Page 54 - Read Online
P. 54

Page 6 of 15                 Sasaki et al. Microbiome Res Rep 2023;2:12  https://dx.doi.org/10.20517/mrr.2023.08

               (Accession no. LC745705). This gene cluster was similar to that in B. longum JCM7052 and comprised
               conserved neighboring putative GH36 enzyme (MCC10289_0426), Lac-I-type transcriptional regulator
               (MCC10289_0430), extracellular solute-binding protein (MCC10289_0427), and ABC transporter
               permeases (MCC10289_0428 and MCC10289_0429) [Figure 1A]. In contrast, the homologous gene
               encoding putative ATPase in B. longum JCM 7052 (BLGA_00350) was not found in the gene cluster of B.
               pseudocatenulatum. AAfase (MCC10289_0425) and putative β-L-arabinopyranosidase (MCC10289_0426)
               exhibited 60% and 26% amino acid sequence identities with GAfase (BLGA_00340) and GH36  α-
               galactosidase (BLGA_00330, BlAga3), respectively. AAfase contains a putative SP, GH39 catalytic domain,
               and three galactose-binding domains according to SignalP and InterPro [Figure 1B]. GAfase is a bacterial
               cell wall anchoring protein, while AAfase has no transmembrane region, suggesting that it is a secreted
               enzyme. Although most of the Bifidobacterium GH36s reported so far have been characterized as α-
               galactosidases, BAD_1528 from Bifidobacterium adolescentis ATCC 15703 was the first GH36 member to be
                                                     [22]
               characterized as a β-L-arabinopyranosidase . A phylogenetic tree of MCC10289_0426 was constructed
               using several sequences belonging to the GH36 family [Figure 1C]. The GH36 family is classified into four
                         [23]
               subfamilies , and the enzymes characterized in bifidobacteria mainly belong to subfamilies-I and -II. The
               phylogenetic tree showed that the GH36 subfamily-I can be further divided into a group with  α-
               galactosidase activity, including many characterized α-galactosidases (Subfamily-I-a group) and a group
               with possible β-L-arabinopyranosidase activity, including BAD_1528 (Subfamily-I-b group). Although
               MCC10289_0426 failed to express as a soluble protein in E. coli, our findings indicate that BBCT_0489 in
               the Subfamily-I-b group is a β-L-arabinopyranosidase (unpublished data). Therefore, we predict that
               MCC10289_0426 of the Subfamily-I-b group is a β-L-arabinopyranosidase and not an α-galactosidase;
               moreover, the neighboring GH39 GAfase homolog was predicted to release β-L-Arap-(1→3)-L-Ara rather
               than α-D-Gal-(1→3)-L-Ara.


               Recombinant AAfase was designed without the N-terminal SP (aa 1-28) and with C-terminal His-tag and
               N-terminal SKIK-tag. E. coli BL21 (DE3) plysS cells transformed with pET23a–MCC10289_0425 were
               cultured at 37 °C for 3 h, induced with IPTG, and cultured at 15 °C for 44 h. The protein was purified using
               the C-terminal His-tag. Purified recombinant AAfase migrated as a single band on SDS-PAGE, with an
               apparent molecular mass of 104 kDa, corresponding to its calculated molecular mass of 104,092 Da
               [Supplementary Figure 1].


               Substrate specificity and general properties of AAfase
               We determined the ability of gum arabic, larch, wheat AGP, and sugar beet arabinan to serve as potential
               substrates for AAfase [Figure 2]. TLC and HPAEC-PAD analyses revealed that AAfase released β-L-Arap-(1
               →3)-L-Ara and α-D-Galp-(1→3)-L-Ara from gum arabic AGP and β-L-Arap-(1→3)-L-Ara from larch AGP.
               AAfase cleaved a small amount of β-L-Arap-(1→3)-L-Ara from sugar beet arabinan, but wheat AGP was not
               utilized as a substrate. The reaction products were identical to those cleaved by GAfase, the chemical
               structures of which had already been determined via enzymatic and NMR analyses in our previous study .
                                                                                                        [8]
               However, HPAEC-PAD analysis showed that the amount of α-D-Galp-(1→3)-L-Ara released from gum
               arabic AGP by AAfase was not equivalent to that released by GAfase [Figure 2C].

               Next, experiments using oligosaccharides derived from gum arabic AGP (S3-GA, S3-AA, and S5-GA) as
               substrates were performed [Figure 3]. Under enzyme-limiting conditions, GAfase released α-D-Galp-(1→3)-
               L-Ara from S3-GA and S5-GA, which corroborated with our previous study results; however, AAfase did
               not release this disaccharide. AAfase can act on S3-AA and release β-L-Arap-(1→3)-L-Ara. Moreover,
               AAfase preferentially released β-L-Arap-(1→3)-L-Ara over α-D-Galp-(1→3)-L-Ara and thus had a different
               substrate specificity compared with GAfase. Based on substrate specificity, this enzyme was named 3-O-β-L-
               arabinopyranosyl-α-L-arabinofuranosidase. The K , k , and k /K  values of AAfase for larch AGP were
                                                                     cat
                                                                        m
                                                           m
                                                              cat
   49   50   51   52   53   54   55   56   57   58   59