Received 11 January 2011
Methyl 2-acetamido-2-deoxy--D-glucopyranoside dihydrate and methyl 2-formamido-2-deoxy--D-glucopyranoside
Methyl 2-acetamido-2-deoxy--D-glucopyranoside (-GlcNAcOCH3), (I), crystallizes from water as a dihydrate, C9H17NO6·H2O, containing two independent molecules [denoted (IA) and (IB)] in the asymmetric unit, whereas the crystal structure of methyl 2-formamido-2-deoxy--D-glucopyranoside (-GlcNFmOCH3), (II), C8H15NO6, also obtained from water, is devoid of solvent water molecules. The two molecules of (I) assume distorted 4C1 chair conformations. Values of for (IA) and (IB) indicate ring distortions towards BC2,C5 and C3,O5B, respectively. By comparison, (II) shows considerably more ring distortion than molecules (IA) and (IB), despite the less bulky N-acyl side chain. Distortion towards BC2,C5 was observed for (II), similar to the findings for (IA). The amide bond conformation in each of (IA), (IB) and (II) is trans, and the conformation about the C-N bond is anti (C-H is approximately anti to N-H), although the conformation about the latter bond within this group varies by 16°. The conformation of the exocyclic hydroxymethyl group was found to be gt in each of (IA), (IB) and (II). Comparison of the X-ray structures of (I) and (II) with those of other GlcNAc mono- and disaccharides shows that GlcNAc aldohexopyranosyl rings can be distorted over a wide range of geometries in the solid state.
Acylation is an important covalent modification that affects the biological functions of saccharides and other biomolecules. Two types of saccharide acylation are common in biological systems, N- and O-acylation, and recent work has shown that different O-acylation patterns affect biological function. For example, this type of covalent control has been described recently for glycopeptidolipids involved in signaling through Toll-like receptors (Sweet et al., 2008). N-Acylation is found in biologically relevant monosaccharides, such as N-acetyl-D-glucosamine, N-acetyl-D-galactosamine and N-acetylneuraminic acid. It has been shown recently that cis-trans isomerization (CTI) of the amide bond in N-acylated sugars can be detected in aqueous solution, with the cis/trans ratio dependent on, among other factors, the anomeric configuration of the saccharide (Hu, Zhang et al., 2010). For example, Ktrans/cis is 60 for methyl N-acetyl--D-glucosaminide and 38 for methyl N-acetyl--D-glucosaminide at 326 K. In support of NMR studies of saccharide CTI and of the parameterization of NMR J couplings within saccharide exocyclic N-acyl fragments (Hu, Carmichael & Serianni, 2010), we undertook the crystallization of methyl 2-acetamido-2-deoxy--D-glucopyranoside dihydrate, (I), and methyl 2-formamido-2-deoxy--D-glucopyranoside, (II). Their crystal structures, reported here, are compared with the structurally related compounds N-acetyl--D-glucosamine (2-acetamido-2-deoxy--D-glucopyranose), (III) (-GlcNAcOH; Mo & Jensen, 1975), -chitobiose [2-acetamido-2-deoxy--D-glucopyranosyl-(14)-2-acetamido-2-deoxy--D-glucopyranose], (IV) (-GlcNAcOH and -GlcNAcOR; Mo, 1979), -chitobiose [2-acetamido-2-deoxy--D-glucopyranosyl-(14)-2-acetamido-2-deoxy--D-glucopyranose], (V) (-GlcNAcOH and -GlcNAcOR; Mo & Jensen, 1978) and methyl -D-glucopyranoside, (VI) (-GlcOCH3; Jeffrey & Takagi, 1977).
Crystals of (I) and (II) were obtained from aqueous solutions by slow evaporation at room temperature. Compound (I) crystallizes with two independent molecules in the asymmetric unit [denoted (IA) and (IB)] and two solvent water molecules (Fig. 1a and 1b), whereas the unit cell of (II) contains only one molecule of the saccharide and no solvent water molecules (Fig. 2).
A comparison of selected structural parameters in compounds (I)-(VI) is shown in Table 1. The average C1-C2 and C2-C3 bond lengths in (IA), (IB) and (II) are 1.532 (1) and 1.534 (4) Å, respectively. These values are 0.01-0.02 Å greater than corresponding values in (VI), which are 1.522 (4) and 1.515 (2) Å, respectively, presumably reflecting the different substitutions at C2. All other corresponding C-C bonds in (IA), (IB) and (VI) have very similar lengths. Within the same three structures, the exocyclic C5-C6 bond appears to be the shortest C-C bond [1.516 (3) Å]. Similar inspections of the C1-O1 and C1-O5 bonds suggest a slight lengthening of the former and a slight shortening of the latter in N-acyl sugars (IA) and (IB) compared with the related bonds in the simple glycoside (VI).
In all structures in Table 1 bearing an N-acyl group, the average C2-N1 bond length is 1.453 (5) Å. This bond is 0.03 Å longer than the average exocyclic (non-anomeric) C-OH bond length in a pyranosyl ring, which is 1.427 (4) Å in the same data set.
Within the N-acyl exocyclic fragment, the C8-O8 bond averages 1.236 (5) Å, with no discernible difference between N-acetyl and N-formyl groups. The C8-C9 bond averages 1.502 (8) Å, which is 0.02-0.03 Å shorter than the endocyclic C-C bonds found in the pyranosyl rings.
The N-acyl side chains in (IA), (IB) and (II) contain the amide bond N1-C8 in a trans conformation, i.e. with C2-N1-C8-C9 torsion angles of ±179.1 (2)° in both (IA) and (IB), and an average C2-N1-C8-O8 torsion angle of 0.77 (2)° in (IA), (IB) and (II). These torsion angles also demonstrate that the amide group is planar. In all three structures, the conformation about the C2-N1 bond is anti, i.e. atom N1H is roughly anti to H2, which is consistent with the behavior reported in solution (Zhu et al., 2006). However, inspection of the C1-C2-N1-C8 and C3-C2-N1-C8 torsion angles in (IA), (IB) and (II) reveals a range of 91-108° in the former and -128 to -144° in the latter, indicating some flexibility about the C2-N1 bond in the solid state.
The C5-O5-C1 bond angles in (I)-(VI) appear to depend on anomeric configuration, with -anomers yielding an average value of 112.0 (6)° and -anomers an average of 114.7 (3)°. Within the full data set, the largest C-C-C bond angle within any given structure (-anomers only) is C4-C5-C6, which averages 113.8 (11)°. The two bond angles that incorporate the carbonyl O atom, O8-C8-C9 and O8-C8-N1, are very similar in each structure and average 122.5 (14)° in the full data set. These results contrast with the remaining angle, N1-C8-C9, which is uniformly smaller than the others in any given structure and averages 116.0 (4)° in the full data set.
Within (IA), (IB) and (II), the endocyclic ring torsion angles vary from 41 to 72° (absolute values), indicating that aldopyranosyl rings containing N-acyl substiutents at C2 are distorted. It is noteworthy that the extreme angles within this group are observed in (II), where the C1-C2-C3-C4 and C1-O5-C5-C4 torsion angles are -41.09 (15) and 71.61 (13)°, respectively. Within (III)-(VI), these torsion angles range from 48 to 68° (absolute values). A more quantitative treatment of these torsion angles is provided by the Cremer-Pople parameters calculated for (I)-(VI) (Table 2; Cremer & Pople, 1975). The most distorted ring within (I)-(VI) is that of (II), which yields a value of 16.59 (14)°. The least distorted ring is found in (Va), where = 0.9 (3)°, and the ring is almost an ideal 4C1 chair. The direction of distortion varies widely within this series of compounds. For (IA) and (II), = 302.0 (12) and 314.4 (5)°, respectively, suggesting distortion towards BC2,C5. In contrast, for (IB) is 0 (2)°, or a C3,O5B distortion. These data show that not only can aldopyranosyl rings be substantially distorted when bearing an N-acyl functionality at C2, but the direction of distortion can also vary widely, with various boat (BC2,C5, C3,O5B and BC1,C4) and twist-boat (C1TBC5, C3TBC1, O5TBC2 and C5TBC1) conformations represented.
The exocyclic hydroxymethyl conformation in each of (IA), (IB) and (II) is gt (C4 anti to O6), but the gg conformation (H5 anti to O6) is observed in each of (III), (IVa), (Va) and (Vb). These data show that gg and gt conformations are favored in GlcNAc/NFm aldopyranosyl rings in the solid state, a behavior which is expected to mimic that in solution, based on related studies of the simpler Glcp anomers (Thibaudeau et al., 2004).
The structure of (I) forms hydrogen-bonded pairs of (IA) and (IB) molecules of through hydroxy atom O6A to atom O6B. In turn, atom O6B forms a hydrogen bond to atom O6Aiii of the next pair related by translation along the a axis. In addition, the amide groups form hydrogen bonds, although this is to the same GlcNAc molecule related by translation along the a axis (all symmetry codes as in Table 3). Amide atom N1A of (IA) has a bifurcated hydrogen bond shared between the adjacent amide carbonyl atom O8Ai and methoxy atom O1Ai. Amide atom N1B of (IB) has a single hydrogen bond to amide atom O8Bi. This association results in chains of pairs of (IA) and (IB) that run through the lattice parallel to the a axis. These chains are hydrogen bonded to other chains via hydroxy-hydroxy (O4BO4Aii) or hydroxy-water interactions. The water molecules are located in a hydrophilic channel lined with amide O atoms (O8B) and hydroxy groups, again parallel to the a axis. One motif that is apparent is a hydrogen-bonded ring formed by four sets of these hydrogen-bonded chains. These rings are oriented around a hydrophobic channel formed by the methyl groups of the methoxy moieties and the acetamide methyl group (Fig. 3; see Table 3 for specific details). Overall, this structure forms a highly hydrogen-bonded three-dimensional network of water and amidosaccharide molecules.
The structure of (II) also forms a three-dimensional network of hydrogen-bonded molecules. The network is formed from sheets of (II) that have hydrogen bonds through atom O4 to atom O8iii running parallel to the b axis (related by the 21 screw axis) and from atom O6 to atom O4iv along the c axis, and thus these sheets lie in the bc plane of the lattice. Propagation of these sheets into the third dimension is via hydrogen bonds from atom O3 to atom O6ii and from atom N1 to atom O8i of sheets in the next layer along the a axis (Fig. 4; all symmetry codes as in Table 4).
| || Figure 1 |
Labeling schemes for (a) molecule (IA) with water molecules and (b) molecule (IB). Displacement ellipsoids are drawn at the 50% probability level.
| || Figure 2 |
The atom-labeling scheme for compound (II). Displacement ellipsoids are drawn at the 50% probability level.
| || Figure 3 |
Hydrogen-bonding scheme for molecules (IA) and (IB), viewed along the a axis. Dashed lines (blue in the electronic version of the paper) represent hydrogen bonds.
| || Figure 4 |
Hydrogen-bonding scheme for (II), viewed along the a axis. Dashed lines (blue in the electronic version of the paper) represent hydrogen bonds.
Compounds (I) and (II) were prepared as described recently in the literature (Hu, Zhang et al., 2010). Both compounds were crystallized from water at room temperature, giving long colorless needle-like crystals that were harvested for X-ray analysis.
The absolute configurations of (I) and (II) were determined both from the known configuration of the starting materials and by comparison of the intensities of Friedel pairs of reflections. However, the Flack parameters were inconclusive [x = -0.4 (2) for (I) and 0.13 (14) for (II); Flack, 1983]. Further confirmation of the configurations was sought by the Hooft analysis, yielding a Hooft y parameter of -0.13 (12) and P2(true) and P3(true) values of 1.000 and 1.000 for (I), and a Hooft y parameter of 0.16 (4) and P2(true) and P3(true) values of 1.000 and 1.000 for (II) (Hooft et al., 2008).
For both structures, the hydroxy, amide and, where applicable, water H atoms were all located from a difference Fourier map and initially included in those positions. The hydroxy and amide H atoms were subsequently constrained to have reasonable geometric X-H bond distances and angles (N-H = 0.88 Å and O-H = 0.84 Å). Where applicable, mild restraints were applied to the water O-H bond distances [0.84 (1) Å]. All C-H bonds were constrained to distances of 0.98-1.00 Å. For all H atoms, Uiso(H) = 1.5Ueq(C) for methyl H atoms or 1.2Ueq(C) for all others.
Data collection: APEX2 (Bruker-Nonius, 2009) for (I); APEX2 (Bruker-Nonius, 2008) for (II). Cell refinement: SAINT (Bruker-Nonius, 2009) for (I); SAINT (Bruker-Nonius, 2008) for (II). Data reduction: SAINT (Bruker-Nonius, 2009) for (I); SAINT (Bruker-Nonius, 2008) for (II). For both compounds, program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008); molecular graphics: XP (Sheldrick, 2008) and POV-Ray (Cason, 2003); software used to prepare material for publication: XCIF (Sheldrick, 2008) and publCIF (Westrip, 2010).
Supplementary data for this paper are available from the IUCr electronic archives (Reference: FA3249 ). Services for accessing these data are described at the back of the journal.
Bruker-Nonius (2008). APEX2 (Version 2008-6) and SAINT (Version 7.53A). Bruker-Nonius AXS Inc., Madison, Wisconsin, USA.
Bruker-Nonius (2009). APEX2 (Version 2009-9) and SAINT (Version 7.60A). Bruker-Nonius AXS Inc., Madison, Wisconsin, USA.
Cason, C. J. (2003). POV-Ray. Version 3.6.2. Persistence of Vision Raytracer Pty. Ltd, Victoria, Australia.
Cremer, D. & Pople, J. A. (1975). J. Am. Chem. Soc. 97, 1354-1358.
Flack, H. D. (1983). Acta Cryst. A39, 876-881.
Hooft, R. W. W., Straver, L. H. & Spek, A. L. (2008). J. Appl. Cryst. 41, 96-103.
Hu, X., Carmichael, I. & Serianni, A. S. (2010). J. Org. Chem. 75, 4899-4910.
Hu, X., Zhang, W., Carmichael, I. & Serianni, A. S. (2010). J. Am. Chem. Soc. 132, 4641-4652.
Jeffrey, G. A. & Takagi, S. (1977). Acta Cryst. B33, 738-742.
Mo, F. (1979). Acta Chem. Scand. Ser. A, 33, 207-218.
Mo, F. & Jensen, L. H. (1975). Acta Cryst. B31, 2867-2873.
Mo, F. & Jensen, L. H. (1978). Acta Cryst. B34, 1562-1569.
Sheldrick, G. M. (2008). Acta Cryst. A64, 112-122.
Sweet, L., Zhang, W., Torres-Fewell, H., Serianni, A. S., Boggess, W. & Schorey, J. (2008). J. Biol. Chem. 283, 33221-33231.
Thibaudeau, C., Stenutz, R., Hertz, B., Klepach, T., Zhao, S., Wu, Q., Carmichael, I. & Serianni, A. S. (2004). J. Am. Chem. Soc. 126, 15668-15685.
Westrip, S. P. (2010). J. Appl. Cryst. 43, 920-925.
Zhu, Y., Pan, Q., Thibaudeau, C., Zhao, S., Carmichael, I. & Serianni, A. S. (2006). J. Org. Chem. 71, 466-479.