Enzymes Required for Maltodextrin Catabolism in Enterococcus faecalis Exhibit Novel Activities

The genes mmgT to GenBank accession no. EFT41964 of E. faecalis JH2-2 are located upstream from an operon which encodes an ABC transporter catalyzing the uptake of maltotetraose and longer maltooligosaccharides (Fig. 1) (9). Genes with sequences nearly identical to those of mmgT to EFT41964 are found in all E. faecalis strains for which the genome has been sequenced, and they are usually annotated as α-amylase genes. We therefore suspected that they might play a role in maltodextrin metabolism. Compared to growth on glucose and maltose, the wild-type strain grew similarly well on maltotriose but significantly slower on maltotetraose (Fig. 2A) and still slower on longer maltooligosaccharides and maltodextrin (data not shown). We subsequently constructed mutants of each of the three genes and tested the effects of the various mutations on the growth behavior on linear and cyclic α-1,4-linked maltooligosaccharides up to maltoheptaose and on maltodextrin. While the EFT41964 mutant grew normally on glucose and maltose, it did not grow at all on maltotriose and maltotetraose (Fig. 2B). In contrast to maltotetraose, but similar to maltose (8), maltotriose is taken up and phosphorylated by the PTS permease MalT. Intracellular maltotriose-6″-P is subsequently dephosphorylated to maltotriose by the phosphatase MapP (9), thus making it a substrate for EFT41964. Complementation of the mutant strain with the EFT41964 wild-type allele expressed from the agmatine-inducible aguB promoter restored growth on maltotriose and maltotetraose, confirming that EFT41964 is essential for their catabolism (Fig. 2C). Among the various linear maltooligosaccharides tested, the wild-type strain was not able to utilize the α-1,6-linked isomaltose and isomaltotriose (data not shown).

In order to determine the activity of EFT41964, we purified the His-tagged protein, as described in Materials and Methods, and incubated maltodextrin with the purified enzyme, which resulted in the production of a major compound comigrating with maltose during thin-layer chromatography and a minor product comigrating with glucose (Fig. 3, lane MD). We subsequently incubated the enzyme with maltose and maltooligosaccharides ranging from maltotriose to maltoheptaose. Thin-layer chromatography suggested that EFT41964 transforms even-numbered maltooligosaccharides into only maltose and odd-numbered maltooligosaccharides into maltose and glucose (Fig. 3, lanes G2 to G7). This assumption was confirmed by following the hydrolysis of maltotriose or maltotetraose in a two-step spectrophotometric assay coupled to the NAD-dependent oxidation of glucose, as described in Materials and Methods. When we used maltotriose as the substrate, we first followed the NAD-dependent oxidation of glucose until the reaction reached equilibrium, with almost no further increase of the absorption at 340 nm. We subsequently added E. faecalis maltose phosphorylase (8) and observed a second increase, which was due to the oxidation of glucose formed from maltose by the action of maltose phosphorylase (Fig. 4). The glucose dehydrogenase-coupled spectrophotometric assay with maltotriose revealed a K_m of 2.7 mM and a V_max of 82 μmol · min⁻¹ · mg of protein⁻¹, which corresponds to a k_cat (per second) of 91. While a similar K_m value (2.4 mM) was obtained with 4-nitrophenyl-α-maltoside as the substrate, the V_max was almost 2-fold higher (176 μmol · min⁻¹ · mg of protein⁻¹). In contrast, 4-nitrophenyl-α-glucoside is not a substrate of EFT41964. When the coupled spectrophotometric assay was carried out with maltotetraose, EFT41964 was not able to release glucose from the tetrasaccharide. An increase in the optical density at 340 nm (OD₃₄₀) was observed only when maltose phosphorylase was added, indicating that EFT41964 forms two molecules of maltose when it hydrolyzes maltotetraose. In addition, only very slight yellow coloration was observed with 4-nitrophenyl-α-maltotrioside, which is probably hydrolyzed to maltose and 4-nitrophenyl-α-glucoside (data not shown). These results unequivocally confirmed that EFT41964 functions as a maltose-producing α-1,4 maltodextrin hydrolase, and we called it MmdH.

MmdH also functions as an endoglucosidase, because it is able to hydrolyze the α-1,4 bonds in α- and β-cyclodextrins. Purified MmdH efficiently produces maltose and a small amount of glucose from α- and β-cyclodextrins (data not shown). The formation of glucose from β-cyclodextrin, which is composed of seven (odd-numbered) glucose residues, was not surprising. However, α-cyclodextrin is composed of six glucose residues and after linearization should have produced only maltose similar to maltohexaose (Fig. 3, lane G6). Why MmdH nevertheless forms glucose is presently unknown. Despite this activity, E. faecalis does not grow on β-cyclodextrin and grows only very slowly on α-cyclodextrin. The slow growth on α-cyclodextrin disappeared when mmdH was inactivated (data not shown). Interestingly, α- and β-cyclodextrins bind with high affinity to the maltodextrin binding protein of Lactobacillus casei and induce the closed active conformation (16). The slow growth of E. faecalis on α-cyclodextrin might therefore be due to slow utilization of the cyclic compound, but a slight contamination with linear maltodextrins cannot be excluded.

Interestingly, a malP mutant defective in maltose phosphorylase, which in E. faecalis is essential for maltose utilization (8, 14), was able to grow on maltotriose, although at a 3-fold-lower rate than the wild-type strain. A likely explanation for this result might be that MmdH hydrolyzes maltotriose to maltose and glucose, and glucose supports the slow growth of the malP mutant. The previously reported accumulation of maltose in a malP mutant incubated with labeled maltose (8) probably prevents more efficient growth.

Several bacteria were reported to utilize isopanose (17, 18), a linkage isomer of maltotriose, in which the reducing glucose residue is bound via its hydroxyl group at C-6 to the C-1 of the preceding glucose moiety (Fig. 5). In contrast to maltotriose, which is mainly taken up by the PTS permease MalT, isopanose is transported by E. faecalis via the MdxEFG-MsmX ABC transporter (9), and the wild-type strain grows on the triose. Interestingly, deletion of mmdH prevented growth on isopanose (Fig. 6). However, complementation of the mmdH mutant only partially restored growth on maltotriose (Fig. 2C) and did not restore growth at all on isopanose (data not shown). As described below, MmdH hydrolyzes isopanose about 6 times more slowly than maltotriose, which might explain why the complementation allowed slow growth on maltotriose but not on its linkage isomer isopanose.

Depending on whether MmdH hydrolyzes the α-1,4 or α-1,6 linkage in isopanose, glucose and either isomaltose or maltose will be formed, respectively (Fig. 5). Incubation of isopanose in the presence of purified MmdH followed by thin-layer chromatography suggested that similar to maltotriose, isopanose is also hydrolyzed to maltose and glucose and not to glucose and isomaltose (Fig. 7). This result suggested that MmdH hydrolyzes not only α-1,4, but also α-1,6 glucosidic linkages. In order to confirm this unexpected result, we carried out a coupled two-step spectrophotometric assay. An increase in OD₃₄₀ and therefore glucose oxidation occurred during both steps, similar to that shown in Fig. 4. MmdH must therefore release maltose from the nonreducing end of isopanose and thus function as a maltose-producing α-1,6 maltodextrin hydrolase (Fig. 5). The V_max of the hydrolysis of isopanose was about 6-fold slower (13 μmol · min⁻¹ · mg of protein⁻¹) and the K_m 2-fold higher (5.9 mM) than the values determined for maltotriose, which is probably partly responsible for the slightly slower growth of E. faecalis on isopanose than on maltotriose (compare Fig. 2A and 6).

A spectrophotometric assay revealed that MmdH also hydrolyzes maltotriose-6″-P, the product formed during the PTS-catalyzed transport of the triose, to glucose and probably maltose-6′-P (data not shown). However, the activity was 30-fold slower than with unphosphorylated maltotriose. In contrast, MmdH was not able to hydrolyze isomaltotriose (Fig. 7) or isomaltose (data not shown), which both contain only α-1,6 linkages. As mentioned before, E. faecalis is not able to grow on these two oligosaccharides.

Among the five maltotriose linkage isomers (19), E. faecalis also utilizes panose (20), in which the first two glucose residues at the nonreducing end are connected via an α-1,6 linkage (Fig. 5). Panose is taken up by the maltodextrin ABC transporter (9). In contrast to the wild-type strain, a mutant deleted for GenBank accession no. EFT41963 was not able to grow on this carbon source (Fig. 8). Complementation of the EFT41963 mutant with the wild-type gene restored growth on panose, confirming that EFT41963 is involved in the catabolism of the triose.

Depending on whether the α-1,6 or α-1,4 linkage is hydrolyzed, panose will be converted to glucose and either maltose or isomaltose, respectively (Fig. 5). In order to distinguish between the two possibilities, we purified EFT41963, as described in Materials and Methods. Incubation of panose in the presence of purified EFT41963, followed by thin-layer chromatography, suggested that the trisaccharide is hydrolyzed to maltose and glucose, thus indicating that EFT41963 hydrolyzes the α-1,6 linkage (Fig. 7). In order to confirm the presumed α-1,6-glucosidase activity, we carried out a coupled two-step spectrophotometric assay (Fig. 4). As observed for the hydrolysis of isopanose by MmdH, an increase in OD₃₄₀ and therefore glucose oxidation occurred during both steps (data not shown). EFT41963 therefore functions as a glucose-producing α-1,6-maltodextrin hydrolase, and we called it GmdH (for glucogenic maltodextrin hydrolase), because it releases glucose from the nonreducing end of panose (Fig. 5) and probably other maltooligosaccharides beginning with an α-1,6 linkage. The V_max of panose hydrolysis by GmdH (12 μmol · min⁻¹ · mg of protein⁻¹) and the K_m (6.2 mM) are very similar to the values determined for isopanose hydrolysis by MmdH, which probably explains why, similar to isopanose, panose is also less efficiently utilized by E. faecalis than maltotriose (compare Fig. 2A and 6 with Fig. 8).

Thin-layer chromatography indicated that GmdH slowly transforms isomaltose into glucose (data not shown) and isomaltotriose into glucose and isomaltose (Fig. 7), although E. faecalis cannot grow on these two compounds. The release of glucose from isomaltose and isomaltotriose by the enzyme GmdH was also confirmed by carrying out the glucose dehydrogenase-coupled spectrophotometric assay (data not shown). The formation of glucose from isomaltotriose was found to be 2.5 times faster than that from isomaltose. In contrast, GmdH was not able to detectably hydrolyze pullulan, a polysaccharide in which maltotriose units are connected via α-1,6 linkages (data not shown).

We also inactivated the third gene in the amylase operon but could not attribute a physiological role to the encoded protein, which is also annotated as α-amylase. The resulting mutant grew identically to the wild-type strain on all maltooligosaccharides tested (data not shown). We nevertheless purified MmgT (GenBank accession no. EFT41962 ), as described in Materials and Methods, and tested whether it might react with any of the maltooligosaccharides utilized by E. faecalis. Subsequent thin-layer chromatography revealed that while the enzyme did not react with maltose (data not shown), it disproportionated maltotriose into maltose and maltotetraose together with a small amount of maltopentaose (Fig. 9). The function of the enzyme therefore resembles that of amylomaltases, except that MmgT does not liberate glucose but transfers it to another maltotriose molecule. A similar glucosyl transfer reaction occurred when MmgT was incubated with maltotetraose. Thin-layer chromatography revealed that the tetrasaccharide is converted to a mixture of maltose, maltotriose, maltopentaose, maltohexaose, and even a small amount of maltoheptaose (Fig. 7). Again, no glucose could be detected among the reaction products. In order to confirm the observed combined α-1,4-glucosidase and glucosyltransferase activities of MmgT, we also carried out two-step spectrophotometric assays coupled to glucose dehydrogenase and either maltose phosphorylase or MmdH. When MmgT was incubated with maltotriose, no increase in the OD₃₄₀ occurred, confirming that the enzyme does not liberate glucose from the trisaccharide. However, when maltose phosphorylase was also present in the assay mixture, the OD₃₄₀ increased, confirming that MmgT liberates maltose from maltotriose and transfers the remaining glucose to another maltotriose molecule. When we carried out spectrophotometric assays with maltotetraose, again, MmgT did not release a detectable amount of glucose from the tetrasaccharide (data not shown). By including MmdH in the assay mixture, we could show that MmgT rather liberates maltotriose from maltotetraose probably by hydrolyzing the α-1,4 linkage at the reducing end of the tetrasaccharide (data not shown). The glucosyl residue probably remains bound to the enzyme before it is transferred to another maltotetraose molecule. MmgT did not react with panose and isopanose, the linkage isomers of maltotriose (data not shown), suggesting that the enzyme hydrolyzes only α-1,4-linked linear maltooligosaccharides.

The E. faecalis strains used in this study are listed in Table 1. Mutants were derived from a maltotetraose-positive clone isolated from our stock of strain JH2-2 (9), which is also called TX4000. E. faecalis strains were routinely grown at 37°C without shaking in 100-ml sealed bottles filled with 20 to 50 ml of Luria-Bertani (LB) medium (Difco, NJ, USA) containing 0.5% [wt/vol] glucose. Growth studies were carried out with carbon-depleted M17/MOPS (morpholinepropanesulfonic acid) medium (M17cc), which was prepared as previously described (30). It was supplemented with 0.3% (wt/vol) glucose, maltose, maltotriose, maltotetraose, maltopentaose, maltohexaose, maltoheptaose, maltodextrin, panose, or isopanose. Erythromycin and chloramphenicol were added, when appropriate, at concentrations of 150 μg · ml⁻¹ and 10 μg · ml⁻¹, respectively.

E. coli strains used for protein purification and cloning experiments were grown aerobically by gyratory shaking at 250 rpm in LB medium at 37°C and transformed by electroporation with a Gene Pulser apparatus (Bio-Rad Laboratories). Growth was followed by measuring the absorption at 600 nm in a Novaspec II spectrophotometer.

The E. faecalismmgT (EFT41962), gmdH (EFT41963), and mmdH (EFT41964) mutants were constructed by introducing two stop codons at the beginning of each open reading frame (ORF) in order to produce small truncated proteins. For amplification of the PCR fragments required for the various genes, we used the primer pairs listed in Table 2. The amplified DNA fragments were cloned into the vector pMAD (31). The resulting plasmids were inserted into the genome of E. faecalis strain JH2-2 by double crossing over. For the identification of transformed plasmids, we used the procedure described by Mokhtari et al. (8).

For complementation of the E. faecalisEFT41963 (gmdH) and EFT41964 (mmdH) mutants, the corresponding wild-type genes were cloned into the vector pAGEnt (32). This vector contains the gene aguR, which encodes the activator of the agmatine operon aguBDAC, and the aguB promoter. The expression of genes cloned into the multiple-cloning site, which follows the aguB promoter in pAGEnt, is therefore induced by agmatine. For complementation of the EFT41964 (mmdH) mutant, the wild-type allele was amplified by PCR, and the resulting product was digested with the restriction enzymes PstI and SpeI and then ligated into the pAGEnt vector cut with the same enzymes. The ligation mixture was used to transform the intermediary host Lactococcus lactis strain IL1403. For complementation of the EFT41963 (gmdH) mutant, the required pAGEnt-derived plasmid was constructed with the NEBuilder HiFi DNA assembly cloning kit (New England BioLabs), according to the supplier's recommendation.

The media used for the growth studies were inoculated with overnight cultures, which were started from the glycerol stocks kept at −20°C. Each overnight culture was washed once with one volume of physiological water, and the final OD at 600 nm was adjusted to 2.0. A 96-well plate was used, and each well was filled with 200 μl of M17cc medium supplemented with different sugars per well and inoculated with 2 μl of the washed overnight culture. Three drops of paraffin oil were added on top of the medium to avoid desiccation and to create anaerobic conditions. A model 680 microplate reader (Bio-Rad Laboratories) was used, and the temperature was kept at 37°C. The absorption at 595 nm was measured every 30 min over a total period of 12 h. Growth studies under all different conditions tested were carried out at least three times.

Maltotriose phosphorylated at the O-6 position of the glucose residue at the nonreducing end was prepared enzymatically using the α-glucoside-specific PTS present in palatinose-grown cells of Klebsiella pneumoniae (33).

In order to purify MmdH (EFT41964), GmdH (EFT41963), and MmgT (EFT41962 ) with an N-terminal His tag, we first amplified the corresponding genes by PCR using E. faecalis JH2-2 DNA as the template and the appropriate primer pair (R-62-SalI and D-62-BamHI, R-63-HindIII and D-63-BamHI, or R-64-HindIII and D-64-BglII in Table 2), creating restriction sites at the ends of the amplified DNA. We subsequently cut the amplified DNA fragments with the appropriate restriction enzymes and cloned them into the His tag expression vector pET-18a (Novagen) restricted with the same enzymes. After verification of the correct sequences of the amplified genes, the resulting plasmids were used to transform E. coli strain BL21(DE3). A transformant for each amplified gene was subsequently grown in 500 ml of LB medium supplemented with ampicillin until the medium reached an OD₅₀₀ of about 0.5. Expression of the different E. faecalis genes was subsequently induced by growing the cells for 3 h at 28°C in the presence of 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG). The cells were harvested by centrifugation, and the preparation of crude extracts and purification of the His-tagged proteins on nickel-nitrilotriacetic acid (Ni-NTA) columns under nondenaturing conditions were carried out as described in reference 34.

Maltose phosphorylase of E. faecalis strain JH2-2 is encoded by the gene with GenBank accession no. EFT41759. The His-tagged enzyme was overexpressed in E. coli from the expression vector pQE30 and purified as previously described (8).

In order to determine the enzymatic activities of the three purified enzymes (MmdH, GmdH, and MmgT), which are all annotated as α-glucosidases, 10 μl of a solution of various maltooligosaccharides (10 μg/ml) in 50 mM Tris-HCl (pH 7.4) was mixed with 10 μl of a solution containing a concentration of 1 μg/ml of one of the enzymes. The assay mixture, which also contained 1 mM CaCl₂, was incubated for 3 h at 37°C. Aliquots of 5 μl were spotted on thin-layer plates, and the products formed from the various maltooligosaccharides were subsequently separated by thin-layer chromatography on silica gel glass plates (Kieselgel 60; DC Fertigplatten, Macherey-Nagel, Germany). The mobile phase for the chromatography was composed of 60% propan-1-ol, 10% propan-2-ol, 10% ethanol, and 20% water. After drying, the plates were first sprayed with Molisch's reagent (5% α-naphthol, 95% ethanol) (35) and subsequently with 0.1 N sulfuric acid, which allowed the visualization of mono- and oligosaccharides.

In order to test whether the purified E. faecalis enzymes form glucose from maltooligosaccharides, we carried out spectrophotometric assays. The detection of glucose was coupled to its NAD⁺-dependent oxidation to glucono-1,5-lactone by the enzyme glucose dehydrogenase (Sigma). The assay mixture contained a total volume of 350 μl of 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 2 mM NAD⁺, maltooligosaccharide at 2.5 mM, and 0.1 mg of glucose dehydrogenase. After preincubation for 5 to 10 min, 2.5 μg of one of the three purified enzymes was added. The formation of NADH was monitored by measuring the OD at 340 nm with a Uvikon 9X3W double-beam spectrophotometer (Kontron Bio-Tek) using the Autorate program.

The formation of maltose from maltooligosaccharides was monitored by coupling the enzymatic reaction to two additional enzymatic steps. First, maltose was phosphorolyzed by purified E. faecalis maltose phosphorylase to glucose and glucose-1-P, and glucose was subsequently oxidized by glucose dehydrogenase, as described above. The assay mixture contained a total volume of 350 μl of 100 mM phosphate buffer (pH 7.0), 5 mM MgCl₂, 2 mM NAD⁺, the maltooligosaccharide at 0.5 mM, 0.1 mg of glucose dehydrogenase, and 20 μg of purified E. faecalis maltose phosphorylase (8). After preincubation for 5 to 10 min, the reaction was started with 5 μg of one of the three purified enzymes.

In order to test whether the purified enzymes can transform panose or isopanose into glucose and maltose, we carried out a two-step spectrophotometric assay. In the first step, we followed the formation of glucose, as described above, until the reaction reached equilibrium (after about 45 min under the employed reaction conditions). We subsequently added 25 μg of maltose phosphorylase to the assay mixture. If during the first reaction maltose had been formed simultaneously with glucose, a further increase in the OD at 340 nm should occur, because the phosphorylase will liberate additional glucose from maltose but not from isomaltose.