Galactooligosaccharides-containing composition

Abstract

The present invention relates to a composition comprising galactooligosaccharides (GOS) wherein the GOS are obtainable by transgalactosylation of lactose using a β-galactosidase comprising an amino acid sequence of SEQ ID NO: 1 or one having at least 60% homology to the amino acid sequence of SEQ ID NO: 1, wherein the β-galactosidase comprises one or more amino acid substitutions in its active site. A use of the composition in improving intestinal health is also disclosed.

Description

[0001] GALACTOOLIGOSACCHARIDES-CONTAINING COMPOSITION

[0002] Field of the Invention

[0003] The present invention relates to a composition comprising galactooligosaccharides (GOS) wherein the GOS are obtainable by transgalactosylation of lactose using a β-galactosidase comprising an amino acid sequence of SEQ ID NO: 1 or one having at least 60% homology to the amino acid sequence of SEQ ID NO: 1 , wherein the β-galactosidase comprises one or more amino acid substitutions in its active site. The invention also relates to use of the composition in improving intestinal health and a method for improving the intestinal health of a subject by administering the composition.

[0004] Background to the Invention

[0005] The gut microbiome is a complex ecosystem that can have a strong interaction with the host and provide beneficial effects. Studies have demonstrated that different compositions of the gut microbiome can be related to either the development or prevention of certain diseases, including metabolic disorders and cardiovascular disease. There are many factors that influence the microbiome balance including diet, antibiotics, infections and stress, which can all alter the stability of the microbial population and thereby have a deleterious effect on the health of the host. Different strategies to favorably modulate the gut microbiome have been proposed. One approach involves the use of probiotics and prebiotics as dietary supplements or ingredients to alter the microbial composition in the gut.

[0006] Galacto-oligosaccharides (GOS) are defined as prebiotic compound mixtures produced by the enzymatic transgalactosylation of lactose by β-galactosidase enzymes. The backbone structure of GOS is based on lactose plus a further residue of galactose to obtain a mixture of sugars with different degrees of polymerization (DP) varying from DP2 to DP9 and the presence of different ratios of glycosidic linkages of [3-(1 →2), β-(1 →3), β-(1 →4), and / or P-(1 → 6). One of the most important properties of GOS is its indigestibility. In vitro and in vivo experiments show that GOS cannot be hydrolysed by enzymes in the human stomach and intestine. Instead, these are digested and fermented by microflora in the gut. Therefore, GOS can feed an intestinal microbiome with a positive impact on health and composition.

[0007] Different GOS mixtures show different prebiotic effects due to the ability of commensal bacteria to digest GOS with a certain degree of polymerization and glycosidic linkages. Recent studies showed for example that within several species of Bifidobacteria and Lactobacilli, certain strains can utilize only DP2 GOS, while others can utilize GOS < DP3.

[0008] P-galactosidases are glycoside hydrolase (GH) enzymes that catalyse the hydrolysis of β- galactosides into monosaccharides through the breaking of a glycosidic bond formed between a galactose unit and glucose. They can also catalyse the transgalactosylation of galactose moieties to produce oligosaccharides. GH enzymes are grouped into 9 families: GH1 , GH2, GH3, GH35, GH39, GH42, GH50, GH59, and GH147. All of these are capable of hydrolysing terminal non-reducing β-D-galactose residues in β-D-galactosides.

[0009] Glycoside hydrolases, exhibiting β-galactosidase activity, are present in many different microorganisms, from Bacteria to Eukaryota, and there are many homologues with similar sequences. Despite having a similar sequence and structure, these enzymes have evolved for different functions. For these reasons, different enzymes from specific organisms can be used, by different techniques, to produce a variety of active molecules.

[0010] There are several different β-galactosidases commercially available, and some are used in both the food industry and pharmaceutical research. The product of lactose conversion depends on the ability of the enzyme to recognise and bind the galactosyl acceptor at the active site in a specific orientation during the catalytic moment to improve the transgalactosylation efficiency and establish the glycosidic bond of the product.

[0011] Lactobacillus rhamnosus GR1 (LGR-1) is one of the most widely used probiotics for woman’s health. It has been shown, in clinical studies, to help restabilise a healthy vaginal microbiome and prevent urinary tract infections. Despite its discovery in the urinary tract and vaginal microbiota, more recent studies have explored the effects of LGR-1 on gut health. For example, it has been demonstrated in mice models that LGR-1 can give a protective effect against airborne allergies and asthma inflammation response by increasing gut microbiome resilience. Furthermore, oral supplementation of LGR-1 , but not L. rhamnosus LGG or L. reuteri RC-14, has shown positive effects on re-stabilising a healthy microbiome after toxic lead (Pb) exposure, and LGR-1 has exhibited the potential to aid recovery from memory and cognitive ability loss. Additionally, LGR-1 has been tested for its role in cardiovascular health and administration of LGR-1 has been found to reduce the leptimadiponectin plasma concentration ratio therefore reducing heart hypertrophy and improving recovery after ischemia- induced heart failure in rats. The mechanism behind the probiotic effects arising from LGR-1 is unclear. It could be due to the production of specific prebiotic carbohydrates (GOS) that, due to their specific structural characteristics, can feed specific beneficial strains. Of particular note, LGR-1 has one gene that encodes for a β- galactosidase, which belongs to the GH35 protein family and it is capable of using lactose as a substrate and thus synthesise GOS from galactose units.

[0012] It is an object of this invention to produce GOS with prebiotic properties comprising varying degrees of polymerization and a range of glycosidic bonds using β-galactosidase enzymes with mutated amino acid residues in their active sites.

[0013] Statements of invention

[0014] In a first aspect of the invention there is provided a composition comprising galactooligosaccharides (GOS) wherein the GOS are obtainable by transgalactosylation of lactose using a β-galactosidase comprising an amino acid sequence of SEQ ID NO: 1 or one having at least 60% homology to the amino acid sequence of SEQ ID NO: 1 , wherein the β-galactosidase comprises one or more amino acid substitutions in its active site.

[0015] P-galactosidases are glycoside hydrolase enzymes that catalyse the hydrolysis of terminal non-reducing β-D-galactose residues in β-D-galactosides. During transgalactosylation the P-galactosidase transfers a galactosyl moiety to lactose forming GOS.

[0016] The active site of β-galactosidase is located in domain 1 of each monomer of the β- galactosidase in any oligomeric state. The active site is in a deep pocket of the triosephosphate isomerase (TIM) barrel. A substrate sugar moiety occupies a pocket in the active site during hydrolysis of the substrate and synthesis of a resulting product. The active site as described here includes any amino acid residues lining the active site binding pocket. The active site also encompasses any amino acid residues that form temporary bonds with a substrate, and any residues that catalyse reaction of the substrate.

[0017] A substitution is a change from one amino acid in the active site amino acid residues of the P-galactosidase to a different amino acid. The change in amino acid may result from a point mutation in the corresponding encoding DNA sequence. SEQ ID NO: 1 is the wild-type amino acid sequence of a β-galactosidase belonging to Lactobacillus rhamnosus GR-1. SEQ ID NO:2 is the gene sequence which encodes the β-galactosidase belonging to Lactobacillus rhamnosus GR-1.

[0018] In embodiments the β-galactosidase may comprise an amino acid sequence having 60- 100% homology to the amino acid sequence of SEQ ID NO:1. The β-galactosidase may comprise an amino acid sequence having 65,%, 70%, 75%, 80%, 85%, 90%, 95% or 99% homology to the amino acid sequence of SEQ ID NO:1. The β-galactosidase may consist of SEQ ID NO:1 and comprise at least one amino acid substitution in its active site.

[0019] The β-galactosidase may comprise any amino acid substitution(s) in the TIM barrel domain.

[0020] The inventors carried out homology modelling of the amino acid sequence encoding for β- galactosidase designated as “SEQ ID NO:1”. They selected a homology model to SEQ ID NO:1 at 60% identity and 100% confidence as the final homology model for further work.

[0021] The lactose may be lactose monohydrate. The concentration of lactose monohydrate may be in a range from 0.2 g / mL to 0.5 g / mL.

[0022] GOS production by the β-galactosidase may be carried out over 4-24 hours at 25-60°C and pH 5.5-7.5. GOS production by the β-galactosidase may occur in the presence of magnesium. GOS production by the β-galactosidase may occur in the presence of magnesium chloride at a concentration of magnesium chloride in a range of 10 mM to 30 mM.

[0023] In embodiments the β-galactosidase comprises one or more amino acid substitutions of any of the amino acids Y52, N56, H58, Y94, I95, A97, E98, D129, N155, E156, G211 , E238, W240, W243, F244, N245, N252, R246, W247, Y275, G279, T281 , N287, G288, S290, L297, P298, D304, Y305, D306, T383, N457, L492 and / or G534 of SEQ ID NO: 1.

[0024] In embodiments the β-galactosidase comprises one or more amino acid substitutions of any of the amino acids Y52, Y94, I95, A97, E98, N155, E156, E238, W243, F244, N245, R246, W247, Y275, G279, N287, G288, S290, L297, P298, D304, Y305, D306, and / or N457 of SEQ ID NO: 1. The amino acid residues Y52, Y94, I95, A97, E98, N155, E156, E238, W243, F244, N245, R246, W247, Y275, G279, N287, G288, S290, L297, P298, D304, Y305, D306, and / or N457 of SEQ ID NO:1 are in the active site of the β-galactosidase.

[0025] In embodiments the amino acid substitutions may be present in any of subsite- 1 , subsite +1 or subsite +2 of the active site. Subsite -1 , subsite +1 and subsite +2 are subsite binding pockets in the active site of β-galactosidase. Subsite -1 , subsite +1 and subsite +2 comprise amino acid residues involved in binding a sugar moiety during hydrolysis and synthesis. The amino acids located in subsite -1 are amino acids Y52, Y94, I95, N155, E156, E238 and / or Y275 of SEQ ID NO: 1. The amino acids located in subsite +1 are amino acids A97, E98, W240, W243, F244, N245, R246, G279, D304, Y305 and / or D306 of SEQ ID NO: 1. The amino acids located in subsite +2 are amino acids W247, N287, G288, S290, L297, P298 and / or N457 of SEQ ID NO: 1.

[0026] In embodiments the substituted amino acid may comprise one or more of the following amino acid residues in subsite -1 : Y52, H58, Y94, I95, D129, N155, E156, G211 , E238 and / or Y275.

[0027] In embodiments the substituted amino acid may comprise one or more of the following amino acid residues in subsite +1 : N56, F244, N245, N252, G279, T281 , D304, Y305 and / or D306. In embodiments the substituted amino acid may comprise one or more of the following amino acid residues in subsite +2: R246, W247, N287, G288, S290, L297, P298, T383, L492 and / or G534.

[0028] In embodiments the substituted amino acid may comprise one or more of the following amino acid residues: Y52, N56, H58, Y94, I95, D129, N155, E156, G211 , E238, F244, N245, R246, W247, N252, Y275, G279, T281 , N287, G288, S290, L297, P298, D304, Y305, D306, T383, L492, and / or G534.

[0029] In embodiments the β-galactosidase comprises substituted amino acids for one or more of the following amino acids: E98, N155, E238, F244, Y275, Y305 and N457.

[0030] E238 (glutamic acid in sequence position 238) has been identified as a catalytic residue in the active site of the β-galactosidase enzyme and is involved in binding the substrate of β- galactosidase. F244 (phenylalanine) and Y305 (tyrosine) have been identified as conferring stability to reaction intermediates and help to create a hydrophobic pocket in the active site. F244 and Y305 are located in subsite +1 .

[0031] The Y305 amino acid may be substituted for any conserved or similar hydrophobic amino acids and / or polar amino acids, such as alanine, glycine, proline, cysteine, valine, isoleucine, leucine, methionine, phenylalanine, tryptophan, serine, threonine, asparagine and glutamine. The Y305 amino acid may be substituted for either alanine or phenylalanine. The inventors have identified that Y305 is important for transgalactosylation and that substituting tyrosine for alanine or phenylalanine resulted in a β-galactosidase able to produce GOS. Advantageously the inventors found that the aromatic ring of phenylalanine, due to its increased hydrophobicity compared to tyrosine, was able to stabilize a hydrophobic pocket in subsite +1 of β-galactosidase. A phenylalanine substitution of Y305 resulted in a β-galactosidase able to produce a greater variety of GOS compared to an alanine substitution at Y305. The longer hydrophobic side chain of phenylalanine may support elongation of the forming saccharide chain and stabilize oligosaccharide products with different linkages.

[0032] Advantageously the inventors have demonstrated that substituting Y305 for phenylalanine resulted in a β-galactosidase which synthesized higher order GOS with degrees of polymerization of DP3, DP4 and DP6. In addition, the GOS had pi → 4 and pi → 6 linkages which are known predictors of the potential to exert a prebiotic effect on administration to a subject. The inventors have shown the importance of Y305, in stabilizing the intermediate and subsequent synthetic products of the reaction.

[0033] The E238 amino acid may be substituted for any amino acid that weakens the H-binding capacity of the β-galactosidase. In embodiments the E238 amino acid is substituted for valine. The inventors have shown that substituting E238 for valine results in a hydrolytic β- galactosidase. Advantageously the substitution to valine resulted in a higher hydrolysis rate. Possibly this may be a result of a reduced H-bonding capacity for a β-galactosidase with a valine substituted for E238.

[0034] The presence of more than one mutation may combine the benefits in transgalactosylation by the Y305 substitutions with the higher hydrolysis rate of the E238 substitution.

[0035] Preferably the described β-galactosidase has an improved hydrolysis rate and / or transgalactosylation rate compared to the β-galactosidase of SEQ ID NO: 1. β- galactosidases are glycoside hydrolase (GH) enzymes that catalyse the hydrolysis of β- galactosides into monosaccharides through the breaking of a glycosidic bond formed between a galactose unit and glucose. They can also catalyse the transgalactosylation of galactose moieties to produce oligosaccharides. The inventors substituted amino acids in the active site of β-galactosidase to increase hydrolysis and / or transgalactosylation compared to the activity of the β-galactosidase of SEQ ID NO: 1 .

[0036] Preferably the described β-galactosidase can also synthesize GOS of a range of degrees of polymerization and with a range of glycosidic linkages which differ in amount and variety from the GOS produced by the β-galactosidase of SEQ ID NO: 1.

[0037] In embodiments the composition may comprise GOS which comprise a mixture of GOS of 3, 4 or 5 degrees of polymerisation. In embodiments the composition may further comprise GOS with degrees of polymerisation of greater than 5. GOS mixtures may vary in saccharide length (degree of polymerization, DPs) and regiochemistry (the bond type / linkage). These differences can determine the prebiotic effects of GOS and their selective fermentation by specific probiotic bacteria. Different bacteria in the gut ferment different sugars, so the prebiotic effect of different GOS can be specific to different microbial targets in the gut. It is known that different probiotic strains produce different GOS mixtures with differing degrees of polymerisation and different linkages, β-galactosidase enzymes can produce GOS comprising β-(1 6), β-(1 — >4) an →d β- (1^3) linkages and that dependent on the source of the β-galactosidase the ratios of these linkages in the GOS they synthesise will vary. These different glycosidic linkages and degrees of polymerisation of the GOS affect the prebiotic effect they elicit when administered to a subject.

[0038] The inventors have demonstrated that by making amino acid substitutions in the active site of a β-galactosidase enzyme GOS of different lengths, varying degrees of polymerisation and with different bond types have been produced. The production of different GOS is beneficial as an increased range of intestinal bacteria can utilise these GOS mixtures and elicit positive therapeutic benefits.

[0039] The composition may comprise a mixture of one or more disaccharides, trisaccharides, tetrasaccharides and / or pentasaccharides. The composition may comprise %w / v disaccharides, %w / v trisaccharides, %w / v tetrasaccharides and / or %w / v pentasaccharides. The composition may comprise relative amounts of disaccharides, trisaccharides, tetrasaccharides and / or pentasaccharides according to the effects desired in a particular subject.

[0040] The composition may comprise GOS which comprise β-(1 →3), P(1-4) and / or P(1-6) glycosidic linkages. GOS with different bond types can be utilised by different bacteria including different probiotic bacteria. GOS with different bond types may be utilised by different bacteria located in the gut, possibly the gut of a mammal such as a human. The utilisation or fermentation of GOS with differing bond types by different gut bacteria may elicit beneficial effects in the gut such as an increase in beneficial bacteria and improved intestinal and digestive function.

[0041] It has been shown that using GOS comprising β-(1 3) linkage →s and lower amounts of β- (1 →4) and β-(1 →6), resulted in higher bifidogenic effects compared to a GOS mixture produced by Bacillus circulans, which consisted mainly of β-(1 4) and β-(1 →6) linkages. → GOS of a high degree of polymerization is generally desirable as these potentially will exert a prebiotic effect in the more distal regions of the colon. The β-galactosidase may be derived from either fungi, yeast, bacteria, animals or plants. The β-galactosidase may be derived from Lactobacillus rhamnosus GR1 (LGR-1). LGR-1 is a widely used probiotic and has been shown to have therapeutic benefits for a healthy vaginal microbiome, urinary tract, microbiome and cardiovascular benefits. The inventors have shown that the wild-type β-galactosidase of LGR-1 was poorly active when tested for GOS production and unstable at high temperatures. SEQ ID NO:1 corresponds to the β- galactosidase amino acid sequence of LGR-1.

[0042] In embodiments the composition may have different uses. For instance, the composition may be for use as a prebiotic composition. The composition may be for use as a dietary supplement. The composition may be for use in maintaining or improving a healthy intestinal microbiome in a subject. The composition may be for use in improving bowel function. The composition may be for use as a medicament.

[0043] The composition may be for use in feeding microbial strains in the gut of a subject, for instance in a human gut. The microbial strains may be able to elicit beneficial effects, such as prebiotic effects, through the fermentation of the GOS of the composition of the present invention. For instance, the composition of the present invention may be for use in feeding Bifidobacterium and / or Lactobacillus microbial strains. The microbial strains may be probiotic microbial strains.

[0044] The composition may be in the form of a solid or liquid dosage form. For instance, the composition may be in the form of a liquid, tablet or capsule. The composition may be in the form of a freeze-dried powder, syrup or pastille.

[0045] The composition may be formulated for a particular administration route. For example, formulations suitable for oral administration include tablets, lozenges, hard or soft capsules, aqueous or oily suspensions, emulsions, dispersible powders or granules, syrups or elixirs. Suitable formulations for topical use include, for example, creams, ointments, gels, or aqueous or oily solutions or suspensions. Suitable formulations for inhalation include, for example, as a fine powder or a liquid aerosol. Suitable formulations for administration by insufflation include, for example, a fine powder. Suitable formulations for parenteral administration include, for example, a sterile aqueous or oily solution for intravenous, subcutaneous, intramuscular or intramuscular dosing or as a suppository for rectal dosing.

[0046] The composition for oral administration may further comprise one or more colourings, sweeteners, flavourings and / or preservative agents. The composition may further comprise a pharmaceutically acceptable carrier, diluent, enhancer or excipient. The composition may further comprise an excipient or carrier compound to modify the release profile of the GOS of the composition through the intestinal environment. The excipient or carrier compound may modify the release profile of the GOS mainly in the small intestine and / or the large intestine.

[0047] The composition may be in the form of a food stuff or food additive.

[0048] The composition may further comprise one or more active ingredients selected from: vitamins, minerals, phytochemicals, antioxidants, and combinations thereof.

[0049] The composition may further comprise one or more probitotics and / or prebiotics.

[0050] In another aspect of the invention there is provided a use of the composition as described in any of the embodiments above in improving intestinal health in a subject. The composition may be part of dairy products, beverages, infant foods, cereals, biscuits, confectionary, cakes, food supplements, dietary supplements, animal feeds, poultry feeds, therapeutic composition or any other food or beverage. The subject may be a mammal for instance a human.

[0051] In another aspect of the invention there is provided a method for improving the intestinal health of a subject comprising administering to a subject an effective amount of any of the compositions described in the embodiments above. The subject may be a human subject.

[0052] The present imnvention also provides a kit comprising the composition comprising the GOS as described herein.

[0053] It will be appreciated that numerous modifications to the above described embodiments may be made without departing from the scope of the invention as defined in the appended claims.

[0054] Figures

[0055] The invention is further described in the following non-limiting figures.

[0056] Figure 1 shows the HPLC trace of Mut3 (Y305F) GOS products. Peak 7 corresponds to the substrate lactose, while peak 1 and 3 indicate the presence of sugars product with DPs between 3 and 5.

[0057] Figure 2 shows the Dionex IC trace of Mut3 (Y305F) GOS products. Peaks 1 , 2 and 3 indicate the hydrolysis products (glucose and galactose) and starting substrate (lactose), respectively. A maltodextrin mixture was used as a standard to identify different sugars products. Peaks 6, 7 and 8 correspond to DP3 GOS, peaks 9-12 correspond to DP4 GOS, peaks 13-16 correspond to DP5 GOS and peaks from 17 to 21 indicate the presence of higher DP GOS.

[0058] Figure 3 shows the Dionex IC trace of Mut4 (Y305A) GOS products. Peaks 1 , 2 and 3 indicate the hydrolysis products (glucose and galactose) and starting substrate (lactose), respectively. A maltodextrin mixture was used as a standard to identify different sugar products. Peaks 8-10 correspond to DP4 GOS and peaks 13 and 14 indicate the presence of higher DP GOS.

[0059] Figure 4 shows a Bgal_LR 3D structure. On the left, the Bgal_LR structure obtained using X-ray crystallography. It shows a three domain structure organised and shown here in a dimeric form. Domain 3 has an alpha-beta TIM barrel architecture, where the active site is located. On the right, a close up image of the active site with a galactose moiety shown bound at the active site. The residues involved in the enzymatic activity are visualised as wires and coloured in black and grey.

[0060] Figure 5 shows Bgal_LR active site. Detailed image of the active site with key residues lining the subsite binding pocket (labelled SS-1 , SS+1 and SS+2, according to where a sugar moiety would occupy during hydrolysis and synthesis). These active site residues are responsible for the dual action of the enzyme; hydrolysis and synthesis of GOS.

[0061] Figure 6 is a 1 D / 1 D NEOSY 1 H NMR spectrum, visualised using TopSpin, for Mut 3 (Y305F) GOS mixture, highlighting the presence of higher DP GOS showing likely beta-(1-4) and beta-(1-6) linkages. Peak assignments correspond to those presented in Leeuwan et al

[0033] ,

[0062] Detailed description

[0063] Examples

[0064] Methods

[0065] Bioinformatics

[0066] The L. rhamnosus β-galactosidase (Bgal_Lr) sequence, retrieved from the NCBI database (Accession code WP_064522546.1), was used to build three-dimensional homology models by IntFOLD and Phyre2. The top 5 models arising from each method were chosen for further analysis and final selection of a single top model for further analysis, based on the following criteria: highest sequence coverage, lowest scoring function, highest confidence level, and highest resolution of the template structure.

[0067] Plasmid design and protein cloning

[0068] Genome analysis from different Lactobacilli highlighted the presence of only one potential P- galactosidase from Lactobacillus rhamnosus GR-1. The selected sequence was reverse transcribed and cloned into pET28a-HisTag plasmid vector by Genscript Biotech Corp (Piscataway NJ). L. rhamnosus GR-1 β-galactosidase mutations were designed using Snapgene 18. For each novel, mutant β-galactosidase, the amino acid sequence was cloned into different expression vectors by Genscript. Each vector is a pET expression vector containing an antibiotic resistance marker, a His6-tag and a cleavage site between the tag and the protein sequence. For vector used see Table 1 .

[0069] TABLE 1 : Vector construct used for protein expression in E. coli competent cells.

[0070] Transformation and over-expression

[0071] The native and mutant constructs were used to transform E. coli competent cell lines (BL21 (DE3) pLysS and Tuner (DE3) for expression trials. For expression trials, 2 pL of each plasmid was used to transform an aliquot of competent cells. Transformed bacteria were plated in agar plates with 15 pg / ml kanamycin and 30 pg / mL chloramphenicol. From each plate, one colony was inoculated in 10 mL of LB with 15 mg / mL of kanamycin and grown overnight at 37°C. After the overnight culture, 1 mL of bacterial culture was inoculated in 50 ml of fresh LB with 15 pg / mL kanamycin and grown at 37°C. When the culture reached ODeoo 0.4 - 0.6, after approximately 2h, it was induced with 1 mM IPTG and the bacterial culture was left to grow, while shaking (250rpm), overnight at different temperatures (20°C, 25°C, 30°C). After the induction, 1 mL of the sample, at each temperature, was collected every hour and then pelleted by centrifugation at 10,000 rpm for 10 minutes to follow cell growth and protein expression. In each case, the cell pellet was then lysed with B-PER lysis buffer and an aliquot of 15 pL was taken to measure the presence of the protein in the insoluble phase using SDS-PAGE. After one hour, the whole bacterial lysate, for each sample, was centrifuged at 16,000 rpm for 20 min to separate the soluble (supernatant) and insoluble (cell pellet) fractions. Protein expression was monitored and assessed using SDS- PAGE.

[0072] Protein purification

[0073] Protein expression was scaled up, 1 L culture in a 2 L flask, to obtain a higher yield of the bacterial pellet (~10 g). The pellet was lysed with homemade lysis buffer (100 mM NaCI, 25 mM TRIS, 10 mM Imidazole for each gram of wet pellet, plus 200 pL lysozyme and 1 EDTA- free protease inhibitor tablet) and further lysed using a cell disruptor (Constant Systems) at 30 kpsi for 4 cycles. The protein of interest was purified with affinity chromatography using a Ni-NTA His-trap column (Cytivia), followed by size exclusion chromatography (SEC) using an AKTA-FPLC system.

[0074] Protein crystallisation and structure determination

[0075] In all cases, purified protein was concentrated using Amicon concentrators (50 mL, 50 kDa cutoff) to obtain approximately 12 mg / ml protein. Initial crystallization screening was performed using the sitting-drop vapour-diffusion method, with commercial crystallization screening kits from Molecular Dimensions Ltd (Table 2) with MRC 2 Lens Crystallisation Plate (SWISSCI), using an Oryx8 robot (Douglas Instruments) in 96-well Greiner plates using 200 nL drops.

[0076] TABLE 2: Commercial crystallization screening kits utilized.

[0077] The optimal conditions found, from the initial screening, were replicated in 24 well-plates, using the hanging-drop vapour diffusion technique, with variable protein / reservoir solution ratios (1 :1 , 1 :0.5, 0.5:1) applied on a coverslip and suspended on top of 400 pL reservoir solution, at room temperature. The crystals, undisturbed, typically grew over a period of 5 days. For co-crystallization with ligands (lactose 200 mg / mL, lactulose 200 mg / ml, galactose 200 mg / mL) crystallization was performed with 1 pL well solution, 1 pL ligand at known concentration and 1 pL protein solution. Any resulting crystals were flash frozen and sent to Diamond Light Source for X-Ray crystallography.

[0078] Functional assay

[0079] Protein activity was tested using a standard oNPG colorimetric test (from Thermo Scientific™). In brief, 40 pl of pure protein sample (at 10mg / mL) or 40 pL of crude lysate were added to 100 pl of 1 M sodium phosphate buffer (pH 6.8) and 250 pL of 20 mM oNPG and 10 pL of 500 mM MgCh. The protein activity was assessed by the evaluation of the solution turning from transparent to yellow in colour, due to the hydrolysis of the substrate and release of the o-nitrophenol product, which is yellow.

[0080] To test the ability of the protein to hydrolyse lactose, a GOPOD glucose test kit (Megazyme) was used, as follows; 500 mg / mL lactose was dissolved in sodium phosphate buffer pH 6.5 and left at40°C for 5 hours after which 0.1 mLwas collected and added to 3.0 mL of GOPOD reagents and incubated at 40-50°C for 20 min. The amount of glucose produced was estimated by reading the absorbance at 510 nm against the reagent blank.

[0081] GOS synthesis and optimisation

[0082] Once protein activity and stability were established, GOS production was tested, using lactose monohydrate 200 mg / mL as a substrate. Three main factors can affect lactose hydrolysis and transgalactosylation: pH, temperature and reaction time. And, in the case of P-galactosidase enzymes, it is known that magnesium ions can further enhance enzyme activation. For this reason, GOS production was optimized under different conditions with variable temperatures, buffers and reaction time, as described in Table 3.

[0083] TABLE 3: GOS production optimization.

[0084] For each reaction, 40 pL of protein was mixed with 10 pL of MgCh 10 mM and 950 pL of appropriate buffer with 2 mg / mL lactose and executed at different temperatures. At discrete time points, 1 mL of the sample was collected. The reaction was stopped with a high temperature (97° C for 3 minutes), filtered through a syringe with 0.2 pm sterile filters and stored in a glass 1 mL vial at 4° C until further analysis.

[0085] Carbohydrate analysis

[0086] Serial dilutions were made from the original samples, from 1 :2 to 1 :1000. Enzyme hydrolytic and transgalactosidic activity were tested in terms of carbohydrate yield and profile, using the methods described below.

[0087] HPLC analysis

[0088] The resulting mono- and oligosaccharide products were analysed using an HPLC system; Hi-Plex Na (Octo), 7.7 x 300 mm, 8 pm and Hi-Plex Na, 7.7 x 300 mm, 10 pm HPLC column (Agilent Technologies). HPLC was run with water as the eluent, at a flow rate of 0.4 ml / min for 30 min. The resulting peaks from the HPLC trace were evaluated by comparison with standard solutions under the same conditions (lactose 500 mg / mL in phosphate buffer, glucose 500 mg / ml, galactose 500 mg / mL and GOS standard mixture).

[0089] HPAEC-PAD analysis

[0090] Samples were analysed using high pressure anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) using a Thermo Scientific ICS-5000+ system; separation was achieved using a PA200-4mm column (Thermo Scientific) with a sodium acetate / sodium hydroxide gradient. Standard solutions of maltodextrin (DP1-7), galactose, lactose and GOS mixture were injected as retention time markers and to ensure response variation was constant throughout the run. Samples solutions were diluted 1 :10 fold with Milli-Q water.

[0091] NMR analysis

[0092] Selected samples, previously analysed with Dionex, were used for GOS characterization by NMR spectroscopy (outsourced to Reading Scientific Services Ltd). One-dimensional 1 H NOESY recorded at a probe temperature of 298 K on a Bruker Neo 600 MHz spectrometer. Samples were dissolved in D2O, containing 3-(trimethylsilyl)propionic- 2,2,3,3-d4 acid, sodium salt, which was used as an internal standard and set to 0 ppm.

[0093] 1 D 1 H NOESY spectra were recorded with 12,500 Hz spectral width at 65k complex data points, using a noesygpprld pulse to suppress the HOD signal. Spectra were processed using TopSpin software (Bruker).

[0094] Results

[0095] Sequence Analysis and Homology Modelling of β-galactosidase

[0096] The amino acid sequence encoding for β-galactosidase was retrieved from the Lactobacillus rhamnosus GR-1 genome sequence from NCBI (Accession code WP_064522546.1). The amino acid sequence is designated as “SEQ ID NO:1”. The β- galactosidase from Lactobacillus rhamnosus GR-1 is referred to here as Bgal_Lr.

[0097] Protein structure organization was studied using Pfam19 (PF01301) and CATH20, which showed that this protein belongs to a GH35 protein family group with the enzymatic ability to catalyse the hydrolysis of terminal non-reducing β-D-galactose residues in β-D- galactosides.

[0098] Three-dimensional homology models, using this sequence, were predicted by IntFOLD and Phyre2. The resulting output, for both methods, gave rise to 5 top homology models of interest with the top model from Phyre2 at 60% identity and 100% confidence being selected as the final homology model.

[0099] The closest structural homologue used as a template is a β-galactosidase from Streptococcus pneumoniae TIGR4 (PDB 4E8C). The structural alignment of this structure with the model produced with Phyre2, showed 100% confidence and 60% ID, with 99% coverage, allowing a valid preliminary structural characterization of Bgal_Lr (Figure 4). The fold of these enzymes is conserved and shows three domains: domain 1 has an alpha-beta TIM barrel structure where the active pocket is located, while domain 2 and domain 3 have a beta-sandwich structure. The active site is in a deep pocket in the TIM-barrel structure in domain 1 and the key residue consists of a Glu156, which is the nucleophile required for the hydrolytic reaction.

[0100] Protein Expression and Purification

[0101] The Bgal_Lr protein sequence was cloned into a pET-28a vector with a kanamycin resistance sequence and a His6-tag at the C-terminus. This protein construct was used to transform BL21-pLysS E. coli competent cells. Protein expression was tested at different temperatures (30°C, 25°C and 20°C) and samples were taken at different time points (1 h, 2h, 3h, 4h, 5h, 6h and overnight after induction) to assess optimal protein expression conditions.

[0102] Optimal protein expression was confirmed by SDS-PAGE and Western blot analyses.

[0103] The expression trials performed show that the optimal conditions to achieve soluble Bgal_Lr enzyme are 20°C, in BL21 pLysS cells, with overnight induction. Under these conditions, it was possible to scale-up protein expression (1 L culture in a 2L flask) and proceed with bacterial lysis, using the optimised lysis buffer (50 mM Tris pH 7.5, 500 mM NaCI, 30 mM Imidazole, 0.2% Tween20, 5% glycerol), 200pL lysozyme and 1 tablet of protease inhibitors EDTA-free. Additionally, cells were lysed using a cell disruptor.

[0104] The SDS-PAGE results indicate that a one-step purification strategy was not sufficient to achieve pure protein (to >90-95%, as desired). Thus 5 mL of protein was used in a subsequent SEC purification step.

[0105] The 2 mL fractions, from the SEC trace corresponding to the peak with the highest absorbance were collected and the effectiveness of the purification again was evaluated using SDS-PAGE analysis.

[0106] The fractions containing pure protein were pooled and concentrated (Amicon concentrators spin column 50 kDa cut-off), and the final protein concentration was evaluated using a NanoDrop One spectrophotometer. After ~12 mg / ml of concentrated protein was obtained, 100 pL of the sample was used to evaluate protein activity with oNPG. The results showed low hydrolytic activity when tested at room temperature for 2 hours.

[0107] Bgal_Lr mutants Due to the low activity of the wild-type Bgal_Lr, functional mutants were designed to increase hydrolysis and transgalactosylation by targeting residues in the active site of the enzyme.

[0108] Initially, 29 single point mutants were designed (Table 4) and expressed in E. coli Tuner competent cells, which were chosen because their lack of E. coli β-galactosidase allows the hydrolytic activity of the whole cell lysate to be tested directly and eliminates the need for initial protein purification to remove E. coli Bgal. This competent cell line was used to test the expression for each of the 29 mutants, at two different temperatures (25°C and 20°C), and samples were taken at different time points (2h, 4h, 6h, and overnight).

[0109] Single-point mutations have been designed targeted to the TIM-barrel domain, specifically involved in the enzymatic reaction. N155D, N155E, E238A, E238V, E238D and E238N, E238T, E238D, and E238Q were designed to increase substrate hydrolysis.

[0110] F244W, F244H, F244A, F244L were designed to lower galactose inhibition and increase substrate hydrolysis and support GOS synthesis, via the transgalactosylation reaction.

[0111] Y305F, Y305A, Y305S, Y305L, Y305V and Y305I were designed to increase transgalactosylation and GOS >DP3 synthesis.

[0112] Y275W, Y275I, Y275L, Y275V, N457F and N457Y were designed to target the hydrophobic pocket and facilitate the entrance of galactose moieties in the active site to support transgalactosylation.

[0113] Y305F / E238A, Y305F / E238N and Y305A / E238N double mutants were designed to increase, simultaneously, hydrolysis of the substrate and synthesis of the products.

[0114] I

[0115] TABLE 4 List of protein mutations and their corresponding constructs used for molecular cloning. “Mutant name” refers to the code used in the laboratory and herein for each mutant.

[0116] The results for the mutant proteins indicated that, under the conditions tested, in each case the mutant protein was highly expressed but in the insoluble fraction. Mut1 (E238V) and Mut5 (Y305S) were used as a model to test the solubility after mechanical lysis and purification. Mut5 (Y305S) expression was scaled-up (4 L of bacterial culture, split as 1 L across 4 x 2L flasks). The resulting cells were lysed with lysis buffer (50 mM Tris pH 7.5, 500 mM NaCI, 30 mM Imidazole, 0.2% Tween20, 5% glycerol), followed by cell disruption using cell disruptor, four times at 30 kpsi. This too resulted in a high yield of insoluble protein.

[0117] Mut1 (E238V) expression also was scaled-up in the same way and the resulting cells were lysed with lysis buffer and a cell disruptor, as for Mut5. After purification using Ni-NTA affinity on the AKTA- FPLC system, some purified protein was evident, and the corresponding fractions were run on SDS- PAGE, which, somewhat surprisingly, revealed relatively pure protein at the expected molecular weight.

[0118] Mut1 (E238V) enzymatic activity was tested with oNPG and GOPOD kit. Despite successful isolation of Mut1 , this mutant exhibited no activity using the oNPG colorimetric test and it was poorly active with GOPOD, when lactose hydrolysis was tested at 40°C for 12 hours.

[0119] Due to the insolubility of most of the mutants, when expressed in Tuner cells, a small-scale expression trial for all mutants was repeated using BL21 pLysS competent cells, as shown to be successful for the wild type Bgal_Lr protein. Mut2 (E238A) expression was tested first, and it appeared to grow optimally reaching an OD600 0.4-0.6 after only 4 hours, showing favourable protein expression at 25°C after overnight induction. These conditions were considered optimal and used to express the other mutants.

[0120] Mut2 (E238A) expression was scaled-up (1 L of culture in 2 L flasks). Cells were lysed with lysis buffer followed by mechanical disruption using a cell disruptor as for the wild-type enzyme and purified using a Ni-NTA pre-packed column on the AKTA-FPLC system, as before. This resulted in 1 mL of 15 mg / mL of pure protein.

[0121] 2 mL fractions containing the targeted protein were concentrated and pooled together to achieve a total of 5 mL protein solution, which was used in a subsequent SEC purification step.

[0122] These optimized conditions (25°C, induced for 5 hours, BL21-pLysS) for Mut2 were then used to express the remaining mutants on a large scale (1 L culture in 2L flasks). Mut3 (Y305F) expressed and purified using the optimized protocol, resulted in 2 mL of ~12 mg / mL pure protein, which was stored at -80°C for later use. The remaining Bgal_Lr mutants were successfully produced with the optimized method described above. An average of 4 L of bacterial culture were grown for each mutant, then the mutants were purified in the same manner as for the wild type and Mut2, from which most yielded an average of 1-2 mL of IQ- 15 mg / mL of pure protein. Results of the mutant protein expression and purification are shown in Table 5.

[0123] TABLE 5: Results obtained from Bgal_Lr mutants for expression and purification screening. Column 3 shows the different results obtained for hydrolytic activity and GOS production. The GOS and resulting synthesis products produced by Mut3 (Y305F) and Mut 4 (Y305A) were analysed using HPLC and Dionex.

[0124] Mut3 (Y305F) showed promising active hydrolytic activity with oNPG at 30°C, pH 6.8. The optimal conditions obtained for GOS production from Mut3 (Y305F) were found to be 30°C for 16 hours, pH 6.8. The GOS mixture obtained was analysed by HPLC (figure 1) and showed a peak corresponding to the lactose substrate (peak 7) and the presence of sugars with DPs of between 3 and 5 (peaks 1 and 3). Table 6 below shows the HPLC analysis of Mut3 (Y305F) showing GOS products corresponding to figure 1.

[0125] Table 6: HPLC analysis of Mut3 (Y305F) showing GOS products, corresponding to figure 1. Additionally, Mut4 (Y305A) showed hydrolytic activity when tested with oNPG, thus a GOS synthesis reaction was performed under optimized conditions, as for Mut3.

[0126] To gain a more detailed analysis of the resulting GOS profiles for Mut3 and Mut4, the same GOS mixtures were analysed by Dionex, outsourced to Reading Scientific Services Ltd (RSSL) (figures 2 and 3, and corresponding tables 7 and 8). The results showed the presence of lactose, glucose and galactose in every sample, indicating incomplete hydrolysis, which is not uncommon (peaks 1 , 2 and 3). A maltodextrin mixture was used as a standard to identify different sugar products.

[0127] Figure 2 showing the GOS produced by Mut3 (Y305F) show the presence of DP3 GOS (peaks 6, 7 and 8), DP4 GOS (peaks 9-12), DP5 GOS (peaks 13-16) and higher DP GOS

[0128] (peaks 17-21). Table 7 below shows the IC analysis of the GOS produced by Mut3 (Y305F).

[0129] Table 7: IC analysis of Mut3 (Y305F) GOS products, corresponding to figure 2.

[0130] Figure 3 shows the GOS produced by Mut 4 (Y305A) and has peaks corresponding to DP4 GOS (peaks 8-10) and higher DP GOS (peaks 13 and 14). Table 8 below shows the analysis of the GOS produced by Mut 4 (Y305A). Table 8: IC analysis of Mut4 (Y305A) GOS products, corresponding to figure 3.

[0131] Due to the uniqueness of the resulting GOS obtained from Mut3 (Y305F), this also was analysed by NMR (outsourced to Reading Scientific Services Ltd). The results obtained were visualised using TopSpin software and compared to the results available in the literature (Figure 6). Signals at ca. 3.4 ppm, between ca. 4.0-4.2 ppm and at ca. 4.5 ppm indicate the presence of GOS materials, despite high concentration of unreacted lactose.

[0132] The spectrum obtained from Mut3 GOS was compared to the Vivinal GOS analysis, as presented in Leeuwen’s paper

[0033] , This highlighted the presence of higher DP GOS with B(1 →4) and B (1 →6) linkages (Figure 6). Crystallisation and structure characterisation

[0133] The optimal conditions for crystallization of wild-type Bgal_Lr were tested, using the commercially available LMB and Wizard crystallization kits and an Oryx8 robot to take advantage of a 96-well high throughput approach. From crystallisation screening, a few protein crystal hits were obtained, under different conditions.

[0134] Crystals from LMB D7 were mounted and sent for data collection on beamline I04 at Diamond Light Source. Data collection yielded a dataset of resolution 2.2 A with 95.1% completeness. Since good diffraction data were obtained from only one crystal (15% PEG 3350 - 0.1 M MES - pH 6.2) these conditions were considered optimal and utilized for subsequent optimisation experiments.

[0135] In an effort to explore any potential conformational changes upon ligand binding, protein- ligand co-crystallisation screening was performed in 24-well plates (using 450 μL of reservoir solution). Different ratios of protein:lactose:reservoir solution were tried. Co- crystals were found within one week and were mounted and sent to the Diamond Light Source for data collection. Data collection yielded a dataset of resolution 2.2 A with 95.1 % completeness. The final ligand: protein structure was retrieved using molecular replacement and structure refinement was performed using PHENIX software, with iterative steps of manual model building using COOT software.

[0136] The same procedure, as for wild-type Bgal_Lr, was used for Mut2 (E238A) crystallization screening, resulting in a few crystals that were mounted and sent to a Diamond Light Source for data collection. A dataset of resolution 1.8 A with 95% completeness was achieved and used to solve the structure using molecular replacement.

[0137] Model building and structure analysis

[0138] The Bgal_Lr protein structure was solved by using the molecular replacement technique, using the homology model obtained from Phyre2. Model building iterated between manual model building and fitting in COOT and refinement in PHENIX using the Phenix, refine. A similar process was repeated for the co-crystallised and mutant proteins. Subsequent structures were solved using the newly acquired wild type Bgal_Lr structure as the molecular replacement model, with modelling and refinement as before.

[0139] The final Bgal_Lr crystal structure (Figure 4) confirms the protein as a heterodimer with a TIM barrel structure in domain 1 of the monomer, as predicted by the homology model from Phyre2 and as observed for the GH35 homologous proteins from Streptococcus sp and Bacillus circulans.

[0140] The protein-ligand bound structure, from the co-crystallisation experiment of Bgal_Lr with lactose, showed the presence of galactose in the active site (Figure 4, highlighting the likely hydrolysis of the substrate, lactose).

[0141] Discussion

[0142] Lactobacillus rhamnosus GR-1 is a probiotic bacterial strain, which is already recognised for its beneficial effects in the human host. The strain belongs to the Firmicutes bacteria phylum, which include one of the most abundant bacterial species that colonize the human gut. It was recently re-classified as a Lactiplantibacillus and is evolutionarily close to other Lactiplantibacilli present in the human microbiome, such as Lactiplantibacillus casei and Lactiplantibacillus paracasei.

[0143] From the L. rhamnosus GR-1 genome, available in the NCBI database, a genetic sequence corresponding to a β-galactosidase was identified and retrieved for this study. This protein was shown to belong to the GH35 enzyme family, from which similar enzymes have been isolated from different microorganisms, such as fungi, yeast, and bacteria and also within higher organisms such as animals and plants.

[0144] Depending on the structure, some GH35 enzymes have a preference towards β-1 ,3-, β-1 , 6- or β-1 ,4- galactoside linkages [1 ,2] showing optimal activity under acidic conditions. These enzymes are widely used in industry for lactose hydrolysis [6] and can potentially synthesise galactooligosaccharides [3-5],

[0145] However, according to the results obtained from this research, the wild type Bgal_Lr enzyme appeared to be poorly active when tested for GOS production, using lactose as substrate and under standard conditions. Its activity was tested with oNPG as a substrate in a simple colorimetric assay and reaction conditions were screened for optimisation under different temperatures and pHs. The enzyme was also unstable when tested under high temperatures, such as above 40°C. There were no signs of lactose hydrolysis or GOS production at 50° C and 60° C, suggesting the enzyme was no longer functional at these temperatures.

[0146] To improve the functionality of the wild-type Bgal_Lr enzyme, 29 mutants were designed utilising the wild-type Bgal_Lr structure, solved using X-ray crystallography. The newly solved Bgal_Lr structure is characterized by a TIM-barrel structure in domain 1 , and stabilized by two domains exhibiting a beta-sandwich fold. The active site sits slightly buried at the ‘top’ (open end) of the TIM barrel. According to other studies, the nucleophile residue is a glutamic acid in position 156 and the other catalytic residue is a glutamic acid in position 238. These residues are conserved in other GH35 enzymes, where they are revealed to have the same function in binding the substrate lactose, according to the structural characterization of these homologous proteins [6-12] and illustrated in the alignment between Bgal_Lr and similar proteins retrieved from the PDB server.

[0147] In addition to the catalytic residues, there are other important residues that contribute to the enzyme activity by binding and stabilising the substrate-protein complex, even temporarily for catalysis or more substantially to promote GOS synthesis (in the case of β- galactosidases). These residues can interact through hydrogen bonds, van der Waals and even hydrophobic (and TT-TT) type interactions. Specifically, these residues consist of Asn155, Glu156, Glu238, in Subsite -1 , Tyr52, Ile95 and Ala97, Glu98 in Subsite +1 , and Tyr305 in Subsite +1 , which show polar contacts with the galactose moiety when bound. A phenylalanine (Phe244) and a tyrosine (Tyr305) in Subsite +1 contribute to conferring stability to the reaction intermediates by creating a hydrophobic pocket. Many of these active site residues are conserved in other GH35 structures, previously characterized. Figure 5 shows the β-galactosidase active site with key residues lining the subsite binding pocket (labelled SS-1 , SS+1 and SS+2, Figure 5) according to where a sugar moiety would occupy during hydrolysis and synthesis. These residues play a role in the hydrolytic and transgalactosylic potential of β-galactosidase.

[0148] In all β-galactosidases, Glu238 (numbered here according to Bgal_Lr sequence; SEQ ID NO:1) is a fundamental residue for hydrolysis and, in this work, shown to be equally important for subsequent transgalactosylation. When mutated into a valine (E238V), some enzymatic activity is retained, while when mutated into an alanine (E238A) there was no hydrolytic activity observed. And, mutation to an aspartic acid (E238D), showed no observed changes in hydrolytic activity but there was no subsequent transgalactosylation activity observed.

[0149] Interestingly, residue Y305 was revealed to be important for transgalactosylation activity, whereas it does not appear to be crucial for hydrolysis. In the Bgal_Lr Mut3 (Y305F) of this work, the aromatic ring of the phenylalanine is rotated (compared to the tyrosine side chain in the wild type) and it is more hydrophobic than tyrosine which stabilise a hydrophobic pocket in subsite +1. In Mut4 (Y305A), on the other hand, when an alanine is substituted for the tyrosine, whereby the alanine has a very small hydrophobic side chain and the distance between this residue and the galactose moiety is more than 7 A, there is no chance of forming hydrogen bonds or hydrophobic interactions with sugar moiety. Despite the loss of hydrogen bond potential and the lack of van der Waals complementarity, both Mut3 (Y305F) and Mut4 (Y305A) appear to synthesise GOS under the conditions tested. However, Mut3 (Y305F) appears to produce a larger variety of GOS, with different degree of polymerisation (DP); DP3, DP4 and DP6 oligosaccharides, likely comprising pi 4 and → pi →6 linkages, according to the Dionex and NMR analyses performed for the first time on this mutant Bgal_Lr enzyme product. These observations highlight the importance of Y305, in stabilizing the intermediate and subsequent synthetic products of the reaction. The presence of phenylalanine with a larger hydrophobic side chain than alanine likely supports the elongation of the saccharide chain and stabilises longer oligosaccharide products with different linkages. Both mutations, however, as expected did not appear to increase the hydrolysis, as indicated by the large amount of lactose remaining (observed by Dionex), following the reaction, which ran for 16 hours.

[0150] GOS mixtures vary in saccharide length (degree of polymerization, DPs) and regiochemistry (the bond type / l inkage). These differences determine the prebiotic effects of GOS and the beneficial effects of GOS are due to their ability to be selectively fermented by specific probiotic bacteria. Different bacteria in the gut ferment different sugars, so the prebiotic effect can be specific to different targets

[0013] , Moreover, different probiotic strains produce different GOS mixtures. For example, Bifidobacterium bifidum NCIMB 41171 produces mainly DP2 and DP3 GOS, whereas the fraction of higher DP GOS is significantly lower

[0014] , Aspergillus oryzae forms mainly trisaccharides

[0015] , a small yield of tetrasaccharides and only traces of pentasaccharides

[0016] , Bacillus circulans has been reported to produce DP2 and DP5 pentasaccharides [17-19],

[0151] Many factors can influence GOS production. Studies on Bacillus circulans and Aspergillus oryzae show the effect of lactose concentration on GOS with DP>3 [8,20], The temperature of the reaction also plays an important role; higher temperatures showed the production of higher DP GOS in Bifidobacterium bifidum

[0014] and Aspergillus aquleatus

[0021] , Another factor influencing the resulting GOS mixture is the reaction time. Hydrolysis and transgalactosylation occur simultaneously but with different ratios throughout the course of the reaction duration; with the balance of the reaction leading toward transgalactosylation, until the GOS yield reaches a critical maximum, after which the GOS products become susceptible to hydrolysis, subsequently reducing the GOS final yield.

[0152] With regard to synthesis, DP3 is usually produced first and acts as an acceptor for molecules to form DP4, which then act as the acceptor to form DP5 oligosaccharides. As a result, at different times it is possible to have different yields of oligosaccharides. Further, the types of linkages formed between the GOS products depend on the enzyme source. For example, a β-galactosidase from A oryzae produces mainly β-(1 →6) linkages followed by β-(1 →3) and β-(1 →4)

[0015] , L. plantarum and L. sakei

[0022] GOS mixtures mainly contain P-(1 →6) linkages and a small amount of β-(1 →3). However, L. reuteri forms mainly β-(1 →6) and β-(1 →3) but not β-(1 →4)

[0023] B. circulans synthesized GOS products linked by β- (1—2), β-(1 →3) β-(1 →4) and β-(1 →6) glycosidic linkages

[0024] , B. bifidum NCIMB 41171 GOS products consist mainly of β-(1 →3) linkages and lower amounts of β-(1 — -4) and β- (1 →6)

[0025] , Importantly, these different glycosidic linkages and DP of GOS have been shown to affect the prebiotic outcome. In human trials, using GOS β-(1 →3) linkages and lower amounts of β-(1 — -4) and β-(1 →6), resulted in higher bifidogenic effects compared to a GOS mixture produced by B. circulans, which consisted mainly of β-(1 →4) and β-(1 →6) linkages

[0025] , GOS of a high degree of polymerization is generally desirable as these potentially will exert a prebiotic effect in the more distal regions of the colon

[0026] ,

[0153] Conclusion

[0154] Bgal_Lr from Lactobacillus rhamnosus GR-1 strain represents a newly characterised β- galactosidase enzyme that can be engineered using any mutagenesis approach, such as any protein engineering approach, to produce novel GOS mixtures with potential prebiotic properties. Despite the native protein having limited enzymatic activity, using protein engineering methods it is possible to mutate specific residues and improve the GOS production, via increased hydrolysis and supporting elongation of the oligosaccharide during synthesis. In fact, Mut3 (Y305F) exhibited a variety of desirable and surprising higher order GOS products (DP3, DP4 and DP6) with β1 → 4 and β1 → 6 linkages: a promising predictor of exerting a prebiotic effect.

[0155] The results highlight the particular importance of residues E238 and Y305 for affecting, respectively, hydrolysis and transgalactosylation.

[0156] References

[0157] 1. Gamauf, C. et al. Characterization of the bga1-encoded glycoside hydrolase family 35 β- galactosidase of Hypocrea jecorina with galacto-β-d-galactanase activity. FEBS J. 274, 1691-1700 (2007).

[0158] 2. Zhang, S. et al. Kinetic mechanism and characterization of human β-galactosidase precursor secreted by permanently transfected Chinese hamster ovary cells. Biochem. J. 304, 281-288 (1994).

[0159] 3. Vera, C., Guerrero, C., Conejeros, R. & lllanes, A. Synthesis of galacto- oligosaccharides by β-galactosidase from Aspergillus oryzae using partially dissolved and supersaturated solution of lactose. Enzyme Microb. Technol. 50, 188- 194 (2012).

[0160] 4. Ishikawa, K. et al. Crystal structure of β-galactosidase from Bacillus circulans ATCC 31382 (BgaD) and the construction of the thermophilic mutants. FEBS J. 282, 2540- 2552 (2015).

[0161] 5. Nguyen, T. L. & Kim, D. H. Genome-Wide Comparison Reveals a Probiotic Strain Lactococcus lactis WFLU12 Isolated from the Gastrointestinal Tract of Olive Flounder (Paralichthys olivaceus) Harboring Genes Supporting Probiotic Action. Mar. Drugs 2018, Vol. 16, Page 140 16, 140 (2018).

[0162] 6. Rico-Diaz, A. et al. Structural features of Aspergillus niger β-galactosidase define its activity against glycoside linkages. FEBS J. 284, 1815-1829 (2017).

[0163] 7. Rojas, A. L. et al. Crystal Structures of β-Galactosidase from Penicillium sp. and its Complex with Galactose. J. Mol. Biol. 343, 1281-1292 (2004).

[0164] 8. Henze, M. et al. Rational design of a glycosynthase by the crystal structure of β- galactosidase from Bacillus circulans (BgaC) and its use for the synthesis of N- acetyllactosamine type 1 glycan structures. J. Biotechnol. 191 , 78-85 (2014).

[0165] 9. Cheng, W. et al. Structural insights into the substrate specificity of Streptococcus pneumoniae P(1 ,3)-galactosidase BgaC. J. Biol. Chem. 287, 22910-22918 (2012).

[0166] 10. RCSB PDB - 3D3A: Crystal structure of a beta-galactosidase from Bacteroides thetaiotaomicron. https: / / www.rcsb.org / structure / 3D3A.

[0167] 11 . Maksimainen, M. M., Lampio, A., Mertanen, M., Turunen, O. & Rouvinen, J. The crystal structure of acidic β-galactosidase from Aspergillus oryzae. Int. J. Biol. Macromol. 60, 109-115 (2013). Schroder, S. P. et al. Towards broad spectrum activity-based glycosidase probes: synthesis and evaluation of deoxygenated cyclophellitol aziridines. Chem. Commun. 53, 12528-12531 (2017). Rabiu, B. A., Jay, A. J., Gibson, G. R. & Rastall, R. A. Synthesis and Fermentation Properties of Novel Galacto-Oligosaccharides by β-Galactosidases from Bifidobacterium Species. Appl. Environ. Microbiol. Q7, 2526-2530 (2001). Osman, A., Tzortzis, G., Rastall, R. A. & Charalampopoulos, D. BbglV Is an Important Bifidobacterium β-Galactosidase for the Synthesis of Prebiotic Galactooligosaccharides at High Temperatures. J. Agric. Food Chem. 60, 740-748 (2012). Urrutia, P. et al. Detailed analysis of galactooligosaccharides synthesis with β- galactosidase from Aspergillus oryzae. J. Agric. Food Chem. 61 , 1081-1087 (2013). Huerta, L. M., Vera, C., Guerrero, C., Wilson, L. & lllanes, A. Synthesis of galactooligosaccharides at very high lactose concentrations with immobilized β- galactosidases from Aspergillus oryzae. Process Biochem. 46, 245-252 (2011). Mozaffar, Z., Nakanishi, K. & Matsuno, R. Continuous production of galactooligosaccharides from lactose using immobilized β-galactosidase from Bacillus circulans. Appl. Microbiol. Biotechnol. 25, 224-228 (1986). Mozaffar, Z., Nakanishi, K., ... R. M.-A. and & 1984, undefined. Purification and Properties of β-Galactosidases from Bacillus circulans. Taylor Fr. 48, 3053-3061 (1984). Palai, T., Mitra, S. & Bhattacharya, P. K. Kinetics and design relation for enzymatic conversion of lactose into galacto-oligosaccharides using commercial grade β- galactosidase. J. Biosci. Bioeng. 114, 418-423 (2012). Gao, X., Wu, J. & Wu, D. Rational design of the beta-galactosidase from Aspergillus oryzae to improve galactooligosaccharide production. Food Chem. 286, 362-367 (2019). Cardelle-Cobas, A., Martinez-Villaluenga, C., Villamiel, M., Olano, A. & Corzo, N. Synthesis of Oligosaccharides Derived from Lactulose and Pectinex Ultra SP-L. J. Agric. Food Chem. 56, 3328-3333 (2008). Iqbal, S., Nguyen, T.-H., Nguyen, T. T., Maischberger, T. & Haltrich, D. β- Galactosidase from Lactobacillus plantarum WCFS1 : biochemical characterization and formation of prebiotic galacto-oligosaccharides. Carbohydr. Res. 345, 1408- 1416 (2010). Splechtna, B. et al. Production of Prebiotic Galacto-Oligosaccharides from Lactose Using β-Galactosidases from Lactobacillus reuteri. J. Agric. Food Chem. 54, 4999- 5006 (2006). Yanahira, S., Kobayashi, T., Suguri, T., Nakakoshi, M. & Miura, S. Formation ofOligosaccharides from Lactose by Bacillus circulans β-Galactosidase. academic.oup.com 59, 1021-1026 (2014). Depeint, F., Tzortzis, G., Vulevic, J., I’Anson, K. & Gibson, G. R. Prebiotic evaluation of a novel galactooligosaccharide mixture produced by the enzymatic activity of Bifidobacterium bifidum NCIMB 41171 , in healthy humans: A randomized, doubleblind, crossover, placebo-controlled intervention study. Am. J. Clin. Nutr. (2008) doi:10.1093 / ajcn / 87.3.785. Manning, T. S. & Gibson, G. R. Prebiotics. Best Pract. Res. Clin. Gastroenterol. 18, 287-298 (2004). Page, R.., et al. Shotgun crystallization strategy for structural genomics: an optimized two-tiered crystallization screen against the Thermotoga maritima proteome. Acta. Crystallogr. D Biol. Crystallogr. 59(Pt6): 1028-1037 (2003). Newman., J et al. Towards rationalization of crystallization screening for small- to medium-sized academic laboratories: the PACT / JCSG+ strategy. Acta. Crystallogr. D Biol. Crystallogr. 61(Pt10):1426-1431 (2005). McPherson., A. A comparison of salts for the crystallisation of macromolecules. Protein Science. 10(2): 418-22 (2001). Gorrec., F. Protein crystallization screens developed at the MRC Laboratory of Molecular Biology. 21(5):819-25 (2016). Fazio., V, Peat., T and Newman., J. A drunken search in crystallization space. Acta. Crystallogr. F Struct Biol. Commun. 70(Pt 10) 1303-11 (2014). Gilliland., G and Davies., D. Protein crystallization: the growth of large-scale single crystals. Methods Enzymol. 104:370-81 (1984). Van Leeuwen., S et al. 1H NMR analysis of lactose / β-galactosidase-derived Galactooligosaccharide components of Vivinal® GOS up to DP. Carbohydr. Res. 400, 59-73 (2014). Sequence information

Claims

Claims1. A composition comprising galactooligosaccharides (GOS) wherein the GOS are obtainable by transgalactosylation of lactose using a β-galactosidase comprising an amino acid sequence of SEQ ID NO: 1 or one having at least 60% homology to the amino acid sequence of SEQ ID NO: 1, wherein the β-galactosidase comprises one or more amino acid substitutions in its active site.

2. The composition of claim 1 wherein the β-galactosidase comprises one or more amino acid substitutions of any of the amino acids Y52, N56, H58, Y94, I95, A97, E98, D129, N155, E156, G211 , E238, W240, W243, F244, N245, R246, W247, N252, Y275, G279, T281 , N287, G288, S290, L297, P298, D304, Y305, D306, T383, N457, L492 and / or G534 of SEQ ID NO: 1.

3. The composition of claim 1 or claim 2 wherein the β-galactosidase comprises substituted amino acids for one or more of the following amino acids: E98, E238, F244, Y275 and Y305.

4. The composition of claim 3 wherein Y305 is substituted for either alanine or phenylalanine.

5. The composition of any previous claim wherein the GOS comprise a mixture of GOS of 3, 4 or 5 degrees of polymerisation.

6. The composition of claim 5 wherein the GOS further comprise GOS with degrees of polymerisation of greater than 5.

7. The composition of any previous claim comprising a mixture of one or more disaccharides, trisaccharides, tetrasaccharides and / or pentasaccharides.

8. The composition of any previous claim wherein the GOS comprise P(1 -4) and / or P(1 -6) glycosidic linkages.

9. The composition of any previous claim wherein the β-galactosidase is derived from either fungi, yeast, bacteria, animals or plants.

10. The composition of claim 9 wherein the β-galactosidase is derived from Lactobacillus species.

11. The composition of claim 10 wherein the β-galactosidase is derived from Lactobacillus rhamnosus GR1.

12. The composition of any previous claim for use as a prebiotic composition.

13. The composition according to claim 12 for use in feeding beneficial microbial strains in the human gut.

14. The composition of claim 13 for use in feeding Bifidobacterium and / or Lactobacillus microbial strains.

15. The composition of any preceding claim, wherein the composition is in the form of a solid or liquid dosage form.

16. The composition of claim 15 wherein the composition further comprises an excipient or carrier compound to modify the release profile of the GOS through the intestinal environment.

17. The composition of claim 15 or claim 16, wherein the composition further comprises one or more active ingredients selected from: vitamins, minerals, phytochemicals, antioxidants, and combinations thereof.

18. Use of the composition of any of claims 1 to 17 in improving intestinal health of a subject.

19. The use of claim 18 wherein the composition is part of dairy products, beverages, infant foods, cereals, biscuits, confectionary, cakes, food supplements, dietary supplements, animal feeds, poultry feeds, therapeutic composition or any other food or beverage.

20. A method for improving the intestinal health of a subject comprising administering to a subject an effective amount of any of the compositions of any of claims 1 to 16.21 . The method of claim 20 wherein the subject is a human subject.

Images