Genome Analysis of Celeribacter sp. PS-C1 Isolated from Sekinchan Beach in Selangor, Malaysia, Reveals Its β-Glucosidase and Licheninase Activities

A halophilic marine bacterial strain, PS-C1, was isolated from Sekinchan beach in Selangor, Malaysia. The 16S rRNA gene sequence analysis indicated that strain PS-C1 was associated with the genus Celeribacter. To date, there have been no reports on enzymes from the genus Celeribacter. The present study reports on the cellular features of Celeribacter sp. PS-C1, its annotated genome sequence, and comparative genome analyses of Celeribacter glycoside hydrolase (GH) enzymes. The genome of strain PS-C1 has a size of 3.87 Mbp and a G+C content of 59.10%, and contains 3739 protein-coding genes. Detailed analysis using the Carbohydrate-Active enZYmes (CAZy) database revealed that Celeribacter genomes harboured at least 12 putative genes encoding industrially important GHs that are grouped as cellulases, β-glucanases, hemicellulases, and starch-degrading enzymes. Herein, the potential applications of these enzymes are discussed. Furthermore, the activities of two types of GHs (β-glucosidase and licheninase) in strain PS-C1 were demonstrated. These findings suggest that strain PS-C1 could be a reservoir of novel GH enzymes for lignocellulosic biomass degradation.


Introduction
Marine environments are home to complex and diverse microorganisms that are yet to be discovered through appropriate microbial investigations. Currently, marinederived halophilic bacteria are being explored to harness their novel enzymes and bioactive compounds as substitutes for many industrial applications [1][2][3][4].
The Rhodobacteraceae family is one of the bacterial lineages in marine ecosystems [5], and currently consists of 191 genera that are archived in the List of Prokaryotic names with Standing in Nomenclature (LPSN) database (https://www.bacterio.net; accessed 1 January 2022). Unlike the major industrial microbes of the Bacillaceae family [6][7][8], reports on the potential applications of members of the family Rhodobacteraceae are relatively scarce [9]. Celeribacter is one of the least studied genera in Rhodobacteraceae, with no reports available on its industrial-related enzymes. Earlier studies using whole cells suggested that Celeribacter spp. could potentially be used in bioremediation processes of heavy metals [10], organic compounds [11][12][13][14], volatile substances [15], and hydrocarbons [16][17][18]. Members of this genus are halophiles and Gram-negative, rod-shaped bacteria that thrive in various marine habitats (e.g., seawater, sea sediments, and mangrove soil) [19][20][21]. To date, 10 types

Sampling Site, Isolation, Taxonomy Identification, and Bacterial Characterisation
Wet sediment and mud samples (uppermost layer until a depth of 15 cm) were collected using a sterilised laboratory scoop at Sekinchan beach in Selangor, Malaysia, on 14 September 2020. At the sampling site, the collected samples were stored in sterile bottles that were closed immediately after sampling. The samples were stored at 25 • C, transferred to the laboratory, and stored at 4 • C until further use. The temperature and pH of the collected samples were measured using a laboratory thermometer and pH meter, respectively.
Strain PS-C1 was isolated from the samples using a previously described ex situ cultivation method [41,42]. Pure colonies of strain PS-C1 were obtained by streaking the cells on marine agar at 30 • C (pH 6.5) for 48 h. Subsequently, genomic DNA was extracted from strain PS-C1 using the Monarch Genomic DNA Purification Kit (New England BioLabs, Ipswich, MA, USA) following the manufacturer's instructions. The 16S rRNA gene was amplified by PCR using the forward primer 27F (5 -AGAGTTTGATCCTGGCTCAG-3 ) and reverse primer 1492R (5 -GGTTACCTTGTTACGACTT-3 ) [43]. Gene sequencing was performed by Apical Scientific Sdn. Bhd. (Sri Kembangan, Selangor, Malaysia). Taxonomic identification was performed by comparing the strain PS-C1's 16S rRNA gene sequence with the available sequences in the NCBI GenBank and EzBioCloud 16S databases [44]. A phylogenetic tree was constructed using the neighbour-joining method with 1000 bootstrap replicates with MEGA11 software [45].
Field emission scanning electron microscopy (FESEM) was used to determine the cell shape and size of the strain PS-C1. The cells were treated and sputtered with gold according to the method established by Yang et al. [46] prior to observation under high-resolution FEI Quanta 650 FEG FESEM (Thermo Fisher Scientific, Hillsboro, OR, USA) operating at 10 kV. Gram staining and endospore detection were performed using the methods described by Stankus et al. [47] and observed under a light microscope (OPTIKA Srl, Ponteranica, Italy). Catalase and oxidase activities and hydrolysis of Tween 20 and Tween 80 were performed and assessed according to the method established by Beveridge et al. [48].
The temperature range and optimum growth of strain PS-C1 were analysed by incubating the cells in marine broth at 10-70 • C with shaking at 200 rpm for up to 3 days. The optimal pH for growth was determined at 30 • C and was tested over a pH range of 5.5-11.0. The salt tolerance of strain PS-C1 was determined in marine broth supplemented with 2-10% (w/v) NaCl. Cell growth was determined by measuring absorbance at 600 nm using an a Ultrospec 2100 pro UV/Visible Spectrophotometer (Cytiva, Marlborough, MA, USA).
The motility test of strain PS-C1 was performed using the Analytical Profile Index (API) M Medium kit (bioMérieux, Marcy-l'Étoile, France). API 20NE and API 20E test strips (bioMérieux) were used to determine the basic biochemical characteristics of strain PS-C1. Carbohydrate utilisation and selective enzyme activity of strain PS-C1 were assessed using the API 50CH and API ZYM test strips (bioMérieux), respectively. All tests using API kits were performed at 30 • C according to the manufacturer's protocol. Unless otherwise specified, all the aforementioned bacterial physiochemical and chemotaxonomic characterisations were performed in triplicate.

Genome Sequencing, Assembly, and Annotation
Strain PS-C1 was grown on marine agar (pH 6.5) at 30 • C for 24 h. Subsequently, strain PS-C1 genomic DNA was extracted from a single colony of cells using the standard protocol of the Monarch Genomic DNA Purification Kit (New England BioLabs). A paired-end library was prepared using the NEBNext Ultra DNA Library Prep Kit for Illumina (New England BioLabs), following the manufacturer's instructions. Sequencing was performed using the NovaSeq 6000 system with 150 bp paired-end reads (Illumina, San Diego, CA, USA). Sequence adaptors and low-quality reads were filtered using Trimmomatic v.0.40 [51]. De novo genome assembly was performed using SOAPdenovo v.2.0.4 [52]. The assembled genome was analysed and annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) v.5.20 [53]. Next, the protein-coding genes were clustered into functional groups using evolutionary genealogy of genes: Non-supervised Orthologous Groups (eggNOG) v.5.0 [54]. Metabolic pathways were predicted using BlastKOALA v.2.2 [55] based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Genome comparison between strain PS-C1 and all 14 available genomes of Celeribacter spp. in the NCBI Genome database (available as of 1 January 2022) was performed using digital DNA-DNA hybridisation (dDDH) in the Genome-to-Genome Distance Calculator (GGDC) v.2.1 [56] and the average nucleotide identity (ANI) function in the EzBioCloud server [57]. Default parameters were used for all software tools unless otherwise specified.

Analysis of CAZymes and Mining of GHs
The putative genes encoding CAZymes present in the genome of strain PS-C1 and all 14 available genomes of Celeribacter spp. were mined using the dbCAN2 meta server [58]. The InterProScan v.5.53-87.0 [59] and PSORTb v.3.0.3 [60] online servers were used to predict the protein domains and localisation of the annotated GHs, respectively. Unless otherwise specified, default parameters were used for all the software tools.

Expression and Determination of BglPS-C1 and LicPS-C1 Activities
Strain PS-C1 was grown on marine agar (pH 6.5) at 30 • C for 24 h. A single colony of strain PS-C1 was inoculated into 50 mL of marine broth in a 250 mL flask and shaken at 200 rpm at 30 • C for 24 h. To induce the expression of both BglPS-C1 and LicPS-C1 enzymes, a 20 mL inoculum (equivalent to 10% v/v) was aseptically transferred into 200 mL of marine broth supplemented with 1.0% (w/v) cellobiose in a 1 L flask. All the flasks were incubated at 30 • C with shaking at 200 rpm. At periodic time intervals, 5 mL of culture medium was sampled for up to 30 h. The absorbance was recorded at 600 nm using the Ultrospec 2100 pro UV/Visible Spectrophotometer (Cytiva), and the cells and cell-free supernatant were separated by centrifugation at 5000× g for 15 min at 4 • C. The cells and cell-free supernatant were stored at −80 • C and −20 • C, respectively, until further use. To obtain BglPS-C1 (intracellular enzyme), the cell pellets were lysed using the B-PER TM Bacterial Protein Extraction Reagent kit (Thermo Fisher Scientific, Rockford, IL, USA), according to the manufacturer's instructions. The cell-free lysate was dialysed against 100 mM sodium phosphate buffer (pH 6.5) for 18 h at 4 • C using SnakeSkin dialysis tubing with a 10 kDa molecular weight cut-off (Thermo Fisher Scientific). Subsequently, βglucosidase activity was determined according to the method described by Chan et al. [61]. A reaction mixture containing 200 µL crude BglPS-C1 and 800 µL 10 mM pNPG in 100 mM sodium phosphate buffer (pH 6.5) was incubated at 50 • C for 15 min. The reaction was stopped by the addition of 1 mL 1 M Na 2 CO 3 . Subsequently, the release of p-nitrophenol was measured at 405 nm using the Ultrospec 2100 pro UV/Visible Spectrophotometer (Cytiva) at 405 nm. As a control, a reaction mixture without the enzyme was incubated and measured under the same conditions. p-Nitrophenol was used as the assay standard. One unit (U) of β-glucosidase activity was defined as the amount of enzyme that produced 1 µmol of p-nitrophenol per min per mL under the assay conditions. The enzyme activities were measured at least in triplicate, unless otherwise specified. To determine LicPS-C1 activity (extracellular enzyme) in the bacterial strain, the cell-free supernatant was allowed to react with pNPG (containing only β-1,4 glycosidic bonds) and β-glucan (containing both β-1,4 and β-1,3 glycosidic bonds). LicPS-C1 activity towards pNPG was determined in a similar manner as that for BglPS-C1. One unit (U) of licheninase activity was defined as the amount of enzyme that produced 1 µmol of p-nitrophenol per min per mL at 50 • C. LicPS-C1 activity towards natural substrates (β-glucan) was determined using the 3,5-dinitrosalicylic acid (DNS) method [62]. 500 µL each of crude LicPS-C1 and 1% (w/v) βglucan dissolved in 100 mM sodium phosphate buffer (pH 6.5) were mixed and incubated at 50 • C for 30 min. DNS (1 mL) was then added to the mixture, followed by boiling for 5 min. Subsequently, the absorbance was measured at 540 nm using the Ultrospec 2100 pro UV/Visible Spectrophotometer (Cytiva). As a control, the unreacted mixture was incubated and analysed under the same conditions. Glucose was used as the assay standard. One unit (U) of licheninase activity was defined as the amount of enzyme that generated 1 µmol of reducing sugar per min per mL at 50 • C. All enzyme assays were performed in triplicate, unless otherwise specified. The reaction products of LicPS-C1 on β-glucan were analysed using ultra-high-performance liquid chromatography with an evaporative light-scattering detector (UHPLC-ELSD). The enzymatic reaction mixture was prepared by incubating crude LicPS-C1 with 1% (w/v) β-glucan in 100 mM sodium phosphate buffer (pH 6.5) at 50 • C for 48 h. At certain time intervals, the sample was withdrawn and the enzymatic reaction was stopped by boiling for 10 min. The insoluble particles were filtered through a 0.22 µm nylon-membrane syringe filter (Millex-GN, Merck Millipore, Darmstadt, Germany). A Shimadzu Nexera X2 UHPLC system with Shimadzu Nexera X2 ELSD (Shimadzu, Kyoto, Japan) and Rezex RSO-Oligosaccharide Ag + column (10 × 200 mm; Phenomenex, Torrance, CA, USA) were used for the UHPLC-ELSD analysis. The column temperature was maintained at 80 • C. The ELSD nebuliser and evaporator temperatures were maintained at 30 • C, and standard N 2 gas flow was maintained at 1.6 standard litres per min. Water (100% v/v) was used as the mobile phase at a flow rate of 0.2 mL/min. Glucose (Dp 1 ), cellobiose (Dp 2 ), and cellotriose (Dp 3 ) were used as standards for the analyses. Unreacted substrate was used as the control.
All results of the enzymatic assays and UHPLC-ELSD analyses were statistically analysed using SYSTAT 12 software (Systat Software, San Jose, CA, USA). A Student's t-test yielded a probability value (p-value) of less than 0.05, confirming that the data were adequate to test all hypotheses.

Sampling Site, Isolation, Taxonomy Identification, and Bacterial Characterisation
Sekinchan beach, located in Selangor, Malaysia, is a hot spot for tourism. The beach landscape is sandy with tiny grains of rocks, and some parts of the beach are muddy and rich in mangrove trees. The temperature and pH of the collected samples were 30-32 • C and pH 5.5-6.5, respectively.

Figure 2.
A field emission scanning electron micrograph of strain PS-C1 at 24,000× magnification. Scale bar: 5 µm. Table 1. Comparison of morphology and biochemical characteristics between strain PS-C1 and type strains of the genus Celeribacter.   Celeribacter spp. cells are singular and not filamentous or chain-shaped [19,23,24]. Similar to members of the genus Celeribacter, strain PS-C1 was positive in the catalase and oxidase tests, but negative in the motility test (Table 1). Except for C. neptunius DSM26471 T , the strain tested negative in the oxidase test [19]. Strain PS-C1 was also found to be a non-spore-forming bacterium, and was unable to hydrolyse Tween 20 and Tween 80.
The growth of strain PS-C1 was observed at temperatures of 30-40 • C and pH 5.5-10.0. The optimal growth conditions were determined to be 30 • C and pH 6.5. These results indicated that strain PS-C1 is a mesophilic and mildly acidophilic bacterium. Celeribacter spp. are known to be mesophiles with optimal growth temperatures of 25-30 • C, except for C. ethanolicus NH195 T , which grows optimally at 37 • C (Table 1). Furthermore, most Celeribacter spp. are moderate alkaliphiles and grow optimally at neutral pH (Table 1). However, the optimal pH for growth of C. baekdonensis DSM27375 T and C. persicus DSM100434 T is pH 5.0 and pH 6.0, respectively, under acidic conditions [23,26]. Salt tolerance studies indicated that strain PS-C1 growth occurred at NaCl concentrations of 1-8% (w/v), with optimal growth at 2-7% (w/v) NaCl. The results showed that strain PS-C1 was a halophile, similar to all members of the genus Celeribacter (Table 1). Strain PS-C1 was found to be a facultative anaerobic bacterium (that could grow with and without oxygen). Among the members of the genus Celeribacter, only C. neptunius DSM26471 T [19] and C. indicus P73 T [24] are facultative anaerobes, while other Celeribacter spp. are strict aerobes (Table 1).
Based on the data obtained from the API 20NE, API 20E, and API 50CH analyses, strain PS-C1 was able to utilise a wide range of chemicals and carbon sources, including nitrate, citrate, urea, D-glucose, L-arabinose, D-fucose, and arbutin (Tables 1 and S1).
Selective enzymatic reactions of strain PS-C1 were identified using API ZYM, and the activities of esterase (C4), leucine arylamidase, acid phosphatase, α-glucosidase, and β-glucosidase were detected (Tables 1 and S1). Other members of the genus Celeribacter can also degrade/uptake various substances. For more information, readers may refer to the literature on Celeribacter type strains, as listed in Table 1. Collectively, the results of phylogenetic analysis and phenotypic and chemotaxonomic properties indicated that strain PS-C1 belongs to the genus Celeribacter.

Genome Sequencing, Assembly, and Annotation
The genome of strain PS-C1 was sequenced, and its genomic features are summarised in Table 2.
The sequencer generated 1,671,186,900 bases from 5,570,623 paired-end reads. The genome was assembled into 40 contigs and showed a coverage of 357-fold. The largest contig was 499,873 bp, with N 50 and N 90 values of 302,457 bp and 80,703 bp, respectively. The draft genome size of strain PS-C1 was determined to be 3,866,278 bp (3.87 Mbp), which is smaller than that of other members of Celeribacter spp., such as C. neptunius DSM26471 T (4.97 Mb), C. indicus P73 T (4.40 Mbp), C. naphthalenivorans EaN35-2 (4.36 Mbp), and C. ethanolicus NH195 T (4.21 Mbp), but larger than C. marinus IMCC12053 T (3.10 Mbp) (Table S2 and Figure 1). An in-depth analysis using the collective information of Celeribacter genomes (Table S2) indicated that Celeribacter has a 44% larger genome size than its closest genus in the family Rhodobacteraceae: Nereida (~2.87 Mbp) [63]. The G+C content of strain PS-C1 was 59.10%, which was slightly lower than that of C. naphthalenivorans EaN35-2 (59.60%) and Celeribacter sp. HF31 (59.90%). Several members of Celeribacter spp. exhibited higher G+C content than strain PS-C1, such as C. indicus P73 T (65.73%), C. neptunius DSM26471 T (61.70%), and C. ethanolicus NH195 T (61.30%) (Table S2). On average, the G+C content of the genus Celeribacter (~59.62%) was higher than that of the closest genus Nereida (~54%) [63]. We then determined the taxonomic affiliation of strain PS-C1 by comparing its genome with all the available genomes of Celeribacter spp. (Table 3). Strain PS-C1 exhibited 18.20-37.50% dDDH and 72.50-89.30% ANI values with members of Celeribacter spp. The closest relative of strain PS-C1 was C. naphthalenivorans EaN35-2 (dDDH, 37.50%; ANI, 89.30%). As the values for dDDH (<70%) [64] and ANI (<96%) [57] were below the corresponding thresholds, these results indicated that strain PS-C1 is a new species of Celeribacter. Based on the NCBI PGAP annotation, the strain PS-C1 genome consisted of 3818 predicted genes, of which 3739 were protein-coding sequences, 54 were noncoding RNA genes (48 tRNAs, 3 ncRNAs, and 3 rRNAs), and 25 were pseudogenes (Table 2). Of these, 1069 protein-coding sequences (28.59% of total protein-coding sequences) were found to be exclusive to the Celeribacter spp. with at least 90% sequence identity (Table S3). In addition, 2112 protein-coding sequences (56.49%) were associated (average sequence identity of 89.70%) with their respective counterparts from other genera in the family Rhodobacteraceae. A small portion of the proteins in the strain PS-C1 genome (96 proteins, 2.57%) were related (~67.72% identity) to their homologues from various bacterial families such as Rhizobiaceae, Ahrensiaceae, Cohaesibacteraceae, and Halomonadaceae. Furthermore, the strain PS-C1's genome was found to encode a total of 462 hypothetical proteins (12.35% of the protein-coding sequences), as they shared low sequence identities with proteins available in the databases, and these proteins are interesting targets for future research. Compared to the closest member to strain PS-C1, the C. naphthalenivorans EaN35-2 genome had a total of 4206 protein-coding sequences (12.48% more than strain PS-C1), 1710 of which were identical (~92.36%) to their homologues from Celeribacter spp. (Table S3).
The protein-encoding sequences of strain PS-C1 were functionally categorised according to Cluster of Orthologous Groups (COGs) analysis, as shown in Table 4. A total of 3702 (87.51%) protein-coding genes were functionally assigned to COGs in the genome of strain PS-C1. Compared to the analysed Celeribacter genomes (Table S4), all strains had at least 77.67% of their total protein-encoding genes annotated with COG functions. In terms of COG assignment profiles, strain PS-C1 exhibited patterns similar to those of all the Celeribacter genomes. These genes were divided into four major functional groups: metabolism (30.27-46.58%), cellular processes and signalling (12.19-18.30%), information storage and processing (12.41-20.09%), and poorly characterised (15.22-18.99%) (Table S4).
An in-depth comparison between the genomes of PS-C1 and its closest relative C. naphthalenivorans EaN35-2 showed a major difference in the number of predicted genes related to the '(E)-amino acid transport and metabolism' classification, with strain PS-C1 representing 324 genes (8.67%) in the genome, which is remarkably higher than C. naphthalenivorans EaN35-2 (three genes, 0.07%). In contrast, the genome of C. naphthalenivorans EaN35-2 (145 genes, 3.45%) had more genes assigned under the category '(L)-replication, recombination, and repair' compared to that of strain PS-C1 (two genes, 0.05%) (Table S4). In terms of COG class for '(G)-carbohydrate transport and metabolism,' a comparable amount of protein-encoding genes were grouped under this category for genomes of strain PS-C1 (219 genes, 5.86%), as well as C. naphthalenivorans EaN35-2 (211 genes, 5.02%). Some of these proteins (β-glucosidase, licheninase, and α-glucosidase) are known to be involved in the degradation of cellulose and starch [35,36,65,66].
Further inspection using KEGG metabolic pathway analysis indicated that the starch and sucrose metabolism pathways of strain PS-C1 and all the analysed Celeribacter genomes were relatively similar ( Figure S1). According to this analysis, all strains encoded β-glucosidase (EC 3.2.1.2), which is necessary for the degradation of cellulose to glucose. Moreover, there were 15 enzymes present in all Celeribacter genomes predicted to be involved in the hydrolysis of starch to maltodextrin, maltose, and glucose; degradation of glycogen to glucose; and conversion of sucrose to D-fructose ( Figure S1). In addition, the genomes of C. baekdonensis B30 and C. indicus P73 T harbour two additional enzymes. The genome of C. baekdonensis B30 was predicted to contain oligo-1,6-glucosidase (EC 3.2.1.10) and β-fructofuranosidase (EC 3.2.1.26), which are involved in the degradation of dextrin/isomaltose to glucose and conversion of sucrose-6-phosphate to glucose-6-phosphate, respectively. In contrast, the C. indicus P73 T genome contained alpha-trehalose-phosphate synthase (EC 2.4.1.15) and trehalose-phosphatase (EC 3.1.3.12) for the conversion of uridine diphosphate glucose to trehalose.

Analysis of CAZymes and Mining of GHs
The dbCAN2 CAZy server was used to identify, predict, and compile the CAZymeencoded genes in strain PS-C1 and 14 other analysed genomes of Celeribacter spp. An overview of the abundance and distribution of CAZymes in each member of the genus Celeribacter is shown in Figure 3.
The Celeribacter spp. encode a total of 50-86 different CAZymes. In terms of CAZyme classification, GHs and glycoside transferases (GTs) were the most dominant groups (average genes of 54.65% GH;~29.17% GT) in all the genomes of Celeribacter spp. Moreover, small amounts of auxiliary activities (AAs) and carbohydrate esterases (CEs) were detected in Celeribacter spp. (~10.47% AAs;~5.72% CE), whereas none of the Celeribacter spp. possessed encoded proteins assigned to polysaccharide lyases (Figure 3). Taken individually, the strain PS-C1 genome encoded a total of 70 CAZymes (including 16 GHs, 41 GTs, 10 AAs, and 3 CEs), whereas its closest relative C. naphthalenivoransEaN35-2 harboured a total of 84 CAZymes (20 GHs, 47 GTs, 12 AAs, and 5 CEs). The presence of various CAZymes in the genomes of Celeribacter spp. suggested that these proteins are likely responsible for the degradation of polysaccharides (i.e., cellulose and starch), and this hypothesis was in agreement with the carbon utilisation profiles of Celeribacter spp. (Table 1). Currently, none of the CAZymes (particularly GHs) from Celeribacter spp. have been biochemically characterised. We analysed and compared the GHs in strain PS-C1 and all the 14 available genomes of Celeribacter. We hope that the analysis provided herein will pave the way for industrial applications of GH enzymes from Celeribacter spp. Table 5 lists the industrially relevant GH enzymes in Celeribacter genomes. The GHs shared among the analysed Celeribacter genomes could be divided into four categories according to the predicted carbohydrate-hydrolysing functions: (i) cellulose-degrading enzyme (one GH); (ii) β-glucan-degrading enzymes (three GHs); (iii) hemicellulose-degrading enzymes (five GHs); and (iv) starch-degrading enzymes (three GHs) ( Table 5).

Expression and Determination of BglPS-C1 and LicPS-C1 Activities
We then expressed and elucidated the activities of two GHs (BglPS-C1 and LicPS-C1) in strain PS-C1. For BglPS-C1, the protein sequence (451 residues) consisted of a GH1 domain located at position R14-R450, as predicted using the InterProScan server (Figure 4a).
The GH1 domain is a catalytic region where the hydrolysis of β-1,4 glycosidic bonds occur [70]. In terms of enzyme localisation based on the PSORTb web server, BglPS-C1 was predicted to be an intracellular enzyme; thus, the cells were lysed prior to the enzymatic assays. As shown in Figure 4b, BglPS-C1 was constitutively produced throughout the 30 h time course, suggesting its important role in the conversion of cello-oligosaccharides in strain PS-C1. The maximum relative enzyme activity was determined after a 14 h incubation period, which was equivalent to the exponential phase of cell growth. Similar to other studies, the optimum β-glucosidase expression in Fusobacterium sp. K-60 and Pseudomonas pickettii were recorded during the exponential phase of growth [71,72].
As LicPS-C1 was deduced to be an extracellular enzyme, the cell-free supernatant was used to measure the enzyme activities of pNPG (containing only β-1,4 glycosidic bonds) and β-glucan (containing both β-1,4 and β-1,3 glycosidic bonds). As shown in Figure 5b, LicPS-C1 could act on both pNPG and β-glucan, indicating that the enzyme was actively hydrolysing β-1,4 and β-1,3 glycosidic linkages. The LicPS-C1 expression pattern on both pNPG and β-glucan was as follows: the relative enzyme activities increased gradually from 0 to 8 h, reached their optimum at 10 h of incubation, and decreased within 12-30 h of incubation (Figure 5b). In separate studies, similar patterns were observed in the production of licheninase from Bacillus subtilis HL-25 and Bacillus subtilis GN156 at 72 h and 24 h of incubation, respectively [75,76]. The end-product (sugar) profile of LicPS-C1 on β-glucan over a 48 h time course was determined using UHPLC-ELSD analysis (Figure 5c). At the beginning of the time plot (6 h), the relative amount of total sugar produced was 77.62%. The total relative amount of sugar then increased to 98.94% at 24 h and reached its optimum (100%) after 36 h of incubation. In terms of types of sugars, glucose (Dp 1 ), cellobiose (DP 2 ), and cellotriose (DP 3 ) were produced at various ratios throughout the analysis. The majority of sugars produced were cellobiose (36.22-62.41%) and cellotriose (19.02-22.79%). Glucose was released at a constant amount (~18.59%) over the 48 h reaction period. Altogether, the GHs of strain PS-C1 (i.e., BglPS-C1 and LicPS-C1) are interesting new enzymes for biomass degradation. Further studies on the gene cloning, purification, and functional biochemical characterisation of these enzymes may reveal their potential biotechnological applications.  Time courses of strain PS-C1 growth and LicPS-C1 production by cultivation in the marine broth supplemented with 1% (w/v) cellobiose. The LicPS-C1 activities were detected using p-nitrophenyl-β-D-glucopyranoside (pNPG) and β-glucan as the enzymatic substrates. The values shown represent the mean ± standard error of triplicate analyses. (c) Analysis of reaction products produced by LicPS-C1 acting on β-glucan at different time intervals using ultra-highperformance liquid chromatography with an evaporative light-scattering detector (UHPLC-ELSD). The amount of total sugars produced at each time point is shown relative to that at 48 h. Dp 1 : glucose; Dp 2 : cellobiose; Dp 3 : cellotriose.

Conclusions
In this report, we described phenotypic, chemotaxonomic, and phylogenetic analyses of strain PS-C1. These results collectively suggested that strain PS-C1 represents a new member of the genus Celeribacter. Additionally, we presented the genomic features of Celeribacter sp. PS-C1 and provided the first comprehensive analysis of the underexplored GHs within the genomes of Celeribacter spp. Furthermore, two GHs from Celeribacter sp. PS-C1 (β-glucosidase and licheninase) were expressed, and their activities were analysed. Based on genomic and experimental data, Celeribacter sp. PS-C1 is an attractive reservoir for novel GH enzymes that might be useful in biomass saccharification.

Supplementary Materials:
The following supporting information can be downloaded at: https: