Mechanism of 4,6-O-benzylidene-directed β-mannosylation as determined by α-deuterium kinetic isotope effects

Article abstract of DOI:10.1002/anie.200453688

Considerable oxacarbenium ion character may be in the transition state of a highly β-selective mannosylation reaction that proceeds via an α-mannosyl triflate. An α-deuterium kinetic isotope effect of 1.2 was measured at -78 °C (=1.1 at 25 °C). This infor

Full text of DOI:10.1002/anie.200453688

Communications
Glycosylation Mechanism
Mechanism of 4,6-O-Benzylidene-Directed b-
Mannosylation as Determined by a-Deuterium
Kinetic Isotope Effects**
David Crich* and N. Susantha Chandrasekera
Glycosylation is one of the most fundamental reactions in
organic chemistry and one that is absolutely critical to the
science of glycobiology.[1] Many diverse types of glycosidic
bonds are found in nature,[2] and an absolutely bewildering
array of methods exist to access them.[3,4] In spite of this,
studies on the mechanism of chemical glycosylation, impor-
tant prerequisites for any rational development of new and
improved methods, are extremely sparse. Most mechanistic
thinking in the area is shaped by the seminal paper of
Lemieux and co-workers advocating an array of contact and
solvent-separated ion pairs as reactive intermediates,[5] but
even that paper, prescient as it may be, contains no
quantitative data. The kinetics of displacement of anomeric
halides by simple amines, alcohols, and thiolates were
measured in the 1950s and 60s,[6] but very few studies have
been conducted with the inclusion of an actual promoter.[7,8]
Thus, “much of the evidence used to substantiate proposed
inter-glycosidic coupling mechanisms is anecdotal or circum-
stantial”.[8d] To some extent this is understandable as, until
recent years, many glycosylation reactions were heteroge-
neous, in other words, used an insoluble promoter. Addition-
ally there is the possibility that the transition state for the
actual glycosylation step is termolecular bringing together the
acceptor alcohol and an activated complex of the donor and
promoter. The enzymic formation and/or cleavage of glyco-
sidic bonds, however, has been very thoroughly studied for a
number of different enzymes, both by kinetic methods and
through the determination of kinetic isotope effects.[9]
Recently, we discovered a very rapid, direct preparation
of b-mannosides,[10] in which a 4,6-O-benzylidene-protected
mannosyl sulfoxide is first activated with triflic anhydride to
give a covalent a-mannosyl triflate.[11] This is then displaced
by the acceptor to give the b-mannoside with excellent yield
and selectivity. In a more recent version, the a-mannosyl
triflate is preformed from a mannosyl thioglycoside and the
combination of triflic anhydride and 1-benzenesulfinyl piper-
idine before addition of the acceptor.[12] This clean, homoge-
neous coupling reaction, which proceeds without promoter,
provides the opportunity to study an actual glycosidic bond-
[*] Prof. Dr. D. Crich, N. S. Chandrasekera
Department of Chemistry
University of Illinois at Chicago
845 West Taylor Street, Chicago, IL 60607-7061 (USA)
Fax: (+1)312-996-4439
E-mail: dcrich@uic.edu
[**] We thank the National Institutes of Health (GM 62160) for support
of our work in this area
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5386
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200453688
Angew. Chem. Int. Ed. 2004, 43, 5386 –5389

Angewandte
Chemie
forming reaction with the possibility of differentiating
between direct SN2 displacement and mechanisms involving
transient contact ion pairs. We report here on the execution of
such a study.
We elected to address the problem by the determination
of kinetic isotope effects (KIEs),[13–15] and, so, synthesized a
2
thiomannoside 4 > 95% enriched in H (D) at C1 by the
method outlined in Scheme 1.[16] Mannoside 4 was then
Scheme 3. The kinetic isotope effect experiment. TTBP=2,4,6-tri-tert-
butylpyrimidine, Tf2O=trifluoromethanesulfonic anhydride).
acetal resonance against the internal standard 7, a trace of the
a-mannoside 12,[22,23] and the methyl b-mannoside 13. After
separation by preparative HPLC, the 1H NMR spectrum of 11
was again recorded and the ratio of the benzylidene acetal to
anomeric protons determined by careful integration. The
complete sequence was repeated three times, giving three
independent sets of data.
The data were processed according to Equation (1), for
the determination of KIEs from reaction products,[13,15]
wherein F is the fractional conversion of the triflate 9 (yield
of 11) and R11 and R6 the ratios of the benzylidene and
anomeric resonances in the product 11 and the thioglycoside
6, corresponding to the D/H ratios in 11 and 6. In this manner
the a-deuterium KIE at ꢀ788C for each of three independent
runs was found to be 1.20, 1.21, and 1.16 (or 1.13, 1.13, and
1.10 after conversion to 258C)[24] the assumption being that
the conversion of 6 to 9 is quantitative.
Scheme 1. Preparation of labeled thioglycoside 4. PTSA=p-toluenesul-
fonic acid.
admixed with an equal quantity of the nondeuteriated
material to give 5, enriched to approximately 50%. This
substance was converted to the benzylidene acetal by trans-
acetalization with benzaldehyde dimethyl acetal, 50%
enriched with deuterium at the acetal position. Standard
benzylation then gave donor 6 incorporating approximately
50% deuterium at the anomeric and, as an internal standard,
the benzylidene acetal position (Scheme 2).
KIE ¼ lnð1ꢀFÞ=ln½1ꢀðFR11=R6Þꢁ
ð1Þ
Scheme 2. Preparation of doubly labeled donor 6. Bn=benzyl,
*H=proton with approximate enrichment of 50% in 2H.
a-Deuterium KIEs of this magnitude correspond well to
those observed in acid-catalyzed hydrolysis of simple methyl
glycosides,[25] and in the hydrolysis of glycosyl fluorides,[26]
leading to the conclusion that the displacement of the triflate
from 9 by the typical carbohydrate acceptor 10 to give 11
proceeds with the development of substantial oxacarbenium
ion character. This may be interpreted either by a fully
dissociative mechanism involving the intermediacy of a
transient contact ion pair (CIP) (Scheme 4, path a), or by a
functionally equivalent mechanism involving an “exploded”
transition state (Scheme 4, path b).[27] In the CIP mechanism
the triflate anion is necessarily closely associated with face of
the oxacarbenium ion from which it has just departed and
shields that face against attack by the incoming alcohol. In the
alternative mechanism there is a loose association of the
nucleophile with the anomeric center as the leaving group
departs. The minor amount of a-anomer 12 formed in these
reactions most likely arises through the intermediacy of a
looser, perhaps solvent-separated, ion pair (SSIP), which is in
Donor 6 and approximately 50 mol% of 4,4,5,5-tetra-
methyl-2-(1-naphthyl)-1,3-dioxolane (7), a convenient inter-
nal standard with which to determine conversion, were
dissolved in CDCl3, the 1H NMR spectrum was recorded,
and the singlets corresponding to the benzylidene acetal and
anomeric hydrogens were integrated against the signals of
7.[17] After removal of the CDCl3, the mixture was taken up in
CH2Cl2 and admixed with tri-tert-butylpyrimidine (TTBP),[18]
and 140 mol% of 8.[19,20] The solution was then cooled to
ꢀ788C and treated with 150 mol% of Tf2O to give the a-
mannosyl triflate 9. Acceptor 10 (50 mol%) was added[21] and
then 1.5 h later MeOH was added before the reaction mixture
was quenched (Scheme 3).
1
The H NMR spectrum of the reaction mixture revealed
the formation of the b-mannoside 11, whose yield was
determined by integration of the corresponding benzylidene
Angew. Chem. Int. Ed. 2004, 43, 5386 –5389
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5387

Communications
[7] a) J. E. Wallace, L. R. Schroeder, J. Chem. Soc. Perkin Trans. 2
1977, 795; b) J. E. Wallace, L. R. Schroeder, J. Chem. Soc. Perkin
Trans. 2 1976, 1632.
[8] In contrast there are numerous quantitative and semiquantita-
tive studies on substituent effects on glycosylation rates: a) N. L.
Douglas, S. V. Ley, U. Lucking, S. L. Warriner, J. Chem. Soc.
Perkin Trans. 1 1998, 51; b) Z. Zhang, I. R. Ollmann, X.-S. Ye, R.
Wischnat, T. Baasov, C.-H. Wong, J. Am. Chem. Soc. 1999, 121,
734; c) C. P. J. Glaudemans, H. G. Fletcher, J. Am. Chem. Soc.
1965, 87, 4636; d) L. G. Green, S. V. Ley, in Carbohydrates in
Chemistry and Biology, Vol. 1 (Eds.: B. Ernst, G. W. Hart, P.
Sinay¨), Wiley-VCH, Weinheim, 2000, p. 427.
[9] a) G. J. Davies, M. L. Sinnott, S. G. Withers in Comprehensive
Biological Catalysis, Vol. 1(Ed.: M. L. Sinnott), Academic Press,
London, 1998, p. 119; b) B. W. Murray, V. Wittmann, M. D.
Burkart, S.-C. Hung, C.-H. Wong, Biochemistry 1997, 36, 823;
c) Y. Tanaka, W. Tao, J. S. Blanchard, E. J. Hehre, J. Biol. Chem.
1994, 269, 32306; d) A. M. MacLeod, D. Tull, K. Rupitz, A. J.
Warren, S. G. Withers, Biochemistry 1996, 35, 13165; e) D. L.
Zechel, S. G. Withers, Acc. Chem. Res. 2000, 33, 11.
Scheme 4. The proposed glycosylation mechanism.
[10] a) D. Crich, S. Sun, J. Org. Chem. 1997, 62, 1198; b) D. Crich, S.
Sun, Tetrahedron 1998, 54, 8321; c) D. Crich in Glycochemistry:
Principles, Synthesis, and Applications, (Eds.: P. G. Wang, C. R.
Bertozzi), Marcel Dekker, New York, 2001, p. 53.
[11] D. Crich, S. Sun, J. Am. Chem. Soc. 1997, 119, 11217.
[12] D. Crich, M. Smith, J. Am. Chem. Soc. 2001, 123, 9015.
[13] Reaction Rates of Isotopic Molecules (Eds.: L. C. S. Melander,
W. H. Saunders), Wiley-Interscience, New York, 1980.
[14] Displacement of triflate from preformed glycosyl triflates is
complete in minutes ꢀ788C,[11] ruling out the use of kinetic
measurements, at least by the NMR methods used to character-
ize these intermediates.
[15] On grounds of the amount of substrate needed, the substrate
solubility, and the instrument time needed, direct determination
of kinetic isotope effects by NMR spectroscopy at natural
abundance (as in a) D. A. Singleton, A. A. Thomas, J . Am.
Chem. Soc. 1995, 117, 9357; b) D. A. Singleton, M. J. Szymanski,
J. Am. Chem. Soc. 1999, 121, 9455; c) J. K. Lee, A. D. Bain, P. J.
Berti, J. Am. Chem. Soc. 2004, 126, 3769) was deemed
impractical.
equilibrium with an initial CIP. The function of the torsionally
disarming[28] benzylidene group is oppose rehybridization at
the anomeric carbon and, so, to shift the complete set of
equilibria toward the covalent triflate and away from the
SSIP, thereby minimizing a-glycoside formation.[29] The
expected chemical shift of an oxacarbenium ion carbon is
dC1 ~ 250 ppm,[30] whereas that measured[11] for the covalent
triflate, the only observable species by NMR spectroscopy, is
dC1 = 104.5 ppm. It is apparent, therefore, that the complete
set of equilibria between the covalent triflate 9, the CIP, and
SSIP lie very heavily toward 9 in complete agreement with
known lifetimes of oxacarbenium ions.[31,32]
The development of significant oxacarbenium ion char-
acter even in the highly stereoselective 4,6-O-benzylidene-
directed b-mannosylation strongly suggests that other, less
selective glycosylation reactions will be similarly dissocia-
tive.[33] The application of the current technique to other
glycosylation methods and stereochemical series is currently
underway.
[16] The critical feature is the liberation of the deuteriated mannose
from 2 with neopentyl glycol and PTSA which minimizes
isomerization to glucose.
[17] 2H NMR spectroscopy was investigated but the linewidths and
chemical shifts were such that the integration was compromised.
[18] D. Crich, M. Smith, Q. Yao, J. Picione, Synthesis 2001, 323.
[19] Compound 8, a more soluble version of 1-benzenesulfinyl
piperidine,[12] permits activation at ꢀ788C rather than the
ꢀ608C needed for the latter.
Received: January 7, 2004
Revised: August 4, 2004 [Z53688]
Keywords: carbohydrates · glycosylation · isotope effects ·
.
reaction mechanisms
[20] An excess of 8 and Tf2O was employed to ensure complete
conversion of 6 to 9, as verified on the NMR spectrum of the
crude reaction mixture following glycosylation, such that any
isotope effects observed do not result from the initial activation
step.
[1] R. A. Dwek, T. D. Butters, Chem. Rev. 2002, 102, 283.
[2] Comprehensive Natural Products Chemistry, Vol. 3 (Eds.:
D. H. R. Barton, K. Nakanishi, O. Meth-Cohn, B. M. Pinto),
Pergamon, Oxford, 1999.
[3] Carbohydrates in Chemistry and Biology, Vol. 1 (Eds.: B. Ernst,
G. W. Hart, P. Sinay¨), Wiley-VCH, Weinheim, 2000.
[4] In 1994 > 700 glycosides were synthesized chemically using
hundreds of combinations of donors and promoters: F. Barresi,
O. Hindsgaul, J. Carbohydr. Chem. 1995, 14, 1043.
[5] R. U. Lemieux, K. B. Hendriks, R. V. Stick, K. James, J. Am.
Chem. Soc. 1975, 97, 4056.
[6] For example see: a) B. Capon, Chem. Rev. 1969, 69, 407; b) A. J.
Rhind-Tutt, C. A. Vernon, J. Chem. Soc. 1960, 4637; c) B. Capon,
P. M. Collins, A. A. Levy, W. G. Overend, J. Chem. Soc. 1964,
3242; d) L. R. Schroeder, J. W. Green, D. C. Johnson, J. Chem.
Soc. B 1966, 447.
[21] P. J. Garegg, H. Hultberg, S. Wallin, Carbohydr. Res. 1982, 108,
97.
[22] The yield of 12 was < 5%, ruling out determination of the KIE
for its formation, but allowing it to be neglected in the
calculation of the KIE of 11.
[23] Ideally, comparison of the KIEs on the a- and b-products in an
unselective coupling would be informative. However, this is very
difficult to achieve as the method employed requires baseline
resolution of the various anomeric and benzylidene acetal
signals. In practice, this has proven to be a limitation with
numerous substrates assayed, particularly for unselective reac-
tions when the number of similar signals is multiplied. Second,
while the reaction conditions can be manipulated[10a,b] to give
5388
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5386 –5389

Angewandte
Chemie
predominantly the a-product it is doubtful whether measure-
ments made under such circumstances are relevant as they
presumably do not proceed via the intermediate glycosyl triflate.
[24] Assuming T(lnkH/kD) is constant and no tunnelling occurs.
[25] a) A. J. Bennet, M. L. Sinnott, J. Am. Chem. Soc. 1986, 108, 7287;
b) D. Indurugalla, A. J. Bennet, J. Am. Chem. Soc. 2001, 123,
10889; c) A. J. Bennet, M. L. Sinnott, W. S. S. Wijesundera, J.
Chem. Soc. Perkin Trans. 2 1985, 1233.
[29] Alternative explanations: a) T. Nukada, A. Berces, D. M. Whit-
field, Carbohydr. Res. 2002, 337, 765; b) H. H. Jensen, L. U.
Nordstrøm, M. Bols, J. Am. Chem. Soc. 2004, 126, 9205.
[30] a) G. A. Olah, D. G. Parker, N. Yoneda, J. Org. Chem. 1977, 42,
32; b) D. A. Forsyth, V. A. Osterman, J. R. DeMember, J. Am.
Chem. Soc. 1985, 107, 818.
[31] a) T. L. Amyes, W. P. Jencks, J. Am. Chem. Soc. 1989, 111, 7888;
b) M. N. Namchuk, J. D. McCarter, A. Becalski, T. Andrews,
S. G. Withers, J. Am. Chem. Soc. 2000, 122, 1270; c) J. Zhu, A. J.
Bennet, J. Am. Chem. Soc. 1998, 120, 3887.
[26] Y. Zhang, J. Bommuswamy, M. L. Sinnott, J. Am. Chem. Soc.
1994, 116, 7557.
[32] The fact that the covalent a-triflate 9 is so heavily favored argues
against the possibility that the effects measured here result from
an equilibrium isotope effect. Likewise, the highly stereoselec-
tive nature of the coupling argues against the effects measured
arising from any significant shift in the equilibrium position as
this would necessarily be associated with reduced selectivity.
[33] On the grounds that the reaction studied is highly stereo-
selective, and alcohol 10 is a typical carbohydrate the results
obtained here may reasonably be considered as representative of
this class of reactions. Of course exceptions may exist, partic-
ularly with extremely selective cases using highly reactive
alcohols, but they are not likely to be representative of the
formation of true interglycosidic bonds.
[27] The distinction between the two mechanisms is a fine one and
cannot be made on the basis of the available data. Exactly
analogous problems arise in the study of the hydrolysis of simple
glycosides,[25] in that of glycosyl fluorides,[26] and in enzymic
glycoside hydrolysis and glycosyl transfer.[9] Comparisons with
KIEs determined for enzymic reactions must, however, be
treated with caution since the enzyme likely binds the substrate
in a nonstandard conformation. For example: G. Sulzenbacher,
H. Driguez, B. Henrissat, M. Schulein, G. J. Davies, Biochemistry
1996, 35, 15280.
[28] C. W. Andrews, R. Rodebaugh, B. Fraser-Reid, J. Org. Chem.
1996, 61, 5280.
Angew. Chem. Int. Ed. 2004, 43, 5386 –5389
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5389