We sampled a coastal area in Weihai, China (37°33′36.0″N, 122°07′12.0″E) on October 15th 2018, and in December 2016 from tidal flat located in Yantai, China (37°33′38.7″N, 121°17′7.6″E). Surface sediment (~5 m depth) were collected in triplicate in sterile plastic bags, kept on ice and transported to the laboratory within 2 h. In total, we sequenced four metagenomes and constructed 16S rRNA gene tag libraries, and isolated 757 bacterial strains. The novel strain RSCT41^T was isolated from the Jia River, whereas strains N1E11^T, N1Y90^T, 1640^T, A049^T, and A043^T were isolated from sediment collected in Weihai, China. The sediment sample (1 g) was suspended in 9 mL sterile seawater and diluted serially. An aliquot (0.1 mL) from the serial dilution was spread on marine agar 2,216 (MA, Hopebio) and incubated at 28°C for 7 days. Colonies were subsequently purified and used for physiological, biochemical, and chemical analyses. All strains were cultivated on MA at 28°C for 60 h. Cultures were maintained frozen at −80°C in sterile water supplemented with 15% (v/v) glycerol and 1% (w/v) NaCl.

Gram staining and hydrolysis assays (agar, starch, gelatin, alginate, Tween 20, 40, 60, 80) were conducted following Smibert’s et al. (1994) protocols. Colonies grown on MA (33°C, 60 h) were examined for morphology. Cell structure and motility were analyzed via light microscopy (Leica ICC50HD) and phase-contrast microscopy (Nikon E600W), with gliding motility assessed using Bowman (2000) methodologies. DNase activity was tested on DNase Agar (Hopebio) per manufacturer guidelines. Temperature-dependent growth was evaluated across a range of 4–45°C on marine agar (MA), using temperature increments of 5°C or smaller. Growth NaCl tolerance and pH were evaluated as follows: NaCl tolerance: 0–12% (w/v) in modified MB (composition: 5 g tryptone, 1 g yeast extract, 3.2 g MgSO₄, 2.2 g MgCl₂, 1.2 g CaCl₂, 0.7 g KCl, 0.2 g NaHCO₃ per liter), pH tolerance: 5.5–9.5 (adjusted with 20 mM buffers: MES for pH 5.5–6.0; PIPES for 6.5–7.0; HEPES for 7.5–8.0; Tricine for 8.5; CAPSO for 9.0–9.5). To examine O₂ metabolism, growth under strictly anaerobic condition was tested on MA with or without 0.1% (w/v) NaNO₃, incubated for 7 days at 33°C. Oxidase activity was determined with an oxidase reagent (bioMérieux). Catalase activity was tested by the observation of gas bubbles after the addition of a few drops of 3% (v/v) H₂O₂ to fresh biomass grown on an agar plate. Various biochemical and additional enzyme activities were harvested using API 20E and API ZYM kits (bioMérieux). Oxidation of different compounds as sole carbon sources was checked in Biolog GEN III MicroPlates. All the API kits and Biolog GEN III MicroPlates tests were carried out according to the manufacturer’s instructions except bacterial biomass for inoculations was suspended in 3% (w/v) NaCl.

Cells for fatty acid profiling were cultivated on marine agar (MA) at 28°C for 60 h. Fatty acid extraction and methyl ester analysis were carried out following the protocol established by Sasser (1990), utilizing the MIDI Sherlock Microbial Identification System (Microbial ID Inc.). For isoprenoid quinone characterization, strain RSCT41^T, N1E11^T, N1Y90^T, 1640^T, A049^T, and A043^T cells were harvested during the late-exponential growth phase in marine broth (MB) at 28°C. Quinones were isolated from lyophilized biomass (300 mg) via chloroform-methanol (2:1, v/v) solvent extraction and purified via high-performance liquid chromatography (HPLC). Polar lipid compositions of strains RSCT41^T, N1E11^T, N1Y90^T, 1640^T, A049^T, and A043^T were resolved by two-dimensional thin-layer chromatography (2D-TLC). The first direction was developed in chloroform:methanol:water (65:25:4, v/v/v), and the second in chloroform:methanol:acetic acid:water (80:12:15:4, v/v/v/v). Appropriate detection reagents were used to identify the spots; molybdophosphoric acid (phosphomolybdic acid reagent, 5% v/v solution in ethanol; Sigma) was used to detect total polar lipids, ninhydrin reagent (0.2% solution; Sigma) was used to detect amino lipids, Zinzadze reagent (molybdenum blue spray reagent, 1.3%; Sigma) was used to detect phospholipids, and Polar lipids were extracted with different ratio of chloroform/methanol/water system and analyzed by two-dimensional silica thin-layer chromatography according to the modified method described previously (Komagata and Suzuki, 1988).

Genomic DNA was isolated from 60-h cultures grown on marine agar (MA) using the TIANamp Bacteria DNA Kit (TIANGEN) according to the manufacturer’s instructions, followed by PCR amplification of the 16S rRNA gene. The genomic DNA G + C content was calculated based on the draft genome sequence, while complete 16S rRNA gene sequences were extracted from individual genomes via Barrnap² and submitted in the GenBank database. Reference strain sequences were acquired from the GenBank/EMBL/DDBJ repositories. 16S rRNA gene sequence similarities were assessed via the EzTaxon-e platform³ (Chalita et al., 2024) and BLASTN analysis. Sequence alignment was conducted using CLUSTALW v1.81, with phylogenetic trees reconstructed through neighbor-joining (NJ) algorithms implemented in MEGA v11.0 (Tamura et al., 2021).

The genomic DNA guanine-cytosine (G + C) content was determined via in silico analysis of the assembled genome sequence. Quality metrics were assessed using CheckM (Chklovski et al., 2023) to evaluate N50, contig count, completeness, and contamination. Average nucleotide identity (ANI) values were calculated with the pyani (v0.2.8) (Pritchard et al., 2016). To infer taxonomic boundaries at the genus level, the average amino acid identity (AAI) (Rodriguez-R and Konstantinidis, 2014) and percentage of conserved proteins (POCP) (Qin et al., 2014) were employed for prokaryotic genus delineation, with evolutionary relationships further clarified through this metric. Previous studies have validated POCP as a robust genomic indicator for prokaryotic genus classification. Phylogenomic reconstruction was performed using GTDB-Tk (Chaumeil et al., 2020), leveraging a curated set of 120 universal single-copy marker proteins for robust evolutionary inference, Algibacter mikhailovii 4-1,052 was selected as the phylogenetic outgroup. Metabolic pathway profiling was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Kanehisa et al., 2016). Carbohydrate-active enzymes (CAZymes) in the genomes were annotated using the dbCAN2 meta-server (Zhang et al., 2018), a specialized tool for automated CAZyme characterization. PULs, PUL-like and other CAZyme-rich gene clusters (CGCs) were predicted by searching annotations of CAZymes and the surrounding upstream and downstream genes, using a seven-gene sliding window such that any CAZyme (excluding glycosyl transferases), sulfatase, TBDT, or susD-like genes would be grouped together if they were within ten genes of another such gene (Francis et al., 2021).

A cumulative sequencing output of 1.4 terabase pairs (Tbp) was obtained, averaging 65 gigabase (Gbp) per metagenomic dataset (Lu et al., 2023). Raw reads underwent quality control through BBDuk v35.14⁴ with FastQC v0.11 for validation. Individual sample assemblies were generated using MEGAHIT v1.2.9 (Li et al., 2015), retaining scaffolds ≥2.5 kbp in length. To facilitate downstream analysis, BBMap v38.86 was employed to map reads back to assemblies, generating BAM files with stringent parameters (minimum identity threshold: 0.99; identity filter: 0.97; fast and nodisk modes enabled). Metagenome-assembled genomes (MAGs) were reconstructed through a hybrid binning strategy within the anvi’o v6.2 platform (Eren et al., 2015), integrating outputs from CONCOCT v0.4.0 (Alneberg et al., 2014), MaxBin2 v2.1.1 (Wu et al., 2016), and MetaBAT2 v0.2 (Kang et al., 2019). Consensus bins were refined using DAS Tool v1.1 (Sieber et al., 2018) to optimize genome completeness. MAG nomenclature follows a hierarchical code: Prefix: Sample origin (B: Saccharina, L: Ulva, H: Grateloupia, R: Gelidium, S: seawater, N: sediment) Season: Numeric designation (1: autumn, 2: winter, 3: spring, 4: summer) Suffix: Binning algorithm and unique identifier.

The 16S high-throughput sequencing data of Microbe Atlas Project (MAP, https://microbeatlas.org/, accessed on 30 July 2023) was used with a 97% sequence similarity threshold. The taxonomy of each operational taxonomic unit (OTU) was identified by search (Rognes et al., 2016) using the silva SSU Ref NR 99138.1 dataset. OTUs with more than 97% 16S rRNA gene sequence similarity to members of the six novel species of family Marinilabiliaceae. The habitat distribution of the six novel species was determined by examining the bacterial abundance in global samples.

Colonies of RSCT41^T, N1E11^T, N1Y90^T, 1640^T, A049^T, and A043^T on MA were smooth and circular. Bacterial cells were rod-shaped and the lengths all exceed 10.0 μm (Supplementary Figure S1). The optimal salinity of six strains was close to the concentration of seawater (2–3%, w/v), and the optimal temperature was around 28°C. Agar and starch hydrolysis were both positive. Additional biochemical/physiological data were given in Supplementary Table S1 and in the species description. As shown in Supplementary Table S2, the common major fatty acids (>5.0%) of strains RSCT41^T, N1E11^T, N1Y90^T, A049^T, and A043^T were iso-C_15:0 and anteiso-C_15:0, which was consistent with other Carboxylicivirga species. Unlike the five species of Carboxylicivirga, the major fatty acids of strain 1640^T (Marinifilum) were C_16:0, C_18:1 ω6c, C_18:0 10-methyl, TBSA and Summed Feature 3. The major respiratory quinone detected was MK-7, which was characteristic of the genus Carboxylicivirga. The major polar lipids were found to be phosphatidylethanolamine (PE) and one unidentified lipid (L) (Supplementary Figure S2) in N1E11^T N1Y90^T A043^T, A049^T, and RSCT41^T. The major polar lipids of N1Y90^T A043^T, and A049^T also include diphosphatidylglycerol (DPG). Major polar lipids of 1640^T include PE and DPG.

The six strains (N1E11^T N1Y90^T A043^T, A049^T, RSCT41^T, and 1640^T) exhibited distinct metabolic profiles (Supplementary Table S1). The six strains exhibit distinct metabolic functional divergence, reflecting adaptations to diverse ecological niches. Proteolytic and lipolytic specialists (C. agarovorans, C. litoralis, and C. caseinilyticus) dominate in protein- and lipid-rich environments, evidenced by strong trypsin, α-chymotrypsin, and esterase/lipase activities (C4/C8 and C14 lipase in C. caseinilyticus), coupled with gelatin hydrolysis (GEL+). These traits align with roles in decomposing organic matter, marine detritus, or dairy waste. In contrast, C. longa specializes in carbohydrate metabolism, uniquely expressing α-galactosidase, β-glucuronidase, α-glucosidase, and α-fucosidase, enabling degradation of plant-derived polysaccharides or host mucins. C. fragile prioritizes amino acid catabolism through diverse arylamidases (leucine, valine, cystine) and tryptophan deaminase (TRP+), suggesting adaptation to peptide-rich, carbon-limited niches like soil organic layers or wound exudates. Meanwhile, M. sediminis displays metabolic inertness, with weak lipase (C14), β-galactosidase, and nitrate reduction (NO₃), indicative of survival strategies in oligotrophic sediments through minimal energy expenditure (Supplementary Table S3). All strains except 1640^T demonstrated starch hydrolysis, while cellulose degradation was restricted to N1E11^T, and 1640^T. Tween utilization varied significantly: N1E11 and N1Y90^T showed broad capabilities (Tween 20/40/60/80 and 20/40/60/80 respectively), whereas RSCT41^T and A049^T failed to utilize any Tweens. Proteolytic activity was observed in N1Y90^T, A043^T, and A049^T through casein hydrolysis, though none displayed gelatinase activity. DNA and agar degradation were exclusive to N1E11^T and RSCT41^T. Oxidase activity was present in N1E11^T and 1640^T but absent in RSCT41^T (data unavailable for others). Catalase activity differed between oxidase-positive strains, being present in N1E11^T but absent in 1640^T. Nitrate reduction occurred in N1E11^T, RSCT41^T, and 1640^T. Notably, N1E11^T demonstrated the most versatile metabolism, while A049^T showed limited enzymatic capabilities except for starch and casein utilization. Strain 1640^T displayed unique combinations of cellulose degradation, Tween 60/80 utilization, and nitrate reduction despite lacking starch hydrolysis capabilities. These metabolic distinctions suggest ecological specialization and potential differences in environmental adaptation among the strains.

The complete 16S rRNA gene sequences of strain N1E11^T, N1Y90^T, A043^T, A049^T, RSCT41^T, and 1640^T were obtained. Analysis of the 16S rRNA gene sequence revealed that strain N1E11^T N1Y90^T A043^T, A049^T, and RSCT41^T belonged to the genus Carboxylicivirga, showing the highest sequence similarity with 97.2% similarity to C. marina N1Y132^T, 95.2% similarity to C. sediminis JR1^T, 98.2% similarity to C. sediminis JR1^T, 98.2% similarity to C. sediminis JR1^T, and 97.9% similarity to C. sediminis JR1^T, respectively, 1640^T belong to the genus Marinifilum showing the highest sequence similarity with 95.9% similarity to Marinifilum breve JC075^T within the order Marinilabiliales. In the neighbor-joining (NJ) phylogenetic tree, strain RSCT41^T, N1E11^T, N1Y90^T, 1640^T, A049^T, and A043^T formed a distinct clade within the genus Carboxylicivirga and Marinifilum. Within this subcluster, strain RSCT41^T, N1E11^T, N1Y90^T, A049^T, and A043^T were grouped together with Carboxylicivirga species, which was supported by high bootstrap values (Figure 1) and 1640^T was grouped together with Marinifilum breve JC075^T.

The order Marinilabiliales was ubiquitous in marine sediment ecosystems, playing critical roles in organic matter decomposition and nutrient cycling. The family Marinilabiliaceae (n = 45) was predominantly found in coastal sediments (71%), followed by marine sediments (22%), and rarely detected in non-marine environments (7%). Prolixibacteraceae (n = 15) exhibited a broad distribution, with the highest prevalence in coastal sediments (53%) and marine sediments (33%). Marinifilaceae (n = 8) specialized in marine sediments (75%) and seawater (25%). Salinivirgaceae (n = 3) exclusively associated with “Other” habitats (e.g., non-marine salt environments). This study aggregates genomic data from diverse environments (coastal sediment, marine sediment, seawater, etc.) to explore how genomic features vary across genera and correlate with environmental adaptation.

Phylogenetic reconstructions showed consistent topologies regardless of the sequences used to reconstruct them. Specifically, (i) the 16S rRNA genes (Figure 1), (ii) the concatenated sequences of 120 conserved single-copy orthologous genes (essential genes; Figure 2). The general features of the genomes were given in Supplementary Table S4. The phylogenomic tree can be roughly divided into 4 families. The largest family was relatively small and contains 58 type strains. It represents a highly diverse group in terms of genomic size and ecological diversity, with strains from various environments. Overall, this division provides valuable insights into the evolutionary, ecological, and genomic diversity among these strains, highlighting the differences in model species representation and the variability in genomic characteristics across difference families. A detailed summary of the overall genome-relatedness indices was provided in Supplementary Table S4. The inconsistencies between the phylogenetic trees based on the 16S rRNA gene sequences and the phylogenomic trees constructed from whole-genome sequence analyses reveal that 16S rRNA gene sequence analyses were insufficient to understand the phylogeny and evolution of the members of the genus Carboxylicivirga. Both the sequence similarities and phylogenetic relationships indicated that strains RSCT41^T, N1E11^T, N1Y90^T, A049^T, and A043^T represent five novel species of the genus Carboxylicivirga.

The dataset exhibits substantial variability in sample sizes across strains and habitat categories, which may influence the reliability and interpretation of the observed ecological patterns (Supplementary Figure S2 and Supplementary Table S6). For instance, Carboxylicivirga strains A049^T, N1E11^T, and A043^T collectively account for thousands of samples in major categories (e.g., n = 9,626 for unknown aquatic habitats), whereas strain N1Y90 has a notably smaller total sample size (n = 98). Among Carboxylicivirga strains, a genus-level consistency was observed for lineages A049^T, N1E11^T, and A043^T, all displaying uniform sediment-associated abundances (8.43%, n = 3,392) despite their primary dominance in seawater (~16.4%, n = 6,610) and unknown aquatic habitats (23.9%, n = 9,626) (Supplementary Figure S2 and Supplementary Table S6). Notably, these strains also shared significant animal associations (~11.9%, n = 4,795), suggesting a dual ecological strategy favoring both aquatic and host-linked niches. Strain N1Y90^T deviated from this pattern, exhibiting an elevated sediment affinity (10.2%, n = 10) within a limited total sample size (n = 98), alongside a disproportionate representation in unknown aquatic habitats (26.5%, n = 26). This anomaly may reflect either localized adaptation to sedimentary conditions or sampling artifacts, necessitating further validation. In contrast, strain RSCT41^T demonstrated minimal sediment colonization (4.02%, n = 18), instead thriving in poorly characterized aquatic environments (35.9%, n = 161) and marine habitats (19.86%, n = 89), with negligible presence in plant rhizospheres (0.22%, n = 1). Comparative analysis revealed genus-specific divergence in sediment utilization. While Carboxylicivirga strains collectively exhibited stronger sedimentary associations (~8.43%) compared to Marinifilum (7.1%), intra-genus variability was evident. For instance, N1Y90^T’s higher sediment prevalence (10.2%) contrasted sharply with RSCT41^T’s low abundance (4.02%), underscoring strain-level adaptability to environmental heterogeneity. These differences may arise from metabolic plasticity or uncharacterized habitat gradients. However, the substantial proportion of “unknown” habitat classifications across all strains (e.g., 29.3% for Carboxylicivirga, n = 11,800) indicates critical gaps in environmental sampling, particularly for sedimentary systems. Similarly, Marinifilum (1640) shows a large sample pool (n = 7,808), but terrestrial niches like soil (n = 208) and plant-associated samples (n = 115) were underrepresented (Supplementary Figure S2 and Supplementary Table S6). The examined bacterial strains exhibited distinct ecological preferences, with sediment environments serving as a secondary but notable niche across most lineages. Marinifilum (Strain 1,640), predominantly associated with aquatic habitats, showed moderate sediment colonization (7.1%, n = 555), ranking below its prevalence in seawater (23.85%, n = 1858) and unknown aquatic environments (30.7%, n = 2,401). In contrast, terrestrial niches such as soil (2.66%, n = 208) and plant-associated samples (1.48%, n = 115) were minimally represented, while animal-derived samples accounted for a minor fraction (4.45%, n = 349).

In conclusion, sediment environments represent a stable yet subordinate niche for most Carboxylicivirga lineages, complementing their dominance in open-water habitats. Marinifilum’s intermediate sediment affinity aligns with its broader aquatic versatility. Future studies should prioritize expanded sampling of sedimentary ecosystems, especially for low-abundance strains like RSCT41^T and N1Y90^T, to unravel their ecological roles and adaptive mechanisms in these understudied environments.

Firstly, we analyzed the similarities and differences in metabolic pathways between strains RSCT41^T, N1E11^T, N1Y90^T, 1640^T, A049^T and A043^T and their closely related strains of the order Marinilabiliales (Figure 4 and Supplementary Table S7). Summary of metabolic pathway completeness at the genus level: The provided KEGG analysis data were grouped at the genus level to assess the completeness of metabolic pathways across taxa. Each value represents the percentage completeness of a specific pathway module for members of a genus, averaged across all species within that genus. The following section summarizes key findings, organized by metabolic category and representative genera. (1) Carbohydrate metabolism key genera: Saccharicrinis: Exhibited high completeness in central carbohydrate metabolism (average 95–100%), including gluconeogenesis (M00003: 92%) and the pentose phosphate pathway (M00004: 100%). Notably, glycogen biosynthesis (M00854: 83%) and degradation (M00855: 90%) were well conserved. Carboxylicivirga: Demonstrated robust pectin degradation (M00081: 85%) and galactose metabolism (M00632: 95%), suggesting adaptation to complex carbohydrate utilization (Figure 4 and Supplementary Table S7). Marinilabilia: Showed variability in “other carbohydrate metabolism” pathways (e.g., D-galacturonate degradation: 60–80%), potentially reflecting niche-specific adaptations. (2) Energy metabolism key genera: Ancylomarina: High completeness in nitrogen fixation (M00175: 75%) and sulfur oxidation (M00984: 90%), indicative of roles in anaerobic environments. Draconibacterium: Dominated dissimilatory nitrate reduction (M00530: 85%) and denitrification (M00529: 70%), aligning with their prevalence in nitrate-rich sediments. Geofilum: Displayed near-complete carbon fixation pathways (CAM pathways: 95–100%), suggesting phototrophic or chemolithoautotrophic capabilities (Figure 4 and Supplementary Table S7). (3) Lipid metabolism key genera: Labilibaculum: Fatty acid biosynthesis (M00083: 98%) and β-oxidation (M00086: 95%) were highly conserved, consistent with lipid-rich marine habitats. Marinifilum: Phosphatidylethanolamine biosynthesis (M00093: 90%) was prevalent, likely supporting membrane stability in fluctuating salinity conditions. (4) Amino acid metabolism key genera: Natronoflexus: Showed completeness in arginine biosynthesis (M00844: 85%) and histidine degradation (M00045: 75%), critical for osmoregulation in hypersaline environments. Prolixibacter: High scores in aromatic amino acid metabolism (M00022: 80%) and polyamine biosynthesis (M00133: 70%), potentially linked to stress response mechanisms (Figure 4 and Supplementary Table S7). (5) Cofactor and vitamin metabolism key genera: Sunxiuqinia: Thiamine biosynthesis pathways (M00127: 65–80%) were prominent, essential for enzymatic cofactor provision in nutrient-limited settings. Mangrovibacterium: Cobalamin biosynthesis (M00924: 75%) and molybdenum cofactor pathways (M00880: 60%) highlighted their role in redox reactions (Figure 4 and Supplementary Table S7). (6) Terpenoid and polyketide biosynthesis key genera: Carboxylicivirga: Terpenoid backbone biosynthesis (M00365: 55–70%) was moderately conserved, suggesting limited secondary metabolite production. Ancylomarina: Low scores in this category (<30%) indicate a metabolic focus on core pathways rather than specialized biosynthesis (Figure 4 and Supplementary Table S7). Saccharicrinis and Carboxylicivirga dominated carbohydrate metabolism, likely facilitating decomposition of organic matter in marine ecosystems. Draconibacterium and Ancylomarina specialized in nitrogen and sulfur cycling, aligning with their prevalence in anoxic sediments. Draconibacterium and Ancylomarina encode dissimilatory nitrate reductases (narGHI) and sulfur oxidation modules, supporting survival in low-oxygen sediments.

Variability in cofactor metabolism (e.g., Sunxiuqinia vs. Marinifilum) suggests niche partitioning driven by micronutrient availability. These findings underscore the metabolic diversity within the order Marinilabiliales and its ecological implications in marine biogeochemical cycles. Data availability: Completeness values were derived from KEGG module annotations (Supplementary Table S7). Statistical analyses were performed using genus-averaged percentages; standard deviations ranged from ±5% (highly conserved pathways) to ±25% (variable pathways).

The genes of degradative CAZymes and predicted PULs in each genome indicates the potential for polysaccharide degradation (Supplementary Table S4). Interestingly, the occurrence of CAZymes gene and PULs exhibited an exponential difference among different families (Figure 5). Notably, Prolixibacteraceae species were annotated with a higher number of PULs, further underscoring their strong potential for polysaccharide degradation. Marinilabiliaceae species were annotated with a second highest number of CAZymes and PULs, followed by Marinifilaceae. In contrast, Salinivirgaceae species had the lowest counts of both CAZymes and PULs. However, the occurrence of PULs did not show a proportional increase with genome size, suggesting no clear correlation between the two variables. As shown in Figure 5, the distribution of sulfatases across families follows a distinct and consistent pattern, suggesting that their presence is non-random and potentially influenced by environmental or ecological factors. We observed a positive correlation between the numbers of CAZymes, sulfatases, and PULs and the different branches of the phylogenetic tree. However, exceptions to this trend were also evident, such as the case of M. sediminis 1640^T, which, despite being located on a more distantly related branch, was annotated with a relatively higher number of CAZymes, sulfatases, and PULs (see Figure 6).

The six novel species identified in the dataset exhibit unique genomic and functional traits that distinguish them from type strains and other members of their respective genera (Figure 6 and Supplementary Table S4). Within the genus Carboxylicivirga (Marinilabiliaceae), A049^T stands out with a genome size of 5.65 Mb, 26 PUL, and 268 CAZymes (Supplementary Table S8)—placing it among the top enzymatic performers in the genus—yet its moderate sulfatase count (14) suggests a focus on non-sulfated carbohydrate degradation compared to extreme marine specialists like Carboxylicivirga_marina (106 sulfatases). Similarly, RSCT41^T (6.58 Mb genome, 29 PUL, 172 CAZymes, 81 sulfatases) combines high sulfatase activity with moderate CAZyme diversity, indicating adaptation to sulfated algal polysaccharides in marine sediments. In contrast, N1Y90^T (5.60 Mb, 20 PUL, 200 CAZymes, 27 sulfatases) displays a generalist profile, lacking specialization in either sulfatases or CAZymes, potentially occupying a transitional niche between marine and non-marine environments. A043^T (5.30 Mb, 27 PUL, 175 CAZymes, 38 sulfatases) (Supplementary Table S9) further highlights Carboxylicivirga’s genomic plasticity, with one of the highest PUL counts in the genus but unremarkable sulfatase levels, suggesting a focus on structural carbohydrate breakdown rather than sulfated polymers (Kertesz, 2000).

Collectively, these novel species reveal patterns of functional divergence within their genera. While some, like RSCT41^T and A049^T, emphasize marine adaptation through elevated sulfatase activity, others (e.g., N1Y90^T) exhibit generalist or niche-specific strategies. Their genomic and enzymatic profiles underscore the importance of integrating taxonomic, metabolic, and environmental data to uncover hidden ecological roles in understudied marine bacteria. These findings highlight the potential for novel species to drive discoveries in carbohydrate metabolism, sulfatase-mediated degradation, and niche partitioning in marine ecosystems.