Stimulatory effects of novel glucosylated lactose derivatives GL34 on growth of selected gut bacteria

Previously we structurally characterized five glucosylated lactose derivatives (F1–F5) with a degree of polymerization (DP) of 3–4 (GL34), products of Lactobacillus reuteri glucansucrases, with lactose and sucrose as substrates. Here, we show that these GL34 compounds are largely resistant to the hydrolytic activities of common carbohydrate-degrading enzymes. Also, the ability of single strains of gut bacteria, bifidobacteria, lactobacilli, and commensal bacteria, to ferment the GL34 compounds was studied. Bifidobacteria clearly grew better on the GL34 mixture than lactobacilli and commensal bacteria. Lactobacilli and the commensal bacteria Escherichia coli Nissle and Bacteroides thetaiotaomicron only degraded the F2 compound α-D-Glcp-(1 → 2)-[β-D-Galp-(1 → 4)-]D-Glcp, constituting around 30% w/w of GL34. Bifidobacteria digested more than one compound from the GL34 mixture, varying with the specific strain tested. Bifidobacterium adolescentis was most effective, completely degrading four of the five GL34 compounds, leaving only one minor constituent. GL34 thus represents a novel oligosaccharide mixture with (potential) synbiotic properties towards B. adolescentis, synthesized from cheap and abundantly available lactose and sucrose. Electronic supplementary material The online version of this article (10.1007/s00253-018-9473-8) contains supplementary material, which is available to authorized users.


INTRODUCTION
The human gut microflora has drawn increasing attention in recent years. It constitutes a very interesting ecosystem that varies in density and functionality in the different gut compartments. 1 These complex ecosystems have a significant impact on host well-being. 2 Strongest interest is focused on understanding what factors cause variations in microbiota composition and how these gut bacteria modulate host health. 3 Our work aims to stimulate the growth of health-promoting probiotic gut bacteria by using newly synthesized non-digestible carbohydrates, i.e. prebiotic compounds.
According to the latest definition, a prebiotic is "a substrate that is selectively utilized by host microorganisms conferring a health benefit". 4 Recently, part of the definition was disputed, since selective stimulation of health-promoting species seems not exclusively necessary to confer health benefits. 5,6 Generally, prebiotics are carbohydrates that are not fully digested by the host. They are fermented by various commensal and health-beneficial gut bacteria, thus promoting their growth and activity which may confer health benefits upon the host. 7,8 To date, the most wellknown prebiotics, supported by good quality data, are human milk oligosaccharides (hMOS), 9 GOS (β-galacto-oligosaccharides), FOS (β-fructo-oligosaccharides), inulin and lactulose. 10,11 All of these prebiotics are also hydrolysed by brush border enzymes, but not completely. 12 Isomalto-/malto-polysaccharides (IMMP), 13,14 xylooligosaccharides (XOS), 15 resistant starch, 16 and soy oligosaccharides also are (emerging) prebiotic oligosaccharides, 17 although more data about their effects on gut health are still needed. Each of these prebiotic compounds may exert specific and selective effects on gut bacteria. The search for new and effective prebiotics combined with specific probiotics (synbiotics) is increasing rapidly. 18,19 Lactose-derived oligosaccharides attract much attention in view of their prebiotic potential. One example is GOS, which are synthesized from lactose by enzymatic trans-galactosylation using β-galactosidases, achieving a degree of polymerization between 3 and 10. 20 This prebiotic has been widely studied and shown to stimulate probiotic bacteria to various extents. 12,21,22 Another commercially available prebiotic in this group is lactosucrose which is hardly utilized by human digestive enzymes and has stimulatory effects on both lactobacilli and bifidobacteria. 23,24 Also the selective bifidogenic effect of 4'-galactosyl-kojibiose, corresponding to compound F2 in our GL34 mixture, 25 on Bifidobacterium breve 26M2 has been reported. 24 These results indicate that there are clear perspectives to further develop and expand this group of lactose-derived prebiotic oligosaccharides. We recently reported synthesis of a mixture of five novel lactose-derived oligosaccharides (F1-F5) using the L. reuteri glucansucrase enzymes Gtf180-ΔN and GtfA-ΔN. 25 Their structural characterization revealed the presence of various glycosidic linkages, (α1→2/3/4), with DP of 3 and 4 (Scheme 1). 25 Four out of these five structures were new and only F2 4'-galactosyl-kojibiose had been reported before. In this work, their resistance to degradation by common carbohydrate degrading enzymes was studied by in vitro incubations. Also the growth of pure cultures of common gut bacteria, including commensal and probiotic strains, on these novel compounds was evaluated and compared with well-known prebiotic mixtures (GOS and FOS). This study provides information about the selective stimulatory effects of these compounds (and glycosidic linkage types) on growth of probiotic bacteria. The GL34 mixture particularly stimulated growth of B.
adolescentis. This is also of interest from an industrial perspective, since these new oligosaccharides with very specific prebiotic effects are produced from low-cost lactose and sucrose, and may be an option for developing synbiotics.  26 The GOS/FOS mixture used is a 90/10 (w w -1 ) mixture of the purified TS0903 GOS and long-chain Inulin (lcInulin, Frutafit TEX, provided by SENSUS, Roosendaal, The Netherlands), also serving as a control for the current prebiotic formula added to infant nutrition. 21

Lactobacillus growth experiments
Lactobacilli were pre-cultured in MRS-medium (Oxoid, Basingstoke, UK) anaerobically (or by using the GasPak system (Becton, Dickinson and Company, Sparks, USA)) under an N 2 atmosphere for up to 2 days at 37°C. 27 Then, 1 ml samples of the pre-cultures were harvested by centrifugation (2,500 x g, 2 min). The bacterial pellets were washed twice with sterile 10% NaCl and diluted 25-fold in 2x mMRS (modified MRS-medium that does not contain a carbon source for Lactobacilli). 28 In separate tubes, carbohydrates were dissolved with Milli-Q water to 10 mg ml -1 and sterilized by filtration using 0.2 μm cellulose acetate filters (the

Intracellular and extracellular activity essays
After growth with GL34 as their only carbon source, the three tested Bifidobacterium strains were harvested by centrifugation at 10,000 x g for 15 min at room temperature. Culture supernatants were sterilized using 0. 45  sec cooling on ice in between. The cytoplasmic extracts were harvested by centrifugation at 10,000 x g for 5 min to remove cell wall fragments, and then concentrated to one-fifth of the initial volume using Amicon Ultra-4 units as above.
The concentrated cell free supernatants and cytoplasmic extracts (10 µg protein for each) were incubated separately with 5 mg mL -1 of the GL34 mixture. All reactions were performed in Milli-Q at 37 °C for 24 h. The progress of the reactions was followed by high-performance-anion-exchange chromatography (HPAEC).

High-pH Anion-Exchange Chromatography (HPAEC)
Samples were analyzed on an ICS-3000 workstation (Dionex, Amsterdam, the Netherlands) equipped with an ICS-3000 pulse amperometric detection (PAD) system and a CarboPac PA-1 column (250 x 2 mm; Dionex). The analytical separation was performed at a flow rate of 0.25 mL min -1 using a complex gradient of eluents A (100 mM NaOH); B (600 mM NaOAc in 100 mM NaOH); C (Milli-Q water); and D: 50 mM NaOAc. The gradient started with 10 % A, 85 % C, and 5 % D in 25 min to 40 % A, 10 % C, and 50 % D, followed by a 35-min gradient to 75 % A, 25 % B, directly followed by 5 min washing with 100 % B and reconditioning for 7 min with 10 % A, 85 % B, and 5 % D. External standards of lactose, glucose, fructose were used for calibration. For the determination of glucosylated lactose compounds with a degree of polymerization (DP) of 3, maltotriose was used as external standard.

Enzymatic hydrolysis of compounds in the GL34 mixture
The GL34 mixture of five compounds was synthesized using glucansucrase Gtf180-ΔN, decorating lactose with one or two glucose units from sucrose as donor substrate, also introducing different types of linkages (Pham et al. 2017). 25 GL34 contains three DP3 compounds and two DP4 compounds, i.e. F1 (4´-glc-lac): α-D- 25 Four types of glycosidic linkages thus occur in this mixture, namely (α1→2), (α1→3), (α1→4), and (β1→4). Only the F2 2-glc-lac compound had been described before. 31 The GL34 mixture also contains glucosyl residues linked (α1→3)/ (α1→4) to the galactosyl residue of the original lactose. In view of the novel composition of this mixture of glucosylated-lactose compounds we tested their resistance or sensitivity to hydrolysis with several commercially available enzymes. Following incubations with the porcine pancreas and Aspergillus oryzae α-amylases (Table   S1), the HPAEC profiles at time 0 and 24 h showed no degradation of the GL34 compounds ( Figure 1). Also various malto-oligosaccharide acting enzymes (αglucosidase, iso-amylase and pullulanase, Table S1) were tested for their ability to hydrolyze GL34 compounds. However, after 24 h incubation, no (monomeric or dimeric) products were detected in the reaction mixtures with these three enzymes ( Figure 1). None of these α-glucose cleaving enzymes thus was active on the GL34 compounds.
Subsequent incubation of the GL34 mixture with the β-galactosidase enzymes from A. oryzae and Kluyveromyces lactis (Table S1) however, did result in (some) hydrolysis. Fig. 1 shows that galactose and kojibiose (a glucose disaccharide with (α1→2)-linkage) were released during incubation with β-galactosidase, especially with the A. oryzae enzyme. Only the peak corresponding to F2 2-glc-lac disappeared, the only GL34 compound with a terminal galactosyl residue. We subsequently studied the utilization of these GL34 compounds for growth by (selected) common intestinal bacteria in more detail. Figure 1: HPAEC profiles of oligosaccharides in 1) the GL34 mixture (1 mg mL -1 , blank) and the hydrolysis products after incubation of GL34 with 2) α-amylase from Porcine; 3) αamylase from A. oryzae; 4) α-glucosidase from yeast; 5) iso-amylase from Pseudomonas sp.; 6) pullulanase type 1 from K. planticola; 7) β-galactosidase from A. oryzae and 8) βgalactosidase from K. lactis.

Fermentation of GL34 compounds by probiotic Bifidobacterium strains
The tested bifidobacterial strains displayed two or more growth phases (Figure 3 and The latter strain appeared to go through different lag phases, adapting to the different carbon sources in GL34, reaching maximal OD after 36 h of incubation ( Figure S1). The experiments were carried out in triplicate, and the average values are shown.

Fermentation of GL34 compounds by commensal gut bacteria
Also the ability of two selected commensal bacteria to grow on the GL34 mixture was studied. B. thetaiotaomicron is a Gram-negative anaerobic bacterium found dominantly in human distal intestinal microbiota. 32   We subsequently identified the specific GL34 compounds utilized by these strains, and products derived, also aiming to elucidate which hydrolytic enzyme activities are involved, with emphasis on β-galactosidases and α-glucosidases.

Hydrolytic activity of commensal bacteria and lactobacilli on the GL34 mixture
The GL34    Alg3 showed very low similarity in protein sequence (between 24 and 31 %) ( Table   S2). The observed accumulation of kojibiose in the growth medium of L.

Bifidobacterium breve DSM 20213
In case of B. breve, four of the five GL34 compounds (F1 4´-glc-lac, F3 3´-glc-lac, F4 4´,2-glc-lac and F5 3´,2-glc-lac) (partly) remained unutilized in culture supernatants after growth (Figure 7-3).  glucosidic linkages in pullulan and starch. 49 Also these 3 enzymes failed to cleave any compounds in the GL34 mixture. Incubations with β-galactosidase enzymes however did result in hydrolysis, but only the F2 2-glc-lac molecule disappeared. This is explained by the ability of these enzymes to catalyze hydrolysis of βglycosidic bonds between galactose and its organic moiety. The combined data thus shows that the GL34 compounds are (largely) resistant to hydrolysis by these common carbohydrate degrading enzymes (Figure 1).

There is abundant clinical evidence for the important roles of Bifidobacterium and
Lactobacillus species in the eco-physiology of the intestinal microbiota, 50 In conclusion, the GL34 mixture promotes growth of the tested bacteria to different extents. The bifidobacteria tested generally were better at degrading GL34 compounds than the lactobacilli and commensal bacteria. The stronger metabolic toolset of bifidobacteria in comparison with lactobacilli also has been observed when comparing their growth on human milk oligosaccharides and other prebiotic oligosaccharides as primary carbon source. 28,44,53,54 The GL34 mixture thus showed potential to shift microbiota composition by specifically stimulating growth of bifidobacteria, particularly B. adolescentis.
Four out of five compounds in this GL34 mixture exerted high and selective growth stimulatory effects towards health-beneficial probiotic bifidobacteria. The combination of monomer composition and linkage type clearly determines the fermentable properties of the GL34 compounds. Individual gut bacteria were able to utilize only specific compounds in the GL34 mixture. Synergistic activities between bacterial species thus are likely to be essential for the utilization of the whole GL34 mixture. In future work this will be studied in more detail e.g. by using faecal bacterial cultures. Only B. adolescentis was able to utilize almost all structures, providing a potential synbiotic combination.

Acknowledgements
The work was financially supported by the University of Groningen/Campus