INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Microbial conversion of precursors obtained by chemical synthesis is an attractive method for drug manufacture because it allows efficient stereoselective biosynthesis. Moreover, if genes involved in microbial conversion are identified, the converting activity can be markedly increased by gene manipulation. This study was undertaken to identify a Penicillium decumbens epoxidase that converts the chemically synthesized substrate cis-propenylphosphonic acid (cPA) into the commercially important antibiotic fosfomycin (FOM) (Fig. 1).

View larger version (8K): [in this window] [in a new window]   FIG. 1.   Conversion of cPA to fosfomycin.

FOM, first discovered in cultures of Streptomyces fradiae ATCC 21096, is structurally characterized by a phosphonate group, a carbon-phosphorus (C-P) bond, and an epoxy ring (2, 5). FOM inhibits the initial reaction in the biosynthesis of prokaryotic peptidoglycans (11) and thus has broad-spectrum antibiotic activity against gram-positive and gram-negative bacteria (24). FOM prepared by chemical epoxidation of cPA (2) yields a racemic mixture and thus requires a costly separation process to isolate the L-stereoisomer. In 1971, White et al. discovered that many Penicillium species could catalyze epoxidation of cPA (23). The microbial process had the advantage that the L-stereoisomer could be produced selectively.

We wanted to employ P. decumbens for commercial application of this bioconversion. However, the natural isolate had low conversion activities which could not be easily improved by mutation and screening. Furthermore, the classical approach of isolating the presumed rate-limiting enzyme and using reverse genetics to identify and overexpress the corresponding gene could not be used since cPA epoxidase was not detected in vitro. As an alternative strategy, we presumed that the enzyme(s) involved would accumulate in cultures most actively carrying out the conversion. Such proteins were identified by two-dimensional (2D) gel analyses. When the corresponding gene, epoA, was cloned by reverse genetics and reintroduced into P. decumbens, the goal of increasing cPA conversion yields was achieved. The similarity of the predicted EpoA protein sequence to those of other oxidases and induction of the corresponding promoter by phosphorous acid suggested that it was the cPA oxidase.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of the proteins induced by cPA and determination of their amino acid sequences. A spore stock of P. decumbens was inoculated into 30 ml of SM-1 medium (glucose, 3%; nutrient broth, 0.8%; yeast extract, 0.2%; malt extract, 0.3%; pH 7.0) and cultured in a 250-ml flask at 28°C for 48 h with shaking. This culture (1 ml) was transferred to 30 ml of P6-1 medium (wheat germ, 3.3%; corn steep liquor, 0.8%; NH₄NO₃, 0.2%; FeSO₄ · 7H₂O, 0.025%; glucose, 10%), and cultured in a 250-ml flask at 28°C for 48 h. Mycelia were harvested by centrifugation and resuspended in 20 mM Tris-HCl buffer (pH 7.5). A cell extract was obtained by sonication (with a Sonifier Cell Disruptor 350; Branson Sonic Power Company) and subjected to 2D gel electrophoresis with a Multiphor II (Pharmacia Biotech) system. Electrophoresis was performed with Immobiline Dry Strip (pH 4 to 7) (Pharmacia Biotech) in the first dimension and with ExcelGelR sodium dodecyl sulfate gradient 8-18 (Pharmacia Biotech) in the second dimension. The protein was transferred to a polyvinylidene difluoride membrane (ProBlott; Perkin-Elmer) and applied to a model 492 protein sequencer (Perkin-Elmer), and the N-terminal amino acid sequence was determined by Edman degradation. The internal amino acid sequence of the protein was determined after digesting the protein adsorbed to the polyvinylidene difluoride membrane with lysyl endopeptidase (Wako Pure Chemical Industries, Ltd.) and fractionating peptides with a model 172 µ-Preparative high-performance liquid chromatography (HPLC) chromatograph (Perkin-Elmer).

Design of the synthetic DNA primer for PCR. The mixed oligonucleotides N-1 (5'-AC[T/A]CCXGA[G/A]CA[G/A]AT[C/T/A]GC-3') and N-2 (5'-AC[C/G]CCXGA[G/A]CA[G/A]AT[C/T/A]GC-3'), corresponding to the N-terminal amino acid sequence of epoA (T-P-E-Q-I-A), were used as forward primers. The mixed oligonucleotides 15-1 (5'-GC[C/T]TG[G/A]AAXCC[G/A]TT[T/A]CC-3') and 15-2 (5'-GC[C/T]TG[G/A]AAXCC[G/A]TT[C/G]CC-3'), corresponding to the internal sequence (G-N-G-F-Q-A), were used as reverse primers for PCR amplification of the nucleotide sequence corresponding to peptide 15. X is inosine.

Cosmid cloning of genomic DNA. P. decumbens HP147 mycelia (about 10 g) grown in SM-1 medium were washed twice with 20 mM Tris-HCl buffer (pH 7.5), frozen with liquid nitrogen, and fragmented with a homogenizer (Nihonseiki Kaisha, Ltd.). Chromosomal DNA was extracted (18), digested partially with Sau3AI, and ligated into the BamHI site of cosmid pCRB8 (constructed and provided by K. Gomi [4a]). Packaging was done with a lambda DNA in vitro packaging kit (Amersham), and libraries were screened with an ECL direct nucleic acid labelling and detection kit (Amersham).

cDNA library construction, screening, and DNA sequencing. Cultures were grown and lysed as described above in "Purification of the proteins induced by cPA and determination of their amino acid sequences." Poly(A)⁺ RNA was prepared with a QuickPrep mRNA purification kit with procedure C provided by the supplier (Pharmacia Biotech). The cDNA library was prepared in lambda gt11 with 5 µg of poly(A)⁺ RNA isolated from cPA-induced cultures. cDNA was made with a TimeSaver cDNA synthesis kit (Pharmacia Biotech). Screening was performed with ECL direct nucleic acid labelling and detection systems with a PCR product as probe. The target fragment from the positive clone was subcloned into the EcoRI site of pUC118 (pFOC2), and its nucleotide sequence was determined.

Transformation of P. decumbens. P. decumbens LP3 spores were inoculated into 40 ml of S-1 medium (glycerol, 3%; nutrient broth, 0.8%; malt extract, 0.3%; yeast extract, 0.2%; sodium glutamate, 0.1%) and cultured in a 250-ml flask at 28°C for 48 h. This culture (2 ml) was diluted into a similar flask containing 40 ml of S-1 medium and grown for an additional 24 h. Mycelia in the culture were collected by centrifugation (this and subsequent centrifugations were carried out at 2,500 × g for 10 min), suspended in 35% sorbitol, and centrifuged again. Protoplasts were prepared by incubating the mycelia at 34°C for 3 h with gentle agitation in a solution containing 3 mg of achromopeptidase (Wako Pure Chemical Industries, Ltd.) per ml, 3 mg of lysing enzyme (Sigma) per ml, and 1 mg of chitinase (Sigma) per ml. After digestion, mycelia were removed by filtration through cotton, and protoplasts were recovered by centrifugation and suspended in 10 ml of 0.5 M sucrose containing 10 mM Tris-HCl buffer (pH 7.5) and 10 mM CaCl₂ (0.5 M sucrose buffer). This protoplast suspension (100 µl) was mixed with 10 µl of DNA (10 µg) and 400 µl of polyethylene glycol solution (60% polyethylene glycol 4000, 10 mM Tris-HCl buffer [pH 7.5], 10 mM CaCl₂). Ten milliliters of 0.5 M sucrose buffer was then added, and the mixture was stirred and centrifuged. The pellet obtained was resuspended in 1 ml of 35% sorbitol and spread on protoplast regeneration medium (potato dextrose agar [Difco], 3.9%; raffinose, 30%; hygromycin B, 25 µg/ml). A soft agar medium (2 ml; potato dextrose agar, 1.3%; raffinose, 30%) was poured over the surface, and the culture was incubated at 26°C for 5 days.

Construction and verification of epoA-disruption mutants. P. decumbens LP3 was transformed (see "Transformation of P. decumbens," above) with pFOM13, a plasmid containing its homologous epoA locus in which the structural gene was disrupted with the hygromycin resistance cassette (hgh). Transformants were selected with hygromycin (see "Transformation of P. decumbens," above), purified, and grown in S-1 liquid medium (see "Purification of the proteins induced by cPA and determination of their amino acid sequences," above). Genomic DNA was extracted and analyzed by Southern hybridization (see "Southern blot hybridization," below).

Evaluation of the cPA-converting activity of transformed strains. P. decumbens was grown as described above ("Purification of the proteins induced by cPA and determination of their amino acid sequences"). After cPA was added to a final concentration of 4 mg/ml, cultivation was continued at 28°C for an additional 7 days. The culture was centrifuged, and FOM was assayed in the supernatant by ion chromatography (column, TSKgel IC-Anion-PWXL; eluent, TSKeluent IC-Anion-A [Tosoh Co.]).

Southern blot hybridization. For Southern hybridization (20), DNA probes were labelled with digoxigenin-11-dUTP (Boehringer Mannheim) by random priming. Hybridization was carried out at 42°C in DIG Easy Hyb solution buffer (Boehringer Mannheim). Membranes were washed twice at 68°C in 2× SSC (1× SSC is 0.15 M NaCl and 15 mM C₆H₅Na₃O₇) and twice in 0.1× SSC containing 0.1% sodium dodecyl sulfate. The hybridized probes were immunodetected with antidigoxigenin according to the procedure provided (Boehringer Mannheim).

Measurement of the bar-inducing activity. A spore stock of P. decumbens LP3/pFOM8 was used to inoculate 30 ml of SM-1 medium. The culture was incubated at 28°C for 48 h with shaking, diluted 1/30 into P6-1 medium, and cultured at 28°C for 48 h. Phosphonates were added, and mycelia were collected 24 h later and used to prepare a cell extract as described above ("Purification of the proteins induced by cPA and determination of their amino acid sequences"). Phosphinothricin acetyltransferase (PAT) encoded by the bar gene was measured with phosphinothricin as the substrate at room temperature (21), and the specific activity (micromoles per minute per milligram of protein) was calculated from the rate of change in molar absorption coefficient of 5,5'-dithio-bis(2-nitrobenzoic acid) (13.6 × 10⁶).

Bacterial strains and reagents. cPA was chemically synthesized and used after purification (98.5% homogeneity). Other phosphonates were provided by Sigma Chemical Company. Concentrated aqueous stock solutions of these compounds were neutralized with NaOH and sterilized with a 0.45-µm-pore-size Millex filter unit.

Nucleotide sequence accession number. The nucleotide sequence described in this paper (2,329-bp HindIII-XbaI fragment) has been submitted to the DDBJ, EMBL, and GenBank nucleotide sequence databases and assigned the accession number D73371.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of proteins induced by cPA and determination of their amino acid sequences. Although repeated attempts to assay cPA-epoxidizing activity in crude extracts were fruitless, two observations suggested a novel approach to identifying the enzyme(s). Bioconversion rates observed in P. decumbens cultures pregrown in cPA were apparently higher than those observed in cultures without cPA preinduction. Secondly, cPA conversion was much more efficient in the improved strain, HP147, than in the original soil isolate. Therefore, 2D gels were used to identify proteins which either were induced by cPA or accumulated at higher levels in HP147 than in its parent (wild-type P. decumbens LP3). Representative gels are shown in Fig. 2.

View larger version (54K): [in this window] [in a new window]   FIG. 2.   cPA induction of P. decumbens protein spots. P. decumbens wild-type LP3 (c), a higher-producing derivative (HP147) (a and b), and LP3 transformed with a plasmid containing epoA (pFOM4) (d) were induced with cPA (a, c, and d) and compared to an untreated HP147 control (b). In other words, panels a to d show HP147 plus cPA, HP147, LP3 plus cPA, and LP3/pFOM4 plus cPA, respectively. Protein spots were silver stained and then scanned to produce the smoothed gel image with an Image Master system (Pharmacia Biotech) (4). cPA-induced proteins are labelled A, B, and C. MW, molecular mass.

View larger version (35K): [in this window] [in a new window]   FIG. 3.   Nucleotide and deduced amino acid sequences of epoA. A mixed PCR probe was designed based on the amino acid residue sequence enclosed by rectangles. The nucleotides with the dotted underlines indicate the DNA sequence amplified by PCR. Nucleotides with the solid underlines indicate TATAAAT and CCAAT motifs. The solid-underlined amino acids indicate lysyl endopeptidase peptides 3, 7, 13, and 15.

Cloning and sequencing of epoA and corresponding cDNA. A cosmid bank was prepared from HP147 genomic DNA and screened with the pep-3 probe. Four hybridization-positive cosmid clones were obtained from about 10,000 colonies. Southern hybridization of cosmid DNA showed that the target gene was located on a HindIII-XbaI fragment of about 2.3 kb (Fig. 4). This was subcloned into pUC119 to generate plasmid pFOM3 (Fig. 4).

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Construction of the plasmid for cloning and disruption of epoA. A HindIII-XbaI fragment (2.3 kb) containing epoA was isolated from cosmid 7 and inserted into pUC119 to generate pFOM3. pFOM4 was prepared by inserting the hph cassette [on an XbaI fragment isolated from pDH25Xba after EcoRI/XbaI linkers were added to pDH25 (3)] into the XbaI site of pFOM3 located downstream of epoA. To prepare pFOM13, pFOM3 was digested to eliminate an HpaI-BglII fragment encoding about 400 bp of the epoA gene. The digested plasmid was filled in and ligated to the hph cassette (isolated from pDH25Xba as a filled-in XbaI fragment). pFOM XbaI/HindIII fragments are labelled A, B, and C (observed in Southern blot of Fig. 5). Abbreviations: PtrpC, the A. nidulans trpC promoter region; TtrpC, the A. nidulans trpC terminator region; hph, the E. coli hygromycin B phosphotransferase gene.

Improved cPA-converting activity of strains transformed with epoA. To examine whether multiple copies of epoA isolated from the efficiently converting strain HP147 would improve the cPA-converting activity in its parental strain, wild-type P. decumbens LP3 was transformed with pFOM4 (containing the intact epoA gene) or pFOM13 (containing an insertionally inactivated epoA allele [see below, "Insertional mutagenesis of epoA," or Fig. 4]). Ten recombinants were cultured, and their cPA-converting activities were compared with that of the parental strain and P. decumbens containing pFOM13 (Fig. 4) as a control. The converting activities of the recombinants were 1.43- to 4.28-fold higher than that of the parental strain (P. decumbens LP3). P. decumbens LP3/pFOM13 was comparable to the parental strain (Table 1).

Overexpression of cPA-induced protein in transformed strains. To examine whether cPA-induced proteins were expressed in large amounts of pFOM4-containing strains with improved converting activities, P. decumbens LP3 and the recombinant that showed the greatest increase in the converting activity (P. decumbens LP3/pFOM4 TF7) were analyzed by 2D gel electrophoresis. Compared to the untransformed host (Fig. 2c), larger amounts of 31-kDa protein spots A, B, and C accumulated in the recombinant (Fig. 2d). Northern blots showed that induction of epoA by cPA was at the transcriptional level (data not shown).

Insertional mutagenesis of epoA. To further elucidate its function, the P. decumbens LP3 chromosomal epoA gene was inactivated by homologous recombination by the following strategy. First, the HpaI-BglII region (about 400 bp) of epoA (Fig. 3) cloned in pFOM3 was deleted and replaced with an hph cassette (pFOM13) (Fig. 4). This plasmid was used to transform P. decumbens LP3, and 50 hygromycin-resistant transformants were examined for their ability to convert cPA. Four of these could no longer epoxidize cPA to FOM.

View larger version (73K): [in this window] [in a new window]   FIG. 5.   Southern blot analysis of P. decumbens genomic DNA. Genomic DNA from P. decumbens transformants was double digested with HindIII/XbaI, blotted, and probed with pFOM13. Lane 1, genomic DNA from P. decumbens LP3; lane 2, genomic DNA from P. decumbens LP3/pFOM13 TF1 (cPA bioconversion-negative strain). lane 3, genomic DNA from P. decumbens LP3/pFOM13 (cPA bioconversion-positive strain); lane 4, pFOM3 double digested with HindIII/XbaI; lane 5, pFOM13 double digested with HindIII/XbaI. Four different HindIII/XbaI-sized bands were observed: 3.6 kb, identified in Fig. 4 as hph-epoA (fragment A); 3.2 kb, the pUC119 vector fragment (Fig. 4, fragment C); 2.3 kb, the nondisrupted epoA locus (Fig. 4); and 1.3 kb, hph (Fig. 4, fragment B).

Investigation of the function of EpoA. To further investigate the function of EpoA from the molecular evolutionary point of view, proteins having similar sequences were identified by using BLASTA. These analyses drew our attention to the similarity of EpoA to L-proline 4-hydroxylase of Dactylosporangium sp. (DDBJ/EMBL/GenBank accession no. D78338). The alignment was refined by careful visual examination (Fig. 6). We then computed homology between the two aligned sequences. Of 240 positions available for comparison, 75 were occupied by the same amino acids. Taking into account evolutionary constraints (Poisson correction), 37% of the sites were conserved. This value indicates that the two sequences had evolved from a common ancestor. Moreover, as the figure shows, there were many sets of continuous residues at which the two sequences have the same amino acid. In addition, both sequences have an H-DXXXXXH motif (Fig. 6), which forms the active site of a family of oxygenases (17). These results together provide independent supporting evidence that epoA encodes an oxygenase.

View larger version (42K): [in this window] [in a new window]   FIG. 6.   Alignment of EpoA and L-proline-4-hydroxylase (D78338) sequences. The sequences are represented by the one-letter amino acid code, and a dash refers to a gap. An asterisk above a residue indicates the same amino acid in both sequences. The underlined H-D....H residues in the two sequences form a motif that resembles the active site of oxygenase.

Regulation of the epoA promoter by phosphorus-containing compounds. To study the regulation of the epoA promoter, the structural gene cloned in pFOM4 was replaced with a reporter gene (bar) which confers bialaphos resistance (bar) to generate pFOM8 (Fig. 7). bar gene expression was measured by the activity of its corresponding enzyme, PAT. P. decumbens LP3/pFOM8 was used to screen compounds for their ability to induce bar expression (Table 2). PAT activity was markedly induced by cPA, FOM, and phosphorous acid (also demonstrated by unpublished Northern blots). Inductions with phosphonoacetic acid, DL-2-amino-3-phosphonopropionic acid, and 2-aminoethylphosphonic acid (9) were much lower (21, 13, and 2%, respectively). No induction of PAT activity was detected with phosphoric acid or phosphocreatine (Table 2). When the putative epoA promoter region ws deleted (pFOM15 [Fig. 7]), PAT activity was not induced by cPA (P. decumbens LP3/pFOM15; Table 2).

View larger version (20K): [in this window] [in a new window]   FIG. 7.   Construction of transcriptional fusions linking the epoA promoter region to the bar reporter gene. Plasmid pFOM5 was prepared from pFOM4 (Fig. 4) by ligating a 2.3-kb HindIII-XbaI fragment including epoA into HindIII-XbaI sites of pUC118. Plasmid pFOM6 was obtained from pFOM5 by introducing a ClaI site upstream of the ATG start codon of epoA by site-specific mutation (12) with a primer (5'-AAATCCCCCATCGATGAAGCCTC-3'). The Streptomyces hygroscopicus bar gene was PCR amplified from pARK22 (19) with two primers, barU (5'-AAGGATCGATATGAGCCCAGAACGACGCCC-3') and barR (5'-GCTTGGATCCTCAGATCTCGGTGACGGGC-3'), which included ClaI and BamHI sites. The termini were cut with ClaI and BamHI and replaced with hph from pDH25 (3) (pDHBAR). Next, a ClaI-XbaI fragment including bar and the trpC terminator (TtrpC) was removed from pDHBAR and inserted into pFOM6 cleaved with ClaI/XbaI so as to replace the epoA structural gene and put the bar reporter gene under the control of the epoA promoter (pFOM7). pFOM8 was prepared by inserting a cassette containing the hph gene into the XbaI site of pFOM7. pFOM15 was prepared by first deleting the putative epoA promoter region from pFOM7 by digestion with HindIII and ClaI and then by using the same procedure to insert the hph cassette. Abbreviations: PepoA, the P. decumbens epoA promoter region; bar, the S. hygroscopicus bialaphos resistance gene.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the fermentation industry, empirical approaches involving repeated rounds of mutation and screening are often required to create modified strains for efficient manufacturing of microbial products. Although the activities of some rate-limiting enzymes or transport proteins are undoubtedly increased in these modified strains (1), corresponding biochemical assays can be difficult to establish. In such cases, 2D gel electrophoresis can be used to identify these proteins (8).

We were not able to assay the enzyme system that converts cPA to FOM in P. decumbens. To overcome this problem, our approach was based on the principle that enzymes for bioconversions are often induced by corresponding substrates. We focused on three protein spots, apparently products of the same gene, which were induced by cPA under conditions where it was being converted to FOM. The fact that these spots also overaccumulated in strains that carry out the bioconversion more efficiently further indicated the important role of the corresponding gene, epoA. When epoA was reintroduced to P. decumbens LP3, the converting activity increased as much as 4.3-fold. P. decumbens lost the ability to convert cPA to FOM when the epoA gene was destroyed by homologous gene replacement (16). Thus, EpoA plays an important role in the epoxidation of this phosphonic acid derivative. cPA can also be converted to FOM in Pseudomonas and Flavobacterium spp.; however, two enzymes (bromoperoxidase and bromohydrin epoxidase) are required (10). Epoxidation of cPA is not the biosynthetic pathway used in the biosynthesis of FOM in Streptomyces wedmorensis (6). Transcriptional regulation of epoA gave some clues as to its role in phosphonate catabolism or anabolism.

In order to study cPA-induced transcriptional control, a construction was made in which the epoA promoter directed expression of the bar gene. This reporter gene was chosen because its product, PAT, is both easy to assay and selectable. The bar gene confers resistance to phosphinothricin and is widely employed as a selectable marker, primarily in plants but also in fungi. Thus, selection for increased phosphinothricin resistance in P. decumbens containing pFOM8 could be used to select for mutants which overexpress the epoA promoter. A subset of these mutants may be second-site, trans-acting alleles which also increase expression of the native epoA promoter directing transcription of this putative epoxidizing activity.

Here, we used the bar reporter construct to screen various substances for their ability to induce expression of the epoA promoter. This is the first report of a promoter whose expression can be regulated by phosphorous acid, phosphonates, and diverse derivatives. cPA-like phosphonates which induced the promoter included those that have no double bonds (FOM), one having a shorter carbon chain (phosphonoacetic acid), and several having amino radicals such as 2-aminoethylphosphoric acid.

Interestingly, while phosphoric acid (H₃PO₄) did not induce expression, phosphorous acid (H₃PO₃) was active. It is reasonable that the reduced, rather than the more oxidized, form of phosphorus induced expression of this promoter, apparently regulating an oxidative conversion. However, we cannot easily rationalize the fact that both the substrate (cPA) and the putative product of epoA (FOM) induced the promoter. Future studies of these curious observations may provide clues to the regulation of phosphonate metabolism, a field where little information is currently available (7, 14).

The TATAAAT sequence characteristic of eukaryotic promoters was found immediately upstream of epoA; the regulatory motif CCAAT appeared repeatedly further upstream. These sequences are likely to be involved in expression and induction by cPA. They play a similar role in regulation of amdS in Aspergillus nidulans (13).

The cPA-converting activity was markedly improved simply by amplifying one gene (epoA). Derivatives containing the cloned epoA gene (Fig. 2d) accumulated increased levels of three related proteins. Although these three proteins had slightly different molecular weights and isoelectric points, comparison of their N-terminal amino acid sequences and peptide maps (data not shown) suggested that they represented posttranslationally modified derivatives of EpoA.

The increase in the cPA-converting activity varied among the recombinants transformed with the same plasmid (Table 1, 1.43- to 4.28-fold). Southern hybridization of genomic DNAs (data not shown) indicated that this phenomenon was probably related to the site of plasmid insertion (a phenomenon commonly observed in fungi) rather than to differences in the number of copies of the plasmid. Further understanding of the complete enzyme system and its coordinated synthesis is expected to allow even greater increases in bioconversion efficiency.