*2.3. Bifidobacterial GH151 Fucosidases*

GH151 enzymes form the smallest group of fucosidases (Table S3) and although there are still doubts about their fucosidase activity, *B. longum* subsp. *infantis* ATCC 15697 counts, with a GH151 enzyme (Blon\_0346) that exhibits probed Fuc-α1,2 Gal activity [21]. Interestingly, bifidobacterial GH151 fucosidases are quite divergent from the fucosidases classified in other GH families [21] and all of them belong to *B. longum* subsp. *infantis* species although they show little differences in their sequences (Figure 4). While no signal peptide or transmembrane helices were observed, CDD architecture analyses revealed AmyAc\_family superfamily and A4\_beta-galactosidase\_middle\_domain, although some sequences are also identified as containing GanA superfamily domain as well (Table S3).

GH151 enzymes probably have domains closest to GH29-BifC fucosidases, identified by containing conserved AmyAc superfamily domain and likely the ability to hydrolyze α glycosidic linkages [31]. However, because GH151 accessory domains shown (Table S3), they could be considered as potential non-specific beta galactosidase enzymes with the capacity to hydrolyze Fuc-α1,2 Gal linkages as occurs with Blon\_0346. Nevertheless, further studies in order to elucidate their subjacent activity, substrate specificity, and conformational structure are needed to understand their role in the hydrolysis of fucosylated carbohydrates.

**Figure 4.** Phylogenetic analysis of bifidobacterial GH151 fucosidases. PCA (**A**) and cladogram tree (**B**) distributions of bifidobacterial GH151 complete fucosidase sequences listed in CAZy, released from Jalview 2.11.1.4 software using the neighbor-joining method.

#### **3. Discussion**

Breast milk, beyond its nutritional function, provides the necessary pillars for the initial establishment of the gut microbiota in newborns. In this regard, FHMOs and FHMGs stand out for their ability to stimulate the growth of bifidobacteria [8,12], which in turn produce SCFAs such as acetate, formate, lactate, and pyruvate [13], stimulating the immune system [14], and serving as an energy source for colonocytes [37].

Although only a few bifidobacterial species have been studied extensively at both cellular and genomics level for their ability to utilize fucosylated carbohydrates such as *B. bifidum* and *B. longum* subsp. *infantis* [22,23], their success in colonizing the gut is due to the different strain-dependent metabolic abilities developed for the use of both FHMOs and FHMGs [24]. Therefore, fucosidases play a key role in the bifidobacterial gut establishment. Concerning to that, *B. bifidum* strains show two extracellular fucosidases belonging to GH29 and GH95 families. Both fucosidases cover the hydrolysis of Fucα1,3Glu; Fuc-α1,3/4GlcNAc; and Fuc-α1,2 Gal linkages [17,18,27,33]. Since *B. bifidum* prefers the utilization of lactose [28], 2 -fucosyllactose could be its target substrate for its extracellular fucosidases, releasing to the environment lactose and fucose, the last could be also liberated from blood Lewis a, b, x, and y antigens [27]. For all the above, *B. bifidum* fucosidases could be considered altruistic and essential for microbial gut establishment through promoting bifidobacterial mutualism and carbohydrate syntrophy in the infant gut [38]. Given that bifidobacteria are able to metabolize lactose, and species such as *B. longum* subsp. *infantis* or *B. breve* can metabolize fucose, their growth is improved under the presence of fucosidases from *B. bifidum*. Thus, Gotoh et al. (2018) suggested that extracellular fucosidases from *B. bifidum* could be crucial during the development of a bifidobacteria-rich microbiota in the breastfed infant gut, by providing fucosylated conjugate degradants [33]. On the other hand, *B. bifidum* fucosidases contribute to the protection of the host through the modification of Lewis antigens [27].

Regarding the catalytic domains of the *B. bifidum* fucosidases, it should be noted that GH29-BifA present orthologous fucosidases in other bifidobacterial species clustered in GH29- BifB/D, and they probably all have a common phylogenetic lineage (Figure 2). However, this statement has only been functionally corroborated through the characterization of the enzymes AfcB (GH29-BifA) and Blon\_2336 (GH29-BifB), due to lack of results of GH29-BifD fucosidases.

Conversely, GH95-BifA fucosidases as well as those grouped in GH95-BifB, and according to CDD database observations (Table S2), could phylogenetically descend from either an evolutionary specialization or non-specification of glycosidases clustered in GH65. Indeed, this in silico observation agrees with the crystallization results obtained for the structure AfcA from *B. bifidum* [18]. According to that, both GH65 and GH95 enzymes share an α/α 6 barrel fold with inverting mechanism and glutamate566 as catalytic proton donor. Moreover, Nagae et al. (2007) compared the structures between families GH65 and GH95, revealing conservation of the general acid residues, except for catalytic acid/base aspartate766, which is shifted in AfcA [18]. That shifting was also found in the rest of the bifidobacterial GH95 fucosidases (data not shown), and agreeing with the above mentioned authors, the reaction mechanisms of bifidobacterial GH95 fucosidases differ from those of the GH65 family [18].

The other species widely studied for its fucosidase activity is *B. longum* subsp. *infantis*. Actually, it is the only species of bifidobacteria that exhibits GH29, GH95, and GH151 fucosidases that have been recombinantly purified and characterized [21]. Those fucosidases allow *B. longum* subsp. *infantis* to use a wide range of substrates, hydrolyzing Fuc-α1,3Glu; Fu-cα1,2/3Gal; and Fuc-α1,3/4/6GlcNAc linkages [21,32]. As previously commented, *B. longum* subsp. *infantis* GH29-BifB fucosidases are orthologous with those classified in GH29-BifA. However, this species also shows GH29-duplicated fucosidases, clustered in the GH29-BifC, with different architecture and paralogs from those of GH29-BifB (Figure 3). Taking into account the fucosidase duplication and in agreement with You et al. (2019), *B. longum* subsp. *infantis* GH29-BifC fucosidases could have evolved from a different

glycosyl hydrolase [30]. According to CDD database observations (Table S1) and because their predicted structure is composed by a β/α 6 barrel fold with retaining mechanism and glutamate as catalytic proton donor, GH29-BifC fucosidases from *B. longum* subsp. *infantis* could descend from GH13 glycosidases (α-amylases).

GH29-BifC fucosidases, similar to GH95-BifB, which is probably phylogenetically originated from GH65 family as described above, need to have their structural crystallization further explored in order to elucidate their origins and evolution pathway. In addition, GH29-BifC fucosidases show similarities with metazoan fucosidases according to the InterPro database (Table S1), including aspartate224 and glutamate270 residues (data not shown), which play the role of the catalytic nucleophile and catalytic acid/base, respectively, in metazoan fucosidases [25].

Finally, GH151 fucosidases are exclusively present in *B. longum* subsp. *infantis*. This fact could suggest a fourth pathway of fucosidases phylogenetic evolution in that species closely related to GH29-BifC fucosidases, since they present a N-terminal α amylase catalytic domain. In addition, Blon\_0346 was originally classified as a member of GH29 family due to their fucosidase activity despite low similarity [21]. However, GH151 enzymes may be the result of a branch in the evolution of GH29-BifC fucosidases, since they show a GH42 beta galactosidase trimerization architecture instead of conserved features of metazoan fucosidases.

#### **4. Materials and Methods**

#### *4.1. Identification and Selection of Fucosidase Sequences*

Complete bifidobacterial fucosidase protein sequences belonging to GH29, GH95, and GH151 families were retrieved from CAZy database [19]. Fucosidase sequences were used as probes in PSI-BLAST searches [39] against the NCBI [40], Swiss-Prot [41], and Ensembl [42] protein databases.

#### *4.2. Protein Sequence, Alignment, and Phylogenetic Analysis of α-L-Fucosidases*

Fucosidase sequences were analyzed using SignalP-5.0 [43], with default options to predict signal peptide sequences: SOSUI [44] and HMMTOP [45] with default parameters for the prediction of transmembrane helices. NCBI Conserved Domains Database (CDD) [46] and InterPro databases (EMBL\_EBI) [47] were used to predict the domain architecture. Inferred fucosidase amino acid sequences were aligned using Clustal Omega web version [48]. All sequences belonging to the same GH families were considered in phylogenetic analyses. Neighbor-joining method cladogram and PCA analyses were performed using the program Jalview 2.11.1.4 [49].

#### **5. Conclusions**

This is the first study that explores phylogenetically the three families of the bifidobacterial fucosidases: GH29, GH95, and GH151, through their conserved architecture, showing that *B. bifidum* and *B. longum* subsp. *infantis* reveal two and four different phylogenetic lineages, respectively, belonging to different fucosidase families. On the other hand, given the differences in the catalytic architecture observed in this work, the bifidobacterial fucosidases belonging to the GH29 and GH95 families could be subclassified into four and two groups, respectively.

Taking into account that the observations described in this work were obtained in silico and supported by current characterization results from some *B. bifidum* and *B. longum* subsp. *infantis* fucosidases, further studies regarding structural characterization and physicochemical properties of more fucosidases identified by computational analysis are needed in order to validate the novel classification of bifidobacterial fucosidases here proposed.

Concerning to *B. longum* subsp. *infantis* fucosidases, which evolved from different GH families such as GH29-BifC, GH95-BifB, and GH151, and given that their conserved architecture presents vestiges of ancestral glycosidases GH13, GH65, and GH42, respectively, as well as *B. Bifidum* GH95-BifA fucosidases phylogenetically descended from GH65,

deepening substrate spectrum analyses could determine their underlying roles in those species. In this context, and since some fucosidases have been used to transfucosylate carbohydrates or glycoconjugates, the application of these evolved and hypothetically nonspecific *B. longum* subsp. *infantis* fucosidases mentioned above can open a new perspective towards the synthesis of novel fucosylated conjugates by using different substrates beyond lactose for synthetizing 2 -fucosyllactose. This vision is oriented towards the supply those novel fucosylated conjugates to adults in combination with fucosidase producer bifidobacteria in order to maintain a healthy microbiota or to reestablish it from dysbiosis states as described previously [50,51]. In this regard, it would be important to elucidate phylogenetically, as well as structurally and physicochemically, the fucosidases of many other gut microorganism genera, as for instance *Lactobacillus, Bacteroides,* and *Akkermansia*, with the aim to reveal the whole gut fucosidase interaction.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/ijms22168462/s1.

**Author Contributions:** Conceptualization, J.A.C.; methodology, J.A.C. and J.M.L.; validation and formal analysis, J.A.C. and Á.P.; investigation, J.A.C., Á.P., J.M.L., S.L. and A.R.d.l.B.; writing—original draft preparation, J.A.C. and J.M.L.; writing—review and editing, J.A.C. and Á.P.; visualization and supervision, J.A.C. and J.L.A.; project administration, J.A.C.; funding acquisition, J.A.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Spanish Ministry of Science and Innovation—Ramón y Cajal program, grant number RYC2019-026368-I.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** J.A.C. has a postdoctoral contract with the research program "Ramón y Cajal" (RYC2019-026368-I) and A.R.d.l.B. is recipient of a Pre-doctoral contract (PRE2018-086293), both from Spanish Ministry of Science and Innovation.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**

