Characterization of β-glucosidase activity of a Lactiplantibacillus plantarum 6-phospho-β-glucosidase

Abstract

Abstract

β-Glucosidases are useful for hydrolysis of glycosidically-bound volatiles (GBV), thereby facilitating the release of aroma chemicals from the fruit matrices. In this study, 10 putative glycosyl hydrolases belonging to GH1 family from Lactiplantibacillus plantarum NCIM 2903 were cloned and recombinantly expressed. Interestingly, only one (LpBgl5) of the nine soluble proteins, previously characterized as a 6-phospho-β-glucosidase showed β-glucosidase activity which was further characterized. The enzyme had an optimum pH and temperature of 6 and 40°C, respectively, and was categorized as aryl-β-glucosidase due to its ability to hydrolyze different natural as well as synthetic glucosides except cellobiose. The enzyme exhibited functional activity across multiple substrates, with relative activity decreasing sequentially from β-xylosidase to β-glucosidase and finally β-mannosidase. The β-xylosidase and β-glucosidase activities of LpBgl5 were stimulated up to 300% and 700% in the presence of 4 M xylose and 4 M glucose, respectively. The enzyme could also hydrolyze GBV from mango. To our knowledge, this is the first recombinant β-glucosidase/β-xylosidase/β-mannosidase from L. plantarum to have potential for aroma enhancement in fruit products.

Key points

• A recombinant β-glycosidase from Lactiplantibacillus plantarum was characterized.

• The enzyme showed higher β-xylosidase activity than β-glucosidase activity.

• The enzyme could also hydrolyze glycosidically bound volatiles from mango.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00253-025-13472-8.

Introduction

β-Glucosidases are the enzymes that hydrolyze the β-1,4-glucosidic bond to release terminal glucose residues present in glucoconjugates such as aryl-β-glucosides, disaccharides, and oligosaccharides. Based on substrate specificities, these enzymes are classified as aryl-β-glucosidases, cellobiases, and broad-range, of which the last one is the most common class (Singh et al. 2016). Based on their sequences, they can be placed into 11 different glycosyl hydrolase (GH) families of which GH1 and GH3 are most reported. The promising applications of β-glucosidase include hydrolysis of aroma glucosides, lignocellulose degradation for biofuel production, and hydrolysis of glucosides such as soybean glucosides to bioactive aglycons (Godse et al. 2021).

Glycosidically-bound volatiles (GBV) are the non-volatile compounds which on hydrolysis by β-glucosidase, release aroma aglycons and a sugar moiety which is generally β-D-glucose. β-Glucosidases are thus claimed for their potential to enhance aroma of the fruit products (Liang et al. 2020). Commercial enzyme formulations are also available in the market for improving the aroma of non-alcoholic (juices and tea) and alcoholic (wines and beers) beverages (de Morais et al. 2023). Although these products are claimed to effectively hydrolyze GBV, most of these enzymes are not true β-glucosidases; instead, they are primarily pectinases with β-glucosidase as side activities.

Many novel β-glucosidases of microbial origin have been characterized for their properties and possible applications (Godse et al. 2021). Considering that most of these enzymes studied for aroma enhancement are from yeast, β-glucosidases from novel bacterial sources could provide useful candidates for the same. Lactiplantibacillus plantarum is an important species of lactic acid bacteria (LAB) that displays greater metabolic features, such as an ability to utilize a wide array of sugars (Fidanza et al. 2021). In congruence with this observation, various enzymes involved in sugar metabolism, such as α-amylases (Jeon et al. 2016; Plaza-Vinuesa et al. 2019) and β-galactosidases (Benavente et al. 2015; Selvarajan and Mohanasrinivasan 2015) have been characterized from L. plantarum and studied for their industrial applications. Hydrolysis of aroma glucosides by L. plantarum strains during fruit juice fermentation has been attributed to their β-glucosidase activities (Wei et al. 2018; Wu et al. 2021), which also corroborates with the existence of several β-glucosidases genes in L. plantarum genomes (Dymarska et al. 2024). However, there are only two studies on L. plantarum β-glucosidase: one where the culture supernatant from the strain USC1 was directly studied for its β-glucosidase activity (Sestelo et al. 2004) and another from the strain B21 in which the bglGPT operon was cloned in Escherichia coli cells and assessed for its β-glucosidase activity (Marasco et al. 2000).

In this study, 10 putative glycosyl hydrolases from L. plantarum NCIM 2903 annotated to GH1 family were cloned and recombinantly expressed in E. coli. The only recombinant enzyme active as β-glucosidase was categorized as aryl-β-glucosidase with higher activity as β-xylosidase than β-glucosidase. Apart from the common substrates, this enzyme was also able to hydrolyze several natural β-glucosides along with GBV from mango.

Materials and methods

Chemicals

The chromogenic substrates viz., p-nitrophenyl-β-D-glucopyranoside (pNPGlc), p-nitrophenyl-β-D-galactopyranoside (pNPGal), p-nitrophenyl-β-D-xylopyranoside (pNPXyl), p-nitrophenyl-β-D-mannopyranoside (pNPMan), p-nitrophenyl-β-D-glucuronide (pNPGlcr), and p-nitrophenyl-α-D-glucopyranoside (pNPαGlc) and the soybean glucosides and aglycons (glycitin, glycitein, daidzin, daidzein, and genistin) were procured from Sigma (Missouri, USA). Bacterial culture media, kanamycin, IPTG, X-glucoside, arbutin, hydroxyquinone, esculin, esculetin, salicin, salicyl alcohol, cellobiose, and glucose were obtained from HiMedia Laboratories (Maharashtra, India). The enzymes for cloning were procured from New England BioLabs (Massachusetts, USA).

In silico analysis

The complete open reading frame of 10 putative glycosyl hydrolases were retrieved from the whole genome data of L. plantarum NCIM 2903 obtained earlier by predicting the CDS in the contigs using ORF finder tool of NCBI (https://www.ncbi.nlm.nih.gov/orffinder/) and annotating them by Prokka v.1.14.6 (Seemann 2014). Molecular weights and pI of the proteins were predicted using ExPASY website (https://web.expasy.org/compute_pi/) while GH family was predicted using dbCAN3 server (Zheng et al. 2023). The percent identity matrix was generated using ClustalO. ClustalO was used to generate alignment, and MEGAX was used to construct the phylogenetic tree with a few recombinant β-glucosidases characterized from lactic acid bacteria.

Bacterial strains, plasmids, and culture conditions

L. plantarum NCIM 2903 was grown in modified de Man Rogosa Sharpe (mMRS) medium supplemented with 0.05% L-cysteine at 37°C under static conditions. E. coli TOP10 was used for cloning, and E. coli BL21 (DE3) was used as the expression host with Luria Bertini (LB) as the growth media. pET-30b was used as the cloning vector, and kanamycin (50 μg ml) was added to the media for selection of transformants wherever required.

Cloning and heterologous expression of recombinant glycosyl hydrolases

Genomic DNA of L. plantarum NCIM 2903 was used as the template for amplifying the putative glycosyl hydrolase genes using Phusion DNA polymerase and respective primers (Table S1). The amplicons were digested with the respective restriction enzymes, ligated with digested vector, and the ligation reactions were transformed in E. coli TOP10 competent cells. The recombinant clones were identified by colony PCR, confirmed by Sanger sequencing, and further transformed into E. coli BL21 (DE3) for recombinant protein expression.

Pre-culture was prepared by inoculating a colony of E. coli BL21 (DE3) having the recombinant plasmid in LB broth with kanamycin (50 μg ml) and grown at 37°C for 18 h aerobically. This pre-culture was inoculated (2%) in LB broth with kanamycin (50 μg ml) and was further grown aerobically at 37°C until the OD₆₀₀ reached 0.4–0.6 and induced using IPTG (1 mM). The induced culture was incubated at 150 rpm for 20 h at 18°C and cells were collected by centrifugation at 12,000 g for 15 min at 4°C. The cells resuspended in 5 ml of sodium phosphate buffer (50 mM, pH 7) were lysed by ultrasonication on ice. Sonication cycles were set to 40% amplitude, 6 s on/off for 20 min. The lysed cells were centrifuged at 12,000 g for 15 min at 4°C, and the supernatant was analyzed for protein over-expression by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (BioRad). E. coli BL21 (DE3) cells containing pET-30b were processed parallelly as a negative control.

Purification of recombinant glycosyl hydrolases

The recombinant 6xHis-tagged proteins were purified by affinity chromatography using Ni²⁺-nitrilotriacetic acid (Ni–NTA) resin from Qiagen (North Rhine-Westphalia, Germany). The lysate was incubated with the activated Ni–NTA agarose for one hour on ice. This slurry was loaded onto a column, and the flow-through was collected. The matrix was washed with sodium phosphate buffer (50 mM, pH 7) containing increasing concentration of imidazole (20 to 80 mM), and the recombinant protein was eluted using sodium phosphate buffer (50 mM, pH 7) containing 150 mM and 250 mM imidazole. The pooled eluted fraction was passed through the Amicon Ultra-4 centrifugal filter by Millipore (Massachusetts, USA) to remove imidazole and traces of non-specific proteins smaller than 30 kDa and to concentrate the protein. The recombinant protein was quantitated using Pierce BCA assay kit by Thermo Fisher Scientific (Massachusetts, USA) with bovine serum albumin as a standard and stored at 4°C for a maximum of 14 days for further studies. No loss of activity was observed during this duration. Lysate of E. coli BL21 (DE3) cells harboring pET-30b was processed parallelly as the negative control.

Enzyme assay

The in vitro enzyme activity of purified enzymes was determined using pNPGlc as the substrate. Almond β-glucosidase was included as a positive control since it is the most widely used β-glucosidase for evaluating the activity of novel β-glucosidases for a range of applications (Godse et al. 2021). A standard 200 μl reaction consisted of 10 mM pNPGlc/pNPXyl and 25 μg enzyme in 50 mM citrate-phosphate buffer (pH 6) incubated for one hour at 40°C. After incubation, the reaction was stopped by adding 250 mM sodium carbonate and the released p-nitrophenol was measured spectrophotometrically at 405 nm. Identical reactions containing 25 μg almond β-glucosidase were set as positive controls. Negative control in all assays was set up by adding volumetrically equal amounts of similarly purified proteins from E. coli BL21 (DE3) carrying empty pET-30b plasmid. All the experiments were carried out in triplicates.

Determination of optimum pH and temperature of LpBgl5

The effect of pH was assessed by performing the enzyme assay in 50 mM citrate-phosphate buffer (pH 3 to 8) and 50 mM glycine-sodium hydroxide buffer (pH 9 and 10) at 37°C. To determine the effect of temperature, the enzyme assay was carried out in 50 mM citrate-phosphate buffer (pH 6) at temperatures ranging from 20-70°C with an interval of 10°C. The rest of the assay composition and conditions were the same as mentioned above.

Determination of substrate specificities of LpBgl5

The activity of LpBgl5 was assessed with various p-nitrophenyl derivatives viz., pNPGal, pNPXyl, pNPMan, pNPGlcr, and pNPαGlc as described for pNPGlc. Activity of LpBgl5 on natural glucosides (esculin, arbutin, salicin, glycitin, daidzin, genistin, and cellobiose) was assessed by setting up 200 µl reaction with 1 mM substrate and 50 μg enzyme in 50 mM citrate-phosphate buffer (pH 6) at 40°C for different time points (2 h, 4 h, 6 h, 24 h, 48 h and 96 h). Positive control for these assays was a similar reaction containing 50 μg almond β-glucosidase instead of LpBgl5. The reactions were assessed by thin layer chromatography using silica gel F₂₅₄ plates from Merck (New Jersey, USA). In the case of all substrates except cellobiose, the mobile phase was toluene: ethyl acetate: formate: methanol (3: 4: 0.8: 0.7), and the plates were visualized under UV after the chromatography. For cellobiose, butanol: ethanol: water (7: 2: 1) was used as the mobile phase and visualized by immersing in 5% methanolic sulfuric acid followed by heating at 120°C for a few minutes until the analytes were visualized.

Determination of kinetic parameters and specific activity of LpBgl5

The kinetic parameters of LpBgl5 were determined in 100 µl reactions containing 5 µg of enzyme and increasing concentrations (2.5-40 mM) of the individual substrates (pNPGlc or pNPXyl). One unit (U) of enzyme activity was determined as the amount of enzyme required to release 1 µM of p-nitrophenol per minute under the assay conditions.

Effect of chemicals and monosaccharides on LpBgl5

The effects of cations (10 mM Na⁺, K⁺ or Mg²⁺), a chelating agent (10 mM EDTA), a detergent (1% SDS), and monosaccharides (0.5 M glucose, xylose, arabinose, rhamnose, or mannose) on the β-glucosidase and β-xylosidase activities of LpBgl5 were individually assessed in the standard reaction. The effects of reaction end-product, viz., glucose and xylose (0.5-4 M with an interval of 0.5 M) on β-glucosidase and β-xylosidase activity, respectively, and that of ethanol (5-20% with an interval of 5%) on both β-glucosidase and β-xylosidase activity of LpBgl5 were similarly assessed.

Ex situ and in situ hydrolysis of mango GBV by LpBgl5

Alphonso mangoes were procured from Sindhudurg district, Maharashtra, India. Extraction of GBV was carried out as described earlier (Godse et al. 2023). The ex situ hydrolysis reactions contained 2 ml of 1 mg ml^-1 enzyme solution and 2 ml of the GBV extract. For in situ hydrolysis, the juice was prepared as per Godse et al. (2023), and the reactions contained 2 ml of enzyme solution (1 mg ml⁻¹) of LpBgl5 and 3.4 ml of juice obtained from 1 g of pulp. Parallel reactions containing volumetrically equal amounts of similarly purified proteins from E. coli BL21 (DE3) cells carrying empty pET-30b plasmid were considered as the negative control. Rest of the reaction conditions and extraction parameters were the same as described earlier (Godse et al. 2023). Gas chromatographic analysis of the reaction products was carried out as described earlier by Godse et al. (2023) with 2 µl of injection volume.

Nucleotide sequences of the putative glycosyl hydrolases (lpbgl1, PQ876087; lpbgl2, PQ876088; lpbgl3, PQ876089; lpbgl4, PQ876090; lpbgl5, PQ462659; lpbgl6, PQ876091; lpbgl7, PQ876092; lpbgl8, PQ876093; lpbgl9, PQ876094; and lpbgl10, PQ876095) have been submitted in NCBI.

Results

In silico analysis of the putative glycosyl hydrolases

The potential glycosyl hydrolases were identified in the whole-genome sequence of L. plantarum NCIM 2903 based on Prokka annotation. The genes were annotated as either 6-phospho-beta-D-glucosidase (lpbgl1, lpbgl8) or as aryl-phospho-beta-D-glucosidase (lpbgl2, lpbgl3, lpbgl4, lpbgl5, lpbgl6, lpbgl7, lpbgl9, and lpbgl10). The molecular weights and pI of all the predicted enzymes ranged between 52.7-57.8 kDa and 4.87-5.64, respectively (Table S2). Based on dbCAN3 annotation, all the enzymes belonged to the GH1 family which has the most characterized β-glucosidases till date. Further, no signal peptide was found in any of the sequences indicating that they were not secretory proteins (Zheng et al. 2023). Percent identity matrix of amino acid sequences revealed the highest similarity between any two enzymes was not more than 72% (Table S3).

Aligning the 10 putative glycosyl hydrolases with a few of the β-glucosidases characterized from LAB showed the same conserved catalytic regions (TF/LNEP/I and T/VENG) around the catalytic glutamate residues in all the sequences which is a characteristic feature of GH1 family enzymes (Fig. 1a).

Fig. 1.

Comparison of amino acid sequences of the 10 putative L. plantarum glycosyl hydrolases with characterized β-glucosidases from lactic acid bacteria. a Amino acid sequence alignment of the 10 putative L. plantarum glycosyl hydrolases with characterized β-glucosidases from lactic acid bacteria (AJ250202, Lactobacillus plantarum B21; JQ706071, Lactococcus lactis FSJ4; YP_193151.1, Lactobacillus acidophilus NCFM) by ClustalO. Residues that are identical (*), conserved (:), or semiconserved (·) in all sequences are indicated. The conserved catalytic regions are denoted in box and the conserved Glu residues are marked in red. b Phylogenetic tree of the 10 putative L. plantarum glycosyl hydrolases and the characterized β-glucosidases from lactic acid bacteria, constructed using neighbor-joining method in MEGAX

Expression and purification of the putative glycosyl hydrolases

All the genes were cloned and overexpressed in E. coli BL21 (DE3). The recombinant proteins of all the glycosyl hydrolases were observed in the soluble fractions (lysate) except for LpBgl2. Attempts were made to allow for soluble expression of LpBgl2 by using a different host strain and varying media composition. However, the protein always remained completely in the insoluble fraction (data not shown). The remaining soluble recombinant proteins were further purified using affinity chromatography and concentrated (Fig. S1).

Enzyme assay with recombinant glycosyl hydrolases

On assessing the activity with pNPGlc, only LpBgl5 but none of the other nine recombinant proteins showed formation of yellow colour confirming in vitro β-glucosidase activity of LpBgl5.

Optimum pH and temperature of LpBgl5

The enzyme activity of LpBgl5 was assessed in the pH range of 3-10 with pNPGlc as the substrate, and the highest activity was found at pH 6 (Fig. 2a). The drastic effect of pH on the activity was evidenced as only 40% activity was retained at pH 5 and 7. The optimum temperature of pNPGlc was found to be 40°C with retention of about 60% activity at 30°C (Fig. 2b). The enzyme activity reduced drastically (by about 80%) above 50°C. All further enzyme assays were carried out at pH 6 and 40°C.

Fig. 2.

Determination of optimum pH (a) and temperature (b) of LpBgl5 using 10 mM pNP-β-glucopyranoside as the substrate. The buffers used were 50 mM citrate-phosphate buffer (pH 3 to 8) and glycine-sodium hydroxide buffer (pH 9 and 10)

Determination of substrate specificities of LpBgl5

Substrate specificity assays of LpBgl5 revealed its activity with various aryl-glycosides to varying extents apart from that with pNPGlc (Table 1). Surprisingly, LpBgl5 hydrolyzed pNPXyl to a greater extent (384%) than pNPGlc. pNPMan (82%) also acted as a substrate; whereas, pNPGlcr was hydrolyzed to a very small extent (6%). No activity was detected with pNPαGlc and pNPGal.

Table 1.

Relative activities (± standard error of measurement) of LpBgl5 with various pNP-glycosidic (chromogenic) substrates

Substrate	Relative activity (%)
p-Nitrophenyl-β-D-glucopyranoside	100 ± 5
p-Nitrophenyl-α-D-glucopyranoside	0
p-Nitrophenyl-β-D-galactopyranoside	0
p-Nitrophenyl-β-D-mannopyranoside	82 ± 6
p-Nitrophenyl-β-D-xylopyranoside	384 ± 10
p-Nitrophenyl-β-D-glucuronide	6 ± 1

Amongst the natural glucosides, LpBgl5 showed complete hydrolysis of glycitin, salicin, and esculin after 2 h, 8 h, and 96 h, respectively. In the case of genistin and daidzin, partial hydrolysis was observed while no hydrolysis of arbutin and cellobiose was seen during 96 h of incubation (Fig. 3).

Fig. 3.

Assessment of the potential of LpBgl5 to hydrolyze natural β-glucosides at different time points by TLC a glycitin (2 h), b salicin (8 h), c esculin (96 h), d daidzin (96 h), e genistin (96 h), f arbutin (96 h), and g cellobiose (96 h). Lane 1, standard substrate; lane 2, standard product (glycitein, salicyl alcohol, esculetin, daidzein, hydroxyquinone, and glucose); lane 3, reaction with almond β-glucosidase (50 µg); lane 4, reaction with protein from E. coli BL21 (DE3) cells with pET-30b; lane 5, reaction with LpBgl5 (50 µg). There are only four lanes for (e), as the standard product (genistein) was not commercially available

Kinetic parameters and specific activities of LpBgl5

Based on the previous literature, initial assays to determine the kinetic properties of LpBgl5 were performed using 2.5-17.5 mM of pNPGlc and pNPXyl as the substrates and 12.5 µg enzyme in a 100 µl reaction. Since no saturation of the enzyme activity was observed at the highest concentration of the substrates (17.5 mM), the kinetics assay was performed with substrate concentration increased up to 40 mM while reducing the enzyme to 5 µg. Even at the highest substrate concentration, LpBgl5 did not show saturation of the activity (Fig. S2). Specific activities of LpBgl5 were 0.06 U mg⁻¹ and 0.19 U mg⁻¹ with pNPGlc and pNPXyl as the substrates, respectively.

Effect of chemicals and monosaccharides on LpBgl5

Neither any metal ions nor the chelating agent (EDTA) had any significant effect on the β-glucosidase and β-xylosidase activities of LpBgl5 (Fig. 4a). Both β-glucosidase and β-xylosidase activities of LpBgl5 were completely inhibited by SDS, which complied with earlier studies on β-glucosidases (Kaushal et al. 2021; Zhong et al. 2016) and β-xylosidases (Salzano et al. 2022).

Fig. 4.

Effects of various chemicals and monosaccharides on the β-glucosidase and β-xylosidases activities of LpBgl5 determined under optimum reaction conditions using pNP-β-glucopyranoside and pNP-β-xylopyranoside as the substrates, respectively. a Effect of metal ions (Na⁺, K⁺, Mg.²⁺) and EDTA assessed at the final concentrations of 10 mM and that of SDS at 1%. b Effect of monosaccharides assessed at a final concentration of 0.5 M. c Effect of glucose on β-glucosidase activity and xylose on β-xylosidase activity assessed in a range of 0.5–4 M. d Effect of ethanol assessed in the concentration range of 5-20%. Asterisks denote values that are significantly different than the control (*p ≤ 0.01, **p ≤ 0.001; ***p ≤ 0.0001) (one-way ANOVA)

Along with various chemicals, the effect of monosaccharides on LpBgl5’s activity was also assessed. None of the tested monosaccharides (glucose, xylose, arabinose, rhamnose, or mannose) inhibited the activities of LpBgl5. Indeed, glucose significantly stimulated the β-xylosidase activity LpBgl5 by 40% (p ≤ 0.001, one-way ANOVA) (Fig. 4b). Glucose also enhanced β-glucosidase activity by 20%, though the difference was not statistically significant. Further, we assessed the effects of increasing concentration of glucose and xylose on the β-glucosidase and β-xylosidase activities of LpBgl5, respectively. Interestingly, both activities were continuously stimulated by the increasing concentrations of sugars, with the relative β-glucosidase and β-xylosidase activities reaching 700% and 300% in the presence of 4 M glucose and 4 M xylose, respectively (Fig. 4c). It could have been interesting to assess the activity at even higher monosaccharide concentrations; however, such an experiment was not feasible as the stock solutions of the sugars were already near the saturation limits. Ethanol tolerance of β-glucosidases is a desired property considering their application in wine-making for aroma enhancement (de Morais et al. 2023). In case of LpBgl5, 5% ethanol led to up to 60% reduction of both the enzymatic activities (Fig. 4d). On increasing ethanol concentration to 20%, the β-glucosidase and β-xylosidase activity reduced to about 10% and 20%, respectively, indicating that the enzyme is not tolerant to ethanol.

Ex situ and in situ hydrolysis of mango GBV by LpBgl5

We further assessed the ability of LpBgl5 to hydrolyze GBV purified from mango (ex situ hydrolysis) as well as to release similar volatiles directly in the fruit juice (in situ hydrolysis). LpBgl5 was found to release a total of 19 and six volatile compounds during ex situ and in situ treatments, respectively (Table 2). These volatiles belonged to three chemical classes: monoterpenoids, phenolics, and norisoprenoids. Five compounds that were common amongst both treatments (ex situ and in situ) were phenolics except one.

Table 2.

Volatile compounds released upon treating the mango GBV extract (ex situ hydrolysis) and mango juice (in situ hydrolysis) with LpBgl5

Compound	Concentrations (μg kg) (± standard error of measurement)
	Ex situ	In situ
Monoterpenoids
cis-Linalool-3,6-oxidea	1.49 (0.08)	nd
trans-Linalool-3,6-oxidea	1.78 (0.32)	nd
Linaloola	1.55 (0.09)	nd
Lavandulol	1.63 (0.04)	nd
Epoxylinalool	1.55 (0.06)	nd
trans-Linalool-3,7-oxide	2.81 (0.02)	nd
Myrtenola	1.75 (0.12)	nd
2,6-Dimethyl-1,7-octadien-3,6-diol	1.91 (0.28)	nd
Isogeraniol	2.82 (0.07)	nd
cis-Verbenol acetate	1.83 (0.11)	nd
p-Mentha-1(7),8(10)-dien-9-ol	49.34 (1.81)b	19.2 (2.6)
Perillyl alcohola	3.48 (0.17)	nd
cis-Carvyl acetate	5.41 (0.12)	nd
Total concentration of monoterpenoids	77	19
Phenolics
Benzeneacetaldehydea	nd	19.5 (4.8)
2-Phenylethanola	3.55 (0.16)	25.3 (3.1)
p-Vinylguaiacol	2.45 (0.03)	23.9 (8.7)
Eugenola	113.35 (5.29)	424.2 (5.8)
Isoeugenol	50.70 (1.62)b	nd
Methoxyeugenol	145.75 (7.87)b	229.8 (0.7)
Total concentration of phenolics	315	722
Norisoprenoids
3-Oxo-α-ionol	2.08 (0.04)	nd
Total concentration of norisoprenoids	2	0
Total GBV concentration	394	741

nd not detected

^aIdentities of these compounds were confirmed using external standards

^bThese compounds were also detected in empty plasmid controls and the indicated value represents the difference between the concentrations observed in LpBgl5 treated and empty plasmid control samples

Discussion

Based on the in silico analysis, the 10 GH1 genes were annotated as either 6-phospho-beta-D-glucosidase or aryl-phospho-beta-D-glucosidase. However, all were selected for cloning and recombinant expression considering that the amino acid residues of the enzymes involved in determining the specificity towards phosphorylated and non-phosphorylated substrates are not known (Michlmayr and Kneifel 2014). Furthermore, a study by Theilmann et al. (2017) reported two genes from Lactobacillus acidophilus NCFM that were identified as 6-phospho-beta-D-glucosidase could utilize β-glucosides as their substrates. Hence, exploring the β-glucosidase activities of the annotated phospho-β-glucosidases for their industrial potential seemed to be a promising strategy.

Amongst the 10 putative glycosyl hydrolases, LpBgl9 showed 100% identity with an earlier reported β-glucosidase from L. plantarum B21 (Marasco et al. 2000). However, since the earlier study only reported β-glucosidase activitiy of the E. coli cells harbouring the bglGPT operon, lpbgl9 was retained in the panel of genes for further characterization. While our study was progressing, Plaza-Vinuesa et al. (2022) reported characterization of multiple 6-phospho-β-glucosidases from L. plantarum WCFS1 which were highly similar (> 98% identity) with the putative glycosyl hydrolases in our study except LpBgl1 (Table S2). All these enzymes from L. plantarum WCFS1 were active on 6-phospho-β-D-glucopyranoside and no activity on pNPGlc (Plaza-Vinuesa et al. 2022). However, amongst our enzymes, LpBgl5 (similar to Lp_3525) was active on pNPGlc as the substrate. The few amino acid residues that varied between LpBgl5 and Lp_3525 were in the non-conserved regions. On further analysis, the difference in the enzyme activities of LpBgl5 and Lp_3525 could be attributed to differences in the reaction conditions (40°C for 1 h and 30°C for 10 min, respectively). The possible influence of temperature on the substrate specificity is also highlighted by the fact that optimum temperature profiles of Lp_3525 for pNP- 6-phospho-β-D-galactopyranoside and pNP- 6-phospho-β-D-glucopyranoside varied greatly from each other (Plaza-Vinuesa et al. 2023). Thus, it is highly possible that LpBgl5 possesses phospho-β-glucosidase and phospho-β-galactosidase activities, and the observed β-glucosidase activity could be a secondary activity. However, considering that we did not find any β-glucosidases activity in the remaining phospho-β-glucosidases, β-glucosidase activity consistently observed with LpBgl5 is definitely not an experimental artifact.

The optimum parameters of LpBgl5 were within the range of optimum pH (4-8) and temperature (30°C- 90°C) reported for most bacterial β-glucosidases (Singh et al. 2016). Microbial β-glucosidases are known to display side activities including β-galactosidase, β-xylosidase, or β-fucosidase (Matsuzawa and Yaoi 2017; Wu et al. 2018). Similar to LpBgl5, a few β-glucosidases from other bacteria also had β-xylosidase and β-mannosidase activities to a varying extent (Cota et al. 2015) with those from Bacillus subtilis PS (Chamoli et al. 2016) and Lactococcus lactis FSJ4 (Fang et al. 2014) possessing higher β-xylosidase activity than the β-glucosidase activity. This phenomenon can be explained by the possibility that the enzymes in polyspecific families such as GH1 might have acquired new specificities during evolutionary events due to the stereochemical resemblance between the substrates (Henrissat 1991). Hence, experimental characterization of enzymes is necessary to confirm the annotation of their respective genes (Acebrón et al. 2009).

Amongst the common natural glucosides (esculin, salicin, and arbutin) assessed for hydrolysis by β-glucosidases, salicin was preferred over esculin for a β-glucosidase purified from L. plantarum USC1 (Sestelo et al. 2004) while for a recombinant metagenomic β-glucosidase from soil (Biver et al. 2014), salicin was reported as the most preferred substrate amongst the three substrates similar to LpBgl5. While arbutin was one of the most preferred substrates for other bacterial and fungal β-glucosidases (Mase et al. 2004; Xie et al. 2022), it was not hydrolyzed by LpBgl5. The trend of substrate preference of LpBgl5 for soybean glucosides was similar to β-glucosidases from Alicyclobacillus herbarius (Delgado et al. 2021) and Paecilomyces thermophila J18 (Yang et al. 2009) with glycitin as the most preferred substrate. Amongst the few bifidobacterial β-glucosidases assessed for hydrolysis of soybean glucosides, all showed hydrolysis of the soybean glucosides, although to varied extents (Guadamuro et al. 2017; You et al. 2015). Since LpBgl5 can hydrolyze different pNP-derivatives as well as natural glucosides but is unable to hydrolyze cellobiose, it can be classified as an aryl-β-glucosidase (Eyzaguirre et al. 2005; Fang et al. 2014). The inability of LpBgl5 to hydrolyze cellobiose is also in congruence with the glycoside conversion ability of lactobacilli (Dymarska et al. 2024).

LpBgl5 did not show saturation of the activity even at the highest possible substrate concentration. The possible low affinity of LpBgl5 with non-phosphorylated substrates aligns with the earlier findings that the highly similar enzyme (Lp_3525) was characterized as a phospho-β-glucosidase/phospho-β-galactosidase (Plaza-Vinuesa et al. 2022). Additionally, a similar trend of the activity not being saturated at the highest substrate concertation was reported for a metagenomic β-glucosidase, and factors such as substrate inhibition or allosteric activation were speculated to affect the glycosidase activity at high substrate concentration (Mo et al. 2024). The specific activity values of LpBgl5 are higher than the bifunctional β-glucosidase/β-xylosidase from L. lactis reported in an earlier study (Fang et al. 2014).

The finding that LpBgl5 is not affected by metal ions is dissimilar to few β-glucosidases from LAB which showed 68–82% reduction in enzyme activity in the presence of the same metal ions (Fang et al. 2014; Zhong et al. 2016). The indifference in the β-xylosidase activity of LpBgl5 in the presence of metal ions is consistent with an enzyme from Aspergillus fumigatus XC6 (Jin et al. 2020). LpBgl5 appears to be a non-metalloprotein since its activity was not affected by EDTA, which is consistent with the bifunctional β-glucosidase/β-xylosidase (Chamoli et al. 2016).

Majority of the characterized β-glucosidases and β-xylosidases are inhibited by monosaccharides and this phenomenon can greatly impede their industrial applications (Cintra et al. 2017; Gomes-Pepe et al. 2016). Cota et al. (2015) reported varying effects of monosaccharides on the activities of β-glucosidases from Pyrococcus furiosus ATCC 43587 (PfBgl1) and Thermotoga petrophila RKU-1 (TpBgl1). For instance, xylose stimulated activities in the case of both enzymes whereas glucose led to reduction in the relative activity of the former enzyme and stimulation in the latter enzyme. The ability of LpBgl5 to retain both of its activities in the presence of all the tested monosaccharides indicates its suitability for industrial processes such as cellulose saccharification which generates various sugars. Based on the classification proposed by Salgado et al. (2018), LpBgl5 falls under class IV β-glucosidases which are not inhibited by high concentrations of glucose. A very few β-glucosidases reported in this class so far include β-glucosidases from Anoxybacillus sp. DT3-1 (Chan et al. 2016) and a metagenomic β-glucosidase (Ariaeenejad et al. 2020) from sheep rumen having tolerance to up to 10 M and 8.8 M glucose, respectively. Most of the β-xylosidases reported earlier are either inhibited by xylose (Knob et al. 2010) or not affected by xylose (Li et al. 2018; Shi et al. 2013). Thus, LpBgl5 is the first β-glucosidase/β-xylosidase stimulated by its respective end-product.

The ethanol tolerance of LpBgl5 was comparable to a metagenomic β-glucosidase having IC₅₀ of 5.8% (Gomes-Pepe et al. 2016). Protein engineering or immobilization can be employed to increase the ethanol tolerance of ethanol-susceptible enzymes (Singh et al. 2013). Using the former approach, a marine metagenomic β-glucosidase with IC₅₀ of 15% was semi-rotationally engineered to have an IC₅₀ of 30% (Fang et al. 2016).

The volatiles released by LpBgl5 from mango are similar to those released by almond β-glucosidase as we reported earlier (Godse et al. 2023). The observations that ex situ hydrolysis is more efficient and phenolic GBV are more readily hydrolyzable than monoterpenoid GBV in juice matrix are also consistent with Godse et al. (2023). Considering that LpBgl5 can also hydrolyze mango GBV as yet another class of natural glucosides, it can be a potential candidate that can be further studied for its application in releasing bioactive compounds and aroma chemicals from plant matrices.

In conclusion, we cloned 10 putative GH1 family glycosyl hydrolases from L. plantarum NCIM 2903 and one showed β-glucosidase activity indicating the importance for biochemical characterization of the annotated enzymes for their suitable applications. Based on the substrate specificities, LpBgl5 was categorized as aryl-β-glucosidase. Its activity as β-glucosidase, β-mannosidase, and β-xylosidase with higher β-xylosidase activity than the other two activities further implies the significance of biochemical characterization of these enzymes for employing them for any suitable use. Although stimulation by glucose and xylose indicates the potential of LpBgl5 in lignocellulose treatment, it was unable to hydrolyze cellulose. Such an impediment can be overcome by techniques such as site-directed mutagenesis to alter substrate specificity. Since the amino acid residues of β-glycosidases involved in specificity towards the glycon moeity and in the monosaccharide tolerance are not known, such a task would demand detailed structural characterization of the enzyme. The ability of LpBgl5 to hydrolyze plant glucosides, including its potential to release aroma compounds from mango juice, underscores the importance of such studies in identifying enzyme candidates for developing value-added and functional foods.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

RG: Data curation, formal analysis, investigation, methodology, visualization, writing—original draft. JMF: Investigation. RK: Conceptualization, Funding acquisition, project administration, formal analysis, writing—review and editing, supervision.

Funding

Open access funding provided by Symbiosis International (Deemed University). The authors would like to acknowledge the financial support from Science and Engineering Research Board (SERB) project number ECR/2017/000758, Government of India and Symbiosis Centre for Research and Innovation, Symbiosis International (Deemed University), India.

Data availability

The data used in the current study will be available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors have consented for publication.

Competing interest

The authors declare no competing interests.