Cloning and Molecular Characterization of an Alpha-Glucosidase (MalH) from the Halophilic Archaeon Haloquadratum walsbyi

We report the heterologous expression and molecular characterization of the first extremely halophilic alpha-glucosidase (EC 3.2.1.20) from the archaeon Haloquadratum walsbyi. A 2349 bp region (Hqrw_2071) from the Hqr. walsbyi C23 annotated genome was PCR-amplified and the resulting amplicon ligated into plasmid pET28b(+), expressed in E. coli Rosetta cells, and the resulting protein purified by Ni-NTA affinity chromatography. The recombinant protein showed an estimated molecular mass of 87 kDa, consistent with the expected value of the annotated protein, and an optimal activity for the hydrolysis of α-PNPG was detected at 40 °C, and at pH 6.0. Enzyme activity values were the highest in the presence of 3 M NaCl or 3–4 M KCl. However, specific activity values were two-fold higher in the presence of 3–4 M KCl when compared to NaCl suggesting a cytoplasmic localization. Phylogenetic analyses, with respect to other alpha-glucosidases from members of the class Halobacteria, showed that the Hqr. walsbyi MalH was most similar (up to 41%) to alpha-glucosidases and alpha-xylosidases of Halorubrum. Moreover, computational analyses for the detection of functional domains, active and catalytic sites, as well as 3D structural predictions revealed a close relationship with an E. coli YicI-like alpha-xylosidase of the GH31 family. However, the purified enzyme did not show alpha-xylosidase activity. This narrower substrate range indicates a discrepancy with annotations from different databases and the possibility of specific substrate adaptations of halophilic glucosidases due to high salinity. To our knowledge, this is the first report on the characterization of an alpha-glucosidase from the halophilic Archaea, which could serve as a new model to gain insights into carbon metabolism in this understudied microbial group.


Introduction
Haloquadratum walsbyi is a squared-shaped, extremely halophilic member of the Euryarchaeota, which has been described as the dominant representative of the microbiota present in aquatic hypersaline (≥32% NaCl) environments [1][2][3]. The organism was first described in 1980 by Walsby [4]. However, subsequent studies employing conventional and molecular techniques have reported The ligation was performed using T4 DNA Ligase (Promega Inc., Fitchburg, WI, USA). The resulting recombinant plasmid was called pET-malH.

Protein Expression and Purification
The recombinant plasmid (pET-malH) was transformed into E. coli Rosetta™ cells, and grown at 37 • C in 4 L of Luria Bertani broth (LB) containing 34 µg/mL chloramphenicol and 30 µg/mL kanamycin. When cultures reached late log phase (OD 600 of 0.6-0.8), they were induced with 1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) for 3 h. Cells were harvested by centrifugation (4000 rpm × 20 min, at 4 • C), resuspended in sodium phosphate buffer (NaH 2 PO 4 , pH 8.0; 3 M NaCl, 10 mM imidazole), and lysed by sonication on ice (100 W, 1 s of sonication vs. 2 s pause, 500 cycles). The cell lysate was then centrifuged (13,000 rpm, 15 min, 4 • C). The resulting supernatant was loaded into a chromatography column packed with Ni-NTA agarose (Qiagen, Venlo, Germany), and washed with sodium phosphate at increasing imidazole concentrations of up to 80 mM. The elution was performed using sodium phosphate containing 250 mM of imidazole. Eluted fractions were tested for alpha-glucosidase activity as described in Section 2.4. The alpha-glucosidase containing fractions were resolved by Polyacrylamide Gel Electrophoresis (SDS-PAGE) using 10% polyacrylamide gels, stained with Bio-Safe™ Coomasie Stain (BioRad Inc., Hercules, CA, USA). Protein concentration was determined using the Pierce BCA Protein Assay (ThermoScientific Inc., Bridgewater, NJ, USA), using bovine serum albumin (BSA) as a standard.

In Silico Functional Characterization of the Hqr. walsbyi Alpha-Glucosidase
The predicted amino acid sequence of the putative glycosyl hydrolase from Hqrw_2071 was screened for the presence of functional domains using the NCBI's CDD/SPARCLE Tools [31]. In order to detect functional features with respect to protein structure, the COBALT program [32] was used to generate an anchored multiple sequence alignment (MSA) based on functional constraints derived from 3D structure information contained in NCBI-curated domains. The alignment was generated by importing COBALT's output into the Graphic View interphase of the BioEdit software package (http://www.mbio.ncsu.edu/BioEdit/page2.html) [27] and manually edited based on CDD annotations in a rich text file using a word processor and included sequences from archaeal alpha-glucosidases previously purified and characterized at the molecular level (Table 1). Additional searches for the presence of functional domains were conducted with the CDD tool using the Pfam and InterPro databases [25,26]. Moreover, structural models of the Hqr. walsbyi alpha-glucosidase were predicted using the Phyre2 www.sbg.bio.ic.ac.uk/~phyre/html/page.cgi?id=index) and (PS)2-V2 (ps2.life.nctu.edu) modelling servers. The resulting structures were visualized using the iCn3D web-based 3D structure viewer (https://www.ncbi.nlm.nih.gov/Structure/icn3d/full.html).

Identification of a Putative Alpha-Glucosidase Gene in the Hqr. walsbyi C23 Genome
Previous studies from our laboratory have demonstrated that Hgm. borinquense has the ability to utilize maltose as a sole carbon source. Phylogenetically, Hqr. walsbyi is closely related to Hgm. borinquense [1], and we hypothesized that these organisms might have a similar physiology for carbohydrate metabolism. Therefore, the genomes of Hqr. walsbyi HBSQ001 (DSM 16790) and Hqr. walsbyi C23 (DSM 16854) were searched for the presence of putative alpha-glucosidase gene sequences using the KEGG database (http://www.genome.jp/kegg/pathway.html). This resulted in the detection of a gene (Hqrw_2071) annotated as a putative alpha-glucosidase in the genome of Hqr. walsbyi strain C23 as well as in that of its homolog (HQ1911A) in strain HBSQ001. Through the use of different databases (Pfam, Expasy Proteomics Server, PROSITE, Inter Pro Scan, NCBI Conserved Domains), it was determined that the region comprised by nucleotide positions 1,126,713 to 1,129,061 of the Hqr. walsbyi C23 chromosome encoded an ORF with a predicted amino acid sequence of 782 residues [25,26,33]. The inferred amino acid sequence of Hqrw_2071 was compared to that of other previously described or annotated alpha-glucosidases from members of the Archaea. Table 1 shows detected domains of Hqrw_2071 from Hqr. Walsbyi, which are 35% identical with respect to those present among representatives of the Family 31 of the glycosyl hydrolases and the family of galactose mutarotase-like 2. Galactose mutarotases act as catalyzers in the interconversion of either αand β-anomers of galactose to glucose [34].

Biochemical Characterization of the Recombinant Alpha-Glucosidase from Hqr. walsbyi
After IPTG induction, crude extracts from Escherichia coli cells were analysed for the detection of alpha-glucosidase activity at salinity concentrations ranging from 0 to 5 M. Crude extracts from cells containing the pET-malH plasmid showed alpha-glucoside activity in assays carried out at 40 • C and supplemented with 3 M NaCl. In contrast, extracts from cells transformed with an empty vector or from cells with no vector were unreactive (data not shown).
Purification of the recombinant alpha-glucosidase was performed by loading 1 mL of concentrated crude cell extract into a Ni-NTA agarose column (Qiagen, Venlo, Germany) and eluted with imidazole as described by the manufacturer. The quality and purity of the recombinant protein was verified using SDS-PAGE ( Figure 1).

Biochemical Characterization of the Recombinant Alpha-Glucosidase from Hqr. walsbyi
After IPTG induction, crude extracts from Escherichia coli cells were analysed for the detection of alpha-glucosidase activity at salinity concentrations ranging from 0 to 5 M. Crude extracts from cells containing the pET-malH plasmid showed alpha-glucoside activity in assays carried out at 40 °C and supplemented with 3 M NaCl. In contrast, extracts from cells transformed with an empty vector or from cells with no vector were unreactive (data not shown).
Purification of the recombinant alpha-glucosidase was performed by loading 1 mL of concentrated crude cell extract into a Ni-NTA agarose column (Qiagen, Venlo Germany) and eluted with imidazole as described by the manufacturer. The quality and purity of the recombinant protein was verified using SDS-PAGE ( Figure 1).
As seen in Figure 2, optimal activity conditions for the recombinant enzyme were observed at 40 °C, pH 6.0, and 3 M KCl. These results are in agreement with Hqr. walsbyi growth conditions [1]. Interestingly, recombinant MalH showed a higher activity when KCl was used in the buffer instead of NaCl. This result is consistent with a cytoplasmic enzyme, as halophilic Archaea accumulate high levels of K + in their cytoplasm to compensate for the high concentration of Na + in their environment [39].  As seen in Figure 2, optimal activity conditions for the recombinant enzyme were observed at 40 • C, pH 6.0, and 3 M KCl. These results are in agreement with Hqr. walsbyi growth conditions [1]. Interestingly, recombinant MalH showed a higher activity when KCl was used in the buffer instead of NaCl. This result is consistent with a cytoplasmic enzyme, as halophilic Archaea accumulate high levels of K + in their cytoplasm to compensate for the high concentration of Na + in their environment [39].

In Silico Functional Chracterization and Phylogenetic Analysis of the Hqrw_2071 Gene Product
With regard to global relationships, structurally-constrained sequence alignments revealed that

In Silico Functional Chracterization and Phylogenetic Analysis of the Hqrw_2071 Gene Product
With regard to global relationships, structurally-constrained sequence alignments revealed that MalH was nearly 50% identical to the partial sequence of an halophilic glycosidase detected in the metagenome of an Australian hypersaline lake [40] and aproximately 36% identical to putative alpha-glucosidases from H. kocurii and H. litoreum [41]. Despite the detection of functional traits shared with xylosidases, MalH only shared 26% amino acid identity with respect to its closest database match with a similar domain architecture, an E. coli YicI alpha-xylosidase [42]. However, similar to theYicI alpha-xylosidases (PDB 2F2H), MalH seems capable of forming homo-multimers since residues potentially associated with binding of homotrimers (T353, G352, and E361) and homohexamers (R485, F488 and E497) were detected using the CDD tool. These findings were in agreement with 3D structure prediction analyses generated by the Phyre2 server, which suggested a folding pattern consistent with that of various alpha-glucosidases (100% of residues modeled at >90% confidence) in which most of the hydrophobic residues were oriented towards the core of the predicted structure ( Figure 3A). Likewise, the 3D model produced by the (PS)2-V2 server revealed strong structural similarities (E-value 4.2 × 10 -28 ) with respect to a homo-multimeric YicI alpha-xylosidase from E. coli (PDB: 1WE5 and 2F2H; Figure 3B-D). Furthermore, the (PS)2-V2 server aligned 98% of the amino acid sequence of MalH with the PDB 2F2H-derived template at a 25.89% of amino acid identity. An identity value similar to that obtained from structurally constrained alignments (26%) using the COBALT tool. ( Figure 3A). Likewise, the 3D model produced by the (PS)2-V2 server revealed strong structural similarities (E-value 4.2 × 10 -28 ) with respect to a homo-multimeric YicI alpha-xylosidase from E. coli (PDB: 1WE5 and 2F2H; Figure 3B-D). Furthermore, the (PS)2-V2 server aligned 98% of the amino acid sequence of MalH with the PDB 2F2H-derived template at a 25.89% of amino acid identity. An identity value similar to that obtained from structurally constrained alignments (26%) using the COBALT tool. Phylogenetic tree reconstructions showed that the Hqr. walsbyi MalH is ≤50% identical to other alpha-glucosidases within the Archaea (Table 1, Figures 4 and 5, Figure S1) although similarities with other glycosyl hydrolases, such as alpha-xylosidases, were observed through multiple sequence alignments. However, the recombinant protein from this study was unable to hydrolize α-PNPX (data not shown). It is suggested that genes encoding xylosidases and glucosidases are homologs, but Phylogenetic tree reconstructions showed that the Hqr. walsbyi MalH is ≤50% identical to other alpha-glucosidases within the Archaea (Table 1, Figures 4 and 5, Figure S1) although similarities with other glycosyl hydrolases, such as alpha-xylosidases, were observed through multiple sequence alignments. However, the recombinant protein from this study was unable to hydrolize α-PNPX (data not shown). It is suggested that genes encoding xylosidases and glucosidases are homologs, but evolutionary changes could have separated them and, therefore, both types of enzymes show degrees of similarity when compared in phylogenetic trees [7,43].
In silico analysis showed two putative transmembrane helices and four putative transmembrane segments. Putative transmembrane helices are also found in eukaryotic alpha-glucosidases. However, since enzymes from eukaryotes do not span the cytoplasmic membrane, the predicted transmembrane segments of the malH product could be involved in enzyme folding rather than membrane attachment [43].   The evolutionary history was predicted using the neighbor-joining method [28]. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (2000 replicates) are shown at the nodes [27]. The phylogenetic tree is drawn to scale, using the same units for branch lengths as those of the evolutionary distances. Evolutionary distances were calculated using the p-distance method and are in the units of the number of amino acid differences per site. The analysis included 32 amino acid sequences. All ambiguous positions were removed for each sequence pair. The final dataset consisted of a total of 2541 positions. Evolutionary analyses were conducted in MEGA 6 [29].
A conserved glutamine residue (Q187; Figure 7) previously described as part of an active site in the tridimensional structures of the N-terminal domain of GH31 alpha-glucosidases (cd14752) from the bacterium Ruminococcus obeum PDB: 3PHA [45] and common beet, Beta vulgaris PDB: 3W37 [46] was detected. However, Q187 emerged as all the other functionally confirmed alpha-glucosidases compared in this alignment had aspartic acid (D213) at this position ( Figure 6). These included enzymes for bacteria and thermophilic archaea including in the conceptual translation of an alpha-xylosidase from the reference genome of Streptomyces coelicolor (GenBank accession No. NP_733521) and in that of a YicI-like alpha-xylosidase from E. coli (PDB: 2F2H_A) [42]. Moreover, the Q187 substitution was an exclusive trait of MalH as well as of sequences provisionally identified as alpha-glucosidases in Halorubrum litoreum and Halorubrum kocurii [33,41], suggesting a signature feature of the N-terminal domain of alpha glucosidases from halophilic archaea ( Figure 7A). All residues (D307, W445, K417, D419, F420, R467, W480, Y516 and H552; Figure 6, MalH numbering system; red-colored positions) comprising the active site and the two catalytic residues (D419 and D483) responsible for the hydrolysis reaction in the YicI-like alpha-xylosidase from E. coli (PDB: 2F2H_A) [42] were also identified in MalH ( Figure 6). Further analyses using the Pfam database [25] showed the presence of domains with functions and coordinates consistent with those detected using the CDD tool. These consisted of a galactose mutarotase-like domain at the N-terminal ( [43]; E-value 4.9 × 10 −12 ) followed by a GH31 domain (E-value 1.4 × 10 −103 ), which were situated at intervals 162-222 and 244-670, respectively.  Four of the 11 active sites described in the tridimensional structure of the Sulfolobus solfataricus alpha-glucosidase MalA (PDB: 2G3N) were identified in MalH ( Figure 6, light blue highlights). Moreover, MalH and S. solfataricus MalA also shared the two catalytic D residues characteristic of GH31 affiliates, which were also present in a thermostable alpha-glucosidase from Sulfolobus tokodaii ( Figure 6, Green highlights) [37]. Nevertheless, MalH displayed a variety of unique substitutions at functional or conserved sites (Figures 6 and 7, pink highlight) that appeared as distinguishing features from enzymes originating from bacterial and thermophilic archaea. Furthermore, in some instances, the same substitutions were shared among putative glycosidases from halophilic hosts, further strengthening the presence of signature traits for this group of proteins (Figure 7, pink residues).
In general, the domain-constrained comparative alignment revealed distinctive substitutions that could account for the inability of the MalH to produce xylose from αPNPX and its high activity on αPNPG. The functionality of representative from the GH31 family is known to be diverse, as this group is comprised of enzymes having functions that include alpha-galactosidase, alpha-glucosidase, alpha-xylosidase, glucoamilase, sucrase-isomaltase, and α-glucan lyase activities [47,48]. Moreover, several GH31 are known to exhibit both glycosidadse and xylosidase activities with different levels of affinity between these substrates [7,49,50].
The importance of halophilic enzymes has been reviewed elsewhere, and is not limited only to food processing, bioremediation, and biosynthesis [8,43,47,48,[50][51][52]. Specifically, glucosidases are studied for their potential in multiple industry processes due to their thermostability [48]-for example, they can be used for the production of biofuels and pharmaceutical products, for enhancing the wine aroma, and for reducing the toxic compounds present in animal feed [48]. The stability of glucosidases under high salt conditions remains poorly understood, and this study provides an example of a novel alpha-glucosidase (MalH) with unique characteristics from the halophilic archaea that could be used to address this gap. MalH might provide insights about glucosidase activity under high salinity conditions. To our knowledge, this is the first report of the cloning and molecular characterization of a novel alpha glucosidase with high salinity requirements, which can help study carbon utilization of Haloquadratum in hypersaline environments.