Conformation of the Galactose Ring Adopted in Solution and in Crystalline Form as Determined by Experimental and DFT 1H NMR and Single-Crystal X-ray Analysis

Article abstract of DOI:10.1021/jo035400u

The solution-state conformations of various galactose derivatives were determined by comparison of the experimental ¹H-¹H vicinal coupling constants to those calculated using density functional theory (DFT) at the B3LYP/cc-pVTZ//B3LYP/6-31G(d,p) level of theory. The agreement between the experimental and calculated vicinal coupling constants for 1,2:3,4-di-O-isopropylidene-α-D-galactopyranose was good, thereby confirming an ^OS₂ skew conformation for it and its derivatives on the basis of their similar observed couplings. Single-crystal X-ray analysis of 1,2:3,4-di-O-isopropylidene-6-O-(3,4, 6-tri-O-acetyl-2-deoxy-2-N-phthalimido-β-D-glucopyranosyl) -α-D-galactopyranose and 1,2,3,4,6-penta-O-acetyl-α -D-galactopyranose provided ^OS₂ and ⁴C ₁ conformations, respectively, for the galactose ring in the solid state. The solid-state structures proved to be suitable starting structures for further DFT structure refinement or for immediate calculation of the coupling constants.

Full text of DOI:10.1021/jo035400u

Con for m a tion of th e Ga la ctose Rin g Ad op ted in Solu tion a n d in
1
Cr ysta llin e F or m a s Deter m in ed by Exp er im en ta l a n d DF T H
NMR a n d Sin gle-Cr ysta l X-r a y An a lysis
Mattias U. Roslund,*,†,‡ Karel D. Klika,*,§ Reko L. Lehtila¨,†,‡ Petri Ta¨htinen,§
Reijo Sillanpa¨a¨,| and Reko Leino†,‡
Department of Organic Chemistry, A° bo Akademi University, Biskopsgatan 8, FIN-20500 A° bo, Finland,
Biotie Therapies Corp., Viikinkaari 9, FIN-00710 Helsinki, Finland, Structural Chemistry Group,
Department of Chemistry, University of Turku, Vatselankatu 2, FIN-20014 Turku, Finland, and
Department of Chemistry, University of J yva¨skyla¨, Survontie 9, FIN-40014 J yva¨skyla¨, Finland
mattias.roslund@abo.fi; karel.klika@utu.fi
Received September 24, 2003
The solution-state conformations of various galactose derivatives were determined by comparison
1
of the experimental H-1H vicinal coupling constants to those calculated using density functional
theory (DFT) at the B3LYP/cc-pVTZ//B3LYP/6-31G(d,p) level of theory. The agreement between
the experimental and calculated vicinal coupling constants for 1,2:3,4-di-O-isopropylidene-R-D-
galactopyranose was good, thereby confirming an OS2 skew conformation for it and its derivatives
on the basis of their similar observed couplings. Single-crystal X-ray analysis of 1,2:3,4-di-O-
isopropylidene-6-O-(3,4,6-tri-O-acetyl-2-deoxy-2-N-phthalimido-â-D-glucopyranosyl)-R-D-galacto-
pyranose and 1,2,3,4,6-penta-O-acetyl-R-D-galactopyranose provided OS2 and 4C1 conformations,
respectively, for the galactose ring in the solid state. The solid-state structures proved to be suitable
starting structures for further DFT structure refinement or for immediate calculation of the coupling
constants.
4
In tr od u ction
tion of the galactose ring in 1 from a C1 chair conformer
to a skew4 (twist-boat) conformer;5 even formation of a
1,2-O-isopropylidene or 3,4-O-isopropylidene derivative
results in a conformational mixture consisting of chair
and skew conformers for the product.5b These structural
changes are evidenced by the 1H-1H vicinal coupling
constants (3J H,H) of the ring protons in these compounds
as well as by their altered reactivities. It is widely
accepted that the ring conformation can readily be
ascertained by application of the Karplus equation,6 and
although the Karplus relationship has stood up well for
almost half a century, there are problems associated7-12
with the method, viz. obtaining good values for the
constants8,10,12,13 in the particular variant of the equa-
Carbohydrates occur widely in all living matter and
function as structural or protective materials, as energy
stores, and even appear to be essential to the process of
infection by certain pathogenic species; as such, they have
a significant biological role.1 The conformational analysis
of carbohydrates, and in particular, the ring conforma-
tion, is of interest as biological activities, as well as
chemical and physical properties, are determined to
various degrees by the adopted conformations of the
substrates.2 The conformation adopted by a sugar ring,
however, can change markedly upon derivatization. For
example, the isopropylidene acetal group is a common
protecting group widely used3 in carbohydrate chemistry
for 1,2-diols and when applied to D-galactose (1), the
formation of 1,2:3,4-di-O-isopropylidene-R-D-galacto-
pyranose (2) is known to change the preferred conforma-
(4) According to IUPAC nomenclature. McNaught, A. D. Pure Appl.
Chem. 1996, 68, 1919-2008.
(5) (a) Cone, C.; Hough, L. Carbohydr. Res. 1965, 1, 1-9. (b)
Midland, M. M.; Asirwatham, G.; Cheng, J . G.; Miller, J . A.; Morell,
L. A. J . Org. Chem. 1994, 59, 4438-4442.
* Corresponding authors. (M.U.R.) Phone: 358-(2)-2154501. Fax:
358-(2)-2154866. (K.D.K.) Phone: 358-(2)-3336762. Fax: 358-(2)-
3336700.
(6) (a) Karplus, M. J . Am. Chem. Soc. 1963, 85, 2870-2871. (b)
Karplus, M. J . Chem. Phys. 1959, 30, 11-15. (c) Karplus, M. J . Phys.
Chem. 1960, 64, 1793-1798.
† A° bo Akademi University.
(7) Buys, H. R. Recueil 1969, 88, 1003-1011.
‡ Biotie Therapies Corp.
(8) Haasnoot, C. A. G.; de Leeuw, F. A. A. M.; Altona, C. Tetrahedron
1980, 36, 2783-2792.
§ University of Turku.
| University of J yva¨skyla¨.
(9) (a) Lambert, J . B. Acc. Chem. Res. 1971, 4, 87-93. (b) Lambert,
J . B.; Papay, J . J .; Khan, S. A.; Kappauf, K. A.; Magyar, E. S. J . Am.
Chem. Soc. 1974, 96, 6112-6118. (c) Lambert, J . B.; Sun, H.-N. Org.
Magn. Reson. 1977, 9, 621-626.
(1) (a) Dwek, R. A. Chem. Rev. 1996, 96, 683-720. (b) Essentials of
Glycobiology; Varki, A., Cummings, R., Esko, J ., Freeze, H., Hart, G.,
Marth, J ., Eds.; Cold Spring Harbor Laboratory Press: Cold Spring
Harbor, NY, 1999.
(2) J ames, L. C.; Roversi, P.; Tawfik, D. S. Science 2003, 299, 1362-
1367.
(3) (a) Rauter, A. P.; Ramoˆa-Ribeiro, F.; Fernandes, A. C.; Figueire-
do, J . A. Tetrahedron 1995, 51, 6529-6540 and references therein. (b)
Clode, D. M. Chem. Rev. 1979, 79, 491-513.
(10) Aliev, A. EÄ .; Sinitsyna, A. A. Bull. Russ. Acad. Sci.sDiv. Chem.
Sci. 1992, 41, 1143-1160.
(11) Haasnoot, C. A. G. J . Am. Chem. Soc. 1993, 115, 1460-1468.
(12) (a) Wu, A.; Cremer, D. J . Phys. Chem. A 2003, 107, 1797-1810.
(b) Wu, A.; Cremer, D.; Auer, A. A.; Gauss, J . J . Phys. Chem. A 2002,
106, 657-667.
10.1021/jo035400u CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/11/2003
18
J . Org. Chem. 2004, 69, 18-25

Conformation of the Galactose Ring
SCHEME 1. Rea ction Sequ en ces Lea d in g to
tion that is to be applied. There are a number of fact-
ors,6a,9,10,13-15 aside from the torsion angle, such as
substituent effects (electronegativity of the substituents),
steric and stereoelectronic factors, hybridization states,
interceding bond angles and lengths, torsional vibrations,
and even solvent effects, etc., that influence the magni-
tude of 3J H,H. Indeed, for each vicinal pair of protons in a
set molecule, a different set of constants can be consid-
ered to be in effect, rendering the approach laborious and
overly complicated even for small molecules where a
conformation might nevertheless be defined by a rela-
tively small number of coupling constants, such as the
conformation of a sugar ring. Thus, the results can be
misleading despite quality attempts8 to surmount the
problem by uniquely defining for each vicinal pair a set
of appropriate constants, wherein the problem quickly
becomes complex and the exercise self-defeating.
Com p ou n d s 2-4 Exa m in ed in Th is Stu d ya
A conformational re-examination of 2, together with
the analogous compounds 1,2:3,4-di-O-isopropylidene-6-
O-(3,4,6-tri-O-acetyl-2-deoxy-2-N-phthalimido-â-D-glu-
copyranosyl)-R-D-galactopyranose (3) and 1,2:3,4-di-O-
isopropylidene-6-O-tert-butyldiphenylsilyl-R-D-galactopy-
ranose (4), was considered worthwhile, as there has been
some doubt and confusion5a regarding the conformation
of 2. In this study, we desired to either confirm the
proposed OS2 conformer as the dominant conformer for
2-4 or assess the system as a confluence of contributing
conformers (i.e., a dynamic equilibrium). A dynamic
conformational equilibrium is not unexpected given that
the energy barrier for pseudorotation can be very low. It
was necessary, therefore, not only to identify the par-
ticular skew form(s) by reliably extracting the J H,Hs but
also to apply a methodology that could faithfully guar-
antee the provision of accurate J H,Hs. A recent methodol-
ogy that has been developed16 and applied17 to confor-
mational analysis is the calculation of J H,Hs using density
functional theory (DFT). One important criterion for
success, however, is the need for good geometry optimiza-
tion, and for this structures obtained from X-ray analysis
were used directly for calculation of the coupling con-
stants as well as for initial starting structures for further
DFT geometry optimization.
a
The 4C1 conformation adopted by both R- and â-D-galactose
(1a and 1b) is altered to an OS2 skew conformation in compounds
2-4. The numbering system in use for compounds 1-5 is also
indicated. Legend: (a) CuSO4, H2SO4, acetone, 24 h, rt, 95%; (b)
3,4,6-tri-O-acetyl-2-deoxy-2-N-phthalimido-R-D-bromoglucose, to-
gether with 2,2,6,6-tetramethylpiperidine, 4 Å molecular sieves,
and AgOTf, 5 h, -50 to 0 °C, 70%; (c) TBDPSCl, DBU, 17 h, 0 °C
to room temperature, 82%.
biologically significant oligosaccharides, an altered pro-
cedure was applied. The introduction of isopropylidene
groups to produce 2 from D-galactose (1), however, can
be accomplished by a number of methods: homoge-
neously using dry acetone and acid18b,21 or a Lewis acid
catalyst such as zinc chloride;22 heterogeneously using
acetone and Dowex23 or Amberlyst ion-exchange resins,24
zeolite catalysts,3a or montmorillonite clay;18e or by acid-
catalyzed transacetalization using an appropriate iso-
Resu lts a n d Discu ssion
(18) (a) Wolfrom, M. L.; Thompson, A. In The Carbohydrates;
Pigman, W. W., Ed.; Academic Press: New York, 1957; pp 236-240.
(b) Horton, D.; Nakadate, M.; Tronchet, J . M. J . Carbohydr. Res. 1968,
7, 56-65. (c) Morgenlie, S. Acta Chem. Scand. 1973, 27, 3609-3610.
(d) J arosz, S.; Krajewski, J . W.; Zamojski, A.; Duddeck, H.; Kaiser, M.
Bull. Pol. Acad. Sci., Chem. 1985, 33, 181-187; Chem Abstr. 1986,
104, 149270m. (e) Asakura, J .; Matsubara, Y.; Yoshihara, M. J .
Carbohydr. Chem. 1996, 15, 231-239 and references therein. (f) Cao,
S.; Herna´ndez-Mate´o, F.; Roy, R. J . Carbohydr. Chem. 1998, 17, 609-
631.
(19) (a) Ito, Y.; Ogawa, T. Tetrahedron Lett. 1988, 29, 1061-1064.
(b) Ito, Y.; Ogawa, T.; Numata, M.; Sugimoto, M. Carbohydr. Res. 1990,
202, 165-175. (c) Kunz, H.; Zimmer, J . Tetrahedron Lett. 1993, 34,
2907-2910.
(20) Nashed, E. M.; Glaudemans, C. P. J . J . Org. Chem. 1987, 52,
5255-5260.
The synthesis of the known compounds 2,3a,18 3,18f,19
and 4,20 followed literature methodology or standard
procedures (Scheme 1). In the case of the protected
disaccharide 3, an important building block for certain
(13) (a) Oh sawa, E.; Ouchi, T.; Saito, N.; Yamato, M.; Lee, O. S.; Seo,
M.-K. Magn. Reson. Chem. 1992, 30, 1104-1110. (b) Barfield, M.;
Smith, W. B. J . Am. Chem. Soc. 1992, 114, 1574-1581. (c) Smith, W.
B.; Barfield, M. Magn. Reson. Chem. 1993, 31, 696-697. (d) Smith,
W. B.; Barfield, M. Magn. Reson. Chem. 1996, 34, 740. (e) Altona, C.;
Francke, R.; de Haan, R.; Ippel, J . H.; Daalmans, G. J .; Westra
Hoekzema, A. J . A.; van Wijk, J . Magn. Reson. Chem. 1994, 32, 670-
678.
(14) Maciel, G. E.; Mclver, J . W., J r.; Ostlund, N. S.; Pople, J . A. J .
Am. Chem. Soc. 1970, 92, 4497-4506.
(21) Schmidt, O. T. Methods Carbohydr. Chem. 1963, 2, 318-325.
(22) Banks, M. R.; Blake, J .; Cadogan, J . I. G.; Dawson, I. M.; Gaur,
S.; Gosney, I.; Gould, R. O.; Grant, K. J .; Hodgson, K. G. J . Chem.
Soc., Chem. Commun. 1993, 1146-1148.
(23) Lorette, N. B.; Howard, W. L.; Brown, J . R., J r. J . Org. Chem.
1959, 24, 1731-1733.
(15) (a) Booth, H. Tetrahedron Lett. 1965, 411-416. (b) Sternhell,
S. Quart. Rev. 1969, 23, 236-270.
(16) (a) Bagno, A. Chem. Eur. J . 2000, 6, 2925-2930. (b) Bagno, A.
Chem. Eur. J . 2001, 7, 1652-1661.
(17) (a) Ta¨htinen, P.; Bagno, A.; Klika, K. D.; Pihlaja, K. J . Am.
Chem. Soc. 2003, 125, 4609-4618 and references therein. (b) Pihlaja,
K.; Ta¨htinen, P.; Klika, K. D.; J okela, T.; Salakka, A.; Wa¨ha¨la¨, K. J .
Org. Chem. 2003, 68, 6864-6869.
(24) (a) J ones, E. R. H.; Meakins, G. D.; Pragnell, J .; Mu¨ller, W. E.;
Wilkins, A. L. J . Chem. Soc., Perkin Trans. 1 1974, 2376-2380. (b)
Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic Synthesis,
3rd ed.; Wiley & Sons: New York, 1999; pp 207-215.
J . Org. Chem, Vol. 69, No. 1, 2004 19

Roslund et al.
propylidene source such as 2,2,-dimethoxypropane.18e
Many other variations on these themes have also been
reported.3a,18e In this work, acid-catalyzed derivatization
in the presence of CuSO4 was utilized without the use of
DMF as a cosolvent at ambient temperature, as this
provided the thermodynamically more stable 1,2:3,4-di-
O-isopropylidene pyranose derivative (2) in virtually
quantitative yield18b in contrast to other conditions3a,18c
where the 1,2:5,6-di-O-isopropylidene furanose derivative
was also obtained in significant amounts as a kinetic
product. In addition to compounds 2-4, the pentaacetate
of 1a (1,2,3,4,6-penta-O-acetyl-R-D-galactopyranose, 5a )
was also prepared using literature methodology for
conformational analysis both in solution and in the
crystalline form. Its anomer, 5b, was also examined in
the solution state.
1b. For the glucose moiety of 3, the observed coupling
constants were also fully consistent with the adoption of
an ideal 4C1 chair conformation by the glucosyl pyranose
ring. However, only very small differences between the
conformations adopted by the compounds 2-4 were in
evidence, as essentially the same vicinal ring coupling
constants were observed for all three compounds for the
galactose moiety, which is clearly not in an ideal chair
1
4C1 conformation, or for that matter, in an inversed C4
conformation. Thus, neither the glucose group in 3 nor
the tert-butyldiphenylsilyl group in 4 appeared to influ-
ence to any great extent the galactose ring conformation
in their respective structures. Thus, it was deemed
appropriate to concentrate the analysis on 2 and refer
the conclusions drawn to 3 and 4 by analogy. Since
substitution at O-6 in the galactose ring for compounds
2-4 has only a negligible effect on the galactose ring
conformation, it is therefore the isopropylidene groups
that are determinant for the galactose ring conformation
in these compounds.
1
The measurement of the H and 13C NMR parameters
for compounds 1-5 followed and the assignments sub-
sequently made using a standard set of correlation
experiments (see general experimental details in Sup-
porting Information). Previous evaluations of the coupling
constants using the Karplus relationship for comparison
to the experimental values have yielded both a poor set
of values5a and an excellent set,5b in both cases purport-
edly for the same OS2 conformer of 2. Indeed, for the latter
study it seems that due to the higher order of the
spectrum, the experimental value for J H4,H5 was under-
estimated (1.3 Hz), and with the better value extracted
here (1.95 Hz), it matches the previously calculated value
(1.85 Hz) more closely. A third study concluded18d a rigid
conformation intermediate between OS2 and B2,5 to be
present. Needless to say, correct assignment of coupling
constants and reliable extraction of their magnitude is
requisite for correct conformational assessment, which
has not always been the case.5a Even at the high fields
currently available, this is not necessarily trivial in some
instances due to severe signal overlap per se or conse-
quential higher order effects precluding manual extrac-
tion, and only by resorting to spin simulation/iteration
techniques (e.g., the program Perch25) could values be
reliably extracted. In some cases (e.g., 5a ), this was
extremely difficult to do from first principles and changes
of solvent were necessary to provide well-separated
signals where the J H,Hs could be reliably extracted and
then used as a basis for the solution of the problematic
spectra. The 1H and 13C NMR parameters for 1-5
obtained under various conditions are listed in Tables
1-3, respectively. Interestingly, a large number of long-
range couplings were extractable, particularly from H-1
of the R-anomer independent of the conformation adopted.
Moreover, it is difficult to clearly categorize many of these
couplings as w-type, sickle,26 etc.
Of all the skew and boat conformations that are
possible for a galactose ring (six distinct conformers for
each category), inspection of the structures permitted the
immediate elimination of three conformers, 1,4B, B1,4, and
1S3, on the basis of the geometrical constraints of the
isopropylidene groups. For these three structures, the two
isopropylidene groups would both be attached to the
galactose ring via an axial and an equatorial linkage,
which is considered unfavorable in comparison to the
other nine structures where at least one isopropylidene
group is attached favorably, either by two axial, two
equatorial, axial and cis-isoclinal, or equatorial and cis-
isoclinal linkages. The standout conformation is the OS2
conformation where both isopropylidene groups are at-
tached by an axial and a cis-isoclinal linkage, and clearly
this conformer should be the most favored. But the
problem of distinguishing between a confluence of con-
tributing conformers (i.e., a time-averaged spectrum of
several conformers), where the number of permutations
involving up to nine structures abound, and a system that
is conformationally biased can be simplified if the ex-
pected coupling constant values based on the dihedral
angles are too divergent from the observed coupling
constants for the vicinal ring protons and used as a basis
for eliminating that conformer. That is to say, if a
particular coupling constant from the set of coupling
constants for one conformer is expected to have a very
large coupling in contrast to that observed (and con-
versely), then that conformer can be eliminated. On this
basis, six more conformers were eliminated as likely
contributing candidate structures, leaving only three
1
candidates: the OS2, S5, and B2,5 conformers (depicted
for 1a in Scheme 2), although the boat conformation is
considered to be an unlikely contributor on energetic
grounds.
For both the galactoses 1a and 1b, the pyranose ring
adopted a C1 chair conformation as clearly evidenced by
the vicinal couplings involving the ring protons, with the
only significant difference being the coupling between H-1
and H-2 due to the differing configuration at position 1.
The same was true also for the pentaacetates of 1a and
4
It was important to limit the number of candidate
structures to maintain a reasonable, and affordable, set
of structures for the calculations. The problem of still
distinguishing between a conformational equilibrium,
even one that is heavily biased, and the total predomi-
nance of one conformer can be accomplished if the
calculations can predict energetically either system and
the calculated coupling constants (population-weighted
in the case of a conformational equilibrium) match the
(25) See, for example: Laatikainen, R.; Niemitz, M.; Weber, U.;
Sundelin, J .; Hassinen, T.; Vepsa¨la¨inen, J . J . Magn. Reson., Ser.
1996, 120, 1-10. See also Peak Research NMR Software;
Perch Solutions, Ltd.: Kuopio, Finland, 2003 (website: http://
www.perchsolutions.com).
A
(26) Mark, V. Tetrahedron Lett. 1974, 15, 299-302.
20 J . Org. Chem., Vol. 69, No. 1, 2004

Conformation of the Galactose Ring
J . Org. Chem, Vol. 69, No. 1, 2004 21

Roslund et al.
TABLE 2. Exp er im en ta l a n d Ca lcu la ted Vicin a l, Gem in a l, a n d Lon g-Ra n ge J H,Hs (Hz) for Com p ou n d s 1-5 a t 25 °Ca
c
compd
1a d exp
sourceb
conf J H1,H2 J H2,H3 J H3,H4 J H4,H5 J H5,H6 J H5,H6 J H6
J H1,H3 J H1,H4 J H1,H5 J H1,H9 J H2,H4 J H8,H9 J H11,H12
A,H6B
A
B
4C1 3.97 10.33 3.42
1.23 4.47
2.59 nc
7.96
nc
-11.61 -0.43e 0.51e -0.59e
-0.37e
-0.54
-0.54
-0.54
-0.40e
1a
1a
1a
DFT (ex 5a ) 4C1 3.44
9.11 4.20
nc
nc
nc
-0.37 0.65 -0.70
DFT (man) 4C1 4.03
8.85 3.53
4.46 6.66
9.95 3.51
2.42 7.94
2.34 7.97
2.30 8.04
2.29 8.01
2.59 7.33
3.19 7.25
2.62 7.87
2.26 7.99
2.38 7.86
2.42 7.91
1.66 nc
3.86 nc
nc
nc
nc
nc
0.73 nc
0.14 nc
DFT (ex 3) OS2 4.20
1bd exp
4C1 7.93
OS2 5.04
OS2 5.04
OS2 5.06
OS2 4.97
1.11 7.90
1.95 7.15
1.89 5.86
1.74 6.54
1.80 6.28
2.26 nc
4.44
4.61
6.47
6.48
6.57
nc
-11.68 ne
ne
ne
2f,g
2h,i
2h,j
exp
exp
exp
-11.65 <|0.2| ne
-0.59e 0.36c ne
-0.76c -0.71c
-10.81 -0.24e 0.30e -0.53e 0.32c -0.17e -0.64c -0.61c
ne
ne
nc
nc
nc
ne
ne
nc
nc
nc
0.25e -0.56e ne
ne
ne
ne
ne
ne
ne
nc
nc
nc
ne
ne
2h,k exp
ne
ne
ne
nc
nc
nc
ne
ne
2
2
X-ray (ex 3) OS2 4.96
DFT (ex 3) OS2 4.79
0.31 nc
0.31 nc
-0.20 nc
-0.28 nc
nc
ne
ne
2.26 nc
nc
2
Karplus
exp
exp
exp
exp
OS2 4.77
OS2 5.04
OS2 5.03
OS2 5.09
1.85 8.10
1.84 6.36
1.79 7.21
1.83 3.61
4.65
6.56
4.76
8.13
2.32
6.25
6.53
6.54
nc
nc
ne
ne
nc
ne
ne
nc
ne
ne
2l
-10.92 ne
2m
3f,n
3f,o
4f
-11.48 ne
-11.96 -0.27e 0.32e -0.54e 0.20c -0.16e -0.57c -0.62c
4C1 8.51 10.69 9.06 10.18 4.56
-12.23 ne
ne
ne
ne
exp
exp
OS2 4.99
2.32 7.97
1.84 7.11
1.44 6.94
1.46 6.57
1.75 nc
1.83 nc
1.23 6.70
-10.02 -0.24e ne
-0.57e 0.29c ne
-0.64c -0.60c
5a f
4C1 3.69 10.94 3.38
4C1 3.72 10.93 3.42
-11.33 -0.46e 0.48e -0.63e
-0.53e
-0.41e
-0.37
-0.45
-0.39e
5a h exp
-11.27 -0.44e 0.49e -0.63e
5a
X-ray
4C1 4.46
8.94 3.27
nc
nc
-0.37 0.56 -0.62
-0.45 0.56 -0.70
5a
DFT
exp
4C1 3.69 11.48 3.69
4C1 8.33 10.43 3.44
nc
6.59
5bf
-11.36 ne
ne
ne
a
b
Legend: nc, not calculated; ne, not extracted by Perch. Experimental values are indicated by exp. X-ray indicates that the X-ray
structure was imported into Gaussian 98W for calculation of the J H,Hs (i.e., geometry optimization was not performed, but one SCF cycle
was run). DFT indicates geometry-optimized structures starting either from the X-ray structure (structure origin is given) or from a
structure that was drawn free-hand and preoptimized by MM2 prior to optimization by DFT (man). In the case of 1a , two different 4C1
conformers were obtained (differing mainly in the disposition of the ring substituents rather than the ring conformation itself) depending
on the starting structure. Karplus indicates that the couplings were calculated using this relationship (from ref 5b). c Sign of J was
d
assumed (evenJ ) -ve, oddJ ) +ve). In D2O. The data for 1a were extracted from a spectrum acquired on a freshly prepared sample of
1a that essentially contained only 1a (i.e., R-D-galactopyranose). The data for 1b were extracted from a spectrum acquired on an equilibrated
sample containing â-D-galactopyranose and R-D-galactopyranose (2.02:1, respectively) in addition to minor amounts of furanose and open-
chain forms, but despite considerable spectral overlap, the values are nonetheless reported with confidence. e Sign of J was taken from
g
h
i
the DFT calculations. f In CDCl3. In some samples, J HO6,H6 could occasionally be observed (ca. 6.4 Hz). In acetone-d6. At 25 °C. A
B
value for J H3,H6 (0.36b Hz) was also extracted. At -75 °C. These values are reported as they are the lowest temperature at which the
j
field homogeneiAty was still good (νtms ) 0.5 Hz). Coupling between HO-6 and the two H-6s could be observed at this temperature, but due
to the complexity of the spectral region containing the H-6s (an overlapped mass consisting of two sets of higher order spin pairs), J HO6,H6
,
A
k
J HO6,H6 , and J H6
could not be reliably extracted. At -100 °C. Although the field homogeneity had deteriorated (νtms 1 ) Hz), the
A,H6B
B
values reported are still considered to be reliable. As at -75 °C, the full spin system could not be accurately simulated. In DMSO-d6.m In
l
toluene-d8. Coupling constants for the galactose ring of 3. o Coupling constants for the glucose ring of 3.
n
TABLE 3. 13C Ch em ica l Sh ifts (p p m ) for Com p ou n d s 1-5 a t 25 °C
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
C-16
1a a
1ba
2b
95.11 71.18 72.01 72.14 73.31 64.02
99.29 74.72 75.63 71.58 77.98 63.82
96.29 70.58 70.74 71.47 68.25 62.10 108.67 24.95 26.02
97.21 71.58 71.60 71.81 69.49 61.83e 108.84 25.18 26.32
96.47 70.28 70.33 70.45 68.52g 60.77h 108.33 24.61 25.67
95.61 69.95 69.95 70.10 68.19 59.92 107.60 24.86 25.92
95.70 70.12 70.23 70.56 68.19 61.14 107.48 23.85 25.13
95.92 70.15 70.65 70.88 67.48 69.41 107.98 24.65 25.34
109.42 24.34
109.42 24.59
108.47 23.46
108.01 24.19
108.22 23.32
109.31 24.24
25.93
26.38
25.67
25.83
25.10
25.88
2c,d
2c,f
2i
2j
3,b gal
3,b,k glu 99.38 54.58 70.71 69.06 71.59 62.07 170.10l 20.47l 169.54l 20.65l 170.74l 20.78l
4b,m
5a b
5a c
5bb
96.30 70.90 70.62 70.68 68.34 62.60 108.43 24.99 26.10
109.00 24.39
25.96
89.72 67.37 66.45 67.43 68.77 61.26 169.06 21.03 170.26 20.80 170.02 20.75 170.29 20.68 170.50 20.78
90.08 67.44 68.05 68.41 69.62 61.92 169.62 20.39 170.44 20.66 170.21 20.49 170.65 20.46 170.44 20.49
92.19 67.88 70.86 66.84 71.74 61.06 168.97 20.97 169.38 20.70 169.96 20.78 170.13 20.81 170.35 20.81
a
In D2O referenced to TSP (0 ppm). The data for 1a were extracted from a spectrum acquired on a freshly prepared sample of 1a that
essentially contained only 1a (i.e., R-D-galactopyranose). The data for 1b were extracted from a spectrum acquired on an equilibrated
sample containing â-D-galactopyranose and R-D-galactopyranose (2.02:1, respectively) in addition to minor amounts of furanose and open-
b
d
chain forms. In CDCl3 referenced to TMS (0 ppm). c In acetone-d6 referenced to TMS (0 ppm). At 25 °C. e Major signal (br), minor
signal 61.7 (ex br). f At -100 °C. Major signal, minor signal 68.49 (integration, 5:3). Major signal, minor signal 60.64. In DMSO-d6
referenced to DMSO-d5 (39.5 ppm). In toluene-d8 referenced to toluene-d7 (20.4 ppm). For the phthalimido moiety: 167.63 (v br, C-7′);
133.74 (C-10′); 132.01 (v br, C-8′); 123.48 (C-9′). Carbons C-11′-C-16′, respectively. For the TBDPS group: 135.71 (C-14A); 135.64
(C-14B); 133.68 (C-13B); 133.64 (C-13A); 129.57 (C-16B); 129.56 (C-16A); 127.59 (C-15B); 127.54 (C-15A); 26.83 (C-18); 19.28 (C-17).
g
h
i
j
k
l
m
experimental values. However, if a conformational equi-
librium is in effect, then examination of the system by
variable-temperature NMR will reveal the presence of
such, either by, in the best case scenario, slowing the
interconversion rate sufficiently on the NMR-time scale
to permit explicit observation of the subspectra of each
of the contributing conformers or, in less than ideal cases,
by perturbation of the spectra in some manner. Com-
pound 2 was thus cooled to -100 °C with spectra recorded
in acetone-d6 at intervals of 25 °C (results for spectra at
25, -75, and -100 °C included in Tables 1-3). However,
immediate interpretation was compromised by the re-
22 J . Org. Chem., Vol. 69, No. 1, 2004

Conformation of the Galactose Ring
SCHEME 2
stricted rotation about the C5-C6 bond; this was readily
attributable by the splitting of C-6 already at 25 °C,
followed subsequently by C-5, which decoalesced at -75
°C into the same 5:3 ratio of conformer signals. Of 2-4,
only for 2 and only in acetone-d6 were the dynamic effects
of the C5-C6 restricted bond rotation evident at ambient
temperature. For both C-5 and C-6, a barely perceptible
broadening continued after decoalescence, indicating that
an additional species was present but was still in fast
exchange with the two, now observable conformers. This
additional species is presumably the third rotamer for
rotation about the C5-C6 bond, as has been previously
described.5b This dynamic process affects the entire
system, although the effects are generally most notable
for the nuclei close to the point of action (hence the basis
for the assignment). Thus, all proton signals, with the
exception of H-1, showed varying degrees of modest to
slight broadening, including even the methyl signals.
Although the H-6s are close to the point of action and
therefore highly subject to change, comprehension of their
behavior was compounded by the slowing of the chemical
exchange of HO-6, which not only became observable
with the decrease in temperature but in fact displayed
coupling to the H-6s, rendering their spin simulation
prohibitive. In the 13C NMR, aside from the decoalescence
of C-5 and C-6, most signals displayed some degree of
broadening due to exchange, although none were dra-
matically broadened. So although the dynamics of the
C5-C6 bond can influence the band shapes of practically
all of the observed nuclei in the molecule, this is a
consequence of the differing chemical shifts of each of the
nuclei in the various rotamers. However, it is unlikely
that the dynamics of the C5-C6 bond can influence the
coupling constants of the vicinal ring protons to any large
degree. Therefore, if a conformational equilibrium was
in effect, while the band shape changes may be masked
by the other dynamic process, any changes in the vicinal
coupling constants for the ring protons should be appar-
ent and, moreover, decisive. Extraction of the coupling
constants at both -75 and -100 °C for 2 revealed that
essentially no change whatsoever was evident for the
determinant vicinal couplings. Therefore, on the basis of
this low-temperature experiment, the system is concluded
to be overwhelmingly dominated by one conformer in the
solution state and it remains only to distinguish between
assume one of two orientations, one of which is depicted
in Figure 1 (the ORTEP plots of both 3 and 5a are
presented in Supporting Information, Figures S1 and S2,
respectively). For 3, an OS2 conformation was found for
the galactose moiety in the crystalline state. By com-
4
parison, a C1 conformation was still adopted by 5a (and
also for the glucose moiety in 3), i.e., peracetylation did
not alter the preferred conformation from R-D-galacto-
pyranose (1a ). Of note in 3 is the torsion angle defined
by H2-C2-C3-H3, which is larger than the H4-C4-C5-
H5 torsion angle (72 vs 46°, respectively), yet the mag-
nitude of the proton coupling in the former set is greater
(2.42 vs 1.83 Hz, respectively).
The X-ray structure coordinates of 3 and 5a were
imported directly into Gaussian 98W.27 To limit the
length of the calculations, the glucosyl moiety attached
to O-6 in 3 was substituted by H, thereby generating
structure 2. DFT calculations (perturbations on a par-
ticular nucleus) at the B3LYP/cc-pVTZ level of theory
were performed in order to calculate the couplings
emanating from the selected proton. (H-2 and H-4 in all
1
cases (1a , 2, and 5a ), thereby obtaining all vicinal H-
1H couplings of the sugar ring and additionally H-1 for
1a and 5a to obtain long-range couplings involving H-1.)
The coupling constant calculations are based on the
assumption that the Fermi contact term dominates the
coupling and the other contributions are either insignifi-
cant (spin-dipole term) or cancel each other out (para-
magnetic and diamagnetic spin-orbit terms).16,17 The
results of the calculations are also presented in Table 2.
In this approach, by using the X-ray structures unal-
tered (except the substitution and with one SCF cycle
run on the geometries to set the electron distributions),
good agreement was obtained with respect to the experi-
mental values for both 2 and 5a (including J H2,H3 and
J H4,H5). The X-ray-based structures were also used as
initial structures for geometry optimization (at the
B3LYP/6-31G(d,p) level of theory) to provide a further
comparison; very similar results were obtained, thus
validating the use of the unaltered X-ray structures and
further confirming that the conformation of the galactose
ring in solution is indeed dominated, overwhelming so,
by the OS2 conformation for 2-4. The two geometry-
(27) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.11; Gaussian, Inc.: Pittsburgh,
PA, 2001.
1
the OS2 and S5 conformers.
Single-crystal X-ray analyses were also performed on
3 and 5a enabling a comparison between the conforma-
tions adopted in the crystalline state and in solution.
Compound 3 crystallized in the hexagonal crystal system,
space group P65, while 5a crystallized in the orthorhom-
bic crystal system, space group P212121. The crystal
structures of both compounds showed normal structural
bonding parameters; however, there was some degree of
disorder for 3 as the C6′-acetyl group fragment could
J . Org. Chem, Vol. 69, No. 1, 2004 23

Roslund et al.
F IGURE 1. Structure of the disaccharide 3 as determined by X-ray analysis. The thermal displacement parameters are drawn
at 20% probability.
optimized structures were also used as the starting
geometries for R-D-galactose (1a ) after replacement of the
O-bound substituents with H. The structure for 1a with
an OS2 conformation maintained its general shape but
provided values, not unexpectedly, inconsistent with the
Con clu sion
In conclusion, the NMR parameters of 1-5, despite the
long prevalence of some of the compounds in the litera-
ture, have not always been properly reported, mainly due
to difficulties associated with overlap and higher order.
The conformation of the diisopropylidene derivatives of
galactose (2-4), despite having been examined previ-
ously, nevertheless represented a good test case for the
recent developments16 that have been made in the
calculation of J H,H using DFT methodology and which is
now proving to be a useful structural tool.17 The calcula-
tion of only coupling constants and not chemical shifts,
despite the economical superiority of the latter, for the
conformational determination was based on previous
work17a where the calculation of δHs was shown to be
inadequate and inferior to the use of J H,Hs for this type
of analysis. A different approach was also tried by using
the structures of two of these compounds, 3 and 5a ,
resulting from X-ray analysis directly for the calculation
of the vicinal coupling constants; it was found that they
not only were appropriate as is but also provided good
starting structures for further refinement. Thus the work,
based on a combination of X-ray analysis, molecular
modeling, and NMR, has confirmed the overwhelming
conformational bias of 2-4 toward an OS2 conformation
both in solution and in the crystalline state. The dem-
onstrated28 efficacy of DFT calculations in evaluating
long-range couplings in saccharides is valuable for dis-
tinguishing between hydrogen bond-mediated pathways
4
observed values in contrast to the C1 conformation (e.g.,
5a ), which provided very consistent values. Two other
structures were also input for geometry optimization.
1
One was a manually generated S5 conformer for 2 (on
the basis that it should also provide couplings similar to
those observed) but which only resulted in the same,
indistinguishable OS2 conformer being obtained as that
resulting from the optimization of the X-ray-based struc-
ture. This result infers that the 1S5 conformation does
not represent a local energy minimum on the PES and
is thus not expected to participate as a discrete entity to
any conformational equilibrium. The other input struc-
4
ture was a manually generated C1 conformation for 1a .
4
Differences were apparent between this C1 conformation
4
and the C1 conformation provided by DFT optimization
of the X-ray-based structure, but the differences reside
mainly in the disposition of the ring substituents rather
than any differences in the ring conformation and reflect
the effects that the attendant substituent orientations
can have on the coupling constants. Of note, in the cases
where long-range coupling constants were able to be
extracted, the DFT calculations provided remarkably
accurate values. These coupling constants in themselves
speak for a conformationally biased system, particularly
when the DFT calculations have been able to reproduce
them so faithfully.
(28) Larsson, E. A.; Ulicny, J .; Laaksonen, A.; Widmalm, G. Org.
Lett. 2002, 4, 1831-1834.
24 J . Org. Chem., Vol. 69, No. 1, 2004

Conformation of the Galactose Ring
and otherwise normal w-type coupling pathways; dif-
ferentiation is important, as erroneous structural inter-
pretations can be based on these observations.28 Evalu-
ation even in small molecules such as those studied here
can nevertheless serve well to model larger systems.
What was further pertinent was the differences calcu-
lated for the two optimized 4C1 conformations of 1a ,
demonstrating that there can be realizable differentiation
due to substituent orientation.
fully acknowledged. The authors also thank the Center
for Scientific Computing (CSC, Finland) for a generous
allocation of computation time.
Su p p or tin g In for m a tion Ava ila ble: Crystallographic
data have been deposited with the Cambridge Crystallographic
Data Centre as supplementary publication nos. CCDC-218904
(3) and CCDC-218905 (5a ), copies of which can be obtained
upon request from the Director, CCDC, 12 Union Road,
Cambridge CB12 1E2, U.K. Computational methodology (es-
sentially the same as that used previously),17 general experi-
mental details, preparative procedures for compounds 2-5a ,
and X-ray crystallographic data (including CIF files, structure
tables, and XYZ files) for 3 and 5a . This material is available
free of charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. The authors wish to thank Mr.
Markku Reunanen for the EIMS measurements. The
assistance of Prof. Alessandro Bagno (Universita` di
Padova), whose advice and guidance with the compu-
tational work has been most influential, is also grate-
J O035400U
J . Org. Chem, Vol. 69, No. 1, 2004 25