The PuhB Protein of Rhodobacter capsulatus Functions in Photosynthetic Reaction Center Assembly with a Secondary Effect on Light-Harvesting Complex 1

Abstract

The core of the photosynthetic apparatus of purple photosynthetic bacteria such as Rhodobacter capsulatus consists of a reaction center (RC) intimately associated with light-harvesting complex 1 (LH1) and the PufX polypeptide. The abundance of the RC and LH1 components was previously shown to depend on the product of the puhB gene (formerly known as orf214). We report here that disruption of puhB diminishes RC assembly, with an indirect effect on LH1 assembly, and reduces the amount of PufX. Under semiaerobic growth conditions, the core complex was present at a reduced level in puhB mutants. After transfer of semiaerobically grown cultures to photosynthetic (anaerobic illuminated) conditions, the RC/LH1 complex became only slightly more abundant, and the amount of PufX increased as cells began photosynthetic growth. We discovered that the photosynthetic growth of puhB disruption strains of R. capsulatus starts after a long lag period, which is due to physiological adaptation rather than secondary mutations. Using a hybrid protein expression system, we determined that the three predicted transmembrane segments of PuhB are capable of spanning a cell membrane and that the second transmembrane segment could mediate self-association of PuhB. We discuss the possible function of PuhB as a dimeric RC assembly factor.

Purple nonsulfur photosynthetic bacteria such as Rhodobacter capsulatus are capable of aerobic respiratory and anaerobic photosynthetic growth. The photosynthetic apparatus includes three membrane-bound pigment-protein complexes: the reaction center (RC), which functions as a light-dependent quinone reductase; light-harvesting complex 1 (LH1), which is adjacent to and forms a ring or arc encircling the RC as part of the so-called core complex that includes the PufX polypeptide (17, 22, 36, 38, 43, 51); and LH2, which is thought to be present in multiple copies of a ring-shaped structure that interconnect core complexes (33). These complexes are located within differentiated invaginations of the cytoplasmic membrane, called the intracytoplasmic membrane system, formed in response to oxygen deprivation (14). The presence of the various photosynthetic complexes can be evaluated by their characteristic light absorption spectra, which reflect the protein environments around bacteriochlorophyll a (BChl); unbound BChl has a far-red absorption peak at 780 nm, whereas this peak of the LH2 BChls of R. capsulatus is at 800 and 850 nm, and BChls in the LH1 complex absorb light at approximately 870 nm. The far-red absorption peaks of the RC are at about 760 nm (bacteriopheophytins), 804 nm (accessory or “voyeur” BChls), and 865 nm (the “special pair” of BChls) (16).

In R. capsulatus, one of the three polypeptides of the RC, RC H, is encoded by the puhA gene that is transcribed as part of the bchFNBHLM-lhaA-puhABC superoperon from two promoters, one 5′ of bchF and the other within the lhaA gene (3, 6, 57). A segment of the puh operon is shown in Fig. 1a. The PuhB protein (formerly known as Orf214) is required for optimal RC/LH1 levels and photosynthetic growth (55). Other than RC H, the remaining RC polypeptides, RC L and RC M, both polypeptides of LH1 (LH1 α and LH1 β), and the associated protein PufX are encoded by the puf operon (Fig. 1b). The puf operon is part of a superoperon that includes pigment biosynthesis genes (53) and also encodes PufQ, a regulatory factor in BChl biosynthesis (4). PufX has been implicated in the exchange of quinone/quinol between the RC and cytochrome b/c1 complexes (28, 29).

FIG. 1.

Organization of puh and puf operons. (a) Genetic and restriction map of the puh operon, with disruptions (shaded boxes) shown below. (b) Genetic and restriction map of the puf operon, with disruptions (shaded boxes) shown below.

In this study, we investigate the basis of the RC/LH1 deficiency and impaired photosynthetic growth of R. capsulatus puhB disruption strains by examining the effects of PuhB on RC assembly in the absence of LH1, LH1 assembly in the absence of the RC, and the abundance of PufX. We also evaluate the ability of each of the three putative transmembrane (TM) segments of PuhB to span the inner membrane and to self-associate. The results indicate that PuhB is a membrane protein, perhaps a dimer, and is required for optimal assembly of the RC polypeptides and cofactors to form a catalytically active RC complex. Thus, the negative effects of puhB disruption on the level of the RC/LH1/PufX core complex appear to result from a primary defect in the assembly of the RC component of the core complex.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The photosynthetically wild-type R. capsulatus strain SB1003 (47), the LH2⁻ mutant strain MW442 (44), the LH2⁻ puhB disruption strain DW23 (55), and the gene transfer agent (GTA) overproducer strain DE442 (56) have been described previously. Escherichia coli DH5α (Life Technologies, GIBCO BRL) and C600 (10) were used for the construction and maintenance of plasmids; strains HB101(pRK2013), S17-1, and TEC5 (13, 46, 48) were used in E. coli-to-R. capsulatus conjugations. E. coli MM39, a kind gift of J. Beckwith, was used for the TOXCAT system (39). Plasmids pUC12, pUC13::EcoF, pUC18, pUC19, pRR5C, pRK767, pTPR9, pTPR8, pStuI, and pXCA6::935 have been described previously (1, 23, 24, 32, 37, 49, 54, 57). Plasmid pUC4::KIXX was from Pharmacia Biotech, Inc. These and other bacterial strains and plasmids produced in this study are described below and listed in Table 1.

TABLE 1.

E. coli and R. capsulatus strains and plasmids used to examine the role of puhB

Strain or plasmid	Descriptiona	Source or reference(s)
Strains
DH5α	Cloning strain	GIBCO BRL
C600	Cloning strain	10
TEC5	Recombination and conjugative transfer of pUC plasmids to R. capsulatus	48
S17-1	Conjugative transfer of IncP and IncQ plasmids to R. capsulatus	46
HB101(pRK2013)	Helper strain to mobilize plasmids from C600 to R. capsulatus	13
MM39	malE mutant for TOXCAT system	J. Beckwith
SB1003	RC+LH1+LH2+ wild-type strain of R. capsulatus	47
MA05	KIXX cartridge inserted into puhB deletion in SB1003	This study
MW442	LH2− derivative of SB1003	44
DW23	puhB deletion in MW442	55
MA01	KIXX cartridge inserted into pufQBALM deletion in MW442	This study
MA03	KIXX cartridge inserted into pufQBALM deletion in DW23	This study
DE442	GTA overproducer	56
Plasmids
pUC12, pUC18, pUC19	Cloning vectors	32, 49
pUCpuf	pUC12 with XbaI-BamHI fragment containing pufQBALMX	This study
pUC13::EcoF	pUC13 with EcoRI fragment containing puhB and flanking sequence	54
pEH214	pUC19 with EcoRI-HindIII fragment containing puhB	This study
pUC4::KIXX	Kanamycin resistance cartridge without transcription termination signal	Pharmacia Biotech Inc.
pRR5C	puf promoter followed by pufQ gene; gentamicin resistance marker	57
pMA20	puf promoter followed by multiple cloning sequence; gentamicin resistance marker	This study
pMA22	puf promoter followed by puhB; gentamicin resistance marker	This study
pRK767	Expression vector; tetracycline resistance marker	23
pTPR9	puf promoter, pufQALMX; tetracycline resistance marker	37
pTPR8	puf promoter, pufQBLMX; tetracycline resistance marker	37
pMA10	puf promoter, pufQLMX; tetracycline resistance marker	This study
pStu I	puf promoter, pufQBAX; tetracycline resistance marker	24
pXCA6::935	puf promoter, pufQ, and pufB fused to lacZ; tetracycline resistance marker	1
pccKAN	Constitutive expression of a ToxR′-(kanamycin resistance cartridge)-MalE hybrid; ToxR-dependent expression of CAT; ampicillin resistance	39
pccGpAwt	pccKAN plasmid expressing wild-type glycophorin A hybrid	39
pccGpA83I	pccKAN plasmid expressing mutated glycophorin A hybrid	39
pccTNM	pccKAN plasmid expressing hybrid without a transmembrane segment	39
pccPuhBTM1	pccKAN plasmid expressing a hybrid of the first transmembrane segment of PuhB	This study
pccPuhBTM2N	pccKAN plasmid expressing a hybrid of the second transmembrane segment of PuhB	This study
pccPuhBTM3N	pccKAN plasmid expressing a hybrid of the third transmembrane segment of PuhB	This study

^aAll KIXX cartridges are oriented so that transcription of the neo gene is in the same direction as that of the operon disrupted.

Growth conditions and media.

E. coli strains were grown in Luria-Bertani (LB) medium (40) or on LB agar plates. Aerobic and semiaerobic cultures of R. capsulatus were grown at 30°C without illumination in RCV medium (7) in Erlenmeyer flasks filled to 20 or 80% of nominal capacity, respectively, and shaken at 300 or 150 rpm, respectively. RCV plates were incubated aerobically at 30°C without illumination. Photosynthetic cultures were grown anaerobically in screw-cap tubes (20 ml) or stoppered Roux bottles (800 ml) inoculated from semiaerobic cultures and filled with RCV or on RCV agar plates placed in BBL GasPak anaerobic jars (Becton Dickinson & Co.). Photosynthetic liquid cultures were incubated at 30°C in an aquarium filled with water and illuminated by halogen (Capsylite; Sylvania) lamps at an intensity of 150 μE/m²/s, measured with a photometer equipped with an LI-190SB quantum sensor (LI-COR Inc.). Culture turbidity was monitored with a Klett-Summerson photometer equipped with a red (no. 66) filter (100 Klett units, 3.3 × 10⁸ CFU/ml). Antibiotic-resistant E. coli and R. capsulatus strains were selected with 25 and 10 μg of kanamycin/ml, 10 and 0.5 μg of tetracycline/ml, and 10 and 2 μg of gentamicin/ml, respectively. Ampicillin was used at 100 μg/ml to select plasmid-bearing E. coli strains.

DNA techniques.

Recombinant and other DNA procedures were carried out essentially as described previously (40). Plasmid DNA was isolated from cells and from agarose gels with kits from QIAGEN. Conjugative transfer of plasmids from E. coli to R. capsulatus was performed by sequential pelleting of cultures of donor and recipient cells in a 1:5 ratio by volume by centrifugation at 15,000 × g for 1 min, resuspension in 50 μl of RCV, and incubation of 10-μl aliquots on an RCV agar plate overnight at 30°C. Donor cells were absent from the negative controls, and one volume of HB101(pRK2013) was added as a helper to transfer pXCA6::935 from a C600 donor strain. Cells from each spot were resuspended in 2 ml of RCV, and 100 μl was spread onto RCV agar plates containing the appropriate antibiotics. Transconjugant colonies were streaked onto rich YPS agar (52) plates to ensure the absence of E. coli donors.

Construction of the pufQBALM disruption strains MA01 and MA03.

A BamHI-to-XbaI fragment containing the puf operon of R. capsulatus (Fig. 1a) was ligated into pUC12 cut with BamHI and XbaI. The resultant plasmid, pUCpuf, was cut with HindIII and XbaI, and the ends were filled in with the Klenow fragment and religated to remove the SalI site from the multiple cloning site. This modified pUCpuf was cut with SalI to delete the pufQBALM coding sequence, which was replaced with an XhoI-cut kanamycin resistance cartridge from plasmid pUC4::KIXX such that the orientation of the cartridge's promoter was the same as that of the puf promoter. The resultant plasmid was conjugatively transferred from TEC5 to DE442, and the disrupted puf operon was transduced with GTA into MW442 and DW23 to produce strains MA01 and MA03, respectively.

Construction of the pufQLMX complementation plasmid pMA10.

Plasmids pTPR9 and pTPR8 contain puf operons with nearly total in-frame deletion mutations of pufB and pufA, respectively (37) (Fig. 1a). To create a plasmid-borne puf operon lacking pufB and pufA that would restore RC expression to MA01 and MA03 in the absence of LH1, the puf operons from these plasmids were excised as KpnI-to-XbaI fragments and ligated into pUC18. An XhoI-to-BseRI fragment of 1,097 bp containing pufQB was removed from pufQB(ΔA)LMX and replaced with the corresponding pufQΔB fragment of 992 bp from pufQ(ΔB)ALMX. The resultant pufQ(ΔBΔA)LMX operon was excised as a KpnI-to-XbaI fragment and ligated into pRK767 cut with KpnI and XbaI, producing plasmid pMA10 (Fig. 1b).

Construction of the puhB disruption mutant strain MA05.

An EcoRI-to-HindIII fragment of 2,734 bp (Fig. 1a) containing puhB and flanking sequences was excised from plasmid pUC13::EcoF (54) and ligated into pUC19, producing plasmid pEH214. BstBI and ClaI were used to excise a 359-bp fragment of puhB (56% of the coding sequence) from pEH214 and to excise the kanamycin resistance cartridge from pUC4::KIXX. The cartridge was ligated into the deletion with the orientation of the cartridge promoter parallel to that of the puh promoter. This plasmid was conjugatively transferred to DE442, and the puhB disruption was transduced into SB1003 to produce strain MA05.

Construction of pMA20 and the puhB complementation plasmid pMA22.

A HindIII-to-EcoRI fragment carrying the puf promoter and pufQ from pRR5C was inserted into pUC18, and pufQ was excised as two fragments with AccI and EcoRI. The annealed oligonucleotides 5′-CTAGATGCATCGATCCGG-3′ and 5′-AATTCCGGATCGATGCATCT-3′ were ligated into the deletion, creating a multiple cloning sequence, and the original HindIII-to-EcoRI fragment within pRR5C was replaced with this modified fragment, bearing only the puf promoter. This plasmid was called pMA20.

Plasmid pMA22 is essentially pMA20 with the puhB gene inserted between the EcoRI and SmaI sites. The history of pMA22 is complicated: epitope-tagged puhB amplicons generated by PCR were sequenced as pUC19 inserts and subcloned into pRR5C, and these plasmids were converted to pMA20 derivatives; finally, HindIII-to-MluI fragments of the pufQ-free plasmids were swapped to produce pMA22, bearing an untagged puhB coding sequence together with its putative ribosome-binding site under control of the puf promoter.

Plasmids pMA20 and pMA22 were conjugatively transferred from E. coli S17-1 to R. capsulatus MA05.

Spectroscopy.

Measurements of the RC special pair (865 nm) and LH1 (880 nm) BChl peak areas were carried out by using triplicate highly aerated cultures of each strain at a low density, which were used as inocula for semiaerobic growth to induce expression of photosynthesis genes in 2-liter flasks. Initial samples of 20 ml and subsequent samples of 10 ml were taken from an initial volume of 1,650 ml at intervals of 1.5 h. Sample turbidity was measured with a Klett-Summerson photometer (see above), and cells were harvested by centrifugation. Absorption spectroscopy of intact cells was performed as described previously (28), and data were collected with the J&M TIDAS II spectrophotometer and analyzed with Spectralys software (World Precision Instruments). Light scattering at 650 nm was used to normalize the spectra. For strains that produced the RC, baselines were drawn from 700 to 830 nm and from 830 to 930 nm, the values along the baselines were subtracted from the spectrum, and the area of the BChl special pair peak from 830 to 900 nm was computed. The baselines for strains that produced LH1 were drawn from 700 to 820 nm and from 820 to 930 nm, and the peak area from 820 to 920 nm was computed.

A Hitachi 557 double-beam spectrophotometer was used for low-temperature absorption spectroscopy. Chromatophores (intracytoplasmic membrane system vesicles) isolated from cultures at 150 Klett units as described previously (3) were mixed with an equal volume of anhydrous glycerol and frozen in liquid nitrogen. Spectra were obtained with the samples chilled by, but not immersed in, liquid nitrogen. Flash spectroscopy was carried out as previously described (28).

β-Galactosidase assays.

Cells from actively growing semiaerobic and photosynthetic R. capsulatus cultures were harvested by centrifugation. Pellets from 10 ml of culture were resuspended in 1 ml of RCV and stored at −80°C before assay as previously described (30). Results are from three independent experiments.

SDS-PAGE and immunoblots.

The amount of protein in chromatophore preparations was determined by a modified Lowry method, with bovine serum albumin as the standard (35). Samples containing 40 μg of protein were mixed with loading buffer, heated at 50°C for 10 min, and electrophoresed in a tricine-SDS-polyacrylamide gel electrophoresis (SDS-PAGE) system (41). Gels were stained with 0.025% Coomassie brilliant blue R-250 in 40% methanol and 10% acetic acid and destained in the same solution minus the dye.

Antibodies were raised against peptides corresponding to the N-terminal and C-terminal regions of “mature” PufX conjugated to keyhole limpet hemocyanin (KLH): (NH₃⁺)-SMFDKPFDYENGSKFC(NH₂)-(KLH) and (KLH)-LPERAHQAPSPYTTEV-(COO⁻) (34). Two rabbits were injected with a combination of both peptides by Genemed Synthesis.

For immunoblots, samples of intact cells (50 μg of protein) were run on 12% polyacrylamide gels and electroblotted at 80 V for 2 h onto nitrocellulose membranes by standard methods (5). The primary antibodies against PufX were used at a 1:2,000 dilution in 20 ml of Tris-buffered saline-Tween (TBS-T) (20 mM Tris-HCl [pH 7.6], 0.8% sodium chloride, 0.1% Tween 20) containing 5% Nestle Carnation skim milk powder. Following overnight incubation with shaking at 7°C with the primary antibody, the membranes were washed three times in 20 ml of TBS-T for 20 min at room temperature. The secondary antibody, horseradish peroxidase-linked donkey anti-rabbit immunoglobulin G (Amersham), was used at a 1:5,000 dilution in TBS-T containing 5% skim milk powder for 1 h at room temperature, followed by three more washes. Chemiluminescence was produced with an ECL kit (Amersham) and detected with BioMax MS film (Kodak).

TOXCAT assays.

The three predicted TM segments of PuhB (amino acid residues 38 to 65, 70 to 95, and 101 to 126, called TM1, TM2, and TM3, respectively) were amplified by PCR from a BamHI fragment containing part of puhB (3). The primers for TM1 were 5′-GTGGATGCTGGCTAGCGACGCGTTCAAG-3′ and 5′-AATAATGCCTGGATCCCGCCTTCCTCGTGCCAG-3′, those for TM2 were 5′-TTGCACCCGAGCTAGCCTGCCCA-3′ and 5′-GTGTAGATCGGGATCCCGGCCTGCGCGAA-3′, and those for TM3 were 5′-CCGTGCCGCGGCTAGCACCATCACCTC-3′ and 5′-GCCAGCGACAGGATCCCGATCACGGTGAAG-3′, respectively (restriction sites are underlined). The amplicons were inserted into pccKAN as NheI-BamHI fragments, and the three plasmids pccPuhBTM1, pccPuhBTM2N, and pccPuhBTM3N were transformed into MM39 cells. The ability of the putative TM segments to span the inner membrane of E. coli was qualitatively tested by streaking the MM39 strains on M9-maltose minimal medium (40) and inspection of colony formation after 24 h. The positive control was MM39(pccGpAwt), expressing a hybrid of the glycophorin A transmembrane segment, and the negative control was MM39 (pccTNM), lacking a transmembrane segment (39).

For the quantitative chloramphenicol acetyltransferase (CAT) (TOXCAT) assay, MM39(pccGpAwt) was the positive control, and the negative control was MM39(pccGpA83I), in which a substitution mutation has reduced the self-association of the glycophorin A transmembrane segment (39). Cultures grown overnight in LB broth were diluted to an optical density at 600 nm of 0.1 and grown for 1.5 h before harvesting the equivalent of 200 μl at an optical density at 600 nm of 0.6. Cells were resuspended in 500 μl of 100 mM Tris HCl (pH 8.0), to which 20 μl of a solution of 50 mM Tris HCl (pH 8.0), 100 mM EDTA, and 100 mM dithiothreitol was added, followed by 50 μl of toluene. After brief vortexing and permeabilization of the cell membranes by incubation at 30°C for 30 min and pelleting of cell debris in a microcentrifuge for 5 min, the cell extract supernatants were placed on ice. Sixty microliters of cell extract was then mixed with 10 μl of substrate (BODIPY FL 1-deoxychloramphenicol in methanol; Sigma) and incubated at 37°C for 5 min. Ten microliters of 9 mM acetyl coenzyme A (Sigma) was added, and samples of 5 μl taken at intervals of 2 min were spotted onto a silica gel plate (Sigma) for thin-layer chromatography. The solvent for chromatography was a mixture of 85 ml of dichloromethane and 15 ml of methanol. Densitometric analysis of a photograph of the plate exposed to UV light for 0.1 s was used to evaluate acetylated product formation over time, and CAT activity was computed as the slope of the line of best fit.

RESULTS

The entire population of PuhB⁻ cells adapts to photosynthetic growth.

In a previous study, an LH2⁻ puhB deletion strain (DW23) was described as incapable of photosynthetic growth, and when growth began after prolonged incubation, it was tentatively attributed to secondary mutations (55). To characterize the function of the puhB gene further, we created an LH2⁺ puhB disruption strain, MA05. Both MA05 and DW23 were consistently observed to begin slow photosynthetic growth after a lag of about 12 h. The growth defect of the puhB disruption strain MA05 was complemented by restoration of puhB under control of the puf promoter on plasmid pMA22 (Fig. 2). The negative control plasmid lacking puhB, pMA20, did not shorten the lag or improve the specific growth rate of MA05. Similar growth patterns were observed in complementation experiments with the LH2⁻ puhB deletion strain DW23 (data not shown).

FIG. 2.

Effect of puhB deletion and restoration in trans on photosynthetic growth. The mean of triplicate cultures is shown for each strain; variation was negligible. The vertical axis gives the culture density in Klett units; the horizontal axis gives the time in hours. SB1003 is the wild-type strain; MA05 is the puhB deletion strain. The puhB gene was restored to MA05 on plasmid pMA22.

When semiaerobically grown MA05 cells were spread onto RCV plates, the number of colonies that appeared under anaerobic illuminated incubation conditions after 3 days was the same (106% ± 18%) as those found under aerobic dark conditions. Therefore, the delayed growth of photosynthetic cultures of the puhB mutant is due to a physiological adaptation in every cell of the population and not the selection of spontaneous secondary mutants.

Disruption of puhB reduces the levels of LH1, the RC, and PufX.

The levels of LH1 and LH2 complexes in MA05 were compared to those of SB1003 in low-temperature absorption spectra of membrane vesicles (chromatophores), in which the difference in LH1 absorption at 870 nm (shoulder of the 850-nm LH2 peak) was visible (Fig. 3). The LH2 absorption peaks were not significantly different. SDS-PAGE of chromatophores from the puhB mutant MA05 and wild-type strain SB1003, grown semiaerobically and photosynthetically, revealed that the amounts of the RC and LH1 proteins per total protein appeared to be reduced due to the puhB disruption (Fig. 4). Thus, the SDS-PAGE data agree with the LH1 absorption spectra.

FIG. 3.

Low-temperature absorption spectra of chromatophores from MA05 and SB1003 grown semiaerobically and photosynthetically.

FIG. 4.

Effect of the puhB deletion on RC and LH1 protein levels in chromatophores isolated from semiaerobic and photosynthetic cultures of MA05 and SB1003 and subjected to SDS-PAGE. The diffuse staining above the RC L protein band in SB1003 may represent incompletely denatured RC L molecules.

The amount of photochemically active RC complex in chromatophores from puhB mutant strain MA05 and wild-type strain SB1003 was measured as the final change in absorbance at 605 minus 540 nm after a train of eight flashes of actinic light (Fig. 5a). This amount of photochemically active RC complex in MA05 was about 17% of that in SB1003 grown under semiaerobic respiratory conditions and 14% of that in SB1003 grown under photosynthetic conditions, consistent with the reduced intensity of RC protein bands in SDS-PAGE (Fig. 3).

FIG. 5.

Flash spectroscopy of chromatophores from semiaerobically and photosynthetically grown strains SB1003 and MA05. (a) Amount of functional RC (605 minus 540 nm). (b) Single-flash carotenoid bandshift (490 minus 475 nm) with and without antimycin. (c) Carotenoid bandshift (540 minus 510 nm for semiaerobic conditions and 490 minus 475 nm for photosynthetic conditions) following eight flashes with and without antimycin. The vertical bars on the left represent a change of 0.00435 absorbance unit at the respective wavelength pairs.

The carotenoid bandshift is an indicator of the transmembrane potential due to light-driven electron and proton translocation (21). The carotenoid bandshift consists of three time-resolved phases; phases 1 and 2 occur within microseconds of an actinic flash and are electrochromic responses to electron and proton transfer reactions in the RC and the attendant c-type cytochromes (21). Phase 3 occurs during ∼100 ms after a flash and is thought to represent the time needed for quinol-mediated transfer of electrons and protons from the RC through the cytochrome b/c₁ complex, which is inhibited by antimycin. The carotenoid bandshift was measured in chromatophores of photosynthetically grown MA05 and SB1003 after a single flash in the presence and absence of antimycin. The overall magnitude of the bandshift was rather less in MA05 than in SB1003, consistent with the reduced RC content discussed above. Furthermore, the third phase of the bandshift (compare the two panels of Fig. 5b) appeared to be reduced in MA05, suggesting that the reduced amount of RC in MA05 is impaired in quinol transfer from the RC to the cytochrome b/c₁ complex, which is characteristic of a PufX deficiency (28, 29).

Even though the puhB disruption strain MA05 had many fewer RCs than the wild type, these RCs were capable of successive turnovers in response to a series of eight flashes and generated progressively greater transmembrane potentials, albeit to a lesser final carotenoid bandshift magnitude than that of wild-type strain SB1003 (Fig. 5c). The difference between the carotenoid bandshifts from these two strains grown photosynthetically appeared to be slightly less than the difference between chromatophores from semiaerobically grown cells. This result indicates that disruption of puhB impairs photosynthetic electron and proton transfer more in semiaerobically grown cells than after puhB mutant cells adapt to anaerobic illuminated conditions, consistent with the growth studies (Fig. 2).

The indications from the flash spectroscopy that the puhB disruption strain MA05 might be deficient in PufX led us to use anti-PufX antibodies to assess the amount of PufX directly. Indeed, an immunoblot revealed that the amount of PufX in PuhB⁻ MA05 cells was extremely low relative to the PuhB⁺ strain SB1003 under semiaerobic growth conditions and under anaerobic illuminated conditions prior to the onset of photosynthetic growth, but the amount of PufX in MA05 increased after cells began to grow photosynthetically (Fig. 6). In contrast, the amount of PufX during semiaerobic and photosynthetic growth was consistently high in SB1003 and consistently moderate when puhB was restored to MA05 on plasmid pMA22.

FIG. 6.

Immunoblot of PufX in 50 μg of total cell protein from SB1003, MA05, and MA05(pMA22) at different times after transfer of semiaerobically incubated cells to anaerobic photosynthetic conditions. The amount of PufX was low in the puhB mutant strain MA05 under semiaerobic conditions (0 h) and during the lag under photosynthetic conditions (12 h) but increased when cells were growing photosynthetically (50 h), whereas PufX was equally abundant in both modes of growth in the puhB⁺ strain SB1003 and remained at a consistent level below that of the wild type when puhB was restored to MA05 on plasmid pMA22.

In summary, the data indicate that puhB disruption leads to diminished core complex levels and furthermore that the reduced amounts of RC/LH1 that are formed may contain a substoichiometric content of PufX that increases when cells adapt to photosynthetic growth conditions.

The puhB disruption strain has normal levels of puf::lacZ gene fusion expression.

To evaluate whether decreased expression of the puf genes, which encode five of the six core complex polypeptides, might account for the reduction in RC/LH1/PufX levels in MA05, the β-galactosidase activities expressed from a translationally in-frame pufB::lacZ gene fusion, which is transcribed from the puf promoter in plasmid pXCA6::935, were determined in the puhB disruption MA05(pXCA6::935) and wild-type SB1003(pXCA6::935) strains. The results (Table 2) indicate that neither initiation of transcription from the puf promoter nor initiation of translation of the LH1 β polypeptide is significantly decreased by the puhB mutation. Therefore, we suggest that the PuhB protein functions posttranslationally, perhaps in assembly, to yield an appropriately high level of the core complex.

TABLE 2.

β-Galactosidase activity expressed from the puf promoter in puhB⁺ SB1003 and puhB mutant MA05 strains grown semiaerobically and photosynthetically^a

Strain (plasmid) and growth conditionb	Mean activity ± SDc
SB1003(pXCA6::935), SA	5.0 ± 1.6
MA05(pXCA6::935), SA	5.0 ± 0.2
SB1003(pXCA6::935), PS	7.3 ± 0.9
MA05(pXCA6::935), PS	7.0 ± 1.0

^aβ-Galactosidase activity is measured as nmol of o-nitrophenol per minute per 10⁸ cells.

^bSA, semiaerobic growth conditions; PS, photosynthetic growth conditions.

^cMeans ± standard deviations from data from three samples are shown.

The puhB disruption impairs RC assembly in the absence of LH1 but not LH1 assembly in the absence of the RC.

Assembly of the RC was measured as the increase in the area of the BChl special pair peak in absorption spectra of intact cells after a shift of cultures from noninducing (high aeration) to inducing (semiaerobic) conditions. One comparison was made between the puhB⁺ MA01(pMA10) and puhB mutant MA03(pMA10) strains, in which both of the chromosomal pufQBALM genes had been deleted and only pufQLMX, without the LH1 genes pufBA, had been restored on plasmid pMA10. Strains MA01(pTPR9) and MA03(pTPR9), which express a puf operon that lacks the LH1 β gene, and MA01(pTPR8) and MA03(pTPR8), which express a puf operon lacking the LH1 α gene, were also compared. The growth kinetics and culture density of each pair were not significantly different at any point in the experiment, and so all cultures experienced similarly semiaerobic conditions. The data indicate an RC assembly deficiency in the MA03 puhB disruption strains (Fig. 7a to c). These results were confirmed by measurement of the area under the RC voyeur or accessory BChl peak (800 nm) (2).

FIG. 7.

Independent production of the RC and LH1 in MA01 (puhB⁺ [filled squares]) and MA03 (puhB mutant [open squares]) strains containing plasmids, following a switch from aerobic to semiaerobic conditions. The BChl special pair peak area was measured in cells containing puf operon variants in the following plasmids: (a) pMA10, which lacks both LH1 α and β genes; (b) pTPR9, which lacks the LH1 β gene; and (c) pTPR8, which lacks the LH1 α gene. (d) The LH1 peak area was measured in cells containing pStuI, which lacks the RC L and M genes.

In a similar experiment, LH1 assembly was compared between the puhB⁺ strain MA01(pStuI) and the puhB disruption strain MA03(pStuI), in which the puf operon restored on plasmid pStuI lacks the pufLM genes encoding RC polypeptides. Disruption of puhB had no effect on LH1 assembly in the absence of the RC L and M proteins (Fig. 7d).

We evaluated RC and LH1 stability (decay rates) in PuhB⁺ and PuhB⁻ cells, as described above, after shifting photocomplex-replete cultures from low to high aeration and did not detect any differences (2). Thus, the differences in Fig. 7a to c, appear to be due to a PuhB-specific function in RC assembly, as opposed to a function in protection from RC disassembly.

PuhB may self-associate through its second transmembrane segment.

Hydropathy plots of PuhB indicate three TM domains. A TOXCAT assay (39) of the three putative TM segments of PuhB assessed whether PuhB is a transmembrane protein and if it might function as a multimer. In this assay, a putative TM segment is fused to an N-terminal ToxR′ dimerization-dependent transcriptional activator domain and a C-terminal MalE maltose-binding domain and is expressed in MM39, a MalE⁻ strain of E. coli. Insertion of the hybrid protein into the membrane with MalE in the periplasm allows the cell to grow with maltose as the sole carbon source, and dimerization of the TM segments is measured by an assay of the CAT enzyme expressed under ToxR control. E. coli MM39 cells expressing hybrids of all three putative TM segments were capable of growth on M9-maltose minimal medium, indicating that all of these segments span the inner membrane in this qualitative measurement (39). However, the cells expressing hybrids of TM1 and TM3 and the positive control, a hybrid of the transmembrane segment of glycophorin A (39), formed colonies approximately 1 mm in diameter, whereas those expressing the ToxR′-PuhBTM2-MalE hybrid were only 0.2 mm in diameter, after overnight incubation. We interpret this result as a difficult insertion of this TM2 hybrid, which in PuhB is predicted by the positive-inside rule (50) to run from periplasm to cytoplasm, but must assume the reverse orientation within a ToxR′-PuhBTM2-MalE hybrid for MalE to reach the periplasm. In support of this interpretation, we observed that when the arginyl residue that follows TM2 in PuhB was included in a ToxR′-TM2-MalE hybrid, MM39 cells expressing this hybrid were unable to grow on M9-maltose minimal medium, indicating a cytoplasmic location of MalE (39). Hence, TM2 of the native PuhB protein appears to have an N-out/C-in topology.

Quantitative measurements of TM self-association yielded high CAT activity in extracts of cells expressing the ToxR′-PuhBTM2-MalE hybrid (Fig. 8), whereas hybrids of the other two transmembrane segments of PuhB, TM1 and TM3, generated little reporter activity. Although this assay may be affected by variations in the level of expression of different hybrids, the results of the maltose growth test suggest that TM1 and TM3 hybrids were expressed at least as well as the TM2 hybrid. These data indicate that the PuhB protein contains three transmembrane segments, and the high self-association affinity of the TM2 segment may link two PuhB proteins as a dimer.

FIG. 8.

TOXCAT analysis of three transmembrane segments of PuhB shows that the second segment (TM2) self-associates. CAT enzyme activity is given relative to the positive control, the glycophorin A transmembrane segment (GpAwt), and the negative control is the glycophorin mutant GpA83I (39). In other experiments, the standard deviation of triplicate samples ranged from 4.5% (GpA83I) to 16.5% (TM2).

DISCUSSION

At the time of the first mutational analysis of the gene called orf214 (herein designated puhB) in R. capsulatus (55), the 3′ region of puhA had been sequenced only in R. capsulatus and Rhodospirillum rubrum (9). Since then, puhB homologues have been found to be located 3′ of puhA in all species for which sequence information is available: Rhodospirillum rubrum, Rhodobacter sphaeroides, Roseobacter denitrificans, Rhodopseudomonas palustris, Rubrivivax gelatinosus, Thiocapsa roseopersicina, and three uncultured marine proteobacteria (8, 12, 20, 25, 27) (Roseobacter denitrificans sequence data from GenBank accession no. AJ132424-1). We also found an incompletely sequenced puhB gene and several 3′ genes of the puh operon in the partially sequenced genome of Magnetospirillum magnetotacticum (data at http://www.jgi.doe.gov). This conservation of both the sequence and the gene order indicates an important function of the puhB gene in purple photosynthetic bacteria.

A previous study noted the RC/LH1 deficiency of a puhB deletion strain and suggested that secondary mutations that suppress this phenotype allowed the strain to grow photosynthetically (55). We investigated this possibility and found that all the cells in an inoculum are capable of both aerobic and photosynthetic growth. Therefore, the growth lag seen upon transfer of puhB mutant cells from semiaerobic to photosynthetic conditions is a physiological adaptation. The exact nature of the adaptation to the photosynthetic mode of growth in the absence of PuhB is unknown, although the increased amount of PufX after adaptation of puhB mutant cells from semiaerobic dark to photosynthetic conditions, without an equivalent increase in RC/LH1, suggests that addition of PufX to RC/LH1 is facilitated, or PufX is somehow stabilized in association with RC/LH1, under anaerobic illuminated conditions.

The PuhB protein is herein shown to be an RC-specific assembly factor with an indirect effect on LH1. This is because no effect of PuhB on LH1 was observed in the absence of RC L and RC M, whereas PuhB enhanced the assembly of the RC in the absence of LH1 proteins. The dependence of LH1 on the RC and PufX has been described before; LH1 absorption is reduced by deletions of either the puhA (55) or pufL and pufM (24) genes and inflated by the absence of pufX (28) in R. capsulatus. Similar effects are seen in other species of purple photosynthetic bacteria as well (11, 15, 31). It is remarkable that disruption of puhB in the presence of RC L and RC M appears to have a more negative effect on LH1 (55) than the absence of RC L and RC M in a PuhB⁺ background (24). This finding suggests that the RC assembled without PuhB not only is less abundant but has a structure disruptive of LH1. Furthermore, the puhB disruption appeared to diminish PufX more than the RC and LH1 polypeptides, consistent with an impaired association of PufX with the core complex. Although the reduced amount of RC assembled in the absence of PuhB was capable of catalytic activity, the reduction of the third phase of the carotenoid bandshift after a single flash is consistent with the low amount of PufX detected.

The functional significance of PuhB TM2 self-association in the TOXCAT assay is not clear. There is imperfect twofold symmetry in the arrangement of BChl, bacteriopheophytin, and quinone molecules in purple bacterial RCs (26). If PuhB is a homodimeric RC assembly factor, PuhB could facilitate the symmetrical assembly of pigments into the RC. However, because RC assembly is not entirely lost in the absence of PuhB, we suggest an alternative, speculative model: homodimeric PuhB could use the structural information from a preexisting RC to lower the energy barrier for assembly of a new RC to facilitate the formation of a symmetrical RC/LH1/PufX dimer. The R. sphaeroides RC/LH1/PufX core complex appears to dimerize in a PufX-dependent manner (18, 19, 42, 45). This, together with our observation of a reduced amount of PufX in the absence of PuhB in R. capsulatus, suggests that the involvement of PuhB in RC assembly could facilitate subsequent RC/LH1/PufX core complex assembly as a dimer.

In conclusion, this paper establishes that PuhB is an RC-specific assembly factor, with proposed secondary effects on LH1 and PufX, and that PuhB may be dimeric in vivo. Further investigation is required to reveal the exact function of PuhB in RC assembly, possible dimerization of core complexes, and adaptation of semiaerobically respiring cells to anaerobic photosynthetic growth.