Journal of Experimental Botany, Vol. 59, No. 7, pp. 15431554, 2008 doi:10.

1093/jxb/ern104 Advance Access publication 9 April, 2008

SPECIAL ISSUE REVIEW PAPER

RuBisCO-like proteins as the enolase enzyme in the methionine salvage pathway: functional and evolutionary relationships between RuBisCO-like proteins and photosynthetic RuBisCO
Hiroki Ashida1,*, Yohtaro Saito1, Toshihiro Nakano1, Nicole Tandeau de Marsac2, Agnieszka Sekowska3, Antoine Danchin4 and Akiho Yokota1 Nara Institute of Science and Technology (NAIST), Graduate School of Biological Sciences, 8916-5 Takayama, Ikoma, Nara, 630-0101 Japan 2 ries, CNRS URA 2172, Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex 15, Unite des Cyanobacte France 3 ne Unite de Ge tique in silico, Institut Pasteur, 25 rue du Docteur Roux, 75724 Paris Cedex 15, France
4 1

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Genetics of Bacterial Genomics, CNRS URA 2171, Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex 15, France

Received 30 January 2008; Revised 26 February 2008; Accepted 28 February 2008

Abstract
Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) is the key enzyme in the xation of CO2 in the Calvin cycle of plants. Many genome projects have revealed that bacteria, including Bacillus subtilis, possess genes for proteins that are similar to the large subunit of RuBisCO. These RuBisCO homologues are called RuBisCO-like proteins (RLPs) because they are not able to catalyse the carboxylase or the oxygenase reactions that are catalysed by photosynthetic RuBisCO. It has been demonstrated that B. subtilis RLP catalyses the 2,3-diketo-5-methylthiopentyl1-phosphate (DK-MTP-1-P) enolase reaction in the methionine salvage pathway. The structure of DK-MTP-1-P is very similar to that of ribulose1,5-bisphosphate (RuBP) and the enolase reaction is a part of the reaction catalysed by photosynthetic RuBisCO. In this review, functional and evolutionary relationships between B. subtilis RLP of the methionine salvage pathway, other RLPs, and photosynthetic RuBisCO are discussed. In addition, the fundamental question, How has RuBisCO evolved? is also considered, and evidence is presented that RuBisCOs evolved from RLPs.

Key words: Bacillus subtilis, CO2 xation, 2,3-diketo-5methylthiopentyl-1-phosphate enolase, methionine salvage pathway, molecular evolution, photosynthesis, RuBisCO, RuBisCO-like protein.

Introduction Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) is the key enzyme in the Calvin cycle. RuBisCO catalyses the carboxylase reaction that xes CO2 on RuBP and produces two molecules of 3-phosphoglycerate (Andrews and Lorimer, 1987; Hartman and Harpel, 1994; Roy and Andrews, 2000). The carboxylase reaction is the initial reaction in the Calvin cycle, and xed CO2 is utilized as the carbon source to synthesize sugars and/or starch for the growth of photosynthetic organisms. Nevertheless, RuBisCO has inefcient enzymatic properties (Andrews and Whitney, 2003). In particular, RuBisCO catalyses an oxygenase reaction that xes O2 into RuBP, and this reaction antagonizes the carboxylase reaction in atmospheres in which O2 is much more abundant than CO2 (Andrews and Lorimer, 1987; Hartman and Harpel, 1994; Roy and Andrews, 2000). The oxygenase reaction is the starting reaction of

* To whom correspondence should be addressed. E-mail: ashida@bs.naist.jp The Author [2008]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

1544 Ashida et al.

photorespiration, which releases CO2 and NH3 and wastes energy. In addition, the turnover of the carboxylase reaction by RuBisCO is a few to several times per second, even when substrates are saturating (Andrews and Whitney, 2003). Therefore, the photosynthetic CO2 assimilation rate in plants can be limited by RuBisCO (Hudson et al., 1992; Von Caemmerer et al., 1997). Plants manage to perform photosynthetic CO2 xation by accumulating large amounts of RuBisCO protein to compensate for the inefcient properties of RuBisCO. Indeed, the RuBisCO protein constitutes ;50% of the soluble proteins in plant leaves and is the most abundant protein on earth (Ellis, 1979). Given these observations, RuBisCO is the critical target for the improvement of photosynthetic efciency and productivity of plants. Manipulation of RuBisCO to increase the carboxylation rate and/or achieve high CO2 and low O2 afnity would promote the efciency of photosynthesis (Andrews and Whitney, 2003). Why has RuBisCO failed to evolve into an efcient enzyme that does not catalyse the oxygenase reaction? Has there been an opportunity to evolve a more efcient enzyme? The answers to these questions should be resolved by a study of the molecular evolution of RuBisCO. RuBisCO is classied into three forms, forms I, II, and III, based on amino acid sequences and protein structures, and all RuBisCOs in these groups catalyse both carboxylase and oxygenase reactions. CO2 and O2 are freely diffusible gasses, and O2 is smaller than CO2, making it difcult to create a physical barrier to dioxygen where CO2 has to be present; but why did RuBisCO not acquire some sort of chemical barrier to prevent the action of O2? This question cannot be answered by only considering the evolutionary relationships among the three forms of RuBisCO. Instead there is a need to understand the emergence of RuBisCO in terms of its molecular evolution. Interestingly, there are proteins that are suitable for the analysis of the molecular evolution of RuBisCO. Bacteria and Archaea possess genes encoding proteins that have considerable sequence identity to the large subunit of photosynthetic RuBisCO. These proteins cannot catalyse either the carboxylase or the oxygenase reactions of RuBisCO, and are named RuBisCO-like proteins (RLPs). RLPs represent a fourth class of the RuBisCO enzyme family (Hanson and Tabita, 2001). Among the Bacteria and Archaea, the function of RLP has been revealed only in Bacillus subtilis. Bacillus subtilis RLP catalyses the 2,3-diketo-5-methylthiopentyl-1-phosphate (DK-MTP-1-P) enolase reaction, which is the fourth step in the methionine salvage pathway (Ashida et al., 2003). Bacillus subtilis RLP catalyses an analogous reaction to photosynthetic RuBisCO since its substrate resembles the structure of RuBP, rationalizing the tight linkage between RLPs and RuBisCO.

Until recently, the molecular evolution of RuBisCO had been discussed using an evolutionary clock involving the emergence of form I and form II RuBisCOs. It was not possible to discuss the molecular evolution of RuBisCO before the enzyme acquired the ability to x CO2 because ancestor proteins of RuBisCO had not yet been found. The nding of RLPs, as a fourth form of the RuBisCO family of enzymes, has enabled discussion of the molecular evolutionary relationship between RLPs and RuBisCO. In Guptas hypothesis for the evolution of Bacteria, low G+C Gram-positive bacteria, including B. subtilis, were predicted to be the most ancient bacteria, which emerged near the root of the phylogenetic tree of life. Considering the molecular evolution of RuBisCO, based on Guptas hypothesis (Gupta, 1998; Gupta et al., 1999), RLPs may still retain features of the ancestor protein of RuBisCO, because low G+C Gram-positive bacteria and Archaea, possessing RLP genes, emerged before cyanobacteria and photosynthetic bacteria, which both utilize RuBisCO in the Calvin cycle. Thus, the discovery of RLPs and determination of RLP function provide a new approach to elucidate the evolution of RuBisCO. However, the methionine salvage pathway, to recycle reduced sulphur in sulphur-deprived environments, may not be a more ancient metabolic pathway, with respect to evolutionary processes. This is because hydrogen sulphide, a reduced form of sulphur, was presumably abundant early on Earth when the most primitive organisms were living with minimal numbers of genes and metabolic pathways (Delano, 2001). It is possible, therefore, that an unknown ancestor protein of RuBisCO already existed before RLPs emerged as the enolase enzyme in the methionine salvage pathway, and that RuBisCO has evolved from this ancestor protein, rather than via RLP in the methionine salvage pathway. Thus, discussion of the evolutionary relationship between RuBisCO and RLPs in the methionine salvage pathway still cannot answer such fundamental questions as: Has RuBisCO evolved via RLPs? and What is the origin of RuBisCO? In this review, functional and evolutionary relationships between RLPs and photosynthetic RuBisCO are discussed, and a hypothesis is also proposed for this fundamental question of RuBisCO evolution. RuBisCO-like proteins are widely distributed in Bacteria and in Archaea Genome projects have revealed that many of them possess genes encoding RLPs that show similarity to the large subunit of RuBisCO. In a phylogenetic tree, produced using deduced amino acid sequences from the large subunits of RuBisCO and RLPs, RuBisCOs are classied into three forms, forms I, II, and III (Fig. 1) (Hanson and Tabita, 2001; Ashida et al., 2003, 2005). Form I consists of eight large and eight small subunits of 5055 kDa and

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

RuBisCO-like proteins 1545

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Fig. 1. Phylogenetic tree constructed from the amino acid sequences of the large subunits of RuBisCO and of RLPs. The multiple sequence alignment and the phylogenetic tree were produced with CLUSTALW and TREE VIEW. Full names of species are as follows: B. subtilis, Bacillus subtilis; B. licheniformis, Bacillus licheniformis; Bacillus sp., Bacillus sp. NRRL B-14911; B. thuringiensis, Bacillus thuringiensis; B. anthracis, Bacillus anthracis; B. weihenstephanesis, Bacillus weihenstephanesis; B. cereus, Bacillus cereus; G. kaustrophilus, Geobacillus kaustrophilus; B. clausii, Bacillus clausii; E. sibiricum, Exiguobacterium sibiricum; Lyngbya sp., Lyngbya sp. PCC8106; M. aeruginosa, Microcystis aeruginosa PCC7806; A. fulgidus, Archaeoglobus fulgidus; H. mobilis, Heliobacillus mobilis; O. tauri, Ostreococcus tauri; O. granulosus, Oceanicola granulosus; N. mobilis, Nitrococcus mobilis; H. halophila, Halorhodospira halophila; R. palustris, Rhodopseudomonas palustris; R. rubrum, Rhodospirillum rubrum; Aurantimonas, Aurantimonas sp. SI85-9A1; F. pelagi, Fulvimarina pelagi; Polaromonas sp., Polaromonas sp. JS666; C. salexigens, Chromohalobacter salexigens; P. putida, Pseudomonas putida; D. acidovorans, Delftia acidovorans; B. bronchiseptica, Bordetella bronchiseptica; M. loti, Mesorhizobium loti; S. meliloti, Sinorhizobium meliloti; A. aurescens, Arthrobacter aurescens; P. luteolum, Pelodictyon luteolum; C. chlorochromatii, Chlorobium chlorochromatii; C. phaeobacteroides, Chlorobium phaeobacteroides; P. aestuarii, Prosthecochloris aestuarii; C. tepidum, Chlorobium tepidum; R. sphaeroides, Rhodobacter sphaeroides; T. denitricans, Thiobacillus denitricans; D. aromatica, Dechloromonas aromatica; N. hamburgensis, Nitrobacter hamburgensis; S. WH8102, Synechococcus sp. WH8102; P. marinus, Prochlorococcus marinus; N. punctifome, Nostoc punctiforme PCC73102; T. elongatus, Thermosynechococcus elongatus; S. PCC6803, Synechocystis sp. PCC6803; S. elongatus, Synechococcus elongatus PCC6301; R. sphaeroides, Rhodobacter sphaeroides; R. gelatinosus, Rubrivivax gelatinosus; B. japonicum, Bradyrhizobium japonicum; M. jannaschii, Methanocaldococcus jannaschii; M. barkeri, Methanosarcina barkeri; M. mazei, Methanosarcina mazei; M. acetivorans, Methanosarcina acetivorans; T. kodakarensis, Thermococcus kodakarensis; H. butylicus, Hyperthermus butylicus; N. pharaonis, Natronomonas pharaonis; P abyssi, Pyrococcus abyssi; P. horikoshii, Pyrococcus horikoshii; P. furiosus, Pyrococcus furiosus; M. marisnigri, Methanoculleus marisnigri.

1218 kDa, respectively, and is distributed widely among photosynthetic organisms such as plants, algae, cyanobacteria, and photosynthetic and chemoautotrophic proteobacteria (Tabita, 1999). Form II is composed only of large subunits and is found mainly in some photosynthetic proteobacteria and chemoautotrophic bacteria (Tabita, 1999). Form III is composed of only large subunits and is found in Archaea (Tabita, 1999). On the other hand, RLPs

form three clades, groups a, b, and c, that are different from the clades of photosynthetic RuBisCO (Fig. 1). Group a includes RLPs from Bacillus species, the cyanobacterium Microcystis aeruginosa, the sulphatereducing Archaeon Archaeoglobus fulgidus, and the photosynthetic bacteria Rhodopseudomonas palustris and Rhodospirillum rubrum. Group b contains RLPs from non-photosynthetic proteobacteria Palaromonas sp.,

1546 Ashida et al.

Bordetella bronchiseptica, and Mesorhizobium loti. RLPs from green non-sulphur bacteria of the genus Chlorobium and the proteobacteria Rhodopseudomonas sp. are classied in group c. Comparing amino acid sequences between B. subtilis RLP and photosynthetic RuBisCO, B. subtilis RLP shows ;23% identity to form I and II RuBisCOs and ;30% identity to form III RuBisCOs. In the same group a, B. subtilis RLP shows 78.5, 62.0, 59.4, 22.7, 35.6, 34.8, 23.2, and 22.7% identity to RLPs of Bacillus licheniformis, Geobacillus kaustophilis, Bacillus anthracis, Bacillus clausii, M. aeruginosa, A. fulgidus, R. rubrum, and R. palustris, respectively. On the other hand, B. subtilis RLP shows 27.5, 26.6, and 23.5% identity to RLPs in group b, from Palaromonas sp., B. bronchiseptica, and M. loti, respectively, and 25.1% and 26.8% identity to RLPs in group c, from R. palustris and Chlorobium tepidum, respectively. Considering this distri-

bution and in the absence of horizontal gene transfer, RLPs should have evolved widely, because the three clades of RLPs show a greater distance from each other compared with that with each form of RuBisCO (Fig. 1). Using X-ray crystallography and point mutational analysis, Hanson and Tabita (2001) proposed 19 amino acid residues of photosynthetic RuBisCO as essential for catalysis (Fig. 2). Alignment of RLPs sequences from RLPs shows conservation of 918 of the 19 residues. Bacillus subtilis RLP has 11 conserved residues (Fig. 2). These observations indicate that RLPs are very unlikely to catalyse the carboxylase and oxygenase reactions that are catalysed by photosynthetic RuBisCO. What is the function of RLPs? Is there a functional and evolutionary relationship between RLPs and photosynthetic RuBisCO? The answers to these questions have interested us with respect to the molecular evolution of RuBisCO.
Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Fig. 2. Partial multiple sequence alignment of the large subunits of RuBisCO and RLPs. Sequences anking the residues involved in the carboxylase reaction catalysed by RuBisCO are aligned. Active site residues conserved in RuBisCO are shown in grey letters on a black background. Residues involved in catalysis and enolization of RuBP are indicated above the alignment by C and E, respectively. Lys201 for carbamylation is indicated by *. 1 and 5 above the alignment indicate residues involved in binding of the phosphate groups on C1 and C5 of RuBP, respectively. White coloured letters indicate residues that are not conserved in RLPs. The alignment is numbered according to the sequence of the large subunit of spinach RuBisCO.

RuBisCO-like proteins 1547

The road to identication of the function of B. subtilis RLP In order to identify the function of RLP, the B. subtilis genome was explored when it was completed. When the analysis began, the gene encoding RLP in B. subtilis was annotated as a gene of unknown function, now named mtnW, coding for a protein similar to the large subunit of RuBisCO. It was initially predicted that B. subtilis RLP could not catalyse the carboxylase and oxygenase reactions of photosynthetic RuBisCOs, because this RLP has only 11 of 19 amino acid residues that are essential for this catalysis (Fig. 2). In order to conrm this prediction, an analysis was carried out to determine whether B. subtilis RLP can catalyse the carboxylase reaction using recombinant protein expressed in Escherichia coli. Indeed, B. subtilis RLP showed no carboxylase activity. This result clearly showed that B. subtilis RLP functions as an enzyme in a different metabolic pathway from that of RuBisCO. The mtnW gene for the RLP of B. subtilis is the rst gene of the mtnWXBD operon (Fig. 3a). The mtnKA operon lies in the vicinity of the mtnWXBD operon

(Fig. 3a). It has been predicted that both operons have a leader mRNA, the S box, known to regulate the expression of the genes involved in methionine metabolism (Murphy et al., 2002; Sekowska and Danchin, 2002) (Fig. 3a). The S box is a riboswitch that regulates gene expression via transcription termination (Winkler et al., 2003; Montange and Batey, 2006), and enhances the expression of target genes in response to methionine starvation by detection of the concentration of Sadenosylmethionine, a metabolite indicator for methionine levels in vivo. In order to analyse the response of the mtnWXBD operon to methionine starvation, lacZ was inserted into the internal region of mtnW. Analysis of the expression of the mtnWXBDlacZ operon showed that expression of lacZ was not induced on growth medium containing methionine but was induced on medium containing NH4Cl, glutamine, and serine (Fig. 4a). In addition, lacZ expression was enhanced by removing methionine from the culture medium after pre-culture on medium including methionine (Fig. 4b). These results clearly suggested that this operon was involved in

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Fig. 3. The methionine salvage pathway and related operons in B. subtilis. The mtnWXBD and mtnKA operons consist of genes involved in the methionine salvage pathway. The expression of genes in both operons is regulated by S boxes. (b) The methionine salvage pathway in B. subtilis.

1548 Ashida et al.

Fig. 4. Induction of the mtnWXBD operon. (a) A Bacillus subtilis mutant with lacZ inserted into the mtnW gene was grown on culture medium containing 40 mg ml1 X-gal and 1 mM NH4Cl, glutamine, serine, or methionine. Blue colonies were produced (black colour in this monochromatic gure) by the expression of a lacZ reporter gene within the mtnWXBD operon, when the mtnWXBD operon was induced. (b) The same mutant was grown in Spizizen minimal medium containing methionine (0.3 mM), split at 0 h, and grown in the presence (open circles) or absence (lled squares) of methionine (0.3 mM). Cells were collected every 1 h, and b-galactosidase activities were measured using the total soluble fraction extracted from cells.

methionine metabolism. Furthermore, in mtnWXBD and mtnKA operons, MtnD showed high similarity to 1,2-dihydroxy-3-keto-5-methylthiopentene dioxygenase, and MtnK has been identied as the methylthioribose (MTR) kinase (Sekowska et al., 2001), suggesting that both operons function in the methionine salvage pathway (Fig. 3b) (Grundy and Henkin, 2002; Murphy et al., 2002; Sekowska and Danchin, 2002). These data indicate that B. subtilis RLP would catalyse a reaction step somewhere in this pathway. Many organisms, including bacteria (Sekowska et al., 2000; Grundy and Henkin, 2002), yeast (Marchitto and Ferro, 1985), plants (Burstenbinder et al., 2007), and mammals (Wray and Abeles, 1995; GarciaCastellano et al., 2002), utilize the methionine salvage pathway to recycle organic sulphur from MTR. In this pathway, organic sulphur is salvaged from MTR, which is produced from methylthioadenosine (MTA), a waste product of polyamine (Sekowska et al., 2000; Grundy and Henkin, 2002), mugineic acid (Ma et al., 1995), and

ethylene synthesis (Burstenbinder et al., 2007). In the methionine salvage pathway that has been proposed in Klebsiella, MTR is phosphorylated to MTR-1-phosphate (MTR-1-P) by the MTR kinase, and then MTR-1-P is isomerized by an isomerase to methylthioribulose-1phosphate (MTRu-1-P) (Saito et al., 2007). MTRu-1-P undergoes a dehydration that is catalysed by a dehydratase and produces DK-MTP-1-P (Furne and Abeles, 1988; Ashida et al., 2003, 2005, 2007; Sekowska et al., 2004). DK-MTP-1-P is converted to 1,2-dihydroxy-3-keto-5methylthiopentene (DHK-MTPene), via the intermediate, 2-hydroxy-3-keto-5-methylthiopentenyl-1-phosphate (HKMTPenyl-1-P), by a bi-functional enzyme, enolase/phosphatase (Balakrishnan et al., 1993; Myers et al., 1993). In the nal step, DHK-MTPene is converted to formate and 2-keto-4-methylthiobutyrate (KMTB) by a dioxygenase (Wray and Abeles, 1993), and KMTB is transaminated to methionine (Berger et al., 2003). In Klebsiella, each step has been predicted based on the analysis of metabolites, but this pathway remains to be fully understood. The structures of MTRu-1-P and of DK-MTP-1-P are similar to the structure of RuBP, and D-glycero-2,3-pentodiulose1,5-bisphosphate, is a by-product of the oxygenase reaction catalysed by RuBisCO (Pearce and Andrews, 2003). It is therefore predicted that B. subtilis RLP is MTRu-1-P dehydratase or DK-MTP-1-P enolase/phosphatase. Growth assays were used to examine whether RLP was involved in the methionine salvage pathway, using an RLP-decient mutant of B. subtilis and pathway metabolites. The wild type grew in the culture medium containing methionine, MTA, or KMTB as a sole source of sulphur (Fig. 5a, b). The RLP-decient mutant grew in culture medium containing methionine or KMTB, but not in medium containing MTA as a sole source of sulphur (Fig. 5a, b) (Ashida et al., 2003). This result clearly showed that B. subtilis RLP catalyses a reaction step between MTA and KMTB in the methionine salvage pathway. To determine this step, enzymes, their substrates, and products in the methionine salvage pathway of B. subtilis were identied, step-by-step, using recombinant proteins encoded in the mtnWXBD and mtnKS operons and employing 1H-NMR and UV-VIS spectroscopy (Ashida et al., 2003). All reaction steps and enzymes were determined in the methionine salvage pathway, consisting of mtnA, mtnW, mtnX, mtnB, and mtnD genes, which encode MTR-1-P isomerase, DK-MTP-1-P enolase, HK-MTPenyl-1-P phosphatase, MTRu-1-P dehydratase, and DHK-MTPene dioxygenase, respectively (Fig. 3b) (Ashida et al., 2003). In this way, B. subtilis RLP was identied as the DK-MTP-1-P enolase. The enolase/ phosphatase activities in other organisms were different from that in the methionine salvage pathway of B. subtilis, in that these steps were catalysed by two enzymes in B. subtilis, RLP for enolization and HK-MTPenyl-1-P phosphatase for deposphorylation (Fig. 3b).

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

RuBisCO-like proteins 1549

micro-organisms utilizing RLPs for methionine salvage and photosynthetic organisms with RuBisCO. However, it is unclear whether this cyanobacterium has retained the gene for RLP from an ancestral organism or acquired it from recent lateral transfer. Further analysis of cyanobacteria possessing both genes for RLP and RuBisCO are needed to settle this question.

What is the comparison of catalysing reactions between B. subtilis RLP and photosynthetic RuBisCO telling us? It seems that there is no linkage between B. subtilis and M. aeruginosa RLPs and photosynthetic RuBisCO, because RLPs function in the methionine salvage pathway, which is far from the Calvin cycle for CO2 xation. However, interesting observations can be made about the catalysing reactions of these RLPs and photosynthetic RuBisCO. DK-MTP-1-P, a substrate of RLPs, has structural similarity to RuBP, the substrate for photosynthetic RuBisCO, except that in DK-MTP-1-P, the OH group and proton on C3 of RuBP are replaced by a carbonyl group (Fig. 6) (Ashida et al., 2005). Additionally, in DK-MTP-1-P, the OH group on C4 and the phosphate group on C5 of RuBP are replaced by a proton and a methylthio group, respectively (Fig. 6). The functional similarities between DK-MTP-1-P enolase and RuBisCO are found not only by comparing substrates, but also by comparing steps in the catalytic cycle. The carboxylase reaction, catalysed by RuBisCO, consists of ve sequential reactions: (i) enolization of RuBP by abstraction of a proton on C3; (ii) carboxylation of C2; (iii) hydration of C3; (iv) cleavage of the C2C3 bond; and (v) protonation of the C2 of the upper aci-acid form of 3-phosphoglycerate (Mauser et al., 2001). The rst catalysing reaction, enolization of RuBP, is the same reaction as enolization of DK-MTP-1-P in RLP, except that the proton on C1 of DK-MTP-1-P is abstracted in the RLP-catalysed reaction. Lys175, Lys201, Asp203, and Glu204 are essential residues for RuBP enolization in RuBisCO (Cleland et al., 1998). In particular, Lys201 is carbamylated by CO2, and this carbamate is stabilized by co-ordination of an Mg2+ ion (Andrews and Lorimer, 1987; Hartman and Harpel, 1994; Roy and Andrews, 2000). The carbamylated Lys201, Asp203, and Glu204 are essential residues to bind the Mg2+ ion, which stabilizes the endiol form of RuBP (Cleland et al., 1998). In the proposed mechanism for the enolization of RuBP, the carbamylated Lys201 is the base that abstracts the proton of C3, while the amino group of Lys175 initially stabilizes the resultant endiol of RuBP and then acts as the acid that protonates the oxygen anion (Cleland et al., 1998) (Fig. 6). Mutation of these residues leads to a lack of RuBP enolization activity in RuBisCO (Estelle et al.,

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Fig. 5. Rescue of the RLP-decient mutant of B. subtilis by the gene for photosynthetic RuBisCO or by cyanobacterium RLP. We used two RLP-decient mutants of B. subtilis, DRLP1 (mtnW::spc amyE::mtnXBD) (Carre-Mlouka et al., 2006) and DRLP2 (mtnW::pMu tinT3) (Ashida et al., 2003). Strain DRLP1 was transformed with the expression vector, pDG148 (Stragier et al., 1998), harbouring the RLP gene from M. aeruginosa PCC7806, yielding DRLP/MaRLP+. The gene for photosynthetic RuBisCO from the photosynthetic bacterium Rhodospirillum rubrum was integrated into the amyE gene in DRLP2, giving strain DRLP/RuBisCO+. Wild type (a), DRLP1 (b), DRLP/MaRLP+ (c), and DRLP/RuBisCO+ (d) were grown for 35 h in Spizizen minimal medium (Anagnostopoulos and Spizizen, 1961) containing 1 mM isopropyl-b-D-thiogalactopyranoside and 0.5 mM MTA as the sole source of sulphur. When required, spectinomycin, erythromycin, kanamycin, and chloramphenicol were added to a nal concentration of 100, 0.5, 5, and 5 mg ml1, respectively.

The function of RLP from the cyanobacterium Microcystis aeruginosa PCC7806 was also identied, based on analysis of B. subtilis RLP. Interestingly, this cyanobacterium possesses genes for both RLP and photosynthetic RuBisCO. The gene for RLP of M. aeruginosa could rescue an RLP-decient mutant of B. subtilis when they were grown in culture medium including MTA as the sole source of sulphur (Fig. 5c) (Carre-Mlouka et al., 2006). The recombinant M. aeruginosa RLP showed DK-MTP1-P enolase activity (Carre-Mlouka et al., 2006), suggest ing that M. aeruginosa RLP is the DK-MTP-1-P enolase in this cyanobacterium. This was the rst report that RLP, functioning in the methionine salvage pathway, co-existed with a photosynthetic RuBisCO in the same organism. In addition, cyanobacteria are considered to be the ancestors of photosynthetic eukaryotic chloroplasts. These facts are interesting, with respect to the evolutionary relationship between RLPs and photosynthetic RuBisCO. Assuming that RLP, in the methionine salvage pathway, is related to the ancestor protein of RuBisCO, based on Guptas bacterial evolution hypothesis, the examination of this cyanobacterium may provide the missing link between

1550 Ashida et al.

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

Fig. 6. Proposed mechanism for the enolization of RuBP by RuBisCO and of DK-MTP-1-P by Bacillus RLP. In enolization of RuBP by RuBisCO, the carbamate group attached to Lys201 is the base that abstracts a proton from C3, while the amino group of Lys175 initially stabilizes the resultant enolate and then acts as the acid that protonates the oxygen anion. In the reaction catalysed by RLP, if RLPs utilize the same residues as RuBisCO, de-protonation at C1 and re-protonation at the oxygen anion are carried out by carbamylated Lys201 and Lys175, respectively. In the reaction mechanism proposed by Imker et al. (2007), Lys123 is the residue for de-protonation at C1.

1985; Hartman et al., 1987; Gutteridge et al., 1988). Interestingly, B. subtilis and M. aeruginosa RLPs conserve these residues that are predicted to be essential residues for the DK-MTP-1-P enolase reaction (Fig. 2). It was easy to predict that Lys201 was also the catalytic residue for abstraction of a proton in B. subtilis RLP. Recently, Imker et al. (2007) reported the X-ray crystal structure of Geobacillus kaustophilus RLP, complexed with an analogue of DK-MTP-1-P, 2,3-diketohexane-1phosphate (DK-H-1-P). Geobacillus kaustophilus belongs to the same clade Bacillaceae as B. subtilis. Geobacillus kaustophilus RLP catalyses the DK-MTP-1-P enolase reaction and has conserved residues Lys175, Lys201, Asp203, and Glu204, which are also conserved in

B. subtilis and M. aeruginosa RLPs (Fig. 2). Lys201 was carbamylated, and carbamylated Lys201, Asp203, and Glu204 were Mg2+ ligands in the 3D structure of G. kaustophilus RLP (Imker et al., 2007). The arrangement of Lys175, carbamylated Lys201, Asp203, and Glu204 in the active site was very similar to that in RuBisCO. Therefore, carbamylated Lys201 was located too far from C1 of DK-H-1-P to abstract a proton, and Lys175 was also located at a distance from C1. In this RLP, Asn123, conserved in RuBisCO, was replaced by lysine (Fig. 2), and Lys123 was appropriately positioned to abstract a proton from C1 of DK-H-1-P. In fact, the enolase activity was retained by alanine substitutions for Lys175 and Lys201, but alanine substitution for Lys123 resulted

RuBisCO-like proteins 1551

in a lack of enolization activity (Imker et al., 2007). Therefore, Imker et al. proposed that a candidate residue to abstract the proton is Lys123, while the carbamylated Lys201 is involved in de-protonation in photosynthetic RuBisCO (Fig. 6). Interestingly, Lys123 is conserved in RLPs of B. subtilis, M. aeruginosa, and G. kaustophilus, which all show DK-MTP-1-P enolase activity (Fig. 2), suggesting that Lys123 is the essential residue for DKMTP-1-P enolase activity in all RLPs. Considering conservation of Lys123, Lys175, and Lys201 in DKMTP-1-P enolase activity, all three residues should contribute to the enolization of DK-MTP-1-P. However, it is unclear whether Lys175 and Lys201 are involved in re- and de-protonation, as they are in photosynthetic RuBisCO (Fig. 6). If Lys123 played a role in the abstraction of protons in DK-MTP-1-P enolase, in spite of the fact that Lys201 was carbamylated, why was Lys201 carbamylated in RLP, as it is in photosynthetic RuBisCO? It was reported by Imkar et al. (2007) that the carbamylation at Lys201 increased the kcat/Km for the enolase reaction of G. kaustophilus RLP. However, the role of carbamylation on Lys201 is still unknown in RLP. Further analysis will reveal the functional relationship of carbamylation on Lys201 between RLPs and photosynthetic RuBisCO. Three RLPs also had conserved residues at Gly401 and Gly403, relative to the P1 phosphate-binding motif of photosynthetic RuBisCO (Fig. 2) (Hartman and Harpel, 1994). In G. kaustophilus RLP, the P1 phosphate oxygens of DK-H-1-P form hydrogen bonds to amide groups of the backbone of Gly401 and Gly403 (Imker et al., 2007). DK-MTP-1-P enolase and photosynthetic RuBisCO utilize the common conserved P1-binding motif to bind the substrate P1 phosphate group.

Are all RLPs acting as the DK-MTP-1-P enolases? Many bacteria possess genes for RLPs, as shown in the phylogenetic tree (Fig. 1). Are all RLPs acting as DKMTP-1-P enolases, as is the case in B. subtilis, M. aeruginosa, and G. kaustophilus? As described above, RLPs from B. subtilis, M. aeruginosa, and G. kaustophilus have conserved residues Lys123, Lys175, Lys201, Asp203, and Glu204 that are essential for the enolization reaction of DK-MTP-1-P enolase (Fig. 2). Other RLPs from group a together with those of Bacillus species and of cyanobacteria also have conserved residues Lys123, Lys175, Lys201, Asp203, and Glu204 (Fig. 2). In addition, the genes for RLPs from B. licheniformis, B. cereus, and B. anthracis, and from other Bacillus species are members of the mtnWXBD operon. The gene for RLP from the cyanobacterium Lyngbya sp. PCC 8106 is also predicted to form this operon with homologous genes for mtnB and mtnE. These observations suggest that

these RLPs in group a also function as the DK-MTP-1-P enolase in the methionine salvage pathway. However, the RLPs of R. rubrum and R. palustris, in the same group a, have conserved residues Lys175, Lys201, and Asp203, but Lys123 and Glu204 are replaced with asparagine and histidine, respectively (Fig. 2). Interestingly, it was reported that R. rubrum and R. palustris utilize the methionine salvage pathway because they can grow using MTA as the sole source of sulphur (Tabita et al., 2007). This result implies that R. rubrum and R. palustris RLPs function as the DK-MTP-1-P enolase in these bacteria. Considering the result of mutational analysis at Lys123 in G. kaustophilus RLP, these RLPs may not be the DK-MTP-1-P enolase, because they do not have conserved Lys123. In R. rubrum and R. palustris, the reaction step of DK-MTP-1-P enolase may be catalysed by a photosynthetic form II RuBisCO and not by RLPs, because R. rubrum RuBisCO can catalyse this reaction, although at a very low rate (described below) (Ashida et al., 2003). As in the case of R. rubrum and R. palustris, group a seem to include RLPs with other functions, distinct from DK-MTP-1-P enolase. Tabita et al. (2007) reported a phylogenetic tree of RLPs produced using a large number of RLP sequences. In Tabitas phylogenetic tree, group a was further classied into two subgroups, RLPs for DK-MTP-1-P enolase and photosynthetic bacterial RLPs. If the phylogenetic tree is examined in detail, RLPs of group a can be divided into two subgroups, a1 including B. subtilis DK-MTP-1-P enolase and a2 including RLPs from photosynthetic bacteria, in agreement with Tabitas classication. This classication of RLPs is in accord with the prediction of function. RLP is the DK-MTP-1-P enolase, based on conservation of predicted essential residues, and, therefore, it is predicted that RLPs in group a1 are DK-MTP-1-P enolases, but group a2 RLPs are not. However, Lys201 is replaced by glutamine in B. clausii RLP of group a1 (Fig. 2) and this micro-organism does not possess homologous genes for other enzymes functioning in the methionine salvage pathway, in spite of B. clausii being a Bacillus species. Likewise, A. fulgidus RLP in group a1 is exceptional in that Lys123 is not conserved (Fig. 2) and, in addition, the homologous genes encoding other enzymes for methionine salvage cannot be found. These facts suggest that these RLPs may not function as DK-MTP-1-P enolase and that some RLPs classied in group a1 possess a different function. On the other hand, in RLPs in groups b and c, Lys123 is replaced by asparagine, and, therefore, these RLP groups should not function as DKMTP-1-P enolase. The function of RLPs from species other than Bacillus species and cyanobacteria is unclear. Hanson and Tabita (2001) reported that C. tepidum RLP, included in RLP group c, is involved in sulphur oxidation and oxidative stress, suggesting that this RLP functions in other biological process.

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

1552 Ashida et al.

Evolutionary and functional linkage between RLP and photosynthetic RuBisCO Evolutionary and functional linkage between RLP and photosynthetic RuBisCO has attracted interest. It is concluded that there is functional linkage of RLP and RuBisCO in the DK-MTP-1-P enolase reaction, because the catalytic reaction and the substrate of DK-MTP-1-P enolase resemble those of RuBisCO. Surprisingly, introduction of the gene for RuBisCO from the photosynthetic bacterium R. rubrum rescued the growth of an RLP-decient B. subtilis mutant on medium containing MTA as the sole sulphur source (Fig. 5d). In addition, recombinant R. rubrum RuBisCO showed a very low, but signicant level of DK-MTP-1-P activity in vitro (Ashida et al., 2003). This suggested that photosynthetic RuBisCO could catalyse the DK-MTP-1-P enolase reaction, and that there is a functional link between RLP and RuBisCO. It is therefore speculated that RLP and photosynthetic RuBisCO evolved from the same ancestor protein. Gupta proposed a hypothesis for linear evolution of bacteria by indel analysis using highly conserved domains of proteins that all bacteria commonly possess (Gupta, 1998; Gupta et al., 1999). In this hypothesis, low G+C Gram-positive bacteria, including Bacillus species, would have features of the most ancient bacteria, rather than Archaea and Bacteria possessing RLPs and RuBisCO. If the molecular evolution of RLP and photosynthetic RuBisCO is discussed following Guptas hypothesis, RLP in the methionine salvage pathway might have evolved earlier than photosynthetic RuBisCO. Furthermore, photosynthetic RuBisCO might have evolved from an ancestral DK-MTP-1-P enolase (Ashida et al., 2005). If this hypothesis is correct, the methionine salvage pathway would be one of the most ancient metabolic pathways. However, ancient organisms might not have utilized this pathway to recycle the reduced sulphur source, because the concentration of hydrogen sulphide would have been high in the environments where the most ancient organisms initially emerged (Delano, 2001). It is therefore thought that the methionine salvage pathway was not part of ancient metabolism and that a catalytically distinct ancestor protein of RLPs, functioning in the methionine salvage pathway, might have already existed. Interestingly, the methionine-salvaging enzymes, MtnK, A, B, and W, from B. subtilis can catalyse sequential reactions for ribose, in addition to the substrate, MTR (Imker et al., 2007). Ribose is converted to 2-hydroxy-3-keto-5-hydroxypent-1ene 1-phosphate by sequential reactions catalysed by four methionine-salvaging enzymes including RLP (Imker et al., 2007). It has been predicted that ribose was formed in the chemical evolution era and was an essential compound in the RNA world (Joyce, 2002). Ribose is one of the most ancient compounds on earth, older than MTR. Considering that methionine-salvaging enzymes can utilize ribose, an ancestral metabolic pathway of the methionine salvage pathway

might have been involved in some form of ribose metabolism, incorporating the sequential reactions catalysed by the methionine-salvaging enzymes. In other words, an ancestral RLP protein might be an enolase enzyme that functioned in ancient ribose metabolism. Concluding remarks Many genome projects have revealed that Bacteria and Archaea possess genes for RLPs that cannot catalyse either the carboxylase or the oxygenase reactions of photosynthetic RuBisCO. These facts provide the fourth form of RuBisCO enzymes. The present studies have revealed that one member of the fourth form, B. subtilis RLP, is the enolase enzyme functioning in the methionine salvage pathway. In addition, RLPs from the cyanobacterium, M. aeruginosa, and G. kaustophilis catalyse the enolase reaction in methionine salvage. Comparing the steps in the catalytic cycle between these RLPs and RuBisCO, it was found that RLPs catalysed a very similar reaction to photosynthetic RuBisCO, for a substrate with a similar structure to RuBP. In addition, both enzymes utilized the same amino acid residues to catalyse each reaction, suggesting that RLP and photosynthetic RuBisCO might have evolved from the same ancestral protein. Furthermore, considering the molecular evolution of RuBisCO based on Guptas evolutionary hypothesis of microorganisms, the hypothesis is proposed that RLP, functioning as an enolase, is the ancestral protein of RuBisCO. Thus, studies on RLPs have suggested an answer to the fundamental question What is the origin of photosynthetic RuBisCO? However, there remains limited information about RLPs and, therefore, further analysis of evolutionary and functional relationships between RLPs and RuBisCO is needed to answer to this question fully. In particular, functional characterization of many RLPs is required to establish the molecular evolution of RuBisCO. Research on RLPs should provide useful information, not only for evolutionary studies, but also to improve RuBisCO efciency. CO2 xation by RuBisCO is the limiting step in plant photosynthesis. Hence, inefcient RuBisCO is the obvious target to improve efciency of photosynthesis in plants. To improve RuBisCO to Super-RuBisCO with high CO2 specicity, depressed reactivity for O2, and a high CO2-xing catalytic rate, it is necessary to know the amino acid residues involved in catalysis of RuBisCO. Analysis of the functional and structural relationships between RLPs and photosynthetic RuBisCO should provide important information on the structure and amino acid residues involved in carboxylase or oxygenase reactions. Acknowledgements
The authors thank Drs Naotake Ogasawara and Kazuo Kobayashi for help and discussion in molecular manipulation of Bacillus. We

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

RuBisCO-like proteins 1553

are grateful to Dr Chojiro Kojima for H-NMR analysis. This work was supported partly by a Grant in Aid (17208031 and 18688021) for Scientic Research from the Japan Society for the Promotion of Science (JSPS), partly by a grant (FY2004-2006) for General Science and Technology from the Asahi Glass Foundation, and partly by a grant (FY2005-2007) from the Nissan Science Foundation.
1

References
Andrews TJ, Lorimer GH. 1987. Rubisco: structure, mechanisms, and prospects for improvement. In: Hatch M, Boardman N, eds. The biochemistry of plants, Vol. 10. New York: Academic Press, 131218. Andrews TJ, Whitney SM. 2003. Manipulating ribulose bisphosphate carboxylase/oxygenase in the chloroplasts of higher plants. Archives of Biochemistry and Biophysics 414, 159169. Ashida H, Danchin A, Yokota A. 2005. Was photosynthetic RuBisCO recruited by acquisitive evolution from RuBisCO-like proteins involved in sulfur metabolism? Research in Microbiology 156, 611618. Ashida H, Saito Y, Kojima C, Kobayashi K, Ogasawara N, Yokota A. 2003. A functional link between RuBisCO-like protein of Bacillus and photosynthetic RuBisCO. Science 302, 286290. Ashida H, Saito Y, Kojima C, Yokota A. 2007. Enzymatic characterization of 5-methylthioribulose-1-phosphate dehydratase of the methionine salvage pathway from Bacillus subtilis. Bioscience, Biotechnology and Biochemistry 71, 20212028. Balakrishnan R, Frohlich M, Rahaim PT, Backman K, Yocum RR. 1993. Appendix. Cloning and sequence of the gene encoding enzyme E-1 from the methionine salvage pathway of Klebsiella oxytoca. Journal of Biological Chemistry 268, 24792 24795. Berger BJ, English S, Chan G, Knodel MH. 2003. Methionine regeneration and aminotransferases in Bacillus subtilis, Bacillus cereus, and Bacillus anthracis. Journal of Bacteriology 185, 24182431. Burstenbinder K, Rzewuski G, Wirtz M, Hell R, Sauter M. 2007. The role of methionine recycling for ethylene synthesis in Arabidopsis. The Plant Journal 49, 238249. Carre-Mlouka A, Mejean A, Quillardet P, et al. 2006. A new Rubisco-like protein coexists with a photosynthetic Rubisco in the planktonic cyanobacteria Microcystis. Journal of Biological Chemistry 281, 2446224471. Cleland WW, Andrews TJ, Gutteridge S, Hartman FC, Lorimer GH. 1998. Mechanism of Rubisco: the carbamate as general base. Chemical Reviews 98, 549562. Delano JW. 2001. Redox history of the Earths interior since approximately 3900 Ma: implications for prebiotic molecules. Origins of Life and Evolution of the Biosphere 31, 311341. Ellis RJ. 1979. Most abundant protein in the world. Trends in Biochemical Sciences 4, 241244. Estelle M, Hanks J, McIntosh L, Somerville C. 1985. Sitespecic mutagenesis of ribulose-1,5-bisphosphate carboxylase/ oxygenase. Evidence that carbamate formation at Lys191 is required for catalytic activity. Journal of Biological Chemistry 260, 95239526. Furne ES, Abeles RH. 1988. Intermediates in the conversion of 5#-S-methylthioadenosine to methionine in Klebsiella pneumoniae. Journal of Biological Chemistry 263, 95989606. Garcia-Castellano JM, Villanueva A, Healey JH, Sowers R, Cordon-Cardo C, Huvos A, Bertino JR, Meyers P, Gorlick R. 2002. Methylthioadenosine phosphorylase gene deletions are common in osteosarcoma. Clinical Cancer Research 8, 782787.

Grundy F, Henkin T. 2002. Synthesis of serine, glycine, cysteine, and methionine. In: Sonenshein A, Hoch J, Losick R, eds. Bacillus subtilis and its closest relatives: from genes to cells. Washington, DC: American Society for Microbiology, 245254. Gupta RS. 1998. Protein phylogenies and signature sequences: a reappraisal of evolutionary relationships among archaebacteria, eubacteria, and eukaryotes. Microbiology and Molecular Biology Reviews 62, 14351491. Gupta RS, Mukhtar T, Singh B. 1999. Evolutionary relationships among photosynthetic prokaryotes (Heliobacterium chlorum, Chloroexus aurantiacus, cyanobacteria, Chlorobium tepidum and proteobacteria): implications regarding the origin of photosynthesis. Molecular Microbiology 32, 893906. Gutteridge S, Lorimer G, Pierce J. 1988. Details of the reactions catalyzed by mutant forms of Rubisco. Plant Physiology and Biochemistry 26, 675682. Hanson TE, Tabita FR. 2001. A ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO)-like protein from Chlorobium tepidum that is involved with sulfur metabolism and the response to oxidative stress. Proceedings of the National Academy of Sciences, USA 98, 43974402. Hartman FC, Harpel MR. 1994. Structure, function, regulation, and assembly of D-ribulose-1,5-bisphosphate carboxylase/oxygenase. Annual Review of Biochemistry 63, 197234. Hartman FC, Soper TS, Niyogi SK, Mural RJ, Foote RS, Mitra S, Lee EH, Machanoff R, Larimer FW. 1987. Function of Lys-166 of Rhodospirillum rubrum ribulosebisphosphate carboxylase/oxygenase as examined by site-directed mutagenesis. Journal of Biological Chemistry 262, 34963501. Hudson GS, Evans JR, von Caemmerer S, Arvidsson YB, Andrews TJ. 1992. Reduction of ribulose-1,5-bisphosphate carboxylase/oxygenase content by antisense RNA reduces photosynthesis in transgenic tobacco plants. Plant Physiology and Biochemistry 98, 294302. Imker HJ, Fedorov AA, Fedorov EV, Almo SC, Gerlt JA. 2007. Mechanistic diversity in the RuBisCO superfamily: the enolase in the methionine salvage pathway in Geobacillus kaustophilus. Biochemistry 46, 40774089. Joyce GF. 2002. The antiquity of RNA-based evolution. Nature 418, 214221. Ma JF, Shinada T, Matsuda C, Nomoto K. 1995. Biosynthesis of phytosiderophores, mugineic acids, associated with methionine cycling. Journal of Biological Chemistry 270, 1654916554. Marchitto KS, Ferro AJ. 1985. The metabolism of 5#-methylthioadenosine and 5-methylthioribose 1-phosphate in Saccharomyces cerevisiae. Journal of General Microbiology 131, 2153 2164. Mauser H, King WA, Gready JE, Andrews TJ. 2001. CO(2) xation by Rubisco: computational dissection of the key steps of carboxylation, hydration, and CC bond cleavage. Journal of the American Chemical Society 123, 1082110829. Montange RK, Batey RT. 2006. Structure of the S-adenosylmethionine riboswitch regulatory mRNA element. Nature 441, 11721175. Murphy BA, Grundy FJ, Henkin TM. 2002. Prediction of gene function in methylthioadenosine recycling from regulatory signals. Journal of Bacteriology 184, 23142318. Myers RW, Wray JW, Fish S, Abeles RH. 1993. Purication and characterization of an enzyme involved in oxidative carbon carbon bond cleavage reactions in the methionine salvage pathway of Klebsiella pneumoniae. Journal of Biological Chemistry 268, 2478524791. Pearce FG, Andrews TJ. 2003. The relationship between side reactions and slow inhibition of ribulose-bisphosphate

Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012

1554 Ashida et al.

carboxylase revealed by a loop 6 mutant of the tobacco enzyme. Journal of Biological Chemistry 278, 3252632536. Roy H, Andrews TJ. 2000. Rubisco: assembly and mechanism. In: Leegood R, Sharkey T, von Caemmerer S, eds. Photosynthesis: physiology and metabolism, Vol. 9. Springer: Dordrecht, The Netherlands, 5383. Saito Y, Ashida H, Kojima C, Tamura H, Matsumura H, Kai Y, Yokota A. 2007. Enzymatic characterization of 5-methylthioribose 1-phosphate isomerase from Bacillus subtilis. Bioscience, Biotechnology and Biochemistry 71, 20212028. Sekowska A, Danchin A. 2002. The methionine salvage pathway in Bacillus subtilis. BMC Microbiology 2, 8. Sekowska A, Denervaud V, Ashida H, Michoud K, Haas D, Yokota A, Danchin A. 2004. Bacterial variations on the methionine salvage pathway. BMC Microbiology 4, 9. Sekowska A, Kung HF, Danchin A. 2000. Sulfur metabolism in Escherichia coli and related bacteria: facts and ction. Journal of Molecular Microbiology and Biotechnology 2, 145177. Sekowska A, Mulard L, Krogh S, Tse JK, Danchin A. 2001. MtnK, methylthioribose kinase, is a starvation-induced protein in Bacillus subtilis. BMC Microbiology 1, 15. Tabita FR. 1999. Microbial ribulose 1,5-bisphosphate carboxylase/ oxygenase: a different perspective. Photosynthesis Research 60, 128. Tabita FR, Hanson TE, Li H, Satagopan S, Singh J, Chan S. 2007. Function, structure, and evolution of the RubisCO-like proteins and their RubisCO homologs. Microbiology and Molecular Biology Reviews 71, 576599. von Caemmerer S, Millgate A, Farquhar GD, Furbank RT. 1997. Reduction of ribulose-1,5-bisphosphate carboxylase/oxygenase by antisense RNA in the C4 plant Flaveria bidentis leads to reduced assimilation rates and increased carbon isotope discrimination. Plant Physiology 113, 469477. Winkler WC, Nahvi A, Sudarsan N, Barrick JE, Breaker RR. 2003. An mRNA structure that controls gene expression by binding S-adenosylmethionine. Nature Structural Biology 10, 701707. Wray JW, Abeles RH. 1993. A bacterial enzyme that catalyzes formation of carbon monoxide. Journal of Biological Chemistry 268, 2146621469. Wray JW, Abeles RH. 1995. The methionine salvage pathway in Klebsiella pneumoniae and rat liver. Identication and characterization of two novel dioxygenases. Journal of Biological Chemistry 270, 31473153.
Downloaded from http://jxb.oxfordjournals.org/ by guest on April 21, 2012