Corallococcus soli sp. Nov., a Soil Myxobacterium Isolated from Subtropical Climate, Chalus County, Iran, and Its Potential to Produce Secondary Metabolites

A novel myxobacterial strain ZKHCc1 1396T was isolated in 2017 from a soil sample collected along Chalus Road connecting Tehran and Mazandaran, Iran. It was a Gram-negative, rod-shaped bacterial strain that displayed the general features of Corallococcus, including gliding and fruiting body formation on agar and microbial lytic activity. Strain ZKHCc1 1396T was characterized as an aerobic, mesophilic, and chemoheterotrophic bacterium resistant to many antibiotics. The major cellular fatty acids were branched-chain iso-C17:0 2-OH, iso-C15:0, iso-C17:1, and iso-C17:0. The strain showed the highest 16S rRNA gene sequence similarity to Corallococcus terminator CA054AT (99.67%) and C. praedator CA031BT (99.17%), and formed a novel branch both in the 16S rRNA gene sequence and phylogenomic tree. The genome size was 9,437,609 bp, with a DNA G + C content of 69.8 mol%. The strain had an average nucleotide identity (ANI) value lower than the species cut-off (95%), and with the digital DNA–DNA hybridization (dDDH) below the 70% threshold compared to the closest type strains. Secondary metabolite and biosynthetic gene cluster analyses revealed the strain’s potential to produce novel compounds. Based on polyphasic taxonomic characterization, we propose that strain ZKHCc1 1396T represents a novel species, Corallococcus soli sp. nov. (NCCB 100659T = CIP 111634T).


Introduction
Myxobacteria are Gram-negative, rod-shaped bacteria belonging to the phylum Myxococcota [1] and are considered unique for their social behavior and complex developmental growth stages. In many myxobacteria, nutrient-limiting conditions enable the vegetative cells to swarm, aggregate, and form multicellular fruiting bodies [2][3][4][5]. Within the fruiting bodies, six days of shaking (160 r.p.m., 30 • C). Swarming was observed on lean media and as well as on Casitone-containing agars. The fruiting bodies and swarm colonies were examined using an Olympus SZX12 stereomicroscope, while the vegetative cells and myxospores were studied using a Zeiss AX10 phase-contrast microscope, photographed using a Zeiss Axiocam MRC camera, and analyzed using AxioVision LE software.
Gram-staining, oxidase, and catalase tests were based on previously described methods [18,28]. The API ZYM ® (bioMérieux) and API ® Coryne reactions were conducted following the manufacturer's instructions. Temperature tolerance of the novel strain was determined at 18,25,30,35,37, and 40 • C, while pH tolerance was tested at pH 5.0-9.0 with intervals of pH 0.5. Both temperature and pH determination were performed in VY/2 agar and were assessed based on the colony growth.
Antibiotic resistance of the novel isolate was tested on VY/2 agar with 50 µg ml −1 antibiotic concentration. The tested antibiotics were ampicillin, amikacin, cefotaxime (Carl Roth), ceftazidime, imipenem, gentamicin, and trimethoprim-sulfamethoxazole (Sigma-Aldrich). All antibiotics were filter-sterilized before being added to the autoclaved agar, which was cooled down to 55 • C before plating.
Microbial predation of the novel myxobacterium was tested using Bacillus subtilis DSM 10 T , Micrococcus luteus DSM 1790, Escherichia coli DSM 1116, and Wickerhamomyces anomalus DSM 6766 T . Predation was evaluated for clearing of the baited strain, which indicated cell lysis. Degradation of cellulose and chitin was determined based on previously described methods [17], while agar degradation was determined in all solid media, using Bacto agar (1.6% w/v) as a solidifying agent.

Genome and Phylogenetic Analysis
For genomic DNA isolation, the cells were obtained from an actively growing CY-H culture and the DNA was extracted following the standard method for Gram-negative bacteria using the Puregene Core Kit A from Qiagen. The amplification of the 16S rRNA gene was performed using the universal primers F27 (5 -GAGTTTGATCCTGGCTCAGGA-3 ) and R1525 (5 -AAGGAGGTGATCCAGCCGCA-3 ) [17]. The amplified PCR products were purified using a Macherey Nagel NucleoSpin Kit, separated by gel electrophoresis (0.8% (w/v) agarose, at 70 V, for 45 min), and subsequently sequenced using primers F27 [18], R1525, R518, F1100, and R1100 [19]. The 16S rRNA gene sequence was aligned using the Cap contig assembly of the BioEdit Sequence Alignment Editor software version 7.0.5 [31].
The 16S rRNA gene sequence phylogenetic analysis was conducted using the GGDC web server (http://ggdc.dsmz.de/, accessed on 25 January 2022) [32]. Pairwise sequence similarities were calculated according to the method of Meier-Kolthoff et al. [33]. Phylogenies were inferred using the phylogenomics pipeline developed by DSMZ [34] adapted to single genes, and the sequence alignment was performed using MUSCLE [35]. Maximum likelihood (ML) and maximum parsimony (MP) trees were constructed using RAxML [36] and TNT [37], respectively. For ML, the autoMRE bootstrapping criterion [38], and a subsequent search for the best tree, was employed. In MP, 1000 replicates from bootstrapping were used in conjunction with tree bisection and reconnection branch swapping, and ten random sequence additional replicates. The sequences were evaluated for a compositional bias using the X 2 test implemented in PAUP* [39].
The genome sequencing of strain ZKHCc1 1396 T was carried out using next-generation sequencing technology (Illumina) with MiSeq 600 cycle v3. De novo genome assembly was performed using a Unicycler [40]. Predicted genes, tRNA genes, rRNA genes, and other characteristics of the genome were annotated using PROKKA [41]. In addition, the annotated data from the Prokaryotic Genome Annotation Pipeline (PGAP) of NCBI [42] were also used for genomic comparisons of all Corallococcus type strain genomes. The possible contamination of the genomic data was evaluated using the ContEst16S algorithm to analyze the 16S rRNA gene fragments (https://www.ezbiocloud.net/tools/contest16s, accessed on 10 March 2020) [43]. The complete 16S rRNA gene sequence of the novel strain was extracted from its genome, and this was used for the phylogenetic analysis and percentage similarity comparisons with the closest type strains. The percentage DNA G + C content was determined based on the strain's genome sequence.
The genomic sequence data of strain ZKHCc1 1396 T was uploaded in the Type Strain Genome Server (TYGS) (https://tygs.dsmz.de, accessed on 25 January 2022) for a wholegenome-based taxonomic analysis [44] with the recently introduced methodological updates and features [32]. Information on nomenclature, synonymy, and associated taxonomic literature was provided by the List of Prokaryotic names with Standing Nomenclature (LPSN, available at https://lpsn.dsmz.de, accessed on 25 January 2022) [45].
The uploaded genome was compared against all type strain genomes in the TYGS database using the MASH algorithm [46]. Additionally, the genome data of Corallococcus silvisoli c25j21 T (JAAAPJ000000000) was also added separately from the NCBI website (https://www.ncbi.nlm.nih.gov/, accessed on 26 January 2022) because it was not yet listed in the TYGS database. The closest type strains were chosen based on the smallest MASH distances and the 16S rRNA gene sequences. Extraction of 16S rRNA gene sequences from the genome was completed using RNAmmer [47], and was subsequently BLASTed [48] against the type strain's 16S rRNA gene sequences available in the TYGS database. This was used as a proxy to search for the top 50 matching type strains (according to the bitscore) and to calculate the distances using the Genome BLAST Distance Phylogeny approach (GBDP) under the "coverage" and distance formula d5 algorithm [49]. These distances were used in determining the closest type strain genomes.
Phylogenomic inference was performed using the GBDP. Intergenomic distances were inferred using "trimming" and distance formula d5 algorithm [49] with 100 distance replicates. The digital DDH values (dDDH) and confidence intervals were calculated using the GGDC 3.0 recommended settings [46,49]. The resulting intergenomic distances were used to infer a balanced minimum evolution tree with FASTME 2.1.6.1 branch support, and include SPR post-processing [50]. Branch support was calculated from 100 pseudobootstrap replicates each. The phylogenetic trees were rooted at the midpoint [51], and was visualized using a PhyD3 program [52].

BiG-SCAPE Analysis
Genome data for all of the type strains of Corallococcus species were downloaded from the NCBI database. The genome data were analyzed using AntiSMASH version 6.0.0 (available at https://antismash.secondarymetabolites.org/, accessed on 2 February 2022) to identify the secondary metabolite gene clusters using the "relaxed" strictness setting [54,55]. All of the predicted biosynthetic gene clusters (BGCs) were then analyzed using the BiG-SCAPE program (version 1.1.2 (3 June 2021)), with the MiBIG database (version 2.1) as a reference [56,57], and Pfam database version 34.0 [58]. Some parameters that were used include a distance cut-off score of 0.3, and the search terms "hybrid" and "mix". Generated networks were visualized with Cytoscape (version 3.8.2) [59].

Extract Production, Antimicrobial Assay, and Extract Analysis
The strain ZKHCc1 1396 T pre-culture was grown in a 100 mL flask containing 20 mL CY-H broth and incubated on a rotary shaker (160 r.p.m.) for 7-14 days at 30 • C. To screen for secondary metabolites, the resultant cultures (20 mL) were transferred in 250 mL flasks containing 100 mL of the production media, including E medium, CY medium, P medium, POL medium, S medium, M medium, and myxovirescin medium (Table S2), and supplemented with 2% (v/v) XAD-16 adsorber resin. After 14 days of incubation at 30 • C, the resins and cells were collected together by filtering through a fine metal mesh. They were extracted with 70 mL acetone for one hour, filtered with filter paper, and concentrated in vacuo at 40 • C using a rotary evaporator (Heidolph, Schwabach, Germany).  [60] to obtain OD 600 of 0.05.
The crude extract, which showed high and interesting biological activity, was chosen for further analysis using HPLC-DAD Agilent 1260 series coupled with a MaXis ESI-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany). Column C18 Acquity UPLC BEH (Ultra Performance Liquid Chromatography Ethylene-Bridged Hybrid, Waters) and two solvents (solvent A: H 2 O + 0.1% formic acid; solvent B: ACN + 0.1% formic acid) were used in the HPLC system. The compound separation was performed with a flow rate of 0.6 mL/min, column temperature of 40 • C, and the gradient condition was as follows: 5% B in 0-0.5 min, 5-100% B in 0.5-20 min, and 100% B in 20-25 min [61,62]. Chromatogram and spectrum analysis was conducted using Compass DataAnalysis Version 4.4 (Bruker Daltonics, Billerica, MA, USA). Fractions were selected based on retention time every two minutes in the range of 1.8-20 min from the base peak chromatogram (BPC). For compound prediction, the detected accurate mass (with ±0.01 Da) was manually searched for using the database of Dictionary of Natural Products version 30.1.

Taxonomic Identification
Strain ZKHCc1 1396 T showed the characteristic features of myxobacteria, which include swarming, fruiting, and myxospore formation ( Figure 1). In all solid agar media, the colony produced a coherent swarm after inoculation at the center of the Petri dish (Figure 1a,b). On VY/2 agar, the colony spread fast and formed dense orange fruiting bodies, while a halo was formed after clearing the Baker's yeast cells (Figure 1a). In contrast, the growth on CY agar was slower and the colony appeared a darker orange due to swarming and some cell aggregates ( Figure 1b). The strain produced a thin and transparent swarm with ripples and a flare-shaped pattern at the colony edges in VY/2 agar (Figure 1c), while veins were commonly produced in Casitone-containing CY agar ( Figure 1d). Swarms remained on the surface of the agar with no diffusing pigment. Yellowish-to-orange, hard, coral-or horn-shaped fruiting bodies were observed in standard, nutrient-lean VY/2 and water agars, and measured 250-543.5 µm, commonly visible to the naked eye (Figure 1e,f). These unique fruiting body structures were not observed in CY and CY-H agars, but instead were replaced by orange cell aggregates. Fruiting bodies contained tightly packed, rounded, and optically refractile myxospores with a thick coat, and measured 1.3-2.2 µm in diameter ( Figure 1g). In contrast, the vegetative cells were nonmotile, phase-dark, flexuous, and nearly spindle-shaped rods that measured 4.0-7.6 µm ( Figure 1h). All these growth stage characteristics fit within the genus Corallococcus [23].
VY/2 and water agars, and measured 250-543.5 µ m, commonly visible to the naked eye (Figure 1e,f). These unique fruiting body structures were not observed in CY and CY-H agars, but instead were replaced by orange cell aggregates. Fruiting bodies contained tightly packed, rounded, and optically refractile myxospores with a thick coat, and measured 1.3-2.2 µ m in diameter ( Figure 1g). In contrast, the vegetative cells were non-motile, phase-dark, flexuous, and nearly spindle-shaped rods that measured 4.0-7.6 µ m ( Figure  1h). All these growth stage characteristics fit within the genus Corallococcus [23].  Strain ZKHCc1 1396 T was catalase-positive, oxidase-negative, and stained Gramnegative. It showed positive API ZYM reactions (+3 to +5) to alkaline phosphatase, C8 esterase lipase, C14 lipase, trypsin, and acid phosphatase, and weak positive reactions (+1 to +2) to C4 esterase, leucine arylamidase, valine arylamidase, cysteine arylamidase, α-chymotrypsin, and naphthol-AS-BI-phosphohydrolase. API ZYM reactions were negative (0) to α-and β-galactosidase, β-glucuronidase, α-and β-glucosidase, N-acetyl-βglucosaminidase, α-mannosidase, and α-fucosidase. In the API ® Coryne, only the gelatin hydrolysis and alkaline phosphatase exhibited positive reactions, which were both similar to those from the same tests using the API ZYM. All these differentiating characteristics with Corallococcus type strains are summarized in Table 1.  Strain ZKHCc1 1396 T exhibited colony growth at 18-35 • C but not at the higher temperatures tested (37 and 40 • C). The optimal growth of the strain was observed at 35 • C; this differs from most of the Corallococcus type strains except for AB050A T , which shows nearly the same pattern. The pH range of the isolate was determined between pH 5.5-10 and was optimal at pH 6.0-8.5, which is in the bracket for most Corallococcus type strains (Table 1).
Cellulose powder, filter paper, chitin, and agar were not degraded or digested, suggesting that the new strain lacks cellulolytic, chitinolytic, and agarolytic activity, respectively. The lysis of the tested bacteria and yeast is not surprising, as it is a common characteristic of the genus Corallococcus and many other myxobacteria in the family Myxococcaceae.
The major cellular fatty acids of strain ZKHCc1 1396 T were iso-C 17:0 2-OH (31.0%), iso-C 15:0 (15.8%), iso-C 17:1 (11.7%), and iso-C 17:0 (9.4%) ( Table 2). The remarkably high amount of branched-chain fatty acids (94.2%) over the straight-chain type agrees with previous myxobacterial fatty acid studies on the genus Corallococcus, and thus differentiates it from the related genera Myxococcus, Pyxidicoccus (Garcia et al., 2011), and Simulacricoccus ). Among the branched-chain fatty acids, iso-C 15 : 0 , iso-C 16 : 0 , iso-C 17 : 0 , and iso-C 17:1 were found to be the most abundant, accounting to 15.8%, 5.6%, 9.4%, and 11.7%, respectively. Moreover, iso-C 15:0 was found in all Corallococcus type species (Table S1), and seemed to be one of the major fatty acids in this genus, and as well as in the whole Myxococcaceae family [29]. The overall fatty acid patterns and their major types indicates that strain ZKHCc1 1396 T belongs to the genus Corallococcus, but differs from other type species in their fatty acid quantities.  The amplified, almost complete 16S rRNA gene sequence of strain ZKHCc1 1396 T was about 1501 bp, while the complete sequence obtained from the genome was 1536 bp, and was determined as identical. Based on the complete 16S rRNA gene sequence, the closest type strain similarities were Corallococcus terminator CA054A T (99.67%) and C. praedator CA031B T (99.17%). Phylogenetic analyses revealed that strain ZKHCc1 1396 T clustered within the Corallococcus clade, but formed a separate branch with the closest type strains (Figure 2). According to the PROKKA annotation, the assembled draft genome of strain ZKHCc1 1396 T (GenBank accession No. JAAIYO000000000) consisted of 9,437,609 bp and was characterized by 69.8 %mol G + C content. The genome was predicted to contain 7535 genes comprising 7453 protein-coding genes, 66 tRNA genes, three rRNA genes, and one copy each of the 5S rRNA, 16S rRNA, and 23S rRNA gene. No signs of contamination were found in the genome based on one copy of the 16S rRNA gene. In contrast, the number of genes, proteins, and RNAs varied in number based on the NCBI PGAP annotation pipeline. For comparison with the type strains, the PGAP annotation was used since all data are available in NCBI (Table 1). Strain ZKHCc1 1396 T differs among type strains of other Corallococcus type species by having the smallest genome size and the least number of genes, proteins, and tRNAs ( Table 1).
The phylogenomic tree supports the novelty of strain ZKHCc1 1396 T as it forms a novel branch in the Corallococcus clade. The closest species type strain appears to be Corallococcus praedator CA031B T and C. terminator CA054A T (Figure 3). Furthermore, the difference of the isolated myxobacterium is indicated by the ANI and dDDH values ( Table  3). All type strains compared have ANI values lower than the species cut-off (95%), and with dDDH scores below the 70% threshold value. According to the PROKKA annotation, the assembled draft genome of strain ZKHCc1 1396 T (GenBank accession No. JAAIYO000000000) consisted of 9,437,609 bp and was characterized by 69.8 %mol G + C content. The genome was predicted to contain 7535 genes comprising 7453 protein-coding genes, 66 tRNA genes, three rRNA genes, and one copy each of the 5S rRNA, 16S rRNA, and 23S rRNA gene. No signs of contamination were found in the genome based on one copy of the 16S rRNA gene. In contrast, the number of genes, proteins, and RNAs varied in number based on the NCBI PGAP annotation pipeline. For comparison with the type strains, the PGAP annotation was used since all data are available in NCBI (Table 1). Strain ZKHCc1 1396 T differs among type strains of other Corallococcus type species by having the smallest genome size and the least number of genes, proteins, and tRNAs ( Table 1).

Comparison and Networking of the Secondary Metabolite Biosynthetic Gene Clusters (BGCs)
All of the eleven Corallococcus strains show a 100% similarity score for the geosmin gene cluster (Table 4), while nine strains exhibited 100% for the rhizomide A/rhizomide B/rhizomide C gene cluster. However, based on AntiSMASH evaluation of all the genomes of the Corallococcus type strains, none of them contained the corallopyronin BGC, including the type strain of Corallococcus coralloides DSM 2259 T . High similarity scores (≥60%) were found in all strains for the BGC of VEPE/AEPE/TG-1, and nine strains for a carotenoid and myxochelin A/myxochelin B. Our results are in accordance with a previous study by Ahearne et al. [64], which suggested that all of the myxobacterial strains from the Myxococcaceae family contained the BGC of geosmin, VEPE/AEPE/TG-1, and a carotenoid. Strain ZKHCc1 1396 T appears to be closely similar in the BGC pattern of Corallococcus terminator CA054A T , but lacks the BGC for icosalide A/icosalide B.  In the nonribosomal peptide synthetase (NRPS) gene cluster, one BGC of strain ZKHCc1 1396 T formed a cluster with numerous edges with the BGCs of other Corallococcus type strains together with the BGC of VEPE/AEPE/TG-1 from the MIBiG (minimum information about a biosynthetic gene cluster) database (Figure 4). Two NRPS gene clusters of strain ZKHCc1 1396 T were found to have one edge. In the type I PKS (polyketide synthases) gene cluster, one BGC of strain ZKHCc1 1396 T had two edges, while in the other PKS gene cluster, it possessed three edges. Strain ZKHCc1 1396 T formed seven gene cluster families (GCFs) with the other strains in the NRPS-PKS hybrid gene clusters, whereas five GCFs were created in the RiPPs (ribosomally synthesized and post-translationally modified peptides) gene cluster that contained BGC of strain ZKHCc1 1396 T . One GCF in the terpene gene cluster comprised one of the BGCs of strain ZKHCc1 1396 T , which was connected with the some BGCs of other Corallococcus type strains, and a carotenoid gene cluster from MIBiG database. For the other BGCs, there were two GCFs containing the BGC of strain ZKHCc1 1396 T . Overall, from the analysis of gene cluster network using the BiG-SCAPE platform, the BGCs of strain ZKHCc1 1396 T could form one or more GCFs to the other type strains of Corallococcus species in various types of BGCs. the terpene gene cluster comprised one of the BGCs of strain ZKHCc1 1396 T , which was connected with the some BGCs of other Corallococcus type strains, and a carotenoid gene cluster from MIBiG database. For the other BGCs, there were two GCFs containing the BGC of strain ZKHCc1 1396 T . Overall, from the analysis of gene cluster network using the BiG-SCAPE platform, the BGCs of strain ZKHCc1 1396 T could form one or more GCFs to the other type strains of Corallococcus species in various types of BGCs. From the various secondary metabolite production media, the cultivation in P medium was shown to have the most bioactive crude extract against Gram-positive bacteria ( Figure 5). The BPC chromatogram of strain ZKHCc1 1396 T obtained from cultivation and extraction in P medium produced more than twenty high peaks (above the 90% relative intensity compared to the highest peak) in the range of 1.8-20 min. (Figure 6). The detected ion masses ranged from 209.1645 Da to 1371.9401 Da (Table 5), suggesting the presence of diverse compounds produced by strain ZKHCc1 1396 T . Interestingly, no hit compound from DNP was found from the genus Corallococcus. Two hits were detected with similar masses to myxobacterial species Chondromyces crocatus. Analysis of fraction 19 showed no hits in the DNP database, suggesting the possibility of discovering a novel compound from strain ZKHCc1 1396 T . Further study is needed to confirm the compounds produced by strain ZKHCc1 1396 T , which can be conducted by the isolation and structure elucidation of the compounds. From the various secondary metabolite production media, the cultivation in P medium was shown to have the most bioactive crude extract against Gram-positive bacteria ( Figure 5). The BPC chromatogram of strain ZKHCc1 1396 T obtained from cultivation and extraction in P medium produced more than twenty high peaks (above the 90% relative intensity compared to the highest peak) in the range of 1.8-20 min. (Figure 6). The detected ion masses ranged from 209.1645 Da to 1371.9401 Da (Table 5), suggesting the presence of diverse compounds produced by strain ZKHCc1 1396 T . Interestingly, no hit compound from DNP was found from the genus Corallococcus. Two hits were detected with similar masses to myxobacterial species Chondromyces crocatus. Analysis of fraction 19 showed no hits in the DNP database, suggesting the possibility of discovering a novel compound from strain ZKHCc1 1396 T . Further study is needed to confirm the compounds produced by strain ZKHCc1 1396 T , which can be conducted by the isolation and structure elucidation of the compounds.         Corallococcus soli (so'li. L. gen. n. soli, of soil, referring to a myxobacterium isolated from a soil sample collected in Iran).

Conclusions
Based on polyphasic characterizations, which include morphological, physiochemical, and genomic studies, strain ZKHCc1 1396 T appears to represent a novel species of Corallococcus, for which we propose the name Corallococcus soli sp. nov. (type strain ZKHCc1 1396 T (=NCCB 100659 T = CIP 111634 T )).

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms10071262/s1. Two supplementary tables are available with the online version of this article. Table S1: Differentiating fatty acid profile of strain ZKHCc1 1396 T against all Corallococcus species. Table S2: Media and ingredients used in this study.