2013 Volume 60 Issue 2 Pages 95-100
In vitro and in vivo studies have demonstrated the prebiotic potential of isomalto-oligosaccharides (IMO), comprising α-(1,6)-gluco-oligosaccharides and panose, which selectively stimulate the growth of probiotic bifidobacteria and lactobacilli. The protein machinery conferring the utilization of IMO by probiotics, however, remains vaguely described. We have used genomic, transcriptomic, enzymatic, and biophysical analyses to explore IMO utilization routes in probiotic lactobacilli and bifidobacteria as represented by Lactobacillus acidophilus NCFM and Bifidobacterium animalis subsp. lactis Bl-04, respectively. Utilization of IMO and malto-oligosaccharide (α-(1,4)-glucosides) appears to be linked both at the genetic and transcriptomic level in the acidophilus group lactobacilli as suggested by reverse transcriptase PCR (RT-PCR) and gene landscape analysis. Canonical intracellular GH13_31 glucan 1,6-α-glucosidases active on IMO longer than isomaltose occur widely in acidophilus group lactobacilli. Interestingly, however, isomaltose, isomaltulose and panose seem to be internalized through a phosphoenoyl pyruvate transferase system (PTS) and subsequently hydrolyzed by a GH4 6-phosphate-α-glucosidases based on whole genome microarray analysis. This sub-specificity within GH4 seems to be unique for lactobacilli mainly from the gut niche, as the sequences from this group segregate from characterized GH4 maltose-6-phosphate-α-glucosidases in other organisms. By comparison, IMO utilization in bifidobacteria is linked to soybean oligosaccharide utilization loci harboring GH36 α-galactosidases, GH13_31 oligo 1,6-α-glucosidases and a dual specificity ATP-binding cassette (ABC) transport system active in the uptake of both classes of α-(1,6)-glycosides. These data bring novel insight to advance our understanding of the basis of selectivity of IMO metabolism by important gut microbiome members.
The human gut hosts a highly diverse microbial community giving rise to the most densely populated ecological niche in nature.1) In recent years, accumulating evidence has substantiated the positive health impact conferred by probiotic gut microbiome members, mainly from the Lactobacillus and Bifidobacterium genera, regarding pathogen protection, allergic disorders, immunomodulation, certain gastro-intestinal disorders, bowel cancer and brain function.2) 3) 4) 5) Utilization of oligosaccharides that are non-digestible by the host has been recognized as an important attribute of probiotic action.6) 7) 8) 9) 10) Currently, only a few oligosaccharides that selectively stimulate the growth and activity of probiotic bacteria are sufficiently validated by both in vitro and human intervention studies to ascribe them prebiotic status e.g., β-galacto-oligosaccharides (GOS) or fructo-oligosaccharides (FOS),11) 12) but several others are emerging. Isomalto-oligosaccharides (IMO) comprising α-(1,6)-glucosides and panose (α-D-Glcp(1–6)-α-D-Glcp(1–4)-D-Glcp), are also potential prebiotics due to their resistance to human digestive enzymes with the exception of isomaltose (α-D-Glcp(1–6)-α-Glcp, IG2), which is hydrolyzed by the sucrase-isomaltase complex.13) Accordingly, the administration of IMO was shown to increase the counts of bifidobacteria in humans,14) 15) and lactobacilli in rats.16) A recent human intervention study in elderly individuals highlighted the prebiotic potential of IMO by showing a decrease in clostridial GIT commensals concomitant with an increase in both lactobacilli and bifidobacteria due to IMO supplementation.17) Moreover, improved bowel function and lower cholesterol levels were observed in the same study. Notably, the molecular basis for the utilization of IMO by probiotic bacteria has not been addressed in detail. This review sheds light on the uptake routes, enzymology and the genetics of IMO utilization in probiotic bacteria based on recent work on the clinically well established and commercially widely used probiotic strains Lactobacillus acidophilus NCFM18) 19) and Bifidobacterium animalis subsp. lactis Bl-04.20) 21)
Enzymology of IMO degradation. Canonical IMO degrading enzymes are assigned into glycoside hydrolase family 13 subfamily 31 (GH13_31)22) which has recently been expanded to include isomaltulose synthases (EC 5.4.99.11) according to the CAZy database classification (http://www.cazy.org/). The hydrolases of GH13_31 (EC 3.2.1.10) catalyze the release of α-(1,6) linked terminal non-reducing glucosyl moieties of IMO (and dextran for some enzymes) with the retention of anomeric configuration. These enzymes are divided into two specificities based on substrate size preference: 1) glucan 1,6-α-glucosidases (G16G) preferring IMO longer than IG2 and active on dextran and 2) oligo 1,6-α-glucosidases (O16G) inactive on dextran and preferring short chain IMO with highest activity on the trisaccharide isomaltotriose (IG3). Gram-positive streptococci are reported to produce 1,6-α-glucosidases displaying the G16G specificity, with the enzyme from S. mutans being the best characterized,23) 24) while only the O16G specificity is reported from Bacillus species.25) 26) 27)
Recently, the first IMO hydrolyzing enzyme from probiotic lactobacilli, the G16G from L. acidophilus NCFM (LaGH13_31), was structurally and biochemically characterized.28) The activity profile of LaGH13_31 was consistent with the predicted G16G specificity with highest catalytic efficiency on panose (Table. 1) . and the activity profile was similar to the homologue from S. mutans (57.6% sequence identity). LaGH13_31 was susceptible to product inhibition by glucose with Ki (4.0 ± 0.18 mM, using para-nitrophenyl α-D-glucopyranoside) being in the same range as the Km for the preferred substrate panose (Table. 1) . Interestingly, LaGH13_31 displayed 15-fold higher activity on dextran and 43 fold higher dextran/panose activity ratio as compared to the S. mutans homologue despite the high similarity of the active sites of two enzymes. A marked difference, however, is observed at subsite +2, where the three residues R212, D213 and N243 in the L. acidophilus NCFM enzyme are likely to form additional substrate contacts at this subsite (Fig. 1), and thus provide better anchoring of the polymeric substrate to the enzyme. The corresponding residues in the S. mutans G16G (V208, S209 and G239) are smaller and are unable to engage in polar interactions with the substrate (Fig. 1) The transglycosylation activity of the enzyme was also assessed by monitoring the rate of liberation of para-nitrophenol and glucose24) and significant transglycosylation activity was observed (JSTAGE Supplementary Material, Fig. S1), which is consistent with a high substrate affinity at the aglycone subsites.
The kinetic parameters of the glucan α-(1,6)-glucosidase (G16G) LaGH13_31 from L. acidophilus NCFM at 37℃ and pH 6.0.
Differences at substrate binding subsite +2 between the glucan 1,6-α-glucosidase from L. acidophilus NCFM (LaGH13_31) and the counterpart from Streptococcus mutans (SmGH13_31).
The figure shows a superimposition of the residues at subsite +2 between LaGH13_31 (PDB entry 4aie; green sticks and labels) and SmGH13_31 (PDB entry 2zid; grey sticks and labels). The catalytic residues of LaGH13_31 are shown in orange sticks, while the ligand isomaltotriose from the structure of SmGH13_31 is shown as grey sticks. The distance between the three LaGH13_31variant subsite +2 residues and the hydroxyl groups of C3 and C4 of the glucosyl moiety bound at subsite +2 are shown to illustrate possible polar contacts between these residues and the substrate at this subsite.
Occurrence of IMO degrading GH13 enzymes in probiotic bacteria. Multiple sequence alignments and phylogenetic analyses suggested that GH13_31 sequences that bear the 1,6-α-glucosidase signature occur widely in lactobacilli and enterococci, which belong to the same taxonomic order, but are lacking in the dominant commensal genus Bacteroides.28) Interestingly, an extracellular GH66 endo-dextranase from Bacteroides thetaiotaomicron VPI-5482 was recently shown to produce isomaltotetraose and longer IMO from dextran.29) Thus, some Bacteroides species produce a dextran PUL (polysaccharide utilization) system conferring extracellular dextran capture, degradation and subsequent uptake of large dextran fragments to the periplasm for further breakdown in a similar fashion to other polysaccharides.30) In addition to the GH66 gene, the putative dextran utilization gene cluster encodes a TonB sensing system, SUS (sugar utilization system) uptake components and a GH31 putative α-glucosidase, which is likely to have a similar role as GH13_31 enzymes in the hydrolysis of dextran degradation fragments in the periplasm. The physiological substrate for this dextran PUL in Bacteroides is likely to be capsular dextran polysaccharides produced by a number of lactic acid bacteria31) 32) that are either commensals in the GIT or are ingested with food.
Only a few gut adapted Clostridium difficile strains possess putative GH13_31 enzymes, but a closer analysis showed that theses sequences lack the important α-1,6 specificity signature comprising a valine residue following the catalytic nucleophile in region II,33) 34) which is consistent with the decrease of clostridia counts upon IMO administration in humans.17) The GH13_31 sequences from acidophilus cluster lactobacilli were all of the G16G specificity, whereas food niche plant material fermenters e.g., Lactobacillus plantarum, Lactobacillus pentosus and Lactobacillus brevis strains possess multiple GH13_31 sequences resembling both the G16G and O16G specificities. By comparison, bifidobacteria produce GH13 enzymes that display an O16G-like activity profile,35) 36) but which can be distinguished from GH13_31 O16G by an insertion of about ten amino acid residues in the β4→α4 loop C-terminal to the catalytic nucleophile in the TIM barrel domain.
Genes mediating the utilization of oligosaccharides in bacteria including probiotics are frequently arranged in functional operons comprising transport systems, hydrolases and transcriptional regulators, e.g., the FOS and soybean oligosaccharide (raffinose) operons in L. acidophilus NCFM.6) 37) We have surveyed the available genomes of mainly lactobacilli and bifidobacteria to bring insight into the organization of genetic loci conferring IMO utilization.
Organization of GH13_31 genes in lactobacilli. In lactobacilli, GH13_31 genes that mediate the utilization of IMO occur frequently within malto-oligosaccharide (α-(1,4)-gluco-oligosaccharide) utilization gene clusters encoding an ABC transport system, a LacI transcriptional regulator, a GH65 maltose phosphorylase38) (EC 2.4.1.8), a GH13_20 maltogenic α-amylase (EC 3.2.1.133) and an uncharacterized GH13 putative enzyme that shows sequence similarity to O16G, but which lacks the α-(1,6) activity motifs.28) This organization is observed in some acidophilus cluster members, e.g., Lactobacillus johnsonii ATCC_33200 (Fig. 2A) and Lactobacillus gasseri JV-V03, but also in other Lactobacillus species, e.g., Lactobacillus casei BL23 (Fig. 2B) as well as in other Lactobacillales, e.g., Enterococcus faecalis OG1RF (Fig. 2C) and S. mutans (not shown). Interestingly, the same organization is observed in other acidophilus cluster members for the malto-oligosaccharide gene cluster, but the G16G GH13_31 is relocated to another locus distant from any carbohydrate utilization genes, e.g., in L. acidophilus NCFM (Fig. 2D) and related strains of L. acidophilus ATCC_4796 as well as in closely related species e.g., Lactobacillus amylovorus GRL1118 and Lactobacillus crispatus ST1 (not shown). This relocation of the GH13_31 genes, which is likely to be a recent event, does not seem to impair growth on IMO or panose, which were shown to sustain the growth of L. acidophilus NCFM.28) 39) Remarkably, both the isolated locus encoding the GH13_31 and the genes within the malto-oligosaccharide operon were upregulated in response to growth on maltotetraose and an IMO mix as analyzed using semi-quantitative RT-PCR28) (Fig. 3)
Organization of representative loci conferring utilization of IMO in probiotic bacteria and related organisms.
Genes encoding transcriptional regulators are in light red, glycoside hydrolases in purple, GH13_31 1,6-α-glucosidases active on IMO are in yellow, GH4 isomaltulose-6-phosphate α-glucosidase is in cyan, ATP-binding cassette (ABC) transporters are in light grey, and the phosphotransferase transport system (PTS) is in dark grey. The gene landscape and locus tag numbers (inside the arrows) are from NCBI, and the annotation is based on BLAST analysis and information from Ref. 28), 38)and 40). A, L. johnsonii ATCC_33200; B, L. casei BL23; C, E. faecalis OG1RF; D-E, L. acidophilus NCFM; F, B. animalis subsp. lactis.
Semi-quantitative RT-PCR transcriptional analysis of L. acidophilus NCFM cells grown on glucose (black bars) as compared to cells grown on isomalto-oligosaccharides (white bars) and maltotetraose (grey bars).
The analysis was performed according to Ref. 28) and the relative fold change compared to cultures grown on glucose was compared for the genes with locus tag numbers annotated as follows: 0264, glucan 1,6-α-glucosidase; 0609, PTS IIA; 1689, isomaltose/isomaltulose-6-phosphate-glucosidase; 1866, malto-oligosaccharide ABC transporter solute binding protein; 1867, malto-oligosaccharide ABC transport system ATP-binding protein; 1872, GH13 enzyme of unknown function, and 2071, 16S r RNA gene as a control.
The genetic co-localization of α-(1,4)- and α-(1,6)-gluco-oligosaccharide utilization genes in several lactobacilli and the transcriptional co-regulation of these genes, even when they reside at distant loci (as in L. acidophilus NCFM), suggests that the metabolism of both α-(1,4)- and α-(1,6)-gluco-oligosaccharide is integrated. Possible advantages of this may involve the use of the same enzymes for common metabolic intermediates e.g., maltose phosphorolysis by the GH65 maltose phosphorylase, or the use of the ABC transport system for the uptake of IMO in addition to malto-oligosaccharides, but additional experimental work is necessary to corroborate this.
A new route for isomaltose uptake and hydrolysis in L. acidophilus NCFM. Recently, another route for the utilization of isomaltose (IG2) has been identified in L. acidophilus NCFM based on whole genome microarray analysis.40) Two loci were highly upregulated upon growth on IG2, one comprising genes for a phosphotransferase system (PTS) and an RpiR transcriptional regulator41) (LBA606-LBA609), whereas the other (LBA1689) encodes a GH4 6-phosphate-α-glucosidase that recognizes phosphorylated disaccharide substrates (Fig. 2E) (Fig. 4) This suggests that uptake of IG2 by the PTS transport system is followed by hydrolysis catalyzed by the specific GH4 enzyme (tentatively annotated as a maltose-6´-phosphate hydrolase). Interestingly the same genes were upregulated by the sucrose isomer isomaltulose (α-D-Glcp-(1,6)-D-Fruf) and the trisaccharide panose,40) indicating that these substrates are metabolized via the same route. The GH4 and the PTS system were also upregulated when L. acidophilus NCFM cells were grown on an IMO commercial mixture (Fig. 3), suggesting that four different loci including two transporters and two hydrolases are involved in the uptake and degradation of complex IMO mixtures including panose. These data represents the first evidence that both IG2 and isomaltulose are transported via the same PTS transporter and degraded by the same GH4 6-phosphate-α-glucosidase. Previously, maltose-6´-phosphate glucosidase activity has been demonstrated in several GH4 members e.g., Clostridium acetobutylicum ATCC 824.42) By comparison, the GH4 from L. casei ATCC 334 (shares 61% amino acid sequence identity with GH4 from L. acidophilus NCFM) has been shown to efficiently hydrolyze isomaltulose and the other sucrose isomers except sucrose per se.43) Importantly, maltose as opposed to sucrose isomers (except sucrose) failed to induce the GH4 and the PTS EIIA based on transcriptional and proteomic data, which prompted the conclusion that this pathway is specific for the uptake and degradation of sucrose isomers.43) The PTS system and the GH4 hydrolase in L. casei ATCC 334 reside on the same locus, designated as a sim (sucrose isomer metabolism) operon. By contrast, the GH4 and the PTS system are encoded by two different loci in L. acidophilus NCFM and related acidophilus complex lactobacilli with the exception of L. johnsonii (e.g., strains DPC 6026 and NCC 533). Despite these differences in the genetic organization, the two GH4 enzymes from L. casei and L. acidophilus NCFM form two adjacent and related clusters in the phylogenetic tree constructed from 100 related GH4 sequences, and they segregate from other characterized maltose-6´-phosphate-α-glucosidases from other organism (JSTAGE Supplementary Material, Fig. S2). This together with the similar transcriptional regulation on isomaltulose suggests that this pathway represents a versatile route for the uptake and metabolism of short IMO and sucrose isomers within mainly GIT niche-lactobacilli.
Volcano plot depicting the differential transcriptome of L. acidophilus NCFM cells grown on isomaltose (IG2) as compared to β-galacto-oligosaccharides (GOS).
Genes are shown as grey dots (∙), and significantly upregulated genes encoding oligosaccharide transport and metabolism are shown as black solid spheres (●). The figure is modified from Ref. 40).
Genes conferring the utilization of α-(1,6)-glucosides and galactosides cluster in a single operon in bifidobacteria. The organization of the O16G genes encoding the degradation of IMO is markedly different between bifidobacteria and lactobacilli. Bifidobacterial O16G genes occur almost invariantly together with an ABC transport system, a NagC type transcriptional regulator44) and a GH36 subfamily I (GH36_I) α-galactosidase gene45) (Fig. 2F). Interestingly, GH36_I enzymes encoded within this gene cluster catalyze the hydrolysis of the α-(1,6)-galactosides abundant in soybeans and shown to have a potential as prebiotics.46) A few Bifidobacterium breve and Bifidobacterium longum strains e.g., B. longum subsp. longum BBMN68 and B. longum subsp. infantis 157F, possess an additional O16G gene residing at different locus together with other ABC transporters. The rationale for this gene duplication is unclear, especially since the enzymes encoded by these O16G genes are almost identical, precluding large functional differences. Indeed, the two O16G from B. breve UCC2003 were shown to have a similar activity profiles, but with different temperature and stability optima.35) The prevalent organization of O16G and GH36_I genes is suggestive that both soybean α-(1,6)-galactosides and IMO α-(1,6)-glucosides are taken up via the same ABC transport system encoded within this operon. Very recently, we provided evidence to this hypothesis as the ABC transport system within this operon was shown to be significantly upregulated when B. animalis subsp. lactis Bl-04 cells were grown on either the soybean α-(1,6)-galactoside raffinose or isomaltose as measured using whole genome microarray transcriptional analysis (Andersen et al., unpublished data). We have also been able to confirm this, as the recombinant solute binding protein of this ABC system showed affinity towards both types of α-(1,6)-galactosides (Andersen et al., unpublished data). This is the first example of a carbohydrate specific ABC transporter with affinity for ligands varying in both saccharide composition and anomeric linkages. Finally, the presence of a gene encoding a GH36 subfamily II (GH36_II)45) putative α-galactosidase within this α-(1,6)-glycoside gene cluster in B. animalis subsp. lactis Bl-04 (Fig. 2F, locus tag number 1596) suggests that the same transporter may also be responsible for uptake of the substrates of this enzyme. Additional work is required to establish the specificity of this GH36_II and to investigate whether the substrates for this putative enzyme are also recognized by the solute binding protein of the ABC transporter.
Mette Pries is acknowledged for technical assistance and Anne Blicher for performing the amino acid analysis. Professor Haruhide Mori from Hokkaido University is acknowledged for valuable discussions on the determinants of 1,6-α-glucosidase activity. This work was supported by a grant from the Danish Strategic Research Council, Committee of Health and Nutrition to the project “Gene discovery and molecular interactions in prebiotics/probiotics systems. Focus on carbohydrate prebiotics”. DTU is acknowledged for PhD stipends for Marie S. Møller and Joakim M. Andersen and for H.C. Ørsted postdoctoral fellowship grant for Avishek Majumder. The North Carolina Dairy Foundation and DuPont are acknowledged for financial support.
