Complex I and Its Involvement in Redox Homeostasis and Carbon and Nitrogen Metabolism in Rhodobacter capsulatus

Abstract

A transposon mutant of Rhodobacter capsulatus, strain Mal7, that was incapable of photoautotrophic and chemoautotrophic growth and could not grow photoheterotrophically in the absence of an exogenous electron acceptor was isolated. The phenotype of strain Mal7 suggested that the mutation was in some gene(s) not previously shown to be involved in CO₂ fixation control. The site of transposition in strain Mal7 was identified and shown to be in the gene nuoF, which encodes one of the 14 subunits for NADH ubiquinone-oxidoreductase, or complex I. To confirm the role of complex I and nuoF for CO₂-dependent growth, a site-directed nuoF mutant was constructed (strain SBC1) in wild-type strain SB1003. The complex I-deficient strains Mal7 and SBC1 exhibited identical phenotypes, and the pattern of CO₂ fixation control through the Calvin-Benson-Bassham pathway was the same for both strains. It addition, it was shown that electron transport through complex I led to differential control of the two major cbb operons of this organism. Complex I was further shown to be linked to the control of nitrogen metabolism during anaerobic photosynthetic growth of R. capsulatus.

Rhodobacter capsulatus is a nonsulfur purple phototrophic bacterium that exhibits a wide range of metabolic capabilities, making it and related organisms probably the most versatile of prokaryotes (28). R. capsulatus grows under dark, aerobic (chemoautotrophic or chemoheterotrophic) conditions using branched respiratory electron transport pathways; in addition, these organisms can grow under anaerobic (photoautotrophic or photoheterotrophic) conditions in the light via cyclic photosynthetic electron transport to generate a proton motive force. Fermentative growth is also an option.

During phototrophic growth of Rhodobacter, redox poise is achieved through the interplay of cyclic photosynthetic electron transport and specific redox-balancing mechanisms of anaerobic metabolism (32). Indeed, it has been suggested that the electron acceptors involved in photosynthetic metabolism function as a sink for excess reducing equivalents to prevent the overreduction of cyclic electron transport; this interaction and control of redox poise and electron transport occur at the level of the ubiquinone pool (12). Under photoheterotrophic growth conditions, the oxidation of C₄-dicarboxylic acids (such as l-malate and succinate) can result in overreduction of the ubiquinone pool (32). The excess reducing equivalents, at the level of the reduced ubiquinone pool, are transferred to NAD⁺ by reverse electron flow mediated by complex I (NADH ubiquinone oxidoreductase) (8). Reducing equivalents stored in the reduced pyridine nucleotide generated by complex I activity may then be dissipated via metabolic systems involved in balancing the intracellular redox state of the organism. These metabolic systems include the Calvin-Benson-Bassham (CBB) reductive pentose phosphate pathway (CBB system) and dinitrogenase system, among others. The dimethyl sulfoxide (DMSO) reductase (DMSOR) system also serves to directly sustain the redox poise of the ubiquinone pool for photosynthetic electron transport (30, 31). Additionally, during photoautotrophic metabolism (in which CO₂ serves as the primary carbon source), molecular hydrogen serves as the reductant, and by means of a membrane-bound hydrogenase, electrons are donated directly into the ubiquinone pool (18, 40, 54).

Complex I is the enzyme that mediates reverse electron flow between the reduced quinone pool and NAD⁺ during photoautotrophic and photoheterotrophic growth in R. capsulatus (7, 20, 29). R. capsulatus contains a single complex I enzyme complex (19), and most of the respective nuo genes have been cloned and sequenced (5, 6, 19). Bacterial complex I of Paracoccus denitrificans (63), Thermus thermophilus (65), Escherichia coli (59), and R. capsulatus (5, 6, 19) contain a minimum of 14 different subunits homologous with mitochondrial complex I (which consists of at least 40 different subunits). Due to its complexity, complex I is the least-understood component of the various known respiratory complexes. Only the R. capsulatus complex I was shown to readily reverse electron flow between the quinone pool and NAD⁺ (7, 9).

Previously, it was established that complex I-deficient mutants of R. capsulatus could not grow photoautotrophically due to overreduction of the quinone pool. That is, electrons donated to ubiquinone from hydrogen oxidation could not be used for NAD⁺ reduction (20). Moreover, mutants of R. capsulatus that lack a functional complex I cannot grow photoheterotrophically with malate as the carbon source; such strains require the addition of an exogenous electron acceptor such as DMSO to overcome this defect (7). Further studies concerning the role of complex I in redox homeostasis under phototrophic growth conditions have not progressed beyond these initial studies (7, 20). During the course of examining the factors involved in regulating CO₂ fixation in R. capsulatus and R. sphaeroides, we isolated mutations that were shown to disrupt complex I. Furthermore, our studies indicated that the control of both carbon and nitrogen metabolism is linked to complex I function in R. capsulatus.

MATERIALS AND METHODS

Bacterial strains and plasmids.

Relevant R. capsulatus and E. coli strains as well as plasmids used or constructed in this study are listed in Table 1.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source or reference
R. capsulatus
B10	Wild type, originally isolated from St. Louis, Mo. (soil)	57
SB1003	Rifr derivative of wild-type strain B10	66
Mal7	Isolated by random transposon mutagenesis, derivative of B10, Kanr	This work
SBC1	nuoF::Gmr derivative of SB1003	This work
E. coli
JM109	recA1 endA1 gyrA96 thi hsdR17 supE44 relA1 Δ(lac-proAB)	64
S17-1λpir	Str Tpr Spcr, λpir and mobilizing factors on chromosome	39
Plasmids
pACYC184	Cmr Tcr, low copy number	4
pUC19	Apr, high copy number	41
pSUP5011	Kmr, suicide vector containing Tn5::mob	47
pRK2013	Kmr, helper plasmid for triparental conjugation	13
PXLB	Tcr, R. capsulatus cbbI translational promoter fusion to lacZ	38
pXFB	Tcr, R. capsulatus cbbII translational promoter fusion to lacZ	38
pUC1318-Gm	pUC1318 containing a Gmr cassette cloned as a 1.6-kb HindIII-SphI fragment from plasmid pFRK-1	S. A. Smith, this laboratory
pJPTC	Tcr, mobilizable suicide vector containing a single EcoRI site located in the multiple cloning site	43
pMa17-1	pACYC184 containing the Tn5::mob site on a 6.0-kb EcoRI fragment of chromosomal DNA from strain Ma17	This work
p19M7-4.8EH	pUC19 containing 3.7 kb of R. capsulatus chromosomal DNA and 1.1 kb of Tn5::mob from plasmid pMa17-1 cloned as a 4.8-kb EcoRI-HindIII fragment	This work
p19M7-3.8ESsp	pUC19 containing 3.7 kb of R. capsulatus chromosomal DNA and ∼0.15 kb of Tn5::mob from plasmid p19M7-4.8EH cloned as a 3.8-kb EcoRI-SmaI fragment into EcoRI and SspI sites	This work
p19M7-2.4ESsp	pUC19 containing 2.3 kb of R. capsulatus chromosomal DNA and ∼0.15 kb of Tn5::mob from plasmid pMa17-1 cloned as a 2.4-kb EcoRI-SspI fragment into EcoRI and SmaI sites	This work
p19-nuoFGM	p19M7-3.8ESsp with a Gmr cassette cloned into the blunt-ended NcoI site of nuoF as a 1.6-kb blunt-ended PstI fragment from plasmid pUC1318-Gm	This work
pJPTC-nuoFGM	p19-nuoFGm cloned into vector pJPTC by linearizing with EcoRI	This work

Transposon mutagenesis.

Transposon mutagenesis was performed on wild-type strain B10 using suicide vector pSUP5011. Mutant selection using R. capsulatus was initiated because it, unlike most nonsulfur purple bacteria, grows well under aerobic chemoautotrophic growth conditions. This additional aerobic autotrophic growth mode was deemed important for isolating genes that function specifically for autotrophy and CO₂ fixation and not general aerobic or anaerobic metabolism. Following transposon mutagenesis, the conjugation mixture was plated onto minimal plates (containing fructose as the carbon source and kanamycin for antibiotic selection) and incubated under dark aerobic conditions; isolated kanamycin-resistant transconjugants were then screened on plates for photoheterotrophic growth (using malate as the carbon source) in the absence or presence of DMSO, as well as chemoautotrophic growth (with CO₂ as the sole carbon source). Strain Mal7 was isolated and selected for further investigation, as its phenotype suggested that it was altered in the regulation of CO₂ fixation.

DNA manipulations and conjugation techniques.

All routine DNA manipulations were carried out following standard protocols (2). R. capsulatus chromosomal DNA was isolated as previously described (15) with the exception that chemoheterotrophic cultures of R. capsulatus were grown in peptone yeast extract (PYE) medium (58) supplemented with 0.4% fructose. Southern blotting and hybridization techniques followed standard protocols (37). Hybridization, labeling of probes, and development of blots were conducted by following protocols of the Vistra ECF fluorescent detection system (Amersham Corporation, Buckinghamshire, England). E. coli strain JM107 (64) containing mobilizable helper plasmid pRK2013 (13) was used in triparental matings (37).

DNA sequencing and analysis.

From chromosomal DNA, an EcoRI fragment containing Tn5::mob was isolated and cloned from strain Mal7, generating plasmid pMal7-1 (Table 1). In this construct, Tn5::mob was flanked by 2.3-kb and 3.7-kb pieces from chromosomal DNA of strain Mal7. The scheme for subcloning and identifying the Tn5::mob transposition site from plasmid pMal7-1 is illustrated in Fig. 1. Nucleotide sequences were determined using an ABI Prism 310 genetic analyzer. A Perkin Elmer Gene Amp PCR System 2400 thermal cycler and Dye Terminator cycle sequencing kit were used as described by the manufacturer (Perkin Elmer, Foster City, Calif.). The M13/pUC sequencing primers forward 17-base sequencing primer (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′) and M13 reverse 24-base primer (5′-AGCGGATAACAATTTCACACAGGA-3′) were used for sequencing reactions (New England Biolabs Inc., Beverly, Mass.).

FIG. 1.

Identification of Tn5 insertion site. To isolate the R. capsulatus DNA flanking the Tn5::mob insertion of the EcoRI fragment of pMal7-1, plasmids p19M7-4.8EH and p19M7-3.8ESsp were generated. Plasmid p19M7-4.8EH contains the 3.7-kb R. capsulatus DNA fragment flanking the Tn5::mob site of insertion, while plasmid p19M7-2.4ESsp contains the 2.3-kb fragment of R. capsulatus DNA from plasmid pMal7-1. Solid lines depict R. capsulatus cloned DNA, while the dotted line indicates Tn5::mob DNA. Restriction sites: E, EcoRI; H, HindIII; S, SspI. Blocked arrows indicate the direction in which nucleotide sequences were obtained using M13/pUC sequencing primers. All subclones were aligned with the genetic map below. The identification and organization of the R. capsulatus genes flanking the Tn5::mob site of insertion were identified by DNA sequencing, with the genes (shown as boxes) designated below. The Tn5::mob site of insertion is in the 5′ part of the gene nuoF; nuoF is the seventh gene of an 18-kb gene cluster encoding 14 known separate subunits of complex I and contains seven unidentified open reading frames as well (6, 9, 19).

Construction of R. capsulatus mutant strain SBC1 by insertional mutagenesis.

Insertional mutagenesis of the nuoF gene of R. capsulatus was performed (Fig. 2). In order to isolate the 3.7-kb chromosomal fragment containing nuoF from low-copy-number plasmid pMal7-1, a 4.8-kb EcoRI-HindIII fragment of plasmid pMal7 (containing 3.7 kb of R. capsulatus chromosomal DNA flanked by 1.1 kb of Tn5::mob DNA) was cloned into plasmid pUC19, generating plasmid p19M7-4.8EH. Plasmid p19M7-3.8ESsp was generated by cloning a 3.8-kb EcoRI-SspI fragment (containing 3.7 kb of R. capsulatus chromosomal DNA flanked by 0.15 kb of Tn5::mob DNA) from plasmid p19M7-4.8EH into the EcoRI and SmaI sites of pUC19.

FIG. 2.

nuo gene cluster of R. capsulatus (A) and sites of gene disruptions in the complex I-deficient strains (B). Solid arrow indicates direction of transcription (A and B). The dotted arrow indicates orientation of the Gm resistance cassette used to inactivate nuoF (B). The genes encoding the 14 subunits of complex I in R. capsulatus (NuoA to NuoN) and seven unidentified open reading frames which might potentially code for factors involved in regulation or subunit assembly (Urf1 to Urf7) are organized in one 18-kb operon (6, 9, 19). EcoRI restriction sites (E) indicate that the 6.0-kb chromosomal fragment isolated from strain Mal7 contains the Tn5::mob site of insertion (A and B). The subclones pMal7-1 and p19-nuoFGm are aligned with the genetic map above. The dotted line indicates orientation of the insertion site of the Gm interposon in nuoF in relationship to the Tn5:: mob insertion (B).

Inactivation of nuoF was accomplished by insertion of a gentamicin resistance cassette as a 1.56-kb PstI-blunt-ended fragment from plasmid pUC1318-Gm into the unique NcoI site of nuoF, blunt-ended in plasmid p19M7-3.8ESsp. This generated plasmid p19-nuoFGm. A cointegrate of plasmid p19-nuoFGm and the suicide plasmid pJPTC was formed after linearizing with EcoRI, resulting in plasmid pJPTC-nuoFGm. E. coli strain S17-λpir (39) was used to mobilize plasmid pJPTC-nuoFGm into R. capsulatus strain SB1003. Of the 1, 100 screened exconjugants, 3 were found to be Gm^r Tet^S and confirmed to be the result of double recombination by Southern blot analysis of chromosomal DNA.

Media and growth conditions.

E. coli strains were grown aerobically on Luria-Bertani medium (2) at 37°C with appropriate antibiotic selection. Phototrophic cultures of R. capsulatus were grown in front of banks of incandescent lights at 30 to 33°C as previously described (11, 37); photoheterotrophic cultures were grown anaerobically in Ormerod's medium (36) supplemented with 0.4% (wt/vol) dl-malate and 1 μg of thiamine/ml. The nitrogen source was provided either as 30 mM ammonia, 6.8 mM l-glutamate, or N₂ gas (95% N₂–5% CO₂). Photoautotrophic cultures were sparged continuously with 1.5% CO₂–98.5% H₂.

Chemoheterotrophic cultures were grown aerobically in the dark in a defined medium containing 0.4% fructose as the carbon source at 30 to 33°C or in PYE medium. The concentrations of antibiotics used for selection of the R. capsulatus strains were as follows: rifampin, 50 μg/ml; kanamycin, 5 μg/ml; trimethoprim, 100 μg/ml; gentamicin (Gm), 10 μg/ml; and tetracycline (Tet), 3 μg/ml for stock cultures or 0.5 μg/ml for plasmid maintenance during phototrophic growth conditions. For E. coli, the antibiotic concentrations for plasmid maintenance were: ampicillin, 100 μg/ml; chloramphenicol, 100 μg/ml; kanamycin, 10 μg/ml; gentamicin, 15 μg/ml; and tetracycline, 6 μg/ml. DMSO was used at a concentration of 30 mM.

Cell extracts and enzyme assays.

Culture samples (10 to 20 ml) from 400-ml bottle cultures (37) were harvested in late exponential phase (optical density at 660 nm [OD₆₆₀] = 0.9 to 1.2), washed in buffer (25 mM Tris-HCl, 1 mM EDTA [pH 8.0]), and disrupted by sonication. The resultant cell debris was removed by centrifugation for 15 min at 18,000 × g at 4°C. β-Galactosidase activity was measured by continuously monitoring the increase in absorbance at 405 nm from the substrate-dependent production of o-nitrophenol (33). Specific activities were calculated using the change in steady-state A₄₀₅ per min (55).

Ribulose 1,5-bisphosphate carboxylase activity was assayed by measuring ribulose 1,5-bisphosphate-dependent ¹⁴CO₂ fixation into acid-stable material (60). Phosphoribulokinase (PRK) activity was measured following previously described protocols (50) except that ribulose 5-phosphate was generated from ribose 5-phosphate by the addition of 5 U of phosphoriboisomerase (Sigma Chemical, St. Louis, Mo.). Protein concentrations were determined using the Bio-Rad protein assay dye-binding reagent (Bio-Rad Laboratories, Hercules, Calif.) using bovine serum albumin as a standard.

RESULTS

Identification of nuoF gene as the site of Tn5::mob transposition.

Subclones derived from plasmid pMal7-1 were generated so that DNA sequence analysis could be performed using M13/pUC forward and reverse sequencing primers (Fig. 1). The site of insertion was determined to be in the 5′ part of the nuoF gene. These sequencing results in strain B10 were compatible with the previous gene order and sequences described by Dupuis et al. (8), also using strain B10.

As shown (Fig. 2), the nuo genes are organized as a single 18-kb gene cluster that encodes 14 subunits of complex I (subunits NuoA to NuoN) (9, 19). The 14 nuo genes of R. capsulatus encode proteins homologous to the 14 major subunits of mitochondrial complex I and complex I from P. denitrificans and E. coli (9, 27, 59). In contrast to the compact nqo operon of P. denitrificans (61) and the nuo operon of E. coli (59), the nuo operon of R. capsulatus contains additional reading frames. These potential gene products are not related to any other subunits of complex I in eukaryotes or bacteria and are speculated to function somehow in regulation and protein assembly (19, 61).

The deduced amino acid sequence of the nuoF product, the site of Tn5::mob insertion in strain Mal7, contains a typical nucleotide-binding sequence motif (GXGXXG) (63). These residues are involved in the binding of NADH. Thus, NuoF is an important component for electron transport to and from the enzyme (63).

Construction of nuoF mutant strains.

To provide confirmation of the phenotype and results obtained with transposon mutant Mal7, strain SBC1 (nuoF) was constructed as described in Materials and Methods in strain SB1003, with the orientation of the Gm resistance cassette depicted (Fig. 2). The Gm interposon insertion in nuoF was situated in position 5403 of the nuo operon (at an NcoI site), approximately 190 bp downstream from the Tn5::mob transposition site (Fig. 2).

Analysis of complex I-deficient mutants of R. capsulatus.

Strain Mal7 could not grow photoautotrophically (1.5% CO₂–98.5% H₂) or chemoautotrophically (H₂–CO₂–air) as well as photoheterotrophically on malate in the absence of DMSO (Table 2). The nuoF site-directed mutant of R. capsulatus, strain SBC1 showed a phototrophic growth phenotype comparable to that of strain Mal7 (Table 2). Photoautotrophic and chemoautotrophic cultures of complex I-deficient strains were supplemented with DMSO, since the addition of DMSO to photoheterotrophic cultures rescued the growth of these strains. Distinct from photoheterotrophic growth conditions, strains Mal7 and SBC1 were not able to achieve growth under autotrophic conditions when DMSO was added. By contrast, wild-type strains B10 and SB1003 grew under all conditions examined (Table 2), and all strains (mutants and wild type) could grow under dark aerobic conditions with either fructose or malate as the carbon source (Table 2).

TABLE 2.

Growth phenotypes of complex I-deficient strains (Ma17 and SBC1) compared to those of wild-type strains (B10 and SB1003) of R. capsulatus

Growth conditions	Growth of strain:
	B10	Ma17	SB1003	SBC1
Chemoheterotrophic
Fructose/O2	+	+	+	+
Malate/O2	+	+	+	+
Chemoautotrophic
CO2/H2/O2	+	−	+	−
CO2/H2/O2/DMSO	+	−	+	−
Photoheterotrophic
Malate	+	−	+	−
Malate/DMSO	+	+	+	+
Photoautotrophic
CO2/H2	+	−	+	−
CO2/H2/DMSO	+	−	+	−

Thus, it is apparent that complex I-deficient strains of R. capsulatus were not able to grow under conditions in which the CBB pathway is required to maintain the redox poise of the cell (photoheterotrophic growth in the absence of DMSO) or under conditions in which CO₂ assimilation provides all the carbon for subsequent cell metabolism (chemo- or photoautotrophic growth). This phenotype is comparable to CBB-deficient strains of R. capsulatus (38, 50) and ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO)-deficient strains of Rhodobacter sphaeroides (11).

Unlike R. sphaeroides, CBB-deficient strains of R. capsulatus do not require an exogenous electron acceptor (DMSO) for photoheterotrophic growth when glutamate is used as the nitrogen source (50). When CBB-deficient strains of R. capsulatus are supplemented with glutamate, the dinitrogenase system is induced, presumably providing a sufficient compensatory electron sink to support photoheterotrophic growth in the absence of a functional CBB cycle. Thus, it was of interest to determine if the dinitrogenase system could support growth of complex I-deficient strains in the absence of DMSO (Table 3).

TABLE 3.

Photoheterotrophic growth with various nitrogen sources in the absence and presence of the ancillary electron acceptor DMSO

Carbon/nitrogen source	Redox balancing system(s) used	Growth of strain:
		B10	Ma17	SB1003	SBC1
Malate/ammonia	CBB	+	−	+	−
Malate/ammonia/DMSO	CBB, DMSOR	+	+	+	+
Malate/glutamate	CBB, nitrogenase	+	−	+	−
Malate/glutamate/DMSO	CBB, nitrogenase, DMSOR	+	+	+	+
Malate/N2	CBB, nitrogenase	+	−	+	−
Malate/N2/DMSO	CBB, nitrogenase, DMSOR	+	+	+	+

Unlike wild-type strains B10 and SB1003, strains Mal7 and SBC1 could not grow in the absence of DMSO with either glutamate or N₂ as the supplied nitrogen source, whereas growth was rescued when cultures were supplemented with DMSO (Table 3). These findings are in stark contrast to a complex I-deficient strain of R. capsulatus generated by Herter et al. (19), which was reported to be able to grow photoheterotrophically in the absence of DMSO (even under nitrogen-fixing growth conditions). Our results do support the findings of Dupuis et al. (7), who suggested that complex I is absolutely essential for the maintenance of redox homeostasis, as previously isolated mutants of complex I (7), as well as the mutants in this study, could not grow under photoheterotrophic conditions in the absence of DMSO.

Control of CBB system in complex I-deficient strains Mal7 and SBC1.

The levels of activity for the key enzymes of the CBB pathway, RubisCO and PRK, of strain Mal7 were determined, and strain Mal7 exhibited levels of activity similar to the wild-type strain B10 during photoheterotrophic growth with DMSO (data not shown). Likewise, levels of RubisCO and PRK activity for strain SBC1 were comparable to those of wild-type strain SB1003 during photoheterotrophic growth with DMSO, and typical enhanced levels for these enzymes (25, 46) were found in wild-type strains under photoautotrophic growth conditions (1.5% CO₂–98.5% H₂) (data not shown).

Complex I-deficient strains Mal7 and SBC1 maintained wild-type control of the cbb_I promoter during photoheterotrophic growth supplemented with DMSO when either ammonia or glutamate was used as the nitrogen source; moreover, levels of cbb_II promoter activity in strain Mal7 were similar to those in wild-type strain B10, and the levels of cbb_II promoter activity in strain SBC1 were comparable to those in its wild-type strain (SB1003) during photoheterotrophic growth with DMSO (Table 4). Thus, like the levels of RubisCO and PRK activity, complex I-deficient strains exhibited wild-type control of the expression of the two CBB pathway operon promoters when grown photoheterotrophically in the presence of DMSO.

TABLE 4.

cbb_I and cbb_II promoter activity measured in wild-type and complex I-deficient strains^a

Growth conditions	β-Galactosidase activity (nmol/min/mg)
	B10 or SB1003		Ma17 or SBC1
	cbbI::lacZ	cbbII::lacZ	cbbI::lacZ	cbbII::lacZ
Photoheterotrophic (B10, Ma17)
Malate/ammonia/DMSO	0	297 ± 35	0	189 ± 26
Malate/glutamate/DMSO	0	342 ± 24	0	287 ± 23
Photoheterotrophic (SB1003, SBC1)
Malate/ammonia/DMSO	0	141 ± 24	0	146 ± 16
Malate/glutamate/DMSO	0	80 ± 7	0	98 ± 12

^aβ-Galactosidase activity was determined from three to four independent cultures assayed in duplicate. Strains were grown under photoheterotrophic conditions with either ammonia or glutamate as the nitrogen source in the presence of DMSO. 

Consistent with the established regulation of the cbb_I operon in R. capsulatus (38, 53), wild-type strains B10 and SB1003 did not exhibit cbb_I promoter activity under photoheterotrophic growth conditions in the absence or presence of DMSO, regardless of the nitrogen source; cbb promoter levels were significantly enhanced, as expected, in wild-type strain B10 after photoautotrophic (1.5% CO₂–98.5% H₂) growth (data not shown).

Obviously, the inability to grow photoautotrophically precludes measurements of either RubisCO/PRK or cbb promoter activity levels in complex I-deficient mutants. Previous results indicated that mere incubation of photoheterotrophically grown photoautotrophic-incompetent strains in an organic carbon-free, photoautotrophic (1.5% CO₂–9.5% H₂) gas environment is sufficient to enable the CBB enzymes and promoters to be induced in the absence of growth (38, 53).

To determine if the complex 1 mutants could induce these enzymes and promoters, the time course for induction of β-galactosidase activity was determined using cbb_I::lacZ and cbb_II::lacZ promoter fusions in wild-type strain SB1003 and mutant strain SBC1 following transition to a photoautotrophic (1.5% CO₂–98.5% H₂–ammonia) environment. The transition to a photoautotrophic growth environment resulted in induction, and a constant level of cbb_I promoter activity and cbb_II promoter activity in wild-type strain SB1003, while cbb_I promoter activity was not induced in strain SBC1 and cbb_II promoter activity was maintained at levels obtained during photoheterotrophic growth in the presence of DMSO (Fig. 3A). Similar results were obtained when total RubisCO and PRK activity was compared in strains B10 and Mal7 (Fig. 3B). It appears that complex I-deficient strains of R. capsulatus were deficient in their ability to transduce the signal that influences the CBB system to be induced under photoautotrophic growth conditions.

FIG. 3.

Time course of β-galactosidase activity induction using form I promoter fusion (cbb_I::lacZ) and form II promoter fusion (cbb_II::lacZ) (A) and total RubisCO and PRK activity (B) following the transition to a photoautotrophic (1.5% CO₂–98.5% H₂–ammonia) environment. Late-exponential-phase (OD₆₆₀ = 0.9 to 1.2) malate-grown photoheterotrophic cultures of wild-type strain SB1003 and strain SBC1 (A) or wild-type strain B10 and strain Mal7 (B) grown with malate-ammonia-DMSO were washed three times with minimal medium and incubated photoautotrophically (1.5% CO₂–98.5% H₂) in the absence of the exogenous electron acceptor DMSO. At time zero, incubation began under photoautotrophic conditions; wild-type strains grew further under these conditions.

DISCUSSION

During photosynthetic growth of Rhodobacter, the interplay between specific metabolic redox-balancing mechanisms (such as the CBB, dinitrogenase, and DMSOR systems, among others) is important to maintain the redox poise of the cyclic electron transport chain (12, 16, 23, 30, 45, 50, 56). This aspect of the physiology of Rhodobacter will undoubtedly contribute to an understanding of interactive control of the diverse metabolic capabilities exhibited by these organisms.

The cyclic photosynthetic electron transport chain of Rhodobacter begins with absorption of light, leading to the oxidation of P-870 and subsequent reduction of ubiquinone in the photosynthetic reaction center. Following this, electron transport through the ubiquinone pool, cytochrome b/c₁ complex, and cytochrome c₂ completes the cycle (for a review, see reference 52 and citations therein). The actual interaction between cyclic electron transport and redox poise was suggested to occur at the level of the ubiquinone pool, since cyclic photosynthesis requires oxidized ubiquinone as an electron acceptor. An overreduction of the ubiquinone pool has been postulated to potentially saturate the photosynthetic pathway (8, 12). Thus, complex I contributes to maintenance of the redox state of the ubiquinone pool by transferring excess reducing equivalents to NAD⁺, generating NADH (7, 20). In this way, complex I functions mainly as an electron sink by keeping part of the ubiquinone pool oxidized, allowing cyclic photosynthesis to continue. Excess reducing equivalents are typically dissipated by CO₂ assimilation via the CBB system or, in some instances, by the dinitrogenase system. Indeed, the control of carbon and nitrogen metabolism and redox homeostasis is linked in Rhodobacter (23, 50). From the results presented here, it is apparent that complex I plays a major role in balancing reducing equivalents generated through carbon and nitrogen metabolism and redox homeostasis in R. capsulatus.

Mutants lacking a functional complex I were not able to grow under photoautotrophic conditions (Table 2) or under photoheterotrophic conditions in the absence of an external electron acceptor (Table 2). This is because electrons transferred to ubiquinone from the oxidation of hydrogen (photoautotrophic metabolism) or from the oxidation of carbon substrates such as l-malate and succinate (photoheterotrophic metabolism) could not be used for NAD⁺ reduction by complex I (7, 20). When substrates such as l-malate are added to cultures, CO₂ is produced from malate metabolism, and this “metabolic” CO₂ subsequently serves an important function, as it is the preferred electron acceptor for reducing equivalents produced during carbon oxidation. Thus, the CBB system, rather than serving as a major means for generating organic carbon (as in photoautotrophic metabolism), plays more of a role to balance the oxidation-reduction potential of the cell during photoheterotrophic growth (12, 26, 56). Therefore, in the absence of a functional complex I enzyme complex, reducing equivalents cannot be provided for CO₂ fixation through the CBB system.

CBB-deficient strains of R. sphaeroides (11, 16, 17) and R. capsulatus (38, 50) exhibited a phenotype similar to that of complex I-deficient strains. Indeed, nuo (Table 2) and cbb (38, 50) disruption mutants were unable to achieve photoheterotrophic growth with a fixed nitrogen source in the absence of an exogenous electron acceptor such as DMSO.

Rhodobacter species are especially sensitive to imbalances in the redox state (30). The permissive effect of DMSO respiration through the DMSOR system, to allow both complex I- and CBB-deficient strains to grow photoheterotrophically, suggests that both types of mutants were affected by an excess of reducing equivalents, most likely generated at the level of the ubiquinone pool. Since the DMSOR system serves to directly sustain the redox poise of the ubiquinone pool for photosynthetic electron transport (30, 31, 45), it is not surprising that the addition of DMSO rescues the growth of both CBB- and complex I-deficient strains. Thus, both the CBB system and complex I are required for redox homeostasis during photoheterotrophic growth conditions in the absence of ancillary electron acceptors.

However, there is a major difference between CBB- (50) and complex I-deficient strains of R. capsulatus in that complex I-deficient strains were unable to grow during photoheterotrophic conditions permissive for the dinitrogenase system (Table 3). Growth was impossible for complex I-deficient strains when glutamate or N₂ was used as the nitrogen source (Table 3). Previous work established that the dinitrogenase system serves as a compensatory electron sink in R. capsulatus in the absence of a functional CBB system (50). In fact, dinitrogenase-catalyzed proton reduction by the hydrogenase-like activity of the dinitrogenase system requires a major energy commitment by the cell, and this activity contributes extensively to redox homeostasis (21, 48). The dinitrogenase system is such an efficient repository of excess reducing equivalents that CBB-deficient strains of R. sphaeroides and Rhodospirillum rubrum (23), plus R. capsulatus (50), derepress synthesis of the dinitrogenase system and abrogate normal control mechanisms in order to achieve photoheterotrophic competency in the absence of exogenous electron acceptors.

The addition of DMSO to cultures of complex I-deficient strains incubated in an ammonia-free environment rescued growth under nitrogen-fixing conditions (Table 3). Whether the permissive effect of DMSO respiration on the ability of complex I-deficient strains to grow under nitrogen-fixing conditions was due to balancing of the oxidation-reduction state of the ubiquinone pool or to sufficient electron flow to ferredoxin I (electron donor to nitrogenase) (24) remains to be determined. In any event, complex I function appeared to be linked to nitrogen metabolism in R. capsulatus, unlike the findings in a previous report (20).

Control mechanisms involved in mediating the function of the nuo operon of R. capsulatus remain to be established. Transcriptional regulation of the nuoA-N gene locus of E. coli involves regulation by the global two-component regulatory system ArcB/A as well as by FNR (references 3 and 51 and citations therein). In R. capsulatus, a global two-component signal transduction system, RegB/A (PrrB/A), has been shown to play a role in the regulation of key operons involved in photosynthesis, including operons that encode structural genes for the light-harvesting complexes (puf and puc) and the photosynthetic reaction center (puh) involved in photosynthetic gene expression (22, 34). In addition, the Reg/Prr system has been implicated in controlling CO₂ fixation (43, 53) as well as nitrogen fixation and H₂ oxidation (10, 23, 44). Thus, during photosynthetic conditions, the Reg/Prr system regulates components of the cyclic electron transport chain as well as systems involved in redox homeostasis.

Complex I appears to provide linkage between the poise of the photosynthetic electron transport chain while balancing the means by which reducing equivalents are generated during carbon metabolism. Conceivably, the Reg/Prr system might also play a role in regulating the nuo operon of R. capsulatus and R. sphaeroides. Additionally, the Reg/Prr system has been shown to be involved in the activation of cbb_I promoter expression as well as maximal expression of the cbb_II promoter under photoautotrophic growth conditions in R. capsulatus (53). Complex I-deficient strains appeared to be blocked in a necessary signal for induction of cbb expression (Fig. 3). Perhaps a redox signal (such as the flow of reductant through the ubiquinone pool) is transduced through a pathway that involves complex I and is then conveyed to the Reg/Prr signal transduction pathway. Possibly the absence of a functional complex I enzyme complex disrupted a necessary redox signal required by the Reg/Prr system, which in turn affected cbb induction.

In R. sphaeroides, the cbb3 cytochrome c oxidase senses the redox state of the quinone pool and transduces a signal to the Reg/Prr system, which in turn regulates expression of photosynthesis genes in response to O₂ (35). Obviously, further studies are required to elucidate a potential linkage between the Reg/Prr signal transduction regulatory system and complex I in R. capsulatus during photoheterotrophic and photoautotrophic growth conditions. Additional control mechanisms could also be involved in mediating the function of the nuo operon of R. capsulatus. Indeed, a characteristic LysR-type consensus DNA-binding motif (T-N₁₁-A) (14) was localized as an inverted repeat separated by a 21-bp spacer in the promoter region of the nuo operon of R. capsulatus (data not shown), raising the possibility that LysR-type transcriptional regulators could also be involved in controlling nuo expression in R. capsulatus. More detailed studies need to be done to address these possibilities.