Results

Identifying unique DNA regions in the genome of B. amyloliquefaciens strain FZB42

Taxonomic classification of Bacillus strains FZB24, FZB37, FZB42, FZB45 and 168

↓56

Bacillus strains FZB24, FZB37, FZB42 and FZB45 have been isolated from plant-pathogen-infested soil and their contribution in plant growth promotion as well as in suppression of plant pathogenic organisms has been documented [5]. Initial studies on growth characteristics and carbon source utilization of those strains showed that they are closely related to the Bacillus subtilis and Bacillus amyloliquefaciens groups [6]. The main criterion for distinguishing between the two Bacillus subspecies was the ability shared within the B. amyloliquefaciens strains to produce lipase and acid from lactose [216]. Thereby it was concluded that FZB24, FZB42 and FZB45 belong to the B. amyloliquefaciens group whereas FZB37 is more related to B. subtilis [5]. However, another study classified strain FZB24 as a member of the Bacillus subtilis group. In order to further verify these results, ribotyping analysis and macrorestriction profiling by PFGE were performed.

The same amount (2 μg) of genomic DNA from FZB24, FZB37, FZB42, FZB45 and B. subtilis 168 was digested overnight at 37°C using the restriction endonuclease EcoRI. After transfer and fixation of the samples on a nylon membrane, overnight hybridization at 55°C was performed (see materials and methods). A DNA fragment, part of the 16S rrnE gene of B. subtilis, was amplified by PCR using primers pRB1601 and pRB1602 [6] and after labelling with DIG-dUTP, it was used as the probe for Southern hybridization. The ribotyping analysis revealed that the patterns obtained for FZB24 and FZB42 were almost identical (Fig. 11). FZB45 displayed a unique riboprint with high similarity to those belonging to FZB24 and FZB42. In contrast, FZB37 and B. subtilis 168 provide profiles that are identical to each other but quite distinct from the ones observed for the rest FZB strains. Comparison of these patterns with a database of known riboprints was performed in DSMZ (Deutsche Sammlung von Mikroorganismen and Zellkulturen GmbH, Braunschweig, Germany). Thereby it was confirmed that FZB24, FZB42 and FZB45 belong to the B. amyloliquefaciens group, whereas FZB37 belongs to the B. subtilis group.

Figure 11: Riboprints of various B. subtilis / B. amyloliquefaciens strains.

Ribotype patterns obtained after digestion of genomic DNA of B. amyloliquefaciens FZB24 (1), FZB37 (2), FZB42 (3), FZB45 (4) and B. subtilis 168 (5) with EcoRI and hybridization with a DIG-labelled 16S rDNA probe. M, EcoRI / HindIII digested phage λ DNA; bands from bottom to top 0,9/1,4/1,6/2/3,5/4,2/5,1/21,2 kb

↓57

Moreover, B. subtilis 168 and B. amyloliquefaciens FZB42 were further investigated by macrorestriction profiling. PFGE is a more analytical approach than riboprinting since it allows better separation of larger DNA fragments. The profiles of B. subtilis 168 and B. amyloliquefaciens FZB42 obtained by PFGE after digestion with the restriction endonuclease SfiI are very distinct from each other, as seen in (Fig. 12). Furthermore, the patterns obtained with the commercially available B. amyloliquefaciens strain FZB24, after digestion with the same restriction endonuclease, were identical to the ones resulting from FZB42 as reported by [6]. Similarly, FZB37 has the same macrorestriction pattern as B. subtilis 168 [6].

Figure 12: Genomic DNA macrorestriction profiles of B. subtilis 168 and B. amyloliquefaciens FZB42.

Genomic DNA of B. subtilis 168 and B. amyloliquefaciens FZB42 was digested with the rare-cutter restriction endonuclease SfiI and then separated by pulsed field gel electrophoresis (PFGE). The obtained macrorestriction profiles of B. subtilis 168 and B. amyloliquefaciens FZB42 can be seen in panels A and B respectively. M, molecular mass marker (MidRange II PFG marker); bands from bottom to top 24,5/48,5/73/97/121,5/145,5/170/194/218,5/242,5/267 kb.

Suppression Subtractive Hybridization (SSH)

Suppression Subtractive Hybridization (SSH) was used as means of identifying extensive gene differentiation between B. amyloliquefaciens FZB42 and B. subtilis 168. At the time point that these experiments were performed, only preliminary data existed for the genome sequence of B. amyloliquefaciens FZB42 (see chapter 3.2 for the updated data), whereas the complete genome of B. subtilis 168 had already been published [7]. Therefore SSH provided us with a rapid but thorough first view of genetic variation between the strains, long before that was possible by direct comparison of both strains’ complete sequences. For these experiments, B. amyloliquefaciens FZB42 was used as tester strain in order to rapidly detect its unique DNA sequences that are absent from B. subtilis 168 (driver strain).

↓58

Sixty-six clones were obtained by this approach. After sequencing analysis, the clones were aligned at nucleotide level to the known genome of B. subtilis 168. Three of these clones appeared twice in the screen and ten displayed more than 60% nucleotide homology to the driver strain. The SSH application can be thus considered as very successful, since 84% of the gained clones contained sequences with low (less than 60%) nucleotide homology to the driver strain. Thereby for the first time fifty-three DNA segments of various lengths present in B. amyloliquefaciens FZB42 but not in B. subtilis 168 were identified. Furthermore, their putative function was deduced by basic alignment search tool analysis (BLAST). The results are presented in table 6.

Most interestingly, eight clones showed high similarity to genes of nonribosomal peptide synthetases and polyketide synthases. In particular, clone cAK6 displayed 81% amino acid homology to MycC, involved in mycosubtilin biosynthesis, whereas clone cAK49 was 98% similar to ItuB, involved in iturin A biosynthesis. These findings suggest that the genome of B. amyloliquefaciens FZB42 may contain an operon for the nonribosomal biosynthesis of an iturin-like antibiotic. In contrast, such operon is not part of B. subtilis 168 genome. In this strain only the peptide sythetase operons encoding for surfactin and fengycin are present. Furthermore, six clones displayed similarity to polyketide synthases. For example, clones cAK24 and cAK48 were respectively 58% and 31% homologous, at amino acid level, to PksR and PksM, proteins that are encoded within the single polyketide synthase operon (pksX) in B. subtilis 168 (Table 6). Low similarities between the pks operon present in the driver strain and the six sequences from the tester strain indicated that B. amyloliquefaciens FZB42 might contain operon(s) responsible for polyketide biosynthesis that differ from the one present in B. subtilis 168.

Clone cAK30 displayed 99% similarity to MrsG, a protein probably involved in the immunity against the lantibiotic mersacidin [36] (see also 1.3.1.2). mrsG is transcribed from the same operon as mrsF and mrsE, in the mersacidin-producer Bacillus sp. strain HIL Y-85,54728. These genes encode an ABC transporter that could be involved in protection against the antibiotic [36]. The question that arose was whether B. amyloliquefaciens FZB42 possessed the whole mrsFGE operon or even the entire biosynthetic gene cluster of mersacidin; none of these genes are encoded in B. subtilis 168. For this purpose, primer walking was performed on the region neighbouring clone cAK30. Thereby the presence of the mrsFGE operon and of mrsR2K2, the two-component regulatory system that controls the operon’s transcription, was demonstrated; the biosynthetic genes of mersacidin were not found in this genomic region.

↓59

A putative IS3-like transposase was identified in clone cAK2, indicating that horizontal gene transfer in B. amyloliquefaciens FZB42 might be achieved by transposases. Interestingly, such phenomena are not observed in B. subtilis 168 which does not contain any transposase in its genome. Moreover, two clones (cAK9 and cAK38) are similar to phage-related proteins. In addition to the phage-related protein, cAK38 harbours a hypothetical protein, conserved in B. licheniformis DSM13 and absent from the driver strain. The remaining clones exhibit similarity to proteins with various functions, such as transcriptional regulators (cAK10), thymidylate synthases (cAK11), membrane or cation efflux proteins (cAK17/cAK20) etc. Moreover, several obtained sequences, parts of ORFs or putative proteins, score the best homology to unknown proteins of B. subtilis 168. Considering that the nucleotide and amino acid homology is very low, it seems plausible that these proteins are new and non-existing in the already sequenced strains of the Bacilli family.

Since SSH is based exclusively on DNA similarity, it was only expected that some of the clones would contain non-coding regions. These regions are of high interest and can have regulatory function since RNA polymerase and various transcriptional regulators might bind to them. Moreover, these can be regions where non-coding RNAs are located. Even though such clones are considered as positive attempts towards identifying sequence variations between B. amyloliquefaciens FZB42 and the driver strain, since they exhibit low nucleotide similarity to B. subtilis 168, they are not included in the table 6; 39,6% of the positive clones mostly contained non-coding region.

Table 6: FZB42 strain-specific SSH clones

Clone

Size (bp)

Nucleotide similarity to B. subtilis 168

Putative function/accession number

Identities(aa level)

Organism

cAK1

258

n.s.

Predicted hydrolase (HAD superfamily)/AAU39182

57/85 (67%)

B. licheniformis DSM 13

cAK2

211

n.s.

Putative transposase part of IS element/YP_080245

37/68 (54%)

B. licheniformis ATCC 14580

cAK4

419

171/419 (40%)

Tmk, Thymidylate kinase/AAU39021

55/132 (41%)

B. licheniformis DSM 13

cAK5

314

n.s.

Acetyltransferase/ZP_00962455

20/38 (52%)

Sulfitobacter sp. NAS-14.1

cAK6

120

n.s.

MycC, Mycosubtilin synthetase C/AAF0879

31/38 (81%)

B. subtilis ATCC 6633

cAK7

191

n.s.

LicR, transcriptional regulator/AAU42903

18/31 (58%)

B. licheniformis DSM 13

cAK9

437

259/437 (59%)

YobO, similar to phage-related pre-neck appendage protein/CAB13795

111/137 (81%)

B. subtilis 168

cAK10

126

n.s

Negative transcriptional regulator/ZP_01185138

28/42 (66%)

B. weihenstephanensis KBAB4

cAK11

377

n.s.

Thymidylate synthase/ZP_00063805

59/96 (61%)

Leuconostoc mesenteroides ATCC 8293

cAK12

526

109/526 (20%)

YqeD, hypothetical protein/CAB14513

60/89 (67%)

B. subtilis 168

cAK13

491

n.s.

Florfenicol-chloramphenicol resistance protein/NP_899167

43/136 (31%)

Staphyloco

ccus sciuri

cAK14

392

n.s.

ykpA, hypothetical protein/CAB13316

31/103 (30%)

B. subtilis 168

cAK17

402

n.s

Putative membrane protein/ZP_00231274

46/126 (36%)

Listeria monocytogenes str. 4b 7858

Clone

Size (bp)

Nucleotide similarity to B. subtilis 168

Putative function/accession number

Identities(aa level)

Organism

cAK18

322

141/322 (43%)

TuaB, Teichuronic acid biosynthesis protein/O32273

87/107 (81%)

B. subtilis 168

cAK20

262

n.s.

cation efflux family protein/ZP_00230783

57/87 (65%)

Listeria monocytogenes str. 4b H7858

cAK22

310

n.s.

TreA, phospho-alpha-(1,1)-glucosidasephospho-alpha-(1,1)-glucosidase/CAA91015

58/91 (63%)

B. subtilis 168

cAK23

383

n.s.

YyaL, hypothetical protein/CAB16119

56/93 (60%)

B. subtilis 168

cAK24

380

106/380 (28%)

PksR, polyketide synthase/CAB13606

72/123 (58%)

B. subtilis 168

cAK26

339

n.s.

conserved hypothetical cytosolic protein/ZP_01182157

70/72 (97%)

B. cereus subsp. cytotoxis NVH 391-98

cAK27

296

n.s.

PksN, polyketide synthase/CAB13604

24/39 (61%)

B. subtilis 168

cAK30

378

n.s.

MrsG, putative ABC-transporter integral membrane protein/AB60256

104/105 (99%)

Bacillus sp. HIL-Y85/54728

cAK36

351

160/351 (45%)

YbdN, hypothetical protein/NP_388086

78/113 (69%)

B. subtilis 168

cAK38

496

n.s.

Phage-like element PBSX protein xkdC/P39782-hypothetical protein BLi01445/AAU40348

32/55 (58%)-24/46 (52%)

B. subtilis 168-B. licheniformis ATCC 14580

cAK39

353

n.s.

hypothetical protein, putative carbohydrate esterase/AAU41672

108/119 (90%)

B. licheniformis ATCC 14580

cAK47

423

174/423 (41%)

YwbD, hypothetical protein/NP_391715

77/129 (59%)

B. subtilis 168

cAK48

261

n.s.

PksM, polyketide synthase/P40872

27/87 (31%)

B. subtilis 168

cAK49

390

n.s.

ItuB, iturin A synthetase B/BAB69699

127/129 (98%)

B. subtilis RB14

cAK53

344

n.s.

PksL, polyketide synthase/Q05470

21/54 (38%)

B. subtilis 168

cAK54

346

n.s.

PksN, polyketide synthase/BG12652

27/102 (26%)

B. subtilis 168

cAK56

493

n.s

YwmC, hypothetical protein / AB03680

62/104 (59%)

B. subtilis 168

Clone

Size (bp)

Nucleotide similarity to B. subtilis 168

Putative function/accession number

Identities(aa level)

Organism

cAK58

443

n.s.

Hypothetical cytosolic protein/ZP_00740431

66/131 (50%)

B. thuringiensis serovar israelensis ATCC 35646

cAK59

497

183/497 (37%)

PksE, polyketide synthesis/ P_389593

108/155 (69%)

B. subtilis 168

Clones obtained by SSH containing sequences of B. amyloliquefaciens FZB42 that exhibit less than 60% nucleotide similarity to B. subtilis 168. The sequence’s size and its exact overall nucleotide similarity to the driver strain are indicated; less than 20% similarity is considered non significant (n.s.). The putative functions of the DNA segments are presented, as derived by BLASTX alignment. Similarities on amino acid level are indicated for the aligned part of the sequences. Clones that mostly comprise of non-coding regions are not included in the table.

↓60

When later the sequencing task of B. amyloliquefaciens FZB42 advanced, the identification of the regions flanking the sequences obtained by SSH was possible. Moreover, in many cases sequencing of the strain was directed by the clones obtained by SSH. Our focus was particularly drawn on those regions that contained genes coding for nonribosomal peptide synthetases and polyketide synthases and further attempts to obtain detailed sequence information of these regions were conducted (mainly by primer walking).

Sequence analysis of these regions revealed the presence of three distinct pks gene clusters (pks1-72442 bp, pks2-54350 bp, pks3-69548 bp) (Fig. 13). Clone cAK24 belongs to pks1 gene cluster whereas clones cAK27, cAK53, cAK54 belong to pks2 gene cluster. pks3 polyketide synthase includes sequences obtained by clones cAK48 and cAK59. From these three gene clusters only the pks1 system from B. amyloliquefaciens FZB42 is similar to the pksX operon present in B. subtilis 168; a strain unable to synthesize polyketides due to a mutation in the sfp gene [174]. The polyketide synthases pks2 and pks3 are novel gene clusters. Various types of mass spectrometry, mutant construction and biological tests were used for verifying the functionality of these gene clusters. These experiments were in majority performed by Xiao-Hua Chen and can been seen in detail in [197].

Finally, clones cAK6 and cAK49 were found to be part of a ~37 kb operon that showed homology to the iturin A operon of B. subtilis RB14. Further experiments were performed for the characterization of this gene cluster (see 3.3 and 3.4).

↓61

Figure 13: Organization of the gene clusters involved in polyketide biosynthesis in B. amyloliquefaciens FZB42 (pks1, pks2, pks3) and B. subtilis 168 (pksX).

The size and location of the three polyketide gene clusters in the genome of B.amyloliquefaciens FZB42 are shown. Filled and bold arrows indicate discrete AT domains and modular PKS respectively. NRPS portions occurring in hybrid NRPS-PKS enzymes are shaded. Gene clusters pks1, pks2 and pks3 are responsible for the biosynthesis of bacillaene, macrolactin and difficidin/oxydifficidin respectively. AT, acyltransferase [197];(K.Schneider and Xiao-Hua Chen, unpublished results).

Sequence analysis of B. amyloliquefaciens FZB42 genome

Sequencing of the B. amyloliquefaciens FZB42 genome was performed as a joint collaboration between our laboratory and the GenoMik Network in Göttingen. The major part of the work and the co-ordination of the whole process were done by Xiao-Hua Chen and myself. Shotgun and fosmid library approaches, primer walking and multiplex PCR were used in order to obtain the complete genetic information encoded in the chromosome of B. amyloliquefaciens FZB42 (for more details see Materials and Methods, chapter 2.6). Sequencing of the whole genome of the strain has been completed whereas the second round of annotation using the GeneSOAP program is currently in process (performed by Xiao-Hua Chen).

The genome of B. amyloliquefaciens FZB42 is a single chromosome consisting of 3916 kb. The G+C content is about 46% and it contains 11 rRNA clusters. Even though the genome annotation is not yet completed, preliminary data revealed the presence of 3931 genes. BLAST comparison with SUBTILIST (a database containing all annotated genes of B. subtilis 168) showed that around 80% of the genes (3125) encoded by B. amyloliquefaciens FZB42 are more than 50% homologous at amino acid level to genes of B. subtilis 168. However, more than 200 of them are located in regions different than in the B. subtilis 168 genome, possibly due to rearrangement events that occurred during evolution of the two genomes. Moreover, co-linear regions exhibiting high similarity to the B. subtilis genome, which are then interrupted by regions of variable length containing genes unique for FZB42 were also detected. The unique genes of B. amyloliquefaciens FZB42 were found distributed in at least 14 DNA islands and islets around the whole genome. In contrast to B. subtilis 168, horizontal gene transfer is achieved not only by phages but also by different types of IS elements which are present in different copy numbers within the FZB42 genome (Table 7). The circular map of the chromosome of B. amyloliquefaciens FZB42 demonstrating some of its basic characteristics as well as an illustrated comparison to the B. subtilis 168 genome are presented in figure 14.

↓62

Table 7: Transposases present in B.amyloliquefaciens FZB42 genome

Transposase type

Position

Size (aa)

Identities (aa level)

Organism/accession number

IS231-related transposase

1999

96

65/80 (81%)

B. weihenstephanensis KBAB4 /ZP_01184801

IS3Bli1

2452

405

208/405 (51%)

B. licheniformis ATCC 14580/AF459921

transposase

3456

199

87/147 (59%)

Staphylococcus epidermidis ATCC 12228/AAO03698

IS3Bli1

3640

405

208/405 (51%)

B. licheniformis ATCC 14580/AF459921

IS3Bli1

3764

405

208/405 (51%)

B. licheniformis ATCC 14580/AF459921

The positions of the transposases on the genome of B. amyloliquefaciens FZB42 are indicated in kb. Similarities are indicated on amino acid level for the aligned part of the sequences, as derived by BLASTX alignment.

Striking is the presence of eight gene clusters encoding for secondary metabolites. In addition to srf, fen and pks1 (bae) operons that are responsible for the synthesis of surfactin, fengycin and bacillaene and are also present in the B. subtilis genome [7, 197], B. amyloliquefaciens FZB42 contains several additional gene clusters coding for peptide/polyketide antibiotics. bmy, pks2 and pks3 operons are involved in bacillomycin D, macrolactin, difficidin / oxydifficidin polyketide synthesis and further information about them can be found in other sections (see 3.3 and 3.1) [196, 197]; K.Schneider and Xiao-Hua Chen, unpublished results). Moreover, B. amyloliquefaciens FZB42 genome contains the bac operon responsible for the biosynthesis of the dipeptide bacilysin [217]. This antibiotic consists of an L-alanine at the N-terminus and a unusual amino acid, L-anticapsin, at the C-terminus and displays antibacterial activity [218]. The unusual epoxy-modified amino acid anticapsin is probably generated through the action of a prephenate dehydratase and an aminotransferase encoded by bacA and ywfG respectively, as a branching off from the prephenate of the aromatic amino acid pathway [219]. Additionally, the dhb operon is present in the genome of B. amyloliquefaciens FZB42 (see also section 3.5). The dhbACEBF operon is involved in the synthesis of 2,3-dihydroxybenzoate (DHB) as well as its modification and esterification to the iron siderophore bacillibactin [78] that enables microorganisms to efficiently scavenge iron [220, 221]. DhbE is a stand-alone adenylation domain that activates DHB in an ATP-dependent reaction. The activated DHB is subsequently transferred to the free thiol group of the co-factor phosphopantetheine of the bifunctional isochorismate lyase/aryl carrier protein DhbB. The third synthetase, DhbF, is a dimodular nonribosomal peptide synthetase that specifically adenylates threonine (and to a lesser extent glycine) as well as covalently loads both amino acids onto the corresponding peptidyl carrier domains [78].

The eight gene clusters encoding peptide/polyketide antibiotics and a siderophore represent about 8% of the total genome and control synthesis of bioactive compounds by processes based on nonconventional translation. Interestingly, three of these gene clusters (bmy, fen, pks1) are localized at the replication terminus, indicating that this region is probably more susceptible to horizontal gene transfer.

↓63

Figure 14: Whole genome map of B. amyloliquefaciens FZB42 (kindly provided by Xiao-Hua Chen).

The scale on the inner circle shows coordinates, in bp. The blue circle is the GC-skew, which is correlated with the replication start point. Total 11 sets of rRNA are shown by the pink arrows. The grey circle represents all 3931 genes detected in FZB42. The colored circle displays the distribution of the homologous genes with B. subtilis 168, detected by BLASTX and BLASTP comparison. Around 80% of the total genes bear similarity of more than 50%. The color code indicates identities greater than 90%, 80%, 70% , 60%, 50%. The green arrows within the outer circle indicate genes which are unique in the FZB42 and might contribute to the plant growth promotion or involved in horizontal gene transfer. The orange GC-content circle shows consistency between horizontal gene transfer and low GC-content.

Lipopeptides produced by B. amyloliquefaciens strain FZB42

Organization of nonribosomal peptide synthetases on the FZB42 chromosome

The SSH experiments revealed the presence of nonribosomal peptide synthetases and polyketide synthases in the genome of B. amyloliquefaciens FZB42. With the later acquirement of the first assembly of the organism’s genome sequence, it became clear that B. amyloliquefaciens FZB42 encodes operons srf, fen and bmy which are responsible for the synthesis of three lipopeptides: surfactin, fengycin and bacillomycin D. This was the first report revealing the coding sequence of bacillomycin D and evidence for its functionality was provided by MALDI-TOF MS analysis (see 3.3.2). B. amyloliquefaciens FZB42 also encodes three polyketide synthases [197].

The cluster of bmy is a FZB42-specific DNA island comprising of 4 genes (bmyD, bmyA, bmyB and bmyC) and is close to the fen operon on the chromosome. Regions flanking the large gene cluster are characterized by DNA rearrangements joining the antibiotic DNA islands with sequences originally present in different regions of the B. subtilis 168 chromosome (Fig. 15A). In particular, right from the 37.2 kb bmy gene cluster two rearranged clusters are situated: yxjCDEF and bioIBDFAW, that are present in B. subtilis 168 at positions 3999 to 4002 kb and 3088 to 3094 kb, respectively. On the left site, regions located in B. subtilis 168 at positions 1910 to 1943 kb (yndG, bglC, ynfJ and xynD) were detected. Interestingly, the bmy operon is inserted at the same position as the iturin A gene cluster in B. subtilis RB14. This “coincidence” and the high homology between bacillomycin D and iturin A made us initially assume (before the MS results; see 3.3.2) that FZB42 could encode the itu operon.

↓64

The fen five-gene cluster (fenA-E) present in B. amyloliquefaciens FZB42 is related to the pps operon in B. subtilis 168 and is situated at the same locus as this, about 25 kb distant from the bmy operon (Fig. 15A). Amino acid similarity between the operons of the two strains is between 60% (fenB) and 65% (fenC). The nonribosomal peptide synthetase directing biosynthesis of fengycin is also present in B. subtilis strains F29-3 [222] and A1/3 [140]. However, in B. subtilis ATCC 6633, mycosubtilin, another iturin-like lipopeptide, is found at the same genetic locus as fengycin [63]. There seems to be an exchange of these two operons between different strains, implying high degree of genetic flexibility in this region. Moreover, this suggests that additional NRPS operons might be integrated in this area either as insertions or as substitutions of already existing NRPS operons.

Figure 15: Organisation of the bacillomycin D, fengycin and surfactin operons in B. amyloliquefaciens FZB42.

In panel A is presented the chromosomal organization of bacillomycin D and fengycin operons in B. amyloliquefaciens FZB42 as well as of operons producing highly homologous antibiotics in other Bacilli strains. In panel B is compared the organisation of the surfactin operon in B. amyloliquefaciens FZB42 with that of B. subtilis 168. The intersecting dotted lines indicate events of insertion or rearrangement in FZB42 compared to the respective B. subtilis 168 genome region, whereas full lines demonstrate conservation of gene order between the two strains. Black-filled boxes indicate genes present in B. subtilis 168 but absent from the respective genome region of B. amyloliquefaciens FZB42. The organisation of the homologous gene clusters in B. subtilis 168 (fengycin operon-pps, surfactin-srf), B. subtilis RB14 (iturin A operon-itu) and B. subtilis ATCC6633 (mycosubtilin operon-myc) are presented according to [7, 49, 63, 101].

The 26.5 kb srf operon present in B. amyloliquefaciens FZB42 genome is organized in a similar manner as in B. subtilis 168 (see introduction). The corresponding genes of these two strains exhibited similarity between 72% (srfAA) and 83% (srfAC) on amino acid level. The genes present at the left flanking region of the operon are hxlBAR, like in the case of B. subtilis 168. However, on the right flank of srfAD, the B. subtilis 168 ycxAB are substituted by two ORFs with unknown function (Fig. 15B). Moreover, the comS gene, encoding a competence signal molecule, is embedded in the srfAB sequence, as already detected for various Bacillus strains and displays 63% homology to its orthologue in B. subtilis 168. The 4'-phosphopantetheinyl transferase Sfp is located 4kb downstream and exhibits 70% amino acid homology to the one encoded by strain 168.

Functional analysis of lipopeptide production in B. amyloliquefaciens FZB42

MS identification of the lipopeptide products of B. amyloliquefaciens FZB42

↓65

In order to check the functionality of the lipopeptide-encoding gene clusters, culture filtrate extracts and whole cells were investigated by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry (MS). The spectra obtained by both methods were found identical and revealed the presence of three different lipopeptides, as three groups of mass peaks were detected (Fig. 16). A summary of their mass numbers is presented in table 8.

Surfactins and fengycins have been identified by comparing their mass data with those previously obtained by MS analysis of numerous B. subtilis strains [211]. Moreover, B. amyloliquefaciens FZB42 produces surfactins and fengycins with fatty acid side chains of 13 to 15 and 15 to 17 carbon atoms, respectively.

Figure 16: MALDI-TOF MS analysis of lipopeptides produced by B. amyloliquefaciens FZB42 (performed in collaboration with Dr. J. Vater).

Detection of surfactin (A), bacillomycin D (A) and fengycin (B) mass peaks in culture filtrate extracts prepared from B. amyloliquefaciens FZB42 grown for 24 hours in Landy medium. Spectra of intact whole cells grown on Landy medium agar plates; detection of surfactin (C), bacillomycin D (C) and fengycin (D) mass peaks. See table 8 for peak identification.

↓66

Table 8: Lipopeptide products of B. amyloliquefaciens FZB42 detected by MALDI-TOF mass spectrometry

Observed peaks m/z

Assignment

Surfactin

1030.8*/1046.8

C13-surfactin [M + Na, K]+

1044.8*/1060.8

C14-surfactin [M + Na, K]+

1058.8*/1074.8*

C15-surfactin [M + Na, K]+

Bacillomycin D

1031.7/1053.7*/1069.7

C14-bacillomycin D [M + H, Na, K]+

1045.7*/1067.7*/1083.7*

C15-bacillomycin D [M + H, Na, K]+

1059.7/1081.7/1097.7*

C16-bacillomycin D [M + H, Na, K]+

1095.7/1111.7*

C17-bacillomycin D [M + Na, K]+

Fengycin

1449.9*/1471.9*/1487.9

Ala-6-C15-fengycin [M + H, Na, K]+

1463.9*/1485.9*/1501.9*

Ala-6-C16-fengycin [M + H, Na, K]+

1477.9*/1499.9/1515.9*

Ala-6-C17-fengycin [M + H, Na, K]+

1491.8/1513.9/1529.9*

Val-6-C16-fengycin [M + H, Na, K]+

1505.8/1527.8/1543.8*

Val-6-C17-fengycin [M + H, Na, K]+

Peaks indicated in figure 16 are marked by an asterisk.

Furthermore, postsource decay (PSD) MALDI-TOF MS revealed that the third produced lipopeptide is bacillomycin D. In the mass spectra obtained for whole cells and surface extracts the mass peaks of sodium and potassium adducts dominate, whereas the protonated species appear with minor intensities. However, the protonated species are preferred for PSD MALDI-TOF MS mediated sequence analysis because they decompose into fragments more readily than the alkali adducts. For example, the lipopeptide with a mass number of m/z 1031.5 produced by B. amyloliquefaciens FZB42 was identified as the protonated form of a bacillomycin D isoform with a fatty acid chain of 14 carbon atoms. Its sequence was determined from a series of bn1-, Yn‘‘(-H2O)-, and proline-directed bn2- fragment ions (Fig. 17). The peptide ring of this bacillomycin D was cleaved both at the peptide bond between its amino fatty acid residue and threonine at position 7 as well as at the N terminus of proline-4. In the first case a series of bn1- and Yn‘‘(-H2O)-fragment ions were detected. In addition, bn2-ions of highly intensity were observed. Based on all these data, this lipopeptide was identified as the protonated form of a C14- bacillomycin D. The obtained sequence was corroborated by bn1- ions of dipeptide fragments at m/z values of 171.4, 212.3, 226.8 and 268.4, indicating nearest-neighbour relationships in the peptide ring of this lipopeptide for ES(-H2O), NP, PE and NY respectively.

Figure 17: In situ structural analysis of the lipopeptide product of B. amyloliquefaciens FZB42 with mass number m/z 1031.5 by PSD-MALDI-TOF-MS (performed in collaboration with Dr. J. Vater).

The structure was derived from a series of f bn1-, Yn‘‘(-H2O)-, and proline-directed bn2- fragment ions. FA, fatty acid

Production of lipopeptides along the growth curve

↓67

The production of lipopeptides during growth in liquid cultures was monitored by MALDI-TOF MS. This type of spectrometry is not suitable for determining their exact concentrations, mainly because of inhomogeneities in the analytical distribution of the crystalline matrix and different ionization efficiencies of the investigated compounds. However, the relative quantities of the three antibiotics at different points of growth can be estimated by calculating the ratios of the intensity values of the peaks corresponding to the various antibiotics.

For this reason, culture filtrate extracts from B. amyloliquefaciens FZB42 grown in ACS medium at 37°C under vigorous shaking for 10, 20, 40 and 60 hours were subjected to MALDI-TOF MS analysis. Thereby it was shown that surfactins and bacillomycins were present at similar intensities but reached their zenith in different stages of growth. Maximum levels of surfactin appeared in samples obtained after 10 to 40 hours of growth, whereas after 60 hours production dropped. On the other hand, bacillomycin D accumulated after 40 to 60 hours of growth. The time course of fengycin resembled that of surfactin, but its intensity was clearly lower compared to the other lipopeptides (Table 9). The same pattern of lipopeptide production was obtained also when B. amyloliquefaciens FZB42 was grown in Landy medium and samples were drawn at 12, 24 and 48 hours. In this case, surfactin peaked already at 12 hours, whereas bacillomycin D at 24 hours. The intensities of fengycin peaks maintained rather low all along growth.

Table 9: Time-dependent production of lipopeptides by B. amyloliquefaciens FZB42 grown in ACS medium

Lipopeptide

m/z

Species

Intensity

at:

10 hours

20 hours

40 hours

60 hours

Surfactin

1044.8

C14 [M + Na]+

15500

10300

10100

2500

1058.8

C15 [M + Na]+

16500

11700

10000

2400

Bacillomycin

1053.7

C14 [M + Na]+

3800

6600

11700

12300

D

1067.7

C15 [M + Na]+

2500

2500

7600

7000

Fengycin

1485.9

C16 [M + Na]+

3500

2950

3630

470

1449.9

C17 [M + Na]+

1800

1810

2510

170

Lipopeptide production was monitored by MALDI-TOF mass spectrometry. Culture filtrate extracts were prepared after 10, 20, 40 and 60 hours of growth in ACS medium. Only the main peaks (m/z) of each lipopeptide were selected for analysis.

Lipopeptide deficient mutants

↓68

In order to confirm that bmy and fen operons are directing bacillomycin D and fengycin biosynthesis, disruption mutants at the bmyA and fenA genes were created. B. amyloliquefaciens FZB42 mutant strains were generated via double-crossover recombination [199] according to a modified protocol that has been originally developed for B. subtilis (see materials and methods).

In detail, a 1,2 kb fragment of bmyA was amplified by PCR using primers bmyAa and bmyAb (Table 4) and cloned into pGEM-T. After digestion with the restriction endonuclease AvaI, an erythromycin cassette was inserted inside the bmyA-fragment, resulting in plasmid pAK2 which was transformed in B. amyloliquefaciens FZB42. Disruption of bmyA in the resistant colonies was demonstrated by PCR with primers bmyAa and bmyAb and by Southern hybridization. Disruption of fenA was achieved in a similar manner by insertion of a chloramphenicol cassette in a fragment obtained by PCR with primers fenAa and fenAb and digested with HindIII and KpnI. Correct integration of the antibiotic cassette was verified by PCR and Southern hybridization. Furthermore, a double mutant of both bmyA and fenA genes was created.

Analysis of the mutant strains by MALDI-TOF MS, verified that strains ΔbmyA::Emr and ΔfenA::Cmr failed to produce bacillomycin D and fengycin respectively, since the corresponding groups of mass peaks were absent (Fig. 18). Moreover, strain AK3 (ΔbmyA::Emr ΔfenA::Cmr) was deficient in the production of both lipopeptides. Disruption of srfAA was performed by Xiao-Hua Chen and resulted in the strain’s inability to produce surfactin (data not shown). Consequently, the NRPS gene clusters that were identified on the chromosome are responsible for the biosynthesis of the respective lipopeptides in B. amyloliquefaciens FZB42.

Biological activity of wild type and mutant strains

↓69

B. amyloliquefaciens FZB42 stimulates plant growth and suppresses plant pathogenic organisms [5, 6], via mechanisms that have not been yet fully characterized. In order to check if the nonribosomal synthesized peptides produced by this bacterium contribute to its biocontrol capacity, the wild type and mutant strains deficient in biosynthesis of lipopeptides were assayed for their biological activities.

For this purpose, growth of various phytopathogenic fungi in the presence of B. amyloliquefaciens FZB42 was investigated. Strain FZB42 was shown to inhibit the growth of Fusarium oxysporum, Gaeumannomyces graminis, Rhizoctonia solani, Alternaria alternate and Pythium aphanidermatum. Moreover, strain AK1 (ΔbmyA::Emr) suppressed growth of all fungi at a smaller extent, suggesting that bacillomycin D contributes to the antifungal activity of B. amyloliquefaciens FZB42. Interestingly, even though growth inhibition in the presence of strain AK2 (ΔfenA::Cmr) was comparable to the one caused by the wild type, no inhibition was observed in the double mutant strain AK3 (ΔbmyA::Emr ΔfenA::Cmr) (Fig. 19A). The fact that fungi could grow uninfluenced by the presence of strain AK3 indicates a synergistic action of bacillomycin D and fengycin, against the target microorganism. The surfactin deficient mutant strain ΔsrfAA::Emr, provided by Xiao-Hua Chen, still retained its antifungal properties.

Figure 18: MALDI-TOF MS analysis of mutant strains in nonribosomal peptide synthetases (performed in collaboration with Dr. J. Vater).

Spectra of culture filtrate extracts prepared from strains grown for 24 hours in Landy medium. A) Bacillomycin D is not produced by mutant AK1 (ΔbmyA::Emr), whereas surfactin and fengycin are. B) Strain AK2 (ΔfenA::Cmr) is deficient in fengycin production while (C) strain AK3 (ΔbmyA::Emr ΔfenA::Cmr) is deficient in both bacillomycin D and fengycin production.

↓70

In parallel, it was observed that B. amyloliquefaciens FZB42 suppresses the growth of B. megaterium. This inhibitory activity was also shared by each one of the strains deficient in lipopeptide biosynthesis (Fig. 19B). Consequently, the antibacterial properties of strain FZB42 are driven by some antibiotic(s) produced by the strain other than nonribosomal peptides [197, 223].

Figure 19: Biological activity of B. amyloliquefaciens FZB42 and lipopeptide deficient mutant strains.

A) Growth of Fusarium oxysporum f.sp.cucumerinum DSMZ 62313 in the presence of B. amyloliquefaciens FZB42 and lipopeptide deficient mutants. B) Growth of B. amyloliquefaciens FZB42 and lipopeptide deficient mutant strains on B. megaterium lawn. (1) wild type strain, (2) AK1 (ΔbmyA::Emr), (3) AK2 (ΔfenA::Cmr), (4) AK3 (ΔbmyA::Emr ΔfenA::Cmr) and (5) CH1 (ΔsrfAA::Emr).

Analysis of functional domains in bmy operon

Sequence analysis of the bmy operon revealed the presence of a cluster of four ORFs designated bmyD, bmyA, bmyB and bmyC, respectively (Fig. 20). bmyD encodes a putative malonyl coenzyme A transacylase and displays strong similarity to FabD, that is involved in fatty acid synthesis. Moreover, the protein is 98% and 80% identical to ItuD and FenF, respectively (ItuD and FenF are the first proteins participating in iturin and mycosubtilin biosynthesis[63, 101]. The ORFs encoding BmyA (3,982 aa), BmyB (5,633 aa) and BmyC (2,619 aa) show high similarity to members of the nonribosomal peptide synthetase family and display the ordered assembly of conserved condensation, adenylation and thiolation domains characteristic for such multienzymes (see introduction). As shown in figure 20, seven amino acid activating modules can be distinguished: one in BmyA (A1), four in BmyB (B1, B2, B3 and B4) and two in BmyC (C1, C2). Modules B1, B2 and C1 contain an epimerization domain, indicating that the activated amino acids are converted into D-configuration. The number of modules corresponds to the number of incorporated amino acids while the location of epimerization domains within the peptide synthetase agrees with the position of D-configurated amino acids in the peptide moiety of bacillomycin D. The last domain of this multienzyme system is a thioesterase domain, which is presumably required for release and cyclization of the synthesized lipopeptide. The organisation of this nonribosomal peptide synthetase is similar to that already described for the closely related lipopeptides iturin A and mycosubtilin (see figure 9C) [63, 101]. Analogously bacillomycin D is also synthesized according to the multicarrier thiotemplate mechanism (see 1.3.2.1).

↓71

Furthermore, the adenylation domains of bacillomycin D were compared to those of iturin A and mycosubtilin (Table 10). In the first case, more than 97% amino acid homology was observed within the first three modules of these synthetases, whereas in the second case the homology was less pronounced (>70%). However, homologies were lower within the adenylation domains responsible for activation of amino acids 4 to 7. This correlates well with the sequence variability between these antibiotics at amino acid positions 4 to 7, as shown in figure 8. The highest homology in this region was obtained for bmy_C1 and myc_C1 (81.4%), which both activate the amino acid serine in the sixth module. Furthermore, comparison of the 10 selectivity-conferring amino acid residues of adenylation domains (see introduction) revealed that Pro, Glu, Ser and Thr are activated by the last four modules, i.e. B3, B4, C1 and C2 respectively.

Table 10: Homologies and selectivity-conferring code of amino acid-specific adenylation domains (A-domains) of the bacillomycin D operon compared to the respective A domains extracted from the iturin A and mycosubtilin gene clusters

Position of selectivity conferring amino acids 1

A-domain

amino acid 2

Identity 3

235

236

239

278

299

301

322

330

331

517

bmy_A1_Asn

Asn (1)

100%

D

L

T

K

I

G

E

V

G

K

itu_A1_Asn

Asn (1)

98.6 %

D

L

T

K

I

G

E

V

G

K

myc_A1_Asn

Asn (1)

80.5 %

D

L

T

K

I

G

E

V

G

K

BacC, TycC4

Asn

D

L

T

K

I

G

E

V

G

K

bmy_B1_Tyr

Tyr (2)

100%

D

A

L

S

V

G

E

V

V

K

itu_B1_Tyr

Tyr (2)

99.5 %

D

A

L

S

V

G

E

V

V

K

myc_B1_Tyr

Tyr (2)

85.2 %

D

A

L

S

V

G

E

V

V

K

TycB, TycC1

Tyr

D

A

L

V

T

G

A

V

V

K

bmy_B2_Asn

Asn (3)

100%

D

L

T

K

I

G

E

V

G

K

itu_B2_Asn

Asn (3)

97.7 %

D

L

T

K

I

G

E

V

G

K

myc_B2_Asn

Asn (3)

80.1 %

D

L

T

K

I

G

E

V

G

K

BacC, TycC4

Asn

D

L

T

K

I

G

E

V

G

K

bmy_B3_Pro

Pro (4)

100%

D

V

Q

F

I

A

H

V

V

K

             

myc_B4_Pro

Pro (5)

44.8 %

D

V

Q

F

I

A

H

V

V

K

             

itu_B4_Pro

Pro (5)

42.7 %

D

V

Q

F

I

A

H

V

V

K

             

Pps44

Pro

D

V

Q

F

I

A

H

V

V

K

             

bmy_B4_Glu

Glu (5)

100%

D

A

K

D

L

G

V

V

D

K

             

myc_B3_Gln

Gln (4)

59.8 %

D

A

Q

D

L

G

V

V

D

K

             

itu_B3_Gln

Gln(4)

58.2 %

D

A

Q

D

L

G

V

V

D

K

             

SrfAA4

Glu

D

A

K

D

L

G

V

V

D

K

             

bmy_C1_Ser

Ser (6)

100%

D

V

W

H

F

S

L

I

D

K

             

Myc_C1_Ser

Ser (6)

81.4%

D

V

W

H

F

S

L

I

D

K

             

Itu_A_C2

Ser (7)

72.4%

D

V

W

H

F

S

L

I

D

K

             

EntF, CdaI4

Ser

D

V

W

H

F

S

L

I

D

K

             

itu_C1_Asn

Asn (6)

43.8%

D

L

T

K

I

G

E

V

G

K

             

bmyC2_Thr

Thr (7)

100%

D

F

W

N

I

G

M

V

H

K

             

FenD, Pps2, PvD4

Thr

D

F

W

N

I

G

M

V

H

K

             

A_C2

Ser (7)

50.2%

D

V

W

H

F

S

L

I

D

K

             

mycC2_Asn

Asn(7)

47.6%

D

L

T

K

I

G

E

V

G

K

             

1As determined by [53]. Domains and conserved residues lining the substrate-binding pockets of adenylation domains of assigned functions are indicated in boldface.2The positions of the activated amino acid within the respective lipopeptides are given in parentheses.3This stands for the overall homology of the whole adenylation domain, about 440 amino acids, compared to the respective domain of the bmy operon. 4Domains and residues lining the substrate-binding pocket as described by [54]

Interestingly, BmyA displays a remarkable complexity, similar to MycA and ItuA. The first amino acid module present in these three nonribosomal peptide synthetases is preceded by several domains with homology to proteins involved in the synthesis of fatty acids and polyketides. Four different domains could be distinguished (Fig. 20). In BmyA, the first domain (AL) shows high similarity to long-chain fatty acid CoA-ligases as well as 98% and 85% homology with the corresponding domains of ItuA and MycA, respectively. Furthermore, two domains similar to acyl carrier proteins (ACP) were recognized as well as one similar to β-ketoacyl synthetases (KS). Finally, one domain homologous to glutamate-1-semialdehyde aminotransferase (AMT) was detected. These domains presumambly play a role in the incorporation of the β-amino fatty acid into the peptide moiety [63]. The condensation domain lying directly upstream of the first adenylation domain in BmyA, responsible for the activation and incorporation of Asn, probably catalyzes the transfer of the β-amino fatty acid to the first amino acid.

↓72

According to the colinearity rule, arrangement of modules within a peptide synthetase determines the order of incorporation of specific amino acids in the peptide moiety. As the multicarrier thiotemplate mechanism proposes, elongation of the peptide occurs stepwise from the N to C end. However, in the case of bacillomycin D there was no experimental evidence proving the consequent activation and incorporation of the seven amino acids in the peptide chain. Therefore in order to verify our assumptions about the biosynthetic pathway of bacillomycin D six mutants were created by disrupting one by one the last six modules of the nonribosomal peptide synthetase. Thereby the multienzyme system was silenced at different points after the incorporation of a new amino acid resulting in intermediate products which reflect the stepwise elongation of the peptide. By identifying the intermediate elongation variants that were produced as the peptide moiety groew, it would be possible to monitor biosynthesis of bacillomycin D.

For this purpose, a chloramphenicol cassette was integrated via double-crossover recombination at the beginning of each adenylation domain (bmy_B1, bmy_B2, bmy_B3, bmy_B4, bmy_C1 and bmy_C2) resulting in the bacillomycin D deficient mutant strains AK15, AK39, AK40, AK41, AK42 and AK43 respectively (Fig. 20, Table 3). The mutant strains were grown in Landy medium at 37°C for 24 hours under vigorous shaking. Culture filtrates and sonificated cell extracts from them were prepared and were subsequently analysed by MALDI-TOF MS. The peaks of the expected products, according to the multicarrier thiotemplate mechanism, were not detected in the spectra obtained from the culture filtrates. This probably means that the intermediates remained attached to the multienzyme system and were not secreted from the cells; only the cyclic compound could be exported from the cell. Having this concept in our mind, we sought for the expected peaks in the spectra obtained from cell extracts. Even in these spectra, those peaks had very low intensity and in most cases were hardly distinguished from the background. This result indicates that the elongation variants were tightly attached to the complex and could be only partially detached from it by sonification (see also Discussion).

Figure 20: Schematic representation of the bacillomycin D operon in B. amyloliquefaciens FZB42.

The operon comprises of four ORFs bmyD, bmyA, bmyB and bmyC; their sizes are given in kb. The number of modules within each protein is indicated in parentheses. A more schematic overview of the four proteins depicts the exact location of the seven modules (A1, B1 etc). In parallel, the domains organisation within the modules is demonstrated. The activated amino acids are depicted within the adenylation domains while their configuration is presented under the respective domains. The arrows indicate the position where the chloramphenicol cassette was introduced within the bmy operon in order to construct strains that produce only intermediate products of bacillomycin D; the names of the obtained strains are also noted. AL, acyl coenzyme A ligase domain; ACP, acyl carrier protein domain; KS, β-ketoacyl synthetase domain; AMT, aminotransferase domain.

Regulation of bacillomycin D production

5'-deletion analysis of the bmy promoter region

Determination of bmy expression in B. subtilis MO1099

↓73

In order to monitor the transcriptional regulation of bacillomycin D of B. amyloliquefaciens FZB42, four reporter fusions of the postulated bmy promoter region (the upstream region of the first gene of the bmy operon, bmyD) to lacZ were generated. A series of nested fragments with a common downstream end (by the 42nd codon of BmyD) and variable upstream ends (400, 183, 120 and 30 bps upstream of the translational start) were amplified by PCR, using primers bmyD1 to bmyD5 (Fig. 21; see also Table 4). The obtained products carried suitable restriction sites (EcoRI and BamHI; embedded on the primers) in order to be cloned into pDG268, a plasmid extensively used for constructing transcriptional reporter fusions in B. subtilis that can be later integrated at the amyE locus of its chromosome [192]. The new pDG268 derivatives (pAK5 to pAK8; see also Table 2) were subsequently used for integrating our series of 5'-deletion bmy promoter variants into the chromosome of B. subtilis MO1099 as single-copies, via double-crossover recombination. The correct chromosomal integration of the transcriptional fusions was verified by Southern hybridization. The new strains contained decreasing lengths of the bmy promoter region fused to lacZ and were named AK4, AK5, AK6, and AK7 (Fig. 21; see also Table 3).

Figure 21: Schematic representation of the 5'-deletion analysis conducted for the bmy promoter region.

The dark arrows indicate the primers used for generating bmyD::lacZ reporter fusions (bmyD1-D5). The 5' and 3' end termini of bmyD promoter regions are indicated by their nucleotide position relative to the translational start. The derivative strains of B. subtilis MO1099 and B. amyloliquefaciens FZB42 that carry the respective fusions are presented on the right side of the scheme (Strains AK16-17 are isogenic to AK9-10 but have a different antibiotic cassette, see Table 3).

The expression of the bmyD::lacZ reporter fusions was determined throughout the growth cycle. All four strains were silent during logarithmic phase, indicating that the promoter(s) of bacillomycin D is not active at this time point (Fig. 22A), in total agreement with the late production of the lipopeptide observed by the MALDI-TOF MS analysis (Table 9). Upon entry into stationary phase, the transcriptional activity of AK4, AK5 and AK6 increased and reached its maximum after 3 hours. On the contrary AK7 remained silent and did not show any β-galactosidase activity all along the growth curve, in consistence with its white colour on LB agar plates containing 40µg/ml X-Gal (all three other strains appeared blue on plates). This result was partially expected, as the DNA promoter region contained in the smallest fusion (AK7) reached only up to 19bp upstream of the potential Shine-Dalgarno site, and therefore could not include an entire promoter site. Furthermore, it is apparent that the smallest DNA fragment containing an intact promoter is that encoded in AK6 (reaching up to 120bp upstream of the bmyD coding region). The fact that the fusions of AK6 and AK5 showed no difference in their expression pattern throughout the whole growth cycle suggested that no additional trans-activating factor binds to the region between –183 and –120bp (in respect with the bmyD translational start). On the other hand AK4 exhibited slightly but reproducibly higher activity than the other two strains (AK6 and AK5) during stationary phase (about 25%; Fig. 22A). This means that the bmyD promoter region between –400bp and –183bp (in respect with the bmyD translational start), harboured only by AK4 (see Fig. 21), possibly carries additional cis-activating elements. The question whether these elements code for a transcriptional activator’s binding-site or for an additional promoter was addressed later.

Determination of bmy expression in B. amyloliquefaciens FZB42

↓74

To monitor the expression pattern of bacillomycin D in its natural environment, plasmids pAK5 to pAK8 (derivatives of pDG268, carrying different 5'-end deletions of the bmy promoter region; see previous paragraph) were attempted to be integrated at the chromosome of B. amyloliquefaciens FZB42. Unfortunately, pDG268 carries parts of the amyE gene of B. subtilis (used for the double-crossover recombination of the fusion to the chromosome) that show relatively low homology (less than 80%) to their corresponding regions of B. amyloliquefaciens FZB42. Therefore, our initial attempts to obtain single-copy bmyD::lacZ fusions as part of the chromosome of B. amyloliquefaciens FZB42 were unsuccessful. In order to overcome this problem, a new plasmid, pAK9, was constructed by replacing the amylase sequences from pDG268 with the respective sequences of B. amyloliquefaciens FZB42. All our bmyD::lacZ transcriptional fusions were further cloned to the new vector and then successfully integrated at the amyE locus of B. amyloliquefaciens FZB42. Therefore the four new strains, AK9, AK10, AK11 and AK12, carry decreasing sizes of the bmy promoter region, in complete analogy to strains AK4-AK7 (note also that strains AK16-17 are isogenic to AK9-10, but have a different antibiotic cassette; see also Table 3. Exchange of the chloramphenicol cassette to a kanamycin one was performed using the marker exchange plasmid pECE73; [193].

The β-galactosidase activity of the four new strains was also examined throughout the growth cycle. The overall expression pattern of the various fusions was quite similar to that observed in the background of B. subtilis MO1099 (compare Fig. 22.A and B), and therefore most of the conclusions drawn in the previous section are also valid here. In other words bmy was only expressed during stationary phase, the entire core promoter was encoded within the first 120bp upstream of the bmyD start codon, and the region directly upstream of that (between –183 and –120) did not play any role in bmy expression. However, the strain carrying the longest upstream promoter region, AK9, exhibited 4-5-fold higher β-galactosidase activity in middle stationary phase (3-4 h after entering stationary phase when the β-galactosidase levels have reached their plateau; Fig. 22B) than the strains carrying the shorter fusions (AK10 and AK11). This difference in the expression levels between AK9 and AK10/AK11 is considerably higher than that observed for the corresponding strains of B. subtilis MO1099 (Fig. 22A), underlining thus the importance of this DNA upstream region in the full transcriptional activation of the bmy operon in its natural environment.

Two straight forward explanations can be provided for the different influence of the far upstream DNA region (between bps –400 and –183) on bmy expression in the two Bacilli strains: the cis-acting element situated at this region is optimally bound i) by a regulator (transcriptional factor or sigma factor) only present in B. amyloliquefaciens FZB42 (in B. subtilis a regulator of the same family only weakly recognises these sequences and offers basal levels of activation) or ii) by a regulator that is significantly less expressed in B. subtilis MO1099. No matter of the nature of this regulator, transcriptional analysis of the bmy operon proceeded further in its natural environment (B. amyloliquefaciens FZB42), despite the practical disadvantages that such a decision had (B. amyloliquefaciens FZB42 is more difficult to genetically manipulate and any regulatory mutant to be tested has to be de novo constructed), so that important regulatory elements of bacillomycin D expression were not to be missed or underestimated.

↓75

Figure 22: Expression of bmyD::lacZ fusions carrying different 5'-deletions of the region upstream of bmyD.

The expression of a series of transcriptional fusions of the bmy operon’s promoter region to lacZ (see also Fig. 21) was monitored both in B. subtilis MO1099 (panel A) and in B. amyloliquefaciens FZB42 (panel B). Strains harbouring single copies of the 5'-deletion bmy promoter variants were grown in Difco medium at 37°C and optical densities (closed symbols) and specific β-galactosidase activities (in Miller Units; open symbols) were determined along the growth curve. The expression patterns shown here represent the average of more than three independent experiments. Squares, strains AK4/AK9, carrying the longest promoter region of bmy (-400, +126 relative to the translational start of bmyD); diamonds, strains AK5/AK10, carrying the bmy promoter region between -183 and +126; triangles, strains AK6/AK11, harbouring the bmy promoter region between -120 and +126; circles, strains AK7/AK12, containing the shortest bmy promoter region between -30 and +126.

DegQ is partially responsible for the differences in bmy expression in B. amyloliquefaciens FZB42 and B. subtilis MO1099

It has been demonstrated that the horizontal transfer of functional gene clusters coding for peptide antibiotics and found in natural Bacilli isolates, in the chromosome of the domesticated B. subtilis 168 requires additional steps for the conversion of the latter into an antibiotic producer strain [172, 173]. Firstly the introduction of a functional sfp is absolutely necessary for peptide antibiotic production by B. subtilis 168, which carries a frame-shift mutation in this gene (see also introduction). Secondly it has been exhibited that increased expression of the pleiotropic regulator DegQ in B. subtilis 168 [224] enhances the antibiotic production [172, 173]. Interestingly most of the natural Bacilli isolates that express peptide antibiotics show significantly elevated degQ expression compared to that of B. subtilis 168, due to the fact that the degQ promoter has a more σA consensus-like -10 hexamer in those strains ( TA C A C T instead of C A C A C T ) [225]. However whether DegQ directly influences the transcriptional regulation of the antibiotic operons or it controls the expression of a post-transcriptional regulator involved in the antibiotic synthesis (for example Sfp) has not been clarified.

B. amyloliquefaciens FZB42 has the promoter version of degQ that yields higher DegQ cellular levels (data not shown). In contrast to that, B. subtilis MO1099, a derivative of the strain 168, carries the defected degQ promoter version. To test whether DegQ is responsible for the differential bmy expression patterns in its host strain and its B. subtilis counterpart, I constructed a plasmid that carried degQ under the IPTG-inducible Pspac promoter, pAK64 (Table 2), using the replicated vector in B. subtilis, pDG148 [191]. pAK64 was subsequently transformed into AK4 and AK5 (the strains harbouring the two longer bmy promoter regions fused to lacZ; see also Fig. 21) and the expression of the reporter fusions was monitored along the growth curve for the following cases: i) AK4 and AK5 without the plasmid (control), ii) AK4 and AK5 with the plasmid uninduced and iii) AK4 and AK5 with the plasmid induced at OD600~0.7. As seen in Fig. 23 the presence of uninduced pAK64 hardly changed bmy expression from the two strains. Upon induction of pAK64 though, a significant 2.5-fold increase could be observed in the activity of AK4 (Fig. 23), a strain that carries the whole upstream promoter region of bmyD. In other words, the B. subtilis strains carrying the two longer bmy promoter fusions, AK4 and AK5, exhibit a more pronounced difference in their activity when DegQ is expressed in higher levels from a plasmid. The magnitude of the difference does not match that observed for the corresponding B. amyloliquefaciens FZB42 strains (AK9 and AK10; Fig 22B), but with higher amounts of DegQ present in the cell, the pattern of bmy expression in B. subtilis approximates more the pattern observed in B. amyloliquefaciens FZB42. Since DegQ levels in B. amyloliquefaciens FZB42 are probably considerably higher than those in B. subtilis MO1099 (see above), DegQ accounts, at least partially, for the different bmy expression in the two host strains. Moreover, this is the first direct evidence that DegQ affects the transcriptional regulation of peptide antibiotics and not a subsequent step of their production. In the case of bacillomycin D the exertion of this effect seems to be dependent on a far upstream region of the promoter. Having in mind that DegQ is not a DNA-binding protein, the effect on bacillomycin D should be mediated in an indirect manner, possibly through modulating the activity of other transcriptional regulator(s) (see also Discussion).

↓76

Figure 23: The effect of DegQ on the expression pattern of bmyD::lacZ fusions in B. subtilis MO1099.

Strains AK4 and AK5, harbouring the longest promoter region of bmy fused to lacZ, were transformed with pAK64 (a replicated plasmid carrying DegQ under an IPTG-inducible promoter). After 3h of growth, pAK64 was induced in two of the cultures by addition of 1mM IPTG. Cells were grown in Difco medium (supplemented with 5 μg/ml kanamycin where applicable) at 37°C and optical densities (closed symbols) and specific β-galactosidase activities (in Miller Units; open symbols) were determined along the growth curve. Squares, strain AK4, carrying the longest promoter region of bmy (-400, +126 relative to the translational start of bmyD), without plasmid; cirles, strain AK5, carrying the bmy promoter region between -183 and +126, without plasmid; diamonds, strain AK60 (AK4+pAK64) with plasmid induced; triangles facing down, strain AK60 with plasmid uninduced; triangles facing up, strain AK61 (AK5+pAK64) with plasmid induced; rectangular triangle, strain AK61 with plasmid uninduced.

Identifying the transcriptional start site of the bmy operon

The transcriptional start site of the bmy operon was determined by primer extension. Total RNA of B. amyloliquefaciens FZB42 was isolated from cultures growing in Difco medium, 2,5-3 hours after their entry into stationary phase, when expression of bmyD promoter peaked, as shown by the lacZ reporter fusions (Fig. 22).

Primer extension using the radiolabelled primer rev1 (binding within the coding region of bmyD; see also Table 4) revealed two overlapping transcriptional start sites for bmyD: an adenine (A) and a thymine (T) nucleotide located 58 and 57 bp upstream of the gene’s initiation codon respectively (see Figs 24 and 25). Sequences resembling the consensus of the -10 and -35 elements of the housekeeping sigma factor, σA, were found upstream of the mapped transcriptional start site (Fig. 24). In particular, the -35 (TATACA) and -10 (TAGGAT) hexamers identified upstream of bmyD carry 4/6 matches to the corresponding consensus sequences of σA [226]. The spacer between them is 18 bp long and within it also lays an extended -10 region, directly upstream of the –10 hexamer, i.e. CATGc (the bold faced nucleotides match the consensus, TRTGn; [227, 228]. Therefore the promoter of bmyD (Pbmy) seems to be recognised and utilised by the vegetative sigma factor, despite the fact that bmy shows stationary-phase induced expression. Consistently the entire core promoter is encompassed within the first 75 bp upstream of the gene’s translational start, as predicted by the deletion analysis of the reporter fusions (see Figs. 22 and 25).

↓77

Figure 24: Mapping of the transcriptional start of the bmy operon by primer extension analysis.

Total RNA was isolated from B. amyloliquefaciens FZB42 cells grown in Difco medium, 2.5-3 hours after their entry into stationary phase. Primer extension was performed using the 5'-end 32P-labelled primer rev1. The first four lanes result from dideoxynucleotide sequencing reactions using the same primer. The positions of the transcripts corresponding to the bmyD transcriptional starts are indicated by arrows. The respective DNA sequences are shown on the right of the picture. Transcriptional starts are highlighted in bold face whereas the -10 hexamer is underlined.

Extensive attempts to identify further upstream-situated promoters that would account for the differences in gene expression between strains AK9 and AK10 failed (data not shown). Consequently, a transcriptional regulator, binding at the region between -400 and -183 bp upstream of the gene’s start codon, is partially responsible for the stimulation of the bmy operon’s expression during stationary phase.

Figure 25: Nucleotide sequence of the bmyD promoter region.

The positions of the two adjacent transcriptional starts (bold face, +1), the translational start (bold face, Met), the putative ribosome binding site (underlined), the -35 and -10 hexamers (boxes) and the extended -10 region (double underlined) are indicated. The annealing site (5'-end) of the oligonucleotides used for primer extension (rev1) and for construction of reporter fusions (bmyD1-D5) are shown with arrows. Site I and II (shaded) represent the two DegU binding sites (see 3.4.4.3). The degenerate forms of the motif AGAA-N11-TTCAG, which was proposed by Dartois et al (1998) as the recognition site for DegU, are indicated by boxes within the shaded sites.

Global regulators control the production of bacillomycin D

Effect of global regulators on the activity of bmyD::lacZ reporter fusions

↓78

In order to obtain further information concerning the transcriptional regulation of bacillomycin D, a series of mutations were introduced in transcriptional regulatory proteins and sigma factors of the bacterium. Some of the genes mutated are also found in the genome of B. subtilis, whereas others are novel members of the genome of B. amyloliquefaciens FZB42. The several mutations were then crossed over into strains AK9 and AK10 (AK11 was excluded from these experiments, since it displays the same pattern of gene expression as AK10). Thereby, it was possible to monitor the behaviour of the mutant strains by calculating their β-galactosidase activities along the growth curve and comparing them to that of the wild type strain.

Using this approach, several players involved in the regulation of the bmy operon were identified. Two two-component response regulator proteins were found to be essential for full activation of the operon: DegU, known to control the expression of degradative enzymes [229, 230] and to be involved in the initiation of competence [231] and ComA, a regulator of late-competence genes [232] and surfactin production [112]. Moreover, the sporulation sigma factor σH [233] regulates expression of bacillomycin D.

The effects of DegU, ComA and σH on the activity of reporter fusions are analytically presented in Figure 26. Inactivation of the genes coding for each of the above proteins resulted in severely impaired promoter activity during stationary phase, especially in strains harbouring the whole promoter region (AK9 derivatives). bmy expression was 3-4-fold lower in the degU, comA and sigH mutant derivatives of AK9 (AK32, AK22 and AK52 respectively) compared to the parental strain AK9 (Fig. 26), and similar to that of wild type strain AK10, carrying the shorter promoter region (its 5'-end is deleted 183 bp upstream of the bmyD translational start). Furthermore, in the absence of DegU, ComA or σH, the activity of AK10 derivatives (AK33, AK23 and AK53 respectively) was also reduced in comparison to that of the parental strain AK10, albeit to a much smaller extent (Fig. 26).

↓79

Therefore DegU, ComA and σH are all required for the full activation of the bmyD promoter and their effects seem to be mostly exerted through the DNA region located between -400 and -183 bps upstream of the gene’s translational start. Despite the severity of the reported effects, the expression of bmy is not completely silenced by any of those mutations (on the contrary to the promoter silencing observed with strain AK12; (Fig. 22B), insinuating that the σA-dependent promoter retains both stationary phase induction and basic levels of bmy expression without any of these regulators (see also Discussion).

Figure 26: Effects of ComA, DegU and σH on the expression of the various bmyD::lacZ fusions.

Strains AK9 and AK10, carrying the two bmyD::lacZ fusions with the longest upstream promoter region and the highest activity (see also Fig. 21, 22), and their degU (A), comA (B) and sigH (C) mutant derivatives were grown in Difco medium at 37°C and optical densities (closed symbols) and specific β-galactosidase activities (in Miller Units; open symbols) were determined along the growth curve. In each panel the expression of the fusions in the wild-type background is provided for direct comparison to that of the mutants. Squares, AK9; diamonds, AK10; triangles, mutant derivatives of AK9 (AK32 in A, AK22 in B, AK52 in C); circles, mutant derivatives of AK10 (AK33 in A, AK23 in B, AK53 in C).

In addition, the effect of σH cannot be associated with the presence of a second upstream promoter, since primer extension experiments ruled out such as a scenario (see above). Consistently, no similarity with the promoter consensus sequence of σH (AGGANNT-15-17bp-GAAT; [234] could be found in the entire bmy promoter region. Thus, the effect of σH exerted on Pbmy is indirect.

↓80

Deletion mutants of several other transcriptional regulators and sigma factors did not significantly influence the expression of the bmy operon, as shown by monitoring the activities of the respective reporter strains along the growth curve (data not shown). The genes/operons tested were: an uncharacterised two-component system RBAM01839-RBAM01840 in the close proximity of bmy operon, present only in FZB42 (the homology to other known two-component systems of B. subtilis is very low; higher scores with yvrGH; [235]; yerPO, YerP belongs to a resistance-nodulation-cell division (RND) family proteins and is involved in surfactin immunity/production [170] whereas YerO is a putative transcriptional regulator encoded adjacent to it; spaR, encoding a response regulator that activates subtilin production and immunity [33], present in B. amyloliquefaciens FZB42 but not in B. subtilis MO1099; sigW, encoding an extracytoplasmic sigma factor involved in antibiotic resistance [236]; and aat, encoding a putative transcriptional regulator direct downstream of srfAD, that is only present in B. amyloliquefaciens FZB42 (see also Fig. 15).

Effects of degU, comA, sigB and sigH mutations on transcriptional initiation by the identified promoter of bmy operon (Pbmy)

Reporter fusions showed that DegU, ComA and σH positively regulate transcription of the bmy operon in B. amyloliquefaciens FZB42. To obtain further proof, the effects of these mutations on the activity of Pbmy were examined by primer extension. In addition the role of σB, the general stress sigma factor in Bacilli [237, 238], was also investigated. Total RNA was extracted from wild type or mutant cells that had reached middle stationary phase, and primer extension was performed as described earlier (see 3.4.2 and materials and methods).

A unique transcriptional start site was identified for the bmy operon in both the wild type and the four mutant strains, and as expected coincided with the one reported in the previous section (Fig. 24). However the intensity of the obtained transcripts varied enormously at the different genetic backgrounds; the wild type strain produced a strong and clearly distinguishable transcript whereas degU, comA, sigH and sigB mutant strains gave only weak but reproducible signals (Fig. 27), a sign of decreased Pbmy promoter activity. This finding is in perfect agreement with our previous results (Fig. 26), identifying DegU, ComA and σH as positive regulators of Pbmy in stationary phase. It also verifies our previous statement that cells lacking those regulators are still able to show basal expression of the σA-dependent Pbmy. Moreover, σB serves as a new positive transcriptional regulator of the bmy operon; its effects were of similar magnitude to those of the other three regulators (Fig. 27), and most probably exerted through an indirect mechanism since there are no sequences in the bmyD promoter region that apparently resemble the promoter consensus sequence of σB (GTTT-15-17bp-GGGWAW, where W stands for A/T; [239] see also section 3.4.5).

↓81

Figure 27: Effects of degU, comA, sigB and sigH mutations on the activity of the bmy operon promoter (Pbmy).

Cells were grown in Difco medium and 2.5-3 hours after entry into stationary phase, total RNA was extracted from B. amyloliquefaciens FZB42 and its mutant degU (TF1), comA (CH23), sigB (CH33) and sigH (AK50) derivative strains. Primer extension was performed using the same amount of total RNA (40 μg) and the 5'-end 32P radiolabelled primer rev1. Sequencing ladders were generated with the same primer. Intensities of the obtained transcripts indicate the effects exerted by the respective mutations on the activity of the Pbmy promoter. 1, B. amyloliquefaciens FZB42; 2, TF1; 3, CH23; 4, AK50; 5, CH33; +1 and the arrow both indicate the transcriptional start.

Several more transcriptional regulators and sigma factors were deleted or disrupted and their role on bmy operon’s transcription was further assessed by primer extension analysis: sig01, an extracytoplasmic sigma factor identified for the first time in FZB42; sigD, the Bacillus sigma factor involved in chemotaxis, flagella synthesis, motility [240]; three more extracytoplasmic sigma factors, encoded by sigX, sigV, sigW and different combinations of them [236, 241, 242]; and codY, encoding a global transcriptional regulator in B. subtilis [243] and a direct repressor of surfactin expression [169]. All these mutant strains produced an equally strong transcript signal as that of the wild type strain, indicating that they do not play a role in transcription of the bmy operon (data not shown).

MALDI-TOF MS analysis of B. amyloliquefaciens FZB42 strains deficient of global regulators that are involved in transcription of the bmy operon; DegU has a post-transcriptional effect on bacillomycin D production

In order to decipher how the reported effects of DegU, ComA, σB and σH on bmy transcription reflect to the end-production of bacillomycin D, we decided to monitor the antibiotic’s synthesis in the different mutant strains by mass spectrometric analysis. MALDI-TOF MS was performed using culture filtrates of the degU, comA, sigH and sigB mutant derivatives of B. amyloliquefaciens FZB42 (grown with aeration for 24h in Landy medium), and the obtained spectra were compared to that of the wild type, in order to evaluate the relative production levels of bacillomycin D in the different genetic backgrounds. Even though this type of analysis does not provide accurate quantitative results, relative production of bacillomycin D can be roughly estimated by comparing the intensity values of the peaks that correspond to bacillomycin D with the intensity values of the surrounding peaks (belonging to other peptide products) in the wild type and mutant strains.

↓82

It has been previously shown that mutations in degU, comA, sigH and sigB severely impair but do not silence transcription of the bmy operon. Consistently, bacillomycin D production was defected but not entirely blocked in the comA, sigH and sigB mutant strains, since the intensities of the peaks reflecting the presence of bacillomycin D were considerably weaker in the spectra of the mutant strains compared to the spectrum of the wild type strain (Fig. 28). However, no bacillomycin D could be detected from a culture filtrate stemming from the degU strain (Fig. 28.E and F), suggesting that the role of DegU in bacillomycin D production goes beyond than the mere activation of the Pbmy promoter.

Figure 28: MALDI-TOF MS analysis of comA, sigB, sigH and degU mutant strains. The absence of DegU deprives the cell of bacillomycin D production (performed in collaboration with Dr. J. Vater and Xiao-Hua Chen).

Spectra of culture filtrate extracts prepared from B. amyloliquefaciens FZB42 and its mutant derivatives, after 24 hours growth in Landy medium. A. In the wild type strain, bacillomycin D (m/z 1053.3, 1097.4) is the antibiotic produced in the highest amounts under these conditions (see also 3.3), judged from the intensities of its peaks, whereas the siderophore bacillibactin (m/z 833, 905 and 921; see also 3.5) and surfactin (m/z 1044) follow in production scale. The peaks corresponding to fengycin production (m/z 1463.7, 1501.6) are the ones with the lowest intensity. For more details concerning the exact number of peaks that define the presence of an antibiotic in the external milieu of B. amyloliquefaciens FZB42 see also Table 8. B-D. The production pattern of the four compounds changes in comA (B), sigB (C) and sigH (D)mutants. In all three cases the levels of bacillomycin D are relatively low in comparison with the other the antibiotics and in most cases hardly exceed them. In addition the intensities of the peaks corresponding to fengycin seem increased in all three mutants, rendering thus fengycin as one of the major antibiotics produced by B. amyloliquefaciens FZB42 in these genetic backgrounds (see also 3.5 and Table 11). As expected, the comA mutant strain (B) is deficient of surfactin production [161, 162]. E-F Bacillomycin D is absent in the degU mutant strain. A zoomed in version of the panel E spectrum is presented in panel F, where only the peaks attributed to surfactin are apparent while those belonging to bacillomycin D (m/z 1053.7, 1067.7, 1069.7) are absent.

↓83

Since in the absence of all four regulatory proteins, i.e. DegU, ComA, σB and σH, Pbmy retained its basal expression (Figs 26 and 27), but only in the absence of DegU the production of bacillomycin D was completely abolished, DegU seems to have an additional “control-point” on bmy regulation, at a post-transcriptional level. To test whether the post-transcriptional effect of DegU was associated with the export of bacillomycin D to the extracellular milieu, sonificated cell extracts of the degU strain (grown with aeration for 24h in Landy medium) were also analysed by MALDI-TOF MS. No bacillomycin D was apparent in the cell extracts too (data not shown), suggesting that DegU influences both transcription of the bmy operon, and also a step post to transcription of bmy but prior to its export to the surrounding environment.

DegU directly binds to the bacillomycin D promoter

DegU was shown to be essential for bacillomycin D production, both at transcriptional and post-transcriptional level. In order to fully decipher its role on the activation of Pbmy, the response regulator was purified as a C-terminal His6-Taq fusion protein [231], and was further used in EMSA and DNAse I experiments. Overexpression and purification of 6xHis-tagged DegU have been already described in detail in material and methods. Analysis by SDS-PAGE (Fig. 29) and Western blot revealed that the purification of native 6xHis-tagged DegU (35 kDa) protein was successful.

Figure 29: Overexpression and purification of the 6xHis-tagged DegU

Overexpression and purification of the 6xHis-tagged DegU protein was performed as described in detail in chapter 2.5.5. Samples collected from different steps of the purification process were analyzed by SDS-PAGE electrophoresis. 6xHis-tagged DegU was the only purified protein. FT, sample from flow through; W, sample from a wash step; E1 to E7, samples from elution steps 1 to 7; M, Prestained Protein Ladder (Fermentas), bands from bottom to top 25, 35, 45, 55, 70, 100, 130, 170 kDa.

EMSA shows that DegU is a direct activator of the bmy promoter

↓84

Our genetic analyses identified a positive role of the response regulator DegU in bmy transcription. To examine whether the bmy promoter is a direct target for DegU, gel retardation mobility shift assays (EMSA) of a 450 bp DNA fragment, harbouring regions between -342 and +108 relative to the transcriptional start (or -400 to +50 relative to the translational start) were performed using increasing concentrations of the response regulator. As seen in Figure 30A, the bmyD promoter indeed contains specific binding sites for DegU. The DNA fragment was already shifted with 0,2 and 0,4 µM unphosphorylated DegU, in an analogous manner to the DegU binding at the comK promoter [231]. Mobility of the 450 bp DNA fragment changed dramatically upon incubation with higher amounts of the protein (0,8 and 1,6 µM), indicating the presence of more than one DegU binding-sites in the promoter region of bmy. The observed shifts are DegU-specific since no migration of the DNA fragment was observed upon incubation with the same amounts of BSA (Fig. 30A).

Since our reporter fusions’ results indicated that DegU influences both expression of the shorter and the longer versions of the bmyD::lacZ fusions (present in strains AK9 and AK10; Fig. 22), we sought to narrow down the regions that DegU binds. In order to accomplish this, gel retardation mobility shift assays were also performed with two smaller DNA fragments. Fragment D1 encompassed the region between -342 to –126 bp relative to the transcriptional start (or -400 to –184 bp relative to the translational start of bmyD), present only in AK9, while fragment D2 encompassed the DNA region between -125 to +108 bp relative to the transcriptional start (or -183 to +50 bp relative to the translational start of bmyD), present in both strains, AK9 and AK10. EMSA experiments with fragments D1, D2 and increasing concentrations of the response regulator revealed that unphosphorylated DegU directly bound to both DNA regions, with similar affinities (Fig. 30B). It is also apparent that higher amounts of DegU (0,4 or 0,8 μM) lead to a supershift of the two DNA fragments (similar to that observed in the EMSA experiments with the DNA fragment encompassing the whole promoter region; Fig. 30A). This indicated that the initial DegU binding to thebmy promoter region might trigger the co-operative binding of more DegU molecules to the promoter and/or change the promoter’s architecture (see also Discussion).

Figure 30: Gel retardation mobility shift assays (EMSA) of the bmyD promoter region

A) Gel retardation mobility shift assays of a 32P-labeled bmyD promoter fragment using increasing concentrations of DegU, as indicated on the top of the gel. The DNA fragment used harbours regions between -342 and +108 bp, relative to the transcriptional start. The incubation of the bmy promoter region with the BSA protein (same amount as that of the highest DegU concentration used) did not cause migration of the DNA fragment.
B) Gel retardation mobility shift assays using two smaller DNA fragments of the bmyD promoter region, D1 and D2, ranging from -342 to -126 bp and -125 to +108 bp respectively (relative to the transcriptional start). Gel mobility shifts were performed using 30ng of DNA fragments D1 (left gel) and D2 (right gel) and increasing concentrations of DegU, indicated on the top of the gels. A control using BSA protein was also performed. Visualization was obtained by ethidium bromide staining.

Mapping the location of the DNA-binding sites of DegU on the bmy promoter region

↓85

DNase I footprinting analysis was performed in order to determine the exact location of the DegU binding-sites in the bmy promoter region. DNase I treatment followed incubation of either a linear DNA fragment encompassing the whole bmyD promoter region or the same promoter fragment cloned in a plasmid, and thus supercoiled, with unphosphorylated DegU (0, 0,8, 1,6 μM DegU). Primer extension with two different primers enabled the visualization of both the template and non-template strand (only the data obtained using the supercoiled DNA can be seen in Fig. 31; similar data were obtained with the linear DNA fragment).

The footprinting patterns obtained with the coding and the non-coding strand lead to similar conclusions. In detail, the presence of DegU lead to the protection of a region spanning from -123 to -106 (relative to the transcriptional start) on the top strand, followed by an extended region of hypersensitive sites ranging from -103 to -85 (Fig. 31). Moreover, directly downstream of a series of hypersensitive sites at around -210 of the top strand, is situated a relative protected region from DegU between -201 and -172. Consistently the bottom strand revealed a strongly protected area from -116 to -99 and a relatively more weakly protected region between -198 and -172 upon DegU addition (Fig. 31). Strong hypersensitive sites could be observed in the region between -98 and –66, and at -201 and -203 of the same strand (Fig. 31).

In conclusion, unphosphorylated DegU binds two distinct sites at the promoter of bmy, i.e. Site I (-123 to -99) and Site II (-201 to -172) (see also Fig. 26), inducing bends and local changes in the DNA architecture adjacently to these sites (seen as hypersensitive sites). The two DNA binding-sites bear the A/T-rich signature of DegU recognition-sites (see also Discussion; [244, 245], and their location is also consistent with the data obtained by the gel retardation assays. It is worth mentioning that phosphorylated DegU (after incubation with acetyl phosphate) produced identical DNase I footprints (data not shown).

↓86

Figure 31: DNase I footprinting analysis of DegU at the bmy promoter region

The DNase I digestion patterns were obtained using primer extension in order to be able to monitor both linear and supercoiled DNA (here shown the results with the latter). On the left panel can be seen the footprint pattern of the coding strand in the presence of increasing concentrations of DegU, whereas on the right panel is presented the footprint pattern of the non-coding strand. The protected and hypersensitive regions are marked with bars and arrows respectively. The sequence reactions of the appropriate DNA strand were used as size markers.

The effect of DegU on bmy transcription is epistatic to that of DegQ

It has been previously observed that DegQ exerts a positive effect on bacillomycin D expression (see Fig. 23 and corresponding text), similarly to its role in the production of other peptide antibiotics such as iturin A and plipastatin [172, 173]. This effect is most probably mediated in an indirect manner (no DNA-binding ability is predicted for DegQ), requiring the existence of a far upstream region in the case of the Pbmy promoter region (between –400 and –183 bps in respect with the translational start of the first gene of the operon, bmyD). Since DegU has a DNA binding-site in this region (see Figs. 25 and 31), a plausible scenario is that the effect of DegQ is mediated through DegU. In addition DegU is known to stimulate the expression of degQ [171], and an alternative scenario would be that the transcriptional effects of DegU on bmy expression are indirect and are due to decreased DegQ levels. To test these hypothesises, the plasmid carrying an IPTG inducible copy of degQ, pAK64, was transformed in the degU mutant derivative of B. amyloliquefaciens FZB42. The strain was grown until middle-late exponential phase (OD600~0.7), pAK64 was induced with IPTG, and cells were harvested after 4h for total RNA preparation. The supernatants were lyophilized and further analyzed by MALDI-TOF mass spectrometry (see Fig. 32B). Primer extension analysis revealed that the Pbmy activity was equally low in both the degU mutant and the degU mutant with an overexpressed DegQ, in contrast to the strong activity that Pbmy showed in the wild type background (Fig. 32A). In addition, bacillomycin D was not detected in the spectrum of the degU mutant strain carrying an induced pAK64 (Fig. 32B), similarly to the spectrum obtained by the degU mutant strain (Fig. 28F). This means that the effect of DegU on bmy expression is epistatic to that of DegQ, and that the latter needs the former in order to exert its role.

Figure 32: Increased DegQ cellular levels cannot restore bacillomycin D production in a degU - background

Strain TF1 (degU -) containing pAK64, which carries the degQ gene under an IPTG inducible promoter (strain AK58), was grown in Difco medium, supplemented with 5 μg/ml kanamycin, until mid-exponential phase. The plasmid was then induced with 1mM IPTG, the strain was further grown for 4 hours and the cells were harvested for total RNA extraction and MALDI-TOF MS analysis.
A. Primer extension analysis was performed as described previously (see Material and Methods and Fig. 27). Intensities of the obtained transcripts indicate the effects exerted on Pbmy promoter activity by the respective mutations. 1, B. amyloliquefaciens FZB42; 2, TF1; 3, AK58 (TF1 with pAK64). Induction of DegQ expression cannot alleviate the defect that the absence of DegU imposes on Pbmy activity.
B. MALDI-TOF MS analysis of the culture filtrate extracts of strain AK58. Bacillomycin D is not synthesized (the obtained peaks are only due to surfactin production, see Table 8). The analysis has been done in collaboration with Dr. J. Vater.

↓87

The fact that ComA is also known to positively regulate the expression of degQ and to have an even more pronounced effect than DegU [171], motivated us to clarify whether the effect of ComA on bmy expression is imposed through DegQ. However extensive efforts to transform a wild type B. amyloliquefaciens FZB42 with pAK64 (that subsequently would be modified to a comA - mutant) were unsuccessful. This result deprived us of testing the truth of this hypothesis, but generated interesting implications considering the cellular role of DegQ (see Discussion).

σB mediates its control on Pbmy by indirectly controlling the repression of a novel member of the Rap protein family

The effects of σB and σH on Pbmyare most probably indirect for reasons explicitly stated above (see chapter 3.4.3.2). Since DegU and ComA are also involved in bmy transcription (the former also acts directly; see above), we reasoned that several Rap proteins should be also indirectly involved in the expression of bacillomycin D. Rap proteins extensively regulate the activity of response regulators in B. subtilis, and five of them (out of eleven present in B. subtilis) have been shown to directly inhibit the DNA-binding ability of either ComA or DegU, i.e. RapC [246], RapF [167], RapK [247], RapG [248] and RapH [249]. Most Rap proteins (including all five members named before) come as pair with a small-sized Phr protein, which after its synthesis is excreted from the cell and processed to a signal pentapeptide [250]. The pentapeptide is re-imported to the cell by an oligopeptide permease [251] and inhibits the activity of its cognate Rap protein [166]. The rap-phr gene pairs usually constitute operons, but the expression of phr is often controlled by an extra σH- dependent promoter ([234]; see also discussion). Recently RghR (formerly YvaN) was reported to repress the expression of rapG and rapH and thereby to have a positive effect in the expression of downstream targets of ComA and DegU [249]. Interestingly yvaN had been previously identified as a member of the σB regulon [252]. These data provided possible links-suggestions to the indirect effects of σB and σH on Pbmy that were further addressed.

B. amyloliquefaciens FZB42 possesses 9 putative Rap proteins (carrying a TRP-tetratricopeptide domain), among which only 6 are highly homologous to their B. subtilis counterparts (see also Discussion). Both rap targets of RghR (rapG and rapH)are not conserved in B. amyloliquefaciens FZB42, but tandem RghR binding-sites [249] are predicted in front of one of the novel rap members of the strain, rapX (data not shown). rapX is located at the same chromosomal locus as rapG in B. subtilis, but is encoded in the opposite strand, carries no obvious phr counterpart (in contrast to rapG) and shares only basic homology to rapG (27% in protein level; a percentage similar to that exhibited to other Rap proteins of B. subtilis; see also Discussion).

↓88

Having in mind that σB exerts an indirect positive effect on Pbmy, we hypothesized that this effect can be mediated through a pathway involving RghR and RapX. To test this hypothesis a rapX - and a rapX - sigB - double mutant strain were constructed and assayed together with the wild type and sigB - strains for Pbmy activity by primer extension (Fig. 33). While the absence of σB (theoretically causes a decrease in RghR levels, and thereby an increase in those of RapX) significantly reduced Pbmy activity, the effect was alleviated when rapX was also mutated. This derepression observed verified that σB and RapX work on the same pathway (with RghR -conserved in B. amyloliquefaciens FZB42- as intermediate) to influence the activity of Pbmy. Nevertheless, the Pbmy activity was only slightly higher in the rapX - background, compared to that of the wild type (Fig. 33), indicating that RapX is not substantially repressing DegU or ComA in stationary phase, where σB activates the expression of RghR (see also Discussion and Fig. 39)

.
Figure 33: σB activates expression of Pbmy due to the repression it exerts on a novel Rap protein found in B. amyloliquefaciens FZB42, RapX

Cells were grown in Difco medium and 2.5-3 hours after entry into stationary phase, total RNA was extracted from B. amyloliquefaciens FZB42 and its mutant derivatives. Primer extension analysis wasperformed as described previously in detail (see Materials and Methods and Fig. 27). 1, B. amyloliquefaciens FZB42; 2, AK59 (rapX::Cmr); 3, CH33 (sigB::Emr); 4, AK57 (rapX::Cmr sigB::Emr). The absence of RapX increases Pbmy activity compared to that of the wild type, whereas introducing a rapX mutation on the sigB mutant strain leads to a de-repression of the Pbmy activity. This indicates that the positive effect of σBon bmy expression is mediated through the repression of RapX, which negatively regulates the promoter activity of Pbmy (see also text).

From the Raps known to inhibit the function of DegU or ComA in B. subtilis, onlyRapC and RapF are also encoded in B. amyloliquefaciens FZB42. Their cognate Phrs (PhrC and PhrF) are both controlled by σH in B. subtilis [234] and the σH-dependent promoter sites are conserved in B. amyloliquefaciens FZB42 (data not shown). This insinuates that the indirect effect of σH in bmy expression can be mediated through PhrC and PhrF (see also Discussion and Fig. 39).

Post-transcriptional effects in bacillomycin D production

Sfp and YczE control bacillomycin D production in a post-transcriptional manner

↓89

Nonribosomal peptide synthetases require posttranslational modification to be functionally active, as already mentioned in the introduction. Sfp, the 4'-phosphopantetheinyl transferase, catalyses conversion of the T-domain from its apo- to its holo-form and unblocks synthesis (see 1.3.2.3). As expected, MALDI-TOF MS analysis revealed that disruption of sfp gene in B. amyloliquefaciens FZB42 (strain CH3) resulted in deficiency in both lipopeptide and polyketide synthesis (Fig. 34B) [197].

Furthermore, yczE encodes a predicted membrane protein of unknown function and is located directly downstream of sfp. Disruption of the yczE gene in B. amyloliquefaciens FZB42 shut down the production of bacillomycin D (and polyketides; [197], whereas synthesis of fengycin and surfactin remained unimpaired, as demonstrated by MALDI-TOF MS analysis of culture filtrate extracts from the corresponding strains (Fig. 34C). Since YczE is similar to a membrane protein, its role could be associated with the export of bacillomycin D. If this were the case, the antibiotic would be synthesized but would be entrapped inside the cell. To test this hypothesis, sonificated cell extracts of the yczE mutant strain were analyzed by MALDI-TOF mass spectrometry. Bacillomycin D was not detected inside the cells (data not shown) indicating that YczE is not involved in the antibiotic’s export but rather in its synthesis in a yet unidentified manner.

Figure 34: MALDI-TOF MS analysis of sfp and yczE mutant strains (performed in collaboration with Dr. J. Vater and Xiao-Hua Chen)

Spectra of culture filtrate extracts prepared from strains grown for 24 hours in Landy medium. A) B. amyloliquefaciens FZB42 produces the lipopeptides bacillomycin D (m/z 1053.7, 1097.4), surfactin (m/z 1030) and fengycin (m/z 1463, 1501) (see also Table 8 and Fig. 16). B. Introduction of an sfp mutation in B. amyloliquefaciens FZB42 causes deficiency in lipopeptide synthesis. C. A yczE mutant strain (CH4) produces surfactin (m/z 1030.7, 1058.7) and fengycin (m/z 1463.8, 1485.7) but no bacillomycin D. A zoomed in version of this spectrum is presented in panel D, where one can clearly see the peaks deriving from surfactin production, but not those expected for bacillomycin D production (m/z 1053.7, 1067.7, 1069.7).

↓90

To test whether the effect of YczE on bacillomycin D synthesis was exerted at the transcriptional level, the yczE::Em r was introduced in strains AK9 and AK10 that carry different sized bmyD::lacZ fusions (see also Fig. 22). The expression of the fusions was not influenced by the presence of the yczE mutation (Fig. 35). Therefore YczE does not have an impact on transcription of bmy operon but exerts its effect post-transcriptionally. The same conclusion was reached after checking the Pbmy activity in the yczE background by primer extension analysis (data not shown).

Figure 35: YczE does not influence the expression of the bmy operon

Strains AK9 and AK10, harbouring lacZ fusions of the bmyD upstream region between -400 and +126 bp, and between -183 and +126 bp, respectively (see also Fig. 21), and their yczE mutant derivatives (AK26, AK27), were grown in Difco medium at 37°C and optical densities (closed symbols) and specific β-galactosidase activities (in Miller Units; open symbols) were determined along the growth curve. Promoter activity was not altered by the presence of a yczE mutation. Squares, parental strain AK9; diamonds, parental strain AK10; triangles, AK26, yczE mutant derivative of AK9; cycles, AK27, yczE mutant derivative of AK10.

The post-transcriptional effect of DegU on bmy production is not mediated through YczE

It has already been demonstrated that both DegU and YczE influence bacillomycin D synthesis in a post-transcriptional manner. Since DegU is a response regulator controlling various post-exponential phase responses, it was postulated that it could be involved in transcriptional activation of YczE. Thereby the post-transcriptional effect of DegU on bacillomycin D production could be also mediated through YczE. In order to test this scenario, the transcriptional site of yczE was mapped by primer extension analysis and the promoter activity was measured in different genetic backgrounds; i.e. wild type strain (B. amyloliquefaciens FZB42), degU mutant (TF1) and comA (CH23) mutant strains. The latter strain was used as a control, since ComA was predicted not to affect yczE transcription, (remember that the disruption of comA does not abolish production of bacillomycin D; Fig. 28).

↓91

The primer extension analysis conducted in the wild type strain revealed three possible overlapping transcriptional starts for yczE, but no apparent promoter recognition sequences could be traced in front of them (Fig. 36). On the other hand a weaker transcriptional start is located directly upstream of these three signals (one and half turn of the DNA helix) and just downstream of a perfect extended -10 promoter recognition sequence for the housekeeping sigma factor σA (Fig. 36). We propose that this is the real transcriptional start and the other signals just derive from 5'-end processing of this main mRNA. Moreover, no significant difference in the intensity of the transcripts was detected in the degU and comA mutant strains (Fig. 36). These findings indicate that the post-transcriptional effects of DegU and YczE on bacillomycin D biosynthesis are mediated through different pathways (see also Discussion).

Figure 36: Mapping of the transcriptional start of yczE by primer extension analysis. DegU and ComA do not influence transcriptional initiation from the identified yczE promoter (PyczE)

Cells were grown and total RNA was prepared from them as described in detail in Fig. 24 and in Materials and Methods. Primer extension was performed using the 5'-end 32P-labelled primer yczeu (see also Table 4). The first four lanes result from dideoxynucleotide sequencing reactions using the same primer. The positions corresponding to the putative yczE transcriptional starts are indicated by arrows. The respective DNA sequences are shown in the outside lanes. Transcriptional starts are highlighted in bold face, the putative -10 hexamer is underlined and highlighted in bold face, whereas the putative extended -10 element is only underlined. Primer extension analysis performed in the wild type strain and its mutant derivatives showed no change in the promoter activity. 1, wild type B. amyloliquefaciens FZB42; 2, TF1 (degU -); 3, CH23 (comA -). Therefore DegU and YczE eliminate bacillomycin D production in B. amyloliquefaciens FZB42 using different pathways.

Global regulators affect the production of surfactin, fengycin and bacillibactin

The MALDI-TOF MS analysis enabled us to monitor the effects that transcriptional regulatory proteins and sigma factors exert on the production of other non-ribosomally synthesized compounds, apart from bacillomycin D. In the spectra obtained by culture filtrate extracts of B. amyloliquefaciens FZB42, fengycin, surfactin, bacillomycin D and the siderophore bacillibactin are visible with distinct peaks (see Table 8 and later on in this paragraph). Although, as already mentioned before, this type of mass spectrometry does not provide accurate quantitative results in respect with the concentrations of the four compounds, a rough estimation of their relative abundance can be achieved by comparing the intensity values of the peaks belonging to each compound with those belonging to the other three in the different genetic backgrounds. Thereby, after 24 hours of growth at 37oC in Landy medium, bacillomycin D is the prominent lipopeptide synthesized by B. amyloliquefaciens FZB42, followed by bacillibactin and surfactin. Fengycin is the antibiotic with the lowest intensity values for its peaks and thus, the lowest abundance (see Figs. 16 and 28A).

↓92

Although the presence of the dhb operon in the genome of B. amyloliquefaciens FZB42 genome was known (see 3.2), it was uncertain whether the biosynthetic template for thecatecholic siderophore 2,3-dihydroxybenzoate-glycine-threonine trimeric ester, bacillibactin [78], was functional. MALDI-TOF mass spectrometry of culture filtrate extracts prepared from B. amyloliquefaciens FZB42, grown for 24 hours at 37°C in Landy medium, verified the production of the compound. Peaks with mass numbers of 883, 905 and 921 m/z (Figs. 28A and 37A) correspond to the protonated [M+H]+ form and the [M+Na, K]+ alkali adducts of the compound, respectively [78].

The peaks attributed to bacillibactin were prominent in the spectrum of the wild type strain (Fig. 37A), showing higher intensity values than those of surfactin (and fengycin; see Fig. 28A) but lower than the intensities of the bacillomycin D peaks. However, the production pattern of the three compounds changed significantly in the degU (TF1) and sigW (AK36) mutants (Fig. 37B and C). In the first case (Fig. 37B), surfactin peaks are significantly stronger than those of bacillibactin, in contrast to the wild type pattern (Fig. 37A), suggesting that the siderophore production is impaired in a degU - genetic background. In consistence with our results, microarray analysis combined with northern blot analysis and reporter fusions activity assays have previously shown that DegU is a positive regulator of the dhb operon in B. subtilis [253]. Moreover, the production of bacillibactin appeared also impaired in the sigW mutant strain (Fig 37C). There the intensity values of the peaks representing bacillibactin repeatedly dropped to the same levels (or lower) of their surfactin counterparts, whereas their difference to the bacillomycin D peaks increased compared to the wild-type strain (compare Figs. 37A and C). These findings suggest that σw might positively regulate the production and/or export of bacillibactin. Interestingly, stress conditions that provoke iron limitation for Bacillus (e.g. high salinity; [221]), trigger both DegU and σw mediated responses in order to aid the cell cope with the new conditions [254]. Iron limitation also triggers the production of bacillibactin [255], which is then secreted from the cell to work as an iron scavenger and then be re-imported into the cell, where its hydrolysis leads to release of cytosolic iron [256]. Therefore, a link between the three components, bacillibactin, DegU and σw seems plausible (see also Discussion).

On the other hand, fengycin levels appeared significantly elevated in many cases where bacillomycin D production was defected. The comA (CH23), sigH (AK50) and sigB (CH33) mutant strains exhibited this behaviour (Fig. 28.B-D and Table 11). However, both the degU (TF1) and bmyD (AK1) mutant did not show any significant increase in fengycin synthesis, even though they are completely unable to produce bacillomycin D (Figs. 18 and 28D). These two results implied that although the same regulatory pathways may differentially act in fengycin and bacillomycin D production, it is not the per se production of bacillomycin D that represses fengycin expression (see also Discussion).

↓93

Finally, most of the introduced mutations on transcriptional regulatory proteins and sigma factors had no apparent effect on the production of surfactin, as observed by MALDI-TOF MS analysis, since the ratios of the intensity values of surfactin peaks to that of other peptide antibiotics were the same for the mutant strains and the wild type. The comA mutant strain (CH23) was deficient in surfactin production (Fig 28B), as already described for B. subtilis [161, 162]. On the contrary, surfactin production was moderately pronounced in the degU mutant strain, in accordance to the reported repressing effect of DegU on srf expression [253, 257].

Figure 37: DegUand σW influence bacillibactin production (performed in collaboration with Dr. J. Vater).

Spectra of culture filtrate extracts prepared from strains grown for 24 hours in Landy medium. A. In the spectrum of B. amyloliquefaciens FZB42, the levels of bacillibactin (m/z 883, 905, 921) are lower than those of bacilllomycin D (m/z 1053, 1067, 1083) and clearly higher than those of surfactin (m/z 1060), as judged by the intensity of their peaks. The ratios of the intensity values between the three compounds’ peaks change in the degU (B) and sigW (C) mutant strains at the cost of bacillibactin. As noticed before, bacillomycin D is not produced (and its peaks are absent) in a degU mutant strain (B), whereas the peaks belonging to surfactin (m/z 1044, 1058, 1074) have higher intensity than those of bacillibactin. In the sigW mutant (C) the levels of bacillibactin are significantly lower than in the wild type strain (A) and are equivalent to those of surfactin (m/z 1058, 1074).

Table 11: MALDI-TOF MS analysis reveals increased production of fengycin in comA, sigB and sigH mutant strains of B. amyloliquefaciens FZB42

Lipopeptide

m/z

Species

Intensity

in:

FZB42

CH23

CH33

AK50

Surfactin

1044.8

C14 [M + Na]+

6000

-

6000

1900

1058.8

C15 [M + Na]+

3000

-

5000

1500

Bacillomycin

1053.7

C14 [M + Na]+

23000

7200

19000

2600

D

1067.7

C15 [M + Na]+

24000

4500

8000

6000

Fengycin

1485.9

C16 [M + Na]+

1600

4000

22000

6000

1449.9

C17 [M + Na]+

1200

3200

18000

3400

Bacillibactin

905

[M+Na]+

9500

4500

7500

1000

921

[M+K]+

11000

3500

6500

1000

Here are presented the intensities of the main peaks corresponding to the four compounds produced by B. amyloliquefaciens FZB42 and its comA (CH23), sigB (CH33) and sigH (AK50) mutant derivatives. The entire spectra can be seen in detail in Fig. 28. Note that the intensity values of the peaks are not comparable between the different MALDI-TOF spectra (this method does not offer quantitative results), but the relative production pattern of the four compounds in the different genetic backgrounds can be directly compared.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
23.01.2007