3720
J. Am. Chem. SOC.1991, 113, 3720-3727
Conformational Behavior of Sucrose and Its Deoxy Analogue in Water as Determined by NMR and Molecular Modeling Catherine Herv6 du Penhoat,t Anne Imberty,*J Nathalie Roques,! V6ronique Michon,+ Julio Mentech,g Gerard Descotes,I and Serge Perez$ Contribution from the Service RMN du Dgpartement de Chimie, U.A. 01 1 1 0. E.N.S., 46 rue d'Ulm. 75230 Paris, France, Laboratoire de Physicochimie des MacromoMcules, INRA, BP527, 44026 Nantes, France, Laboratoire Beghin-Say. E.S.C.I.L.,43 Bd du 1 1 Novembre 1918, 69622 Villeurbanne, France, and Laboratoire de Chimie Organique II. E.S.C.I.L., 43 Bd du 1 1 Novembre 1918, 69622 Villeurbanne, France. Received March 13, 1990
Abstract: The conformational behavior of aqueous sucrose and its 2-deoxy analogue were studied by NMR and computerized molecular modeling. 'H steady-state NOE and NOESY data are reported along with long-range W-'H coupling constants. In modeling calculations, a full force field energy minimization was used to obtain the initial residue geometry, followed by a rigid residue approximation in which the glycosidic dihedral angles and the methoxyl group orientationsare varied. Theoretical steady-state NOEs are calculated by a full spin relaxation matrix method, and 3JC-Hdata are correlated with the glycosidic torsional angle. The data do not support a single conformation model, and only conformational averaging can give a good agreement between theoretical and experimental data. The inclusion of hydrogen bonding in the force field does not affect the statistical weights of calculated NOEs, and the similar values of observed NOEs for sucrose and the 2-deoxy analogue argue against the importance of hydrogen bonding in sucrose conformation.
Introduction The molecular conformation in crystalline sucrose is known unambiguously through X-ray' and neutron diffraction2studies. However, the conformational behavior of sucrose in aqueous solution is still controversial, despite numerous experimentalf" and modelir~g~r'~ studies. In most of those e f f o r t ~ , ~ Jsucrose ~ - ' ~ was assumed to be nearly spherical, similar to its shape in crystalline sucrose, and quite rigid. However, when the crystal structures of sucrase,'J sucrose-salt comple~es,'~J~ and oligosaccharides that contain sucrosyl residueslbZ are examined, the (1 2) glycosidic linkages between a-glucopyranose and 0-fructofuranose exhibit wide ranges (see Table I). Recent molecular mechanics and dynamics s t ~ d i e s identified ~ ~ v ~ ~ three low-energy conformations for sucrose, providing further support for the concept of flexible linkages in this disaccharide. Besides the flexibility of sucrose, another controversy is the persistence in solution of the 0-2g-O-If and 0-5g-O-6f hydrogen bonds found in the crystal. Mathlouti et al.3-Sinterpreted X-ray and Raman data to show that the number of intramolecular hydrogen bonds depends on the concentration of sucrose, with no hydrogen bonding at low concentrations. Bock and Lemieux"J2 argued, supported by modeling and detailed I3C TI and NOE measurements, that dilute aqueous sucrose has one intramolecular hydrogen bond. Finally, based on their N M R study of sucrose in DMSO, Christofides and Davies6v7suggested that two intramolecular hydrogen bonds, namely 0-2g-0-1 f and 0-2g-0-3f, compete with each other. Because of these observations and a report that the crystalline conformation cannot account for all of the N M R data,I0 we decided that further study was needed. Interpreting the solution behavior of sucrose from N M R data without an assumption of rigid conformation is an ambitious task, since NMR parameters reflect only a "virtual" conformati~n.~~ Recently, Cumming and Carverz6 combined N M R data with results from computerized molecular modeling that accommodates conformational flexibility. Further analyses of various di- and trisaccharides have mostly used NOE value^^'-^^ although constants from coupling through the glycosidic linkage were also This modeling procedure is used in the present work. The extent of intramolecular hydrogen bonding is probed in a parallel study of 2-deoxysucrose, a compound that cannot form the 0-1f-O-2g hydrogen bond. Also, the
-
'Service RMN du DEpartement de Chimie. Laboratoire de Physicochimie des MacromolEcules. 1 Laboratoire Beghin-Say. E.S.C.I.L. Laboratoire de Chimie Organique 11. E.S.C.I.L.
*
0002-7863/91/1513-3720$02.50/0
potential energy functions were used with and without a term for hydrogen bonding.
(1) Hanson, J. C.; Sieker, L. C.; Jensen, L. H. Acta Crystallogr. 1973, B29, 797-808. (2) Brown, G. M.; Levy, H. A.; Acta Crystallogr. 1973, 829, 790-791. (3) Mathlouthi, M. Carbohydr. Res. 1981, 91, 113-123. (4) Mathlouthi, M.; Luu, D. V. Carbohydr. Res. 1980,81, 203-212. (5) Mathlouthi, M.; Luu,C.; Meffroy-Biget, A. M.; Luu, D. V. Carbohydr. Res. 1980, 81, 213-233. (6) Christofides, J. C.; Davis, D. B. J. Chem. Soc., Chem. Commun. 1985, 1533-1534. (7) Davies, D. B.; Christofides, J. C. Carbohydr. Res. 1987,163,269-274. ( 8 ) McCain, D. C.; Markley, J. L. Carbohydr. Res. 1986, 152, 73-80. (9) McCain, D. C.; Markley, J. L. J . Am. Chem. Soc. 1986, 108, 4259-4264. (10)Mulloy, B.; Frenkiel, T. A.; Davis, D. B. Carbohydr. Res. 1988,184, 39-46. (11) Lemieux, R. U.;Bock, K. Jpn. J., Antibiot. 1979, XXXII Suppl., S163-Sl72. (12) Bock, K.; Lemieux, R. U. Carbohydr. Res. 1982, 100, 63-74. (1 3) Giacomini, M.; Pullman, B.; Maigret, B. Theor. Chim. Acta 1970, 19, 347-364. (14) Accorsi, C. A.; Bellucci, F.; Bertolasi, V.; Ferretti, V.; Gilli, G. Carbohydr. Res. 1989, 191, 105-116. (15) Accorsi, C. A.; Bertolasi, V.; Ferretti, V.; Gilli, G. Carbohydr. Res. 1989, 191, 91-104. (16) Berman, H. M. Acta Crystallogr. 1970, 826, 290-299. (17) Rohrer, D. C. Acta Crystallogr. 1972, B28, 425-433. (18)Jeffrey, G. A.; Park, Y.J. Acta Crystallonr. 1972, B28, 257-261. (1 9) Avenel, D.; Neuman, A.; Gillier-Pandraud, H. Acta Crystallogr. 1976,832, 2598-2605. (20) Bequart, J.; Neuman, A.; Gillier-Pandraud, H.Carbohydr. Res. 1982, 1 1 1 , 9-21. (21) Ferretti, V.;Bertolasi, V.; Gilli, G . Acta Crystallogr. 1984, C40, 5 3 1-53 5. (22)Gilardi, R.; Flippen-Anderson, J. L. Acta Crystallogr. 1987, C43, 806-808. (23) Tran, V.;Brady, J. W. Biopolymers 1990, 29, 961-976. (24) Tran, V.; Brady, J. W. Biopolymers 1990, 29, 977-997. (25) Jardetsky, 0.Biochim. Biophys. Acta 1980, 621, 227-232. (26)Cumming, D. A,; Carver, J. P. Biochemistry 1987, 26, 6664-6616. (27) Breg, J.; Kroon-Batenburg, L. M. J.; Strecker, G.; Montreuil, J.; Vliegenthart, J. F. G. Eur. J. Biochem. 1989, 178, 727-739. (28) Lipkind, G. M.; Shashkov, A. S.; Nechaev, 0. A.; Torgov, V. I.; Shibaev, V. M.; Kochetkov, N. K. Carbohydr. Res. 1989, 195, 11-25. (29)Carver, J. P.; Mandel, D.; Michnick, S. W.; Imberty, A.; Brady. J. W. In Computer Modeling of Carbohydrate Molecules; French, A. D., Brady, J. W., Eds.; ACS Symposium Series; American Chemical Society: Washington, DC, 1990; pp 266-280.
0 1991 American Chemical Society
Conformational Behavior of Sucrose and Its Deoxy Analogue
J. Am. Chem. SOC.,Vol. 113, No. 10, 1991 3121
Table I. Review of the Crystal Structures of Sucrose, Sucrose-Salts, and Other Molecules Containing the Sucrose Moiety, along with the Geometrical Features of Interest and the Observed Hydrogen Bonds nlucose ._.... fructose ~. ring ring
~
ref 1 2 16 17 18 19 20 21 22
year 1963 1973 1970 1972 1972 I978 1982 1984 1987
compound sucrose
(deg) 114.3
(deg) 107.8
9 (deg)
wg
-44.8
raffinose planteose 1-kestose melezitose I melezitose I1 6-kestose stachyose
122.2 118.9 119.4 119.3 115.3 121.3 117.7
81.7 108.6 84.6 99.8 109.6 89.6 108.8
14 15
1989 1989
NaBr, sucrose 3Na1, 2 sucrose
116.7 118.8 117.9
99.8 79.9 79.6
7,
0.11
GG
TG
11.4 -26.7 -65.9 -30.7 -43.4 -54.5 -47.9
GG GT GG GG GG GT GG
GT GT GT GG GT GT GG
GG GG TG GG GG GG TG
-46.1 -66.6 -61.6
GG GT GT
GG GT GT
GG TG TG
0.61
Figure 1. Sucrose with labels for the atoms and torsional angles of interest. Hydroxyl hydrogen atoms are not shown.
Materials and Methods Material. 2-Deoxysucrose was prepared from 1’,2:4,6-di-Oisopropylidene-3,3’,4’,6’-tetra-Oacetylsucroseaccording to a literature procedure?’ Solutions of this compound and reagent grade sucrose were prepared in D,O (99.96%, SST). Nomenclature. Sucrose and its atomic labels of interest are in Figure 1. The conformations about the glycosidic linkage bonds are described by the following torsion angles
a = e(o-sg-~-ig-o-ig 0.1 s), low senl L sitivity precluded quantitative measurements of the initial NOE Figure 2. Unsymmetrized, phase-sensitive 400-MHz IH NOESY specbuildup rate. In the case of H-2g and H-4g, the two interresidue trum of 0.06 M sucrwe in D20(AmH = 4.8 ppm). On the corresponding NOEs observed in the present work have been reported by Sanders 1D spectrum (above) H-lg, H-4f, H-If, and H-2g are labeled. Summed and Hunters3 for sucrose octaacetate. The value of 3J~.2f-~.lgw1 (insert) and o2(below) subspectra with H-lg on the diagonal are also measured for sucrose, 4.2 Hz, is similar to the one reported by given. Mulloy et al.,'O 3.8 Hz, and differs from the value, 5.3 Hz, predicted for the crystal structure by these authors with use of a unaffected. The crystallographically observed 0-1f-0-2g bond Karplus-type relationship. Again, as N M R parameters reflect appears on the edge of the S2 zone and extends this zone somewhat virtual conformations, this suggests that aqueous sucrose is not compared to the map made without hydrogen bonds. This bond adequately described by the crystal conformation. can occur only when 0 - l f has a TG orientation. Hydrogen bonds Molecular Modeling. Of the primary alcohol group orientations, involving 0-6g or 0-6f are not indicated on the energy maps or only the positions of 0-1f have much effect on the potential energy the molecular drawings. surfaces. The maps for GT-GT-GG are shown in Figure 3a,b, Figure 4 displays the analogous energy maps and structural and the maps for GT-GT-TG are in Figure 3c,d. In Figure 3c,d, drawings for 2deoxysucrose. The main difference between Figures the zones of potential hydrogen bonds are indicated, and the 4a and 3a is the appearance of a new, low-energy region labeled location of each low-energy minimum (Sl-S5) is shown in Figure D6 that can occur when 0-1f has a GG position. The drawing 3c. The main effect of the 0 - l f position is on the S3 minimum. labeled D6GG shows that this conformer exists owing to the On the GT-GT-TG maps, the S3 minimum is isolated from the replacement of the hydroxyl group at C-2g by a less bulky hyother four minima, and the energy is higher than on the corredrogen atom, allowing closer approach of the two residues. Even sponding GT-GT-GG maps. less influence from interresidue hydrogen bonding is observed for Maps computed with and without the hydrogen-bonding po2-deoxysucrose. Only D3 and D4 can form hydrogen bonds tential have the same number and locations of minima. Therefore, (again, hydrogen bonds involving 0-6g and 0-6f are not conit appears that intraresidue hydrogen bonding plays at most a sidered). minor role in determining the conformation of isolated sucrose Combination of NMR and Modeling Results. The observed molecules. This minor role might include deepening of the S3, interresidue NOE values and 3JC-H are compared in Table Iv with S4, and S5 wells, while the depths of the S1 and S2 wells are values computed from the low-energy conformers of sucrose having 0 - l f in GG and TG positions. Table IV also includes some interesting intraresidue NOES. The range (0.1-26%) for the NOE (50) Dais, P.; Perlin, A. S.Ado. Carbohydr. Chem. Biochem. 1987, 45, calculated for H-4f when H-lg is saturated shows how sensitive 125-168. (51) French, A. D.;Tran, V. Biopolymers 1990, 29, 1599-1611. NOES can be to the glycosidic linkage conformation. In contrast, (52) Haasnoot, C. A. G.; De Leeuw, F. A. A. M.; Altona, C. Tetrahedron the intraresidue H-2g NOE is almost constant. 1980, 36, 2783-2792. Of the individual low-energy forms, S3 GG provides calculated (53) Sanders, J. K. M.; Hunter, B.K. In Modern NMR Specrroscopy. A NOEs that agree best with the observed values in Table IV,but Guide for Chemisrs; Oxford Press University: 1987; pp 282-297.
I
I
t!
Heme du Penhoat et al.
3124 J . Am. Chem. SOC.,Vol. 113, No. 10, 1991 100
0 0-2g
....0 5 1
0
Y Y .looO
-100
I
t -2001
@
-zoo-@ I
1
I I I I I I ,
-100
0
-100
0
I
100
200
100
200
100 laat
0
Y
-200
'o,, .loo
,I
, ,
0
100
0
200
-100
-200
0
U
Figure 3. Isoenergy maps of the sucrose molecule as a function of the @ and q torsion angles. The hydroxymethyl groups were fixed in the orientations GT-GT-GG (referring to wg-wf-xf) in panels a and b and GT-GT-TG in panels c and d. Energies in panels b and d include hydrogen bond contributions. In all cases, isoenergy contours are drawn at increments of 1 kcal/mol with respect to the absolute minimum. The drawings of the five low-energy conformers (Sl-S5)referred to @ and q values labeled on map 4c. The crystal structure conformation (SX)also is labeled on map 3d.
Conformational Behavior of Sucrose and Its Deoxy Analogue
J . Am. Chem. SOC.,Vol. 113, No. 10, 1991 3125
to
-m
-100
0
100
200
100
200
0 100
0
Y
Y D1
5
-100
-200
-100
0
0
Figure 4. Isoenergy maps of the 2-deoxysucrose molecule as a function of the and q torsion angles. The different panels have the same legends as in Figure 4. The drawings of the five low-energy conformers corresponding to a TG orientation of xf (Dl-D5) referred to and q values labeled on map 4c; the conformer S6, with a GG orientation of xf, is labeled in Figure 4a.
the 2.6% difference for the H-lP-H-lg NOESis not satisfactory. Figure 5 depicts the conformational zones wherein the values of two observed interresidue NOE would agree with the observed
values, superimposed on the maps of conformational energy for the GG and TG positions of 0 - l f . The failure of these zones to intersect shows that there is no combination that can si-
Herut du Penhoat et al.
3726 J . Am. Chem. SOC.,Vol. 113. No. 10, 1991 100
0
Y
Y
-100
0
-100
100
200
0
-100
0
100
200
0
Figure 5. NOE contour surfaces superimposed on isoenergy maps of the sucrme molecule. The hydroxymethyl groups are in the orientation GT-GT-GG (referring to wg-wf-xf) and GT-GT-TG in panels a and b, respectively. The isoenergy contours are drawn as in Figure 3. Two NOE contour surfaces are displayed. They are limited by two iso-NOE contours corresponding to the observed value f20%. The NOE effect observed for H-If upon saturation of H-lg (0.07) is n l , and n2 is the NOE effect observed on H-lg upon saturation of H-4f (0.015).
Table IV. NOE Values (%) and JH.,,-C.2r Coupling Constant (Hz) Calculated for Some Low-Energy Minima of the Sucrose Molecule and for Average Over All the Potential Energy Surfaces"
SI GG 65 -75
Xf
s2 TG 70 -70
GG 90 -25
s3
TG 90 -25
GG 80 -160
s4
TG 80 -165
s5
TG
GG
AM
AMH
exptl
7.0
TG 150 5
GG 100
100
40
40
0.2 26.2 18.1 0.2 5.8 17.9
0.2 26.1 18.1 0.2 10.1 17.9
0.3 15.7 18.3 0.3 5.9 9.2
0.2 15.7 18.3 0.2 10.3 9.3
7.0 1.1 19.1 11.3 8.6 0.8
19.1 11.3 8.6 0.8
7.0 0.7 18.0 9.5 7.0 1.5
Coupling Constants 3.7 3.7 5.1
5.1
4.0
4.0
3.5
4.1
4.2
150 5
NOE Values
(SIb
H-lg H-lf H-4f
(d)' H-If H-4f H-2g H-lg H-3f H-lg
JH.I,-C:-II
12.2 0.2 18.9 15.3 5.9 0.1
10.6 0.1 19.0 15.1 10.2 0.1
2.4 0.3 19.0 5.8 5.9 0.5
2.2 0.3 19.1 5.3 10.2 0.5
2.3
2.7
4.5
4.5
6.9 0.8 18.9 12.1 5.3 0.4
0.7 1.0 19.2 2.1 9.8 1.0
1.1
"AM and AMH, averaging over all maps without and with the contribution of hydrogen bonds, respectively; exptl, experimental values. b ( s ) saturated proton. (d) detected proton.
Table V. Values ('3%) Calculated for Some Low-Energy Minima of the 2-Deoxysucrose Molecule and for Averaging for All the Potential Energy Surfaces" D1 D2 D3 D4 D5 D6 AM AMH exptl Xf GG TG GG TG GG TG GG TG GG TG GG @ 65 65 90 90 80 80 105 100 170 170 155 -80 -80 -20 -20 -160 -165 30 40 -20 -20 -135 NOE Values
(SIb
(dIC H-If
7.3 0.7 0.2 0.1 0.1 0.0 17.5 6.7 6.7 13.3 9.8 1.8 1.7 29.4 8.2 1.3 1.3 0.2 0.2 0.4 0.4 0.9 1.2 21.6 6.2 0.2 2.0 1.4 2.4 2.4 2.7 2.2 H-2ge 2.7 2.7 2.7 2.7 2.7 2.7 2.7 4.5 4.2 4.5 4.6 4.5 4.6 H-2ga 4.4 4.4 4.4 4.4 4.4 4.4 4.3 0.1 15.3 8.9 8.9 1.5 0.1 0.1 0.1 H-lf H-lg 13.9 12.6 3.2 2.9 10.3 8.7 5.0 8.7 5.2 10.2 5.6 10.4 5.7 10.5 H-3f 5.6 5.7 10.5 10.5 0.7 0.7 0.8 9.5 16.5 4.0 3.2 0.1 H-4f H-lg 0.1 0.5 0.4 0.1 0.5 3.3 3.3 3.2 3.5 3.5 3.7 H-5f 3.5 3.6 3.6 3.6 3.7 3.7 3.3 "AM and AMH, averaging over all maps without and with the contribution of hydrogen bonds, respectively; exptl, experimental values. in Table IV. c(d) as in Table IV. H-lg
H-4f
multaneously generate both enhancements. Therefore, some sort of conformational averaging over the five low-energy conformers is required to explain the N M R results. A simple averaging of the values for the five conformers does not suffice, and complete integration over all the maps, described in the methods section, was then employed. Those averages are given in Table IV under the headings AM and AMH, based on the maps without and with hydrogen bonding, respectively. The NO& were
5.0 2.2 2.5 5.0 6.5 7.0 1.5 4.5 b ( s ) as
not affected by the presence or absence of hydrogen bonding in the maps used for averaging, but the coupling constants fit the experiment better when hydrogen bonding was included. Overall, the AMH values correspond reasonably well with the experimental results. The NOE value observed for H-3f upon irradiation of H-lf is influenced by variations in the 0-1f hydroxymethyl group orientation but not by variations in the conformation at the glycosidic
Conformational Behavior of Sucrose and Its Deoxy Analogue linkage (Table IV). This NOE effect has values of 6% for the GG position and 10% for the TG position, so both positions must be considered to explain the observed value of 7%. The G T position of 0 - I f gives values above 11%, justifying its omission from the above analyses. Table V reports values analogous to those in Table IV for 2-deoxysucrose. Again, there is no satisfactory fit between the values calculated for any single conformer and the experimental results. The bases for this conclusion are the variations in NOES for H-lf and H-4f that are observed when H-lg is saturated and the variation in the NOE for H-lg when H-lf is saturated. The averaged predicted values, invariant with the presence or absence of hydrogen bonding, again provide a better agreement with the experimental results. Hydrogen Bonding. The potential energy surfaces and the N M R data for sucrose are strikingly similar to the surfaces and data of the 2-deoxy analogue. Thus, the crystallographically observed 0-1f-0-2g hydrogen bond must not significantly influence the solution conformation of sucrose. Exchange of hydrogen-bonding partners between the neighboring intramolecular hydroxyl group and solvent water is probably a factor, but sucrose also has an unusual relationship between the energetically accessible conformations of the glycosidic linkage and conformations that allow interresidue hydrogen bonds. The disaccharides maltose and cellobiose have large intersections of energetically accessible areas that can be further stabilized by hydrogen bonds, where Figures 3b,d and 4b,d show that only minor portions of the allowed domains for sucrose are compatible with hydrogen bonding. Figure 3d also demonstrates that, in regard to the hydrogen bonding proposed for DMSO solutions,6 with 0-2g bonding to either 0 - l f or 0-3f, these two bonds cannot occur with the same conformation of the glycosidic linkage. Comparison with Relaxed Residue Mapping. Recently, “relaxed residue” conformational analyses of d i s a ~ c h a r i d e s * ’ ~ ~have ~J~”~ been reported. In relaxed maps, the individual atomic positions are allowed to adjust at each increment of the glycosidic linkage torsion angles, thereby relieving any artificial contribution to the calculated energy that results from residue rigidity. The relaxed (54) Pangborn, W.; Langs, D.; Perez, S. Int. J . Biol. Macromol. 1985, 7, 363-369. ( 5 5 ) Tran, V.;Buleon, A.; Imberty, A.; Perez, S. Biopolymers 1989, 28, 679-690. (56) French, A. D.Carbohydr. Res. 1989,188, 206-21 1. (57) Imberty, A.; Tran, V.; Perez, S. J . Comput. Chem. 1989,11,205-216.
J. Am. Chem. SOC..Vol. 113, No. 10, 1991 3121 map of is consistent with our rigid maps in terms of overall shape and location of the minima but is better for understanding the pathways of conformational interchange among the minima. These intuitively superior, relaxed energy m a p have been criticized4’ for depicting regions of low energy that are too extensive, compared to experimental values in aqueous solution. In studies such as the present one, where solvent is not explicitly considered, it may be that rigid residue analysis provides a more realistic result. Conclusions In regards to the N M R data of sucrose, the major difference between the present study and previous one12 is the observation of a small interresidue NOE. This dipolar interaction, which was observed for a wide range of mixing times, is not compatible with the conformation observed in crystalline sucrose. While the crystalline form has been considered to be the sole conformation in aqueous s ~ l u t i o n ,our ~ ~ studies ~ ~ ’ ~ show that weighted average NOE and coupling constants, computed over all energy maps, provide better agreement with experiment. The influence of hydrogen bonds on the solution behavior of sucrose requires special attention because of its presumed importance as well as the controversial interpretation of the results stemming from different techniques. Because conformational maps varied little regardless of whether hydrogen-bonding potentials were included, we conclude that hydrogen bonding is not important in determining conformations of sucrose in aqueous solution. This was supported by the strong similarity of the conformational maps for sucrose with those of the deoxy analogue which is unable to form the crystallographically observed hydrogen bond. The NOE and coupling constant data observed for the two molecules are also similar, further refuting the importance of hydrogen bonding in aqueous solution. Acknowledgment. The authors are much endebted to Prof. J. Carver and Tom Lew (University of Toronto) for access to the computer programs for the calculation of the NOE values and graphic representation of the energy maps. Appreciation is extended to Dr. A. D.French (USDA, New Orleans) for his careful reading of the manuscript. Financial support from the Pierre et Marie Curie University, the CNRS (UA 11 lo), and the INRA is acknowledged. One of us (N.R.) acknowledges the support from French Ministry of Research and Technology. Registry No. Sucrose, 57-50-1;2-deoxysucrose, 6852-68-2.
