Expression of the Lantibiotic Mersacidin in Bacillus amyloliquefaciens FZB42

Lantibiotics are small peptide antibiotics that contain the characteristic thioether amino acids lanthionine and methyllanthionine. As ribosomally synthesized peptides, lantibiotics possess biosynthetic gene clusters which contain the structural gene (lanA) as well as the other genes which are involved in lantibiotic modification (lanM, lanB, lanC, lanP), regulation (lanR, lanK), export (lanT(P)) and immunity (lanEFG). The lantibiotic mersacidin is produced by Bacillus sp. HIL Y-85,54728, which is not naturally competent. Methodology/Principal Findings The aim of these studies was to test if the production of mersacidin could be transferred to a naturally competent Bacillus strain employing genomic DNA of the producer strain. Bacillus amyloliquefaciens FZB42 was chosen for these experiments because it already harbors the mersacidin immunity genes. After transfer of the biosynthetic part of the gene cluster by competence transformation, production of active mersacidin was obtained from a plasmid in trans. Furthermore, comparison of several DNA sequences and biochemical testing of B. amyloliquefaciens FZB42 and B. sp. HIL Y-85,54728 showed that the producer strain of mersacidin is a member of the species B. amyloliquefaciens. Conclusions/Significance The lantibiotic mersacidin can be produced in B. amyloliquefaciens FZB42, which is closely related to the wild type producer strain of mersacidin. The new mersacidin producer strain enables us to use the full potential of the biosynthetic gene cluster for genetic manipulation and downstream modification approaches.


Introduction
The lantibiotic (i.e. lanthionine-containing antibiotic) mersacidin is an antimicrobial peptide that consists of 20 amino acids. The producer strain of mersacidin, Bacillus sp. HIL Y-85,54728 [1], has not yet been closely characterized. The structural gene (mrsA) and the genes for modification enzymes, transporters and producer self-protection are encoded on a 12.3 kb biosynthetic gene cluster on the chromosome of the producer strain [2]. Moreover, three regulatory genes are present in the gene cluster. The twocomponent regulatory system MrsR2/K2 is mainly involved in immunity and induction of mersacidin biosynthesis in the presence of mersacidin by a quorum sensing mechanism [3]. A further single regulatory protein, MrsR1, is encoded downstream of mrsA and is essential for mersacidin production [4]. Mersacidin inhibits the growth of gram-positive bacteria by binding to the cell wall precursor lipid II and thereby inhibiting cell wall biosynthesis [5].
Production of mersacidin and genetically engineered mersacidin peptides has so far been performed in variants of the original producer strain, Bacillus sp. HIL Y-85,54728. Production of engineered peptides was obtained either in trans after inactivation of mrsA by introduction of a stop codon [6] or in cis after double homologous recombination [7]. However, transformation of the producer strain with exogenous plasmids could only be achieved by protoplast transformation or electroporation and both methods yielded only low transformation frequencies. Therefore, the aim of these studies was to build an expression system for mersacidin in a naturally competent Bacillus strain in synthetic medium and exploit competence transformation as an efficient method for the transfer of the biosynthetic gene cluster and plasmids harboring mrsA.
Successful heterologous production has previously been shown for several lantibiotics, e. g. subtilin and nisin by B. subtilis 168 [8,9], lacticin 3147 by Enterococcus faecalis [10], or epicidin 280 by Staphylococcus carnosus [11]. Very recently, production of active lichenicidin has even been achieved in Escherichia coli [12]. However, heterologous production of a lantibiotic cannot be taken for granted and remains difficult. For example, epicidin 280 and Pep5 are two closely related lantibiotics, but Pep5 shows a higher antibacterial activity than epicidin 280. In contrast to epicidin 280 [11], Pep5 cannot be expressed in S. carnosus, which is susceptible to this agent in the nanomolar range, indicating that the toxicity of the product may be a problem here (G. Bierbaum, unpublished data). Therefore, for successful and high-level expression of mersacidin in a heterologous host, two conditions had to be met, functional producer self-protection and transfer of the biosynthetic gene cluster to the new host. The genome of Bacillus amyloliquefaciens FZB42, a competent plant-growth promoting rhizobacterium, has recently been sequenced. The part of the mersacidin biosynthetic gene cluster, that is devoted to producer self-protection, is already present in this organism [13] making it an ideal candidate for the transfer of the biosynthetic genes. Here we show that production of mersacidin is possible in B. amyloliquefaciens FZB42 and that the mersacidin producer strain itself is a member of this species.

Results and Discussion
B. amyloliquefaciens FZB42 is immune to mersacidin Analysis of the complete genome sequence of B. amyloliquefaciens FZB42 [13] had shown, that mrsFGE, which encode an ABC transporter that inhibits binding of mersacidin to the cells [4], and mrsK2R2, the genes of a two component system that induces expression of mrsFGE in the presence of mersacidin [3], are present in this organism (Fig. 1). These genes are located at the same site as in the original producer strain of mersacidin, i. e. between ycdJ and fbaB [2]. A detailed comparison showed that the encoded proteins share at least 98 % amino acid identity (two exchanges in MrsE and MrsF, one exchange in MrsR2 and four exchanges in MrsK2), with the exception of the N-terminus of MrsG. However, resequencing of the wild type producer indicated that a thymidine residue in position 980 was missing in the original sequence of mrsG. After sequence correction, the MrsG amino acid sequences were identical. In the intergenic region between mrsE and fbaB, a short sequence of 147 bp is inserted in the genome of B. amyloliquefaciens FZB42 (bp 3774591 to 3774738) that is not present in the producer strain of mersacidin and did not yield any hits in the databases. This sequence is flanked by two distinct regions with sequence similarity to the mersacidin gene cluster. The upstream 38 bp region is similar to the sequence found downstream of mrsE, i. e. the sequence upstream of the putative mrsA operator. The 89 bp region found downstream of the 147 bp insert is homologous to the sequence downstream of mrsT, contains the inverted repeat that is thought to delimit the mersacidin biosynthetic gene cluster and is not present in B. amyloliquefaciens DSM 7 T . This might indicate that the biosynthetic part of the gene cluster was lost from B. amyloliquefaciens FZB42 during evolution.
A previous comparison of minimum inhibitory concentration (MIC) values of the wild type producer and an mrsK2R2 knockout clone, which does not express the mersacidin immunity genes mrsFGE, had demonstrated that expression of mrsK2R2FGE increased the resistance to mersacidin about threefold [4]. MIC determinations of the wild type producer (25 mg/l) and B. amyloliquefaciens FZB42 (25 mg/l) demonstrated that B. amyloliquefaciens FZB42 was at least as resistant to mersacidin as the producer strain and therefore this organism was chosen as amenable to mersacidin production.

Reconstitution of mersacidin production in B. amyloliquefaciens FZB42
In order to transfer the biosynthetic part of the mersacidin biosynthetic gene cluster into B. amyloliquefaciens FZB42, chromosomal DNA of Bacillus sp. HIL Y-85,54728 Rec1 was utilized. This strain harbors the complete mersacidin gene cluster including the operator sequence, apart from mrsA and its promoter that have been replaced by an erythromycin resistance cassette [2]. Therefore, the use of erythromycin as a selection marker for successful integration of the gene cluster was possible during transformation experiments. After a competence transformation, 15 erythromycin resistant colonies were isolated and insertion of the biosynthetic part of the mersacidin biosynthesis gene cluster was confirmed by PCR, employing the primers mrsE681.f and ermB665.r annealing in mrsE and ermB as well as mrsT2251.f and fbaB283.r annealing in mrsT and fabB. A PCR of comK, employing primers located in the intergenic region (FZBfor, FZBrev) that did not match the sequence of the wild type producer strain confirmed that the clones were indeed B. amyloliquefaciens FZB42 transformants and did not derive from spores of the producer strain. Subsequently, one of these clones, B. amyloliquefaciens FZB42 mrs1, was transformed with pPAR1 [3], which harbors mrsA and mrsR1, by competence transformation. The regulator MrsR1 is essential for mersacidin biosynthesis [4] and is transcribed from the mrsA promoter (Bierbaum, unpublished results).
In order to demonstrate the expression of mersacidin, MALDI-TOF analysis of the culture supernatant of B. amyloliquefaciens mrs1 pPAR1 was performed and spectra were compared to those of B. amyloliquefaciens FZB42 (wild type) as well as the mrs1 strain, missing mrsA (Fig. 2). The typical mersacidin masses [1826 Da: mersacidin + H, 1848 Da: mersacidin + Na and 1864 Da: mersacidin + K] were detected in the culture supernatant of the strain harboring pPAR1, indicating expression of fully modified mersacidin. In contrast, the spectra of B. amyloliquefaciens FZB42 and of B. amyloliquefaciens mrs1 did not show any mersacidinrelated mass peaks.
On the other hand, in agar well diffusion assays with supernatants of cultures incubated in the absence of chloramphenicol, the inhibition zones of B. amyloliquefaciens mrs1 pPAR1 against M. luteus, S. aureus SG511 and Bacillus megaterium were not significantly larger than those produced by B. amyloliquefaciens FZB42 or B. amyloliquefaciens mrs1. The reason for this observation was that B. amyloliquefaciens FZB42 is able to produce an array of antimicrobial and antifungal substances including polyketides (bacillaene, difficidin, macrolactin), lipopeptides (surfactin, fengy-cin, bacillomycin D), two siderophores (bacillibactin, product of nrs cluster), the antimicrobial dipeptide bacilysin as well as the thiazole/oxazole containing antibiotic plantazolicin [14,15]. Comparative MALDI-TOF spectra of B. amyloliquefaciens FZB42 and its mutant strains indeed indicated the presence of the lipopeptide surfactin and the antifungal compounds fengycin and bacillomycin D in culture supernatants of all tested B. amyloliquefaciens strains (data not shown). The activity of surfactin was also detected by hemolysis on Columbia blood agar plates. In conclusion, the antimicrobial activity of mersacidin was probably masked by the activity of surfactin in the agar well diffusion assays. The production of several antibacterial products by Bacillus strains is far from unusual. For example, secondary antibacterial compounds are also detected in the culture supernatants of the producer strain of mersacidin (Bierbaum, unpublished results) and even B. subtilis 168 -in spite of the mutation in sfp that inhibits production of the lipopeptides encoded in the genome [16] (surfactin [17] and plipastatin/fengycin [18]) -is still able to excrete at least four other antibacterial compounds, i. e. sublancin 168 [19], subtilosin A [20], bacilysocin [21] and bacilysin [22].
In order to demonstrate that the mersacidin produced by B. amyloliquefaciens mrs1 pPAR1 was correctly modified and showed antimicrobial activity, it was partially purified by two consecutive HPLC runs from cultures grown in the presence of chloramphenicol. The HPLC fractions that were active against M. luteus and that showed the typical adsorption spectrum of mersacidin were further analyzed by MALDI-TOF and showed the presence of the typical mass signals (Fig. 3). These results indicated B. amylolique- The producer strain of mersacidin belongs to the species B. amyloliquefaciens Upon comparison of various DNA sequences (yvnB, czcO, hpr, baeD, hemE, and comK) obtained from the producer strain of mersacidin to the sequences of the B. amyloliquefaciens FZB42 genome in NCBI, a high similarity between both strains became obvious. Table 1 demonstrates that all sequences, including intergenic regions, showed about 98.5 % nucleotide sequence identity to those of B. amyloliquefaciens FZB42 and about 93.5 % identity to the sequences of the type strain B. amyloliquefaciens DSM 7 T , whereas the identity to B. subtilis 168 was considerably lower (77.4 %).
During BLAST searches, the genome sequence of B. amyloliquefaciens FZB42 always showed the highest similarity to the sequence of the producer strain. B. amyloliquefaciens DSM 7 scored second, with the exception of the czcO region which seems to be partially missing in this strain, followed by B. subtilis 168 in third position (with the single exception of hemE, here B. subtilis W23 scored third with 168/212 identical bases).
The presence of baeD showed that Bacillus sp. HIL Y-85,54728 carries at least parts of the bacillaene gene cluster that was described for B. amyloliquefaciens FZB42 [13] and B. amyloliquefaciens DSM 7 T [23]. A biochemical identification test demonstrated that, in contrast to B. subtilis 168, B. amyloliquefaciens FZB42 and Bacillus sp. HIL Y-85,54728 both were able to metabolize xylose, lactose and starch and did not grow at 50uC nor in the presence of 10 % NaCl. Furthermore, both strains did not produce acid from trehalose and mannitol. The only difference between B. amyloliquefaciens FZB42 and Bacillus sp. HIL Y-85,54728 was the absence of gelatinase in the latter strain. 16S rRNA sequencing was performed with a PCR product that had been obtained using chromosomal DNA of Bacillus sp. HIL Y-85,54728 as a template. An NCBI BLAST search yielded a close similarity to B. amyloliquefaciens FZB42 16S rRNA (  ( Fig. 4) and, therefore, we propose that the producer strain of mersacidin should be renamed ''B. amyloliquefaciens HIL Y-85,54728''.
B. amyloliquefaciens FZB42 was isolated from the plant rhizosphere and has recently been defined as the type strain of a group of growth-promoting plant associated B. amyloliquefaciens strains (B. amyloliquefaciens subsp. plantarum) [24]. In addition to their ability to colonize plant roots, the members of the plantarum subspecies are discriminated from the subspecies B. amyloliquefaciens subsp. amyloliquefaciens by differences in the gyrA and cheA nucleotide sequence, hydrolysis of cellulose and an increased ability to produce nonribosomal secondary metabolites like fengycin and difficin [24]. In a taxonomic tree that was calculated from the gyrA nucleotide sequences of the mersacidin producer and different members of the genus Bacillus, the producer strain is found in the cluster formed by the members of the plantarum subspecies, suggesting a close association between the mersacidin wild type producer and these strains (Fig. 4). The similarity was not confined to the gyrA gene but was also reflected by the higher overall nucleotide sequence identity (98.5 %) of the mersacidin wild type producer and B. amyloliquefaciens subsp. plantarum FZB42 in comparison to the value reached by B. amyloliquefaciens subsp. amyloliquefaciens DSM 7 T (93.5 %) ( Table 1).
A distinguishing feature of the subspecies B. amyloliquefaciens subsp. plantarum is their ability to colonize Arabidopsis roots [24]. In fact, nearly all members of this subspecies were isolated from plants, plant roots or like B. amyloliquefaciens FZB42 from infested soil. The only exception is represented by strain UCMB5113 that was isolated from soil. In contrast, the producer strain of mersacidin originates from soil of a salt pan in Mulund, India, and unfortunately a plant association of the original isolate was not mentioned in the first report [1]. However, the strain did not grow in the presence of 10 % salt in the laboratory, indicating that the salt pan might not be its natural biotope and that its presence in the sample might rather be due to its ability to form long-lived spores.
In conclusion, we could show here that it is possible to produce mersacidin in B. amyloliquefaciens FZB42. The successful production of fully modified and active mersacidin by this strain provides an appropriate in vivo expression system for the construction and expression of mersacidin analogs. The vast array of antibacterial and antifungal compounds that is already excreted by this organism is thought to provide competitive advantage in the rhizosphere [25]. It also harbors genes which are nearly identical to the immunity genes of mersacidin and which will afford additional protection against competing strains that excrete this lantibiotic. The strategy employed here, i. e. to use an organism that already possesses the immunity genes of a lantibiotic for production of the same substance, proved successful and led to the production of active and fully modified mersacidin.

Strains, plasmids, culture conditions and media
All bacterial strains and plasmids used in this study are listed in table 3. Strains were stored as 50 % glycerol stocks at 280uC. Bacillus strains were cultured in tryptic soy broth (TSB, Oxoid, Wesel, Germany) or on tryptic soy agar. Escherichia coli strains were cultivated in LB. All cultures were maintained at 37uC. For The nucleotide positions correspond to [29]. All other bases are conserved between the four strains. genetically manipulated strains, antibiotics were added to the growth media (ampicillin, 40 mg/ml; erythromycin, 25 mg/ml; chloramphenicol, 20 mg/ml).

Purification of nucleic acids and sequencing
Genomic DNA was prepared using the PrestoSpinD Bug Kit according to the recommendations of the supplier (Molzym, Bremen, Germany). Plasmid DNA was isolated using the Gene-Jet TM Plasmid Miniprep Kit (Fermentas, St. Leon-Rot, Germany). All nucleic acids were analyzed by agarose gel electrophoresis and by spectrophotometry (Nanodrop Technologies, Wilmington, USA). DNA sequencing was performed by Sequiserve (Vaterstetten, Germany) or Seqlab (Göttingen, Germany). Plasmid DNA and

Construction of the mersacidin producing B. amyloliquefaciens FZB42 mutants
The biosynthetic part of the mersacidin gene cluster was transferred to B. amyloliquefaciens FZB42 by competence transformation [26] using genomic DNA of the strain B. sp. HIL Y-85,54728 Rec1, which harbors a selection marker (ermB resistance cassette) instead of mrsA. Transformants were selected on TSA (0.3 mg/ml erythromycin) and were subsequently cultured on TSA containing erythromycin at a final concentration of 25 mg/ml. The transfer of the gene cluster was confirmed by PCR using primer combinations that anneal within ermB (ermB665.r), mrsE (mrsE681.f), mrsT (mrsT2251.f) and the downstream coding ORF fbaB/iolJ (fbaB283.r) (see Fig. 1). To exclude the possibility that these colonies might derive from germinated spores of Bacillus sp. HIL Y-85,54728 Rec1 that had not been eliminated during the gDNA preparation, differential PCRs were performed using the comK primers FZBfor and FZBrev. The 39 ends of these primers anneal to single nucleotide polymorphisms of the sequence of B. amyloliquefaciens FZB42, but not to those of B. sp. HIL Y-85,54728. The resulting clone was named B. amyloliquefaciens mrs1.
For reconstitution of mersacidin production, the mersacidin structural gene (mrsA) was introduced in trans on the plasmid pPAR1 by competence transformation, yielding B. amyloliquefaciens mrs1 pPAR1. The presence of the plasmids as well as the plasmid integrity was analyzed by plasmid isolation and gel electrophoresis.

Producer self-protection against mersacidin
To test the susceptibility of B. amyloliquefaciens FZB42, B. amyloliquefaciens mrs1, and B. amyloliquefaciens mrs1 pPAR1 to mersacidin, the minimal inhibitory concentration (MIC) of mersacidin was determined by arithmetic broth microdilution. Serial twofold dilutions of mersacidin were prepared in polystyrene round bottom microtiter plates (Greiner, Frickenhausen, Germany) using half concentrated Mueller Hinton II broth (Difco, Detroit, USA) containing 1 mM CaCl 2 . An inoculum of 5610 5 CFU/ml was employed in a final volume of 200 ml. The MICs were calculated from the lowest concentration of mersacidin resulting in the complete inhibition of visible bacterial growth after 16 hours of incubation at 37uC and compared to the MIC of the mersacidin producer B. sp. HIL Y-85,54728.

Mersacidin production by Bacillus amyloliquefaciens mrs1 pPAR1
The production of mersacidin by B. amyloliquefaciens mrs1 harboring pPAR1 was assayed in 50 ml synthetic medium (2 x BPM) [2] in the presence of chloramphenicol. The cells were grown for 24 hours at 37uC with agitation (180 rpm). For further analysis the culture supernatant was sterilized by filtration and stored at 220uC. The detection of antimicrobial activity in 50 ml of culture supernatant and HPLC fractions was performed by agar well diffusion assays on Mueller-Hinton agar II plates (Difco, Detroit, USA) seeded with the indicator strains Micrococcus luteus ATCC 4698, Bacillus megaterium KM and Staphylococcus aureus SG511 in wells with a diameter of 7 mm. After incubation at 37uC overnight, the growth inhibition zones were measured.
In order to remove the chloramphenicol and other antibiotics excreted by B. amyloliquefaciens FZB42, 5 ml of the culture supernatant containing 0.1 % trifluoroacetic acid (TFA, Sigma-Aldrich, Taufkirchen, Germany) were applied to a Poros RP-HPLCcolumn (10R2, 10064.6 mm Perseptive Biosystems, Freiburg, Germany) and eluted in a gradient of 30 % to 42 % acetonitrile (containing 0.1 % TFA). The peaks were detected measuring the absorbance at 210 or 220 and 266 nm. The fractions were collected and assayed for the antimicrobial activity against M. luteus ATCC 4698 in agar well diffusion assays. The active fractions of 85 ml of culture supernatant were lyophilized and purified further using an RP C-18 column (Nucleosil-100-C18, 25064.5 mm; Schambeck SFD GmbH, Bad Honnef, Germany) with a gradient of 50 to 65 % acetonitrile (containing 0.1 % TFA). For MALDI-TOF analysis (Bruker Biflex, Bruker Daltonics, Bremen, Germany) of culture supernatant and active HPLC fractions, 20 ml of each fraction were concentrated 1:10 using a rotational Vacuum Concentrator (RVC 2-18, Christ, Osterode, Germany). Then, a 1 ml sample was mixed with 2 ml matrix (alpha-cyano-4-hydroxycinnamic acid in acetonitrile: 0.1 % TFA in water, 1:3). The mixtures were spotted onto the MALDI target and air-dried. Mass spectra were measured in positive ion mode in the range of 500 to 4000 Da and analyzed by Flexanalysis 2.0 (Bruker Daltonics).
Species identification of B. sp. HIL-Y-85,54728 The abilities of the wild type mersacidin producer strain, B. amyloliquefaciens FZB42 and B. subtilis 168 to ferment dextrose, maltose, lactose, sucrose, xylose, trehalose, mannitol and starch, to reduce sulfur and nitrate and to hydrolyze tryptophan and gelatin were compared using a conventional biochemical identification test. All inoculated tubes were incubated at 37uC for 24 h. Furthermore, growth at high salt concentrations (10 %) and high temperature (50uC) as well as motility were tested. The PCR product of the 16S rRNA genes was sequenced using the primers 16s9.f and 16s1550.r [27]. The gene coding for the gyrase subunit A (gyrA) was partially sequenced employing the primers gyrA_F and gyrA_R [28].

Bioinformatic tools and nucleotide sequence accession numbers
The sequences of the chromosome of B. subtilis 168 and B. amyloliquefaciens FZB42 are deposited in the NCBI database CoreNucleotide under the accession numbers NC_000964 and NC_009725, respectively. The sequence of the mersacidin gene cluster has the accession number AJ250862. Further sequences of the mersacidin producer strain have been deposited at NCBI under the following accession numbers: yvn, JF519627; czcO, JF519628; hpr, JF519629; baeD, JF519630; hemE, JF519631; comK, JF519632; gyrA, JF519633.