MA'AT analysis of the O-glycosidic linkages of oligosaccharides using nonconventional NMR J-couplings: MA'AT and MD models of phi

Abstract

MA'AT analysis (Meredith et al., J. Chem. Inf. Model. 2022, 62, 3135-3141) is a new NMR-based method to treat ensembles of redundant NMR spin-coupling constants (J-couplings) to obtain experiment-based probability distributions of molecular torsion angles in solution. Work reported to date on modeling the conformations of O-glycosidic linkages of oligosaccharides using three conventional J-coupling constraints (² J _COC, ³ J _COCH, ³ J _COCC) has shown that the method gives mean torsion angles and circular standard deviations (CSDs) for psi in very good agreement with those obtained by MD simulation. On the other hand, CSDs for phi determined by MA'AT analysis have consistently been much larger than those determined by MD, calling into question either the reliability of MA'AT analysis or MD to accurately predict this behavior. Prior work has shown that this discrepancy does not stem from the limitations of DFT-based J-coupling equation parameterization where secondary conformational dependencies can introduce uncertainties. The present work re-visits this problem by incorporating a new nonconventional J-coupling constraint into MA'AT analyses of phi, namely, a geminal (two-bond) ² J _CCH J-value that exhibits a strong primary dependence on phi. The latter property pertains explicitly to linkages contributed by GlcNAc pyranosyl rings and pyranosyl rings devoid of substituents at C2 (i.e., deoxy residues) where known secondary contributions to ² J _CCH magnitude caused by C-O bond rotation involving the coupled carbon are negligible or absent. The results show that when ² J _CCH values are added to the analysis, phi CSDs reduce considerably, bringing them into better alignment with those obtained by MD simulation. The cause of the discrepancy when only three conventional J-couplings are used to treat phi appears to be associated with the two-bond ² J _COC, which has properties that make it less effective than the non-conventional ² J _CCH as a discriminator of different conformational models of phi.

Scheme 1. (A) Structure of methyl β-d-galactopyranosyl-(1→4)-β-d-glucopyranoside (methyl β-lactoside (1)), showing the O-glycosidic linkage torsion angles, ϕ and ψ. (B) Expansion of the βGal ring of 1 showing ϕ and θ, and the lone-pair orbitals on O2 whose orientation relative to the C2′–C1′–H1′ coupling pathway affects ²J_C2′,H1′. (C) The same structure as in (B) but replacing the hydroxyl group at C2 with an N-acetyl side-chain. Torsion angle θ is more constrained in 3 than in 2, with the H2′–C2′–N2′–H torsion angle approximating 180° based on MA’AT analysis. (D)–(G) Structures of methyl 2-acetamido-2-deoxy-β-d-[1-¹³C]glucopyranosyl-(1→4)-2-acetamido-2-deoxy-β-d-[4-¹³C]glucopyranoside (4^1′,4), methyl 2-acetamido-2-deoxy-β-d-[1-¹³C] and [2-¹³C]glucopyranosyl-(1→4)-β-d-mannopyranoside (5^1′/2′), methyl 2-acetamido-2-deoxy-β-d-[1-¹³C]glucopyranosyl-(1→2)-α-d-[2-¹³C]mannopyranoside (6^1′,2), and methyl 2-deoxy-β-d-[1-¹³C]arabino-texopyranosyl-(1→4)-β-d-[4-¹³C]glucopyranoside (7^1′,4), showing the θ and ϕ torsion angles. In 1 and 4–7, anomeric carbons are labeled as either 1 or 1′. Superscripts on the compound numbers identify the carbons labeled with ¹³C.

Scheme 2. ϕ-Dependent conventional and nonconventional (in bold) J-couplings in disaccharides 4–7. In this work, the four J-couplings in each compound were used in three different combinations in MA'AT analysis of ϕ (Groups I–III). See Tables S1–S4 (ESI†) and the text for more discussion.

Fig. 1. Plots showing the dependence of the calculated ²J_C2′,H1′ value in disaccharides 4–7 on ϕ. (A) 4. (B) 5. (C) 6. (D) 7. For all four two-bond C2′–C1′–H1′ coupling pathways, the calculated geminal ²J_CCH is positive. In each plot, curves corresponding to the trimmed (blue) and constrained (red) eqn [S1]–[S32] (ESI†) are shown.

Fig. 2. Population distributions of ϕ in 4–7 determined by MA’AT analysis using Groups I (red), II (green) and III (black) ϕ-dependent J-couplings, superimposed on the distributions determined by MD simulation (purple hatched). (A) 4. (B) 5. (C) 6. (D) 7. MA’AT analyses were conducted using constrained eqn [S5]–[S8], [S13]–[S16], [S21]–[S24] and [S29]–[S32] (ESI).

Fig. 3. (A) MA’AT-determined mean values (A) and CSDs (B) for ϕ in 4–7 determined using Groups I (red), II (blue) and III (green) ϕ-dependent J-couplings, compared to the mean value and CSDs obtained from MD simulation (black). Constrained eqn [S5]–[S8], [S13]–[S16], [S21]–[S24] and [S29]–[S32] (ESI†) were used in the MA’AT analyses.

Fig. 4. Plots of the dependencies of ²J_C1′,C4 (A) and ²J_C2′,H1′ (B) in 4 on the H1′–C1′–O1′–C4 torsion angle ϕ. Blue circles, trimmed data; red squares, constrained data. The solid lines are plots of the trimmed (blue) and constrained (red) equations for each J-coupling. Point scatter at discrete values of ϕ is caused by the secondary dependence of the J-coupling on ψ.