Strain s54d21^T was isolated from a sediment sample collected from Xinghu National Wetland Park (N 23°4′54″, E 112°28′17″) in Guangdong Province of China. The strain was isolated following the baiting approach with E. coli as prey on water agar (0.1% CaCl₂∙2H₂O, 20 mM HEPES, 1.5% agar), where small portions of soil samples were placed adjacent to the E. coli spot. Until swarming colonies or fruiting bodies observed after incubation at 30°C, purification of the isolate was subsequently completed by repeatedly transferring onto fresh VY/2 agar (0.5% dried baker’s yeast, 0.1% CaCl₂∙2H₂O, 1.5% agar). Finally, the pure strain characterized orange-pigment fruiting bodies on agar and was deposited in Guangdong Microbial Culture Collection Center (GDMCC, No. 1.1945) and Japan Collection of Microorganisms (JCM, no. 39263).

The swarm morphology was observed under a stereomicroscope (Olympus SZX10, Japan). The vegetative cells and myxospores were investigated by using scanning electron microscope (SEM, Hitachi S-3000 N, Japan) and the phase contrast microscope (ZEISS Axioscope 5, Germany), respectively. The Gram reaction was determined using a Gram staining kit (Huankai, China), while Conga red staining was performed as described by McCurdy (1969). The oxidase test was determined using a commercial strip (Huankai, China), while the catalase test was performed using 3% (v/v) H₂O₂. The growth temperature range was determined on VY/2 agar at different temperatures (4, 10, 15, 20, 25, 30, 37, 40, and 42°C). The pH range (4.0–10.0, at intervals of 1 pH unit) for growth was tested on VY/2 agar at 30°C for 14 days and buffered with 100 mM citrate/sodium citrate (pH 4.0–5.0), HEPES (pH 6.0–8.0), and Tris (pH 9.0–10.0). Salt tolerance was evaluated by growing the isolated on VY/2 agar plates supplemented with NaCl to concentrations of 0.5, 1, 1.5, 2, and 3% (m/v). Susceptibility to antibiotics was investigated on VY/2 agar supplemented with various antibiotics at the final concentration of 50 μg/mL at 30°C for 14 days. Sixteen antibiotics were selected for strain s54d21^T: ampicillin, apramycin, spectinomycin, polymyxin B, neomycin, bacitracin B, gentamicin, tetracycline, erythromycin, oxytetracycline, chloramphenicol, nalidixic acid, trimethoprim, streptomycin, rifampin, and kanamycin. Anaerobic growth was tested on VY/2 agar plates in an anaerobic jar at 30°C for 14 days. Hydrolysis of starch, skimmed milk, colloidal chitin, and carboxymethycelluose (CMC) was assessed on tryptone agar (0.2% tryptone, 0.05% MgSO₄, 0.01% CaCl₂∙2H₂O, 1.5% agar; pH 7.2) at the final concentration of 1.0% (m/v), respectively. Hydrolysis of Tween 20 and 80 was tested on tryptone agar supplemented with 1.0% (v/v) Tween 20 and 80, respectively. Additional enzyme activities were tested using the API ZYM and 20NE kits (bioMérieux, France), according to the manufacturer’s instructions.

Genomic DNA of strain s54d21^T was extracted by using Genomic DNA Isolation Kit (Magen, China). The 16S rRNA gene sequence was amplified using the universal primers 27F/1492R (Weisburg et al., 1991). The PCR products were sequenced by Shanghai Majorbio Biopharm Technology Co., Ltd. Sequence alignment was performed at online EZBioCloud server (Yoon et al., 2017a) and NCBI. Phylogenetic tree based on the 16S rRNA gene sequences was reconstructed using software MEGA X (Kumar et al., 2018) with the maximum likelihood (ML) method (Felsenstein, 1981) under the Kimura’s two-parameter model (Kimura, 1980). Bootstrap analysis was conducted based on 1,000 replicates (Hoang et al., 2018).

Genomic sequencing was carried out using an Illumina Hiseq platform at Shanghai Majorbio Biopharm Technology Co., Ltd. The obtained reads were assembled into contigs using SPAdes v3.11.1 (Bankevich et al., 2012). Genome contamination and completeness were assessed using CheckM tool (Parks et al., 2015). Genome was annotated using the Prokka (Seemann, 2014) and RAST server (Aziz et al., 2008). The biosynthetic gene clusters of secondary metabolites were predicted using antiSMASH 6.0 with the default settings (Blin et al., 2021). The carbohydrate activities were identified using the dbCAN2 meta server (Zhang et al., 2018). Average nucleotide identity (ANI) and digital DNA–DNA hybridization (dDDH) values among strain s54d21^T and the closely related type strains were calculated using an online ANI calculator and the Genome-to-Genome Distance Calculator 3.0 (Meier-Kolthoff et al., 2013; Yoon et al., 2017b). Phylogenomic tree was constructed using the up-to-data bacterial core gene (UBCG) pipeline based on 92 core housekeeping genes (Na et al., 2018).

Respiratory quinones of the strain s54d21^T were extracted and analyzed as previously described using HPLC (Agilent 1260) (Collins et al., 1977). For the fatty acid detection, strain s54d21^T and its closely related type strains were incubated on VY/2 agar plates at 30°C for a week. Fatty acids were extracted and identified according to the described protocol using a gas chromatography mass spectrometry (GCMS, Agilent 7890B-5977B) (Gemperlein et al., 2014).

The nearly complete 16S rRNA gene sequence of the strain s54d21^T was obtained with the length of 1,536 bp (GenBank accession number: OR885464). Pairwise comparison of 16S rRNA gene revealed that the strain s54d21^T shared the highest similarity to H. minutum DSM 14724^T (98.2%), followed by H. versicolor H56D21^T (97.8%) and H. gracilis DSM 14753^T (97.7%). Phylogenetic analysis based on 16S rRNA gene sequence showed that the strain s54d21^T located in the genus Hyalangium and formed an individual branch containing all type strains of Hyalangium genus (Figure 1).

The draft genome of the strain s54d21^T was 10.77 Mb in size with 43 contigs, the N50 of 586,569 bp, and the DNA G + C content of 68.5 mol% (Supplementary Table S1). The genome contained 8,621 protein-coding sequences, 70 tRNAs, and 3 rRNAs genes. The UBCG-based phylogenomic tree showed that the strain s54d21^T was closely related to all type species of Hyalangium (Figure 2), which was consistent with the tree topology based on the 16S rRNA genes. The calculated genome similarity values between the strain s54d21^T and other type species in the genus Hyalangium were in the ranges of 81.3–82.7% for ANI and 24.4–25.7% for dDDH (Supplementary Table S2), which were all below the threshold values for species delineation (95.0–96.0% for ANI and 70.0% for dDDH) (Goris et al., 2007; Richter and Rosselló-Móra, 2009). These results suggested that the strain s54d21^T should be considered as a novel species in the genus Hyalangium, for which the name Hyalangium ruber sp. nov was proposed.

The RAST annotation result showed that almost 81.0% of coding sequences in the strain s54d21^T were not assigned to subsystems. The general feature of subsystem category distribution in the strain s54d21^T genome was similar to those in its closest phylogenetic neighbor. No genes involved in motility and chemotaxis were found to be encoded by the strain s54d21^T genome and in the genomes of closely related type strains. However, genes associated with iron acquisition and metabolism were found in the strain s54d21^T more than these in the three closely related strain genomes (Supplementary Figure S1). Based on CAZy database annotation, the genome of the strain s54d21^T contained 23 carbohydrate-binding modules, 33 carbohydrate esterases, 78 glycoside hydrolases, 61 glycosyl transferases, 6 polysaccharide lyases, and 15 auxiliary activities (Supplementary Figure S1). There were most types of carbohydrate enzymes involved in glycoside hydrolases, which can hydrolyze polysaccharides such as starch and peptidoglycan. A total of 32 biosynthetic gene clusters (BGCs) were identified in the strain s54d21^T genome, including three NRPS-like fragments, a PKS-type cluster, four hybrid PKS/NRPS-type cluster, nine RiPP-type clusters, five terpenes-type clusters, two lanthipeptides-type clusters, and others (Supplementary Figure S1).

The swarms of the strain s54d21^T exhibited circular, red, thin film with wavy flared edges on VY/2 agar (Figure 3A). Vegetative cells were slender rods with tapered ends measuring 0.5–0.7 × 3–7 μm (Figure 3B). Sporangioles were spherical, some of which were empty and transparent (Figure 3C). Myxospores were irregular spherical (Figure 3C). Strain s54d21^T showed positive to oxidase, catalase, starch, aesculin hydrolysis, gelatin hydrolysis, alkaline phosphatase, esterase (C4), esterase lipase (C8), lipase (C14), leucine arylamidase, valine arylamidase, acid phosphatase, naphtol-AS-Bl-phosphoamidase, β-glucosidase, and N-acetyl-β-glucosaminidase. The differentiating characteristics with other type strains in the genus Hyalangium are shown in Table 1. Strain s54d21^T could be distinguished from its closely related type strains in temperature for growth, β-galactosidase, cystine arylamidase, trypsin, α-chymotrypsin, β-glucosidase and several antibiotic sensitivities.

The major respiratory quinone detected in the strain s54d21^T was menaquinone-8 (MK-8). The major fatty acids (>5% of the total amounts) contained iso-C_17:0 2-OH (27.8%), iso-C_15:0 (17.4%), iso-C_15:0 DMA (8.2%), iso-C_16:0 (6.8%), C_16:1ω5c (6.6%), and iso-C_15:0 3-OH (6.2%) (Supplementary Table S3). The high levels of iso-C_15:0, C_16:1ω5c, and iso-C_15:0 DMA were found in all type strains in the genus Hyalangium and seemed to be the major fatty acid in this genus. Moreover, the remarkably high amount of iso-C_17:0 2-OH and the low proportions of C_16:1ω5c, C_16:0, C_16:1ω7c, and iso-C_17:0 in the strain s54d21^T could distinguish the strain s54d21^T form its neighbor-type species in the genus Hyalangium.

Hyalangium ruber (ru′ber. L. masc. Adj. ruber red, referring to the color of the colonies).

Cells are gram-negative. Growth occurred at 20–40°C, pH 4.0–10.0 with the NaCl tolerance of 0–0.5% (m/v). Resistant to ampicillin, gentamicin, kanamycin, polymyxin B, and bacitracin B, sensitive to erythromycin, nalidixic acid, rifampin, spectinomycin, streptomycin, tetracycline, apramycin, chloramphenicol, neomycin, trimethoprim, and oxytetracycline. Major cellular fatty acids are iso-C_17:0 2-OH, iso-C_15:0, iso-C_15:0 DMA, iso-C_16:0, C_16:1ω5c, and iso-C_15:0 3-OH. Predominant respiratory quinone is MK-8. The genomic DNA G + C content of the type strain is 68.5 mol%.

The type strain s54d21^T (= GDMCC 1.1945^T = JCM 39263^T) was isolated from the sediment sampled from Xinghu National Wetland Park located in Guangdong Province, China. The GenBank accession numbers for the 16S rRNA gene and whole genome sequences of the type strain are OR885464 and JAXIVS000000000, respectively.

The gene sequence analyses of the strain s54d21^T supposed the production of polyketides and peptides in terms of the presence of PKS and NRPS BGCs. Therefore, the chemical investigation of the culture broth of the strain was performed as described. Here, the structure including absolute configurations of three new compounds (1–3) was elucidated (Figure 4).

Compound 1 was obtained as an amorphous solid, and the molecular formula of C₁₀H₁₆N₂O was determined by the positive mode HRESIMS with a [M + Na]⁺ ion peak at m/z 203.1150, indicating four degrees of unsaturation. The 1D NMR data (Table 2) and HSQC spectrum presented signals of ten carbons, including four methyls (a singlet at δ_H 2.22, two doublets at δ_H 1.20, and a triplet at δ_H 1.17), a methylene at δ_H 2.57 (q, J = 7.5 Hz) and δ_C 26.1, a methine at δ_H 3.34 (m) and δ_C 30.9, and four unprotonated carbons. ¹H–¹H COSY correlations demonstrated the existence of two spin systems of an ethyl group and an isopropyl moiety (Figure 5). Combination analyses of the chemical formula and HMBC correlations (Figure 5) revealed that four unprotonated carbons (C-2, C-3, C-5, and C-6) and two nitrogens were assigned to the 2(1H)-pyrazinone ring. The comparison of those spectroscopic data suggested that structure of 1 was such close to those knowns featuring a 3,5,6-substitued 2(1H)-pyrazinone core structure (Motohashi et al., 2011; Ma et al., 2020). The differences between 1 and known compounds were three substituent groups located at C-3, C-6, and C-5, including an abovementioned ethyl group, an isopropyl group, and a methyl (Me-10), respectively. Subsequent analyses of HMBC spectrum supported those assignments for structure 1, where key HMBC correlations from H-7/Me-8 to C-2, H-7/Me-8/Me-10 to C-3, Me-10/H-11 to C-5, and Me-10/Me-12 to C-6 were observed (Figure 5). Hence, compound 1 was unambiguously determined as a 3,5,6-trisubstituted 2(1H)-pyrazinone and designated as hyalanone A.

Hyalanone B (2) was also isolated as an amorphous solid. The molecular formula was demonstrated to be C₁₀H₁₆N₂O₂ via a positive HRESIMS ion peak at m/z 219.1105 [M + Na]⁺. The ¹H and ¹³C NMR data of 2 was such close to those of 1, except for an oxygenated methine at δ_H 4.89 (q, J = 6.5 Hz, H-11) and δ_C 66.6 (C-11) (Table 2), suggesting a similar 2(1H)-pyrazinone skeleton. A hydroxyethyl group replaced the ethyl at C-6 of 1, which was further confirmed by H–¹H COSY cross-peaks and key HMBC correlations from Me-12 to C-6/C-11 and H-11 to C-5 (Figure 5). There is not enough compound 2 for Mosher’s method to determine the absolute configuration of the chiral center C-11. Therefore, it was inferred from theoretical ECD calculation by using TDDFT methodology at PBE0/def2-TZVP level in CPCM acetonitrile using ORCA5.0.1 (Neese, 2012, 2017). As shown in Figure 6, the experimental ECD curve of 2 was well matched with the calculated one that concluded the absolute configuration of C-11 to be R.

The molecular formula C₁₁H₁₈N₂O₂ of 3 was deduced from a HRESIMS spectrum. Analyses of 1D NMR data (Table 2) and HMBC correlations (Figure 5) of 3 exhibited a highly similarity with those of 2, where the unique difference between them was an extra methoxyl at δ_H 3.22 and δ_C 55.9 (11-OMe) in 3. Moreover, key HMBC correlation from 11-OMe to C-11 indicated the location of a methoxyl at C-11 of 3 rather than a hydroxyl in 2 (Figure 5). By comparing of experimental ECD curves of 2 and 3 (Figure 6), the absolute configuration of C-11 was still approved to be R. Therefore, compound 3 was established and given the name hyalanone C.

2(1H)-Pyrazinones are a type of natural products originally biosynthesized by NRPS gene cluster with two amino acids (Wyatt et al., 2010; Zimmermann and Fischbach, 2010; Wyatt et al., 2012). Based on biosynthetic pathways of pyrazinone analogs, three new compounds are proposed to be derived from the valine and alanine or threonine building blocks. Analyses of the genome sequence of strain s54d21^T by antiSMASH showed three NRPS-like and three PKS-NRPS BGCs. Among them, the NRPS (Region 1.3) and PKS-NRPS (Region 5.2) gene clusters containing a Val module are potent BGCs related to the biosynthesis of those new pyrazinones (Supplementary Figure S2).

The previous report of pyrazinone derivatives described their remarkable cytotoxicity against human cancer cell lines and antimicrobial activity (Jansen et al., 2014). Therefore, cytotoxic activities against human cancer cell lines and antimicrobial and anti-inflammatory effects of all isolates were tested in this study. The cytotoxicity evaluation of compounds 1–4 was performed using the MTS method, where taxol was used as the positive control. Among them, compound 4 displayed moderate cytotoxicity against A549, RKO, and NCM460 with IC₅₀ values at 32.14 ± 2.25, 27.31 ± 8.96, and 31.33 ± 1.71 μM (positive control with IC₅₀ values <0.008 μM), respectively. In addition, their inhibition of NO production in LPS-induced RAW 264.7 cells was tested. All compounds showed weak inhibitory effect (< 10.0%) of NO production at the concentration of 40 μM. The evaluation of their antimicrobial activities against S. aureus, E. coli, P. aeruginosa, and E. faecalis did not show obvious inhibition at the concentration of 200 μM.

Pyrazinones are previously reported to regulate phenotypic changes in the pathogen (e.g., quorum sensing) or act directly on the host (e.g., virulence factors), such as tyrvalin and phevalin are used as regulators of virulence factor gene expression (Wyatt et al., 2010; Secor et al., 2012). Thus, although three new pyrazinones do not exhibit obvious effects in the evaluation of cytotoxic, antimicrobial, and anti-inflammatory activities, the discovery of new molecules provides potential for further research studies of relationships between the small molecule and microorganisms as well as the role of myxobacteria in nature.