Background

β-N-acetylglucosaminidases (GlcNAcases, EC 3.2.1.52) are glycoside hydrolases (GHs) that cut the O-glycosidic bonds formed by N-acetylglucosamine (GlcNAc) residues from the non-reducing terminals of oligosaccharides, such as chitooligosaccharides and muropeptides [1, 2]. Chitooligosaccharides can be produced from the degradation of chitin by hydrochloric acid [3] or chitinases (EC 3.2.1.14) [2]. Chitin is second only to lignocellulose in natural abundance, but more than 80,000 tons of chitin from marine sources per year is unutilized chitinous waste [2]. GlcNAcases hydrolyse chitooligosaccharides to produce GlcNAc, which can be further used for the production of bioethanol [4] and single cell protein [5], the treatment of ulcerative colitis and other gastrointestinal inflammation disorders [6] and pharmaceutical therapy for osteoarthritis [7] and tumours [8]. Muropeptides can be derived from the peptidoglycan of the cell wall [1] and flagella [9]. The hydrolysis of muropeptides by GlcNAcases indicates that the enzymes function in cell wall recycling [1] and flagellum assembly of bacteria [9]. Furthermore, GlcNAcases also exhibit many additional important biological functions and wide range of industrial applications, such as catabolism of ganglioside storage in Tay-Sachs disease [10], induction of chitinolytic enzymes [11] and synthesis of biologically important oligosaccharides [12, 13].

According to a homology comparison of amino acid sequences, GlcNAcases have been classified into GH 3, 20, 73 and 84 (http://www.cazy.org/). With the rapid development of genome sequencing technology, many GlcNAcases have been identified in the genomes of animal tissues, insects, plants, bacteria and fungi in recent years. GH 20 GlcNAcases are mainly found in microorganisms, especially bacteria (http://www.cazy.org/GH20.html). However, the molecular and biochemical characteristics of the majority of these GH 20 GlcNAcases have not been reported (http://www.cazy.org/GH20.html).

Yunnan Province is known as “the Kingdom of Nonferrous Metals”, “the Kingdom of Plants” and “the Kingdom of Animals” in China. In these natural habitats containing metal ions in high concentrations or special products from plants or animals, novel microorganisms and their enzymes may have to evolve, indicating biotechnological potential. Heijing is called the “town of salt” in Yunnan and has produced salts since ancient China. We previously sampled saline soil from an old mine in the town and reported novel enzymes from bacteria isolated from the soil [14–17]. One example is the GH 32 invertase InvHJ14, showing novel low-temperature and alkaline activity and sucrose tolerance [15]; another is the GH 27 α-galactosidase AgaAHJ8, showing novel salt − protease tolerance and transglycosylation activity, from a potential novel species of Pontibacter [16].

In this study, a novel GH 20 GlcNAcase, called HJ5Nag, was mined from the genome of Microbacterium sp. HJ5—a new isolate from the saline soil. The purified recombinant HJ5Nag (rHJ5Nag), which was expressed in E. coli, showed better tolerance to salts and GlcNAc and higher activity than most GlcNAcases reported in the literature, including the GH 3 GlcNAcase we previously reported [18]. Potential molecular adaptations for salt tolerance and the high activity of the enzyme were presumed.

Methods

Strains, vectors and reagents

Strain HJ5 was isolated from the saline soil of an old salt mine located in Heijing Town. The details of the strain isolation and identification were described in our previous study [14]. The pure culture was deposited in the Strains Collection of the Yunnan Institute of Microbiology under registration no. YMF 4.00007. E. coli BL21 (DE3) was purchased from TransGen (Beijing, China) and was used for gene expression.

The following vectors and reagents were used: plasmid isolation and genomic DNA kits (Tiangen, Beijing, China); dNTPs and DNA polymerases (TaKaRa, Otsu, Japan); isopropyl-β-d-1-thiogalactopyranoside (IPTG; Amresco, Solon, OH, USA); Ni2+-NTA agarose (Qiagen, Valencia, CA, USA); Qubit protein assay kit (Invitrogen, Carlsbad, CA, USA); pEASY-E2 vector (TransGen); N,N'-diacetyl chitobiose (GlcNAc2), N,N',N'',N'''-tetraacetyl chitotetraose (GlcNAc4), p-nitrophenyl β-N-acetylgalactosaminide (pNPGalNAc), and p-nitrophenyl β-d-glucopyranoside (pNPGlc) (J&K Scientific Ltd., Beijing, China); peptidoglycan from Bacillus subtilis (Ekear, Shanghai, China); p-nitrophenol (pNP), p-nitrophenyl-β-d-xylopyranoside (pNPXyl), p-nitrophenyl α-d-galactopyranoside (pNPGal), p-nitrophenyl-α-l-arabinofuranoside (pNPAra), p-nitrophenyl β-N-acetylglucosaminide (pNPGlcNAc), mutanolysin from Streptomyces globisporus ATCC 21553, chitin and chitosan (Sigma − Aldrich, St. Louis, MO, USA); and silica gel G plates (Haiyang, Qingdao, China). All other chemicals were of analytical grade.

Gene cloning and sequence analysis

The GlcNAcase-encoding gene, designated hj5Nag, was obtained by genome sequencing performed on a Miseq sequencer (Illumina) in our lab. The HJ5 genome sequencing details were described in our previous study [19].

Open reading frames from the draft genome of HJ5 were predicted using the tools described in our previous study [17]. BLASTN and BLASTP (http://www.ncbi.nlm.nih.gov/BLAST/) online tools were used to search homologous sequences and to calculate the identity values. SignalP (http://www.cbs.dtu.dk/services/SignalP/) and InterPro (http://www.ebi.ac.uk/interpro/) online tools were used to predict the signal peptide and protein domain, respectively. Vector NTI 10.3 (InforMax, Gaithersburg, MD, USA) software was used to calculate frequencies of the amino acid residues. Multiple sequences and structures alignments were generated by ESPript [20].

Structure analysis

The homology model of HJ5Nag was predicted using the SwissModel platform (http://swissmodel.expasy.org/). The proportions of secondary structures were analysed by VADAR [21]. The charge distributions on the surfaces of various GlcNAcases were calculated using Discovery Studio 2.5 software (Accelrys, San Diego, CA, USA).

Heterologous expression of HJ5Nag

The coding sequence of HJ5Nag was amplified by PCR using the primer set rhj5NagEF (5'-GGGTGCAGCCCCGCCGC-3') and rhj5NagER (5'-CTCGGTGGCCCAGTCGATCTCGC-3'). The resulting PCR product was ligated to pEASY-E2 vector, which has a single 3′-T overhang at the insertion site according to the manufacturer. E. coli BL21 (DE3) competent cells were transformed with the plasmid for recombinant enzyme expression. A positive transformant harbouring the recombinant plasmid was confirmed by DNA sequencing performed by Tsingke (Beijing, China). A seed culture of the positive transformant was grown overnight at 37 °C and then inoculated in 1:100 dilutions into fresh Luria–Bertani medium with the addition of 100 μg mL−1 ampicillin. Upon reaching an OD600 nm of approximately 0.7, IPTG at a final concentration of 0.25 mM was added to the culture to induce enzyme expression at 20 °C for approximately 20 h.

Purification and identification of recombinant GlcNAcase

Cultures containing positive transformant cells were centrifuged and resuspended in ice-cold buffer A containing 20 mM Tris–HCl, 0.5 M NaCl, and 10% (w/v) glycerol (pH 7.2). The cells were disrupted by ultrasonication on ice with 100 short bursts of 4 s each at a power output of 150 W. After removing cell debris by centrifugation, the supernatant was loaded onto Ni2+-NTA agarose gel columns to bond the recombinant enzyme. The target recombinant enzyme was eluted with a linear imidazole gradient of 20–500 mM in buffer A. The protein concentration was determined using a Qubit protein assay kit using a Qubit 2.0 fluorometer (Invitrogen).

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) was performed to analyse the expression of the recombinant enzyme and the purity of the eluted fractions. The molecular masses of internal peptides from the single band present in the SDS–PAGE gel were analysed via matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI–TOF MS) and were compared with the molecular masses of the internal peptides from rHJ5Nag.

Enzyme assay and substrate specificity

An enzyme assay of purified rHJ5Nag towards various substrates was performed spectrophotometrically using the pNP and 3,5-dinitrosalicylic acid (DNS) methods, as described in our previous study [18]. The pNP method was applied to determine the activity of rHJ5Nag towards substrates pNPXyl, pNPGal, pNPGlc, pNPAra, pNPGlcNAc and pNPGalNAc. The DNS method was used for the substrates GlcNAc2, GlcNAc4, colloidal chitin, chitosan, peptidoglycan and muropeptides (degradation products of peptidoglycan by mutanolysin treatment).

Biochemical characterization

The pNP method using pNPGlcNAc as substrate was applied to determine the biochemical characteristics of purified rHJ5Nag, unless otherwise noted. McIlvaine buffer was used for pH 5.0–8.0, and 0.1 M glycine–NaOH was used for pH 9.0–12.0.

The pH-dependent activity of rHJ5Nag was determined at 37 °C, with pH values ranging from 5.0 to 9.0. The pH-dependent stability of rHJ5Nag was determined by measuring the residual enzyme activity after incubating the enzyme at pH 5.0–12.0 for 1 h.

The temperature-dependent activity of rHJ5Nag was examined at pH 6.0 and temperatures from 0 to 60 °C. The thermostability of rHJ5Nag was determined by measuring the residual activity of the enzyme after pre-incubation at 30, 45 and 50 °C and pH 6.0 for 1–60 min.

To determine the effects of various metal ions and chemical reagents on purified rHJ5Nag, the enzyme activities were measured after the addition of different metal ions and chemical reagents to the reaction mixture. The metal ions and chemical reagents included 1.0 or 10.0 mM (final concentration) KCl, AgNO3, CaCl2, CoCl2, HgCl2, Pb(CH3COO)2, MgSO4, NiSO4, CuSO4, MnSO4, ZnSO4, FeCl3, EDTA, β-mercaptoethanol and SDS, and 0.5 or 1.0% (v/v) Tween 80 and Triton X-100. The absorption was measured after removing the precipitate by centrifugation when AgNO3, HgCl2 or SDS was added to the reaction mixture. Furthermore, the enzyme activity in 1.0–30.0% (w/v) NaCl was also examined, and the enzyme stability in NaCl was determined by measuring the residual enzyme activity after pre-incubating the enzyme in 3.0–30.0% (w/v) NaCl for 1 h.

The K m, V max and k cat values of purified rHJ5Nag were analysed using 0.1–1.2 mM pNPGlcNAc as substrate at 45 °C and pH 6.0. A nonlinear regression Michaelis − Menten fit was applied to analyse the data using GraphPad Prism (GraphPad Software, San Diego, CA, USA). The extent of GlcNAc inhibition on rHJ5Nag activity was determined adding 5.0 and 10.0 mM GlcNAc to the reaction solution. Then, the β-N-acetylglucosaminidase activity was determined according to the aforementioned kinetic assay. The inhibition constant (K i) was calculated using GraphPad Prism using the nonlinear mixed-model inhibition equation described detailedly in a previous paper [22].

Hydrolytic property of purified rHJ5Nag

The hydrolysis mixture included 160 μL of 0.5% (w/v) GlcNAc2 or GlcNAc4 and approximately 0.1 units of purified rHJ5Nag. After incubating the mixture in McIlvaine buffer (pH 6.0) at 37 °C for 6 h, the hydrolysis products were analysed by thin-layer chromatography (TLC), as previously described [17].

Synergistic action

Commercial chitinase from Streptomyces griseus, designated CtnSg, was purchased from YuanYe Bio-Technology (Shanghai, China). The reaction mixture contained colloidal chitin (0.5%, w/v) and either CtnSg (0.01 U mL−1 reaction system), rHJ5Nag (0.5 U mL−1 reaction system) or both enzymes and was incubated for 120 min. In the time-course study, the reaction mixture containing 0.5% (w/v) colloidal chitin and CtnSg alone was incubated for 30, 60 or 90 min, rHJ5Nag was then added to the reaction and incubated for 90, 60 or 30 min, respectively. According to the manufacturer’s instructions for CtnSg and the biochemical characteristics of rHJ5Nag, degradation of colloidal chitin was performed at 25 °C and pH 6.0 (McIlvaine buffer). The DNS method was used to measure the amount of reducing sugars released from the colloidal chitin. A reaction containing heat-inactivated enzyme (incubating in a boiling water bath for 10 min) was used as a control. The degree of synergy was calculated using the following equations:

$$ D S=\frac{R{S}_{rHJ5 Nag CtnSg}}{R{S}_{rHJ5 Nag}+ R{S}_{CtnSg}} $$

where DS is the degree of synergy, RS rHJ5NagCtnSg is the amount of reducing sugars released from simultaneous or sequential enzyme reactions, and RS rHJ5Nag and RS CtnSg are the amounts of reducing sugars released from individual enzyme reactions.

Accession numbers

The accession numbers of hj5Nag and Microbacterium sp. HJ5 16S rDNA in GenBank are KX400857 and KX400858, respectively.

Results

Strain identification

Based on the results of a BLASTN search, the nucleotide identity was 98.8% between the partial 16S rDNA sequence from HJ5 (1375 bp) and the 16S rDNA sequences from Microbacterium flavescens (accession no. HQ530520), Microbacterium takaoensis (AB201047) and Microbacterium trichothecenolyticum (KU179351). Phylogenetic analysis also placed HJ5 in the Microbacterium cluster but not in the clusters grouped by the species of other genera of Microbacteriaceae (Additional file 1: Figure S1). Therefore, HJ5 belonged to the genus Microbacterium.

Gene cloning and sequence analysis

Sequence data of approximately 606.6 Mbp was generated by genome sequencing for HJ5, and these data yielded a draft genomic DNA sequence of approximately 3.6 Mbp after sequence assembly. The analysis of the draft genome predicted an open reading frame designated gene hj5Nag. hj5Nag has a length of 1608 bp and encodes the 535-residue GlcNAcase HJ5Nag, with a predicted molecular mass of 55.9 kDa.

The alanine (A) and glycine (G) frequencies of the GlcNAcase were 16.3 and 10.8%, respectively, both of which are the highest values among the GlcNAcases shown in Table 1. The frequency of acidic aspartic acid (D) and glutamic acid (E) of the GlcNAcase was 13.5%, which was the third highest (Table 1). However, frequencies of cysteine (C) and lysine (K) in the GlcNAcase were only 0.2 and 1.1%, respectively, both of which were the second lowest (Table 1). Therefore, the ratio of the total frequency of D, E, A and G to the total frequency of C and K of HJ5Nag (31.2) was the highest among the GlcNAcases shown in Table 1.

Table 1 Amino acid residue frequencies of the experimentally characterized GH 20 GlcNAcases
Full size table

The predicted signal peptide of HJ5Nag is cleaved between A25 and A26 to yield a mature polypeptide consisting of the typical N-terminal domain of bacterial GlcNAcases (from V14 to R153; domain signature: IPR015882) and the catalytic domain of GH 20 GlcNAcases (from F154 to V510; IPR015883; Fig. 1, Additional file 1: Figure S2). A BLASTP search for HJ5Nag showed that the enzyme was most similar to multiple putative GH 20 GlcNAcases annotated based on genome analysis. Among these putative GH 20 members, the enzyme from Microbacterium sp. Root180 (accession no. WP_056119055) shared the highest identity of 78.9% with HJ5Nag. Based on the alignment of HJ5Nag with other experimentally characterized GlcNAcases (Fig. 1) [23–25], residue D334 probably functioned as the catalytic nucleophile/base, and E335 probably functioned as the catalytic proton donor/acceptor. In addition, residues D162, R165, D194, D195, W365, W430 and W465 might play an important role in binding the GlcNAcase ligand (Fig. 1) [23–25].

Fig. 1
figure 1

Partial amino acid sequences and structures alignments of HJ5Nag with GH 20 GlcNAcases. Sequence names are shown with PDB IDs (except HJ5Nag) as follows: 1QBA (accession no. AAB03808), the GlcNAcase from S. marcescens [23]; 2GJX (accession no. AAD13932), the GlcNAcase from H. sapiens [24]; and 3GH4 (accession no. BAI63641), the GlcNAcase from Paenibacillus sp. [25]. Asterisks and number signs show the putative catalytic residues and ligand-binding residues, respectively. Dotted purple line shows the N-terminal domain of bacterial GlcNAcases (domain signature: IPR015882). Solid purple line shows the catalytic domain of GH 20 GlcNAcases (domain signature: IPR015883)

Full size image

Structure analysis

The model of HJ5Nag showed a GMQE score of 0.72 and a sequence identity of 48.9% when the template was the crystal structure of the GH 20 GlcNAcase from Streptomyces plicatus (PDB ID: 1 M01). The homology model of HJ5Nag and the crystal structures of 7 GlcNAcases were used to calculate the proportions of secondary structures and the charge distributions on the surfaces. Among these GlcNAcases, the proportions of amino acids used to build the α-helix and β-sheet structures of HJ5Nag were the second lowest, while the proportion of amino acids used to build random coil structures and the ratio of random coils to α-helices of HJ5Nag were both the highest (Table 2). As shown in Fig. 2, almost the whole surface of HJ5Nag was negatively charged. Compared with other GlcNAcases, HJ5Nag had a much larger negatively charged surface (Fig. 2).

Table 2 Proportions of secondary structures of GH 20 GlcNAcases
Full size table
Fig. 2
figure 2

Charge distributions on the surfaces of the GH 20 GlcNAcases. The A-chain of GlcNAcase is used. PDB IDs are presented as the names of GlcNAcases (except HJ5Nag) as follows: 1 M01, the GlcNAcase from S. plicatus [74]; 1QBA, the GlcNAcase from S. marcescens [23]; 2GJX, the GlcNAcase from H. sapiens [24]; 3GH4, the GlcNAcase from Paenibacillus sp. [25]; 3NSM, the GlcNAcase from O. furnacalis [72]; 3RCN, the GlcNAcase from A. aurescens; 4PYS, the GlcNAcase from B. fragilis. The charge distribution on the surface was calculated at pH 7.0. Positive charges are depicted in blue and negative charges in red

Full size image

Expression, purification and identification of rHJ5Nag

As shown by an SDS–PAGE gel (Additional file 1: Figure S3), the recombinant GlcNAcase was successfully expressed in E. coli BL21 (DE3) and purified by Ni2+-NTA affinity chromatography. The MALDI–TOF MS spectrum of the purified band further confirmed that the purified product was rHJ5Nag (Additional file 1: Figure S4).

Biochemical characterization

At pH 6.0 and 45 °C, purified rHJ5Nag showed specific activities of 1773.1 ± 1.1 (Table 3), 146.0 ± 3.1, 481.4 ± 2.3 and 445.0 ± 1.7 μmol min−1 mg−1 towards substrates of pNPGlcNAc, pNPGalNAc, GlcNAc2 and GlcNAc4, respectively. However, no rHJ5Nag activity was detected towards substrates of pNPXyl, pNPAra, pNPGlc, pNPGal, peptidoglycan, muropeptides, colloidal chitin or chitosan. Thus, rHJ5Nag is a GlcNAcase that can hydrolyse chitooligosaccharides but cannot participate in cell wall turnover or recycling of bacteria.

Table 3 Activities of the experimentally characterized GlcNAcases
Full size table

Purified rHJ5Nag had activity in the range of pH 5.5–9.0 and 10 − 50 °C, with apparent optima at pH 6.0 and 45 °C (Fig. 3a, c). The enzyme exhibited good stability at 30 °C and maintained more than 55% of its initial activity after incubation in buffers ranging from pH 6.0 to 10.0 for 1 h (Fig. 3b, d).

Fig. 3
figure 3

Enzymatic properties of purified rHJ5Nag. a pH-dependent activity profiles. b pH-dependent stability profiles. c Temperature-dependent activity profiles. d Thermostability assay. e Effect of NaCl on rHJ5Nag activity. f Stability of rHJ5Nag in NaCl. Error bars represent the means ± SD (n = 3)

Full size image

Purified rHJ5Nag exhibited good salt tolerance. Although the activity of purified rHJ5Nag was completely inhibited by AgNO3, HgCl2 and SDS, the addition of most metal ions and chemical reagents at final concentrations of 1.0 and 10.0 mM or 0.5 and 1.0% (v/v) showed little to no effect (retaining 84.5 − 131.5% activity) on the enzyme activity (Table 4). Furthermore, the enzyme retained more than 80% activity in the presence of 1.0–10.0% (w/v) NaCl and even 53.6% activity in 20.0% (w/v) NaCl (Fig. 3e). The enzyme retained approximately 100% of its initial activity after incubation in 3.0–20.0% (w/v) NaCl for 60 min (Fig. 3f).

Table 4 Effects of metal ions and chemical reagents on the activity of purified rHJ5Nag
Full size table

Without the addition of GlcNAc, the K m, V max, k cat and k cat/K m of purified rHJ5Nag towards pNPGlcNAc were 0.52 ± 0.05 mM, 3097 ± 124 μmol min−1 mg−1, 2886 ± 116 s−1 and 5550 ± 223 mM−1 s−1, respectively (Table 3; Fig. 4). When 5.0 and 10.0 mM GlcNAc was added to the reaction solution, the V max of purified rHJ5Nag towards pNPGlcNAc decreased to 2732 ± 91 and 2606 ± 87 μmol min−1 mg−1, respectively, the k cat decreased to 2546 ± 84 and 2429 ± 81 s−1, respectively, and the k cat/K m decreased to 4106 ± 135 and 3327 ± 111 mM−1 s−1, respectively (Fig. 4). However, the K m of purified rHJ5Nag towards pNPGlcNAc increased to 0.62 ± 0.04 and 0.73 ± 0.05 mM in the presence of 5.0 and 10.0 mM GlcNAc, respectively (Fig. 4). Furthermore, the value of the parameter Alpha in the mixed-model inhibition equation was 3.412 based on the analysis of kinetic data using GraphPad Prism. These results indicate that GlcNAc can inhibit the activity of rHJ5Nag via a mixture of competitive and noncompetitive inhibition. As a result, the enzyme was inhibited by GlcNAc, with a K i of 14.59 mM.

Fig. 4
figure 4

Kinetic characterization of purified rHJ5Nag. Error bars represent the means ± SD (n = 3)

Full size image

Hydrolytic properties of purified rHJ5Nag

As shown in Fig. 5, GlcNAc was the end product of GlcNAc2 and GlcNAc4 hydrolysed by purified rHJ5Nag.

Fig. 5
figure 5

TLC analysis of the hydrolytic product. Lanes: M, 0.5% (w/v) GlcNAc; CK2 and CK4, inactivated rHJ5Nag (incubating in a boiling water bath for 10 min) with 0.5% (w/v) GlcNAc2 and GlcNAc4, respectively; S2 and S4, hydrolysis of 0.5% (w/v) GlcNAc2 and GlcNAc4 by purified rHJ5Nag, respectively

Full size image

Synergistic action

Although purified rHJ5Nag cannot hydrolyse colloidal chitin alone, simultaneous addition of CtnSg and rHJ5Nag improved the degradation of colloidal chitin by 2.02-fold (Table 5). Sequential addition of the two enzymes improved the degradation of colloidal chitin by 1.98-fold, 1.72-fold and 1.29-fold when rHJ5Nag was added and incubated for 90, 60 and 30 min, respectively (Table 5). Furthermore, these results indicate that the synergistic effect benefits from the earlier addition of rHJ5Nag.

Table 5 Synergy between the commercial chitinase CtnSg and rHJ5Nag for the degradation of colloidal chitin
Full size table

Discussion

Microorganisms isolated from extremophilic locations are of great interest and have attracted substantial attention. Microbacterium strains have been found in various extremophilic locations, including Antarctic habitats [26], the Atacama Desert [27], sediment of the Indian Ocean [28], uranium-rich soil [29] and salty soil [30]. To be active in these extremophilic locations, enzymes from local Microbacterium strains may have to evolve. To the best of our knowledge, although some pioneering studies have revealed the important functions of some enzymes from Microbacterium strains in recent years [31–36], few of these enzymes were found to have adapted to extremophilic locations [32]. Furthermore, we found that all the GlcNAcases from Microbacterium strains were annotated based on in silico analysis without biochemical characterization. Therefore, the study may be the first to report an experimentally characterized GlcNAcase from Microbacterium.

Drastic inhibition by Ag+ and Hg2+ appears to be the common property of GlcNAcases, including HJ5Nag, the GH 20 GlcNAcases from Paenibacillus [25] and V. furnissii [37], the GH 3 GlcNAcases from R. miehei [38] and Sphingobacterium sp. [18], the GH 84 GlcNAcase from C. paraputrificum [39] and the GlcNAcase from S. cerradoensis [40]. Ag+ and Hg2+ have irreversible effects on sulfhydryl groups, and Hg2+ can also oxidize indole rings. Therefore, cysteine or tryptophan residues may be involved in the catalytic activity of GH 20 GlcNAcases. Sequence analysis revealed that W365, W430 and W465 of HJ5Nag might affect ligand binding. As shown in Fig. 1, the conserved W491 may also be essential for the activity of HJ5Nag. SDS is an inhibitor of GlcNAcases, in addition to Ag+ and Hg2+. All the GlcNAcases from Microbacterium sp. in this study, as well as R. miehei [38], Sphingobacterium sp. [18] and L. casei [41], are strongly inhibited by SDS. Purified rHJ5Nag showed peak activity at pH 6.0 and 45 °C. These properties are the same or similar to those of many GlcNAcases. For example, the GH 20 GlcNAcase from S. thermoviolaceus presents maximum activity at pH 6.0 and 50 °C [42], the GH 20 GlcNAcase from H. latex presents maximum activity at pH 6.0 and 45–50 °C [43], the GH 3 GlcNAcase from R. miehei presents maximum activity at pH 6.5 and 50 °C [38], and the GH 84 GlcNAcase from Clostridium paraputrificum presents maximum activity at pH 6.5 and 50 °C [39]. Furthermore, the molecular mass of HJ5Nag (55.9 kDa) is similar to that of some GlcNAcases, such as the GH 20 GlcNAcases from a soil-derived uncultured organism (53.4 kDa) [44], C. fimi (54.2 kDa) [45], Paenibacillus sp. (57.5 kDa) [25] and S. thermoviolaceus (58.3 kDa) [46], and the GH 3 GlcNAcase from S. thermophilum (55.9 kDa) [47].

Purified rHJ5Nag exhibited good salt tolerance, as the addition of most salts had little to no effect on the enzyme activity, especially CuSO4 and NaCl. Most GlcNAcases reported to date are strongly inhibited by Cu2+, such as the GH 20 GlcNAcases from Paenibacillus [25], A. caviae [48] and L. casei [41], the GH 3 GlcNAcase from R. miehei [38], the GH 84 GlcNAcase from C. paraputrificum [39] and the GlcNAcase from S. cerradoensis [40]. Furthermore, the GH 20 GlcNAcases from Paenibacillus are strongly inhibited by Ni2+ [25]; the GH20 GlcNAcase from A. caviae is strongly inhibited by Mn2+ [48]; the GH20 GlcNAcase from T. reesei is inhibited by Fe3+ [49]; the GH 84 GlcNAcase from C. paraputrificum is inhibited by Ni2+, Co2+, Mn2+, Zn2+ and Fe3+ [39]; and the GH 3 GlcNAcase from R. miehei is inhibited by Ni2+, Co2+, Zn2+ and EDTA [38]. Salt tolerance has attracted considerable attention because various salts are present or used in industries [14, 16, 18, 50–60], such as various marine materials and foods in China, with NaCl contents of 3.5–20.0% (w/w) [16, 18]. In addition, material processing and fermentation with a high concentration of NaCl is helpful to reduce the overall cost because sterilization is unnecessary [61].

To reveal the potential catalytic adaptations of rHJ5Nag to high concentrations of salt, sequence and structure comparisons to its homologs were performed. (I). Similar to the results from other salt-tolerant enzymes [50, 51, 53–59], statistical analysis of the amino acid residues showed that HJ5Nag presented higher proportions of acidic D and E along with A and G but lower proportions of C and K than its homologs (Table 1). Carboxyl groups from acidic amino acid residues can bind more water molecules and cations than other groups from other amino acid residues, especially K, and thus enable the protein to form a particle stabilized by cooperative hydrated ion networks [57, 58]. Hydrophobicity must be finely balanced to maintain a protein’s flexibility and stability [62]. High NaCl concentrations strengthen hydrophobic interactions, which can affect protein folding [62]. Salt-tolerant proteins are thought to reduce the size of hydrophobic residues to compensate for the increased hydrophobic effect caused by the high salt concentration [58]. The high proportion of A and G, which are small hydrophobic residues [57], is compatible with that function. In addition, a low abundance of C might give a protein more flexibility in high salt environments due to the decrease in rigid disulphide bridges [56]. (II). At the level of the secondary structure, a higher percentage of random coil structures at the expense of α-helices is observed for salt-tolerant proteins [56, 58], including HJ5Nag (Table 2). Greater protein rigidity is usually accompanied by a higher percentage of α-helix structures, while the opposite is observed for a higher percentage of random coils. (III). At the level of the tertiary structure, an excess of negatively charged surfaces is a typical property of salt-tolerant proteins [50, 51, 53, 56–59, 62, 63]. The increase in negative charge on the surface can counteract the lower dielectric constant at high salinity and improve the ability of enzymes to bind water and salt ions, thus improving protein solubility [56–58].

Most characterized GlcNAcases, including GH 3, 20 and 84 members, show specific activities or V max of lower than 1000 μmol min−1 mg−1 towards pNPGlcNAc, as summarized in Table 3. In this study, rHJ5Nag exhibited a specificity activity of 1773.1 ± 1.1 μmol min−1 mg−1 and a V max of 3097 ± 124 μmol min−1 mg−1 towards pNPGlcNAc. Furthermore, rHJ5Nag showed a specificity activity of 481.4 ± 2.3 μmol min−1 mg−1 towards GlcNAc2, which is much higher than the values from other GlcNAcases recorded in the literature, such as the values from Sphingobacterium sp. (0.3 μmol min−1 mg−1) [18], T. spiralis (0.79 nmol min−1 mg−1) [64], R. miehei (1.06 μmol min−1 mg−1) [38], V. alginolyticus (18.1 μmol min−1 mg−1) [65] and S. thermophilum (76.8 μmol min−1 mg−1) [47]. As discussed above, the structure of HJ5Nag may be more flexible because it has higher proportions of A, G and random coil structures but lower proportions of C and α-helix structures than its homologs. Furthermore, HJ5Nag has a higher proportion of D and E, which are not only negatively charged but also flexible residues. High B-factor values, which are believed to provide high structural flexibility, are usually observed for D and E. The high flexibility is presumed to be a factor in the high activity of the GH 20 GlcNAcase rHJ5Nag.

Although the synergistic action of chitinases and GlcNAcases for complete enzymatic degradation of chitin is attractive, there are few reports [18, 66, 67]. This study may be the first to report a synergistic effect between a chitinase and a GH 20 GlcNAcase. Significant synergy was observed between the commercial CtnSg and rHJ5Nag for chitin degradation, indicating that rHJ5Nag has great potential in the bioconversion of chitin waste. The tolerance of rHJ5Nag to GlcNAc may be one contributor to the significant synergy because GlcNAc can inhibit GlcNAcases and result in the accumulation of chitooligosaccharides, which can inhibit chitinases [38]. The GlcNAcases from V. furnissii, Trichoderma harzianum, A. oryzae, S. frugiperda, R. miehei and T. spiralis were competitively or noncompetitively inhibited by GlcNAc, with K i values of 0.21 [68], 0.21 [69], 1.6 [13], 2.0 [70], 9.68 [38] and 15.75 mM [64], respectively. In this study, rHJ5Nag showed relatively high GlcNAc tolerance, with a K i of 14.59 mM.

Conclusions

A GH 20 GlcNAcase was isolated from Microbacterium sp. HJ5 harboured in the saline soil of an abandoned salt mine. Biochemical characterization revealed that the GlcNAcase had novel salt–GlcNAc tolerance and high activity. These characteristics suggest that the enzyme has versatile potential in biotechnological applications, such as bioconversion of chitin waste and processing of marine materials and saline foods. Molecular characterization provides an understanding of the molecular–function relationships for the salt tolerance and the high activity of the GH 20 GlcNAcase.