Conformational Properties of Aryl S-Glucosides in Solution

Abstract

The conformational study of saccharides and glycomimetics in solution is critical for a comprehensive understanding of their interactions with biological receptors and enabling the design of optimized glycomimetics. Here, we report a nuclear magnetic resonance (NMR) study centered on the conformational properties of the hydroxymethyl group and glycosidic bond of four series of aryl S-glucosides. We found that in acetyl-protected and free aryl S-β-glucosides, the rotational equilibrium around the C5–C6 bond (hydroxymethyl group) exhibits a linear dependence on the electronic properties of the aglycone, namely, as the aryl’s substituent electron-withdrawing character increases, the dominance of the gg rotamer declines and the gt contribution rises. Likewise, the conformational equilibrium around the glycosidic C1–S bond also depends on the aglycone’s electronic properties, where glucosides carrying electron-poor aglycones exhibit stiffer glycosidic bonds in comparison to their electron-rich counterparts. In the case of the α anomers, the aglycone’s effect over the glycosidic bond conformation is like that observed on their β isomers; however, we observe no aglycone’s influence over the hydroxymethyl group conformation in the α-glucosides.

Graphical Abstract

INTRODUCTION

Carbohydrate–protein interactions mediate numerous physiological and pathological processes, including homeostasis,¹ cell differentiation,² infection,³ immune response,⁴ fertilization,⁵ tumor progression, and metastasis,⁶ to name a few. A successful carbohydrate–protein recognition depends on both binding partners fulfilling geometrical, stereochemical, and conformational requirements for optimal interaction, with the conformational factor being of special relevance for the intrinsically flexible sugar ligand. Hence, studying the conformational properties of carbohydrates in solution, as well as bound to receptors, is vital to understanding the molecular mechanisms underlying the transduction of the sugar code into biological functions.⁷ Moreover, studying how the molecular structure defines the conformational properties of saccharides can enable the design of conformationally tailored glycomimetics with improved binding affinities to biological receptors. For instance, the synthesis of conformationally locked glycomimetics, aimed to emulate the bioactive conformation of natural carbohydrate ligands, has proven effective for enhancing the binding affinity and biological performance of several drug candidates.⁸

The flexibility of glycosides primarily rests around the glycosidic bond (C1–O1) and the hydroxymethyl group (C5–C6). In hexopyranoses, the glycosidic bond flexibility is described by the torsional angles (O5–C1–O1–Cx) and ψ (C1–O1–Cx–C(x-1)), while the torsion ω (O5–C5–C6–O6) describes the rotation around C5–C6 (Figure 1). The rotation around the glycosidic C1–O1 bond generates three staggered conformations, named exo–syn, exo–anti, and non-exo (Figure 2A),⁹ whose contributions to the rotational equilibrium are largely governed by the exo–anomeric effect,¹⁰ which in α- and β-pyranosides favors the exo–syn and exo–anti rotamers.

Figure 1.

Flexibility of glycosides.

Figure 2.

Conformations around the glycosidic bond (C1–O) and hydroxymethyl group (C5–C6) in glucose; (A) flexibility around the glycosidic bond and staggered conformers exo–syn, exo–anti, and non-exo; (B) flexibility around the C5–C6 bond (hydroxymethyl group) and staggered conformers gauche–gauche (gg), gauche–trans (gt), and trans–gauche (tg).

Similarly, the rotation around the C5–C6 bond generates three staggered rotamers called gauche–gauche (gg), gauche–trans (gt), and trans–gauche (tg) (Figure 2B).¹¹ The contribution of each to the conformational equilibrium depends on various aspects including the stereo configuration of C4,¹² the steric and polar properties of the C4 and C6 substituents,¹³ the polarity of the media,¹⁴ and the anomeric configuration.¹⁵ Furthermore, in alkyl O-,¹⁶ S-,¹⁷ and C-β-glycosides,¹⁸ the rotamer contributions also depend on the steric properties of the aglycone. In these compounds, a gradual increment of the aglycon’s volume leads to a progressive increase in the gt percent contribution and a simultaneous gg or tg population detriment depending on the stereochemical configuration of the sugar, i.e., glucoside, mannoside, or galactoside. In alkyl β-glycosides, the rotamer populations correlate with the respective alkyl substituents’ Taft’s steric parameters, evidencing a direct association between the aglycone’s volume and the rotation around C5–C6 in these derivatives. From a practical standpoint, this structure-conformation relationship identifies the alkyl substituent as a convenient knob to modulate the hydroxymethyl conformation of alkyl β-glycosides. Motivated by exploring this effect in saccharides carrying aglycones other than alkyl groups, we focus here on the conformational properties of aryl S-glucosides. S-Glycosides are common glycosyl donors used in the assembly of glycosides and oligosaccharides,¹⁹ and given their conformational similarity to O-glycosides²⁰ and enhanced stability toward chemical and enzymatic hydrolysis,²¹ they are also an attractive alternative to design glycomimetic-based drug candidates.

RESULTS AND DISCUSSION

Synthesis and Characterization.

As described in Scheme 1, we prepared the aryl β-glucosides 2a–2g by reacting the respective substituted thiophenols with glucose per acetate 1 using BF₃Et₂O as a catalytic donor activator. To obtain the target β-glucosides in high yields, we conducted the reactions in the β-inducing solvent MeCN.²² To increase the yields of the respective α-anomers α2a–α2g, the reactions were conducted in DCM. The subsequent Zemplén transesterification²³ of acetates 2a–2g and α2a–α2g afforded the respective deprotected aryl S-glucosides 3a–3g and α3a–α3g.

Scheme 1.

Preparation of Aryl S-Glucosides^a

^a(A) Aryl-SH, BF₃Et₂O, MeCN, reflux; (b) aryl-SH, BF₃Et₂O, DCM, reflux; (c) NaOMe, MeOH, then Amberlite IR 120 (H form).

We characterized the S-glucosides by 1D (¹H and APT ¹³C{¹H}) and 2D (COSY, HSQC) NMR experiments. The anomeric configurations were determined from the ³J_H1,H2 coupling constants extracted from ¹H NMR spectra (³J_H1,H2 ~ 10.0 Hz for β-glucosides, ³J_H1,H2 ~ 5.8 for α-derivatives). The ³J_H,H coupling constants observed for the rest of the vicinal protons confirmed the sugar ring conformation anchored in the ⁴C₁ chair. To ensure a correct calculation of the hydroxymethyl rotamer populations, we identified the pro-chiral H6R and H6S protons according to literature data.²⁴ Briefly, in ester-protected glycosides, the proton H6R appears deshielded with respect to H6S (δH6R > δH6S), and the opposite correlation is predicted for deprotected derivatives (δH6S > δH6R). Moreover, for glucose adducts, the value of the ³J_H5,H6R coupling constant is typically larger than ³J_H5,H6S on both ester-protected and free derivatives. These chemical shift/coupling constant correlations were confirmed in the four glucoside series. In glucosides where H6R and H6S signals overlapped, we obtained the ³J_H5,H6 values from the H5 signal, which was isolated in all cases where H6R and H6S appeared overlapped.

NMR and Conformational Study of Aryl S-Glucosides.

Table 1 shows selected ¹H and ¹³C{¹H} NMR data and the calculated rotamer population for aryl S-β-glucosides 2a–2g (acetyl series) in CDCl₃ and 3a–3g (deprotected series) in CD₃OD; Table 2 summarizes the respective data for aryl S-α-glucosides α2a–α2g (acetyl series) and α3a–α3g (deprotected series).

Table 1.

Selected NMR Data and Calculated Rotational Populations (%) for per-Acetylated aryl S-β-Glucosides 2a–2g (400 MHz, CDCl₃) and Deprotected Aryl S-β-Glucosides 3a–3g (400 MHz, CD₃OD)

acetyl-protected β series
	R	δC1	δH1	δH6R	δH6S	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg
2a	OMe	85.7	4.48	4.11a	4.11a	4.2	3.2	62	31	7
2b	Me	85.9	4.56	4.15	4.09	4.6	2.3	63	37	0
2c	H	85.8	4.64	4.16	4.11	5.0	2.7	57	41	2
2d	F	85.3	4.53	4.13a	4.13a	4.2	3.2	62	31	7
2e	Cl	85.2	4.57	4.14	4.11	4.7	2.8	60	37	3
2f	CF3	84.9	4.71	4.17	4.12	5.2	2.5	56	44	0
2g	NO2	84.3	4.79	4.19	4.13	5.3	2.4	55	45	0
deprotected β series
	R	δC1	δH1	δH6R	δH6S	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg
3a	OMe	88.5	4.46	3.69	3.87	5.2	1.5	56	44	0
3b	Me	88.2	4.53	3.68	3.88	5.1	1.2	58	42	0
3c	H	88.0	4.62	3.68	3.89	5.1	1.3	58	42	0
3d	F	88.1	4.52	3.67	3.88	5.3	1.6	55	45	0
3e	Cl	87.7	4.59	3.68	3.88	5.4	1.5	54	46	0
3f	CF3	86.9	4.78	3.69	3.92	5.6	1.5	52	48	0
3g	NO2	86.1	4.87	3.67	3.90	6.0	1.9	50	50	0

^a3J_H5,H6R and ³J_H5,H6S values were obtained from the H5 signal.

Table 2.

Selected NMR Data and Calculated Rotational Populations (%) for per-Acetylated aryl S-α-Glucosides α2a–α2g (400 MHz, CDCl₃) and Deprotected Aryl S-α-Glucosides α3a–α3g (400 MHz, CD₃OD)

acetyl-protected α series
	R	δC1	δH1	δH6R	δH6S	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg
α2a	OMe	86.0	5.68	4.21	3.96	5.0	1.8	59	41	0
α2b	Me	85.6	5.77	4.20	3.96	5.1	1.8	58	42	0
α2c	H	85.1	5.85	4.21	3.97	5.1	1.8	58	42	0
α2d	F	85.6	5.76	4.21	3.96	5.2	1.7	57	43	0
α2e	Cl	84.9	5.83	4.20	3.96	5.3	1.7	56	44	0
α2f	CF3	83.9	5.98	4.19	3.96	5.2	1.7	57	43	0
α2g	NO2	83.4	6.07	4.20	3.97	5.2	2.3	56	44	0
deprotected α series
	R	δC1	δH1	δH6R	δH6S	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg
α3a	OMe	90.8	5.34	3.73a	3.73a	4.7	2.5	61	39	0
α3b	Me	91.6	5.44	3.74a	3.74a	4.0	3.0	66	30	4
α3c	H	91.2	5.53	3.75a	3.75a	3.9	3.3	65	27	8
α3d	F	90.3	5.47	3.74	3.81	5.2	2.5	56	44	0
α3e	Cl	89.7	5.53	3.71	3.76	5.0	2.3	58	42	0
α3f	CF3	89.9	5.73	3.71	3.76	4.9	2.5	59	41	0
α3g	NO2	87.6	5.86	3.71	3.76	5.2	2.5	56	44	0

^a3J_H5,H6R and ³J_H5,H6S values were obtained from the H5 signal.

In the four aryl S-glucoside series, the chemical shift of the anomeric proton (H1) progressively moves downfield as the aryl substituent’s electron-withdrawing character increases. Thus, in the acetate series, δH1 ranges from 4.48 (2a, R=OMe) to 4.79 (2g, RNO₂) and from 4.46 (3a, ROMe) to 4.87 (3g, R= NO₂) in the tetraols. In turn, the chemical shift of the anomeric carbon (C1) progressively moves upfield as the aryl ring becomes electron-poor.

The δC1 progresses from 85.7 ppm (2a, R=OMe) to 84.3 ppm (2g, R=NO₂) in the acetates and from 88.5 ppm (3a, R=OMe) to 86.1 ppm (3g, R= NO₂) in the tetraols (Table 1). We observe the same trends in the respective protected and deprotected α-anomalies (Table 2). The chemical shifts for H6R and H6S remained practically constant throughout each series.

Hydroxymethyl Group Conformation.

From the ³J_H5,H6R and ³J_H5,H6S values, we calculated the percent contribution for the gg, gt, and tg rotamers to the rotational equilibrium around C5–C6 using different Karplus-type equations including those developed by Haasnoot–Altona,¹³ Nishida,^24e Bock–Duus,¹² Serianni,³¹ and most recently by Crich.³² We found that, although the calculated populations slightly differ among each equation system (see Tables S1–S6), the conformational trends discussed in this work can be observed in each data set. To facilitate comparison with our previously reported works, here, we base our discussion on the rotamer populations calculated using the equations developed by Serianni et al.^31b The analysis of the calculated hydroxymethyl rotamer populations, presented in Tables 1 and 2, reveal noteworthy trends for each glucoside series. Consistent with previous studies on glycosides,¹⁷ the rotamer equilibria in all glucose derivatives studied here are distributed between gg and gt rotamers, with little to no tg contribution. The gg rotamer predominates in the equilibria in each glucoside; however, the comparison throughout each series reveals interesting conformational trends. For example, in peracetylated aryl S-β-glucosides 2a–2g, the gg predominance declines as the aryl substituent’s electron-withdrawing character increases. This gg detriment comes with an increment of the gt contribution. Thus, derivative 2b, substituted with an electron–donor group (R = Me), shows a Pgg of 63%, the largest gg contribution in this series, which decreases to 55% in compound 2g, which is substituted with a strong electron-withdrawing group (R = NO₂). The gt populations, in turn, increase from 37% (2b) to 45% (2g).

Remarkably, the deprotected aryl S-β-glucosides 3a–3g exhibit the same trend observed in the parent per-acetates (Table 1). Derivatives 3b (R = Me) and 3c (R = H) show the highest gg contribution (58%), which progressively decreases to 50% in compound 3g (R = NO₂). Like in per-acetylated derivatives, the Pgg detriment is balanced by a Pgt increase from 42% (3b and 3c) to 50% (3g).

The hydroxymethyl rotational populations in the α-anomers do not show a marked dependence on the substituents of the aglycone (Table 2). The acetyl-protected α-glucosides α2a–α2g exhibit nearly identical gg and gt contributions (in CDCl₃) across the series, whereas deprotected derivatives α3a–α3g, analyzed in CD₃OD, show disperse gg and gt contributions.

Given the evident connection between the hydroxymethyl’s rotational preferences and the aglycone’s electronic properties in the β-series, we plotted the calculated gg and gt populations vs the Hammett parameters σ_p of the respective aryl substituents (Figure 3). The acetyl-protected β-glucosides 2a–2g exhibit good linear correlations in the P_gg vs σ_p (R² = 0.69) and P_gt vs σ_p (R² = 0.61) plots (Figure 3A), and we observe a stronger linear trend for the tetraols 3a–3g (Figure 3B, P_gg vs σ_p R² = 0.83; P_gt vs σ_p R² = 0.83). In both β-series, the gt rotamer increases its contribution, along with a gg detriment, as the Hammett parameter becomes more positive, i.e., as the aryl substituent’s electron-withdrawing character increases.

Figure 3.

Hydroxymethyl rotamer populations (P) vs the aryl substituent’s Hammett parameter (σ_p). (A) P_gg and P_gt vs σ_p for acetyl-protected aryl β-glucosides 2a–2g in CDCl₃;²⁵ (B) P_gg and P_gt vs σ_p for deprotected aryl β-glucosides 3a–3g in CD₃OD;²⁶ (C) P_gg and P_gt vs σ_p for acetyl-protected aryl α-glucosides α2a–α2g in CDCl₃;²⁷ (D) P_gg and P_gt vs σ_p for deprotected aryl α-glucosides α3a–α3g in CD₃OD.²⁸ (E) P_gg and Pgt vs σ_p for deprotected aryl β-glucosides 3a–3g in D₂O.²⁹ (F) P_gg and P_gt vs σ_p for deprotected aryl α-glucosides α3a–α3g in D₂O.³⁰

For the acetyl-protected α-glucosides α2a–α2g, their gg and gt contributions show acceptable linear fits with σ_p (Figure 3C, P_gg vs σ_p R² = 0.66; P_gt vs σ_p R² = 0.66); however, the small slopes (m = 1.67) evidence a negligible aglycone’s influence over the hydroxymethyl’s conformation in these anomers. The deprotected α-glucosides α3a–α3g (Figure 3D) exhibit poor linear correlations in the Hammett plots (P_gg vs σ_p R² = 0.40; P_gt vs σ_p R² = 0.30) pointing to a null aglycone effect in this series.

To investigate the extent of this effect in more polar and biological-like media, we determined the rotamer populations of tetraols 3a–3g and α3a–α3g in D₂O. Table 3 gathers the ³J_H5–H6R and ³J_H5–H6S values (Hz) and the respective calculated rotamer populations in this solvent. We found that in D₂O the β-tetraols exhibit similar conformational behavior to that observed in CD₃OD, i.e., an increment on the aryl substituent’s electron-withdrawing character leads to larger gt and lower gg contributions through the series. Thus, methoxy derivative 3a shows the highest gg population (55%), which progressively diminishes to 50% in nitro-substituted compound 3g. In parallel, the gt contribution increases from 45% (3a) to 50% (3g). From P vs σ_p plots for β-glucosides 3a–3g in D₂O (Figure 3E), the linear correlations of P_gg and P_gt with σ_p (R² = 0.83) becomes clear. The absolute value of the slope in both trends (m = 5.5) is smaller than the respective slopes observed in CD₃OD (m = 6.4), which exposes the solvent effect over the conformation around C5–C6 and the stronger aglycone’s conformational influence in less polar media. The analysis of α-tetraols α3a–α3g in D₂O (Figure 3F) shows no conformational trend.

Table 3.

³J_H5–H6R, ³J_H5–H6S (Hz), and gg, gt, and tg Rotamer Populations (%) for Aryl S-β-Glucosides 3a–3g and Aryl S-α-Glucosides α3a–α3g in D₂O

R	cmpd.	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg	cmpd.	3 J H5,H6R	3 J H5,H6S	P gg	P gt	P tg
OMe	3a	5.4	2.1	55	45	0	α3a	4.1	2.8	66	32	2
Me	3b	5.2	1.5	57	43	0	α3b	3.6	3.6	66	23	11
H	3c	5.7	2.2	52	48	0	α3c	4.6	2.8	61	36	3
F	3d	5.6	1.6	53	47	0	α3d	4.6	2.6	62	37	1
Cl	3e	5.7	2.1	51	49	0	α3e	4.8	2.7	60	39	1
CF3	3f	5.8	1.8	51	49	0	α3f	4.4	2.9	62	34	4
NO2	3g	5.9	1.6	50	50	0	α3g	4.3	3.2	61	32	7

Glycosidic Bond Conformation.

To assess the conformational flexibility of the glycosidic bond, we performed 1D NOESY (selnogp) experiments on selected glycosides. Figure 4 shows the three staggered conformers around the C1–S bond and the anticipated spatial couplings between the aryl’s ortho hydrogens (o-H) and the sugar ring’s protons. In β-glucosides, the spatial coupling of o-H with H1 and H5 characterizes the exo–syn conformation, whereas coupling with H2 indicates exo–anti contribution.

Figure 4.

Staggered conformers around the glycosidic bond and predicted NOE couplings for aryl β-glucosides (A) and aryl α-glucosides (B).

The spatial coupling of o-H with C2–acetyl’s methyl group reports the non-exo conformation in per–acetylated glucosides. Given the exchangeability of the C2–OH proton with the protic solvents used in this study (CD₃OD and D₂O), the non-exo rotamer is challenging to assess for the tetraol series. In the α-glucosides, the spatial coupling of o-H with H1 and H5 reports the exo–syn conformer, coupling with H5 and H3 indicates the sterically unfavored exo–anti configuration, and coupling with H1 and C2–OCOCH₃ characterizes the non-exo rotamer. Prior to excitation and 1D–NOESY spectrum acquisition, we identified the o-H in each glycoside from the HMBC spectra. Figure 5 shows stacked 1D–NOESY spectra sections for selected acetyl-protected and free β-glucosides. In both series, we observe spatial couplings of the aryl’s o-H with the sugar ring protons H1 and H2, revealing exo–syn and exo–anti contributions. In the acetates, the non-exo rotamer contribution is confirmed by the coupling between o-H and C2–acetate’s methyl group. Additionally, a second NOE with a methyl group belonging to an acetate appears in the spectra of all acetyl derivatives, which is assigned to the spatial coupling with the C6–acetyl’s methyl group from the exo–syn/gt conformation. In deprotected glucosides, as anticipated, it was not possible to confirm or rule out the non-exo contribution from the collected spectra.

Figure 5.

(A) Stacked 1D–NOESY spectra sections for acetyl-protected aryl S-β-glucosides 2a (R = OMe), 2b (R = Me), 2d (R = F), 2e (R = Cl), 2f (R = CF₃), and 2g (R = NO₂) in CDCl₃. (B) Stacked 1D–NOESY spectra sections for deprotected aryl S-β-glucosides 3a (R = OMe), 3b (R = Me), 3d (R = F), 3e (R = Cl), 3f (R = CF₃), and 3g (R = NO₂) in CD₃OD.

The analysis of the integral ratios for the H1, H2, and C2–OCOCH₃ coupling peaks in the 1D–NOESY spectra is a convenient qualitative approximation to assess the exo–syn, exo–anti, and non-exo relative contributions to the conformational equilibrium around the C1–S bond. For acetyl-protected glucosides 2a, 2b, 2d, 2e, 2f, and 2g, analyzed in CDCl₃, the H1 peak integral ratio with respect to H2 and C2–OCOCH₃ increases as the aglycone’s substituent electron-withdrawing character increases. Thus, the H1/H2 integral quotient for glycoside 2a (R = OCH₃) is 1.1, which progressively increases to 2.7 in glucoside 2g (R = NO₂). A similar trend is observed for the H1/C2–OCOCH₃ ratios (C2–OCOCH₃ methyl’s peak was identified from the HMBC spectra) where compound 2a (R = OCH₃) shows an H1/OCOCH₃ ratio of 1.3, whereas the ratio for glucoside 2g (R = NO₂) is 6.2. The respective tetraols exhibit the same trend shown by their acetyl-protected congeners. Derivative 3a (R = OCH₃) shows an H1/H2 integrals ratio of 1.4, which increases to 6.2 in compound 3g (R = NO₂).

The lower H1/H2 ratios observed in glucosides carrying electron-rich aglycones reveal comparable exo–syn and exo–anti contributions and a higher degree of flexibility around the glycosidic bond in these substrates. The increment of the o-H/H1 NOE strength with respect to H2 and C2–OCOCH₃ in glucosides carrying electron-poor aglycones evidence a more restricted rotation around the C1–S bond and an equilibrium shift favoring the exo–syn rotamer.

Tetraols 3a, 3b, 3d, 3e, 3f, and 3g, analyzed in CD₃OD, exhibit a similar trend. As the aryl substituents’ electronwithdrawing character increases, the NOE between the o-H and H2 diminishes and the NOE with H1 predominates, indicating less flexibility around the glycosidic bond and a larger exo–syn contribution. The analysis of tetraols 3b (R = Me) and 3g (R = NO₂) in D₂O (Figure 6) shows conformational behavior equivalent to that observed in CD₃OD. Figure 7 shows stacked 1D NOESY spectra for the selected α-glucosides α2b (R = Me) and α2g (R = NO₂) (Figure 7A) and their respective tetraols α3b and α3g (Figure 7B).

Figure 6.

Stacking of 1D–NOESY spectra sections for tetraols 3b (R = Me) and 3g (R = NO₂) in D₂O.

Figure 7.

Stacking of 1D–NOESY spectra sections for aryl α-glucosides. (A) acetates α2a (R = Me) and α2g (R = NO₂) in CDCl₃; (B) tetraols α3a (R = Me) and α3g (R = NO₂) in CD₃OD.

In the acetates α2b and α2g, analyzed in CDCl₃ (Figure 7), o-H couples with H1, H5, H6S, and C2–acetyl’s CH₃, being the coupling with H1 the strongest, which indicates exo–syn and non-exo contributions. The larger H1/H5 integral ratio (3.9) in the nitro-substituted glucoside α2g compared with α2b (H1/H5 = 1.5), points to a larger contribution of the nonexo conformer (Figure 4B) in the electron-deficient derivative. For tetraols α3b and α3g, analyzed in CD₃OD, we observe the same properties: a larger non-exo contribution, judged from a weaker o-H/H5 NOE, as the aryl substituent’s electron-withdrawing character increases. This observation is also confirmed in D₂O, where tetraols α3b and α3g exhibit the same peak patterns on their 1D NOESY spectra (Figure S1).

DFT Analysis of the Conformational Trends.

A plausible mechanism to explain the increased preference for the exo–syn and gt rotamers as the aglycone becomes more electron-deficient rests on the exo–syn/gt conformer stabilization. To explain the origin of this stabilization, we employed density functional theory (DFT) calculations. The calculations were performed in Gaussian16³⁸ using hybrid exchange-correlation PBE1PBE³³ augmented with D3 dispersion correction³⁴ and a 6–311+G(d,p) basis set. This level of theory yields accurate conformational energies of mono- and disaccharides,³⁵ and has been used to determine structures of carbohydrates and glycosyl intermediates.³⁶ To simulate the effect of the solvent, we selected SMD continuous solvation model due to its extended parametrization, including non-electrostatic effects, for biochemically relevant molecules.³⁷ First, we performed systematic geometry optimizations of nine plausible rotamers of protected and deprotected S-β-glucosides and S-α-glucosides. The optimizations were carried out for each of the substituents and then followed by the derivation of the harmonic free energies.

Each rotamer was sampled in a few orientations of the S–C_Ar and C6–O6 bonds to localize the most stable conformers, and the relative free energies of the most stable conformers of each rotamer were used to derive their Boltzmann populations. To compare with the experimental data, the populations that shared the same orientation of the C6–O6, gt, tg, and gg, bonds were added together. Interestingly, the DFT calculations for the four sets of glucosides showed that the exosyn conformer was consistently the most stable, but they were not able to reproduce the experimental trends with respect to the orientation of the C6 group. In the case of deprotected S-glucosides, the free O6H group formed hydrogen bond interactions, which overstabilize the tg and gt conformations of glucose in continuous solvent (Tables S7 and S8), an effect that has been reported in the literature.¹² The acetyl group in the protected S-glucosides, on the other hand, introduced two additional degrees of freedom, along the S–C_Ar and C6–O6 bonds, which resulted in multiple conformations for each of the rotamers, and complicated the assignment. Thus, to circumvent the problem of explicit hydrogen bonds and multiple conformers, we looked at the populations of methyl-protected derivatives. First, we started by sampling the nine rotamers, which confirmed that the exosyn orientation of aglycone is the most favorable. We also observed that the tg orientation of the C5–C6 bond is the least favorable. Then, we scanned the dihedral angle along the S–C_Ar bond in the exosyn/gg and exosyn/gt orientations. The scan revealed that the most stable orientation of the aryl group depends on the properties of the substituent (Figure S2). For electron-poor aglycones (R = CF₃ and NO₂), the aryl group adopts an orientation in plane with C1 which increases the ring resonance with the S atom and promotes the interactions between o-H and O5. We selected the most stable orientation of the aryl group, calculated the Boltzmann populations of the permethylated S-glucosides of each rotamer, and added them to gg, gt, and tg groups (Figure 8 and Tables S9–S12). The calculations for both S-glucosides follow the experimental trend where an electron-poor aryl ring decreases the stability of the gg conformer and increases the stability of the gt conformer. In the case of S-β-glucosides, the calculations predict that the P_gg is between 50 and 60% for R = CH₃, H, F, Cl, and CF₃, and decreases with increasing σ_p, which is in line with the experimental data. The effect, however, appears overestimated for R = OMe (P_gg of 80%), and for R = NO₂ (P_gg of 10%), possibly due to a stronger resonance effect. In the case of the S-α-glucosides, the overall trend is less clear; in principle, the population of the gg conformer appears to decrease with increasing σ_p, but there is more deviation in the population. This, however, is also consistent with the experimental data, which showed a large deviation for the deprotected S-α-glucosides. Visual inspection of the most stable conformers did not provide a clear answer about potential interactions that would correlate with the stabilization of the gt conformer over gg when a strongly electron-withdrawing group is added to the aglycone. The analysis of the local dipoles, however, reveals an interplay between the orientation of the dipole moment at C6 and at the aglycone. For electron-rich substituents, the electron dipole of the aglycone, in both anomer forms, is aligned with the S atom, and the total dipole moment of the molecule is similar for the gg and gt conformers (Figure 9). However, the strongly electron-withdrawing groups change the direction of the aglycone’s dipole moment. In the gg orientation, the dipole moment of C6 is perpendicular to the aglycone, whereas in the gt orientation, both dipole moments are parallel, which results in an increase of the molecule’s total dipole moment. Overall, this effect increases with the σ_p value, which would explain the better stabilization of the gt conformer in a polar medium.

Figure 8.

Hydroxymethyl rotamer populations (P) vs aryl substituent’s Hammett parameter (σ_P) predicted from DFT calculations. (A) P_gg and P_gt vs σ_P for methyl-protected aryl α-glucosides.³⁹ (B) P_gg and P_gt vs σ_P for methyl-protected aryl β-glucosides.⁴⁰

Figure 9.

(A) Alignment of dipole moments between the C6 rotamer and the electron-poor and electron-rich aglycone. (B) Dipole moment predicted for gg and gt exosyn conformers of methyl-protected aryl α/β-glucosides.⁴¹

CONCLUSIONS

The conformational study of aryl S-glucosides revealed remarkable differences in their flexibility patterns with respect to their alkyl congeners, specifically on how the aglycone modulates the rotation about C5–C6. Our previous studies on protected alkyl S-β-glucosides exposed the alkyl aglycone’s steric properties as a remote conformational modulator where bulky groups lead to increased gt contributions in a conformational equilibrium dominated by gg. In the case of aryl derivatives, it is not the aglycone’s bulkiness but the electronic properties that exert analogous control. In the aryl S-β-glucosides, increased gt contributions are observed in those glucosides carrying electron-deficient aryl aglycones, and we observe this trend in protected and deprotected aryl β-derivatives. Our findings on aryl glucosides challenge the stereoelectronic explanation formulated for alkyl derivatives to explain their structure/conformation relationships; however, we found the alignment of the O5, aglycone, and C6–O6 local dipoles a plausible mechanism to explain the trends.

In the case of the α anomers, the effect of aglycone over the glycosidic bond conformation is similar to that observed on their β isomers; however, no influence of aglycone over the hydroxymethyl group conformation was observed. This could be explained by the increased non-exo conformation.

From a practical standpoint, our findings expand the toolbox to modulate the conformation of saccharides and could prove helpful in designing optimized glycomimetics with applications in medicinal chemistry and chemical biology.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.