*2.7. Deletion of GH19 Chitinase Does Not Significantly Influence the Proteome of S4059*

To investigate the impact of the GH19 chitinase deletion on the proteome of S4059, the GH19 mutant, and the wild type were grown on a chitin-based medium to stationary phase. Supernatant and cells were separated for proteome analysis. A total of 2512 proteins were identified, but no significant changes in any of the detected proteins were observed between wild type and GH19 mutant in either supernatant or cells when grown on crystalline chitin, except that GH19 chitinase could not be detected in the mutant cultures.

#### *2.8. Similar Secondary Metabolome Pattern in Wild Type Strain and GH19 Mutant*

Following the measurement of growth kinetics, the extracts of cultures grown in MMM were analyzed by HPLC-HRMS. Each strain was cultivated in three biological replicates in media containing one of the four carbon sources: mannose, NAG, colloidal chitin, or shrimp chitin. The wild type and the mutant had similar secondary metabolome profiles, as shown in Figure S1A. Screening for known natural products was also undertaken, revealing that only prodigiosin, as well as three of its analogs, could be identified (Figure S1B).

#### **3. Discussion**

Chitinolytic bacteria have been widely studied due to the possible use as antifungal agents in biocontrol [25]. In addition, the involvement of chitin in the ecology of *Vibrio cholerae* and its virulence regulation has been the focus of many studies [26–28]. Recently, a novel perspective on chitin degradation has been seen in some strains of *Vibrionaceae*, where growth on chitin as compared to on glucose leads to enhanced expression of several BGCs [19] along with enhanced antibacterial activity [20]. Chitin is the most abundant

polymer in the ocean, and, hence, many marine bacteria are potent chitin degraders [10]. In silico analysis has demonstrated that pigmented species of the genus *Pseudoalteromonas* have elaborate chitinolytic machinery containing at least one of the otherwise rare GH19 chitinase in their genomes, and also, they devote up to 15% of their genome to BGCs with a vast potential for secondary metabolite production [7]. Here, we demonstrated that the pigmented bacterium *P. rubra* S4059 grew well using both purified chitin (crystalline and colloidal chitin) and natural shrimp shells as a sole source of nutrients. These results were further substantiated by proteome analyses that demonstrated that the strain S4059, indeed, produces an efficient chitinolytic enzyme cocktail, including GH18, GH19, LPMO, and GH20.

GH18 chitinases are the predominant chitinolytic enzymes in bacteria [29–31], and the strain S4059 harbored seven putative GH18 chitinases as determined by genome analysis, four of which were significantly upregulated when grown on chitin. Secretome analysis of the soil-derived and chitinolytic bacterium *Cellvibrio japonicus* indicated that many chitinolytic enzymes, including GH18 chitinases, and GH19 chitinases and LPMOs, were highly upregulated when grown on β-chitin as compared to α-chitin [29]. Out of fourteen putative chitinolytic enzymes in S4059, only the two putative enzymes, GH19 A0A5S3UPT5 and LPMO A0A5S3V4S2, were not detected when S4059 was grown on an α-chitin surface, suggesting that these two enzymes may have other functions. Proteome analysis of *P. rubra* S4059 culture supernatant confirmed the presence of all chitinolytic enzymes except for the two enzymes GH19 A0A5S3UPT5 and LPMO A0A5S3V4S2, and an even higher amount of the enzymes in culture supernatant than in cells (Figure 1C), implying that those detected proteins indeed are secreted enzymes able to degrade chitin. Secretome analysis of *Cellvibrio japonicus* showed that thousands of proteins were detected in the filtered culture supernatant. However, the authors mentioned that cell lysis can lead to overestimating the numbers of secreted proteins [29]. Further, they used the plate-based method that was developed by Bengtson et al. [32] to detect truly secreted proteins, and the result indicated that only 267 secreted proteins were detected in α-chitin containing medium [29]. In our study, due to the insoluble property of chitin, secreted chitinases may bind to small chitin particles, and, thus, the secreted chitinases cannot be passed through sterile filters leading to misestimating the expression of secreted chitinases. Thus, we did not filter the culture supernatant for retaining bound chitinases, which were on chitin particles. We found that 880 proteins were influenced by chitin in the culture supernatant of S4059, as compared to on mannose, and these high numbers of influenced proteins may be due to a combination factor of the unfiltered culture supernatant and cell lysis during centrifugation [33]. Further, the agar plate-based method probably can point out the true numbers of secreted proteins in S4059.

One of the two GH19 chitinases in *P. rubra* S4059 was highly expressed when the bacterium was grown on chitin, and while some GH19 chitinases have been linked to antifungal activity, their broader role in bacteria remains enigmatic [18,34,35]. The results suggesting antifungal activity have mainly relied on heterologous expression of GH19 in hosts such as *E. coli,* and this provides information about the enzyme per se, but not necessarily about the actual role in its native host. We, therefore, chose to delete the GH19 that was highly expressed on chitin to explore its possible role in S4059 further. Both the mutant and wild type displayed similar antifungal activity (data not shown). Despite the high expression of GH19 when S4059 was grown on chitin, the GH19 deficient mutant showed no difference in growth compared to the WT when grown on chitin (Figure 2 and Figure S2). Our results, therefore, indicate that GH19 chitinase is not the predominant chitin degrading enzyme in *P. rubra* S4059, given the wide array of other chitinolytic enzymes. Chitin utilization capacity has been investigated in a chitinolytic soil bacterium *Cellvibrio japonicus* [30]. Through a combination of secretome and genetic manipulation approaches, a highly expressed GH18 chitinase was identified as the enzyme essential for the degradation of chitin in the strain [30]. Proteome analyses in S4059 demonstrated that two GH18 chitinases A0A5S3USH6 and A0A5S3USE2 are the most highly expressed

chitinases in both cells and culture supernatant according to protein abundance when S4059 was grown on chitin (Table S7), suggesting that the GH18 chitinases could be potentially essential enzymes for chitin degradation in S4059 under testing conditions.

In conclusion, the proteomic analysis showed that a highly efficient chitin degradation machinery was identified in pigmented *P. rubra* S4059, and four GH18 chitinases and two GH20 hexosaminidases in S4059 were significantly upregulated when grown on chitin. In contrast to different proteome profiles, the different growth conditions on chitin investigated here did not significantly alter the metabolite profile of S4059, in contrast to what has been reported in other *Vibrio* species and *Streptomyces coelicolor* A3 (2) [19–22]. Although the deletion of GH19 chitinase did not influence chitin degradation activity in S4059, we developed a conjugation-based approach allowing genetic manipulation in this bioactive bacterium S4059. Given the large potential for secondary metabolism as found by genome mining and from uncharacterized compounds produced on especially chitin derived substrates, the genetic approach developed here will allow exploration of the novel chemical space of this organism.
