Arch Microbiol (1998) 170 : 817

Springer-Verlag 1998

O R I G I N A L PA P E R

George C. Paoli F. Robert Tabita

Aerobic chemolithoautotrophic growth and RubisCO function in Rhodobacter capsulatus and a spontaneous gain of function mutant of Rhodobacter sphaeroides

Received: 8 August 1997 / Accepted: 26 December 1997

Abstract Photosynthetic prokaryotes that assimilate CO2 under anoxic conditions may also grow chemolithoautotrophically with O2 as the electron acceptor. Among the nonsulfur purple bacteria, two species (Rhodobacter capsulatus and Rhodopseudomonas acidophilus), exhibit aerobic chemolithoautotrophic growth with hydrogen as the electron donor. Although wild-type strains of Rhodobacter sphaeroides grow poorly, if at all, with hydrogen plus oxygen in the dark, we report here the isolation of a spontaneous mutant (strain HR-CAC) of Rba. sphaeroides strain HR that is fully capable of this mode of growth. Rba. sphaeroides and Rba. capsulatus fix CO2 via the reductive pentose phosphate pathway and synthesize two forms of ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO). RubisCO levels in the aerobic-chemolithoautotrophic-positive strain of Rba. sphaeroides were similar to those in wild-type strains of Rba. sphaeroides and Rba. capsulatus during photoheterotrophic and photolithoautotrophic growth. Moreover, RubisCO levels of Rba. sphaeroides strain HR-CAC approximated levels obtained in Rba. capsulatus when the organisms were grown as aerobic chemolithoautotrophs. Either form I or form II RubisCO was able to support aerobic chemolithoautotrophic growth of Rba. capsulatus strain SB 1003 and Rba. sphaeroides strain HR-CAC at a variety of CO2 concentrations, although form II RubisCO began to lose the capacity to support aerobic CO2 fixation at high O2 to CO2 ratios. The latter property and other facets of the physiology of this system suggest that Rba. sphaeroides and Rba. capsulatus strains may be effectively employed for the biological selection of RubisCO molecules of altered substrate specificity.
G. C. Paoli1 F. R. Tabita ( ) The Department of Microbiology and Plant Molecular Biology/Biotechnology Program, The Ohio State University, 484 West 12th Avenue, Columbus, Ohio 43210-1292, USA e-mail: tabita.1@osu.edu Tel. +1-614-292-4297; Fax +1-614-292-6337 Present address: 1 AFRL/MLQR, Tyndall Air Force Base, FL 32403-5323, USA

Key words Aerobic lithoautotrophic growth Rhodobacter sphaeroides Rhodobacter capsulatus RubisCO Ribulose 1,5-bisphosphate carboxylase/ oxygenase Biological selection Abbreviations RubisCO Ribulose 1,5-bisphosphate carboxylase/oxygenase CAC Chemolithoautotrophic competent RuBP Ribulose 1,5-bisphosphate CBB pathway Calvin-Benson-Bassham reductive pentose phosphate pathway

Introduction
The regulation of photosynthetic carbon metabolism in purple bacteria has been the subject of intense study for several years (Tabita 1988, 1995; Kondratieva 1989; Gibson 1995; Gibson and Tabita 1996;). Phototrophic growth occurs only under anoxic conditions, with CO2 fixed via the Calvin-Benson-Bassham reductive pentose phosphate pathway (CBB pathway). Although not nearly as well studied, some purple bacteria are capable of aerobic chemolithoautotrophic growth in the dark. Several species of Chromatiaceae catalyze aerobic chemolithoautotrophic metabolism using reduced sulfur compounds as the source of reducing power (Kondratieva et al. 1976; Kampf and Pfennig 1980, 1986 a, b; Kondratieva 1989). To date, however, only a few species of nonsulfur purple phototrophs (Rhodospirillaceae) have been shown to grow aerobically as chemolithoautotrophs. Rhodobacter capsulatus (Madigan and Gest 1979; Siefert and Pfennig 1979) and Rhodopseudomonas acidophilus (Siefert and Pfennig 1979) grow aerobically with molecular hydrogen as the electron source. Under these conditions the organisms grow in a completely inorganic medium in the presence of gaseous CO2, O2, and H2, catalyzing the so-called knallgas reaction (2 H2 + O2 2 H2O). Growth under such conditions is basically equivalent to the characteristic growth patterns exhibited by several species of nonphotosynthetic aerobic hydrogen bacteria hereafter referred to as chemoautotrophic growth.

With the exception of thermophilic strains of Hydrogenobacter (Kawasumi et al. 1984; Shiba et al. 1985) and Aquifex (Beh et al. 1993), bacteria that catalyze the knallgas reaction fix CO2 via the CBB pathway (Schlegel 1989), where the key enzyme is ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO). In the presence of oxygen, the metabolism of 2-phosphoglycolate, a product of the oxygenase reaction of RubisCO, results in the net loss of carbon from the cell. In plants, the dissimilation of 2phosphoglycolate proceeds via the photorespiratory pathway, and various estimates predict a loss of 0.5 carbon for each O2 fixed (Gutteridge 1989). In bacteria, the metabolism of phosphoglycolate is not thoroughly understood. In some cases, phosphoglycolate is dephosphorylated and glycolate is excreted directly. If all of the phosphoglycolate produced were metabolized in this manner, the oxygenase reaction would be truly costly in reducing net CO2 fixation by RubisCO. The presence of glycolate oxidase activity in species of phototrophic (Codd and Smith 1974; Codd and Turnbull 1975) and chemoautotrophic bacteria (Codd et al. 1976), and the increase in glycolate excretion when this enzyme is inhibited indicates that some of the glycolate may be further metabolized. The rates () at which RubisCO catalyzes the carboxylase and oxygenase reactions can be expressed by: c/o = ([CO2]/[O2]), where is the substrate specificity factor. is defined by the ratio of the RubisCO carboxylase (c) and oxygenase (o) kinetic constants: = VcKo/VoKc, where V is the Vmax and K is the Michaelis constant for each gaseous substrate (Jordan and Ogren 1981). Thus, the ratio of carboxylation to oxygenation (c/o) in a given atmosphere ([CO2]/[O2]) is determined by . The of form I RubisCO, comprised of eight large and eight small subunits, varies from 25 to 240 depending on the source of the enzyme (Read and Tabita 1992, 1994;
Table 1 Bacterial strains used in this study

Tabita 1995; Uemura et al. 1997), while the of form II RubisCO is ~10 (Tabita 1995). Rhodobacter sphaeroides and Rba. capsulatus, two species of phototrophic nonsulfur purple bacteria, synthesize both form I and form II RubisCO (Gibson and Tabita 1977 a, b, 1985; Shively et al. 1984; Paoli et al. 1995). These facultatively autotrophic and facultatively anaerobic organisms present an interesting potential model system for examining the capacity of form I and form II RubisCO to support aerobic chemoautotrophic growth. Although Rba. capsulatus can grow chemoautotrophically with H2 plus O2, wild-type strains of Rba. sphaeroides apparently lack this ability (Madigan and Gest 1979). We report here the isolation of a spontaneous mutant of Rba. sphaeroides that has acquired the ability to grow chemoatrophically with H2 plus O2. The ability of form I or form II RubisCO to support aerobic chemoautotrophic growth was examined in Rba. capsulatus and Rba. sphaeroides strains synthesizing only one or the other form of RubisCO. The use of these strains for the biological selection of RubisCO with improved substrate specificity may be a direct consequence of these studies.

Materials and methods

Bacterial strains and growth conditions A list of Rba. sphaeroides and Rba. capsulatus strains used in this study, all derived from wild-type strains HR and SB1003, respectively, is given in Table 1. A variety of different wild-type Rba. sphaeroides and Rba. capsulatus strains were also employed for some studies (see Table 5); the latter were the generous gift of Michael T. Madigan (Southern Illinois University, Carbondale, Ill., USA). Aerobic chemoheterotrophic cultures of Rba. sphaeroides and Rba. capsulatus were grown in PYE medium (Weaver

Strain

Relevant characteristics

Source or reference Weaver and Tabita (1983) This work Gibson et al. (1991) Gibson et al. (1991) This work This work Falcone and Tabita (1991) Wang et al. (1993) Gibson and Tabita (1993) Gibson and Tabita (1997) Yen and Marrs (1976) Paoli and Tabita (manuscript in preparation) Paoli and Tabita (manuscript submitted) Paoli and Tabita (manuscript submitted)

Rhodobacter sphaeroides HR SmR derivative of wild-type strain ATCC 17023 HR-CAC Spontaneous mutant of HR able to grow as an aerobic chemolithoautotroph HR-FI cbbLS::KmR derivative of HR HR-FII cbbM::TpR derivative of HR CAC-FI cbbLS::KmR derivative of HR-CAC FII-CAC Spontaneous mutant of strain HR-FII able to grow as an aerobic chemolithoautotroph 16 cbbLS::KmR, cbbM::TpR derivative of HR 16PHC Spontaneous mutant of strain 16 able to grow photoheterotrophically 1312 cbbR::TpR derivative of strain HR 4X cbbX::SpR derivative of strain HR, cbbXYZ Rhodobacter capsulatus SB1003 RifR derivative of wild-type strain B10 SBI cbbL::SpR derivative of SB1003 SBII SBI-II cbbM::KmR derivative of SB1003 cbbL::SpR, cbbM::KmR derivative of SB1003

10 and Tabita 1983) at 30 C. Photoautotrophic and chemoautotrophic cultures were grown in Ormerods minimal medium (Ormerod et al. 1961) supplemented with vitamins (1 g thiamine/ml, 1 g nicotinic acid/ml, and 0.1 g biotin/ml). Photoautotrophic cultures were bubbled continuously with CO2/H2 (1.5 : 98.5, v/v) and incubated in the light as previously described (Falcone and Tabita 1991). Chemoautotrophic growth was carried out on solid media or in small liquid cultures (10 ml) at 30 C in BBL GasPak jars using a H2/CO2 generating system without the palladium catalyst (Becton-Dickinson Microbiology Systems, Cockeysville, Md., USA). Larger aerobic chemoautotrophic liquid cultures (300 ml) were grown with continuous bubbling (50 ml/min) with H2/CO2/air. Gases were prepared by blending a mixture of H2/CO2 (Liquid Carbonic Specialty Gases, Oak Brook, Ill., USA) with air (50 : 50, v/v) using a two-tube rotameter (Matheson Gas Products, Montgomeryville, Pa., USA). For growth at different CO2 concentrations, different mixtures of H2/CO2 were used. In all instances the level of O2 was half the concentration found in air, 10.5%. Photoheterotrophic cultures were grown in Ormerods medium plus vitamins supplemented with 0.4% malate or other carbon sources as specified, in filled screw-capped tubes or by continuously bubbling cultures with argon. Antibiotics were used where appropriate at the following concentrations: for Rba. sphaeroides, kanamycin (30 g/ml), trimethoprim (200 g/ml), and streptomycin (50 g/ ml); for Rba. capsulatus, rifampicin (100 g/ml), kanamycin (5 g/ml), and spectinomycin (5 g/ml). For comparison of the growth of Rba. sphaeroides CAC with that of Rba. capsulatus SB1003 and Rba. sphaeroides HR, streptomycin was used at 50 g/ml and penicillin G was used at 0.1 U/ml. Mutant strain isolation and construction Chemoautotrophic-competent (CAC) strains were selected by incubating cultures for 34 weeks in an H2/CO2/air atmosphere in a BBL GasPak jar using the H2/CO2 generating system without the palladium catalyst. The jar was opened weekly, and a new H2/CO2 generating envelope was used to refresh the atmosphere in the jar. Escherichia coli SM10 (Simon et al. 1983) containing plasmid pSUP::E25::Km (Falcone and Tabita 1991) was used to construct the cbbLcbbS-deletion strain CAC-FI from Rba. sphaeroides strain HR-CAC. The conjugal mating procedure and screening have been described previously (Falcone and Tabita 1991). The construction of the Rba. capsulatus RubisCO-minus strains will be presented elsewhere (Paoli and Tabita, unpublished work). Preparation of cell extracts and RubisCO assays Culture samples (2030 ml) were washed twice in 100 mM TrisHCl (pH 8.0) and 1 mM EDTA and were frozen at 70 C. Thawed pellets were resuspended in 1 ml TEM [50 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 5 mM -mercaptoethanol] and disrupted by sonication at 0 C. Cell debris was removed by centrifugation for 10 min in a microcentrifuge at 4 C. The resultant crude extract was used for enzyme assays. RubisCO activity was assayed as ribulose 1,5-bisphosphate (RuBP)-dependent 14CO2 fixation into acid-stable material (Gibson et al. 1991), and protein concentrations were measured by a modified Lowry procedure (Markwell et al. 1978). Rocket electroimmunoassay The amount of RubisCO protein in Rba. sphaeroides cell extracts was determined by rocket electroimmunoassay using antibodies specific for Rba. sphaeroides form I and form II RubisCO (Jouanneau and Tabita 1986).

Results
Selection and characterization of Rba. sphaeroides strain HR-CAC Under a H2/CO2/air atmosphere in GasPak jars, the growth of Rba. capsulatus was complete by 7 days, at which time the Rba. sphaeroides cultures showed little or no increase in turbidity. This is consistent with the observations of Madigan and Gest (1979), who have investigated the ability of Rba. capsulatus to grow chemoautotrophically; Rba. sphaeroides grew very poorly, if at all, under similar conditions. However, when Rba. sphaeroides cultures were allowed to incubate for longer periods of time, turbidity increased dramatically after 34 weeks of incubation. This was repeated several times with Rba. sphaeroides wild-type strain HR. Each time the culture showed little or no increase in turbidity until approximately 3 weeks had passed, after which time the culture began to show a slow but steady increase to full density (OD660 > 1.5) during the following 7 days. In these initial experiments, it was not possible to monitor growth turbidometrically because the cultures were grown in sealed jars. The lengthy time period required before the onset of chemoautotrophic growth of Rba. sphaeroides suggested that a spontaneous mutant might have been selected. Nevertheless, the possibility still existed that growth occurred only after extensive physiological adaptation. The latter possibility was eliminated by the following experiment: a sample from the chemoautotrophically grown culture was streaked on a peptone yeast extract (PYE) plate; single colonies were picked and used to inoculate malate-supplemented Ormerods mineral media tubes, which were incubated under anoxic conditions in the light. Samples from the photoheterotrophic cultures were used to inoculate minimal medium lacking a fixed carbon source under a H2/CO2/air atmosphere in GasPak jars. After this regimen, chemoautotrophic growth proceeded without the pronounced 3-week lag period previously observed. Growth was complete after approximately 7 days. Thus, the chemoautotrophic competent (CAC) phenotype of this strain, HR-CAC, appeared to be the result of a stable genetic change. The strain has been maintained in frozen glycerol stocks for over 3 years and can be grown under a variety of conditions for months without losing the CAC phenotype. Similar observations were made when Rba. sphaeroides HR cultures were bubbled with a gas mixture of CO2/H2/air (5 : 45 : 50, by vol.; Fig. 1). The culture density of the wild-type strain HR increased only slightly; then, after approximately 3 weeks, exponential growth ensued. Strain HR-CAC grew almost immediately under these conditions. The fact that the CAC strain originated directly from Rba. sphaeroides strain HR was confirmed in several ways. First we ruled out the possibility that strain HRCAC is merely a strain of Rba. capsulatus based upon the response of Rba. sphaeroides and Rba. capsulatus to different carbon sources and their differential antibiotic sen-

11

Time (days)
Fig. 1 Aerobic chemolithoautotrophic growth of Rhodobacter sphaeroides strains HR and HR-CAC. Strains HR (b) and HRCAC ( ) were inoculated from photolithoautotrophically grown cultures into carbon-free Ormerods minimal media. Cultures were incubated at 30 C in the dark with continuous bubbling at 50 ml/min using a gas mixture of CO2/H2/air (5 : 45 : 50, by vol.) Table 2 Comparison of growth of the CAC strain with that of Rhodobacter sphaeroides strain HR and Rhodobacter capsulatus strain SB1003. Cultures were grown photoheterotrophically in the presence of the indicated carbon source (0.4%) and 0.05% bicarbonate. When indicated, streptomycin (Sm) was added to 50 g/ml final concentration and penicillin G (Pc) was added to 0.1 U/ml final concentration. +++ Very good, ++ good, + poor, no growth Strain Growth Malate ProCitrate Sm pionate Rba. sphaeroides HR +++ Rba. sphaeroides HR-CAC +++ Rba. capsulatus SB1003 +++ + + +++ +++ +++ + +++ +++ Pc +++ +++

Growth rates and RubisCO activity of Rba. sphaeroides HR-CAC cultured under various conditions Growth rates and RubisCO activity levels of Rba. sphaeroides strain HR-CAC were determined under various growth conditions and compared to those of parent strain HR and Rba. capsulatus wild-type strain SB1003 (Table 3). Comparisons were made to Rba. capsulatus because the wild-type strain of this species can grow chemoautotrophically (Madigan and Gest 1979). All three strains grew photoheterotrophically in 0.5-l vessels containing 300-ml cultures with argon bubbling at doubling times of 6.57 h; in addition, each strain exhibited similar levels of RubisCO activity. During photolithoautotrophic growth, the two Rba. sphaeroides strains had similar doubling times and RubisCO activity (Table 3), while Rba. capsulatus grew considerably faster, doubling every 10 h, than
Table 3 Growth rates and ribulose 1,5-biphosphate carboxylase/oxygenase (RubisCO) activities in Rhodobacter sphaeroides strains HR and HR-CAC and Rhodobacter capsulatus strain SB1003. Cultures were grown photoheterotrophically in a malate minimal medium bubbled with argon, photolithoautotrophically [CO2/H2 (1.5 : 98.5, v/v)], or chemolithoautotrophically [CO2/H2/air (5 : 45 : 50, by vol.)]. Numbers represent averages of duplicate experiments Strain Doubling time (h) RubisCO activity [nmol min1 (mg protein)1] 34 40 34 109 123 186

sitivity (Weaver et al. 1975; Richardson et al. 1988; Table 2). Since Rba. capsulatus form I RubisCO does not cross-react with antibodies raised against the Rba. sphaeroides form I RubisCO (Gibson and Tabita 1977 b; Paoli et al. 1995), the ability of anti-Rba. sphaeroides form I RubisCO antibodies to cross-react with CAC extracts in Ouchterlony immunodiffusion experiments (not shown) and in rocket immunoelectrophoresis (see Table 4) further differentiated the CAC strain from Rba. capsulatus. In addition, results of Southern hybridization analysis of CAC chromosomal DNA using Rba. sphaeroides cbbL and cbbM gene probes exactly matched the restriction and label patterns observed for Rba. sphaeroides wild-type strain HR (data not shown). All of this evidence indicates that the CAC strain was derived from strain HR, presumably as a result of a spontaneous mutation, resulting in the ability of strain HR-CAC to grow chemoautotrophically.

Photoheterotrophic growth HR 7.0 HR-CAC 6.5 SB1003 6.5 Photolithoautotrophic growth HR 17 HR-CAC 14.5 SB1003 10

Aerobic chemolithoautotrophic growth HR-CAC 27 93 SB1003 22 94

12 Table 4 Ribulose 1,5-biphosphate carboxylase/oxygenase (RubisCO) protein levels in Rhodobacter sphaeroides strain HR-CAC. The amount of form I or form II protein was quantitated by rocket electroimmunoassay. In all cases the data represent averages of duplicate experiments. Cultures were grown as indicated in Table 3 Growth RubisCO level (% soluble protein) Form I Photoheterotrophic Photolithoautotrophic Aerobic chemolithoautotrophic 0.7 4.1 6.6 Form II 1.7 4.2 2.6 Table 5 Aerobic chemolithoautotrophic growth of Rhodobacter sphaeroides and Rhodobacter capsulatus wild-type strains. +++ very good, ++ good, + poor, +/ very poor growth Rba. sphaeroides strain HR HR-CAC 2.4.1 2.4.1 GA DSM158 Geller 21.7 2.4.7 Growth +/ ++ +/ +/ +/ + + ++ Rba. capsulatus strain B10 SB1003 VW2 Kb1 LB2 C3 C4 JH2 SP101 Growth +++ +++ +++ + + +++ +++ +++ +++

the Rba. sphaeroides strains, and its RubisCO activity was somewhat higher (Table 3). In both Rba. sphaeroides and Rba. capsulatus, RubisCO activities of chemoautotrophically grown cells were intermediate to those obtained from photoheterotrophically and photolithoautotrophically grown cells. It should be stressed that RubisCO synthesis was not constitutive in Rba. sphaeroides strain HR-CAC since - as in the case of strain HR - RubisCO activity was only barely detectable when cultures were grown chemoheterotrophically with O2 on minimal medium containing malate as the carbon source (data not shown). The amount of form I and form II RubisCO protein in strain HR-CAC was quantified by rocket electroimmunoassay. Although qualitatively similar to that of Rba. sphaeroides (based on Western immunoblots), quantitation of the amount of RubisCO protein synthesized was not obtained with Rba. capsulatus since antibodies against Rba. capsulatus RubisCO are not yet available. Considerably more form I and somewhat more form II RubisCO were synthesized in Rba. sphaeroides strain HR-CAC during both photolithoautotrophic growth and chemoautotrophic growth as compared to photoheterotrophic growth (Table 4). These observations are consistent with those of previous experiments in which the amounts of form I and form II RubisCO protein were quantitated in strain HR (Jouanneau and Tabita 1986). Chemoautotrophic growth of other Rba. sphaeroides and Rba. capsulatus strains To determine if chemoautotrophic competence is a speciesspecific trait, a number of Rba. capsulatus and Rba. sphaeroides wild-type strains were tested for their ability to grow as chemoautotrophs. The strains were streaked on solid medium and grown in GasPak jars. After 3 weeks, growth was evaluated. With the exception of wild-type strain 2.4.7 and mutant strain HR-CAC, Rba. sphaeroides strains grew poorly, while most of the Rba. capsulatus strains grew well (Table 5). The ability of various Rba. sphaeroides HR mutant strains to acquire the CAC phenotype was also determined. Each strain was incubated under selective conditions (liquid cultures in GasPak jars) to see if chemoautotrophic growth occurred within 4 weeks. Duplicate cultures of each strain were tested. For the

strains where neither replicate grew, a third attempt was made to grow the strain chemoautotrophically. Four strains were unable to acquire the CAC phenotype: strains 16, 16PHC, 1312, and HR-FI (Table 6). Strain 16 is an HR derivative that lacks both form I and form II RubisCO due to insertional inactivation of the respective genes (Falcone and Tabita 1991). Strain 16PHC is a spontaneous mutant of strain 16 that has acquired the ability to grow photoheterotrophically (Wang et al. 1993). This strain does not synthesize RubisCO, but is able to grow photoheterotrophically in the absence of RubisCO using protons as an alternative electron acceptor via nitrogenase (Joshi and Tabita 1996). The inability of strains 16 and 16PHC to grow under the selective conditions reported here is unTable 6 Growth of Rhodobacter sphaeroides and Rhodobacter capsulatus strains under various conditions and the ability of Rba. sphaeroides strains to acquire the CAC phenotype. Strain 16 (pJG336) is strain 16 complemented with the Rba. sphaeroides cbbI operon, and strain 16 (pJG106) is strain 16 complemented with the Rba. sphaeroides cbbII operon. (MAL/CO2 photoheterotrophic growth with malate as a carbon source, PA photolithoautotrophic growth, CA aerobic chemolithoautotrophic growth, and NA not applicable) Strain Growth phenotype MAL/CO2 PA Rba. sphaeroides HR HR-FI HR-FII 16 16 (pJG336) 16 (pJG106) 16PHC 1312 4X Rba. capsulatus SB1003 SBI SBII SBI-II + + + + + + + + + + + + + + + + + + + + CA + + + Ability to acquire CAC phenotype + + + + + NA NA NA NA

13

doubtedly due to the requirement for RubisCO and the CBB pathway for chemoautotrophic growth. Strain 1312 is a mutant strain of HR in which the cbbR gene (encoding an activator of the cbb operons; Gibson and Tabita 1993) is disrupted. This strain grows photoheterotrophically, but at a reduced rate due to decreased cbb gene expression, and is incapable of photolithoautotrophic growth (Gibson and Tabita 1993). The inability of this strain to acquire the CAC phenotype is probably due to its inability to express sufficient levels of RubisCO and other CBB pathway enzymes during autotrophic growth. Rba. sphaeroides strain HR-FI is unable to synthesize form I RubisCO due to the disruption of the cbbLcbbS genes, yet is still able to synthesize form II RubisCO, which supports photoheterotrophic and photolithoautotrophic growth (Gibson et al. 1991). Despite the fact that this strain did not adapt under the selective conditions (Table 6), a form I RubisCO-minus strain constructed from HR-CAC (strain CAC-FI) was able to grow as a chemoautotroph. It is likely that, with additional attempts or an increased adaptation period, a CAC strain of HR-FI could be selected. In addition to Rba. sphaeroides wild-type strain HR, strains HR-FII, 4X, and strain 16 containing cosmids pJG336 or pJG106 were able to gain the capacity for chemoautotrophic growth (Table 6). Strain HR-FII is unable to synthesize form II RubisCO due to the disruption of the cbbM gene (Gibson et al. 1991), but can still synthesize form I RubisCO. The growth of strain 16 complemented with plasmid pJG336 (containing the Rba. sphaeroides cbbI operon) or plasmid pJG106 (containing the Rba. sphaeroides cbbII operon) under selective conditions is consistent with the requirement for the CBB pathway during chemoautotrophic growth. Strain 4X was able to gain the CAC phenotype (Table 6) even though this strain lacks any detectable phosphoglycolate phosphatase activity (Gibson and Tabita 1997). Additionally, strains of Rba. capsulatus that synthesize only one or the other form of RubisCO (strains SBI and SBII) also maintained the ability to grow as chemoautotrophs, while a strain of Rba. capsulatus in which both RubisCO large subunit genes are disrupted (strain SBI-II) was unable to grow (Table 6).

at 2.5% CO2 displayed approximately two-thirds of both RubisCO activity and protein levels as compared to cultures grown at 1.5% CO2 (Jouanneau and Tabita 1986), RubisCO activities were determined for each strain at various CO2 concentrations during chemoautotrophic growth. For strain HR-CAC, the decrease in CO2 concentration from 5.0 to 2.5% resulted in approximately twice the RubisCO activity (Table 7). No further increase in RubisCO activity was observed when strain HR-CAC was grown in an atmosphere containing 0.75% CO2. Disruption of one RubisCO gene resulted in elevated synthesis of the other form of RubisCO when Rba. sphaeroides form I or form II RubisCO-minus strains were grown photoheterotrophically and photolithoautotrophically (Gibson et al. 1991). The relatively high level of RubisCO activity in strain FII-CAC as compared to HR-CAC at each CO2 concentration suggests that the synthesis of form I RubisCO was increased in response to the absence of form II RubisCO during chemoautotrophic growth (Table 7). It must be stressed that RubisCO protein levels were not determined directly, and this change in activity could theoretically be the result of some post-translational modification or alteration of RubisCO to compensate for the lack of form II RubisCO. Such enhancement of the catalytic activity of form I RubisCO has never been observed previously (Tabita 1995). No such compensation in form II RubisCO activity was observed for strain CAC-FI since the levels of RubisCO activity in this strain were uniformly low at each CO2 concentration (Table 7). Since strain CAC-FI was generated from a CAC derivative of strain HR, and strain FII-CAC was selected in an independent experiment, these strains are not necessarily isogenic. Thus, it is conceivable that apparent differences in the compensation of RubisCO activity by the two strains may be illusory and actually reflect the different genetic background of the two strains. The level of RubisCO activity at various CO2 concentrations was also determined for Rba. capsulatus wild-type

Growth and RubisCO activity of RubisCO deletion strains of Rba. sphaeroides and Rba. capsulatus Strains of Rba. sphaeroides and Rba. capsulatus lacking either form I or form II RubisCO were tested for the ability to support chemoautotrophic growth at various CO2 concentrations (0.255%). The oxygen concentration was kept constant at 10.5%. All of the strains grew with a doubling time of 26 3.9 h (results not shown). No decrease in growth rate was observed at the lower CO2 concentrations tested. Rba. sphaeroides HR-CAC and RubisCO-minus CAC derivatives of HR grew chemoautotrophically with O2 at each of the CO2 concentrations tested, indicating that either form of RubisCO can support chemoautotrophic growth. Since early studies indicated that photolithoautotrophic cultures of Rba. sphaeroides grown

Table 7 Ribulose 1,5-biphosphate carboxylase/oxygenase (RubisCO) activity in Rhodobacter sphaeroides and Rhodobacter capsulatus strains during aerobic chemolithoautotrophic growth at varying CO2 levels and constant levels of O2 (10.5%). Average of duplicate experiments; strain SBI grew only after a 12-day lag when grown with 0.25% CO2; ( not determined) Strain RubisCO activity [nmol min1 (mg protein)1] (% CO2) 5.0% Rba. capsulatus SB1003 SBI SBII 82 90 50 2.5% 185 57 224 1.5% 74 57 36 0.75% 183 62 257 0.25% 96 102 63

Rba. sphaeroides HR-CAC 93 CAC-FI 48 FII-CAC 293

14

and RubisCO-minus strains during chemoautotrophic growth (at constant levels of O2). Rba. capsulatus strains SB1003 and SBII grew at each of the CO2 concentrations tested (0.255%). The CO2 concentrations at which the Rba. capsulatus strains were grown were slightly different from those at which the Rba. sphaeroides strains were grown (Table 7). Strain SBI grew at both 5.0% CO2 and 1.5% CO2. At a concentration of 0.25% CO2, the SBI strain showed a markedly reduced capacity to grow; in one instance, the culture began to grow only after a 12-day lag (results not shown). In another experiment, no growth was observed. Strains SB1003 and SBII, on the other hand, grew after 2 days at 0.25% CO2. All of the strains grew with comparable doubling times (26 3.9 h) at each of the CO2 concentrations tested, including the SBI culture that began to grow after a 12-day lag. Whether or not SBI is unable to grow at 0.25% CO2 or if the SBI strain that grew after a long lag is a spontaneous mutant that has acquired the ability to grow under these conditions awaits further experimentation.

Discussion
In an earlier investigation, Rba. capsulatus has been shown to grow aerobically as a chemoautotroph (Madigan and Gest 1979). In the same study, Rhodospirillum rubrum and Rba. sphaeroides were found to be poorly capable or incapable of chemoautotrophic growth. We report here the isolation of a spontaneous mutant of Rba. sphaeroides strain HR that has acquired the ability to grow chemoautotrophically. This is not the first demonstration of such a gain-of-function mutation in nonsulfur purple phototrophic bacteria. Wild-type strains of Rba. capsulatus are unable to utilize glycerol as a carbon source, but Lueking et al. (1973) have isolated a spontaneous mutant of Rba. capsulatus that can assimilate glycerol. The mutant, strain L1, shows an increase in the activities of both glycerokinase and glycerophosphate dehydrogenase. Although the specific mutation that allows strain L1 to utilize glycerol has not been elucidated, the concomitant increase in the activity of two different enzymes suggests that the mutation is in a regulatory element affecting the synthesis or activity of these enzymes. Mutant strains of Rba. sphaeroides in which the cytochrome c2 gene (cycA) is disrupted are unable to grow photosynthetically (Donohue et al. 1988); however, spontaneous mutations that suppress the photosynthetic deficiency (spd mutants) occur, thus allowing cytochrome c2independent photosynthetic growth of the cycA strain (Rott and Donohue 1990). The spd mutants were found to synthesize an alternative cytochrome, referred to as isocytochrome c2, that was necessary and sufficient for cytochrome c2-independent photosynthetic growth (Rott et al. 1993). The gene encoding isocytochrome c2 (cycI) was found to be cotranscribed with adhI (encoding a glutathione-dependent formaldehyde dehydrogenase), suggesting that this alternative cytochrome is normally involved in energy metabolism when cells are grown on a

carbon source whose metabolism generates formaldehyde (Barber et al. 1996). The regulation of the adhI-cycI operon was apparently modified in the spd mutants to allow isocytochrome c2 synthesis, thus permitting photosynthetic growth. Another example of a spontaneous gain-offunction mutation occurred in the Rba. sphaeroides HR RubisCO-minus mutant, strain 16. Strain 16 grows photoheterotrophically only in the presence of an added external electron acceptor since the lack of RubisCO precludes the use of CO2 as an electron acceptor. However, strain 16PHC is a spontaneous mutant derived from strain 16 that retains the RubisCO-minus phenotype, yet is capable of photoheterotrophic growth in the absence of exogenous electron acceptors (Wang et al. 1993). Subsequent studies have shown that strain 16-PHC produces H2 gas and deregulates the synthesis of nitrogenase when grown in the presence of ammonia (Joshi and Tabita 1996). Strain 16PHC is thus able to utilize protons as an alternative electron acceptor, catalyzing their reduction via nitrogenase. The spd strains and strain 16PHC of Rba. sphaeroides have thus gained the ability to grow under nonpermissive conditions by modifying the regulation of genes whose products normally perform other functions in the cell. This represents an interesting form of evolutionary recruitment of existing functions to perform a similar task under new conditions. The ability to adapt to such diverse physiological conditions may be a consequence of the considerable metabolic versatility of the nonsulfur purple phototrophic bacteria. In contrast, growth of Rba. capsulatus strain L1 on glycerol was apparently due to the recovery of a function that had recently been lost. By analogy, since most of the wild-type Rba. capsulatus strains and at least one wildtype strain of Rba. sphaeroides grew chemoautotrophically, the ability to grow under these conditions may have been recently lost in Rba. sphaeroides and regained by strain HR-CAC. The chemoautotrophically competent (CAC) mutant of Rba. sphaeroides, strain HR-CAC, possessed levels of RubisCO activity comparable to that of the wild-type strain under a variety of growth conditions. The ability to acquire the CAC phenotype was dependent upon RubisCO synthesis in the parent strain. In addition, strains of Rba. capsulatus that have lost a functional CBB pathway because of disruptions either in the genes encoding both forms of RubisCO or in the single gene encoding phosphoribulokinase lose the ability to grow chemoautotrophically (Paoli and Tabita, manuscript submitted). Thus, chemoautotrophic growth of nonsulfur purple bacteria is dependent on the CBB pathway for CO2 fixation. In addition to potentially interesting regulatory alterations that accompany the development of the CAC phenotype of Rba. sphaeroides, there are other practical uses for the chemoautotrophic capacity of this strain and Rba. capsulatus. As discussed earlier, the RubisCO substrate specificity factor, , relates the initial velocities of carboxylation and oxygenation to the relative ratio of CO2 and O2: c/o = ([CO2]/[O2])

15

Since the oxygenase reaction causes the eventual loss of carbon from the cell, c/o must exceed some critical value in order for RubisCO to catalyze net carbon assimilation. For example, in higher plants it has been estimated that 0.5 carbons are lost for each O2 fixed (Gutteridge 1989). Therefore, c/o must be greater than 0.5 in order for RubisCO to catalyze net carbon fixation and support autotrophic growth. Accordingly, the [CO2]/[O2] ratio can be adjusted such that c/o would be growth-limiting. Incubation of an organism at the growth-limiting condition should provide a means for the biological selection of RubisCO with increased substrate specificity. The [CO2]/[O2] ratio at which growth limitation would occur would thus depend upon the of the RubisCO synthesized. In this vein, a mutant strain of Synechocystis sp. strain PCC 6803 has been constructed in which the natural RubisCO gene is replaced by the Rhodospirillum rubrum RubisCO gene (Pierce et al. 1989). Growth of this cyanorubrum strain is dependent on increased atmospheric CO2 due to the organisms dependence on the form II RubisCO from Rsp. rubrum for CO2-dependent growth. The cyanorubrum grows poorly at 0.5% CO2 and does not grow at 0.1% CO2 (in 21% O2). This study and recent modifications of this system (Amichay et al. 1993) show definite potential for the selection of mutant RubisCO molecules with increased substrate specificity. The physiological versatility of nonsulfur purple bacteria and their ability to grow photoheterotrophically and chemoheterotrophically, as well as chemoautotrophically (in both the absence and presence of oxygen), present some very interesting possibilities. For example, the strains of Rba. capsulatus and Rba. sphaeroides that synthesize only form I or form II RubisCO differ only in the of their respective RubisCO. A [CO2]/[O2] ratio at which the form-Isynthesizing strain grows but at which the form-II-synthesizing strain does not would be an ideal condition for selecting a form II RubisCO with increased substrate specificity. Analysis of such RubisCO mutants should allow for a greater understanding of the structural basis for the discrimination between the gaseous substrates. Although much further experimentation is required, preliminary experiments described here suggest that concentrations at or near 0.25% CO2 and 10.5% O2 may be a proper gas ratio to effect such selection. For example, the only difference between strains SBII and SBI is in the type of RubisCO each expresses. If the form-I-RubisCO-expressing strain (SBII) is able to grow at 0.25% CO2/10.5% O2, but the form-II-RubisCO-containing strain (SBI) does not or begins to experience difficulty at this and lower CO2 levels, differences in growth should reside in the inherent substrate specificity of the particular RubisCO synthesized by each strain. The RubisCO activity levels of Rba. sphaeroides HRCAC and Rba. capsulatus SB1003 are nearly the same at 5.0% CO2 (Table 7). Unlike Rba. sphaeroides, RubisCO activity in Rba. capsulatus SB1003 did not increase in response to lower CO2 concentrations during chemoautotrophic growth. At each CO2 concentration tested, the level of form I RubisCO activity in strain SBII was approximately half of the form II RubisCO activity in strain

SBI.

This is opposite to the response observed in the RubisCO-minus strains of Rba. sphaeroides in which the form-I-expressing strain displayed higher activities than did the form-II-expressing strain. Indeed, differences in absolute RubisCO activity levels between Rba. capsulatus and Rba. sphaeroides are most pronounced in strains expressing only form I RubisCO, strains SBII and FIICAC, respectively. If these differences reflect form I protein synthesis in both Rba. capsulatus and Rba. sphaeroides, the expression of the cbbI gene of each organism is undoubtedly subject to different modes of control, perhaps reflective of the different evolutionary origins of form I RubisCO from each organism (Paoli et al. 1998). In addition, form I RubisCO from these two species possesses different biochemical properties; subsequent work in our laboratory (K. Horken, personal communication) has shown that the form I enzyme from Rba. capsulatus has an unusually low in keeping with its unique evolutionary lineage. Finally, in addition to the Rba. capsulatus strains expressing only one form of RubisCO, a strain has been constructed in which neither RubisCO is expressed (Paoli and Tabita, manuscript submitted). This strain, SBI-II, does not grow chemoautotrophically, but can be complemented to both photolithoautotrophic and chemoautotrophic growth by a variety of RubisCO genes cloned into a RubisCO expression vector. RubisCO expression constructs could then be randomly mutagenized and introduced into strain SBI-II to screen large numbers of mutant RubisCOs for increased substrate specificity. The metabolic versatility, genetic adaptability, and ease with which the nonsulfur photosynthetic bacteria may be genetically manipulated make the various strains discussed here and the diversity of RubisCO clones available to us an ideal system for the biological selection of RubisCO with increased substrate specificity.

Acknowledgements The authors would like to thank M. Madigan for providing the additional Rba. capsulatus and Rba. sphaeroides wild-type strains and M. Harding, whose initial work led to the isolation of the CAC strain of Rba. sphaeroides. This work was supported by U. S. Public Health Service grant no. GM24497 from the NIH.

References
Amichay D, Levitz R, Gurevitz M (1993) Construction of a Synechocystis PCC 6803 mutant suitable for the study of variant hexadecameric ribulose bisphosphate carboxylase/oxygenase enzymes. Plant Mol Biol 23:465476 Barber RD, Rott MA, Donohue TJ (1996) Characterization of a glutathione-dependent formaldehyde dehydrogenase from Rhodobacter sphaeroides. J Bacteriol 178:13861393 Beh M, Strauss G, Huber R, Stetter KO, Fuchs G (1993) Enzymes of the reductive tricarboxylic acid cycle in the autotrophic eubacterium Aquiflex pyrophilus and the archaebacterium Thermoproteus neutrophilus. Arch Microbiol 160:306311 Codd GA, Smith BM (1974) Glycollate formation by the purple photosynthetic bacterium Rhodospirillum rubrum. FEBS Lett 48:105108 Codd GA, Turnbull F (1975) Enzymes of glycollate formation and oxidation in two members of the Rhodospirillaceae (purple non-sulphur bacteria). Arch Microbiol 104:155158

16 Codd GA, Bowien B, Schlegel HG (1976) Glycollate production and excretion by Alcaligenes eutrophus. Arch Microbiol 110: 167171 Donohue TJ, McEwan AG, Van Doren S, Crofts AR, Kaplan S (1988) Phenotypic and genetic characterization of a cytochrome c2 deficient mutant of Rhodobacter sphaeroides. Biochemistry 27:19181924 Falcone DL, Tabita FR (1991) Expression of endogenous and foreign ribulose 1,5-bisphosphate carboxylase-oxygenase (RubisCO) genes in a RubisCO deletion mutant of Rhodobacter sphaeroides. J Bacteriol 173:20992108 Gibson JL (1995) Genetic analysis of CO2 fixation genes. In: Blankenship RE, Madigan MT, Bauer CE (eds) Anoxygenic photosynthetic bacteria. Kluwer, Dordrecht, pp 11071124 Gibson JL, Tabita FR (1977a) Different molecular forms of Dribulose-1,5-bisphosphate carboxylase from Rhodopseudomonas sphaeroides. J Biol Chem 252:943949 Gibson JL, Tabita FR (1977b) Isolation and preliminary characterization of two forms of ribulose 1,5-bisphosphate carboxylase from Rhodopseudomonas capsulata. J Bacteriol 132:818823 Gibson JL, Tabita FR (1985) Structural differences in the catalytic subunits of form I and form II ribulose 1,5-bisphosphate carboxylase/oxygenase from Rhodopseudomonas sphaeroides. J Bacteriol 164:11881193 Gibson JL, Tabita FR (1993) Nucleotide sequence and functional analysis of CbbR, a positive regulator of the Calvin cycle operons of Rhodobacter sphaeroides. J Bacteriol 175:57785784 Gibson JL, Tabita FR (1996) The molecular regulation of the reductive pentose phosphate pathway in proteobacteria and cyanobacteria. Arch Microbiol 166:141150 Gibson JL, Tabita FR (1997) Analysis of the cbbXYZ operon in Rhodobacter sphaeroides. J Bacteriol 179:663669 Gibson JL, Falcone DL, Tabita FR (1991) Nucleotide sequence, transcriptional analysis and expression of genes encoded within the form I CO2 fixation operon of Rhodobacter sphaeroides. J Biol Chem 266:1464614653 Gutteridge S (1989) Limitations of the primary events of CO2 fixation in photosynthetic organisms: the structure and mechanism of RubisCO. Biochim Biophys Acta 1015:114 Jordan DB, Ogren WL (1981) Species variation in the specificity of ribulose bisphosphate carboxylase/oxygenase. Nature 291: 513515 Joshi HM, Tabita FR (1996) A global two component signal transduction system that integrates the control of photosynthesis, carbon dioxide assimilation, and nitrogen fixation. Proc Natl Acad Sci USA 93:14 51514 520 Jouanneau Y, Tabita FR (1986) Independent regulation of synthesis of form I and form II ribulose bisphosphate carboxylaseoxygenase in Rhodopseudomonas sphaeroides. J Bacteriol 165: 620624 Kampf C, Pfennig N (1980) Capacity of Chromatiaceae for chemotrophic growth. Specific respiration rates of Thiocystis violacea and Chromatium vinosum. Arch Microbiol 127:125 135 Kampf C, Pfennig N (1986 a) Isolation and characterization of some chemoautotrophic Chromatiacea. J Basic Microbiol 26: 507515 Kampf C, Pfennig N (1986 b) Chemoautotrophic growth of Thiocystis violacea, Chromatium gracile and C. vinosum in the dark at various O2 concentrations. J Basic Microbiol 26:517531 Kawasumi T, Igarashi Y, Kodama T, Minoda Y (1984) Hydrogenobacter thermophilus gen. nov., sp. nov., an extremely thermophilic, aerobic, hydrogen-oxidizing bacterium. Int J Syst Bacteriol 34:510 Kondratieva EN (1989) Chemolithotrophy of phototrophic bacteria. In: Schlegel HG, Bowien B (eds) Autotrophic bacteria. Science Tech Publishers, Madison, pp 283287 Kondratieva EN, Zhukov VG, Ivanovsky RN, Petushkova YP, Monosov EZ (1976) The capacity of phototrophic sulfur bacterium Thiocapsa roseopersicina for chemosynthesis. Arch Microbiol 108:287292 Lueking D, Tokuhisa D, Sojka G (1973) Glycerol assimilation by a mutant of Rhodopseudomonas capsulata. J Bacteriol 115: 897903 Madigan MT, Gest H (1979) Growth of the photosynthetic bacterium Rhodopseudomonas capsulata chemoautotrophically in darkness with H2 as the energy source. J Bacteriol 137:524 530 Markwell MAK, Haas SM, Bieber LL, Tolbert NE (1978) A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples. Anal Biochem 87:206210 Ormerod JG, Ormerod KS, Gest H (1961) Light-dependent utilization of organic compounds and photoproduction of molecular hydrogen by the photosynthetic bacteria; relationships with nitrogen metabolism. Arch Biochem Biophys 94:449463 Paoli GC, Strom-Morgan N, Shively JM, Tabita FR (1995) Expression of the cbbLcbbS and cbbM genes and distinct organization of the cbb Calvin cycle structural genes of Rhodobacter capsulatus. Arch Microbiol 164:396405 Paoli GC, Soyer F, Shively J, Tabita FR (1998) Rhodobacter capsulatus genes encoding form I ribulose-1,5-bisphosphate carboxylase/oxygenase (cbbLS) and neighboring genes were acquired by a horizontal gene transfer. Microbiology 144: 219 227 Pierce J, Carlson TJ, Williams JGK (1989) A cyanobacterial mutant requiring the expression of ribulose bisphosphate carboxylase from a photosynthetic anaerobe. Proc Natl Acad Sci USA 86:57535757 Read BA, Tabita FR (1992) A hybrid ribulose bisphosphate carboxylase/oxygenase enzyme exhibiting a substantial increase in substrate specificity factor. Biochemistry 31:519525 Read BA, Tabita FR (1994) High substrate specificity factor ribulose bisphosphate carboxylase/oxygenase from eukaryotic marine algae and properties of recombinant cyanobacterial rubisco containing algal residue modifications. Arch Biochem Biophys 312:210218 Richardson DJ, King GF, Kelly DJ, McEwan AG, Ferguson SJ, Jackson JB (1988) The role of auxiliary oxidants in maintaining redox balance during growth of Rhodobacter capsulatus on propionate and butyrate. Arch Microbiol 150:131137 Rott MA, Donohue TJ (1990) Rhodobacter sphaeroides spd mutations allow cytochrome c2-independent photosynthetic growth. J Bacteriol 172:19541961 Rott MA, Witthuhn VC, Schilke BA, Soranno M, Ali A, Donohue TJ (1993) Genetic evidence for the role of isocytochrome c2 in photosynthetic growth of Rhodobacter sphaeroides spd mutants. J Bacteriol 175:358366 Schlegel HG (1989) Aerobic hydrogen-oxidizing (knallgas) bacteria. In: Schlegel HB, Bowien B (eds) Autotrophic bacteria. Science Tech Publishers, Madison, pp 305329 Shiba H, Kawasumi T, Igarashi Y, Kodama T, Minoda Y (1985) The CO2 assimilation via the reductive tricarboxylic acid cycle in an obligately autotrophic, aerobic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus. Arch Microbiol 141: 198203 Shively JM, Davidson E, Marrs BL (1984) Depression of the synthesis of the intermediate and large forms of ribulose-1,5-bisphosphate carboxylase/oxygenase in Rhodopseudomonas capsulatus. Arch Microbiol 138:233236 Siefert E, Pfennig N (1979) Chemoautotrophic growth of Rhodopseudomonas species with hydrogen and chemotrophic utilization of methanol and formate. Arch Microbiol 122:177182 Simon R, Priefer U, Puhler A (1983) A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram-negative bacteria. Biotechnology 1:784791 Tabita FR (1988) Molecular and cellular regulation of autotrophic carbon dioxide fixation in microorganisms. Microbiol Rev 53: 155189 Tabita FR (1995) The biochemistry and molecular regulation of carbon metabolism and CO2 fixation in purple bacteria. In: Blankenship RE, Madigan MT, Bauer CE (eds) Anoxygenic photosynthetic bacteria. Kluwer, Dordrecht, pp 885914

17 Uemura K, Anwaruzzaman, Miyachi S, Yokota A (1997) Ribulose-1,5-bisphosphate carboxylase/oxygenase from thermophilic red algae with a strong specificity for CO2 fixation. Biochem Biophys Res Commun 233:568571 Wang X, Falcone DL, Tabita FR (1993) Reductive pentose phosphate-independent CO2 fixation in Rhodobacter sphaeroides and evidence that ribulose bisphosphate carboxylase/oxygenase activity serves to maintain redox balance of the cell. J Bacteriol 175:33723379 Weaver KE, Tabita FR (1983) Isolation and partial characterization of Rhodopseudomonas sphaeroides mutants defective in the regulation of ribulose bisphosphate carboxylase/oxygenase. J Bacteriol 156:507515 Weaver PF, Wall JD, Gest H (1975) Characterization of Rhodopseudomonas capsulata. Arch Microbiol 105:207216 Yen HC, Marrs B (1976) Map of genes for carotenoid and bacteriochlorophyll biosynthesis in Rhodopseudomonas capsulatus. J Bacteriol 126:619629