Microcystis aeruginosa FACHB 905 culture was obtained from the Freshwater Algae Culture Collection at the Institute of Hydrobiology (FACHB collection), Chinese Academy of Sciences (Wuhan, China; http://algae.ihb.ac.cn/Default.aspx), and incubated in Hughes Gorham Zehnder (HGZ) medium as described previously (Zhang et al., 2015). To obtain pure cultures of attached bacteria, 0.2-ml aliquots of Maf culture were spread onto International Streptomyces Project 2 (ISP 2) medium and incubated at 28.0°C for 2–5 days. The isolated strain JXJ CY 35^T was stored on ISP 2 at 4.0°C and as glycerol suspensions (30%, v/v) at −80.0°C.

Strain JXJ CY 35^T was cultured using ISP 2, tryptic soy agar (TSA), Middlebrook 7H10 agar, and Löwenstein–Jensen medium (Jensen, 1932) at 28.0°C. Cell morphology was observed using light microscopy (BX53; Olympus) and transmission electron microscope (JEM-2100, JEOL) from a culture grown on ISP 2 at 28.0°C for 3 days. Gram staining was done by using the standard Gram's staining procedure. Catalase activity was determined as described previously by Zhang et al. (2016b). Oxidase activity and motility were performed using standard methods (Kovacs, 1956; Xu et al., 2005). Acid-alcohol-fastness was examined according to the study by Dong and Cai (2001). Physiological and biochemical tests, including hydrolysis of Tweens (20, 40, and 80), starch and cellulose, and nitrate reduction, were done according to the study by Smibert and Krieg (1994). Anaerobic growth was determined after incubation on ISP 2 for 14 days at 28.0°C using the GasPak EZ Anerobe Pouch System (BD). Growth at different temperatures (5.0, 10.0, 15.0, 20.0, 28.0, 30, 32, 34.0, 36.0, 38.0, and 45.0°C), pH range (2.0–12.0 with interval of 1.0), and NaCl contents (0–10.0%, w/v, with interval of 1%) were tested using ISP 2 medium as the basal growth medium (Zhang et al., 2017). Other biochemical characteristics including various enzyme activities were further tested using the API ZYM, API 20NE, and API 50CH systems according to the manufacturer's instructions. Sensitivity to antibiotics was tested using a paper disk according to the method described by Zhang et al. (2016c).

Cells of strain JXJ CY 35^T for chemical analysis were obtained from pure cultures grown on ISP2 agar at 28.0°C for 3 days. Polar lipids were extracted and analyzed according to the methods described by Minnikin et al. (1979). Fatty acid analysis was performed using the Sherlock Microbial Identification System (MIDI). Mycolic acids were analyzed using thin-layer chromatography (Minnikin et al., 1980). Menaquinones were extracted and analyzed according to the methods described by Tindall (1990). The peptide subunit of peptidoglycan and whole-cell sugars were detected according to the methods described by Hasegawa et al. (1983) and Tang et al. (2009).

Extraction of genomic DNA and PCR amplification of the 16S rRNA gene sequence were done as described by Peng et al. (2006) after being cultured on ISP 2 medium at 28.0°C for 3 days. The 16S rRNA gene sequence of the isolate was compared with sequences from EzBioCloud (https://www.ezbiocloud.net/) server databases (Yoon et al., 2017). Phylogenetic trees were generated by neighbor-joining (Saitou and Nei, 1987), maximum-parsimony (Fitch, 1971), and maximum-likelihood (Felsenstein, 1981) algorithms using MEGA version 5.0 (Tamura et al., 2011). The topologies of the resultant trees were evaluated using bootstrap analysis (Felsenstein, 1985) with 1,000 replications.

The genome sequencing, assembling, and annotation of strain JXJ CY 35^T were performed at Sangon Biotech (Shanghai, China) using the Illumina HiSeq 4000 platform. The genome was assembled using SPAdes version 3.5.0 (Bankevich et al., 2012) and corrected using PrInSeS-G. Gene prediction was performed using Prokka software tool (Seemann, 2014). Genomic annotation was performed using NCBI version Blast 2.2.28 with the default parameters. The digital DNA-DNA hybridization (dDDH) and average nucleotide identity (ANI) values between strain JXJ CY 35^T and the related strains were calculated using the Genome-to-Genome Distance Calculator (GGDC, version 3.0) (http://ggdc.dsmz.de/ggdc.php) (Meier-Kolthoff et al., 2013) and JSpeciesWS (JSWS) website (http://jspecies.ribohost.com/jspeciesws/#analyse), respectively. The genomes of type strains M. arabiense JCM 18538^T, M. goodii ATCC 700504^T, M. mageritense DSM 44476^T, M. austroafricanum DSM 44191^T, and Mycobacterium neglectum CECT 8778^T were downloaded from the NCBI genome database (https://www.ncbi.nlm.nih.gov/genome/).

The ability of nitrogen-fixing was determined by culturing in the nitrogen-free medium (glucose, 10 g; KH₂PO₄, 0.2 g; MgSO₄·7H₂O, 0.2 g; NaCl, 0.2 g; CaSO₄·2H₂O, 0.2 g; deionized water, 1,000 ml; pH, 7.0–7.2) at 28°C for 2 days, and the cell density was determined by spread-plate counting. The ability to dissolve insoluble phosphorus was determined according to the method described by Zhang et al. (2016b). Both Ca₃(PO₄)₂ and phytin were used as the insoluble phosphorus at the dosages of 5 g/L.

To understand the interactions between strain JXJ CY 35^T and Maf, we performed the following tests. A pure culture of Maf was obtained, examined, and cultured according to the methods described previously (Zhang et al., 2016b). Maf (about 2 × 10⁶ CFU/ml) was inoculated with strain JXJ CY 35^T at three different inoculums (1 × 10⁵, 1 × 10⁶, and 1 × 10⁷ CFU/ml). The pure culture of Maf served as the control. All of the coculture and controls were tested in triplicates and cultured, as described above, and sampled on days 5 and 10 of cultures. Bacterial densities of the samples were determined by the method of spread-plate counting. The contents of chlorophyll a (chl-a), extracellular microcystin LR (E-MC-LR), and intracellular microcystin LR (I-MC-LR) of the samples were determined according to the methods described previously (Zhang et al., 2015).

We performed the following tests to confirm if strain JXJ CY 35^T could promote the growth of Maf in the absence of available P and N. Maf (2 × 10⁶ CFU/ml) and strain JXJ CY 35^T (1 × 10⁶ CFU/ml) were cocultured using revised HGZ medium as described above. In the revised HGZ medium, KH₂PO₄ was replaced by Ca₃(PO₄)₂, or NaNO₃ was removed (nitrogen-free). The contents of chl-a, I-MC-LR, and E-MC-LR of the cultures, as well as the cell density of strain JXJ CY 35^T were determined as described above on day 12 of culture. Pure cultures of Maf (2 × 10⁶ CFU/ml) in the revised HGZ medium served as the controls.

Strain JXJ CY 35^T was gram-positive, aerobic, and fast-growing. The colonies were orange in color, moist, smooth, and round. It grows well on ISP2 and TSA but poorly on Middlebrook 7H10 agar and Löwenstein–Jensen medium. The cells were short rods with a size of 0.6–0.9 × 1.2–2.4 μm (Figure 1). Growth occurred at 10.0–36.0°C, pH 4.0–10.0, and 0–5.0% (w/v) NaCl, with optimal growth at 28.0°C, pH 7.0–8.0, and 0% (w/v) NaCl. The isolate was found to be positive for catalase and negative for oxidase. It can hydrolyze Tweens (20, 40, and 80) and starch but cannot hydrolyze cellulose. Other differential phenotypic characteristics of strain JXJ CY 35^T and the related type strain M. arabiense DSM 45768^T are listed in Table 1, and the detailed characteristics are given in the type strain-description.

Strain JXJ CY 35^T was found to be sensitive to tetracycline (30.0 μg), rifampicin (5.0 μg), cefoxitin (30.0 μg), gentamicin (10.0 μg), kanamycin (30.0 μg), erythromycin (15.0 μg), neomycin (30.0 μg), chloramphenicol (30.0 μg), streptomycin (300.0 μg), and neomycin (5.0 μg), and it was resistant to ampicillin (10.0 μg), carboxylbenzicillin (100.0 μg), amoxicillin (10.0 μg), cefotaxidine (30.0 μg), and bacitracin (0.04 IU).

The polar lipids of strain JXJ CY 35^T were found to be diphosphatidylglycerol (DPG), phosphatidylethanolamine (PE), glycolipid (GL1, 2, 3), phosphoglycolipid (PGL), phosphatidylinositol (PI), and unidentified lipid (L1) (Supplementary Figure S1). The mycolic acids showed obvious differences from those of the type strain M. arabiense DSM 45768^T (Supplementary Figure S2). Predominant cellular fatty acids (>10.0%) were identified as C_17:1ω7c (37.0%) and C_18:1 ω9c (18.9%). However, the main cellular fatty acids (>10.0%) of the reference strain M. arabiense DSM 45768^T were C_16:0 (15.9%), C_17:1ω7c (17.5%), and C_18:1ω9c (31.7) (Table 1). The predominant menaquinone was found to be MK9. The peptidoglycan of strain JXJ CY 35^T contained aspartic acid, glutamic acid, glycine, and alanine, with mannose, ribose, galactose, and arabinose as whole-cell sugars.

Analysis of the almost-complete 16S rRNA gene sequence (1,514 bp; GenBank accession number: MW723388) indicated that strain JXJ CY 35^T is a member of the genus Mycolicibacterium. It shared 16S rRNA gene sequence similarities of 98.2% with M. arabiense DSM 45768^T (Zhang et al., 2013) and <98.2% with other members of the genus of Mycolicibacterium. The strain formed a distinct clade using three treeing methods (Figure 2; Supplementary Figures S3, S4). However, it formed a clade with M. madagascariense P2^T in three trees using all validly published species of the genus Mycolicibacterium. The digital DNA-DNA hybridization (dDDH) and ANI values between strain JXJ CY 35^T and the closest five type strains Mycolicibacterium arabiense JCM 18538^T, M. goodii ATCC 700504^T, M. mageritense DSM 44476^T, M. austroafricanum DSM 44191^T, and Mycobacterium neglectum CECT 8778^T were 52.1, 20.3, 20.3, 20.6, and 19.8%, and 92.7, 75.5, 75.6, 76.0, and 75.2%, respectively, based on the genomes from the NCBI genome database. Both values were lower than the species boundary recommended for distinguishing novel prokaryotic species (Kim et al., 2014; Chun et al., 2018). Therefore, based on the above data, it was concluded that strain JXJ CY 35^T represents a novel species of the genus Mycolicibacterium, for which the name Mycolicibacterium lacusdiani sp. nov. is proposed.

The genome of strain JXJ CY 35^T was sequenced and submitted to GenBank with the accession JAKCFC000000000. Its draft genome contains 24 contigs, with a total length of 6,138,096 bp and an N₅₀ length of 596,067 bp. It has 5,946 protein-coding genes with a total size of 5,701,250 bp (an average of 958.8 bp per protein). Therefore, the coding ratio is 92.9%. The DNA G + C content was determined to be 68.3% based on the genome. No repeat region was found. The numbers of tRNA and rRNA are 50 and 3, respectively. According to the genome (6,017,160 bp; accession AP022593) obtained from the NCBI genome database, M. arabiense JCM 18538^T has some different genomic characteristics from strain JXJ CY 35^T. It has 5,873 protein-coding genes with a total size of 5,598,840 bp (an average of 953.3 bp per protein). The coding ratio is 93.0%. The DNA G + C content is 68.5%. Unlike strain JXJ CY 35^T, M. arabiense JCM 18538^T has 1,225 repeat regions with a total size of 116,611 bp and a repeat ratio of 1.94%. The numbers of tRNA and rRNA are 54 and 6, respectively. The gene annotation rates of strain JXJ CY 35^T and M. arabiense JCM 18538^T in various databases were diverse, which were about 98% in the TrEMBL and NR databases, but only about 41% in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Supplementary Table S1). More genomic characteristics and gene annotation of strain JXJ CY 35^T are shown in Supplementary Figures S5–S9 and Supplementary Tables S1–S5.

Data from the genes annotated by the GO database showed that strain JXJ CY 35^T has three genes related to the arylsulfatase activity, indicating that this strain can potentially produce arylsulfatase. Data from COG database indicated that strain JXJ CY 35^T has a gene coding tellurite resistance protein and related permeases, indicating that it can grow in the presence of potassium tellurite. Strain JXJ CY 35^T has many genes or gene clusters in favor of coexisting with algae. Data from the genes annotated by the GO database showed that strain JXJ CY 35^T has 62 gene clusters related to symbiotic interaction, and immune response and their regulations (Supplementary Table S2), 34 genes related to phosphatases, 4 genes related to nodulation, 8 genes related to nitrogen fixation (Supplementary Table S3), 19 genes related to the synthesis of plant growth hormones (such as indole-3-acetic acid, auxin, abscisic acid, ethylene, and polyamine), 16 gene clusters related to the synthesis of various vitamins (Supplementary Table S4), 7 genes related to ATP-binding cassette (ABC) transporter complex, 8 genes related to protein secretion, and 29 gene clusters related to signaling and signaling pathway (Supplementary Table S5).

The cell density increased initially from 1.1 × 10⁵ to 3.1 × 10⁶ CFU/ml after 2 days of culture in a nitrogen-free medium, indicating that strain JXJ CY 35^T has the nitrogen-fixing ability. The contents of available phosphate were 1.6 ± 0.1 mg/L for Ca₃(PO₄)₂ and 17.1 ± 0.3 mg/L for phytin on day 4 of culture, indicating that strain JXJ CY 35^T can dissolve insoluble phosphate.

The Chl-a contents of the control increased from initial 0.095 mg/L to 0.388 and 0.445 mg/L on days 5 and 10 of cultures, respectively. Strain JXJ CY 35^T showed no significant influence (P > 0.05) on the growth of Maf during the test at a low inoculation dosage (1 × 10⁵ CFU/ml). However, the chl-a contents of Maf increased by 9.0% (P < 0.05) on day 10 of culture at a moderate inoculation dosage (1 × 10⁶ CFU/ml) of strain JXJ CY 35^T and decreased by 14.9% and 7.2% (P < 0.05, P < 0.01) on days 5 and 10 of cultures, respectively, at a high inoculation dosage (1 × 10⁷ CFU/ml) of strain JXJ CY 35^T (Figure 3). Maf in cocultures probably exhibited influences on the growth of strain JXJ CY 35^T. Cell densities of strain JXJ CY 35^T showed no significant difference (P > 0.05) between days 5 and 10 of cultures in a low inoculation dosage (1 × 10⁵ CFU/ml) group and decreased significantly (P < 0.01) on day 10 of culture in other two higher inoculation dosages (1 × 10⁶, 1 × 10⁷ CFU/ml) groups (Table 2).

Figure 4 shows that strain JXJ CY 35^T exhibited no significant influence on the E-MC-LR content of Maf at the inoculation dosages of both 1 × 10⁵ and 1 × 10⁶ CFU/ml during the test. However, the E-MC-LR content of the coculture group inoculated with 1 × 10⁷ CFU/ml of strain JXJ CY 35^T was 67.4 μg/mg chl-a on day 5 of culture (Figure 4), 21.6% higher than that of the control group (P < 0.01). Next, the E-MC-LR content of this coculture group did not increase significantly (P > 0.05) on day 10 of culture, unlike that of the control and the coculture groups inoculated with 1 × 10⁵ and 1 × 10⁶ CFU/ml of strain JXJ CY 35^T (P < 0.05, P < 0.01). Therefore, higher inoculation dosages of strain JXJ CY 35^T could promote the release of I-MC-LR temporarily.

Different inoculation dosages of strain JXJ CY 35^T exhibited different influences on the I-MC-LR content (Figure 4). Low inoculation dosage (1 × 10⁵ CFU/ml) showed no obvious influence on the I-MC-LR content of Maf in coculture. However, the I-MC-LR contents of moderate inoculation dosage (1 × 10⁶ CFU/ml) were 314.6 and 679.6 μg/mg chl-a on days 5 and 10 of cultures, which were 8.7% and 10.8% lower than that of the controls, respectively; the I-MC-LR contents of high inoculation dosage (1 × 10⁷ CFU/ml) were 411.9 and 1,026.1 μg/mg chl-a on days 5 and 10 of cultures, which were 19.5% and 34.6% higher than that of the controls, respectively.

In Ca₃(PO₄)₂ medium, the chl-a content of Maf inoculated with JXJ CY 35^T was 0.170 mg/L (Table 3) on day 12 of culture, which were 13.3% higher than that of the control group (P < 0.05). In a nitrogen-free medium, the chl-a content of Maf inoculated with JXJ CY 35^T was 0.218 mg/L (Table 3), which were 27.5% higher than that of the control group (P < 0.01). Both E-MC-LR and I-MC-LR contents of Maf were significantly influenced (P < 0.01) by strain JXJ CY 35^T in the absence of available N and P (Table 3). In Ca₃(PO₄)₂ medium, the contents of E-MC-LR and I-MC-LR were 235.0 and 625.0 μg/mg chl-a in the group inoculated with JXJ CY 35^T, which were 15.6 and 43.3% lower (P < 0.01) than that of the control group, respectively. In a nitrogen-free medium, the contents of E-MC-LR and I-MC-LR were 109.2 and 439.5 μg/mg chl-a in the group inoculated with JXJ CY 35^T, which were 32.4 and 29.9% lower (P < 0.01) than that of the control group, respectively.

Mycolicibacterium lacusdiani sp. nov. (la.cus.di.a'ni L. gen. n. lacus, of a lake; N.L. gen. n. diani, of Dian; N.L. gen. n. lacusdiani, of Dian lake).

Strain JXJ CY 35^T is gram-positive, aerobic, fast-growing, nonmotile, and short rod-shaped (0.6–0.9 × 1.2–2.4 μm) and grows well on ISP 2 and TSA. The pH, NaCl content (w/v), and temperature range for growth are 4.0–10.0, 0–5.0%, and 10–36°C, respectively, with optimal growth at 7.0–8.0, 0% (w/v), and 28°C. It is positive for hydrolysis of Tweens (20, 40, and 80), nitrate reduction, and catalase and negative for hydrolysis of cellulose and oxidase. The diagnostic diamino acids in the cell-wall peptidoglycan were aspartic acid, glutamic acid, glycine, and alanine. MK9 was the predominant menaquinone, and polar lipids include diphosphatidylglycerol (DPG), phosphatidylethanolamine (PE), glycolipid (GL 1, 2, 3), PI, PGL, and L1. Major cellular fatty acids (>10.0%) were C_17:1ω7c (37.0%) and C_18:1 ω9c (18.9%).

The type strain, JXJ CY 35^T (=KCTC 49379^T = CGMCC 1.17501^T), was isolated from the culture mass of M. aeruginosa FACHB-905 collected from Lake Dian Yunnan Province, south-west PR China, and has a DNA G + C content of 68.3%. The GenBank accession number for the 16S rRNA gene sequence and draft genome sequence of strain JXJ CY 35^T are MW723388 and JAKCFC000000000, respectively.