MICROBIOLOGICAL REVIEWS. Mar. 1988. p. 50-690146-0749/88/010050-20$0(2 .00/0Copyright g 1988, American Society for Microbiology

Molecular Genetics of Photosynthetic Membrane Biosynthesis inRhodobacter sphaeroides

PATRICIA J. KILEYt AND SAMUEL KAPLAN*Departmenit of Microbiology, University of Illiniois (it Urbana-Chainpaign, Ur-banati, Illiniois 61801

INTRODUCTION .......................................................................... 50

ICM STRUCTURE AND FUNCTION .......................................................................... 51

COMPOSITION OF ICM COMPLEXES .......................................................................... 52

GENETIC ORGANIZATION AND IDENTIFICATION OF'THE STRUCTURAL GENES FORICM COMPONENTS .......................................................................... 55

Chromosomal Mapping ........................................................................... 55

Site-Specific Mutagenesis .......................................................................... 56

Utility of the lac Operon .......................................................................... 56

Gene Expression .......................................................................... 56

Identification of Photosynthetic Genes .......................................................................... 57

REGULATION OF ICM SYNTHESIS .......................................................................... 57

Bchl Synthesis .......................................................................... 58

ICM Assembly in Synchronous Cell Populations .........................................................................59

Induction and Assembly of the 1CM .......................................................................... 60

REGULATION OF PHOTOSYNTHETIC GENE EXPRESSION ......................................................61

Expression of the puf Operon .......................................................................... 61

Expression of the puc Operon.......................................................................... 63

Posttranscriptional Control .......................................................................... 64

CONCLUSIONS .......................................................................... 64

ACKNOWLEDGMENTS .......................................................................... 65

ADDENDUM IN PROOF .......................................................................... 65

LITERATURE CITED .......................................................................... 65

INTRODUCTION

In the last 30 to 40 years, the study of photosyntheticmembranes from the anoxygenic and oxygenic photosyn-thetic procaryotes has provided a wealth of informationpertaining to their biosynthesis and structure-function rela-tionships of the various integral membrane, pigment-proteincomplexes. In particular, photosynthetic membranes de-rived from the gram-negative, purple nonsulfur, photohet-erotrophic bacterium Rhodopseiido,nonias sphaeroicles, re-

cently renamed Rhodohacter spihaer-oides (59), have beenextensively charatcterized with regard to primary photo-chemistry, redox chemistry, bioenergetics, and structure, as

well as in studies which pertain to the assembly of thephotosynthetic membrane and its components (for reviews,see references 32, 41, 48, 63-65, 87, and 98). TIhe aim of thisreview is to relate very recent advances in this field toprevious work, so that we may begin to conceptualize, at themolecular level, how numerous, seemingly diverse syntheticpathways interact and how these pathways are regulalted toultimately allow the orderly assembly of this biologicalmembrane system in response to a variety of environmentalconditions. We focus our discussion on the photosyntheticmembranes of R. spizaeroide,s since, in many respects, thisrepresents the most thoroughly studied system to date and,when appropriate, we also compare and contrast theseprocesses in other well-studied members of the Rhodlospiril-liaceae; several recent reviews discuss specific aspects of

* Corresponding author.t Present address: t)epartment of Biochemistry. University of

Wisconsin, Madison, WI 53706.

this topic in other purple nonsulfur bacteria in more detail(43, 44, 99a, 136).

R. sphlaerioides, as well as other representatives of thepurple nonsulfur photosynthetic bacteria, are capable ofgrowth by aerobic and anaerobic respiration, fermentation,and anoxygenic photosynthesis. In some members of thisgr-oup the ability to fix atmospheric nitrogen permits anintegrated study of these diverse metabolic activities oftenseparated in other biologic systems. When growing chemo-heterotrophically, R. sphi(ieroides has a typical gram-nega-tive cell envelope and growth is supported by aerobicrespiration. When oxygen is removed from such a culture, aseries of events is triggered which results in the differentia-tion of the cytoplasmic membrane (CM) through a process ofinvagination into specialized domains which comprise thephotosynthetic intracytoplasmic membrane system (ICM).'I he ICM is physically continuous with the CM but structur-ally and functionally distinct in that the ICM specificallycontains all of the membrane components required for thelight reactions ot photosynthesis. iTherefore, R. splaeroidlesprovides an excellent model system with which to study bothphotosynthesis and membrane development. An advantageto studying membrane biosynthesis in R. sphaeroides andother closely related facultative photoheterotrophic bacteriais the ability of these bacteria to synthesize photosyntheticmembranes in the absence of light and under conditions inwhich these membranes are otherwise gratuitous for cellgrowth. For example, the photosynthetic membrane is syn-thesized while cells are growing chemoheterotrophicallyunder low-oxygen partial pressures, using 0 as a terminalelectron acceptor, or under conditions of anaerobic respira-tion in the dark, using dimethyl sulfoxide (131), trimethyl-

5t)

Vol. 52. No. 1

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 51

amine-N-oxide (99), or other terminal electron acceptors.The mechanism by which reduced oxygen tensions or anae-

robiosis translates into a biological signal to trigger inductionof the ICM is unknown but may be mediated, as previouslysuggested, through a redox carrier(s) (29), repressor or

activator protein(s), intracellular levels of a small mole-cule(s), or some combination of these effectors. Regardlessof the signal(s), the gratuitous induction of ICM synthesisunder defined physiological conditions allows analysis ofICM formation in both wild-type strains and those contain-ing mutations in structural or regulatory loci which renderthem unable to grow photosynthetically. Moreover, in addi-tion to regulation of ICM synthesis by oxygen tension, themolecular composition and intracellular amount of the 1CMare regulated by incident light intensity. The amount of ICMper cell and the whole-cell specific bacteriochlorophyll a

(Bchl) content increase proportionally as a result of decreas-ing the incident light intensity used for growth (Table 1; see

Fig. 8). Regulation by cultural incident light intensities isalso characterized by the differential synthesis of individualpigment-protein complexes (see subsection, "Induction andAssembly of the ICM"). Therefore, the inducibility of ICMsynthesis by low-oxygen tensions has allowed study of thede novo synthesis of a functional biological membrane; byfurther varying conditions of incident light intensity, ques-

tions pertaining to how ICM synthesis and composition are

physiologically regulated can also be addressed. As de-scribed in this review, de novo synthesis of the ICM refers tothe morphologically distinguishable developmental process

of differentiation of the CM leading to what we term matureICM (see subsection, "Induction and Assembly of ICM").In contrast, derepression in ICM synthesis can also occur as

a result of the insertion of pigment-protein complexes intopreexisting mature ICM which reveals no apparent morpho-logical developmental distinction.

ICM STRUCTURE AND FUNCTIONWhen examined by electron microscopy, the ICM appears

as vesicular invaginations budding from the cytoplasmicmembrane (90). Upon cell disruption, the ICM fractures andcomminutes to form numerous 50- to 60-nm-diameter spher-ical vesicles which are inside-out relative to the CM (52).The ICM vesicles, termed chromatophores, are photochem-ically active and can be easily purified from other cellularcomponents (52). The ability to study the kinetics of light-induced electron transfer reactions in chromatophores has

been central to elucidating the mechanisms of both theprimary reactions of photosynthesis and cyclic photosyn-thetic electron transport in R. sphaeroides (31, 32). For thesereasons, the best-studied functional ICM components are

those involved in light energy capture and subsequent redoxreactions, while parallel studies have followed the synthesisof these components in the ICM. It is not our intent toreview the wealth of biophysical and spectroscopic studiespertaining to chromatophore function but rather to summa-

rize major findings that relate to ICM synthesis and struc-ture.The most abundant chromatophore protein complexes are

the Bchl- and carotenoid-containing light-harvesting (LH)complexes that have been designated B875 (formerly LHI)and B800-850 (formerly LHII), based on their near-Bchlinfrared absorption maxima (25). The composition, function,and organization of LH complexes in the photosyntheticmembranes from several photosynthetic bacteria have re-

cently been reviewed (141, 142). As their name implies, theLH complexes act as antenna to funnel photons to reactioncenter (RC) Bchl-protein complexes in which light energy isconverted to chemical energy by photo-induced oxidation-reduction reactions (98). The funneling of photons to the RCoccurs by a process of exciton transfer, rather than lightenergy emission and reabsorption (98). The arrangement ofthese Bchl-protein complexes in the ICM phospholipid bi-layer must be highly organized to achieve the high efficiencyof exciton transfer, since little light is emitted as fluores-cence from wild-type strains during active photosyntheticgrowth (82). In photosynthetic bacteria such as R. sphaer-oides, which contain both B875 and B800-B850 LH com-

plexes, the B875 complex appears to be an obligatoryintermediate in exciton energy transfer from the B800-850complex to the RC complex (76). Aggregates of B875 com-

plexes surround and possibly interconnect RC complexeswithin the ICM; these aggregates have been termed the fixedphotosynthetic units since the ratio of B875/RC complexes isinvariant at approximately 15:1 (1). Direct evidence for a

symmetrical arrangement of LH complexes surrounding an

RC has been demonstrated only in the photosynthetic bac-terium Rhodopseiudoinonas viridis by immunoelectron mi-croscopy combined with image enhancement (109). Extrap-olation of biophysical measurements suggest a similararrangement in other purple nonsulfur bacteria (44, 82). TheB800-850 complexes are peripherally associated around thefixed photosynthetic unit, and the amount of B800-850

TABLE 1. Effect of light intensity on ICM synthesis

Lht Specific Amt of ICM components in Relative amt of ICM polypeptides (densitometer units)" ICM

Light Generation Bch] (pg/mg crude membranes" vesicles/p.m

intens2t time (h) of cell B800- of cellinW/m-ity time (h) protein) B875" B800-850" RC-H` RC-H RC-M RC-L B875-cx 8500a B875-, membrane

3 10.8 9.1 13.7 18.7 2.2 176 106 105 1,751 2,057 1,348 21.5 + 4.210 3.0 4.6 7.4 9.3 1.4 187 115 112 1.681 1,687 1,202 11.10 + 2.5100 3.0 2.9 7.4 4.0 1.0 136 92 84 1,646 731 1,290 6.47 ± 1.9

"Crude membranes were prepared as described previously (40)." Determined from [3H]tyrosine-labeled chromatophores (67). Densitometer scans of X-ray films of sodium dodecyl sulfate-polyacrylamide gel electrophoresis

resolved ICM isolated from R. sphaeroides 2.4.1 grown at different light intensities. The densitometer units were normalized for the moles of tyrosine perpolypeptide.

' Cell membrane lengths (in centimeters) were measured on x 100,000 or x 150,000 enlargements. using a manual measuring wheel (for x 100,000, 1 cm = 0.1pm; for x 150,000, 1 cm = 0.067 ,m). The results were obtained from measurements of thin sections such as those shown in Fig. 8 (kindly provided by A. Varga)." Determined from the extinction coefficient EX78-820= 73 mM - cm-' for the B875 complex normalized for 2 mol of Bchl a per complex. Similarly, the amount

of B800-850 Bchl was calculated by the A9(,)-,49, using E = 96 mM cm (76), normalized for 3 mol of Bchl a (expressed as nanomoles of spectral complex permilligram of crude membrane protein).'Determined from a Western blot (immunoblot) analysis using anti-RC-H antibody. and quantitated by densitometer scans of X-ray films as described

previously (40).

VOL. 52, 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

52 KILEY AND KAPLAN

complex present in the chromatophore varies inversely withincident light intensity (1).The absorption of light energy by the RC ultimately results

in the photooxidation of a "special pair" of Bchl molecules(46, 85). Then, following a series of intermediate states, aquinone is reduced, diffuses from the QB binding site of theRC complex, and equilibrates with the membrane-solublequinone pool, thereby carrying an electron to the ubiquinolcytochrome c,:oxidoreductase (cyt b/c1) complex (31, 32).Cytochrome C2 (cyt c2) functions as a mobile periplasmicredox carrier in cyclic photosynthetic electron flow bytransferring an electron from the membrane-bound cyt b/clcomplex to reduce the photooxidized RC complex (94). Thekinetic parameters of this cyclic photosynthetic electronflow and the relative topology and distribution of individualredox centers within the complexes in the ICM bilayer arebest understood in R. sphateroides (31, 32), and an illustra-tion depicting these components is shown in Fig. 1.

COMPOSITION OF ICM COMPLEXESSeveral of the supramolecular complexes of the ICM have

been purified to homogeneity, using detergent solubilizationtechniques. The analysis of purified complexes has beencentral to determining the minimal subunits required for, aswell as the structural basis of, their functional activity. TheB800-850 complex has been purified following lauryl di-methylamine-N-oxide detergent solubilization of chromato-phore membranes (24). The minimal unit required for B800-850 spectral activity appears to consist of six molecules ofBchl, three molecules of carotenoid, and two each of twosmall hydrophobic polypeptides designated B800-850-a andB800-850-3 (R. Cogdell, personal communication, 24. 117).Unlike Rhodobacter capsiulatis, there is no 14,000-daltonpolypeptide (45) associated with the purified R. spIl(eroidesB800-850 complex. The primary amino acid sequence andcognate deoxyribonucleic acid (DNA) sequences of B800-850-ox and B800-850-r have been described previously, andthe sizes of these polypeptides have been determined to be5,599 and 5,448 daltons, respectively (116, 117). Previously,Cohen and Kaplan purified and characterized three low-molecular-weight polypeptides from photosynthetic mem-branes, designated 15A, 15B, and 15C, which were thought

to be components of the LH machinery due to their abun-dance in purified chromatophores (27, 28). The polypeptidedesignated 15A does not appear to be a component of any ofthe purified spectral complexes. In addition, the 15A poly-peptide cannot be detected immunochemically in a mutantdeficient in the B800-850 complex (67a). It is possible thatpolypeptide 15A has a functional role associated with assem-bly or synthesis of the B800-850 complex and may beanalogous to the 14-kilodalton polypeptide (45) that can beisolated with the R. capsulatus B800-850 complex. Analyseshave shown that antisera against 15B and 15C cross-reactwith the B800-850-3 polypeptide (M. Morneault, seniorthesis, University of Illinois. Urbana, 1984). Theiler et al.(116, 117) have shown from amino acid sequence analysisthat there is heterogeneity at the amino-terminal end of theR. spliare-oides B800-850-f3 polypeptide; some polypeptidescontain amino-terminal methionine, and the remainder had ablocked amino-terminal threonine. It is possible that thedifference in charge between 15B and 15C may relate to theheterogeneity at the N terminus of the B800-850-, polypep-tide observed by Theiler et al. (116, 117) and that this couldbe the result of posttranslational modification of the B800-850-4 polypeptide.The purified B875 complex contains the B875-4 and

B875-cl apoproteins in a 1:1 stoichiometry, per two mole-cules each of Bchl and carotenoids (10). Both B875-4 andB875-x apoproteins have been sequenced, and their molec-ular weights have been determined to be 5,457 and 6,809,respectively (115). A similar analysis of the LH complexesfrom several other photosynthetic bacteria has been per-formed (141, 142).

Secondary structure predictions of the (x and a subunitsfrom both LH complexes show overall similarities (Fig. 2).'IThe LH polypeptides contain a single conserved histidineresidue which may function in Bchl binding and a centralhydrophobic domain predicted to span the ICM bilayeronce. The conserved histidine residues presumed to be theaxial ligands to the central magnesium atom of Bchl arelocated within this hydrophobic domain. The B800-850-3polypeptides from both R. capsulatus and R. sphalaeroidescontain an additional histidine residue which is also locatedin the hydrophobic domain and may also bind Bchl

FIG. 1. Illustration of a portion of the ICM bilayer depicting the relative orientation of specific ICM components. DH, Dehydrogenase;NADH, reduced nicotinamide adenine dinucleotide; ADP. adenosine 5'-diphosphate: ATP. aidenosine 5'-triphosphate.

MICROBIC)I. REV.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 53

1 21

21

c

21 41

4±

*

41

1 214±

/8-subunit

FIG. 2. Putative structures for the four subunits of R. spli(aeroides LH complexes. The structures are shown as cylinder models, in whichthe cylinders of different diameter represent different types of secondary structure, as indicated in the inset (bottom). The predicted secondarystructure was determined by use of the programs MCF, a modified Chou-Fasman prediction program incorporating Rao-Argos parameters formembrane proteins, and AMPHI, a program to analyze sequences for information about secondary and tertiary structure. The program

includes an extensive set of routines to test for hydropathy, prediction profiles, and amphiphilicity in potential helices and sheets and forplotting profiles, cartoons, helical wheels, and cylinders. Probable membrane-spanning helices are marked by an asterisk. These were

assigned on the basis of the Rao-Argos membrane helix parameter. Sheets and turns outside the membrane are assigned mainly on the basisof the Chou-Fasman rules. The cartoons were produced by using the program A MPHI. The programs MCF and AMPHI are part of a sequence

analysis package available from the Biotechnology Center, University of Illinois at Urbana-Champaign. Courtesy of A. R. Crofts, Universityof Illinois, Urbana.

(141-143). Studies of the iodination and protease sensitivitiesof chromatophores of R. capsul/atus (92, 110) or Rhodospi-rill/um ritbrurn (13) show that the LH polypeptides are

oriented in the bilayer such that the amino terminus pro-trudes into the cytoplasm whereas the carboxy end is locatedin the periplasmic space. A model of the Rhodospirillitin)lhbrlrm B880 complex (142, 143) that demonstrates thisorientation is shown in Fig. 3. Although recent attempts tocrystallize LH complexes have been successful (120), thequality of the crystals has not yet been suitable for high-resolution X-ray diffraction studies. The three-dimensionalstructure of the LH complexes is required to preciselydetermine the location of Bchl molecules and the organiza-tion of the pigments and polypeptides within these com-

plexes.The RC complex from R. sphaer-oides contains three

polypeptides, RC-H, RC-M, and RC-L (based on theirrelative mobilities from sodium dodecyl sulfate-polyacryl-amide gel electrophoresis of 28,000, 24,000, and 21.000daltons, respectively), in a 1:1:1 stoichiometric ratio with 4molecules of Bchl, 2 molecules of bacteriopheophytin, 2molecules of ubiquinone, a nonheme iron, and a singlecarotenoid (48). Calculations of the subunit molecular

weights from the deduced primary sequence of the R.spliier-oides genes reveal that the RC-M (34,265 daltons) andRC-L (31,319 daltons) subunits are actually larger than theRC-H subunit (28,003 daltons). The estimated molecularweight for the whole R. spliaei-oides RC complex includingcofactors is 100,858 (122).The RC complexes from both Rlhodopseludornonas viridis

and R. splhaeroides have been crystallized. The structure ofthe RC from Rliodopseiudornona.s liridis has been deter-mined to a resolution of -0.25 nm (38, 78), whereas that ofR. spli(Ieroides has been determined to -0.28 nm (3, 4, 128).Although the crystallized RC complex from Rliodopseiudo-monas viridis also contains 1 mol of a membrane-bound c

type cytochrome (38, 78. 121), the major structural featuresof the RC from both bacteria have been conserved, and thestructure of the chromophores from the R. splihieoides RCis shown in Fig. 4. The RC-L and -M subunits are bothintegral membrane polypeptides which are in intimate asso-

ciation with the chromophores shown in Fig. 4. The thirdsubunit, RC-H, has a single. amino-terminal, membrane-spanning region, but the majority of the protein is hydro-philic and appears to protrude into the cytoplasm. Since it isknown that the RC-H polypeptide does not bind any cofac-

B800- B850

a- subunit

B875

/3-subunit

a - subunit

Residue numberi

She etCoir sHel1i x Turns

VOL. 52, 1988

j

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

54 KILEY AND KAPLAN

polar head groups _

K G

COOH

N~ (PN ~ COOHPERIPLASM IG G

FIG. 3. Structure model of the transmembrane-oriented LH polypeptides fronm purple photosynthetic bacteria (amino acid residues of thea- and 3-polypeptides from Rhodospirilllunti rulbrill). An (x-helical conformation of the hydrophobic domain within the hydrocarbon tail regionis assumed. The possible Bchl-binding site is at the His residues (one His residue in the x- and P3-polypeptide each, near the periplasmic sideof the membrane, exciton-coupled Bchl pair: single His residue in the 3-polypeptide monomeric Bchl). PK, CH, SA, and TR are sites ofpartial hydrolysis with proteinase K. chymotrypsin. Staplhlococcus aurciis proteinase. and trypsin at the N-terminal domains of the a- and0-polypeptides located at the cytoplasmic side. Reprinted from Zuber (143) with permission of Springer-Verlag.

tors and is not required for some aspects of RC function invitro (2), there has been much speculation as to its role. Forexample, it has been determined that substantial amounts ofRC-H can be found in aerobically grown cells where thereare no detectable RC-M or RC-L polypeptides (16, 40).Moreover, there is an approximately 50% excess of theRC-H polypeptide relative to RC-M and RC-L in the bulkphotosynthetic membrane (P. J. Kiley, Ph.D. thesis, Univer-sity of Illinois, Urbana-Champaign, 1987). Analysis of R.sphaeroides mutants deficient in the RC-H subunit and its

role in RC structure or photosynthetic membrane assemblyor both are discussed later. It might be inferred that theRC-H subunit is not necessary for RC function since somephotosynthetic bacteria (i.e., Rhodocvclhs gelatinosa) con-tain RC complexes with only two subunits (21). When DNAprobes from R. sphlaeroides specific to the RC-H, -M, and -Lgenes were used against Rhodocv8clhis gelatinosa genomicDNA. strong homologies to RC-L and RC-M were observedbut the RC-H probe detected no signal, even at low strin-gencies of hybridization (W. Havelka and S. Kaplan, unpub-

peripIasm

QA ° )tQB QA ° J{)QB_,__ --- - -A Fe _F_cytopl a smn

FIG. 4. Stereoplot of the cofactors of the RC from R. sphaeroidles at a resolution of t).33 nm. The twofold symmetry axis is alignedvertically in the plane of the paper. and the electron transfer proceeds preferentially along the A-branch. The inferred position of themembrane is indicated by dashed lines. Reprinted with permission of Allen et al. (4).

MICROBIOLI REV.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTlHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROII)ES 55

lished results). However, specific antibody to R. spIh(ar-oides RC-H showed a cross-reaction with the RC-L andRC-M polypeptides of Rhodocvclus gelatinosa but nounique signal for RC-H was detected in membrane fractionsfrom Rhodocvclus gelatinosa (J. Hoger and S. Kaplan,unpublished results). Therefore, it may be proposed that inR. splhaieroides, RC-H has functions that may be served inRhodocyclhs gelatinosa by (parts of) RC-L or RC-M or both.

Other protein complexes in the R. sphcieroides ICM thathave been characterized include the cyt b/cl complex whichconsists of equimolar amounts of cyt cl (33 kilodaltons), acyt b (40 kilodaltons) which contains two hemes, and aRieske iron-sulfur protein (24 kilodaltons) (55). It has beenestimated that chromatophores from cells grown at relativelyhigh light intensities contain approximately 0.5 mol of the cytb/cl complex per mol of flash-oxidizable RC (31). Thesoluble periplasmic cyt c2 is also present in the ratio of 0.5mol/mol of RC (31). Cyt c, consists of a single subunit of14,200 daltons with one covalently attached heme (77),although recent studies suggest the existence of multipleelectrophoretic forms of this protein in the R. sphaeroicdesperiplasm (T. J. Donohue. A. E. McEwan, S. van Doren,A. R. Crofts, and S. Kaplan, Biochemistry, in press).Previous studies have suggested the existence of multiple cyt(--type species in other species of purple nonsulfur bacteria(46, 49), although the physiological significance of theseobservations is yet to be determined.

Until recently, it was assumed that the mechanism ofcyclic electron transfer was the same in R. (apsulatus as itwas in R. sphaeroides. As expected from the previouslymentioned pathway of cyclic photosynthetic electron flow,cyt c' (cvcA) mutants from R. splhieroides were incapable ofgrowth by photosynthesis (Donohue et al., in press). Therate of RC reduction in (-vA mutant strains grown underconditions gratuitous for ICM synthesis was approximately10,000-fold slower than in the wild-type strain (Donohue etal., in press). These results provided genetic support forearlier spectroscopic studies which demonstrated that cyt c,was the immediate electron donor to the R. splihaeloidces RC(94). In contrast, (YeA mutants in R. capsulatus werepreviously shown to be capable of photosynthetic growth(33); kinetic studies demonstrate a cyt c,-independent routeof RC reduction which was not dependent on a periplasmicredox protein (95). since this process occurred in sphero-plasts from both wild-type and (eyA mutants. Although thisdifference in photosynthetic electron transport between R.(apsulatlits and R. sphliaroides was initially unexpected, theresults obtained are now seen as consistent with previousstudies on the differential inhibition of light-induced electronflow by antibodies against cyt c, (94). The immediate elec-tron donor to the Rhoclopseudomonoas viridis RC is a mem-brane-bound c type cytochrome which has recently beenshown to be a lipoprotein (121). Michel and co-workers (121)have proposed that the fatty acids function to anchor thecytochrome subunit in the photosynthetic membrane. Eventhough cyt c) is not the immediate electron donor to the RCin Rlhodopslcidoinonasvsiridis, a periplasmic cyt c, appearsto stimulate RC reduction in vitro (103).

GENETIC ORGANIZATION AND IDENTIFICATION OFT'HE STRUCTURAL GENES FOR 1CM COMPONENTS

Chromosomal Mapping

A classical genetic analysis of the R. spllaeroides and R.(apsilatus genomes has utilized chromosome transfer tech-

niques that use promiscuous R plasmids (75, 106, 108, 125).Miller and Kaplan were the first to show the usefulness ofpromiscuous R plasmids to the genetic analyses of photo-synthetic bacteria (81). Since then, several groups have usedthese plasmids to develop genetic techniques in the photo-synthetic bacteria (75, 106, 108, 125). Sistrom and co-workers demonstrated conjugal transfer of chromosomalgenes in R. splitieoides mediated by the broad-host-range,self-mobilizable plasmid R68.45 (106) and, using a variety ofauxotrophic and photosynthetic deficient mutants, orderedseveral alleles on the chromosome through a series of two-and three-factor crosses (Fig. 5). Similarly, Pemberton andco-workers (88, 108), using chromosome transfer techniquesand linkage analyses, mapped six genes involved in carote-noid biosynthesis (crtA, crtB, (ctC. crtD, crtE, and crtF)relative to the chromosomal genes plwe-2, and arg-4 (Fig. 6).Using a replication temperature-sensitive R plasmid, pTH 10,Willison et al. (125) generated a circular linkage map of theR. capsulatas BlO chromosome which demonstrated thatmutations affecting nitrogen fixation were dispersed in sev-eral linkage groups on the chromosome. Fine-structuremapping of several nif mutations was accomplished by usingthe gene transfer agent specific for R. capsulatus (125). Genetransfer agent acts similar to a generalized transducing phagein that it packages random tragments of chromosomal DNA(approximately 5 kilobases IkbI in length), but unlike a moretypical transducing phage it cannot independently replicate(74). Willison et al. (125) further showed that the photosyn-thetic gene cluster maps between one of the nziJ gene clustersand a gene for histidine biosynthesis. It is interesting that anif mutation isolated from R. sphaier-oides also maps near thegenes for carotenoid and Bchl biosynthesis (88).

With the availability of the techniques of molecular genet-ics, more sophisticated techniques of chromosomal mappingare being developed. Dryden and Kaplan (unpublished re-sults) as well as Weaver and Tabita (119) have constructedcosmid banks by using vector systems that can be stablymaintained in R. splhaierloides, and similar chromosomalbanks have been constructed from R. (apsulatus (68). Sincethe average cosmid generated in these studies carries 25 to35 kb of insert DNA, a minimum of 200 random isolates issufficient to represent the entire chromosome assuming agenome size of approximately 5,000 kb. Therefore, thesebanks will be useful for chromosomal mapping studies aswell as for the identification of genes by mutant complemen-tation analysis. By the use of orthogonal field gel electro-phoresis, the R. sphlaerioid(es chromosome can be resolvedinto a minimum of 13 DNA Di-aI restriction enzyme frag-

psn-52b met-I

,(his-2, his-11)(psn-49b, psn-67d)

str-Il-

nal-IlFIG. 5. Linear genetic linkage map of the R. sphlaeroides WS2

chromosome. Courtesy of A. Maculosa and W. R. Sistrom.

Vol. 52. 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

56 KILEY AND KAPLAN

met-5 leu-2I I

trp-1

phe-6 phe-4gly-I arg-2 phe-I rsrI-l

trp-3 str-14 gly-3| leu-1 aro-2 nif-i

rif-10 gly-4ileu-2

ade-3

I-I1- I IIarg-6 leu-3 aro-Ilarg-4phe-5 phe-2

phe-7 --I

*

his-6 I crtA I crtD|crtE IcrtB crtC crtF bChAB

FIG. 6. Linear genetic linkage map of the R. spliaer-oides RS630 chromosome. Courtesy of J. Pemberton and with the kind permission ofPlenum Publishing Corp.

ments ranging in size from approximately 50 to >600 kb(W. Y. Qiang and S. Kaplan, unpublished results). From thisanalysis, the genes encoding the following proteins werefound to reside on a unique approximately 650-kb DraIfragment: cyt c2, phosphoribulokinase, fructose bisphospha-tase, ribulose 1,5-bisphosphate carboxylase/oxygenase(form II), b-aminolevulinic acid (ALA) synthase, and B800-850, B875, and RC-H, RC-M, and RC-L polypeptides.Furthermore, the genes for cyt c, and the B800-850, B875,and RC polypeptides all hybridized to an approximately100-kb region of R. sphaeroides DNA from the R' pWS2(106, see subsection, "Identification of PhotosyntheticGenes"). By use of the cloned genes described in this reviewtogether with a number of Tn5-induced auxotrophic mark-ers, the linkage relationships of those DNA regions contrib-uting to ICM structure, function, and synthesis can bedetermined. Recently, two strains of R. splhaeroides havebeen described that will aid in complementation analysissince, phenotypically, one appears to be devoid of theendogenous restriction enzyme system and the second mu-tant may be deficient in recombination activity (89, 106).

Site-Specific Mutagenesis

Generation of site-specific insertion mutations in the pho-tosynthetic bacteria, designated interposon mutagenesis,was first demonstrated in R. (apsulatius (135). In thesestudies, gene transfer agent was used as the vehicle todeliver genes that had been inactivated or replaced byantibiotic-resistant cartridges back to the R. (apsidatltschromosome by homologous recombination (135). By usinga slightly different approach, site-specific insertion muta-tions have been constructed in R. sp/iaeroides through theuse of the suicide vector system of Simon et al. (104), whichallows for a homologous recombination event to occurbetween the modified gene on an unstable plasmid, such aspSUP202, and the wild-type gene on the R. spll(leroideschromosome. Whereas the pBR replicon appears to beincapable of replicating in R. sphlaeroides (50), it appears tofunction normally in R. capsidatas (135). Such site-specificmutations have been constructed in numerous photosyn-thetic genes with relatively high frequencies in both R.capsl/latlis (33, 35) and R. sphaeroides (37; Donohue et al.,in press). The analysis of such mutants has been instrumen-tal in and will continue to aid in our understanding of themechanisms of photosynthetic electron transfer and theregulation, structure, and function of ICM components, ashas been referred to at various points in this review.

Utility of the lac OperonNano et al. (85) were the first to demonstrate the potenti-

ality of the Escherichia coli /ac system for molecular geneticanalyses in the photosynthetic bacteria by demonstrating theapplication of this system to R. sphaeroides. Since R.spliaeroides is naturally unable to use lactose as a carbonsource, expression of the E. coli lac operon dependent ontranscriptional or translational fusions of R. sphaeroidesDNA to /acZ provides a very powerful genetic tool both tostudy the expression of specific genes and to select formutations that affect expression of various gene fusions.Recently, T. Tai, W. Havelka, and S. Kaplan (submitted forpublication) have prepared a variety of plasmid construc-tions based on the RSF1010 replicon that facilitate cloningand the construction of translational /acZ fusions in R.sp/haeroides.

Gene ExpressionThe recent advances in molecular genetics in the photo-

synthetic bacteria has led to studies of gene expression forthose ICM components described above. Numerous studieshave noted the general lack of expression of R. sphaeroidesgenes in E. Coli (17, 19, 83, 84). The difference in expressionappeared to be at the transcriptional level, which may stemfrom the difference in the genomic G+C content of DNAfrom R. sphaeroides (68 to 70 mol% G+C; 93) and that of E.coli (50 to 51 mol%), although other possibilities have notbeen ruled out. To our knowledge, no E. coli -10 and -35consensus promoter sequences have been found upstream ofany genes so far sequenced from any species of photosyn-thetic bacteria.To study gene expression in these bacteria, two in vitro

systems have been developed. A homologous coupled tran-scription-translation system from R. sphlaeroides has beenused to express R. splhaeroidcs genes from exogenouslyadded DNA templates (17, 19, 40, 42, 57, 66, 67). This samesystem has proven to be of general utility in its ability toexpress DNA sequences from other high-G+C procaryotes(J. Chory, T. Donohue, P. J. Kiley, W. Havelka, and S.Kaplan, unpublished observations). Ribonucleic acid (RNA)polymerase from both R. capsulatus (51) and R. sph/aeroides(J. Kansy and S. Kaplan, manuscript in preparation) havebeen purified, and an in vitro transcription assay system iscurrently being used to identify promoter regions for variousphotosynthetic genes. Although 5' ends of several stablemessenger RNA (mRNA) species have been mapped inseveral species of photosynthetic bacteria (7, 56, 118, 140)and attempts have been made to compare potential upstream

his-3

I Ihis-I mmc-Ihis-2his-4

MICROBIOL. REV.

trp-5

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 57

regulatory sequences, it is our feeling that sufficient data arenot available at this time to unambiguously identify canoni-cal promoter sequences in any of the photosynthetic bacte-ria.

Identification of Photosynthetic Genes

From genetic studies, it has been shown that the genesencoding the Bchl and carotenoid biosynthetic enzymes, andthe subunits for the B875 and RC pigment-protein com-plexes, are clustered in both R. splIaeroides (88, 106) and R.capsulatus (75, 114, 130). This was first demonstrated by theisolation of an R-prime from R. capsulatus carrying anapproximately 50-kb insert of R. capsulatus DNA (pRPS404)which could complement most mutants deficient in variousaspects of photosynthetic function (75). It was later shownthat this plasmid contained the genes for the latter steps incarotenoid biosynthesis (the enzymes required to synthesizethe colored carotenoid species from the ultraviolet light-absorbing phytoene; 100), Bchl biosynthesis (the enzymesthat convert protoporphyrin IX to Bchl), and the structuralgenes for the RC and B875 polypeptides (114, 132). Sistromet al. have subsequently isolated an R-prime, pWS2, from R.sphaeroides which appears to be functionally equivalent tothe R. capsulatus R-prime pRPS404 (106). Pemberton andHarding (88) constructed a cosmid bank from R. sphaeroidesand by complementation analysis have identified cosmidscarrying six different genes involved in carotenoid biosyn-thesis. Recently, two separate structural genes (hemnA andhemT) for ALA synthase have been cloned from R. sphaer-oides (T. Tai and S. Kaplan, manuscript in preparation),using a heterologous probe containing the hemnA gene ofRhizobilirn meliloti (72). ALA synthase is responsible forthe synthesis of ALA, which is the first committed precursorin tetrapyrrole biosynthesis. Neither of these genes areclose to the pluf or puih operons of R. sphaeroides asdetermined by restriction enzyme mapping (see "Chromo-somal Mapping"). Another key biosynthetic pathway nec-essary for ICM formation is isoprenoid synthesis since it is aprecursor of carotenoids, phytol (for Bchl synthesis), andisoprenoid quinones. The key regulatory enzyme in isopre-noid biosynthesis in eucaryotes is the 3-hydroxy-3-methyl-glutaryl coenzyme A reductase. When Dryden and Kaplan(unpublished results) attempted to identify this gene in R.sphaeroides, using synthetic deoxyoligonucleotides basedon the conserved region within the catalytic domain of3-hydroxy-3-methylglutaryl coenzyme A reductase asprobes in a Southern hybridization analysis, three uniqueEcoRI restriction fragments were identified. Each individualfragment was cloned and subsequently shown to contain acommon homologous PvulII restriction fragment which alsocontained homology to the synthetic probes. It is not knownat this time whether this represents three distinct genes forthis enzyme and, if this is the case, whether all three genesare functional and, if so, the nature of their physiologicalroles.

Several of the structural genes for previously describedICM components have been cloned by using either heterol-ogous gene probes or synthetic deoxyoligonucleotides de-signed from the available cognate amino acid sequences. Ingeneral, most of the genes encoding subunits of a particularcomplex are linked in R. sphaer-oides (5, 40, 66, 67. 122-124),R. capsulatus (132, 134), Rhodospirilllin ruibruirn (8), andRhodopselidornonas i'iridis (79, 80), with one very notableexception. In those bacteria that have been examined, thestructural genes for two of the RC polypeptides, RC-L and

RC-M, designated plLfZ and pulM, respectively, are tran-scriptionally linked to the two structural genes for the (x andA subunits of the B875 complex, piifA and pullB, respectively(7, 139, 140). The structural gene for the third polypeptide inthe RC, RC-H (designated puhA), is not linked to the piufoperon, and preliminary data suggest that it maps at least 40kb away in R. spli(weroides, as it does in R. capsulatus (114).The structural genes for the B800-850 subunits cx and a arealso linked to one another and comprise the pilc operon (5,67, 134).The genes encoding the subunits of the cyt b/c 1 complex

have recently been cloned in both R. capsulatus (35, 36) andR. sphaeroides (56) and comprise an operon. The gene orderis Reiske iron-sulfur protein, cyt b, and cyt c1, orfbc as firstdesignated by Gabellini et al. (56). The original identificationand cloning of these genes were reported to be from R.spliaer-oides (56), although it was subsequently shown thatthe source of these genes was most likely R. c(apsulatIls (36).This apparent discrepancy originated from the use of asupposed R. sphaeroides mutant strain, GA, that was latershown to have been most probably derived from R. capsu-Iltuis (36). This green mutant GA used by Gabellini et al.should not be confused with the R. splhaeroides 2.4.1 deriv-ative Ga, isolated by Cohen-Bazire (29, 105), which we haverecently reconfirmed is from 2.4.1 by phage typing (T.Donohue and S. Kaplan, unpublished results). Subse-quently, the Jbc operon has been reisolated and character-ized from both species (36). The gene encoding the solublecyt c', ((v(A) has also been cloned, and its map location isknown in R. sphaeroides (see "Chromosomal Mapping").The membrane-bound cyt c from R. viridis has been identi-fied downstream of the genes for the RC-L and RC-Mpolypeptides (79). Presumably, it is cotranscribed with theR. viridis puif'operon, since the initiation codon for the cyt cgene overlaps with the termination codon for piifM (79). Thisdifference in genetic organization between the pi(foperon ofR. iiridis and that of R. sphaer-oides and R. capsulatus isinteresting in light of the differences in the three-dimensionalcrystal structure of the RC (see "Composition of 1CMComplexes"). In both R. sphaeroides (B. Dehoff and S.Kaplan, unpublished results) and R. capsulatius (132), thereis a different open reading frame downstream of puiiM thatwould code for a hydrophobic 82-amino-acid polypeptide(puJX). This segment of DNA is cotranscribed with the largeplIf operon transcript (see section entitled Expression of theplf operon) in both R. sphaeroides (R. Gyure and T.Donohue, personal communication) and R. capsulatus (138),although the function of this polypeptide is presently un-known. There is no information currently available as to theexistence of a similar open reading frame downstream of thegene for the membrane-bound cyt cofthe R. viridis RC. It ispossible that the pli(X gene product could play a role in RCstructure or assembly together with B875 complexes.

REGULATION OF ICM SYNTHESIS

When considering the biosynthesis and regulation of ICMformation in R. splhaeroides, it is preferable to analyzesteady-state growing cell cultures. Light intensities of 10W/m2 (measured through >650-nm cutoff filters) have beenused for many physiological experiments involving photo-synthetic cells, and cells exposed to a 25% 02 atmosphere(this represents 100% dissolved O0 in the culture medium;16) have been used for aerobic chemoheterotrophic growth.At 10-W/m2 light intensity, R. splhaier-oides 2.4.1 has an

VOL. 52, 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

58 KILEY AND KAP'LAN

optimum growth rate of approximately 2.5 to 3.0 h but stillderepresses both the amount of intracellular ICM and B800-850 complexes relative to cells grown at supersaturating lightintensities such as 100 W/m2 (18). One major problem thathas arisen in studying the effects of changes in oxygenconcentration or light intensity on various aspects of ICMbiosynthesis is the variability in cultural conditions used inindependent laboratories. Therefore, it has been ditficult tocompare the magnitude of the effects observed from study tostudy, and thus one can only consider the relative changes.For example, total repression of ICM synthesis in R. sphacr-oides during active growth requires high oxygen tensions(16): cells grown at high aeration on a gyratory shaker(commonly used laboratory conditions) quite often do notgenerate high enough cultural oxygen tensions to repressICM synthesis even at low cell densities. Only a few studieshave actually characterized the levels of various photosyn-thetic related activities in cells grown in high oxygen ten-sions (16. 40( 42. 67, 138. 139), and therefore we know thebasal level for only a relatively few activities in cells that arecompletely repressed for ICM synthesis. Superimposed onthese difficulties is the tendency to directly compare orequate what are apparently complex physiological responsesto changes in oxygen tension or light intensity in differentspecies of photosynthetic bacteria, which may or may not bevalid.

Previous studies have described the requirement for thecoordinate synthesis of Bchl and protein for assembly of theICM (11, 12, 113). In contrast, concurrent insertion ofphospholipids (102) or carotenoids (29, 34) into the 1CM ofsteady-state cultures is not required for the assembly offunctional pigment-protein complexes. One possible excep-tion to this is that assembly of B800-850 complexes may betightly coupled to synthesis of colored carotenoids since allsimple mutants described to date that lack colored carote-noids do not synthesize wild-type B800-850 complexes (5,34, 84, 105). The coupling between ICM protein and Bchlsynthesis has been best described in synchronous cultures ofR. sphaeraoidles (65) in which the ICM phospholipid/proteinratio varies throughout the cell cycle due to the continuousinsertion of protein into the membrane, with the movementof previously synthesized phospholipid into the ICM occur-ring just prior to cell division. In asynchronous steady-statephotosynthetic cells, the average ICM protein/phospholipidratio in the population remains constant, and this ratio isindependent of the light intensity for growth (129). However,increases in the amount of ICM per cell in response tochanges in light intensity are accompanied by increases inthe cellular amount of phospholipid and Bchl-protein com-plexes (18. 65).

Bchl Synthesis

Coupling between synthesis of a polypeptide and its ligannd(i.e.. Bchl), as in the assembly of the LH and RC complexes,is an integral part of assembling the photosynthetic mem-brane. This is supported by the fact that wild-type strains donot have detectable pools of free Bchl, LH. or RC polypep-tides and that mutants blocked in Bchl synthesis do notsynthesize ICM or accumulate large quantities ot Bchl-binding proteins, even under gratuitous growth conditionsfor ICM synthesis (see above; 11, 12, 70, 113). TIhe regula-tion of Bchl synthesis in R. sphaeroides is complex, sinceintermediates in the early part of the biosynthetic pathwayare used under photosynthetic conditions for synthesis of

two other tetrapyrrole-containing compounds, namely,hemes and corrinoids (71. 96). Moreover, the ability tosynthesize Bchl under photosynthetic conditions withoutinterrupting the flow of intermediates to heme or corrinoidsis further evidence of the complexity of this pathway sincethe Bchl/heme ratio is 10 to 50:1 in photosynthetic cells (71,96). Further compounding this regulatory problem is theabsence of information on how bacteriopheophytin (Mg-freederivative of Bchl, found only in the RC complex) is synthe-sized.

Analysis of individual steps specific to the Bchl branch ofthis pathway has been hampered by the difficulty of assayingand purifying the enzymes involved in Bchl synthesis. Con-sequently. most of the steps in Bchl biosynthesis have beendefined by the characterization of mutants blocked at vari-ous steps in the pathw-ay. This subject has been extensivelyreviewed by Lascelles and co-workers (71, 96) and Jones(62). The complementation experiments of Marrs and co-workers (75. 130. 133) identified the genes required forsynthesis of Bchl trom R. c(apsula(tus which were localizedon the R-prime pRPS404. and a physical-genetic map wassubsequently generated for this cluster of genes (114). Tothis date, the activities encoded by the cloned Bchl geneticloci have not been demonstrated, nor has the DNA sequencefor these proposed genes been determined in any of thephotosynthetic baicteriat: neither of these tasks (especiallythe former) will prove to be easy. Confirmation otf theproposed steps in the Bchl biosynthetic pathway awaits thedemonstration of these activities in vitro.

Since the final product. Bchl. is relatively easy to assay,most of what is known about the regulation of Bchl biosyn-thesis in R. sphaer(eides is known from examining cellularlevels of this end product. For example. chemoheterotrophi-cally grown cells of R. sphaeroidles in high oxygen atmo-sphere show no detectatble Bchl or ICM invaginations (16).Reduction or removal of oxygen ftrom chemoheterotrophi-cally growing cells induces synthesis of both Bchl and theICM. In R. (apsilatus, studies attempting to describe thetranscriptional regulation of the Bchl biosynthetic geneshave shown that hybridization to two Batn HI restrictiontraLgments containing the bO/i. -(C. -I) -H. -K. and -F genesof the Bchl branch of the biosynthetic pathway could beobserved with mRNA isolated from aerobically grown cellsof R. (apsillatus (20. 138). It is not clear whether these cellswere fully repressed for ICM synthesis. Biel and Marrs (9),using transcriptional, lacZ fusions into bclhB, bch(, bch/G,and bclH, showed a twofold induction of 3-galactosidaseexpression when R. capsla(lits wcls grown under low aera-tion versus high aeration, which was similar to the resultsobtained through RNA hybridization studies with specificrestriction fragment probes (20, 138). However, since thedelineated bounditi-ies of these genetic loci can representonly a minimal estimate of both gene size and number, onehas not yet been atble to correlate the levels of mRNA withalny specific enzyme alctivity. especially since the restrictionfragments used in these studies were not specific to onegenetic locus (20, 138) and sufficient information is notavailable on the nuLmber or transcriptional organization ofindividual structuratl genes within these loci.

Several steps have been implicLated as being important tothe control ot Bchl synthesis: (i) the alctivity of ALAsvnthase (ALA is the first committed precursor in thetetrapyrrole pathway): (ii) the branch point where either iron(heme biosynthesis) or magnesium (Bchl biosynthesis) isincorporated into protoporphyrin IX: (iii) the synthesis ofphytol side chain of Bchl; (iv) synthesis of a carrier for Bchl

M ICROBIC)l. RF-V.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 59

intermediates as proposed initially by Lascelles and co-workers (71, 96) and more recently by Bauer and Marrs(C. E. Bauer and B. L. Marrs, Abstr. Mol. Biol. Photosynth.Procaryotes, 1987, p. 2). Early studies by Lascelles andco-workers demonstrated that synthesis of the ALA syn-thase and S-adenosylmethionine-Mg-protoporphyrin meth-yltransferase (encoded by bchH) were repressed by growthof R. sphaeroides under high aeration on a gyratory shaker(71). S-Adenosylmethionine-Mg-protoprotoporphyrin meth-yltransferase is the only enzyme specific to the Bchl branchfor which an enzyme activity has been demonstrated invitro, and from these studies the incorporation of Mg intoprotoporphyrin IX appears to be coupled to the methylationstep (96). Therefore, it has not been possible to compareenzymatically the Mg or ferrochelatase activity in terms of°2 sensitivity, inducibility by anaerobiosis, or their relativeaffinities for protoporphyrin IX. Genetic evidence for theexistence of a Bchl carrier has been recently suggested in R.capsulatus (the proposed function of the Q gene describedby Bauer and Marrs) (see "Expression of the piif Operon"').Although the function and characterization of Q gene prod-uct has not yet been demonstrated in vivo, mRNA specific tothis gene has only been detected under photosyntheticconditions in R. sphaeroides (J. Lee and S. Kaplan, unpub-lished results).

Photosynthetically grown cells have been shown to containmore ALA synthase activity than chemoheterotrophicallygrown cells (71). Two forms of ALA synthase have beendemonstrated in R. sphaeroides and the second form may bemembrane associated (47). One way of sequestering interme-diates for Bchl synthesis could be by compartmentalizingthese at the ICM where Bchl will ultimately be inserted ascomponents of the Bchl-protein complexes. The putativeBchl carrier protein mentioned above might function in thisrole as would the suggested membrane association of ALAsynthase. ALA synthase requires a reducing environment forenzyme activity, and in vitro ALA synthase can be activatedby a thioredoxin/thioredoxin reductase system (22, 23). It hasbeen postulated that the levels of thioredoxin in vivo may berelated to the redox state of the cell and, therefore, thisprotein may play a role in light regulation of Bchl synthesis inphotosynthetic bacteria as has been previously suggested forplants (22, 23, 96). In this regard, since light regulation of Bchlsynthesis has been suggested to be controlled by the availablepool of ALA (86). the amount of reduced thioredoxin inphotosynthetic cells may also be influenced by light-inducedelectron flux under photosynthetic conditions.

Recently, Tai and Kaplan (in preparation) have identifiedtwo unlinked restriction endonuclease fragments that hy-bridize to a Rhizobilim ineliloti structural gene (hemA) forALA synthase. Each of these DNA fragments complementsboth E. coli and R. sphaeroides IlewmiA mutants, and for oneisolate high levels of ALA synthase activity were detected inthe complemented strain. Therefore, it is interesting to specu-late that these may represent two copies of the heinA gene andthat each gene may be regulated differentially with respect tooxygen or light or both. Thus, we would suggest that mutantsrequiring ALA specifically for photosynthetic growth shouldbe unobtainable because loss of membrane-bound ALA syn-thetic activity would be compensated for by the freely availablepool of ALA produced by the soluble ALA synthase. Clearly,elucidating the details of the molecular mechanisms controllingBchl synthesis is imperative to understanding not only Bchlsynthesis itself, but also how the flow of Bchl and its precursorsinteract with other molecules to regulate assembly of thephotosynthetic membrane. The identification and analysis of

the genes and gene products for enzymes in the Bchl (9, 30,130) and carotenoid (88, 100, 130) biosynthetic pathways in R.(apsuIlatuls and R. sphlaeroides will thus be instrumental toeventually understanding the coordinate regulation of thesecomplex biosynthetic pathways.

IC1. Assembly in Synchronous Cell Populations

Under steady-state photosynthetic growth conditions, acell cycle-dependent synthesis of the ICM has been ob-served. This topic has been extensively reviewed (65) and isonly briefly summarized here. Following cell division, LHand RC complexes and other membrane components (e.g.,adenosine triphosphatase), as well as many unidentifiedproteins, are continuously incorporated into preexistingICM (53, 58, 127). Whole-cell phospholipid synthesis, al-though transiently interrupted at the time of cell division,does not result in new phospholipid being incorporated intoICM during the course of the division cycle (14, 73). Imme-diately preceding, or at the time of cell division, previouslysynthesized phospholipids are mobilized and incorporatedinto the replicating ICM as it is being partitioned to daughtercells.The consequences of this cell cycle-specific uncoupling of

protein and lipid insertion into the ICM are manyfold andwell documented, namely: (i) the ICM protein/phospholipidratio fluctuates in a cyclical pattern throughout the cell cycle(54); (ii) the intrinsic density of the ICM fluctuates in asimilar cyclical fashion (54); (iii) the fluidity of the ICMphospholipid bilayer reveals a cyclical pattern (65); (iv) theintramembranous particle distance between photosyntheticunits changes directly with the protein/phospholipid ratio(129); (v) the movement of electrons through the electrontransport chain reveals second-order kinetics at the highprotein/phospholipid ratio, suggesting that this process ismore efficient when the ICM particle density of ICMintramembranous complexes is high (107); (vi) the protonadenosine triphosphatase shows cyclical fluctuations in itsactivity despite the continued increase in adenosine triphos-phatase antigen (58).The enzymes involved in phospholipid synthesis are local-

ized exclusively in the CM in both chemoheterotrophicallyand photoheterotrophically growing cells (15). Thus, the cellcycle-specific accumulation of phospholipid in the ICMdescribed above must be regulated by controlling the move-ment of phospholipid into the ICM as opposed to regulatingphospholipid biosynthetic enzyme activities localized withinthe ICM. Because phospholipid transfer activities have beendiscovered and purified from R. sphlaeroides (26, 112), thesemight also be considered to play a role in phospholipidmovement into the ICM. In this regard, recent studies havedemonstrated that chromatophores with a high protein/phos-pholipid ratio (i.e., those isolated from cells just prior to celldivision) are better substrates in the in vitro phospholipidtransfer activity assay than those purified from cells justafter cell division, which have a lower protein/phospholipidratio (111).From the studies of synchronous cell populations, it is

clear that the number of photosynthetic units per chromato-phore increases during the cell cycle, which suggests thatwithin defined limits the density of ICM fixed photosyntheticunits is related to the protein/phospholipid ratio. Examina-tion of steady-state asynchronously grown cells at differentlight intensities shows that the apparent density of photo-synthetic units per chromatophore is relatively constant(129). This is also apparent from the data in Table 1 which

VOL. 52. 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

60 KILEY AND KAPLAN

demonstrate that the amounts of RC and B875 polypeptidesdo not vary when we compare equal amounts of chromato-phores isolated from R. sphaeroidles 2.4.1 grown at differentlight intensities. Decreases in the incident light intensity ofgrowth result in increases in the cellular content of the 1CMas does de novo ICM synthesis, which occurs when highlyaerated cells are shifted to photosynthetic growth undernongratuitous conditions (see below). In light of these ob-servations, there must be some regulatory mechanism thatcouples the number or formation of "new' 1CM invagina-tions to synthesis of fixed photosynthetic units which is notmerely dependent on the inser-tion of these complexes intothe membrane. The recent demonstration of a mutant of R.sphaeroides which contains invaginations only at the cellpoles and in which new invaginations only appeari- to arise atthe site of cell constriction during division may prove usefulin addressing questions pertaining to the process of ICMformation (67a). However, freeze-fracture analysis of wild-type cells undergoing a transition from aerobic to photosyn-thetic conditions did not show a localization of' ICM forma-tion at the cell poles (16).

Induction and Assembly of the ICM

The use of environmental shifts is a common exper-imentalapproach that has been used to study assembly of the ICM.This approach has been useful in defining, at the cellularlevel, a hierarchy of controls directed by various environ-mental factors. For example, lowering the pO, of a chemo-heterotrophically growing culture (representing a gratuitousshift) is sufficient to trigger the induction of ICM develop-ment (63). Under these conditions. accumulation ot B875and RC complexes (69), as well as the mRNA for thesepigment-protein complexes, is detected prior to synthesis ofthe B800-850 complexes (60). In contrast, both mRNA (Leeand Kaplan, unpublished results) and all three spectralcomplexes appear coordinately in the ICM when chemohet-erotrophically grown cells are shifted to stringent nongratui-tous anaerobic photosynthetic conditions (16). This suggeststhat synthesis of the B800-850 complexes responds differen-tially with respect to cultural oxygen tensions when com-pared with B875 and RC complexes. Therefore. it is ofinterest to understand how synthesis and assembly of theICM and its associated functional complexes are integratedand controlled at both the molecular and genetic levels underthese two distinct conditions of ICM induction. However, ithas only recently been possible to address the control ofICM formation at this level due to the identification of thestructural genes for several ot the ICM components.The occurrence and significance of phospholipid move-

ment into the newly synthesized ICM were demonstratedduring the process of de novo ICM induction under nongra-tuitous conditions (16). In the absence of net phospholipidsynthesis or increase in cell number, new invaginations ofthe CM were seen early during the nearly 10- to 12-h lag priorto the resumption of growth under photosynthetic conditions(16). Moreove,- the first invaginations observed during thislag period were much larger, irregular- in shape. and morelipid enriched than mature ICM. The newly formed ICMinvaginations may be related to a membrane fraction, whichhas been proposed to represent initiation sites for new ICMinvaginations (60). This membrane fraction isolated by Nie-derman and co-workers from cells undergoing a gratuitousshift has been designated as the upper pigmented band(UPB) due to its behavior on sucrose gradients (60). althoughit is not clear what percentage ot the total cell membrCane this

UPB represents. The UPB fraction appears to be preferen-tially enriched in B875 and RC complexes, and it has beensuggested that B800-850 complexes are not coordinatelyinserted with B875 and RC complexes into this membranedomain (60). The kinetics of insertion of pigment-proteincomplexes into the UPB parallel that observed during denovo synthesis of new ICM invaginations under gratuitousconditions (60). Per-haps related to this is the fact that an R.sph(aroidle.s mutant deficient in B800-850 complexes andcarotenoids (RS104) does not make mature ICM, but rathersynthesizes long tubular ICM (67a). It is interesting tospeculate that the tubular ICM formed in this mutant repre-sents unrestricted nascent invaginations derived from theinitical large lipid-enr-iched invaginations described by Choryet al. (16) or that they may be related to the UPB fraction(60). However, the tubular ICM in RS104 differs from thenascent invaginations in wild-type cells in that they extendfrom the cell poles and are unable to "mature," presumablydue to the absence of B800-850 complexes.

Induction of the 1CM, in addition to requiring the preciseregulaktion of specific transcriptional activities, necessitatesthe coordinate assembly of individual spectral complexesprior to or concomitant with their incorporation into theICM. Although these events are ultimately linked by as yetundefined mechanisms in wild-type bacteria, analysis ofmutants defective in various aspects of ICM synthesis re-veals that uncoupling of these processes is possible. Forexample. mutants containing a kanamycin resistance car-tridge in either the pl/il (L. Sockett and S. Kaplan, unpub-lished results) or piif (37) operon of R. sphueroides areunable to grow photosynthetically, but are able to synthesizeICM when grown in gratuitous conditions of anaerobic darkrespiration. Thin-section electron microscopy of these mu-tants showed that the Puf mutant contained somewhatirregulalr- but vesicular shaped 1CM, whereas the Puhmutant appeared more like the wild type grown under similarconditions (Varga and Kaplan, unpublished results). Incontrast. N-methyl-N'-nitro-N-nitrosoguanidine-inducedmutants, which are only lacking one of the LH complexes,are able to grow photosynthetically, and examination of theICM f'rom a B875 sti-ain shows morphologically wild-typeICM invaginations (67a). As mentioned above, mutantslacking B800-850 complexes are still able to form invagina-tions, but these are somewhat distorted and are tubular inappearance. Thus, the analysis of both wild-type and mutantstrCains independently confirms that the signal(s) for phos-pholipid movement required for ICM formation is indepen-dent of the synthesis of spectral complexes and that theprocess of ICM invagination or maturation is separable fromthe insertion of individual intact spectral complexes. Theseobservations also suggest that mature vesicular invagina-tions, characteristic of steady-state photoheterotrophiccells, are highly dependent upon the presence of B800-850complexes and perhatps carotenoids.However, once the process of ICM invagination begins in

wild-type cells, assembly of mature ICM appears to becoupled to insertion of Bchl-protein complexes. That the Hpolypeptide of the RC is present in the CM of chemohetero-trophically grown cells (16) and that its total cellular abun-dance in photoheterotrophically gi-own cells is in excess ofRC-M .and RC-L (Table 1) lead to the suggestion that RC-Hmay serve a critical role in permitting RC-M and RC-L toenter the membrane at a particular point (docking) or toorient themselves (scaffolding) one to the other around theone membrane-spanning helix of the RC-H subunit. Subse-quent events. such as the positioning of the B875 complex

MICROBIOIL. REV.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTIHESIS IN R. SPHAEROIDES 61

relative to the RC, could depend upon the presence of eitheran intact RC complex or components of the RC as well asother posttranscriptional or posttranslational factors. This issupported by the observation that insertion of a kanamycincartridge into a plhA deletion results in the reduction of RC(1%) as well as the loss of B875 complexes but not B800-850complexes (Sockett and Kaplan, unpublished results).Cross-linking studies of R. capsidltitis ICM has independent-ly revealed a relationship between RC-H and B875 subunits(91). Finally, the assembly of the B800-850 complexesaround the fixed photosynthetic unit consisting of B875 andRC complexes can be considered to be independent of RCand B875 spectral complexes, since the amount of B800-850complex in the variable photosynthetic unit is tightly regu-lated by light intensity in wild-type cells and can be uncou-pled from RC and B875 synthesis in mutant strains. More-over, the above-described studies with specific mutantstrains would suggest that the presence of B800-850 com-plexes in the ICM may be linked to the formation of matureinvaginations but not dependent on the presence of RC orB875 complexes. Table 1 and Fig. 8 demonstrate thatchanges in the amount of Bchl per milligram of cell proteinobserved in R. sphaeroides 2.4.1 grown photosyntheticallyunder different light intensities are reflected in both theamount of ICM per cell and the size of the variable photo-synthetic unit. The cellular level of B800-850 complex rela-tive to the B875 complex is greater when comparing cellsgrown at 10 W/m2 versus those grown at 100 W/m2 thanwhen comparing cells grown at 10 W/m2 versus those grownat 3 W/m2. Thus, we suggest that the greatest contribution tothe increase in Bchl per milligram of protein when comparingcells grown at 10 W/m2 and those grown at 3 W/m2 isprimarily the result of an increase in the amount of ICM percell and not the result of changes in the size of the variablephotosynthetic unit. In turn, the presence and abundance ofall spectral complexes in wild-type or mutant strains dependon the steady regulated synthesis of Bchl.

In wild-type cells the observation that the number of RCcomplexes per unit area of ICM is relatively constant andindependent of light intensity (129) suggests that, once theprocess of ICM invagination takes place, the intracellularlevel of such invaginations is tightly linked to the cellularlevel of RCs. However, as previously stated, the ability toform new invaginations during de novo ICM induction doesnot depend upon the presence of intact RC complexes or theRC-H subunit since this appears to occur normally in strainslacking RC-H antigen. Future studies will be required toaddress whether other specific membrane proteins or phos-pholipid domains within the membrane are involved in theformation of ICM invaginations.One other point worth considering is that secondary

structure predictions (122) and X-ray crystallography data

indicate that the bulk of the RC-H polypeptide protrudes intothe cytoplasm of the cell. The orientation of the RC-Hsubunit in the bilayer may make it available for impartingdirectionality for RC-L and RC-M insertion, alignment, andassembly of a functional RC pigment-protein complex. Thisprotrusion of the RC-H subunit could also provide a mech-anism by which the RC complex can signal or communicatedirectly with the cellular interior. For example, if RC com-plexes in an oxidized versus ground state gave rise todifferent conformational states of the cytoplasmic domain ofthe RC-H subunit, such conformational changes could beinvolved in activities such as regulating gene expression orbacterial tactic responses by light and oxygen. Such asensing mechanism would be analogous to the response ofother bacterial membrane proteins involved in responding tochanges in environmental stimuli (97, 126). For example, it ispossible that the RC-H polypeptide may be functionallysimilar to the recently described class of transmembranesensor proteins which provide bacteria a mechanism bywhich they can sense and respond to changes in theirenvironment (i.e., osmolarity changes, phosphate or nitro-gen limitation) (97, 126). By analogy with these other twocomponent systems, there may also be a positive regulatoryprotein which interacts directly with the transmembraneprotein to transmit some oxygen or light-regulated signal,resulting in either direct or indirect transcriptional activationof the genes under its control.

REGULATION OF PHOTOSYNTHETIC GENEEXPRESSION

In an effort to keep this review as current as possible, thissection contains some unpublished results. We trust thereader will distinguish our unpublished observations aspreliminary.The structural genes for the RC and LH polypeptides have

all been placed within transcriptional units in R. spluiewoides(40, 67, 139, 140) and R. capsulatits (7, 137, 138), and recentexperiments in both bacteria have demonstrated that ICMinduction under anaerobic conditions is not solely regulatedat the transcriptional level (see below).

Expression of the puf Operon

The pif operon (Fig. 7) represents the most thoroughlystudied transcriptional unit encoding Bchl-binding proteins,and in R. spIl(roicles it encodes two stable polycistronicmRNA molecules: a high-abundance small transcript (ap-proximately 600 nucleotides Inti), specific for the piufB andpiufA genes, which has a 5' end 104 nt upstream of pufB; anda low-abundance large transcript (2,600 nt) which encodesplIJBALM and which has a 5' end 75 nt upstream of putfB

pufH R Q B A puf LI I=l E00 F- ItD H H c ,HIWH Hco

0 .>C,0

X I >.CL C-o

a- ~-c ya.>0 0

LiiiQICiiL Qii i)I1 11

puf M pufXn 1

H

z

HI

E

a-

a--- V

250 bp

FIG. 7. Restriction map of the pufoperon region from R. sphiaeoidcs 2.4.1. pifBALMX code tor the B875-13 B875-(x. RC-L. and RC-Mpolypeptides, respectively. pl,fX is a gene of unknown function. The Q gene has been proposed to be involved in Bchl biosynthesis (Bauerand Marrs, Abstr. Mol. Biol. Photosynth. Procaryotes. 1987. p. 2). The R gene appears to be involved in assembly of B875 complexes (J.Davis, L. Sockett. and S. Kaplan. unpublished results). bp. Base pairs.

VOL. 52, 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

62 KILEY AND KAPLAN MICROBIOL. REV.

.Vi.*

0XtY. -',':'-.,- '':f1

FIG. 8. Thin-section electron micrographs of R. sphaeroides 2.4.1 grown at (top) high, 100 W/m2; (middle) medium, 10 W/m2; and (bottom)low, 3 W/m2, incident light intensity. The number of ICM vesicles is shown to be inversely related to the incident light intensity (see text andTable 1). Kindly provided by A. Varga.

(140). More recent results demonstrate that the large pluftranscript extends far enough to encode a putative down-stream gene, piJX (Gyure and Donohue, personal commu-

nication), which was first described in R. capsulatus (132).The ratio of small/large piaf operon-specific transcript varies

from 20 to 30:1 in chemoheterotrophically grown R. sphaer-oiles to 8 to 15:1 in photosynthetically grown cells (139,140). Therefore, the obvious question arises as to the originof each of these piuf operon-specific transcripts. Belasco etal. (7) working with R. (capsilatus have suggested specific 3'

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYN'T'HETIC MEMBRANE BIOSYN'T'HESIS IN R. SPHAEROIDES 63

processing of the 2.600-nt-long puIBALMX transcript toyield a steady-state level of 9:1 put BA/pIufBALMX mRNAspecies. In R. sphtieroides it has been proposed that a

stem-loop which could be formed in the intercistronic region

(124 base pairs) between pulBfA and piifLMX serves as a

differential transcription terminator so that the ratio of thesetwo transcripts result from the selective termination ofmRNA species with unique 5' ends upstream of p{f1B (66).Deletion of the purported terminator region in R. sphaer-

oides yielded a mutant that displays a B875- phenotypewhich grows photosynthetically (Dehoff and Kaplan, unpub-lished results). This phenotype is consistent with the inter-cistronic region functioning as a transcription terminator,although more work is required to establish the validity ofthis hypothesis.

Observations from our laboratory have identified a DNAsequence containing a 20-amino-acid open reading frame(orfK) immediately upstream of the puJB structural gene.

This putative "leader'" polypeptide coding region is pre-

ceded by a Shine-Dalgarno sequence, has an ATG-methio-nine, and terminates with a UAG one base prior to the ATGof the plqJB gene. Characteristic of the DNA sequence in thisregion is the use of relatively rare codons for 9 of the 20designated amino acids as well as the use of a UAG codon as

a terminator, which has not been previously observed in any

of the R. sphlatieroidles structural genes sequenced so far.Previous studies have documented differences in transferRNA isoaccepting species in chemoheterotrophically andphotoheterotrophically grown cells (101), and experimentsare under way to test whether this proposed leader peptidefunctions to control translation of pujI'ImRNA in chemohete-rotrophic cells, perhaps by "ribosome stalling," resulting in

the inability to translate pl(f:specific mRNA under aerobicconditions but not under photosynthetic conditions. It is alsopossible that the ratio of small/large pit operon transcriptsmay be influenced by ribosome stalling and could be coupledto differential transcription termination in the intercistronicregion between piujA and pit L. It is evident that the pi(foperon genetic region may be highly regulated by severallevels of control in R. sphaeeroides.

Since the put operon codes for essential gene productsrequired in precise stoichiometries for the light reactions ofphotosynthesis, it is perhaps not surprising that expression

of this operon is highly regulated by physiological condi-tions. In addition to allowing us to study the mechanism(s) ofgene expression in these bacteria, the analysis of howexpression of this operon is regulated has provided theopportunity to study how the stoichiometry, assembly, andorganization of a supramolecular membrane complex are

regulated and, thus, coupled to other aspects of cellularmetabolism required for photosynthesis. The ratio of thesmall/large puJ operon-specific transcripts (approximately 8to 15:1) is essentially constant under all photosyntheticgrowth conditions (Kaplan, unpublished results), and it hasbeen proposed that this excess of small to large pif' tran-

script results in the 10- to 15-fold excess of B875 to RCcomplexes in the ICM (1, 139, 140). Although in both R.splhaeroides and R. capsulauts the 5' ends of the pif(transcripts have been mapped to regions immediately up-

stream of pitB (7, 140), recent genetic evidence suggests thata region several hundred base pairs upstream of pit B in R.(apsulatits is required in cis to obtain regulated expressionof this operon in vivo (6). This region codes for a 77-amino-acid polypeptide, designated the Q gene product (Bauer andMarrs, Abstr. Mol. Biol. Photosynth. Procaryotes. 1987, p.

2). Therefore, in R. (cpsiulatus it has been proposed that the

Q gene is transcriptionally linked to the pt' operon andtranscripts which initiate upstream of Q are processed rap-idly in vivo to produce the observed 5' ends associated witheach stable pi(foperon transcript. Moreover, the absence ofdetectable transcript derived from Q under chemoheterotro-phic growth conditions but the presence of substantialamounts of both pif-specific transcripts under these sameconditions would imply an additional requirement for thedifferential regulation of the mRNA processing functionunder aerobic versus photosynthetic growth (Lee and Kap-lan, unpublished results). To date, there is no physicalevidence linking the Q gene transcript to either of the plf'operon-specific transcripts.Recent genetic evidence in R. capsulatius demonstrates

that insertions into the Q gene result in a dramatic decreaseof Bchl or any precursors in the Mg branch of this pathway(Bauer and Marrs, Abstr. Mol. Biol. Photosynth. Procary-otes, 1987, p. 2). These and other results have led Bauer andMarrs to suggest that this gene product functions as a carrierfor Bchl intermediates, an activity previously suggested byLascelles to exist in R. s.phIaetoides (71, 96). In R. sphlaer-oitdes, deletion of the Q gene resulting in retention ofapproximately 25% of the proximal region of Q resulted inBchl and carotenoid synthesis at greatly reduced levelscompared with that of wild type (37). A second open readingframe designated R in R. sphaeroides maps upstream of theQ gene (Havelka and Kaplan, unpublished results), andcomplementation analysis of a mutant deficient in B875complexes (RS103; 61, 76) showed that this mutant could berestored to wild type with the R gene region in trans (J.Davis and S. Kaplan, unpublished results). Furthermore,complementation experiments have shown that this sameregion can restore the previously described pi,hA deletion/insertion mutant which is phenotypically B875- RC- toB875' when grown anaerobically in the dark on dimethylsulfoxide, independent of restoring functional RC complexes(Sockett and Kaplan, unpublished results). Since we knowthat the RS103 mutant is not affected in pi(f operon transcrip-tion (Davis and Kaplan, unpublished results), these dataimply that the R gene product (or a closely linked region)functions in an as yet uncharacterized role to control syn-thesis or assembly specifically of the B875 complexes andalso suggest that expression of the pbli operon or DNAsequences linked to plili can influence assembly of plufoperon gene products, perhaps by regulating expression ofthe R gene product.

Further complicating the analysis of the expression of theregion upstream of pi(IB is determination of the location ofthe promoter for the pif'operon. If it is just upstream of the5' ends of the stable pitfoperon transcripts (7, 140), then thepromoter would be located in the distal portion of thereading frame of Q. Similarly, the promoter for the Q genewould be within the distal portion of the R coding region.Therefore, if there are (is-acting effects due to the physicallinkage of these genes, it is difficult to distinguish betweenthese effects and transcriptional linkage of these genes withthe plif'operon. Recent results in R. sphleroides reveal thathlac tusions into pi,fB and extending from the distal portionof Q can express 3-galactosidase activity in vivo (Havelkaand Kaplan, unpublished results). Similarly, "runoff' tran-scription assays utilizing highly purified RNA polymerasefrom R. sphlaeroicles map transcription start sites for Q to thedistal region of R and tor pluf just downstream of Q (Kansyand Kaplan. in preparation). In vitro transcription-transla-tion experiments confirmed these results for sequencesrequired for expression of the piuf operon (66).

VOL. 52, 19X88

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

64 KILEY AND KAPLAN

Expression of the puc Operon

Of all of the genes specifying Bchl-binding proteins whichhave been analyzed to date (40, 67, 137-139), transcription ofthe piuc operon appears to be regulated to the greatest extentby changes in physiological conditions. Furthermore, muta-tions which affect the synthesis of the RC and B875 poly-peptides produce increased levels of puc mRNA underanaerobic respiratory conditions with dimethyl sulfoxide,although this is not necessarily accompanied by correspond-ing increases in the levels of B800-850 complexes (37). It isnot known whether derepression of plc mRNA levels inthese mutants is accompanied by increased synthesis ofcellular levels of unliganded B800-850 polypeptides. Thatone sees derepression of puic mRNA levels in such mutantstrains grown under anaerobic respiration in the dark sug-gests that factors other than light intensity can affect piucoperon expression at the transcriptional or posttranscrip-tional level (i.e., mRNA stability). Therefore, regulation ofpt(c operon expression by light intensity in the wild-typestrain could be an indirect effect and may be related moredirectly to one of the following: (i) movement of electronsthrough the electron transport chain, (ii) the redox state ofthe cell; (iii) the levels of piuf orplih gene products in the cell;(iv) the available pool of Bchl; (v) some as yet unknownregulatory precursor.

Posttranscriptional Control

As might have been expected from the analysis of mutantsdescribed above, there are transcriptional and translationaldifferences related to the expression of the structural genesencoding the polypeptides of the different spectral com-plexes. The steady-state levels of the mRNA encoding manyof the photosynthetic genes have been reported from avariety of physiological conditions (40, 67, 137-139). Forexample, the two stable (1,400 and 1,130 nt) piihA (40),plufBA and pi(fBALMX (139) transcripts are readily detect-able in R. spihaeroides grown chemoheterotrophically in a30% 02 atmosphere, whereas pluc (640 nt) mRNA is barelyapparent (67). As mentioned earlier, the pul/iA gene product(RC-H polypeptide) is present in the CM when grown underhigh 02 conditions, although there are no detectable spectralcomplexes (16, 40). Moreover, no RC-M, RC-L (16), B875-ox(67a), or B800-850-3 (16) polypeptides could be detected in aWestern blot (immunoblot) analysis of either soluble ormembrane fractions derived from R. sphaeroides grown in a30% 02 atmosphere. This implicates physiologically regu-lated translational or posttranslational control over the syn-thesis of individual ICM proteins in addition to differenttranscriptional effectors being involved in regulating thesteady-state level of these mRNAs under different physio-logical conditions. One possible explanation for the lack ofthe pigment-binding polypeptides under chemoheterotrophicgrowth conditions is due to instability of the Bchl-bindinggene products in the absence of their ligand, Bchl. Supportfor this model comes from recent pulse-labeling experimentsin R. capsulatus which have demonstrated the existence of atransient pool of pigment-binding polypeptides under condi-tions of induction for ICM synthesis in either Bchl- mutantsor in wild-type cells in the presence of inhibitors of porphy-rin biosynthesis (39, 70). In R. spliaieroides. the lability ofboth the B800-850-3 (67) and B875-3 (Kiley, Ph.D. thesis)polypeptides synthesized in the absence of Bchl in a coupledin vitro transcription-translation system has been demon-strated. Both observations are consistent with a model in

which synthesis of these polypeptides is also controlledposttranslationally, perhaps as the result of the action of aspecific protease which acts on the unliganded pigment-binding polypeptides. However, it still remains to be deter-mined whether pif specific mRNA is translated in vivounder chemoheterotrophic conditions. Although these datasuggest a posttranslational mechanism for the regulation ofthese gene products in the absence of Bchl, additional formsof translational control (e.g., Bchl-dependent translation,ribosome stalling) over photosynthetic gene expression can-not be ruled out.There is at least three- to sixfold more pufJ and phll-

specific mRNA in photosynthetically grown (100 W/m2) R.spihaeroides than in chemoheterotrophically grown cells (40,139). In contrast, the amount of plic-specific mRNA is atleast 100-fold greater in photosynthetically grown cells thanmeasured in chemoheterotrophically grown cells in the pres-ence of 30% 02 (Lee and Kaplan, unpublished results). Theamount of each of these transcripts, in steady-state photo-synthetically grown cells, varies two- to fivefold dependingon the incident light intensity of growth. In general, changesin plc, plt'. and pluli operon-specific mRNAs in wild-typecells grown under steady-state photosynthetic conditionsclosely paralleled the amount of each cognate spectralcomplex present within the ICM (Table 1). However, morework is required before we understand the precise physio-logical relationship between mRNA levels and the cellularabundance of specific spectral complexes in both wild-typeand mutant strains. It is not known whether the regulation ofan individual spectral complex (RC, B875, or B800-850)under photosynthetic conditions at different light intensitiesis due solely to changes in transcription of these operons orwhether the mRNAs observed in wild-type or mutant strains(see above) are due to a combination of transcription initia-tion and differential message stability.

CONCLUSIONS

The recent identification and sequencing of the structuralgenes for the Bchl-binding proteins and components of theelectron transport chain in the photosynthetic bacteria havebeen preceded by a strong foundation in the biochemical,spectroscopic, and genetic analyses of these organisms.These recent accomplishments provide the framework toprobe structure-function relationships of the above-de-scribed membrane complexes, especially when three-dimen-sional crystal structures are available. Use of moleculargenetic techniques in our analysis of the photosyntheticbacteria also has the potential to unravel the complexmolecular mechanisms by which these organisms sense andrespond to environmental stimuli such as oxygen and light.This review highlights the need to expand the pioneeringwork of Lascelles and co-workers on the pathway fortetrapyrrole biosynthesis in these bacteria. Without suchanalyses! it will be difficult, if not impossible, to completelyunderstand how the synthesis of ligands such as heme, Bchl,carotenoids. and bacteriopheophytin is coupled to synthesisor assembly or both of ICM apoproteins into functionalunits.

Further, assembly and function of individual ICM compo-nents cannot be separated from the synthesis of the mem-brane bilayer in which they reside. Many unanswered ques-tions still persist with regard to the processes or eventsinvolved in formation and partitioning of ICM invaginationsduring cell division in both steady-state cells or cells re-sponding to changes in environmental conditions. Although

MICROBIOL. REV.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 65

there is recent genetic evidence suggesting that specificproteins function in assembly of Bchl-protein complexes,much work is needed to precisely define their physiologicalrole(s), subcellular location, and mechanism of action. Therecent crystallization of RC complexes in Rhodopseludoino-nas viridis and R. sphaeroides offers the potential for crys-tallization of other membrane protein complexes which willallow us to understand how individual redox componentsfunction and interact with other such complexes in themembrane. It is perhaps not surprising that specific struc-tural or regulatory gene mutations have pleiotropic effects onthe structure, assembly, and regulation of the ICM since allof these components reside in the same lipid bilayer.

Still one of the most intriguing questions is how all of theseparticular components are targeted to the ICM to create thisspecific membranous domain. The fact that photoheterotro-phically growing cells contain three distinct membrane sys-tems (ICM, cell membrane, and outer membrane), each ofwhich possesses its own unique macromolecular composi-tion and structure, raises important questions. How doproteins unique to each membrane system find their correctlocation and how do proteins which are present in more thanone membrane system partition themselves between thesemembranes? The mode of photosynthetic membrane assem-bly also poses many questions with regards to the physicalstate of the bilayer and the function of specific proteins,lipids, or membrane domains in controlling activity or diffu-sion of components between the cell membrane and theICM.

Studies of the structure, function, and expression ofindividual components within the inducible photosyntheticapparatus of R. sphaeroides and in the photosynthetic bac-teria in general have reached a high degree of sophistication.Although not discussed here, events occurring at the level ofthe ICM must somehow be communicated to regulate theexpression of those soluble activities which we associatewith photosynthetic carbon assimilation (Calvin cycle en-zymes) and with nitrogen fixation. Given the rate of progressover the past several years, the next few years will prove tobe exciting, informative, and, doubtless, somewhat contro-versial. However, it is our hope that we do not lose site ofthe remarkable versatility and metabolic elegance of thephotosynthetic bacteria as models with which to addresssignificant problems in many areas of biological interest.

ACKNOWLEDGMENTS

We acknowledge the support provided by research grants fromthe Public Health Service National Institute of General MedicalSciences, National Science Foundation, U.S. Department of Agri-culture, and Amoco Corpn.We also acknowledge those individuals willing to share their

results with us and to include their data in this review. Finally, wethank those members of this laboratory. past and present. who havemade these studies a great deal more than just work.

ADDENDUM IN PROOF

Recent analysis of an R. sphaeroides mutant containing adeletion of the intercistronic region between pifA and pitfLshows reduced amounts of B875 (7%) and RC (25%) spectralcomplexes (R. Prince, personal communication) and similardecreases in the B875-o (5%) and RC (60%) polypeptides byWestern blot analysis (B. Dehoff and S. Kaplan, unpublishedresults). The physiological consequence of this deletion isthat the ratio of B875 to RC complexes has been decreasedto approximately 1-2:1. This appears to result from altered

stoichiometry of the small to large pl.f operon mRNA,perhaps due to a defect in transcription termination betweenpi,fA and piufL (see text). Recently, Klug et al. have alsoshown that deletion of the analogous intercistronic region inR. capsiulatuis also affects the stoichiometry of the piifoperon gene products, and they conclude that this supports3' processing of the large pluf operon mRNA as the origin ofthe small transcript rather than transcription termination(67b). Such strains will provide the basis for studying themechanism underlying the stoichiometric synthesis of pufoperon gene products (mRNA and polypeptides), as well asthe effect that altered ratios of these Bchl-protein complexeshave on the light reactions of photosynthesis.The region upstream of plufB has been altered by site-

directed mutagenesis such that the putative o)fK gene prod-uct was fused in frame to the piufB protein. The predictedfused gene product can be detected in the R. sphaeroides invitro transcription-translation system, suggesting that o)fKhas the proper signals required for translation. The wild-typepiufB gene product was also synthesized in great molarexcess over the fused gene product in vitro, indicating thatthere could be differences in translation initiation at the oijKand piufB ribosome-binding site. Thus, it is possible thattranslation of the o,fK region can control transcriptiontermination between piujA and pullf (see text). Analysis ofthe physiological effects of mutations in orfK in vivo will aidin addressing the role of the oiJK gene on pl,f operonexpression and synthesis of the fixed photosynthetic unit.

LITERATURE CITED

1. Aagaard, J., and W. R. Sistrom. 1972. Control of synthesis ofreaction center bacteriochlorophyll in photosynthetic bacteria.Photochem. Photobiol. 15:209-225.

2. Agalidis, I., A. M. Nuijs, and F. Reiss-Husson. 1987. Charac-terization of an LM unit purified by affinity chromatographyfrom Rlhodobacter sphlaeroides reaction centers and interac-tion with the H subunit. Biochim. Biophys. Acta 890:242-250.

3. Allen, J. P., G. Feher, T. 0. Yeates, H. Komiya, and D. C.Rees. 1987. Structure of the reaction center from Rhodobacterspliaeroides R-26: the protein subunits. Proc. NatI. Acad. Sci.USA 84:6162-6166.

4. Allen, J. P., G. Feher, T. 0. Yeates, D. C. Rees, J. Deisenhofer,H. Michel, and R. Huber. 1986. Structural homology of reac-tion centers from Rlhodopseludomonas splhaeroides and Rho-dIopseidofllonas liridis as determined by X-ray diffraction.Proc. Natl. Acad. Sci. USA 83:8589-8593.

5. Ashby, M. K., S. A. Coomber, and C. N. Hunter. 1987.Cloning. nucleotide sequence, and transfer of genes for theB800-850 light-harvesting complex of Rhodobacter sphaer-oides. FEBS Lett. 213:245-248.

6. Bauer, C. E., M. Eleuterio, D. A. Young, and B. L. Marrs.1987. Analysis of transcription through the Rhodobacter cap-silautis plif'operon using a translational fusion of pi(fM to theE. coli lacZ gene. p. 699-705. In J. Biggens (ed.), Progress inphotosynthesis research, vol. 4. Martinus Nijhoff Publishers,Dordrecht, The Netherlands.

7. Belasco, J. G., J. T. Beatty, C. W. Adams, A. V. Gabain, andS. N. Cohen. 1985. Differential expression of photosynthesisgenes in R. capsldata results from segmental differences instability within the polycistronic rxcA transcript. Cell 40:171-181.

8. Berard, J., G. Bldanger, P. Corriveau, and G. Gingras. 1986.Molecular cloning and sequence of the B880 holochrome genesfrom Rliodospir-illiuni rlib7-uiru. J. Biol. Chem. 261:82-87.

9. Biel, A. J., and B. L. Marrs. 1983. Transcriptional regulation ofseveral genes for bacteriochlorophyll biosynthesis in Rlhodo-pseudlomlona1o1(s capsdlata in response to oxygen. J. Bacteriol.156:686-694.

10. Broglie, R. M., C. N. Hunter, P. Delepelaire, R. A. Niederman,

VOL. 52, 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

66 KILEY AND KAPLAN

N.-H. Chua, and R. K. Clayton. 1980. Isolation and character-ization of the pigment-protein complexes of Rhodopseudotno-na.5 spliaemoidtes by lithium dodecylsulftite/polyaicrylamide gelelectrophoresis. Proc. Natl. Acad. Sci. USA 77:87-91.

11. Brown, A. E., F. A. Eiserling, and J. Lascelles. 1972. B3acte-riochlorophyll synthesis and the ultratstructure of wild type andmutant strains of Rliodopseulotnona. sphaeroides PlantPhysiol. 50:743-746.

12. Brown, A. E., and J. Lascelles. 1972. Phytol and bacteriochlo-rophyll synthesis in Rhodopseudomonas spIlaeroid(s. PlantPhysiol. 50:747-749.

13. Brunisholz, R. A., H. Zuber, J. Valentine, J. G. Lindsay, K. J.Wooley, and R. J. Cogdell. 1986. The membrane location of theB890-complex from Rlodiospirilliniti rulhbrulm(tand the effect ofcarotenoid on the conformation of its two apoproteins exposedat the cytoplasmic surface. Biochim. Biophys. Acta 849:295-303.

14. Cain, B. D., C. Deal, R. 1. Fraley, and S. Kaplan. 1981. In vivointermembrane transfer of phospholipids in the photosyntheticbacterium Rhodlopseiotiotnis splhaeroides. J. Bacteriol. 145:1154-1166.

15. Cain, B. D., T. J. Donohue, W. D. Shepherd, and S. Kaplan.1984. Localization of phospholipid biosynthetic enzyme activ-ities in cell-free fractions derived from Rhodopseuidomonas.sphiaeroides. J. Biol. Chem. 259:942-948.

16. Chory, J., T. J. Donohue, A. R. Varga, L. A. Staehelin, and S.Kaplan. 1984. Induction of the photosynthetic membraines ofRhodopseudomonas sphaeroidle,s: biochemical and morpholog-ical studies. J. BIacteriol. 159:54(t-554.

17. Chory, J., and S. Kaplan. 1982. The iti litro transcription-translation of DNA and RNA templates by extracts of Rhiodto-p.seIud)tomnas spli(iwroidlc.s. J. Biol. Chem. 257:1 511(0-1 5121.

18. Chory, J., and S. Kaplan. 1983. Light-dependent reguliation ofthe synthesis of soluble and intracytoplasmic membrane pro-teins of Rhlodopseudlotmiotnuis spl(eroidlcs. J. Baicteriol. 153:465-474.

19. Chory, J., E. D. Muller, and S. Kaplan. 1985. DNA-directed invitro synthesis and assembly of the form 11 D-ribulose-1.5-bisphosphalte carboxylase/oxygenase from Rhodopscudoono-tiis sphl(aroidCs. J. Bacteriol. 161:307-313.

20. Clark, W. G., E. Davidson, and B. L. Marrs. 1984. Variation oflevels of mRNA coding for antenna and reaction center poly-peptides in Rliodopseidlottiomonis (apsldaIta in r-esponse tochanges in oxygen concentration. J. Bacteriol. 157:945-948.

21. Clayton, R. K., and B. J. Clayton. 1978. Properties of photo-chemical reaction centers purified from Rhodopsetidoinonasgelatinosa. Biochim. Biophys. Acta 501:470-477.

22. Clement-Metral, J. D. 1979. Activation of Al-A synthase byreduced thioredoxin in Rlhodopseidomlotnais sphitieroidces Y.FEBS Lett. 101:116-120.

23. Clement-Metral, J. D., J.-O. Hoog, and A. Holmgren. 1986.Characterization of the thioredoxin system in the faicultativephototroph Rhodobacter splhaceoides Y. Eur. J. Biochem.161:119-126.

24. Cogdell, R. J., and A. R. Crofts. 1978. Analysis of the pigmentprotein content of an antenna pigment-protein complex fromthree sti-rains of Rh0odlopsNC1udoitnoa(Ns sphaeroide.s. Biochim.Biophys. Acta 502:409-416.

25. Cogdell, R. J., and J. P. Thornber. 1980. Light-harvestingpigment-protein complexes of put-ple photosynthetic bacteria.FEBS Lett. 122:1-8.

26. Cohen, L., D. Lueking, and S. Kaplan. 1979. Intermemib-anephospholipid transfer mediated by cell-free extracts of Rhodo-pseudomnonas sphaeroidels. J. Biol. Chem. 254:721-728.

27. Cohen, L. K., and S. Kaplan. 1981. The non-detergent solubi-lization and isolation of intracytoplasmic membra-ane polypep-tides from Rhodopseudomona.s splh(aroi(dcs. J. Biol. Chem.256:5901-5908.

28. Cohen, L. K., and S. Kaplan. 1981. Characterization of thethree matjor intracytoplasmic membrane polypeptidies isolatedfrom Rhodopseudoinonias splhaeroides. J. Biol. Chem. 256:5909-5915-.

29. Cohen-Bazire, G., W. R. Sistrom, and R. Y. Stanier. 1957.

Kinetic studies of pigmlient synthesis by non-suLItur purplebacteria. J. C'ell. Comp. Phvsiol. 49:2548.

30. Coomber, S. A., NM. F. Ashby, and C. N. Hunter. 1987. Cloningand oxygen regulated expression of genes for the bacteriochlo-rophyll biosynthetic pathway in Rhodopseludonotna .spha(r-oides. p. 737-74(0. In J. Biggens (ed.), Progress in photosyn-thesis research. vol. 4. Martinus NihotYf Publishers.D)ordrecht. The Netherlands.

31. Crofts, A. R., S. W. Meinhardt, and J. R. Bowyer. 1982. Theelectron transport chain of Rhodop.suldoinonas .sphroid(e.s.p. 477-498. In 13. L. Trumpower (ed.), Function of quinones inenergy conserving systems. Academic Press, Inc., Ncw York.

32. Crofts, A. R., and C. A. Wraight. 1983. The electrochemicaldomain of photosynthesis. l3iochim. Biophys. Acta 726:149-185.

33. Daldal, F., S. Cheng, J. Applebaum, E. Davidson, and R.Prince. 1986. Cvtochromne c, is not essential for photosyntheticgrowth of Rho(op.seu(otnotiuas (capsula(ta. Proc. Natl. Acad.Sci. USA 83:2012-2016.

34. Davidson, E., and R. J. Cogdell. 1981. The polypeptide com-position of the B850 light-har-vesting pigment-protein complexfrom Rhiodlopseudotno0as Npitleroide.s R26. 1. FEBS Lett. 132:81-84.

35. Davidson, E., and F. D)aldal. 1987. Primary structur-e ot the bc,complex of Rhodlopseu(otinona.vs (capsII(lta- Nucleotide se-quence of the p)t operon encoding the Rieske. cytochr-ome b,aind cytochrome b/c1 apoproteins. J. Mol. Biol. 195:13-24.

36. Davidson, E., and F. I)aldal. 1987. fhb operion, encoding theRieske Fe-S protein. cytochrome b. and cvtochromic cl apo-proteins previously descr-ibcd tIroim Rhodop.s%eiluotionas.s%plhairoides is from Rliohojd)seudlononas ((J)sul(at(. J. Mol.B3iol. 195:25-29.

37. Davis, J., T. J. I)onohue, and S. Kaplan. 1988. ConstruCtion.characterization, ani complementation of a P'uf mutant otRhzodobatuer(t .spit/oitles. J. Bacteriol. 170:320-(329.

38. Deisenhofer, J., 0. Epp, K. Nliki, R. Huber, and H. Michel.1985. Structure of the protein subunits in the photosyntheticreaction centre of Ro(iopdolcudootntia.s 6irii.&s at 3A resolution.Nature (London) 318:619-624.

39. Dierstein, R. 1984. Synthesis of pigment-hinding protein intoluene-treated Rlhodop.seiononass capsiulita and in cell-freesvstems. Eur. J. l3iochem. 138:50)9-518.

40). Donohue, T. J., J. H. Hoger, and S. Kaplan. 1986. Cloning andexpression of the Rhodobacicr spihaeroitles reiaction center Hgene. J. Bacteriol. 168:953-961.

41. Donohue, T. J., and S. Kaplan. 1986. Synthesis and assemblyof bacterial photosynthetic membranes. p. 6324639. In L. A.Stachelin and C. J. Arnzten (edl.) Photosynthesis Ill: photo-synthetic membranes. Encyclopedia ot plant physiology, newseries, vol. 19. Springer-Verlag. New York.

42. Donohue, 1. ,J., A. G. McEwan, and S. Kaplan. 1986. Cloning.DNA sequLence. and expression of the Rlholodahlter sphaer-oi(ldes cytochrome c gene. J. Bacteriol. 168:962-972.

43. Drews, G. 1986. Adaptation of the bacterial photosyntheticapparatus to diftlerent light intensities. Trends. Biochem. Sci.1:2'55-257.

44. Drews, G., J. Peters, and R. Dierstein. 1983. Molecular organ-ization and biosynthesis of pigment-protein complexes of Rlio-lop.uhmiona.ls (laps.vulata. Ann. Microbiol. ( Paris) 134B:151-

45. Drews, (., and R. Feick. 1978. Isoliation and characterizationof light-harvesting bacteriochlorophyll protein complexes fromRhodop.s-eudotnona,s caIp.slbata. Biochim. Biophys. Acta 501:499-513.

46. I)us, K., T. Flatmark, H. deKlerk, and NI. Kamen. 197(0.Isolation and chemical properties of two c-type cytochromes ofRhodo.spirillni noloIi.% (Ihialnml. Biochemistry 9:1984-1990.

47. Fanica-Gaigner, NI., and J. Clement-Metral. 1973. Cellularcompartmentation of two species ot 6-aminolevulinic acidlsvnthetase in a flacuLatative photoheter-otrophic bacterium Rp.s.ApIM(lerOi(Ies V. B3iochenm. Biophys. Res. Communn . 55:61(0-15.

48. Feher, (., and Ni. Y. Okamura. 1975. Chemnical compositionand properties ot reaiction centers. p. 349-386. In R. K.

MICROBIOl(l. RFv.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 67

Clayton and W. R. Sistrom (ed.). The photosynthetic bacteria.Plenum Publishing Corp., New York.

49. Flatmark, T., K. Dus, H. deKlerk, and M. Kamen. 1970.Comparative study of physiochemical properties of two c-typecytochromes of Rhodospirilliin inolischianuln. Biochemistry9: 1991-1995.

50. Fornari, C. S., and S. Kaplan. 1982. Genetic transformation inRhodopseldoinonas sphlaeroides by plasmid DNA. J. Bacte-riol. 152:89-97.

51. Forrest, M. E., and J. T. Beatty. 1987. Purification of RNApolymerase and its use for in vitro transcription. FEBS Lett.212:28-34.

52. Fraker, P., and S. Kaplan. 1971. Isolation and fractionation ofthe photosynthetic membrane organelles from Rlhodopseudo-inonas splhaeroides. J. Bacteriol. 108:465-473.

53. Fraley, R. T., D. R. Leuking, and S. Kaplan. 1977. Intracyto-plasmic membrane synthesis in synchronous cell populationsof Rhodopseiudornonias sphaeiroides: polypeptide insertion intogrowing membrane. J. Biol. Chem. 253:458-464.

54. Fraley, R. T., G.-S. L. Yen, D. R. Leuking, and S. Kaplan.1979. The physical state of the intracytoplasmic membrane ofRhodopseludoinonats spliaeroides and its relationship to the celldivision cycle. J. Biol. Chem. 254:1987-1991.

55. Gabellini, N., J. R. Bowyer, E. Hurt, B. A. Melandri, and G.Hauska. 1982. A cytochrome bc1 complex with ubiquinol-cytochrome c. oxidoreductase activity from Rlhodopseudoilno-nas spliaeroides. Eur. J. Biochem. 126:105-111.

56. Gabellini, N., V. Harnisch, J. E. G. McCarthy, G. Hauska, andW. Sebald. 1985. Cloning and expression of the thc operonencoding the FeS protein, cytochrome b and cytochrome clfrom the Rliodopseludoinona(is spliieroides b/c1 complex.EMBO J. 4:549-553.

57. Hoger, J., J. C. Chory, and S. Kaplan. 1986. In vitro biosyn-thesis and membrane association of photosynthetic reactioncenter subunits from Rhodopseudolnonas spliaeroides. J. Bac-teriol. 165:942-950.

58. Hoger, J. H., S.-P. Tai, and S. Kaplan. 1987. Membraneadenosine triphosphatase in synchronous cultures of Rliodo-bacter sphaeroides. Biochim. Biophys. Acta 898:70-80.

59. Imhoff, J. F., H. G. Truper, and N. Pfennig. 1984. Rearrange-ment of the species and genera of the phototrophic "purplenonsulphur bacteria." Int. J. Syst. Bacteriol. 34:340-343.

60. Inamine, G. S., J. van Houten, and R. A. Niederman. 1984.Intracellular localization of photosynthetic membrane growthinitiation sites in Rlhodopseludotnonas sphlaer-oides. J. Bacte-riol. 158:425-429.

61. Jackson, W. J., P. J. Kiley, C. E. Haith, S. Kaplan, and R. C.Prince. On the role of the light-harvesting B880 in the correctinsertion of the reaction center of Rhodobacter capsdalutis andRhodohacter sphlaeroides. FEBS Lett. 215:171-174.

62. Jones, 0. T. G. 1978. Biosynthesis of porphyrins. hemes. andchlorophylls. p. 751-777. In R. K. Clayton and W. R. Sistrom(ed.), The photosynthetic bacteria. Plenum Publishing Corp.,New York.

63. Kaplan, S. 1981. Development of the membranes of photosyn-thetic bacteria. Photochem. Photobiol. 34:769-774.

64. Kaplan, S., and C. J. Arntzen. 1982. Photosynthetic membranestructure and function. p. 65-151. In Govindjee (ed.). Photo-synthesis: energy conversion by plants and bacteria. vol. 1.Academic Press, Inc.. New York.

65. Kaplan, S., B. D. Cain, T. J. Donohue, W. D. Shepherd, andG. S. L. Yen. 1983. Biosynthesis of the photosynthetic mem-branes of Rliodopsecdo,nonas spliaeroides. J. Cell. Biochem.22: 15-29.

66. Kiley, P. J., T. J. Donohue, W. A. Havelka, and S. Kaplan.1986. DNA sequence and in vitro expression of the B875light-harvesting polypeptides of Rhodobacter spliaiecroidces. J.Bacteriol. 169:742-750.

67. Kiley, P. J., and S. Kaplan. 1987. Cloning, DNA sequence. andexpression of the Rhodohbacter spli(wrtoidcs light-harvestingB800-850-a and B800-850-P genes. J. Bacter-iol. 169:3268-3275.

67a.Kiley, P. J., A. Varga, and S. Kaplan. 1988. Physiological andstructural analysis of light-harvesting mutants of Rhodobacter

spihieoides. J. Bacteriol. 170:1103-1115.67b.Klug, G., C. W. Adams, J. Belasco, B. Doerge, and S. N.

Cohen. 1987. Biological consequences of segmental alterationsin mRNA stability: effects of deletion of the intercistronichairpin loop region of the Rhodobacter capsulatius pfoperon.EMBO J. 6:3515-3520.

68. Klug, G., and G. Drews. 1984. Construction of a gene bank ofRhodopseudolnionas (apsulata using a broad host DNA clon-ing system. Arch. Microbiol. 139:319-325.

69. Klug, G., N. Kaufmann, and G. Drews. 1985. Gene expressionof pigment binding proteins of the bacterial photosyntheticapparatus: transcription and assembly in the membrane ofRlhodopselidotinonas (apsulata. Proc. Natl. Acad. Sci. USA82:6485-6489.

70. Klug, G., R. Liebetanz, and G. Drews. 1986. The influence ofbacteriochlorophyll biosynthesis on formation of pigment-binding proteins and assembly of pigment protein complexes inRliodopseludoliinotas capsilata. Arch. Microbiol. 146:284-291.

71. Lascelles, J. 1978. Regulation of pyrrole synthesis, p. 795-808.In R. K. Clayton and W. R. Sistrom (ed.), The photosyntheticbacteria. Plenum Publishing Corp., New York.

72. Leong, S. A., G. S. Ditta, and D. R. Helinski. 1982. Hemebiosynthesis in Rliiobiuin. Identification of a cloned genecoding for 6-aminolevulinic acid synthetase from Rhizobilinmtie/iloti. J. Biol. Chem. 257:8724-8730.

73. Lueking, D. R., R. T. Fraley, and S. Kaplan. 1978. Intracyto-plasmic membrane synthesis in synchronous cell populationsof Rliodopseiudomtiionas spihaeroides. J. Biol. Chem. 253:451-457.

74. Marrs, B. 1974. Genetic recombination in Rhodopseiudomonascapsidata. Proc. NatI. Acad. Sci. USA 71:971-973.

75. Marrs, B. 1981. Mobilization of the genes for photosynthesisfrom Rlhodopseuidomnonas (apsldata by a promiscuous plas-mid. J. Bacteriol. 146:1003-1012.

76. Meinhardt, S. W., P. J. Kiley, S. Kaplan, A. R. Crofts, and S.Harayama. 1985. Characterization of light-harvesting mutantsof Rliodopseiudoliiomonas spihieroides. 1. Measurement of theefficiency of energy transfer from light-harvesting complexesto the reaction center. Arch. Biochem. Biophys. 236:130-139.

77. Meyer, T. E., and M. D. Kamen. 1982. New perspectives onc-type cytochromes. Adv. Protein Chem. 35:105-212.

78. Michel, H., 0. Epp., and J. Deisenhofer. 1986. Pigment-proteininteractions in the photosynthetic reaction centre from Rliodo-pseudoinonas viridis. EMBO J. 5:2445-2451.

79. Michel, H., K. A. Weyer, H. Gruenberg, I. Donger, D. Oseter-helt, and F. Lottspeich. 1986. The "light" and "medium"subunits of the photosynthetic reaction centre from Rhodo-pseldomlnonas viridis: isolation of the genes, nucleotide andamino acid sequence. EMBO J. 5:1149-1158.

80. Michel, H., K. A. Weyer, H. Gruenberg, and F. Lottspeich.1985. The 'heavy' subunit of the photosynthetic reactioncentre from Rliodopseludomnionas v,iridis: isolation of the gene,nucleotide and amino acid sequence. EMBO J. 4:1667-1672.

81. Miller, L., and S. Kaplan. 1978. Plasmid transfer and expres-sion in Rhodopseudomnonas spliaier-oides. Arch. Biochem.Biophys. 187:229-234.

82. Monger, T. G., and W. W. Parson. 1977. Singlet-triplet fusionin Rliodopseidclomnonas spihaeroides chromatophores: a probeof the organization of the photosynthetic apparatus. Biochim.Biophys. Acta 460:393-407.

83. Muller, E. D., J. Chory, and S. Kaplan. 1985. Cloning andcharacterization of the gene product of the form 11 ribulose-1.5-bisphosphate carboxylase gene of Rhodopseudornonasspihaeroides. J. Bacteriol. 161:469-472.

84. Nano, F. E., and S. Kaplan. 1984. Plasmid rearrangements inthe photosynthetic bacterium Rliodopseiudomnonias sphaer-oidles. J. Bacteriol. 158:1094-110)3.

85. Nano, F. E., W. D. Shepherd, M. M. Watkins, S. A. Kuhl, andS. Kaplan. 1985. Broad-host range plasmid vector for the invitro construction of transcriptional/translational lac fusions.Gene 34:219-226.

86. Oelze, J. 1983. Control of the formation of bacteriochlorophylland B875- and B85()-bacteriochlorophyll complexes in Rliodo-

VOL. 52, 1988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

68 KILEY AND KAPLAN

pseiudotinonas splhaierloides mutant strain H5. Arch. Microbiol.136:3112-316.

87. Parson, W. W., and D. Holten. 1986. Primary electron transfer-reactions in photosynthetic bacteria: energetics and kinetics oftransient states. p. 338-343. In L. A. Staehelin and C. J.Arntzen (ed.), Photosynthesis III: photosynthetic membranes.Encyclopedia of plant physiology, new series, vol. 19. Spring-er-Verlag, New York.

88. Pemberton, J. M., and C. M. Harding. 1986. Cloning ofcarotenoid biosynthesis genes from Rhodopselidoinonaissphaeroides. Curr. Microbiol. 14:25-29.

89. Pemberton, J. M., and C. M. Harding. 1987. Expression ofRhodopseudoinonas sphaeroides carotenoid photopigmentgenes in phylogenetically related nonphotosynthetic bacteria.Curr. Microbiol. 15:67-71.

90. Peters, G. A., and R. A. Cellarius. 1972. Photosyntheticmembrane development in Rhodopseludomontas splhaer-oides.11. Correlation of pigment incorporation with morphologicalaspects of thylakoid formation. Bioenergetics 3:345-359.

91. Peters, J., J. Takemoto, and G. Drews. 1983. Spatial relation-ships between the photochemical reaction center and thelight-harvesting complexes in the membrane of Rlhodopseuldlo-iionas (capsidala. Biochemistry 22:566(-5667.

92. Peters, J. D., and G. Drews. 1983. The transverse orientation ofthe light-harvesting and reaction centre polypeptides of Rhlo-dopseudomtionais capsilata investigated by surface iodination.FEBS Lett. 162:57-60.

93. Pfennig, N., and H. G. Truper. 1974. The phototrophic bacte-ria, p. 2-75. In R. E. Buchanan and N. E. Gibbons (ed.).Bergey's manual of determinative bacteriology, 8th ed. TheWilliams & Wilkins Co., Baltimore.

94. Prince, R. C., A. Baccarini-Melandri, G. A. Hauska, B. A.Melandri, and A. R. Crofts. 1975. Asymmetry of an energytransducing membrane: the location of cytochrome c. in Rlho-dopseuidoitnona.s sphaeroides and Rlhodopseludoinonal s ((Ipsil-

ltia. Biochim. Biophys. Acta 387:212-227.95. Prince, R. C., E. Davidson, C. E. Haith, and F. Daldal. 1986.

Photosynthetic electron transfer in the absence of cytochromec, in Rhodopseudotnonas (apsidata: cytochrome c. is notessential for electron flow from cytochrome bc, complex to thephotochemical reaction center. Biochemistry 25:5208-5214.

96. Rebeiz, C. A., and J. Lascelles. 1982. Biosynthesis of pigmentsin plants and bacteria, p. 699-780). In Govindjee (ed.). Photo-synthesis: energy conversion by plants and bacterita, vol. 1.Academic Press, Inc., New York.

97. Ronson, C. W., B. T. Nixon, and F. M. Ausubel. 1987.Conserved domains in bacterial regulatory proteins that re-

spond to environmental stimuli. Cell 49:579-581.98. Sauer, K. 1986. Photosynthetic light reactions-physical as-

pects, p. 85-96. In L. A. Staehelin and C. J. Arntzen (ed.).Photosynthesis IlI: photosynthetic membranes. Encyclopediaof plant physiology. new series. vol. 19. Springer-Verlag, NewYork.

99. Schultz, J. E., and P. F. Weaver. 1982. Fermentation and

anaerobic respiration by Rhodospirillhni ridbruml and Rliodo-pseud)ononas (apsld(ta. J. Bacteriol. 149:181-19t).

99a.Scolnik, P. A., and B. L. Marrs. 1987. Genetic research withphotosynthetic bacteria. Annu. Rev. Microbiol. 41:7t)3-726.

1t)0. Scolnik, P. A., NI. A. Walker, and B. L. Marrs. 1980. Biosyn-

thesis of carotenoids derived from neurosporene in Rlio(lo-pseudomnonas capslatma. J. Biol. Chem. 255:2427-2432.

101. Shepherd, W. D., and S. Kaplan. 1978. Changes in Rhodopseu-domnonas spllaeroidles isoaccepting phenyalanyl-tRNA species

during transitions from chemoheterotrophic to photoheter-otro-phic growth. Arch. Microbiol. 116:161-167.

1(02. Shepherd, W. D., and S. Kaplan. 1983. Effect of cer-ulenin on

macromolecule synthesis in chemoheterotrophically and pho-toheterotrophicallIy gr-own Rhodopseudonionas sPhaeroides. JBacteriol. 156:1322-1331.

103. Shill, D. A., and P. NI. Wood. 1984. A role for cvtochrome c.in Rhodopseudomnonas l iridis. Biochim. Biophvys. Acta 764:1-7.

104. Simon, R., V. Priefer, and A. Puhler. 1983. A broad host range

mobilization system for in iiio genetic cngineering: transposon

mutagenesis in Gram-negative bacteria. Bio/Technology 1:37-45.

105. Sistrom, W. R. 1975. List of mutant strains, p. 927-934. InR. K. Clayton and W. R. Sistrom (ed.). The photosyntheticbacteria. Plenum Publishing Corp., New York.

106. Sistrom, W. R., A. Macalusa, and R. Pledger. 1984. Mutants ofRliodopseudoltotiotiis sphaeroides useful in genetic anal ysis.Arch. Microbiol. 138:161-165.

107. Snozzi, M., and A. R. Crofts. 1984. Electron transport inchromatophores from Rhodopseiudottionas spliacroides Gafused with liposomes. Biochim. Biophys. Acta 766:451-463.

108. St. G. Bowen, A. R., and J. M. Pemberton. 1985. Mapping ofthe genome of Rhlodopseiudotmotnoas splae'roides, p. 885. InD. R. Helinski. S. N. Cohen, D. B. Clewell, D. A. Jackson,and A. Hollaender (ed.), Plasmids. Plenum Publishing Corp..New York.

109. Stark, W., F. Jay, and K. Muehlethaler. 1986. Localisation ofreaction centre and light harvesting complexes in the photo-synthetic unit of Rhlodopseiutlotinonais lvirid1is. Arch. Microbiol.146:130-133.

110. Tadros, M. H., R. Frank, and G. Drews. 1986. Localization ofthe exposed N-terminal region of the B8t)0-850-o. and -flight-harvesting polypeptides on the cytoplasmic surface ofRhodop.seldomnonas ((psiulata chromatophores. J. Bacteriol.367:96-100.

111. Tai, S.-P., J. H. Hoger, and S. Kaplan. 1986. Phospholipidtrcansfer activity in synchronous populations of RhodobacterspIl(leroides. Biochim. Biophys. Acta 859:198-208.

112. 'Tai, S.-P., and S. Kaplan. 1984. Purification and properties ofa phospholipid transfer protein from Rhodop.swudoinona,ssplal(eroides. J. Biol. Chem. 259:12178-12183.

113. Takemoto, J., and J. Lascelles. 1973. Coupling between bacte-riochlorophyll and membrane protein synthesis in Rliodtopsei-doinonas spliaelroides. Proc. Natl. Acad. Sci. USA 70:799-8t)3.

114. Taylor, D. P., S. N. (Cohen, W. G. Clark, and B. L. Marrs.1983. Alignment of genetic and restriction maps of the photo-synthesis region of the RhodopseltdoiniontIs C(apsid1aw chromo-some by a conjugation-mediated marker rescue technique. J.Bacteriol. 154:580-590.

115. Theiler, R., F. Suter, J. D. Pennoyer, H. Zuber, and R. A.Niederman. 1985. Complete amino acidi sequence of the B875light-harvesting protein of Rliodtopseiudomtionitis sphaeroidesstrain 2.4.1. FEBS Lett. 184:231-236.

116. Theiler, R., F. Suter, V. Wiemken, and H. Zuber. 1984. Thelight-hairvesting polypeptides of Rhotlopseudotnonas sphiaer-oide.s R-26.1 1. Isolation, purification and sequence analysis.Hoppe-Seylet's Z. Phvsiol. Chem. 365:703-719.

117. Theiler, R., F. Suter, H. Zuber, and R. J. Cogdell. 1984. Acomparison of the primary structures of the two B800-850-apoproteins from wild-type Rhodopseudoinonas s phaeroide.sstrain 2.4.1 and a carotenoi(iless mutant strain R26.1. FEBSLett. 175:231-237.

118. 'I'vbulewicz, V. L. J., G. Falk, and J. E. Walker. 1984.Rhodop.seudonona.s 1)I(sti(a(lIt) operon: nucleotide sequenceand transcription. J. Mol. Biol. 179:185-214.

119. Weaver, K. E., and R. B. Tabita. 1985. Complementation of a

Rlhodopseldomonas sphaeroidels ribulose bisphosphate car-boxylase-oxygenase regulatory mutant trom a genomic library.J. Bacteriol. 164:147-154.

120. Wette, W., T. Wacker, M. Leis, W. Kreutz, J. Shiozawa, N.Gad'on, and G. I)rews. 1985. Crystallization of the photosyn-thetic light-harvesting pigment-protein complex B800-850 ofRhodopseudotnona. (capslalta. FEBS Lett. 182:260-263.

121. WXever, K. A., W. Schafer, F. Lottspeich, and H. Michel. 1987.The cytochrome subunit of the photosynthetic reaction centerfrom Rhodopseuldolmionas riiidis is a lipoprotein. Biochemistry26:2909-2914.

122. WN'illiams, ,l. C., L. A. Steiner, and G. Feher. 1986. Primarystructul-e of the reaction center from RhodopseudotnonavsNsplia(coisdes. Proteins 1:312-325.

123. Williams, J. C., L. A. Steiner, G. Feher, and M. I. Simon. 1984.Primall-y stl-uctule otf the L subunit of the reaction center of

MICROBIOL. REV.

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

PHOTOSYNTHETIC MEMBRANE BIOSYNTHESIS IN R. SPHAEROIDES 69

Rhlodopseutdotniotiis splicroicles. Proc. Ncatl. Acad. Sci. USA81:7303-7308.

124. Williams, J. C., L. A. Steiner, R. C. Odgen, M. I. Simon, andG. Feher. 1983. Primary structure of the M subunit of thereaction center f-om Rhlodopseldomnonas sphiceroides. Proc.Natl. Acad. Sci. USA 80:6505-6509.

125. Willison, J. C., G. Ahombo, J. Chabert, J.-P. Nlagnin, andP. M. Vignais. 1985. Genetic mapping of the Rlio(lopse(ludomlo-nIais capsulata chromosome shows nonclustering of genesinvolved in nitrogen fixation. J. Gen. Microbiol. 131:3001- 3015.

126. Winans, S. C., P. R. Ebert, S. E. Stachel, M. P. Gordon, andE. W. Nester. 1986. A gene family essential for Agrohacteriumnvirulence is homologous to a family of positive regulatory loci.Proc. NatI. Acad. Sci. USA 83:8278-8282.

127. Wraight, C. A., D. R. Leuking, R. T. Fralev, and S. Kaplan.1978. Synthesis of photopigments and electron transport com-ponents in synchronous phototrophic cultures of Rliodopse(l-domonias spihiea oide.s. J. Biol. Chem. 253:465-471.

128. Yeates, T. O., H. Komiya, D. C. Rees, J. P. Allen, and G.Feher. 1987. Structure of the reaction center from Rhlodohactersphaeroides R-26: membrane-protein interactions. Proc. NatI.Acad. Sci. USA 84:6438-6442.

129. Yen, G. S. IL, B. D. Cain, and S. Kaplan. 1984. Cell-cyclespecific biosynthesis of the photosynthetic membrane of Rlio-dopselud1oinionas sphaeroides: structural implications. Bio-chim. Biophys. Acta 777:41-45.

130. Yen, H.-C., and B. Marrs. 1976. Maps of genes for carotenoidand bacteriochlorophyll biosynthesis in Rhoodopseudomonascap.sulata. J. Bacteriol. 126:619-629.

131. Yen, H.-C., and B. L. Marrs. 1977. Growth of Rhodlopsedlo-itnonCas capsidlata under anaerobic dark conditions with di-methylsulfoxide. Arch. Biochem. Biophys. 181:411-418.

132. Youvan, D. C., E. J. Bylina, A. M. Alberti, H. Begusch, andJ. E. Hearst. 1984. Nucleotide and deduced polypeptide se-quences of the photosynthetic reaction center. B870 antennaand flanking sequences from R. c(apsidata. Cell 37:949-957.

133. Youvan, D. C., J. E. Hearst, and B. L. Marrs. 1983. Isolationand characterization of enhanced fluoresence mutants of Rlio-dopseiudontiona(is capsulata. J. Bacteriol. 154:748-755.

134. Youvan, D. C., and S. Ismail. 1985. Light-harvesting 11 (B800-

850 complex) structural genes from Rliodopseuldottiotitis (cap-sulata. Proc. Natl. Acad. Sci. USA 82:5842.

135. Youvan, D. C., S. Ismail, and E. J. Bylina. 1985. Chromosomaldeletion and plasmid complementation of the photosyntheticreaction center and light-harvesting genes from Rhodopseudo-IoaIS(UIscpsiulata. Gene 38:19-30.

136. Youvan, D. C., and B. L. Marrs. 1984. Molecular- genetics andthe light reactions of photosynthesis. Cell 39:1-3.

137. Zhu, Y. S., D. W. Cook, F. Leach, G. A. Armstrong, M.Alberti, and J. E. Hearst. 1986. Oxygen-regulated mRNAs forlight-harvesting and reaction center complexes and for bacte-riochlorophyll and carotenoid biosynthesis in Rhodobacter(apslldati.s during the shift from anaerobic to aei-obic growth.J. Bacteriol. 168:1180-1188.

138. Zhu, Y. S., and J. Hearst. 1986. Regulaltion of expression ofgenes for light-harvesting antenna proteins LHI and LHII;reaction center polypeptides RC-L. RC-M. and RC-H;, andenzymes of bacteriochlorophyll and carotenoid biosynthesis inRhodobacter (apsiulatts by light and oxygen. Proc. Natl.Acad. Sci. USA 83:7613-7616.

139. Zhu, Y. S., and S. Kaplan. 1985. Effects of light. oxygen, andsubstrates on steady-state levels of mRNA coding for ribulose1,5-bisphosphate carboxylase and light-harvesting and reactioncenter polypeptides in Rhodopseudomionas sphaeroidles. J.Bacteriol. 162:925-932.

140. Zhu, Y. S., P. J. Kiley, T. .j. Donohue, and S. Kaplan. 1986.Origin of the mRNA stoichiometry of the pifl operon inRhodobacter spliaiceroi(des. J. Biol. Chem. 261:10366-10374.

141. Zuber, H. 1985. Structure and function of light-harvestingcomplexes and their polypeptides. Photochem. Photobiol.42:821-844.

142. Zuber, H. 1986. Structure of light-harvesting antenna com-plexes of photosynthetic bacteria. cyanobacteriai and red algae.Trends Biochem. Sci. 11:414-419.

143. Zuber, H. 1986. Primary structure and function of the light-harvesting polypeptides from cyanobacteria. red algae. andpurple photosynthetic bacteria. p. 238-251. In L. A. Staehelinand C. J. Arntzen (ed.). Photosynthesis Ill: photosyntheticmembranes. Encyclopedia of plant physiology, new series.vol. 19. Springer-Verlag, New York.

VOL. 512 1'988

on January 21, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from