Construction of truncated forms of dsrE gene.
Plasmids for high-level overexpression of truncated dsrE genes were first constructed, checked by DNA sequencing, and expressed in E. coli.
Truncated DSR-E variants (Table ) were successfully overproduced, as shown on the SDS-3 to 8% PAGE gel (Fig. ). The protein sizes were in agreement with the expected sizes. Enzyme assays of crude cell extracts using dinitrosalicylic acid confirmed that, except for CD2 alone, all the variants possessed a glucosyltransferase activity in the presence of sucrose as a unique substrate. The levels of activity obtained for each crude cell extract were 580 U/liter of culture for the entire DSR-E, 635 U/liter for Δ(VZ), 169 U/liter for Δ(CD2), 38 U/liter for GBD-CD2, and 2 and 0 U/liter for CD1 and CD2 alone, respectively.
FIG. 1. SDS-PAGE profiles of recombinant DSR-E and truncated forms produced by E. coli TOP10. Lanes: M, broad-range prestained precision protein standard (Bio-Rad); 1, DSR-E, 325 kDa; 2, Δ(VZ), 304 kDa; 3, Δ(CD2), 230 kDa; 4, CD1, 116 kDa; 5, (more ...)
Zymograms detecting polymer synthesis activity were analyzed in parallel. They revealed that CD2, with or without GBD, did not produce any polymer. The other forms, except for CD1, were active. The zymograms also revealed the presence of several active bands of lower molecular weight than the expected one. This may be due to proteolytic degradation, mRNA instability, secondary translation sites, or premature release of ribosomes. Western blot experiments using anti-thioredoxin or anti-His antibodies indicated that all degraded forms harbored only one of the two tags, showing that degradation proceeds from both extremities and thus prevents the isolation of the complete forms on the Ni column (data not shown).
However, SDS-PAGE gels (Fig. ) showed that nondegraded recombinant proteins are preponderant in each crude enzyme extract. In addition, no glucansucrases are naturally produced by E. coli. This allowed these extracts to be used for polymer synthesis.
Actions of DSR-E and its truncated forms on sucrose: polymer synthesis.
Polymer synthesis was performed in the presence of 100 g of sucrose liter−1
and the various crude enzyme extracts except for CD1 and CD2 alone, which were not sufficiently active. The glucans obtained were analyzed by 13
C NMR spectroscopy. The various signals were assigned as described by Seymour and Knapp (23
). The relative contents of the different glucosidic linkages of the polymer were then calculated and compared to the data obtained for the glucan synthesized by the native enzymes (Table ).
Glucosidic linkages of dextrans synthesized by different variants of DSR-E, determined by quantitative 13C NMR analysis
All the glucans produced are dextrans, since they possess >50% α-1,6 linkages, ranging from 81% for the complete form to 86% for the Δ(CD2) form.
In addition to α-1,6 glucosidic bonds, α-1,2 linkages were also observed in dextran synthesized by the complete form of DSR-E. Surprisingly, the percentage of α-1,2 linkages was low (5%). Characteristic α-1,3 and α-1,4 13C anomeric signals were also detected. They represented 10 and 3%, respectively, of the overall glucosidic bonds. The variant Δ(VZ) synthesized α-1,2 bonds more efficiently than the complete enzyme: indeed, the polymer produced was composed of 10% α-1,2 glucosidic bonds versus 5% for the complete DSR-E. Moreover, no α-1,2 signal was detected on spectra corresponding to enzymatic forms devoid of the second catalytic domain. Finally, GBD-CD2 was incapable of producing any polymer. Seventy-eight and 16% of the glucosyl residues coming from the sucrose were transferred to water (sucrose hydrolysis) or to previously released fructose to form leucrose (α-1,5 glucosylfructose), respectively (Table ). Then, acceptor reactions were performed to further characterize the capacity of each variant to transfer glucose units on various acceptors and to determine yields and the type of linkage synthesized.
Products formed with or without acceptors by GBD-CD2a
Actions of DSR-E and its truncated forms in the presence of maltose acceptor.
First, GOS synthesis was performed with each variant using 200 g of sucrose liter−1 and 100 g of maltose liter−1. Fig. shows the different HPLC profiles from which the oligosaccharide synthesis yields were calculated (Table ). CD1 and CD2 show very low sucrose consumption rates: only 20 and 7%, respectively, were consumed after 20 h of reaction. GOS with α-1,2 linkages are synthesized only by the variants possessing both catalytic domains. Another observation concerns the variant devoid of the variable zone, for which a twofold-increased α-1,2 GOS yield was observed compared to DSR-E. Surprisingly, variants containing only the second catalytic domain did not produce any GOS from sucrose and maltose; most of the glucosyl moieties were transferred to water (65%) or fructose (25%) (Table ). To understand the absence of GOS synthesis, α-1,6 GOS mixtures were tested as acceptors for GBD-CD2.
FIG. 2. Reverse-phase HPLC chromatograms of the oligosaccharides synthesized by the different crude truncated forms of DSR-E in the presence of sucrose and maltose. The enzymes used were (a) native L. mesenteroides NRRL B-1299 dextransucrase, (b) complete DSR-E, (more ...)
Synthesis yields of α-1,6 GOS, α-1,2 GOS, and total GOS in the presence of 200 g of sucrose liter−1 and 100 g of maltose liter−1a
Action of GBD-CD2 in the presence of α-1,6 GOS and dextran acceptors. (i) α-1,6 GOS acceptor.
The HPAEC chromatograms presented in Fig. show that in the presence of 200 g of sucrose liter−1 and 50 g of α-1,6 GOS acceptors liter−1, the variant GBD-CD2 used almost all the α-1,6 GOS as an acceptor to form products whose retention times are similar to those of α-1,2 GOS. This shows that GBD-CD2 transfers one glucosyl residue to the nonreducing end of each α-1,6 GOS through the formation of an α-1,2 linkage. Moreover, no α-1,6 GOS was produced, as indicated by the calculation of the α-1,2 GOS yield, which reached 31%. This low result can be explained by the high level of glucosyl transfer to water (32%) or fructose (28%) (Table ). However, under these conditions, α-1,2 GOS represented 89% of the total final GOS mixture. Using a lower sucrose/acceptor ratio (75 g of sucrose liter−1 and 122 g of α-1,6 GOS liter−1) reduced hydrolysis and leucrose synthesis. Consequently, an α-1,2 GOS yield of 72% was reached and α-1,2 GOS represented 60% of the total GOS mixture.
HPAEC chromatograms of the oligosaccharides synthesized by the truncated form GBD-CD2 of DSR-E in the presence of sucrose and α-1,6 GOS.
In parallel to the acceptor reactions carried out with the crude extracts, reactions were performed with the constructs purified from SDS gels and thus devoid of any active degraded proteins. The profiles of the synthesized products were similar to those obtained with the crude extracts, confirming that the activity of the degraded forms can be ignored (data not shown).
(ii) Linear dextran acceptor.
α-1,6 dextran with an average molecular weight of 70,000 was used as an acceptor. NMR spectra of the dextran, before and after the acceptor reaction (shown in Fig. , respectively), revealed that GBD-CD2 significantly modified the acceptor. Indeed, spectrum B displayed all the chemical shifts characteristic of the α-1,2 branched dextran. At 97.17 ppm, the anomeric carbon that participates in the α-1,2 linkage appears. At 98.62 and 96.34 ppm, the anomeric carbons corresponding to glucosyl units involved in α-1,6 linkages, carbon 2 free or linked to another glucose unit, respectively, are encountered. The presence of the α-1,2 linkage is further confirmed by the signal at 76.61 ppm, which corresponds to carbon 2 involved in α-1,2 linkages. From the integration of the different anomeric signals, the α-1,2 content was estimated to be 32%.
13C NMR analysis of the dextran synthesized by dextransucrases from L. mesenteroides NRRL B-512F before (A) and after (B) modification by the variant GBD-CD2 of DSR-E.