asynchronous JavaScript calls, so JavaScript sup- port must be enabled at all times. Common gateway interface scripts are implemented in PHP, Perl (using preexisting BioPerl modules [307]), and C++ (select components of the BLAST+ suite, version 2.2.24, NCBI C++ toolkit [308]). For those wishing to access the RF-Cloning design algorithm directly from custom scripts, we have included a Simple Object Access Pro- tocol (SOAP) server to accept incoming XML requests¶. The affiliated WSDL file and sample SOAP clients are also downloadable‖. 3.5 Conclusion Restriction-free cloning is a powerful technique for constructing custom plas- mids, but to date, designing the necessary hybrid primers has been a manual task. We have created rf-cloning.org to help automate hybrid primer design, which is both faster and reduces the likelihood of human error. An empha- sis has been placed on creating an intuitive browser-based user interface, although the underlying design algorithms can be accessed by direct XML requests if a user wishes to create their own custom batch scripts. ¶http://www.rf-cloning.org/classes/rf_cloning_server.php ‖http://www.rf-cloning.org/soap_server.php 81 Chapter 4 Pannexin3 is a novel target for Runx2, expressed by osteoblasts and mature growth plate chondrocytes∗ “We must trust to nothing but facts. These are presented to us by Nature, and cannot deceive. We ought, in every instance, to submit our reasoning to the test of experiment, and never to search for truth but by the natural road of experiment and observation.” -Antoine Laurent Lavoisier 4.1 Introduction Pannexins (Panxs) are a class of Chordate channel proteins initially iden- tified by their homology to the insect gap junction proteins, called innex- ins (Inxs) [90]. Significant sequence similarity does not exist between the Inx/Panx family and the connexins (Cxs), which are the canonical Chor- date gap junction proteins, although the two classes share a high degree of predicted structural similarity [70, 90], suggesting a potential overlap in functionality. Most mammalian tissues express at least one of the three Panx sub-types [104, 106], and they have been shown to influence a number of cel- lular processes, including inflammation, ischemic tolerance in neurons, and cancer [122, 174, 198, 310]. In contrast to channels formed by Inxs and Cxs, Panx channels do not significantly contribute to direct cell-cell gap junc- tional communication, but instead, tightly regulate the transfer of ions and small molecules between the cytoplasm and extracellular space [130, 166]. It ∗A version of this chapter has been previously published [309] 82 has also been suggested that some of the processes originally attributed to Cx ‘hemichannel’ activity may in fact be the result of Panx channels [311]. Of the three isoforms (Panx, 2, and 3), Panx1 has the widest expression profile and demonstrates the clearest channel activity [102]. adenosine-5’- triphosphate (ATP) in particular has been repeatedly reported to traverse the Panx1 channel under physiological conditions [126, 131, 166]. Panx2 is primarily observed in neurons of the central nervous system [102, 173], and contains a large hydrophilic domain on its C-terminal tail that sets it apart from the rest of the innexin superfamily [70]. Functionally, Panx2 has been reported to suppress glioma growth in vivo [174] and to regulate the differentiation of neuronal progenitor cells [112], although its role as a chan- nel still requires further study [107]. Panx3 has received the least attention within the Panx family, most likely owing to its restricted expression profile [91], and a lack of observable channel activity during the initial descriptive experiments reported by Bruzzone et al. in 2003 [102]. Panx3 channel ac- tivity has since been observed in response to mechanical stimulus in vitro [104], but the functional role of Panx3 in vivo is unclear, and its mode of transcriptional regulation is undescribed. Reverse-transcription PCR (RT-PCR) screens indicate that Panx3 mes- senger RNA (mRNA) is present in various tissues, including skin, kidney, spleen, and brain [102, 104], some of which have been verified using im- munofluorescence and Western blot [104]. However, an in silico analysis of gene expression databases reveals a significant enrichment of Panx3 in skele- tal tissues [218]. This has been verified in a number of osteoblast cell lines as well as primary calvarial osteoblasts [125, 169, 170, 312], and visualized more directly by immunofluorescence in the mouse cochlea, where Panx3 was seen exclusively in bone [175]. Interestingly, Panx3 is also expressed in hypertrophic chondrocytes during the development of axial skeleton long bones [125, 169]. Osteoblasts and chondrocytes are derived from a shared osteochondral progenitor cell type, which differentiates in response to local conditions. While the transcriptional programs of the two cell types are quite distinct, there is a convergence late in the maturation process of chondro- cytes destined for endochondral ossification. These chondrocytes, located in the diaphysis of maturing long bones, cease mitosis and become terminally differentiated by switching their gene expression profile to resemble that of osteoblasts [313]. We therefore hypothesized that Panx3 is regulated by one or more of the osteoblastic transcription factors utilized by chondrocytes late in their life cycle. Here, we report the expression pattern of Panx3 in developing bone with emphasis on the high levels observed in both osteoblasts and mature 83 osteogenic chondrocytes. We further show that Runt-related transcription factor 2 (Runx2), a key transcription factor for normal bone formation, binds to and transactivates the Panx3 promoter. These results build upon previous reports of Panx3 expression in osteogenic cell types by qualifying the specific temporal and spatial distribution of this expression, and also identify an important part of the regulatory mechanism controlling its transcription. 4.2 Materials and methods 4.2.1 Animal care All experiments were performed in accordance with the guidelines estab- lished by the Canadian Council on Animal Care and were approved by the University of British Columbia Animal Care Committee. 4.2.2 Antibodies A rabbit anti-Panx3 polyclonal antibody was developed and described previ- ously [104]. All other antibodies used during this study were obtained from the following sources: Goat anti-osteopontin (PA1-25152; Pierce, Rockford, IL, USA); mouse anti-GAPDH (5G4; HyTest, Turku, Finland); rabbit anti- Pthr1 (ab75150; Abcam, Cambridge, MA, USA); mouse anti-Collagen type 1α1, developed by H. Furthmayr [314] (SP1.D8; Developmental Studies Hy- bridoma Bank (DSHB), University of Iowa, Iowa City, IA, USA); mouse anti-Collagen type X, developed by T.F. Linsenmayer (X-AC9; DSHB); rab- bit anti-Runx2 (M-70; Santa Cruz Biotechnology, Santa Cruz, CA, USA); rabbit anti-Cx43 (C6219; Sigma, St. Louis, MO, USA). 4.2.3 Cell culture All cells were grown in a humidified 37◦C incubator with 5% CO2. NIH- 3T3 fibroblasts and HEK-293 cells were purchased from American Type Culture Collection (Manassas, VA, USA), and maintained in DMEM + 10% FBS. MC3T3-E1 (clone 4) pre-osteoblast cells were also purchased from American Type Culture Collection, and maintained in α-MEM + 10% FBS for no more than 15 passages. To induce differentiation, the MC3T3-E1 cells were seeded at 100,000 cells/cm2 and supplemented with β-glycerol phosphate (10 mM) and ascorbate (50 μg/ml) for between three and eight weeks as indicated. Primary osteoblasts were liberated from rat or mouse calvaria using sequential collagenase digestion as previously described [315], 84 seeded at 50,000 cells/cm2 and maintained in α-MEM + 10% FBS until confluent. Differentiation was once again induced with β-glycerol phosphate (10 mM) and ascorbate (50 μg/ml). To confirm differentiation, alkaline phosphatase activity was detected with 0.1mg/ml Naphthol and 0.6mg/ml Fast Red Violet, and mineralization was visualized with Von Kossa staining (2.5% silver nitrate under bright light). 4.2.4 Immunofluorescence At least three CD-1 mouse embryos each, from 13, 14, 15, and 17 days post conception (E13, E14, E15, and E17) were fixed whole for 24-48 hours in 4% paraformaldehyde (PFA) at 4◦C. Metatarsals were harvested from pups 2 days postnatal (P2), and fixed/decalcified in phosphate buffered 12.5% EDTA and 2.5% PFA (pH 7.4) for 48 hours. Older animals (P7–12 months) were first perfused with 4% PFA, and harvested tissues were further fixed in 4% PFA overnight at 4◦C. Fixed tissue was dehydrated sequentially through 15% and 30% sucrose for cryoprotection, then embedded in Tissue-Tek O.C.T. compound (Sakura Finetek, Torrance, CA, USA), and sectioned at 6-8 μm. Metatarsal sections were pre-treated with 0.5% hyaluronidase (Sigma) in Hank’s Balanced Salt Solution (with Mg2+ and Ca2+) for 30 minutes, to free collagen epitopes from extracellular proteoglycans [316]. Sections were blocked with 7.5% BSA fraction V (Invitrogen, Carlsbad, CA, USA) and 10 mM glycine, then permeabilized with 0.3% Triton X-100 be- fore incubation overnight at 4◦C in primary antibody. AlexaFluor 488 or 568 secondary antibodies (Invitrogen) were applied for 1 hour at room tem- perature, followed by mounting in ProLong Gold antifade reagent with DAPI (Invitrogen). Imaging was performed on an Axioplan2 fluorescence microscope fitted with an AxioCam MRm camera (Carl Zeiss, Thornwood, NY, USA). Sequential images of individual sections were aligned using the auto-merge algorithm in Adobe Photoshop CS4 (Adobe Systems, San Jose, CA, USA), and manually adjusted to correct misalignments. Visual inspec- tion of the reconstructed sections was performed, and histological structures were identified by reference to The Atlas of Mouse Development, by M.H. Kaufman [317]. 4.2.5 Microarray RNA was harvested from fresh rat post-natal calvaria (P1) or from cultures derived from P1 rat calvaria. For the cultures, RNA was harvested at con- fluence (D0) or after 9 days of differentiation (D9). Control cultures were 85 also treated with 100 nM all-trans retinoic acid to inhibit differentiation (D9-RA). RNA was collected using Trizol Reagent (Invitrogen) followed by purification through RNAeasy columns (Qiagen, Valencia, CA, USA). RNA was labelled and hybridized to RAT 230 2.0 arrays (Affymetrix, Santa Clara, CA, USA) using the manufacturers recommended protocol, and gene expression profiles were analyzed using MAS 5.0 and GeneSpring GX. 4.2.6 Western blot MC3T3-E1 cell lysates were collected with radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM Tris-HCl pH 8.0, 0.5% Sarkosyl, 1% IGEPAL, 0.1% SDS). Protein concentration was determined using a bicin- choninic acid (BCA) assay kit (Pierce), and 30-50 μg was separated on 10% Tris-glycine SDS-PAGE gels. The protein was transferred to Immuno-Blot PVDF (Bio-Rad, Hercules, CA, USA) and then blocked in 5% non-fat milk + 0.1% Tween20 (NFM-T). The membranes were probed at 4◦C overnight, with primary antibodies diluted in 3% NFM-T, followed by horseradish per- oxidase (HRPO) linked (Sigma) secondary antibodies for 2 hours at room temperature. HRPO activity was visualized by treating the membrane with SuperSignal West Pico Chemiluminescent Substrate (Pierce), and expos- ing/developing Bioflex Econo Film (Clonex, Markham, Ontario, Canada). 4.2.7 Promoter alignment The genomic sequences of the first Panx3 exon, as well as one kilobase 5’ from the start codon, were obtained for all mammalian species with an entry in the Entrez gene database, including mouse, rat, dog, cow, horse, macaca, human, chimp, and opossum. Multiple sequence alignment was performed with ClustalX2.0.3 [318]. MatInspector (Genomatix Software, Munich, Germany) was used to predict putative transcription factor bind- ing motifs within the mouse Panx3 promoter [319], and these results were correlated against conserved regions identified by the ClustalX algorithm. 4.2.8 Plasmids Three fragments from the Panx3 promoter were directionally cloned into the BlgII and BamHI sites of the promoterless pGL4.11 [luc2P] luciferase reporter plasmid (Promega, Madison, WI, USA). These fragments were PCR amplified from mouse genomic DNA, using a reverse primer with a BlgII site designed to cut at +1 (5’ - ATAGATCTGATCAGGCTCTGT- GCC - 3’), and forward primers with BamHI sites designed to cut at -206 86 (5’ - GCGGATCCCACGGCCCTCCT - 3’), -648 (5’ - CCGGATCCAGTG- GTTTCCTCA - 3’), and -1175 (5’ - ATGGATCCCACTGACTCCTTGC - 3’). Two alternate constructs were prepared from the 648bps construct using site-directed mutagenesis, to incorporate nucleotide substitutions into the predicted Runx2 and Msx1/2 binding motifs, from -283 to -275 (AAAC- CACAA to CGTTTAGCT) and from -33 to -28 (TAATTG to GGATCC), respectively. A Runx2 expression vector, pCMV-OSF2, was kindly provided by Gerard Karsenty [320]. A second Runx2 expression vector, pSG5-EGFP- Runx2Δrep, was engineered in house: EGFP was excised from pEGFP-N3 (Clontech, Madison, WI, USA) using BglII and NotI, blunt ended, and lig- ated into pSG5 which had been linearized with BamHI and BglII and also blunt ended. Runx2 was then removed from pCMV-OSF2 with EcoRI (trun- cating the protein by 13 amino acids, starting at residue 583), and ligated into the EcoRI site of pSG5-EGFP. The rat Panx3 cDNA was amplified from a cDNA library using primers with ClaI sites added to the 5’ end, then cloned into the ClaI site of pRK5 for use as a positive control of Panx3 expression in real-time PCR experiments, while pEF-GM-EGFP [321] was used as a negative control. Finally, pCMV-SPORT6-Msx1 was purchased from Open Biosystems. All new constructs were verified by sequencing the relevant regions. 4.2.9 Transcription reporter assay HEK-293 cells were co-transfected with the pGL4.11 [luc2P] plasmids car- rying Panx3 promoter constructs, a constitutively driven transcription fac- tor or control (Runx2, Runx2-Δrep, Msx1, or GFP), and a constitutively driven Renilla luciferase plasmid pGL4.73 [hRluc/SV40] (Promega) to mon- itor transfection efficiency. Primary mouse osteoblasts were co-transfected with only the pGL4.11 [luc2P] promoter plasmids and control pGL4.73 [hRluc/SV40] plasmid. Transfection was accomplished using Fugene6 reagent, as directed by the manufacturer (Roche, Indianapolis, IN, USA). After trans- fection, the cells were maintained in culture for 48 hours, and then total cellular protein was extracted with passive lysis buffer (Promega). Dual- luciferase reporter assays were performed according to the manufacturer’s protocol (Promega), and Luciferase activity was quantified using a 96-well, duel-Luciferase luminometer (Berthold Detection Systems, Pforzheim, Ger- many). Firefly luciferase activity was standardized against Renilla luciferase activity, and relative luciferase units from the Panx3 promoter constructs were normalized to the empty pGL4.11 [luc2P] group (which was arbitrar- ily assigned a value of 100). Each sample was analyzed in triplicate and 87 the values averaged, with error bars representing the standard error of at least three separate experiments. Normalized sample means were compared against one another using two-way ANOVA with subsequent Holm-Sidak pairwise multiple comparison. P-values ≤ 0.05 were considered significant. 4.2.10 Real-time qPCR Total RNA was harvested from cell culture using Trizol Reagent (Invit- rogen) according to the manufacturer’s directions, and 550ng was reverse transcribed into cDNA using SuperScript III (Invitrogen). 10 ng of cDNA was diluted in 2X TaqMan Fast Universal PCR MasterMix (Applied Biosys- tems, Foster City, CA, USA) and TaqMan probes against Panx3, Runx2, or rRNA (Applied Biosystems). Real-time quantitative PCR was performed on the 7500 Fast thermal cycler (Applied Biosystems), using the comparative CT experimental settings of the 7500 V2.0.1 software. Panx3 and Runx2 levels were normalized against rRNA. 4.2.11 Chromatin immunoprecipitation Chromatin was isolated from MC3T3-E1 cells differentiated for four weeks, then queried with anti-Runx2, or non-specific rabbit IgG control as previ- ously described [322]. Briefly, the cells were cross-linked in 1% formalde- hyde at room temperature for 15 minutes, followed by mechanical lysis in a Dounce homogenizer (critical for adequate recovery of MC3T3 cells, which are embedded in dense collagen at the time of collection), and shearing of the chromatin into fragments of 500-1000 bps by sonication. The chromatin was pre-cleared with protein A Sepharose (Pierce) for one hour, followed by incubation with primary antibodies overnight. The immune complexes were subsequently precipitated with protein A Sepharose and reverse cross-linked overnight. The DNA was phenol/chloroform extracted, and then analyzed by PCR. PCR primers were designed to span the putative Runx2 binding site in the Panx3 promoter (Fwd: 5’ - ATCAAATACAGGGCAGTTTCAGGG - ’3, Rev: 5’ - CACTGTGCCTTTATGCTGTCC - ’3). 88 4.3 Results 4.3.1 Panx3 is expressed in both endochondral and intramembranous bone during embryonic development To visualize the global expression pattern of Panx3 during development, E13, E14, E15, and E17 mouse embryos were collected, fixed, sectioned, and surveyed by immunofluorescence using a Panx3 specific antibody [104]. Between E13 and E15.5, Panx3 expression predominantly localizes to mem- branous bones of the face and upper thorax (Fig. 4.1). The mandible and clavicle are two of the first primordia to mineralize during development, forming ossification centers as early as E14 [317], although Panx3 expres- sion appears as early as E13-13.5 (Fig. 4.1). By E14-14.5, the mesenchymal condensations destined to become the parietal and frontal bones of the lat- eral calvaria and the maxilla also expressed Panx3 (Fig. 4.1). All of these structures are derived through intramembranous ossification, whereby con- densed mesenchymal cells differentiate into osteoblasts to fabricate bone directly [323]. Development of the appendicular skeleton and parts of the axial skeleton involves a transition from cartilage to mineralized bone ma- trix. Longitudinal growth of this tissue is achieved through chondrocyte division, followed by hypertrophy of cells situated closest to the diaphysis [324]. Inspection of cartilaginous centres of endochondral ossification (such as the dorsal ends of the ribs and diaphyses of major long bones) reveals Panx3 expression as early as E14-14.5 (Fig. 4.1). After E15, chondrogenic Panx3 expression is easily identified at the interface between terminally dif- ferentiated chondrocytes and the mineralized matrix during endochondral ossification (Fig. 4.1 and 4.2), particularly in appendicular long bones, while continuing to be observed in intramembranous bones as well (Fig. 4.1 and 4.2). The specificity of the Panx3 antibody was confirmed by peptide com- petition (Fig. 4.3). 4.3.2 Panx3 is expressed by pre-hypertrophic chondrocytes, hypertrophic chondrocytes, and mature osteoblasts To better characterize the distribution of Panx3 within developing long bones, further immunofluorescent analyses were performed on metatarsals from P2 mice using antibodies against known markers of chondrocyte differ- entiation (Fig. 4.4). Collagen type X alpha 1 (Col10α1) is a classical marker of terminally differentiated, hypertrophic chondrocytes in the growth plate 89 E15-15.5 1 mm 1 mm 1 mm E13-13.5 1mm B A Q P O L N M G K I J H 1mm F E E14-14.5 1mm D C 50μm 50μm 100μm 100μm 100μm 100μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm Figure 4.1: Panx3 (green) was first observed in large membranous bones, such as the mandible (A) and clavicle (B), as early as E13-13.5. By E14- 14.5, a large number of membranous bones were positive for Panx3, as il- lustrated by the frontal bone (C), mandible (D), and lateral calvaria (E). Panx3 was also found in the diaphyseal regions of long bones at E14-14.5, as demonstrated in the tibia (F). At E15-15.5, Panx3 expression was ro- bust in the osteoblasts of maturing membranous bones, as detected in the orbito-sphenoid bone (G), nasal capsule (I), exoccipital bone (J), mandible (K and M), and maxilla (N). Regions within cartilaginous bones also ex- hibited strong Panx3 expression, such as the neural arch of the C1 vertebra (H), articular and transverse processes of vertebra (L and O), rib tubercle (O), rib shaft (P), and radius diaphysis (Q). Nuclei are stained blue with DAPI 90 E17-E17.5 1mm B C A 100μm 100μm 100μm Figure 4.2: Panx3 (green) expression at E17-17.5 was extensive through- out the developing skeleton, as demonstrated in the basisphenoid bone (A), mandible (B), and vertebra (C). Nuclei are stained blue with DAPI 91 [325, 326]. Panx3 is observed before Col10α1 induction in the developing metatarsal cartilage, indicating that Panx3 expression precedes terminal chondrocyte differentiation (Fig. 4.4). This localization pattern supports recently published in situ data, where a similar relationship between Panx3 and Col10α1 transcripts was reported in the growth plate [169]. Osteoblasts and hypertrophic chondrocytes both express and secrete the glycosylated phosphoprotein osteopontin (Spp1) during long bone formation [327, 328]. Induction of Spp1 expression in the cartilage anlagen closely mimics that of Col10α1, and is also preceded by Panx3 expression. Within the perichon- drium, the inner chondrogenic layer of cells at the periphery of the develop- ing bone collar also express Panx3. Spp1 expression in the perichondrium overlaps that of Panx3 to a limited extent, but Panx3 appears to be reduced at the onset of Spp1 expression in the inner chondrogenic layer. Interest- ingly, Panx3 is absent from the outer fibrous layer of the perichondrium, despite an increase in Spp1 (Fig. 4.4). As expected, the metatarsal model also illustrates Panx3 expression by osteoblasts within the mineralizing di- aphysis. Collagen type I alpha 1 (Col1α1) is secreted by endochondral and perichondrial osteoblasts [329], and Panx3 expressing cells within the miner- alizing diaphysis are surrounded by layers of Col1α1 containing extracellular matrix 4.4. While the spatial relationship between chondrocytes in the growth plate (i.e., increasing maturation from the epiphysis to the diaphysis) simplifies the process of determining when in the chondrogenic life cycle Panx3 is induced, an analogous structure does not exist for osteoblast maturation. To better characterize the temporal profile of Panx3 expression during osteoblast dif- ferentiation, mouse calvarial pre-osteoblast cells (MC3T3-E1) were cultured in differentiation inducing conditions over an 8 week period. A robust in- crease in Panx3 was observed over the first 30 days of differentiation, fol- lowed by a plateau as the cultures were maintained for up to 60 days (Fig. 4.5). This pattern is inversely correlated with parathyroid hormone 1 re- ceptor (Pth1r) expression, and positively correlated with up-regulation of Col1α1 (Fig. 4.5). Pth1r is involved in regulating cell growth and differen- tiation, with expression peaking early in MC3T3-E1 differentiation prior to osteoid secretion [330, 331], while Col1α1 expression is consistent with the expression profile we observed in the metatarsals. This increase in Panx3 protein in MC3T3-E1 cells is mirrored at the mRNA level, as determined by real-time polymerase chain reaction (PCR) (data not shown). This ex- periment was performed twice with similar results, and differentiation was confirmed by mineralization of the cultures (Fig. 4.5). Figure 4.5C shows that the expression pattern of Panx3 in primary rat calvarial cultures, as 92 No 1° antibody α-Panx3 Pre-immune α-Panx3 + Panx3 ppt α-Panx3 + Panx1 ppt 500 μm 500 μm500 μm 500 μm 500 μm Figure 4.3: Sections of neonatal mouse metatarsal were probed with the anti- Panx3 antibody, the Panx3 antibody pre-absorbed against a 10-fold molar excess of cognate Panx3 blocking peptide, the Panx3 antibody pre-absorbed against a blocking peptide designed for an unrelated Panx1 antibody, rab- bit serum taken pre-immunization, and with no primary antibody. The dramatic reduction in immunofluorescence caused by the Panx3 blocking peptide appears to be specific, since the same molar ratio of Panx1 peptide had no effect 93 Panx3 - OPN 200 μm B Panx3 - Col10α1 200 μm A Panx3 - Col1α1 200 μm C 100 μm 50 μm 25 μm Figure 4.4: Endochondral ossification in the metatarsal of neonatal mouse pups is representative of long bone development. Panx3 expression precedes induction of chondrocyte hypertrophy, as delineated by Col10α1 expression (A). Osteoblasts, hypertrophic chondrocytes, and perichondrial fibroblasts all express Spp1 (B). While co-localization of Spp1 with Panx3 appears to occur in the proximal layer of chondrocytes in the bone collar, the outer fibrous layers were negative for Panx3. Col10α1 is present in the ECM layer surrounding Panx3 positive osteoblasts that have infiltrated the diaphysis (C, arrow) 94 determined by mRNA microarray, follows the same trend as was observed in MC3T3-E1 cells, and is very similar to established osteogenic markers (liver/bone/kidney alkaline phosphatase (Alpl), integrin binding sialopro- tein (Ibsp), and osterix (Sp7)). There is abundant expression in freshly cultured calvaria (F-RC), followed by a reduction in transcript levels as mi- totically active progenitor cells quickly dominate the sub-confluent primary cultures established from these isolates (D0). Panx3 expression, along with that of Alpl, Ibsp, and Sp7, is elevated once the cultures become confluent and differentiation proceeds (Day 9, D9). Retinoic acid (RA) signalling plays two distinct roles in osteoblast development, first acting to inhibit differenti- ation of immature progenitor cells, and then later enhancing mineralization during osteogenesis [153]. Consistent with this, treatment of primary rat calvarial cultures with 100 nM RA for 9 days (D9-RA) leads to markedly decreased expression of Panx3, along with the other osteogenic markers (Fig. 4.5). Treating MC3T3-E1 cells with RA produced a similar result (data not shown). These experiments indicate that Panx3 is expressed by mature osteoblasts. 4.3.3 The Panx3 promoter is responsive to Runx2 Multiple sequence alignment of the first 1000 basepairs (bps) immediately 5’ from the Panx3 start codon reveals conserved binding sites for transcription factors associated with bone development, such as Barx2 [332], the vitamin D3 receptor - retinoid-X receptor heterodimer (VDR/RXR) [333], Msx1/2 [334], and Runx2 [335] (Fig. 4.6). Three progressively longer stretches of the Panx3 promoter (206 bps, 648 bps, and 1175 bps), were cloned into the pGL4.11 [luc2P] luciferase reporter plasmid (Fig. 4.7). These plasmids were co-expressed in human embryonic kidney (HEK-293) cells with either full length Runx2, or with Runx2-ΔRep (truncated by 13 amino acids on the C-terminal end to delete a repressive domain [336]). Compared to empty pGL4.11 [luc2P] and the 206 bps truncation construct, there was a 2-3 fold increase in luciferase expression generated by the 648 bps and 1175 bps con- structs when co-transfected with either of the Runx2 plasmids (Fig. 4.8). The longer truncations both contain a highly conserved Runx2 binding motif (AACCACA [337]) 275 bps from the transcriptional start site. To test the significance of the putative Runx2 binding site more directly, its sequence was disrupted in the 648 bps construct. This reduced the Runx2 induced luciferase expression to the same levels recorded from the empty pGL4.11 [luc2P] control vector. A binding motif for the homeobox proteins Msx1 & 2 ((C/G)TAATTG [338]) is also highly conserved in the Panx3 promoter. 95 Pthr1 Panx3 GAPDH Days Days 10 105 15 20 25 30 35 40 20 30 40 50 60A B Col1α1 C Re la tiv e m RN A le ve ls 0 2 4 6 8 10 12 14 16 F -RC D0 D9 D9-RA  Calvarial Isolate Ibsp Sp7 Alpl Panx3 Figure 4.5: (A) Panx3 levels progressively increase over several weeks of MC3T3-E1 differentiation, reaching a maximum at 1 month. There was a positive correlation between the expression of Panx3 and the mature os- teoblast marker Col1α1, and an inverse correlation between Panx3 and Pth1r expression. (B) Differentiation was confirmed by measuring alka- line phosphatase activity (red stain), and mineralization (von Kossa; black). (C) CDNA microarray data shows increased expression of Panx3 in freshly isolated calvarial osteoblasts (F-RC) compared to primary cultures derived from these cells (D0). This is the same pattern observed for classical mark- ers of osteogenesis (Ibsp, Sp7, and Alpl). When the osteoblast cultures are induced to differentiate for 9 days (D9), Panx3, Ibsp, Sp7, and Alpl are all up-regulated, unless they are treated with RA (D9-RA), which inhibits the transition of osteoblast progenitor cells into mature osteoblasts 96 Msx1 is important for normal skeletal patterning and growth, especially in relation to osteoblast activity [339], and can function as both a transcrip- tional repressor [340] or an activator [341]. Luciferase expression appeared elevated upon co-expression of the promoter constructs with Msx1. However, this level was unchanged when the Msx1/2 binding sequence was disrupted, implying that Msx1 on its own does not significantly enhance expression by binding to this particular location. As expected, Runx2 continued to induce Luciferase expression despite disruption of the Msx1/2 binding motif. To ensure that the increased luciferase induction was not an artifact of Runx2 over-expression, primary calvarial osteoblasts were also transfected with the promoter truncation constructs after six days of differentiation (Fig. 4.9). Even without exogenous Runx2, Luciferase expression was enhanced 18 and 20 fold over the promoterless pGL4.11 [luc] control by the 648 bps and 1175 bps constructs. To show a direct interaction between Runx2 and the Panx3 promoter in differentiated MC3T3-E1 cells, a ChIP assay was used. Follow- ing pull down and reverse crosslinking, the DNA was analyzed by PCR using primers spanning the Runx2 binding site in the Panx3 promoter. The pro- moter fragment was highly enriched by incubation with the Runx2 antibody compared to IgG control (Fig. 4.10). To test whether Runx2 is sufficient to induce Panx3 expression, the NIH-3T3 cell line (which does not express appreciable levels of Panx3) was transfected with the Runx2 and Runx2- ΔRep plasmids. Real-time quantitative real-time PCR (qPCR) was used to monitor Runx2 and Panx3 expression, and while high levels of Panx3 mRNA were observed in cells transfected with pRK5-Panx3, there was no measurable increase in Panx3 in the Runx2 transfected cells when compared to controls transfected with green fluorescent protein (GFP) (Fig. 4.11). While the distribution of Runx2 is limited, it is not exclusive to os- teogenic cell types. Sebaceous glands in the skin and lactating mammary gland epithelium also express Runx2, which in turn induces the expres- sion of molecules normally found in developing bone, such as Spp1 and Ihh [342, 343]. Both lactating mammary gland of post-partum females and se- baceous glands in the tail sections of P7 animals tested positive for Panx3 expression (Fig. 4.12A-B). During our global immunofluorescence surveys, we also unexpectedly observed Panx3 expression in the small intestinal ep- ithelium, as shown in a section from a P2 animal (Fig. 4.12C). 97                *   *                     *                              * *                                          * * * * *  * *  * * * * * * * *  * *  * *  * * * * *  * * * * *  * *  * *  * * *  * * * * mouse TGCCTGAAGCTGTCACTCCA- - - - - - - CTGTACCTGGATCGCTTAGGGTAGCATTTTCCTCTGTATCC- - - - CCCCAA- GTCCCCATTCTCAGCAGCATCATGTCGCTCGCACACACTGCTGCAGAGTACATGCTCTCTGATGCCCTGCT rat TGTCACA- - - - - TCACCCCA- - - - - - - CTGTACCTAGATTGCTAAGAGTAGCAGTGTCCTCTGTATTC- - - - CCCCAAAGCCCCCATACTCAGCAGCATCATGTCACTCGCACACACAGCGGCAGAGTACATGCTCTCCGATGCCCTGCT human TGCCTGAAGCTGCCGTCTCC- - - - - - TC- - - - - - - - - - - ATTCCACCATCCCAGGACCCCTGCTGCCA- - - - CCTCTGCACCCCCAAGCTCAGCAGCATCATGTCACTTGCACACACAGCTGCAGAGTACATGCTCTCAGATGCCCTGCT macaca TGCCTGAAGCTGCCATCTCC- - - - - - TC- - - - - - - - - - - ATTCCACCATCCCAGGACCCCTCCTGCCA- - - - CCTCGGCACTCCCAAGCTCAGCAGCATCATGTCACTCGCACACACAGCTGCAGAGTACATGCTCTCGGATGCCCTGCT horse TGCCTGAAGCTGCCATCTCC- - - - - - TC- - - - - - - - - - - ACTCTACCATCCTGGAACCCCTCCTGCTG- - - - CCTCTGCA- - - - - - - - - - - - - - - - - - TAATGTCGCTTGCACACACAGCTGCAGAGTACATGCTCTCAGATGCCCTGCT dog TGCCTGAAGCTGCCATCTCC- - - ATCTCCCCCCAC- - - - ACACCACTGTGCGGGGCCCCTTCCTGCCG- - - - CCTCTGCCCCCCGAAGGTCCGCGGCACCATGTCGCTCGCACACACAGCCGCGGAGTACATGCTGTCCGATGCCCTGCT cow TCCCTGAAGCTGCCACCTACTGGGACTCCACTCTCCGGGACCCCACTCTCCAGGACCCCCGTCTGCCA- - - - GCCCTGGGCCC- - AAGCTCAACAGCAGCATGTCGCTCGCACACACAGCTGCAGAGTACATGCTCTCAGATGCCTTGCT chimp TGCCTGAAGCTGCTGTCTTC- - - - - CTC- - - - - - - - - - - ATTCCACCATCCCAGGACCCCTGCTGCCAGCCACCTCTGCACCCCCAAGCTCAGCAGCATCATGTCACTTGCACACACAGCTGCAGAGTACATGCTCTCAGATGCCCTGCT Opossum AGCCTGAAGCTGCTCTCTGTA- - - - - CCT- - - - - - - - - TCTTCTTATACCTAGGTTTCCTTCCCACGTC- - - TTTTCAGGCCCCCACATTCAAGAGCATCATGTCCCTGGCACACACTGCGGCAGAGTATATGCTCTCAGACGCCCTGCT               *  * *         * *        * * * * *                   *  *     *             * *    * *      *       * *        * * * *    * *                  *     * * *       *  * *             * mouse - GAAA- - - - - - - - TATACTG- TTGAAA- - - - - - - - - - - - - - GTCTAGAATCGGGGAAGGGAGATTAATATCAAATACAGGGCAGTT- - - - TCAGGGAAGAAGTCTGG- ACCTTAAGCAAAACACATCACTAAGACAAAGCTTGCTCTATA rat - GAAAATTGTAAGTATAGTG- TTGAAA- - - - - - - - - - - - - - GACTAGAGTCGGG- AAGGAAAATTAACATCAA- - - CAGGGCAGTT- - - - TCAGGGAAGAAGC- - - - - - - - - - - - - - - AAACACATCACTGGAACAAAGCTTGCTCTAGA human - GGAAATTGTAAATATGTAG- TTGAAACATTTTGGAAAAAAGACCATAGTCAGGGAAGGAAATCTAACACCAAATACAGAAAAGTT- - - - - CTGGGAAAGAGTCTGG- GCCCAGAGCAAAAATCATCACTGGGACAAAGACTACTATGTA macaca - GGAAATTGTAAATATGTAG- TTGAAACATTCTGGAAAAAAGACCAGAGTCAGGGAAGGAAATCTAACACCAAACACAGAAAAGTT- - - - TCTGGGAAAGAGTCTGG- GCCCAGAGCAAAAACCATCACTGGGACAAAGACTACTGTGTA horse TGGAAATTGTAAATATGTGGGTTGAAACATTTTGGGAAAA- GATGAGAGGCAGGGAAGGCAATTTAACACCAAATTCAAAGCAGTT- - - - TCTGGGAAACAGTAGGG- ACCCAAAGCAAAG- TCATCCCTGGGACAAAGAATACCATATA dog TGGAAATTGTAAATACGTGG- CTGAAACATTGTGGGAGAA- GACCAGAGGCAGGGGAAAAAATTGAACAGCAAATATAG- - - - - TT- - - - TCTGGGAAAGAGCCTGG- GCCCGAGGCAAAAATCATCACAGAGACAAAGAATACTATGGA cow AGGAAATTGTAAATATGTGG- TTGAAACATTTTAGAAACA- GACCAGAAGCAGAGCAGGGGACCTAACGCCAAAAACAGATCAGTT- - - - TCTGGGAAGGAGCCTGG- GCCCGAAGCAGAAATCATCACTGGGACAAGTACTATGTTGGA chimp - GGAAATTGTAAATATGTAG- TTGAAACACTTTGGAAAAAAGCCCATAGTCAGGGGAGGAAATCTAACACCAAATACAGAAAAGTT- - - - - CTGGGAAAGAGTCTGG- GCCCAGAGCAAAAATCATCACTGGGACAAAGACTACTATGTA Opossum TGAAATATTAAAGTAAGAAA- - TGAAAGATACAACAGAAA- AATCAAAGGAAGAAGGATAGTTATAATGGCACATGCAGAATGATTGGGTTATGGGAGGGAGCAAAAAGCCCAGAGCTTAAGTTATCTTAAAGATAAAGCAGGCTAGCTA                 *       *    *       *         * *   * *      * * * *    * *  * * * * * *  *  * *   *    *     * *  * * *  * * * *   *   *   * * *  * * * * * * * * *  * * * * *  *     *           *                    * * mouse - - - GAG- TAACCATCCTTGATGAATGCTGGGGGCTCCGTTCCCACCCCAAACCACAAAATGAAAGCCTTTGGCACATTTAAGCAGTCTCCACCCAAATTCCACACCAATTATATGGGGGG- - - - - - CTGAGCC- - - - - - - - - - - - - - CAC rat - - - GAGGTAACCATCCTCGACGAATGCTGGGGGCTCCGTTCCCACCC- AAGCCACAAAATGAGAGCCTTCGACACATTTGAGCAATCTCCACCCAAATTCCACACCAATTACATGAGGGG- - - - - - CTGAGCC- - - - - - - - - - - - - - CAC human - - - GG- GTAGCAATCCCAGCTGGGTCGTGGGGACTCCATTCCCACTCCAAACCACAAAATGAGTGCCTTCAAAACATTTAAGCAATCCTCACCCAAATTCCACACAAATTACATGGAGGGGGGTC- CTGGCTC- - - - - - - - - - - - - - CAC macaca - - - GG- GTAGCAATCCCAGCTGGGTCGCGGGGGCTCGACTCCCACCCCAAACCACAAAATGAGTGCCTTCAAAACATTTAAGCAATCCTCCTCCAAATTCCACACCAATTACATGGGGGGG- - - C- TTGGTTC- - - - - - - - - - - - - - CAC horse - - - GAAGTAGCAGTC- - - - CTGAGTCCCAGGG- CTCCATTCCCACCTCAAACCACAAAACGAAGGCCTTCAACACATTTAAGCAATCTCCACCCAAATTCCACACCAATTACATGGGGGG- - - - C- CCGGCTG- - - - - - - - - - - - - - CAC dog - - - GC- - - AACAGTCCTAGCTGGGTCTTGGGGGCTCCACTCCCACCCTAA- CCACAAGATGAGGGCCTTCAACACGTTTAAGCAATCTCCACCCAAATTCCACACCAATTACACGGGGGG- - - - C- CTGGCTC- - - - - - - - - - - - - - TAC cow - - - GC- - - CGCAACCCTAGCGGAGTCCTGGGGGCTCCATTCCCACCCCAAACCACAAAATGAGGGCCTTCCACACATTTAAGCAATCTCCACCCAGATTCCACACCAATTACACAGGGGG- - - - C- CTG- CTT- - - - - - - - - - - - - - GAC chimp - - - GG- GTAGCAATCCCAGCTGGGTCGTGGGGACTCCATTCCCACTCCAAACCACAAAATGAGTGCCTTCAAAACATTTAAGCAATCCTCACCCAAATTCCACACAAATTACATGGAGGGGGGTC- CTGGCTC- - - - - - - - - - - - - - CAC Opossum CATGGGATACCCAACTCTAGTGAGGCTTGAGGGTTCCATCTCCACCCTAAACCACAAAATGAGGGTGGTCAACACATTTAAGCAAACTCCATCCAAATTCCACACCAATTATATGGGGAATGGGCATTGGCCCTAAAGCTCAATACCCAC                *  *  * * *  *  *   *  *               * * * * * *     * *  * * *  *  * * *  *  *   * * * * * * * *      *   *   *               *                *  * *                  *   *   * * * *   * * mouse GGCCCTCCTTCTCTCCACCA- - - - CT- - GCTGAAAATACTGTATTAATCAGAGGACAGCATAAAGCCACAGTGCTTCGTCTGTTGGGAAACAATCTGGGC- GAGCAGCCTGGCACCAACCACACGGAAAGCCTGACCCTCCTCCAGTGAG rat GGCCCTCCTTCTCTCCACCA- - - - CT- - GCTGAAAATACTGTATTAATCAGAGGACAGCATAAAGCCACAGTGCTTCGTCTGTTGGGAGACAATCTAGGC- AAGCACCCTGGCACCAGACACAAGTAAAGCCTGACCCTCCTCCAGTAAA human AGCCCTCCTTCTCTCCAC- - - - - - - - - - - - TGAAAATACTTTATCAATGACAGGGCAGCATAAAGCCACAGACCTCCATCTGCTCTAGGACAATCCAGA- - CAGCAGCGTGGCACCAGGCCCAGGTAAGGCCCGACCCTCCTCCAGCAAG macaca AGCCCTCCTTCTCTCCACCG- - - - CCCGGCTGAAAATACTTTATTAATGACAGGGCAGCATAAAGCCACAGAACTCCATCTGCTCTAGGATAATCCAGA- - CAACAGTAGGGCACCAGCCCCAGGTGAGGCCCGATCCTCCTCCAGCAAG horse AGCCCTCCTTCTCTCCACCA- - - - CCCCACTAAAAATACTTTATTAATGAGAGGGCAGTATAAAGCCACGGGACTCCATCTGCTCAGGGACAACCCAGAGACAGCAGCCTGGCACCAGCCCCAGCTGAAGCCTGACCCTCCTCCAGCGAG dog AACCCTCCTTCTCTCCACCA- - - - CCCCACCAAAAATACTTTATTAATGAGAGGACAGCATAAAGCCACAGGACTGCATCTGCTCGGGGGCAATCCAGAGCGAGCAGCCTGGCACCAGCCCCAGCTGACACCCGACCCTCCTCCAGTGAG cow AGCCCTCCTCCTCTCCCCCA- - - - CCTCACCAAAAATACTAGATTAATGACAGGGCAGCTTAAAGCCATAGGACTGCGTCTGCCTTCAA- CAGTCCAGAG- - GGCGGCCTGGCACCAGCCCCAGCCGACACCCGACCCTCCTCCAGTAAC chimp AGCCCTCCTTCTCTCCAC- - - - - - - - - - - - TGAAAATACTTTATCAATGACAGGGCAGCATAAAGCCACAGACCTCCATCTGCTCTAGGACAATCCAGAC- - AGCAGCGTGGCACCAGGCCCAGGTGAGGCCCGACCCTCCTCCAGCAAG Opossum AACTCCCCTTCCCATCCCCAACAGCCCTGCAGAAAATAACTCATTAATTAGAGGCCTGCATAAAGCCATGAGACCCCACCCAGGCCCAGGACAGCCACAG- - AGAGAAAAGACAGTTATTCTGT- - - - - - TCTAACACTCCATCATCAGA              *  *  *              *  *   * * * * * * * * *   *  *              *   *       * * *        * *   * *  * * * * *      *     *      *   *                   *   *   *           *   *        * mouse GACTCACCGAGGTT- - ATGCCTTCCTCTAATTGCCAGGGTGTC- - - - - - - TGGCACAGAGCCTGATCAGATATGTAGAGAACCATTTCTTGGTCCTAGCACAGAGAT- - GTGAGGCTGGTGCTGG- GACACTCTCCGCTGAAAA- CTAGC rat GACCCATCAAGGCT- - ATGCCCTCCTCTAATTGCCAGGGCGTC- - - - - - - TGGCTTAGAGCCTGGGCAATTATGTAGAGAACCACTGCCTGGTACTAGCACAGATCT- - GTGAGGCTGGTGCTGC- GACATAGCTGCCTGA- - - - - - AGC human GACCCGCTAAGGCT- - ATGCCTCCCTCTAATTAGCAGG- G- CC- - - - - - - TGGCTCAGGGCCTGGGCAGTTATTTAGAGAACCATCGCCTGCTCCCAGCAGAGATATCCATAAGGCTGGGGTGGCAGGCACTGTCTGCCCAAAG- TCAGC macaca GACCCGCTAAGGCT- - GTGCCTCCCTCTAATTAGCAGG- G- CC- - - - - - - TGGCTCAAGGCCTGGGCAGTTATTTAGAGAACCATCGCCTGCTCCCAGCATAGATATCCATAAGGCTGGGGCAGCAGGCACTGTCTGCTCAAAG- CCAGC horse GACCCACTAAG- CT- - ACGCCTCCCTCTAATTACCAGG- TTCC- - - - - - - TGGCTCAGGGCCTGGCCAGTTATTTAGAGAACCAGCGCCTGCTCCAAGCACAGACATCTACGAGGTTGGGGCTGCAGGCACTATCTCCTCAAAG- CCAGC dog GACCCACTAAT- CT- - CTGCCTCCCTCTAATTACCAGA- TGCC- - - - - - - TGGCTCAGGGCCTGGCTGGTTATTTAGAGAACCAGGGCCTGCTCCTAGCACAGATCTCTGCAGGGCTGGGGCTGCAGGCGCTGTCCACTCAAC- - CCAGC cow GACCCACTGGG- CG- - ATGCCTCCCTCTAATTACCAGG- TGCC- - - - - - - TGGCTCAGGGCCTAGCCAGTTATTTAGAGAACCAGTGCCCGCTCCTCACACAGATACTCGTGAGGCCCAGGCTGCAGGAACCATCTGCTGAAAG- CCTGC chimp GACCCGCTAAGGCT- - ATGCCTCCCTCTAATTAGCAGG- - GCC- - - - - - - TGGCTCAGGGCCTGGGCAGTTATTTAGAGAACCATCGCCTGCTCCTAGCGTAGATATCCATAAGGCTGGGGTGGCAGGCACTGTCTGCTCAAAG- TCAGC Opossum GGCACACCCAAGCCTCAAGTCTCCCTCTAATTAACTGGGTACTGATCAAGTTCCCTGGGTCCTGACCAGTTACTTACAGAACAGTCTCTTACTCATGGCACATTTC- - - - - - ATCCCTGGGCTGCCGACTTGCTTAGCTCAGAGCCCAGT STAT Runx2 Rbp-Jκ FAST1 S8 BARX2 VDR/RXR MSX1/2 STAT Transcription Start Translation Start Figure 4.6: Multiple sequence alignment of the Panx3 promoter reveals sev- eral highly conserved sequences (as predicted by ClustalX), several of which were also identified as transcription factor recognition sites by the MatIn- spector database. Of special interest are the Runx2, BARX2, VDR/RXR heterodimer, and MSX1/2 recognition sequences (green boxes), since they are all implicated in the regulation of bone development. Grey boxes also represent putative recognition sites for transcription factors involved in bone development, although the sequences are less well conserved 98 206 bps 648 bps 1175 bps ΔMSX ΔRunx2 EmptyLuc Luc Luc Luc Luc Luc TranscriptionPanx3 Promoter Putative Runx2 binding motif Putative Msx1/2 binding motif Figure 4.7: Illustration of the Luciferase reporter constructs created to de- termine the effect of Runx2 and Msx1 on the Panx3 promoter. The putative Runx2 and Msx1/2 binding motifs are both present in the 1175 bps and 648 bps constructs, but the 206 bps construct was sufficiently truncated to elim- inate the Runx2 site. The predicted Runx2 and MSX binding site sequences were each modified in the 648 bps plasmid using site directed mutagenesis, to create ΔRunx2 and ΔMSX respectively 99 050 100 150 200 250 300 350 206 bps 648 bps 1175 bps ΔMSX ΔRunx2 Empty **† † †** * R el at iv e lu m in os ity  U ni ts  Panx3 promoter-luciferase constructs Runx2 Runx2-ΔRep Msx1 GFP HEK-293* 206 bps 648 bps 1175 bps ΔMSX ΔRunx2 EmptyLuc Luc Luc Luc Luc Luc TranscriptionPanx3 Promoter Putative Runx2 binding motif Putative Msx1/2 binding motif Figure 4.8: Two-way ANOVA with Holm-Sidak pairwise multiple compar- ison indicates that over-expression of Runx2 in HEK-293 cells significantly enhanced Luciferase expression from the 648 bps, 1175 bps, and ΔMSX plas- mids compared to the empty control, but not from the 206 bps and ΔRunx2 plasmids (∗ p-value < 0.001). Over-expression of Msx1 enhances Luciferase expression from the 648 bps, ΔMSX, and ΔRunx2 constructs over empty control († p-value ≤ 0.003), but there is no difference between these con- structs, indicating that removal of the putative Msx1/2 binding site has no significant effect 100 Panx3 promoter-luciferase constructs 0 500 1000 1500 2000 2500 3000 206 bps 648 bps 1175 bps ΔMsx ΔRunx2 empty Primary Osteoblasts R el at iv e lu m in os ity  U ni ts  206 bps 648 bps 1175 bps ΔMSX ΔRunx2 EmptyLuc Luc Luc Luc Luc Luc TranscriptionPanx3 Promoter Putative Runx2 binding motif Putative Msx1/2 binding motif Figure 4.9: Panx3 promoter constructs were stimulated when placed in primary osteoblasts without exogenous transcription factor over-expression. The 648 bps, 1175 bps, and Δ-MSX constructs all generated significantly more luciferase expression than each of the 206 bps, Δ-Runx2, and Empty constructs, while no difference existed between the 206 bps, Δ-Runx2, and Empty constructs (One-way ANOVA with Holm-Sidak pairwise multiple comparison, p-value < 0.001) 101 Panx3 Inp ut Ru nx 2 IgG Figure 4.10: The Panx3 promoter was shown to directly associate with Runx2 in differentiated MC3T3 cells using a ChIP assay. PCR amplifica- tion of the Runx2 binding region in the Panx3 promoter resulted in a higher intensity band on a DNA gel when the template was chromatin immunopre- cipitated with a Runx2 antibody, as opposed to IgG control 102 Panx3 Runx2 NIH-3T3 Fibroblasts Over-expression constructs pRK5-Panx3 pCMV5-Runx2 pSG5-Runx2-ΔRep pEF-GM-GFP R el at iv e m R N A le ve ls  10000 1000 100 10 1 Figure 4.11: Despite the pro-expression effects Runx2 imparts on the Panx3 promoter, over-expression of Runx2 is insufficient to induce an up-regulation of Panx3 in fibroblast (as assessed by real-time PCR, two-way ANOVA with Holm-Sidak pairwise multiple comparison, p-value > 0.05). Baseline expression of Panx3 and Runx2 was arbitrarily set at a value of 10, using fibroblasts transfected with GFP as a normalizing control. The sensitivity of the Panx3 PCR reaction was confirmed by including a sample transfected with pRK5-Panx3. All experiments were performed at least three times 103 B200 μm100 μm C 50 μm A 50 μm Figure 4.12: Panx3 is present in tissues known to express members of the Runx family. Immunofluorescence revealed Panx3 expression (green, all panels) in lactating mammary tissue (A), in sebaceous glands (B), and in the small intestinal epithelium (C) 104 4.4 Discussion The Pannexin family has received an ever increasing amount of attention since its discovery ten years ago, as it represents a unique class of channel proteins with what appears to be a diverse role in cellular functionality. The majority of work to date has focused on Panx1 [344], and only recently have more detailed studies begun to explore the properties of Panx2 [112, 174] and Panx3 [96, 105, 125, 169]. A number of reports have now demonstrated that Panx3 is expressed by, and plays a role in the differentiation of os- teogenic cell types [125, 169, 170, 175, 312], so the focus of this study was to specifically identify when expression occurs during development, and to understand some of the regulatory underpinnings of this expression pattern. Using immunofluorescence we were able to assess the spatial and tem- poral distribution of Panx3 in embryonic mice during development. Mor- phological analysis reveals Panx3 expression at sites of both intramembra- nous ossification and endochondral ossification just prior to the initiation of mineralization, beginning as early as E13-E13.5. Given this expression pattern, future studies into the function of Panx3 in vivo should be directed at processes subsequent to mesenchymal condensation and skeletal pattern- ing. In the growth plate, we confirmed that induction of Panx3 expression precedes the onset of chondrocyte mineralization by comparing the local- ization of Panx3 protein against other well characterized markers of bone development, such as Col10α1 and Spp1. Growth plate chondrocytes go through a number of distinct differentiation steps, as they first construct a cartilaginous anlage of the future mature bone, followed by hypertrophy, mineralization, and eventually apoptosis [324]. Hypertrophy is best char- acterized at the molecular level by the onset of Col10α1 expression [345], which we only observe in the large centre-most cells of the diaphysis. The detection of Panx3 in the smaller pre-hypertrophic chondrocytes indicates that expression precedes the initiation of mineralization. Spp1 has a similar distribution as Col10α1 in hypertrophic chondrocytes [346], but is also found in the chondrogenic inner layer of the perichondrium. The cells in this re- gion are strongly influenced by soluble Indian Hedgehog (Ihh) diffusing out from prehypertrophic and early hypertrophic chondrocytes. Ihh stimulates the cells in the lower perichondrial layer to switch from a chondrogenic to an osteoblastic phenotype [347], some of which are retained in the peripheral tissue to deposit layers of mineralizing osteoid in the creation of the bone collar, while others migrate inwards to facilitate trabecular bone forma- tion [348]. Both of these osteoblast populations appear to express Panx3, as illustrated by positive staining in a thin proximal layer of cells at the 105 bone collar, as well as in irregularly shaped Col1α1 secreting cells inside the matrix of the diaphysis. Consistently, cultured osteoblasts express greater levels of Panx3 as they differentiate. MC3T3-E1 differentiation involves a well described transition from fibroblast-like pre-osteoblasts to mature, min- eralizing osteoblasts over several weeks [349]. Panx3 was found to increase 6-7 fold during MC3T3-E1 differentiation, and normal maturation of these cells was confirmed by an increase in Col1α1 expression and concomitant decrease in Pth1r. The same pattern was observed in the transcriptome of primary osteoblast cultures, where the expression of Panx3 mirrored that of classical osteogenic markers (Alpl, Ibsp, and Sp7), both during normal induction of differentiation, and when osteogenesis is inhibited with RA. From a functional perspective, it has recently been reported that Panx3 expression strongly influences osteoblast differentiation, acting as an Akt sensitive calcium channel on the endoplasmic reticulum (ER), as an ATP channel at the plasma membrane, and as a gap junction facilitating calcium waves between cells [125]. Panx3 has also been shown to enhance maturation to the terminally differentiated hypertrophic state of cultured chondrocytes [169]. In this case, it was suggested that Panx3 channel activity reduces the intracellular concentration of ATP, thus reducing the synthesis of cyclic adenosine monophosphate (cAMP) upon activation of Pth1r, and ultimately hindering cAMP-responsive element binding protein (CREB) driven prolif- eration [169]. Although not addressed in the aforementioned report, it is worth stating that parathyroid hormone, cAMP, and CREB are also impor- tant mediators of osteoblast proliferation and differentiation [350]. While further work is still required to determine the relative impact each of these mechanisms has in vivo, during both intramembranous and endochondral ossification, it is clear that an intimate link exists between osteogenesis and the presence of Panx3. In developing bone, separate transcriptional programs drive osteochon- dral progenitor cells to mature into osteoblasts or chondrocytes. To become osteoblasts, transcription factors like Msx2, Dlx5/6, Runx2, and Sp7 are important [351], while Sox5, 6, and 9 will push the progenitor cells towards chondrogenesis [239, 352]. Interestingly, many of the molecular markers of differentiated osteoblasts are also expressed by hypertrophic chondrocytes in the epiphyseal growth plate, where a transcriptional switch takes place to initiate ossification [353]. The promoter of Panx3 was scrutinized for highly conserved protein binding motifs that could possibly link its expres- sion to the osteogenic transcriptional regime. Of the motifs identified as being related to the bone transcriptional program, the highly conserved Runx2 binding site at -275 bps stood out, because Runx2 is an important 106 mediator of both osteoblast differentiation [354] and the mineralization of growth plate chondrocytes [355]. The Panx3 promoter was indeed found to be responsive to Runx2 induction, and subsequent mutagenesis confirmed that the putative motif at -275 bps was responsible. Taken together, these observations strongly support our hypothesis that Panx3 expression is reg- ulated by Runx2 activity. However, Runx2 over-expression did not induce Panx3 expression in fibroblasts, suggesting that transcriptional cofactors may be involved. Indeed, Runx2 has previously been reported to work in conjunction with auxiliary transcription factors like MSX2 (inhibitory) and DLX3/5 (excitatory) [322]. This dependence on other factors could also explain why the Panx3 promoter constructs were nearly 10 times more re- sponsive in primary osteoblasts than they were in Runx2 over-expressing HEK-293 cells. As stated earlier, sebocytes and lactating mammary gland epithelium are both non-mineralizing tissues that are regulated by Runx2 [342, 343]. We have now shown that these tissues also express Panx3, supporting the premise that Panx3 is regulated by Runx2 and further suggesting that the correlation between Panx3 expression and mineralization is not necessarily causative. At the very least, it indicates that Panx3 expression outside the context of osteogenic cell types is insufficient to induce mineralization. The presence of Panx3 in the small intestinal epithelium was not an anticipated result, since it is not a known site of Runx2 expression. It does however ex- press a second member of the Runx family (Runx3) [356], and the functional redundancy between these two transcription factors [357] may partly explain the expression observed. It seems clear that Panx3 does not indiscriminately induce ossification, but it may still facilitate significant movement of ATP and calcium within and between cells in vivo, as the recent in vitro work by Iwamoto et al. [169] and Ishikawa et al. [125] suggests. This could have the pro-osteogenic effect of reducing intracellular cAMP, but the possibility remains of an anti-osteogenic effect via activation of P2Y2 receptors, which have been shown to inhibit maturation of osteoblasts [358]. It is antici- pated that these conflicting scenarios will be better assessed through the generation and characterization of a Panx3 null mouse. In conclusion, this study describes the temporal and spatial expression profile of Panx3 in mice during development, with an emphasis on the in- duction pattern observed in maturing osteoblasts and hypertrophic chondro- cytes. This is also the first study to describe the transcriptional regulation of the Panx3 promoter, which we have found to contain a functional enhancer element specific for the transcription factor Runx2. 107 Chapter 5 In vivo role of Panx3 during endochondral ossification “Nature is nowhere accustomed more openly to display her secret mysteries than in cases where she shows traces of her workings apart from the beaten path.” -William Harvey 5.1 Introduction Endochondral ossification (EO) is the primary process leading to the for- mation of long bones in the axial and appendicular skeleton. Mesenchymal cells condense and differentiate into a cartilaginous anlage, which then ex- pands primarily along the longitudinal axis as the chondrocytes first divide and then undergo hypertrophy upon terminal differentiation, forming the classical ‘growth plate’ structure characteristic of EO derived long bones. The hypertrophic chondrocytes deposit mineralizing matrix before undergo- ing apoptosis, and then following vascularisation, osteoclasts and osteoblasts infiltrate these primary ossification centres to begin remodelling the miner- alized cartilage into primary spongiosa, which precedes mature trabecular bone. In contrast, the flat bones that predominantly comprise the head and face are the product of intramembranous ossification (IO), whereby mes- enchymal condensations differentiate directly into osteoblasts that progres- sively deposit layers of mineralizing matrix onto the surface of the growing bone (for reviews, see [359–362]). Chondrocytes and osteoblasts share a common progenitor, but local cues that stimulate expression of transcrip- tion factors such as Runt-related transcription factor 2 (Runx2) and osterix (Sp7) direct these cells to become osteoblasts, while members of the Sry- related HMG box (Sox) family induce chondrogenesis [239, 351, 352]. As growth plate chondrocytes become terminally differentiated however, their transcriptional program switches to resemble that of osteoblasts [353], and this switch is heavily influenced by Runx2 [355]. In chapter 4 Runx2 was 108 shown to regulate expression of the channel protein pannexin 3 in both os- teoblasts and growth plate chondrocytes [309], while in vitro experiments performed elsewhere have suggested that pannexin 3 mediates maturation in both of these cell types [125, 169]. The pannexins (Panxs) are a small family of Chordate proteins (Panx1, Panx2, and Panx3) homologous to the invertebrate gap junction proteins known as innexins [90]. Three pannexin genes have been described in mam- mals [102], four in teleost fish (chapter 2 [267]), and two in tunicates [363]. Unlike the innexins however, pannexins do not readily form intercellular gap junctions under physiological conditions [128]. Instead, they are understood to form large transmembrane pores that facilitate passage of ions and small molecules (such as Ca2+ and ATP) between the intercellular and extracel- lular spaces [104, 131], in response to positive membrane potentials [102], mechanical stimulation [131, 152], intracellular Ca2+ transients [110, 151], caspase cleavage [158, 159], and extracellular K+ [114, 156]. All of these properties have been ascertained while studying Panx1, and although it may be reasonable to hypothesize that Panx3 will behave similarly given that it shares more than 60% identity with Panx1 at the amino acid level, only a small number of observations have been reported in support of this (such as the ability to pass small tracer molecules or ions in response to mechanical stimulation [104, 125]). Furthermore, while misexpression of Panx3 has been implicated in aberrant differentiation of keratinocytes [105], osteoblasts [125], and chondrocytes [169] in vitro, there have yet to be any reports on how Panx3 over-expression or under-expression influences in vivo physiology. Here, the chicken embryo has been used as a model of embryonic limb development to describe the effects of altered Panx3 expression on EO in vivo, by exploiting the Replication Competent ALV LTR with a Splice ac- ceptor (RCASBP) viral vector to deliver exogenous PANX3 messenger RNA (mRNA) or PANX3 -shRNA into early limb buds. Over-expression was well tolerated, producing no overt morphological phenotype, while knock-down reduced the size of the underlying bones in the affected limb and altered expression of late chondrocyte differentiation markers. 5.2 Materials and methods 5.2.1 Plasmid construction The chicken PANX3 (accession #XM 001231502) coding sequence (CDS) was amplified from stage 24 embryo whole body cDNA, and ligated into the 109 SalI and EcoRI sites of pBluescript-SK+ (Agilent Technologies, Santa Clara, CA). PANX3 was then sub-cloned between the attL1 and attL2 sites of pENTR3C (Invitrogen, Carlsbad, CA) using BamHI and XhoI, and restric- tion-free cloning [294, 309] was used to position PANX3 between the β-actin promoter and IRES-EGFP of the pMES expression plasmid [364]. PANX3 was also transferred from pENTR3C into a Gateway compatible variant of the RCASBP vector [365] using the Gateway LR-clonase system (Invitro- gen). It was necessary to include a canonical Kozak consensus sequence (GCCGCC) in front of the Panx3 start codon to achieve appreciable over- expression from the RCASBP virus. To generate small-hairpin RNA (shRNA) delivery vectors, restriction free cloning was used to modify a pENTR3C-mir-30a plasmid [366]. The leader and trailer sequences immediately flanking the native mir-30a hair- pin (miRbase accession #MI0001204) were originally replaced by Chen et. al. with MluI and NcoI restriction sites to facilitate cloning custom hair- pins into the target region, but this required the synthesis of large 99 bp synthetic duplexes for each desired construct. The same group was later able to reduce the duplex size to 78 bp by utilizing SphI and NgoMIV re- striction sites [367], but a restriction-free approach could use even smaller synthetic oligos. As such, the original leader and trailer sequences were re- turned to the mir-30a cassette, leaving a KpnI restriction site as a spacer sequence in place of the native hairpin. In house software was used to de- sign 19-mer siRNA sequences based on previously described criteria [368] (a basic web-interface is freely available to the public at http://www.naus- lab.com/tools/siRNA prediction.php). The 20 nucleotides (nts) native mir- 30a loop sequence was used except for the first two and last two bases, which were changed from CT-to-TA and GG-to-AT respectively because the CT and GG are expected to internally pair when converted to RNA (effectively increasing stem length with non-specific sequence). Based on these param- eters, a 56-mer and 54-mer restriction-free cloning oligonucleotide can be used to synthesize any custom construct; the forward primer consists of 17 nt leader sequence, 19 nt sense sequence, and 20 nt loop sequence, while the reverse primer consists of 15 nt trailer sequence, 19 nt anti-sense se- quence, and 20 nt anti-sense loop sequence. To incorporate the hairpin into pENTR3C-mir-30a, 25 ng of each primer was first mixed into a standard 20 μl PCR reaction using iProof high fidelity DNA polymerase (Bio-Rad, Hercules, CA), and cycled 5 times to anneal and extend the oligos into full duplexes (denature 8 s at 98◦C, anneal 20 s at 50◦C, extend 10 s at 72◦C). Next, 100 ng of the empty pENTR3C-mir-30a plasmid was added to the same tube, and the reaction was allowed to continue for 16 more 110 cycles with an increased extension time of 90 s. The reaction was treated with DpnI (to digest parental plasmid) and KpnI (to linearize any newly synthesized plasmid which failed to gain an insert) for 2 hours, and chemi- cally competent bacteria were transformed with the un-purified reaction mix. Using super-competent bacteria (>108 CFU/μg pUC18) can be beneficial for this step, but sub-cloning grade (106 CFU/μg pUC18) is usually suffi- cient. It should be noted that alternative loop sequences will probably yield equivalent knockdown [369], allowing further reduction in synthetic oligonu- cleotide length should other investigators choose. Three 19-mer sequences from the PANX3 CDS, denoted PANX3 -14 (5’ - ACACGGCTGCTGAG- TACAT - 3’), PANX3 -347 (5’ - TGGTGGCAGTGCTCATGTA - 3’), and PANX3 -620 (5’ - TCATCTACCTCCTGAGGAA - 3’) for their relative po- sition from the start codon, as well as a control from EGFP (EGFP -416, 5’ - GGCACAAGCTGGAGTACAA - 3’), were identified as promising RNAi candidates, and were cloned into the hairpin region of pENTR3C-mir-30a. 5.2.2 Cell culture DF-1 chicken fibroblasts (American Type Culture Collection, Manassas, VA) were cultured in Delbecco’s modified Eagle media, supplemented with 10% fetal bovine serum, and maintained in a humidified 37◦C incubator with 5% CO2. All transfections were performed using FuGENE-6 reagent (Roche, Indianapolis, IN) according to the manufacturer’s directions. The pMES plasmid does not contain selective antibiotic resistance, so to encourage sta- ble integration of pMES-PANX3 into the DF-1 line the cells were subjected to fluorescence-activated cell sorting (FACS) once per week (every 2–3 pas- sages) for four weeks to enrich for EGFP expression, and retention of Panx3 along with EGFP was confirmed using Western blot. 5.2.3 Western blot Cell lysates were collected with radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM Tris-HCl pH 8.0, 0.5% Sarkosyl, 1% IGEPAL, 0.1% SDS), and after determining protein concentration with a BCA assay kit (Pierce, Rockford, IL), 30-50μl of each were separated on 10% Tris- glycine SDS-PAGE gels. The protein was transferred to Immuno-Blot PVDF (Bio-Rad) and then blocked in 5% non-fat milk + 0.1% Tween20 (NFM-T). The membranes were probed with primary antibodies diluted in 3% NFM-T for 2 hours at room temperature or overnight at 4◦C. After washing, HRP- linked secondary antibodies were visualized by treating the membrane with 111 SuperSignal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL), and exposing/developing Bioflex Econo Film (Clonex, Markham, Ontario, Canada). The antibodies used were: Goat anti-Panx3 (N-20, Santa Cruz Biotechnology, Santa Cruz, CA), mouse anti-γ-tubulin (T6557, Sigma, St. Louis, MO), and mouse anti-viral GAG protein [370] (3C2, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA). 5.2.4 Virus preparation All RCASBP constructs used in this study contained the ‘A’ variant of the envelope protein, and viral particles were generated with a modified version of methods described previously [371]. DF-1 cells were transfected with proviral plasmid DNA in a 12-well plate, and then expanded into two 150 mm plates through three passages over 10 days. The culture media was replaced with 20 ml fresh media when the cells reached confluence. This media was then collected and replaced with another 20 ml after 24 hours, and stored at 4◦C until a second round of collection 24 hours later. All collected media was pooled and centrifuged in a Beckman SW-28 rotor at 25000 RPM for 2.5 hours at 4◦C. Note that cell debris is not pre-filtered from these viral preparation, because a higher final titre is recovered without this step. Pellets were drained, and allowed to slowly resuspend in 100 μl of Opti-MEM (Invitrogen) on ice overnight. Debris was removed from the resuspension by centrifuging for 5 minutes at 3000 g, then 5 μl aliquots were flash frozen in liquid nitrogen and stored at -70◦C. 5.2.5 Infection of developing forelimb Fertilized White Leghorn chicken eggs (University of Alberta) were used for all injection experiments, in accordance with a University of British Columbia approved animal care protocol (#A11-0351). The embryos were grown at 38◦C in a humidified incubator until stage 18–21 [372], then 0.05% Fast Green was added to the concentrated virus, and loaded into a pulled glass pipette. A Picospritzer II (General Valve Corp., Fairfield, NJ) was used to deliver virus directly into the developing forelimb bud, and the embryos were allowed to continue developing until the desired stage. 5.2.6 Real-time qPCR Total RNA was isolated from stage 35 limbs using Trizol Reagent (Invitro- gen) according to the manufacturer’s directions. Relative gene expression between samples was determined using the primers listed in table 5.1 on 112 a StepOnePlusTM quantitative real-time PCR (qPCR) machine (Life Tech- nologies, Grand Island, NY). The final PCR mix included 10 ng of total RNA in 10 μl reactions, using the iScriptTM One-Step RT-PCR kit with SYBR Green (Bio-Rad). Expression levels were standardized against 18S rRNA to generate delta-CT (ΔCT), and final relative expression (delta-delta-CT (ΔΔCT)) was calculated as the ratio between the infected and uninfected limbs. Target Primer sequence Anneal. Size PANX3 F: 5’-TGAAGTTCGTTGCAGTGGG-3’ 55.6C 101 bp R: 5’-AGTTGGAAGGCGAGAAACAG-3’ 55.2C ALPL F: 5’-CACGACACTAAGCCTGCG-3’ 62.4C 132 bp R: 5’-GTCCAGCTCATACACCATATCC-3’ 61.9C BGLAP F: 5’-CAAAGGCAAAGCAGACATGAG-3’ 62.0C 145 bp R: 5’-GGCAACTTCCTTTTCAACTGTG-3’ 62.4C COL10A1 F: 5’-ACCAGAAACAGTCCAGCATC-3’ 62.1C 88 bp R: 5’-CCTTGTTCACCCCTCATCTG-3’ 62.6C COL2A1 F: 5’-GCGTGGTATTGTGGGTCTC-3’ 62.2C 90 bp R: 5’-TCAATGCTTTCCAGGTTCTCC-3’ 62.5C IBSP F: 5’-CGGAGGAGAATGCGGTG-3’ 61.7C 86 bp R: 5’-ACTGCCCTTGTAGCGATG-3’ 61.6C IHH F: 5’-TGGGTCTACTACGAGTCCAAG-3’ 62.3C 118 bp R: 5’-CACCGTTCTCCAGCGTC-3’ 61.9C MMP13 F: 5’-CCTCGAAAACTCAAATGGTCAAAC-3’ 62.6C 150 bp R: 5’-TCGTATTCTGGTGAAGTTAAGGG-3’ 61.9C PANX1 F: 5’-TCCATGATAAACTTCAGGTCCG-3’ 54.4C 145 bp R: 5’-ACAACCCAGAGAACATTCCC-3’ 54.8C PTHR1 F: 5’-ATCCATAACTGAAGCACCTCC-3’ 61.8C 139 bp R: 5’-GAAGATGAGGCTGTGGAGATAG-3’ 61.6C SOX9 F: 5’-CTGGGCAAGCTGTGGAG-3’ 62.1C 114 bp R: 5’-GGTTGGTACTTGTAGTCGGG-3’ 61.6C WNT5A F: 5’-ACATCTGCACAGGGTTCATG-3’ 62.5C 129 bp R: 5’-CAATGGCTTCTCAGTACCTCG-3’ 62.2C WNT5B F: 5’-TGGTGTTCATCGGCGAAG-3’ 62.1C 150 bp R: 5’-CGTAGGTGAAAGCGGTCTC-3’ 61.8C Table 5.1: qPCR primers for genes involved in bone development 113 5.2.7 Whole mount in situ hybridization Embryos between stage 30 and 34 were fixed in 4% paraformaldehyde (PFA) at 4◦C for 1–3 days, and then dehydrated through graded alcohol to 100% methanol. The dehydrated embryos were stored at -20◦C for at least 24 hours before performing whole mount in situ hybridization (WISH) as pre- viously described [373]. The cDNAs for PANX3 and RCASBP pol were used to generate Digoxigenin (DIG)-labelled anti-sense RNA probes. An in situ hybridization robot (Intavis Bioanalytical Instruments, Koeln, Germany) was used to automate all hybridization steps and washes prior to colorimet- ric detection of alkaline phosphatase (ALP)-labelled anti-DIG antibody. 5.2.8 Radioactive in situ hybridization on tissue sections Stage 35 embryos were fixed in 4% PFA at 4◦C for three days, and then pro- cessed into paraffin wax. The tissue was sectioned at 7 μm, and mounted on (3-Aminopropyl)triethoxysilane (TESPA) coated slides that were then baked at 60◦C over night. The sections were queried with 35S-labelled anti- sense RNA probes against PANX3, BMP7, IBSP, LEF1, MMP13, PTCH1, RUNX2, WNT5A, or WNT5B, and left to expose in Ilford K5 emulsion for about four weeks before they were fixed and developed. Imaging was per- formed on an Axioplan2 microscope fitted with an AxioCam MRm camera using dark field illumination (Carl Zeiss, Thornwood, NY, USA). 5.2.9 Histological analysis of mineralized forelimb bones Embryos injected with RCASBP::PANX3 -shRNA virus were allowed to de- velop for 17 days (∼stage 41) before the forelimbs were dissected away from the body and fixed in 4% PFA at room temperature for two days, followed by two more days in 10% sucrose buffered with 0.1 M phosphate, pH 7.4. The tissue was dehydrated through graded ethanol to 100%, and then embedded in Immuno-BedTM (Electron Microscopy Science, Hatfield, PA) as per man- ufacturer’s instructions. The tissue was sectioned at 1 μm, and stained with 1% Alcian blue in 1% acetic acid for 1 hour to visualize cartilage, and 0.1% Sirius red in saturated picric acid for 1 hour to label mineralized bone. Each section was then analysed for average trabecular thickness using a custom implementation of the ImageJ image analysis library (kindly provided by Dr. Robert van ’t Hof, University of Edinburgh). Trabecular Thickness = 2 Bone surface area/Bone volume 114 5.2.10 Cell division assay Bromodeoxyuridine (BrdU) (10 mM) was injected directly into the heart of stage 35 embryos, followed by a 2 hour incubation to allow for incorpora- tion. The forelimbs were then harvested and fixed in 4% PFA over-night at 4, processed into wax, and sectioned at 7 μm. The sections were probed for BrdU incorporation with an anti-BrdU antibody, and imaged on a Leica (Nussloch, Germany) TCS SP5II Basic VIS system, using the special pho- tomultiplier R 9624 with low dark current. Montages of the sections were captured using a 20X objective lens, and each visually identifiable cartilage anlage was manually traced in Adobe Photoshop. The anlages were further masked to only contain areas with BrdU positive nuclei, which were then converted to 8-bit grey scale and uniformly thresholded in imageJ. Finally, the ‘analyse particles’ function was called to count the number of dividing cells, so that the number of dividing cells per unit area could be calculated. 5.2.11 Optical projection tomography Forelimbs from stage 39–41 embryos were fixed directly in 100% ethanol for at least 48 hours, followed by Alcian blue (cartilage) and Alizarin red (min- eralized bone) staining as previously described [374]. The specimens were cleared in 1% KOH for 2-4 days, and then any remaining soft tissue was manually dissected away and the humerus and radius/ulna were separated for independent analysis. The bones were next dehydrated through 50% and 100% methanol, and finally transferred to benzyl alcohol/benzyl benzoate (1:2). Each sample was imaged on an optical projection tomography (OPT) scanner (Bioptonics, Edinburgh, Scotland), and the raw images were recon- structed into a 3D model for volumetric measurement using the NRecon and CTan software (SkyScan, Kontich, Belgium). 5.2.12 Statistical analysis One way analysis of variance (ANOVA) was used to test for differences when comparing between multiple treatment groups following OPT scans, and Holm-Sidak pairwise multiple comparison tests were subsequently used to identify the samples with significant dissimilarity (p ≤ 0.05). All other infected:uninfected ratios (i.e., from qPCR, BrdU labelling, and trabecular thickness) were first log2 transformed, and then normalized between 0 and 2 (arbitrarily setting sample minimum and maximum values that were equidis- tant from 0, and encompassed the entire data set) to generate a distribution that could be analysed with one-sample t-tests, using a test mean of 1.0 (null 115 hypothesis that infection has no effect). Shapiro-Wilk tests and f-tests were used to confirm normality and equal variance respectively (passing when p ≥ 0.05). In cases where normality could not be assumed, one-sample signed rank tests were used, with Yates correction for continuity. 5.3 Results 5.3.1 Panx3 is expressed in intramembranous and endochondral bone in chick To ensure chicken PANX3 expression matches that previously described in mouse [169, 309], transcript levels were assessed in the limbs of stages 30, 32, and 34 embryos using WISH (Fig. 5.1). The first detectable signal was observed at stage 32 in both forelimb and hindlimb, and by stage 34, expression was clearly localized to the diaphyses of the long bones. Closer inspection of sectioned stage 35 animals following radio in situ revealed strong expression in the perichondrium and mature chondrocytes closest to the diaphyses, as well as the intramembranous bones in the face. 116 Stage 35Stage 30 Forelimb Hindlimb Stage 34 Forelimb Hindlimb Stage 32 Forelimb Hindlimb Radius Ulna Humerus Forelimb Head 500 μm 250 μm 100 μm 50 μm 250 μm 250 μm 500 μm Figure 5.1: WISH reveals PANX3 expression in the developing limb between stages 20–34. The first sign of expression appears around stage 32 (arrows), and is robust by stage 34. Radio in situ was used to detect PANX3 expression (green) in the forelimb and face at stage 35, clearly illustrating expression in both endochondral and intramembranous bone 117 5.3.2 In vivo knock-down of PANX3, but not over-expression, alters long bone morphology during development The full coding sequence of chicken PANX3 was cloned into the pMES ex- pression plasmid [364] and stably incorporated into DF-1 cells (which do not express detectable levels of endogenous Panx3, not shown) for biochemical analysis. Immunocytochemistry indicates that the protein localizes to the plasma membrane in this cell type (Fig. 5.2A). Proper trafficking of Panx3 to the cell surface in mouse has previously been shown to depend on post- translational glycosylation [103, 104], and the chicken ortholog contains the same predicted glycosylation site on residue N71 (Fig. 1.3). As expected, treating chicken PANX3 over-expressing cellular lysates with PNGase prior to SDS-PAGE reduced the relative protein size by approximately 3 kDa (Fig. 5.2B). To achieve over-expression in the developing embryo, PANX3 was packaged into RCASBP-A [375], and viral particles were injected into the upper limb-bud of stage 20 chicken embryos. The resultant PANX3 over-expression was confirmed via WISH after the embryos were allowed to develop until stage 35 (Fig. 5.2C). RCASBP delivery of custom shRNA sequences into the developing chicken embryo has recently been validated using the chicken mir-30a micro-RNA gene [366, 367]. As outlined in the materials and methods, the mir-30a vec- tor has been redesigned to facilitate rapid restriction-free cloning into the hairpin region, and three 19-mer sequences were chosen from the PANX3 CDS (PANX3 -14, PANX3 -347, and PANX3 -620). These constructs were subsequently used to transfect PANX3 over-expressing DF-1 cells, where all three reduced PANX3 mRNA levels by approximately 75% (Fig. 5.3A). At the protein level there was more variability, with the greatest knock- down observed after transfection with PANX3 -14 (Fig. 5.3B), therefore RCASBP::PANX3 -14 was selected for all subsequent knockdown exper- iments. A control shRNA construct was also generated against EGFP (EGFP -416), and its activity was confirmed in EGFP over-expressing cells byWestern blot (∼70% knockdown, not shown). The efficacy of RCASBP::PANX3 - 14 knockdown in vivo was assessed using WISH and Western blot, both of which indicate excellent reduction of endogenous PANX3 at stage 35 (Fig. 5.3C and D). 118 50 — + Panx3 PNGase γ-tubulin 37 A C B 20 μm 1 mm anti-Panx3 RCASBP::PANX3 EGFP Ventral view injected uninjected Figure 5.2: A) Panx3 over-expression (red) in cultured DF-1 cells reveals migration to the plasma membrane, with very little co-localization with soluble EGFP. B) Treating lysates from Panx3 over-expressing DF-1 cells with PNG-ase (de-glycosylation enzyme) reduces the molecular weight of the protein. C) WISH confirms that infection with RCASBP::PANX3 results in robust misexpression (dark purple) throughout the injected side of the embryo, with a small amount of transfer to the contralateral side 119 PANX3 γ-tubulin Viral GAG 50 37 75 PANX3-14 Infected Uninfected PANX3 γ-tubulin Viral GAG + #1 #2 #3 +- - -+ Replicate animal PANX3-14 Infection 0 25 50 75 100 125 PANX3-14 PANX3-347 PANX3-620 EGFP-416 Empty PA NX 3-1 4 PA NX 3-6 20 PA NX 3-3 47 EG FP -41 6 Em pt y A C D B R e l a t i v e  P A N X 3  m R N A  l e v e l s RCASBP constructs RCASBP constructs Figure 5.3: A) Real-time qPCR indicates that PANX3 mRNA is reduced by ∼75% in PANX3 over-expressing DF-1 cells when they are infected with any of the indicated RCASBP::PANX3 shRNA constructs (compared to uninfected control cells), while a control shRNA against EGFP and the empty virus have no effect. B) Western blot reveals that there is a similar level of knockdown at the protein level, although RCASBP::PANX3 -14 appears to be the most effective. C) The RCASBP::PANX3 -14 virus is also highly effective in vivo, as WISH reveals reduced endogenous expression in the injected limb relative to the uninjected limb (teal stain represents viral mRNA, dark purple is PANX3 ). D) Similarly, Western blot shows a consistent degree of Panx3 knockdown between the infected and uninfected limb, across replicate animals 120 To determine the gross morphological effects of Panx3 expression on the developing forelimb, RCASBP::PANX3, RCASBP::PANX3 -14, and RCASBP- ::EGFP -416 infected embryos were allowed to develop for fifteen days follow- ing injection (until stage 41). Both the infected and contralateral uninfected limbs were then stained with Alizarin red and scanned with OPT to create a digital 3D representation. The injected forelimb of RCASBP::PANX3 - 14 infected animals were visibly smaller than the paired uninjected limbs, and total bone volume was reduced by 27.4% (18.2%–36.7% at 95% confi- dence interval, Fig. 5.4). No such effect was observed in limbs infected with RCASBP::PANX3 or RCASBP::EGFP -416. RCASBP::PANX3 -14 injected forelimbs were also harvested at stage 41 for histology. Fully mineralized sections were stained for cartilage and bone, and the trabecular network was analysed (figure. 5.5A and C). The average thickness of the trabeculae was not statistically different between the infected and uninfected limb. To test for differences in mitosis within the proliferating chondrocytes, stage 35 embryos were treated with BrdU, and the forelimbs were assessed for incorporation. Again, there was no statistically significant change in the number of BrdU labelled nuclei per unit area between RCASBP::PANX3 -14 infected limbs and their contralateral control (figure. 5.5B and C). 5.3.3 Panx3 knockdown has no effect on the expression of genes involved in chondrocyte proliferation or early differentiation Knockdown of Panx3 has previously been correlated with a reduction in col- lagen type II alpha 1 (Col2α1), aggrecan (Acan), and collagen type X alpha 1 (Col10α1) expression in cultured chondrocyte cell lines [169], while putative function blocking has also been attributed with reduced expression of Sp7, alkaline phosphatase (Alpl), and osteocalcin (Bglap) in metatarsal explants [125]. These studies propose a role for Panx3 in halting mitosis and stim- ulating differentiation, so RCASBP::PANX3 -14 infected stage 35 forelimbs were probed with in situ hybridization for the expression of genes known to modulate chondrocyte proliferation and differentiation during EO (Fig. 5.6, 5.7, and 5.8). PANX3 knockdown did not result in obvious changes in ex- pression or distribution of RUNX2, LEF1, PTCH, BMP7, IBSP, WNT5A, or WNT5B, and only the collagenase MMP13 appeared to be modified, with qualitatively smaller foci in the shRNA infected limbs. This decrease in MMP13 was not confirmed by qPCR however, where the mRNA ratio between the infected verses uninfected limb was not significantly divergent 121 0.40 0.70 1.00 1.30 UninfectedInfected RCASBP::PANX3-14 UninfectedInfected RCASBP::PANX3 UninfectedInfected RCASBP::EGFP-416 A B Humerus Radius/Ulna PANX3-14 n=10 n=10 n=8 n=8 n=8 n=9 Ra o  o f b on e vo lu m es  In fe ct ed :U ni nf ec te d (lo g 2  tr an sf or m ed ) GFP-416 PANX3 PANX3-14 GFP-416 PANX3 * NS * NS 400 μm † † Figure 5.4: A) Representative images of humeri (top) and raduses/ulnas (bottom) taken from the infected and contralateral uninfected limb of the same animal. RCASBP::PANX3 -14 caused a discernible reduction in size of all major forelimb bones. B) Following OPT scanning, the bone volume of each sample was compared against its contralateral control to produce an infected:uninfected ratio. A value of 1.0 would indicate no change, as was the case for RCASBP::EGFP -416 and RCASBP::PANX3 infection, but RCASBP::PANX3 -14 caused a significant reduction in this ratio (one sample t-test, ∗ p ≤ 0.05). The RCASBP::PANX3 -14 values were also significantly different from those of RCASBP::EGFP -416 or RCASBP::PANX3 (one-way ANOVA with Holm-Sidak pairwise multiple comparison, † p-value ≤ 0.05, NS = not significant) 122 Ra o  o f b on e pa ra m et er s In fe ct ed :U ni nf ec te d (lo g 2  tr an sf or m ed ) Uninfected Picrosirius red - bone Alcian blue - cartilage Green BrdU labeling - mitotic cell nuclei Infected Uninfected Infected Uninfected Infected Uninfected Infected B A C Stage 35 Stage 41 0.20 0.60 1.00 1.40 Trabecular Thickness Mitoc chondrocyte # n=7n=3 1 mm 100 μm100 μm 500 μm500 μm 200 μm 200 μm Figure 5.5: A) Fully mineralized stage 41 limb sections show no obvious difference in overall cartilage (blue) and bone (red) architecture following PANX3 knockdown by RCASBP::PANX3 -14. B) Similarly, there does not appear to be a change in BrdU incorporation (green nuclei) in proliferating forelimb cartilage at stage 35, following RCASBP::PANX3 -14 infection. C) Box plot illustrating the infected:uninfected ratios for trabecular thickness and number of BrdU labelled nuclei per unit area. The difference of the sample mean from 1.0 was not statistically significant in either case (one sample t-test, p ≤ 0.05) 123 from a value of 1.0. In fact, expression of only the late hypertrophic marker COL10A1 was significantly altered following PANX3 knockdown (0.595– 0.986 at 95% confidence interval), among a panel of genes that also included ALP, BGLAP, CLO2A1, IBSP, IHH, PANX1, PTHR1, SOX9, WNT5A, and WNT5B (Fig. 5.9). 5.4 Discussion Panx3 has been previously reported in osteoblasts and hypertrophic chon- drocytes [309], and its expression in these cell types is thought to be impor- tant for normal physiology/osteogenesis [125, 169]. To date however, any functional descriptions of Panx3 have resulted from in vitro culture work, so it remains unclear how essential Panx3 is to the normal developmental program of an intact embryo. In lieu of a commercially available Panx3 knockout mouse, the chicken was chosen in this study for its long history as a model organism of skeletal development [376], and because it has become amenable to protracted gene knockdown through viral delivery of shRNA [366, 367, 377]. The mouse and chicken Panx3 proteins share 66% identity and 79% similarity at the amino acid level, and the chicken Panx3 pro- moter contains a Runx2 binding site (AACCACA [337]) at position -632 from the start codon. We have shown previously that Runx2 is an impor- tant factor regulating mouse Panx3 expression in osteogenic cells [309], so it seems reasonable that a similar expression profile may exist in the avian system. Indeed, when visualized with WISH, PANX3 mRNA was visible in the cartilage of primordial chicken diaphyses. Expression of PANX3 was also visible in nascent intramembranous bones of the face, presumably in young osteoblasts, further confirming that regulation of the gene is uniform between the mouse and avian system [309]. Previous reports have indicated that forced expression of Panx3 is growth suppressive [105], and enhances differentiation of osteoblasts and chondro- cytes [125, 169]. As such, it was predicted that misexpression in the devel- oping limb could cause significant dysplasia, resulting from reduced mito- sis within the proliferative zone of developing long bone anlages, and from premature differentiation into hypertrophic chondrocytes. Robust over- expression was achieved by cloning the full length chicken PANX3 cDNA into the RCASBP virus, and the recombinant protein was confirmed to be post translationally glycosylated and trafficked to the plasma membrane as previously reported [107]. Surprisingly, PANX3 over-expressing forelimbs developed normally, with no significant change in bone volume when com- 124 pared to contralateral control. Next it was tested whether the loss of PANX3 effected endochondral bone formation, and this time the outcome was imme- diately visible. The total volume of mineralized bone in forelimbs infected with an RCASBP virus carrying PANX3 specific shRNA was reduced by ∼25%, strongly implying that Panx3 is part of either a proliferation or differentiation signalling pathway. The preliminary histological data here indicates that the proliferation rate of periarticular and columnar chondro- cytes are not correlated with the presence or absence of Panx3, nor are there significant changes in trabecular thickness. Some caution must be exercised though, because the sample variance was to high for the given sample sizes to achieve statistical power ≥ 0.8 (i.e., type II errors are more likely). However, an element of legitimacy can be granted to the hypothesis that proliferation is not correlated with PANX3 expression, because subsequent in situ and qPCR analysis also failed to find evidence to the contrary. For example, the five BMPs expressed in the developing long bone generally promote prolifer- ation, in particular by stimulating IHH expression [378]. BMP7 is the only isotype found in proliferating chondrocytes [379, 380], but its expression was not obviously altered following PANX3 knockdown. In fact, there was no change in the expression of any Indian hedgehog pathway genes tested (IHH, PTCH, and PTHR1 ), and this pathway is one of the primary mech- anisms responsible for suppressing differentiation and maintaining mitotic activity to promote longitudinal outgrowth [240, 241, 381]. Furthermore, SOX9 expression (responsible for regulating much of the chondrocyte phe- notype [238]) remained unchanged, as did expression of the Sox9 target gene COL2A1 [382]. The canonical Wnt pathway has been reported to fa- cilitate the transition from a Sox9 controlled transcriptional program to one controlled by Runx2 [243], thus promoting chondrocyte differentiation and hypertrophy [383]. The expression of LEF1 was assessed as a readout of the canonical Wnt pathway [384], and it remained unchanged, but expres- sion of this gene is mostly restricted to the perichondrium and developing bone collar [385], thus it is difficult to make a claim regarding the canon- ical pathway in growth plate chondrocytes. A better choice to test this pathway in the future would be β-catenin or TCF4 [385]. Only three Wnt proteins are actually expressed in the growth plate: Wnt5A, Wnt5B, and Wnt11 [386]. WNT5A and WNT5B were assessed here because WNT5A is required for longitudinal growth of skeletal elements, and over-expressing WNT5B in cartilage causes significant shorting of endochondral elements [387]. As with LEF1 however, there was no change in expression of either of these genes. Almost all of the primary effectors that lead to regulation of proliferation in long bone chondrocytes eventually converge on the G1 phase 125 BMP7 PANX3 InfectedUninfected PTCH 500 μm 500 μm 500 μm Figure 5.6: Stage 35 forelimb radio in situs, probing for PANX3, PTCH, and BMP7. Silver grains within each section are visualized using brightfield, so RNA signal appears black on a white background 126 IBSP RUNX2 LEF1 InfectedUninfected 500 μm 500 μm 500 μm Figure 5.7: Stage 35 forelimb radio in situs, probing for LEF1, RUNX2, and IBSP. Silver grains within each section are visualized using brightfield, so RNA signal appears black on a white background 127 MMP13 WNT5A InfectedUninfected WNT5B 200 μm 200 μm 200 μm Figure 5.8: Stage 35 forelimb radio in situs, probing for WNT5A, WNT5B, and MMP13 (arrows). Silver grains within each section are visualized using brightfield, so RNA signal appears black on a white background 128 RCASBP::PANX3-14 RCASBP::EGFP-416 Ra o  o f m RN A ex pr es sio n In fe ct ed :U ni nf ec te d (lo g 2  tr an sf or m ed ) n=8 * * 0.50 1.00 1.50 n=3 0.40 1.00 1.60 Ra o  o f m RN A ex pr es sio n In fe ct ed :U ni nf ec te d (lo g 2  tr an sf or m ed ) Figure 5.9: Real-time qPCR for bone genes in stage 35 forelimbs, following injection with RCASBP::PANX3 -14 or RCASBP::EGFP -416. Values repre- sent the ratios between the infected and uninfected forelimbs. Only PANX3 and COL10A1 expression in limbs infected with RCASBP::PANX3 -14 were significantly different from the null hypothesis of 1.0 (one sample t-test, ∗ p ≤ 0.05) 129 regulator cyclin D1 [388]. In future follow up work it would be prudent to include this gene in the analysis, as it should provide a more direct measure of cell division. A number of genes associated with osteoblast activity and/or terminal chondrocyte hypertrophy/mineralization were also assessed for changes in expression following PANX3 knock-down. Given that mouse Panx3 is down- stream of Runx2 [309], it is unsurprising that there was no change in RUNX2 expression in the RCASBP::PANX3 -14 infected forelimbs. Nor were there changes in expression of ALP, IBSP, or BGLAP, all of which are classically recognised Runx2 target genes often used to monitor osteoblast differentia- tion (reviewed in [244]). Of all the genes assessed in this study, only MMP13 and COL10A1 showed a discernible change in expression. Comparison of the in situ data for MMP13 suggests a reduction in the nascent diaphysis following PANX3 knockdown, although the difference was not confirmed by qPCR, and in the end COL10A1 was the only gene besides PANX3 to show a quantifiable, statistically significant change in expression. There are morphological long bone defects associated with abnormal trimerization of Col10a1 fibrils, including a reduction in the size of the hypertrophic zone and less newly formed bone [389, 390]. The effects of complete gene abla- tion during embryogenesis are less clear however, ranging from a completely normal phenotype [391] to an abnormal thinning of the trabecular archi- tecture [392], possibly as a result of penetrance differences between genetic stains. Given this discrepancy, it seems premature to infer causation from the correlation between PANX3 and COL10A1 expression, because there is no way of knowing whether loss of either gene product in isolation would be sufficient to reproduce the phenotype in chicken. Although, now that we know that PANX3 expression does not have much effect on gene expression in chondrocytes prior to the induction of COL10A1, it is even more likely that PANX3 is involved in terminal differentiation and hypertrophy, and not in mitosis. It bears repeating that the statistical power associated with most of the quantifiable data in this study is sufficiently low that one must keep an open mind regarding the possibility of false negatives. To gain more confidence in the results it may be beneficial to emulate the work of Shimei Zhu, who used the chick to describe an EO phenotype of similar magnitude to that seen following PANX3 knockdown [393]. Zhu also used an RCASBP based shRNA delivery system, and found that knockdown of the BMP2-induced WD-repeat containing protein WDR5 causes a shortening of the long bones, likely through changes in RUNX2 expression. They chose to use slightly older animals however, focusing on stage 38 instead of stage 35, which was 130 the stage most commonly used in this current study. By stage 38 there is a visible chondrocyte–bone boundary, and vascularisation of diaphyses has begun. This could make it easier to delineate the various chondrocyte pop- ulations within the growth plate, allowing for a more robust assessment of where proliferation is occurring. Furthermore, because the forelimb bones will be considerably larger, they can be isolated from all surrounding skin, muscle, and connective tissue prior to mRNA isolation. This will hopefully reduce some of the inter-sample variation, thus increasing the statistical power of the final analysis. And finally, if Panx3 really does exert its influ- ence only during or after final differentiation/hypertrophy, there should be a greater impact on relevant gene expression in these older embryos. 5.4.1 Conclusion In this study the chicken embryo has been used to analyse, for the first time, the effects of altered Panx3 expression on endochondral bone formation in vivo. Widespread over-expression of the protein was well tolerated, with- out causing obvious defects in forelimb development, while reduced Panx3 expression had a growth suppressive effect. It is still unclear exactly why reduced Panx3 expression results in shorter forelimbs, but it does not ap- pear to be due to impaired chondrocyte proliferation, nor from changes in the Wnt, Bmp, or Ihh/Pthr1/Pthlh pathways. Instead, the effect seems to be just upstream of terminal chondrocyte differentiation and hypertrophy, based on a positive correlation between PANX3 knockdown and reduction of COL10A1 expression. Further work will still be required to fully explore the role of Panx3 in endochondral bone development. 5.5 Acknowledgements We would like to thank Dr. Sheri Holman for kindly providing us with the pENTR3C-mir-30a plasmid. 131 Chapter 6 Extended discussion and concluding remarks “Any new fact or insight that I may have found has not seemed to me as a ‘discovery’ of mine, but rather something that had always been there and that I had chanced to pick up.” -Subrahmanyan Chandrasekhar The number of physiological processes pannexins (Panxs) have been shown to be involved in has steadily increased over the past 7–8 years, and the work presented in this thesis has further expanded this list — Novel Panx isotypes have been identified in teleost fish with distinct expression profiles and channel properties, and the role of Panx3 in bone development has been explored, including its transcriptional regulation along with the developmental consequences of abnormal expression. Each of these findings were discussed in their respective chapters, so will not be exhaustively re- stated here. Instead, each result will be summarized and more critically evaluated, and then the future outlook will be discussed with recommenda- tions as to how the work could be extended. 6.1 Panx1a and Panx1b are probably ohnologs Well accepted statistical methods do not currently exist to test the proba- bility that a given set of paralogs result from a whole genome duplication event, and so the description of panx1a and panx1b in chapter 2 relied on a more qualitative approach to infer their evolutionary relationship — the cir- cumstantial evidence is nonetheless compelling. Without question the entire gene was involved in the duplication, because the reinsertion of a processed mRNA (via retrotransposition) would have needed to have been followed by up to five novel insertion events, in precisely the same intronic position as those in the original gene. The gain of new introns is an extremely rare event 132 in mammals, with no evidence to suggest the situation is any different in fish [394], thus the probability of multiple intron reinsertion events in panx1 (prior to the radiation of contemporary teleosts no less) is essentially zero. In terms of other possible duplication mechanisms, it is conceivable that a more conventional process is responsible (such as in-equal crossover), but as discussed earlier, timing of the event must have coincided very near R3 for both of the panx1 genes to be so well represented in the teleost lineage. This is in addition to further genetic modifications that would be necessi- tated by a non-R3 origin, including the deletion of one extra copy of the gene if the duplication occurred after R3, and deletion of two copies if it happened before. The most parsimonious explanation is an R3 origin. 6.2 Greater functional diversity through gene duplication The physiology of teleost Panx1 has been modified following its duplica- tion about 350 million years ago (MYA). Despite the caveats of C-terminal tagging (discussed in the next section), there is clear evidence that the tran- scriptional regulation of the two genes has changed, and that they are no longer equal in terms of biochemical properties. Evolutionary theory pre- dicts that gene duplication will either result in loss of one copy over time or divergence through neo- or subfunctionalization [272], so these results conform with conventional models, but they also reveal something more specifically about Panx1. Upon first glance, Panx1 does not seem to be critically important; knockout mice are viable and healthy, with no outward signs of congenital deformity [156, 157, 159, 167, 168, 197, 203, 208, 209]. Yet we see retention of this gene not only across all higher chordates stud- ied, but it was conspicuously retained and redeployed following the teleost genome duplication (significant, because >95% of duplicate gene pairs have lost one, or both, copies since that time [273]). Inferentially then, Panx1 must have greater importance than is evident by observations in the tightly controlled laboratory setting. Exciting recent work by Katharina Kranz has even added weight to this hypothesis, by identifying a role for Panx1 in the mouse retina [395]. With the exception of the photoreceptor layer, Panx1 is present throughout all retinal layers, primarily in neurons [109, 110, 176]. Interestingly, the Panx1(−/−) mouse experiences greater response amplitude to light inputs when dark conditioned, which would significantly reduce the resolving power of rod cell receptive fields [395]. This is an excellent example of a phenotype that would be invisible in a laboratory setting, while poten- 133 tially conferring a significant disadvantage in the wild. Furthermore, this could be where subfunctionalization has occurred, decoupling the retinal ef- fects of teleost Panx1b from, for example, potential vasodilatory properties of Panx1a (assuming Panx1a confers similar functionality as Panx1 in mouse [132, 185]). Future knockout work in the zebrafish could go a long way to resolve what the differences are between Panx1a and Panx1b, and by ex- tension to identity how the protein can be subfunctionalized. Furthermore, it could be very interesting to assess the state of panx1b in teleost species that do not rely on vision, such as the blind cave dwelling fish Astyanax mexicanus. While the ohnologs discussed here may allow us to separate the function- ality of teleost Panx1 in the brain and eye from other tissues, could there be even more Panx genes awaiting discovery, with even greater subfunctional- izion? Viable/fertile polyploidy is a fairly rare occurrence in most Metazoan phyla, and yet is surprisingly common in plants. The reason for this dis- crepancy lays in a combination of plants’ relative tolerance to hybridization, their plasticity of body form, and an ability to self fertilize. Hybridization between related species can generate an F1 offspring whose chromosomes are sufficiently divergent to impede pairing during meiosis, but is otherwise healthy. If however, an error occurs during mitosis in a well positioned so- matic cell that results in genomic duplication, that cell has the potential to develop into a complete ‘allotetraploid’ germ tissue that can become self fer- tile [396]. This is obviously not as simple in animals, which do not enjoy the same flexibility in terms of generating new sexual structures from disparate regions throughout their body, and even in the rare case where the germ tissue is affected, there is still the need to encounter and mate with another individual with the same genetic abnormality. This obstacle is not insur- mountable of course, because genome duplications are still evident in higher animals. Xenopus laevis is a well studied example, having undergone no less than six polyploidization events throughout the history of extant lineages [397], at least four whole genome duplication (WGD) events are predicted to have occurred in separate lineages of Acipenseridae (sturgeons) [398], and although controversial, there is even evidence of a genome duplication in a the red vizcacha rat Tympanoctomys barrerae (controversial because poly- ploidy is almost always lethal in mammals) [399]. Irrespective of how these duplications have occurred, they represent a possibility of even greater Panx diversity, and as the genome sequences of these organisms become available, the presence or absence of more Panx ohnologs will be determined. 134 6.3 Effects of epitope tags C-terminal tagging with fluorescent or antigenic sequences is a common practice that has been successfully employed in many connexin (Cx) studies [400–402]. Panxs seem to be somewhat sensitive to these additions how- ever, and while some groups have reported no untoward effect on trafficking or function [133, 134, 172], others have observed partial, or even complete, loss of membrane localization and/or channel activity [111, 137, 153]. In chapter 2, all of the Danio rerio Panxs were tagged with EGFP on their C-terminal tail to facilitate visualization. This decisions was justified, be- cause it allowed for rapid preliminary comparison between each paralog in lieu of comprehensively screening commercial antibodies or generating our own (which would have come at a non-trivial expense, in both dollars and time, with no guarantee of success). To try and control for adverse tagging effects, care was taken to use the same linker sequence to fuse EGFP to the C-terminus of each Panx. While this does help to some degree, in that any biochemical effects between the tag and other components of the cell should be more-or-less uniform across samples, it does very little to prevent vari- able interactions between the tag and each individual isotype. One way to assess whether EGFP has negatively impacted the physiological properties of the Danio Panxs would be to repeat the experiments using different tags, and see if the results are equivalent. For example, HaloTag is about the same size as EFGP and has a number of fluorescent binding ligands that can be used for live cell imaging [403], or tetracysteine, which is very small at only six amino acids, that would allow FlAsH/ReAsH based detection [404]. Until antibodies are identified that can react with these Panxs, tag- ging is about the only way to explore their trafficking pattern. To assess channel activity on the other hand, the best approach would be to study the untagged protein. The lack of easily resolvable opening/closing events from Panx1a-EGFP strongly suggests that the channel is being interfered with, especially when one considers previous reports of easily identifiable unitary Panx1a activity [110, 133]. If removing the tags does not alter the result then our findings will be at odds with this literature (this could perhaps warrant collaboration with these previous authors to try and replicate each others results, using each others constructs), otherwise we will have pro- vided further evidence for the importance of the Panx1 C-terminal tail in channel activity. If indeed the final 8–10 residues on the extreme C-terminus normally line the inner pore of the channel, as suggested by SCAM analysis [98], then it might actually be more surprising that any C-terminal tagged Panx1 forms functional channels. 135 6.4 Regulation of the Panx3 promoter requires, but also transcends Runx2 In chapter 4, Runx2 was clearly shown to influence Panx3 expression in cul- tured cells. Not only did Runx2 over-expression significantly increase the level of Luciferase production when the reporter gene was downstream of the Panx3 promoter, but mutating the putative Runx2 binding site within the promoter was sufficient to reduce Runx2 induced expression by 80–90%; strongly implying that Runx2 acts in cis to the Panx3 gene. The support for this is further enhanced by the fact that a Runx2 antibody was able to enrich the Panx3 promoter from a chromatin preparation. It should be recognized however that Runx2 cannot be the whole story. As discussed in section 4.4, expression of the reporter gene was dramatically higher when the pro- moter constructs were transfected into primary osteoblasts than they were in HEK-293 cells, despite exogenous over-expression of Runx2 in the HEK-293 cultures. This strongly implies that there are cofactors present in osteoblasts that enhance the ability of Runx2 to promote Panx3 transcription, and/or there are repressive factors present in HEK-293 cells that impede transcrip- tion. Evidence for both enhancers and repressors of osteogenic genes abound in the literature, and an excellent example is actually the Runx2 gene itself [322]. Msx2 suppresses Runx2 transcription while osteoprogenitor cells are proliferating, but is then down-regulated when BMP-2 signals for differenti- ation to occur. BMP-2 also stimulates expression of Dlx3 and Dlx5, which are sufficient to activate the Runx2 promoter when they replace the dwin- dling Msx2 (they bind to the same sequence). Newly synthesized Runx2 is then able to recursively enhance its own expression by binding to its own promoter [322]. The osteocalcin promoter provides an even more compelling example, because it contains a vitamin D3 responsive element very similar to that found in the Panx3 promoter. During osteoblast differentiation, VDR/RXR heterodimers greatly enhance the basal level of osteocalcin ex- pression stimulated by Runx2 alone [405]. It would not be surprising to see a very similar effect on Panx3 transcription (Fig. 6.1). Any future initiative to extend the work in chapter 4 should begin by systematically mutating regions of the 648 bps Panx3 promoter construct, including the predicted Barx2 and VDR/RXR binding sites, as well as the highly conserved regions around -70, -145, -175, and -260 bps. We already know that the Runx2 binding site is necessary for Panx3 transcription to proceed in osteoblasts, but testing this expanded set of constructs could also address the question as to whether the Runx2 site is sufficient; i.e., 136 Cbfβ Runx2 Runx 2 VDR/RXR VDR/RXR Msx1/2 Msx 1/2 Dlx3/5 Dlx3/5 Pol II Figure 6.1: Proposed group of transactivating factors that may be respon- sible for driving Panx3 expression in osteogenic cell types 137 if mutations in another locale renders the promoter less responsive, than the answer would be no. Co-expressing the mutant promoter constructs with exogenous Runx2 in HEK-293 cells could also be very informative, because any sudden increases in reporter activity would be characteristic of a repressive domain being eliminated. If other locations on the promoter are identified as regulatory elements, and candidate transcription factors can be rationally selected that may bind to them, then chromatin IP and siRNAs can be used as confirmation, otherwise a yeast two-hybrid screen [406] or quantitative proteomics approach [407] could be implemented to identify new candidate regulatory proteins. 6.5 RCASBP delivery of shRNA in vivo The earliest implementations of RNAi to suppress gene expression in chick- ens involved electroporation of synthetic dsRNA into the spinal cord [408] and U6-driven shRNA plasmids into the brain [409]. These provided a very successful proof of concept, but as with any use of ‘naked’ dsRNA or episo- mally retained plasmid, the effects were transient. Around the same time, another group developed a way to deliver shRNA using a modified RCASBP vector that allowed for prolonged knockdown in the developing limb [410], but they failed to release details about the plasmid construction. A number of subsequent studies have gone on to generate their own viral-shRNA deliv- ery systems, usually using some variant of the U6-promoter to drive hairpin synthesis from the RCANBP vector (there is no splice acceptor for the ex- ogenous sequence in this vector, so expression relies on the addition of a user defined promoter) [393, 411–413], but given how important knockdown and knockout studies have become in other model organisms, there has been surprising little development of this technology. As such, I invested time to extend the work of Mo Chen and colleagues [366, 367, 414], who used the long terminal repeat promoter and splice acceptor of RCASBP to deliver the chicken mir-30a microRNA gene for prolonged knockdown [366, 367, 414]. Mir-30a encodes a 0.5 kb transcript that is processed by the drosha/dicer machinery into siRNA duplexes for incorporation into the RISC complex [415], and because full length pre-miRNA genes have been shown to in- crease knockdown efficiency over conventional ‘naked’ shRNA sequences, I chose to develop this technology over the competing U6 driven options. The mir-30a gene was commandeered to accept user specified hairpins by flank- ing the endogenous hairpin with restriction sites [366, 367], and it was placed into the GateWay compatible entry vector pENTR3C to allow easy transfer 138 into the RCASBP destination vector originally constructed by Stacy Lof- tus [365]. My preference was to use a restriction-free approach however, so the pENTR3C-miR-30a construct was modified as outlined in section 5.2.1 (Fig. 6.2). This new approach provided a number of benefits, including shorter synthetic oligos (∼56 bp verses 78 bp), elimination of possible re- striction site incompatibility with the desired hairpin sequence, more faithful replication of the original miR-30a gene, and the addition of a KpnI recog- nition sequence in the new insertion site that allows for pre-treatment with this enzyme to reduce insert-negative colony growth (assuming the hairpin sequence does not contain a KpnI site). Combined with a small web accessi- ble implementation∗ of the siRNA selection algorithm developed by Angela Reynolds [368], we now have a reliable, rapid, and inexpensive method for targeted knockdown of any chosen gene. To improve the technology even further, the size of the loop-sequence could probably be reduced [369], and addition of GFP to the final RCASBP construct would provide an excellent reporter to ensure proper viral targeting (this was successfully demonstrated in one of the early U6-driven constructs [413]). An enterprising individual could even consider capitalizing on economy of scale if they wanted to try commercializing RCASBP-shRNA construction, since no such service cur- rently exists. 6.6 The effects of Panx3 on endochondral bone development As explained in chapter 1, it was completely unknown at the outset of the work presented in this thesis at to how, or why, Panx3 is expressed in bone, although a few bone-centric reports have been published in the intervening years. The first of these was a simple descriptive study, showing expression of Panx3 in the bones of the ear without any concern for function [175]. The other two both came out of a single lab, and these proposed a physiologi- cal relevance for Panx3 expression in mineralizing cell types [125, 169]. In both cases the conclusion was that Panx3 is osteogenic, whether expressed in osteoblasts or chondrocytes, and that suppressing Panx3 expression in these cells significantly impairs their ability to differentiate. In the first of these reports, Tsutomu Iwamoto and colleagues focus on chondrocyte dif- ferentiation; Panx3 was shown to promote nodule formation (i.e., cartilage) in cultured chondrogenic cell lines and primary chondrocytes, along with concomitant changes in expression of the ECM genes Col2a1, Col10a1, and ∗http://www.naus-lab.com/tools/siRNA_prediction.php 139 pENTR3C-miR-30a 2768 bp KpnI 654 BamHI 508 EcoRI 502 SacII 497 HindIII 491 ClaI 485 SalI 476 NheI 43 MfeI 938 NotI 973 XhoI 982EcoRV 987 PstI 1104 rrnB T1 terminator rrnB T2 terminator attL1 miR-30a attL2 KanR2 pBR322 origin PsiI 1023 PsiI 424 a uc g gaagc gcg cuguaaacaucc gacuggaagcu u a ||| ||||||||||||  ||||||||||| | cgu gacguuuguagg cugacuuucgg g g c -- g uagac gcgaacacggcugcugaguacau ||| | ||||||||||||||||||||||| cgugugugccgacgacucaugua g gaagc ua u a au g g g uagac Original miR-30a hairpin chPanx3-14 miR-30a hairpin Figure 6.2: Entry plasmid used to shuttle custom miR-30 shRNA constructs into the RCASBP expression plasmid. The original miR-30a hairpin se- quence is compared to the hairpin sequence used to knockdown Panx3 in chapter 5 aggrecan. They propose that cell surface Panx3 desensitises the cells to parathyroid hormone 1 receptor (Pth1r) mediated proliferation by decreas- ing intracellular cAMP levels, to ultimately reduce the pro-mitotic effects of phospho-CREB. This model does not easily explain the in vivo results presented here however. If Panx3 is sufficient to slow chondrocyte prolif- eration, then we should expect a reduction in the length of Panx3 over- expressing long bones, as a result of fewer chondrocytes being present in the tissue. Conversely, we could even predict greater long bone volume follow- ing Panx3 knock-down, as the proliferating chondrocytes stay mitotically active for longer before the balance finally tips in favour of differentiation. As we saw in chapter 5 this is essentially opposite to what actually occurs, with Panx3 over-expression having little-to-no effect on the gross morphol- ogy of forelimb bones progressing through endochondral ossification (EO), while knock-down significantly reduces the size of these same bones. When we further consider that most chondrocyte proliferation occurs well before the cells enter the hypertrophic zone [388], and thus is temporally and spa- tially removed from where Panx3 is actually expressed, it seems even more unlikely that the anti-mitotic effects of Panx3 reported in cultured cells play a significant physiological role in vivo. Ishikawa et. al. published the second manuscript from the research group Iwamoto is part of, and reca- 140 pitulated some of the in vivo phenotype reported here, although they chose to use metatarsal explants and a putative Panx3 mimetic peptide to block function instead of knockdown technology [125]. They report a reduction in endochondral outgrowth in the presence of peptide, in association with modest reductions in expression of the differentiation markers Sp7, Alp, and Bglap. Interestingly, they also report a strong increase in expression of these genes when they transduced the explant cultures with an adeno-virus carry- ing exogenous Panx3, as well as a 1.5 fold increase in total length. A similar analysis of gene expression was not pursued in the chicken limbs following PANX3 over-expression in chapter 5, but should perhaps be included in fu- ture work to determine whether the observations of Ishikawa are an artefact of their culture system, or can be recapitulated in vivo. Two general mechanisms could be responsible for forelimb shortening fol- lowing PANX3 knock-down: reduced chondrocyte proliferation, or reduced matrix secretion and hypertrophy. BrdU labelling did not suggest that pro- liferation rates were significantly altered in RCASBP::PANX3 -14 infected limbs, and there were no obvious changes in the expression of markers tradi- tionally associated with proliferating chondrocytes, such as IHH, BMP7, or PTHR1. Unfortunately the variance within the BrdU and qPCR data was quite large, leading to low statistical power and increased chance of type 2 error, so the possibility of a Panx3 mediated paracrine signalling mech- anism that regulates chondrocyte proliferation should not be completely dismissed. Based on the current evidence however, it seems likely that the effect is rooted in the hypertrophic zone. Most compelling is the correla- tion between PANX3 and COL10A1 expression in both the developing limb demonstrated here, as well as in primary chondrocytes elsewhere [169]. This suggests that PANX3 could indeed be involved in terminal differentiation, and the reduction in bone size could thus be the result of impaired hyper- trophy. Referring back once again to the work by Ishikawa and colleagues, Panx3 localization to the endoplasmic reticulum (ER) was shown to correlate with release of calcium into the cytosol [125], which in turn is an important activator of calcium/calmodulin-dependent protein kinase II (CaMKII) in prehypertrophic chondrocytes [416]. CaMKII activation drives chondrocyte hypertrophy, along with expression of genes like IHH and COL10A1, so could be at least partially responsible for the in vivo effects. Going forward, a number of experiments should be performed to clarify whether the hypertrophic zone is in fact the key pool of chondrocytes that are sensitive to Panx3 expression. To begin, Replication Competent ALV LTR with a Splice acceptor (RCASBP)::PANX3 -14 infected embryos should be allowed to mature to stage 36–37, allowing time for primary centres of 141 ossification and identifiable metaphyses to form, thus ensuring that a pool of fully matured hypertrophic chondrocytes is present prior to analysis. To gain more confidence in the cell division data currently presented, BrdU la- belling should be repeated in these samples and the different growth plate zones clearly delineated by co-labelling with Col10a1 and Pthlh antibodies so each can be analysed for mitotic activity independently. Furthermore, aver- age chondrocyte size should be assessed in each zone to determine the effect of PANX3 on hypertrophy, and the distance between chondrocytes can be measured as a rough readout on changes in extracellular matrix (ECM) pro- duction. Chondrocyte disorganization is also often correlated with growth abnormalities caused by mutations in growth plate genes [417, 418], and can be assessed histologically using safranin-orange or Alcian-blue staining coupled with measurement of morphometric parameters of columnar orga- nization [419]. Another set of stage 36–37 embryos should also be used to repeat the qPCR experiments, but instead of harvesting mRNA from the entire limb, which includes a lot of contaminating tissue, the bones should be isolated prior to mRNA extraction. A second PANX3 -shRNA sequence should also be used to confirm that any significant results are in fact specific to PANX3 knockdown, and not from an off-target effect of the PANX3 -14 construct. It might also be worthwhile to assess gene expression changes on a larger scale, using microarray or RNA-Seq instead of trying to predict ahead of time which genes may have changed for analysis by qPCR. A even more ambitious experiment could involve cloning the COL10A1 promoter into the RCANBP vector, and using this to drive expression of PANX3 -shRNA only in the hypertrophic chondrocytes. I am unaware of this technique be- ing used in chicken, but there is a precedent of exploiting a tissue specific RNA polymerase II promoter to drive shRNA expression in mice [420]. The outcome of such an experiment could be quite illuminating, either strongly supporting my hypothesis by producing a phenotype of equal magnitude to that seen with the current shRNA construct, or essentially falsifying it if robust knockdown occurs in the hypertrophic zone without the affiliated phenotype, thus implying that Panx3 expression in the surrounding peri- chondrium is probably involved. An intermediate phenotype would indicate that Panx3 expression is important in both the hypertrophic zone and sur- rounding perichondrium. 142 6.7 Panx3 C-terminal leucine zipper Predicting transmembrane domains from the Panx3 primary sequence us- ing a hidden Markov model [421] reveals four clearly defined membrane spanning motifs, but a fifth sequence within the C-terminal tail also meets several requirements of a transmembrane domain, and is flagged by the algo- rithm (Fig. 6.3A). Closer inspection of a multiple sequence alignment from 24 species of vertebrate (spanning mammals, amphibians, birds, and fish) reveals a ubiquitously conserved group of leucine and isoleucine residues be- tween positions 321–342 (Fig. 6.3B), and if arranged into an α-helix, the leucine side chains form a ‘leucine zipper’ motif [422] (Fig. 6.3C). Two such motifs, when juxtaposed, can interdigitate to form the zipper, which is an effective mode of dimerization between protein subunits [423]. While classi- cally associated with transcription factors, the leucine zipper has also been shown to confer valuable functionality to transmembrane ion channels, in- cluding voltage sensitivity, kinase/phosphatase recruitment, and facilitating oligomerization between channel subunits [424]. Nothing is currently known about the purpose of this structure in Panx3, and it does not exist in the other isotypes, so it could point to an exciting future direction for further inquiry. There have still not been any direct electrophysiological recordings of the Panx3 channel at the cell surface (despite claims of increased dye uptake [104, 107] or calcium release from ER stores [125] in Panx3 over- expressing cells), perhaps because a leucine zipper forms between adjacent monomers in the mature channel, holding it in a closed state. Mutagenesis of the leucine residues may thus allow the channel to open, and would be an entirely novel gating mechanism within either of the Cx or innexin (Inx) superfamilies. 6.8 A place for RF-Cloning.org in the molecular biology community RF-Cloning.org started out as a small script to automate the hybrid primer design process for restriction free cloning projects. The steps involved in rf- cloning primer design are not conceptually very difficult, but keeping track of the insert and destination sequences manually introduces a non-trivial opportunity for human errors. After experiencing such errors first-hand, I wrote the automation program and made it accessible through the internet. The user interface was extremely primitive however, and very confusing for others to use. Over the course of about two years the usability of the site 143 L319 T320 C321 P322 I323 N3124 D3125 L326 N327 V328 I329 L330 L331 F332 L333 R334 A335 N336 S338 E339 L340 I337 a b c d e f g A B C Figure 6.3: A) Putative Panx3 transmembrane domains were predicted using the TMHMM Server v.2.0 (urlhttp://www.cbs.dtu.dk/services/TMHMM). Each of the first four probability peaks represents an actual transmembrane domain (see Fig. 1.3 for schematic), but the fifth peak (green arrow) is unlikely to be a transmembrane domain. B) Multiple pairwise alignment reveals highly conserved leucine and isoleucine residues at the position of the fifth peak. Red arrows indicate canonical leucine zipper residues, each spaced by seven residues (i.e., one turn of an α-helix), while green arrows point to possible supporting leucine and isoleucine residues that may con- tribute to Panx3 oligomerization. C) Helical wheel plot of the putative leucine zipper motif in Panx3, clearly illustrating the alignment of the con- served residues indicated in B 144 was dramatically improved, and it was accepted for publication in the 2012 Nucleic Acids Research web server issue [292]. During review, it was pointed out that pre-existing online tools like Primer3 [425] could also be used to design these hybrid primers if the default settings were modified, but it was also recognized that there is real value in an attractive and intuitive site design. The defining event that allowed RF-Cloning.org to begin its transition from a piece of personal-use computer code to a bona fide publicly accessi- ble resource, was probably the decision to build the user interface around Savvy plasmid maps. Even more important was the generosity of Dr. Malay Basu, who provided the original Savvy source code that made the current version of RF-Cloning.org possible. There are surprisingly few online tools for generating custom plasmid maps, and each has limitations. EZ Plasmid Map† and NetPlasmid‡ both require the user to manually specify all fea- tures, without any automation options based on sequence. This is fine if the user only has a small number of maps they need to build, and having abso- lute control over feature position does ensure desired output, but it comes at the expense of time. Even restriction sites must be annotated manually, which can be very tedious for a single plasmid, let alone a complete library. PlasMapper [301] on the other hand is very feature rich, supporting both automated and manual annotation, but it restricts the number of manually annotated features to six with no option to adjust the look of the output, and provides no option to specify a custom set of restriction sites. Savvy suffered from the same manual annotation issue as EZ Plasmid Map and NetPlas- mid, but because the program that rendered the maps was served through an open common gateway interface, it was possible to reverse engineer the necessary input requirements, and then simply inject them directly into the Savvy server remotely (i.e., I could use the map drawing program with- out having access to its source code). In later versions of RF-Cloning.org, JavaScript (which runs on the clients web browser) was used to make the site much more dynamic, allowing the user to make asynchronous calls back to server side programs on the host machine. This allowed changes to be made to the underlying database, for example, without requiring a browser window refresh (this technology is known as asynchronous JavaScript and XML (AJAX)). Due to security risks associated with this type of client– server interaction however, web browsers do not allow AJAX calls to servers other than the one where the initial content originates, so it was not possible †http://www.infosake.com/plasmid/ ‡http://www.justbio.com/index.php?page=netplasmid 145 to continue remotely accessing the Savvy programs from a truly dynamic user interface. The issue was fortunately solved by contacting Dr. Basu and acquiring the Perl source files. Having control over the code also provided the ability to make improvements. For example, the subroutine controlling where restriction enzyme names are output on the final map has been com- pletely re-written, producing a more legible final product (see Fig. 6.4 for comparison). pExample 4733 bp AgeI 665 BamHI 659 XmaI 655 TspMI 655 SmaI 655 PspOMI 652 ApaI 652 SacII 648 KpnI 644 Acc65I 644 SalI 638 PstI 633 EcoRI 628 HindIII 621 SacI 615 Eco53kI 615 XhoI 612 TliI 612 PaeR7I 612 BglII 608 AfeI 593 NheI 590 BmtI 590 SnaBI 337 NdeI 232 AseI 5 BsrGI 1387 NotI 1399 XbaI 1410 MfeI 1506 HpaI 1517 AflII 1638 FspI 2856 MscI 2836 SfoI 2755 NarI 2755 KasI 2755 ClaI 2595 BspDI 2595 StuI 2575 ApaLI 4360 PciI 4674 RF-Cloning.org pExample 4733 bp AseI 5 NdeI 232 SnaBI 337 BmtI 590 NheI 590 AfeI 593 BglII 608 PaeR7I 612 TliI 612 XhoI 612 Eco53kI 615 SacI 615 HindIII 621 EcoRI 628 PstI 633 SalI 638 Acc65I 644 KpnI 644 SacII 648 ApaI 652 PspOMI 652 SmaI 655 TspMI 655 XmaI 655 BamHI 659 AgeI 665 BsrGI 1387 NotI 1399 XbaI 1410 MfeI 1506 HpaI 1517 AflII 1638 StuI 2575 BspDI 2595 ClaI 2595KasI 2755 NarI 2755 SfoI 2755MscI 2836 FspI 2856 ApaLI 4360 PciI 4674 Original Savvy Figure 6.4: Changes have been made to the original (right) Savvy program used to create plasmid maps on RF-Cloning.org (left) Now that RF-Cloning.org is in full production, it will hopefully serve a dual purpose. First of course, is as a tool for those already familiar with rf-cloning to design primers quickly and accurately. A lot of emphasis was placed on including functionality to the homepage and output page that will make the design process as easy as possible. From the homepage, insert and plasmid sequences can be manipulated using a variety of tools based largely on those found on the Sequence Massager website§, insertion sites can be identified numerically or directly in the sequence, advanced settings pro- vide an option to override the default annealing temperatures and min/max §http://www.attotron.com/cybertory/analysis/seqMassager.htm 146 primer lengths, the plasmid map can be auto-generated as a graphic or as colour-coded text, and regularly used plasmid backbones are accessible from a drop-down menu (which can be added to by users). After the project is executed, the user then has full freedom to dynamically shift the insert sites or extend/decrease the size of the primers, they can BLAST their sequenc- ing results against the predicted plasmid, features and restriction sites can be annotated manually or automatically, and the maps can be exported as scalable vector graphics (SVG) or as GenBank flat files. The entire work- flow is streamlined into an easy to use, easy to understand online resource. The second purpose is to educate more people about rf-cloning as a powerful and rapid tool for generating custom plasmids when convenient restriction sites are not present. 6.9 The future of RF-Cloning.org Pursuant to the author’s agreement for publication in the web server issue of Nucleic Acids Research, RF-Cloning.org must be hosted on a publicly accessible server for the next five years. To avoid any confusion, the look and feel of the primer design tools will not be altered significantly during this time, although the underlying technology can be upgraded to the new HTML5 specification¶. This may improve search engine optimization (to generate more traffic), but the primary motivation would be to streamline rendering of SVG elements. All of the dynamically generated plasmid maps so prominently featured throughout the site are encoded as SVG, but due to the fact that HTML4 made no provisions for the standard, native support was not widely implemented in popular web browsers until recently. Mi- crosoft’s Internet Explorer (prior to IE9) was particularly resistant to sup- porting the format until it was codified in HTML5, so all SVG on the site is currently parsed by a custom function that uses an open source JavaScript library called Raphaël‖ to render the images cross-platform. This makes it slightly more complicated to modify the Savvy class responsible for gener- ating the plasmid maps, and significantly more difficult to add new static SVG images, which are generally produced in vector graphics software like InkScape or Adobe Illustrator that do not output the correct mark-up re- quired by Raphaël. Since early 2011 however, all major web browsers have implemented inline SVG support, thus the Raphaël work-around has become deprecated. ¶http://www.w3.org/TR/html5/ ‖http://www.raphaeljs.com 147 Figure 6.5: Registered users of RF-Cloning.org are able to access a plasmid management dashboard where they can view and modify their saved projects and plasmid backbones. New backbones can also be added to their profile, and these can be made public should to user desire While the core primer design tool should not be substantially modified, lest a disconnect form between the published description of the tool and the live version, there are no such restrictions on the plasmid management system available to registered users. The plasmid management dashboard provides a means of storing/organizing rf-cloning projects and custom plas- mid backbones, but due to the sign-in requirement, was only mentioned in passing in the manuscript (the Nucleic Acids Research author’s agree- ment requires that all services described are free and do not necessitate creation of an account). Figure 6.5 shows a screenshot of this dashboard, which could be the starting point for a much more feature rich DNA ma- nipulation platform. Currently, the leading commercial molecular biology suites include Geneious (Biomatters Ltd.), Vector NTI (Invitrogen), and MacVector (MacVector, Inc.); ranging in cost from 100/year for a student version of MacVector, to >3000 for a static academic license of Vector NTI. Free alternatives exist, including Serial Cloner (Serial Basics), Gene 148 Designer (DNA 2.0), pDRAW32 (AcaClone software), and ApE (M. Wayne Davis) among others, but all of these are desktop applications that require installation. In terms of completely online solutions, LabLife.org provides a number of useful tools for managing, manipulating, and outputting DNA sequences, but their interface does not provide the level of dynamic inter- activity users have with their sequences within the various desktop applica- tions. In the four months following the online release of the RF-Cloning.org manuscript, the amount of traffic to the site more than doubled (Fig. 6.6), and users have begun to send in comments, suggestions, and bug reports. If this growth continues, and it becomes obvious that new tools will be uti- lized if integrated with the plasmid management system, there is a wealth of options for upgrade: DNA-protein translation, codon optimization, for- mat conversion, more flexible sequence alignment, contig assembly, sequence colour coding, event driven marker assignment (e.g., selecting a sequence, right-click, assign sequence feature.), directory-based plasmid organization, integration with public plasmid databases like Addgene, open reading frame identification, prediction of primer-dimer and hairpin artefacts, motif predic- tion. . . There are actually very few restrictions on what could be included, because abstraction would allow most conceivable tools to be built and val- idated in isolation from the other functionality of the site. The knock-on benefit of this of course, is that other interested individuals are able to build tools independently, and would only need my direct involvement for the final phase of lunching the tool on the production server. Hence, RF-Cloning.org is structured to become an open-source community project if the interest develops — only time will tell if it will play out this way though. 6.10 Personal appeal for greater computational literacy in the life sciences I began learning to program in a high-level scripting language (PHP in this case) 2–3 years into my PhD; a number of years before work began on RF-Cloning.org. Aside from an entry level computer science course dur- ing my undergraduate degree, where macro scripting in Visual Basic and primitive web design in static HTML were briefly introduced, this was my first experience with computer code, and I was shocked by the freedom gained from the knowledge. The ability to manipulate data independent of pre-fabricated software is both powerful and liberating, and the confi- dence to use open source software packages/libraries (which may not be de- signed with the slick graphical user interfaces we are generally accustomed 149 20 40 60 80 100 120 140 160 180 Online publication of NAR manuscript # of  u ni qu e vi si to rs  p er  w ee k Figure 6.6: The average number of weekly visits to RF-Cloning.org has more than doubled since publication of chapter 3 in Nucleic Acids Research 150 to) dramatically expands ones tool kit. It is therefore quite unfortunate that biology programs do not widely include an introductory computer science component in their curriculum. At the graduate level in particular, where managing and manipulating data is a daily task, computational literacy is surprisingly under-appreciated. A brief look at the largest 10 cell and de- velopmental biology (or equivalent) programs in the USA (ranked by the National Research Council as average number of graduating students per year between 2002–2006 [426]) reveals that only the Massachusetts Institute of Technology bioengineering program includes a requisite introduction to programming, and fully half offer no introductory computational courses tailored to graduate students even as part of their elective curriculum. If graduate students from biology programs want to remain competitive in an increasingly saturated scientific workforce, I think this trend is going to need to change. In 2001, the human genome draft sequence was published in Nature [427]. The exercise cost the public 3 billion, requiring thousands of personnel who collectively contributed millions of man-hours. A little over a decade later a single technician can now generate the same amount of output in under a week, and it will soon require little more than an afternoon using reagents costing a couple hundred dollars [428]. To put this in perspective, the rate at which ‘next-gen’ sequencing output is increasing per dollar is outpacing Moore’s Law∗∗ by 3–5 fold [429], so the cost of generating sequencing data is quite literally becoming less expensive than the hard-drives it is stored on. We have entered into the ‘-omics’ era where ‘big-data’ is going to drive medical science forward faster then ever before, and those with the tools and expertise to process the flood of information, and to convert it into some- thing tangible, are going to be the ones who succeed. Of course professional computer scientists will still be the ones solving most of the technical chal- lenges of managing all this data (specialization and separation of labour is very important — it would not be very efficient to pursue mastery at both the bench and the keyboard!), but it is going to take biologists to frame the proper questions that will lead to new discovery. Ergo, progress is depending on effective communication between the two disciplines. For those entering into a career of experimental biology, they owe it to themselves, as well as their future computer savvy collaborators, to have a working knowledge of the machines they use to acquire and analyse their data. They should know the limitations of computer hardware, as well as ∗∗The capacity of many computer components, such as transistor density on integrated circuits, doubles approximately every 18–24 months 151 its strengths, so they can design projects efficiently and to an appropriate scale. They should also have the ability to write their own basic software to manipulate logically structured data, especially as the volume of this data continues to grow larger. A single day is sufficient to teach a reasonably intelligent individual (which a graduate student presumably is) the core concepts that will allow them to start writing their own custom computer code in a high-level scripting languages like Python or Ruby. Furthermore, educational programs like Software Carpentry†† already provide an excellent 2–3 day ‘boot-camp’ model for teaching a suite of useful computer skills to those in the physical and life sciences, so there is very little reason to postpone teaching it as part of core curriculum. 6.11 Conclusion Human knowledge What I know{ My PhD Figure 6.7:  Matt Might http://matt.might.net/articles/phd-school-in-pictures/ Reproduced under the terms of the Creative Commons Attribution- NonCommercial 2.5 License The content of this thesis spans a number of distinct disciplines, which ††http://software-carpentry.org/ 152 of course brings with it the danger of discontinuity. My sincerest hope is that I was able to find balance in my presentation, and you were able to flow from one topic to the next without too jarring a transition. I contend that a PhD should be about learning how to learn, and coming to grips with ones interests, strengths, and weaknesses in the academic space. To those ends, I think I have been successful. I have been able to explore genomic dynamics and architecture, phylogenetic relationships through hundreds of millions of years, and some core facets of evolutionary theory. These topics are simply fascinating, and in the process of learning, I was able to contribute another small discovery to the Pannexin community. Over the past four years I have also learned a considerable amount of computer science; not only in software design, but also in data management, system administration, and philoso- phy. I am convinced that there is a tsunami of data rushing towards the life science community, and my hope is that I am now positioned to ride the wave, instead of being crushed by it. It is to early to tell if RF-Cloning.org will blossom into a truly popular web application for molecular biologists, but even if it never reaches ‘viral’ status, and is only ever used by a few dozen people around the world, at least it is there to help those few dozen people. In any case, the final product is perhaps less important than the journey taken to produce it. And finally, my experience with Panx3 has been no less than tumultuous. I have often compared science to baseball, in that a batting average of 300 is actually quite respectable, so I try not to be too discouraged by the other 70% of the time I fail to get on base. As John E. Olerud put it: “You [just] keep swinging”. Between chapters 4 and 5, there are nearly twenty separate techniques described in the materials and methods, and for every result presented herein, three more languish in notebooks — discarded as uninteresting, or impractical, or due to technical failure. Even so, from this tangled web of stagnant and dead experiments has emerged a small, yet tangible contribution to our understanding of biol- ogy. We can now say with confidence that Panx3 is expressed by bone cells; that this expression is regulated in part, but directly, by Runx2; and that disrupting this expression alters the developmental process of endochondral ossification. It is a bit humbling to condense so many years worth of labour into a single sentence, but I think Matt Might has perhaps captured the scope of a PhD best with the schematic above. At first glance the metaphor can look to be disheartening, putting into perspective the relatively small contribution of any one student. Disillusionment is not Mr. Mights intent however, and instead he uses this picture as a call to arms to ‘keep push- ing’ at the boundaries of our knowledge, because lots of people adding little dimples eventually leads to serious progress. I am very happy to have con- 153 tributed my dimple, and now look forward to adding a few more as I move on to the next phase of my life in science. —Stephen R. Bond 154 Bibliography [1] E. J. Furshpan and D. D. Potter. Transmission at the giant motor synapses of the crayfish. J Physiol, 145(2):289–325, Mar 1959. [2] H. H. Dale, W. Feldberg, and M. Vogt. Release of acetylcholine at voluntary motor nerve endings. J Physiol, 86(4):353–80, May 4 1936. [3] F. M. Watt. Unexpected hedgehog-wnt interactions in epithelial dif- ferentiation. Trends Mol Med, 10(12):577–80, Dec 2004. [4] Y. V. Panchin. Evolution of gap junction proteins–the pannexin al- ternative. J Exp Biol, 208(Pt 8):1415–9, Apr 2005. [5] Goran Shl and Klaus Willecke. An update on connexin genes and their nomenclature in mouse and man. Cell Commun Adhes, 10(4-6):173– 180, 2003. [6] Judy K. Vanslyke, Christian C. Naus, and Linda S. Musil. Conforma- tional maturation and post-er multisubunit assembly of gap junction proteins. Mol Biol Cell, 20(9):2451–2463, May 2009. [7] M. M. Dewey and L. Barr. Intercellular connection between smooth muscle cells: the nexus. Science, 137(3531):670–2, Aug 31 1962. [8] J. D. Robertson. The occurrence of a subunit pattern in the unit mem- branes of club endings in mauthner cell synapses in goldfish brains. J Cell Biol, 19:201–21, Oct 1963. [9] M. V. Bennett, E. Aljure, Y. Nakajima, and G. D. Pappas. Electro- tonic junctions between teleost spinal neurons: electrophysiology and ultrastructure. Science, 141:262–4, Jul 19 1963. [10] W. A. Wells and L. Bonetta. Defining gap junctions. J Cell Biol, 169(3):379, May 9 2005. [11] S. Ahmad, J. A. Diez, C. H. George, and W. H. Evans. Synthesis and assembly of connexins in vitro into homomeric and heteromeric 155 functional gap junction hemichannels. Biochem J, 339 ( Pt 2):247–253, Apr 1999. [12] P. R. Brink, K. Cronin, K. Banach, E. Peterson, E. M. Westphale, K. H. Seul, S. V. Ramanan, and E. C. Beyer. Evidence for heteromeric gap junction channels formed from rat connexin43 and human con- nexin37. Am J Physiol, 273(4 Pt 1):C1386–C1396, Oct 1997. [13] V. Valiunas, R. Weingart, and P. R. Brink. Formation of heterotypic gap junction channels by connexins 40 and 43. Circ Res, 86(2):E42– E49, Feb 2000. [14] E. C. Beyer, J. Gemel, A. Martnez, V. M. Berthoud, V. Valiunas, A. P. Moreno, and P. R. Brink. Heteromeric mixing of connexins: compatibility of partners and functional consequences. Cell Commun Adhes, 8(4-6):199–204, 2001. [15] Agustin D. Martinez, Volodya Hayrapetyan, Alonso P. Moreno, and Eric C. Beyer. Connexin43 and connexin45 form heteromeric gap junc- tion channels in which individual components determine permeability and regulation. Circ Res, 90(10):1100–1107, May 2002. [16] G. T. Cottrell and J. M. Burt. Functional consequences of heteroge- neous gap junction channel formation and its influence in health and disease. Biochim Biophys Acta, 1711(2):126–41, Jun 10 2005. [17] V. Cruciani and S-O. Mikalsen. The vertebrate connexin family. Cell Mol Life Sci, 63(10):1125–1140, May 2006. [18] S. A. John and J. P. Revel. Connexon integrity is maintained by non-covalent bonds: intramolecular disulfide bonds link the extracel- lular domains in rat connexin-43. Biochem Biophys Res Commun, 178(3):1312–1318, Aug 1991. [19] S. Rahman and W. H. Evans. Topography of connexin32 in rat liver gap junctions. evidence for an intramolecular disulphide linkage con- necting the two extracellular peptide loops. J Cell Sci, 100 ( Pt 3):567– 578, Nov 1991. [20] C. I. Foote, L. Zhou, X. Zhu, and B. J. Nicholson. The pattern of disulfide linkages in the extracellular loop regions of connexin 32 sug- gests a model for the docking interface of gap junctions. J Cell Biol, 140(5):1187–1197, Mar 1998. 156 [21] G. Dahl, R. Werner, E. Levine, and C. Rabadan-Diehl. Mutational analysis of gap junction formation. Biophys J, 62(1):172–80; discussion 180–2, Apr 1992. [22] P. D. Lampe and A. F. Lau. The effects of connexin phosphorylation on gap junctional communication. Int J Biochem Cell Biol, 36(7):1171– 86, Jul 2004. [23] Ane Kjenseth, Tone Fykerud, Edgar Rivedal, and Edward Leithe. Reg- ulation of gap junction intercellular communication by the ubiquitin system. Cell Signal, 22(9):1267–1273, Sep 2010. [24] B. N. Giepmans. Gap junctions and connexin-interacting proteins. Cardiovasc Res, 62(2):233–45, May 1 2004. [25] V. M. Unger, N. M. Kumar, N. B. Gilula, and M. Yeager. Three- dimensional structure of a recombinant gap junction membrane chan- nel. Science, 283(5405):1176–80, Feb 19 1999. [26] Shoji Maeda, So Nakagawa, Michihiro Suga, Eiki Yamashita, Atsunori Oshima, Yoshinori Fujiyoshi, and Tomitake Tsukihara. Structure of the connexin 26 gap junction channel at 3.5 a resolution. Nature, 458(7238):597–602, Apr 2009. [27] I. Simpson, B. Rose, and W. R. Loewenstein. Size limit of molecules permeating the junctional membrane channels. Science, 195(4275):294–296, Jan 1977. [28] Andrew L. Harris and Darren Locke. Permeability of connexin chan- nels. In Andrew L. Harris and Darren Locke, editors, Connexins: A guide., pages 165–206. Humana Press, 2009. [29] G. S. Goldberg, V. Valiunas, and P. R. Brink. Selective permeability of gap junction channels. Biochim Biophys Acta, 1662(1-2):96–101, Mar 23 2004. [30] Andrew L. Harris. Connexin channel permeability to cytoplasmic molecules. Prog Biophys Mol Biol, 94(1-2):120–143, 2007. [31] R. D. Veenstra, H. Z. Wang, D. A. Beblo, M. G. Chilton, A. L. Har- ris, E. C. Beyer, and P. R. Brink. Selectivity of connexin-specific gap junctions does not correlate with channel conductance. Circ Res, 77(6):1156–1165, Dec 1995. 157 [32] C. Elfgang, R. Eckert, H. Lichtenberg-Frate, A. Butterweck, O. Traub, R. A. Klein, D. F. Hulser, and K. Willecke. Specific permeability and selective formation of gap junction channels in connexin-transfected hela cells. J Cell Biol, 129(3):805–17, May 1995. [33] Rebecca Lewandowski, Junko Shibayama, Eva M. Oxford, Rosy Joshi- Mukherjee, Wanda Coombs, Paul L. Sorgen, Steven M. Taffet, and Mario Delmar. Chemical gating of connexin channels. In Andrew L. Harris and Darren Locke, editors, Connexins, pages 129–142. Humana Press, 2009. [34] Heather S. Duffy, Paul L. Sorgen, Mark E. Girvin, Phyllis O’Donnell, Wanda Coombs, Steven M. Taffet, Mario Delmar, and David C. Spray. ph-dependent intramolecular binding and structure involving cx43 cy- toplasmic domains. J Biol Chem, 277(39):36706–36714, Sep 2002. [35] Bethany J. Hirst-Jensen, Prangya Sahoo, Fabien Kieken, Mario Del- mar, and Paul L. Sorgen. Characterization of the ph-dependent inter- action between the gap junction protein connexin43 carboxyl terminus and cytoplasmic loop domains. J Biol Chem, 282(8):5801–5813, Feb 2007. [36] C. Peracchia. Chemical gating of gap junction channels; roles of cal- cium, ph and calmodulin. Biochim Biophys Acta, 1662(1-2):61–80, Mar 23 2004. [37] Miduturu Srinivas. Pharmacology of connexin channels. In Andrew L. Harris and Darren Locke, editors, Connexins, pages 207–224. Humana Press, 2009. [38] Thaddeus Bargiello and Peter Brink. Voltage-gating mechanisms of connexin channels. In Andrew L. Harris and Darren Locke, editors, Connexins, pages 103–128. Humana Press, 2009. [39] L. C. Barrio, T. Suchyna, T. Bargiello, L. X. Xu, R. S. Roginski, M. V. Bennett, and B. J. Nicholson. Gap junctions formed by connexins 26 and 32 alone and in combination are differently affected by applied voltage. Proc Natl Acad Sci U S A, 88(19):8410–8414, Oct 1991. [40] C. Peracchia, X. G. Wang, and L. L. Peracchia. Is the chemical gate of connexins voltage sensitive? behavior of cx32 wild-type and mutant channels. Am J Physiol, 276(6 Pt 1):C1361–C1373, Jun 1999. 158 [41] Daniel Gonzlez, Juan M. Gmez-Hernndez, and Luis C. Barrio. Molec- ular basis of voltage dependence of connexin channels: an integrative appraisal. Prog Biophys Mol Biol, 94(1-2):66–106, 2007. [42] F. F. Bukauskas, C. Elfgang, K. Willecke, and R. Weingart. Biophysi- cal properties of gap junction channels formed by mouse connexin40 in induced pairs of transfected human hela cells. Biophys J, 68(6):2289– 2298, Jun 1995. [43] A. Revilla, C. Castro, and L. C. Barrio. Molecular dissection of tran- sjunctional voltage dependence in the connexin-32 and connexin-43 junctions. Biophys J, 77(3):1374–1383, Sep 1999. [44] Feliksas F. Bukauskas and Vytas K. Verselis. Gap junction channel gating. Biochim Biophys Acta, 1662(1-2):42–60, Mar 2004. [45] S. Oh, J. B. Rubin, M. V. Bennett, V. K. Verselis, and T. A. Bargiello. Molecular determinants of electrical rectification of single channel con- ductance in gap junctions formed by connexins 26 and 32. J Gen Physiol, 114(3):339–364, Sep 1999. [46] Vytas K. Verselis, Maria P. Trelles, Clio Rubinos, Thaddeus A. Bargiello, and Miduturu Srinivas. Loop gating of connexin hemichan- nels involves movement of pore-lining residues in the first extracellular loop domain. J Biol Chem, 284(7):4484–4493, Feb 2009. [47] E. B. Trexler, M. V. Bennett, T. A. Bargiello, and V. K. Verselis. Voltage gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci U S A, 93(12):5836–5841, Jun 1996. [48] D. L. Paul, L. Ebihara, L. J. Takemoto, K. I. Swenson, and D. A. Goodenough. Connexin46, a novel lens gap junction protein, induces voltage-gated currents in nonjunctional plasma membrane of xenopus oocytes. J Cell Biol, 115(4):1077–1089, Nov 1991. [49] L. Ebihara and E. Steiner. Properties of a nonjunctional current ex- pressed from a rat connexin46 cdna in xenopus oocytes. J Gen Physiol, 102(1):59–74, Jul 1993. [50] Daniel J. Mller, Galen M. Hand, Andreas Engel, and Gina E. Sosinsky. Conformational changes in surface structures of isolated connexin 26 gap junctions. EMBO J, 21(14):3598–3607, Jul 2002. 159 [51] Juan C. Sez, Mauricio A. Retamal, Daniel Basilio, Feliksas F. Bukauskas, and Michael V L. Bennett. Connexin-based gap junc- tion hemichannels: gating mechanisms. Biochim Biophys Acta, 1711(2):215–224, Jun 2005. [52] Juan C. Sez, Kurt A. Schalper, Mauricio A. Retamal, Juan A. Orel- lana, Kenji F. Shoji, and Michael V L. Bennett. Cell membrane per- meabilization via connexin hemichannels in living and dying cells. Exp Cell Res, 316(15):2377–2389, Sep 2010. [53] Michael V L. Bennett, Jorge E. Contreras, Feliksas F. Bukauskas, and Juan C. Sez. New roles for astrocytes: gap junction hemichannels have something to communicate. Trends Neurosci, 26(11):610–617, Nov 2003. [54] K E Ludwig Scheckenbach, Sophie Crespin, Brenda R. Kwak, and Marc Chanson. Connexin channel-dependent signaling pathways in inflammation. J Vasc Res, 48(2):91–103, 2011. [55] Mathieu Vinken, Elke Decrock, Elke De Vuyst, Raf Ponsaerts, Cathe- leyne D’hondt, Geert Bultynck, Liesbeth Ceelen, Tamara Vanhaecke, Luc Leybaert, and Vera Rogiers. Connexins: sensors and regulators of cell cycling. Biochim Biophys Acta, 1815(1):13–25, Jan 2011. [56] Teresita Bellido and Lilian I. Plotkin. Novel actions of bisphosphonates in bone: preservation of osteoblast and osteocyte viability. Bone, 49(1):50–55, Jul 2011. [57] Fiona Hanner, Charlotte Mehlin Sorensen, Niels-Henrik Holstein- Rathlou, and Jnos Peti-Peterdi. Connexins and the kidney. Am J Physiol Regul Integr Comp Physiol, 298(5):R1143–R1155, May 2010. [58] Tetsuji Miura, Takayuki Miki, and Toshiyuki Yano. Role of the gap junction in ischemic preconditioning in the heart. Am J Physiol Heart Circ Physiol, 298(4):H1115–H1125, Apr 2010. [59] Roger J. Thompson and Brian A. Macvicar. Connexin and pannexin hemichannels of neurons and astrocytes. Channels (Austin), 2(2):81– 86, 2008. [60] D. C. Spray, Z. C. Ye, and B. R. Ransom. Functional connexin ”hemichannels”: a critical appraisal. Glia, 54(7):758–73, Nov 15 2006. 160 [61] J. Wang, M. Ma, S. Locovei, R. W. Keane, and G. Dahl. Modulation of membrane channel currents by gap junction protein mimetic peptides: size matters. Am J Physiol Cell Physiol, 293(3):C1112–9, Sep 2007. [62] D. M. Wilson. The connections between the lateral giant fibers of earthworms. Comp Biochem Physiol, 3:274–284, Nov 1961. [63] R. Eckert. Electrical interaction of paired ganglion cells in the leech. J Gen Physiol, 46(3):573–587, Jan 1963. [64] R. D. Penn. Ionic communication between liver cells. J Cell Biol, 29(1):171–174, Apr 1966. [65] N. M. Kumar and N. B. Gilula. Cloning and characterization of human and rat liver cdnas coding for a gap junction protein. J Cell Biol, 103(3):767–76, Sep 1986. [66] D. L. Paul. Molecular cloning of cdna for rat liver gap junction protein. J Cell Biol, 103(1):123–34, Jul 1986. [67] T. M. Barnes. Opus: a growing family of gap junction proteins? Trends Genet, 10(9):303–5, Sep 1994. [68] P. Phelan, J. P. Bacon, J. A. Davies, L. A. Stebbings, M. G. Todman, L. Avery, R. A. Baines, T. M. Barnes, C. Ford, S. Hekimi, R. Lee, J. E. Shaw, T. A. Starich, K. D. Curtin, Y. A. Sun, and R. J. Wyman. In- nexins: a family of invertebrate gap-junction proteins. Trends Genet, 14(9):348–9, Sep 1998. [69] P. Phelan, L. A. Stebbings, R. A. Baines, J. P. Bacon, J. A. Davies, and C. Ford. Drosophila shaking-b protein forms gap junctions in paired xenopus oocytes. Nature, 391(6663):181–184, Jan 1998. [70] M. R. Yen and Jr. Saier, M. H. Gap junctional proteins of animals: the innexin/pannexin superfamily. Prog Biophys Mol Biol, 94(1-2):5–14, May-Jun 2007. [71] Lucy A. Stebbings, Martin G. Todman, Rose Phillips, Claire E. Greer, Jennifer Tam, Pauline Phelan, Kirsten Jacobs, Jonathan P. Bacon, and Jane A. Davies. Gap junctions in drosophila: developmental ex- pression of the entire innexin gene family. Mech Dev, 113(2):197–205, May 2002. 161 [72] Reinhard Bauer, Birgit Ler, Katinka Ostrowski, Julia Martini, Andy Weimbs, Hildegard Lechner, and Michael Hoch. Intercellular commu- nication: the drosophila innexin multiprotein family of gap junction proteins. Chem Biol, 12(5):515–526, May 2005. [73] Z. F. Altun, B. Chen, Z. W. Wang, and D. H. Hall. High resolu- tion map of caenorhabditis elegans gap junction proteins. Dev Dyn, 238(8):1936–50, Aug 2009. [74] Brandon Kandarian, Jasmine Sethi, Allan Wu, Michael Baker, Neema Yazdani, Eunice Kym, Alejandro Sanchez, Lee Edsall, Terry Gaaster- land, and Eduardo Macagno. The medicinal leech genome encodes 21 innexin genes: different combinations are expressed by identified central neurons. Dev Genes Evol, 222(1):29–44, Mar 2012. [75] S. Bunse, M. Schmidt, S. Hoffmann, K. Engelhardt, G. Zoidl, and R. Dermietzel. Single cysteines in the extracellular and transmem- brane regions modulate pannexin 1 channel function. J Membr Biol, 244(1):21–33, Nov 2011. [76] Salli I. Tazuke, Cordula Schulz, Lilach Gilboa, Mignon Fogarty, An- thony P. Mahowald, Antoine Guichet, Anne Ephrussi, Cricket G. Wood, Ruth Lehmann, and Margaret T. Fuller. A germline-specific gap junction protein required for survival of differentiating early germ cells. Development, 129(10):2529–2539, May 2002. [77] Corinna Lehmann, Hildegard Lechner, Birgit Ler, Martin Knieps, Sonja Herrmann, Michael Famulok, Reinhard Bauer, and Michael Hoch. Heteromerization of innexin gap junction proteins regulates epithelial tissue organization in drosophila. Mol Biol Cell, 17(4):1676– 1685, Apr 2006. [78] L. A. Stebbings, M. G. Todman, P. Phelan, J. P. Bacon, and J. A. Davies. Two drosophila innexins are expressed in overlapping do- mains and cooperate to form gap-junction channels. Mol Biol Cell, 11(7):2459–2470, Jul 2000. [79] Chia-Lin Wu, Meng-Fu Maxwell Shih, Jason Sih-Yu Lai, Hsun-Ti Yang, Glenn C. Turner, Linyi Chen, and Ann-Shyn Chiang. Het- erotypic gap junctions between two neurons in the drosophila brain are critical for memory. Curr Biol, 21(10):848–854, May 2011. 162 [80] Y. Landesman, T. W. White, T. A. Starich, J. E. Shaw, D. A. Good- enough, and D. L. Paul. Innexin-3 forms connexin-like intercellular channels. J Cell Sci, 112 ( Pt 14):2391–2396, Jul 1999. [81] L. Bao, S. Samuels, S. Locovei, E. R. Macagno, K. J. Muller, and G. Dahl. Innexins form two types of channels. FEBS Lett, 581(29):5703–8, Dec 11 2007. [82] Adam Depriest, Pauline Phelan, and I. Martha Skerrett. Tryptophan scanning mutagenesis of the first transmembrane domain of the in- nexin shaking-b(lethal). Biophys J, 101(10):2408–2416, Nov 2011. [83] Iain M. Dykes, Fiona M. Freeman, Jonathan P. Bacon, and Jane A. Davies. Molecular basis of gap junctional communication in the cns of the leech hirudo medicinalis. J Neurosci, 24(4):886–894, Jan 2004. [84] J. C. Sez, V. M. Berthoud, M. C. Branes, A. D. Martinez, and E. C. Beyer. Plasma membrane channels formed by connexins: their regu- lation and functions. Physiol Rev, 83(4):1359–400, Oct 2003. [85] Katinka Ostrowski, Reinhard Bauer, and Michael Hoch. The drosophila innexin 7 gap junction protein is required for development of the embryonic nervous system. Cell Commun Adhes, 15(1):155–167, May 2008. [86] T. A. Starich, R. Y. Lee, C. Panzarella, L. Avery, and J. E. Shaw. eat-5 and unc-7 represent a multigene family in caenorhabditis elegans involved in cell-cell coupling. J Cell Biol, 134(2):537–548, Jul 1996. [87] T. A. Starich, R. K. Herman, and J. E. Shaw. Molecular and genetic analysis of unc-7, a caenorhabditis elegans gene required for coordi- nated locomotion. Genetics, 133(3):527–541, Mar 1993. [88] T. M. Barnes and S. Hekimi. The caenorhabditis elegans avermectin resistance and anesthetic response gene unc-9 encodes a member of a protein family implicated in electrical coupling of excitable cells. J Neurochem, 69(6):2251–2260, Dec 1997. [89] S. B. Hedges, J. Dudley, and S. Kumar. Timetree: a public knowledge-base of divergence times among organisms. Bioinformatics, 22(23):2971–2, Dec 1 2006. 163 [90] Y. Panchin, I. Kelmanson, M. Matz, K. Lukyanov, N. Usman, and S. Lukyanov. A ubiquitous family of putative gap junction molecules. Curr Biol, 10(13):R473–4, Jun 29 2000. [91] A. Baranova, D. Ivanov, N. Petrash, A. Pestova, M. Skoblov, I. Kel- manson, D. Shagin, S. Nazarenko, E. Geraymovych, O. Litvin, A. Tiunova, T. L. Born, N. Usman, D. Staroverov, S. Lukyanov, and Y. Panchin. The mammalian pannexin family is homologous to the invertebrate innexin gap junction proteins. Genomics, 83(4):706–16, Apr 2004. [92] P. Phelan. Innexins: members of an evolutionarily conserved family of gap-junction proteins. Biochim Biophys Acta, 1711(2):225–45, Jun 10 2005. [93] D. Fushiki, Y. Hamada, R. Yoshimura, and Y. Endo. Phylogenetic and bioinformatic analysis of gap junction-related proteins, innexins, pannexins and connexins. Biomed Res, 31(2):133–42, Apr 2010. [94] Federico Abascal and Rafael Zardoya. Lrrc8 proteins share a common ancestor with pannexins, and may form hexameric channels involved in cell-cell communication. Bioessays, 37(7):551–60, Jul 2012. [95] S. Li, M. Tomic, and S. S. Stojilkovic. Characterization of novel pan- nexin 1 isoforms from rat pituitary cells and their association with atp-gated p2x channels. Gen Comp Endocrinol, 174(2):202–10, Nov 1 2011. [96] P. Turmel, J. Dufresne, L. Hermo, C. E. Smith, S. Penuela, D. W. Laird, and D. G. Cyr. Characterization of pannexin1 and pannexin3 and their regulation by androgens in the male reproductive tract of the adult rat. Mol Reprod Dev, 78(2):124–38, Feb 2011. [97] G. Zoidl, M. Kremer, C. Zoidl, S. Bunse, and R. Dermietzel. Molecular diversity of connexin and pannexin genes in the retina of the zebrafish danio rerio. Cell Commun Adhes, 15(1):169–83, May 2008. [98] J. Wang and G. Dahl. Scam analysis of panx1 suggests a peculiar pore structure. J Gen Physiol, 136(5):515–27, Nov 2010. [99] C. Ambrosi, O. Gassmann, J. N. Pranskevich, D. Boassa, A. Smock, J. Wang, G. Dahl, C. Steinem, and G. E. Sosinsky. Pannexin1 and pannexin2 channels show quaternary similarities to connexons and 164 different oligomerization numbers from each other. J Biol Chem, 285(32):24420–31, Aug 6 2010. [100] S. Bunse, M. Schmidt, N. Prochnow, G. Zoidl, and R. Dermietzel. Intracellular cysteine 346 is essentially involved in regulating panx1 channel activity. J Biol Chem, 285(49):38444–52, Dec 3 2010. [101] F. Qiu, J. Wang, and G. Dahl. Alanine substitution scanning of pan- nexin1 reveals amino acid residues mediating atp sensitivity. Puriner- gic Signal, 8(1):81–90, Mar 2012. [102] R. Bruzzone, S. G. Hormuzdi, M. T. Barbe, A. Herb, and H. Monyer. Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci U S A, 100(23):13644–9, Nov 11 2003. [103] D. Boassa, C. Ambrosi, F. Qiu, G. Dahl, G. Gaietta, and G. Sosinsky. Pannexin1 channels contain a glycosylation site that targets the hex- amer to the plasma membrane. J Biol Chem, 282(43):31733–43, Oct 26 2007. [104] S. Penuela, R. Bhalla, X. Q. Gong, K. N. Cowan, S. J. Celetti, B. J. Cowan, D. Bai, Q. Shao, and D. W. Laird. Pannexin 1 and pannexin 3 are glycoproteins that exhibit many distinct characteristics from the connexin family of gap junction proteins. J Cell Sci, 120(Pt 21):3772– 83, Nov 1 2007. [105] S. J. Celetti, K. N. Cowan, S. Penuela, Q. Shao, J. Churko, and D. W. Laird. Implications of pannexin 1 and pannexin 3 for keratinocyte differentiation. J Cell Sci, 123(Pt 8):1363–1372, Apr 15 2010. [106] R. Bruzzone, M. T. Barbe, N. J. Jakob, and H. Monyer. Pharmacolog- ical properties of homomeric and heteromeric pannexin hemichannels expressed in xenopus oocytes. J Neurochem, 92(5):1033–43, Mar 2005. [107] S. Penuela, R. Bhalla, K. Nag, and D. W. Laird. Glycosylation reg- ulates pannexin intermixing and cellular localization. Mol Biol Cell, 20(20):4313–4323, Oct 2009. [108] Joell L. Solan and Paul D. Lampe. Connexin43 phosphorylation: structural changes and biological effects. Biochem J, 419(2):261–272, Apr 2009. 165 [109] G. Dvoriantchikova, D. Ivanov, Y. Panchin, and V. I. Shestopalov. Expression of pannexin family of proteins in the retina. FEBS Lett, 580(9):2178–82, Apr 17 2006. [110] N. Prochnow, S. Hoffmann, R. Vroman, J. Klooster, S. Bunse, M. Kamermans, R. Dermietzel, and G. Zoidl. Pannexin1 in the outer retina of the zebrafish, danio rerio. Neuroscience, 162(4):1039–54, Sep 15 2009. [111] R. Bhalla-Gehi, S. Penuela, J. M. Churko, Q. Shao, and D. W. Laird. Pannexin1 and pannexin3 delivery, cell surface dynamics, and cy- toskeletal interactions. J Biol Chem, 285(12):9147–60, Mar 19 2010. [112] L. A. Swayne, C. D. Sorbara, and S. A. Bennett. Pannexin 2 is ex- pressed by postnatal hippocampal neural progenitors and modulates neuronal commitment. J Biol Chem, 285(32):24977–86, Aug 6 2010. [113] P. Pelegrin and A. Surprenant. Pannexin-1 mediates large pore for- mation and interleukin-1beta release by the atp-gated p2x7 receptor. Embo J, 25(21):5071–82, Nov 1 2006. [114] W. R. Silverman, J. P. de Rivero Vaccari, S. Locovei, F. Qiu, S. K. Carlsson, E. Scemes, R. W. Keane, and G. Dahl. The pannexin 1 channel activates the inflammasome in neurons and astrocytes. J Biol Chem, 284(27):18143–51, Jul 3 2009. [115] V. Poornima, M. Madhupriya, S. Kootar, G. Sujatha, A. Kumar, and A. K. Bera. P2x7 receptor-pannexin 1 hemichannel association: ef- fect of extracellular calcium on membrane permeabilization. J Mol Neurosci, 46(3):585–94, Mar 2012. [116] S. Buvinic, G. Almarza, M. Bustamante, M. Casas, J. Lopez, M. Riquelme, J. C. Saez, J. P. Huidobro-Toro, and E. Jaimovich. Atp released by electrical stimuli elicits calcium transients and gene expression in skeletal muscle. J Biol Chem, 284(50):34490–505, Dec 11 2009. [117] S. Bunse, A. Haghika, G. Zoidl, and R. Dermietzel. Identification of a potential regulator of the gap junction protein pannexin1. Cell Commun Adhes, 12(5-6):231–6, Jul-Dec 2005. [118] S. Bunse, S. Locovei, M. Schmidt, F. Qiu, G. Zoidl, G. Dahl, and R. Dermietzel. The potassium channel subunit kvbeta3 interacts with 166 pannexin 1 and attenuates its sensitivity to changes in redox poten- tials. FEBS J, 276(21):6258–70, Sep 24 2009. [119] M. Billaud, A. W. Lohman, A. C. Straub, R. Looft-Wilson, S. R. Johnstone, C. A. Araj, A. K. Best, F. Chekeni, K. Ravichandran, S. Penuela, D. W. Laird, and B. E. Isakson. Pannexin1 regu- lates alpha1-adrenergic receptor-mediated vasoconstriction. Circ Res, 109(1):80–5, Jun 24 2011. [120] Haiying Zhan, Craig S. Moore, Bojun Chen, Xin Zhou, Xin-Ming Ma, Kumiko Ijichi, Michael V L. Bennett, Xue-Jun Li, Stephen J. Crocker, and Zhao-Wen Wang. Stomatin inhibits pannexin-1-mediated whole- cell currents by interacting with its carboxyl terminal. PLoS One, 7(6):e39489, 2012. [121] D. Boassa, F. Qiu, G. Dahl, and G. Sosinsky. Trafficking dynamics of glycosylated pannexin 1 proteins. Cell Commun Adhes, 15(1):119–32, May 2008. [122] C. P. Lai, J. F. Bechberger, R. J. Thompson, B. A. Macvicar, R. Bruz- zone, and C. C. Naus. Tumor-suppressive effects of pannexin 1 in c6 glioma cells. Cancer Res, 67(4):1545–1554, Feb 15 2007. [123] F. Vanden Abeele, G. Bidaux, D. Gordienko, B. Beck, Y. V. Panchin, A. V. Baranova, D. V. Ivanov, R. Skryma, and N. Prevarskaya. Func- tional implications of calcium permeability of the channel formed by pannexin 1. J Cell Biol, 174(4):535–46, Aug 14 2006. [124] Rodolfo M. Iglesias and David C. Spray. Pannexin1-mediated atp release provides signal transmission between neuro2a cells. Neurochem Res, 37(6):1355–1363, Jun 2012. [125] M. Ishikawa, T. Iwamoto, T. Nakamura, A. Doyle, S. Fukumoto, and Y. Yamada. Pannexin 3 functions as an er ca2+ channel, hemichannel, and gap junction to promote osteoblast differentiation. J Cell Biol, 193(7):1257–74, Jun 27 2011. [126] Y. J. Huang, Y. Maruyama, G. Dvoryanchikov, E. Pereira, N. Chaud- hari, and S. D. Roper. The role of pannexin 1 hemichannels in atp release and cell-cell communication in mouse taste buds. Proc Natl Acad Sci U S A, 104(15):6436–41, Apr 10 2007. 167 [127] Gerhard Dahl and Andrew L. Harris. Pannexins or connexins? In Andrew L. Harris and Darren Locke, editors, Connexins, pages 287– 301. Humana Press, 2009. [128] G. E. Sosinsky, D. Boassa, R. Dermietzel, H. S. Duffy, D. W. Laird, B. Macvicar, C. C. Naus, S. Penuela, E. Scemes, D. C. Spray, R. J. Thompson, H. B. Zhao, and G. Dahl. Pannexin channels are not gap junction hemichannels. Channels (Austin), 5(3):193–7, May 1 2011. [129] Silvia Penuela, Ruchi Gehi, and Dale W. Laird. The biochemistry and function of pannexin channels. Biochim Biophys Acta, 1828(1):15–22, Jan 2013. [130] G. Dahl and S. Locovei. Pannexin: to gap or not to gap, is that a question? IUBMB Life, 58(7):409–19, Jul 2006. [131] L. Bao, S. Locovei, and G. Dahl. Pannexin membrane channels are mechanosensitive conduits for atp. FEBS Lett, 572(1-3):65–8, Aug 13 2004. [132] S. Locovei, L. Bao, and G. Dahl. Pannexin 1 in erythrocytes: function without a gap. Proc Natl Acad Sci U S A, 103(20):7655–9, May 16 2006. [133] N. Prochnow, S. Hoffmann, R. Dermietzel, and G. Zoidl. Replacement of a single cysteine in the fourth transmembrane region of zebrafish pannexin 1 alters hemichannel gating behavior. Exp Brain Res, 199(3- 4):255–64, Dec 2009. [134] R. Iglesias, D. C. Spray, and E. Scemes. Mefloquine blockade of pan- nexin1 currents: Resolution of a conflict. Cell Commun Adhes, 16(5- 6):131–7, Dec 2010. [135] W. Ma, V. Compan, W. Zheng, E. Martin, R. A. North, A. Verkhratsky, and A. Surprenant. Pannexin 1 forms an anion- selective channel. Pflugers Arch, 463(4):585–92, Apr 2012. [136] J. S. Davidson, I. M. Baumgarten, and E. H. Harley. Reversible inhibi- tion of intercellular junctional communication by glycyrrhetinic acid. Biochem Biophys Res Commun, 134(1):29–36, Jan 1986. [137] W. Ma, H. Hui, P. Pelegrin, and A. Surprenant. Pharmacological characterization of pannexin-1 currents expressed in mammalian cells. J Pharmacol Exp Ther, 328(2):409–18, Feb 2009. 168 [138] A. Li, C. T. Leung, K. Peterson-Yantorno, W. D. Stamer, C. H. Mitchell, and M. M. Civan. Mechanisms of atp release by human trabecular meshwork cells, the enabling step in purinergic regulation of aqueous humor outflow. J Cell Physiol, 227(1):172–82, Jan 2012. [139] Min Zhang, Nikol A. Piskuric, Cathy Vollmer, and Colin A. Nurse. P2y2 receptor activation opens pannexin-1 channels in rat carotid body type ii cells: potential role in amplifying the neurotransmitter atp. J Physiol, 590(Pt 17):4335–4350, Sep 2012. [140] Scott J. Cruikshank, Matthew Hopperstad, Meg Younger, Barry W. Connors, David C. Spray, and Miduturu Srinivas. Potent block of cx36 and cx50 gap junction channels by mefloquine. Proc Natl Acad Sci U S A, 101(33):12364–12369, Aug 2004. [141] G. A. Ransford, N. Fregien, F. Qiu, G. Dahl, G. E. Conner, and M. Salathe. Pannexin 1 contributes to atp release in airway epithelia. Am J Respir Cell Mol Biol, 41(5):525–34, Nov 2009. [142] Mohamed Lamkanfi, R K Subbarao Malireddi, and Thirumala-Devi Kanneganti. Fungal zymosan and mannan activate the cryopyrin in- flammasome. J Biol Chem, 284(31):20574–20581, Jul 2009. [143] R. Dando and S. D. Roper. Cell-to-cell communication in intact taste buds through atp signalling from pannexin 1 gap junction hemichan- nels. J Physiol, 587(Pt 24):5899–906, Dec 15 2009. [144] Jingsheng Xia, Jason C. Lim, Wennan Lu, Jonathan M. Beckel, Ed- ward J. Macarak, Alan M. Laties, and Claire H. Mitchell. Neurons respond directly to mechanical deformation with pannexin-mediated atp release and autostimulation of p2x7 receptors. J Physiol, 590(Pt 10):2285–304, May 2012. [145] W. Silverman, S. Locovei, and G. Dahl. Probenecid, a gout remedy, inhibits pannexin 1 channels. Am J Physiol Cell Physiol, 295(3):C761– 7, Sep 2008. [146] P. Pelegrin and A. Surprenant. Pannexin-1 couples to maitotoxin- and nigericin-induced interleukin-1beta release through a dye uptake- independent pathway. J Biol Chem, 282(4):2386–94, Jan 26 2007. [147] J. P. Reyes, C. Y. Hernandez-Carballo, G. Perez-Flores, P. Perez- Cornejo, and J. Arreola. Lack of coupling between membrane stretch- 169 ing and pannexin-1 hemichannels. Biochem Biophys Res Commun, 380(1):50–53, Feb 27 2009. [148] N. Montalbetti, M. F. Leal Denis, O. P. Pignataro, E. Kobatake, E. R. Lazarowski, and P. J. Schwarzbaum. Homeostasis of extracellular atp in human erythrocytes. J Biol Chem, 286(44):38397–407, Nov 4 2011. [149] G. Dahl. Gap junction-mimetic peptides do work, but in unexpected ways. Cell Commun Adhes, 14(6):259–64, Nov-Dec 2007. [150] F. Qiu, J. Wang, D. C. Spray, E. Scemes, and G. Dahl. Two non- vesicular atp release pathways in the mouse erythrocyte membrane. FEBS Lett, 585(21):3430–5, Nov 4 2011. [151] S. Locovei, J. Wang, and G. Dahl. Activation of pannexin 1 channels by atp through p2y receptors and by cytoplasmic calcium. FEBS Lett, 580(1):239–44, Jan 9 2006. [152] A. M. Forsyth, J. Wan, P. D. Owrutsky, M. Abkarian, and H. A. Stone. Multiscale approach to link red blood cell dynamics, shear viscosity, and atp release. Proc Natl Acad Sci U S A, 108(27):10986–91, Jul 5 2011. [153] A. Li, C. T. Leung, K. Peterson-Yantorno, C. H. Mitchell, and M. M. Civan. Pathways for atp release by bovine ciliary epithelial cells, the initial step in purinergic regulation of aqueous humor inflow. Am J Physiol Cell Physiol, 299(6):C1308–17, Dec 2010. [154] S. Li, I. Bjelobaba, Z. Yan, M. Kucka, M. Tomic, and S. S. Stojilkovic. Expression and roles of pannexins in atp release in the pituitary gland. Endocrinology, 152(6):2342–52, Jun 2011. [155] H. T. Liu, A. H. Toychiev, N. Takahashi, R. Z. Sabirov, and Y. Okada. Maxi-anion channel as a candidate pathway for osmosensitive atp re- lease from mouse astrocytes in primary culture. Cell Res, 18(5):558– 65, May 2008. [156] M. F. Santiago, J. Veliskova, N. K. Patel, S. E. Lutz, D. Caille, A. Charollais, P. Meda, and E. Scemes. Targeting pannexin1 improves seizure outcome. PLoS One, 6(9):e25178, 2011. [157] Sylvia O. Suadicani, Rodolfo Iglesias, Junjie Wang, Gerhard Dahl, David C. Spray, and Eliana Scemes. Atp signaling is deficient in cul- tured pannexin1-null mouse astrocytes. Glia, 60(7):1106–16, Jul 2012. 170 [158] F. B. Chekeni, M. R. Elliott, J. K. Sandilos, S. F. Walk, J. M. Kinchen, E. R. Lazarowski, A. J. Armstrong, S. Penuela, D. W. Laird, G. S. Salvesen, B. E. Isakson, D. A. Bayliss, and K. S. Ravichandran. Pan- nexin 1 channels mediate ’find-me’ signal release and membrane per- meability during apoptosis. Nature, 467(7317):863–7, Oct 14 2010. [159] Y. Qu, S. Misaghi, K. Newton, L. L. Gilmour, S. Louie, J. E. Cupp, G. R. Dubyak, D. Hackos, and V. M. Dixit. Pannexin-1 is required for atp release during apoptosis but not for inflammasome activation. J Immunol, 189(11):6553–61, Apr 20 2011. [160] Joanna K. Sandilos, Yu-Hsin H. Chiu, Faraaz B. Chekeni, Allison J. Armstrong, Scott F. Walk, Kodi S. Ravichandran, and Douglas A. Bayliss. Pannexin 1, an atp release channel, is activated by caspase cleavage of its pore-associated c terminal autoinhibitory region. J Biol Chem, 287(14):11303–11, Mar 2012. [161] M. C. Kienitz, K. Bender, R. Dermietzel, L. Pott, and G. Zoidl. Pan- nexin 1 constitutes the large conductance cation channel of cardiac myocytes. J Biol Chem, 286(1):290–8, Jan 7 2011. [162] S. Godecke, C. Roderigo, C. R. Rose, B. H. Rauch, A. Godecke, and J. Schrader. Thrombin-induced atp release from human umbilical vein endothelial cells. Am J Physiol Cell Physiol, 302(6):C915–23, Mar 2012. [163] Y. Sumi, T. Woehrle, Y. Chen, Y. Yao, A. Li, and W. G. Junger. Adrenergic receptor activation involves atp release and feedback through purinergic receptors. Am J Physiol Cell Physiol, 299(5):C1118–26, Nov 2010. [164] Ivar von Kgelgen. Pharmacological profiles of cloned mammalian p2y- receptor subtypes. Pharmacol Ther, 110(3):415–432, Jun 2006. [165] R. Iglesias, S. Locovei, A. Roque, A. P. Alberto, G. Dahl, D. C. Spray, and E. Scemes. P2x7 receptor-pannexin1 complex: pharmacology and signaling. Am J Physiol Cell Physiol, 295(3):C752–60, Sep 2008. [166] F. Qiu and G. P. Dahl. A permeant regulating its permeation pore: inhibition of pannexin 1 channels by atp. Am J Physiol Cell Physiol, 296(2):C250–255, Feb 2009. 171 [167] P. Bargiotas, A. Krenz, S. G. Hormuzdi, D. A. Ridder, A. Herb, W. Barakat, S. Penuela, J. von Engelhardt, H. Monyer, and M. Schwaninger. Pannexins in ischemia-induced neurodegeneration. Proc Natl Acad Sci U S A, 108(51):20772–7, Dec 20 2011. [168] I. Lemaire, S. Falzoni, B. Zhang, P. Pellegatti, and F. Di Virgilio. The p2x7 receptor and pannexin-1 are both required for the promotion of multinucleated macrophages by the inflammatory cytokine gm-csf. J Immunol, 187(7):3878–87, Oct 1 2011. [169] T. Iwamoto, T. Nakamura, A. Doyle, M. Ishikawa, S. de Vega, S. Fukumoto, and Y. Yamada. Pannexin 3 regulates intracellular atp/camp levels and promotes chondrocyte differentiation. J Biol Chem, 285(24):18948–58, Jun 11 2010. [170] S. Penuela, S. J. Celetti, R. Bhalla, Q. Shao, and D. W. Laird. Diverse subcellular distribution profiles of pannexin 1 and pannexin 3. Cell Commun Adhes, 15(1):133–42, May 2008. [171] G. Dvoriantchikova, D. Ivanov, A. Pestova, and V. Shestopalov. Molecular characterization of pannexins in the lens. Mol Vis, 12:1417– 26, 2006. [172] R. Gehi, Q. Shao, and D. W. Laird. Pathways regulating the traffick- ing and turnover of pannexin1 protein and the role of the c-terminal domain. J Biol Chem, 286(31):27639–53, Aug 5 2011. [173] A. Zappala, G. Li Volti, M. F. Serapide, R. Pellitteri, M. Falchi, F. La Delia, V. Cicirata, and F. Cicirata. Expression of pannexin2 protein in healthy and ischemized brain of adult rats. Neuroscience, 148(3):653–667, Sep 7 2007. [174] C. P. Lai, J. F. Bechberger, and C. C. Naus. Pannexin2 as a novel growth regulator in c6 glioma cells. Oncogene, 28(49):4402–4408, Dec 10 2009. [175] X. H. Wang, M. Streeter, Y. P. Liu, and H. B. Zhao. Identification and characterization of pannexin expression in the mammalian cochlea. J Comp Neurol, 512(3):336–46, Jan 20 2009. [176] A. Ray, G. Zoidl, S. Weickert, P. Wahle, and R. Dermietzel. Site- specific and developmental expression of pannexin1 in the mouse ner- vous system. Eur J Neurosci, 21(12):3277–90, Jun 2005. 172 [177] A. Vogt, S. G. Hormuzdi, and H. Monyer. Pannexin1 and pannexin2 expression in the developing and mature rat brain. Brain Res Mol Brain Res, 141(1):113–20, Nov 18 2005. [178] A. Zappala, D. Cicero, M. F. Serapide, C. Paz, M. V. Catania, M. Falchi, R. Parenti, M. R. Panto, F. La Delia, and F. Cicirata. Expression of pannexin1 in the cns of adult mouse: cellular local- ization and effect of 4-aminopyridine-induced seizures. Neuroscience, 141(1):167–78, Aug 11 2006. [179] G. Zoidl, E. Petrasch-Parwez, A. Ray, C. Meier, S. Bunse, H. W. Habbes, G. Dahl, and R. Dermietzel. Localization of the pannexin1 protein at postsynaptic sites in the cerebral cortex and hippocampus. Neuroscience, 146(1):9–16, Apr 25 2007. [180] S. Weickert, A. Ray, G. Zoidl, and R. Dermietzel. Expression of neural connexins and pannexin1 in the hippocampus and inferior olive: a quantitative approach. Brain Res Mol Brain Res, 133(1):102–9, Jan 5 2005. [181] S. Locovei, E. Scemes, F. Qiu, D. C. Spray, and G. Dahl. Pannexin1 is part of the pore forming unit of the p2x(7) receptor death complex. FEBS Lett, 581(3):483–8, Feb 6 2007. [182] R. Iglesias, G. Dahl, F. Qiu, D. C. Spray, and E. Scemes. Pannexin 1: the molecular substrate of astrocyte ”hemichannels”. J Neurosci, 29(21):7092–7, May 27 2009. [183] A. Ray, G. Zoidl, P. Wahle, and R. Dermietzel. Pannexin expression in the cerebellum. Cerebellum, 5(3):189–92, 2006. [184] V. M. Berthoud, P. J. Minogue, J. G. Laing, and E. C. Beyer. Path- ways for degradation of connexins and gap junctions. Cardiovasc Res, 62(2):256–67, May 1 2004. [185] M. Sridharan, S. P. Adderley, E. A. Bowles, T. M. Egan, A. H. Stephenson, M. L. Ellsworth, and R. S. Sprague. Pannexin 1 is the conduit for low oxygen tension-induced atp release from human ery- throcytes. Am J Physiol Heart Circ Physiol, 299(4):H1146–52, Oct 2010. [186] H. Jiang, A. G. Zhu, M. Mamczur, J. R. Falck, K. M. Lerea, and J. C. McGiff. Stimulation of rat erythrocyte p2x7 receptor induces 173 the release of epoxyeicosatrienoic acids. Br J Pharmacol, 151(7):1033– 1040, Aug 2007. [187] D. A. Vessey, L. Li, and M. Kelley. Pannexin-i/p2x 7 purinergic re- ceptor channels mediate the release of cardioprotectants induced by ischemic pre- and postconditioning. J Cardiovasc Pharmacol Ther, 15(2):190–5, Jun 2010. [188] D. A. Vessey, L. Li, and M. Kelley. Ischemic preconditioning re- quires opening of pannexin-1/p2x(7) channels not only during pre- conditioning but again after index ischemia at full reperfusion. Mol Cell Biochem, 351(1-2):77–84, May 2011. [189] D. A. Vessey, L. Li, and M. Kelley. P2x7 receptor agonists pre- and postcondition the heart against ischemia-reperfusion injury by opening pannexin-1/p2x channels. Am J Physiol Heart Circ Physiol, 301(3):H881–7, Sep 2011. [190] W. Tang, S. Ahmad, V. I. Shestopalov, and X. Lin. Pannexins are new molecular candidates for assembling gap junctions in the cochlea. Neuroreport, 19(13):1253–7, Aug 27 2008. [191] D. Reigada, W. Lu, M. Zhang, and C. H. Mitchell. Elevated pressure triggers a physiological release of atp from the retina: Possible role for pannexin hemichannels. Neuroscience, 157(2):396–404, Nov 19 2008. [192] T. D. Kanneganti, M. Lamkanfi, Y. G. Kim, G. Chen, J. H. Park, L. Franchi, P. Vandenabeele, and G. Nunez. Pannexin-1-mediated recognition of bacterial molecules activates the cryopyrin inflamma- some independent of toll-like receptor signaling. Immunity, 26(4):433– 43, Apr 2007. [193] N. Marina-Garcia, L. Franchi, Y. G. Kim, D. Miller, C. McDonald, G. J. Boons, and G. Nunez. Pannexin-1-mediated intracellular delivery of muramyl dipeptide induces caspase-1 activation via cryopyrin/nlrp3 independently of nod2. J Immunol, 180(6):4050–7, Mar 15 2008. [194] P. Pelegrin, C. Barroso-Gutierrez, and A. Surprenant. P2x7 receptor differentially couples to distinct release pathways for il-1beta in mouse macrophage. J Immunol, 180(11):7147–57, Jun 1 2008. [195] D. Brough, P. Pelegrin, and N. J. Rothwell. Pannexin-1-dependent caspase-1 activation and secretion of il-1beta is regulated by zinc. Eur J Immunol, 39(2):352–8, Feb 2009. 174 [196] M. Kronlage, J. Song, L. Sorokin, K. Isfort, T. Schwerdtle, J. Leipziger, B. Robaye, P. B. Conley, H. C. Kim, S. Sargin, P. Schon, A. Schwab, and P. J. Hanley. Autocrine purinergic receptor signaling is essential for macrophage chemotaxis. Sci Signal, 3(132):ra55, 2010. [197] P. J. Hanley, M. Kronlage, C. Kirschning, A. Del Rey, F. Di Virgilio, J. Leipziger, I. P. Chessell, S. Sargin, M. A. Filippov, O. Lindemann, S. Mohr, V. Konigs, H. Schillers, M. Bahler, and A. Schwab. Transient p2x7 receptor activation triggers macrophage death independent of toll-like receptors 2 and 4, caspase-1, and pannexin-1 proteins. J Biol Chem, 287(13):10650–63, Mar 23 2012. [198] U. Schenk, A. M. Westendorf, E. Radaelli, A. Casati, M. Ferro, M. Fu- magalli, C. Verderio, J. Buer, E. Scanziani, and F. Grassi. Puriner- gic control of t cell activation by atp released through pannexin-1 hemichannels. Sci Signal, 1(39):ra6, 2008. [199] T. Woehrle, L. Yip, A. Elkhal, Y. Sumi, Y. Chen, Y. Yao, P. A. Insel, and W. G. Junger. Pannexin-1 hemichannel-mediated atp release to- gether with p2x1 and p2x4 receptors regulate t-cell activation at the immune synapse. Blood, 116(18):3475–84, Nov 4 2010. [200] T. Woehrle, L. Yip, M. Manohar, Y. Sumi, Y. Yao, Y. Chen, and W. G. Junger. Hypertonic stress regulates t cell function via pannexin- 1 hemichannels and p2x receptors. J Leukoc Biol, 88(6):1181–9, Dec 2010. [201] Y. Chen, Y. Yao, Y. Sumi, A. Li, U. K. To, A. Elkhal, Y. Inoue, T. Woehrle, Q. Zhang, C. Hauser, and W. G. Junger. Purinergic signaling: a fundamental mechanism in neutrophil activation. Sci Signal, 3(125):ra45, 2010. [202] Chiara Rigato, Nina Swinnen, Roeland Buckinx, Isabelle Couillin, Jean-Marie Mangin, Jean-Michel Rigo, Pascal Legendre, and Herv Le Corronc. Microglia proliferation is controlled by p2x7 receptors in a pannexin-1-independent manner during early embryonic spinal cord invasion. The Journal of Neuroscience, 32(34):11559–11573, 2012. [203] L. Seminario-Vidal, S. F. Okada, J. I. Sesma, S. M. Kreda, C. A. van Heusden, Y. Zhu, L. C. Jones, W. K. O’Neal, S. Penuela, D. W. Laird, R. C. Boucher, and E. R. Lazarowski. Rho signaling regulates pannexin 1-mediated atp release from airway epithelia. J Biol Chem, 286(30):26277–86, Jul 29 2011. 175 [204] Luis A. Cea, Manuel A. Riquelme, Bruno A. Cisterna, Carlos Puebla, Jos L. Vega, Maximiliano Rovegno, and Juan C. Sez. Connexin- and pannexin-based channels in normal skeletal muscles and their possible role in muscle atrophy. J Membr Biol, 245(8):423–436, Aug 2012. [205] Yoshihiro Murata, Toshiaki Yasuo, Ryusuke Yoshida, Kunihiko Obata, Yuchio Yanagawa, Robert F. Margolskee, and Yuzo Ninomiya. Action potential-enhanced atp release from taste cells through hemichannels. J Neurophysiol, 104(2):896–901, Aug 2010. [206] R. A. Romanov, O. A. Rogachevskaja, M. F. Bystrova, P. Jiang, R. F. Margolskee, and S. S. Kolesnikov. Afferent neurotransmission medi- ated by hemichannels in mammalian taste cells. Embo J, 26(3):657–67, Feb 7 2007. [207] R. A. Romanov, O. A. Rogachevskaja, A. A. Khokhlov, and S. S. Kolesnikov. Voltage dependence of atp secretion in mammalian taste cells. J Gen Physiol, 132(6):731–44, Dec 2008. [208] F. Anselmi, V. H. Hernandez, G. Crispino, A. Seydel, S. Ortolano, S. D. Roper, N. Kessaris, W. Richardson, G. Rickheit, M. A. Filip- pov, H. Monyer, and F. Mammano. Atp release through connexin hemichannels and gap junction transfer of second messengers propa- gate ca2+ signals across the inner ear. Proc Natl Acad Sci U S A, 105(48):18770–5, Dec 2 2008. [209] Galina Dvoriantchikova, Dmitry Ivanov, David Barakat, Alexander Grinberg, Rong Wen, Vladlen Z. Slepak, and Valery I. Shestopalov. Genetic ablation of pannexin1 protects retinal neurons from ischemic injury. PLoS One, 7(2):e31991, 2012. [210] Roger J. Thompson, Ning Zhou, and Brian A. MacVicar. Ischemia opens neuronal gap junction hemichannels. Science, 312(5775):924– 927, May 2006. [211] M. Domercq, A. Perez-Samartin, D. Aparicio, E. Alberdi, O. Pam- pliega, and C. Matute. P2x7 receptors mediate ischemic damage to oligodendrocytes. Glia, 58(6):730–40, Apr 15 2010. [212] R. J. Thompson, M. F. Jackson, M. E. Olah, R. L. Rungta, D. J. Hines, M. A. Beazely, J. F. MacDonald, and B. A. MacVicar. Acti- vation of pannexin-1 hemichannels augments aberrant bursting in the hippocampus. Science, 322(5907):1555–9, Dec 5 2008. 176 [213] J. E. Kim and T. C. Kang. The p2x7 receptor-pannexin-1 complex decreases muscarinic acetylcholine receptor-mediated seizure suscepti- bility in mice. J Clin Invest, 121(5):2037–47, May 2 2011. [214] Mark S. Nash, Jonathon M. Willets, Brian Billups, R. A. John Chal- liss, and Stefan R. Nahorski. Synaptic activity augments muscarinic acetylcholine receptor-stimulated inositol 1,4,5-trisphosphate produc- tion to facilitate ca2+ release in hippocampal neurons. J Biol Chem, 279(47):49036–49044, Nov 2004. [215] Y. Fukushi. Heterologous desensitization of muscarinic receptors by p2z purinoceptors in rat parotid acinar cells. Eur J Pharmacol, 364(1):55–64, Jan 1999. [216] Flavio Frhlich, Maxim Bazhenov, Vicente Iragui-Madoz, and Ter- rence J. Sejnowski. Potassium dynamics in the epileptic cortex: new insights on an old topic. Neuroscientist, 14(5):422–433, Oct 2008. [217] M. A. Whittington, R. D. Traub, and J. G. Jefferys. Erosion of inhi- bition contributes to the progression of low magnesium bursts in rat hippocampal slices. J Physiol, 486 ( Pt 3):723–734, Aug 1995. [218] O. Litvin, A. Tiunova, Y. Connell-Alberts, Y. Panchin, and A. Bara- nova. What is hidden in the pannexin treasure trove: the sneak peek and the guesswork. J Cell Mol Med, 10(3):613–34, Jul-Sep 2006. [219] B. A. Bao, C. P. Lai, C. C. Naus, and J. R. Morgan. Pannexin1 drives multicellular aggregate compaction via a signaling cascade that remodels the actin cytoskeleton. J Biol Chem, 287(11):8407–16, Mar 9 2012. [220] M. Skals, N. R. Jorgensen, J. Leipziger, and H. A. Praetorius. Alpha- hemolysin from escherichia coli uses endogenous amplification through p2x receptor activation to induce hemolysis. Proc Natl Acad Sci U S A, 106(10):4030–5, Mar 10 2009. [221] M. Skals, J. Leipziger, and H. A. Praetorius. Haemolysis induced by alpha-toxin from staphylococcus aureus requires p2x receptor activa- tion. Pflugers Arch, 462(5):669–79, Nov 2011. [222] C. Seror, M. T. Melki, F. Subra, S. Q. Raza, M. Bras, H. Saidi, R. Nar- dacci, L. Voisin, A. Paoletti, F. Law, I. Martins, A. Amendola, A. A. Abdul-Sater, F. Ciccosanti, O. Delelis, F. Niedergang, S. Thierry, 177 N. Said-Sadier, C. Lamaze, D. Metivier, J. Estaquier, G. M. Fimia, L. Falasca, R. Casetti, N. Modjtahedi, J. Kanellopoulos, J. F. Mous- cadet, D. M. Ojcius, M. Piacentini, M. L. Gougeon, G. Kroemer, and J. L. Perfettini. Extracellular atp acts on p2y2 purinergic receptors to facilitate hiv-1 infection. J Exp Med, 208(9):1823–34, Aug 29 2011. [223] M. Solle, J. Labasi, D. G. Perregaux, E. Stam, N. Petrushova, B. H. Koller, R. J. Griffiths, and C. A. Gabel. Altered cytokine production in mice lacking p2x(7) receptors. J Biol Chem, 276(1):125–132, Jan 2001. [224] T. D. Hassinger, P. B. Guthrie, P. B. Atkinson, M. V. Bennett, and S. B. Kater. An extracellular signaling component in propagation of astrocytic calcium waves. Proc Natl Acad Sci U S A, 93(23):13268– 13273, Nov 1996. [225] P. B. Guthrie, J. Knappenberger, M. Segal, M. V. Bennett, A. C. Charles, and S. B. Kater. Atp released from astrocytes mediates glial calcium waves. J Neurosci, 19(2):520–528, Jan 1999. [226] Eliana Scemes and Christian Giaume. Astrocyte calcium waves: what they are and what they do. Glia, 54(7):716–725, Nov 2006. [227] Kazuhiko Kawasaki and Kenneth M. Weiss. Evolutionary genetics of vertebrate tissue mineralization: the origin and evolution of the secretory calcium-binding phosphoprotein family. J Exp Zool B Mol Dev Evol, 306(3):295–316, May 2006. [228] T. A. Franz-Odendaal, B. K. Hall, and P. E. Witten. Buried alive: how osteoblasts become osteocytes. Dev Dyn, 235(1):176–90, Jan 2006. [229] William R. Thompson, Clinton T. Rubin, and Janet Rubin. Mechan- ical regulation of signaling pathways in bone. Gene, 503(2):179–193, Jul 2012. [230] R. B. Widelitz, T. X. Jiang, B. A. Murray, and C. M. Chuong. Adhe- sion molecules in skeletogenesis: Ii. neural cell adhesion molecules me- diate precartilaginous mesenchymal condensations and enhance chon- drogenesis. J Cell Physiol, 156(2):399–411, Aug 1993. [231] S. A. Oberlender and R. S. Tuan. Expression and functional involve- ment of n-cadherin in embryonic limb chondrogenesis. Development, 120(1):177–187, Jan 1994. 178 [232] J. Chimal-Monroy and L. Daz de Len. Expression of n-cadherin, n- cam, fibronectin and tenascin is stimulated by tgf-beta1, beta2, beta3 and beta5 during the formation of precartilage condensations. Int J Dev Biol, 43(1):59–67, Jan 1999. [233] A. R. Haas and R. S. Tuan. Chondrogenic differentiation of murine c3h10t1/2 multipotential mesenchymal cells: Ii. stimulation by bone morphogenetic protein-2 requires modulation of n-cadherin expression and function. Differentiation, 64(2):77–89, Jan 1999. [234] B. K. Hall and T. Miyake. All for one and one for all: condensations and the initiation of skeletal development. Bioessays, 22(2):138–147, Feb 2000. [235] V. Lefebvre, W. Huang, V. R. Harley, P. N. Goodfellow, and B. de Crombrugghe. Sox9 is a potent activator of the chondrocyte- specific enhancer of the pro alpha1(ii) collagen gene. Mol Cell Biol, 17(4):2336–2346, Apr 1997. [236] L. C. Bridgewater, V. Lefebvre, and B. de Crombrugghe. Chondrocyte-specific enhancer elements in the col11a2 gene resemble the col2a1 tissue-specific enhancer. J Biol Chem, 273(24):14998–15006, Jun 1998. [237] I. Sekiya, K. Tsuji, P. Koopman, H. Watanabe, Y. Yamada, K. Shi- nomiya, A. Nifuji, and M. Noda. Sox9 enhances aggrecan gene pro- moter/enhancer activity and is up-regulated by retinoic acid in a cartilage-derived cell line, tc6. J Biol Chem, 275(15):10738–10744, Apr 2000. [238] Haruhiko Akiyama, Marie-Christine Chaboissier, James F. Martin, Andreas Schedl, and Benoit de Crombrugghe. The transcription fac- tor sox9 has essential roles in successive steps of the chondrocyte dif- ferentiation pathway and is required for expression of sox5 and sox6. Genes Dev, 16(21):2813–2828, Nov 2002. [239] P. Smits, P. Li, J. Mandel, Z. Zhang, J. M. Deng, R. R. Behringer, B. de Crombrugghe, and V. Lefebvre. The transcription factors l-sox5 and sox6 are essential for cartilage formation. Dev Cell, 1(2):277–90, Aug 2001. [240] B. Lanske, A. C. Karaplis, K. Lee, A. Luz, A. Vortkamp, A. Pirro, M. Karperien, L. H. Defize, C. Ho, R. C. Mulligan, A. B. Abou-Samra, 179 H. Jppner, G. V. Segre, and H. M. Kronenberg. Pth/pthrp receptor in early development and indian hedgehog-regulated bone growth. Sci- ence, 273(5275):663–666, Aug 1996. [241] H. M. Kronenberg. Developmental regulation of the growth plate. Nature, 423(6937):332–6, May 15 2003. [242] Peter Dy, Weihuan Wang, Pallavi Bhattaram, Qiuqing Wang, Lai Wang, R Tracy Ballock, and Vronique Lefebvre. Sox9 directs hy- pertrophic maturation and blocks osteoblast differentiation of growth plate chondrocytes. Dev Cell, 22(3):597–609, Mar 2012. [243] Yu-Feng Dong, Do Y. Soung, Edward M. Schwarz, Regis J. O’Keefe, and Hicham Drissi. Wnt induction of chondrocyte hypertrophy through the runx2 transcription factor. J Cell Physiol, 208(1):77–86, Jul 2006. [244] Toshihisa Komori. Regulation of bone development and extracellular matrix protein genes by runx2. Cell Tissue Res, 339(1):189–195, Jan 2010. [245] Muneaki Ishijima, Nobuharu Suzuki, Kentaro Hozumi, Tomoya Mat- sunobu, Keisuke Kosaki, Haruka Kaneko, John R. Hassell, Eri Arikawa-Hirasawa, and Yoshihiko Yamada. Perlecan modulates vegf signaling and is essential for vascularization in endochondral bone for- mation. Matrix Biol, 31(4):234–245, May 2012. [246] R. Civitelli. Cell-cell communication in the osteoblast/osteocyte lin- eage. Arch Biochem Biophys, 473(2):188–92, May 15 2008. [247] W. A. Paznekas, S. A. Boyadjiev, R. E. Shapiro, O. Daniels, B. Woll- nik, C. E. Keegan, J. W. Innis, M. B. Dinulos, C. Christian, M. C. Hannibal, and E. W. Jabs. Connexin 43 (gja1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet, 72(2):408–18, Feb 2003. [248] C. R. Green, L. Bowles, A. Crawley, and C. Tickle. Expression of the connexin43 gap junctional protein in tissues at the tip of the chick limb bud is related to the epithelial-mesenchymal interactions that mediate morphogenesis. Dev Biol, 161(1):12–21, Jan 1994. [249] R. A. Meyer, M. F. Cohen, S. Recalde, J. Zakany, S. M. Bell, WJ Scott, Jr, and C. W. Lo. Developmental regulation and asymmetric expres- 180 sion of the gene encoding cx43 gap junctions in the mouse limb bud. Dev Genet, 21(4):290–300, 1997. [250] D. L. Becker, I. McGonnell, H. P. Makarenkova, K. Patel, C. Tickle, J. Lorimer, and C. R. Green. Roles for alpha 1 connexin in morpho- genesis of chick embryos revealed using a novel antisense approach. Dev Genet, 24(1-2):33–42, 1999. [251] I. M. McGonnell, C. R. Green, C. Tickle, and D. L. Becker. Con- nexin43 gap junction protein plays an essential role in morphogenesis of the embryonic chick face. Dev Dyn, 222(3):420–438, Nov 2001. [252] F. Lecanda, P. M. Warlow, S. Sheikh, F. Furlan, T. H. Steinberg, and R. Civitelli. Connexin43 deficiency causes delayed ossification, craniofacial abnormalities, and osteoblast dysfunction. J Cell Biol, 151(4):931–44, Nov 13 2000. [253] A. G. Reaume, P. A. de Sousa, S. Kulkarni, B. L. Langille, D. Zhu, T. C. Davies, S. C. Juneja, G. M. Kidder, and J. Rossant. Car- diac malformation in neonatal mice lacking connexin43. Science, 267(5205):1831–1834, Mar 1995. [254] Marcus Watkins, Susan K. Grimston, Jin Yi Norris, Bertrand Guil- lotin, Angela Shaw, Elia Beniash, and Roberto Civitelli. Osteoblast connexin43 modulates skeletal architecture by regulating both arms of bone remodeling. Mol Biol Cell, 22(8):1240–1251, Apr 2011. [255] A. F. Taylor, M. M. Saunders, D. L. Shingle, J. M. Cimbala, Z. Zhou, and H. J. Donahue. Mechanically stimulated osteocytes regulate os- teoblastic activity via gap junctions. Am J Physiol Cell Physiol, 292(1):C545–C552, Jan 2007. [256] Yue Zhang, Emmanuel M. Paul, Vikram Sathyendra, Andrew Davi- son, Neil Sharkey, Sarah Bronson, Sundar Srinivasan, Ted S. Gross, and Henry J. Donahue. Enhanced osteoclastic resorption and respon- siveness to mechanical load in gap junction deficient bone. PLoS One, 6(8):e23516, 2011. [257] Z. Li, Z. Zhou, C. E. Yellowley, and H. J. Donahue. Inhibiting gap junctional intercellular communication alters expression of differenti- ation markers in osteoblastic cells. Bone, 25(6):661–6, Dec 1999. 181 [258] P. C. Schiller, G. D’Ippolito, W. Balkan, B. A. Roos, and G. A. Howard. Gap-junctional communication is required for the matura- tion process of osteoblastic cells in culture. Bone, 28(4):362–369, Apr 2001. [259] Alayna E. Loiselle, Emmanuel M. Paul, Gregory S. Lewis, and Henry J. Donahue. Osteoblast and osteocyte-specific loss of con- nexin43 results in delayed bone formation and healing during murine fracture healing. J Orthop Res, 31(1):147–154, Jan 2013. [260] Lidan You, Sara Temiyasathit, Peling Lee, Chi Hyun Kim, Padmaja Tummala, Wei Yao, Wade Kingery, Amanda M. Malone, Ronald Y. Kwon, and Christopher R. Jacobs. Osteocytes as mechanosensors in the inhibition of bone resorption due to mechanical loading. Bone, 42(1):172–179, Jan 2008. [261] Susan K. Grimston, Daniel B. Goldberg, Marcus Watkins, Michael D. Brodt, Matthew J. Silva, and Roberto Civitelli. Connexin43 deficiency reduces the sensitivity of cortical bone to the effects of muscle paral- ysis. J Bone Miner Res, 26(9):2151–2160, Sep 2011. [262] Shane A. Lloyd, Gregory S. Lewis, Yue Zhang, Emmanuel M. Paul, and Henry J. Donahue. Connexin 43 deficiency attenuates loss of trabecular bone and prevents suppression of cortical bone formation during unloading. J Bone Miner Res, 27(11):2359–2372, Nov 2012. [263] J. G. Laing, R. N. Manley-Markowski, M. Koval, R. Civitelli, and T. H. Steinberg. Connexin45 interacts with zonula occludens-1 and connexin43 in osteoblastic cells. J Biol Chem, 276(25):23051–5, Jun 22 2001. [264] M. Koval, S. T. Geist, E. M. Westphale, A. E. Kemendy, R. Civitelli, E. C. Beyer, and T. H. Steinberg. Transfected connexin45 alters gap junction permeability in cells expressing endogenous connexin43. J Cell Biol, 130(4):987–95, Aug 1995. [265] M. Koval, J. E. Harley, E. Hick, and T. H. Steinberg. Connexin46 is retained as monomers in a trans-golgi compartment of osteoblastic cells. J Cell Biol, 137(4):847–57, May 19 1997. [266] A. Pizard, P. G. Burgon, D. L. Paul, B. G. Bruneau, C. E. Seidman, and J. G. Seidman. Connexin 40, a target of transcription factor tbx5, 182 patterns wrist, digits, and sternum. Mol Cell Biol, 25(12):5073–83, Jun 2005. [267] Stephen R. Bond, Nan Wang, Luc Leybaert, and Christian C. Naus. Pannexin 1 ohnologs in the teleost lineage. J Membr Biol, 245(8):483– 493, Aug 2012. [268] P. Dehal and J. L. Boore. Two rounds of whole genome duplication in the ancestral vertebrate. PLoS Biol, 3(10):e314, Oct 2005. [269] N. H. Putnam, T. Butts, D. E. Ferrier, R. F. Furlong, U. Hellsten, T. Kawashima, M. Robinson-Rechavi, E. Shoguchi, A. Terry, J. K. Yu, E. L. Benito-Gutierrez, I. Dubchak, J. Garcia-Fernandez, J. J. Gibson- Brown, I. V. Grigoriev, A. C. Horton, P. J. de Jong, J. Jurka, V. V. Kapitonov, Y. Kohara, Y. Kuroki, E. Lindquist, S. Lucas, K. Osoe- gawa, L. A. Pennacchio, A. A. Salamov, Y. Satou, T. Sauka-Spengler, J. Schmutz, I. T. Shin, A. Toyoda, M. Bronner-Fraser, A. Fujiyama, L. Z. Holland, P. W. Holland, N. Satoh, and D. S. Rokhsar. The am- phioxus genome and the evolution of the chordate karyotype. Nature, 453(7198):1064–71, Jun 19 2008. [270] O. Jaillon, J. M. Aury, F. Brunet, J. L. Petit, N. Stange-Thomann, E. Mauceli, L. Bouneau, C. Fischer, C. Ozouf-Costaz, A. Bernot, S. Nicaud, D. Jaffe, S. Fisher, G. Lutfalla, C. Dossat, B. Se- gurens, C. Dasilva, M. Salanoubat, M. Levy, N. Boudet, S. Castel- lano, V. Anthouard, C. Jubin, V. Castelli, M. Katinka, B. Vacherie, C. Biemont, Z. Skalli, L. Cattolico, J. Poulain, V. De Berardinis, C. Cruaud, S. Duprat, P. Brottier, J. P. Coutanceau, J. Gouzy, G. Parra, G. Lardier, C. Chapple, K. J. McKernan, P. McEwan, S. Bosak, M. Kellis, J. N. Volff, R. Guigo, M. C. Zody, J. Mesirov, K. Lindblad-Toh, B. Birren, C. Nusbaum, D. Kahn, M. Robinson- Rechavi, V. Laudet, V. Schachter, F. Quetier, W. Saurin, C. Scarpelli, P. Wincker, E. S. Lander, J. Weissenbach, and H. Roest Crollius. Genome duplication in the teleost fish tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature, 431(7011):946–57, Oct 21 2004. [271] C. Roth, S. Rastogi, L. Arvestad, K. Dittmar, S. Light, D. Ekman, and D. A. Liberles. Evolution after gene duplication: models, mech- anisms, sequences, systems, and organisms. J Exp Zool B Mol Dev Evol, 308(1):58–73, Jan 15 2007. 183 [272] A. Force, M. Lynch, F. B. Pickett, A. Amores, Y. L. Yan, and J. Postlethwait. Preservation of duplicate genes by complementary, degenerative mutations. Genetics, 151(4):1531–45, Apr 1999. [273] K. S. Kassahn, V. T. Dang, S. J. Wilkins, A. C. Perkins, and M. A. Ragan. Evolution of gene function and regulatory control after whole- genome duplication: comparative analyses in vertebrates. Genome Res, 19(8):1404–18, Aug 2009. [274] M. Lynch, M. O’Hely, B. Walsh, and A. Force. The probability of preservation of a newly arisen gene duplicate. Genetics, 159(4):1789– 804, Dec 2001. [275] Andreas Wagner. Asymmetric functional divergence of duplicate genes in yeast. Mol Biol Evol, 19(10):1760–1768, Oct 2002. [276] K. Willecke, J. Eiberger, J. Degen, D. Eckardt, A. Romualdi, M. Guldenagel, U. Deutsch, and G. Sohl. Structural and functional diversity of connexin genes in the mouse and human genome. Biol Chem, 383(5):725–37, May 2002. [277] V. I. Shestopalov and Y. Panchin. Pannexins and gap junction protein diversity. Cell Mol Life Sci, 65(3):376–94, Feb 2008. [278] Buxton S Cheung M Cooper A Duran C Field M Heled J Kearse M Markowitz S Moir R Stones-Havas S Sturrock S Thierer T Wil- son A Drummond AJ, Ashton B. Geneious v4.8, available from http://www.geneious.com, 2009. [279] J. Felsenstein. Confidence-limits on phylogenies - an approach using the bootstrap. Evolution, 39(4):783–791, 1985. [280] J. M. Catchen, J. S. Conery, and J. H. Postlethwait. Automated identification of conserved synteny after whole-genome duplication. Genome Res, 19(8):1497–505, Aug 2009. [281] F. F. Bukauskas, A. Bukauskiene, M. V. Bennett, and V. K. Verselis. Gating properties of gap junction channels assembled from connexin43 and connexin43 fused with green fluorescent protein. Biophys J, 81(1):137–52, Jul 2001. [282] L. L. Sharp, J. Zhou, and D. F. Blair. Features of mota proton channel structure revealed by tryptophan-scanning mutagenesis. Proc Natl Acad Sci U S A, 92(17):7946–7950, Aug 1995. 184 [283] M. W. Hahn. Distinguishing among evolutionary models for the main- tenance of gene duplicates. J Hered, 100(5):605–17, Sep-Oct 2009. [284] C. Semple and K. H. Wolfe. Gene duplication and gene conversion in the caenorhabditis elegans genome. J Mol Evol, 48(5):555–64, May 1999. [285] M. Cifuentes, L. Grandont, G. Moore, A. M. Chevre, and E. Jenczewski. Genetic regulation of meiosis in polyploid species: new insights into an old question. New Phytol, 186(1):29–36, Apr 2010. [286] I. J. Leitch and M. D. Bennett. Genome downsizing in polyploid plants. Biological Journal of the Linnean Society, 82(4):651–663, Aug 2004. [287] C. Clair, L. Combettes, F. Pierre, P. Sansonetti, and G. Tran Van Nhieu. Extracellular-loop peptide antibodies reveal a predom- inant hemichannel organization of connexins in polarized intestinal cells. Exp Cell Res, 314(6):1250–65, Apr 1 2008. [288] C. Mayo, R. Ren, C. Rich, M. A. Stepp, and V. Trinkaus-Randall. Regulation by p2x7: epithelial migration and stromal organization in the cornea. Invest Ophthalmol Vis Sci, 49(10):4384–91, Oct 2008. [289] Ulf Geumann, Sina Victoria Barysch, Peer Hoopmann, Reinhard Jahn, and Silvio O. Rizzoli. Snare function is not involved in early endosome docking. Mol Biol Cell, 19(12):5327–5337, Dec 2008. [290] F. F. Bukauskas, K. Jordan, A. Bukauskiene, M. V. Bennett, P. D. Lampe, D. W. Laird, and V. K. Verselis. Clustering of connexin 43- enhanced green fluorescent protein gap junction channels and func- tional coupling in living cells. Proc Natl Acad Sci U S A, 97(6):2556– 2561, Mar 2000. [291] C. Carnarius, M. Kreir, M. Krick, C. Methfessel, V. Moehrle, O. Va- lerius, A. Bruggemann, C. Steinem, and N. Fertig. Green fluorescent protein changes the conductance of connexin 43 (cx43) hemichannels reconstituted in planar lipid bilayers. J Biol Chem, 287(4):2877–86, Jan 20 2012. [292] Stephen R. Bond and Christian C. Naus. Rf-cloning.org: an online tool for the design of restriction-free cloning projects. Nucleic Acids Res, 40(Web Server issue):W209–W213, Jul 2012. 185 [293] J. L. Hartley, G. F. Temple, and M. A. Brasch. Dna cloning using in vitro site-specific recombination. Genome Res, 10(11):1788–95, Nov 2000. [294] A. V. Bryksin and I. Matsumura. Overlap extension pcr cloning: a simple and reliable way to create recombinant plasmids. Biotechniques, 48(6):463–5, Jun 2010. [295] G. J. Chen, N. Qiu, C. Karrer, P. Caspers, and M. G. Page. Re- striction site-free insertion of pcr products directionally into vectors. Biotechniques, 28(3):498–500, 504–5, Mar 2000. [296] F. van den Ent and J. Lowe. Rf cloning: a restriction-free method for inserting target genes into plasmids. J Biochem Biophys Methods, 67(1):67–74, Apr 30 2006. [297] M. Geiser, R. Cebe, D. Drewello, and R. Schmitz. Integration of pcr fragments at any specific site within cloning vectors without the use of restriction enzymes and dna ligase. Biotechniques, 31(1):88–90, 92, Jul 2001. [298] T. Unger, Y. Jacobovitch, A. Dantes, R. Bernheim, and Y. Peleg. Applications of the restriction free (rf) cloning procedure for molecular manipulations and protein expression. J Struct Biol, 172(1):34–44, Oct 2010. [299] S. N. Ho, H. D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene, 77(1):51–9, Apr 15 1989. [300] C. Papworth, J. C. Bauer, and J. Braman. Quickchange site-directed mutagenesis. Strategies, 9:34, 1996. [301] X. Dong, P. Stothard, I. J. Forsythe, and D. S. Wishart. Plasmapper: a web server for drawing and auto-annotating plasmid maps. Nucleic Acids Res, 32(Web Server issue):W660–4, Jul 1 2004. [302] C. W. Dieffenbach, T. M. Lowe, and G. S. Dveksler. General concepts for pcr primer design. PCR Methods Appl, 3(3):S30–7, Dec 1993. [303] N. von Ahsen, M. Oellerich, V. W. Armstrong, and E. Schutz. Appli- cation of a thermodynamic nearest-neighbor model to estimate nucleic acid stability and optimize probe design: prediction of melting points 186 of multiple mutations of apolipoprotein b-3500 and factor v with a hybridization probe genotyping assay on the lightcycler. Clin Chem, 45(12):2094–101, Dec 1999. [304] Jr. SantaLucia, J. A unified view of polymer, dumbbell, and oligonu- cleotide dna nearest-neighbor thermodynamics. Proc Natl Acad Sci USA, 95(4):1460–5, Feb 17 1998. [305] R. M. Horton, H. D. Hunt, S. N. Ho, J. K. Pullen, and L. R. Pease. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene, 77(1):61–8, Apr 15 1989. [306] A. Erijman, A. Dantes, R. Bernheim, J. M. Shifman, and Y. Peleg. Transfer-pcr (tpcr): a highway for dna cloning and protein engineer- ing. J Struct Biol, 175(2):171–7, Aug 2011. [307] J. E. Stajich. An introduction to bioperl. Methods Mol Biol, 406:535– 48, 2007. [308] C. Camacho, G. Coulouris, V. Avagyan, N. Ma, J. Papadopoulos, K. Bealer, and T. L. Madden. Blast+: architecture and applications. BMC Bioinformatics, 10:421, 2009. [309] S. R. Bond, A. Lau, S. Penuela, A. V. Sampaio, T. M. Underhill, D. W. Laird, and C. C. Naus. Pannexin 3 is a novel target for runx2, expressed by osteoblasts and mature growth plate chondrocytes. J Bone Miner Res, 26(12):2911–22, Dec 2011. [310] P. Bargiotas, H. Monyer, and M. Schwaninger. Hemichannels in cere- bral ischemia. Curr Mol Med, 9(2):186–94, Mar 2009. [311] E. Scemes, D. C. Spray, and P. Meda. Connexins, pannexins, innexins: novel roles of ”hemi-channels”. Pflugers Arch, 457(6):1207–26, Apr 2009. [312] Z. Xiao, C. E. Camalier, K. Nagashima, K. C. Chan, D. A. Lucas, M. J. de la Cruz, M. Gignac, S. Lockett, H. J. Issaq, T. D. Veenstra, T. P. Conrads, and Jr. Beck, G. R. Analysis of the extracellular matrix vesicle proteome in mineralizing osteoblasts. J Cell Physiol, 210(2):325–35, Feb 2007. [313] V. Lefebvre and P. Smits. Transcriptional control of chondrocyte fate and differentiation. Birth Defects Res C Embryo Today, 75(3):200–12, Sep 2005. 187 [314] H. G. Foellmer, K. Kawahara, J. A. Madri, H. Furthmayr, R. Timpl, and L. Tuderman. A monoclonal antibody specific for the amino ter- minal cleavage site of procollagen type i. Eur J Biochem, 134(1):183–9, Jul 15 1983. [315] E. McLachlan, I. Plante, Q. Shao, D. Tong, G. M. Kidder, S. M. Bernier, and D. W. Laird. Oddd-linked cx43 mutants reduce endoge- nous cx43 expression and function in osteoblasts and inhibit late stage differentiation. J Bone Miner Res, 23(6):928–38, Jun 2008. [316] H. B. Evans, S. Ayad, M. Z. Abedin, S. Hopkins, K. Morgan, K. W. Walton, J. B. Weiss, and P. J. Holt. Localisation of collagen types and fibronectin in cartilage by immunofluorescence. Ann Rheum Dis, 42(5):575–81, Oct 1983. [317] M H Kaufmann. The Atlas of Mouse Development. Elsevier, London, sixth edition, 2003. [318] M. A. Larkin, G. Blackshields, N. P. Brown, R. Chenna, P. A. McGet- tigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, and D. G. Higgins. Clustal w and clustal x version 2.0. Bioinformatics, 23(21):2947–2948, Nov 2007. [319] K. Cartharius, K. Frech, K. Grote, B. Klocke, M. Haltmeier, A. Klin- genhoff, M. Frisch, M. Bayerlein, and T. Werner. Matinspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics, 21(13):2933–2942, Jul 2005. [320] P. Ducy, R. Zhang, V. Geoffroy, A. L. Ridall, and G. Karsenty. Osf2/cbfa1: a transcriptional activator of osteoblast differentiation. Cell, 89(5):747–54, May 30 1997. [321] Annie Paquin, Fanie Barnab-Heider, Ryoichiro Kageyama, and Freda D. Miller. Ccaat/enhancer-binding protein phosphorylation bi- ases cortical precursors to generate neurons rather than astrocytes in vivo. J Neurosci, 25(46):10747–10758, Nov 2005. [322] M. Q. Hassan, R. S. Tare, S. H. Lee, M. Mandeville, M. I. Morasso, A. Javed, A. J. van Wijnen, J. L. Stein, G. S. Stein, and J. B. Lian. Bmp2 commitment to the osteogenic lineage involves activation of runx2 by dlx3 and a homeodomain transcriptional network. J Biol Chem, 281(52):40515–26, Dec 29 2006. 188 [323] K. Nakashima and B. de Crombrugghe. Transcriptional mecha- nisms in osteoblast differentiation and bone formation. Trends Genet, 19(8):458–66, Aug 2003. [324] R. T. Ballock and R. J. O’Keefe. The biology of the growth plate. J Bone Joint Surg Am, 85-A(4):715–26, Apr 2003. [325] T. M. Schmid and T. F. Linsenmayer. Immunohistochemical localiza- tion of short chain cartilage collagen (type x) in avian tissues. J Cell Biol, 100(2):598–605, Feb 1985. [326] G. Shen. The role of type x collagen in facilitating and regulating endochondral ossification of articular cartilage. Orthod Craniofac Res, 8(1):11–7, Feb 2005. [327] J. Chen, K. Singh, B. B. Mukherjee, and J. Sodek. Developmental expression of osteopontin (opn) mrna in rat tissues: evidence for a role for opn in bone formation and resorption. Matrix, 13(2):113–23, Mar 1993. [328] A. Schnapper and W. Meyer. Osteopontin distribution in the canine skeleton during growth and structural maturation. Cells Tissues Or- gans, 178(3):158–67, 2004. [329] D. Tornehave, B. Teisner, H. B. Rasmussen, J. Chemnitz, and M. Kassem. Fetal antigen 2 (fa2) in human fetal osteoblasts, cul- tured osteoblasts and osteogenic osteosarcoma cells. Anat Embryol (Berl), 186(3):271–4, Aug 1992. [330] N. S. Datta, G. J. Pettway, C. Chen, A. J. Koh, and L. K. McCauley. Cyclin d1 as a target for the proliferative effects of pth and pthrp in early osteoblastic cells. J Bone Miner Res, 22(7):951–64, Jul 2007. [331] L. K. McCauley, A. J. Koh, C. A. Beecher, Y. Cui, T. J. Rosol, and R. T. Franceschi. Pth/pthrp receptor is temporally regulated during osteoblast differentiation and is associated with collagen synthesis. J Cell Biochem, 61(4):638–47, Jun 15 1996. [332] R. Meech, D. B. Edelman, F. S. Jones, and H. P. Makarenkova. The homeobox transcription factor barx2 regulates chondrogenesis during limb development. Development, 132(9):2135–46, May 2005. 189 [333] N. Yagishita, Y. Yamamoto, T. Yoshizawa, K. Sekine, Y. Uematsu, H. Murayama, Y. Nagai, W. Krezel, P. Chambon, T. Matsumoto, and S. Kato. Aberrant growth plate development in vdr/rxr gamma double null mutant mice. Endocrinology, 142(12):5332–41, Dec 2001. [334] J. Han, M. Ishii, Jr. Bringas, P., R. L. Maas, Jr. Maxson, R. E., and Y. Chai. Concerted action of msx1 and msx2 in regulating cranial neural crest cell differentiation during frontal bone development. Mech Dev, 124(9-10):729–45, Sep-Oct 2007. [335] G. S. Stein, J. B. Lian, A. J. van Wijnen, J. L. Stein, M. Montecino, A. Javed, S. K. Zaidi, D. W. Young, J. Y. Choi, and S. M. Pockwinse. Runx2 control of organization, assembly and activity of the regulatory machinery for skeletal gene expression. Oncogene, 23(24):4315–29, May 24 2004. [336] W. Wang, Y. G. Wang, A. M. Reginato, D. J. Glotzer, N. Fukai, S. Plotkina, G. Karsenty, and B. R. Olsen. Groucho homologue grg5 interacts with the transcription factor runx2-cbfa1 and modulates its activity during postnatal growth in mice. Dev Biol, 270(2):364–81, Jun 15 2004. [337] P. Ducy and G. Karsenty. Two distinct osteoblast-specific cis-acting elements control expression of a mouse osteocalcin gene. Mol Cell Biol, 15(4):1858–69, Apr 1995. [338] K. M. Catron, N. Iler, and C. Abate. Nucleotides flanking a con- served taat core dictate the dna binding specificity of three murine homeodomain proteins. Mol Cell Biol, 13(4):2354–65, Apr 1993. [339] S. Orestes-Cardoso, J. R. Nefussi, F. Lezot, M. Oboeuf, M. Pereira, M. Mesbah, B. Robert, and A. Berdal. Msx1 is a regulator of bone formation during development and postnatal growth: in vivo inves- tigations in a transgenic mouse model. Connect Tissue Res, 43(2- 3):153–60, 2002. [340] K. M. Catron, H. Wang, G. Hu, M. M. Shen, and C. Abate-Shen. Comparison of msx-1 and msx-2 suggests a molecular basis for func- tional redundancy. Mech Dev, 55(2):185–99, Apr 1996. [341] F. Zhuang, M. P. Nguyen, C. Shuler, and Y. H. Liu. Analysis of msx1 and msx2 transactivation function in the context of the heat shock 70 190 (hspa1b) gene promoter. Biochem Biophys Res Commun, 381(2):241– 6, Apr 3 2009. [342] Donald J. Glotzer, Elazar Zelzer, and Bjorn R. Olsen. Impaired skin and hair follicle development in runx2 deficient mice. Dev Biol, 315(2):459–473, Mar 2008. [343] C. K. Inman and P. Shore. The osteoblast transcription factor runx2 is expressed in mammary epithelial cells and mediates osteopontin expression. J Biol Chem, 278(49):48684–9, Dec 5 2003. [344] B. A. MacVicar and R. J. Thompson. Non-junction functions of pannexin-1 channels. Trends Neurosci, 33(2):93–102, Feb 2010. [345] P. Lu Valle, M. Iwamoto, P. Fanning, M. Pacifici, and B. R. Olsen. Multiple negative elements in a gene that codes for an extracellular matrix protein, collagen x, restrict expression to hypertrophic chon- drocytes. J Cell Biol, 121(5):1173–9, Jun 1993. [346] T. Barak-Shalom, M. Schickler, V. Knopov, R. Shapira, S. Hurwitz, and M. Pines. Synthesis and phosphorylation of osteopontin by avian epiphyseal growth-plate chondrocytes as affected by differentiation. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol, 111(1):49– 59, May 1995. [347] U. I. Chung, E. Schipani, A. P. McMahon, and H. M. Kronenberg. In- dian hedgehog couples chondrogenesis to osteogenesis in endochondral bone development. J Clin Invest, 107(3):295–304, Feb 2001. [348] H. M. Kronenberg. The role of the perichondrium in fetal bone devel- opment. Ann N Y Acad Sci, 1116:59–64, Nov 2007. [349] R. T. Franceschi, B. S. Iyer, and Y. Cui. Effects of ascorbic acid on collagen matrix formation and osteoblast differentiation in murine mc3t3-e1 cells. J Bone Miner Res, 9(6):843–54, Jun 1994. [350] J. T. Swarthout, R. C. D’Alonzo, N. Selvamurugan, and N. C. Par- tridge. Parathyroid hormone-dependent signaling pathways regulating genes in bone cells. Gene, 282(1-2):1–17, Jan 9 2002. [351] S. Harada and G. A. Rodan. Control of osteoblast function and regu- lation of bone mass. Nature, 423(6937):349–55, May 15 2003. 191 [352] W. Bi, J. M. Deng, Z. Zhang, R. R. Behringer, and B. de Crombrug- ghe. Sox9 is required for cartilage formation. Nat Genet, 22(1):85–9, May 1999. [353] L. C. Gerstenfeld and F. D. Shapiro. Expression of bone-specific genes by hypertrophic chondrocytes: implication of the complex functions of the hypertrophic chondrocyte during endochondral bone development. J Cell Biochem, 62(1):1–9, Jul 1996. [354] T. M. Schroeder, E. D. Jensen, and J. J. Westendorf. Runx2: a master organizer of gene transcription in developing and maturing osteoblasts. Birth Defects Res C Embryo Today, 75(3):213–25, Sep 2005. [355] I. S. Kim, F. Otto, B. Zabel, and S. Mundlos. Regulation of chondro- cyte differentiation by cbfa1. Mech Dev, 80(2):159–70, Feb 1999. [356] Qing Lin Li, Kosei Ito, Chohei Sakakura, Hiroshi Fukamachi, Ken ichi Inoue, Xin Zi Chi, Kwang Youl Lee, Shintaro Nomura, Chang Woo Lee, Sang Bae Han, Hwan Mook Kim, Wun Jae Kim, Hiromitsu Ya- mamoto, Namiko Yamashita, Takashi Yano, Toshio Ikeda, Shigeyoshi Itohara, Johji Inazawa, Tatsuo Abe, Akeo Hagiwara, Hisakazu Yamag- ishi, Asako Ooe, Atsushi Kaneda, Takashi Sugimura, Toshikazu Ushi- jima, Suk Chul Bae, and Yoshiaki Ito. Causal relationship between the loss of runx3 expression and gastric cancer. Cell, 109(1):113–124, Apr 2002. [357] C. A. Yoshida, H. Yamamoto, T. Fujita, T. Furuichi, K. Ito, K. Inoue, K. Yamana, A. Zanma, K. Takada, Y. Ito, and T. Komori. Runx2 and runx3 are essential for chondrocyte maturation, and runx2 regu- lates limb growth through induction of indian hedgehog. Genes Dev, 18(8):952–63, Apr 15 2004. [358] I. R. Orriss, G. E. Knight, S. Ranasinghe, G. Burnstock, and T. R. Arnett. Osteoblast responses to nucleotides increase during differen- tiation. Bone, 39(2):300–9, Aug 2006. [359] B. R. Olsen, A. M. Reginato, and W. Wang. Bone development. Annu Rev Cell Dev Biol, 16:191–220, 2000. [360] Sylvain Provot and Ernestina Schipani. Molecular mechanisms of endochondral bone development. Biochem Biophys Res Commun, 328(3):658–665, Mar 2005. 192 [361] Manuela Wuelling and Andrea Vortkamp. Transcriptional networks controlling chondrocyte proliferation and differentiation during endo- chondral ossification. Pediatr Nephrol, 25(4):625–631, Apr 2010. [362] E. J. Mackie, L. Tatarczuch, and M. Mirams. The skeleton: a multi- functional complex organ: the growth plate chondrocyte and endo- chondral ossification. J Endocrinol, 211(2):109–121, Nov 2011. [363] Y. Sasakura, E. Shoguchi, N. Takatori, S. Wada, I. A. Meinertzhagen, Y. Satou, and N. Satoh. A genomewide survey of developmentally relevant genes in ciona intestinalis. x. genes for cell junctions and ex- tracellular matrix. Dev Genes Evol, 213(5-6):303–13, Jun 2003. [364] M. E. Swartz, J. Eberhart, E. B. Pasquale, and C. E. Krull. Epha4/ephrin-a5 interactions in muscle precursor cell migration in the avian forelimb. Development, 128(23):4669–80, Dec 2001. [365] S. K. Loftus, D. M. Larson, D. Watkins-Chow, D. M. Church, and W. J. Pavan. Generation of rcas vectors useful for functional genomic analyses. DNA Res, 8(5):221–6, Oct 31 2001. [366] M. Chen, A. J. Granger, M. W. Vanbrocklin, W. S. Payne, H. Hunt, H. Zhang, J. B. Dodgson, and S. L. Holmen. Inhibition of avian leukosis virus replication by vector-based rna interference. Virology, 365(2):464–72, Sep 1 2007. [367] M. Chen, W. S. Payne, H. Hunt, H. Zhang, S. L. Holmen, and J. B. Dodgson. Inhibition of marek’s disease virus replication by retroviral vector-based rna interference. Virology, 377(2):265–72, Aug 1 2008. [368] A. Reynolds, D. Leake, Q. Boese, S. Scaringe, W. S. Marshall, and A. Khvorova. Rational sirna design for rna interference. Nat Biotech- nol, 22(3):326–30, Mar 2004. [369] Makoto Miyagishi, Hidetoshi Sumimoto, Hiroyuki Miyoshi, Yutaka Kawakami, and Kazunari Taira. Optimization of an sirna-expression system with an improved hairpin and its significant suppressive effects in mammalian cells. J Gene Med, 6(7):715–723, Jul 2004. [370] W. M. Potts, M. Olsen, D. Boettiger, and V. M. Vogt. Epitope map- ping of monoclonal antibodies to gag protein p19 of avian sarcoma and leukaemia viruses. J Gen Virol, 68 ( Pt 12):3177–3182, Dec 1987. 193 [371] M. Logan and C. Tabin. Targeted gene misexpression in chick limb buds using avian replication-competent retroviruses. Methods, 14(4):407–20, Apr 1998. [372] V. Hamburger and H. L. Hamilton. A series of normal stages in the development of the chick embryo. 1951. Dev Dyn, 195(4):231–72, Dec 1992. [373] H. Shen, T. Wilke, A. M. Ashique, M. Narvey, T. Zerucha, E. Savino, T. Williams, and J. M. Richman. Chicken transcription factor ap-2: cloning, expression and its role in outgrowth of facial prominences and limb buds. Dev Biol, 188(2):248–266, Aug 1997. [374] R. J. Wassersug. A procedure for differential staining of cartilage and bone in whole formalin-fixed vertebrates. Stain Technol, 51(2):131–4, Mar 1976. [375] S. H. Hughes. The rcas vector system. Folia Biol (Praha), 50(3-4):107– 19, 2004. [376] Christopher T. Gordon, Felicity A. Rodda, and Peter G. Farlie. The rcas retroviral expression system in the study of skeletal development. Dev Dyn, 238(4):797–811, Apr 2009. [377] Sanjiv Harpavat and Constance L. Cepko. Rcas-rnai: a loss-of-function method for the developing chick retina. BMC Dev Biol, 6:2, 2006. [378] E. Minina, H. M. Wenzel, C. Kreschel, S. Karp, W. Gaffield, A. P. McMahon, and A. Vortkamp. Bmp and ihh/pthrp signaling interact to coordinate chondrocyte proliferation and differentiation. Development, 128(22):4523–34, Nov 2001. [379] A. Haaijman, E. H. Burger, S. W. Goei, L. Nelles, P. ten Dijke, D. Huylebroeck, and A. L. Bronckers. Correlation between alk- 6 (bmpr-ib) distribution and responsiveness to osteogenic protein- 1 (bmp-7) in embryonic mouse bone rudiments. Growth Factors, 17(3):177–192, 2000. [380] Mary B. Goldring, Kaneyuki Tsuchimochi, and Kosei Ijiri. The control of chondrogenesis. J Cell Biochem, 97(1):33–44, Jan 2006. [381] Tatsuya Kobayashi, Ung-Il Chung, Ernestina Schipani, Michael Star- buck, Gerard Karsenty, Takenobu Katagiri, Dale L. Goad, Beate 194 Lanske, and Henry M. Kronenberg. Pthrp and indian hedgehog control differentiation of growth plate chondrocytes at multiple steps. Devel- opment, 129(12):2977–2986, Jun 2002. [382] D. M. Bell, K. K. Leung, S. C. Wheatley, L. J. Ng, S. Zhou, K. W. Ling, M. H. Sham, P. Koopman, P. P. Tam, and K. S. Cheah. Sox9 directly regulates the type-ii collagen gene. Nat Genet, 16(2):174–178, Jun 1997. [383] Fumiko Yano, Fumitaka Kugimiya, Shinsuke Ohba, Toshiyuki Ikeda, Hirotaka Chikuda, Toru Ogasawara, Naoshi Ogata, Tsuyoshi Takato, Kozo Nakamura, Hiroshi Kawaguchi, and Ung-Il Chung. The canon- ical wnt signaling pathway promotes chondrocyte differentiation in a sox9-dependent manner. Biochem Biophys Res Commun, 333(4):1300– 1308, Aug 2005. [384] J. Behrens, J. P. von Kries, M. Khl, L. Bruhn, D. Wedlich, R. Gross- chedl, and W. Birchmeier. Functional interaction of beta-catenin with the transcription factor lef-1. Nature, 382(6592):638–642, Aug 1996. [385] C. Hartmann and C. J. Tabin. Dual roles of wnt signaling during chondrogenesis in the chicken limb. Development, 127(14):3141–59, Jul 2000. [386] Florian Witte, Janine Dokas, Franziska Neuendorf, Stefan Mundlos, and Sigmar Stricker. Comprehensive expression analysis of all wnt genes and their major secreted antagonists during mouse limb develop- ment and cartilage differentiation. Gene Expr Patterns, 9(4):215–223, Apr 2009. [387] Yingzi Yang, Lilia Topol, Heuijung Lee, and Jinling Wu. Wnt5a and wnt5b exhibit distinct activities in coordinating chondrocyte prolifer- ation and differentiation. Development, 130(5):1003–1015, Mar 2003. [388] Frank Beier. Cell-cycle control and the cartilage growth plate. J Cell Physiol, 202(1):1–8, Jan 2005. [389] O. Jacenko, P. A. LuValle, and B. R. Olsen. Spondylometaphyseal dysplasia in mice carrying a dominant negative mutation in a matrix protein specific for cartilage-to-bone transition. Nature, 365(6441):56– 61, Sep 1993. 195 [390] S. Ikegawa, G. Nishimura, T. Nagai, T. Hasegawa, H. Ohashi, and Y. Nakamura. Mutation of the type x collagen gene (col10a1) causes spondylometaphyseal dysplasia. Am J Hum Genet, 63(6):1659–1662, Dec 1998. [391] R. Rosati, G. S. Horan, G. J. Pinero, S. Garofalo, D. R. Keene, W. A. Horton, E. Vuorio, B. de Crombrugghe, and R. R. Behringer. Normal long bone growth and development in type x collagen-null mice. Nat Genet, 8(2):129–135, Oct 1994. [392] K. M. Kwan, M. K. Pang, S. Zhou, S. K. Cowan, R. Y. Kong, T. Pfordte, B. R. Olsen, D. O. Sillence, P. P. Tam, and K. S. Cheah. Abnormal compartmentalization of cartilage matrix compo- nents in mice lacking collagen x: implications for function. J Cell Biol, 136(2):459–471, Jan 1997. [393] Shimei Zhu, Eric D. Zhu, Sylvain Provot, and Francesca Gori. Wdr5 is required for chick skeletal development. J Bone Miner Res, 25(11):2504–2514, Nov 2010. [394] Jasmin Coulombe-Huntington and Jacek Majewski. Characterization of intron loss events in mammals. Genome Res, 17(1):23–32, Jan 2007. [395] Katharina Kranz, Birthe Dorgau, Mark Pottek, Regina Herrling, Kon- rad Schultz, Petra Bolte, Hannah Monyer, Silvia Penuela, Dale W. Laird, Karin Dedek, Reto Weiler, and Ulrike Janssen-Bienhold. Ex- pression of pannexin1 in the outer plexiform layer of the mouse retina and physiological impact of its knock-out. J Comp Neurol, Sep 2012. [396] Elizabeth A. Kellogg. What happens to genes in duplicated genomes. Proc Natl Acad Sci U S A, 100(8):4369–4371, Apr 2003. [397] Ben J. Evans, Darcy B. Kelley, Richard C. Tinsley, Don J. Melnick, and David C. Cannatella. A mitochondrial dna phylogeny of african clawed frogs: phylogeography and implications for polyploid evolution. Mol Phylogenet Evol, 33(1):197–213, Oct 2004. [398] A. Ludwig, N. M. Belfiore, C. Pitra, V. Svirsky, and I. Jenneck- ens. Genome duplication events and functional reduction of ploidy levels in sturgeon (acipenser, huso and scaphirhynchus). Genetics, 158(3):1203–1215, Jul 2001. 196 [399] E. Y. Surez-Villota, R. A. Vargas, C. L. Marchant, J. E. Torres, N. Khler, J. J. Nez, R. de la Fuente, J. Page, and M. H. Gallardo. Distribution of repetitive dnas and the hybrid origin of the red viz- cacha rat (octodontidae). Genome, 55(2):105–117, Feb 2012. [400] T. Thomas, K. Jordan, and D. W. Laird. Role of cytoskeletal elements in the recruitment of cx43-gfp and cx26-yfp into gap junctions. Cell Commun Adhes, 8(4-6):231–236, 2001. [401] Jorge E. Contreras, Juan C. Sez, Feliksas F. Bukauskas, and Michael V L. Bennett. Gating and regulation of connexin 43 (cx43) hemichan- nels. Proc Natl Acad Sci U S A, 100(20):11388–11393, Sep 2003. [402] Tamsin Thomas, Karen Jordan, Jamie Simek, Qing Shao, Chris Jedeszko, Paul Walton, and Dale W. Laird. Mechanisms of cx43 and cx26 transport to the plasma membrane and gap junction regenera- tion. J Cell Sci, 118(Pt 19):4451–4462, Oct 2005. [403] Georgyi V. Los, Lance P. Encell, Mark G. McDougall, Danette D. Hartzell, Natasha Karassina, Chad Zimprich, Monika G. Wood, Randy Learish, Rachel Friedman Ohana, Marjeta Urh, Dan Simpson, Jacqui Mendez, Kris Zimmerman, Paul Otto, Gediminas Vidugiris, Ji Zhu, Aldis Darzins, Dieter H. Klaubert, Robert F. Bulleit, and Keith V. Wood. Halotag: a novel protein labeling technology for cell imaging and protein analysis. ACS Chem Biol, 3(6):373–382, Jun 2008. [404] Thomas Machleidt, Matt Robers, and George T. Hanson. Protein labeling with flash and reash. Methods Mol Biol, 356:209–220, 2007. [405] Roberto Paredes, Gloria Arriagada, Fernando Cruzat, Alejandro Villagra, Juan Olate, Kaleem Zaidi, Andre van Wijnen, Jane B. Lian, Gary S. Stein, Janet L. Stein, and Martin Montecino. Bone- specific transcription factor runx2 interacts with the 1alpha,25- dihydroxyvitamin d3 receptor to up-regulate rat osteocalcin gene ex- pression in osteoblastic cells. Mol Cell Biol, 24(20):8847–8861, Oct 2004. [406] K. H. Young. Yeast two-hybrid: so many interactions, (in) so little time... Biol Reprod, 58(2):302–311, Feb 1998. [407] Timothy W. Sikorski, Yoo Jin Joo, Scott B. Ficarro, Manor Askenazi, Stephen Buratowski, and Jarrod A. Marto. Proteomic analysis demon- 197 strates activator- and chromatin-specific recruitment to promoters. J Biol Chem, 287(42):35397–35408, Oct 2012. [408] Vladimir Pekarik, Dimitris Bourikas, Nicola Miglino, Pascal Joset, Stephan Preiswerk, and Esther T. Stoeckli. Screening for gene function in chicken embryo using rnai and electroporation. Nat Biotechnol, 21(1):93–96, Jan 2003. [409] Tatsuya Katahira and Harukazu Nakamura. Gene silencing in chick embryos with a vector-based small interfering rna system. Dev Growth Differ, 45(4):361–367, Aug 2003. [410] Yasuhiko Kawakami, Joaqun Rodrguez-Len, Christopher M. Koth, Dirk Bscher, Tohru Itoh, Angel Raya, Jennifer K. Ng, Concepcin Ro- drguez Esteban, Shigeru Takahashi, Domingos Henrique, May-Fun Schwarz, Hiroshi Asahara, and Juan Carlos Izpisa Belmonte. Mkp3 mediates the cellular response to fgf8 signalling in the vertebrate limb. Nat Cell Biol, 5(6):513–519, Jun 2003. [411] Romke Bron, Britta J. Eickholt, Matthieu Vermeren, Ninfa Fragale, and James Cohen. Functional knockdown of neuropilin-1 in the de- veloping chick nervous system by sirna hairpins phenocopies genetic ablation in the mouse. Dev Dyn, 230(2):299–308, Jun 2004. [412] Raman M. Das, Nick J. Van Hateren, Gareth R. Howell, Elizabeth R. Farrell, Fiona K. Bangs, Victoria C. Porteous, Elizabeth M. Manning, Michael J. McGrew, Kyoji Ohyama, Melanie A. Sacco, Pam A. Hal- ley, Helen M. Sang, Kate G. Storey, Marysia Placzek, Cheryll Tickle, Venugopal K. Nair, and Stuart A. Wilson. A robust system for rna in- terference in the chicken using a modified microrna operon. Dev Biol, 294(2):554–563, Jun 2006. [413] Takao Hashimoto, Xiang-Mei Zhang, Brenden Yi-kuang Chen, and Xian-Jie Yang. Vegf activates divergent intracellular signaling com- ponents to regulate retinal progenitor cell proliferation and neuronal differentiation. Development, 133(11):2201–2210, Jun 2006. [414] M. Chen, W. S. Payne, J. R. Dunn, S. Chang, H. M. Zhang, H. D. Hunt, and J. B. Dodgson. Retroviral delivery of rna interference against marek’s disease virus in vivo. Poult Sci, 88(7):1373–1380, Jul 2009. 198 [415] V Narry Kim. Microrna biogenesis: coordinated cropping and dicing. Nat Rev Mol Cell Biol, 6(5):376–385, May 2005. [416] Yuwei Li, Molly J. Ahrens, Amy Wu, Jennifer Liu, and Andrew T. Dudley. Calcium/calmodulin-dependent protein kinase ii activity reg- ulates the proliferative potential of growth plate chondrocytes. Devel- opment, 138(2):359–370, Jan 2011. [417] Jishuai Zhang, Xiaohong Tan, Wenlong Li, Youliang Wang, Jian Wang, Xuan Cheng, and Xiao Yang. Smad4 is required for the normal organization of the cartilage growth plate. Dev Biol, 284(2):311–322, Aug 2005. [418] Molly J. Ahrens, Yuwei Li, Hongmei Jiang, and Andrew T. Dudley. Convergent extension movements in growth plate chondrocytes require gpi-anchored cell surface proteins. Development, 136(20):3463–3474, Oct 2009. [419] Rachel M. Randall, Yvonne Y. Shao, Lai Wang, and R Tracy Ballock. Activation of wnt planar cell polarity (pcp) signaling promotes growth plate column formation in vitro. J Orthop Res, 30(12):1906–1914, Dec 2012. [420] Jeffery C. Giering, Dirk Grimm, Theresa A. Storm, and Mark A. Kay. Expression of shrna from a tissue-specific pol ii promoter is an effective and safe rnai therapeutic. Mol Ther, 16(9):1630–1636, Sep 2008. [421] A. Krogh, B. Larsson, G. von Heijne, and E. L. Sonnhammer. Pre- dicting transmembrane protein topology with a hidden markov model: application to complete genomes. J Mol Biol, 305(3):567–580, Jan 2001. [422] W. H. Landschulz, P. F. Johnson, and S. L. McKnight. The leucine zipper: a hypothetical structure common to a new class of dna binding proteins. Science, 240(4860):1759–64, Jun 24 1988. [423] R. S. Kass, J. Kurokawa, S. O. Marx, and A. R. Marks. Leucine/isoleucine zipper coordination of ion channel macromolecu- lar signaling complexes in the heart. roles in inherited arrhythmias. Trends Cardiovasc Med, 13(2):52–6, Feb 2003. [424] J. T. Hulme, T. Scheuer, and W. A. Catterall. Regulation of cardiac ion channels by signaling complexes: role of modified leucine zipper motifs. J Mol Cell Cardiol, 37(3):625–31, Sep 2004. 199 [425] S. Rozen and H. Skaletsky. Primer3 on the www for general users and for biologist programmers. Methods Mol Biol, 132:365–386, 2000. [426] J.F. Lorden, C.V. Kuh, and J.A. Voytuk. Research-doctorate pro- grams in the biomedical sciences: Selected findings from the nrc as- sessment. 2011. [427] E. S. Lander and The International Human Genome Sequencing Con- sortium. Initial sequencing and analysis of the human genome. Nature, 409(6822):860–921, Feb 2001. [428] S.C. Baker. Next-generation desktop sequencers. Genetic Engineering & Biotechnology News, 15(32), 2012. [429] Lincoln D. Stein. The case for cloud computing in genome informatics. Genome Biol, 11(5):207, 2010. 200 Appendix A Journal club publication — Passing potassium with and without gap junctions 201 Journal Club Editor’sNote:These short reviewsof a recentpaper in the Journal,writtenexclusivelybygraduate studentsorpostdoctoral fellows, are intended to mimic the journal clubs that exist in your own departments or institutions. For more information on the format and purpose of the Journal Club, please see http://www.jneurosci.org/misc/ifa_features.shtml. Passing Potassium with and without Gap Junctions Michael G. Kozoriz, Dave C. Bates, Stephen R. Bond, Charles P. K. Lai, and David M. Moniz Graduate Program in Cellular and Physiological Sciences, Department of Cellular and Physiological Sciences, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z3 Review of Wallraff et al. (http://www.jneurosci.org/cgi/content/full/26/20/5438) Glia, originally thought to play a passive role in the CNS, are now recognized as active regulators of CNS activity. For ex- ample, astrocytes are essential for the maintenance of extracellular ion concen- trations, notably K, at physiological lev- els (Orkand et al., 1966). Any deviation of extracellular K concentration ([K]o) from3mM can affect neural activity. El- evated [K]o occurs during seizure activ- ity, ischemia, and spreading depression, in which [K]o can reach 10–50 mM (Walz, 2000; Somjen, 2002), with conse- quent effects on neuronal excitability and ultimately on cell viability. To limit in- creased [K]o, astrocytes are equipped with a variety of K uptake mechanisms, including theNa-K-ATPase, Na-K- 2Cl cotransporters and voltage- activated K channels. Furthermore, lo- cal elevations in [K]o shift the K  equilibrium potential (EK) to more posi- tive values relative to the membrane po- tential (Vm), thus driving K  into these cells along an electrochemical gradient. K can also be spatially buffered via dif- fusion through the astrocytic cytoplasm to areas of lower [K]o, and then K  is driven back out of the cell at distal sites at which EK is still more negative than the Vm. Spatial redistribution of K  is be- lieved to be enhanced by gap junction coupling between astrocytes, although the experimental evidence for this mecha- nism is minimal (Walz, 2000; Kofuji and Newman, 2004). In their recent paper in The Journal of Neuroscience, Wallraff et al. (2006) examined the role of gap junction coupling between astrocytes inK buffer- ing and its subsequent physiological effect. Using hippocampal slices, the authors examined coupling in transgenic mice from which connexin43 (Cx43), the most abundant astrocytic Cx, had been condi- tionally deleted. Cx43/Cx30 double knock-out (dko) mice were also used be- cause Cx30 is the other major connexin in astrocytes. Dye coupling of the gap junction-permeable tracer biocytin was reduced in the Cx43 knock-out and abol- ished in the dko [Wallraff et al. (2006), their Fig. 1B,D (http://www.jneurosci. org/cgi/content/full/26/20/5438/F1)]. Importantly, the morphology and density of dko astrocytes labeled with GFAP were similar to wild-type (wt) [Wallraff et al. (2006), their Fig. 2B,D (http://www. jneurosci.org/cgi/content/full/26/20/ 5438/F2)]. The dko astrocytes had a slightly more negative resting membrane potential and a greater membrane resis- tance than wt, likely attributable to the lack of gap junction coupling. To detect the current passing through gap junctions (Igj), whole-cell currents (Ic) were re- corded inwt and dko cells before and after coupling was ablated by a low HCO3 /low pH solution. Wt Ic was reduced by 50% when coupling was blocked, and dko Ic was decreased by 20% [Wallraff et al. (2006), their Fig. 3D (http://www.jneurosci.org/cgi/ content/full/26/20/5438/F3)], a differ- ence that the authors attributed to Igj. Al- though low HCO3 /low pH abolished coupling, the 20%decrease in Ic in the dko preparation represents a modification in conductance that was independent of Igj. It would be interesting to see whether gap junctional blockers such as carbenox- olone produce Ic profiles similar to the low HCO3 /low pH solution. To test the involvement of gap junc- tions in limiting extracellular K accu- mulation, the authors used antidromic stimulation of CA1 pyramidal neurons to increase [K]o. The slices were treated withGABAand glutamate receptor block- ers to inhibit synaptic activity, and [K]o was measured with a K-sensitive elec- trode in the middle of the CA1 subfield. Compared with wt, a larger [K]o re- sponse was generated in dko slices after maximal stimulation (elicited by either paired-pulse or high-frequency trains) but not during low-to-moderate stimula- tion [Wallraff et al. (2006), their Fig. 4D,E (http://www.jneurosci.org/cgi/content/ full/26/20/5438/F4)]. During maximal stimulation, [K]o climbed as high as 17 mM in dko slices, whereas wt slices hit a “ceiling” at 12 mM. The rate of decay of [K]o after stimulation was also slower in the dko mice [Wallraff et al. (2006), their Fig. 5B,D (http://www.jneurosci.org/cgi/ content/full/26/20/5438/F5). Together, these data suggest a deficiency in K clearance in the absence of gap junctions, but only when K levels are high. Received June 13, 2006; revised June 23, 2006; accepted June 23, 2006. We thank Drs. Christian Naus and Claudia Krebs for their helpful com- ments and support of our Journal Club. Correspondence should be addressed to Michael G. Kozoriz, Depart- ment of Cellular and Physiological Sciences, Life Sciences Centre, 2350 Health Sciences Mall, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z3. E-mail: kozoriz@interchange.ubc.ca. DOI:10.1523/JNEUROSCI.2500-06.2006 Copyright©2006SocietyforNeuroscience 0270-6474/06/268023-02$15.00/0 The Journal of Neuroscience, August 2, 2006 • 26(31):8023–8024 • 8023 202 To further examine alterations in K clearance between wt and dko slices, rises in [K]o were generated in the stratum pyramidale in the same manner as before, and [K]o was recorded at various dis- tances from the site of stimulation. Up to 300 m away from the stratum pyrami- dale, corresponding to the stratum radia- tum, normalized [K]o remained un- changed between wt and dko slices. Astrocytes in this region have long over- lapping processes, and the authors rea- soned that K can be buffered to neigh- boring astrocytes independently of gap junctions via “indirect coupling” whereby K released from one astrocyte is taken up by another (Fig. 1). However, at a dis- tance of 400–500 m away, correspond- ing to the stratum lacunosummoleculare, [K]o levels were significantly reduced in the dko [Wallraff et al., their Fig. 6B (http://www.jneurosci.org/cgi/content/ full/26/20/5438/F6)]. Astrocytes in this region have short non-overlapping pro- cesses and appear to depend on gap junc- tions to facilitate K buffering (Fig. 1). Because the expression of other proteins (e.g., ion channels) may be altered by knocking out Cxs, testing whether gap junction blockers cause a similar impair- ment in K spatial buffering in the wt would be an important control. Further- more, the authors reported a steep decline in population spike potential as they moved away from the site of stimulation, but direct depolarization of the astrocytic syncytium cannot be ruled out as a poten- tial confounding factor. Last, extracellular field potentials were measured to investigate the potential pathological effect of reduced K clear- ance on neuronal activity. The dko slices developed spontaneous epileptiform events [Wallraff et al., their Fig. 7A,C (http://www.jneurosci.org/cgi/content/ full/26/20/5438/F7)] and experienced sei- zure-like discharges in the CA1 stratum pyramidale at a higher frequency than the wt slices when exposed to Mg2-free per- fusion media. These findings are in con- trast to previous studies that identified gap junction blockers as effective anticon- vulsants. However, the pharmacological agents used in other studies were not able to selectively block astrocytic gap junc- tions, which could indicate that the anti- convulsant effect may be the result of blocked neuronal gap junctions or action on another unidentified target. Wallraff et al. (2006) have demon- strated that K redistribution occurs by both gap junction-dependent and -independent processes, thereby supply- ing evidence for a pathway that has been postulated for decades. Efficient handling of K has relevance to the treatment of pathological conditions, and indeed the authors found an increased susceptibility to epileptiform events in the Cx43/Cx30 dko mice. These findings may extend to other conditions, such as stroke, in which elevated [K]o occurs. More efficient K  buffering by astrocytes could curb exces- sive K accumulation and in turn reduce the potential for neuronal damage. References Kofuji P, Newman EA (2004) Potassium buffer- ing in the central nervous system. Neuro- science 129:1045–1056. Orkand RK, Nicholls JG, Kuffler SW (1966) Ef- fect of nerve impulses on the membrane po- tential of glial cells in the central nervous sys- tem of amphibia. J Neurophysiol 29:788–806. Somjen GG (2002) Ion regulation in the brain: implications for pathophysiology. Neurosci- entist 8:254–267. Wallraff A, Kohling R, Heinemann U, Theis M, Willecke K, Steinhauser C (2006) The im- pact of astrocytic gap junctional coupling on potassium buffering in the hippocampus. J Neurosci 26:5438–5447. Walz W (2000) Role of astrocytes in the clear- ance of excess extracellular potassium.Neuro- chem Int 36:291–300. Figure 1. Overview of K buffering of CA1 neurons in the hippocampus. Neuronal activity in the stratum pyramidale (S.P.) causes a rise in [K]o that is subsequently taken up by astrocytes in the stratum radiatum (S.R.). K  is distributed along the long processes of these cells toneighboringS.R. astrocytesbygap junctions andby indirect coupling (a)wherebyK released fromone cell into the extracellular space is taken up by another neighboring astrocyte. In this case, dko astrocytes can still buffer K because indirect coupling remains intact. K is subsequently distributed via gap junctions (b) in stratum lacunosummoleculare (S.L.M) astrocytes. In this case, K buffering is impaired in dko because astrocytes lack preferential organization for indirect coupling. 8024 • J. Neurosci., August 2, 2006 • 26(31):8023–8024 Kozoriz et al. • Journal Club 203


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items