Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, January 2005, p. 230-240, Vol. 49, No. 1
0066-4804/05/$08.00+0 doi:10.1128/AAC.49.1.230-240.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Microbiology, University of Illinois, Urbana, Illinois
Received 4 June 2004/ Returned for modification 16 August 2004/ Accepted 20 September 2004
|
|
|---|
|
|
|---|
Research interest in PTT biosynthesis stems from both its herbicidal activity and the incorporation of the unique phosphinic acid moiety. Previous studies in either S. hygroscopicus or S. viridochromogenes have shown that the biosynthesis of PTT involves more than 13 discrete enzymatically catalyzed reactions, (Fig. 1) linked to a chromosomal gene cluster about 35 kb in length (50). Many of these biosynthetic steps have been accounted for, including those involved with nonribosomal peptide synthesis (Fig. 1, step XIII), the formation of both C-P bonds (steps I and XIV) and those with remarkable homology to portions of the tricarboxylic acid cycle (steps VIII-X). PTT biosynthesis has consequently become one of the only models for reduced-phosphorus antibiotic production. Important biosynthetic questions, however, remain regarding the nature of PTT biosynthesis despite the elucidation of most of the steps. For example, the genes and corresponding enzymes involved in the stepwise oxidation of phosphonoacetaldehyde, an early intermediate, to phosphonoformate (steps III and IV) are currently unaccounted for. Similarly, the mechanism of carboxyphosphonoenolpyruvate synthesis from phosphonoformate also remains largely uncharacterized (step V). Finally, the entire biosynthetic gene cluster has heretofore never been sequenced in its entirety from either producer.
![]() View larger version (34K): [in a new window] |
FIG. 1. Model of phosphinothricin tripeptide biosynthesis, adapted from Thompson and Seto (50). The steps referred to in the text are indicated by roman numerals. The genes sequenced here with equivalents to those with experimentally assigned function from previous work in S. viridochromogenes or S. hygroscopicus are listed below their corresponding steps.
|
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Microorganisms and plasmids used in this study
|
|
View this table: [in a new window] |
TABLE 2. Oligonucleotides used in this study
|
DNA isolation and manipulation. All cloning was performed by established methods (40). Endonucleases, T4 DNA polymerase, and T4 DNA ligase were purchased from Invitrogen (Carlsbad, Calif.) and New England Biolabs (Beverly, Mass.). Shrimp alkaline phosphatase was purchased from Roche Diagnostics GmbH (Mannheim, Germany). Oligonucleotides were obtained from Integrated DNA Technologies (Coralville, Iowa). DNA fragments for cloning were isolated after gel purification with the Agarace enzyme (Promega, Madison, Wis.) according to the manufacturer's recommendations. Plasmids were isolated by the use of Qiagen (Valencia, Calif.) Miniprep or Maxiprep kits. Fosmids were isolated by CsCl gradient ultracentrifugation. For the isolation of high-molecular-weight chromosomal DNA, cultures of S. viridochromogenes were grown under the conditions published previously for protoplasting (48). These were incubated with vigorous agitation for 50 h prior to harvesting cells for lysis. Cells were lysed in TE25S buffer by combined proteinase K and sodium dodecyl sulfate treatment as outlined by Kieser et al. (24). Protein was removed from DNA suspensions by repeated extraction with phenol-chloroform before ethanol precipitation.
PCR amplifications involving Streptomyces DNA were performed with FailSafe PCR PreMix buffers (Epicentre, Madison, Wis.). Amplifications to screen exconjugants were performed via the colony PCR method established by Van Dessel et al. (51). Exconjugant screenings were routinely performed with Taq polymerase. The oligonucleotide PCR primers used in this study are listed in Table 2.
Construction and screening of an S. viridochromogenes genomic library.
S. viridochromogenes DNA was partially digested by Sau3AI, phosphatase treated, and ligated into the BamHI site of the fosmid vector pJVD1 (Table 1). This fosmid vector contains a low-copy-number origin of replication to ensure insert stability and can be modified, as described below, after cloning to allow introduction of new plasmid functions, e.g., the ability to be moved via conjugation and to integrate into the chromosome of various Streptomyces strains via the
C31 site-specific recombination system. Adding these functions to the plasmid vector after library construction maximizes the size of plasmid inserts obtainable with this fosmid vector, which is limited by the amount of DNA that can be contained in phage lambda particles.
Fosmid constructs were packaged with Gigapack III XL (Stratagene, La Jolla, Calif.), and the resulting phage were used to transduce E. coli DH10B to chloramphenicol resistance. Clones containing genes associated with PTT biosynthesis were isolated by PCR screening with the degenerate primer sets PnPy F2 and R1 and Pepmut F1 and R2 to detect the phosphonopyruvate decarboxylase and phosphoenolpyruvate phosphomutase genes, respectively. Positive clones were further characterized by PCR to detect DNA fragments internal to previously published S. viridochromogenes gene sequences associated with the PTT biosynthetic pathway, namely phsA, phsB, pmi, ppm, and pat with primers phsA FOR and phsA REV, phsB 5' end FOR and phsB 5' end REV, pmi FOR and pmi REV, pepmut FOR and pepmut REV, and pat FOR and pat REV, respectively (Table 2). Fosmids containing PTT biosynthetic genes were then sequenced through the cloning junction with the BigDye terminator kit v. 3.0 (ABI Prism, Foster City, Calif.), to screen for constructs where cloning did not disrupt known PTT biosynthetic genes.
DNA sequencing and analysis. To provide priming sites for DNA sequencing reactions, fosmid 5-9G (Table 1) was mutagenized with transposon mini-Mu-JK4740 in in vitro reactions with Mu transposase as recommended (MJ Research, San Francisco, Calif.). Mini-Mu-JK4740 carries two antibiotic resistance markers that are functional in Streptomyces, the kanamycin resistance gene (aph) of Tn5 obtained from Supercos1 (Stratagene) and the thiostrepton resistance gene (tsr) of pIJ702 (23), as well as the conditional oriV replication origin to allow increasing the copy number of plasmids into which the transposon inserts (54). The transposon is carried on pJK95 and is flanked by BglII sites to allow its excision for use in in vitro transposition reactions. The construction of pJK95 is described in Table 1, and the sequence of the plasmid has been deposited in GenBank.
Sequencing reactions were performed at the W. M. Keck Center for Comparative and Functional Genomics at the University of Illinois from primers (Mu-SEQ L1 and Mu-SEQ R1 or Mu-SEQ R5, Table 2) that read out of each side of the transposon. The sequence data from each insertion were compiled with Sequencher 4.0 (Gene Codes Co., Ann Arbor Mich.), and deduced open reading frames were analyzed with the BLAST (2), FASTA (37), and InterProScan (59) programs at NCBI and EMBL.
Construction of the pJVD9/5-9G and pJVD9/5-9G
(orf416-orf571)::kan integrating fosmids.
Fosmid clones derived from pJVD1 were modified to allow their insertion into the chromosome of appropriate Streptomyces species by retrofitting with plasmid pJVD9 (Table 1). Plasmid pJVD9 carries the Streptomyces phage
C31 integration and apramycin resistance determinants, oriT site from the conjugal plasmid RP4, and E. coli HK022 phage attachment site (attP), whereas pJVD1 carries the E. coli phage HK022 bacterial attachment site (attB). Site-specific recombination between HK022 attB and attP results in cointegration of the two plasmids and was performed in vitro as previously described for lambda phage recombination assays (33), except that HK022 integrase was substituted for the lambda equivalent. HK022 integrase was provided by cell extract from WM3321 (Table 1).
Extracts were obtained by inducing mid-log-phase cells of WM3321 with 1 mM isopropylthiogalactopyranoside (IPTG) for 4 h, resuspension of the cells in 50 mM Tris HCl-10% sucrose buffer at pH 7.4, and cell lysis with a French pressure cell. Lysates were cleared by centrifugation at 13,000 x g for 30 min and stored at 70°C. Recombination between pJVD9 and fosmid 5-9G gave rise to pJVD9/5-9G. The deletion of open reading frames 416 to 571' in pJVD9/5-9G was carried out in E. coli with the PCR-mediated gene replacement technique of Datsenko and Wanner as described (6) with primers F5-9G and downstream KO Forward and Reverse (Table 2). Deletion of the desired region from pJVD9/5-9G gave rise to pJVD9/5-9G
(orf416-orf571')::kan, the structure of which was confirmed by PCR with primers RS1Check F and RS1Check R.
Construction of S. lividans heterologous PTT-producing strains.
Plasmids pJVD9/5-9G and pJVD9/5-9G
(orf416-orf571')::kan were transformed into the E. coli conjugal donor strain WM3780. Intergenic conjugation between E. coli donors and S. lividans germinating spores, after heat shock, was performed as previously described by Wohlleben and Pielsticker (57), except that 100 µl of an E. coli mid-logarithmic-phase culture was used as the donor and conjugation was allowed to proceed on nitrocellulose filter disks overnight at 37°C. S. lividans exconjugants were selected on TYE plates supplemented with apramycin and nalidixic acid (50 µg/ml each). Exconjugants were purified at least twice on selective medium with the same antibiotics before growing nonselectively for PTT production bioassays.
Detection of heterologous PTT production by S. lividans. Broth cultures of S. lividans carrying PTT biosynthetic genes were assayed daily by disk diffusion bioassay as previously described (1). For subsequent analysis, cells were removed from bioactive cultures by centrifugation, and methanol was added to the supernatant to 70%. After chilling on ice, particulate matter was removed via centrifugation, and the supernatant was concentrated in vacuo. 31P nuclear magnetic resonance (NMR) analyses of concentrated supernatants were carried out in 20% D2O at the Varian Oxford Center for Excellence in NMR laboratory (University of Illinois at Urbana-Champaign) with a 5-mm Nalorac Quad probe equipped with a Varian Unity U500 spectrometer tuned for phosphorus at 202.28 MHz. The 31P NMR shift values reported have been externally referenced to an 85% phosphoric acid standard (0 ppm). The presence of PTT in bioactive supernatants was confirmed by the addition of genuine PTT and phosphinothricin (Research Products Incorporated, North Prospect, Ill.) to a final calculated concentration of 400 µg/ml each.
Nucleotide sequence accession number. The sequence of the fosmid 5-9G insert containing the PTT biosynthetic gene cluster has been deposited in GenBank under accession number AY632461. The nucleotide sequences of pJK202 and pJK95 were compiled from known sequences and have been deposited in GenBank under accession numbers AY741093 and AY738638, respectively.
|
|
|---|
Heterologous expression of PTT biosynthetic genes in Streptomyces lividans.
To find if fosmid 5-9G contained the intact PTT biosynthetic gene cluster, we inserted the plasmid into the Streptomyces lividans chromosome (after retrofitting with
C31 integration functions and an apramycin resistance determinant as described above) and assayed the recombinant strain for PTT biosynthesis. Bioassays showed that S. lividans with fosmid 5-9G integrated into the chromosome (WM4367) produced a bioactive compound that was not produced by S. lividans containing the vector alone (WM4366) (Fig. 2, plate I). To show that the bioactive compound was PTT and not another S. lividans natural product, the supernatant from this strain was tested against a PTT-resistant Bacillus subtilis 6633 mutant, which was not inhibited by either WM4367 supernatant or authentic PTT (Fig. 2, plate II). Authentic phosphinothricin (no alanyl residues) did not inhibit either strain.
![]() View larger version (85K): [in a new window] |
FIG. 2. Bioassay of WM4367 and WM4368 culture supernatants against PTT-sensitive (plate I) and PTT-resistant (plate II) Bacillus subtilis indicator strains. Disks designated 1 are soaked with WM4366 (pJVD9) supernatant; disks designated 2 are soaked with WM 4367 (pJVD9/5-9G) supernatant; disks designated 3 are soaked with WM 4368/pJVD9/5-9G (orf416-orf571')::kan supernatant; disks designated 4 are soaked with phosphinothricin (10 µg/ml); and disks designated 5 are soaked with PTT (10 µg/ml).
|
DNA sequence analysis of fosmid 5-9G. The 40,241-bp insert of fosmid 5-9G was found to contain 29 complete open reading frames (ORFs) and one partial ORF after double-strand DNA sequencing (Fig. 3). BLAST searches against GenBank revealed that most of these ORFs had been previously sequenced in either S. viridochromogenes or S. hygroscopicus or both. A portion of one ORF had been previously sequenced in S. hygroscopicus. Each gene was assigned a php locus name (for phosphinothricin tripeptide production), except that previously assigned names based upon experimental evidence were preserved. The ORFs found on fosmid 5-9G and their identity scores to Swiss-Prot homologs are presented in Table 3.
![]() View larger version (24K): [in a new window] |
FIG. 3. Open reading frame map of the fosmid 5-9G sequenced insert. ORFs with sequences previously published from S. viridochromogenes analysis alone are shown in light gray. ORFs with previously published sequences from S. hygroscopicus alone are shown in dark gray. ORFs with sequences previously published in both producers are shown in black. ORFs with sequences unique to this study are shown with dark outlines. The nonsequential base pair numbering at the end of phpL and the beginning of phpM indicates that the sequences of these genes overlap by 4 bp.
|
|
View this table: [in a new window] |
TABLE 3. Summary of fosmid 5-9G open reading frames
|
![]() View larger version (18K): [in a new window] |
FIG. 4. 31P NMR spectra of (A) concentrated WM4368 culture supernatant, (B) the same sample spiked with PTT with a concomitant gain in signal intensity at the same frequency, and (C) the sample shown in B spiked with phosphinothricin, showing the acquisition of a new phosphorus signal.
|
|
|
|---|
33.8-kb cluster of genes from S. viridochromogenes that confers the biosynthesis of PTT on a heterologous Streptomyces host. The results presented here indicate that most, if not all, genes required for PTT production are present on fosmid 5-9G. However, it is possible that other genes not present in our clone may have a role in PTT biosynthesis in the native host. If so, such genes would also have to be replaced by functionally equivalent genes in the heterologous host. Likewise, we cannot rule out the possibility that additional genes affect the level of PTT production. Indeed, a host carrying our plasmid with a deletion of the genes downstream of phpR appears to produce less PTT than a host with the full-length insert (Fig. 2). We are uncertain whether this is a function of the particular heterologous strain or whether a deleted ORF could have influenced PTT production in a nonessential manner. Many of the genes and enzymes of the proposed PTT biosynthetic pathway have been previously characterized and are readily identifiable in our sequence (Fig. 1 and 3). Other steps have not yet been solved, although possible candidates are suggested based on previous genetic studies and our analysis of as yet uncharacterized genes in our sequence. Together these data allow a plausible reconstruction of the PTT biosynthetic pathway. The S. viridochromogenes genes and enzymes responsible for the initiation of phosphinothricin biosynthesis and production of phosphonoacetaldehyde (steps I and II), involving genes ppm and ppd, have been described previously (17, 32, 42) and were readily identified in our sequence; however, relatively little is known about the next two steps, the conversion of phosphonoacetaldehyde to hydroxymethylphosphonate (step III) and the subsequent oxidation of hydroxymethylphosphonate to phosphonoformate (step IV). S. hygroscopicus mutants blocked at both steps in the pathway could be complemented in trans by an unsequenced region of DNA closely linked to the ppm and ppd genes from either S. hygroscopicus (21, 31) or S. viridochromogenes (13). The arrangement of the complementing genes within the unsequenced region was refined by restriction to at least two separate ORFs upstream of ppm, with step III complementing DNA localizing directly upstream from the step IV complementing sequence (31).
Three open reading frames that localized to this region of the PTT cluster were identified by our analysis, phpC, phpD, and phpE. The location of phpE corresponds to the step IV complementing region and is a phosphoglycerate dehydrogenase homolog which could conceivably play a role in the oxidation of hydroxymethylphosphonate to phosphonoformate. Either or possibly both phpC and phpD could be assigned a function, based upon location, in the conversion of phosphonoacetaldehyde to hydroxymethylphosphonate (step III). phpC is an alcohol dehydrogenase homolog, whereas phpD is not homologous to any known proteins. Kuzuyama and Seto hypothesized that the unusual biochemistry involved in step III could be achieved by Baeyer-Villiger oxidation of phosphonoacetaldehyde (44), but homology to enzymes known to catalyze such reactions was not discovered here.
CPEP biosynthesis, hypothesized by Hidaka et al. (19) to involve the direct replacement of the phosphate group of PEP by phosphonoformate, was found to involve the bcpE gene product by DNA complementation of a S. hygroscopicus blocked mutant (27). The S. viridochromogenes homolog of this gene was found to be phpH.
Carboxyphosphoenolpyruvate (CPEP) is the substrate for CPEP phosphonomutase in a reaction that yields phosphinopyruvate as a product (combined steps VI and VII). It is unknown whether the decarboxylation of the presumed carboxyphosphinopyruvate intermediate (step VII) of this reaction is the result of enzyme catalysis or inherent product instability despite in vitro study of the enzyme (7); further enzymatic studies may help clarify the mechanism. CPEP phosphonomutase is encoded by bcpA in S. hygroscopicus (18), corresponding to phpI in S. viridochromogenes.
Phosphinopyruvate was found to be converted into phosphinomethylmalate by the addition of an acetate group (step VIII) from arising from acetyl coenzyme A by phosphinomethylmalate synthase (PmmS) in S. hygroscopicus (20, 46), a homolog of the S. viridochromogenes pms gene product (55). Phosphinomethylmalate isomerase, the pmi gene product, was previously shown to rearrange the structure of phosphinomethylmalate (15) (step IX) for subsequent oxidation and decarboxylation (step X) by an unknown enzyme into deamino-
-keto-demethylphosphinothricin (DAKDMPT). Comparison of the products of the genes found in the cluster based on homology to proteins of known function failed to identify a possible candidate for the enzyme responsible for the reaction predicted in step X. It was previously predicted that this reaction would take place by an enzyme similar to (or perhaps identical to) isocitrate dehydrogenase (15). Likewise, an aminotransferase homolog was also not found in the PTT biosynthetic gene cluster, which would likely be required for the conversion of DAKDMPT to demethylphosphinothricin (step XI). Unpublished results cited by Seto and Thompson (50) indicate that both of these steps could be catalyzed by microorganisms that do not produce PTT; thus, it is probable that these steps are catalyzed by ubiquitous, generic enzymes that can be found in most microorganisms.
It has previously been shown that the acetylation of demethylphosphinothricin (step XII) is catalyzed by demethylphosphinothricin N-acetyltransferase, corresponding to the pat (47) or the homologous S. hygroscopicus bar gene product (49). N-Acetyltransferase activity provides the substrate for the alanylation steps (collectively shown in Fig. 1 as step XIII, further discussed below) as well as a mechanism of detoxification against free phosphinothricin (25) (56) that may be produced within the cell. After N-acetylphosphinothricin tripeptide is nonribosomally synthesized, it has been shown that the phpK homolog from S. hygroscopicus, bcpD, encodes the P-methyltransferase that creates the second C-P bond (16, 22), yielding N-acetylphosphinothricin tripeptide (step XIV). The final step in PTT biosynthesis was found to be the deacetylation of N-acetylphosphinothricin tripeptide (step XV) by the dea gene product (39, 55) to produce the intact PTT molecule.
The addition of the two alanine residues to phosphinothricin, producing PTT, has been shown to occur by a nonribosomal peptide synthesis mechanism (9), and a large segment of the minimal gene cluster is dedicated to nonribosomal peptide synthesis activities (Fig. 1, step XIII) (for a review of nonribosomal peptide synthesis, see reference 28). The product of the first ORF found in the insert of fosmid 5-9G, phpA, formerly published in GenBank without analysis as orf1 by Schwartz et al. (accession number Y17268), is highly homologous to the product of an mbtH-like gene. MbtH homologs are typically found in association with nonribosomal peptide synthesis gene clusters, though the functions of these small peptides are not currently known. Thus, the presence of such a gene is expected in the PTT biosynthetic gene cluster, especially given its proximity to phsB and phsC. The proteins that correspond to these two genes, phosphinothricin tripeptide synthetase (PTTS) III and II, respectively, have been biochemically characterized and suggested by Grammel et al. (9) to be the alanine-activating enzymes required for nonribosomal peptide assembly with the phosphinothricin precursor N-acetyldemethylphosphinothricin.
InterProScan analysis of the peptide sequence of the nonribosomal peptide synthesis proteins revealed that PTTS III has phosphopantetheine-binding domains near the amino and carboxy termini that flank a condensation domain and adenylation/activation domain. PTTS II has an identical domain arrangement except that it lacks the amino-terminal phosphopantetheine-binding domain. The enzyme responsible for the activation of N-acetyldemethylphosphinothricin is PTTS I (9), encoded by phsA (41), a gene which is not in close proximity to phsB and phsC in the PTT biosynthetic gene cluster (Fig. 3).
The arrangement of the nonribosomal peptide synthesis genes on fosmid 5-9G confirms the predictions of Hara et al. (12) and Schwartz et al. (41), who noted the likely spatial separation of PTT nonribosomal peptide synthesis genes after complementation analysis of multiple independent nonribosomal peptide synthesis-deficient S. hygroscopicus and S. viridochromogenes mutants, respectively. PTTS I appears to function only as an activating module, because analysis of phsA revealed characteristic adenylation and phosphopantetheine domains (41) but no conserved condensation domain; thus, peptide bond synthesis most likely takes place in the condensation domains of PTTSs II and III. It is interesting that InterProScan domain searches of these proteins did not locate a strongly conserved thioesterase domain in any of the three peptide synthetases. This implies that thioesterase activity, required to release the mature tripeptide peptide from the terminal nonribosomal peptide synthesis module, is likely provided by another enzyme in trans; phpL and phpM are thioesterase gene homologs that could be involved in this step. This observation bolsters the prediction of Raibaud et al. (39) that the equivalents of these genes from S. hygroscopicus, orf1 and orf2, could encode nonribosomal peptide synthesis-related thioesterase activity.
A PTT-specific transport protein is predicted to be associated with the PTT gene cluster because Kyte-Doolittle hydropathy plotting of the protein encoded by dea, responsible for the deacetylation of inactive N-acetylphosphinothricin tripeptide to the active PTT molecule, indicates that this enzyme probably does not span the membrane extensively, if at all, and thus would not have transporter activity. This implies that PTT is released inside the cell in the bioactive form and then exported. Two open reading frames in the PTT biosynthetic gene cluster could have a role in PTT export. phpN has homology to the S. hygroscopicus orf3 gene, a transmembrane transporter homolog which had been previously suggested to be a PTT exporter, or involved in the uptake of biosynthetic substrates (39). phpB has homology to a putative membrane protein, and protein family predicted structure analysis lists it as a member of the OPT oligopeptide transporter protein family.
The production of PTT biosynthetic intermediates in S. hygroscopicus has previously been shown to be under the transcriptional control of a protein encoded by brpA, which has been analyzed by sequencing and mutagenesis (3, 39). A homolog of brpA found in S. viridochromogenes was cloned, sequenced, and found to have 62% identity at the protein level to the S. hygroscopicus gene according to Thompson and Seto (50) in a citation of unpublished data. The same source also indicates that the sequenced gene from S. viridochromogenes could also complement a brpA mutant. From our sequence analysis of fosmid 5-9G, phpR has homology to LuxR-type transcriptional regulators, and its translated product was found to have 63.9% protein identity to S. hygroscopicus brpA; phpR was therefore designated the PTT transcriptional regulatory gene in S. viridochromogenes. brpA was previously found to contain a TTA leucine codon at amino acid position 250, leading to the implication that brpA may be under translational regulation by bldA expression (39). We similarly located a TTA codon within phpR, although the position of that codon, at amino acid position 131, does not correspond with that found in brpA.
Sequence comparison of the S. viridochromogenes PTT biosynthetic cluster to published portions of the corresponding S. hygroscopicus cluster further substantiate previous mapping efforts that predicted the architecture of the two gene clusters were identical (13). The amino acid identity of S. viridochromogenes genes compared to published S. hygroscopicus homologs range from very highly conserved at 95.5% for the bcpA/phpI homolog pair from the core region of the PTT cluster to 63.9% for the phpR/brpA homologs located at the far downstream flank of the gene cluster. The significance of the apparent divergence of the downstream genes cannot be ascertained at this time.
Eleven of 24 ORFs in the gene cluster do not have experimental evidence to support the roles currently assigned. Some, for example the phpL and phpM thioesterase homologs, have proposed functions based on homology. Others cannot currently be assigned biosynthetic roles including phpJ, an aldehyde dehydrogenase homolog and phpG, a bisphosphoglycerate mutase homolog (Table 3). One of the most interesting discoveries resulting from previous studies on PTT biosynthesis was the analogy drawn between the reactions involved in the stepwise conversion of phosphinopyruvate to deamino-alphaketo-demethylphosphinothricin (steps VII through X, Fig. 1) to those corresponding closely to steps of the tricarboxylic acid cycle involved in the stepwise conversion of oxaloacetate to
-ketoglutarate (15). Evolutionary implications regarding the origin of these enzymes were discussed previously by Thompson and Seto (50), who noted that enzymes common to central metabolism may have been modified for specialized secondary metabolic reactions.
We noted two putative gene products in the PTT cluster with homologs involved in central metabolism, namely the glycolytic reactions leading to the production of PEP; PhpG is a homolog of bisphosphoglycerate mutase and PhpH a homolog of enolase. The respective predicted stop (TGA) and start (ATG) nucleotide sequences of the corresponding genes overlap, possibly indicating cotranscription. It is interesting that a mutant of the phpH homolog in S. hygroscopicus, bcpE, was found to be deficient in the production of carboxyphosphonoenolpyruvate (27), a phosphonoenolpyruvate homolog. Biosynthesis of the structural homologs phosphonoenolpyruvate and carboxyphosphonoenolpyruvate by enzymes also showing homology suggests that carboxyphosphonoenolpyruvate biosynthesis may be more complex than the direct phosphate replacement mechanism postulated by Hidaka et al. (19). Further investigation into carboxyphosphonoenolpyruvate biosynthesis is warranted to determine if the phpG product is involved.
We thank M. J. Thomas (University of Wisconsin, Madison) for plasmid pOJ436, E. Yagil (Tel Aviv University, Israel) for strain WM3321, and V. Mainz (University of Illinois, Urbana-Champaign) for invaluable NMR instruction and advice.
|
|
|---|
plasmids: compilation and comparative analysis. J. Mol. Biol. 239:623-663.[CrossRef][Medline]
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»