Structural and Functional Similarities between a Ribulose-1,5-bisphosphate Carboxylase/Oxygenase (RuBisCO)-like Protein from Bacillus subtilis and Photosynthetic RuBisCO*

The sequences classified as genes for various ribulose-1,5-bisphosphate (RuBP) carboxylase/oxygenase (RuBisCO)-like proteins (RLPs) are widely distributed among bacteria, archaea, and eukaryota. In the phylogenic tree constructed with these sequences, RuBisCOs and RLPs are grouped into four separate clades, forms I-IV. In RuBisCO enzymes encoded by form I, II, and III sequences, 19 conserved amino acid residues are essential for CO2 fixation; however, 1-11 of these 19 residues are substituted with other amino acids in form IV RLPs. Among form IV RLPs, the only enzymatic activity detected to date is a 2,3-diketo-5-methylthiopentyl 1-phosphate (DK-MTP-1-P) enolase reaction catalyzed by Bacillus subtilis, Microcystis aeruginosa, and Geobacillus kaustophilus form IV RLPs. RLPs from Rhodospirillum rubrum, Rhodopseudomonas palustris, Chlorobium tepidum, and Bordetella bronchiseptica were inactive in the enolase reaction. DK-MTP-1-P enolase activity of B. subtilis RLP required Mg2+ for catalysis and, like RuBisCO, was stimulated by CO2. Four residues that are essential for the enolization reaction of RuBisCO, Lys175, Lys201, Asp203, and Glu204, were conserved in RLPs and were essential for DK-MTP-1-P enolase catalysis. Lys123, the residue conserved in DK-MTP-1-P enolases, was also essential for B. subtilis RLP enolase activity. Similarities between the active site structures of RuBisCO and B. subtilis RLP were examined by analyzing the effects of structural analogs of RuBP on DK-MTP-1-P enolase activity. A transition state analog for the RuBP carboxylation of RuBisCO was a competitive inhibitor in the DK-MTP-1-P enolase reaction with a Ki value of 103 μm. RuBP and d-phosphoglyceric acid, the substrate and product, respectively, of RuBisCO, were weaker competitive inhibitors. These results suggest that the amino acid residues utilized in the B. subtilis RLP enolase reaction are the same as those utilized in the RuBisCO RuBP enolization reaction.

Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) 4 catalyzes the carboxylation and oxygenation reactions of ribulose 1,5-bisphosphate (RuBP) in photosynthesis (1)(2)(3)(4). This enzyme is the sole CO 2 -fixing enzyme in plants; however, it has certain inefficiencies. It has a very low turnover rate, a low affinity for the substrate, CO 2 , and low specificity between the carboxylation and oxygenation reactions (5)(6)(7). Thus, the intrinsic enzymatic properties of RuBisCO are inadequate for efficient incorporation of CO 2 into organic matter in photosynthesis (7). However, plants have overcome these disadvantages by investing a huge amount of leaf nitrogen in RuBisCO synthesis (8).
In nature, there are wide variations in the properties and primary sequences of RuBisCO among different photosynthetic organisms (9 -12). The primary sequences vary as much as 73% without loss of activity. The relative specificity ranges from ϳ0.5 in a small subunitless RuBisCO to 238 in a red algal, hexadecameric RuBisCO (13,14). The affinity for CO 2 varies some 100-fold (15). Comparisons between these kinetic parameters and the primary sequences are expected to reveal promising strategies for improving the enzyme, and many studies have been conducted on this topic (7, 16 -18).
A RuBisCO-like protein (RLP) with no CO 2 -fixing activity was first demonstrated in Chlorobium tepidum (19), and a similar protein in Bacillus subtilis was found to be involved in the methionine salvage pathway (20). These findings have pointed to a new direction in RuBisCO research (17,21). The phylogenetic tree of the catalytic subunits of RuBisCOs and their homologs shows four major clusters, forms I-III, and form IV (Fig. 1A). Form I and II RuBisCOs are involved in photosynthetic or chemosynthetic CO 2 fixation, whereas the metabolic function of form III RuBisCOs remains unclear, although they can fix CO 2 on RuBP (9,22). Forms I-III conserve almost all 19 amino acid residues that are essential for CO 2 fixation in RuBisCO (Fig. 1B). The form IV cluster in the phylogenetic tree consists of RLPs that show ϳ20% homology to plant form I or bacterial form II RuBisCOs (12,20,21,(23)(24)(25). There are 8 -18 RuBisCO-essential residues that are conserved in RLPs (Fig.  1B). Form IV RLPs are further subdivided into four groups; ␣1, ␣2, ␤, and ␥ (21). The RLP of B. subtilis is classified in ␣1 and catalyzes the enolization reaction of 2,3-diketo-5-methylthiopentyl 1-phosphate (DK-MTP-1-P) but not the carboxylation of RuBP ( Fig. 2A) (20,21,23). The absence of CO 2 -fixing activity in the B. subtilis RLP may be ascribed to changes in 8 of the 19 amino acid residues essential for CO 2 fixation in RuBisCO (Fig. 1B). Several of these residues are located at the C-terminal domains of B. subtilis RLP and RuBisCO. The dimeric RuBisCO from Rhodospirillum rubrum catalyzes the DK-MTP-1-P enolase reaction with very low activity (20). These findings, together with the similarity in the chemical structures of substrates for B. subtilis RLP and RuBisCO ( Fig. 2A), suggest that they may have a close evolutionary relationship (12,21,(23)(24)(25).
The RuBisCO reaction starts with the abstraction of the C3 proton from RuBP to form the cis-enediol(ate) of RuBP ( Fig.  2A) (26). Using the spinach numbering format to identify RuBisCO and RLP residues, the carbamate formed on the ⑀amino group of Lys 201 may be the general base to abstract the proton, and the cis-enediol(ate) form of RuBP is stabilized in the combination of side chains from Lys 175 and His 294 (27). Asp 203 , Glu 204 , and the carbamate Lys 201 of the enzyme active site stabilize the cis-enediol(ate) and CO 2 through the Mg 2ϩ ion (26). The B. subtilis RLP abstracts the C1 proton of its substrate DK-MTP-1-P to start the DK-MTP-1-P enolization reaction (12,21,23). The ⑀-amino group of Lys 123 is thought to be required for the abstraction of the 1-proS proton in the Geobacillus kaustophilus RLP, which belongs to group ␣1, together with the B. subtilis RLP (Fig. 2B) (25). Lys 123 is conserved among DK-MTP-1-P enolases and resides very near the C1 of 2,3-diketohexane 1-phosphate (DK-H-1-P), a structural analogue of DK-MTP-1-P. As is the case in RuBisCO, the enolate intermediate is stabilized by Mg 2ϩ and several amino acid residues: Lys 175 , Asp 203 , Glu 204 , His 294 , and the carbamylated Lys 201 .
The results of these studies suggest that the DK-MTP-1-P enolase is structurally and functionally related to photosynthetic RuBisCO. However, research on the G. kaustophilus RLP revealed that the proton-abstracting, reaction-starting residues differed between the DK-MTP-1-P enolase and RuBisCO (25). It has been reported that when lysine at 201 is substituted with an alanine in the G. kaustophilus RLP, the enzyme is still capable of catalyzing enolization of DK-MTP-1-P (25). This result raises a question about the above hypothesis on the close evolutionary relationship between the RLP and RuBisCO, because a carbamylated lysine residue would be required at this position to form the Mg 2ϩ -chelating triad linkage together with Asp 203 and Glu 204 and to stabilize the reaction intermediate in the RuBP enolization reaction of RuBisCO.
Evolutionary relationships of genes with similar sequences are deduced by comparing gene sequence homology of the genes and amino acid sequence homology of the predicted proteins and by analyzing conservation of functional motifs of the predicted proteins in silico. Comparison of protein structures at the active sites also provides important information. However, it may difficult to predict their mutual evolutionary relationship more precisely when they catalyze different reactions in indi-vidual metabolic pathways. The present research adopted a new method to resolve such an issue.
We studied the structural and functional interrelationships of RLP and RuBisCO after enzymological characterization of B. subtilis RLP as the DK-MTP-1-P enolase enzyme. The results showed that DK-MTP-1-P enolase activity was limited to some RLPs in the cluster, including B. subtilis in form IV RLPs. All of the catalytic residues for the RuBisCO reaction were also indispensable for DK-MTP-1-P enolase activity. The architecture of the B. subtilis RLP substrate-binding residues stereospecifically stabilized the transition state analog in CO 2 fixation of RuBisCO. The fact that the transition state analog of RuBisCO interacts with the active site of Bacillus RLP strongly supports their evolutionary proximity.

EXPERIMENTAL PROCEDURES
Materials-Xylulose 1,5-bisphosphate (XuBP) was synthesized from dihydroxyacetone phosphate and glycol aldehyde phosphate by the method of Yokota (29). The product was separated on a Super-Q Toyopearl column (Tosoh) and eluted with a linear gradient of LiCl. XuBP fractions were pooled. The concentration of XuBP was measured as reported elsewhere (29).
The racemic mixture of 2-carboxy-D-arabinitol 1,5-bisphosphate (CABP) and 2-carboxy-D-ribitol 1,5-bisphosphate (CRBP) was synthesized by the method of Pierce et al. (30). Enantiomers were separated from each other by ionic chromatography on a MonoQ 5/50 column (GE Healthcare) and eluted with a 0 -0.4 M linear gradient of LiCl in 3 mM HCl. Fractions with peaks in UV absorbance at 215 nm were assayed for total phosphate (31). The phosphorous peak fractions were separately pooled, and Li ϩ was replaced with Na ϩ using a Dowex 50 column (Na ϩ form). Concentrations of CABP and CRBP were calculated from their phosphate content. Other chemicals were analytical grade and were obtained from commercial sources.
Preparation of Recombinant RLPs-R. rubrum (National Institute of Technology and Evaluation Biological Resource Center number 3986) was purchased from the National Institute of Technology and Evaluation (Tsukuba, Japan). R. rubrum was cultured in 702 medium (pH 7.0) containing 10 g/liter polypeptone, 2 g/liter yeast extract, and 1 g/liter MgSO 4 ⅐7H 2 O at 30°C. Genomic DNA of R. rubrum was extracted with a DNeasy tissue kit (Qiagen). Genomic DNA from Chlorobium tepidum was kindly provided by Professor Hirozo Oh-oka (Osaka University). The gene for the Microcystis aeruginosa PCC 7806 RLP was obtained from Dr. N. Tandeau de Marsac (Pasteur Institute) (24).
The primers for PCR were as follows: CATCATATGACG-GACAGACTGCG and ACCGGATCCCTTGGCGACCTT-GAC for R. rubrum; GACCGGATCAACATATGAATGCT-GAAGACG and GCAGCGGATCCTTTCAGTCCTGCTTC for C. tepidum; and CAGGGTGCTCATATGACTATAATTG and CAAATTGGCGGGATCCAAAGACTCAC for M. aeruginosa. Restriction sites for NdeI and BamHI are underlined. The gene for the B. subtilis RLP was cloned into pDG148 plasmid (32) using an upstream primer introducing a SalI site (underlined), ACGCGTCGACACGCGATTGCAGTTTGAAGAG, and a downstream primer introducing an SphI site (underlined), ACATGCATGCAGGTTTTCGAGTCGTCATACG.
Genes for RLPs of Rhodopseudomonas palustris CGA009 rlp2 (IV-2 in Fig. 1) and Bordetella bronchiseptica RB50 were purchased from GenScript Corp. These genes were synthesized so that codon usage was optimized for Escherichia coli and were flanked by NdeI and BamHI sites. The DNA sequences for RLPs were cloned into pET15b vectors (Novagen, Tokyo, Japan). The pET15b-RLP genes were transformed into E. coli BL21 (DE3) (Novagen).
Site-directed Mutagenesis-The gene for the B. subtilis RLP was mutagenized in a site-directed manner to give K123N, K123I, K123E, K175I, K175E, K201I, and K201E using a QuikChange XL site-directed mutagenesis kit (Stratagene, Paris, France). The mutagenic primer sets used for PCR mutagenesis are listed in Table S1. The thermocycling program was as follows: denaturation for 1 min at 95°C, followed by 18 cycles of denaturation at 95°C for 50 s, annealing at 60°C for 50 s and elongation at 68°C for 14 min, and termination with an additional elongation for 7 min.
Expression and Purification of Recombinant RLPs and Mutated Proteins-Recombinant RLPs and mutated proteins were expressed and purified as reported previously (33,34).
E. coli BL21 (DE3) was used for expression of recombinant RLPs and mutated B. subtilis RLPs. Histidine-tagged RLPs were purified with His-Bind resin (Novagen) according to the manufacturer's instructions. After purification, the RLP fraction was passed through a PD-10 column (GE Healthcare) equilibrated with 50 mM Tris-HCl buffer (pH 8.2) containing 150 mM NaCl, 2.5 mM CaCl 2 , and 1 mM MgCl 2 , and the protein fractions were collected. The His 6 tag was excised with 2.6 g of thrombin/mg of recombinant protein for 3 h at 22°C. Recombinant RLPs and mutated B. subtilis RLPs were further purified on a Superose 6 10/300 GL column (GE Healthcare) equilibrated with 50 mM Tris-HCl buffer (pH 8.2) containing 150 mM NaCl, 1 mM MgCl 2 , and 10% (v/v) glycerol. The active fractions were pooled and stored at Ϫ80°C until use. Native PAGE and SDS-PAGE were carried out as reported previously (35,36).
Assay of DK-MTP-1-P Enolase-DK-MTP-1-P enolase activity was assayed in the coupling reaction with methylthioribulose-1-phosphate (MTRu-1-P) dehydratase (20,24,33,34). The substrate and the coupling enzyme were prepared as described previously (33). The reaction was initiated by adding 1.3 g of methylthioribulose-1-phosphate dehydratase to a mixture containing 50 mM Tris-HCl buffer (pH 8.2), 1 mM MgCl 2 , 2 mM MTRu-1-P, and 0.9 g of DK-MTP-1-P enolase. The final reaction volume was 100 l and the reaction temperature was 35°C unless otherwise stated. The concentration of the reaction substrate, DK-MTP-1-P, was calculated on the basis of the reaction characteristics between 2,3-butanedione and o-phenylenediamine (37,38). The concentration of the reaction product, 2-hy-droxy-3-keto-5-methylthiopentenyl 1-phosphate was monitored spectrophotometrically at 280 nm (20,24,33,34). Enolase activity was calculated from the absorbance change within the initial seconds of the reaction. Within this time, the consumption of substrate, estimated based on the amount of the product, was less than 10%.
Changing of the CO 2 concentration in the reaction mixture, where specified, was done by removal of dissolved inorganic carbon, as reported by Matsuda and Colman (39). A mixture of DK-MTP-1-P enolase (0.1 mg/ml) and methylthioribulose dehydratase (2 mg/ml) was added to a column (0.7 ϫ 26 cm) of Sephadex G-25 (GE Healthcare) equilibrated with 50 mM Tris-HCl buffer (pH 8.2) containing 1 mM MgCl 2 , which had previously been equilibrated with N 2 (N 2 Ͼ 99.99%) (Iwatani) for 30 min at 100°C. The exact concentration of CO 2 in the reaction mixture after adding CO 2 -free MTRu-1-P was calculated from the concentration of the dissolved inorganic carbon that was measured with a CO 2 analyzer (LI-6252; LI-COR, Lincoln, NE) (40).

Distribution of DK-MTP-1-P Enolase Activity among RLPs-
Although some bacteria, archaea, and eukaryota have been reported to possess RLP genes, the only known function of these RLPs is enolization of DK-MTP-1-P to 2-hydroxy-3-keto-5-methylthiopentenyl 1-phosphate in RLPs from B. subtilis, G. kaustophilus, and M. aeruginosa classified in group ␣1 (20,24,25). However, there is no information regarding the distribution of this function among other RLPs in other groups or regarding functional analysis of the amino acid residues that may be involved in the DK-MTP-1-P enolization reaction.
To this end, we compared DK-MTP-1-P enolization activity (Table 1) and conservation among 19 amino acid residues in form IV RLPs (Fig. 1B). Activity was detected in RLPs from B. subtilis and M. aeruginosa in group ␣1, as reported elsewhere (20,25). However, enolase activity was not detectable in several other form IV RLPs, including those of R. rubrum from group ␣2, B. bronchiseptica from group ␤, and R. palustris IV-2 and C. tepidum from group ␥.
Of the 19 amino acid residues required for the RuBisCO reaction, Lys 175 , Gly 196 , Asp 198 , Lys 201 , Asp 203 , Glu 204 , His 294 , Ser 379 , Gly 381 , Gly 403 , and Gly 404 were conserved both in RuBis-COs and in RLPs with enolase activity. Gly 60 , Lys 123 , and Pro 295 were conserved in RLPs with enolase activity, but these residues are Glu 60 , Asn 123 , and Arg 295 in RuBisCOs, respectively. Except for Arg 295 , which is conserved in R. rubrum IV, these Gly 60 and Asn 123 residues are not conserved in form IV RLPs that show no enolase activity (Fig. 1B and Table 1).  Fig. 1 for grouping. b The activity with wild type normalized to 100%, which corresponded to 101.6 mol min Ϫ1 mg of protein Ϫ1 .

Mutational Analysis of the B. subtilis DK-MTP-1-P Enolase-
The RuBP-carboxylation reaction of RuBisCO is composed of five partial reactions: enolization of RuBP, carboxylation of the C2 carbon, hydration of the reaction intermediate, cleavage of the bond between the C2 and C3 carbons, and protonation of the aci-acid form of 3-phosphoglycerate (PGA) ( Fig. 2A) (26). The enolization of RuBP is initiated with the abstraction of the C3 proton of RuBP by carbamylated Lys 201 together with Lys 175 , and its cis-enediol(ate) is formed between C2 and C3 of RuBP. On the other hand, the C1 proton of DK-MTP-1-P, a structural analogue of RuBP, is abstracted, and the enediol(ate) is formed between C1 and C2 carbons in the DK-MTP-1-P enolase reaction (20). Binding and/or catalysis for RuBP enolization of RuBisCO is associated with four essential residues: Lys 175 , Lys 201 , Asp 203 , and Glu 204 . These are completely conserved among DK-MTP-1-P enolases (Fig. 1B). The importance of these residues has been shown by a structural comparison between the CABP-bound Spinacia oleracea RuBisCO and the G. kaustophilus RLP, which was reconstructed assuming its binding of DK-MTP-1-P in place of the reported compound DK-H-1-P (Fig. 2B) (25,41).
These four essential residues and the Lys 123 that is specific to the DK-MTP-1-P enolase were mutated to other amino acids to evaluate their functional significance for B. subtilis DK-MTP-1-P enolase activity. All mutant proteins were expressed in E. coli as soluble proteins and were purified in the same way as the wild-type recombinant protein (Fig. S1A). Native PAGE (Fig. S1B) and gel chromatography (data not shown) suggested that all mutant proteins retained the dimer structure. This implies that the mutations of these residues near the active sites did not cause any fundamental structural change to the B. subtilis RLP.
Changing Lys 123 to alanine, aspartate, isoleucine, or glutamate resulted in complete loss of enolase activity. The substitutions K175I, K175E, K201I, K201E, and K201A also resulted in complete loss of enolase activity (Table 2). In contrast, 7% of enolase activity was retained when Asp 203 was mutated to glutamate, and 84% enolase activity was retained when Glu 204 was mutated to aspartate. Changing Asp 203 and Glu 204 to asparagine and glutamine, respectively, caused complete loss of activity.
The kinetic properties of D203E and E204D were compared with those of the wild-type enzyme (Fig. 3). k cat and K m for DK-MTP-1-P of the wild-type enzyme were 153.5 s Ϫ1 and 11.7 M, respectively, under optimum conditions. Changing the carboxylmethyl group of Asp 203 to a carboxyethyl group lowered k cat and K m for DK-MTP-1-P to 10.8 s Ϫ1 and 9.1 M, respectively. However, shortening the side chain of Glu 204 to a carboxylmethyl group caused a much smaller decrease in k cat (128.9 s Ϫ1 ) and a slight increase in K m (22.4 M). These results suggest that Asp 203 and Glu 204 are indispensable for enolase activity, but the location of the carboxyl group of Asp 203 has a stronger impact on catalytic efficiency than that of Glu 204 in the B. subtilis RLP.
Effects of pH, CO 2 , and Metal Ions on B. subtilis DK-MTP-1-P Enolase Activity-Activation of RuBisCO involves CO 2 and Mg 2ϩ binding to the unprotonated amino group of Lys 201 (42). The resulting carbamate-Mg 2ϩ complex participates in deprotonation of C3 or enolization of RuBP. We examined the activation and/or deprotonation systems of the B. subtilis DK-MTP-1-P enolase through kinetic studies.
The pH optimum of the DK-MTP-1-P enolase reaction of the B. subtilis RLP was pH 8.2; activity at pH 7-7.2 was one-tenth that at the optimum pH (Fig. 4A). This suggests that a residue with pK a over pH 7.0 and pH 8.2 exerts a significant role in catalysis.
Since it was impossible to remove dissolved inorganic carbon in an alkaline buffer, we measured the exact amount of dissolved inorganic carbon in the reaction mixture by the reported traditional CO 2 removal method (39). The measured dissolved inorganic carbon after the removal was 0.21 mM, corresponding to 1.77 M CO 2 at the pH, ionic strength, and temperature used in these experiments (40). These values were 9.34 mM and 80.64 M when 10 mM bicarbonate was added to the N 2 -purged reaction mixture. The enolase activity in the low CO 2 buffer (1.77 M) decreased to ϳ77% of that of the high CO 2 buffer (80.64 M) (Fig. 4B). The purified RLP enzyme was stripped of metal ions as follows. After treatment with 10 mM EDTA for 30 min at 4°C, the enzyme was applied to a NAP-5 column (GE Healthcare) equilibrated with 50 mM Tris-HCl buffer (pH 8.2). Activity was assayed with no metals or in the presence of 1 mM various divalent metal ions. Enolase activity was undetectable after EDTA treatment, but 1 mM Mg 2ϩ recovered the original activity and saturated activity of the purified enzyme under our assay conditions (Fig. 4C). Mg 2ϩ was inhibitory at 10 mM (data not shown). Very low activities were detected with Co 2ϩ , Ni 2ϩ , and Ca 2ϩ , but Mn 2ϩ , Fe 2ϩ , Cu 2ϩ , and Zn 2ϩ were inactive (data not shown).

Similarity between Active Site Structures of B. subtilis RLP
and RuBisCO-Another way to analyze the functional similarity between two related enzymes is to analyze enzymatic kinetics. Even if the substrates of the two enzymes are different but similar, analyzing the effects of structural analogs of one enzyme on the catalytic activity of the other can yield information about the similarity of their substrate-binding structures. Table 3 shows the effects of competitive inhibitors against RuBisCO on DK-MTP-1-P enolase activity of the B. subtilis RLP. No inhibitory effects were observed with sodium sulfate, inorganic phosphate, gluconate, and 6-phosphogluconate under these experimental conditions. However, the substrate of the RuBisCO reaction, RuBP, and its product, PGA, showed low but significant inhibitory activities. The inhibition by RuBP and PGA was competitive with respect to DK-MTP-1-P (Fig. 5, A  and B). XuBP, an enantiomer of RuBP in relation to C3, is a potent inhibitor of RuBisCO but is carboxylated very slowly    (29). This structural analog also weakly inhibited the RLP reaction. A much stronger inhibitory effect was observed with CABP, the transition state analog of RuBisCO (Fig. 5C). The inhibition constant of CABP, 103 M, was much lower than those of PGA and RuBP (2.29 and 10.2 mM, respectively). It should be noted that there was no significant inhibitory effect with the enantiomer of CABP with respect to C2, CRBP (Table  3). We confirmed that the activity of methylthioribulose-1-phosphate dehydratase used as the coupling enzyme in our RLP assay, was not inhibited by these inhibitors under the same conditions (data not shown).

DISCUSSION
There are some important similarities between RLPs and RuBisCOs, although RLPs do not fix CO 2 with RuBP. To date, the only metabolic reaction documented for RLPs is the DK-MTP-1-P enolization in RLPs from B. subtilis, M. aeruginosa, and G. kaustophilus in group ␣1 (20,24,25). RLPs from R. rubrum, B. bronchiseptica, R. palustris, and C. tepidum had no detectable DK-MTP-1-P enolase activity (Table 1). This is the first report analyzing the distribution of DK-MTP-1-P enolase activity among RLPs. The results clearly suggest that DK-MTP-1-P enolase activity is limited to RLPs in group ␣1 and that RLPs of other groups have evolved widely into enzymes catalyzing different reactions. All DK-MTP-1-P enolases have a conserved lysine at residue 123, but this residue is substituted with asparagine, glutamate, or other amino acids in RLPs with no enolase activity ( Fig. 1B and Table 2). Mutations of Lys 123 to alanine, asparagine, isoleucine, or glutamate in B. subtilis RLP caused loss of activity (Table 2). Thus, Lys 123 is thought to be the general base to abstract the C1 proton of DK-MTP-1-P to initiate its enolization (25). However, the R. rubrum RuBisCO has asparagine at residue 123 but still catalyzes the DK-MTP-1-P enolase reaction, albeit at a low rate (Fig. 1B) (20). This implies that Lys 123 is not the universal general base for the DK-MTP-1-P enolase reaction in RLPs or that another residue is involved in the enolase reaction in R. rubrum RuBisCO.
B. subtilis RLP showed slightly lower activity at 1.8 M CO 2 than at 80.6 M CO 2 (Fig. 4B), and Mg 2ϩ and Mg 2ϩ -binding carboxyl groups of Asp 203 and Glu 204 were essential for DK-MTP-1-P enolase activity (Table 2 and Fig. 4C). The present study revealed that Lys 201 is essential for the DK-MTP-1-P enolase reaction of the B. subtilis RLP (Table 2) and that CO 2 exerts a similar function in catalysis to that found in the RuBisCO reaction (Fig. 4B). Lys 201 was carbamylated in the  a The DK-MTP-1-P concentration was 100 mM, and the inhibitor concentration was 1 mM.