Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies on inter-species expression of photosynthesis genes in Rhodobacter capsulatus Zilsel, Joanna 1990

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

Item Metadata

Download

Media
831-UBC_1991_A6_7 Z54.pdf [ 5.88MB ]
Metadata
JSON: 831-1.0098553.json
JSON-LD: 831-1.0098553-ld.json
RDF/XML (Pretty): 831-1.0098553-rdf.xml
RDF/JSON: 831-1.0098553-rdf.json
Turtle: 831-1.0098553-turtle.txt
N-Triples: 831-1.0098553-rdf-ntriples.txt
Original Record: 831-1.0098553-source.json
Full Text
831-1.0098553-fulltext.txt
Citation
831-1.0098553.ris

Full Text

STUDIES ON INTER-SPECIES EXPRESSION OF PHOTOSYNTHESIS GENES IN RHODOBACTER CAPSULATUS by Joanna Zilsel B. Sc., University of British Columbia, 1983 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Microbiology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1990 © Joanna Zilsel, 1990 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 it 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. Department of Microbiology The University of British Columbia Vancouver, Canada D a t e September 25th,. 1990 DE-6 (2/88) ii ABSTRACT The primary amino acid sequences of the L, M, and H photosynthetic reaction center peptide subunits from a number of purple non-sulfur bacteria, including Rhodopseudomonas viridis, Rhodobacter sphaeroides, and Rhodobacter capsulatus have been previously shown to be highly homologous, and detailed X-ray crystallographic analyses of reaction centers from two species of purple non-sulfur bacteria, Rps. viridis and R. sphaeroides have shown that all recognized structural and functional features are conserved. Experiments were undertaken to determine whether genes encoding reaction center and light harvesting peptide subunits from one species could be functionally expressed in other species. Plasmid-borne copies of R sphaeroides and Rps. viridis pigment binding-peptide genes were independently introduced into a photosynthetically incompetent R. capsulatus mutant host strain, deficient in all known pigment-binding peptide genes. The R. sphaeroides puf operon, which encodes the L and M subunits of the reaction center as well as both peptide subunits of light harvesting complex I, was shown to be capable of complementing the mutant R. capsulatus host. Hybrid reaction centers, comprised of R. sphaeroides-encotied L and M subunits and an R. capsulatus-encoded H subunit, were formed in addition to the R. sphaeroides-encoded LHI complexes. These hybrid cells were capable of photosynthetic growth, but their slower growth rates under low light conditions and their higher fluorescence emission levels relative to cells containing native complexes, indicated an impairment in energy transduction. The Rps. viridis puf operon was found to be incapable of functional expression in the R. capsulatus mutant host. Introduction of a plasmid-borne copy of the Rps. viridis puhA gene, which encodes the H subunit of the reaction center, into host cells already containing the Rps. viridis puf iii operon, such that all structural peptides of the Rps. viridis reaction center were present, still did not permit stable assembly of Rps. viridis photosynthetic complexes. RNA blot analysis demonstrated that the barrier to functional expression was not at the level of transcription. Differences between Rps. viridis and R. sphaeroides that may account for their differing abilities to complement the R. capsulatus mutant host strain are discussed. iv TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables vi List of Figures vii Abbreviations and symbols x Acknowledgements xi INTRODUCTION 1 MATERIALS AND METHODS 15 1. Growth and manipulation of bacterial strains 15 2. Plasmid constructions 17 3. RNA isolation 28 4. RNA blotting and probing 30 5. Fluorescence detection 32 6. Spectrophotometric analyses 32 7. Bete-galactosidase assays 33 RESULTS 34 1. Rhodobacter sphaeroides study 34 a. Absorption spectroscopy 34 b. SDS-PAGE analysis of chromatophore membrane proteins.... 36 V c. Fluorescence emission 36 d. Photosynthetic growth rates 39 2. Rhodopseudomonas viridis study 39 a. Oxygen and ammonia effects on nifHDK promoter activity in expression vector pNIF215 39 b. Absorption spectroscopy 43 (i) Rps. viridis put operon expression in U43(pJZ1) 43 (ii) Rps. viridis puhA gene expression in U43(pJZ1)+(pJZ6) 48 c. RNA blot analysis. 52 (i) Rps. viridis put operon expression in U43 52 (ii) Rps. viridis puhA gene expression in U43 55 d. R. capsulatus LHI complex levels in cells with and without reaction center complexes as determined by absorption spectroscopy 56 DISCUSSION 60 1. R. sphaeroides study 61 2. Rps. viridis study 64 CONCLUSIONS 78 REFERENCES 81 APPENDIX 94 vi List of Tables Table I: Percent amino acid identities in alignments of reaction center and light harvesting subunits from ft capsulatus, ft sphaeroides and Rps. viridis 13 Table II: Plasmids and bacterial strains used 29 Table III: Befa-galactosidase specific activities in ft capsulatus strains U43(pJZ4) and U43(pJZ5) grown in ammonia-free or ammonia-supplemented media under high O2 . low O 2 and anaerobic conditions 43 Table IV:' Putative Shine-Dalgarno sites in ft capsulatus and Rps. viridis pufoperon and puhA genes 68 vii List of Figures Figure 1: Schematic diagram of Rps. viridis reaction center and o/Ci complex, showing cofactor arrangements and electron transfer times between primary reactants' 2 Figure 2: Schematic diagram of the photosynthetic apparatus of R. capsulatus and relevant genes and corresponding Rps. viridis genes 9 Figure 3: Expression vectors pJAJ9 and pNIF215 19 Figure 4: Plasmids pCT1 and pTB999 20 Figure 5: Construction of plasmids pJZ1 and pJZ2 21 Figure 6 : Construction of plasmids pJZ4 and pJZ5 24 Figure 7: Construction of plasmid pJZ6 26 Figure 8: Absorption spectra of intact cells of R. capsulatus strains U43, U43(pJAJ9), U43(pCT1), and U43(pTB999) 35 Figure 9: SDS-polyacry lamide gel electrophoresis of chromatophore vesicles from R. capsulatus strains U43, U43(pCT1) and U43(pTB999) 37 viii Figure 10: Relative fluorescence emission levels of U43(pTB999) and U43(pCT1) 38 Figure 11: Relative growth rates of ft capsulatus B10, and strain U43 containing plasmids (pJAJ9), U43(pTB999) and U43(pCT1) 40 Figure 12: Absorption spectra of membranes prepared from YPS-grown cells of R. capsulatus strains U43, U43(pJAJ9), and U43(pJZ1), with second derivatives inset 46 Figure 13: Absorp t ion s p e c t r a of m e m b r a n e s pre-pared from RCWDMSO/pyruvate/fructose-grown cells of R. capsulatus strains U43, U43(pJAJ9), U43(pJZ1) and U43(pJZ1+pJZ6) 49 Figure 14: A. Agarose/formaldehyde gel of RNA from photosynthetically grown Rps. viridis and low 02 grown (+/- NH4) ft capsulatus U43 strains containing various plasmids, and B. autoradiogram of an [a- 3 2P]-labelled Rps. viridis put DNA-probed blot of this gel 53 Figure 15: Autoradiogram of an [a- 3 2P]-labelled Rps. viridis puhA DNA-probed blot of the gel described in Figure 14 57 Figure 16: Absorption spectra of intact cells of ft capsulatus strains U43(pRC77) and U43(pTB999) 59 ix Figure 17: Molecular structures of pigment and cofactor molecules of the reaction centers of ft capsulatus and Rps. viridis 70 X ABBREVIATIONS AND SYMBOLS Bchl bacteriochlorophyll bp base pair BSA bovine serum albumin DNA deoxyribonucleic acid DMSO dimethyl sulfoxide EDTA ethylenediaminetetra-acetic acid kan kanamycin kb kilobase LH light harvesting mRNA messenger RNA ORF open reading frame PAGE polyacrylamide gel electrophoresis PSI I photosystem II of higher plants pufA, pufB structural genes of LHI a and p peptides pufL, pufM structural genes of the RC L and M subunits pufQ, pufX genes of unknown function, required for photosynthetic growth in R. capsulatus and R. sphaeroides puhA structural gene of the RC H subunit RC reaction center RCV* ammonia-free RCV medium RNA ribonucleic acid rRNA ribosomal RNA SDS sodium dodecyl sulfate SSC standard sodium citrate TBE tris borate EDTA Tc tetracycline UV ultraviolet xi Acknowledgements I thank Tom Beatty, thesis supervisor extraordinaire, for his patience, encouragement, and enthusiastic guidance throughout the course of my research. A rare atmosphere of good will, co-operation and excitement permeates his lab. It has been a privilege to work with Beatty lab members Cheryl Wellington, Tim Lilburn, Farahad Dastoor and Heidi LeBlanc. Their contributions to this thesis have been considerable, and their friendships wonderful. I especially thank Tim Lilburn, who did the protein work in the R. sphaeroides study. I am grateful to Dan Walker for help with the statistical approach used in the appendix. To my parents, who have helped me in so many ways over the years, I express heart-felt gratitude. This thesis is dedicated to my partner, Andre Sobolewski, who put his own career "on hold" for two years, allowing me to return to school while he took primary responsibility for the care of our son, Daniel. Without his constant support and encouragement, this thesis would not have been possible. 1 INTRODUCTION Photosynthesis is a fundamental biological process by which light is converted to chemical energy. The minimal structural unit capable of performing this energy conversion is an integral membrane pigment-peptide complex termed the photosynthetic reaction center (RC). Within the RC, charge separation across the membrane occurs upon absorption of light. The process can be simply summarized as follows (see below and Fig. 1): absorption of light energy by pigment molecules located near the periplasmic side of the RC causes an electron to be transferred, through several discrete intermediates, to an acceptor molecule located near the cytoplasmic side. After a complex series of reactions, the electron is returned to the photo-oxidized initial donor. This cyclical electron flow is coupled to proton translocation across the membrane. The net effect is the production of a trans-membrane pH and electrical potential, which can be used to drive ATP synthesis (23,34,52,70). For many years, researchers have sought to elucidate the precise molecular mechanisms by which energy conversion in the RC and associated light harvesting (LH) complexes occurs. One group of photosynthetic organisms, the purple non-sulfur bacteria, have been extraordinarily useful as experimental systems for the study of such fundamental problems of photosynthesis. Their photosynthetic apparatus is relatively simple, although clearly structurally and functionally related to photosystem II of higher plants (7,55,89). In addition, they are facultative phototrophs whose photosynthesis genes can be fully induced during non-photosynthetic growth. Thus, essential photosynthesis genes can be mutated and the effects studied in living cells expressing the genes encoding the 2 Figure 1. A simplified schematic diagram of the Rps. viridis reaction center and cytochrome b/c\ complex. Electron flow between cofactors in the reaction center is shown by arrows, with transfer times indicated. Dashed arrows outside the reaction center show movement of protons, electrons and cofactors. Abbreviations: Cyt, cytochrome subunit of the reaction center; L, M and H, positions of the respective reaction center subunits; Cyt. b/c-\, cytochrome b/c\ complex; Cyt c 2 , cytochrome c 2 ; P, special pair; A, accessory bacteriochlorophyll; BP, bacteriopheophytin; QA, primary quinone; QB, secondary quinone; Q B H 2 j ubiquinol (doubly-reduced and doubly protonated secondary quinone); Fe, non-haem iron. For a more detailed explanation see the text. P E R I P L A S M MEMBRANE C Y T O P L A S M LIGHT 4 photosynthetic apparatus. Major advances in understanding photosynthetic energy transduction have occurred in the past two decades, largely through experiments involving purple non-sulfur bacteria. Three species in particular - Rhodobacter capsulatus, Rhodobacter sphaeroides, and Rhodopseudomonas viridis have been extensively characterized. The first purified photosynthetic RCs were obtained from R. sphaeroides by Reed and Clayton in 1968 (66). Within two years, isolated RCs from a number of other purple non-sulfur bacteria were available (40,81). The peptides and associated cofactor molecules were characterized. In most cases the complexes were found to be comprised of three peptide subunits, designated L (light), M (medium), and H (heavy) based on their apparent molecular weights according to S D S - P A G E . The RC from Rps. viridis was found to contain a fourth peptide subunit, a cytochrome, and in some species (e.g. Rhodopseudomonas gelatinosa) the RCs were reported to contain only L and M subunits (40). The RC-associated cofactors were generally found to be 4 molecules of bacteriochlorophyll a (Bchl a), two molecules of bacteriopheophytin a, one carotenoid molecule, two molecules of ubiquinone and one non-haem iron. Rps. viridis again proved exceptional, containing bacteriochlorophyll b (Bchl b) rather than a, and one menaquinone rather than two ubiquinones. It has since been determined that the Rps. viridis RC in fact contains a loosely bound ubiquinone (as well as the more tightly bound menaquinone). Much information about RC function was obtained through various biophysical and biochemical studies on intact complexes. By the middle of the 1970s it had been established that the primary electron donor is a Bchl dimer (31,48,59,60), that a 5 bacteriopheophytin molecule is transiently reduced (31,75,84) and that the two quinones act in series as primary and secondary acceptors (33,48,61). It was recognized, however, that a full understanding of the mechanism of energy transduction within the RC would require a detailed knowledge of the spatial arrangement of all of the components. In a landmark achievement in 1982, Harmut Michel and co-workers succeeded in producing large, well ordered photochemically active crystals of RCs from Rps. viridis (24,57). Within two years, good quality crystals of R. sphaeroides RCs had also been obtained (3). Recently, high resolution X-ray analysis of these crystals (to 2.3 and 2.8 A for Rps. viridis and R. sphaeroides respectively) has permitted detailed three dimensional maps of these RCs to be constructed. Thus the precise configuration of all of the peptide and cofactor components of RCs from two species is now known. In conjunction with X-ray crystallography, knowledge of the primary amino acid sequences of the peptide subunits was required for an accurate determination of the three dimensional structures of the RCs. The poor solubility of RCs had made it difficult to determine the amino acid sequences by traditional methods. However in 1983, a year after the crystals were obtained from Rps. viridis, the first molecular cloning of photosynthesis genes from any organism was reported for the species Rhodobacter capsulatus (80). Subsequent ly , molecular genetic techniques were successfully applied to other species, including R. sphaeroides (89,90), and (to a lesser extent) Rps. viridis (54,56,87). The genes encoding the peptide subunits of the RCs from all three species were sequenced, and knowledge of the deduced amino acid sequences assisted in construction of the three dimensional map. 6 In addition to sequence information, molecular genetic approaches have yielded insights into fundamental questions of function and regulation of a wide variety of photosynthesis genes (73). The biogenesis of the photosynthetic apparatus is highly complex, requiring the co-ordinate regulation of approximately 30 known genes in R. capsulatus, including those encoding cytochrome C2, the cytochrome b/c-j complex components, the RC and LH structural peptides, and the enzymes involved in Bchl and carotenoid biosynthesis. Induction of the photosynthetic apparatus occurs in response to reduced oxygen concentration (18), with some components additionally regulated by light intensity (38,97). Regulation has been found to occur transcriptionally (44), post-transcriptionally (1,11,97), and post-translationally (25,26,87,88). The synthesis of the pigment and peptide components is coupled, such that free Bchl does not accumulate in mutant strains deficient in pigment-binding proteins, and conversely, strains deficient in Bchl biosynthesis do not accumulate pigment-binding peptides (45). The first step of photosynthesis, the absorption of incident light, generally does not take place in the RC itself, but in a second category of pigment-peptide complexes termed antenna or light harvesting (LH) complexes. Light energy absorbed by pigment molecules in these complexes migrates to the RC by delocalized exciton or inductive resonance transfer (70). All photosynthetic organisms contain LH complexes in vast molar excess over RCs. These serve not only to increase the number of pigment molecules available for photon absorption (on average, ~95% of total photosynthetic pigments are associated with LH complexes) (81), but also to increase the probability of capturing a photon of a wavelength not absorbed efficiently by the RC. Although the pigment molecules associated with LH and RCs may be identical, the wavelengths which they maximally absorb is strongly influenced by their microenvironment - i.e. by pigment-pigment and pigment-7 peptide interactions. Thus, LH complexes function to broaden the spectrum of absorbable light. Among purple non-sulfur bacteria, there are species with one, two or three different types of LH complexes (81). Each species has a characteristic absorption spectrum, determined by the particular combination of LH and RC complexes present. Within the membrane, the LH and RC complexes are organized in a specific fashion with respect to each other, permitting energy transfer to the RC to occur with excellent efficiency (70,81). Their arrangement in the photosynthetic membrane of ft capsulatus is represented schematically in Figure 2. The L and M subunits of the RC are each comprised of 5 trans-membrane a-helices (not shown), to which the Bchl a and other cofactor molecules are non-covalently bound. Although the function of the H subunit is not yet understood, it forms a cap on the cytoplasmic side of the membrane, which is anchored by a single membrane-spanning a-helix. The RC absorbs light maximally at 800 and 870 nm (apart from the Soret band 760 nm, shared with all Bchl a species). Each RC is surrounded by LH complexes termed LHI with a fixed stoichiometry of ~12:1 (29). LHI complexes are comprised of Bchl a and carotenoid molecules noncovalently bound to two peptides, designated a and p, which form single trans-membrane a-hel ices. LHI complexes absorb light maximally at approximately 875 nm. A second, more peripheral LH complex termed LHII is comprised of three peptides, designated a, p and y. As in LHI complexes, the LHII a and p peptides form single membrane-spanning a-helices to which pigment molecules are non-covalently bound. The function of the y peptide, which is largely hydrophilic, is not known. LHII complexes have two absorption maxima, at 800 and 850 nm. The number of LHII complexes in the membrane varies, being inversely proportional to 8 both light intensity and oxygen concentration. The R C , LHI and LHII complexes comprise the structural components of the photosynthetic apparatus that participate in absorption of light in R. capsulatus. Below the diagram of the complexes in Figure 2, the chromosomal organization of the genes encoding these components is schematically represented. The H subunit of the RC is encoded by the puhA gene (92), which is believed to form part of a larger transcriptional unit. Recently, an open reading frame (ORF1696) immediately upstream of puhA and transcribed in the same direction, has been shown to be necessary for assembly of LHI complexes (9). Approximately 40 kb away, and transcribed in the opposite orientation, is an operon designated put, which is comprised of 6 genes (92). The 5' and 3' most genes, designated pufQ and pufX respectively, have both been shown to be required for photosynthetic growth, but their exact functions are as yet unknown (1,32). The pufB and A genes encode the p and a subunits of the LHI complex respectively, whereas pufL and M encode the L and M subunits of the RC (92). The a, p\ and y subunits of the LHII complex are encoded by genes B A and E of the puc operon, which is located over 100 kb away on the chromosome (93). The pucC gene, which has high sequence identity with ORF1696, has been show to be required for assembly of LHII complexes (83). The ca. 40 kb interval separating the puh and put operons is occupied by operons that encode enzymes involved in carotenoid and bacteriochlorophyll biosynthesis (92). Figure 2 depicts the structural and genetic organization of the photosynthetic apparatus of R. capsulatus, but the same figure, unchanged, would serve to describe the photosynthetic apparatus of R. sphaeroides. The photosynthetic apparatus of Rps. viridis differs at the gross 9 Figure 2. Schematic diagram showing the organization of reaction center and light harvesting I and II complexes in the photosynthetic membrane of R. capsulatus and R. sphaeroides. The organization of pigment-binding peptide operons of R. capsulatus and R. sphaeroides, with structural genes shaded to match the specific subunits they encode, is depicted schematically below. A diagramatic representation of the Rps. viridis puf operon is enclosed in a box. Abbreviations: RC, reaction center; LHI, light harvesting I complex; LHII, light harvesting II complex; L, M and H, light, medium and heavy subunits of the reaction center respectively; a , p and y, the subunits of the light harvesting complexes; puc, operon encoding the subunits of LHII; puf, operon encoding the L and M subunits of the reaction center and the a and p subunits of light harvesting I complex; puhA, gene encoding the H subunit of the reaction center. Arrows show direction of transcription. The dotted line connecting ORF1696 and the puf operon indicates ~40 kb of DNA occupied by operons encoding genes involved in bacteriochlorophyll and carotenoid biosynthesis. Note that a y peptide has not been reported to co-purify with the a and p peptides of the R. sphaeorides LHII complex. 10 PERIPLASM LHI LHII CYTOPLASM 8 A ORFC ORFD E Put Q B A M put B A M C mmm Rps. viridis 11 structural level from that of R. capsulatus and R. sphaeroides in two principle ways. Firstly, as mentioned earlier, the Rps. viridis RC contains a cytochrome subunit in addition to L, M and H. Secondly, Rps. viridis contains only one LH complex, termed B1015, which is comprised of three trans-membrane peptides designated a, p and y (14). The organization of the existing structural subunits as well as the genes encoding them in Rps. viridis, however, is essentially the same as in the above two species. A schematic representation of the Rps. viridis pufoperon is enclosed in a box in Figure 2. The Rps. viridis pufB, A, L, and M genes correspond exactly to their counterparts in R. capsulatus and R. sphaeroides. Notable differences are the absence of puf genes Q and X ( 8 8 ) . The Rps. viridis pufC gene encodes the RC cytochrome subunit. The location of the gene encoding the B1015 y subunit is not known (88). As in R. capsulatus and R. sphaeroides, the H subunit of the Rps. viridis RG is encoded by a gene (also designated puhA) located in an operon separate and distant from the puf operon (not shown in Figure 2). It is clear from the above that the fundamental organization of the photosynthetic apparatus, both structurally and genetically, is similar in all three species. Perhaps even more striking is the similarity between the RC structures of Rps. viridis and R. sphaeroides as determined by X-ray crystallography. Figure 1 shows a schematic representation of the spatial arrangement of cofactors in the Rps. viridis RC. The cofactors are non-covalently bound to the L and M subunits, such that two nearly symmetrical branches are formed. The axis of symmetry runs perpendicular to the membrane, from the Bchl dimer (termed the "special pair", or "primary donor") near the periplasmic surface, to the non-haem iron molecule near the cytoplasmic surface. Upon excitation by absorbed light energy, the special pair rapidly transfers an electron to the 12 bacteriopheophytin molecule on the L branch. It is as yet unclear whether the intervening Bchl molecule is involved in electron transfer, thus it is termed the "voyeur" or "accessory" Bchl . The bacteriopheophytin is only transiently reduced, rapidly passing the electron to a tightly bound menaquinone molecule (Q A), termed the "primary acceptor", on the periplasmic side of the L branch. (Note that near the periplasm, helices of the L and M subunits "cross-over", such that the Q A binding-site is comprised of M subunit residues, while the Q B binding site on the M branch, is in fact comprised of L subunit residues. [Not shown in Figure 2]) From Q A the electron is transferred past the non-haem iron to the ubiquinone (Q B), termed the "secondary acceptor", on the M branch. The reduced Q B becomes protonated to Q B H , and the photo-oxidized special pair is re-reduced by the cytochrome subunit. After a second photo-oxidation event, Q B H is again reduced and subsequently protonated to form a hydroquinone, Q B H 2 . Q B H 2 dissociates from the RC and diffuses toward the periplasmic side of the membrane, being replaced at the Q B site by a quinone from the bulk cytoplasmic pool. After a complex series of reactions involving a cytochrome b/c\ complex, the electrons are returned to the cytochrome subunit (for subsequent donation to the oxidized special pair), and protons are translocated to the periplasm. This detailed structural information largely confirmed earlier spectroscopic findings regarding the identities of the primary reactants. The apparently symmetrical configuration of cofactors was unexpected however, as it raises the question of why electrons flow selectively along the L branch. Differences in the primary amino acid sequences of the L and M subunits clearly exert a profound influence on electron transfer. Thus, the detailed binding of the cofactors to the peptide backbone is different in the two branches (23). Recently, genetically engineered R C s have been 13 artificially symmetrized (67), and progress in understanding the structural basis for unidirectional electron flow is rapidly being made. In the RC of R. sphaeroides, the fundamental cofactor arrangement and path of electron flow is identical, with a periplasmic soluble cytochrome performing the function of the firmly bound cytochrome in Rps. viridis (4,5). There is good spectroscopic evidence that the cofactor arrangement is fundamentally the same in all purple non-sulfur bacteria so far analyzed (23). Furthermore, the primary amino acid sequences of RC peptide subunits from a number of purple non-sulfur bacterial species have been determined, and a high degree of similarity is evident. Table I shows the per cent identities in alignments of amino acid residues of R. capsulatus, R. sphaeroides and Rps. viridis RC ( and LHI) complexes (89,96). It is noteworthy that the L subunit, which forms the "active branch" of the RC, is the most conserved, while the H subunit, which apparently does not participate directly in the charge separation reaction, is the least conserved. Table I: Percent Identities from alignments of RC and LH Subunits R capsulatus/R. sphaeroides R. capsulatus/Rps. viridis RC L: 78 59 RCM: 76 50 RCH: 64 38 LHI a: 78 37 LHI p: 76 32 14 It is perhaps appropriate to mention at this point the extent to which RCs from purple non-sulfur bacteria are similar to those of photosystem II (PSII) in chloroplasts and cyanobacteria. The fundamental processes in both are entirely analogous: In PSII, a photo-oxidized chlorophyll special pair donates an electron first to a pheophytin, and then to two quinones acting in series. The secondary quinone is exchangeable with the bulk quinone pool (7). The cofactors in PSII are associated with peptide subunits, designated D1 and D2, which show weak but significant sequence similarities with the bacterial L and M subunits (55). Furthermore, RCs from purple non-sulfur bacteria and PSII are both sensitive to herbicides of the s-triazine type, which act by displacing Q B from its binding site. Mutations conferring herbicide resistance in purple non-sulfur and PSII RCs and have been shown to change homologous residues (16,62). These structural and functional similarities strongly suggest that purple non-sulfur bacterial RCs and PSII of green plants, algae and cyanobacteria evolved from a common ancestor. Given these striking similarities, the possibility existed that components of the photosynthetic apparatus from one organism might be capable of functional expression in other organisms. As a first step in addressing this possibility, I wished to test whether LH and RC genes from one species of purple non-sulfur bacteria could be functionally expressed in another species. Specifically, I wished to introduce heterologous genes into an R. capsulatus mutant host strain, deficient in all known pigment-binding peptide genes. It was hoped that characterization of inter-species hybrid complexes might shed light on structure-function relationships in the photosynthetic apparatus. Furthermore, if functional expression were shown to be possible, it would imply that R. capsulatus, which is relatively 15 easily manipulated genetically, might be used as a host for molecular-biological studies of a wide variety of photosynthesis genes from genetically recalcitrant species. I report here the results of two independent sets of experiments. In the first, I demonstrate that a plasmid-borne copy of the R. sphaeroides put operon is capable of genetically complementing a puf- puc- R. capsulatus host. Hybrid RCs, comprised of R. sphaeroides-encotieti L and M subunits, and an R. capsulatus-encoded H subunit, assemble in vivo and are capable of supporting photosynthetic growth. Cells containing hybrid RCs are compared to cells containing native RCs with respect to absorption spectroscopy, peptide subunit content, infra-red fluorescence and photosynthetic growth rates. In the second set of experiments, I show that the Rps. viridis puf operon does not functionally complement the same R. capsulatus mutant host strain. Furthermore, stable heterologous pigment-peptide complexes do not form even when a plasmid-borne copy of the Rps. viridis puhA gene is introduced along with the Rps. viridis puf operon. The implications in each case are discussed. The RC and LH genes from the vast majority of photosynthetic organisms are as yet uncharacterized. In an attached appendix, I briefly describe a technique which could be used to detect such uncharacterized genes in R. capsulatus. Theoretical considerations regarding construction of plasmid expression libraries and screening in R. capsulatus are included. 16 MATERIALS AND METHODS 1. Growth and manipulation of bacterial strains. The R. capsulatus strains used in this study were: wild type strain B10 (50), puf'puc' strain U43 (74), and puf strain ARC6 (1). Strain U43 was made by deletion of a 2,778 bp DNA fragment spanning part of pufQ and all of pufBALMX from the puc- R. capsulatus strain MW422 (74). The nature of the puc mutation in this strain is unknown. All strains were routinely grown in YPS medium (86). Oxygen-limited (low O 2 ) cultures were grown in Erlenmeyer flasks filled to 80% of their nominal capacity and shaken on a rotor and shaker at 150 RPM. Under these conditions the expression of photosynthesis genes is induced. Cultures to be used for transfer to anaerobic photoheterotrophic growth conditions were first grown to stationary phase under oxygen-limited conditions. Aliquots were diluted to 20 Klett units (about 6 X107 colony forming units/ml) in completely filled 20 ml screw cap tubes. Photosynthetic growth occurred in aquaria illuminated with light sources of varying intensities. Dark anaerobic cultures were grown in RCV medium (10) supplemented with 20 mM fructose, 30 mM dimethylsulphoxide (DMSO), and 5% pyruvate. DMSO served as an electron acceptor for anaerobic respiration (37). When induction of the R. capsulatus nifHDK promoter was required, cultures were grown anaerobically or under low O 2 conditions in ammonia-free RCV (RCV*) supplemented with 7 mM sodium-glutamate as a nitrogen source. For fluorescence measurements colonies of cells were 9 r o w n 17 aerobical ly on R C V medium supplemented with 1.5% agar. Photosynthetic growth on plates was obtained in anaerobic jars in the aquarium. Plasmid-containing strains were grown in media supplemented with antibiotics as follows: pJAJ9 and its derivatives, tetracycline (0.5 ng/ml); pNIF215 and its derivatives, kanamycin (10 ng /ml); pTB999, tetracycline (0.5 jig/ml); pRC77, kanamycin (10 ng/ml). Antibiotics were omitted from photosynthetically grown cultures. All cultures were grown at 34° C. Rhodobacter sphaeroides strain 2 .4 .1 . was grown photosynthetically in Y P S medium as described for R. capsulatus, with ca. 20 W / m 2 illumination provided by incandescent lamps. Rhodopseudomonas viridis strain DSM 133 was grown photosynthetically as above, in medium comprised of 50% RCV / 50% Y P S supplemented with 0.2 ug/ml para-amino benzoic acid. Escherichia coli strains C600r-m+ (13) and HB101 (pRK2013) (71) were used to deliver plasmids by triparental conjugation (27) to R. capsulatus, and were cultured at 37° C in LB medium (49) supplemented when appropriate with tetracycline or kanamycin (10 ug/ml). 2. Plasmid constructions. Flow charts for the relevant plasmid constructions are given in Figures 8 - 1 1 , and all plasmids are listed in Table II. Two broad host range plasmids were used throughout this study for the 18 expression of heterologous cloned genes in ft capsulatus. These were plasmid pJAJ9 and plasmid pNIF215 (Fig. 3). Plasmid pJAJ9 is an RK2 (71) derivative containing the R. capsulatus puf operon promoter upstream of a multiple cloning site (43). Plasmid pNIF215 is an RSF1010 derivative (72) containing the R. capsulatus nifHDK operon promoter upstream of a multiple cloning site (Cheryl Wellington, unpublished). Unless stated otherwise, all heterologous genes were inserted into these vectors in an orientation such that their expression was driven by the puf or nif promoters. Plasmid pCT1 carries a complete wild-type copy of the R. sphaeroides puf operon, and was constructed as follows. Plasmid pJW1 (90) contains the ft sphaeroides puf operon on a 4.5 kb Pst\ fragment. Digestion with Psfl released this fragment, which was then gel purified and ligated into the unique Pst\ site of pJAJ9 (see Fig. 4a). Plasmid pTB999 is a pRC11 (17) derivative carrying the entire ft capsulatus puf operon (Fig. 4b). Plasmid pJZ1, which carries a complete wild-type copy of the Rps. viridis puf operon, was constructed as follows (see Fig. 5a). Plasmid pKVS1 (gift of Joe Farchaus) which contains the Rps. viridis puf operon, was linearized by digestion with Hpa\. Oligodeoxynucleotide linkers containing an Sst\ recognition site (5 ' -CGAGCTCG-3 ' ) (synthesized by Tom Atkinson, Biochemistry Dept., U.B.C.) were ligated to the Hpa\ - generated blunt ends. The resultant plasmid was digested with Sst I, releasing a 6.7 kb fragment containing the Rps. viridis puf operon. The fragment was gel purified and ligated into the unique Sst\ site in pJAJ9. Plasmid pJZ2, which carries a complete wild-type copy of the 19 B Figure 3. A, A representation of expression vector pJAJ9. The open arrow around pufQB' indicates the direction of transcription initiated at the R. capsulatus puf promoter. The R. capsulatus pufQ gene, as well as the first 20 codons of the pufB gene are present in this vector. Tc indicates the approximate position of the tetracyline resistance determinant, transcribed in the direction shown by the arrow. All unique restriction sites following the puf promoter are indicated. B, A representation of expression vector pNIF215. The open arrow around NIFHDK indicates the direction of transcription initiated at the nifHDK promoter. kan and sm designate kanamycin and streptomycin resistance determinants respectively, with arrows showing direction of transcription. Unique restriction sites following the nifHDK promoter are shown. 20 Figure 4. A. A representation of plasmid pTB999. The solid black line represents the enitreR capsulatus puf operon, transcribed in the direction shown by the arrow. Tc designates the tetractycline resistance determinant. B. A representation of plasmid pCT1. The solid black line represents 4.5 kb of R sphaeroides DNA including the complete puf operon and flanking sequences, inserted into the unique Pst\ site in pJAJ9 and transcribed in the direction shown by the arrow. A detailed description of the construction can be found in section 2 of Materials and Methods. 21 Figure 5. Construction of A. plasmid pJZ1 and B. pJZ2. The thin lines represent vector DNA. The open arrows around pufQB' are as described in Figure 3B. The thick black lines represent Rps. viridis puf operon DNA, transcribed in the direction indicated by the arrow above. The thick shaded lines represent Rps. viridis DNA flanking the puf operon. Tc and amp designate tetracycline and ampicillin resistance determinants respectively, with arrows indicating direction of transcription. A detailed description of both constructions can be found in section 2 of Materials and Methods. 22 ligation 23 Rps. viridis puf operon plus an additional 2.1 kb of upstream Rps. viridis sequence, was constructed as follows (see Fig. 5b). Plasmid pKVS1 was digested with H / n d l l l , releasing an 8.8 kb fragment containing the Rps. viridis puf operon plus the upstream sequence. The fragment was gel purified and inserted into a unique H m d l l l site in plasmid pJAJ9, which was generated as follows. Plasmid pJAJ9 was linearized by digestion with S a m H I . Overhanging ends were filled in by DNA polymerase I (Klenow fragment) treatment to generate blunt ends. The termini were ligated to synthetic oligodeoxynucleotide linkers containing a H / n d l l l recognition site (5'-CCAAGCTTGG-3') (P-L Biochemicals, Inc., Milwaukee, Wis.). This generated a unique H / n d l l l site in pJAJ9, while duplicating the original (unique) B a m H I site. Plasmid pJZ4 carries the entire E. coli lac operon plus an additional ca. 0.9 kb upstream trp operon sequence, and was constructed as follows (see Fig. 6a). Plasmid pMC903, which contains the E. coli lac operon plus the upstream trp operon sequence, was digested with S a m H I and BglW, releasing a ca. 7.6 kb fragment containing the above described lac operon and trp sequences. This fragment was gel purified and inserted into the unique S a m H I site in pNIF215. Plasmid pJZ5 is identical to plasmid pJZ4 except that the 7.6 kb BamH\-Bgl\\ fragment has been inserted in the opposite orientation with respect to the nif promoter (see Fig. 6b). Plasmid pJZ6 carries a complete wild type copy of the Rps. viridis puhA gene, and was constructed as follows (see Fig. 7). Plasmid pDG4B ( gift of Joe Farchaus), which contains the Rps. viridis puhA gene, was digested with Xma\\\, and overhanging ends were filled in by treatment with Klenow fragment. The linearized 24 Figure 6. Construction of A. plasmid pJZ4 and B. pJZ5. The thin lines represent vector DNA. The thick lines represent E. coli lac operon DNA, inserted into the unique BarnHI site in pNIF215. The direction of transcription initiated at the nifHDK promoter is indicated by the open arrow. The orientation of the insert (5' - 3') is shown by the arrow above. A detailed description of the construction can be found in section 2 of Materials and Methods. 25 BomHI 26 Figure 7. Construction of pJZ6. Thin lines indicate vector DNA. Thick black lines indicate the Rps. viridis puhA gene DNA. Thick shaded lines indicate Rps. viridis flanking DNA. amp, kan and sm designate ampicillin, kanamycin and streptomycin resistance determinants respectively, with arrows indicating direction of transcription. A detailed description of the construction can be found in section 2 of Materials and Methods. 27 28 resultant plasmid was then cut with EcoRI , releasing the Rps. viridis puhA gene on a ca. 0.1 kb blunt - EcoRI fragment. The fragment was gel purified and inserted into plasmid pNIF215 which had been digested with Hin6\\\, blunt-ended by filling in with Klenow fragment, and subsequently digested with EcoRI. Plasmid pRC77 is a pRC11 derivative carrying a complete but modified R. capsulatus puf operon (17). The modification involves insert ion of a synthet ic o l igodeoxynuc leot ide (5'-G C C C A C C G G C A G C T G C C G G T G G G C - 3 ' ) immediately following the naturally occurring hairpin downstream of the pufBA genes. The insert functions as a strong transcriptional terminator, such that there is little or no transcription of the downstream pufLMX genes in this construct. DNA was purified from agarose gels by adsorption to glass beads, using the commercially available "Gene Clean" kit (BioRad). Digestion of DNA with restriction endonuclease enzymes, agarose gel electrophoresis, DNA ligations, transformation of E. coli and other recombinant DNA procedures were performed according to standard procedures (49). 4. RNA isolation. RNA was harvested as described (85) from cultures grown as follows. Rps. viridis strain DSM 133 was grown photosynthetically as described in section 1. R. capsulatus strains U43, U43(pJAJ9), U43(pJZ1), U43(pJZ1+pJZ6) and U43(pJZ1+pNIF215) were all grown under low O2 or anaerobic conditions in two different media: R C V V D M S O / f r u c t o s e supplemented with 10mM N H 4 , a n d RCWDMSO/fructose supplemented with 7 mM sodium-glutamate. U43(pJZ1+pJZ6) was also grown in YPS medium. Antibiotics were added as required. In all cases, cells were harvested at the mid to late log phase of growth. 29 TABLE II: Plasmids and bacterial strains used Plasmids Description (source or reference) pRK2013 pJAJ9 pNIF215 pTB999 Mobilizing plasmid (27) Expression vector utilizing puf promoter (contains pufQ gene and 1st 20 codons of pufB) (43; Fig. 3a) Expression vector utilizing nifHDK promoter(C. Wellington; Fig. 3b) pRC11 derivative missing the EcoR\-XhoW segment 5' of the puf operon promoter. Contains R. capsulatus puf operon (95; Fig. 4A) pCT1 p J A J 9 derivative containing sphaeroides puf operon (95; Fig. 4B) R. pJZ1 pJAJ9 derivative with 6.7 kb Hpa\-Sst\ fragment containing Rps. viridis puf operon (this work; Fig. 9A) pJZ2 pJAJ9 derivative with 8.8 kb H/ndlll fragment containing Rps. viridis puf operon plus 2.1 kb of upstream sequence (this work; Fig. 9B) pJZ4 pNIF215 derivative with 7.6 kb BamH\-BglW fragment containing E. coli lac operon (this work; Fig. 10A) 30 TABLE II (continued) Plasmids Description (source or reference) pJZ5 pNIF215 derivative with E. coli lac operon in opposite orientation (this work; Fig. 10) pJZ6 pNIF215 derivative with 0.1 kb XmaIII-EcoRI fragment containing Rps. viridis puhA gene (this work; Fig. 11) pRC77 pRC11 derivative containing R. capsulatus puf operon with transcriptional terminator following puf A (17) Bacterial strains Description (source or reference) E. coli C600 r-m+ Donor strain in conjugations (13) E. coli HB101 -Helper plasmid host strain in triparental conjugations (27) R. capsulatus B10 wild type (50) R. capsulatus U43 puf- puc- (74) R. capsulatus ARC6 puf- (1) R. sphaeroides 2.4.1. wild type (gift of Joanne Williams) Rps. viridis DSM 133 wild type (gift of Joe Farchaus) 31 5. RNA blotting and probing. Prior to loading on 1.4 % agarose/formaldehyde gels, a 34 u.g aliquot of Rps. viridis RNA and 17 \ig aliquots of RNA from all plasmid containing R. capsulatus strains were denatured in the presence of ethidium bromide (final cone. 0.04 |ig/ml) as described (68,69). Each denatured sample was then divided into two equal portions, and two gels were loaded identically - each with 17 u\g of Rps. viridis RNA and 8.5 ng of R. capsulatus RNA. After blotting, one gel was used to probe for the presence of Rps. viridis puf transcripts, the other for the presence of Rps. viridis puhA transcripts. Gels were run for 3 hours at 100V, rinsed for 5 minutes in dhteO and photographed through a U.V. transilluminator to determine the position of the rRNA with respect to the molecular weight markers (see Fig. 14A). Transfer to nylon membranes (ICN Bio-Trans) was accomplished by electroblotting at 30 V in 0.5X TBE buffer (49) for about 16 hours at 4 ° C. The membranes were then baked for 2 hours at 8 0 ° C, U.V. cross-linked for 3 minutes and again photographed through a U.V. transilluminator to determine quality and extent of transfer. Membranes were prehybridized with heat-denatured salmon sperm DNA at a final concentration of 500 ug/ml (95° C, 10') in 5 X SSC (49), 1% SDS, 50% formamide and 10 mM EDTA for a minimum of two hours at 42 ° C. Heat denatured DNA probes (a- 3 2P-labelled by the random oligonucleotide primer method [36]) were then added to the prehybridization mixture. When probing for Rps. viridis puf transcripts the probe DNA used was identical to the DNA sequence inserted into plasmid pJAJ9 to create pJZ2 (see Fig. 5B). When probing for Rps. viridis puhA transcripts, the probe used was identical to the DNA sequence inserted into plasmid pNIF215 to 32 create pJZ6 (see Fig. 7). Hybridization occurred overnight at 42° C. Membranes were then washed twice for 10 minutes at room temperature in 2 X S S C + 0.1% S D S , twice for 10 minutes at 50° C in 2 X S S C + 0.1 % SDS and twice for 5 minutes at 55° C in 0.2 X S S C + 0.1 % S D S . They were exposed to X-ray films for 2 to 13 days at -80° C with intensifier screens. 6. Fluorescence detection. The infra-red fluorescence of cells was evaluated as described (95). R capsulatus strain U43(pJAJ9) was routinely used as a negative control, and R capsulatus strain ARC6(pJAJ9) served as a positive control. 7. Spectrophotometric analyses. Absorption scans were obtained using a Hitachi U2000 spectrophotometer. All cells were grown under low O2 conditions to induce synthesis of the photosynthetic apparatus, and were harvested at the mid to late log phase of growth. In the R sphaeroides study, R capsulatus strains U43, U43(pJAJ9), U43(pCT1) and U43(pTB999) and were grown in Y P S medium. Intact cells (ca. 1.8 X 10 9 cells suspended in 22.5% BSA [77] in Y P S medium) were scanned. In the Rps. viridis study, R capsulatus s t ra ins U 4 3 , U43(pJAJ9), U43(pJZ1), U43(pJZ1+pJZ6) and U43(pJZ1 +pNIF215), were grown both in RCV* medium supplemented with 7mM^ sodium-g lu tama te /DMSO/ f r uc tose /py ruva te and in R C V * medium supplemented with 10mM NH4 + / D M S O / f r u c t o s e / p y r u v a t e . Membranes were prepared as follows. Forty ml of cells were 33 harvested, washed by resuspension in 10 ml RCV medium, and finally resuspended in 750 |il of RCV. Membranes were released by sonication with a Branson microtip probe (2 X 15" treatment at setting "2" on a Sonifier cell disrupter 350 sonicator, Branson Sonic Power Company). The samples were centrifuged for 60" to pellet cellular debris, and supernatant fluids were scanned. R. capsulatus strains U43, U43(pJAJ9) and U43(pJZ1) were also grown in YPS medium, and membranes, prepared as described above, were scanned (Fig. 12). The supernatant fluids were then recovered from the spectrophotometer cuvettes, diluted to 4 ml in RCV medium and centrifuged for 30 minutes at 100,000 rpm. The resultant supernatant fluids and pellets (resuspended in 750 uJ RCV) were both scanned. 8. Beta-galactosidase assays. The amount of (3-galactosidase activity present in R. capsulatus strains U43(pJZ4) and U43(pJZ5) was assayed essentially as described (58). Specifically, 20 - 30 ml of R. capsulatus cells were pelleted and resuspended in 1 ml of Z-buffer (58). The resuspended samples were transferred to Eppendorf tubes and sonicated as above. Various proportions of supernatant fluid (containing the p-galactosidase enzyme) and Z-buffer to give a total volume of 800 ul were added to a 1 ml cuvette. The substrate o-nitrophenyl-p-D-galactoside (ONPG; 200 jil of a 4 jig/jil solution in Z-buffer) was then added and the sample mixed well by repeated inversion. Increase in absorbancy at 420 nm was followed for 1 - 2 minutes. 34 RESULTS 1. Rhodobacter Sphaeroides Study. a. Absorption spectroscopy Figure 8 shows an overlay of typical absorption spectra of intact cells of R. capsulatus strains U43, U43(pJAJ9), U43(pTB999) and U43(pCT1). U43(pTB999), which contains a plasmid-borne copy of the R. capsulatus puf operon, has absorption maxima at 800 and 875 nm. These peaks result from RC (800 and 870 nm) and LHI complex (875 nm) absorbancy respectively. U43(pCT1), which contains plasmid pJAJ9 into which the R. sphaeroides puf operon has been inserted, has an absorption profile qualitatively very similar to that of U43(pTB999), whereas U43(pJAJ9), which contains the expression vector alone, does not. Thus, the presence of the R. sphaeroides puf genes in R. capsulatus U43 results in the formation of stable heterologous LH and hybrid RC complexes with superficial spectroscopic properties very similar to native R. capsulatus complexes. In general, peak amplitudes in absorption spectra are roughly proportional to quantities of absorbing pigment-peptide complexes. Measurements of peak to baseline ratios in these spectra showed about 10 - 30% greater levels of pigment-peptide complexes in U43(pTB999) cells relative to U43(pCT1). There are a number of possible reasons for this, including differences in mRNA accumulation, translation, protein stability or assembly barriers (see Rps. viridis and Discussion Sections).. Note that the absorption spectrum of U43(pJAJ9) shows a slight increase in absorbancy at approximately 800 nm not seen in U43 35 Figure 8. Absorption spectra of intact cells of R. capsulatus strain U43 containing various plasmids. Traces: a. cells containing pTB999 (R. capsulatus puf operon); b. cells with pCT1 (R. sphaeroides puf operon); c. cells with pJAJ9 (expression vector lacking; insert); d. cells with no plasmid. 36 alone. This is believed to be associated with the R. capsulatus pufQgene present on pJAJ9 (see the Rps. viridis and Discussion sect ions). b. SDS-PAGE analysis of chromatophore membrane proteins Confirmation of the presence of all the peptide components of the LHI and RC complexes was obtained by S D S - P A G E , the results of which are shown in Figure 9. The presence of protein bands with mobilities predicted for the subunits of the RC and LHI complexes was evident in samples from cells of U43 containing either the R. capsulatus or R. sphaeroides puf genes, whereas those bands were absent in samples of U43 host cells. It is noteworthy that the R. sphaeroides RC L and M subunits have slightly different mobilities compared with the equivalent R. capsulatus peptides. b. Fluorescence emission! The relative- efficiency of light energy transduction in the native R. capsulatus; and hybrid] photosynthetic complexes was evaluated! by comparison! of their intensities of fluorescence. As cart be seen in Figure 10, U43(pCT1) cells containing R sphaeroides LHI complexes and hybrid R C s , emit significantly greater levels of f luorescence than cells of U43(pTB999) (containing native R. capsulatus pigment-peptide complexes). Wild type R. sphaeroides strain. 2.4.1 emits marginally more fluorescence than, does wild, type R. capsulatus- strain B10/ (data not shown), however the significantly enhanced; fluorescence of U43(pCT1) is suggestive of a : dysfunction in- energy transduction specific to cells containing; the> hybrid RC. In general* cells that contain functionally impaired' LH or. R C complexes are unable to convert absorbed light to electrochemical energy as efficiently as wild-type cells, and emit some of the absorbed light as fluorescence. In R. capsulatus 37 1 2 3 29 18.4 6.2-Figure 9. SDS-po lyac ry lamide gel e lectrophoresis of chromatophore v e s i c l e s . L a n e s : 1. c h r o m a t o p h o r e s f rom st ra in U 4 3 ; 2, c h r o m a t o p h o r e s f r om U 4 3 c o n t a i n i n g p l a s m i d p C T 1 ; 3 , chromatophores from U43(pTB999) . The bands cor respond ing to reaction center subunits H, M and L are indicated on the right, as are the LHI comp lex bands (des ignated B875) . The pos i t ions of molecular mass markers (in kDa) are shown on the left. B875 i 38 Figure 10. Relative fluorescence of colonies of R. capsulatus ce l ls . A. cells of strain U43 containing plasmids pTB999 viewed with visible light; B. U43(pTB999) viewed through the infra-red filter; C. U43(pCT1) viewed with visible light; D. U43(pCT1) viewed through the infra-red filter. 39 U43(pCT1) , efficient transfer of light energy from the R. sphaeroides LHI complexes to the hybrid RC may in some way be impeded, and/or creation of a stable charge separation by the hybrid RC may be impaired (see Discussion). c. Photosynthetic Growth Rates The data given above indicated. that hybrid photosynthetic RCs , comprised of R. sphaeroides L and M subunits and an R. capsulatus H subunit, were capable of stable assembly in R. capsulatus U43 along with R. sphaeroides-encoded LHI complexes. I wished to determine whether this assemblage was functional enough to permit photosynthetic growth. Figure 11 shows the photosynthetic growth kinetics of U43(pCT1) and U43(pTB999) grown at various light intensities. At 20 W / m 2 (considered a moderate light intensity), no significant difference in the growth rates of the two strains was observed, while U43(pJAJ9) was shown to be incapable of photosynthetic growth (Fig. 11 A). When the light intensity was dropped to 4 W / m 2 (considered a low light; intensity),; U43(pCT1) grew somewhat more slowly than U43(pTB999) and, due to the absence of LHII complexes in, U43(pCT1) and U43(pTB999), they both grew more slowly than wild-type R. capsulatus; strain B10 (Fig.11B). At 2 W / m 2 , U4'3(pCT1'> was unable to grow photosynthetically, whereas U43(pTB999) began to grow after a lag of 163 hours (Fig. 11C). At 1 W / m 2 , neither U43(pTB999) nor U43(pCT1) were capable of photosynthetic growth (Fig.. 11 D). 2'.;. Rhodooseudomonas viridis S tudy . a. Determination of oxygen and ammonia effects on regulation of R. capsulatus nifHDK promoter activity in plasmid pNIF:215 The R. capsulatus nifHDK operon promoter was introduced 40 Figure 11. Comparison of photosynthetic growth of R. capsulatus strains containing either the hybrid or native reaction center. A. 20 W/m2; B. 4 W/m?; C. 2 W/m.2; D. 11 W/m.2. Klett Units Klett Units O O Klett Units M O O 42 into plasmid pJRD215 by Cheryl Wellington (as described inMaterials and Methods) to create an expression vector (pNIF215) capable of co-replicating along with the pre-existing fr capsulatus expression vector pJAJ9, which utilizes the R. capsulatus puf promoter for expression of cloned genes (see Fig. 3B). It has been shown (64) that transcription from the nifHDK promoter is totally repressed in the presence of ammonia, but that high levels of transcription are obtained in ammonia-free medium supplemented with (e.g.) glutamic acid as a nitrogen source. Furthermore, transcription from the R. capsulatus puf promoter is independent of the nitrogen source (64). Both the puf and nif promoters are known to be repressed by oxygen. Thus, it should be possible to obtain transcription of cloned genes from both expression vectors by growing cells either anaerobically or under reduced aeration in ammonia-free medium. In order to determine the optimal conditions for expression of heterologous genes introduced into pNIF215, I first created a lac fusion to the nifHDK promoter (pJZ4). R. capsulatus U43(pJZ4) was grown under a variety of conditions (various combinations of high/low/anaerobic oxygen and presence/absence of ammonia), and nifHDK promoter activity under each of these conditions was determined by monitoring p-galactosidase activities. Parallel experiments were performed on U43 cells containing plasmid pJZ5, a pNIF215 derivative containing the lac operon inserted in the opposite orientation to pJZ4, such that expression of the lac genes is not driven by the nifHDK promoter. In this way, any lac expression derived from the lac DNA fragment could be accounted for. Table III shows that, as expected, the highest levels of nifHDK promoter activity were obtained under anaerobic, ammonia-free conditions (1,854 nm ONPG/min/mg). However significant activity 43 (598 nm ONPG/min/mg) was also obtained under low O2, ammonia-free conditions. U43 cells containing the pNIF215 derivative pJZ6, which contains the Rps. viridis puhA gene, were initially grown in R C V * s u p p l e m e n t e d w i t h s o d i u m -glutamate/DMSO/fructose/pyruvate under both low O2 and anaerobic conditions. RNA blots of U43(pJZ1+pJZ6) probed for puhA expression showed strong transcription under both low O2 and anaerobic conditions (data not shown). Because growth under low O2 is considerably faster and more reproducible than non-photosynthetic anaerobic growth, U43(pJZ1+pJZ6) was routinely grown under low O2 conditions. b. Absorption spectroscopy (i) Rps. viridis puf operon expression in U43(pJZ"n Absorption spectra of intact cells of U43(pJZ1) were indistinguishable from those of U43(pJAJ9) (data not shown). To increase resolution of any small peaks that might be present, membranes prepared from U43, U43(pJAJ9), and U43(pJZ1) cells grown under low O2 conditions in YPS medium were scanned. Figure 12A shows a typical membrane absorption spectrum of U43. Although whole cell scans of U43 showed no absorption peaks ( Fig. 5), membrane scans showed peaks at -754, 803 and, at the limits of detection, 864 nm. The precise locations of these peaks were determined in 2nd order derivative scans (shown as insets in Fig. 12), in which valleys correspond to peaks in the original scan, and vice versa). The origin of these peaks is not known. The 754 nm peak is probably due to free bacteriopheophytin a (a break-down product of Bchl a); the 803 nm and 864 nm peaks may be associated with some as yet uncharacterized pigment-peptide complex(es) such as the putative assembly peptide encoded by ORF1696 (see Discussion). Alternatively, the observed peaks may result from very low level "leaky" LHII expression in U43. As stated in 44 Table III: Specific activities of p - g a l a c t o s i d a s e in cultures grown under high O2, low O2 or anaerobic condi t ions in either ammonia-free or ammonia-supplemented media AMMONIA-FREE: HIGH 0 2 LOW 0 2 ANAEROBIC U43(pJZ4) 1 197 3 591 1,854 U43(pJZ5) 2 22 36 48 AMMONIA-SUPPLEMENTED 4: LOW 0 2 ANAEROBIC U43(pJZ4) 22 161 U43(pJZ5) 22 12 1 pJZ4 is the nifHDK-lac fusion in the correct orientation for lacZ expression 2 pJZ5 is the nifHDK-lac fusion in the incorrect orientation 3 Specific activities are expressed as nmoles ONPG cleaved per minute per mg protein 4 Final ammonia concentration = 10mM 45 Materials and Methods, the nature of the mutation in U43 resulting in the LHII" phenotype has not been determined (74). A typical absorption spectrum of U43(pJAJ9) cell membranes is shown in Figure 12B. The amplitude of all absorption peaks was increased compared to U43 cells, the 864 nm peak was consistently blue-shifted to 855 nm, and a small peak at ~688 nm appeared. The increased absorbancy in U43(pJAJ9), which was sufficient to be detectable in whole cell scans as a slight rise at ~800 nm (see Fig. 8d), is believed to be associated with the pufQ gene present on pJAJ9. The pufQ gene is known to be required for Bchl biosynthesis (see below and Discussion). In addition to pufQ, plasmid pJAJ9 contains a truncated pufB gene, encoding the first 20 amino acids of the LHI p peptide (see Fig. 4a). It is unlikely that the peaks observed in U43(pJAJ9) resulted from Bchl associated with a peptide encoded by the 5' end of pufB fused to the downstream lac sequence on pJAJ9 because histidine residue 21, part of the highly conserved Ala-X-X-X-His sequence believed to be involved in Bchla binding (96), is not encoded on pJAJ9 (see Discussion). A representative absorption spectrum of cell membranes from U43(pJZ1), which contains plasmid pJAJ9 into which the R. viridis puf operon has been inserted, is shown in Figure 12C. The amplitudes of both the 754 nm and 685 nm peaks were consistently reduced in U43(pJZ1) relative to U43(pJAJ9). The reduction in 754 nm absorbancy suggests that Bchl a might be bound to peptides more tightly in U43(pJZ1) cells, thus reducing the amount of the free Bchl a break-down product, bacteriopheophytin a. It is evident that the peaks observed in U43(pJZ1) do not reflect the stable formation of Rps. viridis LHI or RC complexes. However, the 46 Figure 12. Absorption spectra of cell membranes prepared from YPS-grown cells of R. capsulatus strain U43, with or without plasmids. A. U43; B. U43 containing pJAJ9 (expression vector lacking insert); G. U43 containing pJZ1 (Rps. viridis puf operon). Second derivatives are inset, with valleys in the second derivative scans corresponding to peaks in the first. 47 Wavelength (nm) 48 absorption profile ofU43(pJZ1) was consistently, though subtly, distinguishable from that of U43(pJAJ9), suggesting that Rps. viridis pu/-encoded peptides interacted in some way with Bchl a. (ii) Rps. viridis puhA gene expression in U43(pJZ1+pJZ6) C e l l membranes prepared from R C W s o d i u m g lu tamate /DMSO/ f ruc tose /pyruva te -grown cel ls of U43 , U43(pJAJ9), U43(pJZ1) and U43(pJZ1+pJZ6) were scanned. Figure 13 shows that the absorption profiles of U43, U43(pJAJ9) and U43(pJZ1) grown in this medium appear quite different from those grown in YPS (see Fig. 12). The media-dependency of peak location and amplitude in these strains is not understood, although it has been observed by other workers (Doug Youvan, personal communication). U43(pJZ1+pJZ6) was grown on supplemented RCV* in order to induce the nifHDK promoter driving expression of the Rps. viridis puhA gene. The other strains were grown on this medium in order to allow valid comparisons of absorption profiles to be made. The absorption spectra of U43(pJZ1) and U43(pJZ1+pJZ6) were found to be essentially indistinguishable from each other. (Figure 13C and D). Thus, although genes encoding all of the structural peptides of the Rps. viridis RC (i.e. the L, M, H and cytochrome subunits) as well as the a and (3 subunits of the LH complex were present in U43(pJZ1+pJZ6), no stable assembly of Rps. viridis-encoded complexes occurred. A number of possible explanations for this lack of functional expression were considered, some of which were readily amenable to experimental verification. Several experiments, described below, were thus undertaken. To test whether stable assembly of Rps. viridis-encoded pigment-binding peptides might occur in an R. capsulatus host 49 Figure 13. Absorption spectra of cell membranes prepared from RCV*/sodium-glutamate/DMSO/pyruvate/fructose-grown cells of ft capsulatus strain U43 containing various plasmids. A. U43 alone; B. U43(pJAJ9); C. U43(JZ1); D. U43(pJZ1+pJZ6). 50 Wavelength (nm) 51 strain synthesizing LHII complexes, plasmids pJZ1 and pJZ6 were introduced into the puf puc+ R. capsulatus strain ARC6. However, absorption spectroscopy indicated that heterologous pigment-peptide complexes were not formed in this strain (data not shown). At the time of this study, the DNA region upstream of the Rps. viridis pufB gene had not been characterized. The Rps. viridis puf operon was known to be comprised of five genes, as follows: B A L M C. The R. capsulatus puf operon is comprised of six genes in the following order: Q B A L M X. The exact functions of the pufQ and pufX gene products in R. capsulatus are as yet unknown. However, as discussed above, the pufQ gene product has been shown to be required for Bchl a biosynthesis and thus, indirectly, for pigment-peptide complex formation. There is evidence to suggest that it may be a "carrier protein", necessary at each step of the Bchl biosynthetic pathway, as well as for delivery of mature Bchl a to the pigment-binding peptides of the RC and LH complexes (1,8). The possibility existed that an Rps. viridis pufQ gene equivalent might be present upstream of the pufB gene on the Rps. viridis chromosome, and that its presence in R. capsulatus U43(pJZ1+pJZ6) might be required for delivery of Bchl a to the Rps. viridis- en coded pigment-binding peptides. Hence, plasmid pJZ2 was constructed (see Fig. 5B). As described in Materials and Methods, pJZ2 is a pJAJ9 derivative carrying the entire Rps. viridis puf operon plus an additional 2.1 kb of upstream DNA. Plasmid pJZ2 was conjugated along with pJZ6 into R. capsulatus strains U43 and ARC6 , but absorption spectra of intact cells of U43(pJZ2+pJZ6) and ARC6(pJZ2+pJZ6) showed that inclusion of this upstream region had no discernible effect- on absorption profiles (data not shown). Recently, the sequence the Rps. viridis pufB and A genes plus an 52 additional 958 base pairs of the upstream DNA region has been reported (88). No open reading frame corresponding to the R. capsulatus pufQ gene was found, although an ORF (designated ORF R), with the same transcriptional orientation as the puf operon was found. ORF R shows a high sequence similarity to the R. capsulatus bchA gene, which encodes a bacteriochlorophyll biosynthetic enzyme, and which lies immediately upstream of the R. capsulatus puf operon (92). c. RNA blot analysis. (i) Rps. viridis ouf operon expression in U43 Because absorption spectroscopy demonstrated lack of stable Rps. Wr/d/s-encoded complex formation, it was essential to determine if in fact the Rps. viridis puf operon and puhA genes were being transcribed from their respective expression vectors in the R. capsulatus U43 host. RNA from Rps. viridis DSM 133 and from U43 strains containing various combinations of plasmids pJZ1, pJZ6, pJAJ9 and pNIF215 was therefore size fractionated on agarose gels, electro-blotted onto nylon membranes and hybridized with [oc-3 2P]-labelled probes specific for either the Rps. viridis puf or the Rps. viridis puh operons (see Materials and Methods). The size of ribosomal RNA transcripts differs in Rps. viridis and R. capsulatus. Both species contain two major rRNA classes, but whereas Rps. viridis contains the standard 16s and 23s rRNA classes found in most prokaryotes, very little 23s rRNA is observed in preparations of R. capsulatus RNA. Instead, a 14s rRNA (in addition to a 16s rRNA) is observed. It is believed that cleavage of R. capsulatus 23s rRNA gives rise to a 16s rRNA (that co-migrates with the standard 16s rRNA) and the 14s rRNA (39,44). The location of the rRNA bands was determined by photographing the RNA gel prior to blotting. Figure 14A lane 6 shows rRNA bands at ~2.4 and 53 Figure 14. A. An ethidium bromide-stained agarose/formaldehyde gel of RNA from photosynthetically grown Rps. viridis (17 \ig, lane 6) and low 02-grown R. capsulatus strain U43 containing various plasmids (8.5 ng, all other lanes) grown either in YPS or in RCV* with (+) or without (-) ammonia; and B. Autoradiogram of a blot of the above gel, probed with [oc- 3 2P]-labelled Rps. viridis puf DNA. Lane 1, U43(pJAJ9) (+); lane 2, pJAJ9 (-); lane 3, pJZ1 (+); lane 4, pJZ1 (-); lane 5, pJZ1 (YPS-grown); lane 6, Rps. viridis; lane 7, pJZ1+pJZ6 (+); lane 8, pJZ1+pJZ6 (-); lane 9, pJZ1+pJZ6 (YPS-grown); lane 10, pJZ1+pNIF215 (+); lane 11, pJZ1+pNIF215 (-). The numbers on the right refer to the approximate size in kilobases of A. rRNA bands and B. hybridizing transcripts. 54 1 2 3 4 5 6 7 8 91011 2.4 1.4 1.1 1 2 3 4 5 6 7 8 9 10 11 55 1.4 kb in Rps. viridis (corresponding to 23s and 16s rRNAs respectively) in contrast to bands of ~1.4 and 1.1 kb in R. capsulatus (corresponding to 16s and 14s rRNAs respectively, all other lanes). This information was essential for subsequent interpretation of the autoradiograms, where the presence of large quantities of rRNA was found to block hybridization to co-migrating messages. The autoradiogram of a [a- 3 2P]-labelled Rps. viridis puf DNA-probed blot of this gel is shown in Figure 14b. Lanes 1 and 2 show that the R. viridis puf probe did not hybridize to mRNA present in U43 (pJAJ9) (host strain containing the expression vector without insert). However hybridization to transcripts of ~4.2, 3.5, 0.82 and 0.68 kb occurred in all U43 strains containing plasmid pJZ1 (expression vector into which the Rps. viridis puf operon had been inserted) (Fig. 14b, lanes 3-5 and 7-11). Interestingly, this transcript size pattern is similar to that observed in Rps. viridis itself, where transcripts of ~3.5 and ~0.68 kb occur (Fig. 14b lane 6). Recent RNA blot analysis and end-mapping experiments (88) have revealed four classes of puf operon transcripts in Rps. yiridis. It was hypothesized that a major transcript of 3.6 kb encoding the entire puf operon is processed to give rise to a more abundant 0.62 kb transcript encoding only the pufBA genes. Two minor mRNAs of 3.7 and 0.76 kb were postulated to arise from a second, upstream puf promoter. Processing of puf operon transcripts seems to occur in several species of purple non-sulfur bacteria, and was originally reported in R capsulatus (11) where major transcripts of 2.7 and 0.5 kb are found (see Discussion). The 3.5 and 0.68 kb transcripts in U43(pJZ1) almost certainly correspond to their counterparts in Rps. viridis. The -4.2 and 0.82 kb transcripts seen in U43(pJZ1) may be fusion mRNAs, derived from 56 ft capsulatus pufQB'v.Rps. viridis pufBALMC sequences. (ii) Rps. viridis puhA gene expression in U43 Unlike the Rps. viridis puf probe, the Rps. viridis puhA probe did hybridize weakly to transcripts present in the U43 host strain (Fig. 15, lanes 1-5, 7 and 9-11). ft capsulatus strain U43, although mutated at both the puf and puc loci, is wild type with respect to the puhA gene. Figure 15 indicates that the low level of homology between Rps. viridis and ft capsulatus puhA genes (38% at the amino acid level) was sufficient for weak hybridization, even under the stringent conditions employed. However U43(pJZ1+pJZ6) cells grown under conditions that induce the ft viridis puhA gene (low O2, - N H 4 ; Fig. 15 lane 8) contained transcripts that hybridized much more strongly to the Rps. viridis puhA probe than either this strain under non-inducing conditions (low O 2 , + N H 4 ; lane 7) or than U43 strains lacking the Rps. viridis puhA gene (lanes 1-5, 10 and 11). A transcript of ~1.0 kb is faintly detectable in YPS-grown U43(pJZ1+pJZ6) (lane 9). The ammonia concentration in YPS is evidently low enough to allow low-level transcription of the puhA gene in pJZ6. In both U43(pJZ1+pJZ6) (lanes 8 and 9) and Rps. viridis (lane 6) the major transcript size was ~1.0 kb. d. Absorption spectroscopy of U43(pRC77) and U43(pTB999). The above data show that both the Rps. viridis puf operon and puhA genes were transcribed in ft capsulatus strain U43(pJZ1+pJZ6) when grown under the appropriate conditions. Yet absorbance spectra showed that stable pigment-peptide complexes did not assemble in this strain. Because both LH and RC complex peaks were absent, I wished to investigate the possibility that, in general, LH and RC complexes may exhibit some form of inter-Figure 15. Autoradiogram of a blot of the gel shown in Figure 14A, probed with [ a - 3 2 P] - label led Rps. viridis puhA DNA. All designations are as for Figure 14B. 58 dependence with regard to assembly and/or stable insertion into photosynthetic membranes of R capsulatus. I therefore conjugated plasmid pRC77 into R capsulatus U43. As described in Materials and Methods, plasmid pRC77 contains a modified R capsulatus puf operon, with a strong transcriptional terminator inserted downstream of the pufBA genes. No detectable transcription of the pufLMX genes, which lie downstream of the inserted terminator, occurs in this construct (11). Introduction of plasmid pRC77 into R capsulatus U43 therefore provided a means to determine whether stable assembly of LHI complexes would occur in the absence of RCs , and if so at what levels compared to strains synthesizing both RC and LHI complexes. In R capsulatus, RCs have two near infra-red absorbance maxima, at 800 and 870 nm. The LHI complex absorbs maximally at 875 nm. Typical absorption spectra of intact cells of U43(pRC77), which contains the modified puf operon, and L)43(pTB999), which contains a wild-type R capsulatus puf operon, are shown in Figure 16\ Although the 875 nm peak characteristic of LHI complexes was present in both strains, measurements of peak to baseline ratios showed a - 5 0 % reduction in amplitude, in U43(pRC77) relative to U43(pTB999). In order to accurately quantitate differences in LHI complex levels between the two strains however, the RC absorbance in U43(pTB999) must be taken into consideration. The 870 hm R C . peak is "buried" under the 875 hm peak. There are -12 LH is complexes per RC in the photosynthetic membrane of R capsulatus (11), so that a maximum of 10% of the 875 nm peak in U43(ptB999) could be attributed to RC absorbance. If the peak amplitude at 875 nm in U43(pTB999) were reduced by 10%, it would still be 30% higher than in U43(pRC77). Thus, stable assembly of LHI complexes in the photosynthetic membrane of U43(pRC77) was reduced by at least r : ! r T 1 -400 600 800 1000 Wavelength (nm) Figure 16. Absorption spectra of intact cells of R. capsulatus strain U43 containing A, plasmid pRC77 and B, plasmid! pTB999. 60 30% relative to U43(pTB999). Reductions of 50 to 70% in LHI complex levels in R C - strains of both fr capsulatus and R. sphaeroides have recently been reported (32,46). DISCUSSION I have described two independent sets of experiments involving inter-species expression of pigment-binding peptide genes from two purple non-sulfur bacterial species, R. sphaeroides and Rps. viridis, in cells of a third species, R. capsulatus. Genes encoding the peptide subunits of the RC and LH complexes from all three species had been cloned, and were readily available for manipulation. The first set of experiments involved introduction of the R. sphaeroides puf operon, which encodes the L and M subunits of the RC as well as both peptide subunits of the LHI complex, into an R capsulatus mutant host deficient in all known structural components of the photosynthetic apparatus, except the H subunit of the RC and the: cytochrome b/C\ complex. The results showed that hybrid RCs , comprised! of R sphaeroides-encoded L and M! subunits and an R capsulatus-encoded H subunit, assemble in vivo a long with the ft sphaeroides-encoded LHI complex. Cel ls containing these hybrid complexes are capable of photosynthetic growth, but: are less efficient at photosynthetic energy transduction thani either of the wild type parental strains. In a subsequent set of experiments,: the? Rps. viridis puf. operon was introduced! into; the R capsulatus mutant host strain. In this; case no stable assembly of Rps. viridis-encoded LHI! complexes occurred1, nor did! Rps. viridis - R capsulatus hybrid RCs form'. Introduction of the Rps. viridis puhA gene, which: encodes the H subunit of the RC, along, with the Rps. viridis puf operon still did 61 not result in stable RC complex formation, although the possibility existed that native Rps. viridis RCs could have assembled, because genes encoding all of the Rps. viridis RC peptides were present in the R. capsulatus host. When this project was initially undertaken, a wide spectrum of results was theoretically possible. The results actually obtained fall at the extreme ends of this spectrum. Among the possibilities were, for example, the stable assembly of nonfunctional hybrid RCs along with LHI complexes (functional or not) or assembly of only LHI complexes (again, functional or not). The results of both the R. sphaeroides and Rps. viridis studies provide insights into requirements for functional expression of the structural components of the photosynthetic apparatus. They point to a distinction between requirements for complex assembly in contrast to requirements for assembled complex function. Both requirements were met in the R. sphaeroides study, although the hybrids RCs. were functionally impaired. Because the requirement for assembly was not met in the Rps, viridis study, it can not be determined whether assembled complexes would or would not have1 been functional-. The impairment of the R. sphaeroides - R. capsulatus hybrid: RC results from interactions between the R. sphaeroides L and Mi subunits with the R. capsulatus H subunit The absence of stable) heterologous complexes in the Rps. viridis study points to hitherto unknown assembly requirements. The two sets of experiments are discussed individually below. 1. Rhodobacter sphaeroides study. U43(pCT1) cells containing the hybrid RC were shown to 62 suffer some degree of impairment in photosynthetic energy transduction by two methods: 1) under low light conditions they grew more slowly than U43(pTB999) cells (their counterparts containing native R. capsulatus complexes), and 2) they emitted significantly more fluorescence than cells containing either native R. capsulatus or native R. sphaeroides complexes. In principle, there are two stages at which the impairment could occur. First, efficient energy transfer from LH to RC complexes requires that the two types of complexes be correctly oriented with respect to each other in the membrane. If the conformation of the hybrid RC caused a perturbation in the normal arrangement of the surrounding LHI complexes, transfer of excitation energy might be impaired. Second, within the RC efficient electron transfer requires the highly precise spatial arrangement of all participating components. Excitation of the special pair within the hybrid RC might not produce a stable charge separated state as efficiently as in a native RC. The phenotype observed with the hybrid RC may result from impairment at either or both stages. Preliminary characterization of purified hybrid RCs by Doug Youvan's laboratory indicated that photobleaching of the 850 nm special pair absorbancy band (which is indicative of charge separation - i.e. loss of the electron from the special pair) was incomplete. Complete bleaching was obtained upon addition of exogenous quinone, suggesting a problem in either QA or QB binding (or both). In order to determine if the Q A site was affected, the rate of the "back-reaction" (electron flow in the opposite direction,, back to the special pair) was measured. Figure 1 shows that excitation of the special pair results in extremely rapid electron transfer to the bacteriopheophytin (~3 X 63 10 - 1 2 seconds at room temperature ). The subsequent transfer step, from the bacteriopheophytin to Q A takes ~2 X 10' 1 0 seconds. The transfer time between Q A and Q B is considerably slower, requiring ~1 X 10"4 seconds. The back-reaction rates are 10"8, 10"1 and 1 seconds respectively. Thus, if an electron is successfully transferred to Q A , and then is impeded in reaching Q B , the back-reaction takes between 1 and 0.1 seconds. An electron impeded in reaching Q A re-reduces the special pair in less than 0.1 seconds. The latter rate was observed by Youvan etal. (personal communication), indicating that electron flow in the hybrid RC was disrupted between the bacteriopheophytin and Q A . Thus, the conformation of the Q A binding pocket, which is formed predominantly by residues of the R. sphaeroides-encoded M subunit, is apparently perturbed by the presence of the R. capsulatus-encoded H subunit. It is not known whether the effect is localised, or whether long-range non-specific conformational changes have occurred. The role of the H subunit in RC function is not yet fully understood. It has been postulated to act as a nucleus around which the L and M subunits aggregate (76), and cross linking experiments have indicated that the H peptide may interact with LHI complexes (30). Interestingly, the conformation of the Q B site has been shown to be influenced by the H subunit (20), but whether this site is abberant in the hybrid reaction center has not been established. It has recently been shown that Q B " is protonated prior to the second photo-oxidation cycle, after which it is reduced and protonated at a second site. Protons are believed to reach QB~ and Q B H _ from the cytoplasm via two different pathways. The pathways have not yet been traced back toward the cytoplasm, but they may well proceed through the H subunit (see Fig. 1). 64 The primary amino acid sequences of the R. capsulatus and R. sphaeroides H subunit peptides are 64% identical after introduction of gaps to maximize alignment (89). This means that the 254 amino acid R. capsulatus H subunit differs from the 260 amino acid R. sphaeroides H subunit at 101 residues. It would be interesting to substitute ft capsulatus H subunit residues in computer graphics representations of the R. sphaeroides RC to see if any of the differing residues might potentially alter the hybrid RC conformation, especially around the quinone binding sites. Hypotheses based on such an analysis could be tested by site-directed mutagenesis. 2. Rhodopseudomonas viridis study. RNA blot analyses showed that both the puf operon and puhA genes were transcribed when U43(pJZ1+pJZ6) was grown under inducing conditions. Because Shine-Dalgarno sequences are highly conserved among prokaryotes, translational blocks are unlikely (see below). Therefore, the lack of stable LH and RC complex formation in U43(pJZ1+pJZ6), in which the genes encoding all structural components of the Rps viridis reaction center as well as the a and p subunits of the B1015 complex were present, suggested that some fundamental requirement(s) for assembly of these complexes was not met in the R. capsulatus host strain. In wild type R. capsulatus cells, puf operon mRNA is processed to give rise to two major classes of transcripts, 2.7 and 0.5 kb in length (11). The 2.7 kb message encodes the pufB, A, L, M and X genes. The smaller transcript, which is nine times more abundant, encodes pufB and A genes only. The relative abundance of pufBA transcripts accounts for the ~ 12:1 stoichiometric relationship between LHI and RC complexes in the membrane. Recently, it has been shown that Rps. viridis puf operon mRNA is processed in an 65 analogous manner, giving rise to two major transcripts of 3.5 and 0.6 kb, with the smaller transcript being more abundant (88). In both species the comparatively stable 5' segment of the transcript ends in a large stem-loop structure, which has been shown in R. capsulatus to be necessary (but not sufficient) for stabilization of the upstream mRNA ( ). Figure 14 shows that Rps. viridis puf operon mRNA in R. capsulatus strains U43(pJZ1) and U43(pJZ1+pJZ6) was processed in a manner similar to that in both parental strains. Thus, not only were the Rps. viridis genes transcribed in the R. capsulatus host, but it appears that the puf operon transcripts were appropriately processed as well. Table IV shows putative Shine-Dalgarno sites on both Rps. viridis and R. capsulatus puf and puhA transcripts. The Rps. viridis sequences are compatible with R. capsulatus ribosome binding since they are very similar to the R. capsulatus sequences and complementary to the 3' terminus of the R capsulatus 16S rRNA (92). Thus, it is unlikely that translation of the Rps. viridis transcripts in U43(pJZ1+pJZ6) would not occur. I did not attempt to verify the presence of the Rps. viridis peptides in U43(pJZ1+pJZ6) because studies in R capsulatus have shown that in the absence of stable pigment-peptide complex formation, the peptide subunits are rapidly degraded, and thus not detectable on SDS-PAGE (16). Joe Farchaus' attempts to detect the Rps. viridis M subunit in U43(pJZ1) with Western blots probed with anti-M antibody were unsuccessful (personal communication). Membranes prepared from R capsulatus strain U43 grown in YPS medium show absorption peaks at ~754, 803 and, at the limits of detection, 864 nm. As mentioned in the Results Section, the origin 66 of these peaks is not known, although they may be associated with some as yet uncharacterized pigment-peptide complex(es) such as the putative assembly peptide encoded by ORF1696. ORF1696 contains four histidine residues. One is located four residues from an alanine (Ala-X-X-X-His) and two are located four residues from a glycine (Gly-X-X-X-His). The Ala-X-X-X-His sequence has been shown to be highly conserved in LH complex peptides of purple non-sulfur bacteria, including R. capsulatus, R. sphaeroides and Rps. viridis, and is believed to be involved in Bchl binding (15). It is postulated that the histidine residue provides a ligand for the central Mg2+ ion of Bchl, while the alanine side chain may be in van der Waals contact with the Bchl ring (15). It has been shown by site-directed mutagenesis that Ala can be functionally replaced with Gly (16). In order to determine whether Ala-X-X-X-His or Gly-X-X-X-His sequences occur frequently in integral membrane proteins not involved in Bchl a binding, the amino acid sequences of the PufX protein, the amino terminal membrane-spanning helices of the RCH protein and the LHII y peptide, as well as the cytochrome b/c\ complex cytochrome b protein of R. capsulatus were analyzed. No Ala-X-X-X-His or Gly-X-X-X-His sequences were found in the PufX, H or y peptides. However, 2 of 11 histidine residues in R. capsulatus cytochrome b protein occur 3 amino acids away from an alanine. Thus, although Ala-X-X-X-His sequences are not obligatory Bchl binding sites, they seem to occur more frequently in pigment-binding peptides, where they have been shown to be necessary for Bchl binding (15). ORF1696 clearly has the potential to bind Bchl, and could account for the low level absorbancy observed in U43. Introduction of expression vector pJAJ9 into U43 cells was shown to result in significantly increased absorbancy (see Fig. 8d 67 for whole cell scans and Fig. 12B for membrane scans). This increase in peak amplitude is believed to be associated with the pufQ gene on pJAJ9, which has been shown to be required for Bchl biosynthesis. Klug et al. (46) studied the effects of pufQ expression in RCV-grown U43, and reported no far-red absorbancy specific for pigment-peptide complexes in R. capsulatus. Peak amplitudes in RCV-grown cells (Fig. 13B) are considerably reduced relative to YPS-grown cells (Fig 12B). Absorption scans of membranes prepared from YPS-grown U43(pJZ1), which contains plasmid pJAJ9 into which the Rps. viridis puf operon has been introduced, were consistently distinguishable from equivalent scans of U43(pJAJ9), suggesting that the Rps. viridis encoded pigment-binding peptides interacted in some way with Bchl a (Fig. 13C), although clearly stable Rps. viridis-encoded complexes did not form. The above data suggest that the barrier to functional expression of the Rps. viridis LH and RC complexes in R. capsulatus U43 is at the level of assembly. It is important to note that neither B1015 complexes nor RCs are assembled. The simplest explanation is that some fundamental "assembly requirement(s)", common to both types of complexes is not met in R. capsulatus strain U43. However it is also possible that each complex has unique requirements, neither of which are met. Lastly, it may be that only one of the complexes has an assembly requirement that is not met, but that formation of the two complexes is inter-dependent, such that absence of one prohibits stable assembly of the other. The following structural features of the Rps. viridis 68 Table IV: Comparison of putative Shine-Dalgarno sites in R. capsulatus and Rps. viridis puf and puhA genes R. capsulatus Rps. viridis pufQ GGAAGG N 1 2 A T G pufB GGAGG N 5 ATG GAGGG N 7 ATG puf A AGGAG N 9 ATG GGAGG N 8 ATG pufL GGAG N 8 ATG GGAGG N 8 ATG pufM AGGAGG N 5 ATG GGAG N 1 5 ATG pufX AGGAG N 8 ATG pufC GGAG N 7 ATG puhA AGGAGG N 6 ATG GGAGG N 5 ATG 69 photosynthetic apparatus, which differ from both R. capsulatus and R sphaeroides, are the most obvious potential sources of assembly problems. The photosynthetic pigment in R capsulatus and R. sphaeroides is Bchl a, whereas Rps. viridis utilizes Bchl b\ The primary acceptor (Q A ) in R capsulatus and R sphaeroides is ubiquinone whereas Rps. viridis utilizes menaquinone; the RCs of R capsulatus and R sphaeroides are comprised of three subunits, whereas the Rps. viridis RC has an additional cytochrome subunit; the photosynthetic membranes of R capsulatus and R sphaeroides are comprised of vesicular invaginations in the inner membrane, whereas in Rps. viridis the photosynthetic membranes are lammellar. In addition, it has recently been determined that several R p s . viridis RC and LH peptide subunits are post-translationally modified (87,88). In principle, any of the above factors could be responsible for or contribute to the lack of functional expression of Rps. viridis photosynthetic complexes in R capsulatus. I will attempt to assess each individually, beginning with a discussion of factors which could affect formation of both LH and RC complexes. a. Rps. viridis LH and RC peptides may be unable to bind Bchl a. Figure 17 shows the molecular structures of Bchl a and b. They differ only at ring II, where Bchl b carries an exocyclic double bond on carbon 4. All available evidence suggests that substituents on ring II do not participate in binding to peptides of the RC and LH complexes. In the RC, the central M g 2 + ions of the special pair are liganded to histidine residues on the L and M subunits. Hydrogen bonding to amino acid side chains on the L and M subunits involves the acetyl groups at carbon 2 on ring I, the keto carbonyl group at carbon 9 on ring V and possibly the ester carbonyl group at carbon 10 on ring V (53). The M g 2 + ions of the accessory Bchls are also liganded to histidine residues, with apparently no hydrogen bonds to C D Figure 17. Molecular structures of A. bacteriochlorophyll a; B. bacteriochlorophyll b, with an arrow pointing to sole structural difference between the two molecules, the exocyclic double bond on carbon 4 of ring II. Functional groups on rings I, IV and V have been implicated in hydrogen bond formation to light harvesting and reaction center subunits; C. ubiquinone; D. menaquinone. 71 the surrounding amino acid residues (53). The bacteriopheophytins appear to be hydrogen bonded to residues on the L and M subunits via the carbonyl groups on ring V (53). A number of recent in vitro studies by Paul Loach and coworkers provide strong experimental evidence that functional groups on Bchl rings I and V are responsible for hydrogen bonding to the highly conserved Ala-X-X-X-His sequence occurring in both a and p subunits of LHI complexes in Rps viridis, R. capsulatus and R. sphaeroides as well as many other species (15). Various Bchl analogues were tested for their ability to bind the LHI a and p peptides from Rhodospirillum rubrum, and it was determined that alterations to the functional groups on rings I and V prevent binding (47). Furthermore, Parkes-Loach et al. (42,63), recently succeeded in obtaining stable R. rubrum LHI a/Rps. viridis LHI p h y b r i d complexes associated with Bchl a, and Loach has obtained both R. rubrum LHI-Bchl b complexes, as well as Rps. viridis LHI p peptide-Bchl a complexes (personal communication). Thus, although the bchl a-b difference remains a formal possibility for the barrier to stable assembly of Rps. viridis photosynthetic complexes in R. capsulatus, all available evidence suggests that it is unlikely. b. The pufQ gene has been shown to be required for Bchl biosynthesis in R. capsulatus, and its product is postulated to act as a "carrier protein" involved in delivery mature Bchl a to the pigment-binding peptides of the RC and LH complexes (1,8). If this latter step involves protein-protein interaction (e.g. Q/a and p or Q/L and M ), then the R. capsulatus-eucoded Q gene product may not be able to deliver Bchl a to the Rps. viridis pigment-binding peptides for steric reasons. The Rps. viridis puf operon does not encode a Q gene equivalent (88). However, since the Bchl a and Bchl b biosynthetic pathways are likely to be very similar, an Rps. 72 viridis Q gene equivalent may well be located elsewhere on the chromosome. It is possible that assembly of Rps. viridis complexes in R. capsulatus U43 would be possible if this putative Rps. viridis pufQ gene were present. If either of the above two possibilities is in fact correct, then the assembly of both LHI and RC complexes would be prohibited. I will now consider possible barriers to assembly of the RC only. c. As mentioned earlier, the Rps. viridis RC contains a tightly bound menaquinone in the Q A site, whereas both R. capsulatus and R. sphaeroides RCs contain ubiquinone (see Fig. 17 for molecular structures). Although it is theoretically possible that Rps. viridis RCs require menaquinone for stable assembly, there is strong experimental evidence to suggest that this is not the case. An extensive and systematic study of the effect of replacing the native ubiquinone in purified reaction centers from R. sphaeroides with a variety other quinones was undertaken by Gunner et al. (41). All quinones used (including menaquinone) were shown to fully reconstitute the Q A -dependent flash-activated electron transfer reactions in the isolated RC. That is, not only did the R C s remain intact, but they were functional with unnatural quinones in the Q A site. Although the reciprocal experiment has not been carried out with Rps. viridis reaction centers, Gunner's work suggests that ubiquinone and menaquinone would be functionally interchangeable at the Q A site of the Rps. viridis RC. d. The fourth subunit of the Rps. viridis RC, the cytochrome subunit, may well be necessary for its stable assembly. Although the cytochrome is encoded by the pufC gene present on plasmid pJZ1, it is post-translationally modified in ways that may be significant for assembly. 73 (i) The DNA sequence of the cytochrome subunit contains a typical bacterial signal peptide of 20 amino acids which is not present in the mature subunit (87). It is the only RC subunit to contain such a sequence. It is also the only RC subunit which does not have an intramembraneous peptide region. Both the L and M subunits are firmly integrated into the membrane with 5 trans-membrane cc-helices each, and the H subunit is anchored by a single membrane-spanning helix. In 1987 Weyer et al. (87) determined that the cytochrome subunit is firmly anchored to the membrane by the post-translational addition of two fatty acids, covalently bound to the amino terminus of the protein via S-glycerocysteine. This post-translational modification is undoubtedly highly specific, and there is no reason to believe that it would occur in R. capsulatus. The stability of the Rps. viridis RC may require anchoring of the cytochrome subunit in the membrane. (ii) Four haems, two high potential and two low potential, are post-translationally inserted into the cytochrome. In general, haem lyases are known to be highly specific. This step in the post-translational modification of the cytochrome would thus presumably not occur in R. capsulatus. An attempt to express the gene encoding the Desulfovibrio vulgaris cytochrome c 3 , which also contains four covalently bound haems, in R. capsulatus was unsuccessful (65). The lack of functional expression was postulated to be due to the absence of a specific haem insertion system in R. capsulatus. Absence of the haem groups from the Rps. viridis cytochrome would clearly impair or destroy the function of the RC, but whether it would interfere with its assembly is not clear. e. It is possible that stable assembly of Rps. viridis RCs requires assembly of B1015 complexes. I will turn now to possible explanations for absence of assembled B1015 complexes. There are several important considerations. 74 f. Purified B1015 complexes from Rps. viridis contain 3 peptide subunits, a , p and y, in a 1:1:1 stoichiometry. The a and p subunits are membrane-spanning peptides known to bind Bchl b and carotenoid molecules. They are encoded by the B and A genes of the puf operon, which are present in U43(pJZ1). The location of the y subunit gene is unknown. Unless it happens to be located within the unsequenced 2.1 kb region upstream of the puf operon in plasmid pJZ2, then it is not present in any of the constructs used in these experiments. The function of the y subunit has not been determined. It has an unusually high proportion of aromatic residues relative to the a and p peptides, and has been postulated to be involved in the formation of regular arrays of LH complexes within the photosynthetic membrane of Rps. viridis (14). The possibility that it is required for assembly and/or stabilization of B1015 complexes in R. capsulatus U43 cannot be ruled out, although Paul Loach has shown that it is not required for in vitro formation of Rps. viridis LHI complexes (personal communication). g. The recently sequenced Rps. viridis puf B and A genes have been shown to have carboxy terminal extensions of 13 and 10 amino acids for the a and p subunits respectively, which are not present in the polypeptides isolated from the photosynthetic membranes (88). The carboxy terminal extensions of the precursor proteins may be required for their correct insertion into the membrane. In situ proteolytic degradation of the a and p peptides in inside-out chromatophore vesicles from Rps. viridis, R. capsulatus and R. sphaeroides resulted in splitting off of parts of the amino terminal domains (78,79,96). Thus, although the LHI genes of the latter two species do not encode carboxy terminal extensions, the a and p peptides are oriented in all three species such that the N terminus protrudes into the cytoplasm. 75 The mechanism by which the extensions are cleaved in Rps. viridis is not known. Since R capsulatus LHI subunit peptides are not processed in this way, the requisite enzyme(s) may not be present, and stable assembly of unprocessed Rps. viridis a and p peptides may not be possible. h. There is strong evidence that an as yet uncharacterized open reading frame(s) around the puhA gene is required for assembly of LHI complexes in both R sphaeroides and R capsulatus. It has been shown (76) that deletion of a -675 bp DNA fragment extending from 140 bp upstream of the start site of the R sphaeroides puhA gene to bp 535 within the puhA gene resulted not only in the loss of photosynthetic competence (due to absence of the H subunit of the RC), but also in loss of LHI complexes. Complementation in trans with the puhA gene alone restored photosynthetic competence but not LHI complexes, whereas both LHI and RCs were restored to the puhA deletion mutant when complemented with the puhA gene plus flanking sequences. RNA blot analysis of the puhA deletion strain compared to wild type R sphaeroides strain 2.4.1 showed that puf operon transcript levels are identical in the two strains. Thus, the loss of LHI complexes in this puhA deletion strain is not associated with loss of pufB A transcripts. Bauer and co-workers have recently determined that insertion of a kan r cartridge into ORF1696, located immediately upstream of the puhA gene in R capsulatus, results in an LHI* phenotype (9, in press). It is interesting to note that the amino acid sequence of the putative ORF1696 peptide shows a high degree of similarity to the pucC gene product, which has been shown to be required for LHII complex assembly (83). These data strongly suggest that expression of a gene or genes flanking puhA in both R capsulatus and R sphaeroides is necessary for assembly of LHI peptide subunits. It is possible that 76 assembly of ftps, viridis B1015 complexes has a similar requirement, and that the R. capsulatus ORF1696 gene product cannot substitute for the Rps. viridis equivalent. This may be a barrier to stable assembly of B1015 complexes in ft capsulatus. i. It is possible that formation of B1015 complexes requires the stable assembly of RCs. My data from the U43(pRC77) study (see Fig. 16 and the results section), as well as work from other laboratories (32,46), show that LHI levels decrease by 30 - 70% in the absence of RCs. Although detectable levels of LHI complexes are present in these native studies, it is possible that such inter-dependence might be more pronounced in a heterologous system. Furthermore, all of the above mentioned studies were done in puhA+ strains. No studies on LHI assembly have been undertaken in a purely puhA- background, although cross linking studies in ft capsulatus have indicated that H may interact with LHI complexes (28). There is evidence to suggest that in ft sphaeroides the H subunit is inserted into the membrane first, and acts as a "nucleus" around which the L and M subunits and LHI complexes assemble (76). If the Rps. viridis H subunit is incapable of insertion into the ft capsulatus membrane, e.g. because of membrane structure differences or because of "competition" with the ft capsulatus H subunit present in U43, then the Rps. viridis LHI complexes may be unable to assemble. In summary, numerous factors may be responsible for or contribute to lack of functional expression of Rps. viridis-encoded photosynthetic complexes in ft capsulatus. Of those discussed above, the post-translational modifications to the RC cytochrome and B1015 peptide subunits, the absence of the B1015 y subunit, and the possible requirement for "assembly peptides" seem to be the most likely obstacles to stable complex formation. 77 CONCLUSIONS The biogenesis of the photosynthetic apparatus in purple non-sulfur bacteria is highly complex, involving transcriptional, post-transcriptional and post translational regulation. In the past 3 years, post-translational modifications significant for assembly of RC and LHI peptides have been discovered in Rps. viridis (87,88), and very recently an ORF(s) adjacent to the puhA gene in both R. capsulatus and R. sphaeroides has been shown to be necessary for formation of LHI complexes (9,76). My experiments have shown that R. sphaeroides-encoded RC and LHI complexes assemble in R. capsulatus U43, whereas Rps. viridis-encoded RC and LHI complexes do not. In retrospect, it is perhaps not surprising that the R. sphaeroides and Rps. viridis experimental results fell at extreme ends of the wide range of possible outcomes. Based on comparisons of primary amino acid sequence identities of LHI peptide subunits, Zuber (96) proposed the following relative phylogenetic relationships among those purple non-sulfur bacterial species characterized: R. sphaeroides - R. capsulatus - Rps. gelatinosa - R. rubrum - Rps. viridis. R. sphaeroides and R. capsulatus show the highest degree of similarity (78% and 76% for the a and (3 peptides respectively) while R. sphaeroides and Rps. viridis show the least (39% and 28% respectively). It is interesting to note that R. rubrum, which by Zuber's comparisons is most closely related to Rps. viridis, has recently been shown to also have carboxy terminal extensions of 13 and 9 amino acids on the LHI a and (3 peptides respectively (12). R. rubrum and Rps. viridis are the only two purple non-sulfur bacterial species in which such post-translational modification of 78 the LHI peptides has been reported. Further studies on inter-species expression of pigment-binding peptides may be very instructive, from the standpoint of understanding assembly requirements as well as hybrid complex function. Several experiments immediately come to mind. It would be very interesting to introduce the Rps. viridis puhA gene on plasmid pJZ6 into an R. capsulatus H- mutant host. Assembly constraints in such a host should be minimal, as the LHI, LHII and RC L and M subunits would be R. capsulatus-encoded. Thus, an opportunity to study the function of an Rps. viridis-R. capsulatus hybrid RC might be possible. When an R. capsulatus puf-puc-puhA- multiple deletion strain becomes available (work is in progress) it might be interesting to introduce plasmid-borne copies of both the R. sphaeroides puf operon (pCT1) and the R. sphaeroides puhA gene into this host. Our studies with U43(pCT1) indicate that there would be no barrier to assembly of native R sphaeroides complexes. If, however, the R. capsulatus-encoded ORF1696 gene product does not assemble the R sphaeroides LHI complexes correctly, the cells would be expected to emit more fluorescence than wild type R sphaeroides cells. Insights into the nature of the assembly process might be gained from such a study. It would be extremely interesting to introduce pigment-binding peptide genes from R gelatinosa into various R capsulatus mutant hosts. On the basis of amino acid sequence identity comparisons it seems that, except for R sphaeroides, R. gelatinosa is the species most closely related to R capsulatus (96). The likelihood that the R gelatinosa complexes would assemble in R capsulatus is therefore greater than, for example, for Rps. viridis. Furthermore, 79 the RC in R. gelatinosa is reported to be comprised of L and M subunits only (40). Studies in both R. capsulatus H- and H+ background hosts might provide insight into the role of the H subunit in RC function. The RC and LH complex peptides from the vast majority of photosynthetic organisms, both prokaryotic and eukaryotic, have not yet been characterized. The technique by which fluorescence emission levels from U43(pCT1) cells was evaluated could in principle be applied to screen for the assembly of any heterologous pigment-binding peptides in U43, nothwithstanding barriers such as those encountered with the Rps. viridis puf genes. In the attached appendix, theoretical considerations regarding construction of an expression library and conjugation into R. capsulatus are discussed. 80 REFERENCES 1. Adams, C. W., M. E. Forrest, S. N. Cohen and J . T. Beatty. 1989. Structural and functional analysis of transcriptional control of the Rhodobacter capsulatus puf operon. J . Bacteriol. 171: 473-482. 2. Agalidis, I., A. M. Nuijs and F. Reiss-Husson. 1987. Characterization of an LM unit purified by affinity chromatography from Rhodobacter sphaeroides reaction centers and interactions with the H subunit. Biochim. Biophys. Acta 890: 2422-250. 3. Allen, J . P. and G. Feher. 1984. Crystallization of the reaction center from Rhodopseudomonas sphaeroides: Preliminary characterization. Proc. Natl. Acad. Sci. USA 81: 4795-4799. 4. Allen, J . P., G. Feher, T. O. Yeates, H. Komiya, and D. C. Rees. 1987. Structure of the reaction center from Rhodobacter sphaeroides R-26: the cofactors. Proc. Natl. Acad. Sci. USA 84: 5730-5734. 5. Allen, J . P., G. Feher, T. O. Yeates, D. C. Rees, J . Deisenhofer, H. Michel and R. Huber. 1986. Structural homology of reaction centers from Rhodopseudomonas sphaeroides and Rhodopseudomonas viridis as determined by X-ray diffraction. Proc. Natl. Acad. Sci. USA. 83: 8589-8593. 6. Arkin, A. P., E. R. Goldman, S. J . Robles, C. A. Goddard, W. J. Coleman, M. M. Yang and D. C. Youvan. 1990. Applications of imaging spectroscopy in molecular biology II. Colony screening based on absorption spectra. Bio/Technology 8: .746-749. 81 7. Arntzen C. J . and H. B. Pakrasi. 1986. Photosystem II reaction center: polypeptide subunits and functional cofactors. In L. A. Staehelin and C. J . Arntzen (eds.) Photosynthesis III Photosynthetic membranes and light harvesting systems. Springer-Verlag, Berlin. Heidelberg. 8. Bauer, C. E. and B. L Marrs. 1988 Rhodobacter capsulatus puf operon encodes a regulatory protein (PufQ) for bacteriochlorophyll biosynthesis. Proc. Natl. Acad. Sci. USA 85: 7074-7078. 9. Bauer, C. E. personal communication 10. Beatty, J . T. and H. Gest. 1981. Generation of succinyl Co-A in photosynthetic bacteria. Arch. Microbiol. 129: 335-340. 11. Belasco, J . G., J . T. Beatty, C. W. Adams, A. Von Gabain and S. N. Cohen. 1985. Differential expression of photosynthesis genes in R. capsulatus results from segmental differences in stability within the polycistronic rxcA transcript. Cell 40: 171-181. 12. Berard, J . G. Belanger, P. Corriveau, and G. Gingras. 1986. Molecular cloning and sequence of the B880 holochrome gene from Rhodospirillum rubrum. J . Biol. Chem. 261: 82-87. 13. Bibb, M.J. and S.N. Cohen. 1982. Gene expression in Streptomyces: construction and application of promoter-probe plasmid vectors in Streptomyces lividans. Mol. Gen. Genetics 187: 265-..277. 14. Brunisholz, R. A., F. Jay, F. Suter, and H. Zuber. 1985. The light-harvesting polypeptides of Rhodopseudomonas viridis. Biol. Chem. Hoppe-Seyler. 366: 87-98. 82 15. Brunisholz, R. A. and H. Zuber. 1988. Primary structure analyses of bacterial antenna polypeptides: - correlation of aromatic amino acids with spectral properties - structural similarities with reaction center polypeptides. In (H. Scheer and S. Schneider, eds.), Photosynthetic light-harvesting systems. Organization and Function. Walter de Gruyter, New York. 16. Bylina, E. J . , S. J . Robles and D. C. Youvan. 1988. Directed mutations affecting the putative bacteriochlorophyll-binding sites in the light harvesting I antenna of Rhodobacter capsulatus. Israel J . of Chem. 28: 73-79. 17. Chen, C-Y. A., J . T. Beatty, S. N. Cohen and J . G. Belasco. 1988. An intercistronic stem-loop structure functions as an mRNA decay terminator necessary but insufficient for puf mRNA stability. Cell 52: 609-619. 18. Clark, W. G., E. Davidson and B. L. Marrs. 1984. Variation of levels of mRNA coding for antenna and reaction center polypeptides in Rhodopseudomonas capsulata in response to changes in oxygen concentration. J . Bacteriol. 157: 945-948. 19. Coleman, W. J . and D. C. Youvan. 1990. Spectroscopic analysis of genetically modified photosynthetic reaction centers. Annu. Rev. Biophy. Biophys. Chem. 19: 333-367. 20. Debus, R. J . , G. Feher, and M. Y. Okamura. 1985. LM complex of reaction centers from Rhodopseudomonas sphaeroides R-26: characterization and reconstitution with the H subunit. Biochemistry 24: 2488-2500. 21. Deisenhofer, J . and H. Michel. 1989. The photosynthetic 83 reaction centre from the purple bacterium Rhodopseudomonas viridis. EMBO Journal 8: 2149-2169. 22. Deisenhofer, J . and H. Michel. 1989. The photosynthetic reaction center from Rhodopseudomonas viridis. Biophys. J . 55: Abstra. 1a. 23. Deisenhofer, J . R. Huber, and H. Michel. 1989. The structure of the photochemical reaction center of Rhodopseudomonas viridis and its implications for function. In G . D. Fasman (ed.), Prediction of protein structure and the principles of protein conformation. Plenum Press, New York. 24. Deisenhofer, J;, O. Epp, K. Miki, R. Huber and H. Michel. 1985. Structure of the protein subunits in the photosynthetic reaction centre of Rhodopseudomonas viridis at 3A resolution. Nature 318: 618-624. 25. Dierstein, R. 1983. Biosynthesis of pigment-protein complex polypeptides in bacteriochlorophyll-less mutants of Rhodopseudomonas capsulata YS. FEBS Lett. 160: 281-286. 26. Dierstein, R. 1984. Synthesis of pigment-binding protein in toluene-treated Rhodopseudomonas capsulata and in cell-free systems. Eur. J . Biochem. 138: 509-518. 27. Ditta, G., T. Schmidhauser, E. Yakobsdon, P. Lu, X.-W. Liang, D. R. Flnlay, D. Guiney and D. R. Helinski. 1985. Plasmids related to the broad host range vector pRK290 useful for gene cloning and for monitoring gene expression. Plasmid 13: 149-153. 28. Drews, G. and J. Oelze. 1981. Organization and differentiation 84 of membranes of prototrophic bacteria. Adv. Microb. Physiol. 22: 1 -92. 29. Drews, G. 1985. Structure and functional organization of light-harvesting complexes and photochemical reaction centers in membranes of phototrophic bacteria. Microbiol. Rev. 49: 59-70. 30. Drews, G. J . Peters and R. Dierstein. 1983. Molecular organization and biosynthesis of pigment-protein complexes of Rhodopseudomonas capsulata. Ann. Microbiol. (Inst. Pasteur) 134B, 151-158. 31. Fajer, J . , D. C. Brune, M. S. Davis, A. Forman and L. D. Spaulding. 1975. Primary charge separation in bacterial photosynthesis: oxidized chlorophyll and reduced pheophytin. Proc. Natl. Acad. Sci. USA 72: 4956. 32. Farchaus, J . W., H. Gruenberg and D. Oesterhelt. 1990. Complementation of a reaction center-deficient Rhodobacter sphaeroides pufLMX deletion strain in trans with pufBALM does not restore the photosynthesis-positive phenotype. J . Bacteriol. 172: 977-985. 33. Feher, G., A. J . Hoff, R. A. Isaacson and L. C. Ackerson. 1975. Endor experiments on chlorophyll and bacteriochlorophyll in vitro and in the photosynthetic unit. Ann. NY Acad. Sci. 244: 239-259. 34. Feher, G. J . P. Allen, M. Y. Okamura and D. C. Rees. 1989. Structure and function of bacterial photosynthetic reaction centres. Nature 339: 111-116. 35. Feher, G., M. Y. Okamura and J . D. McElroy. 1972. Identification 85 of an electron acceptor in reaction centers of Rhodopseudomonas sphaeroides. Biochim. Biophys. Acta 267 : 222-226. 36. Feinberg, A. P. and B. Volgelstein. 1983. A technique for radiolabeling DNA restriction fragments to high specific activity. Anal. Biochem. 132 : 6-13. 37. Ferguson, S. J . , J . B. Jackson and A. G. McEwan. 1987. Anaerobic respiration in the Rhodospirillaceae: characterization of pathways and evaluation of roles in redox balancing during photosynthesis. FEMS Microbiol. Rev. 46 : 177-143. 38. Garcia, A., G . Venturoli, N. Gad'on, J . Fernandez-Velasco, B. A. Melandri and G. Drews. 1987. The adaptation of the electron transfer chain of Rhodopseudomonas capsulata to different light intensities. Biochim. Biophys. Acta 890: 335-345. 39. Gray, Ernest. 1978. Ribosomes and RNA metabolism. In R. Clayton and W. Sistrom (eds.), The photosynthetic bacteria. Plenum Press, New York. 40. Gringas, G. 1978. A comparitive review of photochemical reaction center preparations from photosynthetic bacteria. In R. Clayton and W. Sistrom (eds.), The photosynthetic bacteria. Plenum Press, New York. 41. Gunner, M. R., B. S. Braun, J. M. Bruce and P. L. Dutton. 1985. The characterization of the QA binding site of the reaction center of Rhodopseudomonas sphaeroides. In M. E. Michel-Beyerle (ed.), Antennas and reaction centers of photosynthetic bacteria. Springer-Verlag, Berlin Heidelberg. 42. Heller, B. A., P. S. Parkes-Loach, M. C. Chang and P. A. Loach. 86 1990. Comparison of the structural subunit, B820, of core light-harvesting complexes of photosynthetic bacteria. In M. Baltscheffsky (ed) Current Research in Photosynthesis Vol. II Kluwer Academic Publishers, the Netherlands. 43. Johnson, J . A., W. K. R. Wong, and J. T. Beatty. 1986. Expression of cellulase genes in Rhodobacter capsulatus by use of plasmid expression vectors. J . Bacteriol. 167: 604-610. 44. Klug, G., N. Kaufmann and G. Drews. 1985. The expression of genes encoding proteins of B800-850 antenna pigment complex and ribosomal RNA of Rhodopseudomonas capsulata. FEBS Lett. 177: 61-65. 45. Klug, G., R. Liebetanz, and G. Drews. 1986. The influence of bacteriochlorophyll biosynthesis on formation of pigment-binding proteins and assembly of pigment protein complexes in Rhodopseudomonas capsulata. Arch. Microbiol. 146: 284-291. 46. Klug, G. and S. N. Cohen. 1988. Pleiotropic effects of localized Rhodobacter capsulatus puf operon deletions on production of light-absorbing pigment-protein complexes. J . Bacteriol. 170: 5814-5821. 47. Loach, P. A., T. Michalski and P. S. Parkes-Loach. 1990. Probing the bacteriochlorophyll binding site requirements of the core light-harvesting complex of photosynthetic bacteria using bchl analogs. In M. Baltscheffsky (ed.), Current Research in Photosynthesis, Vol. II. Kluwer Academic Publishers, the Netherlands. 48. Loach, P. A. and R. L. Hall. 1972. The question of the primary 87 electron acceptor in bacterial photosynthesis. Proc. Natl. Acad. Sci . USA 6 9 : 786. 49. Maniatis, T., E. F. Fritsch and J . Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 50. Marrs.B.L. 1974. Genetic recombination in R. capsulata. Proc. Natl. Acad. Sci . USA. 71: 971-973. 51. McElroy, J . D., G. Feher, D. C. Mauzerall. 1969. On the nature of the free radical formed during the primary process of bacterial photosynthesis. Biochim. Biophys. Acta 172: 180-183. 52. Michel, H. and J . Deisenhofer. 1988. Structure and function of the photosynthetic reaction center from Rhodopseudomonas viridis. Pure and Appl. Chem. 7: 953-958. 53. Michel H., O. Epp and J . Deisenhofer. 1986. Pigment-protein interactions in the photosynthetic reaction centre from Rhodopseudomonas viridis. EMBO Journal 5: 2445-2451. 54. Michel, H., K. A. Weyer, H. Gruenberg and F. Lottspeich. 1985. The 'heavy' subunit of the photosynthetic reaction centre from Rhodopseudomonas viridis: isolation of the gene, nucleotide and amino acid sequence. EMBO Journal. 4: 1667-1672. 55. Michel, H., and J . Deisenhofer. 1988. Relevance of the photosynthetic reaction center from purple bacteria to the structure of photosystem II. Biochemistry 27: 1-7. 56. Michel, H. K. A. Weyer, H. Gruenberg, I. Dunger, D. Oesterhelt 88 and F. Lottspeich. 1986. The 'light' and 'medium' subunits of the photosynthetic reaction centre from Rhodopseudomonas viridis: isolation of the genes, nucleotide and amino acid sequence. EMBO Journal 5: 1149-1158. 57. Michel, H. J . Deisenhofer. 19 Three dimensional crystals of a membrane protein complex. The photosynthetic reaction centre from Rhodopseudomonas viridis. Mol. Biol. 158: 567-572. 58. Miller, J . H. 1972. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. 59. Norris, J . R., R. A. Uphaus and H. L. Crespi. 1971. Electron spin resonance of chlorophyll and the origin of signal I in photosynthesis. Proc. Natl. Acad. Sci. USA 68: 625-628. 60. Norris, J . R., H. Scheer and J. J . Katz. 1975. Models for antenna and reaction center chlorophylls. Ann. NY Acad. Sci. 244: 260-280. 61. Okamura, M. Y., R. A. Isaacson and G. Feher. 1975. Primary acceptor in bacterial photosynthesis: obligatory role of ubiquinone in photoactive reaction centers of Rhodopseudomonas sphaeroides. Proc. Natl. Acad. Sci. USA 72: 3491-3495. 62. Paddock, M. L , S. H. Rongey, E. C. Abresch, G. Feher and M. Y. Okamura. 1988. Reaction centers from three herbicide-resistant mutants of Rhodobacter sphaeroides 2.4.1: sequence analysis and preliminary characterization. Photosynthesis Research 17: 75-96. 63. Parkes-Loach, P. S., B. A. Heller, M. C. Chang, W. J . Bass, J . A. Chanatry and P. A. Loach. 1990. Reconstitution of the core light-harvesting complex of photosynthetic bacteria with selected 89 polypeptides. In M. Baltscheffsky (ed.) Current Research in Photosynthesis Vol. II. Kluwer Academic Publishers, the Netherlands. 64. Pollock, D., C. E. Bauer and P. A. Scolnik. 1988. Transcription of the Rhodobacter capsulatus nifHDK operon is modulated by the nitrogen source. Construction of plasmid expression vectors based on the nifHDK promoter. Gene 65: 269-275. 65. Pollock, W. B. R., P. J . Chemerika, M. E. Forrest, J. T. Beatty and G. Voordouw. 1989. Expression of the gene encoding cytochrome C3 from Desulfovibrio vulgaris (Hildenborough) in Escherichia coli: export and processing of the apoprotein. J . of Gen. Microbiology 135: 2319-2328. 66. Reed. D. W. and R. K. Clayton. 1968. Isolation of a reaction center fraction from Rhodopseudomonas sphaeroides. Biochim. Biophys. Acta 30: 471-475. 67. Robles, J . Breton and D. C. Youvan. 1990. Partial symmetrization of the photosynthetic reaction center. Science 248: 1402-1405. 68. Rosen, K. M. and L. Villa-Komaroff. 1990. An alternative method for the visualization of RNA in formaldehyde agarose gels. Focus 12: 23-24. 69. Rosen, K. M., E. D. Lamperti and L. Villa-Komaroff. 1990. Optimizing the Northern blot procedure. BioTechniques 8: 398-403. 70. Sauer, K. Photosynthetic light reactions - physical aspects. 1986. In L A. Staehelin and C. J . Arntzen (eds.), Photosynthesis III Photosynthetic membranes and light harvesting systems. Springer-90 Verlag, Berlin Heidelberg. 71. Schmidhauser, T. J . and D. K. Helinski. 1985. Regions of broad host range plasmid RK2 involved in replication and stable maintainance in nine species of gram-negative bacteria. J . Bacteriol. 164: 446-455. 72. Scholz, P., V. Haring, B. Wittmann-Liebold, K. Ashman, M. Bagdasarian, and E. Scherzinger. 1989. Complete nucleotide sequence and gene organization of the broad-host-range plasmid RSF1010. Gene 75: 271-288. 73. Scolnik, P. A. and B. L. Marrs. 1987. Genetic research with photosynthetic bacteria. Annu. Rev. Microbiol. 41: 703-726. 74. Scolnik, P. A., D. Zannoni and B. L .Marrs. 1980. Spectral and functional comparisons between the carotenoids of the two antenna complexes of Rhodopseudomonas capsulata. Biochim. Biophys. Acta 593: 230-240. U43 75. Shuvalov, V. A. and V. V. Klimov. 1976. The primary photoreaction in the complex cytochrome P-890 P-760 (bacteriopheophytin 760) of Chromatium minutissimum at low redox potentials. Biochim. Biophys. Acta 440: 587. 76. Sockett, R. E., T. J . Donohue, A. R. Varga and S. Kaplan. 1989. Control of photosynthetic membrane assembly in Rhodobacter sphaeroides mediated by puhA and flanking sequences. J . Bacteriol. 171: 436-446. 77. Sojka, G. A., H. H. Freeze and H. Gest. 1970. Quantitative estimation of bacteriochlorophyll in situ. Arch. Biochem. Biophys. 136: 5788-580. 91 78. Stark, W. F. Jay and K. Muehlethaler. 1986. Localisation of reaction centre and light harvesting complexes in the photosynthetic unit of Rhodopseudomonas viridis. Arch. Microbiol 146: 130-133 79. Tadros, M. H., D. Spormann and G. Drews. 1988. The localization of pigment-bionding polypeptides in membranes of Rhodopseudomonas viridis. FEMS Microb. Lett. 55: 243-248. 80. Taylor, D. P., S. N. Cohen, W. G. Clark and B. L. Marrs. 1983. Alignment of genetic and restriction maps of the photosynthesis region of the Rhodopseudomonas capsulata chromosome by a conjugation-mediated marker rescue technique. J . Bacteriol. 154: 580-590. 81. Thornber, J . P. 1986. Biochemical characterization and structure of pigment-proteins of photosynthetic organisms. In Photosynthesis III Photosynthetic membranes and light harvesting systems. Springer-Verlag, Berlin Heidelberg. 83. Tichy, H.V., B. Oberle, H. Stiehle, E. Schiltz, and G. Drews. 1989. Genes downstream from pucB and pucA are essential for formation of the B800-850 complex of Rhodobacter capsulatus. J . Bacteriol. 171: 4914-4922. 84. Tiede, D. M., R. C. Prince, G. H. Reed and P. L Dutton. 1976. EPR properties of the electron carrier intermediate between the reaction center bacteriochlorophylls and the primary acceptor in Chromatium vinosum. FEBS Lett. 65: 301 85. von Gabain, A., J. G. Belasco, J. L. Schottel, A. C. Y. Chang and S. N. Cohen. 1983. Decay of mRNA in Escherichia coli: investigation of 92 the fate of specific segments of transcripts. Proc. Natl. Acad. Sci. USA 80: 653-657. 86. Weaver, P. F., J. D. Wall and H. Gest. 1975. Characterization of Rhodopseudomonas capsulata. Arch. Microbiol. 105: 207-216. 87. Weyer, K. A., F. Lottspeich, H. Gruenberg, F. Lang, D. Oesterhelt and H. Michel. 1987. Amino acid sequence of the cytochrome subunit of the photosynthetic reaction centre from the purple bacterium Rhodopseudomonas viridis. EMBO Journal 6: 2197-2202. 88. Wiessner, C. I, Dunger, and H. Michel. 1990. Structure and transcription of the genes encoding the B1015 light-harvesting complex |3 and a subunits and the photosynthetic reaction center L, M, and Cytochrome c subunits from Rhodopseudomonas viridis. J. Bacteriol. 172: 2877-2887. 89. Williams, J. C, L. A. Steiner, and G. Feher. 1986. Primary structure of the reaction center from Rhodopseudomonas sphaeroides. Proteins 1: 312-325. 90. Williams, J. C, L. A. Steiner, G. Feher, and M. I. Simon. 1984. Primary structure of the L subunit of the reaction center from Rhodopseudomonas sphaeroides. Proc. Natl. Acad. Sci. USA 81: 7307-7307. 91. Yang, M. M. and D. C. Youvan. 1988. Applications of imaging spectroscopy in molecular biology: I. screening photosynthetic bacteria. Bio/Technology 6: 939-942. 92 Youvan, D. C, E. J. Bylina, M. Alberti, H. Begusch and J. E. Hearst. 1984. Nucleotide and deduced polypeptide sequences of the 93 photosynthetic reaction center, B870 antenna and flanking polypeptides from R. capsulata. Cell 37: 949-957. 93. Youvan.D.C. and S. Ismail. 1985. Light harvesting II (B800-850 complex) structural genes from Rhodopseudomonas capsulata. Proc. Natl. Acad. Sci. USA 82: 58-62. 94. Youvan, D. C , S. Ismail and E. J . Bylina. 1985. Chromosomal deletion and plasmid complementation of the photosynthetic reaction center and light-harvesting genes from Rhodopseudomonas capsulata. Gene 38: 19-30. 95. Zilsel, J . , T. G. Lilburn and J . T. Beatty. 1989. Formation of functional inter-species hybrid photosynthetic complexes in Rhodobacter capsulatus. FEBS Lett. 253: 247-252. 96. Zuber, H. Comparative biochemistry of light-harvesting systems. In Photosynthesis III Photosynthetic membranes and light harvesting systems. L. A. Staehelin and C. J . Arntzen (eds.), Springer-Verlag, Berlin Heidelberg. 97. Zucconi, A. P. and J . T. Beatty. 1988. Post-transcriptional regulation by light of the steady-state levels of mature B800-850 light-harvesting complexes in Rhodobacter capsulatus. J . Bacteriol. 170: 877-882. 94 APPENDIX R. capsulatus strain U43 cells that contain expression vector pJAJ9 emit very low levels of fluorescence. If pigment-binding peptide genes inserted into pJAJ9 are functionally expressed in U43, f luorescence emission levels increase significantly (as in U43[pCT1]; see section 1 of Results). Thus, the possibility exists for using fluorescence emission to screen pJAJ9-derivative expression libraries for expression of heterologous RC and LH complex genes in U43, even when photosynthetic growth is not possible. Several important factors to be considered when constructing expression libraries for this purpose are discussed below. Total genomic DNA from the organism of interest can be partially digested with a restriction endonuclease, and the resultant fragments ligated into a compatible site downstream of the puf promoter on pJAJ9. The collection of ligated plasmids would be transformed into E. coli, after which it would be conjugated into U43. In order to minimize the possibility that transcriptional terminators might be present between the coding sequences of interest and the pJAJ9 promoter, the fragment sizes should correspond roughly to the predicted size of the gene(s) of interest. The optimal fragment size could vary considerably, depending on the source of the heterologous DNA. The DNA fragment containing the gene of interest must be inserted into pJAJ9 in the correct orientation, and possibly in the correct reading frame. Because prokaryotic Shine-Dalgarno sites 95 are genera l l y highly c o n s e r v e d , funct ional e x p r e s s i o n of heterologous prokaryotic genes may not require translationally in-frame insertion. The probability (p) that a given recombinant vector will contain a fragment with the gene(s) of interest is given by the expression: (1) p _ average fragment size (in kb) size of genome (in kb) Assuming that the fragment can be inserted into the vector in two equally probable orientations, the probability (p*) that the gene of interest will be present in the correct orientation for transcription from the puf promoter is p/2, i.e.: (2) p* = average fragment size (in kb) size of genome (in kb) X 2 NB: if insertion in the correct reading frame is also required, then the probability is 1/3 the above, i.e. the denominator must be multiplied by 3. For the remainder of this discussion, probabilities will be calculated only considering the requirement for correct orientation, although translationally in-frame insertions could be required for expression of eukaryotic (nuclear c D N A copies of) genes in R. capsulatus. It is necessary to determine how many recombinant vectors (i.e. t ransformed E. coli colonies) must be obtained to ensure a high probability that at least one contains a fragment carrying the gene(s) of interest in the correct orientation. If p* represents the probability that a given recombinant vector contains the gene of interest in the correct orientation, then the probability (P) that at 96 least one copy of this vector will be present in a collection of N transformed colonies is: (3) P = 1 - (1-p*)w The derivation of this formula is perhaps most easily explained by analogy. For example; given a randomly mixed collection of coins comprised of one million each of pennies, nickles, dimes and quarters, how many coins must be picked to ensure a high probability that at least one dime will be picked in the heads-up (HU) orientation? The biological counterpart would be that genomic DNA from a culture comprised of 1 million cells is cut into 4 equal-sized fragments, designated "a" - "d". If the 4 million resultant fragments are present in a ligation mixture with linearized pJAJ9, how many recombinant vectors (transformed E. coli colonies) must be screened to have a high probability that at least one will contain fragment "a" in the correct orientation? The probability of picking a dime, 1/4, corresponds to "p" in (1). The probability that if a dime is picked it will be HU is 1/2, so that the probability of picking a HU dime (HUD) is 1/4 X 1/2 = 1/8. This corresponds to "p*" in (2). The probability of picking at least one HUD in N tries, is equal to one minus the probability of picking up no. HUDs (derivation to follow). Note that, because one is starting with four million coins and making relatively few picks, one can assume for the sake of simplicity, that removal of coins in the picking process has no significant effect on the composition of the collection. Thus, if one picks two coins, the probability that the first is a HUD is 1/8, and the probability that the second is a HUD is 1/8. (1 97 million minus 2 ~= 1 million, so that climes still comprise ~1/4 of the total coin types.) The probability that the first QL the second o_r both are HUDs (i.e. at least one is a HUD) can be calculated as follows: The probability that the first is HUD and the second is not is equal to 1/8X7/8 = 7/64. The independent probability that the second is a HUD and the first is not is 7/8X1/8 = 7/64. The independent probability that both are HUDs is 1/8X1/8 = 1/64. The probability that the first or the second or both are HUDs is equal to the sum of their individual probabilities 7/64 + 7/64+1/64 = 15/64. The probability that neither are HUDs is 7/8 X 7/8 = (7/8)2 = 49/64. Thus, 1 minus the probability that neither are HUDs (64/64 -49/64 = 15/64) is equal to the probability that at least one is a HUD (15/64). Stated generally, the probability of picking at least one (one or more) of an item is equal to one minus the probability of picking none. 98 Notice that the probability of n M picking a HUD on a try, 7/8, is equal to 1 minus the probability of picking a HUD (1 - 1/8), which in the general form = 1 - p*. Furthermore, in the above example two coins were picked. If three coins are picked, the probability of getting at least one HUD = 1 - (7/8)3; in four picks it equals 1 -(7/8)4 and in N picks it equals 1 - (7/8)N. Therefore, the probability P of picking at least one HUD in N tries is equal to 1 - {7/8)N = 1 - ( l -p*)^. The forgoing is the derivation of the formula: (3) P = 1 - (1-p*)w After substitution (by analogy) of DNA fragments (of a size such that they comprise a fraction p* of the total genome), and numbers of E. coli colonies (A/), for the entities used above, the number of colonies required to have a high probability of having at least one of a given gene in the correct orientation, can be determined as follows: (4) 1-P = (1-p*)w and (5) In (1-P) = N In (1-p*) so that (6) N = »n (1-P) In (1-p*) In applying this formula, one must decide a) a suitable size range 99 of fragments to use, and b) the desired value of P. The size range of fragments will vary depending on the size of the genome of the organism that is tested. In the purple non-sulfur bacteria that have been characterized, pigment-binding peptide genes are generally organized into operons, which range in size from ~1 - ~4 kb. To generate an expression library from an as yet uncharacterized purple non-sulfur bacterial species, fragment sizes ranging from 0.5 - 5 kb would probably be reasonable. The average fragment size would thus be 2.75 kb, assuming equal distribution over the range given above. If the average bacterial genome is -4.0 X 10 3 kb, then p = 2.75/(4.0 X 103), and p * = 2.75/(2 X 4.0 X 103) = 3.4 X 10-4. If one desires a 99% probability of obtaining at least one copy of a given fragment in the correct orientation then, substituting into the formula (6) above gives: N = In (0.01) In (1 - 3.4 X 10-4) = 13,544 colonies Thus, using an average DNA fragment size of 2.75 kb, 13,544 colonies transformed with a recombinant vector are required to ensure a 99% probability that at least one colony will contain a fragment with the gene of interest in the correct orientation for expression. There are two important points to be made. Firstly, 13,544' transformed colonies does not ensure a 99% probability of having a c o m p l e t e e x p r e s s i o n library of the organism. This is true notwithstanding the fact that one can choose to screen for any desired fragment with this collection of transformed colonies in a 100 given experiment, and have a 99% probability that the chosen fragment will be present in the correct orientation. If one wishes to create a complete expression library that can be frozen and used repeatedly for screening (or selecting) for any gene, then one requires a high probability that at least one of each of all fragments is present in the correct orientation in the vector. This is given by the formula: (7) P ( i 1 o f a l l ) = ([1 - {1-p*>]".)<1/p) In the above example, 34,970 cells would be required to ensure a 99% probability of having at least one copy of each of the fragments in the correct orientation. This formula can be readily derived from (3) above, as follows: Using the analogy of the four coins again, one now wishes to have a high probability of obtaining at least one of each of a heads up penny (HUP), a heads-up nickel (HUN), a heads-up dime (HUD) and a heads-up quarter (HUQ). The probability of picking at least one HUD in N tries has already been shown to be (8) P(>1 HUD)= 1 " (7/8)" The probability of picking at least one HUP is also 1 - (7/8)^ (see [4]), as is the probability of picking at least one HUN or of picking one HUQ. The probability of picking at least one of each, is the product of their individual probabilities, namely (9) P(>J [HUP,HUN,HUD,HUQ]) = (I1 " f 7' 8}]^) 4 Generally, 7/8 = 1 - p*, and 4 = 1/p. Thus, we have derived the 101 formula (7) P(>1 of all) = ([1-{1-P*}] A') 1 / P The second important point to be made concerns creation of expression libraries specifically for use in R capsulatus. Because a method for direct transformation of R capsulatus does not yet exist, the transformation must be done into E. coli cells, which can then be used as donors for conjugation into R capsulatus. Liquid cultures of donor, helper and recipient cells are mixed in equal proportions, spotted onto small sterile millipore filters placed on RCV agar plates, and left to incubate at 37° C for > 6 hours to allow plasmid transfer to occur. In order to recover all R capsulatus ex-conjugant cells, the filters are resuspended by exhaustive vortexing in selective medium (RCV supplemented with 0.5 n g / m l tetracycline), followed by overnight incubation to allow plasmid-containing cells (which are Tc resistant) to multiply. The O . D . 6 5 0 of this overnight culture is determined, and suitable serial dilutions are made such that spread-plates of cultures grown on RCV/Tc will have ~ 200-500 colonies per plate. These plates are then used for fluorescence screening. Therefore, one is again faced with a statistical question; namely "How many U43 ex-conjugants must be plated in order to ensure a 99% probability that at least one of each of the original 13,544 recombinant plasmid types is present?" This is given by the formula just derived in (8): (7) P(>1 of all) = 0 " [ 1 - p ] " ) 1 ^ Note that in this case p = p*, because orientation is irrelevant. Thus, using the number N obtained above: 102 0.99 = (1 - [ 1-1/13,544 3,544 In order to solve this for N, it is useful to express 0.99 as (1-.01). Thus, In (1-0.01) = 13,544 ln(1 - [ 13,543/13,544 ]") Since ln(1-x) ~= -x when x is very small, then -0.01 ~= 13,544 (-13.543/13,544)" and ln(0.01/13,544) = N ln( 13,543/13,544 ) so N = ln(0.01/13.544) ln(13,543/13,544) = 190,541 colonies. Thus, given the constraints assumed above, to ensure a 99% probability that at least one R. capsulatus ex-conjugant contains the desired recombinant plasmid, 190,541 R. capsulatus ex-conjugant colonies must be screened. Recently, the technology has been developed to allow rapid screening of up to 500 colonies per plate, so that about four hundred plates would be required to screen ~200,00 cells (6). If many more cells were required (e.g. if a cDNA expression library of nuclear transcripts from a eukaryotic organism was to be screened), it 103 might be possible to use fluorescence-activated cell sorting to separate cells with enhanced fluorescence, which could then be spread on plates and screened as above. 

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.831.1-0098553/manifest

Comment

Related Items