Biomacromolecules 2000, 1, 360-364
360
Epimerization of Nonnatural Uronans with Mannuronan C-5-Epimerases To Obtain Alginatelike Polysaccharides Vittorio Crescenzi,‡ Martin Hartmann,† Arjan E. J. de Nooy,‡ Vania Rori,‡ Giancarlo Masci,‡ and Gudmund Skjåk-Bræk* Department of Biotechnology, Norwegian University of Science and Technology, Sem Sælands vei 6-8, 7491 Trondheim, Norway; and Department of Chemistry, University “La Sapienza”, P.le Aldo Moro 5, 00185 Rome, Italy Received April 5, 2000
Different polysaccharides containing D-mannose residues have been C-6-oxidized by a selective TEMPOmediated hypohalite oxidation to obtain the corresponding uronans. These have been treated with various recombinant mannuronan C-5-epimerases and the resulting products were analyzed by 1H NMR spectroscopy. Oxidized konjac mannan could be epimerized to obtain a uronan with a content of about 12% R-L-gulopyranuronate (G) residues. On prolonged epimerization, β-elimination was observed. The oxidized galactomannan locust bean gum could only be scarcely epimerized, probably due to steric effects exerted by its 26% R-D-galacturopyranosyl side groups. Oxidized, galactose-depleted guar gum with a R-D-galactosyl content of 11% could be epimerized to a G content of about 15%. With oxidized cellulose as a substrate, mainly β-elimination was observed. It thus seems that the mannuronan C-5-epimerases employed recognize glucuronate residues and abstract proton-5 but are unable to perform the second epimerization step and instead yield β-eliminated products. Introduction Alginate is a linear polysaccharide consisting of various proportions of (1f4)-linked β-D-mannopyranuronate (M) and R-L-gulopyranuronate (G) residues. The various proportions of these two monomers found in natural alginate depend on its source and may range from 20 to 85% M for alginates isolated from seaweeds to 10-100% M for alginates isolated from bacteria. Alginates, until now exclusively from algal sources, are widely used in various industrial applications for their water holding capacity and for their gelling, viscosifying, and stabilizing properties. These properties are in large part determined by the proportion and the distribution of the two monomers. Alginates with G-blocks are able to form reversible gels on addition of divalent cations such as Ca2+, Ba2+, or Sr2+, whereas alginates rich in M-blocks are weak gelling agents. M-rich alginates stimulate cytokine production and have a much higher antitumor activity than alginates with a high fraction of G-blocks. Therefore, it would be desirable to be able to modify the chemical composition of alginate or alginate-like polysaccharides in such a way that the properties of the ensuing products can be finely tuned to the desired application. In the biosynthetic pathway of alginate mannuronan is formed first, then G-residues are introduced by C-5-epimerization of M-residues within the polymer, yielding alginate.1 The exact chemical composition (proportion and distribution * Corresponding author:
[email protected]. † Norwegian University of Science and Technology. ‡ University “La Sapienza”.
of the two monomers) of the resulting alginate depends on the action of the epimerase. Recently, it has been found that the genome of the alginate-producing bacterium Azotobacter Vinelandii encodes at least seven different mannuronan C-5epimerase genes. These genes have been sequenced and were cloned and expressed in Escherichia coli and the enzymes thus produced have been designated AlgE1-AlgE7.2-4 When these enzymes are used for the epimerization of M-rich alginate or mannuronan, the reaction products differ significantly in proportion and distribution of the two monomers. While AlgE4 predominantly forms alginates with MGblocks,5 AlgE23 and AlgE64 introduce stretches of G-blocks into the polymer. We have investigated the possibility of epimerizing nonnatural uronans using various mannuronan C-5-epimerases. The nonnatural uronans were obtained by a selective TEMPO-mediated hypohalite oxidation6 of the corresponding neutral polysaccharides. Abundant, industrially important polysaccharides containing mannose, such as locust bean gum (LBG) and konjac mannan (KM) were used. Locust bean gum consists of a straight chain of (1f4)-linked β-Dmannopyranosyl units with (1f6)-linked R-D-galactopyranosyl units bound to, on the average, every third to fourth main chain unit. To obtain some insight in the influence of the density of the GalA side chains on the epimerization, guar gum (a polysaccharide with a similar structure as locust bean gum, but with an R-D-galactopyranosyl unit attached to, on the average, every second main chain unit) was treated with R-galactosidase, an enzyme that splits off the R-D-Gal units to obtain a galactomannan with a lower galactose content than LBG. Konjac mannan is a linear (1f4)-linked
10.1021/bm0055394 CCC: $19.00 © 2000 American Chemical Society Published on Web 06/01/2000
Biomacromolecules, Vol. 1, No. 3, 2000 361
Epimerization of Nonnatural Uronans
β-D-glucomannan with a ratio mannose/glucose of 1.6. Apart from these D-ManA containing polysaccharides, the reactivity of the epimerases toward C-6-oxidized cellulose consisting of (1f4)-linked β-D-glucuronate was studied. Experimental Section Materials. The Konjac mannan ([η] ) 3.1 dL/g) was a Nutricol RE 2273 “special research grade” material from FMC Europe, Food Ingredients Division (Brussels, Belgium). Guar gum, locust bean gum, and cellulose (sigmacell) were obtained from Sigma (St. Louis, MO) and were used as received. R-Galactosidase was supplied as an ammonium sulfate suspension in 0.02% NaN3 by Megazyme (Wicklow, Ireland). High molecular weight mannuronan was produced from an epimerase-negative mutant (AlgG-) of Pseudomonas fluorescens grown on agar plates, and was deacetylated by mild alkaline hydrolysis, as described earlier.7 All other chemicals were commercially available products and were used without further purification. Isolation of Recombinant Mannuronan C-5-Epimerases. The mannuronan C-5-epimerases were produced by fermentation of these recombinant E. coli strains: AlgE1 and AlgE2 in JM 109,2,8 AlgE4 in JM 105,8 and AlgE6 in SURE.4 The enzymes were partially purified by ion-exchange chromatography on Q-sepharose FF (Pharmacia, Uppsala, Sweden) and hydrophobic-interaction chromatography on phenyl sepharose FF (Pharmacia). The activity of the enzymes was assayed by measuring the release of tritium to water, when the enzymes were incubated with 3H-5 labeled mannuronan.7 Oxidation of the Polysaccharides. Except for cellulose, the oxidation procedure adopted was as follows. The polysaccharide (2.5 g) was suspended in 1700 mL of distilled water and was heated for 15 min at 90 °C after which the solution was allowed to cool. When the temperature had reached ∼50 °C, the solution was filtered to remove insoluble parts. The solution was allowed to cool to ambient temperature and TEMPO (2,2,6,6-tetramethylpiperidin-1oxyl, Sigma) (0.05 g) and NaBr (0.2 g) were added. The resulting solution was cooled overnight to 4 °C. Then, cold sodium hypochlorite, previously brought to pH 9, was added. Without further cooling the reaction was monitored while 0.5 M NaOH was added to maintain a pH of 9-9.5. After 1-1.5 h, when the pH was decreasing only very slowly, MeOH was added to the reaction mixture to consume NaOCl still present, then NaBH4 (0.2 g) was added. The reaction was left overnight at ambient temperature and was terminated by adding 4 M HCl, bringing the pH to ∼5. The resulting polysaccharide was precipitated by addition of acetone (2.5 L), collected, redissolved in water, and dialyzed. Yields of oxidized polysaccharide were about 80-90%. Cellulose was oxidized in a different manner, since it is completely insoluble in water. To a suspension of 4 g of cellulose in 200 mL of water were added 0.1 g of TEMPO and 0.2 g of NaBr. The polysaccharide was partly oxidized at ambient temperature, by slowly adding NaOCl and maintaining the pH at 9 by addition of 0.5 M NaOH. Within 1 h, about 70% of the primary alcohol groups of the polysaccharide were oxidized, and NaBH4 was added to reduce the carbonyl groups. The solution was brought to pH 5 by addition of 4 M HCl and insoluble parts were filtered. Subsequently, NaOCl was added to oxidize the remaining primary alcohol groups as described above. After a second reduction with NaBH4 the oxidized polysaccharide was desalted by dialysis and freeze-dried. The yield of oxidized product was about 50-60%. r-Galactosidase Treatment and Oxidation of Guar Gum. D-Galactose depleted guar gum was prepared following a previously
Table 1. Materials and Conditions Used for Epimerization of C-6-Oxidized Konjac Mannan epimerase type, amt (mg)
KM (mg)
Ca2+ (mM)
H2O (mL)
reaction time (h)
AlgE1, 5.0 AlgE2, 5.0 AlgE4, 5.0 AlgE4, 20.0 AlgE6, 2.5
50 50 150 800 50
1.0 1.0 4.0 4.0 3.0
75/37.5a 75/37.5a 150 800 120/60a
144/480 144/480 72 168 144/480
a
Samples were taken after the first reaction time.
optimized procedure.9 A 2 g sample of guar gum in 200 mL of 0.1 M acetate buffer (pH 4.6) was treated with R-galactosidase (900 nkat; 1 nkat is defined as the amount of enzyme required to release 1 nmol of p-nitrophenol/s at pH 4.5 and 40 °C) at 35-40 °C for 48 h. During the reaction, a polysaccharide precipitate appeared due to the decreased solubility of D-galactose-depleted guar gum. The suspension was brought to 95 °C for a short time to inactivate the enzyme and was then cooled to room temperature. TEMPO (0.04 g) and NaBr (0.2 g) were added, and the suspension was cooled to 4 °C. The oxidation-reduction was performed as described above for the glucomannans. During the oxidation, the polysaccharide was drawn almost completely in solution. After dialysis, the oxidation-reduction was repeated, yielding a clear solution. The product was finally isolated in a yield of 60%. The content of galacturonate in the final product was determined by 1H NMR and was found to be 11% of the uronates. Epimerization of the TEMPO-Oxidized Polysaccharides and Mannuronan. All samples were incubated in aqueous solution at pH 7.0 at 40 °C under magnetic stirring. No buffer was added. Each of the polymers was incubated without enzymes as a blind test. The reactions were stopped by cooling and chelation of Ca2+ by addition of Chelex 100 (iminodiacetic acid insolubilized on polystyrene beads, Sigma, St. Louis, MO). The samples were shaken gently for about 12 h and the beads removed by filtration (5 µm filter). The pH was adjusted to 7.0, and the samples were dialyzed and freeze-dried. Three 25-mg portions of C-6-oxidized cellulose were dissolved with 2.5 mg of AlgE1, AlgE2, or AlgE6 in 35 mL of 1 mM CaCl2 solution. After incubation for 144 h the samples were treated as described for C-6-oxidized konjac glucomannan. Then 150 mg of either C-6-oxidized cellulose, oxidized locust bean gum, or oxidized Gal-depleted guar gum each were incubated with 5.0 mg of AlgE4 in 150 mL of 4 mM CaCl2 solution. After 168 h the reactions were stopped and the samples treated as described above. As a positive control mannuronan was epimerized with AlgE4. First, 150 mg of mannuronan were incubated with 6 mg of AlgE4 in 300 mL of 2 mM CaCl2 solution for 12 h. The samples were then treated as described above. Viscosity. The intrinsic viscosity of the polymers was determined on a Schott-Gera¨te apparatus with automatic dilution, using an Ubbelohde capillary (Φ ) 0.53 mm) at 25 °C and an added salt concentration of 0.1 M NaCl. 1H NMR Spectroscopy. Spectra were recorded on a Bruker DPX 400 spectrometer. The polymers were slightly degraded by mild acid hydrolysis to reduce their viscosity, as described for alginate.6 Here, 5 mm sample tubes were used; temperature was 90 °C, and 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid Na salt (Aldrich, Milwaukee, WI) was used as internal standard for the ppm scale. Then, 20 µL of 0.3 M TTHA (triethylenetetraamine hexaacetate, Sigma) was added to chelate any remaining Ca2+ from the epimerization. The peaks were assigned according to Grasdalen10 and Heyraud et al.11
362
Biomacromolecules, Vol. 1, No. 3, 2000
Figure 1. Anomeric region of the 1H NMR spectra (400 MHz) of C-6oxidized locust bean gum, galactosidase-treated C-6-oxidized guar gum, and mannuronan, (a-c) before and (d-f) after treatment with the recombinant mannuronan C-5-epimerase AlgE4. G and M denote internal GulA and ManA residues, the numbers denote which proton is causing the signal, and the nonunderlined letters refer to the neighboring residues. The spectra were recorded at a polymer concentration of 10 mg/mL in D2O at 90 °C.
Results and Discussion The degree of oxidation of the oxidized products was determined, apart from the consumption of NaOH during the oxidation, by 1H NMR. Generally, the oxidation of the primary alcohol groups had proceeded to at least 90%, and virtually no other monomers than the corresponding uronates could be detected. The epimerases used on the oxidized polysaccharides were AlgE1, AlgE2, AlgE4, and AlgE6. Out of these only AlgE4 and AlgE6 were found to be active. As stated above, the epimerases are able to introduce different distributions of G when applied to high M-alginate. AlgE4 selectively introduces MG-blocks, implying that a maximum of 50% G may be introduced, whereas AlgE6 introduces mainly G-blocks. In Figure 1 the anomeric regions of the 1H NMR spectra of oxidized locust bean gum, oxidized galactose-depleted guar gum, and pure mannuronan before and after epimerization by AlgE4 are shown. From the fact that mainly two anomeric resonances are seen in the 1H NMR spectrum of oxidized LBG (Figure 1a: at 4.7 ppm for M and 5.05 ppm for GalA), it may be concluded that the oxidation was complete. In the spectrum of oxidized-epimerized LBG (Figure 1d), the small shoulder partially under the H-1 galacturonate resonance stems from G, obtained from epimerization of M residues. The intensity of this resonance is quite low and amounts to approximately 2% G of the final polymer. It was thought that the low amount of epimerization was due to the presence of about 26% R-D-galacturonate branches on the main chain. To produce a galactomannan with a lower galactose content, guar gum was treated with R-galactosidase. This yielded a
Crescenzi et al.
galactomannan with a galactose content of only about 11%, which, due to this low content of galactose branches, was not completely soluble in water. For this reason, this product was not isolated, but directly oxidized as described in the Experimental Section. Clearly, the GalA content in this oxidized galactomannan (Figure 1b) is much lower than in the oxidized LBG. Furthermore, even a second oxidation (see Experimental Section) does not seem to have a negative influence on the selectivity of the oxidation toward the primary alcohol group at C-6, since mainly galacto- and mannopyranuronate anomeric resonances are observed. The resulting uronan was treated with AlgE4 and it can be seen that appreciably more epimerization occurred in this case (about 15% of G-residues in the product, indicating 30% MG-blocks, Figure 1e) which seems to justify the thought that the low amount of epimerization in oxidized LBG was mainly due to steric hindrance caused by the GalA-branches. In the case of mannuronan, where no GalA-branches are present (Figure 1c), the epimerization proceeded to a final G-content of about 48%, almost the theoretical maximum for an alternating MG sequence (Figure 1f). Another mannosyl-containing polysaccharide used in this study was the linear glucomannan konjac mannan. Although the mannose proportion is not as high as for the galactomannans used, it is known that the distribution of mannose and glucose along the chain is not regular, giving rise to mannose-blocks. This feature may make KM, after oxidation, particularly suited for the mannuronan C-5-epimerases. Of the four epimerases tested, only AlgE4 and AlgE6 were active with oxidized KM as substrate. AlgE1 and AlgE2 left the polymer unchanged, even after a prolonged incubation time of 20 days. In Figure 2 the anomeric regions of the 1H NMR spectra of oxidized KM and the two oxidizedepimerized KM samples are shown. The 1H NMR spectrum of oxidized KM only shows two anomeric resonances: M (H-1, 4.70 ppm) and GlcA (H-1, 4.55 ppm), confirming a high selectivity for the primary alcohol groups and a high degree of oxidation. After epimerization with AlgE4 and AlgE6, Figure 2 parts b and a, the proportions of G introduced are comparable and amount to approximately 10-15%. However, two important differences can be noted. In Figure 2a a resonance appears at 4.45 ppm which is caused by H-5 of G next to another G-unit, indicating that G-blocks have been introduced, which is not seen in Figure 2b (or Figure 1). As expected, a relatively intense MG-block signal (MG-1) appears in Figure 2b at 4.75 ppm. These results confirm earlier reports that AlgE6 is able to introduce G-blocks, whereas AlgE4 specifically introduces alternating MG structures. In the spectra in Figures 2b and 1, no resonances have been found at 5.7-5.9 ppm (region not shown in Figure 1), caused by H-4 of the 4,5-unsaturated residue 4-deoxy-Lerythro-hex-4-enepyranosyluronate (∆-4). This residue is formed from M and G units when degradation by β-elimination occurs. A minor signal is found in Figure 2a, indicating that some degradation has taken place. The absence of enzymatic degradation was confirmed by measurements of the intrinsic viscosity of oxidized KM (5.3 dL/g) and the two oxidized-epimerized KM samples (AlgE4: 5.5 dL/g,
Epimerization of Nonnatural Uronans
Figure 2. Anomeric region of the 1H NMR spectra (400 MHz) of C-6oxidized konjac glucomannan treated with the recombinant mannuronan C-5-epimerases AlgE6 and AlgE4 for 3 days. G, M, and ∆ denote internal G and M residues and 4-deoxy-L-erythro-hex-4enepyranosyluronate. The numbers denote which proton of the underlined residue is causing the signal, and the nonunderlined letters refer to the neighboring residues. The spectra were recorded at a polymer concentration of 10 mg/mL in D2O at 90 °C.
AlgE6: 4.9 dL/g), all measured in 0.1 M NaCl. Since even a slight degradation would have a large effect on the intrinsic viscosity of the polymers, it may be assumed that after 3 days of incubation the epimerization by AlgE4 has proceeded virtually without any elimination, while the extent of degradation by AlgE6 is very low. Furthermore, the intrinsic viscosities of the polyelectrolytes are substantially higher than that of the uncharged parent polymer KM (3.1 dL/g, 0.1 M NaCl). Although it is known that the TEMPO-mediated hypohalite oxidation always is accompanied by some chemical degradation, depending on the reaction conditions and the polysaccharide used,12 it appears that for KM this degradation is very limited. We found that the oxidized KM polymer was heavily degraded by β-elimination by both AlgE4 and AlgE6, when the incubation time was increased from 3 to 7 days. The β-elimination mechanism splits the glycosidic bond, leaving an 4,5-unsaturated monomer at the nonreducing end of the polymer and creating a new reducing end. Both types of monomers give distinct signals in 1H NMR spectra, as can be seen in the spectra of the degraded polymers shown in Figure 3. For comparison, the spectrum of an alginate degraded with the recombinant alginate lyase AlgL from A. Vinelandii, published earlier,13 is shown in Figure 3c.
Biomacromolecules, Vol. 1, No. 3, 2000 363
Figure 3. Anomeric region of the 1H NMR spectra (400 MHz) of C-6oxidized konjac glucomannan treated with the recombinant mannuronan C-5-epimerases AlgE4 (a) and AlgE6 (b) for 7 days, the same material as in Figure 2. For comparison, a 300 MHz spectrum of an alginate degraded with the recombinant alginate lyase AlgL from Azotobacter vinelandii, published earlier13 is shown in (c). G, M, Gred, Mred, or ∆ denote internal G or M residues, reducing end G or M residues, or 4-deoxy-L-erythro-hex-4-enepyranosyluronate at the nonreducing end. The numbers denote which proton of the underlined residue is causing the signal, and the nonunderlined letters refer to the neighboring residues. The double peak at 5.82 ppm between ∆-4G and ∆-4M probably stems from the trimer ∆-4-M-Mred.11 The spectra were recorded at a sample concentration of 10 mg/mL in D2O at 90 °C.
Clearly, after 7 days of incubation several elimination products can be seen, especially for oxidized KM treated with AlgE6 (b). The 4,5-unsaturated monomers are indicated by their signals of H-4 between 5.7 and 5.9 ppm and by signals resulting from H-1, found at 5.23 (∆-G) and 5.18 (∆-M). Since ManA and GulA yield the same unsaturated unit, the differences in the chemical shift values of ∆-4 in the alginate spectrum (c) are due to the effects of neighboring units. The additional dublett at 5.88 ppm in (b) could be due to H-4 of the unsaturated unit resulting from GlcA, 4-deoxy-L-threo-hex-4-enepyranosyluronate. The C-5-epimerization of mannuronan appears to be a twostep process, where the first step of the reaction is the abstraction of a proton at C-5, after which either epimerization, β-elimination, or reprotonation from the same side is possible. Deprotonation at C-5 is also the first step in the reaction of alginate lyases, a type of enzymes that degrades alginate by β-elimination. The common mechanism of alginate lyases and epimerases was first suggested by Gacesa.14 Recently, lyase activity on mannuronan has been reported for two recombinant C-5-epimerases from A.
364
Biomacromolecules, Vol. 1, No. 3, 2000
Crescenzi et al.
left the polymer unchanged. In Figure 4 the anomeric regions of the 1H NMR spectra of C-6-oxidized cellulose before (4a) and after treatment with AlgE4 (4b) are shown. From the 1H NMR spectrum in Figure 4a, it is clear that side products of the oxidation are present, probably nonoxidized glucose units (H-1, 4.9 ppm); however, most of the material consists of GlcA as is seen from the intense resonance of H-1 at 4.55 ppm. After treatment with AlgE4 for 7 days, various new resonances arise, indicating that the mannuronan epimerase AlgE4 also is active on a β-glucuronan. It remains to be investigated if any R-L-iduronate is formed by epimerization of GlcA, however it is evident that β-elimination takes place, as was seen in the case of oxidized konjac mannan. This indicates that the mannuronan C-5epimerase was able to recognize glucuronan as a substrate and to perform a β-elimination. Whether modification of the enzymes or the reaction conditions could result in epimerization remains to be shown. In conclusion it is clear that the mannuronan epimerases AlgE4 and AlgE6 are able to epimerize β-D-ManA units in various nonnatural uronans obtained by a TEMPO-mediated hypohalite oxidation. This two-step process, starting from mannose-containing polysaccharides, is thus a convenient method to synthesize a family of uronans, which might have alginatelike properties. Figure 4. Anomeric region of the 1H NMR spectra (400 MHz) of C-6oxidized cellulose (glucuronate) before (a) and after incubation with the recombinant mannuronan C-5-epimerase AlgE4 for 168h (b). δ denotes the 4-deoxy-L-threo-hex-4-enepyranosyluronate at the nonreducing end. The spectra were recorded as explained in the legend to Figure 1.
Vinelandii, AlgE7 (high activity, degradation to approximately pentamers)4 and AlgE2 (low lyase activity, 1-3 eliminations per 1000 epimerizations).15 Lyase activity has not previously been observed for AlgE4 and AlgE6; however, full epimerization of mannuronan with these enzymes usually is obtained after an incubation time of only 6 h, and a very slow β-elimination could have gone unnoticed. To investigate the reactivity of the mannuronan C-5epimerases on glucuronan (mannuronans 2-epimer) and to investigate possible lyase action on GlcA, the (1f4)-β-Dglucan cellulose was oxidized and incubated with the different epimerases. The oxidation of cellulose was a nontrivial task, since in all cases some insoluble material remained. Several methods of activation of cellulose before the oxidation were tried, but none of them yielded a clear solution after a single oxidation cycle. For example, activation with either DMF at high temperatures or activation in concentrated NaOH did not give improved results. Also oxidation for 24 h at room temperature did not yield a clear solution. Finally the procedure described in the Experimental Section was adopted to obtain a clear solution of oxidized cellulose with a high degree of oxidation. Of the four epimerases tested, only AlgE4 was active with oxidized cellulose as substrate. AlgE1, AlgE2, and AlgE6
Acknowledgment. This work was supported by FMC Biopolymer, Drammen, Norway, by EU Grant No. QLK 3-CT, 1999-00034, and by the Italian Ministry for Universities and Scientific Research, MURST (“cofinanciamento” 1999). References and Notes (1) Larsen, B.; Haug, A. Carbohydr. Res. 1971, 20, 225. (2) Ertesvåg, H.; Doseth, B.; Larsen, B.; Skjåk-Bræk, G.; Valla, S. J. Bacteriol. 1994, 176, 2846. (3) Ertesvåg, H.; Høidal, H. K.; Skjåk-Bræk, G.; Valla, S. J. Biol. Chem. 1998, 273, 30927. (4) Svanem, B. I. G.; Skjåk-Bræk, G.; Ertesvåg, H.; Valla, S. J. Bacteriol. 1999, 181, 68. (5) Høidal, H. K.; Ertesvåg, H.; Skjåk-Bræk, G.; Stokke, B. T.; Valla, S. J. Biol. Chem. 1999, 274, 18, 12316. (6) de Nooy, A. E. J.; Besemer, A. C.; van Bekkum, H. Carbohydr. Res. 1995, 269, 89. (7) Ertesvåg, H.; Skjåk-Bræk, G. In Methods in Biotechnology; Bucke, C., Ed.; Humana Press Inc.: Totowa, NJ, 1999; Vol. 10, p 71. (8) Ertesvåg, H.; Høidal, H. K.; Hals, I. K.; Rian, A.; Doseth, B.; Valla, S. Mol. Microbiol. 1995, 16, 719. (9) Pulpin, P. V.; Gidley, M. J.; Jeffcoat, R.; Underwood, D. R. Carbohydr. Polym. 1990, 12, 155. (10) Grasdalen, H. Carbohydr. Res. 1983, 118, 255. (11) Heyraud, A.; Gey, C.; Leonard, C.; Rochas, C.; Girond, S.; Kloareg, B. Carbohydr. Res. 1996, 289, 11. (12) de Nooy, A. E. J.; Besemer, A. C.; van Bekkum, H.; van Dijk, J. A. P. P.; Smit, J. A. M. Macromolecules 1996, 29, 6541. (13) Ertesvåg, H.; Erlien, F.; Skjåk-Bræk, G.; Rehm, B. H. A.; Valla, S. J. Bacteriol. 1998, 180, 3779. (14) Gacesa, P. FEBS Lett. 1987, 121, 199. (15) Ramstad, M. V.; Ellingsen, T.; Josefsen, K. D.; Høidal, H. K.; Valla, S.; Skjåk-Bræk, G.; Levine, D. V. Enzyme Microb. Technol. 1999, 24, 636.
BM0055394
