Background

β-Galactosidases (EC 3.2.1.23) are enzymes that catalyze the hydrolysis of β-galactosidic bonds in oligosaccharides and transgalactosylation reactions. These enzymes have attracted increasing biotechnological interest, particularly in the dairy industry, where they are used to produce low-lactose and lactose-free milk. In addition, these enzymes could be used to hydrolyze lactose in whey, a by-product of the cheese industry, to produce various valuable products, such as galactose-containing compounds (including galacto-oligosaccharides, which are used as prebiotics), ethanol, syrup sweeteners, and lactic acid [1,2,3]. Currently, most β-galactosidases used in the food and pharmaceutical industries derive from fungi and yeasts, with Kluyveromyces sp. and Aspergillus sp. being the main sources [4]. Commercial β-galactosidases generally exhibit optimal activity at temperatures between 35 °C and 70 °C. However, identifying β-galactosidases with novel properties, such as high activity at lower temperatures, which can help reduce production costs and minimize microbial contamination in industrial processes, remains an active area of research. Potential sources of cold-active β-galactosidases are microorganisms living in environments with consistently low temperatures, such as polar regions, high mountains, and deep-sea habitats [5, 6]. As can be seen in Table 1, the majority of β-galactosidase-producing microorganisms correspond to bacteria from cold environments, with reported optimal enzymatic temperatures (Topt) ranging from 15 °C to 45 °C, but some with reported Topt as high as 60 °C (Marinomonas sp. BSi20414, isolated from the Arctic Ocean). The β-galactosidases reported from fungi generally have an elevated Topt (50 °C or higher) but, in some cases, maintain high activity at lower temperatures (30–35 °C), such as those from Penicillium chrysogenum NCAIM 00237 and Cladosporium tenuissimum URM 7803.

Table 1 Microorganisms producing ß-galactosidases with reported optimal temperature for enzymatic activity (Topt)
Full size table

The performance of an enzyme with respect to temperature is closely related to its structural properties, and it has been described that cold-active enzymes generally exhibit increased local and/or global flexibility compared to their mesophilic counterparts. A high structural flexibility of a protein can be achieved by several factors, such as a lower content of secondary structures, longer and more abundant hydrophilic loops, a larger hydrophobic surface, a smaller hydrophobic core, and a lower number of ionic-electrostatic interactions, hydrogen bonds, and salt bridges [28,29,30,31]. Reduced proline content in bacterial cold-active proteins [32, 33] has been proposed as an adaptation to mitigate the negative effect of proline isomerization on protein folding [34]. Structural properties of the active site, such as larger opening catalytic sites and longer linkers that can adopt different conformations to facilitate substrate accessibility, have also been proposed as key factors in the increased activity of cold-active enzymes at low temperatures [35,36,37]. It is important to note that no single “structural strategy” or adaptation is common to all cold-active enzymes, and each may exhibit a particular combination of the above characteristics.

This work used 16 ORFeomes corresponding to eight Antarctic fungi to search for potential genes encoding β-galactosidases. A putative coding sequence for β-galactosidase was found in Tetracladium sp. and expressed in Saccharomyces cerevisiae. The recombinant β-galactosidase was characterized in terms of its structural properties and hydrolytic activity under various conditions.

Methods

Strains, plasmids, and culture conditions

Escherichia coli DH5α (F-, Φ80dlacZ∆ M15, ∆ (lacZYA-argF) U169, deoR, recA1, endA1, hsdR17 (rk-, mk+), phoA, supE44, λ-, thi-1,gyrA96, relA1) was routinely cultured in LB medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl) supplemented with 0.2% glucose at 37 °C. S. cerevisiae INVSc1 (MATa, his3D1, leu2, trp1-289, ura3-52) was grown in YM medium (0.3% yeast extract, 0.3% malt extract, 0.5% bactopeptone) supplemented with 1% glucose at 30 °C, with orbital shaking at 170 r.p.m. The SC medium (0.85% yeast nitrogen base without amino acids and with ammonium sulfate, 0.5% casamino acids, 0.4% NaOH, 2% glycerol, 1% succinic acid, 2% glucose, 0.005% histidine, 0.01% leucine, 0.01% uracil, 0.01% adenine hemisulfate) was used for S. cerevisiae transformant selection. Agar was added at a final concentration of 1.5% for semi-solid media. Plasmid pUC57 was used for cloning experiments in E. coli, and transformants were selected on LB-agar plates supplemented with 100 µg/ml ampicillin. For molecular cloning and gene expression in S. cerevisiae, plasmid pYES3/CT was used.

Standard molecular and biochemical methods

Standard methods such as plasmid DNA and protein extractions, agarose and polyacrylamide gel electrophoresis, conventional PCR, and restriction digestions were performed as previously described [38] unless otherwise noted. T4 DNA ligase, PfuUltraII Fusion HS DNA polymerase, restriction endonucleases, DNase I, RNase A, and T4 polynucleotide kinase were purchased from Agilent Technologies, Thermo Scientific, and Life Technologies and were used according to the manufacturer’s instructions. Transformations of competent cells were performed by electroporation using a GenePulser XcellTM (BioRad, Hercules, CA, USA) equipment with settings of 1.8 kV, 25 µF, 200 Ω for E. coli, and 1.5 kV, 25 µF, 200 Ω for S. cerevisiae. When required, sample absorbance at different wavelengths was measured using Jasco UV-Vis spectrophotometer V-630 and Epoch 2 Microplate Spectrophotometer (Biotek, Winooski, VT, USA).

Prediction of sequences encoding β-galactosidases

The protein sequences of β-galactosidases were downloaded from the UniProt database and used to build a local database using Geneious Prime software. The translated ORF sequences, predicted from the transcriptomes of eight Antarctic yeasts [39], were compared to the local database using BLASTp with the “Bin into ‘hashit’ vs. ‘no hit’” option in the Geneious software. The “hit sequences” were then selected and compared by BLASTp to the UniProt database for annotation, applying a similarity threshold of ≥ 30%, using the Blosum62 cost matrix. Sequences annotated as related to β-galactosidase were selected for further analysis. The prediction of domains, families, or functional sites was performed using the InterProScan server (https://www.ebi.ac.uk/interpro/search/sequence/). The DeepLoc-2.0 web server (https://services.healthtech.dtu.dk/services/DeepLoc-2.0/) was used to predict subcellular localization.

Maximum likelihood evolutionary analysis

Protein sequences were aligned using MAFFT [40] and trimmed using trimAl [41]. The phylogeny was inferred using IQ-TREE [42] with automatic model selection (MFP), and branch support was assessed using 1,000 ultrafast bootstrap replicates [43].

3D structural modeling and comparison

The 3D protein models were constructed using the SWISS-MODEL server (https://swissmodel.expasy.org) with the best ortholog identified as a template considering the following parameters: coverage ≥ 50%, similarity ≥ 30%, Global Model Quality Estimate (GMQE) ≥ 0.8. The quality of the models was assessed using VERIFY 3D [44, 45] available at UCLA-DOE LAB-SAVES v6.1 (https://saves.mbi.ucla.edu/).

Protein structural properties related to flexibility, such as the number of hydrogen bonds, salt bridges, apolar solvent-accessible surface area (apoSASA), and content of secondary structures [46], were calculated in each model. The properties and parameters used in the ChimeraX software [47, 48] for these calculations were: apoSASA (radius = 1.4 Å, peak density = 2), hydrogen bonds, and salt bridges (radius = 0.075 Å, dashes = 6, distance tolerance = 0.4 Å, angle tolerance = 20º). The secondary structure content was determined from PDB files using the Pfeature web server (https://webs.iiitd.edu.in/raghava/pfeature/sec.php). The percentage of residues in regions predicted to be rigid (0), flexible (1), or very flexible (2) was calculated using MEDUSA [49]. Comparative superimposition analysis of protein 3D structures was performed using the Dali server (http://ekhidna2.biocenter.helsinki.fi/dali/; [50]).

Cloning, expression, and purification

The ORF sequence of the selected β-galactosidase was modified in silico according to the codon usage of S. cerevisiae and then synthesized by Gene Universal (https://www.geneuniversal.com/), including cloning into the vector pUC57 (named pUC57Bgal1) and transformed into E. coli DH5α by electroporation. The pUC57Bgal1 plasmid was purified from E. coli culture and used as a template in a PCR reaction using primers Bgal1KpnI (5’-GTACggtaccATGGCTTCTTCTGATAAAAACTTCCCT-3’) and Bgal1ApaI (5’-GTACgggcccTTCATCAGCTTCCAAAGAGTAAAC-3’), which contained restriction sites for KpnI and ApaI (indicated in lower case), respectively. The PCR products were resolved on a 1% agarose gel. The amplicon band of the expected length (1,709 bp) was purified from the gel using the Gene Clean protocol [51], ligated into the pYES3/CT vector, and transformed by electroporation into E. coli DH5α. The recombinant plasmid pYES3/CTbgal1 was purified from E. coli cultures and transformed into S. cerevisiae electrocompetent cells by electroporation. Transformants were selected on SC agar plates after incubation at 30 °C for 16 h and verified by colony PCR using primers Gal1 (5’-AATATACCTCTATACTTTAACGTC-3’) and CYC1 (5’-GCGTGAATGTAAGCGTGAC3’). A S. cerevisiae transformant clone was cultured in 50 ml SC medium for 48 h at 30 °C. The culture was centrifuged at 5,230 g for 10 min, the cell pellet was suspended in 50 ml sterile distilled water, and centrifuged at 5,230 g for 5 min. The cell pellet was then inoculated into SC medium supplemented with 2% galactose to achieve an OD600nm of 0.4, incubated at 30 °C, and culture aliquots of 50 ml were collected after 8, 24, and 48 h. These aliquots were centrifuged at 5,230 g for 10 min, the supernatant was discarded, and the cell pellet was stored at -50 °C until further processing. Five ml of the cell pellet was suspended in one volume of 10 mM Tris-HCl, 0.5 M NaCl, and 10 mM imidazole, pH 8.0, and one ml of cell suspension was aliquoted into lysis tubes. Two hundred µl of 0.5 mm diameter glass beads (Biospec, Bartlesville, OK, USA) were added to each tube, and the cells were disrupted using a Mini-Beadbeater-16 (Biospec, Bartlesville, OK, USA) with 7 cycles of agitation for 3 min, followed by 3 min of incubation on ice. The supernatant obtained after centrifugation of the sample at 19,745 g for 15 min was mixed with 100 ml of 10 mM Tris-HCl, 0.5 M NaCl, 10 mM imidazole, pH 8.0, and then loaded into a HisTrap FF crude column (GE Healthcare, Chicago, IL, USA). Samples were eluted using an Akta prime plus equipment (GE Healthcare, Chicago, IL, USA) with a mobile phase of 20 mM Tris-HCl, 0.5 M NaCl, pH 8.0, at a flow rate of 1 ml/min and an imidazole gradient from 10 to 500 mM. One ml aliquots were collected for subsequent analyses.

Enzyme activity assays

Enzyme assays were performed under different temperatures and pH conditions. The pH was adjusted using two sets of buffers: (i) 0.1 M sodium acetate buffer (pH 5.0 and 5.5); 100 mM phosphate buffer (pH 6.0 to 7.0), containing 10 mM KCl, 1 mM MgSO4; (ii) phosphate-citrate buffer (pH 5.0 to 7.0).

For the β-galactosidase activity determination [52], 7.5 µl of protein sample (25 µg/ml) was mixed with 200 µl of the appropriate pH buffer, and 70 µl of 4 mg/ml o-nitrophenyl-β-D-galactopyranoside (ONPG) was added. The mixture was incubated at different temperatures, and the reaction was stopped at different times by adding 500 µl of 1.0 M Na2CO3. The release of o-nitrophenol from ONPG was measured by absorbance at 420 nm. For lactose hydrolysis determination, a solution of 5% lactose was used, the reaction was stopped by incubation at 100 °C for 10 min, and the release of glucose was quantified using the Glucose Assay Kit abcam (Cambridge, UK), according to the manufacturer’s instructions.

For the α-glucosidase activity determination [53], 7.5 µl of protein sample (25 µg/ml) was mixed with 50 µl of buffer pH 5.0, pH 6.0, or pH 7.0 and 17.5 µl of 1.5 mg/ml of 4-nitrophenyl-α-D-glucopyranoside (pNPG). The reaction mixture was incubated at 35 °C for 1 h and then stopped by adding 125 µl of 1 M NaHCO3. The enzyme activity was determined by measuring absorbance at 405 nm.

For the β-glucuronidase [54] and β-glucosidase [55] activity determinations, 25 µl of enzyme sample was mixed with 150 µl of buffer pH 5.0, pH 6.0, or pH 7.0, and 25 µl of 1 mg/ml 4-methylumbelliferyl-β-D-glucuronide (4MBG) or 4-methylumbelliferyl-β-D-glucoside (4MBDG), respectively. The reaction was incubated at 35 °C for 1 h and stopped by adding 1.3 ml of 0.1 M glycine-NaOH buffer pH 10.4. The reaction was then exposed to UV light using a CUV40A transilluminator (ClinX, Shanghai, China) to observe fluorescence emission, indicating the enzymatic release of 4-methylumbelliferone.

Protein thermal unfolding kinetics

Thermal protein unfolding kinetics were assessed as previously described [56] using a CFX96 Real-Time System (BioRad, Hercules, CA, USA) in 96-well PCR microplates. In each well, 13 µl of the appropriate buffer at the desired pH, 10 µl protein sample (1 µg/µl), and 2 µl 125X concentrated SYPRO Orange were added. The temperature was increased from 5 °C to 95 °C with a ramp of 1 °C per min. The excitation wavelength was set to 470 nm, and the emission was registered at 569 nm.

Results

Prediction of β-galactosidases in the ORFeomes of Antarctic fungi and structure of the putative gene and enzyme

By mining the transcriptomes of eight Antarctic fungi for ORFs encoding β-galactosidases, a putative ORF of 1692 nt was identified in the transcriptome of Tetracladium sp. and mapped to the corresponding genome to determine the gene structure. The encoding gene tspbgal spans 1884 nt (NCBI accession number PQ310115) and consists of four exons, resulting in a coding sequence of 1692 nt (Fig. 1A). A strong Kozak sequence was identified, including the translation start codon ATG. Tspbgal was predicted to be a cytoplasmic protein (70% probability), and the predicted domains were glycosyl hydrolase family 42, a transglycosylase, and a domain of unknown function 5597 (DUF5597).

Fig. 1
figure 1

Gene and protein structure of β-galactosidase from Tetracladium sp. A Scheme of the protein functional domains (above) and gene exon-intron structure (below). The translation start codon in the Kozak sequence is indicated. B Superposition of the crystal structure of β-galactosidase from C. japonicus (cyan) and the predicted model of the β-galactosidase from Tetracladium sp. (brown). The iminosugar 1-deoxygalactonojirimycin (DGJ), is shown in pink. The major structural differences between both structures are labeled SD1 to SD8 and colored in blue for the enzyme from Tetracladium sp. and yellow for the one from C. japonicus. Zoom of the active site (discontinuous circle) indicating the conserved residues in both models and showing those corresponding to C. japonicus in parentheses. The non-conserved residue in C. japonicus is written in red, and the putative alternative residues in Tetracladium sp. are in blue, including their distance to DGJ

Full size image

Multiple alignment and phylogenetic analyses were performed with Tspbgal and 89 related β-galactosidases, with sequences from 558 to 601 residues identified through BLASTp searches in the NCBI Protein Reference and Swiss-Prot databases. Tspbgal grouped with β-galactosidases from the species Baudoinia panamericana, Cyphellophora europaea, Exophiala dermatitidis, Capronia coronata, and Pseudogymnoascus sp., but with low bootstrap support (Figure S1).

A three-dimensional model of Tspbgal was constructed by homology modeling using Cellvibrio japonicus Ueda107 tetrameric β-galactosidase (Bgl35A, PDB ID: 4D1I) as the template, the best ortholog found in the Swiss Model Server. Superposition analysis of the monomers from the predicted model of Tspbgal and Bgl35A using the Dali server aligned 492 C-alpha atoms showed a sequence identity of 34%, a Z-score of 61.1, and an r.m.s.d. value of 1.1 Å. Several structural differences (labeled SD1 to SD8 in Fig. 1B) were observed by superimposing the three-dimensional structures of the monomers from the crystal structure of Bgl35A and the predicted model for Tspbgal. Four loops in Tspbgal (SD1, SD2, SD3, and SD8) were 4, 21, 11, and 18 residues longer than the corresponding loops in Bgl35A, respectively. Loops SD5 and SD7 in Tspbgal were 13 and 8 residues shorter than their counterparts in Bgl35A, respectively. The helix SD4 and the helix-loop-helix SD6 structures observed in Bgl35A were present as loops in Tspbgal and were shorter by 3 and 13 residues, respectively. To analyze the active sites, the crystal structure of Bgl35A was soaked with the iminosugar 1-deoxygalactonojirimycin (1,5-dideoxy-1,5-imino-D-galactitol, DGJ) [57]. Residues N67, K134, N135, and N204, which coordinate DGJ in the active site cavity of Bgl35A, are conserved in Tspbgal, corresponding to residues N36, K104, N105, and N179, respectively. Residue N383 of Bgl35A, which interacts with the O6 of DGJ at a calculated distance of 3.2 Å, is absent in Tspbgal. However, other nearby residues in Tspbgal may potentially interact with this atom. Although residue Y334 is too distant (4.3 Å) to form a canonical hydrogen bond, its location within a predicted flexible region suggests that it could still contribute to O6 coordination. Similarly, residues Q35 and R358, which are located even farther from O6 (7.8 Å and 8.4 Å, respectively), are situated in the same flexible region and may therefore also be considered potential candidates for its coordination. The proposed catalytic residues E205 and E349 in Bgl35A are structurally conserved in Tspbgal, corresponding to residues E180 and E356, respectively. The conservation of these residues can be observed in other 89 fungal β-galactosidases related to Tspbgal, as shown in Fig. 2 and Figure S2, including the putative active site residues of Tspbgal that are conserved relative to Bgl35A, as well as residue Y334 in Tspbgal, which likely coordinates O6 of DGJ.

Fig. 2
figure 2

Sequence alignment of fungal β-galactosidases. Only one representative from each genus is shown, along with the regions containing conserved residues. The full alignment is shown in Figure S2. Conserved predicted domains are shown at the top. Residues in the active site, as shown in Fig. 1B, are indicated by blue boxes

Full size image

Purification and enzymatic characterization of recombinant Tspbgal

As described in the Materials and Methods section, the S. cerevisiae INVSc1 strain was transformed with the pYES3/CTbgal1 vector by electroporation, and transformants were selected by PCR with the Gal1/CYC1 primer pair (amplicon size ~ 2,000 bp; see Figure S3A). Tspbgal expression was induced, culture aliquots were collected at 8, 24, and 48 h post-induction, proteins were extracted from cell pellets, and analyzed by SDS-PAGE and for β-galactosidase activity. A protein band of the expected size (63 kDa) was observed in the SDS-PAGE, and β-galactosidase activity was detected in all protein samples (Figure S3B). Recombinant Tspbgal was purified by His-tagged affinity chromatography on nickel columns using an imidazole gradient (Fig. 3A). A protein with the expected molecular weight (63 kDa) eluted in fractions between 8 and 16 min, corresponding to the fractions displaying β-galactosidase activity (Fig. 3B). The activity of the purified β-galactosidase was evaluated at temperatures ranging from 10 °C to 60 °C and pH levels ranging from 5.0 to 7.0, with and without a buffer containing MgSO₄. The enzyme activity was higher at temperatures between 35 °C and 40 °C and pH levels from 5.5 to 6.6, with the highest activity at 35 °C and pH 6.0 (Fig. 3C). The performance of β-galactosidase activity across pH levels was not affected in buffers with or without MgSO₄. The enzyme maintained at least 60% of its maximum activity at temperatures ranging from 25 °C to 40 °C, and pH levels between 5.0 and 7.0 (Fig. 3D). It also retained 33% of its maximum activity at 10 °C and pH 6.0. The enzyme showed high stability at 35 °C and pH 6.0, retaining 87% of its maximum activity after 4 h and 67% after 24 h at these conditions (Fig. 3E).

The specificity of recombinant Tspbgal was evaluated in assays performed using pNPG (for α-glucosidase activity), 4MBG (for β-glucuronidase activity), and 4MBDG (for β-glucosidase activity) as substrates at 35 °C and pH 5.0, 6.0, and 7.0. No activity was detected for any of these substrates under these conditions (data not shown).

The ability of Tspbgal to hydrolyze lactose was tested using a 5% solution of lactose at 35 °C and a pH of 6.0, conditions that yield the maximum enzyme activity. As shown in Fig. 4, the hydrolysis of approximately 40–50% of the lactose was achieved after 3–4.5 h of incubation, and nearly 70% after 11.5 h. When the lactose hydrolysis was tested at different temperatures at pH 6.0, the Tspbgal retained a residual activity of 52% and 59% at 22 °C and 45 °C, respectively. The highest lactose hydrolysis by Tspbgal was observed at pH 5.0 at 35 °C, with residual activity of around 80% at pH 6.0 and 7.0.

Fig. 3
figure 3

Purification and characterization of recombinant β-galactosidase. A Protein elution profile measured at 280 nm (blue) and imidazole gradient (red). B SDS-PAGE analysis of the collected fractions, along with colorimetric detection of β-galactosidase activity (yellow; bottom). The arrow indicates the protein band with the expected relative molecular weight (63 kDa). St, protein standard; R, crude protein extract; 5 to 16, elution fractions 5 to 16 min. C Recombinant β-galactosidase activity on ONPG at different temperatures and pH adjusted with two sets of buffers: sodium acetate and phosphate containing 1 mM MgSO4 (AP) and phosphate citrate (PC). D Percentage of activity relative to the maximum activity shown in C, represented by circle size and color. E Stability assay conducted over time; the relative enzyme activities were assessed after incubation at 35 °C and pH 6.0 from 0 to 24 h (curves of two independent experiments with three replicates each). In panels C and E, the average of triplicates is shown, and the bars represent the standard deviation

Full size image
Fig. 4
figure 4

Lactose hydrolysis by Tspbgal. Assays were performed using a 5% lactose solution, and glucose release was quantified to calculate the percentage of lactose hydrolysis. A % of lactose hydrolysis over time at 35 °C and pH 6.0. B Residual lactose hydrolysis activity at pH 6.0 at different temperatures. C Residual lactose hydrolysis activity at 35 °C and different pH levels. Values represent the mean of three independent experiments, and error bars indicate the standard deviation

Full size image

Thermal unfolding of Tspbgal and relationship between the structural properties of fungal β-galactosidases and their optimal activity temperature

The thermal unfolding of Tspbgal was assessed at pH levels ranging from 5.0 to 9.0. A variation in the thermal unfolding curves was observed, with lower thermal stability occurring at pH values above 7.0 (see Fig. 5A). The highest melting temperatures (Tm) were observed at pH 5.5 (Tm = 49 ± 0.2 °C) and pH 6.0 (Tm = 49 ± 0.4 °C), which were slightly lower at pH 5.0 (Tm = 46 ± 0.1 °C) and pH 6.6 (Tm = 47 ± 0.2 °C). The lowest Tm values were observed at pH 8.0 to 9.0, ranging from 35 °C to 38 °C. The β-galactosidase activity was evaluated at pH levels from 5.0 to 7.0 and temperatures from 10 °C to 60 °C and expressed as a percentage with respect to the maximum activity. As shown in Fig. 5B, the effect of temperature on enzyme activity correlated with its effect on protein structure, since maximum enzyme activity was observed under conditions when still no protein unfolding was detected (35 °C to 40 °C), and the interpolated 50% of enzyme activity corresponded with 50% of protein unfolding.

Fig. 5
figure 5

Thermal unfolding and activity of Tspbgal at different pH levels. A Thermal protein unfolding assay performed at different pH values. Relative unfolding was calculated as the percentage of fluorescence at each point relative to the maximum fluorescence detected in each curve. The melting temperature at each pH is shown in the inserted box. B Comparison of the percentage of enzyme activity (act) and protein unfolding (unf) at different temperatures and pH. Values represent the mean of three independent experiments, and error bars indicate the standard deviation

Full size image

The structural properties associated with protein flexibility that have been proposed to enhance enzyme activity at lower temperatures were determined in 47 β-galactosidases, for which data on their optimal temperature for enzyme activity (OTEA) and the optimal temperature for growth (OTG) of the producer organism were available, and analyzed by principal component analysis (PCA). As shown in Fig. 6A, the two principal components explain 53.7% of the variance, and the OTG was the parameter most related to OTEA. Some positive relations to OTEA were observed regarding structural protein properties, including solvent accessible surface area (SASA), the percentage of residues classified as medium flexible (Med1), and the percentage of α-helix (Alpha). Negative relations were observed for the number of hydrogen bonds (Hbond) and salt bridges (SaltBrid), apolar solvent accessible surface area (apoSASA), and the percentage of β-sheet. Spearman’s correlation was calculated for each parameter versus OTEA (see Fig. 6B). Positive correlations with a p-value of ≤ 0.05 were found for OTG (0.6) and the percentage of α-helix (0.4). Negative correlations were found for apoSASA (-0.4) and the percentage of β-sheet (-0.4).

Fig. 6
figure 6

Relationship between the structural properties of β-galactosidases and their optimal temperature for activity. A Principal component analysis (PCA) of data from 47 β-galactosidases, and B data dispersion of each parameter versus temperature of the enzyme. Spearman’s correlations between parameters and optimal temperature for activity with a p-value ≤ 0.05 are indicated in parentheses. OTEA, optimal temperature for enzyme activity; OTG, optimal growth temperature; Alpha, α-helix; Beta, β-sheet; Med1, % of residues classified as medium flexibility by MEDUSA; Med2, % of residues classified as high flexibility by MEDUSA; Hbond, number of hydrogen bonds; SaltBrid, number of salt bridges; SASA, solvent accessible surface area; apoSASA, apolar solvent accessible surface area

Full size image

Discussion

The search of the ORFeomes from eight Antarctic fungi revealed an ORF encoding a β-galactosidase in the Tetracladium sp. ORFeome. Genome mapping defined an 1884-nt gene consisting of four exons with a 1,692-nt coding sequence. The predicted 564-residue β-galactosidase from Tetracladium sp. showed low sequence identity with other described β-galactosidases; its closest relatives were found in the genera Hyaloscypha, Scytalidium, Hyphodiscus, Helotiales, Mollisia, and Cadophora. These are all mycorrhizal and saprophytic filamentous fungi [58,59,60], and one species, Cadophora malorum, has also been isolated from a cold environment, specifically from lakes on the Antarctic Peninsula [61]. The most suitable template for three-dimensional structure modeling of the β-galactosidase from Tetracladium sp. was the GH35 β-galactosidase Bgl35A from Cellvibrio japonicus. Although the overall structural organization of the two enzymes was conserved, the most prominent differences were localized to the unstructured loop regions, which were generally longer in TspBgal. Extended loops and turns frequently generate flexible surface regions, a characteristic that has been associated with the enhanced structural plasticity of psychrophilic enzymes [62]. In this work, the analysis of structural features commonly associated with protein flexibility in β-galactosidases revealed a negative correlation between the optimal temperature for enzymatic activity and the apolar solvent-accessible surface area. Regarding secondary structures, a positive correlation with α-helix content and a negative correlation with β-sheet content were observed. The increased presence of surface-exposed hydrophobic groups may contribute to protein destabilization by reducing entropy, as water molecules reorganize around these hydrophobic side chains [28, 30].

According to the results obtained with oNPG and lactose as substrates, the recombinant Tetracladium sp. β-galactosidase exhibits optimal activity at temperatures between 25 °C and 40 °C and within a pH range of 5.5 to 7.0. Notably, the enzyme retains approximately 25% of its activity at 10 °C, highlighting its functionality under cold conditions. These characteristics are comparable to those reported for several bacterial β-galactosidases. For example, enzymes from Alteromonas species show maximal activity at pH 6.0 to 8.0 and temperatures between 25 °C and 45 °C [8,9,10]. Similarly, β-galactosidases from Arthrobacter species function within a pH range of 7.0 to 8.0, with optimal temperature between 18 °C and 28 °C [11,12,13,14, 63, 64]. Enzymes from Pseudoalteromonas species exhibit similar profiles, displaying activity between pH 6.0 and 8.5 and at temperatures ranging from 23 °C to 45 °C [19, 65, 66]. Additionally, β-galactosidases from Antarctic bacteria such as Exiguobacterium antarcticum B7, Micrococcus antarcticus, and Rahnella spp. show optimal activity at pH 6.5 to 7.0 and temperatures of 25 °C to 35 °C [15, 18, 20]. In contrast, β-galactosidases from fungi generally exhibit different biochemical properties. Most reports correspond to Aspergillus species, with optimal activity observed at acidic pH values ranging from 3.5 to 6.0 and at higher temperatures of 40 °C to 60 °C [22, 24,25,26]. Other examples include β-galactosidases from Penicillium chrysogenum, with optimal activity at pH 4.0 and 30 °C [27], and Cladosporium tenuissimum, which displays activity at pH 3.0 to 4.5 and temperatures between 35 °C and 50 °C [23].

It has been reported that the activity of β-galactosidases is influenced by the presence of various ions. For example, enzymes from Kluyveromyces lactis and Kluyveromyces fragilis require Mn²⁺, Na⁺, and Mg²⁺ for optimal activity [3], while the enzyme from Arthrobacter oxydans SB showed improved activity in the presence of Mn²⁺ or Fe²⁺ [67]. In this work, we evaluated the effect of Mg2+ on the activity of the β-galactosidase from Tetracladium sp., and observed comparable activity in its presence or absence. These results suggest that, unlike some other β-galactosidases, the enzyme from Tetracladium sp. may not be strongly dependent on Mg²⁺ for catalysis.

These properties of Tetracladium sp. β-galactosidase are promising for its potential application in the treatment of milk to reduce lactose content at lower temperatures, particularly desirable in industrial applications, as it reduces energy costs and helps preserve heat-sensitive substrates. Cold-active β-galactosidases are relevant in dairy processing, where they enable the production of lactose-free milk suitable for individuals with lactose intolerance, while maintaining the natural flavor and nutritional quality of the product. In addition, these enzymes play an important role in the manufacture of ice cream and condensed milk, preventing lactose crystallization and simultaneously enhancing sweetness and creaminess [5, 68, 69]. In addition, galacto-oligosaccharides are widely applied in the food and cosmetic industries as functional additives and natural sweeteners [5, 70]. Cold-active β-galactosidases are also promising for biosensor development, where their high activity at low temperatures facilitates integration into analytical devices for lactose detection [6, 68, 71].

Moreover, the pH range at which Tetracladium sp. β-galactosidase exhibits high activity corresponds closely to the pH of whey. As a byproduct of the dairy industry, whey can be acidic (pH ~ 5.0) or sweet (pH 6.0 to 7.0), depending on the processing method, and has been proposed as a substrate for the production of whey protein hydrolysates and other value‑added products [1]. The typical whey composition contains 4.5% to 5.0% w/v lactose, which can be used for biofuel production. For example, the combined use of β-galactosidase from Kluyveromyces marxianus and S. cerevisiae was used to produce 28.9 g/L ethanol during fermentation at 35 °C [72]. Furthermore, enzymes with β-galactosidase activity can also carry out transgalactosylations to produce prebiotics, such as galacto-oligosaccharides, which can be used to support the growth of beneficial microorganisms in the human gastrointestinal tract [69, 73, 74].

Regarding psychrophilic enzymes, most are inactivated by temperature before their structure unfolds, which reveals the heat lability of the active site. This is different in mesophilic enzymes, where thermal inactivation correlates with protein unfolding [35]. As Tetracladium sp. is a cold-adapted fungus that was isolated from Antarctic soil, it was expected that the behavior of its β-galactosidase would resemble that of psychrozymes. However, the enzyme inactivation by temperature coincides with protein unfolding. Therefore, Tetracladium sp. β-galactosidase is structurally more like enzymes from mesophiles, such as C. japonicus, which has optimal activity at pH 6.5 and thermal stability between 35 °C and 45 °C, with a rapid decrease in activity at 55 °C and above [75].

Conclusions

A novel β-galactosidase was identified in the Antarctic fungus Tetracladium sp., which was successfully expressed in S. cerevisiae. The recombinant enzyme showed structural and thermal stability properties similar to mesophilic enzymes but with improved performance at temperatures below 35 °C, a property desirable in the dairy industry to reduce production costs and microbial contamination.