Carboxylicivirga marinus sp. nov., Isolated From Marine Sediment, and Genome Insight of the Genus Carboxylicivirga

A Gram-stain-negative, facultatively aerobic, beige and slender bacterium, designed N1Y132T, was isolated from a sediment sample taken from coastal zone of Weihai, China (37°34′31.44″ N, 122°9′15″ E). Growth occurred at 15-33°C (optimal 25–28°C), at pH 6.5–8.5 (optimal pH 7.5), in 0–9% (w/v) NaCl (optimal 2.0–2.5%). According to the phylogenetic analysis based on the 16S rRNA gene sequences, strain N1Y132T showed the highest sequence similarity (97.9%) with Carboxylicivirga taeanensis JCM 19490T, followed by C. sediminis JR1T (97.4%), C. mesophila JCM 18290T (96.4%) and C. linearis FB218T (95.1%). Thus, strain N1Y132T was a member of the genus Carboxylicivirga. The average nucleotide identity (ANI) values between strain N1Y132T and C. sediminis JR1T, C. linearis FB218T and C. mesophila JCM 18290T were 74.8%, 71.1%, and 74.9%, respectively. Chemotaxonomic analysis showed that the sole respiratory quinone was MK-7 and the major fatty acids (> 5.0%) included iso-C15:0, anteiso-C15:0, iso-C13:0, iso-C14:0, iso-C15:0 3-OH, and C15:1ω6c. The polar lipids were consisted of a phosphatidylethanolamine, two phosphoaminolipids and six unidentified lipids. The DNA G + C content was 38.1 mol%. Based on the evidence presented in this study, strain N1Y132T represents a novel species of the genus Carboxylicivirga, for which the name Carboxylicivirga marinus sp. nov. is proposed. The type strain is N1Y132T (= KCTC 72934T = MCCC 1H00431T).


INTRODUCTION
The genus Carboxylicivirga, a member of family Marinilabiliaceae of the phylum Bacteroidetes, was originally proposed by Yang et al. (2014). Bacteria of the genus Carboxylicivirga are Gramstain-negative and slender, and the major respiratory quinone is MK-7. At the time of writing, 5 species are assigned to the genus, including C. mesophila (Yang et al., 2014), C. taeanensis (Yang et al., 2014), C. linearis (Wang et al., 2015), C. flava (Wang et al., 2016), and C. sediminis (Wang F. Q. et al., 2018). Without exception, members of the genus Carboxylicivirga are isolated from marine sediments. The 16S rRNA gene sequencing revealed that Carboxylicivirga was the core genera in the gut of marine invertebrates, such as the mud crab Scylla paramamosain (Wei et al., 2019) and the lobster Homarus gammarus (Holt et al., 2019). However, the function of this newly recognized genus in the environment or in the host are less known. In this study, a beige, non-motile and facultatively aerobic bacterium, strain N1Y132 T , is proposed as representing a novel species of genus Carboxylicivirga. Comparative analysis of genomes of strain N1Y132 T and reference strains C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T are contributed to understand the function of genus Carboxylicivirga.

Bacterial Isolation and Cultivation
Strain N1Y132 T was isolated from a sediment sample taken from coastal zone of Weihai, China (37 • 34 31.44 N, 122 • 9 15 E), in October 2018. For the bacterial isolation, the sediment sample was homogenized and 10-fold serially diluted to 10 −5 with sterile seawater and 100 µl aliquot of each dilution was spread on marine agar 2216 (MA; Becton Dickinson) and incubated at 28 • C for 21 days. The strain, designed N1Y132 T , was incubated on MA at 28 • C and was stored at −80 • C in sterile 1% (w/v) saline supplemented with 15% (v/v) glycerol. The type strains C. mesophila JCM 18290 T (obtained from Japan Collection of Microorganisms, Japan), C. taeanensis JCM 19490 T (obtained from Japan Collection of Microorganisms, Japan), C. sediminis JR1 T (generously provided by Wang FQ, Wang F. Q. et al., 2018) and C. linearis FB218 T (storage in our lab) were used as reference strains and cultivated routinely on MA at 28 • C.

Enzyme activity
Alkaline phosphatase methods from Smibert and Krieg (1994). Nitrogen fixation ability was tested by using a nitrogen free medium comprised of 0.1% starch and 1.8% agar, prepared with artificial seawater (0.32% MgSO 4 , 0.23% MgCl 2 , 0.12% CaCl 2 , 0.07% KCl and 0.02% NaHCO 3 , all w/v) at 28 • C for 2 weeks in aerobic and anaerobic conditions (Wang H. B. et al., 2018). Additional physiological and biochemical characteristics were determined using API 20E and API ZYM kits (bioMérieux). Tests of acid production from carbohydrates were detected using the API 50CHB fermentation kit (bioMérieux). Oxidation of substrate was examined in Biolog GEN III MicroPlates. All API tests were performed according to the manufacturer's instructions (except for salinity, which was adjusted to 3%).

16S rRNA Gene Sequence Analysis
The 16S rRNA gene was amplified by PCR using the Taq DNA polymerase with the universal primers 27F and 1492R (Lane, 1991). The purified PCR product with 3 -A overhangs was cloned into a pMD18-T vector (Takara) based on the TA cloning method, and the ligation product was transformed into Escherichia coli DH5α cells. The positive clones were selected and sequenced bi-directionally with primers M13-47 and RV-M at BGI Co. Ltd (Qingdao, China) using the ABI 3730XL system. Thus, a near-complete 16S rRNA gene sequence (1500 bp) of strain N1Y132 T was gained. The determined 16S rRNA gene sequence for strain N1Y132 T was submitted to the National Centre for Biotechnology Information (NCBI) GenBank database. The 16S rRNA gene similarities were calculated using the EzBioCloud Database (Yoon et al., 2017). Multiple sequences were aligned using the CLUSTAL_X program (Thompson et al., 1997). In order to determine the phylogenetic position of strain N1Y132 T , phylogenetic trees were constructed using the neighbor-joining (NJ) (Saitou and Nei, 1987), maximum-likelihood (ML) (Fitch, 1971) and maximum-parsimony (MP) (Felsenstein, 1981) methods in MEGA version 7.0 (Kumar et al., 2016). The reliability of relationships was ensured by performing bootstrap analyses based on 1000 replications.

Chemotaxonomic Characterization
In order to determine the chemotaxonomic features, the cells of strain N1Y132 T and C. sediminis JR1 T , C. linearis FB218 T , C. mesophila JCM 18290 T , and C. taeanensis JCM 19490 T were harvested by centrifuge and subjected to freeze-drying after cultivation in marine broth 2216 medium for 3 days at 28 • C. Fatty acid methyl esters (FAMEs) were extracted from 30 mg freeze-dried biomass according to the standard protocol of MIDI (Sherlock Microbial Identification System, version 6.1). Fatty acids were prepared and analyzed by using an Agilent 6890N gas chromatograph. Peaks were automatically integrated, and fatty acid names and percentages were determined using the TSBA40 database of the Microbial Identification system (Sasser, 1990). Respiratory quinones of strain N1Y132 T were obtained from 300 mg freeze-dried cells using the two-stage method described by Tindall et al. (2007) and analyzed by HPLC (Kroppenstedt, 1982). Polar lipids were extracted from 50 mg of freeze-dried cell material using a chloroform/methanol system according to the procedures described by Komagata and Suzuki (Komagata and Suzuki, 1988). Polar lipids were analyzed by two-dimensional thin layer chromatography according to Minnikin et al. (1984).

Genomic Analysis
Genomic DNAs of strain N1Y132 T and reference strains C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T were extracted by using a genomic DNA extraction kit (Takara) according to the manufacturer's recommendations. The draft genome sequence of N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T were sequenced on the Illumina HiSeq PE150 platform at Beijing Novogene Bioinformatics Technology (Beijing, China). The  Filled circles indicate that the corresponding nodes were also recovered in the trees generated with the maximum-likelihood and maximum-parsimony algorithms.
Open circles indicate that the corresponding nodes were also recovered in either the maximum-likelihood or maximum-parsimony algorithm. Bacteroides acidifaciens JCM 10556 T (GenBank accession number, AB510696) was used as an outgroup. Bar, 0.02 substitutions per nucleotide position.
genome sequence was deposited in the DDBJ/GenBank/EMBL database. The G + C content of the chromosomal DNA was calculated using genome sequence. Homologous genomic regions of four Carboxylicivirga species were analyzed using progressiveMauve algorithm (Darling et al., 2004). Core genes and specific genes of four strains were analyzed by the CD-HIT (Fu et al., 2012) rapid clustering of similar proteins software with a threshold of 40% pairwise identity and 0.4 length difference cutoff in amino acid. Then, the Venn figure was drawn to show their relationships among the samples. The genes of involved in N1Y132 T metabolic pathways were analyzed using the KofamKOALA (Aramaki et al., 2020 1 ) and the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) (Tatusova et al., 2016 2 ). Carbohydrate-active enzymes of strain N1Y132 T and three reference strains were annotated using the dbCAN2 meta server . The presence 1 https://www.genome.jp/tools/kofamkoala/ 2 https://www.ncbi.nlm.nih.gov/genome/annotation_prok/ of gene clusters encoding secondary metabolites of strain N1Y132 T was predicted by using the antiSMASH 6.0 database (Blin et al., 2021). The average nucleotide identity (ANI) was calculated using ANI calculator (Lee et al., 2016 3
Frontiers in Marine Science | www.frontiersin.org (C4), esterase lipase (C8), α-galactosidase, H 2 S production and D-adonitol. Given these positive results, the ability of strain N1Y132 T to utilize substrate was weaker than that of these three reference strains. Detailed phenotypical features were listed in species description as well as in Table 1

Phylogenetic Analysis of 16S rRNA Gene Sequences
After PCR amplification and sequencing, a near-complete 16S rRNA gene sequence (1500 bp) of strain N1Y132 T was obtained and subjected to comparative analysis. BLAST searches revealed that strain N1Y132 T shared the highest 16S rRNA gene sequence similarity with C. taeanensis JCM 19490 T (97.9%, Figure 1A). In addition, strain N1Y132 T shared the 16S rRNA gene sequence similarity of 96.4% with C. mesophila JCM 18290 T , which was the type species of the genus Carboxylicivirga. In the resulting NJ tree, strain N1Y132 T was placed within the clade comprising the genus Carboxylicivirga and was located nearest to C. sediminis JR1 T (Figure 1B). MP and ML trees also showed similar tree topologies. Both the sequence similarities and phylogenetic relationships indicated that strain N1Y132 T may represent a novel species of the genus Carboxylicivirga.

Basic Characteristics of Four Genomes
The genome length of strain N1Y132 T , C. mesophila JCM 18290 T , C. linearis FB218 T , and C. sediminis JR1 T were less differentiating these four strains ( Table 3). Among them, C. linearis FB218 T harbors the largest genome (6,050,047 bp, JAGUCO000000000), while the new strain N1Y132 T holds the smallest one (5,649,616 bp, JAENRR000000000). The sequencing depth of coverage and the contig numbers are different for these four draft genomes. However, some basic genome characteristics have been analyzed. The DNA G + C content of strain N1Y132 T FIGURE 2 | Distribution of orthologous genes in the genomes of the four Carboxylicivirga species (strain N1Y132 T , C. mesophila JCM 18290 T , C. linearis FB218 T , and C. sediminis JR1 T ) tested by the CD-HIT rapid clustering of similar proteins software. 2082 core gene clusters that were represented in all four genomes. Strain N1Y132 T , C. mesophila JCM 18290 T , C. linearis FB218 T , and C. sediminis JR1 T contained 1139, 366, 1766, and 553 species-specific gene clusters, respectively. The clustering results were used to generate Venn's diagram (Supplementary Data 1). was 38.1 mol%, which was higher than that of C. linearis FB218 T (36.4 mol%), but lower than that of C. mesophila JCM 18290 T (41.2 mol%) and C. sediminis JR1 T (40.9 mol%).
A total of 4,463 genes were predicted in the genome of strain N1Y132 T by the NCBI PGAP, including 4,383 proteincoding genes, 59 tRNAs, 3 rRNAs, and 3 ncRNAs. The amino acid sequence of strain N1Y132 T are annotated by KEGG (Supplementary Figure 3), and 1976 genes (39.1%) could be assigned a putative function, including carbohydrate metabolism (225 genes), signaling and cellular processing (213 genes), genetic information processing (194 genes), amino acid metabolism (126 genes), environmental information processing (115 genes) and metabolism of cofactors and vitamins (107 genes), glycan biosynthesis and metabolism (78 genes) and energy metabolism (70 genes). The results of C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T and the comparison with strain N1Y132 T were shown in Supplementary Figure 3.
Four Carboxylicivirga genomes are similar in the gene number and protein function prediction. The homologous gene regions have been showed by the progressiveMauve algorithm (Supplementary Figure 4). Subsequently, core genes and specific genes of four strains were analyzed with a threshold of 40% pairwise identity and 0.4 length difference cutoff in amino acid (Figure 2). These four strains shared 2082 core genes. The genome of C. linearis FB218 T seem to hold the most unique genes of 1766. The resulting dDDH values between the genomes of strain N1Y132 T and C. sediminis JR1 T , C. mesophila JCM 18290 T , and C. linearis FB218 T are 18.5%, 18.5%, and 15.4%, respectively ( Figure 1A). The dDDH values between strain N1Y132 T and the reference strains are less than 20%, which are far below the threshold (70%) for new species identification (Goris et al., 2007). The ANI values between strain N1Y132 T and C. mesophila JCM 18290 T , C. sediminis JR1 T , and C. linearis FB218 T are 74.9%, 74.8%, and 71.1%, respectively ( Figure 1A). These values are below the proposed cut-off for a species boundary of 95-96% ANI (Richter and Rossello-Mora, 2009). These results support that the new isolate represents a novel species of the genus Carboxylicivirga.

Metabolic Pathways
Many complete metabolic pathways have been found in the genomes of these four strains (Figure 3). All four strains hold the complete lipid metabolism, nucleotide metabolism and glycan metabolism. For instance, a phosphatidylethanolamine biosynthesis pathway (M00093) has been identified in these four genomes, which is in accordance with the polar lipids results of the genus Carboxylicivirga (Supplementary Figure 2). For the carbohydrate metabolism, the Carboxylicivirga strains possess complete pathways of glycolysis, gluconeogenesis, pyruvate oxidation, citrate cycle, pentose phosphate pathway (M00006), and phosphoribosyl diphosphate (PRPP) biosynthesis. However, the C. sediminis JR1 T and C. linearis FB218 T are deficient in the pentose phosphate pathways (M00004 and M00007), and the C. sediminis JR1 T and C. mesophila JCM 18290 T are devoid of the other carbohydrate metabolism (M00631 Dgalacturonate degradation). For the amino acid metabolism pathways, only the strain N1Y132 T has the ability of lysine biosynthesis, and both C. sediminis JR1 T and C. mesophila JCM 18290 T contain the tryptophan biosynthesis pathway. For the metabolism of cofactors and vitamins, the strain N1Y132 T is lack of riboflavin, biotin and cobalamin biosynthesis, which differentiate between strain N1Y132 T and other three Carboxylicivirga strains. The energy metabolism pathways exhibited the most striking difference among these four Carboxylicivirga strains. All strains have two carbon fixation metabolism pathways [CAM (Crassulacean acid metabolism), dark and light]. In many bacteria, CAM achieves carbon fixation through phosphoenolpyruvate carboxylase (EC: 4.1.1.31), malate dehydrogenase (EC:1.1.1.37), and PPDK (EC:2.7.9.1), the specific process is shown in Figure 4A. Only the strain N1Y132 T has a phosphate acetyltransferase-acetate kinase pathway, which is one of carbon fixation metabolism pathways. However, both C. linearis FB218 T and C. mesophila JCM 18290 T have a complete assimilatory sulfate reduction pathway, wherein sulfate is reduced to cysteine and sequentially incorporated into the biomass. Assimilatory sulfate reduction of C. linearis FB218 T and C. mesophila JCM 18290 T is accomplished through four steps sequentially mediated by sulfate adenylyltransferase (EC 2.7.7.4;cysNC,cysN,cysD), adenylyl-sulfate kinase (EC 2.7.1.25;cysNC), PAPS reductase (EC 1.8.4.8;cysH) and assimilatory sulfite reductase (NADPH) (EC 1.8.7.2; cysJ, cysI) ( Figure 4B). In the genomes of the strains N1Y132 T and C. sediminis JR1 T , a homologous gene cluster contain cysN, cysD, cysH, cysJ, and cysI, but lack of the cysNC gene. Additionally, C. linearis FB218 T has a complete nitrogen fixation pathway (Figure 3). C. linearis FB218 T could grow on the mineral salts medium plates containing starch and agar but lacking NH 4 + , NO 3 − , or organic nitrogen, while other Carboxylicivirga strains could not grow.

Prediction of Secondary Metabolites
Based on secondary metabolite analysis predicted by antiSMASH, the genome of strain N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T and C. mesophila JCM 18290 T share four gene clusters encoding for the RRE (RRE-element containing cluster), the terpene cyclase, the NRPS-like (non-ribosomal peptides) and RiPP-like (Other unspecified ribosomally synthesized and post-translationally modified peptide product) cluster (Supplementary Figure 5). The strain N1Y132 T have 7 putative biosynthetic gene clusters (BGCs). These clusters include one RRE-containing protocluster, one RiPP-like biosynthetic cluster, two NRPS-like clusters, two terpene cyclase clusters and one hserlactone region. For bacteria living in coastal waters, the possession of multiple BGCs that can produce secondary metabolites may help them grow better (Lim et al., 2020). Besides, strain N1Y132 T and C. linearis FB218 T are found to have unique gene cluster, strain N1Y132 T has a hserlactone region (Homoserine lactone cluster) and C. linearis FB218 T has Type I PKS (Polyketide synthase).

Carbohydrate-Active Enzymes (CAZymes)
Given the polysaccharides utilization of marine Bacteroidetes bacteria (Krüger et al., 2019), these four Carboxylicivirga genomes have been analyzed by the CAZy database. Strain N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T and C. mesophila JCM 18290 T contain 255, 296, 333 and 287 carbohydrate-active enzymes, respectively (Figure 5). Among these carbohydrateactive enzymes, glycoside hydrolases (GHs) are the greatest number of enzymes (more than 60% of the identified enzymes were assigned to the GH family in the genomes of four strains). Compared with three reference strains, strain N1Y132 T harbor 255 carbohydrate-active enzymes including 172 GHs, which were obviously lower than C. sediminis JR1 T , C. linearis FB218 T and FIGURE 5 | Histogram of predicted carbohydrate-active enzymes (CAZymes) in strain N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T . Numbers of each enzyme detected in the genome were shown in the map.
C. mesophila JCM 18290 T . The results obtained from the CAZy database suggest that strain N1Y132 T and three reference strains may have a metabolic potential for polysaccharide utilization. The experimental results show that strains N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T and C. mesophila JCM 18290 T can utilize various carbon sources (Table 1 and Supplementary Table 1), which are consistent with the genome annotation results.
According to the polyphasic taxonomic characterization results, strain N1Y132 T could be classified as a new member of the genus Carboxylicivirga, for which the name Carboxylicivirga marinus sp. nov. was proposed.
Description of Carboxylicivirga marinus sp. nov.
The GenBank accession number for the 16S rRNA gene sequence of strain N1Y132 T is MW532703 and the draft genome data has been deposited in GenBank under the accession number JAENRR000000000.

DATA AVAILABILITY STATEMENT
The GenBank accession number for the 16S rRNA gene sequence of strain N1Y132 T is MW532703 and the whole genome shotgun project of strain N1Y132 T , C. sediminis JR1 T , C. linearis FB218 T , and C. mesophila JCM 18290 T have been deposited at DDBJ/ENA/GenBank under the accession number JAENRR000000000, JAGTAR000000000, JAGUCO000000000, JAGUCN000000000, respectively.

AUTHOR CONTRIBUTIONS
D-CL isolated the Strain N1Y132 T . T-SZ, Y-NZ, YG, and Z-JD performed material preparation, experimental operation, data collection, and analysis. T-SZ, YG, and Z-JD wrote the manuscript. YG and Z-JD performed project guidance and critical revision of manuscripts. All authors contributed to the article and approved the submitted version.