Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cooperative binding mechanisms leading to the specific androgen receptor regulation of target genes Reid, Kimberly Jane 2001

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

Item Metadata

Download

Media
831-ubc_2001-0604.pdf [ 6.28MB ]
Metadata
JSON: 831-1.0090177.json
JSON-LD: 831-1.0090177-ld.json
RDF/XML (Pretty): 831-1.0090177-rdf.xml
RDF/JSON: 831-1.0090177-rdf.json
Turtle: 831-1.0090177-turtle.txt
N-Triples: 831-1.0090177-rdf-ntriples.txt
Original Record: 831-1.0090177-source.json
Full Text
831-1.0090177-fulltext.txt
Citation
831-1.0090177.ris

Full Text

COOPERATIVE BINDING MECHANISMS L E A D I N G TO THE SPECIFIC A N D R O G E N RECEPTOR R E G U L A T I O N OF TARGET GENES by K I M B E R L Y J A N E RE ID B.Sc. Double Major, Royal Roads Military College, 1989 B.Sc. Honours, The University of Victoria, 1997 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR A M A S T E R OF SCIENCB€?SSi^-~-in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Experimental Pathology and Laboratory Medicine) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A June 2001 © Kimberly Jane Reid, 2001 UBC Special Collections - Thesis Authorisation Form http://www.library.ubc.ca/spcoll/thesauth.html In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department 1 of 1 8/27/2001 2:37 PM Abstract Genes uniquely regulated by the androgen receptor (AR) typically contain multiple androgen response elements (AREs) that in isolation are of low D N A binding affinity and transcriptional activity. However, specific combinations of A R E s in their native promoter context results in highly cooperative D N A binding by A R and high levels of transcriptional activation. Within this study, we demonstrate that the natural androgen-regulated promoters of P S A and Probasin contain two classes of A R E s dictated by their primary nucleotide sequence that function to mediate cooperativity. Class I AR-binding sites display conventional guanine contacts. Class II AR-binding sites have distinctive atypical sequence features and upon binding to A R the D N A structure is dramatically altered through allosteric interactions with the receptor. Class II sites stabilize AR-binding to adjacent Class I sites and result in synergistic transcriptional activity and increased hormone sensitivity. The specific nucleotide variation within the androgen receptor binding sites dictate was determined to dictate differential functions to the receptor. The potential role of individual nucleotides within Class II sites and predicted consensus sequences for Class I and II sites was also identified. Our data suggest that this may be a universal mechanism by which A R achieved unique regulation o f target genes through complex allosteric interactions dictated by primary binding sequences. Table of Contents Abstract p. i i Table of Contents p. i i i List o f Figures p. v i List o f Tables p. ix List of Abbreviations p. X Acknowledgements p. x i i i Preface p. x iv Chapter 1. Introduction p. 1 1.1 The Androgen Receptor and Prostate Cancer p. 1 1.2 Normal Androgen Act ion p. 2 1.3 The Steroid Receptor Subfamily Makes Common and Discriminatory Protein-DNA Contacts p. 3 1.4 The Steroid Receptors Cooperatively B ind to a Common Steroid Response Element and Homodimerize Upon D N A Binding p. 8 1.5 Five Levels Through Which AR-Specific Gene Regulation Can Occur p. 11 1.6- The Architectural Context of AR-Regulated D N A Regions Underlies A R Cooperative Binding p. 13 1.7 Cooperativity Arises From Protein-Protein and Protein-DNA Interactions p. 19 1.8 The Probasin Promoter and the P S A Enhancer Possess Multiple Cooperatively Interacting A R E s p. 23 1.8.1 The Rat Probasin Promoter p. 23 1.8.2 The Human P S A Enhancer p. 25 1.9 Objectives of this Study p. 27 Chapter 2. Experimental Procedures p. 29 2.1 Methylation Protection and Methylation Intereference o f t h e A R D B D p. 29 2.1.1 Preparing Recombinant D N A Binding Protein p. 32 2.1.2 Generation of Radioactively Labelled D N A Probes p. 33 2.1.2a Generating Single-End Labelled D N A Probes Using P C R Amplification p. 34 2.1.2b Generating Single-End Labelled D N A Probes Through Successive Enzymatic Digestion of Recombinant Plasmid Vectors p. 36 2.1.2c Generating Single-End Labelled D N A Probes by Annealing Synthesized Oligos p. 38 2.1.3 Methylation Protection Assay p. 3 8 2.1.4 Methylation Interference Assay p. 41 2.2 Determining the Binding Affinities of the Probasin Promoter A R E s p. 42 2.2.1 Generating Radioactively Labelled Probasin Promoter Fragments With High Specific Activi ty p. 42 2.2.2 Determining Binding Affinity Through E M S A p. 43 2.3 Transfection Studies to Determine Effects of Probasin Promoter Fragments on Transcriptional Activation p. 44 Chapter 3. Results and Discussion p. 46 3.1 Methylation Interference Assay Failed to Footprint a Cooperative System p. 46 3.2 A Modified Methylation Protection Assay Revealed Atypical A R Binding Sites p. 46 3.3 Local D N A Structural Distortion Resulting From A R D B D Allosteric Binding was Intrinsic to the D N A Sequences of the G - l and G-2 A R E s p. 53 3.4 Hypersensitive Guanines in Class II A R E s are Necessary for A R D B D Binding 3.5 The Human P S A Enhancer Contained Class I and Class II A R E s and the Two Classes Possess Distinctive Nucleotide Features 3.6 Mutation of K e y Nucleotides Transforms the M e P Pattern of a Class II Half-Site to a Class I Half-Site 3.7 Specific Arrangements of Class I and Class II A R E s Increase Cooperative Binding 3.8 Class II A R E s Contribute to Synergistic Transactivation as Observed in Transient Transfection Assays 3.9 Cooperative Protein-Protein Interactions Between A R Homodimers Could Involve Stabilizing Disulfide Bridg 3.10 Recombinant Sox-4 Protein Increases the Number of A R D B D Monomers that Bind the Probasin Promoter 3.11 M e P of Other AR-Regulated Promoter and Enhancer Regions Reveal Novel Protection Patterns that Suggest Monomer and Direct Repeat A R - D N A Binding Chapter 4. Future Directions and Conclusion 4.1 Summary of Project Results 4.2 Future Directions 4.3 Conclusion Bibliography List of Figures Figure 1. Figure 2 A and B . Figure 3 A and B . Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. The process of androgen action in a prostate epithelial cell p. 4 Rat androgen receptor D B D structure and D N A half-site contacts p. 5 Structure of the E R homodimer bound to an E R E and the specific A R contacts with the S R E p. 7 The transcription factor-DNA multicomplex or enhanceosome recruits the basal transcriptional machinery p. 14 The rat probasin promoter from -426 to +28 base pairs p. 16 The human P S A enhancer from -4267 to -4026 base pairs p. 17 Schematic of the methylation protection assay p. 30 Schematic of the methylation interference assay p. 31 Failed M e P of AR-His tag on the probasin -426 to +28 base pair promoter fragment using the conventional M e P assay p. 48 1.0 M P-mercaptoethanol disrupted the A R DBD-probasin promoter complex p. 49 Modified M e P of the A R DBD-His tag on the probasin promoter from -269 to -77 base pairs p. 51 The probasin promoter sequence from -426 to +24 base pairs p. 52 M e P of the probasin promoter sequence from -269 through -77 base pairs p. 54 M e P of the probasin promoter sequence from -269 through -164 base pairs p. 55 vi Figure 15. M e P of the probasin promoter sequence from -150 through -77 base pairs Figure 16. M e P of the individual G-site probasin promoter A R E s p. 58 Figure 17. M e l of the G-1 and A R E 2 from the probasin promoter p. 59 Figure 18 A . M e P of the human P S A enhancer from -4267 to -4062 base pairs p. 61 Figure 18B and C. M e P of the P S A enhancer A R E IIIA and the P S A enhancer sequence from -4267 to -4062 base pairs p. 62 Figure 19A-D The change in the relative D M S methylation hypersensitivity and protection pattern of the probasin G - l element after site-directed mutagenesis p. 66 Figure 20. Binding Affinity Curves p. 70 Figure 2 1 A - C . Transient Transfection Curves p. 72 Figure 22 A and B . E M S A showing P-mercaptoethanol disruption of the A R D B D - D N A binding complex p. 76 Figure 22C. 1.0 M of P-mercaptoethanol had negligible effect on A R D B D binding to the canonical S R E p. 77 Figure 23. Sox-4 protein increases the occupancy o f A R D B D on the probasin promoter p. 79 Figure 24A. Methylation protection of the A R D B D on the m E - R A B P promoter from -543 to-166 base pairs p. 82 v i i Figure 24B. Methylation protection of the A R D B D on the m E - R A B P promoter from -195 to +26 base pairs p. 83 Figure 25. M e P of the A R D B D on the rat A R enhancer from 2028 to 2333 base pairs p. 84 Figure 26. M e P of the human p21 promoter from -285 to -1 base pairs p. 85 Figure 27. Methylation protection of the A R D B D on the upstream proximal probasin promoter from -705 to -426 base pairs p. 86 v i i i List of Tables Table 1. Nuclear receptor P Box and D box sequence comparison Table 2. Alignment of A R E s into proposed Class II and Class I consensus sequences Abbreviations A l a -alanine A R -androgen receptor A R E -androgen receptor response element B L 1 3 -human transitional cell bladder carcinoma B M R -BioMax Resolution (Kodak film) B M S -BioMax Sensitive (Kodak film) C A T -chloramphenicol acetyl transferase C H O -Chinese hamster ovary cell line crp -calf intestinal phosphatase D B D - D N A binding domain D B o x -Distal Box D H T -dihydrotestosterone D M S -dimethyl sulfate dNTP -dideoxynucleotide triphosphate DU145 -prostate cancer metastasis to the brain cell line E M S A -electrophoretic mobility shift assay E R -estrogen receptor E R E -estrogen receptor response element F C S -fetal calf serum F R E T -fluorescence resonance energy transfer G R -glucocorticoid receptor HepG2 -human hepatic cell line hGK-1 -human glandular kall ikreinl HisTag -histidine tagged H M G -high mobility group Hsp70 -heat shock protein 70 Hsp90 -heat shock protein 90 IFN-p -interferon beta I g - K B -immunoglobulin-KB X K L K - 1 L B D L N C a P L P B M a t a / M C M 1 / S T E 6 M C F - 7 M e l m E - R A B P M e P M M T V M R N A P S N F A T / F o s - J u n / A R R E 2 N F - K B / H M G I/PD II Oct-1 P A G E P B o x PC3 P C R P N K P R Probasin P S A R L U -kallekrein-1 -ligand binding domain -Lymph Node (metastases) Prostate Cancer cell line -large probasin promoter (~12 kbp) -multicomplex of mating-type a protein, minichromosome maintenance-1 protein and the a-cell-specific transporter gene -human breast carcinoma cell line -methylation interference -mouse epididymus retinoic acid binding protein -methylation protection -mouse mammary tumour virus -mineralocorticoid receptor -nucleic acid and protein synthesis -multicomplex of the nuclear factor of activated T cells, the nuclear phosphoproteins Fos and Jun, and the A R R E 2 element within the interleukin-2 promoter -multicomplex of the nuclear factor-KB protein, the high mobility group I protein, and the positive regulatory domain II within the endothelial leukocyte adhesion molecule 1 gene promoter -octamer-1 protein -polyacrylamide gel electrophoresis -proximal box -prostate cancer 3 cell line -polymerase chain reaction -polynucleotide kinase -progesterone receptor -prostatic basic protein -postate specific antigen -relative luciferase unit x i R1881 -synthetic androgen methyltrienolone Sip -sex limited protein Sox-4 -Sry-related H M G box protein 4 SRC-1 -steroid receptor coactivator-1 SRE -steroid receptor? response element SV40 -simian virus 40 T -testosterone Thr -threonine TR -thyroid hormone receptor WT -wildtype 293 -human embryonic kidney cell line Acknowledgements and Dedication I would like to gratefully acknowledge and thank my supervisor, Dr. Colleen Nelson whose enthusiasm and dedication to science has been a tremendous example. Within and without the Prostate Centre I would like to thank the following people for their encouragement, ever-present support and willingness to critically read this thesis: Jody Saito, Doug Hoffart, Stephen Hendy, Helen Cheng, Dr. Pepita Gimenez-Bonafe, M i r a Rao, L i l l i an Yeung, Dr. Jason Read, Dr. Paul Rennie, Tanis Whitmore and Dr. Sujata Presard. This thesis is dedicated to my husband, Phil l ip A . Rennison, for his limitless patience and encouragement. Preface. Results and conclusions contained within this study have been published previously in three separate journals. The methylation protection, binding affinity and transient transfection results pertaining to the proximal probasin promoter and upstream P S A enhancer were published in the Journal of Biological Chemistry. The methylation protection results pertaining to the mouse epididymus-retinoic acid binding protein promoter were published in the journal, Biology of Reproduction. Finally, the novel methylation protection technique described in this study was previously published in the journal, Biotechniques. The publications in detail are as follows: Reid K J , Hendy SC, Saito J, Sorensen P, Nelson C C . 2001. Two Classes of Androgen Receptor Elements Mediate Cooperativity through Allosteric Interactions. J B i o l Chem Jan 26;276(4):2943-52 Lareyre JJ, Reid K , Nelson C, Kasper S, Rennie PS, Orgebin-Crist M C , Matusik R J , 2000. Characterization of an androgen-specific response region within the 5' flanking region of the epididymal retinoic acid binding protein gene. B i o l Reprod Dec;63(6):1881-92 Reid K J , Nelson C C . 2001. Improved methylation protection-based footprinting to reveal structural distortion of D N A transcription factor binding. Biotechniques Jan;30(l):20-2 xiv 1. Introduction 1.1 The Androgen Receptor and Prostate Cancer The prostate gland is a male accessory sex organ that is extremely sensitive to androgens, such as testosterone and dihydrotestosterone (DHT), and to their cognate receptor, the androgen receptor (AR) (Petrow et al, 1986; Frick et al, 1991). A correctly functioning A R and sufficient androgen hormone are necessary for the normal growth and development of prostate epithelial tissue (Isaacs et al, 1994; Thomas et al, 2001). Mediated by their action through A R , androgens appear to control prostate epithelial cell number by both stimulating the rate of cell proliferation and inhibiting the rate of cell death (Isaacs et al, 1994). Removal of androgens w i l l invoke programmed cell death and, because of their reliance on androgens for survival, prostate epithelial cells are referred to as androgen dependent (Isaacs et al, 1994). From the above, it is apparent that androgens and the A R are critically involved in a complex cell-signaling network that is responsible for prostate epithelial cell growth, development and maintenance. L ike healthy prostate, cancerous prostate epithelial cells require androgens for proliferation. Although a causative role for the androgen/AR complex in the onset of prostate cancer has yet to be shown, upregulation and increased mutation frequency of the A R , leading to a gain in function, is a hallmark of advanced staged disease and could contribute to metastasis and progression (Hobisch et al, 1995; Taplin et al, 1995; Marcell i et al, 2000). Mutated A R can be activated by steroids or agonists other than androgens: within the cultured cell line L N C a P , which is derived from a prostate cancer metastasis to the lymph node, the endogenous A R contains a Thr—»Ala mutation at position 877 that permits a broadened specificity beyond androgens for other steroids and steroid-like molecules such as progesterone, oestradiol and hydroxyflutamide (MacDonal et al, 2000). This particular A R mutation and others that lead to ligand promiscuity, were found in several prostate cancer metastases (Suzuki et al, 1993; Gaddipato et al, 1994). In addition, the A R was amplified in prostate cancers such that low levels of androgens are able to induce a biologically active response (Hobisch et al, 1995; Koivisto et al, 1995; Visakkorpi et al, 1995; Linja et al, 2001). The adrenal androgens, androstenedione and dehydroepiandrosterone, normally contribute less than 10% of the precursor androgens to the prostate gland, however this low level may be enough to be biologically active i f the number of A R molecules is highly amplified (Culig et al, 1996, Tan et al, 1997). A R activity appears to be critical to not only normal prostate growth but also to the growth and progression of prostate cancer. The primary action of steroid hormones, such as androgens, is to stimulate target gene expression by stimulating rates of m R N A production (Higgins and Gehring, 1978). The selective and specific binding of steroid hormone-receptor complexes to target gene regulatory regions regulates transcription of m R N A (Gronemeyer and Pongs, 1980; Payvar et al, 1981). Surprisingly, it is not clear how the A R specifically and uniquely regulates the transcription of target genes that ultimately controls the growth, development and homeostasis of prostate cells. Bayes' Theorem states that the more we know about a system, the better we can predict its outcome. If we apply Bayes' Theorem to the role o f androgens and the A R in prostate cancer, then the more we know about normal androgen action, the better we w i l l gain insight in how prostate cancer progresses and how it may be controlled. 1.2 N o r m a l Androgen Act ion Androgens, such as testosterone, belong to the steroid hormone family that includes progesterone, estradiol, glucocorticoids and mineralocorticoids. Androgens are responsible for the normal development, differentiation, growth and homeostasis of several organs including, among others, the human male prostate. It is accepted within the steroid field that androgen action is primarily responsible for the above cellular effects. Androgen action is a molecular and biological process whereby the extracellular androgenic signal culminates in direct transcriptional regulation by the A R , a transcription factor that primarily employs the androgens, testosterone or D H T , as activating ligands (Brinkmann et al, 1989; Govindan et al, 1991). Due to their lipophilic nature, androgens are able to easily enter the plasma membrane that separates the cell's cytoplasm from the extracellular environment. How lipophilic androgens enter into the aqueous environment of the cytosol is not clearly understood. However, once androgens do enter the cell they are often converted to a more active form (Figure 1). Within prostate epithelial cells the enzyme 5a-reductase converts testosterone to 2 D H T , which has a higher affinity for the ligand-binding domain ( L B D ) of the A R than its precurser, testosterone. The cognate receptor for androgens is the A R , which is a ligand-activated transcription factor that belongs to the superfamily of nuclear hormone receptors. The A R can be further assigned to the steroid receptor subfamily that includes the receptors for glucocorticoid, mineralocorticoid and progesterone (respectively G R , M R and PR) (Laudet et al, 1992). L ike the G R , correctly folded A R is normally chaperoned within the cytoplasm by the heat shock proteins hsp70 and hsp90 along with other co-chaperone molecules (Ohara-Nemoto et al, 1988; Veldscholte et al, 1992) (Figure 1). After D H T binds the A R L B D , the A R / D H T complex is thought to dissociate from the chaperone multiplex and translocate to the nucleus. Once inside the nucleus the A R regulates transcription by specifically binding to androgen response elements (AREs) located within the promoter and enhancer regions of target genes. 1.3 The Steroid Receptor Subfamily Makes C o m m o n and Discr iminatory Protein-D N A Contacts A l l steroid receptors are capable of homodimerizing in a DNA-dependent manner through contacts mediated by their D N A binding domain and some are capable of homodimerizing in solution through contacts mediated by their L B D or D N A hinge region (Tsai et al, 1988; Dahlman-Wright et al, 1990; de Vos et al, 1993; Savory et al, 2001). In solution and not bound to D N A , the A R is able to homodimerize at concentrations wel l above physiological conditions (> 0.2 uM) , however, A R primarily exist as a monomer in solution and below stochiometric concentrations (Liao et al, 1999). After activation by ligand, a steroid receptor monomer located in the nucleus w i l l recognize a six base pair D N A half site with the consensus sequence A G A A C A (Scheidereit et al, 1983; Cato et al, 1988; Roche et al, 1992). A l l four steroid receptors share a highly homologous D N A binding domain (DBD) that contains two zinc-binding motifs where each motif is closely followed by a short amphipathic a-helix (Figure 2A). The a-helix immediately following the first zinc motif is called the D N A recognition helix and contains the steroid receptor D N A recognition motif G S C K V called the Proximal Box (P Box) (Umesono and Evans, 1989). 3 A T Prostate Epithelial Cell Figure 1. The process o f androgen action. Testosterone (T) enters the cytoplasm, is converted to D H T , which in turn binds the A R . The Hsp70/90 chaperone complex releases the A R / D H T complex, which then translocates to the nucleus where it binds an A R E and regulates gene transcription. c G H rAR S A E D G \ C I L / C T R D H V L P I D Y Y F P P Q K I Y G A L T C D B o x Z n 524 G S c K V K D I T ND R S R R K N MC \ / Z n / \ F F K R A A E G K Q K Y L P S R L R K P B o x C Y E I H S V T M K Q S P D E T P S G A S S N E G E E Q L K L N G L K K L K R A G L T M G A F 648 B A R P-Box NH3-Steroid Receptor Half-site G S C |K][v1 F F K [R] - C O O H 5 ' A [G] A \ A C / A 3 ' 3 , T C T If l lGl T 5 , Figure 2. Rat androgen receptor DBD structure and DNA half-site contacts. A. The D N A recognition P Box and dimerization D Box are outlined with the contacting and dimerizing residues indicated in bold (Umesono and Evans 1989). B. A n A R monomer w i l l interact with a steroid receptor canonical D N A half-site. The lysine (K) and arginine (R) residues make hydrogen bonds with the indicated guanines within the half-site. The Van der Waal contact between the valine (V) residue and the thymidine is discriminatory for the steroid receptor subfamily. In general, the zinc motif and oc-helical structure of the D B D is highly conserved among members of the nuclear receptor superfamily. Upon recognition of the A G A A C A half site, the steroid receptor D N A recognition a-helix makes base pair contacts using amino acid residues that are universally conserved throughout the nuclear receptor superfamily. The conserved lysine located in the steroid receptor P Box G S C K V makes a hydrogen bond essential for binding with the first guanine in the half site A G A A C A (Figure 2B) (Umesono and Evans, 1989). This conserved lysine is also present in the estrogen receptor/thyroid hormone receptor (ER/TR) subfamily of nuclear receptors within the P Box sequences E G C K A , E G C K G , E G C K S , and E A C K A . A second essential protein-DNA contact conserved throughout the nuclear receptor superfamily is made by an arginine, which is located outside of the P-box but within the D N A recognition a-helix. This arginine makes a hydrogen bond to a conserved guanine base paired to cytosine in the steroid receptor half site A G A A C A (Figure 2B). The X-ray crystallographic structure of the A R bound to its D N A half site is still forthcoming, however, such an assembly has been achieved for the G R although the resolved structure contained D N A half sites that were separated by four nucleotides rather than the requisite three (Luisi et al, 1991). Since the steroid receptors, except E R , share a highly homologous D N A recognition a-helix and appear to recognize and bind to the same D N A half site, it is feasible to extrapolate from the G R / D N A crystal structure those protein-DNA contacts that are specific to the steroid receptor subfamily. The crystal structure of the G R homodimer bound to D N A has shown that each G R monomer D N A recognition a-helix lies within the major groove at right angles to the D N A helix and makes contacts with D N A through weak interactions such as hydrophobic and hydrogen bonds (Luisi et al, 1991). The same recognition a-helix/major groove orientation has been demonstrated for the E R / D N A crystal structure (Figure 3A) (Schwabe et al, 1993). The protein-DNA interactions revealed by both the G R / D N A and the E R / D N A crystal structures include the lysine-guanine and arginine-cytosine hydrogen bonds that are common to all nuclear receptors (Figure 2B). In addition, the G R / D N A structure illustrated that a valine found within the steroid receptor P Box G S C K V makes a specific and discriminating van der Waals contact with the methyl group of the thymidine base paired to A in the half site A G A A C A (Luisi et al, 1991; Schwabe et al, 1993). 6 Zinc motifs containing D Box dimerization loops cx-helices and P Boxes Major groove B AR P-Box NH3-S R E G S C K V 5 ' G P G T F F K R - C O O H N N N N N N T G T T C T 3-C C A T G' T > C O O H - I " | V i AR P-Box Figure 3. A . Structure o f the E R homodimer bound to a consensus E R E . The E R recognition a-helices lie at right angles to the D N A axis within the major groove. The D Boxes located within the second zinc motif mediate dimerizing salt bridges between the two E R monomers. Structure was visualized using Chime software and obtained from the Protein Data Bank (Schwabe et al, 1993). B . Two A R monomers bind two half sites within the S R E . The K and R residues make hydrogen bonds with the indicated guanines and the V residue makes a hydrophobic interaction with the thymidine methyl group. This valine-thymidine contact is discriminating for the steroid receptor subfamily because the alanine within the E R P Box E G C K A sterically restricts the E R from binding to a D N A sequence containing an adenine in the fourth position of the half site A G A A C A . 1.4 The Steroid Receptors Cooperatively B i n d to a C o m m o n Steroid Response Element and Homodimerize Upon D N A Bind ing Members of the E R / T R subfamily of nuclear receptors also recognize a common half site A G G T C A . Specificity of response, however, is primarily obtained through the spacing and orientation of pairs of half sites found within gene regulatory regions (Umesono and Evans, 1989, Umesono et al, 1991). In contrast, the steroid receptors all commonly recognize, bind to, and homodimerize on an imperfect palindrome comprised o f two six base pair half sites in reverse or head-to-head orientation. These two half sites are usually separated by three base pairs as represented by the steroid response element (SRE) 5' G G T A C A n n n T G T T C T 3', which is an optimized sequence derived from the mouse mammary tumor virus ( M M T V ) promoter (Figure 3B) (Beato et al, 1987; Ham et al, 1988). Upon binding to the first half site in the S R E , the receptor monomer undergoes a conformation change exposing the dimerizing Distal Box (D Box) (Holmbeck et al, 1998; van Tilborg et al, 2000) (Figure 2A) . Unlike the E R / T R subfamily, the amino acid residues within the D Box are highly homologous among members of the steroid receptor subfamily, which may account for the rigidity of half site spacing and orientation (Table 1). The allosteric conformation change that results from steroid receptor binding to a D N A half site subsequently facilitates the binding of a second monomer to the neighbouring half site in reverse orientation. Once bound to the S R E , the two receptors homodimerize by forming salt bridges between the arginine and aspartic acid residues found in the newly exposed D Boxes within each steroid receptor monomer, similarly depicted in the D N A -bound E R homodimer crystal structure (Figure 3A) (Luisi et al, 1991; Schwabe et al, 1993). 8 Sequence of P and D Boxes from select members of the steroid receptor and E R / T R subfamilies Receptor P B o x D B o x Steroid Receptors G R , M R , P R A R G S C K V G S C K V A G R N D A S R N D E R / T R Subfamily E R T R a E G C K A E G C K G P A T N Q K Y D S C Table 1. Nuclear receptor superfamily P Box and D Box sequence comparison. Amino acid sequence variation and similarity of the P Boxes and D Boxes among select members of the nuclear receptor superfamily are shown (adapted from Umesono and Evans, 1989). The facilitation of D N A binding between two steroid receptor monomers to an S R E is termed cooperative binding. Essentially, cooperative interactions between the two steroid receptor monomers are facilitated by the two D N A half sites within an S R E , resulting in greater-than-additive protein occupation (Tsai et al, 1988; Holmbeck et al, 1998). Cooperative binding has the effect of increasing overall binding affinity since protein occupation of the D N A binding site w i l l occur more rapidly than i f occupation was merely additive or linear. Furthermore, the DNA-facilitated cooperative binding o f two steroid receptor monomers w i l l ensure that the two half sites are saturated at a lower concentration of protein than i f there was no facilitation between D N A half sites. A l l four steroid receptors ( A R , PR, G R and M R ) w i l l bind cooperatively to an individual S R E with high affinity and sensitivity (Majors et al, 1983; Ham et al, 1988; Rundlett et al, 1995). There appears to be, therefore, a universal mechanism by which the members of the steroid receptor family recognize and cooperatively bind to a common D N A binding site. Early data on the D N A binding characteristics of the steroid receptors G R , M R , P R and A R revealed that all four are capable of binding in vitro with high affinity to the S R E and all four are capable of transactivating from the S R E in transient transfection assays (Majors et al, 1983; Ham et al, 1988; Rundlett et al, 1995). These observations should not come as a surprise since the steroid receptor D N A binding domains (DBDs) possess remarkably high homology to each other. A s well , the M M T V viral promoter, from which the S R E is derived, would likely benefit from exploiting a variety of endocrine signaling pathways. Such promiscuity o f response is l ikely a desirable trait for a viral promoter in order to capitalize on a host's cellular system: However, in mammalian cells each of the four steroid receptors specifically target a unique set of genes and the resulting gene expression culminates in quite different phenotypes. The dilemma, therefore, is in understanding how the A R specifically and uniquely regulates genes when it appears to recognize and bind to the same D N A binding element as the other three steroid receptors. 10 1.5 Five Levels Through W h i c h AR-specif ic Gene Regulation C a n Occur . There are at least five levels through which AR-specific gene regulation can occur in light of the above binding site dilemma. They are as follows: (1) the availability o f constituents such as activated hormone and its activated cognate hormone receptor; (2) the role of tissue specific coregulators; (3) the accessibility to chromatin; (4) the role of other transcription factors, and finally; (5) the role of the D N A architecture of natural A R -regulated promoter and enhancer regions. Firstly, the availability of constituents would be an obvious solution to AR-specific activation. I f only androgen and activated A R were available within a cell then receptor competition for D N A binding sites would be nonexistent. Reality denies this simple solution, however, since PR, G R and A R and their cognate ligands are found to coexist in many cell environments and activated G R tends to be present in a much higher concentration than A R (Stumpf et al, 1976; Chaudhuri et al, 1981; Saatok et al, 1984; Konrad et al, 1998; Yamashita et al, 2001). The second solution is that within the cellular environment, coregulators exist that are capable of directing an AR-specific response. In other words, unique coregulators direct and modulate not only A R - D N A interactions but also the subsequent protein contacts that are necessary for an AR-specific transcriptional response. The search for unique AR-specific and tissue-specific coregulators is the current focus of a tremendous amount o f research within the steroid receptor field. Many proteins have been found that interact with the A R although there has yet to be discovered a coregulator that uniquely and specifically interacts with the A R . Furthermore, the majority of research thus far involves using molecular techniques that fail to take into consideration the contribution of D N A to the AR-coregulator interaction. From inference from other nuclear receptors such as E R and G R , we know that upon binding to D N A the A R undergoes a conformation change that has functional significance, primarily the effect of receptor homodimerization and cooperative binding to an A R E (Liao et al, 1999). Using an immunoprecipitation assay or GST-pul l down assay w i l l not reveal critical coregulators that interact with the A R while it is bound to an A R E . Furthermore, such assays w i l l not reveal critical coregulator interactions while the A R and the D N A are in different and transcriptionally relevant conformations. 11 One study, however, has examined the effect of the DNA-bound A R conformation on subsequent coregulator and other transcription factor interactions. This study showed that the A R w i l l interact with the octamer transcription factor (Oct-1) when both transcription factors are bound to their respective response elements within the mouse sex limited protein (Sip) gene and that this interaction is dependent on the DNA-bound state o f both proteins (Gonzalez and Robins, 2001). Furthermore, this group showed that the steroid receptor coactivator-1 (SRC-1) is actively recruited to the O c t - l / A R / D N A complex. These results were specific to the A R since the DNA-bound G R was unable to interact with Oct-1 and subsequently failed to recruit SRC-1 even though G R was capable o f binding the Sip A R E . While these results show a role for coregulators in effecting and modulating an AR-specific response they also suggest that the D N A context is essential for the correct hormonal response. Therefore, the exclusive role of coregulators in AR-specific transactivation has yet to be shown and is probably unlikely. The third solution pertains to D N A accessibility. AR-specific regulation of target genes relies on the accessibility of regulatory regions. G R , P R or A R can moderately transactivate from the M M T V promoter in transient transfection assays where the promoter is subject to only randomly arrayed nucleosomes (Richard-Foy H and Hagar G L 1987, Archer T K 1992, Truss M 1995). However, when the M M T V promoter is stably transfected, encased within the genome, and assembled with an ordered nucleosome array, the accessibility and subsequent transactivation response can be narrowed to a single steroid receptor type depending on the site of integration (Lambert and Nordeen, 1998; List et al, 1999) . This suggests that chromatin and accessibility to regulatory regions can play at least a restrictive role in a selective response. The fourth solution pertains to the role of other transcription factors in effecting an AR-specific response. Careful examination of the regulatory regions o f several AR-regulated genes has revealed that the binding of transcription factors other than the A R are critical to the AR-specific response (Scarlett et al, 1995; Darne et al, 1997; N i n g et al, 1999; L u et al, 2000) . The binding of the transcription factor S p l to the p21 promoter has been shown to be essential to the AR-specific regulation of this gene (Lu et al, 2000). L ike coregulators, however, the likelihood is small that a specific transcription factor w i l l be found that is solely responsible for the observed AR-specific response. So far, the transcription factors like Oct-12 1 and S p l that interact with the A R are generally ubiquitous among cell types and involved in the regulation of several genes. The final proposed solution to the observed AR-specific response lies within the architectural context of the natural promoter and enhancer regions found within A R -regulated genes. Architectural context refers to the primary and flanking nucleotide sequence, spacing, and relative orientation of the protein binding sites found within gene regulatory regions. The 'talking D N A ' hypothesis suggests that the architectural context of D N A conveys information to a bound protein through allostery that w i l l dictate conformation and function (Stephen Hendy, personal communication). Allosteric interactions are necessarily two-way, which means that a specific protein bound to D N A w i l l also impart a D N A conformation change that has functional significance. This hypothesis incorporates the roles of coregulators and other transcription factors in generating an AR-specific response: D N A sequence could direct not only the specific transcription factors that bind to neighbouring sites but also the consequent correct conformation of the protein-DNA transcription regulatory multiplex, termed the enhanceosome, which is necessary to recruit coregulators that effect maximal transactivation of the gene in question (Figure 4) (Thanos et al,-1995; Mer ika and Thanos, 2001). These D N A sequence-directed allosteric interactions could function to increase protein binding affinity, sensitivity and specificity through cooperative binding, which w i l l ultimately lead to the synergistic transcriptional activation or repression o f targeted genes (Carey et al, 1998; El lwood et al, 1999; Wang et al, 1999). 1.6 The Architectural Context of AR-Regulated DNA Regions Underlies A R Cooperative Binding When natural promoter and enhancer regions of known AR-regulated genes are examined they are generally found to comprise of multiple A R E s that individually vary in primary nucleotide sequence from the high affinity S R E (Kasper et al, 1994; Cleutjens et al, 1996; Grad et al, 1999; L i n et al, 2000). A n A R E found in a mammalian gene regulatory region possesses nucleotide divergence that results in a relatively low affinity for the A R when compared to the high affinity S R E (Kasper et al, 1999; Verrijdt et al, 2000): This is seemingly a contradictory observation since intuition would suggest that individual A R E s 13 Enhanceosome complex Basal transcriptional machinery Figure 4. The enhanceosome recruits the basal transcriptional machinery. The enhanceosome complex comprises of transcription factors and coregulators cooperatively recruited to the D N A regulatory region.The multicomplex that results is termed the 'enhanceosome', which w i l l recruit the basal transcriptional machinery to the promoter. found within the promoter and enhancer regions of AR-regulated genes would naturally have a high affinity for the A R . This interesting contradiction was the basis of a theory that the nucleotide divergence found within an A R E could have evolved to increase AR-specificity in a competitive steroid receptor environment but at a cost to binding affinity (Nelson et al,. 1999). It was proposed that natural A R E s not only evolved through the selection of nucleotides that contributed additional bond energy but also through the selection of nucleotides that discouraged binding of inappropriate receptors. A binding site selection assay showed that, when in direct competition with either G R or PR, the A R would selectively bind to lower affinity response elements consequently restricting specificity to the A R (Nelson et al, 1999). This study and others also showed that i f the third nucleotide in the core sequence A G A A C A is altered from an adenine to a thymidine or a guanine, then A R - and PR-mediated transcriptional activity increased despite a lowering of D N A binding affinity for both A R and P R (Tan et al, 1992; Lieberman et al, 1993; Nelson et al, 1999). Together these observations suggest that nucleotides, particularly those that are evolutionarily conserved, that do not engage in base-specific bonds can nevertheless affect function. In other studies it has been shown that particular nucleotides are discriminated against. This is likely due to the incompatibility o f a receptor D B D structure with a given sequence, which provides another level of discrimination (Nelson et al, 1999; Zilliacus et al, 1994; Gewirth et al, 1995; Nelson et al, 1995). Overall, there appears to be general core nucleotide requirements for an A R E , but there is also nucleotide variation within the core binding site and flanking sequence that provides function apart from receptor binding affinity. The A R E s that are found within natural AR-regulated regions individually have relatively low affinity albeit high specificity for the A R . Interestingly, like other steroid receptor response elements, when multiple A R E s are examined collectively and in their natural promoter context they cooperatively interact to dramatically increase overall affinity for the A R (Tsai et al, 1989; Schule et al, 1988; Strahle et al, 1988). The rat probasin gene proximal promoter and the human prostate specific antigen (PSA) gene enhancer are two well-studied A R - and prostate-specific gene regulatory regions that contain multiple cooperatively interacting A R E s (Figure 5 and 6) (Kasper et al, 1994; Huang et al, 1999). A s observed with other 15 5 ' A A G C T T C C A C AAGTGCATTT A G C C T C T C C A GTATTGCTGA T G A A T C C A C A -3 77 GTTCAGGTTC AATGGCGTTC A A A A C T T G A T CAAAAATGAC C A G A C T T T A T - 3 2 7 A T T C T T A C A C C A A C A T C T A T CTGATTGGAG GAATGGATAA T A G T C A T C A T - 2 77 G T T T A A A C A T C T A C C A T T C C AG TT AAGAAA ATATGATAGC A T C T T G T T C T - 2 2 7 • * ARE1 T A G T C T T T T T CTTAATAGGG A C A T A A A G C C C A C A A A T A A A A A T A T G C C T G - 1 7 7 AAGAATGGGA CAGGCATTGG GCATTGTCCA TGCCTAGTAA A G T A C T C C A A - 1 2 7 • ARE2 G A A C C T A T T T GTATACTAGA TGACACAATG TCAATGTCTG T G T A C A A C T G - 7 7 CCAACTGGGA TGCAAGACAC TGCCCATGGC A A T C A T C C T G A A A A G C A G C T - 2 7 A T A A A A A G C A GGAAGCTACT CTGCACCTTG TCAGTGAGGT CCAGATACCT A C A G 3 ' Figure 5. The rat probasin promoter from -426 to +28 base pairs. Two A R E s were discovered that cooperatively interacted to increase overall binding affinity for the A R and they are A R E 1 at position -241 to -227 base pairs and A R E 2 at position -136 to -122 base pairs (Kasper et al, 1994). 16 - 4 2 67 A A A C C T G A G A TTAGGAATCC T C A A T C T T A T ACTGGGACAA CTTGCAAACC • < A R E V - 4 2 1 7 TGCTCAGCCT TTGTCTCTGA TGAAGATATT A T C J T C A ^ G A TCTTGGATTG A R E I V - 4 1 6 7 A A A A C A G A C C TACTCTGGAG GAACATATTG TATCGATTGT CCTTGACAGT A R E I I I - 4 1 1 7 A A A C A A A T C T GTTGTAAGAG A C A T T A T C T T T A T T A T C T A G GACAGTAAGC A R E I I I A - 4 0 6 7 AAGCCT Figure 6. The human P S A enhancer from -4267 to -4062 base pairs. Four A R E s were discovered that cooperatively interact to increase A R binding affinity and to direct A R -specific P S A expression. These A R E s are as follows; IIIA from -4065 to -4079 base pairs; III from -4133 to -4148 base pairs; IV from -4175 to -4189 base pairs; and V from -4220 to -4134 base pairs (Huang et al, 1999). 17 steroid receptors, the binding of one A R homodimer to an A R E w i l l facilitate the binding of A R homodimers to neighbouring A R E s (Tsai et al, 1989; Kasper et al, 1994; Cleutjens et al, 1996; Grad et al, 1999; Huang et al, 1999; L i n et al, 2000). The cooperative assembly of protein-DNA complexes is widely observed in both prokaryotes and eukaryotes. It is generally accepted that cooperative binding is universally involved in mediating the assembly of specific transcription factor complexes, which culminate in the specific, synergistic and maximal transcriptional activation o f gene expression (Ankenbauer et al, 1988; Tsai et al, 1989; Ohmori et al, 1997; Vashee et al, 1998 pp 530; Vashee et al, 1998 pp452; El lwood et al, 1999; Merika et al, 2001). The specific facilitation that characterizes cooperativity exhibits three features that contribute to synergistic transactivation and they are as follows: (1) Increased binding affinity of the transcription factor for its DNA binding element. D N A binding sites that exhibit cooperativity facilitate the binding of transcription factors and architectural proteins such that the binding of one protein to its D N A response element assists the binding of other proteins to neighbouring D N A sites. This facilitation results in greater-than-additive or non-linear binding site occupation and directly translates into increased binding affinity (Mao et al, 1994; L i u et al, 1998; Senear et al, 1998); (2) Increased sensitivity of the DNA binding site to transcription factor concentration. Cooperatively interacting sites become occupied at a much lower concentration of transcription factor compared to binding sites that are occupied in a linear and independent manner (Mao et al, 1994; Senear et al, 1998). Independent or unassisted occupation of a D N A binding site would occur at a much higher protein concentration than that required for those proteins assisted by neighbouring DNA-bound proteins. Additionally, this increased sensitivity functions to serve as a molecular switch where small changes in factor concentration w i l l result in the full cooperative occupation of the D N A regulatory region (Liu et al, 1998). Increased sensitivity is a hallmark of true cooperative interactions and its observation w i l l help discriminate cooperative binding from D N A binding sites that possess high affinity but do not interact cooperatively (Carey et al, 1998); and (3) Increased transcription factor specificity for DNA binding elements. Cooperative interactions between transcription factors are typically specific suggesting that only the correct combination of transcription factors w i l l occupy the gene regulatory region (Mo et al, 1998; Bhoite et al, 1998; Senear et al, 1998; Szymczyna and Arrowsmith, 2000). 18 This specificity of interaction arises from allostery. When a protein factor and D N A interact there is a conformation change in both the protein and the D N A that w i l l affect the type of subsequent protein-protein and protein-DNA interactions (Holmbeck et al, 1998; Kerppola et al, 1998). The resultant protein-DNA conformation change would ensure that only the correct factors are recruited and encouraged to bind to D N A . 1.7 Cooperativity Arises From Protein-Protein and Protein-DNA Interactions There are several well-studied eukaryotic cooperative systems that include the Matcc2/MCM1/STE6 yeast a-specific gene promoter complex, the N F A T / F o s - J u n / A R R E 2 mammalian cytokine gene promoter complex, the N F - K B / H M G I/PDII human interferon-p gene (IFN-P) promoter complex, and the E R / E R E vitellogenin gene promoter complex (Thanos and Maniatis, 1992; Zhong et al, 1997; Diebold et al, 1998; Wood et al, 1998; Zhang and Verdine, 1999). A l l o f these cooperative systems have been shown to share several characteristics and they are: (1) protein and D N A conformational changes resulting from allosteric protein-protein and protein-DNA interactions; (2) a flexible protein-protein interaction interface; and (3) an extended protein-DNA interaction interface. In general, investigations of the molecular mechanisms underlying cooperative binding of protein to D N A have focused on the role of protein-protein interactions. Target D N A nucleotide sequence is often described as a protein-docking site. A n y D N A structural distortion that results from the DNA-protein interaction has usually been explained as a consequence of protein-protein interactions rather than concluding that D N A sequence is a contributing mechanism to the resulting protein/DNA structure. For example, the D N A bending observed in both the yeast M A T a 2 / M C M l / S T E 6 and mammalian N F A T l / F o s -Jun /ARRE2 cooperative complexes have been considered the result of cooperative protein binding rather than as an underlying contributing molecular mechanism (Diebold et al, 1998; Kerppola et al, 1998). It is thought that upon binding D N A , N F A T 1 undergoes a conformation change that compels Fos-Jun to bind D N A through protein-protein interactions. Due to this interaction with N F A T 1 , the Fos-Jun heterodimer apparently adopts a new conformation that in turn forces the D N A to bend to accommodate specific Fos-Jun/DNA contacts. 19 This protein-centric perspective arises from the static crystallographic method that is often employed to examine protein-DNA complexes and the dogma that proteins are the functional molecules whereas D N A is merely the genetic code. A crystallographic X-ray structure invariably fails to convey the dynamic and temporal contribution of each molecule to the formation of the overall complex. Additionally, functional studies typically examine the effect of protein mutations on complex formation and often ignore the contribution of D N A sequence. If we consider the protein-DNA interactions that occur within cooperative complexes to be allosteric, then the D N A would conceivably also undergo a conformation change upon protein binding and this D N A conformation change likely has biological significance (Senear et al, 1998). It has also been suggested that cooperative binding of transcription factors in vivo can occur by several mechanisms, some of which do not require direct protein-protein interactions (Vashee et al, 1999). A few researchers have examined the contribution of D N A nucleotide sequence to the formation o f cooperative complexes and subsequent synergistic transcriptional activation. The in vitro selection of non-consensus D N A binding site sequences in the presence of known cooperatively interacting proteins has shown that variation in nucleotide sequence affects the formation of cooperative complexes. A n in vitro binding site selection assay constrained the transcription factor M A T a 2 to select D N A half sites in the presence of the cooperatively interacting M C M 1 protein (Zhong et al, 1997). This assay revealed that the selected Mata2 half sites varied in nucleotide sequence from the consensus high affinity M A T a 2 half site. Intriguingly, the half sites selected in the presence o f M C M 1 resembled the asymmetric natural M A T a 2 half sites found within the cooperative M A T a 2 - M C M l binding sites of yeast a-site genes, such as STE6. The M A T a 2 half site nucleotides selected in this assay did not necessarily make specific contacts with MAToc2 although their mutation significantly decreased both the in vitro cooperative formation of the M A T a 2 / M C M l / S T E 6 complex and its in vivo synergistic transcriptional effects. These results suggest that nucleotides that do not specifically contact M A T a 2 and are not essential to M A T a 2 binding nevertheless contribute to the formation of a cooperative M A T o c 2 / M C M l protein-DNA complex and affect synergistic transactivation. 20 The N F - K B transcription factor is an example of how variation in D N A binding site sequence can structurally alter the bound protein and that this protein conformation change can affect function. N F - K B binds D N A as a dimer that can consist of any combination of five different subunit proteins that are members of the Re l family (p50, p65, p52, c-Rel, and RelB) (Lenardo and Baltimore, 1989; Gilmore et al, 1992). Binding studies of N F - K B binding to D N A revealed that the N F - K B / D N A complex is reliant on a specific ion concentration that is dependent on the combination of N F - K B dimer subunits for a given D N A binding element (Menetski et al, 2000). The apparent dissociation constant of the protein complex is a function of ion concentration and p H . Measuring the ion release after the protein-DNA complex has come to equilibrium provides information on the structural differences for the same N F - K B dimer bound to D N A sites that vary in nucleotide sequence (deFfaseth et al, 1977, Menetski et al, 2000). When the formation of a p50 homodimer N F -K B / D N A complex was compared using two different D N A binding site sequences, the number of ions released from the protein-DNA complex was significantly different depending on the D N A binding sequence (Metetski et al, 2000). This change in ionic strength upon complex formation implied that there was a change in structure of the bound p50 homodimer protein that depended on the sequence of the D N A binding site (Menetski et al, 2000). N o changes were observed in D N A structure upon N F - K B binding, although the technique used (the net change in end-to-end distance of an oligonucleotide upon N F - K B binding) may not have been sensitive enough to detect local structural changes in D N A . Protein structural changes that were dependent on the nucleotide sequence of the D N A binding site were also observed within the p65 homodimer N F - K B - D N A X-ray crystal structure (Chen et al, 1998). The variable effect of D N A sequence on N F - K B protein structure could explain why some N F - K B DNA-binding sites but not others promoted cooperative interactions with other transcription factors (Thanos and Maniatis, 1992). The human IFN-P and immunoglobulin-K B ( I g - K B ) are two genes that are regulated by N F - K B . However, cooperative interactions between N F - K B and the architectural protein H M G I are observed within the IFN-P but not the I g - K B regulatory region. H M G I belongs to the high mobility group ( H M G ) protein superfamily whose members recognize and bind to either specific D N A sequence or D N A 21 tertiary structure (Laudet et al, 1993). H M G I is a sequence-specific architectural protein that specifically binds AT- r i ch D N A but does not have any intrinsic transactivation function. Instead these architectural proteins bind the minor groove of AT- r i ch D N A and intercalate a hydrophobic residue into the D N A structure, which causes the D N A to bend towards the minor groove (Churchill and Travers, 1991; Werner et al, 1995; Love et al, 1995). This architectural rearrangement of the D N A functions to open the major groove facilitating the access of a transcription factor, such as N F - K B (Yie et al, 1997). H M G I minor groove binding to the E F N - P promoter greatly facilitates the binding of N F - K B to the major groove (Thanos andManiatis, 1992). This cooperative interaction was dependent on the AT- r i ch flanking and core binding sequence of the H M G I DNA-bind ing site within the IFN-fi promoter (Zhang and Verdine, 1999). This AT- r i ch sequence is absent within the N F - K B core binding site and flanking regions of the I g - K B gene, which does not display cooperativity between N F - K B and H M G I. H M G I and N F - K B have been shown to interact using a G S T pull-down assay. The mutation of several H M G I alanine residues, which are not necessary for D N A binding, w i l l not only disrupt the interaction between H M G I and N F - K B but also the cooperative D N A -binding between these two proteins (Du et al, 1993; John et al, 1995; Leger et al, 1995; Zhang and Verdine, 1999). Intriguingly, the separate N M R structures o f the H M G I bound to its D N A element and the N F - K B bound to its D N A element contradict the above N F K B -H M G I interaction data. These two separate N M R structures reveal that the proposed N F K B -H M G I interactions would be highly improbable i f the proteins were within proximity of each other and bound to a common D N A molecule (Huth et al,1997; Zhang and Verdine, 1999). Therefore, the protein-protein interactions that appear to be necessary for cooperative binding likely arise from the D N A structural changes that result from both N F - K B and H M G I binding to D N A (Zhang and Verdine, 1999). The crystal structure of N F - K B and H M G I together bound to the same D N A molecule has yet to be published, yet the above observations suggest that N F - K B and H M G I could mutually cooperate by reciprocally altering D N A structure. This implies that the specific nucleotide sequence o f the binding site w i l l favor N F K B - H M G I cooperative interactions whereas a different nucleotide sequence would fail to adopt the correct structure. 22 Within the nuclear receptor field, D N A binding site sequence has been shown to dictate the conformation of a bound E R homodimer (Wood et al, 1998). Partial digestion of the bound E R homodimer using trypsin and chymotrypsin revealed distinct cleavage patterns that varied with the estrogen response element (ERE) nucleotide sequence. Additionally, using naturally occurring EREs , three different ER-specific antibodies interacted differentially with the E R depending on the nucleotide sequence o f the E R E . These data show that the nucleotide sequence of the E R E dictated the conformation o f the bound E R . While functional studies were not done on these variant bound E R conformations, the results suggest that nucleotide sequence could dictate function. Conceivably the variations in bound E R conformations could permit or restrict subsequent transcription factor and/or coregulator interactions that w i l l ultimately contribute to transcriptional regulation. 1.8 Probasin Promoter and the P S A Enhancer Possess M u l t i p l e Cooperatively Interacting A R E s 1.8.1 The Rat Probasin Promoter Rat prostatic basic protein (probasin) was first isolated from to the dorsolateral lobes of the rat prostate (Matuo et al, 1982). This region of the rat prostate is analogous to the peripheral zone of the human prostate, which is the most common site for prostate cancer. Since its first isolation, a homologue of rat probasin was found in mouse dorsolateral prostate although a human homologue has yet to be found (Genbank AF005204). Probasin expression was further delineated to the lateral lobe where it has the highest expression then, in descending order of expression, the dorsal lobe, anterior lobe, ventral lobe, and seminal vesicles (Matuo et al, 1986, Matusik et al, 1986). Although probasin is primarily secreted in high levels, weak protein staining was found in the nuclei o f luminal epithelial cells (Matuo et al, 1985). There is only one m R N A transcribed from the probasin gene but there are alternate translation initiation codons that explain the observed different protein localizations. One translation initiation codon permits the translation of a signal peptide permitting membrane localization whereas the second initiation codon does not. After post-translation modification, however, the mature secreted and nuclear probasin proteins are identical 23 suggesting a bifunctional role for the protein that is dependent on location rather than amino acid sequence (Spence et al, 1989). A minimal probasin promoter region from -426 to +28 base pairs was enough to direct expression specifically to prostate cancer cell lines such as L N C a P , PC3 and DU145, whereas reporter gene expression was very low in non-prostate cell lines such as HepG2, B L 1 3 (bladder cancer) C H O , M C F - 7 (breast), 293 and M R C 5 (Brookes D E 1998). Expression driven by the -426 to +28 base pair probasin promoter fragment was shown to be androgen-specific in the PC-3 human prostate cancer cell line in that cotransfected A R with androgen, but not G R with dexamethasone, resulted in chloramphenicol acetyl transferase ( C A T ) gene expression (Rennie et al, 1993; Kasper et al, 1994). However, in L N C a P cells reporter gene expression from this probasin promoter fragment was inducible using G R with dexamethasone, although levels of transcription were much lower than those achieved with equivalent concentrations of A R and R1881, the synthetic androgen methyltrienolone (Rennie et al, 1993). The proximal probasin promoter and the large probasin (LPB) promoter (~12 kilobase pairs) were also shown to be prostate-specific in vivo. Using transgenic mice, the -426 to +28 base pair probasin promoter region or L P B promoter successfully targeted several reporter gene products, (such as the bacterial reporter gene C A T or the simian virus 40 (SV40) large T-antigen), to mouse prostate epithelial cells (Greenberg et al, 1994; Green et al, 1998). Probasin gene expression is androgen regulated in that castration results in a dramatic drop in probasin m R N A levels (Dodd et al, 1983). Transgene expression driven by the L P B promoter increased with increasing serum androgen levels and stabilized when the mice reached sexual maturity. Like probasin m R N A expression levels, transgene expression levels driven by the L P B promoter dropped off after castration and levels were rescued after treatment with androgens (Greenberg et al, 1994). DNase I protection assays revealed two androgen response elements (AREs) within the probasin promoter between positions -236 to -233 and -140 to -117 base pairs ( A R E l and A R E 2 respectively) (Figure 6) (Rennie et al, 1993; Kasper et al, 1994; Kasper et al, 1999). These A R E s varied considerably from the canonical S R E nucleotide sequence and Scatchard analyses revealed that the A R E s individually have weak A R binding affinity 24 compared to the high affinity S R E , although A R E 2 had markedly higher binding affinity for the A R relative to A R E 1 (K<i values of 6.7 n M versus 20 n M respectively) (Kasper et al, 1994). Together, the probasin A R E 1 and A R E 2 were shown to interact cooperatively such that the A R binding affinity was increased dramatically and occupation of these two sites occurred at a much lower concentration of A R relative to either A R E 1 or A R E 2 individually (Kasper et al, 1994; Kasper et al, 1999). A R E 2 was shown in particular to be specifically and uniquely recognized by the A R although G R was capable of binding both A R E 1 and A R E 2 but with extremely low binding affinity (Kasper et al, 1999; Schoenmakers et al, 1999). When the nucleotide sequence o f the steroid response element within the C(3) gene was changed to that of the probasin A R E 2 , then C(3) expression narrowed from dual A R / G R control to AR-specific control. This phenomenon suggested that nucleotides within A R E 2 conferred AR-specificity by preventing G R from binding (Claessens et al, 1996). Another interpretation of this data is that the nucleotide sequence of A R E 2 within the C(3) promoter context dictated specific function to the bound A R , but not to a bound G R , which permitted subsequent coregulator and/or other transcription factor interactions that mediated the observed AR-specific transcriptional response. Disruption of either A R E 1 or A R E 2 destroyed cooperativity, substantially reduced A R binding to both elements, and also reduced androgen-dependent expression by more than 95% from a transiently transfected probasin promoter (-426 to +28 base pairs) driven C A T . reporter gene (Kasper et al, 1994; Kasper et al, 1999). Finally, removal of the intravening wildtype sequence also disrupted androgen-dependent expression in transfection studies suggesting that either the spacing between the A R E s was critical or that another binding site existed that was necessary for an androgen-directed synergistic transcriptional response (Kasper et al, 1994). 1.8.2 The Human PSA Enhancer Human prostate specific antigen (PSA) is expressed at high levels in the prostate luminal epithelial cells and is either absent or expressed at very low levels in other tissues (Aumuller et al, 1990). Clustered with two other members of the glandular kallikrein gene 25 subfamily of serine proteases (hGK-1 and K L K - 1 ) on chromosome 19q, P S A gene expression is tightly androgen-regulated (Riegman et al, 1989; Evans et al, 1988; Young et al, 1992; W o l f et al, 1992). Within the proximal P S A gene promoter, A R E I was discovered at position -170 base pairs (Reigman et al, 1991) and a 35 base pair androgen responsive region was identified starting at position -400 base pairs, which contains A R E II at position -392 base pairs (Cleutjens et al, 1996). This proximal P S A promoter region conferred androgen-responsiveness, however, it was not AR-specific. G R with dexamethasone was able to stimulate transcription from the P S A promoter fragment to levels similar to those achieved with A R and R1881in transient transfection assays although A R and G R induced expression was reasonably specific to the prostate cancer L N C a P cell line (Cleutjens et al,1996). In contrast to the proximal probasin promoter, the proximal P S A promoter starting at -630 base pairs was inadequate to confer androgen-responsiveness or target gene expression to the prostate of transgenic mice (Cleutjens et al, 1997 ppl256). A n enhancer element located approximately 4.2 kilobase pairs upstream o f the P S A gene transcription start site, in concert with the proximal P S A promoter, was found to be necessary for both androgen-responsiveness and prostate-specific expression in transgenic mice (Cleutjens et al, 1997 ppl48; Schuur et al, 1996; Pang et al, 1997; Cleutjens et al, 1997 ppl256). The core AR-regulated enhancer region spans 455 base pairs and contains at least four A R E s (III, IIIA, IV and V ) at positions -4079, -4143, -4179, -4225 base pairs (Figure 7) (Cleutjens et al, 1997 ppl48; Huang et al, 1999). Each A R E was found to intrinsically possess weak affinity for the A R as determined by DNase I quantitative footprinting, however, collectively these A R E s cooperated to enhance AR-binding affinity and sensitivity to A R concentration and to contribute to synergistic transactivation (Huang W 1999). When any single A R E out of the four was mutated to destroy A R binding, transcriptional activity decreased by 50-75% relative to the wildtype enhancer. The P S A enhancer region, when coupled with the proximal P S A promoter, was determined to be AR-specific in that A R was able to induce transcriptional activity in both prostate cancer cell lines and non-prostate cell lines whereas the P R failed to induce transcription from this construct (the G R was not tested) (Huang et al, 1999). 26 1.9 Objectives of this Study The cooperativity studies described in section 1.7 illustrate that nucleotide variation within transcription factor binding sites can dictate conformation of the bound protein and that the nucleotides involved need not make base-specific contacts to do so. The objective of this project was to characterize the allosteric protein-DNA interactions that underlie the cooperativity observed between A R E s and to determine the critical architectural A R E context that results in synergistic AR-specific transcriptional regulation. The probasin promoter and the P S A enhancer are specifically regulated by the A R and specifically expressed in the prostate. Both gene regulatory regions possess multiple A R E s that cooperatively interact to ensure an AR-specific response. These characteristics recommend both the probasin promoter and P S A enhancer as excellent systems for the study of cooperative A R - A R E interactions. The allosteric protein-DNA interaction that results in a protein conformation change w i l l also effect a conformation change in the D N A . Studying protein-DNA interactions usually involves resolving the complex either through X-ray crystallography or through employing solution-based chemical or enzymatic footprinting techniques. Crystallographic studies are time consuming and labour intensive but nonetheless provide critical information concerning the weak interactions between protein and D N A . The limitations of crystallography however, concern the resulting static nature of the protein-DNA complex and the confined conditions that are necessary to generate a resolved structure. Chemical or enzymatic probing of protein-DNA contacts, in contrast, permit the resolution of weak interactions in a realistic solution-based and dynamic system. Enzyme probes, such as DNase I, can be employed to probe global structural distortion and show protected and hypersensitive regions of D N A that result from protein binding. Chemical probes, such as dimethylsulfate ( D M S ) , can be employed to probe specific nucleotide contacts that are involved in protein binding in both the major and minor grooves of D N A . Additionally, employing D M S to probe protein-DNA contacts also provides information on D N A structural distortion that results from protein-binding (Chan et al, 1990, Clark et al, 1990, Frappier et al, 1992; Espinas et al, 1995; Ramesh and Nagaraja, 1996; Reid and Nelson, 2001). 27 To examine cooperative protein-DNA interactions within the rat probasin promoter and human P S A enhancer, methylation protection (MeP) and interference (Mel) footprinting • assays were employed. In the process of footprinting the cooperative probasin promoter from -426 to +28 base pairs, it was discovered that neither M e l nor the conventional M e P assay was successful in generating a footprint. A n alternative and novel M e P assay was devised, which permits the visualization of protein-DNA contacts within highly cooperative systems (Reid and Nelson, 2001). Using this modified M e P protocol, two novel A R E s were discovered within the proximal probasin promoter. The novel A R E s , called G - l and G-2, displayed an unusual protection and D M S hypersensitivity pattern compared to conventional A R E s that suggested a structural change in D N A upon A R binding. Binding assays were employed to determine the dissociation constants of individual A R E s and the relative binding affinity for multiple A R E s . Transient transfection assays, using an ARE-dr iven luciferase reporter construct in L N C a P cells, were employed to determine whether the in vitro results had any significance in vivo. Additional work concerning A R - A R E interactions was completed that contributes to the evolving understanding of how A R specifically regulates its target genes. Upon closer examination of the nucleotide sequence of both G - l and G-2, there exists a potential overlapping binding sequence for the architectural protein Sox-4, which belongs to the sequence-specific H M G - b o x architectural protein superfamily described earlier. Applying the M e P assay revealed that the addition of Sox-4 to the A R binding reaction increased the number of A R occupied sites within the probasin promoter. Lastly, the A R - D N A contacts within other AR-regulated gene regions such as the mouse epididymous retinoic-acid binding protein ( m E - R A B P ) promoter, the enhancer region within the rat A R gene, the upstream rat probasin promoter, and the human p21 proximal promoter were examined using the modified M e P assay. 28 2. Experimental Procedures 2.1 Methylation Protection and Methylation Interference of the A R DBD In vitro methylation-based footprinting (MeP) is used to identify the protein-DNA contacts at the nucleotides guanine or adenine, made by a DNA-bind ing protein (Figure 7) (Brunelle and Schleif, 1987). M e P can also be used to detect local structural D N A distortion caused by allosteric interactions with the transcription factor (Chan et al, 1990; Clark et al, 1990; Frappier et al, 1992; Espinas et al, 1995; Ramesh and Nagaraja, 1996; Reid and Nelson, 2001). Dimethylsulfate (DMS) is typically employed since this chemical w i l l uniformly methylate guanines in the N 7 position in the major groove and, less efficiently, adenines in the N3 position in the minor groove, unless the nucleotides are successfully protected from methylation by protein contact. Protection from D M S methylation does not directly measure weak interactions between amino acids and nucleotides. However, since the D M S molecule is very small, i f its methylation activity is blocked then a weak interaction such as a hydrogen bond is inferred. Using M e P , guanines and adenines are protected from D M S attack after the protein-DNA complex is formed. Local structural distortion that results from protein binding w i l l manifest as hypersensitivity to D M S methylation (Chan et al, 1990; Clark et al, 1990; Frappier et al, 1992; Espinas et al, 1995; Ramesh and Nagaraja, 1996; Reid and Nelson, 2001). In vitro methylation-interference (Mel) is used to identify those guanine or adenine contacts that are necessary for protein binding (Figure 8). In M e l the D N A is pre-methylated with D M S using one-hit kinetics, purified, then bound to the protein of interest. The free and bound D N A are segregated using an electrophoretic mobility shift assay ( E M S A ) . If a particular pre-methylated guanine or adenine is necessary for the DNA-prote in interaction then that D N A molecule w i l l be segregated to the free fraction of D N A and w i l l be correspondingly rejected from the bound fraction of D N A . Unlike M e P , M e l does not provide information about the D N A structural distortion that results from allosteric interactions with transcription factors since the D N A was methylated before its introduction to protein. In addition, M e l cannot always be used for cooperative D N A binding interactions because a solitary disrupted DNA-protein interaction can be 29 32P-labelled D N A probe A R D B D protein D M S J J CH3. C H 3 1 CH3 Piperidine o r N a O H Cleavage Unbound Bound A R D B D Footprint Figure 7. Schematic of the methylation protection assay. The M e P assay involves the methylation of protein-bound D N A followed by chemical cleavage of methylated guanines or adenines. Cleaved D N A is visualized using denaturing P A G E . 30 Pre-methylated C H 3 32P-labelled ^ £22 D N A probe ^ CId3_ C H 3 A R D B D protein j J ,CH3 M l C H 3 C H 3 Separation of bound and free D N A populations by P A G E followed by piperidine or N a O H cleavage 1 Unbound Free Bound — A R D B D Footprint Figure 8. Schematic of the methylation intereference assay. The M e l assay requires the D N A to be pre-methylated using D M S then the purified probe is used in a protein-binding assay. Bound and free D N A populations are separated by P A G E followed by chemical cleavage. Cleaved D N A is visualized using denaturing P A G E where those nucleotides important for protein-DNA binding are enhanced in the free population. compensated for by protein-protein and other neighbouring protein-DNA interactions. Therefore, the methylation of a single necessary guanine or adenine is not enough to dislodge the protein from the D N A . For multiple DNA-binding elements that exhibit cooperative binding, M e l w i l l not provide DNA-protein contact information whereas M e P is very effective at investigating cooperative D N A binding complexes. 2.1.1 Preparing Recombinant DNA Binding Protein The rat A R D B D (amino acids 524 through 648) and the full length human Sox-4 protein (52 kDa) were cloned into the EcoRl and BamHI sites of the plasmid vector pTrcHisC (Invitrogen) and transfected into Escherichia Coli (strain JM109) cells. To express the recombinant six-histidine N-terminal fusion protein, 5 mLs o f L-Broth with 100 pg/ p L of ampicillin were inoculated with transformed JM109 cells containing the recombinant pTrcHisC vector. After incubating overnight while shaking at 250 R P M at 37°C, 200 p L (50:1) of the overnight culture were sub-cultured into 100 mLs o f L-Broth with 100 pg/ p L of ampicillin and incubated at 37°C shaking at 250 R P M . When log growth was reached (approximately 3.0 hours), as measured by an optical density (O.D.600) measurement of 0.6, expression of the recombinant protein was induced by adding 0.1 M of I P T G (isopropylthio-p-D-galactoside). After 3.0 hours induction, the JM109 cells were pelleted and resuspended in 3.0 mL/gram Lysis Buffer (50 m M N a H 2 P 0 4 (pH 8.0); 300 raM N a O H ; 10 m M imidazole) with 1.0 mg/mL Lysozyme (Gibco). Cells were lysed by sonication, pelleted, and then the cleared lysate was resuspended in 1.0 m L of N i - N T A slurry (Qiagen) for every 4.0 mLs of cleared lysate. To purify the recombinant FfisTag protein from the other proteins in the cell extract, a nickel column is used to specifically bind the histidine residues located on the N-terminus of the recombinant protein of interest. To elute the recombinant histidine tagged protein off the column, imidazole is used to compete for the nickel-binding sites. Briefly, the cleared lysate and N i - N T A slurry was mixed gently for 1.0 hour at +4°C then transferred to an empty column. The flow-through was permitted to exit by gravity flow, then the column was washed once with 50 m M N a H 2 P 0 4 (pH 8.0), 300 m M N a C l , and 20 m M imidazole to rid the column of non-specifically bound proteins. The column was washed once with 50 m M 32 N a H 2 P 0 4 (pH 8.0), 100 rriM KC1 , and 20 m M imidazole and then once again with 20 m M H E P E S (pH 7.9), 100 m M KC1, and 20 m M imidazole. The recombinant protein was finally eluted off the n ick le -NTA column using 20 m M H E P E S (pH 7.9), 100 m M K C 1 , 25% glycerol and 250 m M imidazole. Ce l l extract, washes and recombinant protein fractions were visualized and confirmed by S D S - P A G E followed either by staining with Coomassie Brilliant Blue ( B R L ) or Western blotting using a polyclonal antibody (Invitrogen) directed against the six histidines located at the N-terminus of the recombinant protein. Protein concentrations were determined using the Bradford Assay and B C A protein assay and the protein fractions were stored at -85°C (BioRad). 2.1.2 Generation of Radiactively Labelled DNA Probes To generate single-end labelled probes of the rat probasin promoter, the human P S A enhancer, the rat A R enhancer, the human p21 promoter, or the m E - R A B P promoter one of three general methods was employed. The different protocols were as follows: (1) Single-end labelled fragments were generated using polymerase chain reaction (PCR) amplification with one of the two oligonucleotide primers end-labelled with 3 2 P (see section 2.1.2a). (2) Single-end labelled probes of the probasin promoter, p21 promoter, P S A enhancer, or m E - R A B P promoter were generated by sequentially cutting the fragment of interest out of a recombinant plasmid vector (see section 2.1.2b). The plasmid was initially linearized using a restriction endonuclease then end-labelled w i t h 3 2 P using either polynucleotide kinase (PNK) (NEB) or Klenow D N A polymerase (NEB) . Lastly, the fragment insert was cut out using another restriction endonuclease; or (3) Finally, single end-labelled D N A probes of either the wildtype or mutated G - l element derived from the rat probasin promoter were generated by annealing synthesized complementary oligos. Wildtype or mutated G - l oligos were synthesized at the Nucleic Acids and Protein Synthesis laboratory (N A PS) at the University of British Columbia. Oligos were end-labelled w i t h 3 2 P and annealed to a cold complementary oligo to generate a single-end labelled double-stranded D N A probe. A l l three end-labelling protocols employed polyacrylamide gel electrophoresis ( P A G E ) to purify the probes. 33 2.1.2. a Generating Single-end Labelled DNA Probes Using PCR Amplification Eight different probasin promoter probes and one rat A R enhancer probe were generated using P C R amplification. The -426 to +28 base pair probasin promoter insert in the pBluescript plasmid vector and the entire rat A R gene (~3.4 kbp) in the pBluescript plasmid vector were used as D N A templates (Stratagene). The probasin promoter fragments generated by P C R were also inserted into the p T K L u c expression plasmid and used to generate the cloned versions of the fragments and for transfection studies (refer to section 2.3). The P C R primers were synthesized at N A P S U B C . The primers and resulting fragments are as follows {note: to facilitate cloning, the probasin fragments that were subsequently cloned into the p T K L u c expression vector were P C R generated, using the following primers synthesized with BamHVHinDlll linkers): (1) Probasin A R E 1 - G 1 - A R E 2 -G2: -269 through -77 base pairs using forward and reverse primers 5' C A T C T A C C A T T C C A G T T A A G 3' and 5' C A G T T G T A C A C A G A C A T T G A 3' (2) Probasin G 1 - A R E 2 -G2: -229 through -77 base pairs using forward and reverse primers 5' T C T T A G T C T T T T T C T T A A T A G G 3' and 5' C A G T T G T A C A C A G A C A T T G A 3' (3) Probasin A R E 1 -G l : -269 through -164 base pairs using forward and reverse primers 5' C A T C T A C C A T T C C A G T T A A G 3' and 5' C T G T C C C A T T C T T C A G G C 3' (4) Probasin A R E 2 - G 2 : -150 through -77 base pairs using forward and reverse primers 5' G T C C A T G C C T A G T A A A G 3' and 5' C A G T T G T A C A C A G A C A T T G A 3' (5) Probasin A R E 1 : -269 through -210 base pairs using forward and reverse primers 5' C A T C T A C C A T T C C A G T T A A G 3' and 5' T A T T A A G A A A A A G A C T A 3' (6) Probasin A R E 2 : -150 through -105 base pairs using forward and reverse primers 5' G T C C A T G C C T A G T A A A G 3' and 5' C A T C T A G T A T A C A A A T A 3' (7) Probasin G - l : -229 through -164 base pairs using forward and reverse primers 5' T C T T A G T C T T T T T C T T A A T A G G 3' and 5' C T G T C C C A T T C T T C A G G C 3' (8) Probasin G-2: -121 through -77 base pairs using forward and reverse primers 5' T A T T T G T A T A C T A G A T G 3' and 5' C A G T T G T A C A C A G A C A T T G A 3'; and (9) rat A R Enhancer 2028 to 2333 base pairs using forward and reverse primers 5' C A T T G A A G G C T A T G A A T G T C 3' and 5' C G G T A C T C A T T G A A A A C C A G 3'. Before generating single-end labelled P C R generated fragments, either the forward or the reverse primer had to be end-labelled using P N K (New England Biolabs (NEB)) . Briefly, 34 10 u M of primer was incubated with 50 p C i of y P - A T P (Amersham) with I X P N K buffer (NEB) and 1.5 Units of P N K (NEB) for 1 hour at 37°C. The end-labelling reaction was then incubated at 75 °C for 20 minutes to heat denature the kinase. To rid the reaction of unincorporated radionucleotides, the sample was centrifuged (2800Xg for 2 minutes) through a G25 tris-sepharose column (Amersham). To P C R amplify each probe, 100 ng of the D N A template was added to a 25 p L volume containing 200 p M of deoxynucleotide triphosphates (Gibco), 1.5 m M MgC12, I X P C R Buffer (Gibco), 10 p M each of the radioactively labelled primer and the cold reverse primer, and 2.5 Units of Taq D N A polymerase (Gibco). Using the M J Research D N A Engine Thermocycler, each sample was initially denatured at 95°C for 2.0 minutes then for 25 cycles the following steps: 95°C for 1.0 minute, 50°C for 30 seconds, and 72°C for 30 seconds. After all the cycles were completed the reaction was incubated at 72°C for 1.0 minute to ensure all fragments were fully extended. After adding loading buffer containing bromophenol blue and xylene cyanol, the entire reaction volume was loaded onto a 0.75 mm 5% P A G E (29:1 acrylamide:bis (BioRad)) with I X T B E (90 m M Tris Base, 90 m M Boric A c i d , and 2 n M Disodium E D T A ) . The P A G E was run at 16 V / c m at room temperature until the expected fragment size had run well into the gel as measured by the migration of the marker dyes - bromophenol blue and xylene cyanol. The gel was covered in plastic film and exposed to Kodak B i o M a x Sensitive ( B M S ) film for approximately two minutes at room temperature. Using the developed film as a template, the fragments were cut out of the gel and the gel slices were placed into a fresh 1.7 m L silanized tube and crushed using a blue P1000 pipette tip. To elute the labelled D N A from the gel slice, 500 p L o f Maxam and Gilbert Elution Buffer (0.6 M ammonium acetate, 1 m M E D T A and 0.1% SDS) was added to the crushed gel slice and the tube was rotated overnight at room (Maxam A M and Gilbert W 1977). The following day the acrylamide gel slice was briefly pelleted in a tabletop centrifuge and the eluant was transferred to a fresh 1.7 m L silanized tube using a blue PI000 pipette tip with its end clipped off. The eluant was then centrifuged through a BioRad miniprep spin column to filter out any remaining acrylamide pieces. To ethanol-precipitate the end-labelled D N A , 1.0 m L of ice-cold 95% ethanol was added to the eluant. The tube was inverted to mix and placed in a -20°C freezer for ten minutes to help precipitate the 35 D N A . The tubes were then centrifuged at maximum R P M (>10 OOOXg) on a tabletop centrifuge for thirty minutes at 4°C. The ethanol was carefully removed and the D N A pellet was washed twice with 70% ethanol with five-minute centrifuges between washes. After careful removal of all traces of ethanol, the D N A pellet was air-dried briefly then resuspended in D N A Binding Buffer D (20 m M H E P E S (pH 7.9), 100 m M KC1, and 10% glycerol). The radioactive disintegrations per minute (dpm) for 1.0 u L of end-labelled D N A fragment were determined using a Beckman scintillation counter. 2.1.2.b Generating Single-end Labelled DNA Probes Through Successive Enzymatic Digestion of Recombinant Plasmid Vectors To generate end-labelled rat probasin promoter probes (derived from the -426 to +28 base pair fragment insert), the pBluescript or p T K L u c plasmid (plasmid courtesy o f Dr. S. Nordeen, University of Colorado) constructs had to be initially linearized using either HinDIll or BamBI (NEB) . To generate the -705 to -426 base pair end-labelled rat probasin promoter probe from the -705 to +28 base pair fragment insert, the pGEMteasy plasmid construct (courtesy of Dr. R. Matusik, Vanderbilt University, Te) had to be linearized using either Nhel or HinDIll (NEB) . To generate the downstream end-labelled m E - R A B P promoter from the -198 to +28 base pair fragment insert, the pBluescript plasmid construct (construct courtesy Dr. J. Lareyre, Vanderbilt University, Te) had to be linearized using either HinDIll or Xbal (NEB) . To generate the upstream end-labelled m E - R A B P promoter from the -543 to -166 base pair fragment insert, the pBluescript plasmid construct (Dr. Lareyre) had to be linearized using either HinDIll or Xbal (NEB) . If end-labelling using P N K , the 5' phosphorous had to be removed and this was done post-linearization by adding 1.5 Units of calf intestinal phosphotase (CIP) (NEB) to the restriction digest reaction. The dephosphorylation reaction was incubated for one hour at 37°C then the CIP and restriction enzymes were heat denatured for 20 minutes at 75°C. To help rid the sample of enzyme protein, a phenol/chloroform extraction was employed. The digest was brought to a volume of 100 u L using sterilized distilled deionised water (sddH20) and then 100 p L o f phenol/chloroform/isoamyl alcohol (29:1:1) was added. The organic and aqueous solutions were vigorously vortexed then centrifuged 10 OOOXg for five minutes. 36 The top aqueous layer was transferred to a fresh 1.7 m L silanized tube and 10 p L o f 3.0 M sodium acetate (pH 5.2) was added (300 m M final sodium acetate concentration). The linearized dephosphorylated D N A was ethanol-precipitated as described above. After the D N A pellet was air-dried, it was resuspended in 20 p L of sddH20. To end-label using P N K , the linearized and dephosphorylated recombinant plasmid vector containing the promoter or enhancer fragment of interest was incubated with 50 p C i of y 3 2 P - A T P (Amersham), I X P N K buffer (NEB) , and 1.5 Units of P N K ( N E B ) at 37°C for 30 minutes. The P N K enzyme was then heat denatured at 75°C for 20 minutes to destroy P N K activity. To rid the reaction of unincorporated radionucleotides, the labelling reaction was passed through an S200 spin column (Amersham). The labelled D N A was ethanol-precipitated as described above then resuspended in 25 p L of sddH20. If labelling using Klenow D N A polymerase, no dephosphorylation step was required. The linearized recombinant vector (~ 4 pmoles of 5' ends) was incubated with 50 p C i (16 pmoles) each of a 3 2 P - d A T P , a 3 2 P - d T T P , a 3 2 P - d G T P , and a 3 2 P - d C T P (Amersham), I X EcoPol Buffer ( N E B ) , and 2.0 Units of Klenow (NEB) at 25°C for 15 minutes. To ensure that the 5' overhangs were completely filled-in with nucleotides, a cold-chase o f 200 p M of dNTPs was added and incubated at 25°C for 5 minutes. The Klenow enzyme was heat denatured at 75°C for 20 minutes. To rid the reaction of unincorporated radionucleotides, the labelling reaction was passed through an S200 spin column (Amersham). The labelled D N A was ethanol-precipitated as described above then resuspended in 34 p L o f sddH20. To generate a single-end labelled fragment, the -426 to +28 probasin probe was excised from the plasmid vector using a second restriction enzyme - either BamHI or HinDIII (NEB) or in the case of the -705 to -426 base pair probasin fragment, either Nhel or HiriDHL. To generate a single-end labelled fragment, the m E - R A B P probes were excised from the plasmid vector using a second restriction enzyme - either HinDlll or Xbal. After a two-hour digest at 37°C, 10 p L of 4 X loading buffer containing 5% bromophenol blue and xylene cyanol was added to the reaction. The single-end labelled D N A fragments were purified by P A G E and eluted as described above. After resuspension in Buffer D , the dpm for 1.0 p L of resuspension was counted using a Beckman scintillation counter. 37 2.1.2.C Generating Single-end Labelled DNA Probes by Annealing Synthesized Oligos Six 29mer oligos were synthesized at N A P S U B C along with their complementary strand. The sequence o f these oligos corresponds to either the wildtype (WT) G - l element from the probasin promoter region (-209 to -195 base pairs) or mutated derivations of this element. The synthesized oligos, with their mutated nucleotides underlined, are as follows: (1) G - l W T : 5' C T T A A T A G G G A C A T A A A G C C C A C A A A T A A 3' (2) G - l A195C: 5' C T T A A T A G G G A C A T A A A G C C C C C A A A T 3'; (3) G - l C197T: 5' C T T A A T A G G G A C A T A A A G C T C A C A A A T A A 3'; (4) G - l G207A: 5' C T T C C T A G G A A C A T A A A G C T C A C A A A T A A 3'; (5) G - l T215G: 5' C G T A A T A G G G A C A T A A A G C C C A C A A A T A A 3'; (6) G - l Spacer mutant: 5' C T T A A T A G G G A C A C G G A G C C C A C A A A T A A 3'. Single-end labelled double-stranded oligos were generated first by incubating either 32 the single-stranded top strand or complementary bottom strand with 50 u C i o f y P - A T P (Amersham), I X P N K buffer (NEB) , and 1.5 Units of P N K ( N E B ) in a reaction volume of 30 u L for one hour at 37°C. The P N K enzyme was heat denatured at 75°C for 20 minutes. Unincorporated radionucleotides were removed by centrifuging the labelling reaction through a G25 column (Amersham). To generate a single-end labelled double-stranded probe, the cold complementary oligo was annealed to the 3 2 P-labelled oligo. The annealing step involved a slow temperature ramp from 90°C to 4°C at l°C/second. Once the labelled oligo was annealed to its unlabelled complement, the double-stranded probe was isolated and purified using P A G E as described above. The air-dried labelled oligo was resuspended in Buffer D and the dpm/uL counted as described above. 2.1.3 Methylation Protection Assay A s mentioned earlier, the M e P assay employs D M S to methylate the D N A using one-hit kinetics after the protein-DNA complex has come to equilibrium. Two different protocols were employed with contrasting degrees of success. Both methods add D M S to the protein-D N A complex in solution. The difference between the protocols is in the method employed to stop the methylation reaction. The first protocol follows the conventional method and 38 employs P-mercaptoethanol to quench D M S activity after the reaction has proceeded to the desired point. The second protocol employs P A G E to separate the negatively charged protein-bound D N A and unbound (protein-free) D N A from the neutral D M S molecule (Reid K J and Nelson C C 2001). For both protocols, the protein-DNA binding reaction and D M S methylation procedures were the same. Briefly, the recombinant HisTag protein (7.2 pg or 13.6 p M of AR-Ffistag, HMG-His tag , or Sox-4-Histag) was incubated at room temperature for 15 minutes with 2.0 pg Poly dl-dC (Amersham) in D N A Binding Buffer ( D B B : Buffer D and 1 m M D T T freshly added). To each reaction, 350 000 dpm (26.5 fM) of 3 2P-single-end labelled D N A probe was added to the final volume of 30 p L and the binding reaction was brought to equilibrium at room temperature for 10 minutes. Once binding equilibrium was reached, 3.0 p L of a 2% D M S solution was added to the binding reaction to a final D M S concentration of 0.18% (19 m M ) and incubated at room temperature for exactly 2 minutes. D N A treated in the same manner, but without HisTag protein, was used as a control. For the first protocol the D M S methylation activity was quenched with p-mercaptoethanol after 2 minutes by adding 50 p L o f D M S Stop Solution (1.5 m M sodium acetate (pH 7.0) and 1.0 M p-mercaptoethanol). After vortexing vigorously, the reaction was ethanol-precipitated as described above. After the last of the ethanol was carefully removed, the pellet was air-dried then resuspended in sddH20. The volume of water used depended on the ensuing cleavage reagent. For the second protocol loading the reaction onto a P A G E stopped the D M S methylation activity. After the 2 minutes incubation time each methylation reaction was loaded onto a 5% (29:1 acrylamide:bis) 0.5X T B E P A G E while the current was running at 16 V / c m . Since the D M S is an overall neutral molecule, the negatively charged D N A w i l l move away from the D M S and into the gel. Additionally, P A G E permitted the separation of bound and unbound D N A populations. After the AR-bound D N A had moved wel l into the gel, the wet gel was exposed to Kodak B M S film for 1 hour at room temperature. Using the film as a template, the bound (protein) and unbound (no protein) D N A gel fractions were excised and the D N A eluted and ethanol precipitated as described above. Again, the volume of water used to resuspend the D N A depended on the ensuing cleavage reagent. 39 Piperidine is a chemical that w i l l specifically cleave those guanines that are methylated in the N7 position. If piperidine was used then the air-dried pellet was resuspended in 90 p L of sddH20. The cleavage reaction required 1.0 M piperidine, therefore 10 u L of stock piperidine was added to the resuspended methylated D N A probe to achieve this concentration. The methylated guanines were cleaved by incubating the piperidine reaction at 90°C for 30 minutes. Since piperidine is a volatile chemical, lyophilization is a good method to remove piperidine from the sample. The 100 u L reaction volume was snap-frozen in liquid nitrogen and lyophilised under vacuum for ~1.5 hours in a Savant speedvacuum emplying an oi l condensation pump. The dried pellet was washed twice with 100 u L of sddH20, snap-frozen, and lyophilised for ~1.5 hours in between washes. The final dried pellet was resuspended in 10 u L of formamide buffer overnight at 4°C. The following day, 1.0 u L of the piperidine-cleaved samples was counted using a Beckman scintillation counter. The samples were diluted to 2000 dpm/uL, heated to 90°C for five minutes, placed on ice, and then 2.0 u L o f each sample (4000 dpm per lane) was loaded onto apre-run (20 V / c m for 1.0 hour) 6% (29:1 acrylamide:bis), I X T B E , 8.3M Urea P A G E . The denaturing P A G E was run at 20 V / c m at room temperature until the bromophenol blue and xylene cyanol reached the appropriate distance for the D N A fragment size of interest. For those fragments over 100 base pairs, a salt gradient gel was used to slow down the faster and lighter fragments, preventing them from running off the gel while allowing the heavier fragments to move farther into the gel, thereby improving band resolution. In contrast to the regular I X T B E denaturing acrylamide gel, the salt gradient gel utilized 0.5X T B E and the bottom chamber buffer, contained 100 m M sodium phosphate (pH 8.0). To establish the salt gradient and ensure that the salt-front would be ahead of the D N A fragments, the acrylamide gel was pre-run at 20 V / c m for exactly 40 minutes before loading the samples. For both types o f P A G E , the denaturing gel was dried under vacuum at 80°C then exposed at -80°C to Kodak B i o M a x Sensitive ( B M S ) or B i o M a x Resolution ( B M R ) film in a High Energy Transcreen enhancer cassette or a regular cassette respectively. N a O H can also be employed to cleave modified nucleotides. This cleavage method is more flexible than piperidine since N a O H w i l l also cut methylated adenines as well as guanines. Since adenines are methylated by D M S less efficiently than guanines, adenines 4 0 w i l l necessarily appear fainter than guanines on an autoradiography To cleave using N a O H , the dried pellet of methylated D N A probe was resuspended in 100 p L of sddH20. After 100 p L of 20 m M of sodium phosphate (pH 7.0) was added, the reaction was vortexed vigorously and incubated at 95°C for 15 minutes. The reactions were placed briefly on ice and 20 p L o f 1.0 M fresh N a O H was added to a final concentration of 90 m M , vortexed vigorously, and incubated at 95°C for 1.0 hour. To neutralize the reaction, 20 p L of 1.0 N HC1 was added and the reaction was vortexed vigorously. To help precipitate the D N A fragments, 10 pg o f t R N A was added to the mixture. To reduce the salt concentration the volume was increased to 500 p L with sddH20. The N a O H cleavage reaction was then ethanol-precipitated as described above. The air-dried D N A fragments were resuspended in 10 p L o f formamide buffer, counted using a scintillation counter, and separated using a 6% denaturing P A G E as described above. Acrylamide gels were dried and exposed to autoradiograph film as described above. To compare relative band intensity, the exposed autoradiograph film was scanned using a Hewlitt Packard 1200 dpi resolution scanner and the bands quantified using ImageQuant software. 2.1.4 Methylation Interference Assay A s mentioned, the M e l assay requires the D N A to be pre-methylated and purified before use in the protein-DNA binding reaction. Briefly, the labelled D N A fragment of interest (200 fM) in D M S Buffer (50 m M sodium cacodylate (ph 8.0), 10 m M MgC12, and 1.0 pg of calf thymus D N A ) was incubated with 45 m M D M S for exactly two minutes. The reaction was stopped with 50 p L of D M S Stop Solution and the D N A was ethanol-precipitated as described above. The methylated D N A probe was then used in a protein-binding reaction with A R D B D - H i s T a g exactly as described in the M e P protocol. After separating the bound (protein-bound), free (rejected in the presence o f protein), and unbound (no protein) D N A populations using P A G E , the bands were excised from the acrylamide gel and the D N A probes eluted as described above. The selected D N A populations were then ethanol-precipitated and cleaved at the methylated guanines and/or adenines either using piperidine or N a O H cleavage methods as described in the M e P protocol. Unbound, free, and 41 bound populations of D N A probes were visualized using denaturing P A G E and autoradiography as described in the M e P protocol. 2.2 Determining the Binding Affinities of the Probasin Promoter AREs Note to Reader: Stephen Hendy, Research Associate with the Prostate Centre at Vancouver General Hospital, kindly performed the following EMSA and binding assays to determine the relative and actual binding constants for the probasin promoter AREs. To determine the binding affinity of individual A R E s located within the probasin -426 to +28 base pair region, the dissociation constant (Kd) for each binding site was determined using Scatchard Analyisis. A constant amount of AR-His tag protein was incubated with increasing concentration of radiolabelled D N A . The probe concentration from the bound and free fraction, as determined by the radioactive signal, at each titration point was plotted according to the Scatchard plot; [PX]/[X] versus [PX] where [PX] is the protein-bound D N A concentration and [X] is the free D N A concentration. The slope, -K, provides the equilibrium constant. 2.2.1 Generating Radioactively Labelled Probasin Promoter Fragments With High Specific Activity For quantitative analysis of D N A binding affinity it is necessary to radioactively label the D N A probe with high efficiency and specific activity. Using a labelled probe with high specific activity ensures that variation between experiments is minimized and duplication of results is possible because the effects of cold or unlabelled probe on binding specificity w i l l be minimized. For this reason, the D N A probe used in these experiments was labelled by incorporating radioactive a 3 2 P - d C T P into the amplified strand through the P C R extension 42 step. This contrasts the labelling method employed for the M e P and M e l assays where pre-labelling one o f the primers prior to amplification was necessary to generate single-end labelled D N A probes. Using an isotope-incorporation method, the amplification conditions can be manipulated such that each generated D N A probe w i l l have a high average radioactive specific activity. Using the following P C R amplification conditions, D N A probes were generated with approximately 2 incorporated radioactive cytosines for every 20 cytosines present within the D N A with a specific activity of 20 p C i / m M o l . To generate radioactively labelled D N A probes for binding studies, the P C R amplification reaction contained 100 ng of the template plasmid -286 to +28 probasin promoter in pBluescript with 200 p M of dNTPs, 20 p C i of a 3 2 P - d C T P (specific activity of 3000 Ci /mMol ) , I X P C R Buffer, 1.5 m M of MgC12, 1.0 p M of forward and reverse primers (refer to section 2.1.1 for primer sequence and fragments generated), and 2.5 Units of Taq D N A polymerase. Using the M J Research D N A Engine Thermocycler, the amplification program was 95°C for 1.0 minute then 25 cycles of 95°C for 30 seconds, 45°C for 30 seconds, and 72°C for 30 seconds. A final extension step at 72°C for 5 minutes ensured that all the amplification products were fully extended. Labelled amplified D N A probes were purified using P A G E and ethanol-precipitated as described above. The air-dried D N A pellet was resuspended in Buffer D and the dpm/pL was determined using a Beckman scintillation counter. 2.2.2 Determining Binding Affinity Through EMS A E M S A was employed to determine the A R - H i s T a g binding affinity for the individual probasin promoter A R E s and various combinations of these A R E s . Briefly, 10 pmol of A R -HisTag was pre-incubated for 15 minutes at room temperature with 1.0 pg of Poly dl-dC (Amersham) in D B B in a 10 p L volume. Increasing amounts of the radiolabelled probe in a 2.0 p L volume of D B B was added and the binding reactions were further incubated at room temperature for 10 minutes. After the reaction was brought to equilibrium, the titrations were loaded onto a pre-run (16 V / c m for 3 minutes) 5% (29:1 acrylamide:bis) 0.5X T B E P A G E and run at 16 V / c m until the bound and free D N A populations are wel l separated. The gels were transferred to Whatman chromatography paper and dried under vacuum at 80°C 43 then exposed to Kodak B M R film. Using the developed film as a template, the bands corresponding to the bound protein-DNA complex and free D N A were excised from the dried gel. To determine the activity of each band, the dpm of each dried gel slice was counted using a scintillation counter. For the D N A probes containing only one A R E , the binding constants were determined by Scatchard analysis. For the D N A probes containing more than one A R E , the relative binding affinity was compared by plotting bound versus free values averaged from three independent binding experiments for each probe. 2.3 Transfection Studies to Determine Effect of Probasin Promoter Fragments on Transcriptional Activation Note to Reader: Jody L. Saito, UBC Department of Pathology PhD student with the Prostate Centre at Vancouver General Hospital, kindly performed the following transfection experiments to determine the contribution of specific probasin promoter fragments to transcriptional activation. To determine whether the in vitro footprinting and binding affinity assay results had any relevance within an in vivo system, transient transfection assays were carried out using the prostate cancer cell lines L N C a P and PC-3 . Briefly, luciferase reporter plasmids were created by introducing BamHUHinDlll ends to each fragment of the probasin promoter by P C R (please refer to section 2.1.2a) followed by cloning the amplified product into the BamHUHinDlll sites of the p T K - L u c expression plasmid. L N C a P or PC-3 cells were grown to 60% confluence in 24-well plates in R P M I media (Gibco) supplemented with 5% fetal calf serum (FCS). Cells were then transfected with 0.2 ug o f luciferase reporter plasmid, 1.2 ug o f rat A R in the p R L T K expression plasmid (Invitrogen), and 8 ng of renilla expression plasmid p R L T K (Promega) using Lipofectin (Gibco). Cells were incubated for 22 hours in 5% charcoal-stripped medium with 0 to 5 n M of R1881 (Perkin Elmer Gibco). Cells were washed and harvested with passive lysis buffer (Promega) and the luciferase activity of 20 u L cell lysate aliquots were determined using the Dual-Luciferase Reporter assay system (Promega) on a luminometer (Berthold, Germany). Luciferase activity was normalized for transfection efficiency using 44 renilla activity. Experiments were done in triplicate, averaged, and expressed both in relative luciferase units (REU) and as fold induction. 45 3. Results and Discussion 3.1 Methylation Interference Assay Failed to Footprint a Cooperative System Initial attempts at footprinting the rat probasin promoter (-426 through +28 base pairs) using M e l were problematic because of the cooperativity that occurred between the A R E s within the promoter fragment. D M S methylation was carried out at one-hit kinetics, which meant that the methylation rate was empirically established at an average of one methylated guanine or adenine per molecule of D N A . Success using M e l was predicated on the theory that i f a major groove guanine or minor groove adenine was necessary for protein binding and was methylated, then this D N A molecule would be rejected from the bound D N A population and relegated to the free D N A population. In a cooperative system, however, methylation of a single nucleotide necessary for protein binding would not guarantee the relegation of the D N A molecule to the free population. This was because other protein-protein or protein-DNA contacts involved in neighbouring cooperative interactions could compensate for any individual nucleotide disturbed through methylation. Using M e l to determine those guanines or adenines necessary for A R D B D binding to the cooperative probasin promoter unfortunately resulted in no information at all because of the failure of this assay to separate free and bound D N A populations (data not shown). 3.2 A Modified Methylation Protection Assay Revealed Atypical A R Binding Sites B y switching to the M e P assay it was possible to overcome the cooperativity-based problem intrinsic to the M e l assay since M e P did not require the separation of bound and free D N A populations. Instead, M e P involved the D M S methylation o f the protein-DNA complex after the complex had reached equilibrium, therefore, all guanine and adenine contacts that were involved were detectable and cooperative interactions did not interfere with the information obtained. Initial attempts at footprinting the probasin promoter region, from -426 to +28 base pairs, used the conventional method of employing (5-mercaptoethanol to quench the methylation activity of D M S (Brunelle A and Schleif R F 1987). p-mercaptoethanol is a 46 hydrophobic molecule that is able to quench D M S activity but it is also able to disrupt protein-protein interactions - in particular disulfide bridges. The footprint result after employing P-mercaptoethanol was uninformative as no protected nucleotide contacts were observed when the protein-bound reaction (bound) was compared to the protein-free reaction (unbound) (Figure 9). There were three possibilities for this null result. Firstly, it was possible that the A R D B D - D N A contacts were not specific and therefore failed to be protected from the methylation activity of the small and sensitive D M S molecule. Secondly, it was possible that the experimental conditions were such that only a small proportion of the D N A population was bound by protein; therefore, the overwhelming unbound D N A population masked any protein-DNA contacts. Finally, it was possible that either D M S or P-mercaptoethanol disrupted the protein-DNA complex permitting D M S to freely and uniformly methylate the released D N A . P-mercaptoethanol is not entirely efficient at quenching D M S methylation activity, therefore, i f the protein-DNA complex was disrupted due to the presence of P-mercaptoethanol, there was enough residual D M S remaining to methylate the released D N A fragments (Brunell A and Schleif R T 1987). To determine which of the above possibilities was responsible for the null footprinting result, the A R D B D - D N A complex was examined by E M S A after exposure to D M S and/or P-mercaptoethanol. I f the protein-DNA complex is disrupted by either D M S or P-mercaptoethanol then an E M S A would allow visualization of the disrupted complex. Additionally, an E M S A would reveal the proportion of D N A bound by the A R D B D relative to the free population. The E M S A revealed that the entire D N A probe was bound by A R D B D for the given experimental conditions and that D M S minimally disturbed this particular protein-DNA complex (Figure 10). However, 1.0 M p-mercaptoethanol completely disrupted the A R D B D - D N A complex either when it was added after the protein-DNA complex was brought to equilibrium or after the addition of D M S . A n alternate method, therefore, was required to stop D M S methylation activity without disrupting the A R D B D - D N A complex. Using E M S A to visualize the A R D B D - D N A complex provided an alternative method to quench D M S activity and recover the bound D N A population. Since D M S is overall a neutral molecule, E M S A permitted the electrophoretic separation o f the negative 47 UNBOUND BOUND A R D B D r A R E 1 ARE2 Figure 9. Failed MeP of AR-Histag on the probasin -426 to +28 base pair promoter fragment using the conventional MeP assay, which employs (3-mercaptoethanol to quench DMS activity. The ' U N B O U N D ' lane refers to methylated D N A that has not been bound by the A R DBD protein. The 'BOUND A R D B D ' lane refers to D N A that has been bound by the A R DBD prior to DMS methylation. The positions of ARE1 and ARE2 are indicated. A R DBD-Histag DMS(19mM) P-mercaptoethanol (1.0 M) Probasin D N A A R DBD-DNA Complex Free D N A probe Figure 10. 1.0 M P-mercaptoethanol disrupted the A R DBD-probasin promoter (-426 to +28 base pairs) complex whereas 19 m M DMS had little or no effect on the protein-DNA complex. 49 D N A from D M S relinquishing the need to add a quenching chemical such as (3-mercaptoethanol. Using this modified M e P method revealed A R DBD-guanine contacts within the -426 to +28 base pair probasin promoter fragment (Figure 11). A s anticipated, the previously identified A R E s (ARE1 at -241 to -227 base pairs and A R E 2 at -136 to -122 base pairs) were protected in the pattern expected, with guanines in both half sites making the obligatory major groove contacts with the A R D B D . The sequences of the known probasin promoter A R E s with the resulting underlined protein contacts as revealed through M e P are as follows: A R E 1-5' A T A G C A T C T T G T T C T 3' and A R E 2 - 5 ' A G T A C T C C A A G A A C C 3' . Unexpectedly, two additional A R E s were identified using the M e P assay at positions -209 to -195 base pairs and -107 to -93 base pairs (Figure 11 and 12). These novel A R binding sites within the probasin promoter displayed an unusual methylation pattern in that the guanines located at either end of each respective A R E were hypersensitive to, rather than protected from, D M S methylation. Due to this unusual hypersensitivity these A R E s were termed G-sites reflecting the hypermethylated guanines present within the A R E . The sequences of these novel A R E s identified using M e P are as follows (with the protected guanines underlined and the hypermethylated guanines in bold): G - l - 5' G G G A C A T A A A G C C C A 3' and G-2 - 5' A T G A C A C A A T G T C A A 3'. The hypersensitivity observed within G - l and G-2 belied the uniform methylation activity normally observed for D M S . For a single guanine to be subjected to a higher D M S methylation rate than neighbouring guanines inferred that the guanine was more available for methylation than its neighbour. This has been observed with other protein-DNA interactions employing D M S as a chemical probe (Espinas M L 1995, Ramesh V and Nagaraja V 1996). Local structural distortion of the D N A resulting from A R D B D binding could have exposed guanines to a higher methylation rate. This suggested that the allosteric binding of A R D B D to either the G - l or G-2 A R E resulted in a local D N A structural conformation change. 50 -241 bp B O U N D U N B O U N D A R - D B D ARE1 G-l c ARE2 r G-2 B O U N D A R - D B D U N B O U N D ARE1 G-l ARE2 G-2 Figure 11. Modified MeP of the A R DBD-Histag on the probasin promoter from -269 to -77 base pairs revealed the previously described ARE1 and ARE2 (Rennie PS 1994) and two additional novel AREs; G - l and G-2, which displayed hypersensitivity to DMS. 51 a a g c t t c c a c a a g t g c a t t t a g c c t c t c c a g t a t t g c t g a t g a a t c c a c a -377 g t t c a g g t t c a a t g g c g t t c aaaacttgat caaaaatgac c a g a c t t t a t -327 a t t c t t a c a c c a a c a t c t a t c t g a t t g g a g gaatggataa t a g t c a t c a t -277 AREh o O O' g t t t a a a c a t c t a c c a t t c c agttaagaaa atatgjatagc a t c t t g t t c t G - l oo# o o • • o o aagaatggga caggcattgg g c a t t g t c c a t g c c t a g t a a |agtactccaa G-2 o o • o o g a a c c t a t t t gtatactag|a tgacacaatg t c a a t g t c t g t g t a c a a c t g -77 ' • < •227 t a g t c t t t t t c t t a a t a g g g acataaagcc cajcaaataaa a a t a t g c c t g -177 " " A R E 2 •127 ccaactggga tgcaagacac t g c c c a t g c c a a t c a t c c t g aaaagcagct -27 ataaaaagca ggaagctact c t g c a c c t t g tcagtgaggt c c a g a t a c c t +24 Figure 12. The probasin promoter sequence from -426 to +24 base pairs. M e P revealed two novel A R E s , G - l and G-2, at positions -195 to -209 base pairs and -93 to -107 base pairs respectively, which displayed unconventional protection patterns and hypersensitivity to D M S . Protected guanines are indicated by an open circle, whereas hypersensitive guanines are indicated by a dark circle. 52 3.3 Local DNA Structural Distortion Resulting From A R DBD Allosteric Binding was Intrinsic to the DNA Sequences of the G - l and G-2 AREs Local D N A structural distortion, as measured by the D M S hypersensitive sites in the M e P assay, could manifest from one of two possible mechanisms. The probasin promoter fragment examined by M e P contained four A R E s that cooperatively interact where the protein-DNA interface likely comprises of protein-protein and protein-DNA interactions. The first mechanism that could underlay the observed guanine hypersensitivity was that the resultant local D N A structural distortion had arisen from the collective interaction of neighbouring DNA-bound A R D B D homodimers. This buttressing effect of the A R D B D homodimers could have modified the A R D B D contacts at either G - l or G-2. Secondly, the specific nucleotide sequence of both G - l and G-2 could have affected how the A R D B D bound D N A , adopting a different conformation than that observed with conventional A R E s . This last proposed mechanism implies that the primary D N A nucleotide sequence dictates the conformation of the bound A R homodimer and therefore its function. To determine which proposed mechanism was responsible for the observed methylation hypersensitive sites found within G - l and G-2, the individual probasin promoter A R E s were removed from the larger promoter context and the influence o f neighbouring A R E s and subjected to the M e P assay. If the hypersensitivity observed was.due to the buttressing effect of neighbouring DNA-bound A R D B D homodimers, then we would expect the G - l and G-2 sites to lose their hypersensitivity and revert to a more conventional protection pattern. If the hypersensitivity to D M S methylation was intrinsic to the nucleotide sequence of G - l and G-2, then we would expect that the hypersensitive guanines would be maintained despite having removed G - l and G-2 out of the larger context o f the probasin promoter. The M e P assay was applied to the fragments of the probasin promoter containing combinations of A R E s in their natural promoter context. These combinations were G l -A R E 2 - G 2 , A R E 1 - G l and A R E 2 - G2 (Figure 13, 14 and 15). The results from these smaller combinations of probasin promoter A R E s did not alter from the original observation in that A R E l and A R E 2 invariably revealed.conventional A R D B D contacts whereas G - l 53 ARE2 -136br> r G-2 ARE2 ] G-2 -93 bp Figure 13. MeP of the probasin promoter sequence from -269 through -77 base pairs. MeP of this fragment revealed that the hypersensitivity to DMS methylation within G - l and G-2 was retained in the presence of ARE2. 54 BOUND U N B O U N D A R DBD BOUND A R DBD UNBOUND Figure 14. MeP of the probasin promoter sequence from -269 through -164 base pairs. MeP of this fragment revealed that the hypersensitivity to DMS methylation within G - l was retained in the presence of A R E l . 55 BOUND BOUND U N B O U N D A R DBD A R DBD UNBOUND Figure 15. MeP of the probasin promoter sequence from -150 through -77 base pairs. MeP of this fragment revealed that the hypersensitivity to DMS methylation within G-2 was retained in the presence of ARE2. 56 and G-2 maintained their unconventional hypersensitivity to D M S methylation. Applying the M e P assay to the individual G - l and G-2 sites revealed that hypersensitivity to D M S was also maintained (Figure 16A and 16B). These results support the proposal that the nucleotide sequence of both G - l and G-2 dictate how the A R D B D binds D N A and, consequently, imply that nucleotide sequence of the A R E dictates function to the bound A R homodimer. This was the first time that two different classes of A R E s have been described based on the allosteric effects of A R binding. A R E s that display conventional guanine contacts were termed Class I A R E s whereas those A R E s that displayed hypersensitivity to D M S methylation in place of protected guanines were termed Class II A R E s . 3.4 Hypersensitive Guanines in Class II A R E s are Necessary for A R D B D Bind ing A M e P assay w i l l generally show those guanines that make weak interactions with the bound protein. What was not clear from the M e P results was whether the observed G - l and G-2 hypersensitive guanines had any functional role in the A R D B D - D N A interaction. The hypersensitive guanines could either be integral to the A R D B D binding to D N A or incidental in that their presence was not required for the stability or conformation of the bound homodimer. To determine the necessity of the hypersensitive guanines, a M e l assay was performed on the individual G - l and G-2 A R binding sites. A successful M e l assay w i l l enhance in the free D N A population, those nucleotides required for protein-DNA interactions whereas these rejected nucleotides w i l l be absent from the bound D N A population (Figure 8). M e l involved pre-methylating the D N A before performing the protein-DNA binding reaction, followed by a separation of the A R D B D bound from the free D N A populations using E M S A . A s described earlier, the M e l assay was unsuccessful in probing nucleotide contacts of A R E s that interact cooperatively because of the buttressing effect of neighbouring A R D B D homodimers. Individual A R E s , however, would have no such buttressing effect; therefore, the M e l assay remained a useful tool. The M e l results revealed that the G - l guanines that were found to be hypersensitive to D M S methylation in the M e P assay were also relegated to the free D N A population in the M e l assay just as the conventionally protected guanine contacts in A R E 2 were found to be necessary and therefore enhanced in the M e l assay (Figure 17). These results revealed that 57 B BOUND UNBOUND A R DBD 3' -107 bp 5' ' A I G-2 T A C T o G T G T T A C A • G T T 5' A C A c A A T G O T C A A BOUND A R DBD UNBOUND G-2 -93 bp 3' Figure 16. MeP of the individual G-site probasin promoter AREs. A . The promoter fragment from -229 through -164 base pairs reveals that the hypersensitive guanines were intrinsic to the G - l nucleotide sequence. B . The promoter fragment from -121 through -77 base pairs reveals that the hypersensitive guanines were intrinsic to the G-2 nucleotide sequence. 58 Figure 17. Me l of the G - l and A R E 2 from the probasin promoter. To differentiate from the MeP assay, those guanines found to be necessary for A R DBD-Histag binding are indicated by an open triangle. The G - l guanines found to be hypersensitive in the MeP assay, indicated in bold, are clearly important for AR-Histag binding since they are absent from the bound fraction and slightly enhanced in the free fraction. 59 the hypersensitive guanines were indeed necessary for the A R D B D interaction with Class II A R E s . Therefore, in addition to the unusual sequence-dependent D M S hypersensitivity revealed in the M e P assay, these guanines in Class II A R E s appear to also play a functional role in binding site recognition, subsequent conformation change, and/or stabilization of the A R - D N A interaction. It is possible that the guanines are initially involved in hydrogen bonds with the A R in the recognition of Class II sites, as they are in Class I sites. After this initial recognition, there is an allosteric conformational change to the Class II binding sites in which the interaction with the guanines is altered and subsequently become hypersensitive to D M S methylation. 3.5 The Human PSA Enhancer Contained Class I and Class II AREs and the Two Classes Possess Distinctive Nucleotide Features The results pertaining to Class II A R E s , which displayed local structural distortion upon A R D B D binding, were thus far limited to the rat probasin promoter region. It is possible that the hypersensitive sites were merely artefacts of the probasin promoter and not translatable to other androgen-regulated gene regions. To investigate the universality of Class II A R E s and their trademark hypersensitive guanines, the human P S A enhancer region from —4267 to -4062 base pairs was examined using the M e P assay modified for cooperative systems. This region was shown to harbour at least four A R E s , referred to in an earlier study as V , IV , III, and Il ia . A l l four A R E s interact in a cooperative manner and each site contributes to full androgen-specific induction (Figure 6) (Huang W 1999). M e P results showed that the P S A enhancer region possess both Class I and Class II A R E s respectively as indicated by conventional A R D B D - A R E contacts and hypersensitivity to D M S methylation (Figure 18A, 18B and 18C). The previously characterized A R binding sites, V and Il ia , within the P S A enhancer possessed the same distinctive hypersensitive guanines as reported above for the probasin promoter. The high affinity P S A enhancer site III made similar conventional contacts as observed within Class I elements. Two classes of A R E s , therefore, were present in at least two known androgen-regulated promoters that display cooperative binding. The P S A enhancer site IV appeared similar to a Class II element 60 -4234 bp g. B O U N D U N B O U N D A R D B D V IV III 5' -4218 bp 3 y -4189 bp 4175 bp -4148 bp \\ c G O c G O T A T A O G C T i r A A T T A A T A t k T C G O A T T A A T X G C C G ' T A A T A T . c G 7 A T • G C G C 5' B O U N D A R D B D U N B O U N D V IV III -4125 bp Figure 18A. MeP of the human PSA enhancer from -4267 to - 4 0 6 2 base pairs. The previously identified AREs III and IV possess conventional protection pattern whereas A R E V clearly possesses unconventional hypersensitive guanines. 61 B BOUND U N B O U N D A R DBD IIIA -4081 bp 5' C -4062 bp BOUND A R DBD U N B O U N D IIIA V o o o o -4267 aaacctgaga ttaggaatcc tcaatcttat ac gggacaa cttgcaat cc IV o o -4217 tgctcagcct ttgtctctga tgaagatajtt atcttcatgalcttggattg o n o HI _ o o o o -4167 aaaacagacc tactctggajg gaacatattg tatclgattgt ccttgacagt • 4 IIIA o • o o • -4117 aaacaaatct gttgtaagag acattatctt tattatc^ag"gacagtaagcaag]cct Figure 18B and 18C. B . MeP of the PSA enhancer A R E IIIA possesses hypersensitivity that resembles the probasin G - l and G-2 AREs. C . The PSA enhancer sequence from -4267 to -4062 base pairs with protected guanines indicated by open circles and hypersensitive guanines indicated by dark circles. but did not have enough guanines to confidently classify. Finally, M e P analysis of the P S A A R E s V and I l ia further revealed that additional guanines flanking Class II sites were hypersensitive suggesting that structural distortion resulting from A R D B D binding could extend over an area of at least 17 base pairs (Figure 18A and 18B). A comparative table was drawn up to determine a consensus sequence for each of the two classes of A R E s (Table 2). Class II A R E s from the probasin promoter and the P S A enhancer were aligned in order to determine whether these novel A R E s possessed specific representative or consensus nucleotides when compared to aligned conventional Class I A R E s . Along with verified Class I and II A R E s from the probasin promoter and P S A enhancer, proposed Class I and II A R E s based on sequence similarity from known A R E s from other androgen-regulatory regions such as the mouse Sip promoter, mouse E - R A B P promoter and proximal P S A promoter were also used to generate a consensus sequence. The nucleotide numbering scheme employed to describe the A R E sequence assigns the number 0 to the central nucleotide in the spacer region of the inverse palindrome. From this starting point the nucleotide position number ascends to +7 towards the 3'-half site and descends to -7 towards the 5'-half site (Table 2). A comparison of the two consensus sequences reveals common and distinctive features that distinguish Class I and Class II A R E s . Both Classes have a cytosine at position -3 and a guanine at +3 where the guanines at these positions in each half site make a conserved hydrogen bond with the arginine residue found within the A R recognition a-helix (Figure 3B and Table 2). Both Class I and II A R E s possess an adenine at position - 4 in the 3' half site where a conserved V a n der Waals contact is made between the thymidine in this position and a valine found within the A R recognition P Box (Figure 3 and Table 2). Unlike the Class I consensus sequence, there is no conserved adenine-thymidine base pair at position +4 in the 5' half site of the Class II consensus sequence. The absence o f this V a n der Waal's contact in the 5' half site suggests that, in general, the A R homodimer binds asymmetrically to the Class II A R E . A pronounced feature of a Class II site is a unique guanine at position -5 and cytosine at position +4 and/or +5. The most prominent feature o f Class II sites is a purine at position +7 whereas Class I sites posses a highly conserved pyrimidine at this 63 Probasin G-l -209GGGACA-TAA -AGCCCA 196 Probasin G-2 -107ATGACA-CAA -TGTCAA 93 PSA Enhancer V 4 2 3 4 GGGACA-ACT -TGCAAA- 4220 PSA Enhancer IIIA -4079 AGGACA-GTA -AGCAAG" 4065 PSA Enhancer IV 4 1 7 5 AGATCA-TGA -AGATAA" 4189 SLP 2 +142AGAACT- GGC -TGACCA*12* • • • • CONSENSUS CLASS II RGGACA-NNA -AGCCAA -7 -3 0 + 3 +7 CONSENSUS CLASS I RGAACA-NGN -TGTNCT o o Probasin ARE1 "241ATAGCA-TCT -TGTTCT 227 Probasin ARE2 136AGTACT-CCA -AGAACC 122 PSA Enhancer III - 4 1 4 8 GGAACA-TAT -TGTATT" 4134 PSA ARR -390GGATCA- GGG -AGTCTC 376 PSA ARE ~167AGAACA - GCA -AGTGCT 153 SLP 3 +144AGAACA- GGC -TGTTTC*15* Table 2. Alignment of A R E s into proposed Class II and Class I consensus sequences. A R E s from the probasin promoter, P S A enhancer, P S A promoter, Sip enhancer, and m E - R A B P promoter were employed to generate consensus. The consensus nucleotides that are common to both Class I and Class II sites are underlined. The consensus nucleotides that are distinctive to each class type are indicated by an open circle for Class I nucleotides and a black-filled circle for Class II nucleotides. Italicized elements are proposed by sequence similarity to the Class designation. 64 position. Outside of the Class II consensus sequence, a thymidine was found conserved at position -13 in all cases and an adenine was found at position +13 in all but one case. Finally, there is a tendency to have a higher AT- r i ch flanking region and spacer in Class II APvEs in comparison to Class I A R E s . 3.6 Mutation of Key Nucleotides Transforms the MeP Pattern of Class II Half Site to a Class I Half Site The nucleotide sequence of Class II A R E s was shown to dictate the bound A R D B D conformation and the results from the M e l assay had shown that the hypersensitive guanines appear to be critical for D N A binding site recognition, subsequent allosteric conformational changes, and/or stability of the protein-DNA complex. To further delineate the role of specific nucleotides within the sequence of Class II A R E s , key residues within the Class II element G - l were altered such that the modified sequence partially resembled a Class I consensus sequence. The M e P assay was then used to determine whether these mutations affected the D M S methylation pattern and consequently the bound-AR D B D conformation. The distinctive features of Class II A R E s include a relatively A T - r i c h spacer region, a thymidine at position -13 relative to the central spacer nucleotide, and a purine at position +7 in the 3' half site in contrast to the consensus pyrimidine found at this position in Class I sites. These distinct features were examined for their contribution to the D M S hypersensitivity pattern through site-directed mutagenesis. The results from the D M S M e P assays demonstrate that conversion of the thymidine to a guanine at position -13 , six base pairs upstream of the 5' half site, decreased the guanine hypersensitivity in the 5' half site compared to wildtype G - l but did not affect the binding pattern of the 3' half site (Figure 19A and 19B). When the adenine at position +7 of the 3' half site was converted to a cytosine, the D M S protection pattern of the 3' half site reverted to the pattern observed with a conventional Class I binding site but did not alter the hypersensitivity in the 5' half site (Figure 19C). Finally, converting the palindrome spacer nucleotides from T A A to C G G also decreased the hypersensitivity in the 5' half site alone (Figure 19D). These results suggest that Class II type binding by the A R is primarily distinguished by the non-consensus purine at position +7 within the 3' half site but is also 65 co + U u rj O ro *« + < < O EH i < O O LT) C D o S3 C D CO . 1 o C D loi> g (o/„) ^ISUajUI <D '+-» o u z <D > s < < u u u 32: O T 3 ' - * - » o JJ "o Z U -4-* o <D +-» o P L , < < < H H U u i n as < CO <D *-*-» o ID Z > c < < U < u u o o o u < o o 1-PH < H H O o CD O o CM O CM O (o/Q) A>ISU3JUT <D o u 3 < 0 > c <; < U o u < 3 ID r° O 1 ) i — i o 13 B o ID +-» o l-c PH H < < H H U o o o o o o o o o o o o o o o o o o o o o o o ° § § § C O C M ° ° C M C O § C M O O O C D T t C M CM CO 0 0 CO a 4 0 o . S - i ft C D rg C+H O C D C O ft $3 o • 1—1 -f-> o C D -*-> o (-1 ft u > CO $3 C D CO (-1 C D a $3 _o cd 1 C D GO C D > u C D 43 C D bO u 4 3 o <u 43 H ON S-i C D bO S3 o • 1"I 4H 43 0 0Q o > CO $3 CD CO V-i C D ft GO C D c bO CO O 43 S3 C D CO C D S-J OH C D V-i CO S3 43 Q co C D c C D bo cs 5 s T3 C D O C D 1-T3 C D 13 C D S3 O • • i—i rn . H ^ CO O o ft bfl •»-| aj O C D S3 C D 4 3 T3 C D Q S3 O o +-> C D > 13 »-< S3 O I C D B Q T3 C D - » - > o C D . co CD S3 bO CD CO o 43 co ^ C D (o/0) ^jisuajui (%) ^!SU3JUI 3AIJBPH 2. « 13 S3 C D co C D ft C D (-1 co 43  C D cd C D ; - i C D 43 O bo a o u 2^  o o u C D S3 1 . bO 03 O +-» T3 C D n C D S^  o o co cd CO S3 O co O ft +J CS CD S3 l-H T3 C D C D I-C D 43 O bO • i-H O T3 C D •c C D > S3 o o co cd S3 O bo C D • C D C D ft co < H C D 43 +-» C D l-i C D 1 o bO 8 o CD O ccJ ft 0 0 S3 • 1—» CO O -4-» >^  o cd T3 C D C D -4-" CO cd 2 - ^+ 66 influenced by the identity of other nucleotides flanking the A R E and within the spacer region. 3.7 Specific Arrangements of Class I and Class II A R E s Increase Cooperative B ind ing Within the steroid receptor field, it is accepted that there is a conventional steroid receptor D N A response element whose nucleotide sequence w i l l vary slightly in nature depending on receptor subtype, but nevertheless w i l l maintain conventional contacts between the steroid receptor and the D N A response element. Contrary to this supposition, we have discovered two classes of A R E s based on the primary nucleotide sequence within the probasin promoter and P S A enhancer. The newly discovered Class II A R E s display unconventional local D N A structural distortion along with the absence of previously assumed obligatory protein-DNA contacts. We have shown that the nucleotide sequence of an A R E dictates how the A R D B D binds in that, upon A R D B D binding, specific guanines within Class II A R E s were hypersensitive to D M S methylation and, furthermore, that these guanines were necessary for the A R D B D - D N A interaction. The next question was whether the nucleotide sequence, which dictated the conformation of the bound A R D B D homodimer, also dictated function. Mammalian AR-regulated promoters such as the probasin and the P S A gene contain multiple low-affinity A R E s that cooperatively interact to dramatically increase overall affinity for the A R . The molecular mechanisms that underlie cooperative binding are unknown. However, the existence of two classes of A R E s that result in different conformations of bound A R homodimers imply that the interplay between these two classes o f A R E s could provide a molecular basis for cooperative binding. It is accepted that cooperative binding likely involves both protein-protein and protein-DNA interactions (Senear D F 1998). To determine the contribution o f protein-DNA interactions to cooperativity, the relative binding affinities of multiple binding sites were compared to the specific binding affinity of individual binding sites. A distinctive feature of cooperative systems is that a much lower concentration 67 of protein is required to fully occupy the multiple binding sites that make up a cooperative system compared to a single binding element (Mao C 1994, Kerppola T 1998, L i u Z 1998, Senear D F 1998). Using this feature, it was possible to determine the contribution of each A R E class type to cooperativity by comparing binding affinities. Subsequently this comparison can determine the contribution o f nucleotide variations in each A R E class type and the concomitant differences in bound-AR D B D conformation to the cooperativity function. It has been established that the concentration of protein required to fully occupy the multiple A R E s found within the -426 to +28 base pair probasin promoter fragment, was at least 10 fold lower than that required to individually occupy either A R E 1 or A R E 2 (Rennie PS 1994). From the above data, there exists two additional A R E s within the probasin promoter and it is apparent that two different classes of A R E s likely contribute to the cooperativity of this system. To determine the contribution of these two classes of A R E to cooperativity, the specific binding affinity o f each of the four individual A R E s was determined using Scatchard Analysis. These analyses were performed using a constant amount of purified D N A in an E M S A to avoid inconsistencies caused by protein dilution over large ranges of concentration. This affinity measurement was then compared to the relative binding affinity of different combinations of A R E s that contained both Class I and Class II A R E s in their natural promoter context. The Scatchard Analysis revealed that the four A R E s found within the probasin promoter individually had a weak specific binding affinity for the A R D B D (Figure 20). The relative binding affinities of the A R D B D were 6.3 n M for A R E 2 , 11 n M for A R E 1 , 14 n M for G - l , and 18 n M for G-2. Although the isolated Class I sites of the probasin promoter have slightly higher affinity than the Class II sites, all elements displayed measurable binding activity. The low affinity for the A R D B D for individual A R E s was expected since many of the multiple A R E s found within androgen-regulated promoters individually displayed weak affinity for the A R . Within the probasin promoter context, the natural arrangement of the A R E s alternated the two classes of A R E s in the following order: 5' A R E 1 - G l - A R E 2 -G2 3' (Figure 12). To determine the functional contribution of Class II A R E s to 68 cooperative A R D B D binding to the probasin promoter, a series of D N A fragments were created containing combinations of A R E 1 , G - l , A R E 2 , and G-2 in their native context. These combinations of Class I and II elements were analyzed to determine the relative contribution of individual binding sites to cooperativity. This was performed by measuring A R D B D - D N A complex formation as a function of increasing D N A concentration. In all combinations tested a substantial level of cooperativity was observed in that the concentration of A R D B D required to fully occupy multiples of A R E s was lower than that needed to occupy any of the individual A R E s (Figure 20). These analyses revealed that the combination of A R E 1 and G - l (with R v a l u e s of 11 n M and 14 n M in isolation respectively) resulted in a cooperative interaction that half-saturated the D N A at more than 10-fold lower concentration (0.9 nM) . Similarly A R E 2 and G-2 (with Kd values of 6.3 n M and 18 n M in isolation respectively) interacted in a cooperative manner to shift the rate of complex formation an order of magnitude lower in concentration. The promoter fragment containing G - l , A R E 2 , and G-2 interacted with the highest degree of D N A binding cooperativity resulting in half-saturated binding at 0.11 n M (Figure 20). This last probasin promoter fragment containing G 1 - A R E 2 - G 2 displayed over 50 times stronger binding compared to A R E 2 , the individual element with the highest D N A binding affinity. The combination of A R E 1 - G 1 - A R E 2 - G 2 resulted in half-saturated binding at 1.2 n M , which intriguingly demonstrates that the addition of A R E 1 to the extremely cooperative G 1 - A R E 2 - G 2 promoter fragment weakened the overall strength of the D N A binding complex (Figure 20). Further examination of the A R E 1 - G 1 - A R E 2 - G 2 promoter fragment revealed that the binding curve was biphasic. The order of complex formation appears to be occupation of G 1 - A R E 2 - G 2 at low concentrations of protein, followed by additional binding to A R E 1 as protein concentration increases. Overall these results indicate that Class II elements are instrumental in providing the dramatic level of D N A binding cooperativity observed on the probasin promoter. Furthermore it is apparent that the four individual elements interact in a complex manner where G - l can interact individually with A R E 1 or A R E 2 - G 2 . 69 1 0.75 A 0.5A 0.25 A A A • 0 A 9 UfUX&Z n Pig"?' 0.01 0.1 10 O • AREl (-241 to -227) • ARE2 (-136 to -122) • Gl (-209 to -196) • G2 (-107 to -93) A AREl Gl (-268 to -163) A ARE2 G2 (-141 to -76) O Gl ARE2 G2 (-229 to -76). • AREl Gl ARE2 G2 (-268 to -76) Free (nM) Figure 20. Indivudual A R E s displayed weak binding affinity for the A R . Cooperative binding was displayed between the A R E s within the A R E 1 - G 1 and A R E 2 - G 2 probasin promoter fragments. The highest cooperative probasin promoter fragment contained G 1 - A R E 2 - G 2 . The addition of A R E l to this highly cooperative region decreased overall binding affinity and cooperativity for the fragment. The X axis denotes amount of free D N A probe in n M whereas the Y axis denotes the log of A R D B D bound probe i n n M . 7 0 However, i f all four elements are present on the same D N A fragment then the interaction between G - l and A R E 2 - G 2 is weakened, presumably through G - l ' s interaction with A R E l , thereby affecting the stability of the overall complex. 3.8 Class II A R E s Contribute to Synergistic Transactivation as Observed in Transient Transfection Assays Thus far the probasin proximal promoter-AR D B D binding characteristics have been described using in vitro methods of analysis. To determine whether the in vitro data had biological significance, it was necessary to examine the above results in the context of an in vivo system. Cooperative binding of transcription factors to a promoter region did not necessarily imply that AR-specific regulation or synergistic transcription would result. This is because cooperative binding and synergistic transcription are two different albeit linked processes. The presumption is that a specifically regulated promoter segment would not only be occupied by the correct transcription factors as promoted by cooperative binding, but that the resultant conformation of the transcription factor-DNA multicomplex would recruit essential coactivators that in turn assist in recruiting the basal transcriptional machinery. The resultant transcription factor-coactivator-DNA complex, termed the 'enhanceosome' and its formation likely depends on the primary nucleotide sequence of the gene regulatory region, which not only dictates the specific type of transcription factor that binds but also the resultant conformation of the recruited protein-DNA multicomplex. Transient transfection assays were employed to determine whether the cooperative binding observed for the various combinations of A R E s used in the in vitro binding assays, translated into AR-mediated synergistic transcriptional activation. Using the human prostate cancer cell lines L N C a P and P C 3 , promoter constructs of individual A R E s or the four different A R E combinations described above were attached to a luciferase reporter gene and transiently transfected along with increasing concentrations of R l 881, a synthetic androgen. A s expected the individual A R E s displayed minimal luciferase expression (less than 2000 R L U s ) that 71 a) 2500 0.0 [R1881] nM AREl (-241 to -227) ARE2 (-136 to -122) G l (-209 to-196) G2 (-107 to -93) b) c) 5000 4000 3000 2000 -\ 1000 -4 0.0 0.01 0.1 [R1881] nM AREl Gl (-268 to -163) ARE2 G2 (-141 to -76) AREl Gl ARE2 G2 (-268 to -76) G l ARE2 G2 (-229 to -76) 0.0 0.01 0.1 [R1881] nM Figure 21. The X axis denotes R1881 concentration in n M . The Y axis denotes relative luciferase units corrected for Renilla expression. A . The individual A R E s all displayed weak transactivation ability. B . The combination of either A R E 1 - G 1 or A R E 2 - G 2 resulted in additive transactivation relative to the individual A R E s . C . The most synergistic combination was the A R E 1 - G 1 - A R E 2 - G 2 promoter fragment. 72 were maximal at 0.5 n M R1881, which corresponds to 12-, 23-, 25-, and 33- fold induction for G - l , A R E l , G-2, and A R E 2 respectively (Figure 21 A ) . Combining Class I and Class II sites resulted in an increase in sensitivity to hormone concentration which was maximal at ~0.5 n M R1881 (Figure 21B). However, the magnitude of transcriptional activation of A R E 1 - G 1 and A R E 2 - G 2 was at most additive when compared to the activity observed for the individual elements. The addition of G - l to the A R E 2 - G 2 binding element displayed a dramatic increase in transcriptional activation to a maximum of 169-fold induction (-20 000 R L U s ) , which can be described as synergistic (Figure 21C). The synergistic transcriptional activation observed with the G 1 - A R E 2 - G 2 promoter fragment is consistent with the increase in D N A binding cooperativity observed when G - l was combined with A R E 2 - G 2 in the in vitro binding assays (20B). The addition of A R E l to the highly cooperative G 1 - A R E 2 - G 2 D N A binding region resulted in a biphasic curve from 0.01 to 0.05 n M of R1881 followed by a sharp increase in the level of activation with increasing hormone concentration. The A R E 1 - G 1 - A R E 2 - G 2 promoter fragment culminated in a 5-fold enhanced transcriptional response (-100 000 R L U s ) compared with the G 1 - A R E 2 - G 2 fragment and more than 30-fold greater than any individual A R E (Figure 21C). The observed high relative transcriptional activity of the A R E l - G 1 - A R E 2 - G 2 promoter fragment contrasts the cooperative binding data, which demonstrated that the addition of the A R E l element decreased the overall binding strength of the complex (Figure 20). Together the in vitro binding data and the in vivo transient transfection data suggest that a biphasic curve of activity arises from the highly cooperative D N A binding complex at low concentrations of R1881, presumably corresponding to concentration of activated A R in the nucleus. At high levels of R1881 (or activated A R ) , however, the inclusion of the A R E l binding site to the probasin promoter fragment is able to provide highly synergistic levels of transcriptional activation. Overall, these data imply that unique combinations of Class I and Class II elements are required for maximal transcriptional activation by the A R and that each A R E plays an unique functional role in transcriptional activation, D N A affinity, and complex stability. Intriguingly, these observations suggest that the A R E s within the 73 probasin promoter function as a rheostat sensitive to A R concentration. A slight increase in androgen signal ensures that the highly cooperative arrangement of Class I and Class II A R E s are occupied and transcription is activated. Another slight increase in androgen signal ensures full occupation of neighbouring A R E s that result in maximal transcriptional activation. In summary, the above analyses demonstrate that the A R binds to two structurally and functionally distinct classes of A R E s that are directed by allosteric interactions of the binding complex. A R binding to conventional Class I sites have been previously recognized and employ conventional nucleotide contacts that are used by other nuclear receptors. In contrast, Class II binding sites result in D N A -structural alterations that display D M S hypersensitivity to guanines that are normally contacted in Class I sites, where they are protected from D M S methylation activity. In isolation both classes of A R binding sites are of low D N A binding affinity and transcriptional activity. Unique combinations of Class I and Class II sites result in dramatic cooperative D N A binding and a highly synergistic effect upon transcriptional activity. This complex and composite function that facilitates cooperative D N A binding and achieves a synergisyic level of transcriptional response is dictated by the primary nucleotide sequence to which A R binds. 3.9 Cooperative Protein-Protein Interactions Between A R Homodimers Could Involve Stabilizing Disulfide Bridges Initial attempts to employ P-mercaptoethanol to quench D M S methylation activity resulted in the total disruption of the cooperative A R D B D - D N A complex that formed on the probasin promoter segment (Figure 10). P-mercaptoethanol is a reducing agent that has been employed to reduce or disrupt disulfide bonds that can form between free cysteine residues located within or between proteins. The A R D B D possesses two presumably free cysteine residues: one in the first zinc motif and one in the hinge region (Figure 2 A ) . Neither of these cysteines has been shown to be necessary for A R D B D recognition or binding to an individual A R E , however, it is 74 conceivable that either or both residues could be involved in cooperative interactions with neighbouring DNA-bound A R D B D s . If P-mercaptoethanol disrupts A R protein-protein interactions that are necessary for cooperative D N A binding but does not disrupt the binding o f an A R D B D homodimer to an individual A R E , then this can be examined using an E M S A . P-mercaptoethanol at 1.0 M concentration disrupted the highly cooperative G l -A R E 2 - G2 probasin promoter fragment as observed earlier with the -426 to +28 probasin promoter fragment (Figure 10 and 22A). To a lesser extent, l . O M p -mercaptoethanol also disrupted the binding of the A R D B D complex to the individual Class II G - l element (Figure 22B). Intriguingly, A R D B D binding to the canonical S R E was not disturbed by 1.0 M P-mercaptoethanol (Figure 22C). These data suggest that novel protein-protein interactions arise from A R D B D binding to natural A R E s but do not arise from A R D B D binding to the promiscuous S R E , supporting the argument that nucleotide sequence of natural A R E s affects the conformation of the bound A R D B D . 3.10 Recombinant Sox-4 Protein Increases the Number of A R DBD Monomers that Bind the Probasin Promoter The primary and flanking nucleotide sequence of the probasin promoter Class II A R E s G - l and G-2 also matched the sequence that potentially could be recognized by an H M G - b o x architectural protein from the Sox family (Matlnspector 2.2). Sox proteins possess an H M G box domain that binds the minor groove o f AT- r i ch sequence and intercalates a hydrophobic residue into the D N A structure. This intercalation results in bending the D N A towards the minor groove effectively opening the major groove and thereby facilitating transcription factor access. Architectural proteins have been implicated in the facilitation of many cooperatively interacting protein-DNA systems such as N F - K B protein and the I F N a promoter (see Introduction, Section 1.7). 75 A . A R D B D H i s t a g [P-mercaptoethanol (M)] G 1 - A R E 2 - G 2 D N A A R D B D Histag bound D N A G 1 - A R E 2 - G 2 Figure 22A and 22B. E M S A showing P-mercaptoethanol disruption of the A R D B D - D N A binding complex. A . 1.0 M of P-mercaptoethanol disrupted A R D B D binding to the highly cooperative G 1 - A R E 2 - G 2 probasin promoter fragment. B . 1.0 M P-mercaptoethanol moderately disrupted A R D B D binding to the Class II G-2 element. 76 . A R D B D - H i s t a g - + + + + + + + [P-mercaptoethanol(M)] - 2 1 .75 .5 .25 .1 -Canonical S R E + + + + + + + + A R DBD-His tag Bound D N A S R E Figure 22C. 1.0 M of P-mercaptoethanol had negligible effect on A R D B D binding to the canonical S R E . Database research revealed that Sox-4 should recognize the minor groove sequence that coincides with both G - l and G-2 (Matlnspector 2.2). To determine whether the Sox-4 protein would specifically contact the downstream proximal probasin promoter, a M e P assay was employed using purified recombinant Sox-4 histidine tagged protein. Since piperidine would only cleave methylated guanines in the major groove, its use would not yield information on Sox-4 minor groove contacts. Instead, N a O H was employed since it w i l l cleave both methylated guanines in the major groove and adenines in the minor groove. D M S methylates adenines at the N3 position in the minor groove thus exposing Sox-4 - D N A interactions. Initial footprinting results were negative despite Sox-4 successfully shifting the G 1 - A R E 2 - G 2 probasin promoter fragment in an E M S A (data not shown). When Sox-4 was combined with the A R D B D in an M e P assay however, the number of A R D B D monomer binding sites within the proximal probasin promoter fragment increased (Figure 23). These additional binding sites were assumed to be A R D B D and not Sox-4 since the protected sites were guanines and therefore in the major groove and, furthermore, the nucleotide binding sequence resembled A R E half-sites. 3.11 MeP of Other AR-Regulated Promoter and Enhancer Regions Reveal Novel Protection Patterns that Suggest Monomer and Direct Repeat AR-DNA Binding In addition to the —426 to +28 base pair probasin promoter and the P S A enhancer, three other AR-regulated regions were examined using the modified M e P assay. The m E - R A B P promoter from -553 to +28 base pairs revealed two Class I-type A R E s at -432 and -418 base pairs and -74 and -91 base pairs termed A R binding site-1 (ARBS-1) and A R binding site-0 (ARBS-0) respectively (Figure 24A and B) (Lareyre JJ 2000). 78 UNBOUND BOUND AR-DBD BOUND AR-DBD UNBOUND and SOX4 G-l Urn* (P* an • ARE2 G-2 - M M . W » 3' C c c T ,OG T A T T T c G • G • G T 5' • 209 bp G -G G A C A T A A A 196 bp 136 bp GO' C C c A 3' 5' -122 bp 3 3'T A C T OG T G T T A C A G T T 5' •107 bp 5 A BOUND AR-DBD BOUND and S0X4 UNBOUND AR-DBD UNBOUND G-l -93 bp ARE2 2 W • ;; m---*-^. G-2 Figure 23. Sox-4 protein increases the occupancy of A R DBD on the probasin promoter. Protected guanines are indicated by open circles whereas hypersensitive guanines are indicated by dark circles. In the presence of Sox-4, additional guanines protected by the A R DBD and hypersensitive to DMS are indicated by an asterisk and an 'FT respectively. 79 The rat A R enhancer from position 2028 to 2333 base pairs revealed a Class II-type A R E at 2165 to 2179 base pairs that was previously described as an A R -responsive intravening sequence (IVS) along with a potential Class I-type A R E at 2073 to 2087 base pairs that was previously described as a putative A R E (Grad J M 1999) (Figure 25). Within the A R exonic enhancer there are two more previously described A R E s ( A R enhancer A R E l and A R E 2 ) , which require greater resolution to determine the exact A R D B D contacts, however, the M e P information acquired in this study suggests that A R E 2 at 2250 to 2272 base pairs may comprise of at least two direct repeats of the A R D B D separated by four base pairs (Grad J M 1999) (Figure 25). The proximal p21 promoter from position -288 to -1 base pairs revealed A R D B D protected and hypersensitive guanines at the previously described p21 A R E at -209 to -195 base pairs and a novel Class II-type A R E at -39 to -25 base pairs (Lu S 1999) (Figure 26). A n additional novel yet unresolveable A R E or cluster o f A R E s between -177 and -141 base pairs was also discovered although, like the rat A R enhancer, greater resolution is required to confirm and establish the specific A R D B D contacts within this region of the p21 proximal promoter (Figure 26). Finally, the -705 to —426 base pair upstream proximal probasin promoter revealed an intriguing cluster of A R D B D monomer and homodimer binding sites between -529 and -509 base pairs, which corresponds to the A R binding site-3 (ARBS-3) (Matusik R, personal communication) (Figure 27). A n additional novel cluster of A R D B D binding sites was revealed using the M e P assay between -577 and -554 base pairs (Figure 27). The M e P results for the rat A R enhancer, human p21 promoter, and the rat upstream proximal probasin promoter (-705 to -426 base pairs) are preliminary results at best (Figures 25, 26, and 27). Smaller D N A fragments spanning each A R D B D binding region w i l l yield greater resolution and, therefore, greater information on the specific A R D B D contacts within these A R D B D binding regions o f D N A . Having stated that more detailed work is required, the preliminary results from these other AR-regulated D N A regions suggest that the A R D B D may not be restricted to a 80 homodimer structure bound to an inverse D N A palindrome. Others have suggested that the A R could bind as a homodimer to a direct repeat although this has yet to be shown either in an X-ray crystal structure or in functional studies (Zhou Z 1997, Claessens F 2001). The M e P results for the m E - R A B P promoter, rat A R enhancer, human p21 promoter and the upstream proximal probasin promoter also suggest that the A R D B D could bind as a monomer (Figures 24A and B , 25, 26, and 27). It is not clear from this study whether this observation is the result of DNA-bound A R D B D homodimers recruiting A R D B D monomers to the D N A or i f A R D B D monomer binding occurs regardless of the larger D N A context. The G R D B D is known to bind as a monomer to D N A half site sequences whereas the full length G R fails to bind as a monomer (Chalepakis G 1990, Segard-Maurel 1 1996). Therefore, the A R D B D monomer binding observed in this study may be a similar artefact. Overall, these intriguing results suggest that A R may be able to bind in unconventional ways although this has yet to show any functional significance. 81 Figure 24 A. Methylation protection of the A R DBD on the mE-RABP promoter from -543 to -166 base pairs. The previously described A R binding site-1 (ARBS-1) is indicated (Lareyre JJ 2000). Protected guanines are indicated with an open circle. 82 B O U N D 5 6 , B O U N D U N B O U N D A R D B D 3' 5' A R D B D U N B O U N D A R B S - 0 Figure 24B. Methylation protection of the A R D B D on the m E - R A B P promoter from -195 to +26 base pairs. The previously described A R binding site-0 ( A R B S - 0 ) is indicated with the protected guanines represented by an open circle (Lareryre JJ 2000). I 83 BOUND U N B O U N D A R DBD ARE-2 f IVS Putative A R E 2253 bp 2179 bp 3' 5' A T C GO A T T A O G C T 1 ' A T A C G G C A J i T C GO A T • G C A T OG C 5' 3' 2165 bp 2102 bp 3' 5' G C - G c C G C G G C T A OG c T r A G C T A G c G c T A • G C A T G C OG c A 1 r T C G C G G C A i T OG C 5' 3' BOUND A R DBD U N B O U N D ARE-2 IVS Putative A R E Figure 25. MeP of the A R DBD on the rat A R enhancer from 2028 to 2333 base pairs. Previously established ARE-2, IVS (intravening sequence), and the putative A R E reveal DMS protected and hypersensitive guanines (Grad J M 1999). 84 BOUND U N B O U N D A R DBD p21 p A R E B O U N D A R D B D U N B O U N D h p 2 1 A R E Figure 26. M e P of A R D B D on the p21 promoter from -285 to -1 base pairs. Previously described p21 A R E is indicated ( L u S 1999). 85 mm Figure 27. Methylation protection of the A R D B D on the upstream proximal probasin promoter from -705 to -426 base pairs. A R B S - 3 is indicated (Matusik R personal communication). 86 4. Future Direction and Conclusion 4.1 Summary of Project Results Previous observational research has shown that the A R w i l l cooperatively bind to multiple A R E s within the transcriptional regulatory regions of targeted genes such as the rat probasin promoter and the human P S A enhancer (Tsai et al, 1989; Schule et al, 1988; Strahle et al, 1988; Kasper et al, 1994; Huang et al, 1999). There are several previously studied systems that use cooperative binding o f transcription factors to D N A as a mechanism to increase transcription factor specificity and sensitivity to transcription factor concentration (Thanos and Maniatis, 1992; Zhang and Verdine, 1999; Wood et al, 1998). Close examination of these systems has revealed that both protein-protein and protein-DNA allosteric interactions underlie cooperative binding and subsequent specificity and sensitivity. The observation that A R cooperatively binds to multiple A R E s is an important contribution to understanding androgen action. Current knowledge of steroid receptors cannot explain how the A R specifically regulates its own set of unique genes while simultaneously sharing a common D N A binding site with other steroid receptors. It is logical to propose that cooperative binding contributes to AR-specific transcriptional regulation since cooperative binding has been implicated in the specificity and sensitivity of other transcription factor systems (Thanos and Maniatis, 1992; Zhang and Verdine, 1999; Wood et al, 1998). However, unlike other cooperative systems where interactions occur between different types of transcription factors, the A R interacts cooperatively with other A R molecules. This observation emphasizes the role of protein-DNA rather than protein-protein interactions in cooperative A R binding to D N A since variation in type of DNA-b ind ing protein cannot be a contributing factor. Additionally, it is evident from this study and others that promoter context and nucleotide sequence contributes to specific A R - D N A interactions since specific combinations of A R E s in their native context, which individually vary in primary nucleotide sequence from the canonical S R E , are 87 necessary for cooperative A R binding and AR-specific transcriptional regulation (Tan et al, 1992; Lieberman et al, 1993; Nelson et al, .1999). In order to determine the contribution of nucleotide sequence to the underlying mechanisms of cooperative binding, this study examined in detail the A R E s found within two AR-regulated gene regions; the rat probasin promoter and the human P S A enhancer. Initial attempts at using M e l to examine specific protein-DNA contacts between the A R D B D and multiple A R E s within the probasin promoter were ineffectual. Failure likely occurred because the buttressing effect of neighbouring protein-protein and/or protein-DNA interactions within the A R D B D - D N A cooperative complex compensated for any single nucleotide disrupted through D M S methylation. The conventional M e P assay, which employs the reducing agent P-mercaptoethanol to quench D M S methylation activity, was also ineffectual in probing A R D B D - A R E contacts within the probasin promoter from -426 to +28 base pairs. This failure to footprint A R D B D on the probasin promoter was due to the perturbing effects of p-mercaptoethanol on the A R D B D - D N A cooperative complex presumably through the disruption of protein-protein interactions. In contrast, P-mercaptoethanol did not disturb an A R D B D homodimer bound to the canonical S R E suggesting that protein-protein interactions such as dilsulphide bridges are involved in cooperative A R D B D binding to the multiple A R E s found within the probasin promoter. Unlike the canonical S R E , P-mercaptoethanol also disturbed A R D B D binding to the Class II A R E , G - l . This last observation tentatively suggests that protein-protein interactions that are disruptable by P-mercaptoethanol arise from A R D B D homodimer binding to individual A R E s found in AR-specifically regulated genes but do not arise from A R D B D binding to the promiscuous viral-derived S R E . This last supports the argument that nucleotide sequence of natural A R E s affects the conformation of the bound A R D B D . Employing E M S A to stop D M S methylation activity instead of P-mercaptoethanol provided A R D B D - D N A contact information for the highly cooperative probasin promoter. In addition to the previously described A R E l and A R E 2 , two novel A R E s were discovered and were labelled G - l and G-2 because of their unusual hypersensitivity to D M S methylation at guanines where protection from 88 methylation was expected. For a guanine to be hypersensitive to D M S methylation implied local structural distortion of the D N A that resulted from A R D B D binding (Espinas et al, 1995; Ramesh and Nagaraja, 1996; Reid and Nelson, 2001). The hypersensitivity to methylation was found to be intrinsic to the nucleotide sequence of both G - l and G-2 and the hypersensitive guanines were found to be necessary for the A R D B D - D N A interaction. Here was physical evidence that primary nucleotide sequence of an A R E w i l l dictate how the A R D B D binds D N A and there appeared to be at least two classes of A R E s according to their M e P pattern upon A R D B D binding. Class I A R E s are represented by the probasin promoter A R E l and A R E 2 , which displayed conventional A R D B D contacts. Class II A R E s are represented by the probasin promoter G - l and G-2 binding elements, which display unusual hypersensitivity to D M S methylation upon A R D B D binding. Both Class I and Class II A R E s were found within the human P S A enhancer implying that differential binding of the A R D B D to A R E s that vary in primary nucleotide sequence is not merely a phenomenon isolated to the rat probasin proximal promoter. When the nucleotide sequences of established Class I or Class II A R E s were aligned, it became apparent that there were distinctive as well as common features between the class consensus sequences. Both classes of A R E s possess a cytosine at position -3 and a guanine at +3 where the arginine within each A R D N A recognition a-helix within the A R D B D homodimer is expected to hydrogen bond to guanine. Additionally both classes possess an adenine at position - 4 where a V a n der Waal's contact is expected between valine and the methyl group of a thymidine. Relative to Class I A R E s , Class II A R E s distinctively possessed an AT- r i ch spacer region, a thymidine at position -13 (outside of the primary sequence of the A R E ) , and a purine at position +7 in the 3' half site in contrast to the highly consensus pyrimidine found within Class I A R E s at this position. Site directed mutagenesis of these distinctive features revealed Class II type binding by the A R is primarily distinguished by the consensus purine at position +7 where its conversion to a pyrimidine changed the 3' half site M e P pattern to that of a Class I A R E but did not affect the 5' half site M e P pattern. Similarly, site directed mutagenesis of the Class II A R E spacer region or conversion of the -13 thymidine to a guanine appeared to decrease the 89 hypersensitivity to D M S methylation only within the 5' half site. Overall, the site directed mutagenesis effects on the D M S protection and hypersensitivity pattern o f G - l were confined to one particular half site implying that an A R D B D monomer, within the larger A R D B D homodimer, w i l l recognize and bind uniquely to its own half site. A R E s found within, AR-specifically regulated genes are invariably asymmetric inverse palindromes suggesting that each half site nucleotide sequence conveys specific and functionally relevant information to its own bound A R monomer. The asymmetry noted in mammalian A R E s could itself dictate function to the bound homodimer apart from variation in nucleotide sequence of the A R E or its relative location. Scatchard analysis of the individual proximal probasin promoter A R E s revealed that each had a weak specific binding affinity for the A R D B D . When combinations of Class I and Class II A R E s were examined, the highest cooperatively interacting permutation was the promoter fragment containing G 1 - A R E 2 - G 2 . Half-saturated binding occurred at 0.11 n M for this highly cooperative fragment, which was over 50 times stronger binding than A R E 2 , the individual element with the highest D N A binding affinity. Intriguingly, adding A R E l to this combination weakened overall binding affinity and the binding curve for the larger A R E 1 - G 1 -A R E 2 - G 2 fragment appeared to be biphasic: at low concentrations, A R D B D appears to first occupy G 1 - A R E 2 - G 2 followed by occupation of A R E l at higher A R D B D concentration. These data suggest that Class I and Class II A R E s interact in a complex manner and that both classes are essential for the dramatic cooperativity observed on the probasin promoter. Transient transfection analyses of the probasin promoter fragment combinations augmented and complemented the above binding affinity data. Synergistic transcriptional activation was observed with the G 1 - A R E 2 - G 2 and A R E 1 -G1 - A R E 2 - G 2 promoter fragments. A R E 1 - G 1 - A R E 2 - G 2 exhibited a biphasic curve that resulted in a 5-fold higher transcriptional response relative to G 1 - A R E 2 -G2. Complementing the binding affinity data, at low concentrations o f R1881 the lower part of the biphasic curve suggests that G 1 - A R E 2 - G 2 , the region with the highest D N A binding affinity, is fully occupied by activated A R and synergistic 90 transcription ensues. A t higher concentrations of R1881, however, the higher part o f the biphasic curve suggests that A R E l is now occupied and maximal transcription is permitted. Like the binding affinity data, these transfection data imply that both Class I and Class II A R E s are required for maximal transcriptional activation by the A R . Four supplementary AR-regulated promoter or enhancer regions were examined using the modified M e P assay: the m E - R A B P promoter, the rat A R enhancer, the proximal human p21 promoter, and the upstream rat probasin promoter from -705 to —426 base pairs. In addition to Class I and Class II A R D B D patterns of methylation protection and hypersensitivity, other novel patterns of A R D B D binding were discovered. A R D B D monomer half site binding was observed along with clusters of nested A R D B D binding, particularly within the -705 to -426 base pair probasin promoter. These A R D B D footprints revealed both conventional and expected A R D B D - D N A contacts as well as unconventional hypersensitivity to methylation such as that observed within Class II A R E s . Although the above novel A R D B D binding patterns have yet to show functional significance either in affinity binding or transfection studies, these last results imply that the A R D B D may not be restricted to binding D N A regulatory regions as a homodimer in a traditional head-to-head orientation. Instead, other A R D B D configurations and orientations may be possible depending on D N A sequence and promoter context. When the sequence of the probasin promoter from -426 to +28 base pairs was analyzed using the internet based software, Matinspector 2.0, it was discovered that the nucleotide sequence of the Class II A R E s , G - l and G-2, coincided with the near perfect binding sequence for the H M G - b o x architectural protein Sox-4. While the E M S A of recombinant full-length histidine tagged Sox-4 resulted in a shifted protein-D N A complex, no footprint was discerned using the modified M e P assay using N a O H cleavage. When Sox-4 was combined with the A R D B D , however, the number of D N A half sites occupied by A R D B D increased. Further work is required to determine whether Sox-4 makes specific contacts with the probasin promoter and whether the apparent Sox-4 enhancement of A R D B D occupation has an authentic effect on the regulation of prostate-specific and AR-specific genes. 91 4.2 Future Projections While other research groups have successfully purified active full length A R , this has yet to be accomplished in our lab although the ability to do so is close at hand. The problem appears to be solubility and maintaining long-term activity of the full length A R after purification from bacterial, mammalian or insect protein purification systems. Once purified, applying the M e P assay to the full length A R w i l l augment the data we have generated using the A R D B D especially with respect to the observed A R D B D monomer D N A binding. When the D N A binding abilities of the full length G R are compared to the binding abilities of the G R D B D , it is clear that the G R D B D is more capable of binding D N A as a monomer than the full length G R (Chalepakis et al, 1990; Segard-Maurel et al, 1996). Given the high sequence homology between the D B D domains of the A R and G R it is reasonable to expect similar results with the full length A R . Illustrating that full length A R generates a similar M e P pattern as the A R D B D w i l l also confirm our results. To determine the in vivo protein occupation of the probasin promoter and other AR-regulated gene regions one could employ in vivo footprinting techniques using either DNase I or D M S supplemented with ligation-mediated P C R ( L M - P C R ) (Mueller and Wold , 1989). So far there has yet to be a published footprint revealing in vivo A R - D N A interactions although other DNA-bind ing proteins have been footprinted in vivo and shown to interact with AR-regulated gene regions (Scarlett and Robins, 1995). The lack of data on A R interactions with D N A may be because these in vivo interactions are transient and therefore unable to be captured using present footprinting techniques or the A R - D N A complex may be easily disrupted in vivo by DNase I or D M S treatment. However, in vivo footprinting using L M - P C R is a challenging technique to master and its use in probing in vivo A R - D N A interactions may yet prove informative. A n alternative to L M - P C R in vivo footprinting is the Chromatin Immunoprecipitation (ChTP) assay, which permits the examination of specific protein-DNA interactions within an in vivo environment (Strahl-Bolsinger et al, 1997). ChIP involves the initial covalent crosslinking of proteins bound to D N A in 92 vivo, usually using ultraviolet light. After covalently linking the proteins o f interest to the D N A , the D N A is removed from the cell, purified from other nuclear and cellular components, and sheared into relatively large fragments. A t this point the protein-DNA fragments are enriched for a particular protein of interest by immunoprecipitating with a specific antibody. The specific protein-DNA fragments are then subject to protease digestion and the released D N A fragment that was occupied by the immunoprecipitated protein is P C R amplified using primers specific to the D N A region of interest. The sequence of the amplified D N A reveals the in vivo binding location of the protein factor. Unlike in vivo footprinting, the CHIP assay requires prior knowledge of the type of proteins expected to bind to a given D N A sequence because it relies on a protein-specific antibody to pull out or enrich for the D N A binding sequence of that protein. The in vitro binding affinity assays carried out in this study illustrate the important contribution of both Class I and Class II A R E s to A R D B D cooperative D N A binding to the proximal probasin promoter. The in vivo transient transfection studies carried out in L N C a P and PC3 cells in this study also point to the important contribution of Class I and II A R E s in AR-specific synergistic transcription. Similar contributions to cooperative binding and AR-specific synergistic transcription by Class I and II A R E s located within the human P S A enhancer have also been previously demonstrated using full length A R (Huang et al, 1999). Comparing binding affinity results with the in vivo transcription results indicates a strong association between AR-specific cooperative binding to the proximal probasin promoter and the AR-specific synergistic transcription that results from these cooperatively interacting D N A binding elements. The next step in probing this association between cooperative AR-specific binding and AR-specific synergistic transcription is to mutate the individual A R E s found with the proximal probasin promoter and examine the effects of mutation on both cooperative binding and synergistic transcription. Mutational analysis of the Class II A R E , G - l , was accomplished in this study but only so far as to examine the change in M e P pattern of the G - l half sites upon A R D B D binding. The binding affinity and transient transfection assays suggest that 93 altering Class II sites.to resemble Class I sites likely w i l l have an affect on observed cooperative binding and synergistic transcription. This is apparent since the D N A fragment that possessed the highest cooperativity was a unique combination o f two Class II sites flanking a Class I site; G 1 - A R E 2 - G 2 . Furthermore, the D N A fragment that displayed the highest synergistic transcription was the unique combination of A R E 1 - G 1 - A R E 2 - G 2 . Generating the larger probasin promoter fragments that contain site directed mutations such as the spacer mutant (ATT>CGG) , the +7 purine-pyrimidine mutant (A195C) and the -13 thymidine to guanine mutant (T215G), w i l l provide interesting data on their effects on cooperative binding in the presence o f the other unaltered probasin A R E s and their effects on synergistic transcription. The hypersensitivity to D M S methylation upon A R D B D binding to Class II A R E s is indicative of structural D N A distortion, however, it would be useful to probe the relative effects of Class I and Class II A R E nucleotide sequences on the D N A -bound conformation of the A R D B D homodimer. This could be accomplished by applying the trypsin/chymotrypsin protein digest assay or the differential antibody-protein interaction assay perfected in Dr. A n n Nardulli 's lab, which applied these techniqes to the DNA-bound E R (Wood et al, 1998). If binding to a Class I A R E resulted in a different A R D B D conformation compared to the conformation of the A R D B D bound to a Class II A R E then one would expect a variation in trypsin/chyotrypsin digest pattern or antibody-interaction. This variation in A R D B D conformation would be dependent on the nucleotide sequence o f the A R E . Examining the allostery between the A R and D N A from the perspective o f the protein conformation change would support the thesis that D N A sequence dictates specific protein conformation and therefore function. A n even more conclusive albeit labour-intensive approach would be to generate separate X-ray crystal structures of the A R D B D bound to a Class I A R E and to a Class II A R E . Such structures would reveal not only the weak interactions that are made between the A R D B D and a specific class of A R E , but also the specific A R D B D - D N A conformation that results from A R D B D binding to a specific A R E class type. From the resultant A R E class-specific conformation structure, one may even be able to deduce function. 94 While the M e P assay was able to generate a D M S protection and hypersensitive pattern visible on an autoradiograph film, exact orientation of A R D B D homodimers was not always resolvable using this assay. Recent published work has suggested that the A R was capable of binding D N A as a direct repeat homodimer (Zhou et al, 1997; Claessens et al, 2001). To obtain such an orientation and conformation, the A R D B D would have to homodimerize through protein-protein interactions that involve amino acid residues outside the dimerizing D Box . Such dimerizing ability has been observed for the thyroid hormone nuclear receptor but not for members of the steroid receptor subfamily (Umesono et al, 1991). Within the steroid receptor subfamily, dogma states that the head-to-head steroid receptor homodimer is the primary mode of occupying D N A binding sites, with salt bridges linking monomers through their D Boxes. For the most part, the data in this study support the convention of steroid receptor homodimers occupying D N A binding sites that contain two half sites in inverse orientation separated by three base pairs. However, A R D B D monomer and direct repeat binding could not be ruled out within some M e P patterns generated for various AR-regulated gene promoter and enhancer regions. Orientation and dimerization of the A R D B D bound to D N A could be resolved using X-ray crystallography. However, the large nucleotide variation observed in A R E s found within AR-regulated gene regions would prohibit the generation of an A R D B D - A R E X-ray crystal structure for every permutation of nucleotide sequence. Alternatively, a modification of the fluorescent resonance energy transfer (FRET) assay could be used to determine the orientation of the A R D B D bound to a specific A R E (Ramirez-Carrozzi and Kerppola, 2001). T h e A R D B D and the D N A strand containing the specific A R E could be tagged at one end using two different fluorescent molecules. After the A R D B D - D N A binding reaction has been brought to equilibrium, the change (or lack of change) in fluorescent signal can be measured. A change in wavelength would occur when the two fluorescent molecules are in close proximity and therefore interacting. Close proximity between fluorescent molecules would mean a specific orientation, such as parallel A R D B D -D N A orientation (depending on the fluorescent molecule location). I f there were no 95 change in wavelength one could assume that the fluorescent molecules are distant from one another and therefore the A R D B D - D N A orientation would be antiparallel, again depending on the location of the fluorescent molecule tags. Using fluorescent tagging in this way requires the FluorlmagerFSI fluorescence scanner (Molecular Dynamics) or something similar. This method has been successfully applied to determine the D N A binding orientation of Fos-Jun heterodimers (Ramirez-Carrozzi and Kerppola, 2001). 4.3 Conclusion The A R and its cognate ligand, androgen, are intricately involved in the growth, homeostasis, and development of normal prostate epithelial tissue. The complex control that androgens and the A R have over cell proliferation is also critical to the growth and proliferation of malignant prostate tissue. Understanding exactly how androgen action works to control proliferative molecular responses in prostate epithelium and other androgeh-regulated tissues is essential to understanding how prostate cancer initiates, progresses and, most importantly, how it may be controlled or cured. This study aimed to gain insight into how the A R specifically regulates target genes despite apparently recognizing the same D N A binding site as other steroid receptors. From the results of this study it is apparent that the architectural context of the promoter and enhancer regions of AR-regulated genes significantly contribute to AR-specific recognition and binding to D N A regulatory regions and to AR-specific transcriptional regulation. This architectural context refers to the nucleotide sequence of both the primary D N A binding element and the flanking sequence along with the relative spacing and orientation of A R E s within the D N A regulatory region. Nucleotide sequence and D N A binding site orientation can dictate function to a bound transcription factor and the subsequent allosteric exchange between protein and D N A w i l l permit the specific cooperative recruitment of other transcription factors and coregulators. The resultant protein-DNA enhanceosome then is capable of recruiting the transcriptional machinery and this interaction between critical D N A -96 bound proteins and the transcriptional apparatus culminates in synergistic transcription. Understanding how diseases like prostate cancer are able to harness molecular processes like androgen action w i l l lead to novel therapies and prevention strategies that could either manage the disease or perhaps cure. Conceivably, basic research into the architectural components of gene regulatory regions that direct an AR-specific and tissue-specific response could result in complex and intricately engineered gene driven therapies that target malignant prostate tissue. A t the very least, understanding how androgens and the A R control gene expression in prostate epithelial tissue w i l l contribute to knowledge of molecular mechanisms that factor in prostate cancer onset and progression. 97 Bibliography Ankenbauer W , Strahle U and Schutz G . 1988. Synergistic action of glucocorticoid and estradiol responsive elements. Proc Natl Acad Sci U S A ; 85(20): 7526-30. Archer T, Lefebvre P, Wolford R and Hager G . 1992. Transcription factor loading on the M M T V promoter: a bimodal mechanism for promoter activation. Science; 2556(5051): 1573-6. Aumuller G , Seitz J, Li l ja H , Abrahamsson P and Scheit K . 1990. Species- and organ-specificity of secretory proteins derived from human prostate and seminal vesicles. Prostate; 17(1): 31-40. Beato M , Arnemann J, Chalepakis G , Slater E and Willmann T. 1987. Gene regulation by steroid hormones. J Steroid Biochem; 27(1-3): 9-14. Bhoite L and Stillman D . 1998. Residues in the Swi5 zinc finger protein that mediate cooperative D N A binding with the Pho2 homeodomain protein. M o l Ce l l B i o l ; 18(11): 6436-46. Brinkmann A , Faber P, van Rooij H C , Kuiper G , Ris C, Klaassen P, van der Korput J, Voorhorst M , van Laar J H and Mulder E . 1989. The human androgen receptor: domain structure, genomic organization and regulation of expression. J Steroid Biochem; 34(1-6): 307-10. Brookes D , Zandvliet D , Watt F , Russell P and M o l l o y P. 1998. Relative activity and specificity of promoters from prostate-expressed genes. Prostate; 35(1): 18-26. Brunelle A and Schleif R. 1987. Missing contact probing of DNA-prote in interactions. Proc Natl Acad Sci U S A ; 84(19): 6673-6. Carey M . 1998. The enhanceosome and transcriptional synergy. Ce l l ; 92(1): 5-8. Cato A and Weinmann J. 1988. Cato A and Weinmann J. 1988. Mineralocorticoid regulation of transcription of transfected mouse mammary tumor virus D N A in cultured kidney cells. J Ce l l B i o l ; 106(6): 2119-25. Chalepakis G , Schauer M , Cao X and Beato M . 1990. Efficient binding o f glucocorticoid receptor to its responsive element requires a dimer and D N A flanking sequences. D N A Cel l B i o l ; 9(5): 355-68. Chan B , Spassky A and Busby S. 1990. The organization of open complexes between Escherichia coli R N A polymerase and D N A fragments carrying promoters either with or without consensus -35 region sequences. Biochem J; 270(1): 141-8. 98 Chaudhuri P, Walker M , Beattie C and Gupta T D . 1981. Distribution o f steroid hormone receptors in human soft tissue sarcomas. Surgery; 90(2): 149-53. Chen Y , Ghosh S and Ghosh G . 1998. A novel D N A recognition mode by the NF-kappa B p65 homodimer. Nat Struct B i o l ; 5(1): 67-73. Churchill M and Travers A . 1991. Protein motifs that recognize structural features of D N A . Trends Biochem Sci; 16(3): 92-7. Claessens F , A l e n P, Devos, Peeters A , Verhoeven G and Rombauts W . 1996. The androgen-specific probasin response element 2 interacts differentially with androgen and glucocorticoid receptors. J B i o l Chem; 271(32): 19013-6. Claessens F , Verrijdt G , Schoenmakers E , Haelens A , Peeters B , Verhoeven G and Rombauts W . 2001. Selective D N A binding by the androgen receptor as a mechanism for hormone-specific gene regulation. J Steroid Biochem M o l B i o l ; 76(1-5): 23-30. Clark L , Matthews J and Hay R. 1990. Interaction of enhancer-binding protein EBP1 (NF-kappa B) with the human immunodeficiency virus type 1 enhancer. J V i r o l ; 64(3): 1335-44. Cleutjens K , van Eekelen C C , van der Korput H A , Brinkmann A and Trapman J. 1996. Two androgen response regions cooperate in steroid hormone regulated activity of the. prostate-specific antigen promoter. J B i o l Chem; 271(11): 6379-88. Cleutjens K , van der Korput H , van Eekelen C E , Sikes R, Fasciana C , Chung L and Trapman J. 1997. A 6-kb promoter fragment mimics in transgenic mice the prostate-specific and androgen-regulated expression of the endogenous prostate-specific antigen gene in humans. M o l Endocrinol; 11(9): 1256-65. Cleutjens K , van der Korput H A , Eekelen C V , van Rooij H , Faber P and Trapman J. 1997. A n androgen response element in a far upstream enhancer region is essential for high, androgen-regulated activity of the prostate-specific antigen promoter. M o l Endocrinol; 11(2): 148-62. Cul ig Z , Hobisch A , Cronauer M , Cato A , Hittmair A , Radmayr C, Eberle J, Bartsch G and Klocker H . 1993. Mutant androgen receptor detected in an advanced-stage prostatic carcinoma is activated by adrenal androgens and progesterone. M o l Endocrinol; 7(12): 1541-50. Cul ig Z , Stober J, Gast A , Peterziel H , Hobisch A , Radmayr C , Hittmair A , Bartsch G , Cato A C , Klocker H . 1996. Activation of two mutant androgen receptors from human prostatic carcinoma by adrenal androgens and metabolic derivatives of testosterone. Cancer Detect Prev; 20(1): 68-75. 99 Dahlman-Wright K , Siltala-Roos H , Carlstedt-Duke J and Gustafsson J. 1990. Protein-protein interactions facilitate D N A binding by the glucocorticoid receptor DNA-bind ing domain. J B i o l Chem; 265(23): 14030-5. Darne C, More l L , Claessens, Manin F, Fabre S, Veyssiere G , Rombauts W and Jean C. 1997. Ubiquitous transcription factors NF1 and S p l are involved in the androgen activation of the mouse vas deferens protein promoter. M o l Ce l l Endocrinol; 132(1-2): 13-23. DeFfaseth P, Lohman T and Record M T Jr. 1977. Nonspecific interaction of lac repressor with D N A : an association reaction driven by counterion release. Biochemistry; 16(22): 4783-90. DeVos P, Claessens F, Peeters B , Rombauts W , Heyns W and Verhoeven G . 1993. Interaction of androgen and glucocorticoid receptor DNA-bind ing domains with their response elements. M o l Ce l l Endocrino; 90(2): R l 1-6. Diebold R, Rajaram N , Leonard D and Kerppola T. 1998. Molecular basis of cooperative D N A bending and oriented heterodimer binding in the N F A T 1 - F o s - J u n - A R R E 2 complex. Proc Natl Acad Sci U S A ; 95(14): 7915-20. Dodd J, Sheppard P and Matusik R. 1983. Characterization and cloning o f rat dorsal prostate m R N A s . Androgen regulation of two closely related abundant m R N A s . J B i o l Chem; 258(17): 10731-7. D u W , Thanos D and Maniatis T. 1993. Mechanisms of transcriptional synergism between distinct virus-inducible enhancer elements. Ce l l ; 74(5): 887-98. El lwood K , Huang W , Johnson R and Carey M . 1999. Multiple layers o f cooperativity regulate enhanceosome-responsive R N A polymerase II transcription complex assembly. M o l Ce l l B i o l ; 19(4): 2613-23. Espinas M , Roux J, Pictet R and Grange T. 1995. Glucocorticoids and protein kinase A coordinately modulate transcription factor recruitment at a glucocorticoid-responsive unit. M o l Ce l l B i o l ; 15(10): 5346-54. Evans B , Y u n Z , Close J, Tregear G , Kitamura N , Nakanishi S, Callen D , Baker E , Hyland V and Sutherland G . 1988. Structure and chromosomal localization of the human renal kallikrein gene. Biochemistry; 27(9): 3124-9. Frappier L and O'Donnell M . 1992. E B N A 1 distorts oriP, the Epstein-Barr virus latent replication origin. J V i r o l ; 66(3): 1786-90. Frick J and Aul i tzky W . 1991. Physiology of the prostate. Infection; 19 Suppl(3): SI 15-8. 100 Gaddipati J, M c L e o d D , Heidenberg H , Sesterhenn I, Finger M , M o u l J and Srivastava S. 1994. Frequent detection of codon 877 mutation in the androgen receptor gene in advanced prostate cancers. Cancer Res; 54(11): 2861-4. Gewirth D and Sigler P. 1995. The basis for half-site specificity explored through a non-cognate steroid receptor-DNA complex. Nat Struct B i o l ; 2(5): 386-94. Gilmore T. 1992. Role of rel family genes in normal and malignant lymphoid cell growth. Cancer Surv; 15(: 69-87. Gonzalez M and Robins D . 2001. Oct-1 preferentially interacts with androgen receptor in a DNA-dependent manner that facilitates recruitment of S R C - 1 . J B i o l Chem; 276(9): 6420-8. Govindan M . 1990. Specific region in hormone binding domain is essential for hormone binding and trans-activation by human androgen receptor. M o l Endocrinol; 4(3): 417-27. Grad J, Dai J, W u S and Burnstein K . 1999. Multiple androgen response elements and a M y c consensus site in the androgen receptor (AR) coding region are involved in androgen-mediated up-regulation of A R messenger R N A . M o l Endocrinol; 13(11): 1896-911. Green J, Greenberg N , Ashendel C, Barrett J, Boone C, Getzenberg R, Henkin J, Matusik R, Janus T and cher H . 1998. Workgroup 3: transgenic and reconstitution models of prostate cancer. Prostate; 36(1): 59-63. Greenberg N , DeMayo F, Sheppard P, Barrios R, Lebovitz R, Finegold M , Angelopoulou R, Dodd, Duckworth M and Rosen J. 1994. The rat probasin gene promoter directs hormonally and developmentally regulated expression of a heterologous gene specifically to the prostate in transgenic mice. M o l Endocrinol; 8(2): 230-9. Gronemeyer H and Pongs O. 1980. Localization of ecdysterone on polytene chromosomes of Drosophila melanogaster. Proc Natl Acad Sci U S A ; 77(4): 2108-12. Ham J, Thomson A , Needham M , Webb P and Parker M . 1988. Characterization of response elements for androgens, glucocorticoids and progestins in mouse mammary tumour virus. Nucleic Acids Res; 16(12): 5263-76. Higgins S and Gehring U . 1978. Molecular mechanisms ofsteroid hormone action. A d v Cancer Res; 28(: 313-97. Hobisch A , Cul ig Z , Radmayr C, Bartsch G , Klocker H and Hittmair A . 1995. Distant metastases from prostatic carcinoma express androgen receptor protein. Cancer Res; 55(14): 3068-72. 101 Holmbeck S, Dyson H and Wright P. 1998. DNA-induced conformational changes are the basis for cooperative dimerization by the D N A binding domain of the retinoid X receptor. J M o l B i o l ; 284(3):533-9 Huang W , Shostak Y , Tarr P, Sawyers C and Carey M . 1999. Cooperative assembly o f androgen receptor into a nucleoprotein complex' that regulates the prostate-specific antigen enhancer. J B i o l Chem; 274(36): 25756-68. Huth J, Bewley C , Nissen M , Evans J, Reeves R, Gronenborn A and Clore G . 1997. The solution structure of an H M G - I ( Y ) - D N A complex defines a new architectural minor groove binding motif. Nat Struct B i o l ; 4(8): 657-65. Isaacs J, Furuya Y and Berges R. 1994. The role of androgen in the regulation of programmed cell death/apoptosis in normal and malignant prostatic tissue. Semin Cancer B i o l ; 5(5): 391-400. John S, Reeves R, L i n J, Chi ld R, Leiden J, Thompson C and Leonard W . 1995. Regulation of cell-type-specific interleukin-2 receptor alpha-chain gene expression: potential role of physical interactions between Elf-1 , H M G - I ( Y ) , and NF-kappa B family proteins. M o l Ce l l B i o l ; 15(3): 1786-96. Kasper S, Rennie P, Bruchovsky N , L i n L , Cheng H , Snoek R, Dahlman-Wright K , Gustafsson J, Shiu R, Sheppard P and Matusik R. 1999. Selective activation of the probasin androgen-responsive region by steroid hormones. J M o l Endocrinol; 22(3): 313-25. Kasper S, Rennie P, Bruchovsky N , Sheppard P, Cheng H , L i n L , Shiu R, Snoek R and Matusik R. 1994. Cooperative binding of androgen receptors to two D N A sequences is required for androgen induction of the probasin gene. J B i o l Chem; 269(50): 31763-9. Kerppola T. 1998. Transcriptional cooperativity: bending over backwards and doing the flip. Structure; 6(5): 549-54. Koivisto P, Kononen J, Palmberg C, TAmmela T, Hyytinen E , Isola J, TrapmanJ J, Cleutjens K , Noordzij A , Visakorpi T and Kall ioniemi O. 1997. Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation therapy failure in prostate cancer. Cancer Res; 57(2): 314-9. Konrad L , Weber M , Groos S, Albrecht M and Aumuller G . 1998. Paracrine interaction in testicular somatic cells. Ital J Anat Embryol; 103(4 Suppl 1): 139-52. Lambert J and Nordeen S. 1998. Steroid-selective initiation of chromatin remodeling and transcriptional activation of the mouse mammary tumor virus promoter is controlled by the site of promoter integration. J B i o l Chem; 273(49): 32708-14. 102 Lareyre J, Reid K , Nelson C, Kasper S, Rennie P, Orgebin-Crist M and Matusik R. 2000. Characterization of an androgen-specific response region within the 5' flanking region of the murine epididymal retinoic acid binding protein gene. B i o l Reprod 2000 Dec;63(6):; 63(6): 1881-92. Laudet V , Hanni C, C o l l J, Catzeflis F and Stehelin D . 1992. Evolution of the nuclear receptor gene superfamily. E M B O ; 11(3): 1003-13. Laudet V , Stehelin D and Clevers H . 1993. Ancestry and diversity of the H M G box superfamily. Nucleic Acids Res; 21(10): 2493-501. Leger H , Sock E , Renner K , Grummet F and Wegner M . 1995. Functional interaction between the P O U domain protein Tst-l/Oct-6 and the high-mobility-group protein H M G -I /Y. M o l Ce l l B i o l ; 15(7): 3738-47. Lenardo M and Baltimore D . 1989. NF-kappa B : a pleiotropic mediator o f inducible and tissue-specific gene control. Ce l l ; 58(2): 227-9. Liao M , Zhou Z and Wilson E . 1999. Redox-dependent D N A binding o f the purified androgen receptor: evidence for disulfide-linked androgen receptor dimers. Biochemistry; 38(30): 9718-27. Lieberman B , Bona B , Edwards D and Nordeen S. 1993. The constitution of a progesterone response element. M o l Endocrinol; 7(4): 515-27. L i n C, Lau A , Y e h C, Chang C and Lue T. 2000. Upregulation o f L-plastin gene by testosterone in breast and prostate cancer cells: identification of three cooperative androgen receptor-binding sequences. D N A Cel l B i o l ; 19(1): 1-7. Linja M , Savinainen K , Saramaki O, Tammela T, Vessella R and Visakorpi T. 2001. Amplification and overexpression of androgen receptor gene in hormone-refractory prostate cancer. Cancer Res; 61(9): 3550-5. List H , Lozano C, L u J, Danielsen M , Wellstein A and Riegel A . 1999. Comparison of chromatin remodeling and transcriptional activation of the mouse mammary tumor virus promoter by the androgen and glucocorticoid receptor. Exp Ce l l Res; 250(2): 414-22. L i u Z and Little J. 1998. The spacing between binding sites controls the mode of cooperative DNA-protein interactions: implications for evolution of regulatory circuitry. J M o l B i o l ; 278(3): 331-8. Love J, L i X , Case D , Giese K , Grosschedl R and Wright P. 1995. Structural basis for D N A bending by the architectural transcription factor L E F - 1 . Nature; 376(6543): 791-5. 103 L u S, Jenster G and Epner D . 2000. Androgen induction of cyclin-dependent kinase inhibitor p21 gene: role of androgen receptor and transcription factor S p l complex. M o l Endocrinol; 14(5): 753-60. Luis i B , X u W , Otwinowski Z , Freedman L , Yamamoto K and Sigler P. 1991. Crystallographic analysis of the interaction of the glucocorticoid receptor with D N A . Nature; 352(6335): 497-505. Majors J and Varmus H . 1983. A small region of the mouse mammary tumor virus long terminal repeat confers glucocorticoid hormone regulation on a linked heterologous gene. Proc Natl Acad Sci U S A ; 80(19): 5866-70. Mao C, Carlson N and Little J. 1994. Cooperative DNA-protein interactions. Effects of changing the spacing between adjacent binding sites. J M o l B i o l ; 235(2): 532-44. Marcell i M , Ittmann M , Mariani S, Sutherland R, Nigam R, Murthy L , Zhao Y , DiConcin i D , Puxeddu E , Esen A , Eastham J, Weigel N and Lamb D . 2000. Androgen receptor mutations in prostate cancer. Cancer Res; 5(60): 944-9. Matuo Y , Nishi N , Muguruma Y , Yoshitake Y , Kurata N and Wada F. 1985. Localization of prostatic basic protein ("probasin") in the rat prostates by use o f monoclonal antibody. Biochem Biophys Res Commun; 130(1): 293-300. Matuo Y , Nishi N , Negi T, Tanaka Y and Wada F. 1982. Isolation and characterization of androgen-dependent non-histone chromosomal protein from dorsolateral prostate of rats. Biochem Biophys Res Commun; 109(2): 334-40. Matuo Y , Nishi N , Tanaka Y , Muguruma Y , Tanaka K and Wada F. 1986. Lobe-specific distribution of a 20,000-dalton nonhistone protein in the dorsolateral prostate of rats. Prostate; 8(2): 195-206. McDonald S, Brive L , Agus D , Scher H and E l y K . 2000. Ligand responsiveness in human prostate cancer: structural analysis of mutant androgen receptors from L N C a P and C W R 2 2 tumors. Cancer Res; 2317-22(9): 60(9):2317-22. Menetski J. 2000. The structure of the nuclear factor-kappaB protein-DNA complex varies with DNA-bind ing site sequence. J B i o l Chem 2000 Mar 17;275(11):; 275(11): 7619-25. Merika M and Thanos D . 2001. Enhanceosomes. Curr Opin Genet Dev; 11(2): 205-8. M o Y , aessen B , Johnston K and Marmorstein R. 1998. Structures of SAP-1 bound to D N A targets from the E74 and c-fos promoters: insights into D N A sequence discrimination by Ets proteins. M o l Ce l l ; 2(2): 201-12. Mueller P and W o l d B . 1989. In vivo footprinting of a muscle specific enhancer by ligation mediated P C R . Science; 246(4931): 780-6. 104 Nelson C, Hendy S and Romaniuk P. 1995. Relationship between P-box amino acid sequence and D N A binding specificity of the thyroid hormone receptor. The effects of half-site sequence in everted repeats. J B i o l Chem; 270(28): 16981-7. Nelson C, Hendy S, Shukin R, JCheng H , Bruchovsky N , Koop B and Rennie P. 1999. Determinants of D N A sequence specificity of the androgen, progesterone, and glucocorticoid receptors: evidence for differential steroid receptor response elements. M o l Endocrinol; 13(12): 2090-107. Ning Y and Robins D . 1999. A M L 3 / C B F a l p h a l is required for androgen-specific activation of the enhancer o f the mouse sex-limited protein (Sip) gene. J B i o l Chem; 274(43): 30624-30. Sato N and Ota M . 1988. Characterization of the nontransformed and transformed androgen receptor and heat shock protein 90 with high-performance hydrophobic-interaction chromatography. J Steroid Biochem; 31(3): 295-304. Ohmori Y , Schreiber R and Hamilton T. 1997. Synergy between interferon-gamma and tumor necrosis factor-alpha in transcriptional activation is mediated by cooperation between signal transducer and activator of transcription 1 and nuclear factor kappaB. J B i o l Chem; 272(23): 14899-907. Pang S, Dannull J, Kaboo R, Y i e X , Tso C, Miche l K , deKernion J and Belldegrun A . 1997. Identification of a positive regulatory element responsible for tissue-specific expression of prostate-specific antigen. Cancer Res; 57(2): 495-9. Payvar F , Wrange O, Carlstedt-Duke J, Okret S, Gustafsson J and Yamamoto K . 1981. Purified glucocorticoid receptors bind selectively in vitro to a cloned D N A fragment whose transcription is regulated by glucocorticoids in vivo. Proc Natl Acad Sci U S A ; 78(11): 6628-32. Petrow V . 1986. The dihydrotestosterone (DHT) hypothesis of prostate cancer and its therapeutic implications. Prostate; 9(4): 343-61. Ramesh V and Nagaraja V . 1996. Sequence-specific D N A binding o f the phage M u C protein: footprinting analysis reveals altered D N A conformation upon protein binding. J M o l B i o l ; 260(1): 22-33. Ramirez-Carrozzi V and Kerppola T. 2001. Dynamics of F o s - J u n - N F A T l complexes. Proc Natl Acad Sci U S A ; 98(9): 4893-8. Reid K and Nelson C. 2001. Improved methylation protection-based D N A footprinting to reveal structural distortion of D N A upon transcription factor binding. Biotechniques; 30(1): 20-2. 105 Rennie P, Bruchovsky N , Leco K , Sheppard P, McQueen S, Cheng H , Snoek R, Hamel, Bock M and MacDonald B . 1993. Characterization of two cis-acting D N A elements involved in the androgen regulation of the probasin gene. M o l Endocrinol; 7(1): 23-36. Richard-Foy H and Hager G . 1987. Sequence-specific positioning o f nucleosomes over the steroid-inducible M M T V promoter. E M B O J; 6(8): 2321-8. Riegman P, Vlietstra R, Klaassen P, van der Korput J, van Kessel A , Romijn J and Trapman J. 1989. The prostate-specific antigen gene and the human glandular kallikrein-1 gene are tandemly located on chromosome 19. F E B S Lett; 247(1): 123-6. Riegman P, Vlietstra R, van der Korput J, Brinkmann A and Trapman J. 1991. The promoter of the prostate-specific antigen gene contains a functional androgen responsive element. M o l Endocrinol; 5(12): 1921-30. Roche P, Hoare S and Parker M . 1992. A consensus DNA-bind ing site for the androgen ' receptor. M o l Endocrinol; 6(12): 2229-35. Rundlett S and Miesfeld R. 1995. Quantitative differences in androgen and glucocorticoid receptor D N A binding properties contribute to receptor-selective transcriptional regulation. M o l Ce l l Endocrinol; 109(1): 1-10. Saatok T, Dahlberg E , Bylund P, Eriksson E and Gustafsson J. 1984. Steroid hormone receptors, protein, and D N A in erector spinae muscle from scoliotic patients. C l i n Orthop; 183(: 197-207. Savory J, Prefontaine G , Lamprecht C, Liao M , Walther R, Lefebvre Y and Hache R. 2001. Glucocorticoid receptor homodimers and glucocorticoid-mineralocorticoid receptor heterodimers form in the cytoplasm through alternative dimerization interfaces. M o l Ce l l B i o l ; 21(3): 781-93. Scarlett C and Robins D . 1995. In vivo footprinting of an androgen-dependent enhancer reveals an accessory element integral to hormonal response. M o l Endocrinol; 9(4): 413-23. Scheidereit C , Geisse S, Westphal H and Beato M . 1983. The glucocorticoid receptor binds to defined nucleotide sequences near the promoter of mouse mammary tumour virus. Nature; 304(5928): 749-52. . Schoenmakers E , A l e n P, Verrijdt G , Peeters B , Verhoeven G , Rombauts W and Claessens F. 1999. Differential D N A binding by the androgen and glucocorticoid receptors involves the second Zn-fmger and a C-terminal extension of the DNA-bind ing domains. Biochem J; 341(Pt 3): 515-21. Schule R, Mul ler M , Kaltschmidt C and Renkawitz R. 1988. Many transcription factors interact synergistically with steroid receptors. Science; 242(4884): 1418-20. 106 Schuur E , Henderson G , Kmetec L , Mi l l e r J, Lamparski H and Henderson D . 1996. Prostate-specific antigen expression is egulated by anupstream enhancer. J B i o l Chem; 271(12): 7043-51. Schwabe J, Chapman L , Finch J and Rhodes D . 1993. The crystal structure o f the estrogen receptor DNA-bind ing domain bound to D N A : how receptors discriminate between their response elements. Ce l l ; 75(3): 567-78. Segard-Maurel I, Rajkowski K , Jibard N , Schweizer-Groyer G , Baulieu E and Cadepond F. 1996. Glucocorticosteroid receptor dimerization investigated by analysis of receptor binding to glucocorticosteroid responsive elements using a monomer-dimer equilibrium model. Biochemistry; 35(5): 1634-42. Senear D , Ross J and Laue T. 1998. Analysis of protein and DNA-mediated contributions to cooperative assembly of protein-DNA complexes. Methods; 16(1): 3-20. Spence A , Sheppard P, Davie J, Matuo Y , Nishi N , McKeehan W , Dodd J and Matusik R. 1989. Regulation of a bifunctional m R N A results in synthesis of secreted and nuclear probasin. Proc Natl Acad Sci U S A ; 86(20): 7843-7. Strahl-Bolsinger S, Hecht A , Luo K and Grunstein M . 1997. SIR2 and SIR4 interactions differ in core and extended telomeric heterochromatin in yeast. Genes Dev; 11(1): 83-93. Strahle U , Munsterberg A , Mestril R, K lock G , Ankenbauer W , Schmid W and Schutz G . 1988. Cooperative action of the glucocorticoid receptor and transcription factors. Co ld Spring Harb Symp Quant B i o l ; 53(Pt 2): 835-41. Stumpf W and Sar M . 1976. Steroid hormone target sites in the brain: the differential distribution of estrogin, progestin, androgen and glucocorticosteroid. J Steroid Biochem; 7(11-12): 1163-70. Suzuki H , Sato N , Watabe Y , Masai M , Seino S and Shimazaki J. 1993. Androgen receptor gene mutations in human prostate cancer. J Steroid Biochem M o l B i o l ; 46(6): 759-65. Szymczyna B and Arrowsmith C. 2000. D N A binding specificity studies of four E T S proteins support an indirect read-out mechanism of protein-DNA recognition. J B i o l Chem; 275(37): 28363-70. Tan J, Marschke K , Ho K , Perry S, Wilson E and French F. 1992. Response elements of the androgen-regulated C3 gene. J B i o l Chem; 267(7): 4456-66. 107 Tan J, Sharief Y , Hami l K , Gregory C, Zang D , Sar M , Gumerlock P, de Vere R, Pretlow T, Harris S, Wilson E , Mohler J and French F. 1997. ehydroepiandrosterone activates mutant androgen receptors expressed in the androgen-dependent human prostate cancer xenograft C W R 2 2 and L N C a P cells. M o l Endocrinol; 11(4): 450-9. Taplin M , Bubley G , Shuster T, Frantz M , Spooner A , Ogata G , Keer H and Balk S. 1995. Mutation of the androgen-receptor gene in metastatic androgen-independent prostate cancer. N Engl J Med ; 332(21): 1393-8. Thanos D and Maniatis T. 1992. The high mobility group protein H M G I(Y) is required for NF-kappa B-dependent virus induction of the human IFN-beta gene. Ce l l ; 71(5): 777-89. Thanos D and Maniatis T. 1995. Virus induction of human I F N beta gene expression requires the assembly o f an enhanceosome. Ce l l ; 83(7): 1091-100. Thomson A . 2001. Role of androgens and fibroblast growth factors in prostatic development. Reproduction. Reproduction; 121(2): 187-95. Truss M , Bartsch J, Schelbert A , Hache R and Beato M . 1995. Hormone induces binding of . receptors and transcription factors to a rearranged nucleosome on the M M T V promoter in vivo. E M B O J; 14(8): 1737-51. Tsai S, Carlstedt-Duke J, Weigel N , Dahlman K , Gustafsson J, Tsai M and O'Malley B . 1988. Molecular interactions of steroid hormone receptor with its enhancer element: evidence for receptor dimer formation. Ce l l ; 55(2): 361-9. Tsai S, Tsai M and O'Malley B . 1989. Cooperative binding of steroid hormone receptors contributes to transcriptional synergism at targetenhancer elements. Ce l l ; 57(3): 443-8. Umesono K and Evans R. 1989. Determinants of target gene specificity for steroid/thyroid hormone receptors. Ce l l ; 57(7): 1139-46 Umesono K , Murakami K , Thompson C and Evans R. 1991. Direct repeats as selective response elements for the thyroid hormone, retinoic acid, and vitamin D3 receptors. Ce l l ; 65(7): 1255-66. VanTilborg M , Lefstin J, Kruiskamp M , Teuben J, Boelens R, Yamamoto K and Kaptein R. 2000. Mutations in the glucocorticoid receptor DNA-bind ing domain mimic an allosteric effect of D N A . J M o l B i o l ; 301(4): 947-58. Vashee S, Melcher K , Ding W , Johnston S and Kodadek T. 1998. Evidence for two modes of cooperative D N A binding in vivo that do not involve direct protein-protein interactions. C u r r B i o l ; 8(8): 452-8. 108 Vashee S, Wi l l i e J and Kodadek T. 1999. Synergistic activation of transcription by physiologically unrelated transcription factors through cooperative DNA-binding . Biochem Biophys Res Commun; 247(2): 530-5. Veldscholte J, Berrevoets C, Zegers N , van der Kwast T, Grootegoed J and Mulder E . 1992. Hormone-induced dissociation of the androgen receptor-heat-shock protein complex: use of a new monoclonal antibody to distinguish transformed from nontransformed receptors. Biochemistry; 31(32): 7422-30. Verrijdt G , Schoenmakers E , Haelens A , Peeters P, Verhoeven G , Rombauts W and Claessens F. 2000. Change of specificity mutations in androgen-selectiveenhancers. Evidence for a role of differential D N A binding by the androgen receptor. J B i o l Chem; 275(16): 12298-30165. Visakorpi T, Hyytinen E , Koivisto P, Tanner M , Keinanen R, Palmberg C, Palotie A , Tammela T, Isola J and Kall ioniemi O. 1995. In vivo amplification o f the androgen receptor gene and progression of human prostate cancer. Nat Genet; 9(4): 401-6. Wang J, El lwood K , Lehman A , Carey M and She Z . 1999. A mathematical model for synergistic eukaryotic geneactivation. J M o l B i o l ; 286(2): 315-25. Werner M , Bianchi M , Gronenborn A and Clore G . 1995. N M R spectroscopic analysis of the D N A conformation induced by the human testis determining factor S R Y . Biochemistry 1995 Sep 19;34(37):; 34(37): 11998-2004. W o l f D , Schulz P and Fittler F. 1992. Transcriptional regulation of prostate kallikrein-like genes by androgen. M o l Endocrinol; 6(5): 753-62. Wood J, Greene G and Nardulli A . 1998. Estrogen response elements function as allosteric modulators of estrogen receptor conformation. M o l Ce l l B i o l ; 18(4): 1927-34. Yamashita S. 2001. Histochemistry and cytochemistry o f nuclear receptors. Prog Histochem Cytochem; 36(2): 91-176. Y i e J, Liang S and Merika M . 1997. Intra- and intermolecular cooperative binding o f high-mobility-group protein I(Y) to the beta-interferon promoter. M o l Ce l l B i o l ; 17(7): 3649-62. Young C, Andrews P, Montgomery B and Tindall D . 1992. Tissue-specific and hormonal regulation of human prostate-specific glandular kallikrein. Biochemistry; 31(3): 818-24. Zhang X and Verdine G . 1999. A small region in H M G I(Y) is critical for cooperation with NF-kappaB on D N A . J B i o l Chem; 274(29): 20235-43. 109 Zhong H and Vershon A . 1997. The yeast homeodomain protein MATalpha2 shows extended D N A binding specificity in complex with M c m l . J B i o l Chem; 272(13): 8402-9. Zhou Z , Corden J and Brown T. 1997. Identification and characterization of a novel androgen response element composed of a direct repeat. J B i o l Chem; 272(13): 8227-35. Zilliacus J, Carlstedt-Duke J, Gustafsson J and Wright A . 1994. Evolution of distinct D N A -binding specificities within the nuclear receptor family of transcription factors. Proc Natl Acad Sci U S A ; 91(10): 4175-9. 110 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090177/manifest

Comment

Related Items