First synthetic carbohydrates with the full anticoagulant properties of heparin

Article abstract of DOI:10.1002/(SICI)1521-3773(19981116)37:21<3009::AID-ANIE3009>3.0.CO;2-F

A single disaccharide building block is required to obtain synthetic carbohydrates that reproduce the anticoagulant activity of heparin and inhibit thrombin (n > 6) and/or factor Xa (n ≥ 2; see reaction scheme). Thus, there is evidence that heparin fragme

Full text of DOI:10.1002/(SICI)1521-3773(19981116)37:21<3009::AID-ANIE3009>3.0.CO;2-F

COMMUNICATIONS
Toluene was distilled from a Na/K alloy and degassed. [Be{N(SiMe3)2}2]
was prepared according to a known procedure.[7] Caution: [Be{N(SiMe3)2}2]
is a known carcinogen and should be handled only in a well-ventilated fume
hood with proper precautions. Benzenethiol was dried over activated 4-Š
molecular sieves. [18]Crown-6 was dissolved in hexane, and freshly cut
potassium was added. The mixture was briefly heated to reflux, and the
remaining potassium removed by filtration. The dried crown ether was
isolated by crystallization from hexane.
[7] For the synthesis and solid-state structure of [Be{N(SiMe3)2}2], see
a) H. Bürger, C. Forker, J. Goubeau, Monatsh. Chem. 1965, 69, 597;
b) A. H. Clark, A. Haaland, J. Chem. Soc. Chem. Commun. 1969, 912.
[8] Pyridine was used to obtain a family of soluble magnesium phenyl-
thiolates: S. Chadwick, U. Englich, M. O. Senge, B. C. Noll, K.
Ruhlandt-Senge, Organometallics 1998, 17, 3077.
[9] R. A. Anderson, G. E. Coates, J. Chem. Soc. Dalton Trans. 1975, 1244.
[10] A. J. Amoroso, A. M. Arif, J. A. Gladysz, Organometallics 1997, 16,
6032, and references therein.
1: HSPh (0.21 mL, 2.0 mmol) was added dropwise to [Be{N(SiMe3)2}2]
(0.33 g, 1.0 mmol) and [18]crown-6 (0.26 g, 1.0 mmol) in toluene (20 mL) at
room temperature. A heavy white precipitate formed after stirring for
several minutes. Pyridine (0.1 mL, 1.2 mmol) was added dropwise, and a
homogeneous light yellow solution was obtained. After the mixture had
been stirred at room temperature for 30 min, it was filtered through a
Celite-loaded frit and stored at 08C. Colorless plates gradually grew over
several days, and 0.20 g (55% yield) were collected from the first
crystallization.[17] The white powder shrank slightly when heated above
808C and then irreversibly melted to a yellow oil in the range of 145 ±
1538C. 1H NMR (300 MHz, 258C, [D8]THF): d ˆ 8.54 ± 6.71 (broad over-
lapping signals, 40H), 3.52 (s, 24H), 2.31 (s, 6H); IR (Nujol): nÄ ˆ 3302m,
3197m, 3043w, 2926s, 1643m, 1610m, 1576s, 1463s, 1377m, 1352m, 1317s,
1286m, 1216m, 1137m, 1105s, 1070m, 1050m, 956s, 837m, 785m, 743s,
696s, 648w, 580m, 478 cm 1 m.
[11] D. Labahn, F. M. Bohnen, R. Herbst-Irmer, E. Pohl, D. Stalke, H. W.
Roesky, Z. Anorg. Allg. Chem. 1994, 620, 41.
[12] A. P. Purdy, A. D. Berry, C. F. George, Inorg. Chem. 1997, 36, 3370.
[13] M. A. Beswick, J. M. Goodman, C. N. Harmer, A. D. Hopkins, M. A.
Paver, P. R. Raithby, A. E. H. Wheatley, D. S. Wright, Chem. Com-
mun. 1997, 1879.
[14] D. C. Bradley, H. Chudzynska, M. E. Hammond, M. B. Hursthouse,
M. Motevalli, W. Ruowen, Polyhedron 1992, 11, 375.
[15] C. Burns, personal communication.
[16] [Be(SC6F5)2(NH3)(H2NSiMe3)] was synthesized by treatment of
[Be{N(SiMe3)2}2] with two equivalents of HSC6F5 in toluene. The
resulting colorless crystals were analyzed crystallographically. How-
ever, further spectroscopic analysis was not possible due to rapid
decomposition of the sample. We were unable, even after several
attempts, to trap the intermediate and obtain spectroscopic data. S.
Chadwick, K. Ruhlandt-Senge, unpublished results.
[17] The maximum theoretical yield of this reaction is 66% taking into
account the 1:1 ratio of [Be(SC6H5)2] and NH3 in the final product.
Accordingly, the experimental yield was 55%. Since the limiting
reagent in this reaction is HSC6H5, the reaction was repeated with six
equivalents of HSC6H5, as indicated in Scheme 1. Instead of isolating
an increased amount of 1, only [(NH4)(SC6H5)(py)]1 was identified in
the solid state.
Crystal structure data for 1: C60H76Be2N4O6S4, Mr ˆ 1095.51, triclinic, space
Å
group P1, a ˆ 10.432(2), b ˆ 13.5950(3), c ˆ 22.0275(4) Š, a ˆ 79.374(1),
b ˆ 77.760(1), g ˆ 87.1608, Vˆ 3000.5(1) Š3, Tˆ 150 K, Z ˆ 2, m(MoKa) ˆ
0.210, crystal dimensions 0.42 Â 0.20 Â 0.20 mm. Of 12886 independent
reflections collected (2.56 ꢀ 2q ꢀ 56.008) on a Siemens SMART system
with a three-circle goniometer and a CCD detector operating at 548C,
8176 were observed (I > 2s(I)). Crystal decay was monitored by repeating
a set of initial frames at the end of data collection and comparing the
duplicate reflections; no decay was observed. An absorption correction was
applied with the program SADABS.[18] The crystal structure was solved by
direct methods with SHELXTL. Missing atoms were located in subsequent
difference Fourier maps and included in the refinement. The structure of 1
was refined by full-matrix least-squares refinement on F 2.[19] Hydrogen
atoms with the exception of the NH protons were placed geometrically and
refined by using a riding model with Uiso constrained at 1.2Ueq of the carrier
[18] G. M. Sheldrick, SADABS, Program for Absorption Correction Using
Area Detector Data; Universität Göttingen, Germany, 1996.
[19] G. M. Sheldrick, SHELXTL, Version 5. Siemens Analytical X-ray
Instruments, Madison, WI, 1994.
C
atom. All non-hydrogen atoms were refined anisotropically. NH3
hydrogen atoms were located in difference maps and included in the
refinement by using distance restraints. A center of symmetry, suspected in
the center of the crown ether molecule could not be confirmed even after
various symmetry checks and transformation of the suspected inversion
center to the origin of the unit cell. The absence of significant correlations
also confirms the correct symmetry. R1 ˆ 0.0575 for data with I > 2s(I), and
wR2 ˆ 0.1234 for all data. Crystallographic data (excluding structure
factors) for the structure reported in this paper have been deposited with
the Cambridge Crystallographic Data Centre as supplementary publication
no. CCDC-101292. Copies of the data can be obtained free of charge on
application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax:
(‡44)1223-336-033; e-mail: deposit@ccdc.cam.ac.uk).
First Synthetic Carbohydrates with the Full
Anticoagulant Properties of Heparin**
Maurice Petitou,* Philippe Duchaussoy,
Pierre-A. Driguez, Guy Jaurand, Jean-P. Herault,
Â
Jean-C. Lormeau, Constant A. A. van Boeckel, and
Jean-M. Herbert
Heparin, a major drug for the prevention and treatment of
cardiovascular diseases, exerts its activity through activation
of the serine proteinase inhibitor antithrombin III (AT III),
Received: March 31, 1998 [Z11664IE]
German version: Angew. Chem. 1998, 110, 3204 ± 3206
the main physiological inhibitor of blood coagulation.[1, 2]
A
unique pentasaccharide sequence[3] in the polysaccharide
binds to the protein in a highly specific way, inducing a
Keywords: alkaline earth metals ´ amides ´ hydrogen bonds
´ protonations ´ S ligands
[*] Dr. M. Petitou, Dr. P. Duchaussoy, Dr. P.-A. Driguez, Dr. G. Jaurand,
[1] a) K. Ruhlandt-Senge, Inorg. Chem. 1995, 34, 3499; b) K. Ruhlandt-
Senge, Comm. Inorg. Chem. 1997, 19, 351; c) S. Chadwick, U. Englich,
B. Noll, K. Ruhlandt-Senge, Inorg. Chem. 1998, 37, 4718.
[2] N. A. Bell in Comprehensive Organometallic Chemistry, Vol. 1 (Eds.:
G. Wilkinson, F. G. A. Stone, E. W. Abel), Pergamon, Oxford, 1995,
p. 35.
[3] K. Ruhlandt-Senge, R. A. Bartlett, M. M. Olmstead, P. P. Power,
Inorg. Chem. 1993, 32, 1724.
[4] M. Niemeyer, P. P. Power, Inorg. Chem. 1997, 36, 4688.
[5] H. Nöth, D. Schlosser, Chem. Ber. 1988, 121, 1711.
[6] a) G. E. Coates, A. H. Fishwick, J. Chem. Soc. 1968, 635; b) G. E.
Coates, A. H. Fishwick, J. Chem. Soc. 1968, 640.
Â
Dr. J.-P. Herault, Dr. J.-C. Lormeau, Dr. J.-M. Herbert
Sanofi Recherche
Haemobiology Research Department
195, route dꢁEspagne, F-31036 Toulouse Cedex (France)
Fax : (‡ 33)5-61-16-22-86
E-mail: maurice.petitou@sanofi.com
Prof. Dr. Constant A. A. van Boeckel
N.V. Organon, Oss (The Netherlands)
[**] This work is part of a collaborative project between N.V. Organon
(The Netherlands) and Sanofi Recherche (France) on antithrombotic
oligosaccharides.
Angew. Chem. Int. Ed. 1998, 37, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3009 $ 17.50+.50/0
3009

COMMUNICATIONS
bind at either end of the molecule,
and thrombin can be attracted by
the rest of the chain (Scheme 2).
c) For thrombin to efficiently com-
pete with AT III for binding to a
continuum of ABDs, the affinity of
the AT III/ABD couple must pref-
erably be in the same range of
affinity as the thrombin/heparin
coupleÐthat is, in the micromolar
range[11] and therefore two to three
Scheme 1. An anticoagulant heparin molecule contains an antithrombin binding domain (ABD) prolonged
at both ends by thrombin binding domains (TBDs). An ABD is a unique pentasaccharide sequence
characterized by the presence of b-d-glucuronic acid linked to a N-sulfo-3,6-di-O-sulfo-a-d-glucosamine;
TBDs mainly consist of repeated trisulfated disaccharides made of 2-O-sulfo-a-l-iduronic acid linked to N-
sulfo-6-O-sulfo-a-d-glucosamine.
conformational change that allows adequate presentation of
the inhibitory peptidic loop of AT III to the active sites of
serine proteinases. The mere interaction with activated AT III
is sufficient to inhibit coagulation factor Xa. This is at
variance with thrombin inhibition, which requires the for-
mation of a ternary complex involving heparin, AT III, and
thrombin.[2] A longer heparin chain is then needed (Scheme 1)
which contains the above-mentioned pentasaccharide that is
prolonged at both ends by repeated trisulfated disaccharide
units.[4] Thrombin is electrostatically attracted through its
anion binding site exosite II,[5] and slides along the chain until
it finally hooks itself on the inhibitory loop of activated
AT III.
Thus, an oligosaccharide molecule containing both an
AT III binding domain (ABD) and a thrombin binding
domain (TBD) should be able to mimic the full anticoagulant
activity of heparin. Various biochemical studies aimed at
determining the size of such molecules pointed to tetradeca-,
octadeca-, and eicosasaccharides.[6] It is noteworthy that when
AT III is docked on its binding domain, thrombin, sliding
along the heparin chain, must approach from the correct side
to be inhibited. This implies that only one of the two possible
ways of elongating the ABD with a TBD (reducing or
nonreducing ends, Scheme 2) will allow thrombin inhibition.[7]
Synthesizing an oligosaccharide of this size, and reproduc-
ing the exact structure of the hypothetical heparin fragment,
would be practically unfeasible.[9] In a first approach to
overcome this difficulty, two saccharide molecules, an ABD
and a TBD, were bridged by a non-carbohydrate linker.[8] The
resulting conjugates displayed inhibition of thrombin and
factor Xa, but in some assays the thrombin inhibitory potency
was weaker than that of heparin. In the approach reported
here we exploited the ªnon-glycosaminoº glycan series of
heparin analogues,[10] where N-sulfonato substituents are
replaced by O-sulfonato groups and hydroxyl groups are
alkylated. These structural modifications fully preserve the
specific binding to AT III, yet they dramatically simplify the
synthesis.
The structures of the target molecules stem from the
following considerations: a) In contrast to the highly specific
interaction between AT III and heparin, the interaction of
thrombin with the polysaccharide merely results from an
electrostatic attraction governed by the density of negative
charges on the polysaccharide backbone.[4] Consequently the
ABD, which contains at least six critical negative charges
(Scheme 3), can also attract thrombin and serve as a TBD.
b) A continuum of ABDs would necessarily display the
correct relative position of ABD and TBD, since AT III can
Scheme 2. Inhibition of thrombin by AT III and heparin. In an active
anticoagulant molecule, AT III binds to heparin through the ABD
pentasaccharide (see Scheme 1, symbolized in 1a, b by black rectangles
and triangles) and exposes its activated inhibitory loop (light grey triangle)
to the catalytic site of thrombin (also called factor IIa). Thrombin, attracted
by the negative charges of the TBD (white rectangles and triangles), slides
along the chain until it hooks itself on the exposed loop of the inhibitor.
Heparin and AT III dock themselves in a unique, unknown way (1a or 1b),
and thrombin apparently has to approach from the correct side to be
inhibited. Among these two possible arrangements of ABD and TBD (both
are present in a heparin molecule, see Scheme 1), the efficient one was
unknown.[7] We therefore designed an hybrid molecule (dark grey
rectangles and triangles) which is a continuum of ABDs. AT III can bind
at either end of the molecule (2a ± d), and, whatever the efficient
arrangement (2a or 2b), thrombin can be attracted and inhibited (2a or
IIc) provided the saccharide chain is long enough to accommodate both
proteins.
orders of magnitude lower than the highest affinity reached
with some synthetic pentasaccharides. From this latter con-
sideration it follows that we are, to a certain extent, allowed to
adapt the structure of the ABD to our synthesis strategy
(provided we preserve enough binding affinity). This is not a
minor advantage for the synthesis of the targeted long
saccharide fragments.
The structure of a pentasaccharide with high affinity for
AT III[12] is shown in Scheme 3. The presence of two types of
uronic acids having the d-gluco and the l-ido configurations is
a structural features of high-affinity compounds. From this
structure we reasoned that hexasaccharides consisting of a
3010
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3010 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 21

COMMUNICATIONS
Scheme 4. a) Me3SiOSO2CF3 (TMSOTf), toluene, 4-Š MS,
208C,
30 min, 62%; b) 1m NH2NH2/H2O in pyridine/AcOH (3/2), 15 min, 87%;
c) as for a), 48%; d) H2, 10% Pd/C, MeOH, 24 h; e) 0.5m NaOH, 5 h (91%
based on 7); f) pyridine:SO3, DMF, 558C, 24 h, 98%.
were deprotected and sulfated. Thus, reaction of stoichio-
metric amounts of the known[14] glycosyl donor 4 and the
acceptor 5 (prepared as reported for the methyl ester
analogue[10a]
)
yielded the tetrasaccharide 6[15] (62%,
Scheme 4). Selective cleavage of the levulinic ester gave the
corresponding glycosyl acceptor (87%), which was again
condensed with 4 to afford the fully protected hexasaccharide
7 (a 5:1 mixture of the a and b anomers was formed, from
which 7 was isolated in 39% yield after chromatography on
silica gel). Catalytic hydrogenolysis of the benzyl groups was
followed by saponification of the esters (91%). Complete
removal of the protecting groups was monitored by high-field
1H NMR spectroscopy. Subsequent sulfation and lyophiliza-
tion gave 1 as a white powder (98%).[15] Compound 2[15] was
obtained in a similar manner (Scheme 5) from 4 and the
acceptor 8, which was prepared as reported for its methyl
ester counterpart.[12] Finally, a similar series of reactions
(Scheme 6) gave 3.[15] The sequence started from 12 and 13,
which were both are derived from a common precursor, the
disaccharide 11, prepared as reported for its methyl ester
counterpart.[12]
The affinity of 1 ± 3 for AT III was determined by fluo-
rescence spectroscopy[15] and compared to that of the
pentasaccharide in Scheme 3 (Kd ˆ 1.4 Æ 0.2nm). The intro-
duction of a third trisulfated glucose unit at the nonreducing
end of the pentasaccharide sequence proved to be fully
compatible with the recognition of AT III (Kd(1) ˆ 0.8 Æ
0.3nm). In contrast, substituting d-glucuronic acid for l-
iduronic acid resulted in a dramatic loss of affinity (Kd(2) ˆ
3.4 Æ 0.3mm). The opposite change (substituting l-iduronic
acid for d-glucuronic acid), although it also decreased the
affinity for AT III, nicely fitted our ªrelative affinity crite-
rionº discussed above (c) since the affinity of 3 for AT III
(Kd ˆ 0.35 Æ 0.01mm) was very close to that reported[11] for
thrombin and heparin (1mm). We expected 3 (which like
Scheme 3. Simplifying the structure of a high-affinity binding pentasac-
charide sequence. The critical anionic groups involved in the interaction
with AT III are labeled with an asterisk. The proposed changes (second
line) introduce another sulfate group at position 3 of the third glucose unit
and alter the configuration at C5 of the uronic acids. Such changes would
allow synthesis from a single disaccharide building block (third line).
Compounds 1 ± 3 were synthesized to test these various possibilities: 1
contains a third trisulfated glucose residue, 2 contains only glucuronic acid,
and 3 contains only iduronic acid (the latter was obtained from a single
disaccharide building block).
single repeated disaccharide unit might display significant
affinity for AT III, possibly in the desired micromolar range.
Such hexasaccharides, obtainable from a single disaccharide
synthon, meet our criteria for an easy synthesis of long
fragments (the synthesis of the ultimate target longer
saccharides would also be possible from the same basic
synthon). To test this possibility we prepared the hexasac-
charides 1 ± 3 depicted in Scheme 3. Hexasaccharide 1 was
synthesized to check the influence of adding one more
trisulfated glucose unit, 2 to test whether the easily available
d-glucuronic acid could substitute for l-iduronic acid, and 3 to
obtain a compound in which l-iduronic acid is the only uronic
acid present. The affinity of 1 ± 3 for AT III was determined.
All three hexasaccharides were obtained using strategies
(Scheme 4) where disaccharide building blocks were prepared
first and then repeatedly condensed to each other by the
imidate glycosylation procedure;[13] the final hexasaccharides
Angew. Chem. Int. Ed. 1998, 37, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3011 $ 17.50+.50/0
3011

COMMUNICATIONS
Having identified an ABD that is easily obtainable from a
single disaccharide building block (11), we attempted the
synthesis of longer fragments consisting of a single repeating
basic disaccharide unit. The strategy used to prepare 3 was
further developed, as shown in Scheme 7. The imidate 12 was
used as glycosyl donor to add disaccharide units to the new
Scheme 5. a) TMSOTf, CH2Cl2, 4-Š MS, 208C, 30 min, 69%; b) 1m
NH2NH2/H2O in pyridine/AcOH (3/2), 15 min, 95%; c) as for a), 63%;
d) H2, 10% Pd/C, MeOH, 24 h; e) 0.5m NaOH, 5 h, 99% (based on 10);
f) pyridine:SO3, DMF, 558C, 24 h, 95%.
Scheme 7. a) TMSOTf, toluene, 4-Š MS, 208C, 30 min; b) 1m NH2NH2/
H2O in pyridine/AcOH (3/2), 15 min; c) H2, 10% Pd/C, MeOH, 24 h;
d) 0.5m NaOH, 5 h; e) pyridine:SO3, DMF, 558C, 24 h.
glycosyl acceptor resulting from cleavage of the levulinoyl
group of the oligosaccharide obtained at the preceding step.[17]
The process was reiterated until an eicosamer was obtained.
The structure of the fully protected intermediates (16 ± 22)
was ascertained by high-field 1H NMR spectroscopy and mass
spectrometry.[15] The yield of the successive coupling steps was
about 60% (not optimized), and no b-coupled product was
detected. The large difference in size between the glycosyl
donor disaccharide on the one hand, and the glycosyl acceptor
and the product of the reaction on the other hand, made it
very easy to recover, by gel permeation in dichloromethane/
ethanol, a mixture containing exclusively the unchanged
acceptor and the product. This mixture was occasionally
submitted to another glycosylation reaction to improve the
yield based on the expensive acceptors involved in the final
steps of chain elongation. Deprotection and sulfation of the
oligosaccharides (deca- to eicosamer) obtained at each step
involved, as above for the preparation of hexasaccharides,
hydrogenolysis, saponification, and sulfation. Complete re-
Scheme 6. a) CF3COOH/Ac2O/AcOH, 608C, 4 h, 67%; b) HOCH2-
CH2NH2, THF, RT, 74%; c) CCl3CN, K2CO3, CH2Cl2, RT, 1.5 h, 94%;
d) 1m NH2NH2/H2O in pyridine/AcOH (3/2), 15 min, 91%;
e) TBDM!SOTf, toluene, 4-Š MS, 208C, 30 min, 54%; f) as for d),
99%; g) as for e), 64%; h) H2, 10% Pd/C, MeOH, 24 h; i) 0.5m NaOH, 5 h;
j) pyridine:SO3, DMF, 558C, 24 h, 60% (based on 15).
heparin contains four negative charges per disaccharide unit)
to have a similar affinity for thrombin. Interestingly, 3 also
1
displays significant anti-factor Xa activity (325 unitsmg ).
We were pleased that the synthesis of 3 involves repeated
glycosylation at position 4 of a l-iduronic acid derivative,
since in our experience such a glycosylation always afforded
almost exclusively the desired a-coupled product when an
imidate was used as the glycosyl donor.
3012
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3012 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 21

COMMUNICATIONS
Table 1. Properties of the sulfated iduronic acid containing oligosaccharides.
3
23
24
25
26
27
28
heparin
saccharide units
molecular weight
factor Xa inhibition[a]
6
10
3606
405
12
4301
360
14
4995
310
16
5690
359
18
6384
270
20
7078
236
10 ± 50
ca. 15000
180
2217
325
(Æ16)
(Æ32)
(Æ29)
(Æ16)
(Æ29)
(Æ23)
(Æ19)
thrombin inhibition[b]
> 10
> 10
> 10
> 10
130
23
6.7
3.3
(113 ± 133)
(13 ± 30)
(3 ± 9)
(3 ± 4)
[a] [unitsmg 1] (standard deviation) (n ˆ 3). [b] IC50 [ngmL 1] (95% confidence interval).
Biochem. J. 1984, 218, 725 ± 732; d) A. Danielsson, E. Raub, U.
Lindahl, I. Björk, J. Biol. Chem. 1986, 261, 15467 ± 15473.
moval of the protecting groups was monitored before
sulfation by high-field H NMR spectroscopy. The desired
1
[7] At the start of this project the correct relative position of the ABD and
the TBD was not known. Molecular modeling studies then suggested
that the ABD should be prolonged at the nonreducing end.[8a]
[8] a) P. D. J. Grootenhuis, P. Westerduin, D. Meuleman, M. Petitou,
C. A. A. van Boeckel, Nature Struct. Biol. 1995, 2, 736 ± 739; b) P.
Westerduin, J. E. M. Basten, M. A. Broekhoven, V. de Kimpe,
W. H. A. Kuijpers, C. A. A. van Boeckel, Angew. Chem. 1996, 108,
339 ± 342; Angew. Chem. Int. Ed. Engl. 1996, 35, 331 ± 333.
[9] Reviews on the synthesis of heparin fragments: a) M. Petitou, C. A. A.
van Boeckel, Prog. Chem. Org. Nat. Prod. 1992, 60, 143 ± 210;
b) C. A. A. van Boeckel, M. Petitou, Angew. Chem. 1993, 105,
1741 ± 1761; Angew. Chem. Int. Ed. Engl. 1993, 32, 1671 ± 1690; c) M.
Petitou, C. A. A. van Boeckel, Pure Appl. Chem. 1997, 67, 1839 ± 1846.
[10] a) G. Jaurand, J. Basten, I. Lederman, C. A. A. van Boeckel, M.
Petitou, Bioorg. Med. Chem. Lett. 1992, 2, 897 ± 900; b) J. Basten, G.
Jaurand, B. Olde-Hanter, M. Petitou, C. A. A. van Boeckel, Bioorg.
Med. Chem. Lett. 1992, 2, 901 ± 904; c) J. Basten, G. Jaurand, B. Olde-
Hanter, P. Duchaussoy, M. Petitou, C. A. A. van Boeckel, Bioorg.
Med. Chem. Lett. 1992, 2, 905 ± 910.
sulfated oligomers 23 ± 28 were obtained as white powders
after lyophilization in yields of 49 to 86% over the three steps.
Their purity was assessed by capillary electrophoresis, and
1
their structure was clarified by H NMR spectroscopy and
electro spray ionization mass spectrometry (ESI-MS).[15]
These compounds were then submitted to biological tests to
assess their ability to catalyze inhibition of factor Xa[18] and
thrombin[19] by AT III (Table 1). All the compounds contain
an ABD, which explains their ability to bind to AT III and
their anti-factor Xa activity. The hexa-, deca-, dodeca-, and
tetradecasaccharide were inactive in the thrombin inhibition
assay, whereas the activity of the hexadeca-, octadeca-, and
eicosasaccharide increased with size in this assay. The IC50
values show that the eicosamer is half as potent as standard
heparin. Increasing the chain length would probably improve
the thrombin inhibitory potency since the higher activity of
the longer chains (compare 26 ± 28) reflects the well-known[2]
greater ability of longer heparin fragments to electrostatically
attract thrombin.
[11] S. T. Olson, J. Biol. Chem. 1988, 263, 1698 ± 1708.
[12] P. Westerduin, C. A. A. van Boeckel, J. E. M. Basten, M. A. Broek-
hoven, H. Lucas, A. Rood, H. van der Heijden, R. G. M. van Am-
sterdam, T. G. van Dinther, D. G. Meuleman, A. Visser, G. M. T.
Vogel, J. B. L. Damm, G. T. Overklift, Bioorg. Med. Chem. 1994, 2,
1267 ± 1280.
[13] R. R. Schmidt, W. Kinzy in Adv. Carbohydr. Chem. Biochem., Vol. 50
(Ed.: D. Horton), Academic Press, London, 1994, pp. 21 ± 123.
[14] M. Petitou, P. Duchaussoy, G. Jaurand, F. Gourvenec, I. Lederman, J.-
Using the imidate glycosylation procedure,[13] we have
synthesized for the first time oligosaccharide molecules
displaying the dual (anti-factor Xa and anti-thrombin) activity
of heparin mediated by AT III. Regarding the long standing
question[6] about the minimum size of heparin fragments able
to catalyze thrombin inhibition, the present data restrict the
choice to a pentadeca- or hexadecasaccharide.
Ã
Â
Â
M. Strassel, T. Barzu, B. Crepon, J.-P. Herault, J.-C. Lormeau, A.
Bernat, J.-M. Herbert, J. Med. Chem. 1997, 40, 1600 ± 1607.
[15] All new compounds were analyzed by 1H NMR spectroscopy, by mass
spectrometry, and occasionally by HPLC. Combustion analyses were
systematically performed on disaccharidic building blocks only.
Selected analytical data: 1H NMR data were collected at 500 MHz
in D2O (external standard TSP; for polar compounds) or in CDCl3
(internal standard TMS). The NMR data are given for anomeric
protons of the nonreducing end (NR) to the reducing end (R) units.
Mass spectrometry data were collected with liquid secondary ion mass
spectrometry (LSI-MS) or ESI-MS. 1: 1H NMR: d (J1,2 [Hz]) ˆ 4.64
(7.5; NR), 5.52 (3.7; NR-1), 4.63 (8.1; NR-2), 5.48 (3.5; R-2), 5.10 (6.5;
R-1), 5.10 (3.5; R); LSI-MS: m/z: 2280.4 [M Na] ; [a]D ˆ ‡30 (c ˆ 1
in H2O). 2: 1H NMR: d (J1,2 [Hz]) ˆ 4.66 (ca. 8; NR), 5.54 (3.7; NR-1),
4.64 (ca. 8; NR-2), 5.54 (3.7; R-2), 4.63 (ca. 8; R-1), 5.14 (3.6; R); LSI-
MS: m/z: 2192.8 [M Na] ; [a]D ˆ ‡43 (c ˆ 0.7 in H2O). 3: 1H NMR:
d (J1,2 [Hz]) ˆ 5.10 (ca. 1; NR), 5.41 (ca. 3; NR-1), 5.06 (2.2; NR-2),
5.41 (ca. 3; R-2), 5.09 (ca. 1; R-1), 5.15 (3.7; R); LSI-MS: m/z: 2192.6
[M Na] ; [a]D ˆ ‡28 (c ˆ 1 in H2O). 7: 1H NMR: d (J1,2 [Hz]) ˆ 4.16
(8; NR), 5.43 (3.5; NR-1), 4.03 (8; NR-2), 4.88 (ca. 3; R-2), 5.47 (6.1;
Received: May 25, 1998 [Z11894IE]
German version: Angew. Chem. 1998, 110, 3186 ± 3191
Keywords: drug research ´ enzyme inhibitors ´ heparin ´
oligosaccharides ´ serine proteinases
[1] Heparin (Eds.: D. A. Lane, U. Lindahl), Edward Arnold, London,
1989.
[2] S. T. Olson, I. Björk, Semin. Thromb. Hemost. 1994, 20, 373 ± 409.
[3] a) U. Lindahl, G. Bäckström, L. Thunberg, I. G. Leder, Proc. Natl.
Acad. Sci. USA 1980, 77, 6551 ± 6555; b) B. Casu, P. Oreste, G. Torri, G.
Zoppetti, J. Choay, J.-C. Lormeau, M. Petitou, P. SinayÈ, Biochem. J.
1981, 197, 599 ± 609; c) J. Choay, J.-C. Lormeau, M. Petitou, P. SinayÈ, J.
Fareed, Ann. N. Y. Acad. Sci. 1981, 370, 644 ± 649; d) L. Thunberg, G.
Bäckström, U. Lindahl, Carbohydr. Res. 1982, 100, 393 ± 410.
[4] B. Casu, Adv. Carbohydr. Chem. Biochem. 1985, 43, 51 ± 134.
[5] Review: M. T. Stubbs, W. Bode, Trends Biochem. Sci. 1995, 20, 23 ± 28.
[6] a) T. C. Laurent, A. Tengblad, L. Thunberg, M. Höök, U. Lindahl,
Biochem. J. 1978, 175, 691 ± 701; b) G. M. Oosta, W. T. Gardner, D. L.
Beeler, R. D. Rosenberg, Proc. Natl. Acad. Sci. USA 1981, 78, 829 ±
833; c) D. A. Lane, J. Denton, A. M. Flynn, L. Thunberg, U. Lindahl,
1
R-1), 5.05 (ca. 3; R); [a]D ˆ ‡76 (c ˆ 0.3 in CH2Cl2). 10: H NMR: d
(J1,2 [Hz]) ˆ 4.17 (ca. 8; NR), 5.39 (3.8; NR-1), 4.20 (ca. 8; NR-2), 5.33
(3.6; R-2), 4.14 (ca. 8; R-1), 4.56 (3.5; R); [a]D ˆ ‡86 (c ˆ 0.6 in
CH2Cl2). 15: 1H NMR: d (J1,2 [Hz]) ˆ 4.90 (4.3; NR), 5.10 (3.9; NR-1),
5.23 (6.5; NR-2), 5.12 (3.9; R-2), 4.89 (6.8; R-1), 4.56 (ca. 3; R); [a]D
ˆ
‡23 (c ˆ 0.5, CH2Cl2). The 1H NMR spectra of 16 ± 22 were very
similar in terms of chemical shifts, and, as expected, only the relative
intensities of the signals varied between compounds. Resonances for
the distinguishable anomeric protons: d (J1,2 [Hz]) ˆ 4.92 (4 ± 4.3;
Angew. Chem. Int. Ed. 1998, 37, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3013 $ 17.50+.50/0
3013

COMMUNICATIONS
NR), 5.10 (3.3 ± 3.5; NR-1), 5.26 (6 ± 7; R-1), 4.56 (3.5; R). Optical
rotations of 16 ± 22 were all in the range ‡20 ± 22 (c ˆ 0.5 ± 1 in H2O).
Positive-mode LSI-MS (thioglycerol ‡ NaCl/thioglycerol ‡ KF): 16:
m/z: 2771/2787; 17: m/z: 3402/3418; 18 (after cleavage of the
levulinoyl group): m/z: 3934/3950; 19: m/z: 4664/4680; 20: m/z: not
determined/5310; 21: m/z: not determined/5940. The 1H NMR spectra
of 23 ± 28 were very similar in terms of chemical shifts, and, as
expected, only the relative intensities of the signals varied between
compounds. Resonances for the distinguishable anomeric protons: d
(J1,2 [Hz]) ˆ 5.08 ± 5.11 (1 ± 2; NR), 5.38 ± 5.41 (3 ± 4; NR-1), 5.04 ± 5.09
(1 ± 2; CNR), 5.40 ± 5.42 (3 ± 4; CR), 5.05 ± 5.09 (1 ± 2; R-1), 5.10 ± 5.15
(3.5; R). Optical rotations were all in the range ‡27 ± 34 (c ˆ 0.4 ± 0.6
in H2O); ESI-MS (monoisotopic mass/average mass/experimental
mass Æ standard deviation): 23: m/z: 3603.5/3606.3/3605.13 Æ 0.9; 24:
m/z: 4297.5/4300.7/4296.8 Æ 0.9; 25: m/z: 4991.4/4995.2/4993.0 Æ 2.2;
26: m/z: 5685.3/5689.6/5687.6 Æ 2.3; 27: m/z: 6379.2/6384.1/6381.4 Æ
3.2; 28: m/z: 7073.1/7078.5/7077.3 Æ 3.2.
controlling the stereochemistry of the C(2) carbon center
attached to the heteroatom (P or N).
In accordance with the definition of Mislow and Siegel,[5]
the center C(2) is not a chiral center, but rather a chirotopic
center (a center in a chiral environment). This represents a
great synthetic advantage since it avoids the necessity to
control the stereochemistry at C(2). We chose the carboxylic
acid 2 as the key intermediate for the preparation of ligands of
type 1. The esterification of commercially available (S)-3-
phenylbutyric acid (3) with isobutene provides the tert-butyl
ester 4, which was alkylated in a clean SN2-reaction with (S)-1-
bromoethylbenzene (88% ee)[6] (Scheme 1). Simple recrys-
tallization of (S,S)-tert-butyl ester 5 from pentane/acetone
[16] D. H. Atha, J.-C. Lormeau, M. Petitou, R. D. Rosenberg, J. Choay,
Biochemistry 1987, 26, 6454 ± 6461
[17] Alternatively, a tetrasaccharide donor or a hexasaccharide donor
obtained from the same basic disaccharide was used.
[18] A. N. Teien, M. Lie, Thromb. Res. 1977, 10, 399 ± 410.
Â
[19] J.-M. Herbert, J.-P. Herault, A. Bernat, R. G. M. van Amsterdam,
G. M. T. Vogel, J.-C. Lormeau, M. Petitou, D. G. Meuleman, Circ. Res.
1996, 76, 590 ± 600.
Scheme 1. a) Isobutene, H2SO4 cat., CH2Cl2, 15 !258C, 14 h, 85%;
b) LDA, THF, 788C, 1.5 h, then HMPA (1.0 equiv), (S)-PhCH(Br)CH3
(1.5 equiv), 20 !258C, 65%; c) p-TosOH cat., toluene, 1108C, 14 h,
95%. HMPA ˆ hexamethyl phosphoramide; p-TsOH ˆ para-toluenesul-
fonic acid.
New Chiral Ligands with Nonstereogenic
Chirotopic Centers for Asymmetric Synthesis**
Claus-Dieter Graf, Christophe Malan, and
Paul Knochel*
removed the 10% contamination of the meso isomers, and
provides the pure ester 5 (65% yield). The free carboxylic
acid 2 is then obtained by the acid catalysed cleavage of the
tert-butyl ester 5. This simple reaction sequence has been
performed on a 10-g scale with an overall yield of about 50%.
The carboxylic acid 2 was readily converted into the
corresponding alkyl chloride 6 by a radical decarboxylation.[7]
Thus, 2 was converted into its corresponding acid chloride
followed by treatment with the sodium salt of 2-mercapto-
pyridine-N-oxide with simultaneous photolysis (Scheme 2).
The design of new chiral ligands for asymmetric synthesis is
an active field of research.[1] To obtain an efficient transfer of
the stereochemistry the chiral information of the ligand
should be as close as possible to the reaction center. This
concept led to the development of P-chiral phosphanes[2] and
to the ingenious design of ªchiral pocketsº[3] that give
impressive enantioselectivities. Alternatively, ligands with
chiral secondary organic groups linked to phosphorus or
nitrogen atoms have also been used with excellent results in
enantioselective reactions.[4] However, many of these ligands
involve a challenging synthesis that requires the linkage of a
sterically demanding secondary chiral carbon center to a
heteroatom (P or N). Stimulated by the work of Mislow and
Siegel on local chirality[5] we report the preparation of several
new pseudo-C2-symmetric ligands of type 1 and their utility in
asymmetric synthesis. All these ligands avoid the difficulty of
[*] Prof. Dr. P. Knochel, Dipl.-Ing. C.-D. Graf, Dr. C. Malan
Fachbereich Chemie der Universität
Hans-Meerwein-Strasse, D-35032 Marburg (Germany)
Fax: (‡49)6421-28-21-89
E-mail: knochel@ps1515.chemie.uni-marburg.de
[**] We thank the DFG (SFB 260, Leibniz-program) and the Fonds der
Chemischen Industrie for generous support. We thank BASF,
BAYER, Chemetall, and SIPSY (France) for the generous gift of
chemicals.
Scheme 2. a) SOCl2, 908C, 3 h; b) NaC5H4NOS, DMAP cat., CCl4, 808C,
hn (300 W), 2 h; c) lithium 4,4''-di-tert-butylbiphenylide, THF, 788C,
5 min; d) (PhS)2, THF,
78 !258C; e) MePCl2, THF,
78 ! 258C;
f) BH3 ´ SMe2. DMAP ˆ 4-dimethylaminopyridine.
3014
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3721-3014 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 21