On the Role of Neighboring Group Participation and Ortho Esters in β-Xylosylation: 13C NMR Observation of a Bridging 2-Phenyl-1,3-dioxalenium Ion

Article abstract of DOI:10.1021/jo990424f

The role of ortho esters in the formation of 2,3,4-tri-O-benzoyl-β-xylopyranosides from various donor/ promoter pairs has been investigated. It is concluded that for the activation of sulfoxides with Tf₂O, thioglycosides with PhSOTf, and bromides with AgOTf the anomeric configuration of the donor is of no consequence on the outcome of the reaction. In all methods studied, the presence or absence of a non-nucleophilic, hindered base is of crucial importance with ortho esters only being discernible in its presence. S-Phenyl 2,3,4-tri-O-benzoyl-1-deoxy-1-thia-β-D-xylopyranoside was synthesized enriched with ¹³C at each of the three carbonyl carbons. Activation of this thioglycoside with PhSOTf in CD₂Cl₂ at -78 °C with or without the base permits, for the first time, the observation by ¹³C NMR spectroscopy of a bridging dioxalenium ion as an intermediate in a neighboring group directed glycosylation. Quenching of this cation in the presence of the base leads to the ortho ester, whereas in the absence of the base the glycosides are the only products detected.

Full text of DOI:10.1021/jo990424f

5224
J . Org. Chem. 1999, 64, 5224-5229
On th e Role of Neigh bor in g Gr ou p P a r ticip a tion a n d Or th o Ester s
in â-Xylosyla tion : 13C NMR Obser va tion of a Br id gin g
2-P h en yl-1,3-d ioxa len iu m Ion
David Crich,* Zongmin Dai, and Ste´phane Gastaldi
Department of Chemistry, University of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois 60607-7061
Received March 9, 1999
The role of ortho esters in the formation of 2,3,4-tri-O-benzoyl-â-xylopyranosides from various donor/
promoter pairs has been investigated. It is concluded that for the activation of sulfoxides with
Tf2O, thioglycosides with PhSOTf, and bromides with AgOTf the anomeric configuration of the
donor is of no consequence on the outcome of the reaction. In all methods studied, the presence or
absence of a non-nucleophilic, hindered base is of crucial importance with ortho esters only being
discernible in its presence. S-Phenyl 2,3,4-tri-O-benzoyl-1-deoxy-1-thia-â-D-xylopyranoside was
synthesized enriched with 13C at each of the three carbonyl carbons. Activation of this thioglycoside
with PhSOTf in CD2Cl2 at -78 °C with or without the base permits, for the first time, the observation
by 13C NMR spectroscopy of a bridging dioxalenium ion as an intermediate in a neighboring group
directed glycosylation. Quenching of this cation in the presence of the base leads to the ortho ester,
whereas in the absence of the base the glycosides are the only products detected.
Sch em e 1
In tr od u ction
The concept of neighboring group participation and the
formation and isolation of ortho esters are intimately
linked and central to the field of carbohydrate chemistry.1
The synthesis of the 1,2-trans-type glycosidic bond, as
typified by the â-glucopyranosides, relies heavily on the
use of ester protecting groups for O-2 such that participa-
tion ensures the formation of the desired stereoisomer
(Scheme 1). Ortho esters are frequent byproducts in this
type of chemistry owing to the trapping of the intermedi-
ate bridging cation by the nucleophile as opposed to
attack at the anomeric cation. On the other hand, the
bridging cation may be entered from the ortho ester
manifold under certain conditions, which give rise to the
well-known ortho ester glycosylation method.2-6 The
essential question, therefore, in neighboring group-as-
sisted trans-glycosylation is which conditions and esters
favor attack at the anomeric position and minimize ortho
ester formation. Direct trapping of the bridging cation,
with ortho ester formation, brings about a change in
hybridization and an increase in steric hindrance. For
this reason, bulky pivalate esters have been a popular
choice for minimizing ortho ester formation but they by
no means eliminate it altogether,7-12 and under some
conditions the tert-butyl ortho ester may even predomi-
nate.13 Very recently, Whitfield has published calcula-
tions that address the mechanism of neighboring group
participation and that support the notion of a bridging
cation as intermediate.14 Furthermore, these calculations
suggest, at least for the 2,6-di-O-acyl-3,4-O-isopropyl-
idene-D-galactopyranosyl system studied, that the kinetic
mode of attack by methanol is at the bridging cation.14
Our interest in this area was stimulated by the observa-
tion15 that activation of either thioglycoside 2, or its
sulfoxide (3), with benzenesulfenyl triflate, or triflic
(1) Bochkov, A. F.; Zaikov, G. E. Chemistry of the O-Glycosidic Bond;
Pergamon: Oxford, 1979.
(2) Kochetkov, N. K.; Khorlin, A. J .; Bochkov, A. F. Tetrahedron
1967, 23, 693-707.
(3) Kochetkov, N. K. Izv. Akad. Nauk. SSSR, Ser. Khim. 1984, 243-
255.
(4) Kochetkov, N. K. Pure Appl. Chem. 1984, 56, 923-938.
(5) Wang, W.; Kong, F. J . Org. Chem. 1998, 63, 5744-5745.
(6) Roberts, C.; Madsen, R.; Fraser-Reid, B. J . Am. Chem. Soc. 1995,
117, 1546-1553.
(7) Kunz, H.; Harreus, A. Liebigs 1982, 41-48.
(8) Harreus, A.; Kunz, H. Liebigs 1986, 717-730.
(9) Walker, S. Ph.D. Thesis, Princeton, 1992.
(10) Laurich, V. M. Senior Thesis, Princeton, 1997.
(11) Yan, L. Ph.D. Thesis, Princeton, 1996.
(12) Thompson, C.; Ge, M.; Kahne, D. J . Am. Chem. Soc. 1999, 121,
1237-1244.
(13) Lemieux, R. U.; Morgan, A. R. Can. J . Chem. 1965, 43, 2199-
2204.
(14) Nukada, T.; Berces, A.; Zgierski, M. Z.; Whitfield, D. M. J . Am.
Chem. Soc. 1998, 120, 13291-13295.
(15) Crich, D.; Dai, Z. Tetrahedron 1999, 55, 1569-1580.
10.1021/jo990424f CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/16/1999

Neighboring Group Participation in â-Xylosylation
J . Org. Chem., Vol. 64, No. 14, 1999 5225
anhydride, respectively, at -78 °C in the presence of 2,6-
di-tert-butyl-4-methylpyridine (DTBMP), followed by ad-
dition of cyclohexanol provided the ortho ester 1 in very
high yield. This was in contrast to the use of the bromide
4 with activation by silver triflate, which provided the
â-xyloside 5 smoothly. Here, we report the results of our
investigation into the causes of this dichotomy as well
as our direct observation, using 13C-labeled analogues of
2-4, of an intermediate bridging cation in CD2Cl2 solu-
tion.
stereochemistry. Thus, Lemieux reported that two ano-
meric, ester-protected glucopyranosyl bromides were
converted to ortho esters at different rates with the
â-anomer being the more reactive;13 in the presence of
added tetraalkylammonium bromide, ortho ester forma-
tion from the R-bromide is facilitated due to in situ
formation of the more reactive â-bromide. Paulsen has
similarly demonstrated that pentaacetyl-â-D-glucopyra-
nose is readily converted to a crystalline bridging dioxo-
lenium salt on treatment with SbCl5, whereas the
R-anomer leads to no such species.23 However, Paulsen
also noted that both anomers of acetochloroglucose
proceeded rapidly to the same crystalline salt on exposure
to SbCl5,23 which demonstrates that the dependence of
such reactions on anomeric configuration is a function
of the leaving group and method of activation. Neverthe-
less, the possibility that the dichotomy might be a
function of the differing anomeric configurations of 3 and
4 had to be entertained. To test this hypothesis, we
required the â-xylopyranosyl bromide 6, the R-xylopyra-
nosyl thioglycoside 7, and its sulfoxide 8. Bromide 6 was
prepared uneventfully, according to the literature, by
reaction of â-xylopyranose tetrabenzoate with TiBr4.24
Exposure of 6 to thiophenol and triethylamine then
cleanly afforded the R-thioglycoside 7. Oxidation of 7 with
either Oxone or m-CPBA provided a single diastereomer
of the required sulfoxide 8, which we assign the SR
Resu lts a n d Discu ssion
In connection with the synthesis of a trisaccharide, we
needed to form a â-xylopyranosyl linkage to a secondary
alcohol. Following the literature precedent with a closely
related target,16 we achieved this successfully through
use of the bromide 4 and activation with silver triflate
in dichloromethane at -78 °C.15,17 Our interest18-20 in
Kahne’s sulfoxide method,21,22 however, prompted us also
to investigate its use in â-xylosylation. Accordingly,
displacement of bromide from 4 with thiophenate pro-
vided the â-thioglycoside 2, which was oxidized with
Oxone to provide a separable mixture of diastereoiso-
meric sulfoxides 3. Activation of either diastereomer of
3 in the presence of DTBMP in CH2Cl2 with triflic
anhydride, followed by addition of cyclohexanol, as a
model secondary acceptor, provided a crude reaction
mixture, which, on inspection by 1H and 13C NMR
spectroscopy, was found to be almost exclusively the ortho
ester 1.15 Conversely, activation of bromide 4 in dichlo-
romethane at -40 °C with silver triflate in the presence
of the alcohol smoothly provided the â-xyloside 5.15 At
the time, ortho esters had not been widely reported in
the sulfoxide method, but their formation as byproducts
had been noted by Kahne and co-workers9-11 as well as
by ourselves.19 The very clean formation of 1 from 3 was
therefore somewhat unexpected and unusual and promp-
ted us to take a closer look at the underlying chemistry.
It could be argued that the likelihood of observing ortho
esters is maximized by working in the more conforma-
tionally labile pentopyranose series with benzoate esters,
but that does not explain the dichotomy of the two donors
3 and 4 giving such markedly different product distribu-
tions.
configuration on the basis of literature precedent.25
A
byproduct having a very similar Rf to 7 in these oxida-
tions was isolated and shown to be the corresponding
sulfone 9, which adopts a twist-boat conformation in
order to minimize 1,3-diaxial interactions. Reaction of
thioglycoside 7 with PhSOTf in CH2Cl2 at -78 °C in the
presence of DTBMP followed by treatment with cyclo-
hexanol led to a crude reaction mixture consisting of a
90/10 mixture of the ortho ester 1 and the â-xyloside 5.
A similar result was obtained with sulfoxide 8 on activa-
tion with Tf2O in the presence of DTBMP, followed by
addition of cyclohexanol. On the other hand, the â-bro-
mide 6 afforded the â-xyloside 5 on treatment with AgOTf
and cyclohexanol in CH2Cl2. Clearly, in these particular
reactions the anomeric configuration of the donor has
little or no effect on outcome of the reaction, with both
bromides leading to the â-xyloside and both sulfoxides
and both thioglycosides providing mainly the ortho ester.
Exactly analogous results were obtained when methanol
was used as glycosyl acceptor.
Attention was next focused on the influence of the base
(DTBMP), which was present in all reactions of the
sulfoxides 3 and 8 and of the thioglycosides 2 and 7, but
not of the bromides 4 and 6. Coupling of the â-bromide 6
with methanol in CH2Cl2 at -40 °C in the presence of
DTBMP, with activation by AgOTf, provided a 36/64
mixture of the ortho ester26 10 and the â-xyloside 11,15
as judged by NMR spectroscopy of the crude reaction
mixture. On the other hand, activation of the thioglyco-
sides or sulfoxides with PhSOTf or Tf2O, respectively, in
the absence of base followed by reaction with methanol
led to relatively complex reaction mixtures containing the
In a given series, the rate of ortho ester formation has
been found in some instances to be a function of anomeric
(16) Lichtenthaler, F.; Schneider-Adams, T.; Immel, S. J . Org. Chem.
1994, 59, 6735-6738.
(17) Crich, D.; Dai, Z. Tetrahedron Lett. 1998, 53, 1681-1684.
(18) Crich, D.; Sun, S. Tetrahedron 1998, 54, 8321-8348.
(19) Crich, D.; Sun, S. J . Am. Chem. Soc. 1997, 119, 11217-11223.
(20) Crich, D.; Sun, S. J . Org. Chem. 1997, 62, 1198-1199.
(21) Kahne, D.; Walker, S.; Cheng, Y.; Engen, D. V. J . Am. Chem.
Soc. 1989, 111, 6881-6882.
(23) Paulsen, H.; Herold, C.-P. Chem. Ber. 1970, 103, 2450-2462.
(24) Luger, P.; Durette, P. L.; Paulsen, H. Chem. Ber. 1974, 107,
2615-2625.
(25) Crich, D.; Mataka, J .; Sun, S.; Lam, K.-C.; Rheingold, A. R.;
Wink, D. J . J . Chem. Soc., Chem. Commun. 1998, 2763-2764.
(26) Bochkov, A. F.; Obruchnikov, I. V.; Kochetkov, N. K. Zh. Obsh.
Khim. 1972, 42, 2758-2766.
(22) Yan, L.; Kahne, D. J . Am. Chem. Soc. 1996, 118, 9239-9248.

5226 J . Org. Chem., Vol. 64, No. 14, 1999
Crich et al.
â- and R-xylosides 11 and 12.27 It is therefore evident that
the decisive factor in the chemoselectivity of all of the
above reactions is the hindered base. It buffers the
reaction conditions and permits the isolation of ortho
esters; otherwise, the â-xylosides are obtained. The same
conclusion was reached previously by the Garegg and
Bundle groups, both working in the glucopyranose se-
ries.28,29 More recently, Kahne and co-workers, conscious
of the requirement for the hindered base in their sulfox-
ide method but needing to avoid the formation of ortho
ester byproducts, introduced the combination of DTBMP
and BF3.12 In this system, DTBMP serves to buffer any
Bronsted acidity and to activate phenolic nucleophiles but
is too hindered to complex the Lewis acid, which,
therefore, provides the necessary acidity to promote the
rearrangement of any ortho esters to the glycosides.12
It has been suggested several times that ortho esters
are the kinetic products in silver- and mercury-promoted
glycosylation reactions, using ester-protected glycosyl
halides as donors and that, in the absence of base, these
then rearrange in situ to the 1,2-trans-glycosides.28-30 In
the case of mercuric cyanide promoted reactions of 2-O-
acetyl-3,4,6-tri-O-methyl-R-D-glucopyranosyl bromide with
cyclohexanol, it has been conclusively demonstrated that
the kinetic product is the ortho ester and that, in the
absence of buffer, it subsequently rearranges to the
â-glucoside.31 The obvious, and oft-suggested, corollary
to this mechanistic hypothesis is that a bridging dioxo-
lenium ion is an intermediate on the pathway.28-31 The
recent calculations of Whitfield support this notion.14
Paulsen has demonstrated that such bridging cations
thioglycosides 2 and 7, on activation by preformed
PhSOTf, and with bromide 4 and AgOTf gave comparable
spectra but with differing relative intensities of the
various multiplets, suggesting that the position of any
equilibrium is significantly affected by the reaction
conditions. When sulfoxide 8 was activated at -78 °C
with Tf2O, in the presence of DTBMP, and the probe
allowed to warm gradually in 10 °C steps, monitoring by
1H NMR indicated the appearance of a further set of
signals that grew in beginning gradually around -38 °C.
Between -18 and -8 °C, complete decomposition of the
original species occurred in favor of the new set of signals
which then completely dominated the spectra, indicating
that all of the species formed at low temperature were
eventually channeled into the one product. This sub-
stance was subsequently isolated and shown to be the
elimination product 15. In a preparative-scale experi-
ment, 15 was isolated in 64% yield following activation
of â-thioglycoside 2 at -78 °C with the PhSOTf/DTBMP
combination and warming to room temperature.
-
may be precipitated from solution as their SbCl6 salts
on treatment of the appropriate glycosyl chloride with
SbCl5, as, for example, in the isolation of 13 from a CCl4
solution of 14 in 85% yield.23 This type of experiment,
however, cannot reflect on the relative proportions of the
various species in solution as any equilibrium is neces-
sarily driven by the precipitation of the insoluble salt.
Thus, we were prompted to probe the nature of any
intermediates in the above chemistry by means of low-
Acquisition of the 13C NMR spectra in the above low-
temperature NMR experiments was largely futile as the
signal-to-noise ratio was so poor as to obliterate any
resonances due to the ring carbons. The only observable
signals in these 13C NMR traces were those of the base
and the benzoate rings. The poor signal-to-noise ratios
again suggested dynamic equilibria, with considerable
broadening on the 13C NMR time scale. To address this
problem, the triply 13C-labeled thioxyloside 16 was
synthesized by Zemplen deprotection of 2 followed by
esterification with benzoyl chloride mono-13C-labeled in
the carbonyl carbon. Activation of 16 with PhSOTf/
DTBMP in CD2Cl2 at -78 °C in the probe of the NMR
1
temperature H and 13C NMR spectroscopy.
A series of low-temperature NMR experiments were
conducted in which the R-sulfoxide (8) was dissolved in
CD2Cl2 in the presence of DTBMP and cooled to -78 °C
before treatment with precooled Tf2O. The NMR tube was
then immediately inserted in the precooled probe (-78
1
spectrometer provided an H NMR spectrum analogous
1
°C) of the spectrometer, and the H and 13C spectra were
to that obtained in the unlabeled series. On the other
hand, the 13C NMR spectrum (Figure 1a) now revealed
a cluster of signals between δ 163 and 166 and several
smaller signals between δ 127 and 135. The only upfield
signals corresponded to those of the tert-butyl and methyl
groups of the pyridine base and its salts. In addition to
the above groups of signals, all attributable to the base,
its salts, and the benzoate groups, an intense new signal
at δ 180.3 was observed. We assign this resonance to the
cationic carbon in the bridging cation 17, the chemical
shift being fully consonant with the value of δ 182.9 found
for C-2 in the authentic related dioxolenium ion 18 and
with those of other related R,R-dialkoxybenzyl cations.32
The spectrum also contains a smaller but significant
resonance at δ 120.9 that may represent the ortho ester
carbon in covalent bridging triflate 19 or, more likely,
the hemiortho ester 20 arising from the introduction of
recorded. The 1H NMR spectrum obtained in this manner
was far more complex than had been anticipated and
indicated, aside from the complete consumption of the
substrate, the formation of several new xylose-based
1
species. The complexity of the H NMR spectrum with
its multiple and broad signals suggests the possibility of
a dynamic equilibrium involving several species. A
comparable result was obtained with the anomeric â-sul-
foxide (3). Similar NMR experiments conducted with the
(27) We suspect that the R/â mixture of xylosides obtained in this
experiment is a function of scrambling of anomeric stereochemistry
under the strongly acidic conditions rather than of any lack of
selectivity in the initial glycoside-forming reaction.
(28) Banoub, J .; Bundle, D. R. Can. J . Chem. 1979, 57, 2091-2097.
(29) Garegg, P. J .; Konradsson, P.; Kvarnstrom, I.; Norberg, T.;
Svensson, S. C. T.; Wigilius, B. Acta Chem. Scand. B 1985, 39, 569-
577.
(30) Lemieux, R. U.; Hendriks, K. B.; Stick, R. V.; J ames, K. J . Am.
Chem. Soc. 1975, 97, 4056-4062.
(31) Wallace, J . E.; Schroeder, L. R. J . Chem. Soc., Perkin Trans. 2
1977, 795-802.
(32) Childs, R. F.; Frampton, C. S.; Kang, G. J .; Wark, T. A. J . Am.
Chem. Soc. 1994, 116, 8499-8505.

Neighboring Group Participation in â-Xylosylation
J . Org. Chem., Vol. 64, No. 14, 1999 5227
F igu r e 1. (a) Partial 13C NMR spectrum obtained in CD2Cl2 on activation of 16 with PhSOTf and DTBMP at -78 °C. (b) Partial
13C NMR spectrum obtained in CD2Cl2 on activation of 16 with PhSOTf and DTBMP at -78 °C followed by addition of methanol.
(c) Partial 13C NMR spectrum obtained in CD2Cl2 on activation of 16 with PhSOTf at -78 °C.
adventitious water. Indeed, inspection of the carbonyl
region suggests that at least two different sugars are
present in agreement with the multiplicity of peaks seen
step but is decisive in the subsequent reaction with an
alcohol. In the presence of base, the acid-sensitive ortho
ester is the major product, whereas in its absence the
glycosides are formed directly. Unfortunately, owing to
the rapidity of the reactions, even at -78 °C, we are
unable to say whether the glycosides are formed directly
from the bridging cation 17 or result from a very rapid
acid-catalyzed rearrangement of an initial ortho ester.
These conclusions are in good agreement with the recent
computations of Whitfield14 and the earlier inferences of
several groups derived from isolation experiments.28-31
1
in the H NMR experiments. Addition of cold methanol
to this -78 °C solution results in the immediate disap-
pearance of the signal at δ 180.3 in favor of a strong
signal at δ 121.1 (Figure 1b), which we assign to the ortho
ester carbon of 21.
The position of any equilibrium involving a bridging
cation (Scheme 1), whatever the point of entry, will
depend on the reaction milieu and, hence, on the method
of activation, on the nature of the substituents around
the pyranose ring, and especially on the group R directly
bonded to the dioxolenium ion. Evidently, in working
with the conformationally labile xylopyranose series,
which is readily able to accommodate the fused ring
system, and with the benzoate esters, leading to benzylic
stabilization, we chose conditions most likely to favor the
bridging cation. That does not mean that in general the
use of a benzoate ester is most likely to lead to isolation
of an ortho ester, as benzoate-derived ortho esters are
the ones most likely to heterolyze to the bridging cation
and rearrange to the glycoside if the conditions are
sufficiently ionizing. Indeed, Garegg and co-workers
previously noted that metal-promoted glycosidations with
benzoate-protected glycosyl bromides gave higher yields
of glycoside than the corresponding acetates.29 More
recently, Danishefsky and co-workers have reported that
ortho ester formation was less of a problem when
benzoates were employed than acetates in methyl triflate
promoted couplings with thioglycoside donors.33
Figure 1c is a 13C NMR spectrum taken immediately
after activation of 16 with PhSOTf in CD2Cl2 at -78 °C
in the absence of DTBMP; self-evidently the bridging
cation 17 is again formed in high yield. However, on
addition of methanol at -78 °C, the ortho ester is not
seen, indicating that in the absence of base the cation
either goes directly to the glycosides or the ortho ester
rearranges extremely rapidly under the acidic reaction
conditions. These experiments, with observation of the
bridging cation 16, provide for the first time a direct,
experimental insight into neighboring group participa-
tion, in glycosylation reactions. Both in the presence and
absence of base the 2-O-benzoate group captures any
activated anomeric species leading to a dynamic equilib-
rium of which a major component is a bridging cation.
Importantly, the base plays no significant role in this first
(33) Seeberger, P. H.; Eckhardt, M.; Gutteridge, C. E.; Danishefsky,
S. J . J . Am. Chem. Soc. 1997, 119, 10064-10072.

5228 J . Org. Chem., Vol. 64, No. 14, 1999
Crich et al.
Cou p lin g of â-Xylosyl Br om id e 6 w ith Meth a n ol in th e
P r esen ce of DTBMP . Methanol (7 µL, 0.17 mmol) and
DTBMP (52 mg, 0.26 mmol) were dissolved in dichloromethane
(2 mL) in the presence of 4Α molecular sieves (20 mg) and
cooled with stirring under N2 to -40 °C. Bromide 6 (45 mg,
0.085 mmol) in dichloromethane (1 mL) was then added,
followed by AgOTf (44 mg, 0.17 mmol) in toluene (1 mL), and
stirring was continued for 20 min before the reaction mixture
was allowed to warm to 0 °C and worked up by addition of
saturated aqueous NaHCO3. Ether extraction, washing with
brine, drying, and evaporation (MgSO4) provided a crude
reaction mixture that was shown to consist of a 36/64 mixture
of the known ortho ester26 10 and the â-xyloside 11, which
was identical with an authentic sample.15 Attempted isolation
of this ortho ester by silica gel chromatography resulted in
extensive decomposition, mainly to 11, such that a pure sample
could not be obtained.
Cou p lin g of â-Su lfoxid e 3 w ith Meth a n ol in th e Ab-
sen ce of DTBMP . The â-sulfoxide 3 (40 mg, 0.07 mmol) was
dissolved in dichloromethane (1 mL) and cooled with stirring
under N2 to -78 °C. Tf2O (14 µL, 0.084 mmol) was then added
followed, after stirring for 5 min, by methanol (6.9 µL, 0.17
mmol). The reaction mixture was then allowed to come to room
temperature with stirring before it was diluted with saturated
aqueous NaHCO3 and extracted with ether. The extracts were
dried (Mg2SO4) and concentrated to give a crude reaction
mixture that consisted of a 41/59 mixture of the â- and
R-xylosides 1115 and 12,15 respectively.
Exp er im en ta l Section
Gen er a l Meth od s. Solvents were dried and distilled by
standard methods. All reactions were conducted under an
atmosphere of dry nitrogen or argon. 1H and 13C NMR
experiments were conducted at 300 and 75 MHz, respectively,
with a Bruker AC 300 spectrometer equipped with a QNP
probe and a VT attachment. Chemical shifts are in ppm
downfield from tetramethylsilane. Microanalyses were con-
ducted by Midwest Microlabs, Indianapolis, IN. Compounds
2,15 3,15 4,34 and 624 were prepared according to the literature
methods.
S-P h en yl 2,3,4-Tr i-O-ben zoyl-1-d eoxy-th ia -r-D-xylop y-
r a n osid e (7). A solution of bromide 6 (917 mg, 1.7 mmol) and
PhSH (1.93 g, 17 mmol) in CH2Cl2/Et3N (28 mL/15 mL) was
stirred for 1 night at room temperature, after which time the
reaction mixture was diluted with 3 M NaOH and extracted
with Et2O. The combined extracts were washed with NH4Cl
and then brine, dried (MgSO4), and evaporated. The residue
was purified via flash chromatography (eluent: 0/100 to 10/
90 AcOEt/hexane), providing 820 mg (87%) of the R-thiogly-
1
coside. [R]20 ) +68.8° (c ) 0.9, CHCl3). H NMR (CDCl3) δ:
D
4.20 (dd, J ) 11.5, 5.4 Hz, 1H); 4.42 (dd, J ) 11.5, 9.6 Hz,
1H); 5.43 (m, 1H); 5.55 (dd, J ) 9.4, 5.2 Hz, 1H); 6.03 (d, J )
5.0 Hz, 1H); 6.07 (t, J ) 9.1 Hz, 1H); 7.21-7.60 (m, 14H); 7.91-
8.10 (m, 6H). 13C NMR δ: 60.5, 69.6, 69.9, 71.4, 86.0, 127.7,
128.4, 128.4, 128.5, 129.1, 129.7, 129.8, 130.0, 131.9, 133.3,
133.4, 133.5. Anal. Calcd for C32H26O7S: C, 69.30, H, 4.73.
Found: C, 69.14, H, 4.74.
S-P h en yl 1-Deoxy-1-th ia -â-D-xylop yr a n osid e. â-Thiogly-
coside 2 (278 mg, 0.5 mmol) was dissolved in a NaOMe solution
prepared by adding Na (1 mg, 0.05 mmol) to dry MeOH (10
mL). After being stirred for 24 h, the reaction was quenched
with HCl (1 M in Et2O) and concentrated to dryness. The
residue was crystallized from Et2O providing 100 mg (83%) of
the title compound,35 which was used without further purifica-
S-P h en yl 2,3,4-Tr i-O-b en zoyl-1-d eoxy-1-t h io-r-D-xyl-
op yr a n osid e S-Oxid e (8). Oxidation of the R-thioglycoside
(7) with MCPBA at -78 °C in CH2Cl2 gave the title sulfoxide
72% yield as a single anomer. [R]20 ) +9.3° (c ) 3.3, CHCl3).
D
1H NMR (CDCl3, 300 MHz) δ: 3.96 (1H, dd, J ) 13.2, 2.3 Hz,
H-5), 4.33 (1H, br d, J ) 13.2 Hz, H-5), 4.67 (1H, d, J ) 2.2
Hz, H-1), 5.16 (1H, br d, J ) 2.5 Hz, H-4), 5.68 (1H, t, J ) 2.8
Hz, H-2), 5.97 (1H, t, J ) 2.9 Hz, H-3), 7.10 (2H, t, J ) 7.0
Hz), 7.35-7.70 (10H, m), 7.85 (4H, m), 8.05 (2H, d, J ) 7.5
Hz), 8.25 (2H, d, J ) 7.5 Hz). 13C NMR (CDCl3, 75 MHz) δ:
66.3, 66.7, 67.0, 67.5, 93.3, 125.4, 128.4, 128.7, 128.8, 129.3,
129.4, 130.0, 130.1, 130.6, 131.8, 133.6, 133.8, 134.0, 141.8,
164.3, 165.3. Anal. Calcd for C32H26O8S‚1/2H2O: C, 66.30; H,
4.70. Found: C, 66.41; H, 4.70.
1
tion. H NMR (methanol-d4) δ: 3.22 (m, 2H); 3.36 (t, J ) 9.2
Hz, 1H); 3.48 (m, 1H); 3.95 (dd, J ) 11.2, 5.0 Hz, 1H); 4.58 (d,
J ) 9.2 Hz, 1H); 7.30 (m, 3H); 7.51 (m, 2H). 13C NMR δ: 70.4,
71.0, 73.8, 79.2, 90.2, 128.7, 130.0, 133.2, 135.0.
S-P h en yl 2,3,4-Tr i-O-b en zoyl-1-d eoxy-1-t h ia -â-D-xyl-
op yr a n osid e-13C3 (16). To a cooled (0 °C) solution of S-phenyl
1-deoxy-1-thia-â-D-xylopyranoside (140 mg, 0.36 mmol) and
DMAP (43 mg, 0.43 mmol) in pyridine (2 mL) was added a
solution of 13C-benzoyl chloride36 (monolabeled in the carbonyl
C) in CH2Cl2 (1 mL). After being stirred overnight at room
temperature, the reaction was diluted with water and ex-
tracted three times with Et2O. The organic layer was washed
with saturated aqueous CuSO4 and brine, dried (MgSO4), and
evaporated. The residue was purified via flash chromatography
(eluent: 0/100 to 20/80 AcOEt/hexane) providing 60 mg (37%)
of 16. 1H NMR (CDCl3) δ: 3.82 (dd, J ) 12.3, 6.5 Hz, 1H); 4.51
Cou p lin g of r-Th ioxylosid e 7 w ith Cycloh exa n ol in th e
P r esen ce of DTBMP . AgOTf (51 mg, 0.2 mmol) was sus-
pended in dichloromethane (3 mL) under N2 at -78 °C and
treated with PhSCl (19 µL, 0.17 mmol). After the mixture was
stirred vigorously for 5 min, a solution of 7 (37 mg, 0.067 mmol)
and DTBMP (41 mg, 0.2 mmol) in dichloromethane (1.5 mL)
was added followed, after 5 min, by cyclohexanol (15 µL, 0.14
mmol). The reaction mixture was allowed to warm with
stirring to -30 °C before it was diluted with saturated aqueous
NaHCO3 and then allowed to come to room temperature.
Extraction with Et2O, washing with brine, drying (MgSO4),
and concentration gave a crude reaction mixture that, on
inspection by 1NMR spectroscopy and with the aid of authentic
samples,15 was shown to be a 90:10 mixture of ortho ester 1
and â-xyloside 5.
Cou p lin g of â-Xylosyl Br om id e 6 w ith Cycloh exa n ol
in th e Absen ce of DTBMP . Cyclohexanol (17 µL, 0.16 mmol)
was dissolved in dichloromethane (2 mL) and stirred with 4Α
molecular sieves (20 mg) for 10 min and then cooled to -40
°C followed sequentially by addition of the â-bromide 6 (42
mg, 0.08 mmol) in dichloromethane (2 mL) and then AgOTf
(41 mg, 0.16 mmol) in toluene (1 mL). After being stirred for
0.5 h, the reaction mixture was allowed to come to 0 °C before
it was washed with aqueous Na2SO3 and then brine, dried
(MgSO4), and concentrated to give a crude reaction mixture
consisting mainly of the â-xyloside 5. Chromatography on silica
gel (eluent: AcOEt/hexane 9/1) enabled the isolation of 5 (30
mg, 69%) identical to an authentic sample.15
(dd, J ) 12.3, 4.0 Hz, 1H); 5.30 (m, 1H); 5.30 (d superimposed,
3
J ) 6.1 Hz, 1H); 5.49 (td, J ) 5.0 Hz and J
) 3.8 Hz, 1H);
C-H
5.80 (td, J ) 6.6 Hz and 3J
) 3.7 Hz, 1H); 7.30-7.45 (m,
C-H
9H); 7.46-7.60 (m, 5H); 7.96-8.10 (m, 6H). 13C NMR (CDCl3)
δ: 165.3, 165.4, 165.7.
1,5-An h yd r o-2,3,4-t r i-O-b en zoyl-D-th r eo-p en t -1-en it ol
(15). To a stirred solution of AgOTf (69 mg, 0.27 mmol) in CH2-
Cl2 (2 mL) at -78 °C was added PhSCl (32 mg, 0.22 mmol).
After the mixture was stirred for 5 min at -78 °C, a solution
of 2 (49 mg, 0.088 mmol) and DTBMP (55 mg, 0.27 mmol) in
CH2Cl2 (1 mL) was added, and the reaction mixture was
allowed to warm to room temperature before quenching with
NaHCO3 and extracting with Et2O. The combined extracts
were dried (MgSO4) and evaporated. The residue was purified
via flash chromatography (eluent: 0/100 to 10/90 AcOEt/
hexane) providing 25 mg (64%) of the title compound.26 1H
NMR (CDCl3) δ: 4.21 (bd, J ) 11.4 Hz, 1H); 4.51 (dt, J ) 12.4,
2.2 Hz, 1H); 5.42 (m, 1H); 5.91 (m, 1H); 7.05 (s, 1H); 7.35-
7.61 (m, 9H); 8.02-8.20 (m, 6H). 13C NMR δ: 64.0, 65.5, 68.1,
(35) Lopez, R.; Fernandez-Mayoralas, A. J . Org. Chem. 1994, 59,
737-745.
(34) Fletcher, H. G.; Hudson, C. S. J . Am. Chem. Soc. 1947, 69, 921-
(36) Benzoic acid 99% enriched with 13C in the carbonyl carbon was
obtained from Cambridge Isotope Laboratories, Woburn, MA 01801.
924.

Neighboring Group Participation in â-Xylosylation
J . Org. Chem., Vol. 64, No. 14, 1999 5229
127.7, 128.6, 128.7, 129.4, 129.5, 129.6, 130.0, 130.3, 133.5,
133.6, 133.7, 142.1. Anal. Calcd for C26H20O7: C, 70.27; H, 4.54.
Found: C, 70.04; H, 4.71.
Activa tion of th e 13C-La beled Th ioglycosid e 16 in th e
P r esen ce of DTBMP : Obser va tion of Ca tion 17 a n d of
Or th o Ester 21. AgOTf (23 mg, 0.09 mmol) and PhSCl (6.7
µL, 0.06 mmol) were stirred in CD2Cl2 (0.5 mL) under Ar at
-78 °C and then transferred quickly to a -78 °C solution of
the labeled thioglycoside 16 (16 mg, 0.028 mmol) and DTBMP
(11.9 mg, 0.06 mmol) in CD2Cl2 (0.5 mL) under Ar. The tube
was shaken and inserted into the probe of the spectrometer
precooled to -78 °C, and the 13C NMR spectrum illustrated
in Figure 1a was recorded. The tube was removed from the
probe and cold (-78 °C) methanol (2 µL) added. The tube was
then immediately replaced in the cold (-78 °C) probe, and the
13C NMR spectrum presented in Figure 1b was recorded.
Activa tion of th e 13C-La beled Th ioglycosid e 16 in th e
Absen ce of DTBMP : Obser va tion of Ca tion 17. A NMR
tube experiment was conducted exactly as described above
except for the omission of the base. The spectrum recorded
immediately after mixing of 16 with the preformed PhSOTf
is given in Figure 1c. A further 13C NMR spectrum was
recorded after the addition of methanol; no ortho ester signals
were detected, and all of the intensity was in the region of the
carbonyl carbons.
Va r ia ble Tem p er a tu r e NMR Exp er im en t w ith r-Su l-
foxid e 3 in CD2Cl2. Sulfoxide 3 (10.3 mg, 0.018 mmol) and
DTBMP (7.4 mg, 0.036 mmol) were dissolved in CD2Cl2 (0.5
mL) and cooled to -78 °C in the probe of the NMR spectrom-
eter. A 1H NMR spectrum was recorded, and then the tube
was removed and stored in a dry ice/acetone bath while Tf2O
(3.95 µL, 0.023 mmol, precooled in the same bath) was added.
The tube was vigorously shaken and quickly reinserted in the
cold probe, and a further 1H NMR spectrum was taken. The
spectrum showed multiple, poorly resolved signals, and no
attempt at assignment was made. Nevertheless, it was clear
that the sulfoxide had been consumed. The 13C NMR spectrum,
recorded immediately afterward, had a very poor signal-to-
noise ratio, with the only discernible peaks being those of the
base and its salts. The probe was gradually warmed and a 1H
NMR spectrum recorded every 10 °C. At the lower tempera-
tures, the main changes were gradual sharpening of the
signals and a slow change in their relative intensities. Signals
for the pentenitol 15 (see above isolation) grew in slowly
beginning between -48 and -38 °C with complete decomposi-
tion to 15 occurring between -18 and -8 °C. This behavior
was typical of that observed in similar VT-NMR experiments
with the thioglycosides and the bromides except for the relative
intensities of the signals, which varied with the different donor/
promoter pairs.
Ack n ow led gm en t. We thank the NIH (GM 57335)
for support of this work.
J O990424F