A Eukaryotic Alanine Racemase Gene Involved in Cyclic Peptide Biosynthesis
2000; Elsevier BV; Volume: 275; Issue: 7 Linguagem: Inglês
10.1074/jbc.275.7.4906
ISSN1083-351X
AutoresYi‐Qiang Cheng, Jonathan D. Walton,
Tópico(s)Plant tissue culture and regeneration
ResumoThe cyclic tetrapeptide HC-toxin is an essential virulence determinant for the plant pathogenic fungusCochliobolus carbonum and an inhibitor of histone deacetylase. The major form of HC-toxin contains thed-isomers of Ala and Pro. The non-ribosomal peptide synthetase that synthesizes HC-toxin has only one epimerizing domain for conversion of l-Pro to d-Pro; the source ofd-Ala has remained unknown. Here we present the cloning and characterization of a new gene involved in HC-toxin biosynthesis,TOXG. TOXG is present only in HC-toxin-producing (Tox2+) isolates of C. carbonum. TOXG is able to support d-Ala-independent growth of a strain ofEscherichia coli defective in d-Ala synthesis. A C. carbonum strain with both of its copies ofTOXG mutated grows normally in culture, and although it no longer makes the three forms of HC-toxin that containd-Ala, it still makes a minor form of HC-toxin that contains Gly in place of d-Ala. The addition ofd-Ala to the culture medium restores production of thed-Ala-containing forms of HC-toxin by the toxGmutant. The toxG mutant has only partially reduced virulence. It is concluded that TOXG encodes an alanine racemase whose function is to synthesize d-Ala for incorporation into HC-toxin. The cyclic tetrapeptide HC-toxin is an essential virulence determinant for the plant pathogenic fungusCochliobolus carbonum and an inhibitor of histone deacetylase. The major form of HC-toxin contains thed-isomers of Ala and Pro. The non-ribosomal peptide synthetase that synthesizes HC-toxin has only one epimerizing domain for conversion of l-Pro to d-Pro; the source ofd-Ala has remained unknown. Here we present the cloning and characterization of a new gene involved in HC-toxin biosynthesis,TOXG. TOXG is present only in HC-toxin-producing (Tox2+) isolates of C. carbonum. TOXG is able to support d-Ala-independent growth of a strain ofEscherichia coli defective in d-Ala synthesis. A C. carbonum strain with both of its copies ofTOXG mutated grows normally in culture, and although it no longer makes the three forms of HC-toxin that containd-Ala, it still makes a minor form of HC-toxin that contains Gly in place of d-Ala. The addition ofd-Ala to the culture medium restores production of thed-Ala-containing forms of HC-toxin by the toxGmutant. The toxG mutant has only partially reduced virulence. It is concluded that TOXG encodes an alanine racemase whose function is to synthesize d-Ala for incorporation into HC-toxin. non-ribosomal peptide synthetase HC-toxin synthetase branched-chain amino acid aminotransferase kilobase Michigan State University base pair(s) Non-ribosomally synthesized peptides are a large class of medically and biologically important secondary metabolites produced by bacteria and fungi. They are biosynthesized by a special class of large, multifunctional enzyme called non-ribosomal peptide synthetases (NRPS).1 NRPSs are organized into domains, each of which is ∼600 amino acids in length and is responsible for the activation by aminoacylation and thioesterification of one constituent amino acid. NRPSs containing 1 to more than 11 domains are known. Some NRPS domains also catalyze amino acid modifications such as α-carbon epimerization/racemization and N-methylation (1.De Crécy-Lagard V. Marlière P. Saurin W. C. R. Acad. Sci. (Paris). 1995; 318: 927-936PubMed Google Scholar, 2.Kleinkauf H. von Döhren H. Eur. J. Biochem. 1996; 236: 335-351Crossref PubMed Scopus (283) Google Scholar, 3.Marahiel M.A. Stachelhaus T. Mootz H.D. Chem. Rev. 1997; 97: 2651-2673Crossref PubMed Scopus (901) Google Scholar). The filamentous fungus Cochliobolus carbonum (anamorphHelminthosporium carbonum or Bipolaris zeicola) synthesizes the cyclic tetrapeptide HC-toxin, the major form (HC-toxin I) of which has the structure cyclo(d-Pro-l-Ala-d-Ala-l-Aeo), where Aeo stands for 2-amino-9,10-epoxi-8-oxo-decanoic acid (4.Walton J.D. Earle E.D. Gibson B.W. Biochem. Biophys. Res. Commun. 1982; 107: 785-794Crossref PubMed Scopus (61) Google Scholar, 5.Walton J.D. Plant Cell. 1996; 8: 1723-1733Crossref PubMed Scopus (259) Google Scholar). Three minor forms (HC-toxins II, III, and IV) differ slightly in amino acid composition (6.Kim S.-D. Knoche H.W. Dunkle L.D. McCrery D.A. Tomer K.B. Tetrahedron Lett. 1985; 26: 969-972Crossref Scopus (45) Google Scholar, 7.Rasmussen J.B. Host-selective Toxins from Helminthosporium carbonum: Purification, Chemistry, Biological Activities, and Effect on Chlorophyll Synthesis in Maize Ph.D. thesis. Michigan State University, 1987Google Scholar, 8.Rasmussen J.B. Scheffer R.P. Plant Physiol. (Bethesda). 1988; 86: 187-191Crossref PubMed Google Scholar, 9.Tanis S.P. Horenstein B.A. Scheffer R.P. Rasmussen J.B. Heterocycles. 1986; 24: 3423-3431Crossref Google Scholar). HC-toxin is an essential determinant of virulence and specificity in the pathogenic interaction betweenC. carbonum and its host, maize (Zea mays L.). Resistance to the fungus and insensitivity to HC-toxin are conditioned by the maize Hm1 gene, which encodes an enzyme, HC-toxin reductase, that specifically detoxifies HC-toxin by reducing the 8-carbonyl group of the Aeo side chain (10.Meeley R.B. Walton J.D. Plant Physiol. (Bethesda). 1991; 97: 1080-1086Crossref PubMed Scopus (47) Google Scholar, 11.Meeley R.B. Johal G.S. Briggs S.P. Walton J.D. Plant Cell. 1992; 4: 71-77Crossref PubMed Google Scholar, 12.Johal G.S. Briggs S.P. Science. 1992; 258: 985-987Crossref PubMed Scopus (361) Google Scholar). Like other Aeo-containing cyclic tetrapeptides such as trapoxin, HC-toxin inhibits histone deacetylases in yeast, maize, insects, protozoans, and mammals (13.Brosch G. Ransom R.F. Lechner T. Walton J.D. Loidl P. Plant Cell. 1995; 7: 1941-1950Crossref PubMed Scopus (178) Google Scholar, 14.Darkin-Rattray S.J. Gurnett A.M. Myers R.W. Dulski P.M. Crumley T.M. Allocco J.J. Cannova C. Meinke P.T. Colletti S.L. Bednarek M.A. Singh S.B. Goetz M.A. Dombrowski A.W. Polishook J.D. Schmatz D.M. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 13143-13147Crossref PubMed Scopus (505) Google Scholar, 15.Kijima M. Yoshida M. Sugita K. Horinouchi S. Beppu T. J. Biol. Chem. 1993; 268: 22429-22435Abstract Full Text PDF PubMed Google Scholar, 16.Ransom R.F. Walton J.D. Plant Physiol. (Bethesda). 1997; 115: 1021-1027Crossref PubMed Scopus (45) Google Scholar). The central enzyme in HC-toxin synthesis is HC-toxin synthetase (HTS), a 570-kDa NRPS with four amino acid-activating domains (17.Panaccione D.G. Scott-Craig J.S. Pocard J.-A. Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6590-6594Crossref PubMed Scopus (124) Google Scholar, 18.Scott-Craig J.S. Panaccione D.G. Pocard J.-A. Walton J.D. J. Biol. Chem. 1992; 267: 26044-26049Abstract Full Text PDF PubMed Google Scholar). Like other NRPSs, HTS activates its substrate amino acids (l-Pro, l-Ala, and d-Ala, but notd-Pro) by aminoacylation and thioesterification (19.Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 8444-8447Crossref PubMed Google Scholar). Partially purified HTS was also reported to racemize l-Pro and l-Ala to the corresponding d-amino acids (20.Walton J.D. Holden F.R. Mol. Plant-Microbe Interact. 1988; 1: 128-134Crossref Google Scholar). Subsequent analysis of the amino acid sequence of HTS deduced from the DNA sequence of its encoding gene, HTS1, revealed the presence of a motif in domain A of HTS that was a candidate to be an epimerization “signature” for NRPS domains. This motif, (S/A)RTXGWFT(T/S), was exclusively present in other known epimerizing NRPS domains, which at that time included only the single-domain enzymes gramicidin synthetase I and tyrocidine synthetase I and domain C of aminoadipyl-cysteinyl-valinyl (ACV) synthetase, a three-domain NRPS (18.Scott-Craig J.S. Panaccione D.G. Pocard J.-A. Walton J.D. J. Biol. Chem. 1992; 267: 26044-26049Abstract Full Text PDF PubMed Google Scholar). However, this epimerase motif is present only once in HTS, in domain A, whereas the ability of partially purified HTS to epimerizel-Ala as well as l-Pro (20.Walton J.D. Holden F.R. Mol. Plant-Microbe Interact. 1988; 1: 128-134Crossref Google Scholar) would appear to require that there be another copy of the motif in domain C. Epimerization domains from many additional NRPSs have now been sequenced, and exhaustive alignment analyses indicate that there are definitely characteristic signature motifs (e.g.[I/V]HHLVVDXVSW) in all such domains. Clearly, HTS has only one epimerase motif, in l-Pro-activating domain A (1.De Crécy-Lagard V. Marlière P. Saurin W. C. R. Acad. Sci. (Paris). 1995; 318: 927-936PubMed Google Scholar, 2.Kleinkauf H. von Döhren H. Eur. J. Biochem. 1996; 236: 335-351Crossref PubMed Scopus (283) Google Scholar, 3.Marahiel M.A. Stachelhaus T. Mootz H.D. Chem. Rev. 1997; 97: 2651-2673Crossref PubMed Scopus (901) Google Scholar). Thus, it appears that the d-Ala of HC-toxin could not be made by HTS itself. The fact that HTS uses d-Ala (but notd-Pro) as a substrate in the ATP/PPi exchange assay suggested that the d-Ala in HC-toxin might be synthesized by a different enzyme. The original report that HTS could epimerize l-Ala could have been due to partial co-purification of this unknown enzyme with HTS (20.Walton J.D. Holden F.R. Mol. Plant-Microbe Interact. 1988; 1: 128-134Crossref Google Scholar). Cyclosporin is an undecapeptide containing d-Ala. Analysis of the primary sequence of its NRPS, cyclosporin synthetase, indicated that it, like HTS, also lacks an epimerization signature motif in itsd-Ala-activating domain (domain A) (1.De Crécy-Lagard V. Marlière P. Saurin W. C. R. Acad. Sci. (Paris). 1995; 318: 927-936PubMed Google Scholar, 2.Kleinkauf H. von Döhren H. Eur. J. Biochem. 1996; 236: 335-351Crossref PubMed Scopus (283) Google Scholar, 3.Marahiel M.A. Stachelhaus T. Mootz H.D. Chem. Rev. 1997; 97: 2651-2673Crossref PubMed Scopus (901) Google Scholar, 21.Weber G. Schörgendorfer K. Schneider-Scherzer E. Leitner E. Curr. Genet. 1994; 26: 120-125Crossref PubMed Scopus (222) Google Scholar). In this case, the producing fungus Tolypocladium niveum(Tolypocladium inflatum) has an alanine racemase that converts l-Ala to d-Ala for incorporation into cyclosporin by cyclosporin synthetase (22.Hoffmann K. Schneider-Scherzer E. Kleinkauf H. Zocher R. J. Biol. Chem. 1994; 269: 12710-12714Abstract Full Text PDF PubMed Google Scholar). Therefore, it is possible that C. carbonum also has an alanine racemase to produced-Ala. HC-toxin production is controlled by a complex Mendelian locus,TOX2, in C. carbonum (23.Ahn J.-H. Walton J.D. Plant Cell. 1996; 8: 887-897PubMed Google Scholar). All of the known genes comprising TOX2 (which, to date, include HTS1,TOXA, TOXC, TOXE, and TOXF) are present only in HC-toxin-producing (Tox2+) isolates, and, with one exception in some isolates, are all physically linked within a ∼600 kb region. However, the TOX2 genes are not tightly clustered; for example, the two copies of HTS1 are separated by ∼300 kb (23.Ahn J.-H. Walton J.D. Plant Cell. 1996; 8: 887-897PubMed Google Scholar, 24.Ahn J.-H. Walton J.D. Mol. Plant-Microbe Interact. 1997; 10: 207-214Crossref PubMed Scopus (47) Google Scholar, 25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar). Using a bacterial artificial chromosome library to find new genes involved in HC-toxin biosynthesis, we recently describedTOXF, which encodes a putative branched-chain amino acid aminotransferase (BCAT) (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). Immediately adjacent to TOXFwas another gene, now called TOXG. This study describes the characterization of TOXG and biochemical and genetic evidence that it encodes an alanine racemase involved in HC-toxin biosynthesis. SB111 (ATCC 90305) and SB114 are standard Tox2+ (toxin-producing) and Tox2− (non-toxin-producing) laboratory strains of C. carbonum, and 164R10 is a Tox2+ progeny of a cross between them (24.Ahn J.-H. Walton J.D. Mol. Plant-Microbe Interact. 1997; 10: 207-214Crossref PubMed Scopus (47) Google Scholar). Fungal growth conditions have been described previously (27.Van Hoof A. Leykam J. Schaefer H.J. Walton J.D. Physiol. Mol. Plant Pathol. 1991; 39: 259-267Crossref Scopus (34) Google Scholar). C. carbonum DNA and RNA isolation was described previously (28.Pitkin J.W. Panaccione D.G. Walton J.D. Microbiology. 1996; 142: 1557-1565Crossref PubMed Scopus (225) Google Scholar). DNA and RNA blotting, probe labeling, hybridization, cDNA and genomic library screening, and DNA subcloning were done following standard procedures (29.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). Oligonucleotides were synthesized by the MSU Macromolecular Facility. Automated DNA sequencing was done at the Yale University Keck Foundation Biotechnology Resource Laboratory and at the MSU DNA Sequencing Facility. Sequences were assembled and analyzed with the DNAStar Software package (DNAStar Inc., Madison, WI). Protein sequence alignments were generated with the ClustalW Program (30.Thompson J.D. Higgins D.G. Gibson T.J. Nucleic Acids Res. 1994; 22: 4673-4680Crossref PubMed Scopus (55405) Google Scholar). The transcriptional start site of TOXG was determined by 5′ random amplification of cDNA ends using a kit from Life Technologies, Inc. (31.Frohman M.A. Dush M.K. Martin G.R. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 8998-9002Crossref PubMed Scopus (4329) Google Scholar). The primer for reverse transcription (GSP1) was 5′-CGATTCATTTTAGGGTGTGCCAGAT-3′, and the nested primer for polymerase chain reaction amplification (GSP2) was 5′-TGTTTCGACTAACCGGTAGCAGGG-3′. Escherichia coli strain TKL10 (dadX alr-ts, CGSC #5466) (32.Wijsman H.J. Genet. Res. 1972; 20: 269-277Crossref PubMed Scopus (33) Google Scholar) was obtained from theE. coli Genetic Stock Center at Yale University and maintained in LB medium supplemented with 200 μg/mld-Ala. The TOXG expression vector pARE was constructed by releasing a 1.2-kb fragment of the coding region ofTOXG (lacking 14 amino acids at the 5′-end of the open reading frame) from pGC1, which contains a full-length cDNA copy ofTOXG, by digestion with BamHI and KpnI and subcloning into pQE31 (Qiagen, Valencia, CA). The sequence of pARE was verified and then transformed into TKL10 by electroporation (2.5 kV, 25 microfarads, 200 Ω, 0.1 cm cuvette) using a Bio-Rad Gene Pulser (Bio-Rad, Richmond, CA). Transformants were selected on LB agar plates containing 100 μg/ml ampicillin at 42 °C. Two independent transformants were designated TKL10T1 and TKL10T2. To assay alanine racemase in transformed E. coli, the cells were grown in LB medium with 100 μg/ml ampicillin at 37 °C for 12 h and collected by centrifugation. The pellet was resuspended in extraction buffer (50 mm Tris, pH 8.7, 10% (v/v) glycerol, 4 mm EDTA, 20 mm dithiothreitol, and 30 μm pyridoxal phosphate) and incubated with lysozyme (1 mg/ml) for 30 min. The cells were then broken by sonication and centrifuged at 20,000 × g for 30 min. Alanine racemase was assayed in the supernatant (22.Hoffmann K. Schneider-Scherzer E. Kleinkauf H. Zocher R. J. Biol. Chem. 1994; 269: 12710-12714Abstract Full Text PDF PubMed Google Scholar). The primary reaction contained 50 mm Tris (pH 8.7), 50 mml-alanine, 20 mm dithiothreitol, 30 μm pyridoxal phosphate, and an appropriate amount of protein extract (typically 10 μg) in a total volume of 1 ml. After incubation for 1–4 h at 42 °C, the reaction was terminated by heating at 95 °C for 10 min. The reaction mixture was centrifuged, and 500 μl of the supernatant was transferred to a new tube. Five hundred μl of secondary reaction mixture (50 mm Tris, pH 8.7, 0.5 unit ofd-amino acid oxidase (Sigma), 25 units of lactate dehydrogenase (Sigma), 2 mm UDP-galactose, and 0.2 mm NADH) was added to the primary reaction mixture and incubated at 37 °C for 2 h. The decrease of absorbance at 340 nm compared with a blank (boiled enzyme) reaction was used to calculate the relative racemase activity. Fungal protoplast preparation, transformation, selection, and single-spore isolation of transformants were performed as described previously (17.Panaccione D.G. Scott-Craig J.S. Pocard J.-A. Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6590-6594Crossref PubMed Scopus (124) Google Scholar, 28.Pitkin J.W. Panaccione D.G. Walton J.D. Microbiology. 1996; 142: 1557-1565Crossref PubMed Scopus (225) Google Scholar). Most strains of C. carbonum contain three copies ofTOXG. Strain 164R10, which is Tox2+, contains two copies and was used for the mutation studies (24.Ahn J.-H. Walton J.D. Mol. Plant-Microbe Interact. 1997; 10: 207-214Crossref PubMed Scopus (47) Google Scholar). The two copies were sequentially mutated, the first by gene replacement mediated by double-crossover homologous recombination, and the second by gene disruption mediated by single-crossover homologous integration. To make the replacement vector pTOXGR1, a genomic clone ofTOXG (from plasmid pAATG1) was trimmed with EcoRV and HpaI to eliminate the TOXF coding region and re-ligated to make the intermediate construct pARM2. A 420-bpApaI/PstI fragment (containing the 3′-end of theTOXG cDNA) from pGC1 was subcloned into theBamHI/PstI sites of pARM2 to make another intermediate construct, pARM3. The 512-bp internalSphI/PstI fragment (corresponding to +362 to +874 bp in the genomic sequence) of pARM3 was replaced by a hygromycin resistance cassette (composed of the E. coli hph gene encoding hygromycin phosphotransferase driven by the Aspergillus nidulans trpC promoter from plasmid pCB1003; Ref. 33.Carroll A.M. Sweigard J.A. Valent B. Fungal Genet. Newslett. 1994; 41: 22Google Scholar) to make the final replacement vector pTOXGR1 (5.9 kb). The fragment (2.5 kb) containing the hph cassette plus flanking TOXGDNA was released from pTOXGR1 by digestion with NotI andHindIII and used for transformation. To make the disruption vector pTOXGD1, an 8-bp BamHI linker was inserted into the unique SacI site of pARM3 to make pARM4. The central part of the coding region from pARM4 (a 512-bpSphI/PstI fragment plus an 8-bp linker) was then subcloned into the HindIII and XbaI sites of the pBC-phleo vector (Cayla, Toulouse, France), which contains the A. nidulans ZEO gene cassette for phleomycin resistance, to obtain the final construct pTOXGD1 (6.6 kb). Vector pTOXGD1 was linearized with BamHI before transformation. Strain 164R10 was first transformed with the 2.5-kb fragment from pTOXGR1. Strain T697 with one copy of TOXG replaced with thehph cassette was selected in medium containing 120 μg/ml hygromycin. Strain T697 was subsequently transformed with linearized pTOXGD1. Strain T698 with the second copy of TOXG disrupted was selected in medium containing 50 μg/ml phleomycin. HC-toxin was extracted from culture filtrates with chloroform and analyzed by TLC on Si250-PA plates (J. T. Baker, Phillipsburg, NJ). The solvent system was acetone:dichloromethane (1:1 by volume). HC-toxin was detected using an epoxide-specific reagent. The plates were sprayed with 2% (w/v)p-nitrobenzylpyridine in acetone, heated for 10 min at 110 °C, and sprayed again with 10% (w/v) tetraethylenepentamine in acetone (10.Meeley R.B. Walton J.D. Plant Physiol. (Bethesda). 1991; 97: 1080-1086Crossref PubMed Scopus (47) Google Scholar, 34.Hammock L.G. Hammock B.D. Casida J.E. Bull. Environ. Contamin. Toxicol. 1974; 12: 759-764Crossref PubMed Scopus (73) Google Scholar). For mass spectrometry, the locations on the TLC plate of compounds of interest were identified by comparison to plates run in parallel that had been sprayed with the epoxide detection reagent, scraped from the TLC plate, suspended in ethanol, and centrifuged briefly. The supernatant was dried under vacuum and redissolved in glycerol-HCl solution for fast atom bombardment mass spectrometric analysis by the MSU-NIH Mass Spectrometry Facility. Fungal pathogenicity was assayed on maize inbred Pr (genotypehm1/hm1) by spraying with conidia (17.Panaccione D.G. Scott-Craig J.S. Pocard J.-A. Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6590-6594Crossref PubMed Scopus (124) Google Scholar, 28.Pitkin J.W. Panaccione D.G. Walton J.D. Microbiology. 1996; 142: 1557-1565Crossref PubMed Scopus (225) Google Scholar). Infected maize leaves were photographed every day from day 2 through day 6 after inoculation. Modified Fries' liquid medium containing 2% (w/v) sucrose (27.Van Hoof A. Leykam J. Schaefer H.J. Walton J.D. Physiol. Mol. Plant Pathol. 1991; 39: 259-267Crossref Scopus (34) Google Scholar) was inoculated with conidia of C. carbonum and grown at room temperature (25 °C) in still culture for 3 days before the addition of d-Ala to a final concentration of 1, 10, 25, or 50 mm. Cultures were allowed to grow for an additional 12 days before collecting the culture filtrates for HC-toxin analysis by TLC. We previously reported the cloning of TOXF on the basis of its physical linkage to knownTOX2 genes (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). Within a 2.9-kbPstI/PstI genomic DNA fragment (cloned in pAATG1) containing TOXF, we detected a partial open reading frame that started immediately in the 5′ direction of TOXF and was transcribed in the opposite orientation (Fig.1). A 1.1-kbPstI/HpaI fragment from pAATG1 was used as a probe to obtain a full-length cDNA clone (pGC1) of the corresponding gene, called TOXG. The insert of pGC1 was used as a probe to obtain the 3′-end of a genomic copy of TOXG(pGG2). The inserts of plasmids pGC1 and pGG2 were sequenced on both strands (Fig. 2). TOXG has one intron of 52 bp. The transcriptional start site is at −46 bp. The distances between the transcriptional and translational start sites ofTOXF and TOXG are 195 and 299 bp, respectively.TOXG encodes a predicted protein (ToxGp) of 389 amino acids with a calculated M r of 42,700 and a pI of 6.5. ToxGp contains no predicted signal peptide or glycosylation sites (Fig.2). The sequence of ToxGp showed strong similarity to threonine aldolases from many species (e.g. Gly1p of Saccharomyces cerevisiae, GenBank P30831) and to a sequence (the product ofcssB) in the patent data base (GenBank A40406) annotated as encoding alanine racemase (EC 5.1.1.1) from the cyclosporin-producing fungus T. niveum (35.Kocher, H. P., Schneider-Scherzer, E., Schörgendorfer, K., and Weber, G. (1994) European Patent WO9425606Google Scholar). The sequence of ToxGp is 42% identical overall to the product of cssB and 32% identical to Gly1p. The identity is spread throughout the proteins, with no bias toward any region. Two lysine residues, which are candidates to be the pyridoxal phosphate binding site and are also present in the same relative locations in the product of cssB and in Gly1p, were identified at positions 220 and 235 in ToxGp (Fig. 2) (36.Liu J.-Q. Dairi T. Kataoka M. Shimizu S. Yamada H. J. Bacteriol. 1997; 179: 3555-3560Crossref PubMed Google Scholar, 37.Monschau N. Stahmann K.P. Sahm H. McNeil J.B. Bognar A.L. FEMS Microbiol. Lett. 1997; 150: 55-60Crossref PubMed Scopus (70) Google Scholar). TOXG was present in three copies in all Tox2+ C. carbonum isolates tested except 164R10, which had two copies, and was completely absent in all Tox2− strains (data not shown). Like TOXF, TOXG mRNA expression was dependent on the regulatory gene TOXE (Fig.3). TOXE encodes a protein containing a bZIP DNA binding domain at its N terminus and four ankyrin repeats at its C terminus (25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar). TOXE is also required for expression of TOXA, TOXC, and TOXD(25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar). Recent unpublished results from this laboratory 2K. F. Pedley and J. D. Walton, unpublished results. indicate that ToxEp is a transcription factor that binds to specific consensus sequences in the promoters of the genes that it regulates. To test the putative function of ToxGp as an alanine racemase, we constructed a TOXGexpression vector, pARE, and transformed it into bacterial strain TKL10, which is defective in d-Ala biosynthesis and therefore requires exogenous d-Ala for survival (32.Wijsman H.J. Genet. Res. 1972; 20: 269-277Crossref PubMed Scopus (33) Google Scholar). Two random transformants, TKL10T1 and TKL10T2, were selected from a LB plate containing 100 μg/ml ampicillin at 42 °C. The transformants survived without d-Ala supplementation (Fig.4). The alanine racemase activity in cell-free extracts of TKL10T1 and TKL10T2 was 2- to 3-fold above background. These experiments indicated that TOXG encodes a functional alanine racemase. To investigate the role of TOXG in HC-toxin biosynthesis, both copies of TOXG in isolate 164R10 were mutated by transformation-mediated gene replacement or disruption. Homologous integration of the plasmid constructs into strains T697 (one copy disrupted) and T698 (both copies disrupted) was confirmed by Southern analysis (Fig. 5). The 5.2-kbXbaI band, corresponding to copy 1 of TOXG, was gone in both T697 and T698, as predicted for gene replacement (Fig.5 A). Subsequent transformation and homologous integration of pTOXGD1 into the 4.8-kb XbaI band (corresponding to copy 2 of TOXG) in strain T697 resulted in the disappearance of this band and the appearance of a new band of 18.0 kb, as seen in strain T698 (Fig. 5 A). This was predicted for homologous tandem integration of two copies of pTOXGD1 (17.Panaccione D.G. Scott-Craig J.S. Pocard J.-A. Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6590-6594Crossref PubMed Scopus (124) Google Scholar). When the same blot was re-probed with the full-length cDNA clone pGC1, a 6.1-kb band hybridized in both T697 and T698, as predicted for the gene replacement event (Fig. 5 B). The putative toxG mutant was tested by RNA blotting. BecauseTOXF and TOXG are tightly clustered,TOXF was included in the analysis to be sure that mutation of one gene did not affect the expression of the other. Total RNA from 164R10 (wild type), the toxF null mutant D/R (renamed T696; Ref. 26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar), and the toxG null mutant T698 was analyzed on three identical blots by probing with TOXF, TOXG, or GPD1 (encoding glyceraldehyde-3-phosphate dehydrogenase as a RNA loading control). The results indicated that thetoxF and toxG mutants had no detectableTOXF or TOXG mRNA, respectively, and that disruption of one gene did not influence the expression of the other (Fig. 6). Both the single (T697) and null (T698) toxG mutants showed normal growth and development on agar plates and in liquid culture. Along with its complete absence from Tox2− isolates ofC. carbonum, this indicated that TOXG has no essential housekeeping function. HC-toxin was extracted from culture filtrates of the wild type and T698 mutant grown with or without d-Ala supplementation (Fig.7). Some wild type isolates of Tox2+ C. carbonum produce four forms of HC-toxin that can be detected on TLC plates using an epoxide-specific reagent (8.Rasmussen J.B. Scheffer R.P. Plant Physiol. (Bethesda). 1988; 86: 187-191Crossref PubMed Google Scholar). The three minor forms (HC-toxin II, III, and IV) contain Gly in place of d-Ala, d-trans-3-hydroxyPro in place of d-Pro, and 8-hydroxy-Aeo in place of Aeo, respectively. HC-toxin forms I, II, and III are active at 0.2, 0.4, and 2.0 μg/ml, respectively, whereas form IV is active only at ∼20 μg/ml (6.Kim S.-D. Knoche H.W. Dunkle L.D. McCrery D.A. Tomer K.B. Tetrahedron Lett. 1985; 26: 969-972Crossref Scopus (45) Google Scholar, 7.Rasmussen J.B. Host-selective Toxins from Helminthosporium carbonum: Purification, Chemistry, Biological Activities, and Effect on Chlorophyll Synthesis in Maize Ph.D. thesis. Michigan State University, 1987Google Scholar, 8.Rasmussen J.B. Scheffer R.P. Plant Physiol. (Bethesda). 1988; 86: 187-191Crossref PubMed Google Scholar, 9.Tanis S.P. Horenstein B.A. Scheffer R.P. Rasmussen J.B. Heterocycles. 1986; 24: 3423-3431Crossref Google Scholar). The C. carbonum wild type isolate 164R10 produced HC-toxin forms I, II, and IV but undetectable levels of the most polar minor form, HC-toxin III. The apparent absence of HC-toxin III probably reflects natural variation among isolates of C. carbonum. The toxG null strain T698 did not produce forms I or IV but still produced form II (Fig. 7). Form II differs from the others in having Gly in place of d-Ala (6.Kim S.-D. Knoche H.W. Dunkle L.D. McCrery D.A. Tomer K.B. Tetrahedron Lett. 1985; 26: 969-972Crossref Scopus (45) Google Scholar). The identities of HC-toxin II and IV were confirmed by mass spectrometry of samples scraped from the TLC plates. HC-toxin II had an experimental MH+ mass of 423.35 (theoretical for C20H30N4O6:422.2165), and HC-toxin IV had an experimental MH+ mass of 439.39 (theoretical for C21H34N4O6:438.2478). When T698 was grown in medium supplemented with d-Ala, production of HC-toxins I and IV was restored (Fig. 7). Higherd-Ala concentrations resulted in higher levels of thed-Ala-containing forms of HC-toxin (Fig. 7). These results indicate that TOXG has a specific role in the production of forms of HC-toxin that contain d-Ala. To test the affect of thetoxG mutation on virulence, conidia of C. carbonum wild type 164R10, single toxG mutant T697, and null toxG mutant T698 were spray-inoculated onto leaves of susceptible maize. On plants inoculated with the wild type, lesions became visible in 2 days, and after 6 days, the entire infected leaf was totally brown and dessicated (Fig.8). Mutation of one copy ofTOXG had no apparent affect on virulence (Fig. 8). However, mutation of both copies of TOXG slowed disease development. A large percentage of the infections initiated by T698 failed to develop into full lesions, and the lesions that did develop were consistently smaller than those caused by the wild type up to the first 6 days of disease development (Fig. 8). At day 8, the toxGmutant caused dessication and death of the leaf (data not shown). The modest decrease in virulence of the toxG mutant is in contrast to the results seen with mutants of other TOX2genes, such as HTS1, TOXC, TOXE, andTOXF, which are completely avirulent (17.Panaccione D.G. Scott-Craig J.S. Pocard J.-A. Walton J.D. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6590-6594Crossref PubMed Scopus (124) Google Scholar, 24.Ahn J.-H. Walton J.D. Mol. Plant-Microbe Interact. 1997; 10: 207-214Crossref PubMed Scopus (47) Google Scholar, 25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar, 26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). The most likely explanation for this is that HC-toxin II, which is still made by the toxG mutant, is sufficient for almost full virulence. Many non-ribosomally synthesized peptides containd-amino acids. Here we present evidence thatd-Ala in the cyclic peptide HC-toxin is synthesized from the primary metabolite l-Ala by an alanine racemase encoded by TOXG. Like the other genes of the TOX2 locus,TOXG is specifically dedicated to HC-toxin biosynthesis because mutants of TOXG have no phenotype other than a failure to produce forms of HC-toxin that contain d-Ala. Furthermore, TOXG is clustered with another gene dedicated to HC-toxin production, TOXF, which itself was discovered by being within ∼70 kb of other genes previously shown to be required for HC-toxin synthesis (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). Mapping studies have shown that in isolate SB111, all copies of all of the TOX2 genes, including all three copies of TOXF/TOXG, are distributed over ∼600 kb on the same chromosome, except for one copy ofTOXE (23.Ahn J.-H. Walton J.D. Plant Cell. 1996; 8: 887-897PubMed Google Scholar, 25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar).2 Hoffmann et al. (22.Hoffmann K. Schneider-Scherzer E. Kleinkauf H. Zocher R. J. Biol. Chem. 1994; 269: 12710-12714Abstract Full Text PDF PubMed Google Scholar) reported the purification and characterization of an alanine racemase involved in cyclosporin biosynthesis by the fungus T. niveum. The encoding gene (cssB) was isolated on the basis of peptide sequences obtained from the purified protein (35.Kocher, H. P., Schneider-Scherzer, E., Schörgendorfer, K., and Weber, G. (1994) European Patent WO9425606Google Scholar). ToxGp and the product ofcssB share significant similarity (42% overall amino acid identity). The involvement of specific alanine racemases in eukaryotic non-ribosomal peptide synthesis is therefore not an isolated occurrence. TOXG and cssB are unrelated at the primary amino acid level to any known bacterial alanine racemase (36.Liu J.-Q. Dairi T. Kataoka M. Shimizu S. Yamada H. J. Bacteriol. 1997; 179: 3555-3560Crossref PubMed Google Scholar, 37.Monschau N. Stahmann K.P. Sahm H. McNeil J.B. Bognar A.L. FEMS Microbiol. Lett. 1997; 150: 55-60Crossref PubMed Scopus (70) Google Scholar, 38.Jansonius J.N. Curr. Opin. Struct. Biol. 1998; 8: 759-769Crossref PubMed Scopus (359) Google Scholar, 39.Alexander F.W. Sandmeier E. Mehta P.K. Christen P. Eur. J. Biochem. 1994; 219: 953-960Crossref PubMed Scopus (344) Google Scholar). Nonetheless, TOXG could complement ad-Ala-deficient mutant of E. coli. In addition to amino acid polymerization and peptide cyclization, some multifunctional NRPSs catalyze α-carbon epimerization. For example, gramicidin synthetase I and tyrocidine synthetase I, both of which are single-domain NRPSs, epimerize l-Phe to d-Phe by a mechanism involving pantothenic acid as co-factor and not, like known amino acid racemases, pyridoxal phosphate (40.Stein T. Kluge B. Vater J. Franke P. Otto A. Wittmann-Liebold B. Biochemistry. 1995; 34: 4633-4642Crossref PubMed Scopus (55) Google Scholar). However, at least one other NRPS domain, domain A of cyclosporin synthetase, does not catalyze epimerization and relies instead on a separate enzyme to supply the required d-amino acid. HTS is the only known example of a NRPS that uses both types of epimerization mechanisms. Thed-Pro of HC-toxin is produced directly by the first domain of HTS (18.Scott-Craig J.S. Panaccione D.G. Pocard J.-A. Walton J.D. J. Biol. Chem. 1992; 267: 26044-26049Abstract Full Text PDF PubMed Google Scholar, 20.Walton J.D. Holden F.R. Mol. Plant-Microbe Interact. 1988; 1: 128-134Crossref Google Scholar), whereas the d-Ala of HC-toxin is synthesized by a separate protein, ToxGp, and is then activated and incorporated into HC-toxin by the third domain of HTS. HC-toxin biosynthesis differs from other non-ribosomal peptides in using two mechanisms to produce the necessary d-amino acids. The organization of the HC-toxin biosynthetic genes is also unusual compared with other fungal secondary metabolites. Genes controlling biosynthesis of fungal secondary metabolites such as penicillins in Penicillium, aflatoxin/sterigmatocystin inAspergillus, trichothecenes in Fusarium, and gibberellins in Gibberella are highly clustered (41.Brown D.W., Yu, J.-H. Kelkar H.S. Fernandes M. Nesbit T.C. Keller N.P. Adams T.H. Leonard T.J. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1418-1422Crossref PubMed Scopus (423) Google Scholar, 42.Trapp S.C. Hohn T.M. McCormick S. Jarvis B.B. Mol. Gen. Genet. 1998; 257: 421-432Crossref PubMed Scopus (76) Google Scholar, 43.Tudzynski B. Hölter K. Fungal Genet. Biol. 1998; 25: 157-170Crossref PubMed Scopus (164) Google Scholar, 44.Yu J. Chang P.-K. Cary J.W. Wright M. Bhatnagar D. Cleveland T.E. Payne G.A. Linz J.E. Appl. Environ. Microbiol. 1995; 61: 2365-2371Crossref PubMed Google Scholar). In contrast, although some of the genes involved in HC-toxin biosynthesis (i.e. HTS1 with TOXA andTOXF with TOXG) are clustered (defined as being within ∼3 kb of each other) (Refs. 26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar and 28.Pitkin J.W. Panaccione D.G. Walton J.D. Microbiology. 1996; 142: 1557-1565Crossref PubMed Scopus (225) Google Scholar; this study), the other genes of HC-toxin biosynthesis are only loosely physically linked over ∼600 kb (23.Ahn J.-H. Walton J.D. Plant Cell. 1996; 8: 887-897PubMed Google Scholar). The HC-toxin genes are also unusual in being present in multiple functional copies in all Tox2+ isolates that have been tested, with the exception of some laboratory-generated strains such as 164R10 and 243-7 (25.Ahn J.-H. Walton J.D. Mol. Gen. Genet. 1998; 260: 462-469PubMed Google Scholar). Despite their being only loosely physically linked, there is little or no crossing over, even between the TOX2 genes that are more than 280 kb apart (28.Pitkin J.W. Panaccione D.G. Walton J.D. Microbiology. 1996; 142: 1557-1565Crossref PubMed Scopus (225) Google Scholar). This explains why the trait of HC-toxin biosynthesis appears to segregate as a single Mendelian locus, TOX2 (5.Walton J.D. Plant Cell. 1996; 8: 1723-1733Crossref PubMed Scopus (259) Google Scholar, 45.Scheffer R.P. Nelson R.R. Ullstrup A.J. Phytopathology. 1967; 57: 1288-1289Google Scholar). TOXG is tightly clustered with TOXF, which is also specifically dedicated to HC-toxin biosynthesis (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar).TOXF is predicted to encode a BCAT, whose substrates arel-amino acids. Although the overall amino acid similarity is low between known BCATs and ToxFp, all of the nine essential active site residues are present (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). We hypothesized that ToxFp is responsible for aminating a precursor of Aeo (e.g.2-oxodecanoic acid). BCATs are structurally related tod-amino acid aminotransferases (44.Yu J. Chang P.-K. Cary J.W. Wright M. Bhatnagar D. Cleveland T.E. Payne G.A. Linz J.E. Appl. Environ. Microbiol. 1995; 61: 2365-2371Crossref PubMed Google Scholar), which is intriguing in light of the fact that TOXG encodes an enzyme that makes d-Ala. Insofar as clustered genes might encode enzymes that act in concert, this raises the possibility that ToxF is ad-amino acid aminotransferase rather than a BCAT. However, if d-Ala, produced by ToxGp, is the essential substrate for ToxFp, which in turn is essential for the production of Aeo, then mutation of ToxGp would be expected to eliminate Aeo synthesis and hence all forms of HC-toxin, not just those that containd-Ala. Therefore, it seems unlikely that d-Ala is the amino donor substrate for ToxFp, and the reaction catalyzed by this protein remains to be established (26.Cheng Y.-Q. Ahn J.-H. Walton J.D. Microbiology. 1999; 145: 3539-3546Crossref PubMed Scopus (30) Google Scholar). We are grateful to Kurt Schörgendorfer and his co-workers (Novartis Biochemie GmbH, Kufstein-Schaftenau, Austria) for their generous exchange of information and DNA samples and to Douglas Gage of the MSU Mass Spectrometry Laboratory for the HC-toxin analyses. Oligonucleotides were synthesized by the MSU Macromolecular Structure Facility.
Referência(s)