ExomEtabolitEs of EndosporE-forming bactEria of Bacillus gEnus idEntifiEd by gEnomic-mEtabolomic profiling

a number of unique bioactive metabolites produced by marine bacilli is already known but the metab olomic of these bacteria is underinvestigated. the aim of this work was to carry out the comparative analysis of metabolomic and genomic traits of Bacillus velezensis Onu 553, Bacillus pumilus Onu 554, Bacillus sub tilis Onu 559 strains isolated from Black sea bottom sediments. Organic extracts of each strain were ana lyzed using high-resolution liquid chromatography-mass spectrometry. General annotation of genomes was performed using patRic, search for secretory signals in the primary structure of selected proteins with using signal ip analysis. the search of biosynthetic gene clusters was performed using antisMash, pRisM 3 and BiG-scapes, the reconstruction of metabolites- with pRisM 3 and transator analyzes. the study allowed to find and identify 90, 33 and 43 metabolites in the strains Bacillus velezensis Onu 553, Bacillus pumilus Onu 554 and Bacillus subtilis Onu 559, respectively. the compounds found in metabolome were subdivided into two groups: those which are known members of the genus Bacillus and those new to both genus and prokaryotes in general. among the secondary metabolites of studied strains, the variants of the nonribosomal peptide class surfactins (anhteisoc16-surfactin, surfactin B2-me ester), gageostatins, fengycins and amicou -macins, and the secreted protease inhibiting pentapeptide GpFpi were identified. the biosynthetic clusters of lipopeptides of the pumilacidin subgroup and amicoumacin antibiotic ai-77a were identified for the first time with the use of bioinformatic tools. the data obtained deepen the understanding of the marine bacilli biosynthetic potential.

a number of unique bioactive metabolites produced by marine bacilli is already known but the metabolomic of these bacteria is underinvestigated. the aim of this work was to carry out the comparative analysis of metabolomic and genomic traits of Bacillus velezensis Onu 553, Bacillus pumilus Onu 554, Bacillus subtilis Onu 559 strains isolated from Black sea bottom sediments. Organic extracts of each strain were analyzed using high-resolution liquid chromatography-mass spectrometry. General annotation of genomes was performed using patRic, search for secretory signals in the primary structure of selected proteins with using signal ip analysis. the search of biosynthetic gene clusters was performed using antisMash, pRisM 3 and BiG-scapes, the reconstruction of metabolites-with pRisM 3 and transator analyzes. the study allowed to find and identify 90, 33 and 43 metabolites in the strains Bacillus velezensis Onu 553, Bacillus pumilus Onu 554 and Bacillus subtilis Onu 559, respectively. the compounds found in metabolome were subdivided into two groups: those which are known members of the genus Bacillus and those new to both genus and prokaryotes in general. among the secondary metabolites of studied strains, the variants of the nonribosomal peptide class surfactins (anhteisoc16-surfactin, surfactin B2-me ester), gageostatins, fengycins and amicoumacins, and the secreted protease inhibiting pentapeptide GpFpi were identified. the biosynthetic clusters of lipopeptides of the pumilacidin subgroup and amicoumacin antibiotic ai-77a were identified for the first time with the use of bioinformatic tools. the data obtained deepen the understanding of the marine bacilli biosynthetic potential. k e y w o r d s: Bacillus, metabolomics, genomics, bioinformatic analysis, Lc-Ms. I n our century the search for previously unknown secondary metabolites is one of the main problems of the microorganisms biochemistry, especially bacteria. This allows us to solve some both applied and fundamental issues in prokaryotic biology. Particularly interesting is the study of microorganisms isolated from inaccessible marine biotopes with extreme conditions, which were understudied on earlier stages of metabolite screening researches. And the representatives of the genus Bacillus are promising and relatively under-researched sources of new metabolites [1,2].
The study of bacteria of the genus Bacillus includes the search for new biologically active, including antimicrobial, substances using both meta-bolic and genomic methods. Recently, an integrated genomic-metabolic approach has been actively developed, which provides information on the structure of these compounds and their biosynthesis [3].
The following major groups of antibiotics are already known for members of the Bacillus genus: nonribosomal peptides, including lipopeptides, polyketides, amicoumacins, and their derivatives [4,5]. These molecules are synthesized on polyketide synthetases (PKS) and nonribosomal polypeptides synthetases (NRPS), large molecular complexes where they are assembled from low molecular weight mono mersandmodifiedaccordingtothesequence of corresponding modules -groups of protein domains that have corresponding biosynthetic func-doi: https://doi.org /10.15407/ubj92.06.154 tions. In many cases, the synthesized precursor moleculeundergoessignificantmodificationsalso after leaving the synthetase (for example, conjugates with fatty acids residues). The diversity of synthetase modules leads to the fact that nonribosomal peptides and polyketides form groups that are similar in structurebutdifferinseveralmonomersorinresidues conjugates such as fatty acids [6]. The systematics ofthesecompoundsisstillinconsistent,andquite often certain groups of these compounds with moderate "homology" to one class are considered separately. Besides, even organisms that have only one cluster of a particular polypeptide in the genome are capable to synthesize a whole spectrum of variants, thereof [7]. This is due to a certain indeterminacy oftheprocessitself,suchastheeffectofthermal motion on the work of massive synthetase. In addition to the complex compounds in Bacillus, there are also numerous low molecular weight secondary metabolites, with high biological activity usual, such as butirosin,bacilysin,3,3′-neotrehalosadiamine,and the like [2,4,8].
Representatives of this genus are traditionally considered to be inhabitants of the soil, but they are also regularly isolated from aquatic biotopes, including the sea. The metabolomics of marine bacilli is understudied, but many metabolites have already been known to be isolated from marine bacilli [9].
Aim of this work was in the comparative analysis of metabolomic and genomic traits of antagonistic strains of the genus Bacillus isolated from bottom sediments of the Black Sea. material and methods strains. Antagonistically active strains of Bacillus velezensis ONU 553, Bacillus pumilus ONU 554, Bacillus subtilis ONU 559, which were selected from more than 100 strains of endospore-forming facultative anaerobic bacteria isolated from deep-sea bottom sediments of the Black Sea were studied [10].
Metabolites extraction. The exometabolite profilingwasdonebyusingthemethodfrom [11].For this, the bacteria were grown on tryptic soy broth (Sigma-Aldrich) and M medium of the next composition: soy meal -20 g/l, potato starch -3 g/l, yeast extract -1 g/l, (NH 4 ) 2 SO 4 -1 g/l, MgSO 4 -1 g/l, KH 2 PO 4 -0,6 g/l, pH-7) for 7 day on a ro-taryshakerfortemperature28°Сandrotationspeed 180rpm.Themetabolitesfromthecultureliquidof the studied strains were extracted with ethyl acetate and butanol, evaporated to dryness and dissolved in methanol for further manipulations. The total obtained extracts were fractionated by column size-exclusionchromatographyusinga1mcolumnfilled with LH 20 Sephadex (Sigma-Aldrich). The elution was carried out with methanol as the mobile phase at aflowrateof1ml/min.Thefractionswerecollected every 15 min term, concentrated by evaporation and dissolved in 0.5 ml of methanol.
Lc-Ms. The component composition of the total extracts and their active fractions were analyzed by high-resolution liquid chromatography-mass spectrometry on an amaZon (Brucker, Daltonics) liquid chromatographic mass spectrometer, BEH C1850×2.1mmcolumn,1.7μm(Waters,Germany), mobile phase A -water with 0.1% formic acid and B -acetonitrile with 0.1% formic acid. A linear gradient from 5 to 95% was used for 9 min [11].
Genome sequencing. DNA was isolated using amodifiedmethod [12].Thedailybacterialculture was grown on the LB broth medium at 30°C on an orbital shaker at 160 rpm. Sequencing was done using two libraries with short and long inserts on НiSeq1500equipment(Illumina).Thede novo read assembly is done using the Newbler 2.8 (Roche) assembler.

results and discussion
By using liquid chromatography-mass spectrometry in the culture Bacillus velezensis ONU 553wereidentified90metabolites,Bacillus pumilus ONU 554 -33 and Bacillus subtilis ONU 559 -43, respectively. These compounds can be divided in two groups: those which are known as a products of members of the genus Bacillus and new to both this genus and prokaryotes in general. The compounds of the second group are subject to further study and identification.
Integrated data, obtained by the LC-MS investigationofsamplesfromdifferentmethodsofgrowing and extraction of cultures if studied strains are presented on tables 1, 2, and 3.
Among the secondary metabolites of the studied strains,wereidentifiedvariantsofthebeststudied class of nonribosomal peptides -surfactins (Tables 1, 2, 3). The metabolites of this class are already known in the representatives of B. velezensis, B. pumilus, B. subtilis [2,4,21], and this was confirmedinourresults.Surfactinsareacylatedcyclic nonribosomal heptapeptides with strong surface activity as well as hemolytic and antimicrobial (bacteriostatic and fungicidal) properties [6]. They are characterizedbythestructuralpatternGluLeuLeu d (Val/Ala/Iso)AspLeu d Leu and contains β-hydroxy fatty acid with a chain length in the range of 11-14 atoms. The lichenisins subgroup is characterized by the substitution of glutamic acid for glutamine; besides, they are much more likely than isoleucine to be substituted for leucine at positions 2 and 7 rather thaninclassicalsurfactіns.Itisknownthatthesurface activity of lichenisin is higher than that of surfactin. It should be noted that the biological activity of numerous variants of surfactin has not been thoroughly studied to date.
The structure of the surfactin genetic cluster, is very well known, in contrast. It includes three nonribosomal polypeptide synthetase (srfaa , sfraB, srfac) subunit genes and several relatively small ancillary open reading frames (ORFs), including the obligatory polyketide and NRPS phosphopantethenyl transferase [6].
In strains B. velezensis ONU 553 and B. subtilis ONU 559 were found (Table 1, 3) already known for Bacillus cereus and Bacillus subtilis plipastatins A1, A2, B1, B2 [23], which are phospholipase inhibitors. Detected by systems fengycin IX is a derivative of plipastatinA1anditdiffersfromthelatterbythe presence of bound K + ion [24]. These strains also possess agrastatin A (which source is Bacillus subtilis [22]), an antifungal agent active against many phytopathogens. Lipodepsipeptides SNA 60-367-17 and SNA 60-367-23 are found only in B. subtilis strain ONU 559. They are poorly characterized lipopeptides known for Bacillus sp.anddifferonlyinthe type of fatty acid. They are known for their ability to inhibit aromatase [25].
Genomicstudiesconfirmthepresenceofbiosynthetic clusters similar to the surfactin and fengycin cluster in all strains tested.
Interesting is the issue of the ability to determinethesynthesisofdifferentcompoundsbythe same genetic cluster. The most likely solution to this problem is the known phenomenon of domain skipping, which leads to the formation of molecules "processed"bydifferentsetsofdomainsonthesame synthetase [7].
In the culture of the strain Bacillus velezensis ONU 553 also revealed the presence of three types of gageostatins -A, B, C (Table 1). These compounds have the same peptide skeleton and differ in fatty acid residues at the N-terminus, the recruitment of which in nonribosomal peptides is usually determined by the specificity of the thioesterase domain. The very presence of gageostatins in the strain under study is problematic to explain since a specificgeneticclusterfrompolyketidesynthetase domains, that could synthesize them in the genome ofthisspecies,hasnotbeenidentified.Gageostatins  [26]. These are linear lipopeptides whose mecha nism of biosynthesis has not yet been established, but from the outset, it seems plausible that these are classic nonribosomal peptides; however, a cluster that could be responsible for their biosynthesishasnotbeenidentified.Weemphasizethefact that these compounds were previously detected in strainsofmarineoriginofsimilarspecies.ForBacillus subtilis109GGC020strainalsoweredescribed two other types of antimicrobial peptides, gageobactins, and gageotetrins, but they were not detected in this case. Based on impressive structural similarity, a cautious assumption can be made regarding the synthesis of gageostatin on surfactin synthetase with the skipping of epimerization steps for the corresponding leucineresiduesandthefinaladditionoffattyacidwithoutfinalproductcyclization.
We conducted a comparative analysis of the domain composition of the surfactin-like cluster of strain B. velezensisONU553 (Fig.2)andshowed that the size and location of the domains in this cluster and the reference cluster of surfactin biosynthesis arenotdifferent.
Anothermysteriousfindingwasthepentapep-tideGPFPI(dipeptidylcarboxypeptidaseinhibitor known for Bacillus subtilis) ( Table 3). This relatively poorly known compound was discovered in 1992 [27]. It can be a by-product of metabolism, for example, a residue of the signal peptide, or to be a pheromone. A direct search of this oligopeptide in the genome revealed that such a pentapeptide is part of a hypothetical YvkN peptide, the length of which is 76 amino acid residues. This peptide is included in the cluster with three hypothetical membrane peptides and one hypothetical non-membrane peptide (Table 4). It can be assumed that among these peptides are those that provide propeptide processing  with the formation of pentapeptide and its transport. Search with Signal IP 5.0 does not detect any secretory signals in this peptide.
Muchmorespecific,inbiochemicalterms,is the strain Bacillus pumilus ONU 554 ( Table 2). The presence of pumilacidins C and D in its exometabolomehasalreadybeenmentioned.Genomicmining made it possible to identify the gene cluster that en-codestheNPPSofthislipopeptide (Fig.1).
A much less studied compound, lipoamide A, was also found in this strain. Previously, it was already reported for the marine strain Bacillus pumilus SP21 [28]. It is characterized by relatively weak (minimal inhibitory concentration is about 100µg/ ml) antimicrobial activity against s. aureus and p. aeruginosa. Lipoamides are very simple compounds and are an asparagine conjugate with anteiso-fatty acids with chain lengths in the range 11-13. [28]. It is important not to confuse these compounds with lipoic acid amides. The capabilities of modern methods of genomics at this time do not allow to localize in the genome the determinants of lipoamide A biosynthesis.
Another interesting and specific for Bacillus pumilus is the compound AI-77A, active against gram-positive and gram-negative bacteria, which is cytotoxic, also having anticholesterol and antiinflammatory properties [29]. AI-77A belongs to the class of amicoumacins, antibiotics that are very characteristic of bacilli. Only three such compounds are known for actinobacteria and almost thirty for bacilli, which is not characteristic of any other class of antibiotics [5]. The substances of this group consist of a conservative dihydroisocoumarin core and thetype-specificmoietycoupledwiththecorebyan amide bond.
Compound AI-77A has antibacterial activity against the causative agent of gastric ulcer helicobacter pylori. Its mechanism of action is to bind the molecule simultaneously with the mRNA and 16S RNAs of the small subunit of the ribosome, causing the termination of protein synthesis [5]. Biosynthetic clusters for amicoumacin-like antibiotics have been poorly studied. By comparison, the homology of the hybrid (NRPS /PKC) synthetase cluster from the Bacillus pumilus ONU 554 genome and the cluster of the well-studied antibiotic of this group -xenocoumacin (Fig.3)wasdemonstratedbyEasyfigsoftware.
The analysis of the remaining metabolites (for Bacillus velezensis ONU 553 -58, for Bacillus pumilus ONU 554 -22, Bacillus subtilis ONU 559 -28) is challenging. As was established by the TransA-Tor analyzer, the structure of the trans-AT polyketide synthetase products, which were detected in the genomesstudiedandidentifiedbyantiSMASHas macrolactin,dificidin,andbacillaenedoesnotcorrespond to the known structure for these compounds [4]. Therefore, the actual products of these genetic complexesneedfurtheridentification. Fig. 3. comparison of ai-77a cluster form genome Bacillus pumilus Onu 554 (above) and xenocoumacin cluster from genome Xenomonas sp. (below) [19]. Red color corresponds loci with homology >90% conclusions. As a result of the conducted researches it is established that all studied strains of Black Sea bacilli synthesize compounds, the structureofwhichfitsintoasmallnumberofstructural categories, and the diversity of individual compounds strongly depends on the species of the producer. Even strains of relatively phylogenetically related species B. velezensis and B. subtilisdiffersignificantlyinthesetofmetabolites.Itshouldalsobe notedthatthedetectedproportionofidentifiedcompoundsissmall,andnotyetidentifiedcompounds arethesubjectforfurtherstudyandidentification.
In the studied strains of bacilli isolated from bottom sediments of the Black Sea, biosynthetic clustershavebeenidentifiedanddescribedthatare interesting from a fundamentally biological point of view and may be useful for biotechnological research. Of particular interest may be the peptidase inhibitor pentapeptide B. subtilis ONU 554 and the B. velezensis ONU 553 surfactin cluster, the study of which can shed light on the little-studied issue of biosynthesis of various compounds on a single polypeptide synthetase complex.
acknowledgements. The author's team expresses gratitude to doctor Andryi Luzhetsky from Helmholtz Institute for Pharmaceutical Research Saarlandforscientificadviceandtechnicalsupport in carrying out this work.