Marine arabinofuranosidases: analysis of GH51 homolog sequences from a metagenomic dataset of an extreme environment

DIONISI, HEBE M.; CAMPOS, ELEONORA

virtual

Workshop; Third Meeting & First Workshop of the Argentine Network of Enzymatic Technology (TEz Network); 2021

Red Argentina de Tecnología Enzimática

Xylan is a ramified hetero-polysaccharide that is highly abundant in lignocellulosic biomass and of great interest as feedstock for biofuels, biomaterials and value added products, such as prebiotics. In addition to xylanases, de-ramification enzymes are required for the complete deconstruction of xylan. In particular, enzymes with α-L-arabinfuranosidase (ABF) activity (EC 3.2.1.55) are crucial for the hydrolysis of the glycosidic bond between α-1,2 and α-1,3 L-arabinofuranosyl side chains of hemicelluloses. Therefore, the identification and characterization of ABFs to use in enzymatic cocktails is highly needed, in particular enzymes with high activity under the harsh conditions that are often used in many industrial processes. Most enzymes with ABF activity belong to GH51 and GH62 CAZy families, with a few enzymes of the GH51 family having β-1,4-endoglucanase activity. The goal of this work was to evaluate the diversity, potential hosts and structural properties of GH51 homolog sequences identified in a metagenomic dataset of an extreme coastal environment, in order to select sequences for heterologous expression and characterization. The metagenomic dataset used in this study was generated by sequencing a fosmid library constructed from intertidal sediments obtained near a fuel storage facility in Ushuaia Bay, Tierra del Fuego, Argentina. These microorganisms are adapted to extreme conditions, including low temperatures, high UV-B radiation levels, the presence of various environmental pollutants, as well as variable conditions due to effect of tidal cycles, such as periods of drought and rapid changes in temperature, salinity and nutrient availability. Using Hidden Markov Models specific for the GH51 family, 28 sequences were identified out of 6.8 x 105 protein-coding sequences in the metagenomic dataset. GH51 homologs were highly diverse, and only four sequences shared >60% identity at the amino acid level with members of the GH51 family. When compared with the NCBI nr database using blastp with default parameters, 12 sequences shared >60% identity with the closest match, while the rest showed moderate to low identity values or had no hits with sequences of the database. The taxonomic binning of the metagenomic sequences suggested members of the Bacteroidetes (27%), Planctomycetes (27%), Chloroflexi (19%), Proteobacteria (11%) or Kiritimatiellaeota (8%) phyla as their most probable taxonomic origin, while 8% of the sequences could not be assigned at the phylum level. The two sequences assigned to the recently recognized Kiritimatiellaeota phylum were located in the same 33 Kb scaffold. The majority of the sequences of the scaffold could be related to polysaccharide assimilation processes, including six additional sequences annotated as GHs, three as sulfatases, one as a major facilitator superfamily transporter and one as a rhamnose/proton symporter. Up to date, no enzyme of the GH51 family from organisms belonging to the Planctomycetes or the Kiritimatiellaeota phyla have been characterized, although these organisms have shown potential for the degradation of multiple complex polysaccharides. Out of the 25 full-length GH51 homolog sequences, 21 presented overall 3D models consistent with structures of members of the GH51 family, with the characteristic (β⁄α)8 TIM barrel catalytic domain and the jelly roll C-terminal domain, as well as the presence of the conserved Glu catalytic residues. The length of the β2α2 and the β7α7 loops near the active site were highly variable in the 21 models, with differences of up to 13 residues when compared with the GH51 ABF enzyme of Thermobacillus xylanilyticus. These differences could impact the catalytic properties of these putative ABFs, as it has been proposed that the β2α2 loop plays a key role in substrate binding and catalysis, and that the movement of this loop could be hindered in enzymes with a longer β7α7 loop. Other important feature of the β2α2 loop, the conserved Trp residue shown to interact with the bound arabinosyl moiety, was present in nine of the models. On the other hand, two sequences, assigned to Bacteroidetes and Planctomycetes and sharing only 18% identity, had modeled structures similar to the ABF enzyme from Meripilus giganteus, a fungi, with a CBM4-like N-terminal domain. Three additional GH51 sequences had between 687 and 1,183 residues, although the potential function of the additional domains could not be determined.Overall, we observed a high diversity in the structural features of the metagenomic GH51 homologs, as well as a high proportion of understudied taxonomic groups represented in the study. All these characteristics will be considered for the selection of sequences for heterologous expression and characterization.