INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

It is well established that the synthesis of both form I and form II ribulose 1,5-bisphosphate (RuBP) carboxylase/oxygenase (RubisCO) from the nonsulfur purple bacterium Rhodobacter capsulatus is highly dependent on the environmental growth conditions (8, 33). Yet, aspects of the molecular biology of autotrophic carbon dioxide metabolism and RubisCO synthesis in this organism have only recently been stressed (26-28), and there is little information available on the enzymology of RubisCO in this versatile organism. This neglect has largely been due to the fact that the regulation and biochemistry of CO₂ assimilation are highly studied in the closely related organism Rhodobacter sphaeroides (6, 16, 31, 36). Early immunological studies, however, indicated that there were substantial and unexpected differences between the form I enzymes of R. capsulatus and R. sphaeroides (8), which was later substantiated by DNA hybridization results (34) and additional immunological studies (26). The distinctness of the form I RubisCO of R. capsulatus was recently confirmed during a recent study of the organization and regulation of the CO₂ fixation genes of this organism (28). During this investigation, it was noted that deduced amino acid sequences of the large and small subunits of form I RubisCO from R. capsulatus and R. sphaeroides were surprisingly different, despite the close phylogenetic relatedness of the two organisms based on other molecular indicators (27). In fact, the sequence results placed the R. capsulatus form I RubisCO among members of what has been termed the green-type RubisCO group (4, 38), which is comprised of // chemoautotrophic proteobacterial and cyanobacterial enzymes (27). It is apparent that all genes of the cbb_I regulon of R. capsulatus are derived from chemoautotrophic bacteria via some form of horizontal gene transfer (27), whereas genes of the cbb_II regulon of this organism closely resemble those from its close relative R. sphaeroides (21, 28). Because of the potential uniqueness of form I RubisCO of R. capsulatus and the availability of recombinant clones (26, 28), it was suggested that detailed study of this enzyme might yield useful insights concerning important molecular properties of a class of RubisCO that has not been highly studied to this point (36). Such studies are important because the determinants responsible for discrimination between either CO₂ or O₂ as gaseous substrate are largely unknown. The capacity to preferentially use either CO₂ or O₂ is a property that has great physiological relevance, and yet it can vary significantly depending on the evolutional form of the enzyme (36). Thus, the potential to use the great genetic capacity of R. capsulatus to learn more of the molecular basis for CO₂/O₂ specificity for an unusual type RubisCO about which little is known was an important consideration in initiating such studies. Finally, despite the relative unrelatedness of R. capsulatus and R. sphaeroides form I RubisCO amino acid sequences (58% identity), the large subunit genes from each organism (of similar GC content) are colinear over the majority of the coding region. These sequences would thus be quite amenable for the construction of molecular chimeras or large subunit-small subunit hybrids without disrupting the reading frame. Such studies would be important if kinetic and other properties of the two form I enzymes, especially the CO₂/O₂ substrate specificity, proved to be different. For these reasons, an investigation of the molecular properties and enzymology of form I RubisCO from R. capsulatus was initiated. Because the enzymological properties of this enzyme were found to be unique, preliminary efforts were also begun to produce hybrid large subunit-small subunit R. capsulatus-R. sphaeroides form I enzymes and molecular chimeras, with the eventual goal of pinpointing molecular determinants responsible for key properties.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Growth of R. capsulatus and E. coli strains. R. capsulatus wild-type strain SB1003 and form II RubisCO-deficient mutant strain SBII (28) were grown photoautotrophically in Ormerod's minimal medium (25) supplemented with kanamycin (5 µg/ml) as previously described (26). Cells were grown to late log phase (optical density at 600 nm of >1.0) in 0.5-liter batch cultures, harvested by centrifugation, and washed with TEDM (25 mM Tris-Cl [pH 8.0], 1 mM EDTA, 10 mM MgCl₂, 2 mM dithiothreitol [DTT]).

Enzyme purification. Crude lysates were prepared from washed cell pellets resuspended (1 ml/g [wet weight]) in TEDM adjusted to 1 mM phenylmethylsulfonyl fluoride (added just before cell breakage). Cell suspensions were passed through a French pressure cell (SLM Instruments, Inc., Urbana, Ill.) at 16,000 lb/in² three times, and cell debris was removed by centrifugation at 12,000 × g to yield a clear cell lysate. Contaminating chromosomal DNA in the clear cell lysate was hydrolyzed by incubating the extract with DNase I (100 µg/ml) for 30 min at room temperature. All subsequent purification steps were carried out at 4°C unless otherwise indicated. A soluble protein extract was then obtained after centrifugation at 100,000 × g for 1 h; the supernatant was retained and RubisCO was partially purified from this fraction by Green A (Amicon, Beverly, Mass.) dye affinity chromatography using a 2.5- by 4-cm column equilibrated in TEMMB (25 mM Tris-HCl [pH 7.2], 1 mM EDTA, 5 mM -mercaptoethanol, 10 mM MgCl₂, 50 mM NaHCO₃). The Green A column was washed with equilibration buffer until the A₂₈₀ of the eluent returned to baseline (~250 ml); RubisCO was then eluted with a 300-ml linear salt gradient (0 to 1 M KCl) in TEMMB. Peak fractions were pooled and concentrated in an Amicon Centriplus 30 concentrator. RubisCO was purified to homogeneity by loading the concentrated Green A peak onto a Pharmacia Mono Q HR 5/5 column equilibrated with 10 mM KCl in TEM (TEDM minus DTT) and subjected to a 5-ml 10 to 500 mM KCl gradient. RubisCO-containing fractions were desalted using a Bio-Rad (Richmond, Calif.) DG10 column equilibrated with 20% glycerol in TEM (at room temperature), quickly frozen with liquid nitrogen, and stored at 70°C.

Subcloning RubisCO large and small subunit genes. To simplify future R. capsulatus-R. sphaeroides RubisCO chimera constructions involving EcoRI, all genes were subcloned into a derivative of the pK19 expression vector (30), pK19E1, which lacks an EcoRI site (this study). The cbbS gene of R. sphaeroides (cbbS_s) was subcloned as an NruI fragment from pRSF1A, which contains both the large and small subunit genes from R. sphaeroides, into the SmaI site of pK19E1, generating plasmid pRsS-17 (Fig. 1). Proper orientation of the insert with respect to the lac promoter was confirmed by DNA sequencing with the universal forward primer. The R. sphaeroides RubisCO large subunit gene (cbbL_s) was PCR amplified by using pRSF1A as a template and custom forward (5'-GGAAGCTTA TGCTTCGAAGATCACCG-3') and reverse (5'-GGTCTAGAAGCAGCCTTGGGTGA TGC-3') primers (Operon Technologies, Inc., Alameda, Calif.) that introduced unique 5' HindIII and 3' XbaI sites (underlined), respectively. Optimal synthesis of the desired 1.6-kb product was obtained under the following reaction conditions: 1× U.S. Biochemical (Cleveland, Ohio) Taq polymerase buffer, 200 µM deoxynucleoside triphosphates, 1.7 pmol of template (5.22 kb), 0.4 pmol of primers, 1.0 mM MgCl₂, and 5 U of Taq polymerase (U.S. Biochemical). The PCR program was 3 min at 95°C for the denaturation step; 30 cycles of 1 min at 95°C, 1 min at 60°C, and 3 min at 72°C; and 7 min at 72°C for the extension step. The resultant 1.6-kb PCR product was subcloned into pUC19 (39) as an HindIII/XbaI fragment, generating plasmid pRsL-1 (Fig. 1).

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Diagrammatic representation of hybrid and wild-type RubisCO holoenzyme expression plasmids. The plasmids were named such that the origin of each subunit gene is indicated by the subscripted first letter of the species from which it originated. The bold arrow indicates the location of the lac promoter with respect to the cbb genes, which are all in the proper transcriptional orientation for controlled expression with this promoter. Open and shaded rectangles represent coding regions originating from R. sphaeroides and R. capsulatus, respectively. Abbreviations for restriction sites: H, HindIII; X, XbaI; B, BamHI; and K, KpnI.

Construction of wild-type, hybrid, and chimera RubisCO expression plasmids. The restriction site incorporation and cloning steps described above facilitated the simple construction of wild-type (used as positive controls in expression experiments) and hybrid holoenzyme expression plasmids (diagrammatically represented in Fig. 1). To reconstruct an R. sphaeroides cbbL-cbbS expression plasmid (pL_sS_s), the large subunit gene (cbbL) was directionally subcloned as a HindIII/XbaI fragment from pRsL-1 into pRsS-17 (Fig. 1). Similarly, to express a hybrid between R. sphaeroides large and R. capsulatus small subunits, cbbL_s was subcloned upstream of R. capsulatus cbbS (cbbS_c) in pRcS-4, producing plasmid pL_sS_c (Fig. 1). Restriction digests and DNA sequencing of the 5' and 3' ends of both plasmid inserts (pL_sS_s and pL_sS_c) with universal and reverse primers confirmed that the desired clones were made (data not shown). In addition, a large subunit chimera construct was made between R. sphaeroides and R. capsulatus, using a colinear XhoI restriction site in both genes (Fig. 1). Plasmid pQ, containing the R. capsulatus cbbL and cbbS genes, was digested with HindIII/XhoI and ligated to the ~1.0-kb HindIII/XhoI fragment (result of a partial digest with XhoI) of pRsL-1 (containing the R. sphaeroides cbbL gene), producing plasmid ps/cL-Xho1, which contains a chimeric cbbL gene and a wild-type R. capsulatus cbbS gene (Fig. 1). Restriction digests and DNA sequencing of the insert's 5' and 3' ends confirmed that the desired clone was obtained (data not shown). In addition, an approximately 650 bp EcoRI fragment of ps/cL-Xho1 (which would not exist in the parent plasmid, pQ) that encompasses the splice site of the chimeric cbbL_s/c was subcloned and sequenced to confirm that the desired chimera was constructed.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

RubisCO gene expression and enzyme purification. The highest levels of form I RubisCO (~15% of soluble protein) were obtained from R. capsulatus SBII, a strain containing an inactivated form II (cbbM) RubisCO gene. This was particularly evident when strain SBII was grown under photolithoautotrophic conditions to late log phase (data not shown). Since previous studies (28) had indicated that the absence of form II RubisCO leads to a compensatory increase in form I RubisCO synthesis, these results were not surprising. However, it was further observed that 90 to 95% of the RubisCO activity was found in the particulate fraction when cell extracts were subjected to centrifugation at 100,000 × g. This phenomenon was observed when the form I RubisCO was prepared both from R. capsulatus (SB1003 and SBII) and from E. coli FIPB; thus, the potential formation of carboxysomes in R. capsulatus to account for these results was ruled unlikely. This is in agreement with Southern hybridization data that showed no evidence for a ccmK homologue (a gene product associated with carboxysome structures) when R. capsulatus chromosomal DNA was probed at very low stringency (25a). Western analysis (results not shown) and staining of SDS-polyacrylamide gels containing extracts from photolithoautotrophically grown R. capsulatus SB1003 (a growth condition where both forms of RubisCO are expressed), before and after centrifugation at 100,000 × g, established that the sedimentation phenomenon was specific for the form I enzyme. This is indicated by a decrease in the relative amount of form I protein in the 100,000 × g fraction (Fig. 2). Further experimentation demonstrated that this enzyme had an unusually high affinity for DNA, causing it to cosediment with chromosomal DNA during high-speed centrifugation. The majority of the enzyme activity in cell extracts could be recovered in the 100,000 × g supernatant if the extract was pretreated with DNase I to hydrolyze contaminating chromosomal DNA. Other pretreatments, such as incubation in the presence of 1% Nonidet P-40, 0.5 M NaCl, and 10 mM EDTA, did not prevent sedimentation (data not shown). Application of the SBII soluble protein extract to a Green A dye affinity column, washing with TEMMB (pH 7.2), and elution with a KCl gradient (Table 1) resulted in a 1.6-fold purification. Mono Q anion-exchange chromatography of the Green A peak material yielded an enzyme preparation that was judged greater than 95% pure by SDS-PAGE (results not shown) and exhibited a specific activity of >3.0 U/mg under optimal assay conditions. N-terminal sequencing, after elution of SDS-PAGE-purified large subunits from gels, corresponded well with the deduced amino acid sequence (27) except for a missing N-terminal methionine residue as seen with other sequenced RubisCO large subunits (19). The purified enzyme was desalted, adjusted to 20% glycerol, quickly frozen in liquid N₂, and stored without loss of activity at 70°C for months.

View larger version (63K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE of R. capsulatus SB1003 cell extracts and purified form I RubisCO. Proteins were electrophoresed in a 15% denaturing polyacrylamide resolving minigel; 15 µg of cell extract (lanes 1 and 2) and 1 µg of purified holoenzyme (lane 3) were loaded. Small subunits were electrophoresed off this gel to enhance the resolution between form I and form II large subunits (LSU; bands indicated on the left). Lanes 1 and 2 were loaded with 10,000 × g and 100,000 × g supernatant fractions, respectively.

Enzymatic characterization of R. capsulatus form I RubisCO. The activity of purified R. capsulatus form I RubisCO was particularly sensitive to low protein concentrations (<5 µg/ml), as only about 30% of its optimal activity was obtained. Activity increased by 50% with the addition of DTT (5 mM) and increased by 40 to 370% (depending on the amount of RubisCO assayed) when both DTT and bovine serum albumin (100 µg/ml) were included in reaction mixtures (results not shown). When the enzyme was carbamylated to convert the enzyme to an active ternary complex (24), i.e., by preincubation in the presence of Mg²⁺ and CO₂, the highest activities were obtained at 30°C in the presence of 5 mM DTT (activity increased by 152% over the untreated enzyme). The presence of DTT improved activity at all temperatures but was maximally effective at 30°C. Preincubation temperatures of 37 and 48°C progressively inhibited enzyme activity (activity decreased by 92% in the absence of DTT).

View larger version (11K): [in this window] [in a new window]   FIG. 3.   Preincubation and effect of RuBP on the activation of R. capsulatus form I RubisCO. Enzyme was desalted with a Bio-Rad DG10 column and preincubated either in the presence of standard concentrations of magnesium and bicarbonate or in the presence of RuBP for 5 min before initiation of the reaction with the components needed to complete the assay mixture (23). At the indicated times, aliquots of the assay mixture were acidified and then counted by liquid scintillation spectroscopy; 0.5 µg of purified enzyme was used in the reaction.

Expression of wild-type, hybrid, and chimeric RubisCOs in E. coli. Initial steps were made to begin uncovering relevant structural differences between the form I RubisCOs from R. capsulatus and R. sphaeroides. Although the primary sequences are somewhat different (27), large subunits and small subunits still show 58 and 34% identity at the amino acid level. Few hybrid and chimeric RubisCO mutants have been expressed in vivo to study structure-function relationships. However, vastly dissimilar recombinant small subunits from a eukaryotic nongreen alga could assemble with cyanobacterial large subunits to form a holoenzyme with a substrate specificity factor enhanced by 60% (32). Despite a 20-fold loss in carboxylase activity, this hybrid provided insights concerning individual subunit involvement in regulation by sugar phosphate metabolites (32). Not all attempts to make hybrid RubisCOs have been so fruitful. Lee et al. (22) described a hybrid consisting of cyanobacterial large subunits and small subunits from the chemoautrophic bacterium Ralstonia eutropha (Alcaligenes eutrophus) which exhibited similar kinetic properties (except a 85% reduction in k_cat) to the natural cyanobacterial holoenzyme. Catalytically active hybrids have also been successfully made in vitro by combining partially purified large and small subunits (1, 2, 12, 14), but in most cases Michaelis constants and other enzymological properties were not investigated. In any case, the previous results cited above suggested that in vivo hybrid and chimera construction experiments with the requisite genes of R. capsulatus and R. sphaeroides, which are more closely related than the above algal-cyanobacterial hybrid components, might be quite interesting. In addition, the sequences from R. capsulatus and R. sphaeroides do present some advantages for such constructions, including the ability to make facile chimeras without disrupting the reading frame. Identifying regions of the RubisCO molecule that are germane to the various properties discussed above may help explain the disparity in specificity values and other kinetic parameters. Eventually, it is hoped that these types of studies will suggest the participation of new residues or domains that have so far gone unnoticed in studies using current X-ray structural models. Considering the differences in amino acid sequence (27) and kinetic properties (Table 2), there was the potential to observe dramatic effects upon construction of hybrids and chimeras with the genes encoding large and small subunits of the R. capsulatus and R. sphaeroides enzymes.

Concluding remarks. High-yield purification of R. capsulatus form I RubisCO has been described. Enzymological properties determined here firmly establish its green-like nature, and the results corroborate the suggestion that the genes for this enzyme were likely transmitted to an ascendant of R. capsulatus via a horizontal gene transfer event (27). Of particular importance are the low CO₂/O₂ substrate specificity, the high affinity for CO₂, and the inhibition by RuBP. Failure to successfully assemble recombinant hybrid holoenzymes in vivo with subunits from another purple nonsulfur bacterium further indicate the vast differences of the R. capsulatus protein, as does a preliminary attempt to construct a chimera between large subunits of these proteins. Finally, observed similarities in primary structure of the R. capsulatus and cyanobacterial form I enzymes, and the fact that the K_CO2 is so different for the two enzymes, suggest that the construction of in vivo chimeras between these two proteins might be a worthwhile endeavor.