In Silico Analysis of PKS and NRPS Gene Clusters in Arisostatin- and Kosinostatin-Producers and Description of Micromonospora okii sp. nov.

Micromonospora sp. TP-A0316 and Micromonospora sp. TP-A0468 are producers of arisostatin and kosinostatin, respectively. Micromonospora sp. TP-A0316 showed a 16S rRNA gene sequence similarity of 100% to Micromonospora oryzae CP2R9-1T whereas Micromonospora sp. TP-A0468 showed a 99.3% similarity to Micromonospora haikouensis 232617T. A phylogenetic analysis based on gyrB sequences suggested that Micromonospora sp. TP-A0316 is closely related to Micromonospora oryzae whereas Micromonospora TP-A0468 is an independent genomospecies. As Micromonospora sp. TP-A0468 showed some phenotypic differences to its closely related species, it was classified as a novel species, for which the name Micromonospora okii sp. nov. is proposed. The type strain is TP-A0468T (= NBRC 110461T). Micromonospora sp. TP-A0316 and M. okii TP-A0468T were both found to harbor 15 gene clusters for secondary metabolites such as polyketides and nonribosomal peptides in their genomes. Arisostatin-biosynthetic gene cluster (BGC) of Micromonospora sp. TP-A0316 closely resembled tetrocarcin A-BGC of Micromonospora chalcea NRRL 11289. A large type-I polyketide synthase gene cluster was present in each genome of Micromonospora sp. TP-A0316 and M. okii TP-A0468T. It was an ortholog of quinolidomicin-BGC of M. chalcea AK-AN57 and widely distributed in the genus Micromonospora.


Introduction
Actinomycetes are Gram-positive filamentous bacteria and its members are recognized as a rich source of bioactive secondary metabolites, many of which have been utilized for pharmaceutical purposes [1]. Although soil is the main habitat of actinomycetes, including the genus Streptomyces, marine environments such as sea water have been identified as sites for the isolation of actinomycetal strains producing new bioactive compounds. Members of the genus Micromonospora are often isolated from marine environments and have been found to produce diverse secondary metabolites [2]. In our previous studies, Micromonospora sp. TP-A0316 and Micromonospora sp. TP-A0468 were isolated from sea water by the membrane filter method, followed by their cultivation on an agar plate [3,4]. Micromonospora sp. TP-A0316 produces novel compounds, named arisostatins A and B, in addition to tetrocarcin A [3] whereas Micromonospora sp. TP-A0468 produces kosinostatin [4]. Arisostatins are new members of the tetrocarcin class of antibiotics (Figure 1a), providing antibiotic activity against Gram-positive bacteria and demonstrating antitumor activity [3]. Although the tetrocarcin A-biosynthetic gene cluster (BGC) was identified antitumor activity [3]. Although the tetrocarcin A-biosynthetic gene cluster (BGC) was identified in Micromonospora chalcea NRRL 11289 [5], the arisostatin-BGC of Micromonoposra sp. TP-A0316 has not yet been identified. Kosinostatin is a new quinocycline antibiotic ( Figure 1b) with antibacterial, anti-yeast and antitumor activities [4]. Kosinostatin-BGC have already been reported in Micromonospora sp. TP-A0468. Tetrocarcin A and kosinostatin are synthesized via type-I polyketide synthase (PKS) and type-II PKS pathways, respectively [5,6]. Polyketides are biosynthesized by the assembly of acyl-CoA units. Type-I PKSs are large modular enzymes composed of multiple catalytic domains and synthesize polyketide chains based on the co-linearity rule of assembly lines. The mechanism resembles that of nonribosomal peptide synthetase (NRPS) pathways, as nonribosomal peptides are biosynthesized by the assembly of amino acid units and NRPSs are also large modular enzymes composed of multiple catalytic domains and synthesize peptide chains according to the co-linearity rule of assembly lines [7]. In type-II PKS pathways, a set of three enzymes, ketosynthase  (KS), KS (chain length factor), and acyl carrier protein (ACP), iteratively catalyzes the elongation of polyketide chains. The products are mainly aromatic compounds [8]. Approximately half to three quarters of secondary metabolite-BGCs, in the genomes of actinomycetes, are associated with PKS or NRPS pathways. This suggests that polyketides, nonribosomal peptides, and their hybrid compounds, which are synthesized by hybrid PKS/NRPS gene clusters, are major secondary metabolites in actinomycetes [9]. These compounds are structurally diverse and often exhibit useful pharmaceutical activities. Hence, nowadays, genome analyses focusing on PKS and NRPS gene clusters are often conducted to evaluate the potential use of actinomycete strains as a source for novel secondary metabolites [10][11][12].
In this study, we investigated the taxonomic positions of Micromonospora sp. TP-A0316 and Micromonospora sp. TP-A0468, since the classification of antibiotic producers at the species level is important to understand the relationship between species and products. Next, we sequenced whole genomes of these two strains to reveal their potential in producing diverse secondary metabolites such as polyketides and nonribosomal peptides. Consequently, Micromonospora sp. TP-A0468 was considered to be a novel species, for which we propose Micromonospora okii sp. nov. Additionally, we observed a wide distribution of quinolidemicin-BGCs in the genus Micromonospora and classified ten Arisostatin A, R 1 = NO 2 , R 2 = CH(CH 3 ) 2 ; arisostatin B, R 1 = NH 2 , R 2 = CH(CH 3 ) 2 ; tetrocarcin A: Polyketides are biosynthesized by the assembly of acyl-CoA units. Type-I PKSs are large modular enzymes composed of multiple catalytic domains and synthesize polyketide chains based on the co-linearity rule of assembly lines. The mechanism resembles that of nonribosomal peptide synthetase (NRPS) pathways, as nonribosomal peptides are biosynthesized by the assembly of amino acid units and NRPSs are also large modular enzymes composed of multiple catalytic domains and synthesize peptide chains according to the co-linearity rule of assembly lines [7]. In type-II PKS pathways, a set of three enzymes, ketosynthase α (KSα), KSβ (chain length factor), and acyl carrier protein (ACP), iteratively catalyzes the elongation of polyketide chains. The products are mainly aromatic compounds [8]. Approximately half to three quarters of secondary metabolite-BGCs, in the genomes of actinomycetes, are associated with PKS or NRPS pathways. This suggests that polyketides, nonribosomal peptides, and their hybrid compounds, which are synthesized by hybrid PKS/NRPS gene clusters, are major secondary metabolites in actinomycetes [9]. These compounds are structurally diverse and often exhibit useful pharmaceutical activities. Hence, nowadays, genome analyses focusing on PKS and NRPS gene clusters are often conducted to evaluate the potential use of actinomycete strains as a source for novel secondary metabolites [10][11][12].
In this study, we investigated the taxonomic positions of Micromonospora sp. TP-A0316 and Micromonospora sp. TP-A0468, since the classification of antibiotic producers at the species level is important to understand the relationship between species and products. Next, we sequenced whole genomes of these two strains to reveal their potential in producing diverse secondary metabolites such as polyketides and nonribosomal peptides. Consequently, Micromonospora sp. TP-A0468 was considered to be a novel species, for which we propose Micromonospora okii sp. nov. Additionally, we observed a wide distribution of quinolidemicin-BGCs in the genus Micromonospora and classified ten Micromonospora strains for which whole genome sequences have been published, although species names have been unclear.
Next, we reconstructed a phylogenetic tree based on DNA gyrase subunit B gene (gyrB) sequences ( Figure 3), because 16S rRNA gene-based phylogenies of the genus Micromonospora did not always agree with other taxonomic characteristics, and the gyrB sequence has been reported to be suitable for phylogenetic classification and identification [13]. Micromonospora sp. TP-A0316 formed a clade with the type strain of M. oryzae and their gyrB sequences are identical. This suggests that Micromonospora sp. TP-A0316 is likely M. oryzae. On the other hand, the position of Micromonospora sp. TP-A0468 was deep branched and monophelic, suggesting its phylogenetical independency. Although Micromonospora sp. TP-A0468 formed a clade with the type strains of M. oryzae, Micromonospora carbonacea, M. harpali and M. haikouensis, its gyrB sequence similarities to the four strains were 94.9%, 94.9%, 94.9% and 94.7%, respectively. It has been reported that a 98.5% gyrB-sequence similarity corresponds to 70% DNA-DNA relatedness [13,14]. As the gyrB sequence similarities are well below 98.5%, Micromonospora sp. TP-A0468 is considered as an independent genomospecies.  Next, we reconstructed a phylogenetic tree based on DNA gyrase subunit B gene (gyrB) sequences ( Figure 3), because 16S rRNA gene-based phylogenies of the genus Micromonospora did not always agree with other taxonomic characteristics, and the gyrB sequence has been reported to be suitable for phylogenetic classification and identification [13]. Micromonospora sp. TP-A0316 formed a clade with the type strain of M. oryzae and their gyrB sequences are identical. This suggests that Micromonospora sp. TP-A0316 is likely M. oryzae. On the other hand, the position of Micromonospora sp. TP-A0468 was deep branched and monophelic, suggesting its phylogenetical independency. Although Micromonospora sp. TP-A0468 formed a clade with the type strains of M. oryzae, Micromonospora carbonacea, M. harpali and M. haikouensis, its gyrB sequence similarities to the four strains were 94.9%, 94.9%, 94.9% and 94.7%, respectively. It has been reported that a 98.5% gyrB-sequence similarity corresponds to 70% DNA-DNA relatedness [13,14]. As the gyrB sequence similarities are well below 98.5%, Micromonospora sp. TP-A0468 is considered as an independent genomospecies.
Additionally, we conducted a multilocus sequence analysis (MLSA) using 85 housekeeping genes ( Figure 4). Although Micromonospora sp. TP-A0468 formed a clade with M. haikouensis DSM 45626 T , Micromonospora sp. TP-A0316 and M. carbonacea DSM 43168 T , its evolutionally relationships with them are not as close as the relationships that exist among the three strains ( Figure 4). The DNA-DNA relatedness between Micromonospora sp. TP-A0468 and these three members was found to be between 33.5% and 33.8% (data not shown). These results also suggest Micromonospora sp. TP-A0468 to be an independent genomospecies.
Phenotypic differences were observed between Micromonospora TP-A0468 and its closely related phylogenetic neighbors such as M. oryzae, M. carbonacea, M. harpali and M. haikouensis as listed in Table 1. Unlike these neighbors, Micromonospora TP-A0468 includes galactose within the whole-cell sugar. Its growth ranges and utilization pattern of carbon sources are different from those of the other listed species. Although M. oryzae may appear to show a similar utilization pattern of carbon sources, except for D-xylose, it produces soluble pigment and liquefies gelatin, which is different to Micromonospora TP-A0468. Thus, we classified Micromonospora TP-A0468 as a novel species, for which the name Micromonospora okii sp. nov. is proposed. The type strain is TP-A0468 T (=NBRC 110461 T ).  Additionally, we conducted a multilocus sequence analysis (MLSA) using 85 housekeeping genes ( Figure 4). Although Micromonospora sp. TP-A0468 formed a clade with M. haikouensis DSM 45626 T , Micromonospora sp. TP-A0316 and M. carbonacea DSM 43168 T , its evolutionally relationships with them are not as close as the relationships that exist among the three strains ( Figure 4). The DNA-DNA relatedness between Micromonospora sp. TP-A0468 and these three members was found to be between 33.5% and 33.8% (data not shown). These results also suggest Micromonospora sp. TP-A0468 to be an independent genomospecies.   Phenotypic differences were observed between Micromonospora TP-A0468 and its closely related phylogenetic neighbors such as M. oryzae, M. carbonacea, M. harpali and M. haikouensis as listed in Table 1. Unlike these neighbors, Micromonospora TP-A0468 includes galactose within the whole-cell sugar. Its growth ranges and utilization pattern of carbon sources are different from those of the other listed species. Although M. oryzae may appear to show a similar utilization pattern of carbon sources, except for D-xylose, it produces soluble pigment and liquefies gelatin, which is different to Micromonospora TP-A0468. Thus, we classified Micromonospora TP-A0468 as a novel species, for which the name Micromonospora okii sp. nov. is proposed. The type strain is TP-A0468 T (=NBRC 110461 T ).     (28) pH for growth (optimum) 6-10 (7-8) 5-8.5 5-10 (7) nd 6-10 (7) NaCl tolerance (%)

PKS and NRPS Gene Clusters in Micromonospora sp. TP-A0316 and M. okii TP-A0468 T
Fifteen gene clusters for secondary metabolites such as polyketides and nonribosomal peptides were observed in the genomes of Micromonospora sp. TP-A0316, as listed in Table 2. Type-I PKS gene cluster 1 (t1pks-1) resembled the tca gene cluster responsible for tetrocarcin A synthesis in M. chalcea NRRL 11289 [5] ( Figure 5). As arisostatins are congeners of tetrocarcin A, and Micromonospora sp. TP-A0316 is reported to produce tetrocarcin A in addition to arisostatins A and B [3], t1pks-1 was considered as the BGC for arisostatins and tetrocarcin A. Furthermore, t1pks-2 was found to be a large cluster of >200 kb and include 33 modules. This was considered as an ortholog of BGC for quinolidomicin (qnm), the largest known macrolide [18], according to the similar gene and domain organizations (Table 3). However, its module number is different from that of qnmA because t1pks-2 lacks module 4. The product is likely a quinolidomicin congener, but its polyketide skeleton is presumed to be different from quinolidomicin A 1 [18]. In contrast, t1pks-3 was not found to be multimodular, but harbored only a single module. This gene cluster was predicted to be involved in sporolide synthesis [19]. As t1pks-4 was not completely sequenced, its product could not be predicted. Products of t2pks-1 were not predicted by our bioinformatic analysis. However, it is generally known that type-II PKS pathways are responsible for the synthesis of aromatic compounds. Additionally, t3pks-1 showed similarity to agq, a type-III PKS gene cluster for alkyl-O-dihydrogeranyl-methoxyhydroquinone [20]. Five NRPS gene clusters in this strain did not show high similarities to other known NRPS gene clusters, suggesting them to be orphan, although nrps-4 was not completely sequenced. They were predicted to synthesize pentapeptide, tripeptide, tetrapeptide and dipeptide, respectively, as listed in Table 2. Four hybrid PKS/NRPS gene clusters, pks/nrps-1, -2, -3 and -4, were also orphan and were predicted to synthesize heptapeptide, tripeptide and pentapeptide with polyketide moieties and hexaketide with a glycine molecule, respectively.

Distribution of Quinmuinolidomicin-BGC Orthologs in the Genus Micromonospora
Unexpectedly, both Micromonospora sp. TP-A0316 and M. okii TP-A0468 T possessed an ortholog of qnm gene cluster, which is the largest type-I PKS gene cluster identified to date [18]. We investigated its distribution in genome sequence-published strains of the genus Micromonospora. Among the 74 strains shown in Figure 6, 34 strains were found to harbor the ortholog. Among them, 23 strains were phylogenetically close to Micromonospora aurantiaca or to the two strains studied here. However, the remaining 11 strains are phylogenetically diverse, suggesting the ortholog is widely distributed in the genus Micromonospora. Although the 16S rRNA gene sequences between Micromonospora sp. B006 and Micromonospora tulbaghiae DSM 45142 T were identical, it was found that Micromonospora sp. B006 harbors the ortholog [23] while M. tulbaghiae DSM 45142 T does not. Because it is reported that members in the same species possess similar sets of PKS and NRPS gene clusters [24], we examined DNA-DNA relatedness values, which were estimated using digital DNA-DNA hybridization (DDH) among strains showing high 16S rRNA gene sequence similarities to clarify their taxonomic relationships. As noted in Figure 5, the DNA-DNA relatedness value between Micromonospora sp. B006 and M. tulbaghiae DSM 45142 T was 51%, which is below the species cut-off value of 70% defined in the bacteria systematics [25], suggesting them to be different species.  Although the 16S rRNA gene sequences between Micromonospora sp. B006 and Micromonospora tulbaghiae DSM 45142 T were identical, it was found that Micromonospora sp. B006 harbors the ortholog [23] while M. tulbaghiae DSM 45142 T does not. Because it is reported that members in the same species possess similar sets of PKS and NRPS gene clusters [24], we examined DNA-DNA relatedness values, which were estimated using digital DNA-DNA hybridization (DDH) among strains showing high 16S rRNA gene sequence similarities to clarify their taxonomic relationships. As noted in Figure 5, the DNA-DNA relatedness value between Micromonospora sp. B006 and M. tulbaghiae DSM 45142 T was 51%, which is below the species cut-off value of 70% defined in the bacteria systematics [25], suggesting them to be different species.

Discussion
The relationships that exist between taxonomic species and secondary metabolites are still unclear because many strains that produce bioactive secondary metabolites have not been classified at species level. This study aimed to elucidate the taxonomic positions of both Micromonospora sp. TP-A0316, a producer of arisostatins, and Micromonospora sp. TP-A0468, a producer of kosinostatin, at the species level. We concluded that Micromonospora sp. TP-A0316 is closely related to M. oryzae, and that Micromonospora sp. TP-A0468 should be classified as a novel species, for which we propose M. okii sp. nov. These two strains each harbor 15 PKS and NRPS gene clusters in their genomes. We characterized these gene clusters bioinformatically. Among the 15 clusters of each strain, only 4 were conserved between the strains. This is because Micromonospora sp. TP-A0316 and M. okii TP-A0468 T are different species.
Our genome analysis revealed that, alongside the two strains that have not been reported as quinolidomicin-producers, diverse Micromonospora strains harbor orthologs of the qnm gene cluster. Members in the genus Micromonospora are known to include producers of aminoglycoside antibiotics such as gentamicin [26], mutamicin [27], netilmicin, retymicin, sisomicin [28], verdamicin and turbinmicin [29]. Quinolidomicins may be one of the representative products, although the report is limited [30] by the difficulties associated with its structure [18,31].
In addition to Micromonospora sp. TP-A0316 and Micromonospora sp. TP-A0468, many genome sequence-published Micromonospora strains have not been classified at species level. Digital DDH conducted in this study clarified the taxonomic positions as follows: Micromonospora sp. L5, Micromonospora sp. RV43, Micromonospora sp. WMMB235, Micromonospora sp. CNZ297, Micromonospora sp. CNZ296 and M. globosa NRRL B-2673 are M. aurantiaca; Micromonospora sp. M42 and Micromonospora sp. DSW705 are M. chalcea; Micromonospora sp. NRRL B-16802 is M. profundi. Although the strain NRRL B-2672 has been published as Micromonospora purpureochromogenes, we found this to not be true, because they are phylogenetically distant as shown in Figure 5 and its DNA-DNA relatedness to M. purpureochromogenes DSM 43827 T was only 27% (data not shown). It may be possible to classify Micromonospora sp. CNZ322 as M. saelicesensis since their DNA-DNA relatedness was found to be 71%. In contrast, Micromonospora sp. WMMA2032, Micromonospora sp. TSRI0369 and Micromonospora sp. B006 are likely to be classified as independent genomospecies, since their DNA-DNA relatedness to each phylogenetic neighbor was 37%, 64% and 51%, respectively.
We stated that Micromonospora sp. TP-A0316 is likely to be classified as M. oryzae in the results section. However, this strain and M. haikouensis JXNU-1, which is not the type strain of M. haikouensis, unexpectedly shared the same 16S rRNA gene sequence as shown in Figure 6. Strain JXNU-1 may not be M. haikouensis but M. oryzae. Our digital DNA-DNA hybridization suggested that M. oryzae and M. haikouensis may be identical because the members showed DNA-DNA relatedness values of >74%, as shown in Figure 5, although whole genome sequence of M. oryzae type strain is not available. If it is considered that M. oryzae and M. haikouensis are synonym, Micromonospora sp. TP-A0316 may be classified as M. haikouensis based on the priority rule of the International Code of Nomenclature of Bacteria.

Description of Micromonospora okii sp. nov.
Micromonospora okii (o.ki'i. N.L. gen. n. okii of Oki, named in honor of the late Professor Toshikazu Oki, a celebrated actinomycete biologist who organized the study on strain TP-A0468).
The description provided is based on data obtained in a previous study [4]. Aerobic and Gram stain-positive filamentous actinomycete. Spores are singly formed on substrate mycelium. The spore shape and size are oval and range from 0.8 to 1.2 mm, respectively. The colors of vegetative mycelium and the reverse side are yellowish or grayish white to grayish brown on sucrose-nitrate agar, white light orange on glucose-asparagine agar, yellowish brown on Bennett's agar, light orange to dark gray on nutrient agar, light or light yellowish brown to grayish brown on oatmeal agar, dark brown to dark yellowish brown on inorganic salts-starch agar, and white on glycerol asparagine agar. Vegetative mycelium and the reverse side are, respectively, beige white to light grayish brown and white on glucose-nitrate agar, soft orange to olive gray and light yellowish brown to medium gray on yeast extract-malt extract agar, beige gray to light yellowish brown and grayish white to yellowish brown on tyrosine agar. Vegetative mycelium acts well on nutrient agar, Bennett's agar, yeast extractmalt extract agar, oatmeal agar, and inorganic salts-starch agar, but poorly on sucrose-nitrate agar, glucose-nitrate agar, glucose-asparagine agar, glycerol asparagine agar and tyrosine agar. Aerial mycelium and diffusible pigments are not formed. Starch hydrolysis, milk coagulation and milk peptonization are positive. The temperature range for growth is 13 to 41 • C and the optimum temperature is from 25 to 39 • C. D-Glucose, sucrose maltose, L-rhamnose, Dmannose, D-fructose, L-arabinose, and D-galactose are utilized for growth. Inositol, D-mannitol, raffinose and D-xylose are not utilized. Whole-cell hydrolysates contain meso-diaminopimelic acid as the diagnostic diamino acid, and galactose, xylose, arabinose and glucose as the whole-cell sugars. The phospholipid type is the PII pattern, and phosphatidylethanolamine and phosphatidylinositol are present. The type strain produces kosinostatin.
The type strain is TP-A0468 T (=NBRC 110461 T ). The DNA G+C content of the type strain is 73.9% (determined by whole genome-sequencing). Accession numbers of the draft genome sequence of the type strain are BBZF01000001-BBZF01000036.