INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many Bacillus subtilis strains produce a small peptide(s) with a long fatty moiety, the so-called lipopeptide antibiotics. The peptide portions of these compounds contain -amino acids with a D configuration and are produced nonribosomally with templates of the multifunctional peptide synthetases. As in the synthesis of the peptide antibiotic gramicidin S, the peptide chain grows in a defined sequence by moving on the template of the multifunctional peptide synthetase (18, 35, 44). On the basis of the structural relationships, the lipopeptides that have been identified in B. subtilis are generally classified into three groups: the surfactin group (27), the plipastatin-fengycin group (16, 41, 42), and the iturin group (17). The members of the surfactin and plipastatin-fengycin groups are composed of one -hydroxy fatty acid and 7 and 10 -amino acids, respectively, while the members of the iturin group consist of one -amino fatty acid and 7 -amino acids. The presence of the -amino fatty acid is the most striking characteristic of the iturin A group and distinguishes this group from the other two groups. The operons that encode surfactin (3), plipastatin-fengycin (16, 37, 38, 40), and mycosubtilin (4), which is a member of the iturin A group, have been sequenced and characterized. In particular, a study of the mycosubtilin operon of B. subtilis ATCC 6633 (4) showed that MycA, a novel template enzyme which has functional domain homology to -ketoacyl synthetase and amino transferase and amino adenylation, was present, which implied that MycA is responsible for incorporation of the -amino fatty acid.

Recently, biological control agents for plant diseases have received considerable attention as alternatives to chemical pesticides (32). B. subtilis RB14, which has a suppressive effect against several phytopathogens, is expected to be used as a biocontrol agent (1, 5). We previously demonstrated that the biocontrol activity of RB14 can be attributed mainly to production of iturin A (Fig. 1) (10, 26). Iturin A, as well as mycosubtilin, is a member of the iturin group. The amino acid compositions of iturin A and mycosubtilin are almost identical, except that the sixth and seventh amino acids are inverted, as shown in Fig. 1. In our previous study, we cloned a gene, lpa-14, that encodes the 4'-phosphopantheteinyl transferase required for maturation of the template enzyme of iturin A (6, 8, 15). For further investigation of the details of the synthesis steps and because of the interesting evolutionary relationship between iturin A and mycosubtilin, cloning and sequencing of the complete iturin A synthetase operon are essential.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Structures of iturin A and mycosubtilin. R indicates an alkyl moiety (generally C14 to C17). The arrows represent peptide bonds in the -CO-NH- direction. The differences between the two lipopeptides are indicated by underlining.

In this study, we examined the features of the iturin A synthetase operon based on the nucleotide sequence and gene disruption. By comparing the iturin A operon with the mycosubtilin operon, we found that the difference between the two operons may be a result of intragenic swapping of amino acid adenylation domains. We also obtained genetic evidence of probable horizontal transfer of the iturin A operon. In addition, a promoter replacement experiment whose goal was construction of a iturin A hyperproducer is also described below.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains and media. The strains and plasmids used in this study are listed in Table 1. Luria-Bertani (LB) medium was used for cultivation of Escherichia coli and B. subtilis (30). When necessary, antibiotics were added at the following concentrations: ampicillin, 50 µg/ml; chloramphenicol, 5 µg/ml; erythromycin, 10 µg/ml; tetracycline, 20 µg/ml; and neomycin, 20 µg/ml.

Transformation and DNA manipulation. B. subtilis RB14 was transformed by electroporation as described previously (19, 39). B. subtilis MI113 was transformed by a competent cell method as described previously (39). Chromosomal DNA of B. subtilis strains were prepared by using a method developed by Itaya and Tanaka (12). Routine DNA manipulation and E. coli transformation were performed as described previously (39). Plaque and Southern hybridizations were performed by a digoxigenin enzyme-linked immunosorbent assay using a DNA labeling and detection kit (Roche) as recommended in the instruction manual.

Transposon technique. Transposon-harboring plasmid pHV1249 (25) was introduced into RB14 by electroporation, and random mutagenesis of mini-Tn10 was performed by a previously described method (25). Cloning of the mini-Tn10-disrupted gene was carried out as follows. To construct a B. subtilis MI113 plasmid library of the RB14 chromosome, chromosomal DNA of the transposon-containing mutant was digested with HindIII, ligated at the HindIII site of plasmid pTB522 which could be replicated in B. subtilis (9), and then transformed into B. subtilis MI113 competent cells. The plasmid of the chloramphenicol-resistant transformant resulting from cloning of a chromosomal fragment containing mini-Tn10 was obtained from the library and designated pTB1006. The HindIII insert of pTB1006 was transplanted into the HindIII site of pBR322 by transformation of E. coli JM109, resulting in pBRHd8k (Fig. 2).

View larger version (27K): [in this window] [in a new window]   FIG. 2.   ORF organization of iturin A (A) and mycosubtilin (B) operons. The positions of the plasmid and phages sequenced are shown above the iturin A operon. The sequences of derivative strains RIA1 and R-PM1 associated with ituD are shown in a box. The intersecting dotted lines indicate the difference in the amino acid adenylation domain arrangement between the two operons. The mycosubtilin operon was drawn by referring to reference 4.

Phage library construction and screening. The RB14 chromosome was digested partially with EcoRI and then was separated by electrophoresis to obtain homogeneous fragments that were 10 to 20 kb long. The fragments obtained were used for lambda DASH II library construction with a Lambda DASH II/EcoRI vector kit (Stratagene) as recommended in the instruction manual. First, screening by plaque hybridization in which the 1.9-kb NdeI-PstI fragment of pBRHd8k was used as a probe identified a positive phage with a 19.2-kb insert, designated 12 (Fig. 2). By using the internal 1.7-kb EcoRI-HindIII fragment of 12 as a probe, 66, harboring a 16.0-kb insert, was obtained (Fig. 2). The end of the 1.9-kb EcoRI fragment of 66 was employed in the next screening, and 69 with a 12.4-kb insert was obtained (Fig. 2).

Disruption of ituD by the plasmid pop in-pop out method. A 2-kb Sau3AI-PstI fragment from pBRHd8k, which harbored the entire ituD gene, was inserted at the BamHI-PstI site of pUC19, generating pSC24Pst. The SalI-NsiI fragment of pSC24Pst was removed, and the SalI-PstI fragment of the neomycin resistance gene cassette (neo) from pBEST502 was inserted (11), resulting in pSC24PstNm. The HindIII-KpnI fragment of pSC24PstNm, which harbored the disrupted ituD gene, was inserted at the PstI site of pE194 by blunt-end ligation. The ligated mixture was transformed into MI113. We selected a neomycin-resistant colony and thus obtained pE24PstNm. The ituD coding region in RB14 was disrupted by using the thermosensitive replication origin of pE194, as described previously (40). First, pE24PstNm was transformed into RB14 by electroporation. Strain RB14(pE24PstNm) was plated onto LB agar containing neomycin and erythromycin and then incubated at 48°C. The resulting strain, RB14::pE24PstNm, was cultivated in LB medium without selective pressure at 30°C for 10 generations. The culture was diluted and plated onto LB agar to obtain single colonies. The neomycin resistance and erythromycin resistance of the resulting colonies were assayed. Finally, the disrupted mutant, RIA1, was isolated by screening for neomycin-resistant and erythromycin-sensitive colonies.

Primer extension analysis. Total mRNA was prepared from a stationary-phase culture of RB14 cultivated in Number 3 medium for 17 h. Four milliliters of the culture was centrifuged and washed with STE buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 0.1 M NaCl). The pellet was then resuspended in 50 µl of TE buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA), and 50 µl of TE buffer containing lysozyme (20 mg/ml) was added. The suspension was incubated at room temperature for 10 min. Further purification was performed with an RNeasy mini kit (Qiagen). For primer extension analysis, IRD41-labeled primer ITU-PREX (5'-ATCGCATCGCTCGCTTCTTCAAAC-3') (Fig. 3), which was complementary to the sequence from nucleotide 96 to nucleotide 119 downstream of the putative start codon of ituD, was purchased from NisSHINBO Co. (Tokyo, Japan). Total RNA was ethanol precipitated and dissolved in 20 µl of hybridization buffer [80 mM piperazine-N, N'-bis(2-ethanesulfonic acid) (PIPES) buffer (pH 6.4), 2 mM EDTA, 800 mM NaCl, 50 % formamide]. The solution was supplemented with 1.8 µl of primer ITU-PREX (1 pmol/µl), denatured at 80°C for 15 min, and then cooled to 30°C with gentle shaking. After ethanol precipitation, the pellet was dissolved in extension buffer (4 µl of Moloney murine leukemia virus reverse transcriptase [Toyobo Inc., Osaka, Japan], 8 µl of 5× buffer for Moloney murine leukemia virus reverse transcriptase, 16 µl of a solution containing each deoxynucleoside triphosphate at a concentration of 2.5 mM, 10 µl of H₂O, 2 µl of RNase inhibitor) and then incubated at 42°C for 1 h. The resulting cDNA was subjected to Li-cor dNA automated DNA sequencing.

View larger version (42K): [in this window] [in a new window]   FIG. 3.   Determination of the transcription initiation site of ituD by primer extension analysis. (A) The position of the band corresponding to the ituD-specific primer extension product is indicated by arrows. (B) Nucleotide sequence of the ituD promoter region (PituD). The positions of the ituD transcription initiation site (highlighted, +1), putative 10 and 35 regions (highlighted), and putative ribosome binding site (RBS) (underlined) are indicated. Putative protein sequences of YxjF and ItuD are shown below the DNA sequence (shaded). The sequence complementary to the oligonucleotide used for primer extension analysis (ITU-PREX) is double underlined. The putative -independent terminator downstream of yxjF is indicated by > and < above the sequence.

Promoter exchange. A 0.8-kb fragment of the 5' region of ituD, which contained up to the ribosome binding site but not the promoter, was obtained by PCR. PCR amplification with primers ITUP4-F (5'-CCCCTGTTCTAGATGATCGGAGGAATCTC-3'; underlining indicates an XbaI site, and italics indicates substituted bases) and ITUP5-R (5'-TGCATCGATTCTGTCCATCTAACCGGCATC-3'; underlining indicates a ClaI site) was performed with the RB14 chromosome by using KOD DNA polymerase (Toyobo Inc.). The fragment obtained was double digested with XbaI and ClaI and then inserted between the XbaI and ClaI sites of pUC19. The resulting plasmid was confirmed to have the correct sequence and was designated pKODP4P5. P_repU accompanied by the neomycin resistance gene cassette (P_repU-neo) was excised from pBEST502 (11) by digestion with XbaI and then inserted into the XbaI site of pKODP4P5. A plasmid with P_repU-neo, whose direction of transcription is the same as that of ituD, was selected and designated pP4P5Nm. The 0.8-kb NdeI-TthHB8I fragment, which was upstream of the ituD transcription start site prepared from plasmid subclone pSC53EcoT of the phage library, was blunt ended and inserted into the SmaI site of pP4P5Nm, generating pUCIPNm. The HindIII-KpnI fragment of pUCIPNm, which was P_repU-neo accompanied by the P_itu flanking region, was inserted into the PstI site of pE194 by blunting and ligation and then transformed into MI113, which resulted in plasmid pEIPNm2. Replacement of P_itu of RB14 by P_repU-neo by the plasmid pop in-pop out method using pEIPNm2 was performed by the method described above for ituD disruption. Promoter replacement was confirmed by PCR (data not shown).

Quantitative analysis of iturin A and surfactin. A culture of B. subtilis in 40 ml of Number 3 medium was acidified to pH 2.0 with 12 N HCl. Then the precipitate was collected by centrifugation and extracted with methanol. Iturin A and surfactin in the extracted solution were quantified by reversed-phase high-performance liquid chromatography (HPLC), as described previously (1, 40).

Nucleotide sequence analysis. Double-stranded DNA cloned in pUC19 was sequenced with a Li-cor dNA 4000L DNA sequencer by using an IRD41 dye-labeled primer (Nisshinbo Co., Tokyo, Japan) and a Thermo Sequenase cycle sequencing kit (Amersham Pharmacia Biotech Co.). A partial sequence of the insert of the screened lambda DASH II phage was determined by Nippon Seifun Co., Atsugi, Japan. Motif retrieval for proteins was performed by using the PROSITE package. Harrplot analysis was performed with the Genetyx package (Software Development Co., Tokyo, Japan). Multiple-alignment analysis and phylogenetic analysis were performed by using Clustal W.

Nucleotide sequence accession number. The nucleotide sequence of the iturin A operon of RB14 has been deposited in the DDBJ, EMBL, and GenBank nucleotide sequence databases under accession number AB050629.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning, sequencing, and analysis of iturin A operon. To identify the genes responsible for iturin A production, we performed transposon mutagenesis with mini-Tn10. About 5,000 transposon-containing colonies were replicated on an LB agar plate containing F. oxysporum to assay for iturin A production deficiency. Fifteen colonies exhibited no antifungal activity on the plate, and none of these colonies had the ability to produce iturin A, as determined by HPLC. Ten of the colonies were selected at random, and chromosomal DNA were prepared, digested by HindIII, and subjected to Southern hybridization. All 10 colonies contained mini-Tn10 in the same 8-kb HindIII fragment (data not shown). We selected one colony, designated strain 1006, for further study. No iturin A production by 1006 was observed (Fig. 4). The 8-kb HindIII fragment with a transposon was screened from the plasmid library of the strain 1006 chromosome by using the chloramphenicol resistance marker of mini-Tn10.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Qualitative HPLC analysis of the lipopeptides produced by RB14 (A), 1006 (B), and RIA1 (C). Peaks corresponding to iturin A (IT) and surfactin (SF) are indicated. Plipastatin-like peaks (not identified) (PL) are also indicated.

View larger version (48K): [in this window] [in a new window]   FIG. 5.   Alignment of iturin A operon of RB14 and xynD regions of strain 168. Shaded uppercase letters indicate the sequence of the xynD region of strain 168 (based on the complementary sequence from positions 164, 250 to 163, 691 of the accession no. Z99113 sequence). Lowercase letters indicate the iturin A operon sequence (accession no. AB050629) that has homology to the strain 168 sequence. Sequences encoding iturin A synthetase are indicated by arrow-shaped boxes. Identical nucleotides in the two sequences are indicated by vertical lines. The numbers in the boxes are the positions in the deposited sequence. The lines indicate linkage between discrete homologous segments of the iturin A operon. Underlining indicates coding regions of yxjF and xynD. 35 and 10 are promoter sites of ituD, and +1 is a transcription start site of ituD. The alignment was constructed based on the results of a BLAST search. RBS, ribosome binding site.

Disruption of the ituD gene to derive the iturin A-deficient phenotype. To confirm that ituD is responsible for iturin A synthesis, ituD was disrupted (Fig. 2). The resulting disruptant, RIA1, was inoculated into Number 3S medium and cultivated at 30°C for 60 h to assay for iturin A production. As shown in Fig. 4, analytical HPLC of an RIA1 culture extract resulted in no iturin A peak, while the production of the lipopeptide surfactin by RIA1 was the same as the production by the wild type, indicating that ituD disruption only resulted in an iturin A deficiency. Based on these results, we concluded that ituD is essential specifically for iturin A synthesis.

Promoter analysis of the ituD gene. The transcription start site of ituD was determined by the primer extension method. Total RNA of RB14 was extracted from a 17-h culture in Number 3 medium. The total RNA obtained was hybridized with oligoDNA by using a fluorescent probe that was designed to hybridize with nucleotides 96 to 119 of ituD from the translation initiation codon. cDNA was synthesized with reverse transcriptase, and the resulting cDNA was then sequenced. As shown in Fig. 3, the transcription start site of ituD was found to be an A residue 56 bp upstream of the first residue of the ituD initiation codon. Upstream of this start site, we found a TATACACA-16 bp-TAGGAT sequence that exhibited low levels of homology to the consensus 10 and 35 (TTGACA-17 bp-TATAAT) sequences of ^A (Fig. 3). This promoter was designated P_itu. We also searched for other transcription start sites up to the putative rho-independent terminator of yxjF but did not find any other such sites.

Promoter replacement of the iturin A operon for high iturin A production. Because the mutant with disrupted ituD had an iturin A deficient phenotype and there is no coding region between ituD and ituC, which implies that there is a promoter, it is possible that P_itu governs expression of the iturin A operon. To confirm this, we replaced P_itu with a constitutively expressed promoter by using the neomycin resistance gene cassette, P_repU-neo, of pBEST502 (10). P_repU-neo is the chimera of the promoter of the replication protein gene (P_repU) and the neomycin resistance gene (neo) of plasmid pUB110 (20). Since P_repU of P_repU-neo lacks autoregulation of transcription by RepU (24), constitutive expression of P_repU is expected. Moreover, P_repU is sufficiently strong that neomycin resistance can be conferred by a single copy of neo in the chromosome, and there is no terminator to stop transcription starting from P_repU. The presence of the neomycin resistance gene cassette in a transformant, therefore, indicates that P_repU has been introduced and implies that constitutive and enforced expression of genes located downstream of P_repU-neo occurs in a polycistronic manner. The neomycin resistance gene cassette was ligated with a PCR-amplified fragment of the 5' portion of ituD at the ribosome binding site of ituD and then attached to the upstream fragment of P_itu in front of the neomycin resistance gene. The resulting three-fragment array was ligated to the thermosensitive plasmid pE194 and transformed by electroporation, and then the promoter was replaced by the pop in-pop out method, as described above for ituA disruption. The resulting strain, R-PM1, contained the inserted P_repU-neo instead of 250 bp of the P_itu sequence (Fig. 2).

View larger version (22K): [in this window] [in a new window]   FIG. 6.   Time courses for production of two lipopeptides, iturin A produced by RB14 () and R-PM1 () and surfactin produced by RB14 () and R-PM1 ().

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we identified the 42-kb region of B. subtilis RB14 that contains the complete iturin A synthetase operon, which is more than 38 kb long. The iturin A operon is composed of four ORFs, ituD, ituA, ituB, and ituC, in that order. The iturin A operon closely resembles the mycosubtilin operon. ItuD, as well as FenF of B. subtilis ATCC 6633 (4) and F29-3 (2), exhibits homology to the malonyl CoA-transacylase FabD, which participates in fatty acid synthesis in E. coli (43) and B. subtilis 168 (23). It is thought that in some Streptomyces spp. malonyl CoA-transacylase may be responsible for not only fatty acid synthesis but also type II polyketide antibiotic synthesis (28, 36). To clarify whether ItuD is involved in iturin A production and/or fatty acid synthesis, we disrupted ituD, which resulted in a specific iturin A deficiency. Since neither the growth of nor surfactin production by strain RIA1, a strain in which ituD was disrupted, was inhibited, we concluded that ituD is indispensable for iturin A production. This is the first evidence that a gene related to fenF participates in synthesis of an iturin group antibiotic. On the other hand, ItuA exhibits homology to -ketoacyl synthetase, amino transferase, and peptide synthetase in one molecule and is probably responsible for synthesis of a -amino fatty acid accompanied by ItuD dipeptide (-amino fatty acid-Asn) formation, as has been proposed for the mycosubtilin synthetase MycA (4). These features are consistent with the report that cerulenin, which inhibits -ketoacyl synthetase, is more active in iturin A -amino fatty acid synthesis than in fatty acid synthesis, which implies that -ketoacyl synthesis occurs during iturin A production independent of fatty acid synthesis (7). Other genes, ituB and ituC, encode large peptide synthetases that build peptide chains on the precursor from ItuA. The level of homology between ituC and mycC, the counterpart of ituC in the mycosubtilin operon, is relatively low (64%) compared to the levels of homology for other groups (79%), reflecting the difference in the amino acid arrangements in ItuC and MycC. As shown in Fig. 2, ItuC is predicted to be responsible for the D-AsnL-Ser portion of iturin A, while MycC is thought to synthesize the D-SerL-Asn part of mycosubtilin. Therefore, it is not unreasonable to assume that swapping between two adenylation domains occurs. To examine the possibility that intragenic swapping occurs in the adenylation domain, the sequences of ituC and mycC were compared. The results of a dot matrix analysis of the nucleotide sequences of ituC and the other relevant synthetase genes (mycC, ituA, and ituB), including ituC itself, are shown in Fig. 7. Clearly, two regions of ituC exhibit discrete homology to mycC; the homology junction is roughly 10 amino acids upstream of L(TS)xEL (A1 motif sequence [18]) and the C terminus of GRxDxQVKIRGxRIELGEIE (A8 motif sequence [18]). Homologies are observed between asparagine adenylation domains (ItuC2 and MycC1) and between serine adenylation domains (ItuC1 and MycC2), implying that domain swapping occurs. Since there are three asparagine adenylation domains (ItuA1, ItuB2, and ItuC1) in iturin A synthetases (Fig. 2), it is likely that the asparagine adenylation domain of ItuC1 exhibits the highest level of homology to ItuA1 or Itu2. However, as shown in Fig. 8, phylogenetic analysis of 14 nucleotide sequences encoding the adenylation domain of iturin A and mycosubtilin revealed that the sequence most similar to the sequence encoding the asparagine adenylation domain of ItuC1 is the sequence encoding the asparagine adenylation domain of MycC2. All iturin A domains also show phylogenetic similarities to their counterparts in mycosubtilin domains. These results imply that iturin A and mycosubtilin have a common ancestor. Therefore, we propose a model in which ituC or mycC swapped nucleotide sequences encoding adenylation domains after a common ancestor became established.

View larger version (23K): [in this window] [in a new window]   FIG. 7.   Dot matrices for ituC and ituA (A), ituB (B), ituC (C), and mycC (D). Dots were placed at locations with identical nucleotides when more than 45 of 60 nucleotides were identical. Amino acid adenylation domains of functional modules of synthetases are indicated by black boxes.

View larger version (12K): [in this window] [in a new window]   FIG. 8.   Phylogenetic tree constructed by the neighbor-joining method based on the sequences of the adenylation domains (from part of the A3 motif, PKG to the A6 motif, GELC[Y]) of iturin A synthetase and mycosubtilin synthetase. The numbers are bootstrap values based on 1,000 replicates.

Although the overall sequence of the iturin A operon is highly homologous to that of the mycosubtilin operon, the flanking region of the iturin A operon is significantly different from that of the mycosubtilin operon. It has been shown that the mycosubtilin operon of ATCC 6633, which is not a plipastatin producer, is between pbp and yngL (4), while the plipastatin operon (167° to 171°), instead of the mycosubtilin operon, is present in strain 168 (40). In the case of RB14, the iturin A operon lies between yxjF and xynD. Since strain RB14 is also a producer of a plipastatin-like compound (Fig. 4), it is reasonable to conclude that the iturin A operon of RB14 is in a region other than the plipastatin region. However, in the case of strain 168, yxjF (341°) is located on the side opposite xynD (166°) in the chromosome (14). This can be explained by the fact that the SfiI digestion pattern of the RB14 genome, as determined by pulsed-field gel electrophoresis, was significantly different from that of strain 168 (data not shown), suggesting that the entire genomic structure of RB14 has little similarity to that of strain 168, including the location of yxjF and/or xynD.

However, unexpectedly, as shown in Fig. 5, the promoter region of the iturin A operon has significant homology to the upstream sequence of xynD of strain 168. Since the downstream region of the iturin A operon has two tandem repeat DNA segments with significant homology to the xynD region of strain 168, it is highly probable that the coding region of the iturin A operon is transferred into the promoter of xynD and resides in this promoter, rather than that the promoter of xynD is simply transferred into the iturin A operon. The possibility that such a large antibiotic gene transfer occurs was also suggested for the tyrocidine operon, in which some peptide synthetase genes are integrated into the parental gramicidin S operon (21). In terms of horizontal transfer, it is possible that large DNA transfers occur frequently. The duplicated copy of xynD may indicate transient existence of a tandem repeat of early iturin A operons that were intermediates in the adenylation domain swapping mentioned above.

Our attempt to introduce a repU promoter instead of an intrinsic promoter into RB14 resulted in increased production of iturin A (Fig. 6), and the advantage of using the P_repU-neo cassette was identified. This method might be useful for generating bacteria that act as effective biological control agents.

At present, we are able to compare very similar operons, such as surfactin operons of B. subtilis (3) and two lichenysin operons of Bacillus licheniformis (13, 45), plipastatin (37, 38) and fengycin (16) operons, and iturin A and mycosubtilin (4) operons. These comparisons may provide good techniques for elucidating the boundary of the functional domain that directs engineered peptide synthesis (22, 31, 33, 34).