Ribitol in Solution Is an Equilibrium of Asymmetric Conformations

molecules

Article Ribitol in Solution Is an Equilibrium of Asymmetric Conformations

Shiho Ohno 1, Noriyoshi Manabe 1, Takumi Yamaguchi 2 , Jun Uzawa 3 and Yoshiki Yamaguchi 1,3,*

1 Division of Structural Glycobiology, Institute of Molecular Biomembrane and Glycobiology, Tohoku Medical and Pharmaceutical University, 4-4-1 Komatsushima, Aoba-ku, Sendai 981-8558, Miyagi, Japan; [email protected] (S.O.); [email protected] (N.M.) 2 School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi 923-1292, Ishikawa, Japan; [email protected] 3 Structural Glycobiology Team, RIKEN (The Institute of Physical and Chemical Research), 2-1 Hirosawa, Wako 351-0198, Saitama, Japan; [email protected] * Correspondence: [email protected]; Tel.: +81-22-727-0208

Abstract: Ribitol (C5H12O5), an acyclic sugar alcohol, is present on mammalian α-dystroglycan as a component of O-mannose glycan. In this study, we examine the conformation and dynamics of ribitol by database analysis, experiments, and computational methods. Database analysis reveals that the anti-conformation (180◦) is populated at the C3–C4 dihedral angle, while the gauche conformation (±60◦) is seen at the C2–C3 dihedral angle. Such conformational asymmetry was born out in a solid-state 13C-NMR spectrum of crystalline ribitol, where C1 and C5 signals are unequal. On the other hand, solution 13C-NMR has identical chemical shifts for C1 and C5. NMR 3J coupling constants and OH exchange rates suggest that ribitol is an equilibrium of conformations, under the inﬂuence of hydrogen bonds and/or steric hinderance. Molecular dynamics (MD) simulations allowed us to discuss such a chemically symmetric molecule, pinpointing the presence of asymmetric Citation: Ohno, S.; Manabe, N.; conformations evidenced by the presence of correlations between C2–C3 and C3–C4 dihedral angles. Yamaguchi, T.; Uzawa, J.; Yamaguchi, These ﬁndings provide a basis for understanding the dynamic structure of ribitol and the function of Y. Ribitol in Solution Is an ribitol-binding enzymes. Equilibrium of Asymmetric Conformations. Molecules 2021, 26, Keywords: ribitol; conformation; dynamics; NMR; MD simulation; hydrogen bond 5471. https://doi.org/10.3390/ molecules26185471

Academic Editor: Hideyuki Takeuchi 1. Introduction

Received: 9 August 2021 Ribitol (C5H12O5) is an acyclic sugar alcohol and a component of teichoic acid found Accepted: 6 September 2021 in Gram-positive bacteria [1] (Figure1). Ribitol is also found in riboflavin (vitamin B2) and 0 Published: 8 September 2021 flavin mononucleotide (riboflavin 5 -phosphate). Ribitol phosphate (D-ribitol 5-phosphate) is a component of O-mannose glycans on mammalian α-dystroglycan [2]. A genetic Publisher’s Note: MDPI stays neutral deficiency of ribitol phosphate transferase leads to muscular dystrophy, highlighting the with regard to jurisdictional claims in essential role of ribitol phosphate in the development of skeletal muscle [3]. Laminin is published maps and institutional affil- known to interact with ribitol-containing O-mannose glycans, and the ribitol residues are iations. likely acting as a hinge in bridging laminin and α-dystroglycan. Hence, the structure and dynamics of ribitol are now receiving much attention. However, compared with the cyclic hemiacetal sugars, our knowledge of the structural properties and dynamics of such an acyclic sugar alcohol is rather limited. Copyright: © 2021 by the authors. Jeffrey et al. investigated the crystal structures of nine alditols, including ribitol [4]. Licensee MDPI, Basel, Switzerland. In the crystalline state, the carbon chain of alditols tends to adopt an extended, planar This article is an open access article zigzag conformation when the configurations at alternate carbons are different (e.g., Cn distributed under the terms and and Cn+2 will be D and L or L and D). When the configurations of alternate carbons are the conditions of the Creative Commons same (D and D or L and L), the Cn-O and Cn+2-O bonds align in parallel. This arrangement Attribution (CC BY) license (https:// causes steric hindrance such that the carbon chain changes its conformation to bent and creativecommons.org/licenses/by/ non-planar. Ribitol adopts the same configuration at C2 and C4 (D and D), and the steric 4.0/).

Molecules 2021, 26, 5471. https://doi.org/10.3390/molecules26185471 https://www.mdpi.com/journal/molecules Molecules 2021, 26, 5471 2 of 10

hindrance brought by O2 and O4 is avoided by rotating the C2–C3 or C3–C4 bonds by 120◦. In fact, the crystal structure of ribitol adopts a bent conformation, avoiding the steric Molecules 2021, 26, 5471 2 of 11 hindrance induced by a stretched planar zigzag structure. Likewise, the ribitol moiety of riboﬂavin exhibits a bent conformation in crystals [5,6].

Figure 1. ChemicalChemical structures structures of ofthe the compou compoundsnds examined examined in this in thisstudy. study. (a) Ribitol, (a) Ribitol, (b) ribitol (b) ribitol phos- phosphatephate (D-ribitol (D-ribitol 5-phosphate 5-phosphate is analyzed is analyzed in this instud thisy, study,which whichis the form is the found form in found nature), in nature),and (c) and1,3,5-pentanol (c) 1,3,5-pentanol (reference (reference compound). compound). Carbon and Carbon oxygen and numbering oxygen numbering is indicated is in indicated the figure. in The the ﬁgure.main-chain The main-chaindihedral angles dihedral are defined angles areas φ deﬁned1 (O1–C1–C2–C3), as ϕ1 (O1–C1–C2–C3), φ2 (C1–C2–C3–C4),ϕ2 (C1–C2–C3–C4), φ3 (C2–C3–C4–ϕ3 (C2–C3–C4–C5),C5), and φ4 (C3–C4–C5–O5) and ϕ4 (C3–C4–C5–O5) throughout this throughout manuscript. this manuscript.

HawkesJeffrey et et al. al. investigated performed solutionthe crystal NMR structures analysis of to nine determine alditols, the including structure ribitol of ribitol [4]. basedIn the oncrystalline3J(H,H) values state, [the7]. Ribitolcarbon equilibrates chain of alditols between tends a twist to adopt at C2–C3 an andextended, one at C3–C4,planar removingzigzag conformation the O2/O4 interactionwhen the configurations in the planar chain at alternate form. It shouldcarbons be are noted different that 3J (H2,H3)(e.g., Cn 3 and CJn+2(H3,H4) will be of D ribitol and L areor L indistinguishable and D). When the in configurations solution NMR. of The alternate relationship carbons between are the C2–C3same (D and and C3–C4 D or Ltorsion and L), anglesthe Cn-O isnot and definitive Cn+2-O bonds solely align from in solution parallel. NMR This analysis.arrangement causesGarrett steric andhindrance Serianni such synthesized that the carbon13C-labeled chain changes sugar alcohols, its conformation including toD-[1- bent 13andC] 3 ribitol,non-planar. for NMR Ribitol analysis adopts [8 the]. J same(C1,C4) configuration is 1.7 Hz, which at C2 is and intermediate C4 (D and betweenD), and the≈1.3 steric Hz (gauchehindrance conformations) brought by O2 and and≈ O42.3 Hzis avoided (anti conformation), by rotating the indicating C2–C3 or that C3–C4 the C2–C3bonds by or C3–C4120°. In linkage fact, the is crystal a mixture structure of extended of ribitol and adopts bent forms. a bent Franks conformation, et al. found avoiding that the the3J (H,H)steric valueshindrance are significantlyinduced by a different stretched in planar water andzigzag organic structure. solvents, Likewise, suggesting the ribitol that hydrationmoiety of andriboflavin hydrogen exhibits bonding a bent affect conformation the conformation in crystals of ribitol[5,6]. [9]. KleinHawkes et al.et performedal. performed MD solution and Monte NMR Carlo analys (MC)is to simulations determine forthe the structure tetrasaccharide- of ribitol ribitolbased on unit 3J(H,H) in teichoic values acid [7]. [ 9Ribitol]. They equilibrates found that between the GalNAc-ribitol a twist at C2–C3 linkage and is one more at flex-C3– ibleC4, removing than other the glycosidic O2/O4 interaction linkages, but in thethe ribitolplanar chain chain itself form. is It not should completely be noted flexible. that Hatcher3J(H2,H3) et and al. parameterized 3J(H3,H4) of ribitol theCHARMm are indistinguishable force field forin solution acyclic sugarNMR. alcohols The relationship and per- formedbetween MD C2–C3 calculations and C3–C4 for torsion sugar alcohols,angles is no includingt definitive ribitol solely [10 ].from Other solution than these, NMR thereanal- areysis. few studies on the conformational dynamics of ribitol. OurGarrett present and Serianni study aims synthesized to build 13 onC-labeled previous sugar knowledge alcohols, onincluding the conformation D-[1-13C] ribi- of ribitoltol, for and NMR to deepen analysis our [8]. understanding 3J(C1,C4) is of1.7 static Hz, andwhich dynamic is intermediate structures between of ribitol through≈1.3 Hz database(gauche conformations) analysis, experimental and ≈2.3 NMRHz (anti analysis, conformation), and MD simulation.indicating that the C2–C3 or C3– C4 linkage is a mixture of extended and bent forms. Franks et al. found that the 3J(H,H) 2. Results and Discussion values are significantly different in water and organic solvents, suggesting that hydration 2.1. Database Analysis of Ribitol and Ribitol Phosphate and hydrogen bonding affect the conformation of ribitol [9]. CrystalKlein et structuresal. performed of oligosaccharides MD and Monte Carl boundo (MC) to proteinssimulations provide for the good tetrasaccharide- insight into theribitol more unit stable in teichoic conformations acid [9]. They of eachfound glycosidic that the GalNAc-ribitol linkage [11]. linkage Following is more this flexible notion, wethan extracted other glycosidic the conformation(s) linkages, but of ribitolthe ribitol and ribitolchain itself phosphate is not fromcompletely the Cambridge flexible. CrystallographicHatcher et al. parameterized Data Centre the (CCDC) CHARMm and the forc Proteine field Data for acyclic Bank (PDB), sugar inalcohols total, six and ribitol per- andformed four MD ribitol-phosphate calculations for structures sugar alcohols, (Figure including2). ribitol [10]. Other than these, there are few studies on the conformational dynamics of ribitol. Our present study aims to build on previous knowledge on the conformation of ribitol and to deepen our understanding of static and dynamic structures of ribitol through database analysis, experimental NMR analysis, and MD simulation.

Molecules 2021, 26, 5471 3 of 11

2. Results and Discussion 2.1. Database Analysis of Ribitol and Ribitol Phosphate Crystal structures of oligosaccharides bound to proteins provide good insight into the more stable conformations of each glycosidic linkage [11]. Following this notion, we extracted the conformation(s) of ribitol and ribitol phosphate from the Cambridge Crys- tallographic Data Centre (CCDC) and the Protein Data Bank (PDB), in total, six ribitol and Molecules 2021, 26, 5471 four ribitol-phosphate structures (Figure 2). 3 of 10

Figure 2. Representative 3D structure of ribitol and ribitol phosphate extracted from the database. Figure(a) Crystal 2. Representative structure of 3D ribitol structure (PDB of ID: ribitol 5IAI). and (b) ri Ribitolbitol phosphate phosphate extracted (PDB ID: from 6H4F). the database. (c) Ribitol (ainteracting) Crystal structure with water of ribitol molecules (PDB (PDB ID: ID:5IAI). 4A0S). (b) Ribitol The structures phosphate are shown(PDB ID: in stick6H4F). representation. (c) Ribitol in- In teracting(a,b), the with electron water density molecules map (PDB is depicted ID: 4A0S). in gray The mesh structures contoured are shown at 2.0 σ inlevel. stick Proteins representation. are omitted In (afor,b), clarity. the electron In (c), density water molecules map is depicted hydrogen in gray bonded mesh with contoured ribitol at are 2.0 shown σ level. in Proteins cyan sphere. are omitted for clarity. In (c), water molecules hydrogen bonded with ribitol are shown in cyan sphere. The dihedral angles of the carbon main chain and the C1–C5 distance of extracted structuresThe dihedral are listed angles in Tableof the1 .carbon Although main thechain data and are the limited, C1–C5 onedistance can examineof extracted the structuresstable conformations are listed in forTable each 1. dihedralAlthough angle. the data In each are structure,limited, one the can dihedral examine angles the suggeststable conformationsstaggered conformations. for each dihedral Except angle. for one In exampleeach structure, (6H4M), the the dihedral four dihedral anglesangles suggest (ϕ stag-1–ϕ4) gereddo not conformations. all simultaneously Except assume for one anti-conformations; example (6H4M), somethe four prefer dihedral bent conformations. angles (φ1–φ4) do not all simultaneously assume anti-conformations; some prefer bent conformations. ◦ Table 1. Main chain dihedral anglesWhen ( ),φ C1–C52 (C1–C2–C3–C4) distance (Å), andis in resolution the anti-conformation, (Å) of ribitol and ribitolφ3 (C2–C3–C4–C5) phosphate extracted adopts from the CCDC and the PDB.gauche NA; notand available. vice versa. These results indicate that φ2 and φ3 correlate with each other. C1–C5 distances were also calculated to ascertain the degree of carbon chain stretch. The CCDC ϕ1 ϕ2 ϕ3 ϕ4 Deposition C1–C5 distance of the fully extended ribitol measuredC1–C5 5.0 Å. Distance The averageResolution value of Reference the C1– Number/DB ID (O1–C1–C2–C3) (C1–C2–C3–C4) (C2–C3–C4–C5) (C3–C4–C5–O5) C5 distance of 10 extracted ribitol and ribitol phosphate structures is 4.6 Å, suggesting ribitol 1249410 171 that ribitol− 62and ribitol phosphate−172 are not comp 71letely extended, 4.5 possibly to - avoid the[12 con-] 662559 171 −61 −171 73 4.6 - [13] 1015979 171 tinuous anti-conformations−61 −170 that induce O2/O4 73 repulsion. In 4.6 high-resolution - structures[14] of 5IAI 173ribitol (PDB 166 ID; 5IAI and 4Q0S), 60 hydrogen 170 bonding with surrounding 4.5 water 1.6 molecules NA is 4Q0S −176 −62 −167 176 4.5 1.9 [15] 4F2D −150 observed− (Figure174 2c). This 77suggests that −intermolecular147 interaction 4.5 with 2.3water molecules NA Ribitol phosphate may be involved in the various conformations of ribitol. It seems that the phosphorylation 6H4F 149 94 83 156 4.4 2.2 [16] 6H4M 131 of ribitol −does136 not significantly−143 change the−158 ribitol conformation, 5.0 although 2.7 it needs further[16] 6HNQ 166 160 69 155 4.6 2.4 [16] 6KAM −75detailed study. 164 In this paper, 96 we mainly −focus168 on ribitol. 4.8 2.5 [17]

Table 1. Main chain dihedral angles (°), C1–C5 distance (Å), and resolution (Å) of ribitol and ribitol phosphate extracted from the CCDC and the PDB. NA; Whennot available.ϕ2 (C1–C2–C3–C4) is in the anti-conformation, ϕ3 (C2–C3–C4–C5) adopts gauche and vice versa. These results indicate that ϕ2 and ϕ3 correlate with each other. CCDC φ1 φ3 C1–C5 distancesφ2 were also calculatedφ to4 ascertain theC1–C5 degree of carbon chain stretch. The Deposition (O1–C1–C2–C1–C5 distance of the(C2–C3– fully extended ribitol measured 5.0 Å. TheResolution average value Reference of the C1–C5 (C1–C2–C3–C4) (C3–C4–C5–O5) Distance Number/DB ID C3) distance of 10 extractedC4–C5) ribitol and ribitol phosphate structures is 4.6 Å, suggesting that ribitol ribitol and ribitol phosphate are not completely extended, possibly to avoid the continuous 1249410 171 anti-conformations−62 that−172 induce O2/O471 repulsion. In 4.5 high-resolution - structures[12] of ribitol 662559 171 (PDB ID;− 5IAI61 and 4Q0S),−171 hydrogen bonding73 with surrounding 4.6 water -molecules is[13] observed 1015979 171 (Figure2−c).61 This suggests−170 that intermolecular73 interaction 4.6 with water - molecules[14] may be 5IAI 173 involved166 in the various conformations60 170 of ribitol. It seems4.5 that the phosphorylation1.6 NA of ribitol does not signiﬁcantly change the ribitol conformation, although it needs further detailed 4Q0S −176 −62 −167 176 4.5 1.9 [15] study. In this paper, we mainly focus on ribitol. 4F2D −150 −174 77 −147 4.5 2.3 NA 2.2. NMR Analysis 2.2.1. Solid-State 13C-NMR Analysis of Crystalline Ribitol

We can conclude from the database analysis that stable conformations of ribitol are likely asymmetric. To investigate the conformation of crystalline ribitol from another aspect, solid-state 13C-NMR was performed. When the conformation of ribitol is asymmetric, carbon signals will be observed separately in the NMR spectra. Indeed, a 13C CP-MAS NMR spectrum of crystalline ribitol shows four separate signals (Figure3a). By comparison with the solution 13C-NMR spectrum of ribitol (Figure3b), partial assignments of 13C Molecules 2021, 26, 5471 4 of 11

Ribitol phosphate 6H4F 149 94 83 156 4.4 2.2 [16] 6H4M 131 −136 −143 −158 5.0 2.7 [16] 6HNQ 166 160 69 155 4.6 2.4 [16] 6KAM −75 164 96 −168 4.8 2.5 [17]

2.2. NMR Analysis 2.2.1. Solid-State 13C-NMR Analysis of Crystalline Ribitol We can conclude from the database analysis that stable conformations of ribitol are likely asymmetric. To investigate the conformation of crystalline ribitol from another aspect, solid-state 13C-NMR was performed. When the conformation of ribitol is asymmetric, 13 Molecules 2021, 26, 5471 carbon signals will be observed separately in the NMR spectra. Indeed, a C CP-MAS4 of 10 NMR spectrum of crystalline ribitol shows four separate signals (Figure 3a). By comparison with the solution 13C-NMR spectrum of ribitol (Figure 3b), partial assignments of 13C signals were possiblepossible forfor 7272 andand 7373 ppmppm toto C2/C3/C4 C2/C3/C4 and and the the peaks peaks at 60 and 62 ppm toto C1/C5. Interestingly, Interestingly, C1/C5 shows shows two two separate separate peaks. peaks. This This observation observation may indicate that C1 and C5C5 ofof ribitolribitol areare in in different different microenvironments. microenvironments. As As seen seen above, above, ribitols ribitols deposited deposited in CCDCin CCDC have have an asymmetrican asymmetric structure structure (Table (Table1). It 1). isplausible It is plausible that ribitol that ribitol has an has asymmetric an asym- structuremetric structure in the solid in the state solid as state well, as which well, gives which different gives different chemical chemical shifts. C2/C3/C4 shifts. C2/C3/C4 signals appearsignals aroundappear thearound same the frequency; same frequency; therefore, therefore, interpretation interpretation here is impossible here is impossible without a completewithout a assignment.complete assignment.

Figure 3. Comparison of solid-state and solution 13C-NMR spectra of ribitol. (a) Solid-state 13C 13 13 CP-MASFigure 3. -NMRComparison spectrum of solid-state of crystalline and ribitol. solution (b ) SolutionC-NMR spectra13C-NMR of ribitol. spectrum (a) ofSolid-state ribitol dissolved C CP- MAS -NMR spectrum of crystalline ribitol. (b) Solution 13C-NMR spectrum of ribitol dissolved in in D2O. D2O. In contrast, the solution 13C-NMR of ribitol showed that C1 and C5, C2 and C4 have 13 the sameIn contrast, chemical the shifts solution (Figure C-NMR3b). These of ribitol results showed suggest that that C1 ribitol and C5, transits C2 and rapidly C4 have in solutionthe same between chemical stable shifts asymmetric (Figure 3b). conformations, These results suggest giving risethat to ribitol averaged transits signals. rapidly in solution between stable asymmetric conformations, giving rise to averaged signals. 2.2.2. Solution NMR Analysis of Ribitol Based on Coupling Constant 2.2.2.Then, Solution the NMR dihedral Analysis angle of of Ribitol the ribitol Based main on Coupling chain was Constant analyzed from the coupling constantsThen, obtained the dihedral by solution angle of NMR. the Weribitol initially main triedchain to was obtain analyzed3J(H,H) from from the the coupling splitting ofconstants the signals, obtained assuming by solution a ﬁrst-order NMR. We approximation. initially tried However,to obtain 3 theJ(H,H) assumption from the splitting was not validof the due signals, to strong assuming coupling a first-order in ribitol, approximation. even at the 1H observationHowever, the frequency assumption of 600 was MHz. not Molecules 2021, 26, 5471 3 1 5 of 11 Therefore,valid due toJ strongvalues coupling were estimated in ribitol, by even comparing at the H simulated observation and frequency experimental of 600 spectra MHz. iteratively.Therefore, 3 ToJ values improve were the estimated accuracy, 1D-by comparing1H NMR measurements simulated and were experimental performed spectra at the 1 it-H observationeratively. To frequencies improve the of accuracy, 270, 400, and1D-1H 600 NMR MHz, measurements and the iterative were comparison performed was at the done 1H atobservationat eacheach frequencyfrequency frequencies (Figure(Figure of4 4).). 270, 33J(C,H) 400, and was 600 estimated MHz, and fromfrom the thethe iterative HR-HMBCHR-HMBC comparison methodmethod was [[18].18]. done TheThe observedobserved chemicalchemical shiftsshifts andand couplingcoupling constantsconstants areare summarizedsummarized inin TableTable2 .2.

FigureFigure 4.4.Solution Solution1 H-NMR1H-NMR spectra spectra of of ribitol ribitol (upper (upper) at) 1atH 1 observationH observation frequencies frequencies of 600 of (600left ),(left 400), ( middle400 (middle), and), 270 and MHz 270 (MHzright ).(right Simulated). Simulated NMR spectraNMR spectra at each at1 Heach frequency 1H frequency are shown are shown in the in lower the lower panel. panel.

Table 2. 1H and 13C chemical shift and coupling constant of ribitol determined in this study.

Chemical Shift (ppm) H1R(pro-R) 3.79 H1S(pro-S) 3.64 H2 3.81 H3 3.68 C1 65.1 C2 74.8 C3 74.9 Coupling constant (Hz) 3J(H1R,H2)(pro-R) 3.00 3J(H1S,H2)(pro-S) 7.20 3J(H2,H3) 6.50 2J(H1R,H1S) −12.70 3J(C1,H3) 3.8 * 3J(C3,H1S) 2.9 * *3J(C,H) may include errors due to the presence of strong 3J(H,H) coupling.

Chemical shifts and 2,3J(H,H) coupling constants are in substantial agreement with those of previous studies [7,8,10,19]. 3J(C1,H3) was measured as 3.8 Hz in this study, similar to a previous report (3.7 Hz) [8]. Then, these coupling constants are used for estimating the distribution of each conformer using the Haasnoot equation, which includes a set of empirical constants [20]. Regarding the O1–C1 dihedral angle (φ1 = O1–C1–C2–C3), the population of three conformers (φ1 = 180°, −60°, and +60°) was calculated using 3J(H1R,H2)(pro-R) and 3J(H1S,H2)(pro-S) with reference to the method of Hawkes [7]. The φ1 ratio was estimated as 180°:−60°:+60° = 64:36:0. This closely conforms to the result of Hawkes [7]. For the C1–C2 dihedral angle φ2 (φ2=C1–C2–C3–C4), the conformations can be estimated from 3J(H2,H3) and 3J(C1,H3). The population of each φ2 conformation was calculated as 180°:−60°:+60° = 2:46:52. These data suggest the presence of a rapid transition (>>10 Hz in terms of coupling constant) around the C1–C2 axis and a relatively small population of the anti-conformation. It should be noted that φ1 and φ4 or φ2 and φ3 cannot be discussed separately because their NMR signals (e.g., H1 and H5, H2 and H4) overlap in solution.

2.2.3. NMR Analysis of Hydroxy Protons Franks et al. found that the 3J(H,H) values of ribitol differ significantly in water and organic solvents [19], and this suggests that hydration and hydrogen bonding affect the

Molecules 2021, 26, 5471 5 of 10

Table 2. 1H and 13C chemical shift and coupling constant of ribitol determined in this study.

Chemical Shift (ppm) H1R(pro-R) 3.79 H1S(pro-S) 3.64 H2 3.81 H3 3.68 C1 65.1 C2 74.8 C3 74.9 Coupling constant (Hz) 3J(H1R,H2)(pro-R) 3.00 3J(H1S,H2)(pro-S) 7.20 3J(H2,H3) 6.50 2J(H1R,H1S) −12.70 3J(C1,H3) 3.8 * 3J(C3,H1S) 2.9 * * 3J(C,H) may include errors due to the presence of strong 3J(H,H) coupling.

Chemical shifts and 2,3J(H,H) coupling constants are in substantial agreement with those of previous studies [7,8,10,19]. 3J(C1,H3) was measured as 3.8 Hz in this study, similar to a previous report (3.7 Hz) [8]. Then, these coupling constants are used for estimating the distribution of each conformer using the Haasnoot equation, which includes a set of empirical constants [20]. Regarding the O1–C1 dihedral angle (ϕ1 = O1–C1–C2–C3), the population of three conformers (ϕ1 = 180◦, −60◦, and +60◦) was calculated using 3J(H1R,H2)(pro-R) and 3J(H1S,H2)(pro-S) with reference to the method of Hawkes [7]. The ϕ1 ratio was estimated as 180◦:−60◦:+60◦ = 64:36:0. This closely conforms to the result of Hawkes [7]. For the C1–C2 dihedral angle ϕ2 (ϕ2 = C1–C2–C3–C4), the conformations can be estimated from 3J(H2,H3) and 3J(C1,H3). The population of each ϕ2 conformation was calculated as 180◦:−60◦:+60◦ = 2:46:52. These data suggest the presence of a rapid transition (>>10 Hz in terms of coupling constant) around the C1–C2 axis and a relatively small population of the anti-conformation. It should be noted that ϕ1 and ϕ4 or ϕ2 and ϕ3 cannot be discussed separately because their NMR signals (e.g., H1 and H5, H2 and H4) overlap in solution.

2.2.3. NMR Analysis of Hydroxy Protons Franks et al. found that the 3J(H,H) values of ribitol differ significantly in water and organic solvents [19], and this suggests that hydration and hydrogen bonding affect the conformation of ribitol. In this study, we analyzed the hydroxyl group of ribitol by directly observing the hydroxyl proton by 1H-NMR. The assignments of OH signals are shown in Figure5. At 0 ◦C, two sharp peaks are observed at 5.9 ppm and 5.8 ppm, and these correspond to OH3 and OH2/OH4 (Figure5a). The OH3 signal gives a relatively sharp peak at 0 ◦C, suggesting that OH3 is stabilized by hydrogen bonding. To validate the exchange of the hydroxyl proton, the temperature coefficient (ppb K−1) was estimated as 11 ppb K−1 for ribitol OH3. According to Sandströ et al., the temperature coefficient is greater than 11 ppb K−1 when the OH proton is fully hydrated [8]. This suggests that OH3 is mostly exposed to solvent, but some transient hydrogen bond may exist. The exchange rate constant of OH3 was also estimated from 2D EXSY. From the build-up curve −1 of exchange peak volume, it was estimated as kex = 193 s . To estimate the effect of an intramolecular hydrogen bond on the temperature coefficient and OH exchange rate, the same experiment was performed using the model compound 1,3,5-pentanol (Figure1), which does not have hydroxyl groups at the 2 and 4 positions. 1,3,5-Pentanol gives a broad OH peak around 5.7 ppm at 0 ◦C. This peak can be assigned to OH1/OH3/OH5, and the peak almost disappears at 25 ◦C (Figure5b). We reasoned that this is due to the lack of a hydrogen bond, and hence, OH3 shows faster proton exchange. Molecules 2021, 26, 5471 6 of 11

conformation of ribitol. In this study, we analyzed the hydroxyl group of ribitol by directly observing the hydroxyl proton by 1H-NMR. The assignments of OH signals are shown in Figure 5. At 0 °C, two sharp peaks are observed at 5.9 ppm and 5.8 ppm, and these correspond to OH3 and OH2/OH4 (Figure 5a). The OH3 signal gives a relatively sharp peak at 0 °C, suggesting that OH3 is stabilized by hydrogen bonding. To validate the exchange of the hydroxyl proton, the temperature coefficient (ppb K−1) was estimated as 11 ppb K−1 for ribitol OH3. According to Sandströ et al., the temperature coefficient is greater than 11 ppb K−1 when the OH proton is fully hydrated [8]. This suggests that OH3 is mostly exposed to solvent, but some transient hydrogen bond may exist. The exchange rate constant of OH3 was also estimated from 2D EXSY. From the build-up curve of exchange peak volume, it was estimated as kex = 193 s−1. To estimate the effect of an intramolecular hydrogen bond on the temperature coefficient and OH exchange rate, the same experiment was performed using the model compound 1,3,5-pentanol (Figure 1), which does not have hydroxyl groups at the 2 and 4 positions. 1,3,5-Pentanol gives a broad OH peak around 5.7 ppm at 0 °C. This peak can be assigned to OH1/OH3/OH5, and the peak almost disappears

Molecules 2021, 26, 5471 at 25 °C (Figure 5b). We reasoned that this is due to the lack of a hydrogen bond, 6and of 10 hence, OH3 shows faster proton exchange.

1 FigureFigure 5. 5.PartPart of of1D- 1D-1H-NMRH-NMR (OH (OH region) region) spectra spectra of ribitol of ribitol and and1,3,5-pentanol 1,3,5-pentanol measured measured at 0, at5, 10, 0, 5, ◦ 15,10, 20, 15, 25, 20, and 25, 30 and °C in 30 10C mM in 10sodium mM sodiumacetate buffer, acetate pH buffer, 6.0 (H pH2O:D 6.02O (H = 21:1)O:D.2 (Oa) =Ribitol. 1:1). ((ab)) Ribitol.1,3,5- Pentanol.(b) 1,3,5-Pentanol. x = impurity. x = impurity.

ForFor comparison, comparison, the the temperature temperature coefficient coefﬁcient of 1,3,5-pentanol of 1,3,5-pentanol was was calculated calculated as 13 as −1 ppb13 ppbK−1 for K OH1/3/5.for OH1/3/5. Similarly, Similarly, the exchange the exchange rate of OH1/3/5 rate of OH1/3/5 was estimated was estimated from EXSY from to −1 beEXSY kex = 650 tobe s−1.k Theseex = 650 observations s . These suggest observations that ribi suggesttol OH3 thattends ribitol to form OH3 water-mediated tends to form orwater-mediated intramolecular orhydrogen intramolecular bonds. hydrogen bonds. 2.3. MD Simulations 2.3. MD Simulations Distribution of Main Chain Dihedral Angles Distribution of Main Chain Dihedral Angles Since solution NMR does not distinguish between 3J(H2,H3) and 3J(H3,H4), MD 3 3 simulationsSince solution were performedNMR does tonot shed distinguish light on between the dihedral J(H2,H3) angles and of ϕ J2(H3,H4), and ϕ3 separately.MD sim- ulationsWe used were three performed ribitol coordinates to shed light in PDB on the (PDB dihedral ID: 5IAI, angles 4Q0S, of 4F2D)φ2 and as φ an3 separately. initial structure We used(Supplementary three ribitol coordinates Table S1). Independent in PDB (PDB MD ID: simulations 5IAI, 4Q0S, (referred 4F2D) as to an as initial Run #1,structure #2, and (Supplementary#3, respectively) Table were S1). performed Independent for 100 MD ns forsimulations each initial (referred structure. to as The Run MD #1, simulation #2, and #3,results respectively) are essentially were performed the same; therefore,for 100 ns thefor followingeach initial discussion structure. is The based MD on simulation the results resultsof Runs are #1–#3. essentially Figure the6 showssame; therefore, the results the of MDfollowing simulations discussion of ribitol is based (left, on Run the #1)results and of1,3,5-pentanol Runs #1–#3. Figure (right) 6 for shows each dihedralthe results angle of MDϕ1 simulations to ϕ4. The MD of ribitol simulation (left, resultsRun #1) of and Run 1,3,5-pentanol#2 and #3 are (right) shown for in Supplementaryeach dihedral angle Figure φ1 S1.to φ4. The MD simulation results of Run #2 and #3 are shown in Supplementary Figure S1. Table 3. Comparison of calculated and experimental coupling constants of ribitol. Calculated J values are obtained from the average of three independent MD simulations (mean ± SD).

Calculated J Value (Hz) Experimental J Value (Hz) 3J(H1R,H2)(pro-R) 3.11 ± 0.05 3.00 3 J(H1S,H2)(pro-S) 7.57 ± 1.02 7.20 3J(H2,H3) 6.41 ± 0.09 6.50 3J(C1,H3) 4.20 ± 0.36 3.9 3J(C3,H1S) 2.67 ± 0.64 3.1 It seems that ϕ2 and ϕ3 tend to share the same dihedral angle, i.e., (ϕ2, ϕ3) = (+60◦, +60◦) or (−60◦, −60◦) during the MD trajectory. This differs from the database analysis, in which (ϕ2, ϕ3) mostly adopts (anti, gauche) or (gauche, anti). This discrepancy may come from the intermolecular contacts in the crystal lattice. The combination of ϕ2 and ϕ3 (+60◦, +60◦ or −60◦, −60◦) can avoid steric hindrance between O2 and O4, and this arrangement seems reasonable. MD simulations of 1,3,5-pentanol were also performed to investigate the role of OH2 and OH4 in ribitol (Figure5, right). In 1,3,5-pentanol, ϕ1 and ϕ4 assume all three conformations (180◦, +60◦, −60◦) almost equally (Supplementary Table S2). ϕ2 and ϕ3 also assume all three conformations, but −60◦ is slightly less populated for ϕ2 and +60◦ for ϕ3. Overall, transitions are frequent among conformers in 1,3,5-pentanol, while they are less so in ribitol. These results suggest that the presence of OH groups at positions 2 and 4 affects the frequency of conformational transitions. It is likely that this is because ribitol Molecules 2021, 26, 5471 7 of 10

Molecules 2021, 26, 5471 7 of 11 OH2 and OH4 form water-mediated or intramolecular hydrogen bonds that restructure the molecule.

Figure 6. Dihedral angle distribution of ribitol (left, Run#1) and 1,3,5-pentanol (right) in MD simula- Figuretions. The6. Dihedral results ofangle each distribution dihedral angle of ribitol are shown (left, Run#1) after 10 and ns. The1,3,5-pentanol dihedral angles (right) are in deﬁnedMD simu- as lations.ϕ1 (O1–C1–C2–C3), The results ofϕ each2 (C1–C2–C3–C4), dihedral angleϕ are3 (C2–C3–C4–C5), shown after 10 andns. Theϕ4 (C3–C4–C5–O5).dihedral angles areϕ1 defined and ϕ4 as of φribitol1 (O1–C1–C2–C3), are mostly distributed φ2 (C1–C2–C3–C4), in anti (180◦ )φ and3 (C2–C3–C4–C5), gauche (−60◦ for andϕ 1φ and4 (C3–C4–C5–O5). +60◦ for ϕ4) (Supplementary φ1 and φ4 of ribitolTable S2).are mostly This is consistentdistributed with in anti the (180°) NMR and result, gauche in which (−60° the for populationsφ1 and +60° offor dihedral φ4) (Supplementary angle ϕ1 are Table S2). This is consistent with the NMR result, in which the populations of dihedral angle φ1 are estimated as 180◦:−60◦:60◦ = 64:36:0. For ϕ2 and ϕ3 of ribitol, gauche conformations (−60◦ and estimated as 180°:−60°:60° = 64:36:0. For φ2 and φ3 of ribitol, gauche conformations (−60° and +60°) +60◦) are dominant, but the anti-conformation is also populated. This is consistent with the NMR are dominant, but the anti-conformation is also populated. This is consistent with the NMR result ◦ ◦ ◦ 3 3 (180°:result− (18060°:+60°:−60 = 2:46:52).:+60 = 2:46:52).To validate To validate the MD the results MD resultsquantitatively, quantitatively, 3J(H,H) Jand(H,H) 3J(C,H) and Jare(C,H) calcu- are latedcalculated from fromthe population the population of anti of and anti ga anduche gauche conformations, conformations, and the and calculated the calculated J valuesJ values are com- are paredcompared with with experimental experimental values values (Table (Table 3).3 ).The The calculated calculated J valuesJ values are are mostly mostly consistent consistent with with the experimentalexperimental values, values, suggesting suggesting that that MD MD simulation simulation quantitatively reflects reﬂects the energy level of each conformerconformer based based on on the the Boltzmann distribution.

Molecules 2021, 26, 5471 8 of 10

3. Materials and Methods 3.1. Database Analysis of Ribitol and Ribitol Phosphate Three-dimensional (3D) crystal structures of ribitol and ribitol phosphate (resolution less than 2.7 Å) were extracted from the Cambridge Crystallographic Data Centre (CCDC, as of May, 2020) and the Protein Data Bank (PDB, as of April 2020). Coordinates of ribitol phosphate were extracted that had at least one terminal CH2OH group. When analyzing tandem ribitol phosphate, the terminal ribitol with the CH2OH group was chosen. The dihedral angles of ribitol and ribitol phosphate are deﬁned as ϕ1 (O1–C1–C2–C3), ϕ2 (C1–C2–C3–C4), ϕ3 (C2–C3–C4–C5), and ϕ4 (C3–C4–C5–O5).

3.2. Solution NMR Analysis Ribitol was purchased from Tokyo Chemical Industry (Tokyo, Japan), and 1,3,5- pentanol was purchased from ChemCruz. Solution NMR analyses were performed using JNM-ECZ600R/S1, JNM-ECZ400S/L1 and EX270 spectrometers (JEOL, Tokyo, Japan). The probe temperature was set from 273 to 303 K. A sample (18–30 mg) was dissolved in 650 uL of D2O for signal assignment or 10 mM sodium acetate buffer, pH 6.0 (H2O:D2O = 1:1) for detection of OH chemical shifts. 1H and 13C chemical shits were reportedly related to the internal standard of 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS, 0 ppm). NMR chemical shifts of ribitol were assigned by analyzing 1D-1H, 1D-13C, 2D-DQF-COSY, NOESY, TOCSY, and 1H-13C HSQC spectra. NMR chemical shifts of ribitol phosphate were assigned by analyzing 1D-1H, 1D-13C, 2D-1H-13C HSQC, and 2D-1H-13C HMBC spectra. Stereospeciﬁc assignment of pro-R and pro-S protons in ribitol was conducted with the aid of 1D-1H-NMR spectral simulation by Mnova 14.1.1 (Mestrelab Research, Santiago, Spain). Due to the presence of a second-order term, the 3J(H,H) and 2J(H,H) coupling constants of ribitol were obtained by iterative comparison of simulated NMR spectra with observed NMR spectra obtained at 1H observation frequencies of 270, 400, and 600 MHz. 3J(C,H) coupling constants were obtained from HR-HMBC spectra [18]. A scaling factor (n) of 25 was used, and the digital resolutions for f1 (13C) and f2 (1H) were 4.4 and 0.7 Hz/point, respectively. The conversion from coupling constant 3J(H,H) to dihedral angle was done using the Haasnoot equation, which includes a correction for electronegativity of substituents and the orientation of each substituent relative to the coupled protons [20]. 3J(C,H) is also applied to an empirical prediction equation [21,22]. The exchange rate of hydroxyl protons with water was calculated from 2D chemical exchange spectroscopy [23,24]. Mixing times of 3 to 24 ms with increments of 3 ms were used. The exchange rate constant was calculated from the initial build-up rate of the diagonal peak volume over the exchange cross-peak. The temperature coefﬁcient (ppb K–1) of hydroxyl protons in ribitol and 1,3,5-pentanol was measured by collecting a series of 1D-1H-NMR spectra at different temperatures (273, 278, 283, 288, 293, 298, and 303 K). NMR data processing was performed using Delta5. 3. 1 (JEOL, Tokyo, Japan), and NMR spectral analyses were performed using Mnova.

3.3. Solid-State 13C-CP-MAS NMR 13C CP-MAS spectra of crystalline ribitol (Tokyo Chemical Industry, Tokyo, Japan) were obtained using a Bruker Avance III 500 spectrometer at a 1H frequency of 500 MHz. A VPN probe (4 mm) was used at the spinning rate of 15,000 Hz or 8000 Hz with a spectral width of 300 ppm or 200 ppm, respectively. The data point was set to 2k. 13C chemical shifts were referenced to carboxyl carbon of glycine (176.03 ppm).

3.4. MD Simulation MD simulations were performed using Discovery Studio 2019 [25]. The coordinates of ribitol (PDB ID: 5IAI, 4Q0S and 4F2D) were used as the initial structure (Supplementary Table S1), and the three simulations were performed independently. CHARMm36 was assigned as the force field [10]. Hydrogens were generated using the “Add Hydrogens” protocol in Discovery Studio. The simulation time was set to 100 ns. An explicit periodic boundary was used as Molecules 2021, 26, 5471 9 of 10

the solvation model. An orthorhombic cell shape was used in the explicit periodic boundary solvation model. The minimum distance from the periodic boundary was set to 7.0 Å. For each ribitol coordinate, 182–184 water molecules were explicitly placed based on the protocol in Discovery Studio 2019. CHARMm36 was optimized using the TIP3P water model [26], which was used as the force field template. Explicit waters were 190 for 1,3,5-pentanol. Minimization of the initial coordinate was done in two steps. The first step is to eliminate distortion of the entire structure with the steepest descent algorithm. In the second step, minimization was performed with the adopted basis Newton–Raphson (NR) algorithm. Heating was done at 500 K. After equilibration, the time step was set to 2 fs, and production was done under nPT ensemble. MD simulations were also performed for 1,3,5-pentanol under the same protocol. The initial structure of 1,3,5-pentanol is a zigzag form with each dihedral angle of the carbon main chain being 180◦.

4. Conclusions Until the present study, it has been assumed that the C-C bonds of alditols are rather unstructured, and the molecule exhibits little or no stable folded structure. Here, NMR and MD simulation of ribitol reveal that ribitol is a dynamic conformational equilibrium of staggered conformations. Then, ribitol, is rather structured compared with model compound 1,3,5-pentanol. Conformational asymmetry may prevail in ribitol due to repulsion between OH2 and OH4 and to transient weak hydrogen bonds involving OH3. These ﬁndings provide a basis for understanding the dynamic structure of ribitol and the function of ribitol-related enzymes involved in ribitol biosynthesis and glycan chain elongation.

Supplementary Materials: The following are available online, Figure S1: Dihedral angle distribution of ribitol in MD simulations Run #2 and #3, Table S1: The initial XYZ coordinates for MD simulations, and Table S2: The population (%) of dihedral angles in MD simulations of ribitol (Runs #1–#3) and 1,3,5-pentanol. Author Contributions: Manuscript conception, writing and original draft preparation, S.O., Y.Y.; NMR analysis, S.O., T.Y., Y.Y.; MD simulations, S.O.; editing, data analysis, and interpretation, S.O., N.M., J.U., Y.Y. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported in part by grants-in-aid of KAKENHI 19H05648 and 19H03362. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Acknowledgments: We acknowledge Ohgi Takahashi (Shonan University of Medical Sciences) for technical advices on computational calculations, and Hiroshi Manya and Tamao Endo (Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology) and Jun-ichi Tamura (Tottori University) for intriguing discussions and encouragement. We also express our thanks to Tomoyuki Matsuki and Shin-ichi Sato (Tohoku Medical and Pharmaceutical University) for NMR measurements. This work was partly supported by Nanotechnology Platform Program (Molecule and Material Synthesis) of MEXT. Conﬂicts of Interest: The authors declare no conﬂict of interest. Sample Availability: Samples of the compounds are not available from the authors.

References 1. Brown, S.; Santa Maria, J.P., Jr.; Walker, S. Wall teichoic acids of gram-positive bacteria. Annu. Rev. Microbiol. 2013, 67, 313–336. [CrossRef] 2. Kanagawa, M.; Kobayashi, K.; Tajiri, M.; Manya, H.; Kuga, A.; Yamaguchi, Y.; Akasaka-Manya, K.; Furukawa, J.I.; Mizuno, M.; Kawakami, H.; et al. Identiﬁcation of a post-translational modiﬁcation with ribitol-phosphate and its defect in muscular dystrophy. Cell Rep. 2016, 14, 2209–2223. [CrossRef][PubMed] 3. Endo, T. Mammalian O-mannosyl glycans: Biochemistry and glycopathology. Proc. Jpn. Acad. Ser. B Phys. Biol. Sci. 2019, 95, 39–51. [CrossRef][PubMed] 4. Jeffrey, G.A.; Kim, H.S. Conformations of the alditols. Carbohydr. Res. 1970, 14, 207–216. [CrossRef] Molecules 2021, 26, 5471 10 of 10

5. Tanaka, N.; Ashida, T.; Sasada, Y.; Kakudo, M. The crystal structure of riboflavin hydrobromide monohydrate. Bull. Chem. Soc. Jpn. 1969, 42, 1546–1554. [CrossRef][PubMed] 6. Tanaka, N.; Ashida, T.; Sasada, Y.; Kakudo, M. The crystal structure of riboflavin hydrobromide monohydrate. Bull. Chem. Soc. Jpn. 1967, 40, 1739. [CrossRef][PubMed] 7. Hawkes, G.E.; Lewis, D. 1H nuclear magnetic resonance spectra and conformations of alditols in deuterium oxide. J. Chem. Soc. Perk. Trans. 1984, 2073–2078. [CrossRef] 8. Garrett, E.C.; Serianni, A.S. (1-13C)Alditols: Elimination of magnetic equivalence in 1H- and 13C-n.m.r. spectra of symmetric compounds through (13C)-substitution. Carbohydr. Res. 1990, 208, 23–35. [CrossRef] 9. Klein, R.A.; Hartmann, R.; Egge, H.; Behr, T.; Fischer, W. The aqueous solution structure of the tetrasaccharide-ribitol repeat-unit from the lipoteichoic acid of Streptococcus pneumoniae strain R6 determined using a combination of NMR spectroscopy and computer calculations. Carbohydr. Res. 1994, 256, 189–222. [CrossRef] 10. Hatcher, E.; Guvench, O.; Mackerell, A.D., Jr. CHARMM additive all-atom force field for acyclic polyalcohols, acyclic carbohy- drates and inositol. J. Chem. Theory Comput. 2009, 5, 1315–1327. [CrossRef] 11. Imberty, A. Oligosaccharide structures: Theory versus experiment. Curr. Opin. Struct. Biol. 1997, 7, 617–623. [CrossRef] 12. Kim, H.S.; Jeffrey, G.A.; Rosenste, R.D. Crystal structure of ribitol. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1969, 25, 2223–2230. [CrossRef] 13. Madsen, A.O.; Larsen, S. Insight into solid-state entropy from diffraction data. Angew. Chem. Int. Ed. Engl. 2007, 46, 8609–8613. [CrossRef][PubMed] 14. Madsen, A.O.; Mattson, R.; Larsen, S. Understanding thermodynamic properties at the molecular level: Multiple temperature charge density study of ribitol and xylitol. J. Phys. Chem. A 2011, 115, 7794–7804. [CrossRef][PubMed] 15. Yoshida, H.; Yoshihara, A.; Teraoka, M.; Terami, Y.; Takata, G.; Izumori, K.; Kamitori, S. X-ray structure of a novel L-ribose isomerase acting on a non-natural sugar L-ribose as its ideal substrate. FEBS J. 2014, 281, 3150–3164. [CrossRef][PubMed] 16. Gerlach, D.; Guo, Y.; De Castro, C.; Kim, S.H.; Schlatterer, K.; Xu, F.F.; Pereira, C.; Seeberger, P.H.; Ali, S.; Codee, J.; et al. Methicillin- resistant Staphylococcus aureus alters cell wall glycosylation to evade immunity. Nature 2018, 563, 705–709. [CrossRef][PubMed] 17. Kuwabara, N.; Imae, R.; Manya, H.; Tanaka, T.; Mizuno, M.; Tsumoto, H.; Kanagawa, M.; Kobayashi, K.; Toda, T.; Senda, T.; et al. Crystal structures of fukutin-related protein (FKRP), a ribitol-phosphate transferase related to muscular dystrophy. Nat. Commun. 2020, 11, 303. [CrossRef] 18. Furihata, K.; Tashiro, M.; Seto, H. High resolution-HMBC (HR-HMBC), a new method for measuring heteronuclear long-range coupling constants. Magn. Reson. Chem. 2010, 48, 179–183. [CrossRef] 19. Franks, F.; Kay, R.L.; Dadok, J. A nuclear magnetic resonance study of isomeric pentitols in aqueous and non-aqueous solutions. J. Chem. Soc. Farad. Trans. 1988, 84, 2595–2602. [CrossRef] 20. Haasnoot, C.A.G.; Deleeuw, F.A.A.M.; Altona, C. The relationship between proton-proton NMR coupling constants and substituent electronegativities. 1. An empirical generalization of the Karplus equation. Tetrahedron 1980, 36, 2783–2792. [CrossRef] 21. Palermo, G.; Riccio, R.; Bifulco, G. Effect of electronegative substituents and angular dependence on the heteronuclear spin-spin coupling constant 3J(C-H): An empirical prediction equation derived by density functional theory calculations. J. Org. Chem. 2010, 75, 1982–1991. [CrossRef] 22. Aydin, R.; Günther, H. 13C, 1H spin–spin coupling. X—Norbornane: A reinvestigation of the karplus curve for 3J(13C, 1H). Magn. Reson. Chem. 1990, 28, 448–457. [CrossRef] 23. Dobson, C.M.; Lian, L.Y.; Redfield, C.; Topping, K.D. Measurement of hydrogen-exchange rates using 2D NMR-spectroscopy. J. Magn. Reson. 1986, 69, 201–209. [CrossRef] 24. Sandstrom, C.; Baumann, H.; Kenne, L. The use of chemical shifts of hydroxy protons of oligosaccharides as conformational probes for NMR studies in aqueous solution. Evidence for persistent hydrogen bond interaction in branched trisaccharides. J. Chem. Soc. Perk. Trans. 1998, 2385–2393. [CrossRef] 25. BIOVIA Discovery Studio. Discovery Studio 2019; Dassault Systèmes: San Diego, CA, USA, 2019. 26. Jorgensen, W.L.; Chandrasekhar, J.; Madura, J.D.; Impey, R.W.; Klein, M.L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983, 79, 926–935. [CrossRef]