|Home | About | Journals | Submit | Contact Us | Français|
Sequencing of the genome of Clostridium botulinum strain Hall A revealed a gene (CBO0515), whose putative amino acid sequence was suggestive of the rare enzyme N5-(1-carboxyethyl) ornithine synthase. To test this hypothesis, CBO0515 has been cloned, and the encoded polypeptide was purified and characterized. This unusual gene appears to be confined to proteolytic strains assigned to group 1 of C. botulinum.
In the late 1980s, high concentrations of two unknown ninhydrin-reactive compounds were discovered in the amino acid pool of Lactococcus lactis, an organism used extensively for the manufacture of cheese in the dairy industry. The two compounds, subsequently identified as N5-(l-1-carboxyethyl)-l-ornithine [N5-(CE) ornithine] and N6-(l-1-carboxyethyl)-l-lysine [N6-(CE) lysine], were purified and characterized, and their stereochemical structures were established by chemical syntheses and nuclear magnetic resonance spectroscopy (11, 14, 17). These N-carboxyalkyl derivatives are formed enzymatically via a reductive condensation between pyruvic acid and the ω (side chain) amino groups of ornithine and lysine, respectively (Fig. (Fig.11).
In L. lactis, the biosyntheses of N5-(CE) ornithine and N6-(CE) lysine are catalyzed by a unique tetrameric NADPH-dependent enzyme, N5-(carboxyethyl)-ornithine synthase (CEOS; EC 22.214.171.124.) (7, 13, 16). The gene encoding this protein (ceo) has a chromosomal locus and, in the case of L. lactis strain K1, ceo is present on a large transposon (Tn5306) that also encodes the requisite genes for sucrose metabolism and nisin biosynthesis (6, 7, 19). Since its purification in 1989, CEOS has not been reported in other microorganisms and, until recently, no gene(s) with significant similarity to ceo had been found in any of the hundreds of currently sequenced bacterial genomes. It was therefore of considerable interest to find that the recently sequenced genome (12) of Clostridium botulinum strain Hall A encodes a gene, CBO0515 (designated bceo), whose translated polypeptide by comparative sequence alignment using CLUSTAL W2 (20) exhibits 50% identity with the amino acid sequence of CEOS from L. lactis. The L. lactis enzyme (Mr = 35,323; pI = 5.73) is assigned accession no. P15244 (UniProt/Swiss-Prot database). The C. botulinum polypeptide (YP_001253058) (Mr = 35,849; pI = 5.77) is designated A5HZ59 (UniProt/TrEMBL database). It seemed plausible that the clostridial protein could exhibit properties similar to those of the lactococcal enzyme. Testing this hypothesis is the basis for the study described here.
Regulatory restrictions precluded study of C. botulinum species in our laboratory. Accordingly, the gene CBO0515 from C. botulinum strain Hall A was synthesized by BlueHeron Biotechnology and assigned GenBank accession no. EF628474. Restriction sites for endonucleases NcoI and EcoRI were added to the 5′ and 3′ ends, respectively, as follows: 5′ end, CCATGGGTATGAAATTAGGATTTTTG (the ceo sequence is in boldface, and the NcoI restriction site is underlined); and 3′ end, GGAAATGATAGTATAAGAATTC (the ceo sequence is in boldface, and the EcoRI restriction site is underlined). The purified product (~1 kb) was ligated to the similarly digested and purified expression vector pTrcHis2B (Invitrogen). The recombinant plasmid pTrcHis2Bbceo was transformed into E. coli TOP10 competent cells, and colonies were selected on Luria-Bertani agar plates containing 150 μg of ampicillin ml−1. The cloned fragment was sequenced to confirm that no mutations had been introduced during the synthesis and amplification procedures.
Cells were grown in 800 ml of Luria-Bertani broth containing ampicillin (150 μg ml−1) in 2-liter baffled flasks at 37°C on a rotary shaker. At an A600 of ~0.5 U, 0.1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) was added to each culture, and growth was continued for 3 h. Cultures were harvested by centrifugation (13,000 × g for 10 min at 5°C), and the cells were washed by resuspension and centrifugation from Tris-HCl buffer (25 mM, pH 7.5). Washed cells (18 g, wet weight) were resuspended in 45 ml of Tris-HCl buffer (pH 7.5), and the organisms were disrupted (at 0°C) by two 1.5-min periods of sonic oscillation. The cell extract was clarified by ultracentrifugation (180,000 × g for 2 h at 5°C), and the enzyme was purified by TrisAcryl-M DEAE anion-exchange and phosphocellulose P-11 chromatography, as described previously (13). The yield was 48 mg of highly purified bCEOS. Throughout the purification bCEOS activity [in the N5-(CE) ornithine-forming direction] was measured by the rate of oxidation of NADPH at A340 (see Fig. Fig.1).1). The standard reaction mixture of 1 ml contained Tris-HCl buffer (0.1 M, pH 7.5), NADPH (0.1 mM), l-ornithine (20 mM), and sodium pyruvate (2 mM). Assays at ambient temperature were started by the addition of bCEOS preparation, and the decrease in A340 was monitored in a Beckman DU 640 recording spectrophotometer. The initial rates of NADPH utilization were calculated by using the kinetics program of the instrument. A molar extinction coefficient () of 6,220 M−1 cm−1 was assumed for calculation of NADPH consumed [N5-(CE) ornithine formed]. bCEOS activity was expressed as μmol of N5-(CE) ornithine formed min−1 mg of protein−1. Kinetic parameters were determined from Eadie-Hofstee plots generated by the dogStar software (D. G. Gilbert) version 1.0c kinetics program. Protein concentrations of cell extracts and chromatography fractions were measured by a BCA assay procedure (Pierce). The concentration of purified bCEOS was determined spectrophotometrically by using a theoretical molar extinction coefficient of 34,965 M (monomer)−1 cm−1 and an A2800.1% of 0.975.
Analysis of chromatographically purified bCEOS by SDS-PAGE (Fig. (Fig.2A,2A, lane 4), revealed a single polypeptide of the size (~36 kDa) expected for the full-length product encoded by bceo. Thin-layer chromatographic (TLC) analysis of an assay containing purified bCEOS, pyruvate (2 mM), l-ornithine (20 mM). and NADPH (0.1 mM) revealed N5-(CE) ornithine as the sole reaction product (Fig. (Fig.3A)3A) There was no detectable formation of N2-(CE) ornithine (octopinic acid), and the two stereoisomers were readily separable by TLC (Fig. (Fig.3B).3B). The measured activity of the enzyme was 6.6 μmol of N5-(CE) ornithine formed min−1 mg of protein−1. In the presence of 25% (wt/vol) ammonium sulfate, bCEOS was stable for several months at −20°C, and there was no appreciable loss of activity after three periods of freezing and thawing of a sample (on successive days) from this temperature. However, in the presence of 20% glycerol, freezing and thawing from −20°C resulted in the loss of all enzymatic activity. The homogeneity of the preparation was confirmed by the unambiguous determination of the first 29 residues from the N terminus of the protein (G)MKLGFLIPNHPNEKRVALLPEHVKGFNN. Except for the presence of glycine in the first cycle (added during the cloning of bceo), the experimentally determined sequence of amino acids was precisely that predicted from translation of the gene. The bCEOS monomer mass was obtained through deconvolution of the electrospray spectra obtained from a single quad mass spectrometer (Agilent MSD) after desalting and steep gradient separation of the protein by reversed-phase chromatography (Zorbax 300SB-C3). The molecular weight of bCEOS determined by electrospray-MS (35,907) was 57 mass units greater than the theoretical molecular weight (35,849) calculated from the amino acid sequence of bceo. This anomalous finding is explained by the presence of a single Gly residue that precedes the N-terminal start methionine and is thus consistent with the data from microsequence analysis. Although SDS-PAGE and electrospray-mass spectrometry analyses yielded an Mr of ~36 kDa, when passed through a calibrated AcA-44 gel filtration column all enzymatic activity was recovered in a single peak corresponding to a protein of Mr ~125 kDa. The absence of A280 peaks of lower Mr (36 or 72 kDa) suggests that in solution bCEOS (like the lactococcal protein) exists as a homotetramer. However, our data do not preclude the formation of a catalytically active trimer or (less likely), the existence of an unusually asymmetric dimer. Conformational and structural similarities between the lactococcal and clostridial proteins were evident from the results of a Western blot (Fig. (Fig.2B),2B), which revealed a strong immunoreaction between polyclonal rabbit antibody to CEOS from L. lactis, and denatured bCEOS from C. botulinum.
The kinetic parameters of bCEOS determined in Tris-HCl buffer (0.1 M, pH 7.5) were as follows: Km(pyr) = 0.15 ± 0.01 mM; Km(orn) = 29.5 ± 1.7 mM; Km(NADPH) = 2.95 ± 0.31 μM; Vmax = 22.4 ± 0.9 μmol of N5-(CE) ornithine formed min−1 mg of protein−1, and Kcat = 13.4 s−1. Of a variety of α-keto acids tested, only seven served as substrates in the condensation reaction (Table (Table1).1). Other carbonyl-containing compounds, including β-hydroxypyruvate, acetoacetate, α-ketovaleric, and α-ketocaproic acids, were ineffective. Of the ω-amino acids tested, l-ornithine and l-lysine were the most acceptable substrates but, surprisingly (and in contrast with CEOS from L. lactis), bCEOS also accepted the d-isomers of the two basic amino acids as (albeit poor) substrates (Table (Table22).
The reactions catalyzed by bCEOS and l-alanine dehydrogenase (l-AlaDH, EC 126.96.36.199) are mechanistically similar, and both enzymes require a reduced nucleotide (NADH/NADPH) and pyruvate as two of their three substrates. Whereas bCEOS catalyzes the NADPH-dependent reductive condensation of pyruvate to yield N5-(CE) ornithine, l-AlaDH catalyzes the NADH-dependent reductive amination of pyruvate to form alanine. The crystal structure of l-AlaDH from Phormidium lapideum has been determined (3), and the pyruvate binding-site of this enzyme comprises six essential residues: E13, R15, K74, Y93, H95, and E117. Remarkably, a comparative alignment of the N-terminal portions of bCEOS and l-AlaDH (Fig. (Fig.4)4) shows that these same amino acids (or conservative substitutions) are positionally retained in the C. botulinum enzyme: E13, R15, K71, W90, H92, and D114. To assess the role(s) of these residues in bCEOS activity, each of the six amino acids was changed by site-directed mutagenesis of the gene in plasmid pTrcHis2Bbceo to yield: E13A, R15K, K71R, W90S, H92L, and D114A. Confirmation for the relevant base changes was provided by DNA sequence analysis. Plasmids containing the desired mutations were transformed into E. coli TOP10-competent cells, and the organisms were grown in LB broth containing ampicillin. After IPTG induction, extracts were prepared, and the specific activities of the six mutant proteins were measured with l-ornithine, pyruvate, and NADPH as substrates. Each mutation resulted in a severe reduction (>95%) or complete loss of bCEOS activity.
Two sets of primer pairs were designed to target different regions of the 954-bp bceo gene. The primer bCEO-43F (5′-AGAGTTGCTTTATTACCAGAA-3′) paired with bCEO-519R (5′-CACCTTGAGATACGTTACC-3′) generates a 476-bp product, and bCEO-459F (5′-TGCAGTTATGAATTATGGAA-3′) paired with bCEO-920R (5′-CCTCTACTTCTGCAATAACAT-3′) yields a 461-bp product in PCR experiments. The primers served as gene probes of 70 DNA preparations, representing the seven C. botulinum serotypes (A to G). Control primer pairs were included for each template in order to amplify either the botulinum neurotoxin gene (bont) for serotype A, B, and E strains, or the 16S rrn gene (533F, 5′-CCAGCMGCCGCGGTAA-3′; P3MODrc, 5′-GGACTACHAGGGTATCTAAT-3′) in other serotypes. PCR amplification was performed in a final reaction volume of 10 μl containing 1× PCR buffer, MgCl2 (0.2 mM), deoxynucleoside triphosphates (0.25 mM), 0.4 U of AmpliTaq Gold (Applied Biosystems, Inc.) forward and reverse primers (90 nM), and 1 ng of template DNA. The PCR amplicons were visualized by combining 2 μl of the reaction with 2 μl of 2× Blue JuiceGel loading buffer (Invitrogen) in a 2% Lonza SeaKem LE Agarose gel stained with 0.1 μg of ethidium bromide ml−1. Table Table33 shows that PCR amplicons of the expected size were generated when tested with each of the C. botulinum group I proteolytic strains. No amplicons were generated from the other C. botulinum group II to IV strain DNA preparations. The experimental results were confirmed by querying several of the completed C. botulinum genomic sequences available in the GenBank database. BLAST (1) analysis provided a 100% match for the bceo gene within C. botulinum group I strains of ATCC 3502 (Hall strain A, BoNT/A1 subtype), ATCC 19397 (BoNT/A1 subtype), Kyoto (BoNT/A2 subtype); Loch Maree (BoNT/A3 subtype), NCTC 2916 [BoNT/A1(B)], Bf (unknown strain with BoNT/B and BoNT/F), and Langeland (proteolytic BoNT/F). In addition, BLAST analysis showed a significant match with gene CLOSPO_01202 from Clostridium sporogenes ATCC 15579 (EDU38340). Comparison of the C. botulinum bceo gene with that of C. sporogenes showed that they share 95.9% identity with 98.1% consensus at the amino acid level. Queries of two group II BoNT/E-producing strains of Beluga and Alaska E43 were negative for bceo. These results support data (Table (Table3)3) showing that strains positive for bceo are present in group I.
Our studies confirm the hypothesis that gene CBO0515 of C. botulinum strain Hall A encodes N5-(CE) ornithine synthase. Why this rare enzyme should be present in such disparate Gram-positive bacteria as L. lactis and C. botulinum is intriguing but currently unclear. Whereas L. lactis is used on a global scale in the dairy industry, C. botulinum produces a Zn2+-containing protease that is the most potent of all known microbial toxins, BoNT (2, 8). The evolutionary origins of CEOS are not discernible from our investigations with L. lactis and C. botulinum. Although presence of the gene within the group I clostridia may be the result of horizontal gene transfer event(s), it is also possible that bceo was retained from an ancestral gene common to both species. There is no evidence to suggest a correlation between bCEOS activity and pathogenicity, toxin synthesis or survival of C. botulinum. Indeed, as our investigation shows, bceo is not present in all members of the four established groups of C. botulinum nor does ceo occur in all species of L. lactis (16). Interestingly, the experimental and genomic sequence analyses reveal that the CEOS gene is currently found only within highly proteolytic strains of L. lactis and within the proteolytic strains that comprise group I of C. botulinum. A 16S rDNA gene analysis of the group I C. botulinum strains shows that they are very similar or identical to C. sporogenes (5, 9), the other clostridial species found to contain bceo. Group I C. botulinum strains are considered to have a C. sporogenes background that also contain and express BoNT (4). Strain Hall A (ATCC 3502) is included within the group I C. botulinum, as well as strains that express any of the BoNT/A subtypes and some of the BoNT/B and BoNT/F subtypes. In C. botulinum, the discordant phylogeny of the seven serologically distinct toxin genes (BoNT/A-G) in comparison to their bacterial host 16S rDNA gene sequences indicates that the BoNT genes have been horizontally transferred into different bacterial host backgrounds.
In the absence of structural data for bCEOS, the binding domains for the three substrates (nucleotide, amino acid, and keto acid) cannot be ascertained. However, comparative sequence alignment with l-AlaDH (Fig. (Fig.4),4), and results from site-directed mutagenesis provide experimental evidence for a putative pyruvate-binding domain in bCEOS. As for l-AlaDH, the side chain amine of K71 in bCEOS is likely to hydrogen bond with the C=O group of pyruvate, and it is probable that the ω-NH2 groups of both K71 and R15 bond to the carboxyl oxygens of the α-keto acid. Finally, in bCEOS (as in l-AlaDH), hydrogen bonding may also occur between the imino moiety of H92 and the carbonyl oxygen of pyruvate. This extensive hydrogen bond network presumably facilitates the orientation of pyruvic acid within the active site, thereby imposing a selective restraint upon the molecular size of α-keto acids that can serve as substrates in the condensation reaction (Table (Table1).1). The catalytic functions of CEOS, and the biochemical role(s) of the Nω-(carboxyalkyl) products are currently unknown (15, 18). Indeed, it is possible that neither ornithine nor lysine are the physiological in vivo targets of the reductive condensation reaction. In 1993 Krook et al. (10) reported the first enzyme-catalyzed carboxyalkylation of a specific lysine residue in a protein, by their discovery of N6-(CE) lysine in NADP+-linked prostaglandin dehydrogenase/carbonyl reductase. Whether CEOS can catalyze the posttranslational modification of lysyl residues in proteins of L. lactis and C. botulinum, has yet to be determined.
We thank Ricardo Dreyfuss for assistance with photography and computer graphics, Nga Nguyen for microsequence analyses, and Thomas Macdonald for excellent assistance with the PCR experiments. We appreciate the numerous attempts of Patrick Baker, Fiona Rodgers, and Svetlana Sedelnikova (University of Sheffield, Sheffield, United Kingdom) to determine the crystal structure of bCEOS.
This investigation was supported by the Intramural Research Program of the National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD.
Published ahead of print on 20 November 2009.