Characterization of the Different Dextransucrase Activities Excreted in Glucose, Fructose, or Sucrose Medium by Leuconostoc mesenteroides NRRL B-1299

Abstract

When grown in glucose or fructose medium in the absence of sucrose, Leuconostoc mesenteroides NRRL B-1299 produces two distinct extracellular dextransucrases named glucose glucosyltransferase (GGT) and fructose glucosyltransferase (FGT). The production level of GGT and FGT is 10 to 20 times lower than that of the extracellular dextransucrase sucrose glucosyltransferase (SGT) produced on sucrose medium (traditional culture conditions). GGT and FGT were concentrated by ultrafiltration before sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. Their molecular masses were 183 and 186 kDa, respectively, differing from the 195 kDa of SGT. The structural analysis of the dextran produced from sucrose and of the oligosaccharides synthesized by acceptor reaction in the presence of maltose showed that GGT and FGT are two different enzymes not previously described for this strain. The polymer synthesized by GGT contains 30% α(1→2) linkages, while FGT catalyzes the synthesis of a linear dextran only composed of α(1→6) linkages.

Dextransucrase is a glucosyltransferase (E. C. 2.4.1.5) that catalyzes the transfer of glucosyl residues from sucrose (S) to dextran polymer and liberates fructose (F) according to the following equation (8): n S → n F + dextran (glucose)_n.

Dextran is a high-molecular-mass (10⁷ to 10⁸ Da) glucan (26). It is composed of a linear chain of glucosyl residues all linked through α(1→6) glucosidic bonds and several α(1→2), α(1→3), or α(1→4) branched linkages (27). The frequency and nature of the branch points mainly depend on the origin of the dextransucrase (i.e., the producing microorganism) (9). A single enzyme can catalyze the synthesis of several types of linkages, thus permitting, on its own, the formation of a branched polymer (20, 31). On the other hand, certain bacterial strains have been shown to produce dextrans of various structures, and this was attributed to the excretion by the microorganism of different dextransucrases (1, 5, 33). In other words, each dextran structure is characteristic of a given dextransucrase (9).

Our purpose is to study more precisely Leuconostoc mesenteroides NRRL B-1299 dextransucrase, whose industrial utility was proved a few years ago for the synthesis of small α(1→2) glucooligosaccharides from maltose and S by acceptor reaction (22). L. mesenteroides NRRL B-1299 native dextran, directly synthesized from S in the culture broth, has been separated, by alcohol precipitation, into five fractions with differing solubilities by Kobayashi et al. (12, 13). The polymer was found to be very similar in all of the fractions. Nuclear magnetic resonance (NMR) analysis showed that it contains about 30% α(1→2) linkages and a small amount of α(1→3) linkages (about 5%) (3, 29). The dextransucrase responsible for this dextran formation is extracellular. Production of the enzyme is induced by its own substrate, S (14). As a result, during bacterial culture on S, dextransucrase is always produced in the form of a dextran-enzyme complex, thus making it very difficult to obtain a pure catalyst preparation (25, 30). And, due to the close association between enzyme and dextran, the exact number of enzymes involved in dextran and α(1→2) oligosaccharide synthesis has never been established. Two forms, one soluble and the other insoluble, have often been isolated from culture (3, 14, 25, 30). However, they are generally assumed to be different forms of the same protein. In fact, they present different kinetic parameters, but this was shown to be due to the solubility of the dextran with which they are associated (3). Moreover, the two enzyme preparations synthesize the same dextran and oligosaccharides having the same structure (3, 25).

Recently, dextransucrase constitutive mutants of L. mesenteroides NRRL B-1299 have been isolated (11). Three major active bands were detected at 173, 184, and 240 kDa after sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) analysis of the crude enzyme preparation produced by one of the mutants after cultivation on glucose (G) medium. Results obtained with the wild-type strain after cultivation on S medium were less clear. At the same time, in our laboratory, the gene coding for an intracellular dextransucrase, DSRA, never reported before for this strain was cloned and sequenced (18). The 146-kDa enzyme synthesizes a dextran bearing 87% α(1→6) linkages and 13% α(1→3) linkages. These data indicate that L. mesenteroides NRRL B-1299 possesses different genes coding for dextransucrases. However, these studies give no idea of the conditions under which each enzyme is expressed, making it impossible to determine whether they are involved in B-1299 native dextran synthesis.

We recently showed that the production of the dextransucrase that synthesizes the α(1→2) oligosaccharides was negatively regulated when L. mesenteroides NRRL B-1299 cells coconsumed S and F (4). Dextransucrase production was doubled by growing L. mesenteroides NRRL B-1299 on S medium supplemented with a low G concentration (2). This led us to check if dextransucrases were produced when growing the bacteria on G or F as the sole carbon source. The present study describes, for the first time, the two enzymes released by wild-type L. mesenteroides NRRL B-1299 when grown in such media devoid of S.

MATERIALS AND METHODS

Bacterial strains and culture media.

L. mesenteroides NRRL B-1299 was obtained from the Northern Regional Research Center (Peoria, Ill.) culture collection. One liter of the standard culture medium used consisted of 20 g of yeast extract, 20 g of K₂HPO₄, 0.2 g of MgSO₄ · 7H₂O, 0.01 g of MnSO₄ · H₂O, 0.01 g of NaCl, 0.02 g of CaCl₂, and 0.01 g of FeSO₄ · 7H₂O supplemented with either 20 g of G (G medium), 20 g of F (F medium), or 40 g of S (S medium). The pH of the phosphate buffer was adjusted to 6.9 with orthophosphoric acid. The carbohydrate source, yeast extract, phosphate, and additional salts were sterilized separately. All of the components of the medium were of analytical grade. Yeast extract was analyzed at 200 g/liter, and no traces of S were detected.

Adaptation of L. mesenteroides to G or F medium.

Cells were stored frozen in S medium and glycerol. To eliminate all of the S-induced dextransucrase linked to the cells and to adapt the cells to the other substrates, three successive precultures were carried out on either G or F medium. Erlenmeyer flasks were inoculated (1%, vol/vol) and incubated on a rotary shaker (200 rpm) at 30°C for 12 h. The third preculture was then used to inoculate (1%, vol/vol) a fresh culture (carried out at 27°C, 200 rpm, for 7.5 h), in which dextransucrase production was monitored.

Culture on S medium.

For culture on S medium, a single preculture was made (inoculation of 1% [vol/vol], 27°C, 200 rpm, 6 h) and used to inoculate (1% vol/vol) the fresh culture (carried out at 27°C), in which dextransucrase production was monitored. The final activity level was measured after 6 h of culture. For SDS-PAGE analysis, a sample was taken after 4 h of culture (soluble activity, 0.15 U · ml⁻¹) to analyze a preparation with a low dextran content.

Biomass measurements.

Bacterial growth was measured by a turbidimetric method at 650 nm and calibrated against cell dry-weight measurements as previously described (4). A change of 1 U of optical density at 650 nm was equivalent to 0.4 g of dry matter · liter⁻¹. This method was not convenient for determining biomass during cultures on S medium because of the presence of dextran, which invalidated both optical density and gravimetry measurements. Therefore, cell enumeration was carried out by using a graduated Thoma cell and phase-contrast microscopy; the data were converted into cell dry weight (grams per liter) by means of a calibration graph set up during culture on G medium. A change of 10⁹ cells · ml⁻¹ was equivalent to 0.58 g of dry matter · liter⁻¹.

Electrophoresis analysis.

The culture medium was first centrifuged for 20 min at 7,000 × g and 4°C. The supernatant was then filtered with a Sartorius membrane (0.2-μm cutoff) to ensure the total absence of cells in the supernatant. The filtered supernatants of G and F cultures were concentrated 10-fold by using Centricon tubes (10,000-Da cutoff; Amicon). S medium supernatant was used without concentration because the presence of dextran prevented ultrafiltration. SDS-PAGE was then carried out by the method of Laemmli (15) with 7% (wt/vol) acrylamide gels. Samples containing at least 4.5 mU were deposited on minigels (Minicells system; Hoeffer Instruments). After migration, proteins were stained with Coomassie brillant blue R-250.

For in situ detection of dextransucrase activity, the gel was first washed three times with 20 mM sodium acetate buffer (pH 5.4)–0.05 g of CaCl₂ per liter containing 0.1% (vol/vol) Triton X-100 at 4°C to eliminate the SDS. It was then incubated in the same buffer supplemented with 100 g of sucrose per liter at room temperature for 72 h; the active bands were detected by the appearance of dextran polymer as described by Miller and Robyt (17).

Dextransucrase purification.

For all media (G, F, and S), the cells were pelleted by centrifugation (7,000 × g, 20 min, 4°C).

Extracellular soluble dextransucrase GGT produced on G and FGT produced on F.

With G and F media, culture supernatants were first filtered with a Sartorius membrane filter (0.2-μm cutoff) to ensure complete elimination of cells. Adsorption on exogenous dextran was then used to isolate and concentrate the dextransucrase. Dextran T70 (5 g/liter; Sigma) or sterilized soluble native dextran from L. mesenteroides NRRL B-1299 (5 g/liter) was added to the filtered culture supernatant and left overnight at 4°C. The enzymes were then concentrated by using phase partition as described below.

Extracellular soluble dextransucrase SGT produced on S medium.

With S medium, the supernatant was centrifuged a second time (7,000 × g, 20 min, 4°C). This supernatant could not be filtered because of the presence of native dextran, which immediately clogged the filtration membrane. The native dextran had a final concentration of 5 g/liter, and the supernatant was directly used for phase partition.

Phase partition.

Dextransucrase was concentrated from the supernatants by using aqueous two-phase partition between dextran (native or exogenous) and polyethylene glycol (PEG) (23). After addition of PEG 1500 (15%, vol/vol), the dextran-rich phase containing dextransucrase was concentrated by centrifugation at 7,000 × g for 20 min at 4°C, collected in the pellet, and diluted in 20 mM sodium acetate buffer (pH 5.4). Dextran addition enabled both efficient precipitation of dextransucrase and enzyme stabilization during product synthesis (23).

Dextransucrase standard assay.

One unit of dextransucrase is defined as the amount of enzyme that catalyzes the production of 1 μmol of F per min at 30°C in 20 mM sodium acetate buffer, pH 5.4, with 100 g of S per liter, 0.05 g of CaCl₂ per liter, and 1 g of NaN₃ per liter. It was ascertained that the reducing power measured by DNS assay was due to dextransucrase and not to levansucrase, invertase, or S phosphorylase activity as described by Dols et al. (2).

Dextran synthesis and analysis.

Dextran was synthesized by incubating 1 U of SGT per liter in 100 g of S per liter–0.05 g of CaCl₂ per liter–20 mM sodium acetate buffer (pH 5.4)–1 g of NaN₃ per liter. The same conditions were used with GGT and FGT except that the S concentration was 150 g/liter and the dextransucrase activity was 0.05 U/ml. When the S was depleted, the polymer synthesized was precipitated with ethanol at a 75% (vol/vol) final concentration, washed twice with ultrapure water, and freeze-dried. The ¹³C-NMR spectra of the dextrans were recorded with a Bruker AC 300 spectrometer operating at a frequency of 75.468 MHz. Samples were examined as solutions in D₂O at 70°C in 5-mm-diameter spinning tubes (internal standard; ¹³CH₃ at 31.5 ppm relative to tetramethyl silicon).

Oligosaccharide synthesis and analysis.

Oligosaccharide synthesis in the presence of a maltose acceptor was carried out by incubating 0.05 U of dextransucrase per ml at 25°C in 125 g of total sugar (S and maltose) per liter–20 mM sodium acetate buffer–0.05 g of CaCl₂ per liter at an S/maltose ratio of 2. The oligosaccharides produced (degree of polymerization ranging from 3 to 7) were analyzed by high performance liquid chromatography using a C₁₈ column and a Hewlett-Packard 1050 series system. Oligosaccharides were detected by using an HP1047A refractometer. The products were identified in the chromatograms as described by Remaud-Simeon et al. (25).

RESULTS

After 7.5 h of growth, the cultures were stopped at biomasses of 1.27 and 1.03 g/liter for G and F media; respectively. After 4.5 h, the biomass concentration reached 1.21 g/liter in S medium. The extracellular and soluble dextransucrase activities produced on G medium (GGT) and F medium (FGT) were analyzed and compared to that obtained on S medium (SGT).

Electrophoresis analysis.

The proteins contained in the supernatant of each medium (G, F, or S) were separated by SDS-PAGE. The SDS was removed from the gels by washing, and dextran-forming activity was detected in situ. The results are shown in Fig. 1. For all media, two active bands were observed, one at around 180 kDa and the other at a molecular mass higher than 230 kDa, probably corresponding to a delayed form of the first one. Such a higher-molecular-mass form (at 200 kDa) of dextransucrase was also detected in addition to the expected band at 180 kDa after SDS-PAGE of a recombinant dextransucrase expressed from a single gene expressed in Escherichia coli (19). Actually, dextransucrase has a strong tendency to form aggregates (6, 21, 32) that decrease the efficiency of SDS as a denaturing agent. This hypothesis of two bands corresponding to a single protein is supported by the observation of the same difference between the molecular masses of (i) the nondelayed proteins in the lanes containing G, F, and S media and (ii) the delayed proteins in the same lanes. From the positions of the bands, the molecular masses of the enzymes produced on G, F, and S media were found to be 183, 186, and 195 kDa, respectively, while the delayed bands corresponded to proteins of 235 (G), 238 (F); and 245 (S) kDa. The electrophoresis analysis suggests that the dextransucrases produced are different for each carbon source used to support growth.

FIG. 1.

Dextransucrase molecular mass analysis by SDS-PAGE. The bands were revealed by in situ dextran synthesis. Lanes: 1, G medium; 2, F medium; 3, S medium.

Level of activity reached.

As the level of dextransucrase activity reached with G or F was very low, we used the affinity between dextran and dextransucrase to concentrate the enzyme. With G and F media the purification procedure was run by using two different dextrans to separate any coexisting dextransucrases by means of their affinities for different dextran structures. Two polymers were used: the dextran from L. mesenteroides NRRL B-512F (T70; Sigma) and a soluble native dextran from L. mesenteroides NRRL B-1299. The difference between the two polymers comes from the presence in B-1299 dextran of 32% α(1→2) and 5% α(1→3) branch linkages, while the T70 dextran contains only 5% α(1→3) branch linkages.

Whatever the medium, the initial activity of the broth (just after inoculation) was too low to be measured, even after dextran-PEG precipitation, and was ignored. The soluble activities excreted in G (GGT) and F (FGT) media reached 0.023 and 0.054 U · ml⁻¹ after 7.5 h of culture (Table 1). They were, respectively, 20 and 10 times lower than the soluble activity produced on S medium after a 6-h culture. Whatever the dextran used for precipitation (i.e., T70 or dextran from L. mesenteroides NRRL B-1299), similar levels of activity were recovered.

TABLE 1.

Standard activities of glucosyltransferases produced on G, F, and S media

Carbon source (enzyme) and dextran used for precipitation	Vol (ml) of crude supernatant	Vol (ml) of filtered supernatant	Final vol (ml)	Enzyme activity (U/ml) in:
				Concentratea	Filtered supernatantb
G (GGT)
T70c	31	30	3	0.222	0.023
B-1299c	31	30	5	0.143	0.024
F (FGT)
T70c	31	30	3	0.505	0.054
B-1299c	31	30	7.5	0.262	0.051
S (SGT), B-1299d	50		5.1	4.305	0.439

^aAfter aqueous two-phase partition between dextran and PEG. 

^bCalculated from the activity of the concentrated preparations and the volume ratio. For SGT, the calculated activity of 0.439 U · ml⁻¹ corresponds to that of the supernatant, since no filtration was done. 

^cDextran was added. 

^dNative dextran was present in the crude extract because it was produced by the extracellular dextransucrase during cultivation on S medium (6 h). 

Polymer characterization.

Dextran synthesis was carried out with GGT, FGT, and SGT, all concentrated with B-1299 native dextran. The concentration of the synthesized polymer represented 40 times the amount of dextran coming from phase partition and introduced with the enzyme preparation in the reaction medium. All of the polymers were analyzed by ¹³C-NMR spectrometry. The various signals were assigned as described by Seymour et al. (28) and Remaud et al. (24). As shown in Fig. 2A, B, and C, the structures of the polymers synthesized by the three preparations were different. With GGT, the polymer synthesized had a structure comparable to that of the polymer produced by SGT, except that there was no signal corresponding to α(1→3) linkages at 99.4 to 99.5 and 82.98 ppm (Fig. 2A and C). The chemical shifts characteristic of the α(1→2) branched dextran were present. At 96.55 ppm, the anomeric carbon that participates in the α(1→2) linkage appears. At 96.06 and 95.77 ppm, the carbons corresponding to a G residue involved in α(1→6) linkages and substituted on carbon 2 were encountered. Finally, the signal associated with the α(1→6) linkage resonated at 98.32 and 98.07 ppm. The presence of the α(1→2) linkage was confirmed by the signal at 76.75 ppm which corresponded to carbon 2. From the integration of the different signals, the α(1→2) linkage content was estimated to be 30%.

FIG. 2.

¹³C-NMR analysis of the dextran synthesized by dextransucrases produced on G, F, and S media. A, GGT; B, FGT; C, SGT.

The spectrum of the polymer synthesized by FGT (Fig. 2B) did not display the various signals associated with the α(1→2) or α(1→3) glucosidic bonds and was characteristic of a linear dextran exclusively formed by α(1→6) glucosidic bonds.

Oligosaccharide synthesis.

In the presence of S and maltose (acceptor), GGT synthesized the same mixture of products as SGT (Fig. 3A and C). This was observed whatever the dextran used for GGT precipitation. The products observed were glucooligosaccharides formed of only α(1→6) linkages with a maltosyl residue at the reducing end (OD_n series), together with glucooligosaccharides that contain maltose α(1→6) and α(1→2) linkages (B_n series) (25).

FIG. 3.

Chromatograms of the oligosaccharides synthesized in the presence of S and maltose by dextransucrases produced on G, F, and S media. A, GGT; B, FGT; C, SGT.

With FGT, the oligosaccharides synthesized all belonged to the OD_n series. No oligosaccharides containing α(1→2) linkages were detected, even when B-1299 dextran was used to purify the enzyme (Fig. 3B). Therefore, FGT synthesized only α(1→6) linkages in the presence of S and maltose, as well as in the presence of S alone.

DISCUSSION

This paper clearly demonstrates the excretion of two distinct dextransucrases (GGT and FGT) after cultivation of wild-type L. mesenteroides NRRL B-1299 with either G or F as the carbon source. This has never been reported before for a wild-type strain of the Leuconostoc genus, whose dextransucrases had always been reported to be inducible only by S (14, 20, 26).

The molecular masses of the dextransucrases produced are comparable to those generally described for this type of enzyme, including the delayed form (6, 7, 10, 11). The obtainment of a single nondelayed active band suggests that, in each case, a single dextransucrase is produced. Besides, the different mobilities observed during electrophoresis for GGT, FGT, and SGT suggest that the proteins are distinct. After cultivation on S medium, we obtained a single nondelayed active band at 195 kDa. In the same region, Kim and Robyt (11) obtained a very large active band for their parental strain of L. mesenteroides NRRL B-1299 and concluded that two concomitant bands were present at 184 and 173 kDa. However, they grew the bacteria for 24 h and we carried out a 4-h culture. The difference in culture duration could explain the presence of an additional band corresponding to the proteolytic degradation of SGT or a late induced form corresponding to FGT (see below).

GGT and FGT are hardly detectable without prior concentration and for this reason were never isolated or described before. Their level of production is 10 to 20 times less than the level of soluble activity obtained with S medium. This could result from rapid denaturation of this type of enzyme in the absence of dextran, which is known to stabilize dextransucrase (16, 32). However, T70 dextran addition to the initial culture medium did not improve the level of activity detected in the supernatant after growth (data not shown). The method used to concentrate the enzymes relies on their high affinity for dextran and on the immiscibility of dextran and PEG aqueous solutions. The level of GGT or FGT activity precipitated was exactly the same, whatever the structure of the exogenous dextran used. Moreover, the nature of the enzyme precipitated was exactly the same; i.e., the products synthesized had the same structure. This tends to confirm the presence of a single dextransucrase in both G and F supernatants.

GGT synthesizes both α(1→2) and α(1→6) linkages in proportions similar to those obtained with SGT. However, the absence of α(1→3) linkages in the dextran, together with the molecular mass determination, tends to show that GGT is different from the dextransucrase excreted during growth on S medium. On the other hand, the characterization of the products synthesized by FGT clearly demonstrates that this enzyme differs from the various forms of dextransucrase previously described for L. mesenteroides NRRL B-1299 (3, 9, 13, 25). It synthesizes a linear dextran only composed of α(1→6) linkages from S, and there are no traces of α(1→2) or α(1→3) linkages, even in the oligosaccharides produced on maltose and S.

The control of the synthesis of GGT, FGT, and SGT is probably very complex. In previous papers, we showed the importance of G, S, and F as growth substrates and elements of regulation of dextransucrase production (4). During cultivation on S, the disaccharide is transported into the cell, where it is split into (i) glucose-1-phosphate, which is used by the cell to support growth and dextransucrase production, and (ii) F, which is released outside the cell. However, when F becomes more abundant than S in the medium, it is slowly consumed by the cell together with glucose-1-phosphate. Concomitantly, a drop in the energy yields in the cell is observed and dextransucrase production stops. However, in the presence of G (added by the operator to the S medium), F catabolism is prevented and dextransucrase production lasts until S exhaustion (2). At the time of these studies, we had no idea of the existence of GGT and FGT. Now, new hypotheses can be formulated concerning the system of regulation of dextransucrase production. Due to the low level of activity measured, GGT and FGT could have been regarded as SGT leaks in the absence of S, but no leak of dextransucrase activity was detected on galactose (20 g/liter) medium, even after dextran-PEG phase partition (data not shown). Moreover, from the SDS-PAGE results and product analysis, the synthesis of each dextransucrase appears to be induced solely by the corresponding substrate (e.g., glucose for GGT). And during S catabolism, SGT is the only enzyme expressed: a single nondelayed active band was detected at 195 kDa. It is possible that near the end of S fermentation, F catabolism prevents SGT production and stimulates FGT synthesis. This could be an explanation for the presence of two bands in the gels of Kim and Robyt (11) in supernatant from a long (24-h) S fermentation. And due to S exhaustion before FGT starts being produced, SGT alone is responsible for the synthesis of B-1299 native dextran. GGT and FGT are new enzymes never described before and not involved in B-1299 native dextran synthesis.

Further studies on sugar catabolism and dextransucrase gene organization are necessary to reach a more precise idea of the regulation of dextransucrase production.