Mobile Cytochrome c₂ and Membrane-Anchored Cytochrome c_y Are Both Efficient Electron Donors to the cbb₃- and aa₃-Type Cytochrome c Oxidases during Respiratory Growth of Rhodobacter sphaeroides

Abstract

We have recently established that the facultative phototrophic bacterium Rhodobacter sphaeroides, like the closely related Rhodobacter capsulatus species, contains both the previously characterized mobile electron carrier cytochrome c₂ (cyt c₂) and the more recently discovered membrane-anchored cyt c_y. However, R. sphaeroides cyt c_y, unlike that of R. capsulatus, is unable to function as an efficient electron carrier between the photochemical reaction center and the cyt bc₁ complex during photosynthetic growth. Nonetheless, R. sphaeroides cyt c_y can act at least in R. capsulatus as an electron carrier between the cyt bc₁ complex and the cbb₃-type cyt c oxidase (cbb₃-C_ox) to support respiratory growth. Since R. sphaeroides harbors both a cbb₃-C_ox and an aa₃-type cyt c oxidase (aa₃-C_ox), we examined whether R. sphaeroides cyt c_y can act as an electron carrier to either or both of these respiratory terminal oxidases. R. sphaeroides mutants which lacked either cyt c₂ or cyt c_y and either the aa₃-C_ox or the cbb₃-C_ox were obtained. These double mutants contained linear respiratory electron transport pathways between the cyt bc₁ complex and the cyt c oxidases. They were characterized with respect to growth phenotypes, contents of a-, b-, and c-type cytochromes, cyt c oxidase activities, and kinetics of electron transfer mediated by cyt c₂ or cyt c_y. The findings demonstrated that both cyt c₂ and cyt c_y are able to carry electrons efficiently from the cyt bc₁ complex to either the cbb₃-C_ox or the aa₃-C_ox. Thus, no dedicated electron carrier for either of the cyt c oxidases is present in R. sphaeroides. However, under semiaerobic growth conditions, a larger portion of the electron flow out of the cyt bc₁ complex appears to be mediated via the cyt c₂-to-cbb₃-C_ox and cyt c_y-to-cbb₃-C_ox subbranches. The presence of multiple electron carriers and cyt c oxidases with different properties that can operate concurrently reveals that the respiratory electron transport pathways of R. sphaeroides are more complex than those of R. capsulatus.

The purple nonsulfur facultative phototrophic bacteria of Rhodobacter species, Rhodobacter capsulatus and its close relative Rhodobacter sphaeroides, are excellent model organisms for studying cellular energy transduction (46). They have versatile growth modes and are capable of both anoxygenic phototrophic (Ps) and vigorous respiratory (Res) growth in the presence of oxygen. They contain a photochemical reaction center (RC) (43, 44), a ubihydroquinone:cytochrome c (cyt c) oxidoreductase (cyt bc₁ complex) (6, 45), and a c-type cytochrome as a mobile electron carrier (7, 8), but their Ps and Res electron transport chains are different (46, 47).

During the Ps growth of R. capsulatus, electrons are conveyed from the cyt bc₁ complex to the RC either by the well-studied mobile electron carrier cyt c₂ (7) or the more recently discovered membrane-anchored electron carrier cyt c_y (17), encoded by cycA and cycY, respectively. On the other hand, in R. sphaeroides, the Ps electron transport pathway can be mediated only by a mobile electron carrier like cyt c₂ (8) or its functional analogues such as isocyt c₂ (33) (Fig. 1). Although Ps-grown cells of R. sphaeroides also harbor a membrane-anchored cyt c_y, this electron carrier is not able to support Ps growth (27). Yet the Ps growth inability of R. sphaeroides mutants lacking cyt c₂ can be restored readily by genetic introduction of R. capsulatus cyt c_y (17, 18). These findings indicate that R. sphaeroides cyt c_y is the culprit for the Ps growth inability of this species in the absence of cyt c₂ (27), but the molecular basis of this observation remains unknown.

FIG. 1.

Photosynthetic and respiratory electron transport pathways of R. sphaeroides. DH, Q_pool, bc₁, RC, Q_ox, cyt c₂, cyt c_y, aa₃-C_ox, and cbb₃-C_ox correspond to the respiratory NADH and succinate dehydrogenases, membrane quinone pool, cyt bc₁ complex, photochemical reaction center, hydroquinone oxidase, soluble electron carrier cyt c₂, membrane-attached electron carrier cyt c_y, aa₃-type cyt c oxidase, and cbb₃-type cyt c oxidase, respectively. Dotted arrows highlight the electron transfer pathways catalyzed by cyt c_y. Antimycin A (Ant A) and myxothiazol (Myx) are inhibitors of the cyt bc₁ complex.

Respiratory electron transport pathways of R. capsulatus and R. sphaeroides are also similar but not identical (46) (Fig. 1). Both species contain a respiratory electron transport pathway that is branched after the quinone pool. One of these branches contains hydroquinone oxidases (Q_ox) commonly found in microbes, while the other one harbors a mitochondrial-like cyt c oxidase(s) (C_ox) converting oxygen to water (13). While R. capsulatus contains only one cbb₃-type C_ox (cbb₃-C_ox) (14, 19), both an aa₃-type C_ox (aa₃-C_ox) (16, 36), and a cbb₃-C_ox (12, 42) are present in R. sphaeroides. In the latter species, the cbb₃-C_ox is predominant in microaerobic or Ps-grown cells, while the aa₃-C_ox becomes a major component under high O₂ tension (37). In both species, the nature of the Q_ox is not well known. Recent works including genome sequences (4, 11) suggest the presence of one Q_ox in R. capsulatus (21, 48; K. Zhang and F. Daldal, unpublished data) and possibly two Q_ox in R. sphaeroides.

Recently, we have established that R. sphaeroides cyt c_y can function as an electron carrier between the cyt bc₁ complex and the cbb₃-C_ox in R. capsulatus (27). This finding raised the issue of whether it can also act as an electron carrier during the Res growth of R. sphaeroides and, if so, whether it can donate electrons to both the aa₃-C_ox and the cbb₃-C_ox. Obviously, the presence of two c-type electron carriers and two cyt c oxidases complicates this analysis. To facilitate it, appropriate mutations inactivating either cyt c₂ or cyt c_y were introduced into R. sphaeroides strains lacking either cbb₃-C_ox or aa₃-C_ox activity. The double mutants thus obtained converted the crisscrossed respiratory electron transport pathways into simpler linear pathways (Fig. 1) and allowed estimation of their efficiencies. The data demonstrated for the first time that in R. sphaeroides either cyt c₂ or cyt c_y can act as an electron carrier between the cyt bc₁ complex and either the cbb₃-C_ox or the aa₃-C_ox. Furthermore, they indicated that electron flow via the cyt c₂ → cbb₃-C_ox and cyt c_y → cbb₃-C_ox subbranches appears to be greater under semiaerobic growth conditions. During these studies, it was noted that in the absence of the latter respiratory pathway, light-harvesting protein complexes, which are usually absent in the presence of oxygen, were also produced.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Escherichia coli strains were grown on Luria broth (LB medium) supplemented with kanamycin, spectinomycin, or tetracycline at a final concentration of 50, 50, or 12.5 μg per ml, respectively. R. sphaeroides strains were grown in Sistrom's minimal medium (Med A) (40) supplemented as needed with 10, 10, or 2.5 μg of spectinomycin, kanamycin, or tetracycline, respectively, per ml. For semiaerobic growth, 2-liter flasks were filled with 1 liter of growth medium and shaken at 150 rpm with a rotary shaker (New Brunswick Scientific Co. Inc., Rutherford, N.J.). Photo- or chemoheterotrophic growth phenotypes and ability to catalyze the Nadi reaction (N′,N′-dimethyl-p-phenylenediamine + α-naphthol → indophenol blue + H₂O) (24) were determined on solid media using cells grown in anaerobic jars containing H₂- and CO₂-generating gas packs (BBL) or an aerobic dark incubator at 35°C in the presence of atmospheric oxygen. Growth rates of various strains were determined using liquid cultures inoculated with freshly grown cells at appropriate dilutions. As needed, cultures were first incubated overnight at ambient temperature without shaking and then transferred to a rotary shaker at 35°C, and the turbidity was monitored using side-arm flasks and a Klett-Summerson colorimeter. All strains and plasmids used are listed in Table 1. The parental R. sphaeroides strain Ga (40) is referred to as wild type with respect to its cytochrome profile and its Ps and Res growth properties.

TABLE 1.

Bacterial strains and plasmids used in this work

Strain or plasmid	Genotype	Relevant phenotype	Source or reference
E. coli
HB101	F proA2 hsdS20 recA1 ara-14 lacY1 galK2 rpsL20 supE44 xyl-5 mtl-1	rB− mB−	Stratagene
S17.1	thi pro hsdR recA ara RP4::2-Tc::Mu::kan:Tn7 integrated into chromosome	rB− mB+ Tpr Smr	39
R. sphaeroides
Ga	crt	Wild type	40
Gadc2	crt Δ(cycA::spe)	Cyt c2−	3
Gadcy	crt Δ(cycY::kan)	Cyt cy−	17
Gadc2cy	crt, Δ(cycA::spe) Δ(cycY::kan)	Cyt c2− Cyt cy−	17
MT001	crt Δ(ccoNO::kan)	cbb3-Cox−	42
KD02	crt Δ(cycA::spe) Δ(ccoNO::kan)	Cyt c2−cbb3-Cox−	This work
KD04	crt Δ(cycY::spe) Δ(ccoNO::kan)	Cyt cy−cbb3-Cox−	This work
JS100	crt Δ(ctaD::spe)	aa3-Cox−	36
KD03	crt Δ(cycY::kan) Δ(ctaD::spe)	Cyt cy−aa3-Cox−	This work
KD05	crt Δ(cycA::kan) Δ(ctaD::spe)	Cyt c2−aa3-Cox−	This work
ME127	crt Δ(ctaD::spe) Δ(ccoNO::kan)	cbb3-Cox− aa3-Cox−	42
BC17	crt Δ(petABC(fbcFBC)::kan)	Cyt bc1−	45
Plasmids
pBSII	pBlueScript II KS(+)	Ampr	Stratagene
pHP45Ω	spe cartridge	Sper	31
pHP45Ω-Km	kan cartridge	Kanr	9
pSup202	Suicide plasmid	Tetr Cmr Ampr	39
p4.2	cycY::kan in PstI site of pSup202		This work
pcycA::spe	cycA::spe in PstI site of pSup202		This work
pKD10.2	2-kb spe cartridge of pHP45Ω in BamHI site of cycY carried by pU2000 (in opposite orientation of cycY)		This work
pU2000	1.8-kb BglII-XbaI R. sphaeroides chromosomal DNA carrying cycY in pBSII KS(+)		51
pKD11	3 kb cycY::spe fragment in EcoRI site of pSup202		This work
pSupC2 P2.71::kan	Δ(cycA::kan) at StuI site of cycA carried at PstI site of pSup202		8

Molecular genetic techniques and construction of various R. sphaeroides mutants.

Two different insertion-deletion alleles of cycA and cycY, carrying either spectinomycin resistance (Spe^r) or kanamycin resistance (Kan^r)-conferring gene cartridges, were necessary for constructing various double mutants lacking cyt c₂ or cyt c_y and cbb₃-C_ox or aa₃-C_ox. Thus, a cycY::kan allele was constructed similarly to cycY::spe described previously (27), except that a Kan^r-conferring cartridge instead of a Spe^r-conferring cartridge from pHP45Ω-K (9) was used. The desired allele was transferred into the suicide plasmid pSup202, conferring tetracycline resistance (Tet^r), and via the E. coli donor strain S17.1 (39) conjugated into the appropriate R. sphaeroides strains, selecting for Spe^r or Kan^r as required. Among the transconjugants, those that were Tet^s, and hence carrying a chromosomal allele replacement via a double-crossover event, were retained for further analyses (3). Introduction of appropriate cycA and cycY alleles into the R. sphaeroides Δ(ccoNO::kan) (MT001) and Δ(ctaD::spe) (JS100) mutants yielded the Δ(cycA::spe) Δ(ccoNO::kan) (KD02) and Δ(cycY::spe) Δ(ccoNO::kan) (KD04) and the Δ(cycY::kan) Δ(ctaD::spe) (KD03) and Δ(cycA::kan) Δ(ctaD::spe) (KD05) double mutants, respectively (Table 1).

Biochemical and spectroscopic techniques.

R. sphaeroides chromatophore membranes from washed cells grown semiaerobically in Med A were prepared using a French pressure cell at 18,000 lb/in² and ultracentrifugation either in 50 mM 3-(N-morpholino)propanesulfonic acid (MOPS) buffer (pH 7.0) containing 1 mM KCl, EDTA, and phenylmethylsulfonyl fluoride for polyacrylamide gel electrophoresis (PAGE) analyses (18, 20) or in 50 mM MOPS buffer (pH 7.2) containing 5 mM MgCl₂ for spectroscopic studies (50). Membrane fragments were washed twice, resuspended at a known protein concentration in the same buffer, and stored frozen at −80°C until further use. Protein content of the samples was determined by the method of Bradford (2) or Lowry et al. (22), using bovine serum albumin as a standard. Bacteriochlorophyll concentration was measured by the optical absorption of acetone-methanol (7:2, vol/vol) extracts, using an extinction coefficient ɛ₇₇₅ of 75 mM⁻¹ cm⁻¹ (5). Sodium dodecyl sulfate (SDS)-PAGE was performed using 16.5% acrylamide Tris-Tricine gels as described by Schägger and von Jagow (35). Samples were denatured for 5 min at 37°C in SDS loading buffer prior to electrophoresis, and gels were stained with Coomassie brilliant blue to visualize the polypeptides. The c-type cytochromes were revealed via the intrinsic peroxidase activity of their heme group, using 3,3′,5,5′-tetramethylbenzidine (TMBZ) and H₂O₂ (41). The amounts of cytochromes in chromatophore membranes were estimated by recording at room temperature reduced (with ascorbate or dithionite)-minus-oxidized (with ferricyanide) optical difference spectra. Either a Hitachi 3400 or Jasco 7800 spectrophotometer and the extinction coefficients ɛ_604–630 of 23 mM⁻¹ cm⁻¹, ɛ_561–575 of 22 mM⁻¹ cm⁻¹, and ɛ_551–540 of 29 mM⁻¹ cm⁻¹ were used for a-, b-, and c-type cytochromes, respectively.

Cyt c oxidase activities in membrane fragments were determined using either reduced horse heart cyt c as a substrate or by monitoring O₂ consumption. In the first case, immediately prior to the assay, horse heart cyt c was reduced by addition of sodium dithionite, excess of which was eliminated via chromatography through a small column containing Sephadex G-10 as a matrix (14, 20). Oxidation of cyt c was monitored at 550 nm in a stirred cuvette, thermostated at 25°C, containing 50 mM Tris (pH 8.0) buffer supplemented with 0.1 mg of dodecyl maltoside per ml, 5 μM myxothiazol (Myx), and 40 μM reduced cyt c. As needed, 2 mM potassium cyanide (KCN) was used as an inhibitor (34). In the second case, respiratory rates were measured by polarography at 28°C with a Clark-type oxygen electrode (Yellow Springs Instruments Inc., Yellow Springs, Ohio). These assays used 0.25 to 0.30 mg of membrane fragments per ml resuspended in 50 mM MOPS buffer (pH 7.2) containing 5 mM MgCl₂, 2 mM sodium ascorbate, and 200 μM N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) or 50 μM horse heart cyt c (15, 48). As needed, 100 μM KCN was used to determine the amount of C_ox-independent O₂ consumption during the assay, and enzymatic activities were expressed in moles of substrate consumed per minute per milligram of protein. TMPD was also substituted with 2,3,5,6-tetramethyl-1,4-phenylenediamine (DAD) or 3,6-dichlorophenol indophenol (DCIP) as an electron donor and KCN with NaN₃ as an inhibitor (34). Kinetic changes in the absorption of c-type cytochromes were monitored at 551 minus 540 nm at 28°C, using a Jasco V-550 dual-wavelength spectrophotometer equipped with a Jasco ETC-505T rapid-mixing and temperature control apparatus (15).

To estimate amounts of the light-harvesting complexes in membrane fragments of various R. sphaeroides mutants, 50-ml cultures were grown aerobically with vigorous shaking in 250-ml flasks to an optical density at 630 nm of less than 0.4. Cells were collected by centrifugation and resuspended in 50 mM potassium phosphate buffer (pH 7.0) containing 10 mM EDTA and leupeptin, pepstatin A, and Pefabloc SC (Roche Inc., Indianapolis, Ind.) as protease inhibitors at concentrations suggested by the supplier. After addition of 1/50 volume of lysozyme (10 mg/ml) and 1 h of incubation on ice, cells were frozen at −80°C for 10 min, thawed, and sonicated on ice for 5 min at a 40% relative output in a sonicator (model 150; Dynatech Laboratories Inc., Farmingdale, N.Y.). Cell debris was removed by centrifugation at 13,000 rpm for 30 min at 4°C, and the supernatants were centrifuged at 150,000 × g for 1 h. Membrane fragments thus obtained were washed with 50 mM potassium phosphate (pH 7.0) buffer containing 1 mM EDTA, resuspended in the same buffer, and stored at −80°C until further use. Optical absorption spectra were recorded between the wavelengths of 380 and 1,000 nm with a Beckman DU640 spectrophotometer, using an amount of R. sphaeroides membranes corresponding to a total protein content of 0.2 mg per ml in 10 mM potassium phosphate (pH 7.2) buffer containing 1 mM EDTA.

Chemicals.

All chemicals, including the redox mediators, were of analytical grade and obtained from commercial sources as described previously (15). n-Dodecyl β-d-maltoside was from Anatrace; horse heart cyt c (type VI) and TMPD were from Sigma-Aldrich Chemical Co. (St. Louis, Mo.).

RESULTS

R. sphaeroides strain Ga cells grown under semiaerobic conditions contain both cbb₃-C_ox and aa₃-C_ox activities.

Both cbb₃-C_ox and aa₃-C_ox activities in the wild-type R. sphaeroides strain Ga can be fully inhibited by cyanide (CN⁻) and azide (N₃⁻) anions. However, while CN⁻ affects both C_ox activities similarly, N₃⁻ inhibits them differently (34). Thus, to assess whether cells grown under semiaerobic conditions contained both of these enzymes, ascorbate/cyt c-dependent C_ox activity was measured in the presence of increasing concentration of these inhibitors, using membranes of wild-type strain Ga (Fig. 2). The cbb₃-C_ox activity was totally inhibited by 100 μM N₃⁻, whereas the aa₃-C_ox activity required 2 to 3 mM N₃⁻ for full inhibition. Moreover, the inhibition by CN⁻ was monophasic (90% inhibition at 20 μM KCN), while that by N₃⁻ was biphasic, with a decrease in slope around 100 μM NaN₃, corresponding to about 60% inhibition of total activity. Thus, the R. sphaeroides cells used in this study contained both of the C_ox activities simultaneously.

FIG. 2.

Inhibition of the cyt c oxidase activity found in R. sphaeroides membranes. Inhibitory effects of CN⁻ (circles) and N₃⁻ (triangles) on ascorbate and cyt c oxidase activities found in membrane fragments of R. sphaeroides wild-type strain Ga were determined as described in Materials and Methods. aa₃ and cbb₃ correspond to the aa₃-C_ox and cbb₃-C_ox, respectively. See the text for details.

R. sphaeroides mutants with linear, c-type cytochrome-dependent respiratory electron transport pathways.

In R. sphaeroides, c-type cytochromes containing branches of the respiratory electron transport pathways are complicated due to the presence of the cyt c₂, cyt c_y, cbb₃-C_ox, and aa₃-C_ox (Fig. 1). To simplify these crisscrossed pathways, four double mutants containing only one electron carrier (cyt c₂ or cyt c_y) and one C_ox (cbb₃-C_ox or aa₃-C_ox) were obtained (Table 1). Growth phenotypes of these mutants confirmed that among the electron transport components mutated, only cyt c₂ was required for Ps growth (8) (Table 2). Their doubling times under Res growth conditions (about 100 to 120 min on Med A at 35°C) were similar to those of their parents, with the exception of the cbb₃-C_ox⁻ aa₃-C_ox⁻ double mutant (ME127), which grew slightly slower (doubling time of approximately 150 min). In particular, Res growth of the cyt bc₁⁻ (BC17), cyt c₂⁻ cyt c_y⁻ (Gadc2cy), and cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127) mutants were consistent with the presence of an alternate respiratory pathway, involving a Q_ox (13). It was also noted that mutants lacking the cbb₃-C_ox (MT001, KD02, and especially KD04) formed highly pigmented colonies under respiratory growth conditions.

TABLE 2.

Electron transport chain components of various R. sphaeroides mutants, and their cytochrome contents in membranes^a

Strain	Phenotype	Electron carrier(s)/cyt c oxidase(s)	Cyt c profile					Cyt a	Cyt b	Cyt c	Cyt b/Cyt c ratio
			cp	c1	co	cy	c2
Ga	Ps+ Res+ Nadi+	Cyt c2 + cyt cy/cbb3-Cox + aa3-Cox	+	+	+	+	+	0.8	2.2	3.8	0.57
Gadc2	Ps− Res+ Nadi+	Cyt cy/cbb3-Cox + aa3-Cox	+	+	+	+	−	0.34	1.0	1.3	0.77
Gadcy	Ps+ Res+ Nadi+	Cyt c2/cbb3-Cox + aa3-Cox	+	+	+	−	+	0.37	1.0	1.45	0.68
Gadc2cy	Ps− Res+ Nadi+	None/cbb3-Cox + aa3-Cox	+	+	+	−	−	0.24	1.0	0.95	1.05
MT001	Ps+ Res+ Nadislow	Cyt c2 + cyt cy/aa3-Cox	−	+	−	+	+	0.48	1.15	1.14	1.0
KD02	Ps− Res+ Nadislow	Cyt cy/aa3-Cox	−	+	−	+	−	0.42	0.93	0.7	1.3
KD04	Ps+ Res+ Nadislow	Cyt c2/aa3-Cox	−	+	−	−	+	0.42	0.87	0.69	1.26
JS100	Ps+ Res+ Nadi+	Cyt c2 + cyt cy/cbb3-Cox	+	+	+	+	+	NDb	1.2	1.9	0.63
KD03	Ps+ Res+ Nadi+	Cyt cy/cbb3-Cox	+	+	+	−	+	ND	1.42	1.65	0.86
KD05	Ps− Res+ Nadi+	Cyt cy/cbb3-Cox	+	+	+	+	−	ND	2.2	1.56	0.87
ME127	Ps+ Res+ Nadi−	Cyt c2 + cyt cy/none	−	+	−	+	+	ND	1.28	1.06	1.2
BC17	Ps− Res+ Nadi+	Cyt c2 + cyt cy/cbb3-Cox + aa3-Cox	+	−	+	+	+	0.5	1.05	2.05	0.5

^aCyt a, cyt b, and cyt c values are expressed as nanomoles of heme per milligram of protein, using appropriate extinction coefficients, as described in Materials and Methods. Cyt c profiles are tabulated from the data in Fig. 4. 

^bND, not detected. 

Next, the ability of all strains to perform the Nadi reaction (24), described in Materials and Methods, was tested as an indication of their C_ox activities. Among them, only the cbb₃-C_ox⁻ aa₃-C_ox⁻ mutant (ME127) exhibited a complete Nadi⁻ phenotype (i.e., no blue color formation after more than 15 min of staining) (Table 2). The cbb₃-C_ox⁻ mutant (MT001) and its cyt c_2⁻cbb₃-C_ox⁻ (KD02) and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) derivatives showed various degrees of Nadi^slow phenotypes (i.e., formation of a bluish color within a few minutes of staining), with KD04 being the most defective in this respect. The remaining mutants were Nadi⁺ (i.e., formed a blue color in less than 1 min of staining), like the wild-type R. sphaeroides strain Ga. These observations indicated that absence of either the aa₃-C_ox (JS100), cyt c₂ (Gadc2), cyt c_y (Gadcy), or their combination (Gadc2cy, KD03, and KD05) did not affect the Nadi phenotype. Thus, in R. sphaeroides, the Nadi⁺ phenotype correlated better with the presence of an active cbb₃-C_ox rather than an aa₃-C_ox. It was also noted that the Nadi phenotype of various mutants was affected by the presence or absence of the electron carriers cyt c₂ and cyt c_y and also by the growth conditions used. For example, the cyt c_y⁻ cbb₃-C_ox⁻ mutant KD04 exhibited a stronger Nadi⁺ phenotype if younger colonies were stained, but conversely, the cyt c₂⁻ cbb₃-C_ox⁻ mutant KD02 became as Nadi⁻ as the cbb₃-C_ox⁻ aa₃-C_ox⁻ mutant ME127 when grown in enriched YCC medium (30).

Cytochrome contents of various R. sphaeroides mutants.

The contents of the a-, b-, and c-type cytochromes in chromatophore membranes, and in membrane supernatants, of various mutants grown by respiration in Med A were estimated using reduced-minus-oxidized optical absorption difference spectra (Fig. 3 and Table 2). Comparison of the reduced-minus-oxidized spectra of the cbb₃-C_ox⁻ (MT001) and cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127) mutants with other strains revealed significant decreases in the amounts of the c-, b-, and a-type cytochromes (peaks around 550, 560, and 605 nm, respectively) (Fig. 3A). In addition, contribution of the b- and c-type cyt subunits of the cyt bc₁ complex to total cytochrome contents of membranes was readily visible in the 550- to 560-nm region when the cyt bc₁⁻ mutant BC17 was compared to other mutants. Importantly, comparison of the cyt c₂⁻ cbb₃-C_ox⁻ (KD02), cyt c_y⁻ aa₃-C_ox⁻ (KD03), cyt c_y⁻ cbb₃-C_ox⁻ (KD04), and cyt c₂⁻ aa₃-C_ox⁻ (KD05) double mutants with their respective parents (cbb₃-C_ox⁻ [MT001] and aa₃-C_ox⁻ [JS100]) and with the wild-type strain Ga revealed that the 605-nm broad peak was absent in mutants carrying ctaD::spe and that the 550-nm peak was decreased significantly in those containing ccoNO::kan, as expected based on their genotypes (Fig. 3B).

FIG. 3.

Reduced-minus-oxidized optical absorption difference spectra of various R. sphaeroides mutants. The spectra were recorded between 510 and 630 nm at 25°C, using membrane fragments as described in Materials and Methods. (A) Spectra obtained with the cyt c₂⁻ (Gadc2), cyt c_y⁻ (Gadcy), cyt c₂⁻ cyt c_y⁻ (Gadc2cy), cyt bc₁⁻ (BC17), cbb₃-C_ox⁻ (MT001), and cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127) mutants; (B) those obtained with the wild-type strain Ga and the aa₃-C_ox⁻ (JS100), cyt c_y⁻, aa₃-C_ox⁻ (KD03), cyt c₂⁻ aa₃-C_ox⁻ (KD05), cyt c₂⁻ cbb₃-C_ox⁻ (KD02), and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants. Absorption peaks corresponding to the a-, b-, and c-type cytochromes are indicated by arrows. Note that the absorbance scale used in panel A is twice as large as that in panel B.

In all mutants, the amount of a-type cytochromes was lower (e.g., as low as 0.24 ± 0.01 nmol/mg of protein in the cyt c₂⁻ cyt c_y⁻ mutant Gadc2cy) than in the wild-type strain Ga (0.8 ± 0.04 nmol/mg of protein) (Table 2). The b-type heme content was also low in almost all mutants (around 1.2 ± 0.3 nmol/mg of protein); an exception was the amount in the cyt c_y⁻ aa₃-C_ox⁻ (KD03) mutant, which was similar to that in Ga (2.2 ± 0.1 nmol/mg of protein). Further, the amounts of c-type cytochromes dropped from 3.8 ± 0.2 nmol/mg of protein in Ga to 0.7 ± 0.02 nmol/mg of protein in cyt c₂⁻ cbb₃-C_ox⁻ (KD02) and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants as a result of mutations eliminating various c-type cytochromes. Elimination of the cbb₃-C_ox decreased the total amount of the b-type cytochromes by about one-half and that of the c-type cytochromes by more than two-thirds. Consequently, the cyt b-to-cyt c ratios increased from 0.5 ± 0.04 in membranes of the cyt bc₁⁻ mutant (BC17) to 1.3 ± 0.1 in those of the cyt c₂⁻ cbb₃-C_ox⁻ mutant (KD02). Thus, clearly the contents of the a-, b-, and c-type cytochromes in membrane fragments from various strains varied as a function of the missing components.

The c-type cytochrome contents of chromatophore membrane supernatants of the mutants were also determined using ascorbate-reduced-minus-ferricyanide-oxidized optical difference spectra (not shown). The data indicated that R. sphaeroides mutants carrying a cycA knockout allele had about 1/10 of the 550-nm absorption peak found in a wild-type strain. The absorption maximum of the remaining ascorbate-reducible c-type cyt was shifted to 552 nm, consistent with it being isocyt c₂ (33). Moreover, supernatants of the cyt bc₁⁻ mutant BC17 contained about threefold more cyt c₂ than those of the wild-type strain Ga. The presence of increased amounts of cyt c₂ in supernatants of mutants lacking the cyt bc₁ complex has been observed previously in R. capsulatus (32).

Cyt c profile of mutants lacking components of the c-type cytochrome-dependent respiratory branch of R. sphaeroides.

Chromatophore membrane proteins of the mutants were further analyzed using TMBZ–SDS-PAGE to determine the profiles of their c-type cytochromes. At least five major TMBZ-stained bands of molecular masses ranging from approximately 30 to 12 kDa were discernible in the wild-type strain Ga (Fig. 4). Under the separation conditions used here, the cyt c_p subunit of the cbb₃-C_ox and cyt c₁ subunit of the cyt bc₁ complex run together as an unresolved doublet, which was followed by the cyt c_o subunit of the cbb₃-C_ox and cyt c_y. Cyt c₂ ran ahead of the other c-type cytochromes, but not being membrane anchored, its amount varied in different samples. Additional c-type cytochromes of unidentified nature running behind cyt c_p were also visible, and one of them is likely to correspond to the dimethyl sulfoxide-inducible dorC product (25, 38). As expected, cyt c_y was absent in Gadcy, Gadc2cy, KD03, and KD04; similarly, Gadc2, Gadc2cy, KD02, and KD05 (Table 1) lacked cyt c₂ (18, 27). Equally, cyt c_o was missing in ME127 (cbb₃-C_ox⁻ aa₃-C_ox⁻) and in MT001 (cbb₃-C_ox⁻) and its derivatives. On the other hand, JS100 (aa₃-C_ox⁻) was like the wild-type strain Ga with respect to its c-type cytochrome content. The overall spectral and TMBZ–SDS-PAGE data established that the cyt c profiles of various mutants were in agreement with their genotypes.

FIG. 4.

c-type cytochromes detected in various R. sphaeroides mutants. Membrane fragments of various mutants were analyzed by SDS-PAGE, and c-type cytochromes were revealed via their peroxidase activities using TMBZ as described in Materials and Methods. c₁+c_p, c_o, c_y, and c₂ correspond to the cyt c₁ and c_p subunits of the cyt bc₁ complex and the cbb₃-C_ox, cyt c_o subunit of the cbb₃-C_ox, membrane-attached electron carrier cyt c_y, and soluble electron carrier cyt c₂, respectively. Additional unidentified c-type cytochromes running behind the doublet corresponding to cyt c_p+c₁ are also visible.

Respiratory electron flow and C_ox activities in various R. sphaeroides mutants.

Using the above-described mutant, we probed first for the presence of the C_ox activity in chromatophore membranes of various mutants, using reduced horse heart cyt c as a substrate. As expected, in semiaerobically grown cells of all strains except the cbb₃-C_ox⁻ aa₃-C_ox⁻ mutant ME127, substantial amounts of C_ox activity were detected, although the cbb₃-C_ox⁻ mutants exhibited lower levels (data not shown). Next, a detailed characterization of the electron flow through the different components of the respiratory chain was undertaken by using specific substrates and inhibitors. It is known that while addition of NADH or succinate activates electron transport via the respiratory dehydrogenases, inhibitors like antimycin A (Ant A) or Myx block it at the level of the cyt bc₁ complex (Fig. 1) (46). Respiratory capabilities of various mutants studied here were assessed directly by monitoring the rates of O₂ consumption depending on NADH or succinate, and in the presence or absence of exogenous cyt c, as performed previously with various R. capsulatus mutants (15).

Membrane fragments of the wild-type strain Ga rapidly oxidized various substrates, and the ensuing O₂ uptake was sensitive to Myx, dropping from 20.6 to 5.8 μmol of O₂ consumed/mg of protein/min in the presence of 2 μM of Myx (Table 3). Addition of 50 μM horse heart cyt c stimulated the rate of NADH oxidation to 30.9 μmol of O₂ consumed/mg of protein/min. This suggested that exogenous cyt c substituted well for the shortage of endogenous cyt c₂, which could have been depleted during cell disruption and membrane fragments preparation. Exogenous cyt c also greatly increased the rate of NADH oxidation (from 0.9 to 2.6 μmol of O₂ consumed/mg of protein/min) or ascorbate-C_ox activity (from 6.6 to 23.4 μmol of O₂ consumed/mg of protein/min) in the cyt c₂⁻ cyt c_y⁻ (Gadc2cy), cyt c_y⁻ aa₃-C_ox⁻ (KD03), and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants lacking the membrane-attached cyt c_y (Table 3). Moreover, pairwise comparison of various strains containing or lacking cyt c_y (i.e., Gadc2 versus Gadcy, KD02 versus KD04, and KD05 versus KD03) indicated that the presence of cyt c_y led almost in all cases to higher O₂ consumption activities. Thus, cyt c_y plays an important role as an electron carrier in the respiratory electron transport of R. sphaeroides, for in its absence the respiratory activities become more dependent on exogenous cyt c.

TABLE 3.

Respiratory activities in membranes of various R. sphaeroides mutants grown by respiration^a

Strain	Activity (μmol of O2 consumed/mg of protein/min)
	NADH				Succinate		Ascorbate
	Alone	+Cyt c	+Myx	+Cyt c, +CN	Alone	+Cyt c	+DCIP	+Cyt c
Ga	20.6	30.9	5.8	6.2	6.9	8.3	56.0	57.0
KD02	5.2	7.7	2.2	2.3	4.7	7.5	19.3	34.0
KD04	2.3	4.9	1.1	1.2	2.0	6.0	3.6	43.0
KD03	4.8	12.0	4.1	4.4	4.8	8.8	23.0	88.0
KD05	15.3	19.2	5.0	5.2	4.8	8.0	88.0	92.0
JS100	9.6	11.7	0.6	0.5	4.8	7.5	44.0	35.0
MT001	8.9	11.1	1.2	1.1	5.2	8.8	26.0	33.0
Gadc2cy	0.9	2.6	1.0	0.9	0.9	2.6	6.6	23.4
Gadc2	7.7	7.9	1.2	1.3	2.7	3.4	22.7	33.6
Gadcy	2.0	3.6	1.4	1.3	2.3	3.7	25.4	22.2
ME127	5.3	6.3	5.1	5.3	4.6	5.6	1.3	2.6
BC17	3.8	3.8	3.7	3.7	3.8	3.8	39.0	41.8

^aIn these assays, 0.2 mM NADH, 5 mM succinate, 2 mM ascorbate, 0.2 mM DCIP, 50 μM horse heart cyt c, 50 μM CN⁻, and 2 μM Myx were used as appropriate. Strain genotypes are given in Table 1. 

Remarkably, the artificial electron donor DCIP reduced cyt c_y better than cyt c₂, as indicated by the lower oxidative rates observed in its presence in appropriate mutants (e.g., cyt c_y⁻ cbb₃-C_ox⁻ [KD04] versus cyt c₂⁻ cbb₃-C_ox⁻ [KD02] mutants [3.6 versus 19.3 μmol of O₂ consumed/mg of protein/min], or cyt c_y⁻ aa₃-C_ox⁻ [KD03] versus cyt c₂⁻ aa₃-C_ox⁻ [KD05] mutants [23.0 versus 88.0 μmol of O₂ consumed/mg of protein/min]) (Table 3). The small amounts of C_ox activities detected in membranes of the cyt c_y⁻ cbb₃-C_ox⁻ (KD04) and cyt c₂⁻ cyt c_y⁻ (Gadc2cy) mutants using reduced DCIP (3.6 and 6.6 μmol of O₂ consumed/mg of protein/min, respectively) also differed from the increase of activity seen upon addition of cyt c (43.0 and 23.4 μmol of O₂ consumed/mg of protein/min, respectively). Thus, DCIP did not act as a good electron donor when cyt c_y and cbb₃-C_ox (KD04) or both cyt c₂ and cyt c_y (Gadc2cy) were absent, whereas horse heart cyt c acted as a good substitute for cyt c₂.

Interestingly, no more than one-fifth of the total O₂ consumption activity remained upon addition of either 2 μM Myx or 50 μM CN⁻ to membrane fragments of the wild-type strain Ga (Table 3). This remnant activity was in line with NADH oxidation detected in membranes of the cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127), cyt c₂⁻ cyt c_y⁻ (Gadc2cy), or cyt bc₁⁻ (BC17) mutant. Virtually no C_ox was detected in ME127 under all assay conditions, monitoring either NADH-, succinate-, or ascorbate-dependent O₂ consumption, and both in the presence and in the absence of exogenous cyt c (Table 3). The limited O₂ consumption activity detected was not inhibited by 0.1 mM CN⁻ and possibly corresponded to other some kind of terminal oxidase, such as Q_ox (13). On the other hand, the cyt bc₁⁻ mutant BC17 contained C_ox activities (39 or 42 μmol of 0.1 mM CN⁻-sensitive O₂ consumed/mg of protein/min when assayed with ascorbate plus DCIP or exogenous cyt c, respectively) comparable to those seen with the wild-type strain Ga (56 or 57 μmol of O₂/mg of protein/min using DCIP or cyt c, respectively). Lastly, the cbb₃-C_ox⁻ (MT001) and aa₃-C_ox⁻ (JS100) mutants contained about one-half of the total C_ox activity found in the wild-type strain Ga (11.7 and 11.1 μmol of O₂/mg of protein/min of NADH oxidase activity in the presence of exogenous cyt c in JS100 and MT001, respectively). Clearly, elimination of either the cbb₃-C_ox or the aa₃-C_ox decreased the total amount of the C_ox activity but did not abolish it completely (37).

Most striking findings were obtained using the cyt c₂⁻ cbb₃-C_ox⁻ (KD02), cyt c_y⁻ aa₃-C_ox⁻ (KD03), cyt c_y⁻ cbb₃-C_ox⁻ (KD04), and cyt c₂⁻ aa₃-C_ox⁻ (KD05) mutants, which contained simpler electron transport pathways beyond the cyt bc₁ complex (Fig. 1). The data indicated that higher O₂ consumption activities were found in mutants containing the cbb₃-C_ox (KD03 and KD05), and among them that harboring only cyt c_y (i.e., KD05) exhibited the highest C_ox activity. Conversely, lower O₂ consumption abilities were encountered in mutants lacking the cbb₃-C_ox (KD02 and KD04), and among them that lacking cyt c_y (i.e., KD04) exhibited the lowest activity (Table 3). Therefore, both cyt c₂ and cyt c_y donated electrons to both the cbb₃-C_ox and the aa₃-C_ox in vitro, but to different extents. The data suggested that the respiratory electron transfer pathways via the cbb₃-C_ox may be more active than those via the aa₃-C_ox subbranches in cells grown under semiaerobic conditions. However, the amounts of various components varied between various mutants (Table 2), precluding firmer conclusions. Further dissection of the C_ox activities into fractions corresponding to the cbb₃-C_ox and aa₃-C_ox activities exclusively linked to cyt c₂ or cyt c_y was impossible due to the assays used being unable to discriminate between the enzymes and their substrates.

NADH-induced cyt c oxidoreduction kinetics in various R. sphaeroides strains.

NADH-induced cyt c oxidoreduction kinetics in membranes from wild-type and various R. sphaeroides mutants were monitored in the presence or absence of oxygen in order to assess directly the electron transfer abilities of cyt c_y and cyt c₂ to the cbb₃-C_ox and aa₃-C_ox enzymes. In the wild-type strain Ga, approximately 50% of the total c-type cytochromes in membranes were rapidly reduced (half-life [t_1/2] of <1 s) upon addition of NADH (Fig. 5). This respiration-dependent steady-state cyt c reduction (i.e., first reduction phase) lasted for several minutes, depending on the amount of membrane fragments used, and was sensitive to the cyt bc₁ complex inhibitors Myx and Ant A (2 μM each) (Fig. 5). Thus, a large fraction of the c-type cytochrome complement of membranes from wild-type cells is in rapid equilibrium with the respiratory electron transport system. Similar but slower oxidoreduction kinetics (t_1/2 = 1 to 2 s) were observed using membranes from the aa₃-C_ox⁻ (JS100), cyt c₂⁻ aa₃-C_ox⁻ (KD05), and cyt c_y⁻ aa₃-C_ox⁻ (KD03) mutants. Even slower cyt c reduction kinetics (2 s, <t_1/2 <10 s) were seen with membranes from the cbb₃-C_ox⁻ (MT001), cyt c₂⁻ cbb₃-C_ox⁻ (KD02), and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants.

FIG. 5.

Cyt c reduction kinetics in R. sphaeroides wild-type strain Ga and various mutants. Cyt c reduction kinetics were monitored at 551 to 540 nm using membrane fragments of various mutants as described in Materials and Methods. In all experiments, the first reduction wave was initiated by addition of 0.2 mM NADH at the time indicated by the arrow, while the second reduction wave followed the onset of anaerobiosis the time of which was dependent on the rate of oxygen consumption. Dashed traces represent cyt c reduction patterns observed upon addition of Ant A or Myx (2 μM each) before the addition of NADH. Protein concentrations of membrane fragments used were 0.8 mg/ml and 2 to 3 mg/ml for the wild-type strain Ga and its mutant derivatives, respectively. Below the traces, schemes depicting the available electron transport pathways are shown.

Upon depletion of O₂, and hence inhibition of the C_ox activities, cyt c reduction further increased in the wild-type strain Ga and in various aa₃-C_ox⁻ mutants to about 95% (second reduction phase). In the cbb₃-C_ox⁻ mutants, the amounts of the c-type cytochromes reduced under steady-state respiration were between 30 and 45% of their total content, reaching a maximum of about 50 to 60% under anaerobic conditions (Fig. 5). This second phase of cyt c reduction revealed the amount of cyt c that had been oxidized during the first phase when cyt c oxidases were active and hence illustrated at least partly electron conduction to these enzymes by their electron carriers. Therefore, the data obtained with mutants containing linear electron transport pathways between the cyt bc₁ complex and the cyt c oxidases demonstrated that both cyt c₂ and cyt c_y indeed donated electrons to either the cbb₃-C_ox or aa₃-C_ox enzyme. However, the cyt c_y → aa₃-C_ox (KD02) and cyt c₂ → aa₃-C_ox (KD04) pathways appeared to be less efficient (i.e., smaller second phases), suggesting that under our experimental conditions the aa₃-C_ox played in O₂ consumption a less prominent role than the cbb₃-C_ox.

The light-harvesting antenna complexes are induced in the presence of oxygen in R. sphaeroides mutants lacking either both cyt c₂ and cyt c_y or the cbb₃-C_ox activity (28, 29). During our studies we also noted that R. sphaeroides strain MT001 lacking the cbb₃-C_ox and its derivatives lacking in addition either cyt c₂ or cyt c_y (KD02 or KD04) formed highly pigmented colonies under Res growth conditions. This pigmentation was especially pronounced in a mutant lacking both cyt c_y and the cbb₃-C_ox (KD04). Indeed, spectral examination of chromatophore membranes of various mutants confirmed the presence of prominent absorption peaks in the 800- to 875-nm region (Fig. 6). Thus, the absence of the cbb₃-C_ox and its electron donors modulated the production of the light-harvesting complexes in the presence of oxygen. Remarkably, the pigment complexes were either absent or much less pronounced in the other R. sphaeroides Ga derivaties used in this work, including the cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127), cyt bc₁⁻ (BC17), and cyt c₂⁻ cyt c_y⁻ (Gadc2cy) mutants.

FIG. 6.

Absorption spectra of membrane fragments of various R. sphaeroides mutants. Membrane fragments were prepared from cells grown by respiration at an optical density (OD) at 630 nm of 0.4, and their absorption spectra were recorded as described in Materials and Methods. Spectra were obtained for the wild-type strain Ga and the cyt c₂⁻ (Gadc2), cyt c_y⁻ (Gacy), cyt c₂⁻ cyt c_y⁻ (Gadc2cy), cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127), and cyt bc₁⁻ (BC17) mutants (A), the wild-type strain Ga and the cbb₃-C_ox⁻ (MT001), cyt c₂⁻ cbb₃-C_ox⁻ (KD02), and cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants (B), and the wild-type strain Ga and the aa₃-C_ox⁻ (JS100), cyt c_y⁻ aa₃-C_ox⁻ (KD03), and cyt c₂⁻ aa₃-C_ox⁻ (KD05) mutants (C). See the text for further details.

DISCUSSION

In this work we used a biochemical genetic approach, which consisted of disrupting simultaneously one of the two electron carrier c-type cytochromes and one of the two cyt c oxidases, to dissect the complicated respiratory electron transport pathways of R. sphaeroides (Fig. 1). The double mutants thus obtained simplified the crisscrossed pathways and enabled us to analyze the abilities of the different electron carriers to convey electrons from the cyt bc₁ complex to the terminal cyt c oxidases. Prior to this work, it was well known that the soluble cyt c₂ donates electrons to the aa₃-C_ox; indeed, it could be assumed correctly that it also donates electrons to the cbb₃-C_ox (15, 27). On the other hand, not much was known about the electron carrier properties of the more recently discovered membrane-anchored cyt c_y in R. sphaeroides (27, 51). Studies reported here confirmed for cyt c₂, and demonstrated for the first time for cyt c_y, that both are efficient electron donors to both the cbb₃-C_ox and the aa₃-C_ox during the respiratory growth of R. sphaeroides. Neither of the two electron carriers is restricted to function solely with either of the cyt c oxidases, and clearly all four subbranches appear to be functional (Fig. 1). Whether or not all of these pathways are also proficient in vivo to support respiratory growth of R. sphaeroides in the absence of its Q_ox-dependent alternate branch(es) remains to be seen.

At least under the semiaerobic growth conditions used here, the cbb₃-C_ox subbranches appear to be more prominent in catalyzing electron transfer between the cyt bc₁ complex and the cyt c oxidases in vitro. However, whether this prominence is due to the larger amounts of the components of these pathways or to their better catalytic abilities remains unknown until determination of their amounts in each case. Among the electron carriers, while cyt c₂ can diffuse between the RC and the cyt bc₁ complexes of the photosynthetic electron transfer chains (10), cyt c_y is conceivably restricted to interact with a limited number of partners due to its membrane anchor (26, 27). Nonetheless, it is noteworthy that the alpha group proteobacterium Rickettsia prowazekii, thought to be closely related to mitochondria, lacks a soluble cyt c but contains a membrane-attached cyt c_y homologue (1, 23).

Availability of otherwise isogenic mutants lacking different c-type cytochromes of R. sphaeroides allowed us to identify each of these proteins unambiguously and confirmed our previous assignments of the cytochromes c₂, c_y, c₁, c_o, and c_p in R. sphaeroides (18, 27). It is noteworthy that in this bacterium cyt c_y runs ahead of the cyt c_o subunit of the cbb₃-C_ox (Fig. 3), which is different from what has been observed with R. capsulatus cyt c_y (26). Whether the inability of R. sphaeroides cyt c_y to donate electrons to the RC is linked to its smaller size remains to be seen (27).

The mutants described here, eliminating systematically various respiratory components, allowed us to estimate the amounts of various b- and c-type cytochromes present in membranes of R. sphaeroides cells grown under semiaerobic conditions. It appears that in a wild-type strain like Ga, elimination of the cbb₃-C_ox decreased roughly one-half of the total amount of the b-type cytochromes and two-thirds of the c-type cytochromes in membranes. Similarly, deletion of the cyt bc₁ complex also decreased about one-half of the total amount of both b- and c-type cytochromes. These estimations are in agreement with the subunit composition of the cbb₃-C_ox and the cyt bc₁ complex and suggest that these complexes may be present at similar amounts in R. sphaeroides membranes under the growth conditions used here. However, it should be noted that the amounts of various cytochromes change in different mutants as a function of the presence or absence of various components of the electron transport chain. For example, the absence of electron-carrying c-type cytochromes or the cyt bc₁ complex also decreased the amounts of the cbb₃-C_ox or the aa₃-C_ox (Table 3). Thus, caution should be exercised in extrapolating the results obtained using membrane fragments to intact cells grown under various physiological conditions.

Estimation of the C_ox-independent respiratory pathway in the wild-type strain Ga and in cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127), cyt bc₁⁻ (BC17), and cyt c₂⁻ cyt c_y⁻ (Gadc2cy) mutants indicated that in R. sphaeroides about one-fifth of the respiratory capacity is insensitive to 100 μM CN⁻ and is possibly mediated via a Q_ox enzyme(s). This situation is unlike that observed in R. capsulatus (15), where the analogous CN-resistant NADH-dependent respiration represents about 75 to 85% of NADH oxidation rate. Moreover, in R. sphaeroides, restraining the mitochondrial-like respiratory electron flow by eliminating some of its components does not affect the levels of CN-resistant respiration. Membranes of an R. sphaeroides mutant lacking the cyt c₂ and cyt c_y (Gadc2cy) exhibit very low NADH oxidation activity (Table 3), unlike its R. capsulatus counterpart FJ2 (cyt c₂⁻ cyt c_y⁻), which has a CN-resistant NADH oxidation activity roughly three times higher than that of its parental wild-type strain MT1131 (15). This phenomenon was suggested to reflect a regulatory response of one of the respiratory branches over the other (49), but the molecular basis of this effect is unknown. No similar response is obvious in R. sphaeroides, for the mutants analyzed here exhibited a CN-resistant respiratory activity similar to that present in the wild-type membranes. The marked inefficiency of the alternate respiratory pathway, in addition to the photosynthetic incompetence of cyt c_y and the presence of an aa₃-C_ox, constitutes yet another striking difference between the growth abilities of the closely related species R. capsulatus and R. sphaeroides.

Increased synthesis of the photosynthetic apparatus in the presence of oxygen in mutants lacking the cbb₃-C_ox has been documented previously in R. sphaeroides strain 2.4.1 (28, 52). Recently, a similar situation has also been described in derivatives of the same strain lacking either the cyt bc₁ complex or the cytochromes c₂ and c_y (29). In these mutants, expression of the photosynthetic apparatus is not repressed by O₂, unlike in the wild-type strain, thus leading to the accumulation of increased amounts of light-harvesting antenna complexes and photosynthetic pigments. In R. sphaeroides strain Ga and its derivatives used here, the light-harvesting complexes were also highly induced in the cbb₃-C_ox⁻ (MT001), cyt c₂⁻ cbb₃-C_ox⁻ (KD02), and especially cyt c_y⁻ cbb₃-C_ox⁻ (KD04) mutants, while this induction was less apparent in the cyt bc₁⁻ (BC17), cbb₃-C_ox⁻ aa₃-C_ox⁻ (ME127), and cyt c₂⁻, cyt c_y⁻ (Gadc2cy) mutants. Why similar mutants derived from R. sphaeroides strains 2.4.1 and Ga behave differently in this respect is unclear. In any event, close examination of the electron transport pathways (Fig. 1) suggests that either restricting the electron flow via the cbb₃-C_ox subbranches, as proposed earlier (29, 52), or forcing this flow via the aa₃-C_ox subbranches, or even a combination of these possibilities, may lead to the accumulation of the pigment proteins. Clearly, the creation of an imbalance between the aa₃-C_ox and cbb₃-C_ox subbranches without taking into account the availability of O₂ might explain at least partly the pigmentation phenotypes observed in various mutants used in this work.

In summary, the work presented here establishes for the first time that both of the electron carriers cyt c₂ and cyt c_y are efficient electron donors to both the cbb₃-C_ox and the aa₃-C_ox during the respiratory growth of R. sphaeroides. It remains to be seen whether or not the electron flow via each of the subbranches defined here is alone sufficient to support the Res growth of R. sphaeroides under physiological growth conditions.