**2. Results and Discussion**

*2.1. Genomic Comparison of Zobellia Species*

The characteristics of the ten publicly available *Zobellia* genomes and two unpublished ones (*Z. russellii,* KMM 3677T, and *Z. barbeyronii*, KMM 6746T, our data) were compared. Their genome sizes and GC content ranged from 4.92 to 5.52 Mb and from 36.7% to 42.8%, respectively. To clarify, the phylogenomic relationships of the *Zobellia* species, a phylogenetic tree of these genomes together with two other genomes of type strains from related genera, were built using PhyloPhlAn [22] based on 400 concatenated proteins. According to the genome tree (Figure 1a), all of the *Zobellia* strains clustered together, and the five subclades could be distinguished. The first subclade included a type strain of *Z. uliginosa* together with *Z. galactanivorans* strains, while other subclades consisted of type strains of *Z. laminariae* and '*Z. barbeyronii'*, as well as *Zobellia* sp. strains (Asnod2-F07-B and Asnod3-E08-A). The separated subclades were formed by strains of *Z. russellii*, *Z. amurskyensis* strains, and *Zobellia* sp. (Asnod1-F08 and Asnod2-F02-B). This clustering indicates a closer genome sequence similarity of the strains within subclades. Interestingly, the subclade *Z. uliginosa*/*Z. galactanivorans* was at the base of the *Zobellia* clade, implying an evolutionary divergence from other species of the genus.

**Figure 1.** The phylogeny and CAZome of the genus *Zobellia*: (**a**) maximum-likelihood phylogeny based on 400 universal markers selected by PhyloPhlAn3.0 and reconstructed by RAxML with non-parametric bootstrapping using 100 replicates; (**b**) the bar plot showing the number of carbohydrate-active enzyme genes according to the CAZy classification for each strain. GH, glycoside hydrolase; GT, glycosyltransferase; PL, polysaccharide lyase; CE, carbohydrate esterase; AA, auxiliary activity; CBM, carbohydrate-binding module.

> Previous studies have revealed that the *Zobellia* genomes are abundant with CAZymes genes, which encode the ability to efficiently degrade complex polysaccharide substrates [14,23]. It has been known that carbohydrate degraders are characterized by a high proportion of CAZymes found in their genomes (more than 5% of the predicted protein-coding genes). To identify CAZymes in the *Zobellia* genomes, we used the dbCAN2 meta server (http://cys.bios.niu.edu/dbCAN2, December 2020) [24].

We found that in all *Zobellia* genomes CAZymes account for more than 6% (from 6.4% in *Zobellia* sp. Asnod3-E08-A to 7.6% in *Z. galactanivorans* DsijT), which reflects the outstanding specialization of representatives of this genus in the degradation of polysaccharides. The total statistics of CAZyme classes predicted across the genomes are shown in Figure 1b. We found that *Z. galactanivorans* DsijT possesses the highest number of CAZymes (336), followed by *Z. amurskyensis* MAR 2009 138 (320), *Z. uliginosa* DSM 2061<sup>T</sup> (315), and *Z. galactanivorans* OII3 (311). The smallest numbers of CAZymes were in *Zobellia* sp. Asnod3-E08-A (263) and Asnod2 B07BT (266). Notably, the genome of Z. *galactanivorans* Dsij<sup>T</sup> encodes the largest and most diverse CAZYme repertoire, with approximately 60.7 CAZyme genes per Mb, in contrast to *Zobellia* sp. Asnod3-E08-A (52.5 CAZyme genes per Mb). However, these values indicate again a broad degradation potential conserved at the genus level [23].

*Z. galactanivorans* is well known to degrade alginates of brown algae [25] due to its alginolytic system including alginate lyases of distinct polysaccharide lyase families. PLs are a group of enzymes that cleave uronic acid-containing polysaccharides via a βelimination mechanism [26]. In order to elucidate the alginolytic potential of *Zobellia*, we calculated the content of the PL genes in these genomes (Figure 2). The heat map shows the frequency of the PL genes dedicated to PL families. Notably, the genomes of *Z. amurskyensis* MAR 2009 138 and Z. *galactanivorans* Dsij<sup>T</sup> encode 25 and 24 PLs, while the genomes *Zobellia* sp. Asnod3-E08-A and Asnod2 B07BT encode only 17 and 16 PLs, respectively. Among the identified PLs, the genus-specific ones were PL1, PL6, PL7, PL14, PL17, and PL40. Since all *Zobellia* genomes contain PL genes of families PL6, PL7, PL14, and PL17, this indicates that they all have a functional alginate utilization pathway. PL7 is an important enzyme in the utilization of alginate. Surprisingly, PL7 genes are the most abundant PLs in *Zobellia* genomes, accounting for one to five copies (Figure 2).

**Figure 2.** The distribution of polysaccharide lyase families across *Zobellia* genomes. The heat map shows the number of genes assigned to individual PL families. Rows are clustered using Euclidean distances. ZspF08, *Zobellia* sp. Asnod1-F08; ZspB02B, *Zobellia* sp. Asnod2-B02-B; ZrusT, *Z. russellii* KMM 3677T; ZgOII3, *Z. galactanivorans* OII3; ZulT, *Z. uliginosa* DSM 2061T; ZgaT, *Z. galactanivorans* DsijT; ZspB07B, *Zobellia* sp. Asnod2-B07-B; ZspE08A, *Zobellia* sp. Asnod3-E08-A; Zammar, *Z. amurskyensis* MAR 2009 138; ZbarT, *Z. barbeyroni'* KMM 6746T; ZlamT, *Z. laminariae* KMM 3676T; ZamT, *Z. amurskyensis* KMM 3526T.
