Synthesis of tailor-made glycoconjugate mimetics of heparan sulfate that bind IFN-γ in the nanomolar range

FULL PAPER

Synthesis of Tailor-Made Glycoconjugate Mimetics of Heparan Sulfate That

Bind IFN-g in the Nanomolar Range**

Andrÿ Lubineau,^[a]Hugues Lortat-Jacob,^[b]Ollivier Gavard,^{[a, c]}Stÿphane Sarrazin,^[b]and

David Bonnaffÿ*^[a]

Abstract: We have recently described

the preparation of three building

blocks for the combinatorial synthesis

of heparan sulfate (HS) fragments.

Herein we show that one of these

building blocks (disaccharide 4) allows

the preparation, in high yields and with

total a stereoselectivity, of tetra-, hexa-

and octasaccharides from the heparin

(HP) regular region, by using 2+2, 2+

4 and 4+4 glycosylation strategies, re-

spectively. These oligosaccharides were

processed into sulfated derivatives

bearing an allyl moiety in the anomeric

position. The UV-promoted conjuga-

tion of these compounds with a,w-bis-

(thio)poly(ethylene glycol) spacers of

three different lengths allowed us to

prepare nine benzylated glycoconju-

gates. After final deprotection, the gly-

coconjugates 1a c, 2a c and 3a c were

obtained and their ability to inhibit the

interaction between IFN-g and HP was

tested by using surface plasmon reso-

nance detection. Compound 3b, con-

taining two HP octasaccharides linked

by a 50-ä linker was able to inhibit the

IFN-g/HP interaction with an IC₅₀

value of approximately 35 nm. In addi-

tion, the nine glycoconjugates were

perfect tools in the study to ascertain

the topology of the IFN-g binding site

on HS. Compounds 1a c, 2a c and

3a c, by mimicking the alternating sul-

fated and nonsulfated regions found in

HS, thus comprise the first example of

a library of synthetic HS mimetics

giving access to the ™second level of

molecular diversity∫ found in HS.

Keywords: heparin ¥ combinatorial

chemistry ¥ cytokines ¥ glycosyla-

tion ¥ oligosaccharides

Introduction

cellular matrix, interact and regulate the activity of numer-

ous proteins, such as growth factors, cytokines, chemokines,

viral proteins or coagulation factors.^[1,2]HS is one of the

most heterogeneous biopolymers since the uronic acid may

have either the d-gluco or l-ido configuration and various

sulfation patterns (sulfoforms) may occur along the

Heparan sulfate (HS) is a member of the glycosaminoglycan

(GAG) family and close in structure to heparin (HP), which

is clinically used for its antithrombotic activity. It is a linear

sulfated polysaccharide whose basic disaccharide unit is

composed of a uronic acid 1,4-linked to a 2-deoxy-2-amino-

glucose. HS chains, either at the cell surface or in the extra-

5]

chain.^[3There is growing evidence that the formation of

different HS structures is tightly controlled during biosyn-

thesis, with the presumed goal of generating sequences with

biological specificity.^[6,7]Indeed, the anti-factor-Xa property

of heparin is linked to the presence of a pentasaccharide in

the HP chain with a precise uronic acid content and sulfa-

tion patterns that are able to bind with high affinity and spe-

cificity to antithrombin III (AT III).^[8,9]The chemical synthe-

sis of this pentasaccharide has led to the development of

Arixtra, a Food and Drug Administration (FDA) and Euro-

pean Medicines Agency (EMEA) approved drug against

deep-vein thrombosis.^[10]

[a] Prof. A. Lubineau, Dr. O. Gavard, Prof. D. Bonnaffÿ

Laboratoire de Chimie Organique Multifonctionnelle

UMR CNRS-UPS 8614 ™Glycochimie Molÿculaire∫

Bat. 420, Universitÿ Paris Sud, 91405 Orsay Cedex (France)

Fax : (+33)169-154-715

E-mail: david.bonnaffÿ@icmo.u-psud.fr

[b] Dr. H. Lortat-Jacob, S. Sarrazin

Laboratoire d×Enzymologie Molÿculaire

Institut de Biologie Structurale, UMR CNRS-CEA-UJF 5075

41rue Horowitz, 38027 Grenoble Cedex 01(France)

Impressive improvements in the synthesis of HS frag-

ments have been made in the last decade and have allowed

HS fragments with a large structural variety to be synthe-

[c] Dr. O. Gavard

Current address:

20]

sised.^[10In this regard, we have shown that combinatorial

Cerise Chemtech, rue de Strasbourg 5, 1140 Bruxelles (Belgium)

synthesis is an ideal tool to generate all the sulfoforms of

the basic disaccharide of chondroitin sulfate (another

GAG).^[21]More recently we have published the synthesis of

[**] IFN-g=g-interferon.

Supporting information for this article is available on the WWW

under http://www.chemeurj.org/ or from the author.

Chem. Eur. J. 2004, 10, 4265 4282

DOI: 10.1002/chem.200306063

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4265

FULL PAPER

protected disaccharide building blocks, including compound

4,^[22,23]which are the basis of a combinatorial approach

toward the synthesis of HS fragment libraries.^[24]However,

tight and specific binding between HS and a given protein is

not solely regulated by these microheterogeneities, which

represent only the first level of molecular diversity in HS

chains (Figure 1). In fact, the polymer, typically composed

on-activation normally T-cell expressed and secreted pro-

teins (RANTES) (9 68)^[29]or macrophage inflamatory pro-

tein (MIP) 1a,^[30]a single NS domain is not sufficient for

binding and a longer fragment, including an NA domain, is

needed for an efficient interaction. In fact, the proteins cited

above are either multimeric in solution or multimerised

upon binding to HS and the requirement for two distant NS

domains reflects the fact that at least two basic domains of

different subunits have to interact with the HS polymer to

reach high binding constants. Recently, it has been proposed

that the HS chain may adapt its conformation, and thus the

distance between NS domains, in order to meet the needs of

recognition of a protein (Figure 2a).^[31]However, this will

Figure 1. Dual molecular diversity in HS chains. a) Microheterogeneities

resulting from the various uronic acid and sulfation motifs generate a

first level of molecular diversity. b) The charge distribution due to the al-

ternation of NS and NA NA/NS domains of variable length leads to a

second level of molecular diversity.

Figure 2. Model of binding on the HS polymer for a protein needing

more than a single NS domain. a) Accommodation of the HS polymer

conformation to meet the needs of an HS binding protein. b) Two NS do-

mains linked by a spacer of optimum length should be a functional mi-

metic of the HS binding site of the protein.

of 50 200 disaccharide units (25 100 kD), is not fully hetero-

geneous. Quite regular N-acetylated regions (NA domains),

mainly composed of d-glucuronic acid and N-acetylated glu-

cosamine, and thus with low global charge, separate domains

rich in l-iduronic acid and N-sulfated glucosamine (NS do-

mains), which are hypervariable and highly charged (Fig-

ure 1a). The latter are typically three eight disaccharides

long, while the NA domains are more regular and encom-

pass a larger area, around 15 disaccharides in length.^[25]In

between, mixed NA/NS regions of variable length make the

transition between the NA and NS domains. Thus, in addi-

tion to the first level of molecular diversity discussed above,

HS presents a second level of diversity, since the various

combinations of NS and NA NA/NS domains generate mul-

tiple charge distribution along the polymer chain (Fig-

ure 1b).

The primary interaction between HS and a protein is an

attraction between the highly negatively charged NS do-

mains and clusters of basic residues at the protein surface,

mainly arginines and lysines. In some cases, for example,

with AT-III,^[8,9]fibroblast growth factors (FGFs)^[1]or stromal

cell-derived factor (SDF-1),^[26]a single NS domain is suffi-

cient to allow a high-affinity interaction, the specificity of

which is then linked to the uronic acid and sulfation pattern

of the NS domain. However, with other proteins, such as g-

interferon (IFN-g),^[27]platelet factor 4 (PF-4),^[28]regulated-

have an enthalpic and entropic cost and the binding of do-

mains with the correct preformed geometry will thus be fa-

voured. In this regard, HS glycoconjugates, in which a linker

of an appropriate length separates two NS domains, should

be a selective functional mimetic of the binding site of a

given protein for HS (Figure 2b). Curiously, it is in the HP

field that this approach has been tested, although HP is

more rigid than HS and supports much less conformational

accommodation.^[31]HP mimetics containing a thrombin and

34]

an AT III binding site, linked by a carbohydrate^[32

or a

noncarbohydrate^[35,36]spacer, have been prepared. These

compounds were the first synthetic carbohydrates or mimet-

ics displaying anti-factor-IIa and anti-factor-Xa properties.

Moreover, compounds with the full anticoagulant properties

of heparin but unable to bind PF-4 and thus devoid of one

of the most severe side effects of HP were prepared by

using this approach.^[33]More recently, the ™head-to-head∫

cross-linking of natural HP fragments, with an ethylenedi-

amine linker, has allowed the preparation of mimetics able

to bind RANTES with a greater affinity than a natural HP

fragment with an equivalent length to the mimetic.^[29]

4266

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

Chem. Eur. J. 2004, 10, 4265 4282

Glycoconjugate Mimetics of Heparan Sulfate

4265 4282

Herein we describe a strategy that opens up easy access

to libraries of HS functional mimetics by combining HS syn-

thetic fragments and a,w-bis(thio)poly(ethylene glycol)

spacers of different lengths (Figure 3b). To our knowledge,

this is the first time that a methodology has been developed

to address the question of the second level of molecular di-

versity found in HS chains. This project was started some

years ago in order to prepare glycoconjugates mimicking HS

and able to bind IFN-g with high affinity and specificity.^[37]

IFN-g is a Th-1cytokine mainly secreted by NK and T cells.

It was originally discovered on the basis of its antiviral activ-

ity but was later found to play a key role in the immune re-

sponse and inflammatory processes. The binding of IFNg to

HS was recognised some years ago^[38]and was found to con-

trol the blood clearance^[39]and the subsequent tissue target-

ing, accumulation and localisation of the cytokine.^[40]Thus,

the discovery of compounds able to modulate the activity of

IFN-g could open the way to new immunotherapies,^[41]espe-

cially in the field of Crohn×s disease and ankylosing spondyl-

arthropathies.^[42]

IFN-g is a C₂-symmetric homodimer in solution and binds

to HS by virtue of basic residues located at the C terminus

of the two subunits (Figure 3a).^[43]A fragment of HS poly-

mer, retaining the activity of the full-length polysaccharide,

has been isolated by using a protection approach (an enzy-

matic digestion of HS polymer in the presence of IFN-g).^[27]

The biochemical analysis of this fragment revealed that it

was composed of two small hexa- to octasaccharide NS do-

mains, separated by a more extended NA domain of

15 16 disaccharides (130 140 ä;^[44]Figure 3a). A model of

the interaction between IFN-g and HS was thus proposed in

which the KRKR domains,^[45]located at the C terminus of

each subunit of the IFN-g dimer, would interact with the

highly charged NS domains of the fragment.^[27,39]However,

Figure 3. a) Molecular organisation of the HS binding site of IFN-g. b) Glycoconjugates 1 3, prepared by condensation of NS synthetic fragments bearing

an allyl group in the anomeric position on bis(thio)PEGs, as functional mimetics of the IFN-g binding site on HS.

Chem. Eur. J. 2004, 10, 4265 4282

www.chemeurj.org

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4267

D. Bonnaffÿ et al.

FULL PAPER

X-ray diffraction crystal structures of C-termini-truncated

dimers of IFN-g,^[46,47]reveal a distance of around 23 ä be-

tween the two Ala123 residues of the IFN-g subunits; this

distance is much shorter than the NA domain length (Fig-

ure 3a). This indicates that in this interaction the IFN-g

dimer has to adapt the HS chain conformation to its needs.

To comfirm this hypothesis and to obtain compounds able

to bind selectively and with high affinity to IFN-g, we decid-

ed to prepare glycoconjugates in which two NS domains,

ranging from tetra- to octasaccharides, would be linked by

their reducing end to a spacer of the poly(ethylene glycol)

(PEG) type (Figure 3b). In a model study toward this goal,

we have shown that a,w-bis(thio)PEGs add in high yields to

allyl anomeric derivatives to give such C₂-symmetric glyco-

conjugates.^[37]We chose PEG linkers for their water solubili-

ty and their availability in dif-

moved by hydrogenolysis with the benzyl ethers in the last

step of the synthesis or, selectively, by oxidation or acidic

treatment to furnish a disaccharide acceptor for chain elon-

gation. The anomeric positions are protected as an allyl gly-

coside that can be removed to give, after activation of the

anomeric position, a disaccharide donor.

Synthesis of tetrasaccharide 7, hexasaccharide 10 and octa-

saccharide 11: A 2+2, 2+4 or 4+4 strategy was planned

for the preparation of tetrasaccharide 7, hexasaccharide 10

and octasaccharide 11; this strategy required that the para-

methoxybenzyl group and the allyl group of disaccharide 4

and tetrasaccharide 7 could be removed orthogonally with-

out affecting the other protecting groups. This was indeed

done readily by using standard procedures (Scheme 1). The

ferent lengths. In this regard,

we have shown that, in the syn-

thesis of the a,w-bis(thio)PEG

intermediate, a,w-bis(bromo)-

PEGs may be purified to near

homogeneity by silica gel chro-

matography.^[37]The use of

PEGs as linkers has been criti-

cised since their flexibility leads

to lower binding constants than

with more rigid spacers, due to

the entropic cost needed for

their organisation. However,

such flexibility may be impor-

tant to mimic NA domains

whose conformational plasticity

is thought to be important for

their biological role.^[31]More- Scheme 1. a) DDQ, CH₂Cl₂, room temperature, 3 h, 83%; b) 1. H₂-activated [Ir^I(C₈H₁₄)(MePh₂P)₂]PF₆, THF,

room temperature, 2 h; 2. HgO/HgCl₂, acetone/H₂O (9:1), room temperature, 2 h; 3. Cl₃CCN, K₂CO₃, CH₂Cl₂,

over, PEGs are essentially non-

immunogenic and their incor-

poration into a drug is not a

room temperature, 3 h, 88% (two steps); c) TBDMSOTf, CH₂Cl₂, ꢀ40!08C, 90%. All=allyl, Bn=benzyl,

DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone, pMBn=para-methoxybenzyl, TBDMSOTf=tert-butyldi-

methylsilyl trifluoromethanesulfonate, THF=tetrahydrofuran.

problem from the pharmacolog-

ical point of view.

para-methoxybenzyl group in the 4’ position of the known

disaccharide 4^[22,23]was cleaved with DDQ, in wet CH₂Cl₂,

Results and Discussion

to give the disaccharide acceptor 5^[22]in 83% yield. For the

deallylation of compound 4, PdCl₂/AcONa in aqueous

AcOH^[49]was tested first but gave only moderate isolated

yields of the expected hemiacetal (60%). An isomerisation

of the allyl group by using hydrogen-activated [Ir^IC₈H₁₄-

(MePh₂P)₂]PF₆,^[50]followed by mercuric salt promoted cleav-

age of the resulting anomeric 2-propenyl, gave better re-

sults.^[51]The resulting hemiacetal was then treated with tri-

chloroacetonitrile and K₂CO₃in CH₂Cl₂to give the imidate

6 as a 60:40 a/b mixture, in an overall yield of 88% from 4.

The b anomer of the disaccharide imidate 6 has already

been prepared, by using a different protecting group strat-

egy. However, it was only used as capping building block at

the end of an elongation protocol in order to avoid sulfata-

tion of the nonreducing-end 4-OH group of the prepared

oligosaccharides.^[48]

The preparation of the glycoconjugates 1 3, relies upon the

oligomerisation and functionalisation of the key disaccha-

ride building block 4, for which we have developed efficient,

multigram syntheses.^[22,23]These compounds already contain

the l-iduronic moiety protected as methyl esters, in order to

avoid oxidation steps that could be troublesome on elabo-

rated material at the end of the synthesis. We opted for a

protective-group pattern allowing the same building block

to be transformed into a glycosyl donor or acceptor and

making clear distinction between the positions to be sulfated

and those remaining as free hydroxy groups. Thus, acetates

protect the positions to be sulfated while benzyl ethers are

used as permanent protection. A para-methoxybenzyl group

at the 4’ position avoids the need for the preparation of a

special building block for capping the nonreducing end at

the end of the elongation process.^[19,48]This group will be re-

The condensation of the disaccharide acceptor 5 with the

disaccharide donor 6, in CH₂Cl₂at ꢀ408C and with

4268

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

Chem. Eur. J. 2004, 10, 4265 4282

Glycoconjugate Mimetics of Heparan Sulfate

4265 4282

TBDMSOTf as a promotor, gave the desired tetrasaccharide

7 in excellent yield (90%) and total a stereoselectivity,^[52]as

expected for a donor with a nonparticipating group on the

C-2 position (Scheme 1). Such a high stereoselectivity is a

general tendency when 2-azidoglucose trichloroacetimidate

mono- or oligosaccharide donors are condensed onto the 4-

OH group of l-iduronyl acceptors.^{[10,16,19,48,53]}This has been

attributed previously to the conformations adopted by the

2n+2m oligomerisation strategy, thereby alleviating the

need for a special building block to cap the nonreducing

end.

Preparation of the sulfated building blocks 21, 22 and 23:

We have previously shown that the condensation of a,w-

bis(thio)PEGs onto a-allyl glucosaminyl derivatives pro-

ceeds better in water than in an organic solvent.^[37]Thus, the

coupling reaction has to be performed with water-soluble O-

and N-sulfated oligosaccharides. Since the complete func-

tionalisation and deprotection of the oligosaccharides 7, 10

and 11 would result in the reduction of the allyl moiety

during the last hydrogenolysis step, we chose to condense

the a,w-bis(thio)PEGs onto the benzylated, but water-solu-

ble, oligosaccharides 21 23. To this end, the oligosaccharides

7, 10 and 11 were first deacetylated in near quantitative

yields by using K₂CO₃in MeOH (Scheme 4). The next step

was to perform the reduction of the azido group without re-

ducing the anomeric allyl moiety. The Staudinger reac-

tion,^[56]which worked well in model reactions performed on

disaccharide 4, gave a mixture of products when used on

larger oligosaccharides. A catalytic reduction, with hydrogen

and Lindlar catalyst in the presence of quinoline,^[57]was

tried next but led to concomitant reduction of the allyl

group. SmI₂has been described as an efficient and selective

reagent for the reduction of azides^[58]but, in our case, treat-

ment of tetrasaccharide 7 with a solution of SmI₂led to a

mixture of products. However, 1,3-dithiopropane^[59]allowed

a clean reduction of the azido group in oligosaccharides 12

14 without any reduction or thiol addition on the allyl

moiety when performed in the absence of light. Thus, the

free-amino-containing tetrasaccharide 15, hexasaccharide 16

and octasaccharide 17 were obtained in 92, 87 and 84%

yields, respectively. The sulfation of both amino and hydroxy

functions in compounds 15 17 was performed by using stan-

dard conditions. Thus, after treatment with the pyridine¥SO₃

complex in pyridine, the sulfated oligosaccharides 18 20

were obtained in 73 96% yield. Saponification of the

methyl esters was performed by using 2m LiOH in 18%

H₂O₂. The reaction on tetrasaccharide 18 proceeds smoothly

at room temperature and went to completion within 24 h.

With hexasaccharide 19 and octasaccharide 20, the reaction

did not go to completion within 24 h at room temperature.

The temperature was thus

1

l-iduronyl ring, which is in equilibrium between the C₄and

²S₀conformations, allowing the C-4 hydroxy group to partly

occupy an axial position.^[16]Indeed, in HP-fragment synthe-

sis, lower diastereoselectivities have been observed mostly

when 2-azidoglucose trichloroacetimidate donors were con-

4

densed onto d-glucuronyl acceptors that adopt the C₁con-

formation.^[16,24,54]Tetrasaccharide 7 was then converted into

tetrasaccharide acceptor 8 and trichloroacetimidate 9 in

81% and 87% yields, respectively, by using the same proto-

cols as described above (Scheme 2). The condensation of

Scheme 2. a) DDQ, CH₂Cl₂, room temperature, 3 h, 81%; b) 1. H ₂-acti-

vated [Ir^I(C₈H₁₄)(MePh₂P)₂]PF₆, THF, room temperature, 2 h; 2. HgO/

HgCl₂, acetone/H₂O (9:1), room temperature, 2 h; 3. Cl₃CCN, K₂CO₃,

CH₂Cl₂, room temperature, 3 h, 87% (two steps).

disaccharide and tetrasaccharide donors 6 and 9 with the

tetrasaccharide acceptor 8, under the same conditions as for

the preparation of tetrasaccharide 7, gave hexasaccharide 10

(68%) and octasaccharide 11 (93%) with total a stereose-

lectivity^[55](Scheme 3). This demonstrates that the single di-

saccharide 4 may be efficiently used as building block in a

raised to 378C to get reason-

able reaction rates. Under these

conditions, the saponification of

compounds 19 and 20 was com-

plete in 24 h without detectable

epimerisation, b elimination or

fragmentation. After purifica-

tion by C-18 reversed-phase

chromatography, the desired

tetra-, hexa- and octasaccha-

rides 21, 22 and 23 were ob-

tained in 73, 91and 80% yields,

respectively.

Scheme 3. a) TBDMSOTf, CH₂Cl₂, ꢀ40!08C, 68 % (10), 93% (11).

Chem. Eur. J. 2004, 10, 4265 4282

www.chemeurj.org

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4269

D. Bonnaffÿ et al.

FULL PAPER

Scheme 4. a) K₂CO₃, MeOH, room temperature, 93% (12), 88% (13), 91% (14); b) HS(CH₂)₃SH, NEt₃, MeOH, room temperature, 92% (15), 87%

(16), 84% (17); c) pyridine¥SO₃, pyridine, 24 h, room temperature, then 24 h, 558C, 73% (18), 95% (19), 96% (20); d) LiOH, H₂O₂, 73% (21), 91%

(22), 80% (23); e) Pd(OH)₂/C, H₂, phosphate buffer (pH 7.0), 48 h, room temperature, 94% (24), 85% (25), quant. (26).

Preparation of glycoconjugates 1a c, 2a c and 3a c: For

the syntheses of glycoconjugates 1a c, 2a c and 3a c,

almost homogeneous a,w-bis(thio)PEGs 27 (m=5) and 28

(m=10) were prepared from commercial PEG-300 and

PEG-600, and a more polydisperse sample 29 (m=32) was

prepared from PEG-1500.^[37]These PEG lengths were

chosen because, after conjugation with the allyl group, they

should give linkers of different lengths (33, 50 and 114 ä),

thereby allowing study of the effect of the distance between

the sulfated regions on the biological activity of the glyco-

conjugates. The first two linkers should restrict the access of

the NS domains to the first basic region of the IFN-g C ter-

minus (KRKR domain), while the longer linker should

allow their access to the furthest basic region (RGRR

domain; Figure 3a). The conditions that we had optimised

in our preliminary studies^[37]were first applied to the prepa-

ration of the nine targeted glycoconjugates 30a c, 31a c

and 32a c. A mixture of tetrasaccharide 21 and a,w-bis-

(thio)PEGs 27 was thus irradiated, in a quartz vessel, with

the light of a medium-pressure Hg lamp (Scheme 5). How-

ever, the results were disappointing since HPLC analyses of

the crude reaction mixture indicated that the major product

was only formed in 40%. Benzyl protecting groups have

been reported recently to be reactive under short-wave UV

irradiation.^[60]The discrepancies between our model experi-

ments and those performed with compounds 21 could, there-

fore, originate from the benzyl moieties. The coupling reac-

tions were thus performed by using a filter with maximum

transmittance at 360 nm. Under these conditions, the pro-

portion of the desired glycoconjugates 30 32 rose to 70%.

Glycoconjugates 30a, 30c, 31a, 31c and 32a c were then

purified by semipreparative RP-18 HPLC because we were

not certain that the side products only included compounds

containing changes in the number of benzyl groups. This

yielded the conjugates in yields of 34 66% and purities

ranging from 92 98%^[61](Scheme 5). These compounds

were unambiguously characterised by ESI-MS mass analyses

1

and H NMR spectroscopy. In addition, HMQC experiments

were performed on compounds 30a, 31a and 32b. In our in-

itial synthetic scheme the last step in the synthesis should

have been the hydrogenolysis of the permanent benzyl pro-

tecting groups. However, due to the presence of the sulfide

linkage this reaction failed. The thioether functions were

thus oxidised to sulfones by using oxone,^[62]to give the ex-

pected glycoconjugates 33a, 33c, 34a, 34c and 35a c

(Scheme 5), unfortunately, as a mixture. Indeed products re-

sulting from partial cleavage of the para-methoxybenzyl

group were obtained, as shown by HPLC and ¹H NMR anal-

yses. Compounds 30b and 31b were not purified at the thio-

ether stage, but the coupling reaction mixtures were directly

oxidised with oxone, and the resulting glycoconjugates were

purified by semi-preparative RP-18 HPLC to give 33b and

33c in 21 31% yield. These low yields indicate that, along

with the side products arising from the UV irradiation, com-

pounds that had the lost para-methoxybenzyl group were re-

moved during the HPLC purification. To our knowledge,

there is no report of oxidative removal of a para-methoxy-

benzyl group with oxone. However, the abstraction of a

para-methoxybenzyl benzylic hydrogen atom by KHSO₅,

with a mechanism similar to DDQ, cannot be excluded and

4270

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

Chem. Eur. J. 2004, 10, 4265 4282

Glycoconjugate Mimetics of Heparan Sulfate

4265 4282

Scheme 5. a) hn (360 nm), 34 66%; b) KHSO₅(oxone), K₂HPO₄, (pH 7), 23 31% (combined yields for steps (a) and (b)); c) Pd(OH)₂/C, H₂, phosphate

buffer (pH 7.0), 96 h, room temperature, 62% to quant. (combined yields for steps (b) and (c)).

would explain the observed side reaction. Excess oxone

allows efficient and quick conversion of sulfide into sulfone

without detectable sulfoxide intermediates.^[62]Thus, since

partially deprotected side products should lead after hydro-

genolysis to the same product as the fully protected ones,

we decided to use this reagent in order to avoid uncomplet-

ed sulfide oxidation and formation of barely detectable sulf-

oxides. The final hydrogenolysis was thus performed directly

after desalting (PD-10) the reaction mixture of the previous

step. The oxidised glycoconjugates were thus obtained in

62% to quantitative yields^[63]and were shown to be homo-

anionic HS- and HP-derived oligosaccharides is still a

67]

matter of intensive research.^[64

The product signals are

generally low for large oligosaccharides, even in the negative

mode, since they poorly ionise. Thus, as expected, the re-

cording of MS data on the free glycoconjugates was difficult

even when using ESI techniques. A marked difference was

indeed found between the signals of the debenzylated glyco-

conjugates and the ones obtained with the benzylated glyco-

conjugates, for which ESI-MS data in accordance with the

proposed structures were obtained. However, correct ESI-

MS data were obtained at least for compounds 1b, 2b and

3b. PAGE analysis showed that synthetic tetra-, hexa- and

octasaccharides 24 26 (lanes 2, 5 and 8, respectively, in

Figure 4) have the same migration pattern and are more

pure than the corresponding oligosaccharides obtained from

natural sources (lanes 1, 4 and 7, respectively). Conjugation

to PEG linkers is clearly evidenced by the shift in migration

observed for 1b (lane 3), 2b (lane 6) and 3a c (lanes 9, 10

and 11, respectively). Moreover, the difference in the migra-

tion of the glycoconjugate 3a (33 ä linker, lane 9) and the

glycoconjugate 3b (50 ä linker, lane 10) demonstrates the

high-resolution power of this PAGE technique, since there

is only a difference of 220 gmol^ꢀ1(ꢁ4%) in molecular

weight between the two compounds.

1

geneous by PAGE analysis (Figure 4), H NMR spectrosco-

py and HSQC experiments. The MS analysis of large poly-

Figure 4. PAGE analysis of the glycoconjugates. Heparin-derived oligo-

saccharides (prepared as in ref. [25]) and glycoconjugates (1 mg each)

were run through a 30% polyacrylamide gel and stained with azure A.

Lane 1: heparin-derived tetrasaccharide; lane 2: 24; lane 3: 1b; lane 4:

heparin-derived hexasaccharide; lane 5: 25; lane 6: 2b; lane 7: heparin-

derived octasaccharide; lane 8: 26; lane 9: 3a; lane 10: 3b; lane 11: 3c.

Biological properties of the glycoconjugates 1a c, 2a c, 3a c

and 24 26: Surface plasmon resonance was used here as a

detection system to analyse the ability of the nine glycocon-

jugates, 1a c, 2a c and 3a c, and the three nonconjugated

controls, 24 26, to inhibit the binding of IFN-g to heparin.

Chem. Eur. J. 2004, 10, 4265 4282

www.chemeurj.org

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4271

D. Bonnaffÿ et al.

FULL PAPER

Injection of IFN-g (7.5 nm) over a Biacore sensor chip con-

taining streptavidin and heparin produced a binding re-

sponse of 175 resonance units (RU) at equilibrium. A re-

sponse of 5 10 RU was observed with a similar injection

over a streptavidin sensor chip, used as a blank surface (not

shown). To obtain accurate glycoconjugate concentration

data a colorimetric determination of the uronate content

was performed.^[68]The obtained values were in accordance

with the weighted values. All the experiments were per-

formed in duplicate or triplicate by using two or three

batches of glycoconjugates prepared independently (stan-

dard errors were within 8 15% of the mean). Preliminary

experiments demonstrated that among the nine glycoconju-

gates, only 3a c displayed inhibition activity (not shown).

The activities of the three octasaccharide glycoconjugates

3a c and the monomeric octasaccharide control 26 were

first compared. For that purpose, IFN-g (7.5 nm) was prein-

cubated with 150 nm 3a c and 26 and injected over the hep-

arin surface. Representative sensorgrams are shown in Fig-

ure 5A and demonstrate that the glycoconjugates 3a c

strongly inhibit the IFN-g/HP interaction, while the inhibi-

tion by the sole octasaccharide 26 is low. Moreover, there is

a dependence of the activity on the linker length. Com-

pound 3b (50-ä linker length) is the most potent inhibitor

while 3c (114-ä linker length) is the least active and 3a (33-

ä linker length) has medium activity. By using the same

binding assay, the concentration dependence of the inhibi-

tion by the most active glycoconjugate (3b) was then stud-

ied. Figure 5B shows that the activity of 3b is dose depen-

dant with an IC₅₀value of approximately 35 nm, that is,

within the range of the dissociation constant, K_d, of IFN-g

on HS.^[38]

These results are consistent with the view that the KRKR

domains of the IFN-g C termini are critically involved in the

interaction. Moreover, it seems that the 33 ä linker is too

small and needs to induce constraints in the IFN-g dimer to

allow binding, while the 114 ä linker is too long and has to

adapt to the protein. Glycoconjugate 3b, with a 50-ä linker,

fits best to the IFN-g dimer geometry and is thus the better

functional mimetic of the HS binding site of IFN-g. The op-

timal linker length is thus two to three times smaller than

the NA domain found in the natural IFN-g HS binding site.

This confirms the hypothesis that the conformation of the

natural HS binding site has to be adapted to IFN-g geome-

try in order to allow an efficient binding. The C₂symmetry

of compound 3b may also be an important factor for its ac-

tivity by fitting better with the symmetry of the IFN-g dimer

than the natural HS fragment. It is thus possible that, in

order to allow efficient binding, the NA domain of the natu-

ral fragment has to adopt a conformation allowing a more

C₂-symmetric ™head-to-head∫ presentation of the NS do-

mains. Such a factor has already been found to be important

in the binding of RANTES on HP.^[29]The preparation of

glycoconjugates with other symmetry will help to confirm

this hypothesis.

Finally, as shown in Figure 5B, our results also point out

the importance of the oligosaccharide size for binding.

Indeed, the glycoconjugates containing tetrasaccharide (1b)

or hexasaccharide (2b) NS domains, although linked with

the optimal linker (50 ä), are unable to significantly interact

with IFN-g, even at the highest concentration used (150 nm).

Compound 3b is thus a compound on which potential spe-

cific inhibitors of IFN-g activity may be constructed.

Conclusion

We have thus demonstrated that disaccharide 4 is a potent

building block for the preparation of oligosaccharides of dif-

ferent lengths deriving from the heparin regular region.

From these oligosaccharides, we have optimised a methodol-

ogy for the preparation of glycoconjugates mimicking the

HS binding site of IFN-g. We have thus found that com-

pound 3b was able to inhibit the IFN-g/HP interaction with

an IC₅₀value of approximately 35 nm. In addition, we have

shown that the nine glycoconjugates, prepared with various

oligosaccharide and PEG-linker lengths, are perfect tools to

define the topology of the IFN-g binding site on HS. How-

ever, these results give rise to new intriguing questions,

since it is entirely unclear if other linker systems, linking the

sugar moiety by the nonreducing end, longer oligosaccha-

rides or modification of the sulfation and uronic acid pat-

terns would show better properties for the exploration of

this kind of interaction. We are currently developing the

Figure 5. Inhibition of the IFNg/heparin binding by different glycoconju-

gates. A) IFNg (7.5 nm) was preincubated with 150 nm glycoconjugates

and then injected (arrow 1, t=0) over a heparin-activated surface, as de-

scribed in the experimental section. At the end of the injection phase

(arrow 2, t=240 s), the formed complexes were washed with running

buffer. The binding response (in RU) corresponding to the injection of

IFNg (a), IFNg + 26 (b), IFNg + 3c (c), IFNg + 3a (d), IFNg + 3b

(e), and plain buffer (f) was recorded as a function of time. B) IFNg

(7.5 nm) was preincubated with increasing concentrations (0 150 nm) of

~

&

*

3b ( ), 2b ( ) or 1b ( ) and injected over a heparin-activated surface as

described in the Experimental Section. The level of IFNg bound to the

heparin surface at the end of the association phase was recorded and the

results were expressed as a percentage of inhibition.

4272

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

Chem. Eur. J. 2004, 10, 4265 4282

Glycoconjugate Mimetics of Heparan Sulfate

4265 4282

IFN-g/heparin binding analysis by using surface plasmon resonance:

Binding analysis were performed on a Biacore 2000 apparatus, equilibrat-

ed with HBS-EP buffer (10 mm 2-[4-(2-hydroxyethyl)-1-piperazinyl]etha-

nesulfonic acid (HEPES), 0.15m NaCl, 3 mm ethylenediaminetetraacetate

(EDTA), 0.05% surfactant P20, pH 7.4), at 258C. Size-defined heparin

(9 kDa) was biotinylated at the reducing end and immobilised on a Bia-

core sensor chip.^[26]For this purpose, two flow cells of a Biacore F1

sensor chip were activated with a mixture (50 mL) of 0.2m 3-(3-dimethyl-

aminopropyl)-1-ethylcarbodiimide (EDC)/0.05m N-hydroxysuccinimide

(NHS) before injection of streptavidin (50 mL; 0.2 mgmL^ꢀ1in 10mm ace-

tate buffer, pH 4.2). Remaining activated groups were blocked with 1m

ethanolamine (50 mL; pH 8.5). Typically, this procedure, performed at a

rate of 5 mLmin^ꢀ1, permitted coupling of approximately 2000 2500 RU of

streptavidin. Biotinylated heparin (5 mgmL^ꢀ1) in HBS-EP buffer was

then injected at a rate of 5 mLmin^ꢀ1over one of the two surfaces to

obtain an immobilisation level of 50 60 RU. The other surface (streptavi-

din) was left as a blank surface. Flow cells were then conditioned with

several injections of 1m NaCl. For binding inhibition assays, IFN-g

(7.5 nm in HBS-EP buffer), either alone or preincubated with the oligo-

saccharides (0 150 nm; see legend of Figure 5), was injected at a rate of

50 mLmin^ꢀ1over the two surfaces (streptavidin and streptavidin/heparin)

for 4 min, after which the complexes formed were washed with buffer.

The sensor-chip surface was regenerated at a rate of 50 mLmL^ꢀ1with a 4-

min pulse of 1m NaCl.

tools needed to answer these questions and the results will

be reported in due course.

Experimental Section

General procedures: All moisture-sensitive reactions were performed

under an argon atmosphere by using oven-dried glassware. All solvents

were dried over standard drying agents^[69]and freshly distilled prior to

use. Evaporations were performed under reduced pressure. UV irradia-

tion was performed using a 150 W medium-pressure Hg lamp (TQ 150,

Heraeus) equipped with a midrange/longwave UV filter (45% transmit-

tance at 360 nm, cut-off at 275 and 475 nm); during irradiation the sam-

ples were air cooled. Reactions were monitored by TLC on glass silica

gel 60 F₂₅₄plates with detection by UV light at 254 nm and by charring

with 5% ethanolic H₂SO₄or orcinol reagent^[70]for diluted solutions.

Flash column chromatography was performed on Silica Gel 60 A.C.C.

6 35 m (SDS) or on LiChroprep RP-18 (Merck). HPLC was performed

by using a Waters Spherisorb ODS-2 5 m C18 250î4.6 mm column and

UV detection at 220 nm; HPLC purity is given assuming that the major

compound and all the impurities have the same molar absorbtion at

220 nm. The elution was performed at a rate of 1mLmin ^ꢀ1with a linear

gradient of 10 mm AcOH-NEt₃buffer (pH 7.0)/CH₃CN. Semipreparative

HPLC was performed by using a Waters Spherisorb ODS-2 5 m C18 250î

20 mm column, eluting at 20 mLmin^ꢀ1with a linear gradient of 5 mm

AcOH-NEt₃buffer (pH 7.0)/CH₃CN. Melting points were determined

with a B¸chi capillary apparatus and are uncorrected. Optical rotations

were measured on a Jasco DIP 370 digital polarimeter. NMR spectra

were recorded at room temperature with Bruker AC200, AC250,

AM250, AM360 or DRX400 spectrometers. Chemical shifts (d) are given

in parts per million (ppm) relative to an internal Me₄Si reference, solvent

signals (CDCl₃: d(¹³C)=77.0 ppm) or acetone in D₂O (d(¹H)=2.225 ppm

and d(¹³C)=30.5 ppm). The allyl group carbon atoms are identified in

Allyl (methyl 2-O-acetyl-3-O-benzyl-a-l-idopyranosyluronate)-(1!4)-O-

6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopyranoside (5): DDQ

(415 mg, 1.5 equiv) was added to a solution of disaccharide 4 (1.00 g,

1.2 mmol) in CH₂Cl₂saturated with water (15 mL). After 3 h at room

temperature, Et₂O (150 mL) was added and the resulting solution was

successively washed with ice-cold satd aq NaHCO₃solution (50 mL) and

water (2î50 mL), filtered, dried (MgSO₄) and concentrated. Flash chro-

matography of the residue (silica gel, petroleum ether/AcOEt (1:1!4:6))

gave disaccharide 5 (700 mg, 83%), whose ¹H and ¹³C NMR data were

identical to those previously reported.^[22]

(Methyl 2-O-acetyl-3-O-benzyl-4-O-(4-methoxybenzyl)-a-l-idopyranosyl-

uronate)-(1!4)-O-6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-d-glucopyra-

noside trichloroacetimidate (6): The iridium catalyst (C₈H₁₄(MePh₂P)₂Ir^I-

PF₆; 20 mg, 0.024 mmol, 1.3 mol%) was added to a solution of disaccha-

ride 4^{[22, 23]}(1.49 g, 1.8 mmol) in THF (30 mL). The mixture was degassed,

the Ir catalyst was activated with H₂as previously described,^[51]and, after

2 h at room temperature, the reaction was concentrated. HgO (565 mg,

2.16 mmol, 1.2 equiv) and HgCl₂(538 mg, 1.98 mmol, 1.1 equiv) were

then added to a solution of the residue in acetone/H₂O (9:1; 40 mL).

After 2 h stirring at room temperature, the reaction mixture was filtered

over a pad of celite 545 and concentrated. The residue was dissolved in

Et₂O (150 mL) and the resulting solution was successively washed with

10% (w/v) aq KI solution (2î100 mL), satd aq Na₂S₂O₆(50 mL) and

water (2î50 mL), filtered, dried (MgSO₄) and concentrated. Flash chro-

matography of the residue (silica gel, toluene/AcOEt (8:2!6:4)) gave

the anticipated hemiacetal (1.26 g, 90%) whose ¹H and ¹³C NMR data

were identical to those previously reported.^[48]This compound was dis-

solved in CH₂Cl₂(2.5 mL) and trichloroacetonitrile (970 mL, 9.7 mmol,

6 equiv) was added, followed by potassium carbonate (400 mg, 3.2 mmol,

2 equiv). After being stirred for 3 h at room temperature, the reaction

mixture was directly applied to the top of a flash chromatography

column filled with silica gel and eluted (toluene/AcOEt (9:1!8:2) with

0.1% NEt₃) to give imidate 6 (1.51 g, 98%) as an a/b (4:6) mixture.

Along with the previously described b anomer^[48](60%), we also ob-

tained the a anomer (40%): ¹H NMR (250 MHz, CDCl₃): d=8.73 (s,

1H; NH), 7.40 7.22 (m, 10H; Ph), 7.11 (d, J=8.5 Hz, 2H; PhOMe), 6.82

(d, J=8.5 Hz, 2H; PhOMe), 6.37 (d, J_1,2=3.5 Hz, 1H; H-1^A), 5.26 (d,

ꢀ

the following way: O-C_aH₂C_bH=C_cH₂; the two protons on C-c are iden-

tified as H-cc, for the one cis to H-b, and H-ct, for the one trans to H-b.

Saccharide units in oligosaccharides are identified alphabetically from

the reducing end. For compounds 10 17, spin systems, identified with

COSY experiments, were attibuted to a monosaccharide unit based on

¹H and ¹³C chemical shifts; for compounds 18 20, this attibution was con-

firmed by using HMBC experiments. Apodisations with Gaussian func-

tions (LB=ꢀ1to ꢀ2 Hz and GB=50%) were used, thereby allowing

measurement of coupling constants. COSY, gradient-enhanced COSY,

HMBC, HMQC and HSQC experiments were performed by recording

256 FID measurements with 1024 complex data points and using standard

Bruker programs. Prior to Fourier transformation, the data were zero

filled in the t₁dimension to 1024 points and multiplied with a nonshifted

sinebell function in both dimensions for COSY experiments; for HMBC

experiments, the data were zero filled in the t₁dimension to 2048 points

and an exponential multiplication (LB=3 Hz) in the t₁dimension and a

p/2-shifted squared sinebell function in the t₂dimension were used; for

HSQC experiments, the data were zero filled in the t₁dimension to

2048 points and multiplied with a p/3-shifted squared sinebell function in

both dimensions. MS spectra were recorded in the positive or negative

mode on a Finnigan MAT 95S spectrometer by using electrospray ionisa-

tion. IR spectra were recorded on a Fourier transformation Bruker IFS66

apparatus. Elemental analyses were performed at the CNRS (Gif sur

Yvette, France).

Polyacrylamide gel electrophoresis: PAGE analysis was performed essen-

tially as previously described.^[71]Oligosaccharides (1 mg) in 20% glycerol

were run initially through a stacking gel (8% acrylamide) at a constant

current (15 mA), then through a resolving gel (30% acrylamide) at

20 mA, until the Phenol Red marker (applied to a separate lane as the

electrophoresis marker) had reached the bottom of the gel. The discon-

tinuous buffer system used comprised 0.125m tris(hydroxymethyl)amino-

methane (Tris)/HCl (pH 6.8) in the stacking gel, 0.375m Tris/HCl

(pH 8.8) in the resolving gel and 25 mm Tris/0.192m glycine (pH 8.3) in

the tank buffer. After electrophoresis, oligosaccharide bands were stained

with 0.08% aqueous Azure A for a few minutes under constant agitation.

Excess dye was removed by washing the gel in water.

J

_1,2=4.5 Hz, 1H; H-1^B), 4.92 (d, J=10.5 Hz, 1H; CH₂Ph), 4.90 (t, J_2,3

=

_2,1=4.5 Hz, 1H; H-2^B), 4.75 4.64 (m, 3H), 4.63 (d, J=10.5 Hz, 1H;

CH₂Ph), 4.45 (d, J=11.5 Hz, 1H; CH₂PhOMe), 4.39 (dd, J_6a,6b=12.5,

_6a,5=2.0 Hz, 1H; H-6_a^A), 4.37 (d, J=11.5 Hz, 1H; CH₂PhOMe), 4.20

J

(dd,

J_6b,6a=12.5, J J_4,5=10.0, J_4,3=

_6b,5=3.5 Hz, 1H; H-6_b^A), 4.11 (dd,

9.0 Hz, 1H; H-3^Aor H-4^A), 3.98 (ddd, J_5,4=10.0, J_5,6b=3.5, J_5,6a=2.0 Hz,

1H; H-5^A), 4.02 (dd, J_3,2=10.0, J_3,4=9.0 Hz, 1H; H-3^Aor H-4^A), 3.95

3.90 (m with s at d=3.80, 5H; H-4^B, H-3^Band PhOMe), 3.69 (dd, J_2,3

=

10.0, J_2,1=3.5 Hz, 1H; H-2^A), 3.54 (s, 3H; COOMe), 2.04 (s, 3H; CH₃

OAc), 2.03 (s, 3H; CH₃OAc) ppm; ¹³C NMR (62.5 MHz, CDCl₃): d=

Chem. Eur. J. 2004, 10, 4265 4282

www.chemeurj.org

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4273

D. Bonnaffÿ et al.

FULL PAPER

ꢀ

170.5, 170.0, 169.8 (C=O), 160.3 (C=NH), 159.4 (C OMe pMBn), 137.7

(C_{quat arom}), 129.6, 129.3, 128.4, 128.1, 128.0, 127.9, 127.5 (C_arom), 113.7 (C_m

pMBn), 98.0 (C-1^B), 94.4 (C-1^A), 78.2, 75.2, 75.0, 74.7, 74.4, 73.0, 72.4,

71.8, 70.8, 70.2, 62.8 (C-2^A), 61.6 (C-6^A), 55.2 (CH₃pMBn), 51.7 (CH₃

COOMe), 20.9, 20.6 (CH₃OAc) ppm.

(brd, J_1,2=1.0 Hz, 1H; H-1^D), 4.95 (d, J_1,2=3.5 Hz, 1H; H-1^A), 4.93 (d,

J

_5,4=2.5 Hz, 1H; H-5^D), 4.92 (d, J_1,2=3.5 Hz, 1H; H-1^C), 4.93 (m, 1H; H-

2^D), 4.83 (d, J_5,4=2.5 Hz, 1H; H-5^B), 4.82 (d, J=10.5 Hz, 1H; CH₂Ph),

4.77 (d, J=11.0 Hz, 2H; CH₂Ph), 4.75 (t, J_2,3=J_2,1=3.5 Hz, 1H; H-2^B),

4.71(d, J=12.0 Hz, 1H; CH₂Ph), 4.70 (d, J=11.0 Hz, 1H; CH₂Ph), 4.69

(d, J=11.0 Hz, 1H; CH₂Ph), 4.62 (d, J=12.0 Hz, 1H; CH₂Ph), 4.62 (d,

Allyl [(methyl 2-O-acetyl-3-O-benzyl-4-O-(4-methoxybenzyl)-a-l-idopy-

ranosyluronate)-(1!4)-O-(6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-

glucopyranoside)-(1!4)-O-(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyra-

nosyluronate)]-(1!4)-O-6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glu-

copyranoside (7): Acceptor 5 (460 mg, 0.66 mmol) and imidate 6 (785 mg,

0.85 mmol, 1.3 equiv) were azeotropically dried with toluene and dis-

solved in CH₂Cl₂(2 mL). Powdered 4-ä molecular sieves (1.86 g) were

added and the mixture was then stirred for 30 min at room temperature.

The solution was cooled to ꢀ408C and TBDMSOTf (0.1m in CH₂Cl₂,

1.3 mL, 0.13 mmol, 0.2 equiv) was added. The reaction was stirred at this

temperature for 30 min and then the temperature was raised to 08C over

30 min. The reaction was then quenched with a solution of NEt₃(0.1m in

CH₂Cl₂, 400 mL, 0.4 mmol, 0.6 equiv), directly applied to the top of a

flash chromatography column filled with silica gel and eluted (CH₂Cl₂/

AcOEt/petroleum ether (85:10:5!70:25:5)) to give tetrasaccharide 7

(860 mg, 90%): [a]_D³⁰=1 7 c(=1.05, CH₂Cl₂); ¹H NMR (400 MHz,

CDCl₃): d=7.37 7.24 (m, 20H; Ph), 7.13 (d, J=8.5 Hz, 2H; Ph-OMe),

J=10.5 Hz, 1H; CH₂Ph), 4.41(brd,

(brd, J_6b,6a=12.0 Hz, 1H; H-6_b^A), 4.21(ddt,

_a,ct=1.5 Hz, 1H; H-a), 4.17 (dd, J_6b,6a=12.0, J_6b,5=2.5 Hz, 1H; H-6_b^C),

J

_6a,6b=12.0, 2H; H-6_a^A, H-6_a^C), 4.26

_gem=13.0, J_a,b=5.0, J_a,cc

J

=

J

4.05 (ddt, J_gem=13.0, J_a’,b=6.0, J_a’,cc=J_a’,ct=1.5 Hz, 1H; H-a’), 4.03 3.96

(m, 2H), 3.96 (m, 6H), 3.78 (brd, J_5,4=9.5 Hz, 1H; H-5^C), 3.73 3.58 (m,

3H), 3.48 (s, 3H; COOMe), 3.42 (s, 3H; COOMe), 3.39 (dd, J_2,3=10.0,

J

_2,1=3.5 Hz, 1H; H-2^A), 3.28 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-2^C), 2.59

(d, J=10.5 Hz, 1H; OH), 2.11 (s, 3H; CH₃OAc), 2.07 (s, 3H; CH₃

OAc), 2.06 (s, 3H; CH₃OAc), 2.05 (s, 3H; CH₃OAc) ppm; IR (thin

ꢀ

film): n˜ =3470 (O H_arom), 3088, 3064, 3033 (C H_arom), 2978, 2955, 2932,

ꢀ

2908, 2869 (C H_aliph), 2111 (N₃), 1749, 1733 (C=O), 1500, 1498, 1455,

1432, 1373, 1304, 1244 cm^ꢀ1; ESI HRMS: m/z calcd for C₆₅H₇₆N₆O₂₈Na

[M+Na]⁺: 1363.4758; found: 1363.4758.

[(methyl 2-O-acetyl-3-O-benzyl-4-O-(4-methoxybenzyl)-a-l-idopyranosyl-

uronate)-(1!4)-O-(6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopy-

ranoside)-(1!4)-O-(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyranosyluro-

nate)]-(1!4)-O-6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopyrano-

side trichloroacetimidate (9): Tetrasaccharide 7 (322 mg, 0.22 mmol) was

treated first with [Ir^IC₈H₁₄(MePh₂P)₂]PF₆(3 mg, 3.5 mmol, 1.6 mol%) in

THF (3.5 mL) and then with HgO (69 mg, 0.26 mmol, 1.2 equiv) and

HgCl₂(67 mg, 0.24 mmol, 1.1 equiv) as described for the preparation of

6. Flash chromatography (silica gel, toluene/AcOEt (70:30!55:45)) gave

the anticipated hemiacetal (298 mg, 81%) whose ¹H NMR spectrum

showed the disappearance of the allyl protons. ESI HRMS: m/z calcd for

C₇₀H₈₀N₆O₂₆Na₂[M+2Na]²⁺/2: 733.2459; found: 733.2459. This com-

pound was then dissolved in CH₂Cl₂(400 mL); trichloroacetonitrile

(125 mL, 1.25 mmol, 6 equiv) was then added, followed by potassium car-

bonate (51mg, 0.42 mmol, 2 equiv). After 1h stirring at room tempera-

ture the reaction mixture was directly applied to the top of a flash chro-

matography column filled with silica gel and eluted (toluene/AcOEt

(9:1!7:3) with 0.1% NEt₃) to give imidate 9 (338 mg, 98%) as an a/b

(4:6) mixture: ¹H NMR (250 MHz, CDCl₃): d=8.76 (s, 0.4H; NH), 8.73

(s, 0.6H; NH), 7.40 7.24 (m, 20H; Ph), 7.13 (d, J=8.5 Hz, 2H; Ph-

OMe), 6.82 (d, J=8.5 Hz, 2H; Ph-OMe), 6.39 (d, J_1,2=3.5 Hz, 0.4H; H-

1^Aa), 5.60 (d, J_1,2=8.5 Hz, 0.6H; H-1^Ab), 5.28 (d, J_1,2=5.0 Hz, 1H; H-1^D),

5.25 (d, J_1,2=3.5 Hz, 0.4H; H-1^Ba), 5.22 (d, J_1,2=3.5 Hz, 0.6H; H-1^Bb),

6.82 (d, J=8.5 Hz, 2H; Ph-OMe), 5.94 (dddd, J_b,ct=17.0, J_b,cc=10.5, J_b,a’

6.0, J_b,a=5.0 Hz, 1H; H-b), 5.36 (dq, J_ct,b=17.0, J_ct,a=J_ct,a’=J_gem=1.5 Hz,

1H; H-ct), 5.28 (d, J_1,2=5.0 Hz, 1H; H-1^B), 5.26 (dq, J_cc,b=10.5, J_cc,a

=

J

_cc,a=J_gem=1.5 Hz, 1H; H-cc), 5.24 (d, J_1,2=3.5 Hz, 1H; H-1^D), 4.94 (d,

_1,2=3.5 Hz, 1H; H-1^A), 4.93 (dd, J_2,3=4.0, J_2,1=3.5 Hz, 1H; H-2^D), 4.91

(d, J_1,2=3.5 Hz, 1H; H-1^C), 4.90 (d, J=10.5 Hz, 1H; CH₂Ph), 4.86 (dd,

J

_2,3=5.5, J_2,1=5.0 Hz, 1H; H-2^B), 4.82 (d, J=10.5 Hz, 1H; CH₂Ph), 4.73

(s, 2H; CH₂Ph), 4.69 (d, J=10.5 Hz, 1H; CH₂Ph), 4.68 (d, J_5,4=4.0 Hz,

1H; H-5^D), 4.66 (s, 2H; CH₂Ph), 4.63 (d, J=10.5 Hz, 1H; CH₂Ph), 4.59

(d, J_5,4=5.0 Hz, 1H; H-5^B), 4.46 (d, J=11.5 Hz, 1H; CH₂PhOMe), 4.42

(d, J=11.5 Hz, 1H; CH₂PhOMe), 4.40 (brd, J_6a,6b=12.0 Hz, 1H; H-6_a^A),

4.34 (dd, J_6a,6b=12.5, J_6a,5=2.0 Hz, 1H; H-6_a^C), 4.24 (dd, J_6b,6a=12.0,

J

_6b,5=3.0 Hz, 1H; H-6_b^A), 4.20 (ddt,

1.5 Hz, 1H; H-a), 4.16 (dd, J_6b,6a=12.5,J_6b,5=3.0 Hz, 1H; H-6_b^C), 4.05

J_gem=13.0, J_a,b=5.0, J_a,cc=J_a,ct=

(ddt, J_gem=13.0, J_a’,b=6.0, J_a’,cc=J_a’,ct=1.5 Hz, 1H; H-a’), 3.98 (dd, J_4,3

=

5.0, J_4,5=4.0 Hz, 1H; H-4^D), 3.95 (dd, J_4,5=10.0, J_4,3=9.0 Hz, 1H; H-4^C),

3.92 (dd, J_3,4=5.0, J_3,2=4.0 Hz, 1H; H-3^D), 3.90 3.85 (m, 3H; H-3^A, H-4^A

and H-5^A), 3.82 3.78 (m with s at d=3.80, 5H; H-4^B, H-5^Cand PhOMe),

3.76 (t, J_3,2=J_3,4=5.5 Hz, 1H; H-3^B), 3.65 (dd, J_3,2=10.0, J_3,4=9.0 Hz,

1H; H-3^C), 3.56 (s, 3H; COOMe), 3.49 (s, 3H; COOMe), 3.39 (dd, J_2,3

=

10.0, J_2,1=3.5 Hz, 1H; H-2^A), 3.29 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-2^C),

2.12 (s, 3H; CH₃OAc), 2.06 (s, 3H; CH₃OAc), 2.03 (s, 3H; CH₃OAc),

2.02 (s, 3H; CH₃OAc) ppm; ¹³C NMR (100.6 MHz, CDCl₃): d=170.6,

4.47 (d, J=11.5 Hz, 1H; CH₂PhOMe), 4.41(d,

CH₂PhOMe), 4.45 (brd, J_6a,6b=12.5, 0.4H; H-6_a^Aa), 4.44 (brd, J_6a,6b=12.5,

0.6H; H-6_a^Ab), 4.36 (brd, J_6a,6b=12.5 Hz, 1H; H-6_a^C), 4.23 (dd, J_6b,6a

J=11.5 Hz, 1H;

=

ꢀ

170.5, 169.9, 169.8, 169.7, 169.3 (C=O), 159.4 (C OMe pMBn), 137.8,

12.5, J J_6b,6a=12.5, J_6b,5=4.0 Hz,

_6b,5=4.0 Hz, 0.6H; H-6_b^Ab), 4.22 (dd,

137.7, 137.6, 137.3 (C_{quat arom}), 133.0 (C-b), 129.6, 129.2, 128.4, 128.3, 128.0,

127.8, 127.6, 127.5, 127.4 (C_arom), 118.3 (C-c), 113.7 (C_mpMBn), 97.9 (C-

1^D), 97.8 (C-1^B), 97.4 (C-1^C), 96.5 (C-1^A), 78.3 (C-3^A), 77.8 (C-3^C), 75.7

(C-4^A), 75.5 (C-4^C), 75.4 (C-3^B), 75.1(C-4 ^B), 74.9 (CH₂Ph), 74.7 (CH₂Ph),

73.7 (C-3^D), 73.5 (CH₂Ph), 73.3 (CH₂Ph), 72.8 (C-4^D), 72.5 (CH₂pMBn),

71.1 (C-5^B), 70.9 (C-2^B), 69.6 (C-5^C), 69.3 (C-5^D), 69.1(C-5 ^A), 68.8 (C-2^D),

68.7 (C-a), 63.2 (C-2^A), 62.8 (C-2^C), 62.1(C-6 ^A), 61.6 (C-6^C), 55.2 (CH₃

pMBn), 51.9 (CH₃COOMe), 51.7 (CH₃COOMe), 20.8, 20.7, 20.6 (CH₃

0.4H; H-6_b^Aa), 4.17 (dd, J_6b,6a=12.5, J_6b,5=2.0 Hz, 1H; H-6_b^C), 4.04 3.88

(m, 5H), 3.88 3.74 (m with s at d=3.79, 5H; PhOMe), 3.74 3.58 (m,

3H), 3.55 (s, 3H; COOMe), 3.50 (s, 1.8H; COOMe^Bb), 3.49 (s, 1.2H;

COOMe^Ba), 3.43 (t, J_2,3=J_2,1=8.5 Hz, 0.6H; H-2^Ab), 3.30 (dd, J_2,3=10.0,

J

_2,1=3.5 Hz, 1H; H-2^C), 2.11 (s, 1.8H; CH₃OAc^b), 2.10 (s, 1.2H; CH₃

OAc^a), 2.06 (s, 3H; CH₃OAc), 2.04 (s, 6H; CH₃OAc) ppm.

Allyl [(methyl 2-O-acetyl-3-O-benzyl-4-O-(4-methoxybenzyl)-a-l-idopy-

ranosyluronate)-(1!4)-O-(6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-

glucopyranoside)-(1!4)-O-(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyra-

nosyluronate)-(1!4)-O-(6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glu-

copyranoside)-(1!4)-O-(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyranosyl-

uronate)]-(1!4)-O-6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopy-

ranoside (10): Acceptor 8 (150 mg, 0.11 mmol) and imidate 6 (135 mg,

0.15 mmol, 1.3 equiv) were azeotropically dried with toluene and dis-

solved in CH₂Cl₂(360 mL). Powdered 4-ä molecular sieves (330 mg)

were added and the mixture was then stirred for 30 min at room temper-

ature. The solution was cooled to ꢀ408C and TBDMSOTf (0.1m in

CH₂Cl₂, 220 mL, 0.022 mmol, 0.2 equiv) was added. The reaction was stir-

red at this temperature for 30 min and then the temperature was raised

to 08C over 30 min. The reaction was then quenched with a solution of

NEt₃(0.1m in CH₂Cl₂, 63 mL, 0.063 mmol, 0.6 equiv), directly applied to

the top of a flash chromatography column filled with silica gel and eluted

(CH₂Cl₂/AcOEt/petroleum ether (85:10:5!70:25:5)) to give hexasac-

charide 10 (210 mg, 68%): [a]³_D⁰=1 7 c(=1.50, CH₂Cl₂); ¹H NMR

(400 MHz, CDCl₃): d=7.40 7.23 (m, 30H; Ph), 7.13 (d, J=8.5 Hz, 2H;

ꢀ

OAc) ppm; IR (thin film): n˜ =3083, 3058, 3024 (C H_arom), 2949, 2933,

ꢀ

2866, 2833 (C H_aliph), 2109 (N₃), 1740 (C=O), 1613, 1514, 1455, 1438,

1372, 1295, 1245 cm^ꢀ1; elemental analysis: calcd (%) for C₇₃H₈₄N₆O₂₈

(1461.5 gmol^ꢀ1): C 59.99, H 5.79, N 5.75, O 28.46; found: C 60.22, H 5.76,

N 5.66, O 28.16.

Allyl [(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyranosyluronate)-(1!4)-

O-(6-O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopyranoside)-(1!4)-

O-(methyl 2-O-acetyl-3-O-benzyl-a-l-idopyranosyluronate)]-(1!4)-O-6-

O-acetyl-2-azido-3-O-benzyl-2-deoxy-a-d-glucopyranoside (8): Tetrasac-

charide

7 (465 mg, 0.32 mmol) was treated with DDQ (144 mg,

0.63 mmol, 2 equiv) in wet CH₂Cl₂(10 mL) as described for the prepara-

tion of 5. Flash chromatography (silica gel, toluene/AcOEt (8:2!6:4))

gave compound 8 (370 mg, 81%) along with unreacted starting material

(44 mg, 9%): ¹H NMR (250 MHz, CDCl₃): d=7.40 7.25 (m, 20H; Ph),

5.95 (dddd, J_b,ct=17.0, J_b,cc=10.5, J_b,a’=6.0, J_b,a=5.0 Hz, 1H; H-b), 5.36

(dq, J_ct,b=17.0, J_ct,a=J_ct,a’=J_gem=1.5 Hz, 1H; H-ct), 5.27 (dq, J_cc,b=10.5,

J

_cc,a=J_cc,a=J_gem=1.5 Hz, 1H; H-cc), 5.20 (d, J_1,2=3.5 Hz, 1H; H-1^B), 5.06

4274

¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

Chem. Eur. J. 2004, 10, 4265 4282

Glycoconjugate Mimetics of Heparan Sulfate

4265 4282

Ph-OMe), 6.82 (d, J=8.5 Hz, 2H; Ph-OMe), 5.94 (dddd, J_b,ct=17.0,

6.83 (d, J=8.5 Hz, 2H; Ph-OMe), 5.94 (dddd, J_b,ct=17.0, J_b,cc=10.5, J_b,a’=

J

_b,cc=10.5, J_b,a’=6.0, J_b,a=5.0 Hz, 1H; H-b), 5.36 (dq, J_ct,b=17.0, J_ct,a

=

6.0, J_b,a=5.0 Hz, 1H; H-b), 5.36 (dq, J_ct,b=17.0, J_ct,a=J_ct,a’=J_gem=1.5 Hz,

1H; H-ct), 5.31 (d, J_1,2=5.0 Hz, 1H; H-1^Dor H-1^F), 5.30 (d, J_1,2=6.0 Hz,

2H; H-1^B, H-1^Dor H-1^F), 5.26 (dq, J_cc,b=10.5, J_cc,a=J_cc,a=J_gem=1.5 Hz,

1H; H-cc), 5.24 (d, J_1,2=3.5 Hz, 1H; H-1 ^H), 4.98 (d, J_1,2=3.5 Hz, 1H; H-

1^E(C)), 4.97 (d, J_1,2=3.5 Hz, 1H; H-1^C(E)), 4.94 (d, J_1,2=3.5 Hz, 1H; H-1^A),

4.93 (t, J_2,3=J_2,1=3.5 Hz, 1H; H-2^H), 4.93 (d, J_1,2=3.5 Hz, 1H; H-1^E),

_ct,a’=J_gem=1.5 Hz, 1H; H-ct), 5.30 (d, J_1,2=5.0 Hz, 1H; H-1^D), 5.29 (d,

_1,2=5.0 Hz, 1H; H-1^B), 5.27 (dq, J_cc,b=10.5, J_cc,a=J_cc,a=J_gem=1.5 Hz,

1H; H-cc), 5.24 (d, J_1,2=3.5 Hz, 1H; H-1^F), 4.96 (d, J_1,2=3.5 Hz, 1H; H-

1^C), 4.94 (d, J_1,2=3.5 Hz, 1H; H-1^A), 4.93 (t, J_2,3=J_2,1=3.5 Hz, 1H; H-2^F),

4.93 (d, J_1,2=3.5 Hz, 1H; H-1^E), 4.90 (dd, J_2,3=6.0, J_2,1=5.0 Hz, 1H; H-

2^D), 4.90 (d, J=10.5 Hz, 1H; CH₂Ph), 4.87 (t, J_2,3=5.0 Hz, 1H; H-2^B),

4.84 (d, J=10.5 Hz, 1H; CH₂Ph), 4.82 (d, J=10.5 Hz, 1H; CH₂Ph), 4.77

(d, J=11.5 Hz, 1H; CH₂Ph), 4.76 (d, J=11.0 Hz, 1H; CH₂Ph), 4.72 (d,

J=11.0 Hz, 1H; CH₂Ph), 4.70 (d, J_5,4=4.0 Hz, 1H; H-5^F), 4.69 (d, J=

10.5 Hz, 1H; CH₂Ph), 4.69 (d, J=11.5 Hz, 1H; CH₂Ph), 4.69 (d, J=

12.0 Hz, 1H; CH₂Ph), 4.65 (d, J=10.5 Hz, 1H; CH₂Ph), 4.65 (d, J=

4.91(d, J=10.0 Hz, 1H; CH₂Ph), 4.90 (m, 2H; H-2^D, H-2^F), 4.88 (t, J_2,3

=

6.0 Hz, 1H; H-2^B), 4.86 (d, J=10.0 Hz, 1H; CH₂Ph), 4.85 (d, J=10.0 Hz,

1H; CH₂Ph), 4.83 (d, J=11.0 Hz, 1H; CH₂Ph), 4.78 (d, J=11.0 Hz, 1H;

CH₂Ph), 4.76 (d, J=11.0 Hz, 1H; CH₂Ph), 4.74 (s, 2H; CH₂Ph), 4.73 (d,

J

_5,4=4.0 Hz, 1H; H-5^H), 4.70 (brd, J=11.0 Hz, 3H; 3îCH₂Ph), 4.69 (d,

J=12.0 Hz, 1H; CH₂Ph), 4.67 (d, J=10.0 Hz, 1H; CH₂Ph), 4.66 (d, J=

10.0 Hz, 1H; CH₂Ph), 4.65 (d, J=12.0 Hz, 1H; CH₂Ph), 4.64 (d, J=

12.0 Hz, 1H; CH₂Ph), 4.64 (d, J=10.5 Hz, 1H; CH₂Ph), 4.59 (d, J_5,4

=

5.0 Hz, 1H; H-5^B), 4.56 (d, J_5,4=4.5 Hz, 1H; H-5^D), 4.47 (d, J=11.5 Hz,

1H; CH₂PhOMe), 4.43 (d, J=11.5 Hz, 1H; CH₂PhOMe), 4.40 (brd,

10.0 Hz, 1H; CH₂Ph), 4.59 (d, J_5,4=5.0 Hz, 1H; H-5^B), 4.58 (d, J_5,4

=

5.0 Hz, 1H; H-5^Dor H-5^F), 4.56 (d, J_5,4=5.0 Hz, 1H; H-5^Dor H-5^F), 4.47

J

_6a,6b=11.5 Hz, 1H; H-6_a^A), 4.37 (dd, J_6a,6b=12.0, J_6a,5=2.0 Hz, 1H; H-

(d, J=11.5 Hz, 1H; CH₂PhOMe), 4.43 (d, J=11.5 Hz, 1H; CH₂PhOMe),

6_a^E), 4.34 (dd, J_6a,6b=12.5, J_6a,5=2.0 Hz, 1H; H-6_a^C), 4.25 (dd, J_6b,6a=11.5,

4.40 (d, J_6a,6b=12.0 Hz, 1H; H-6_a^A), 4.37 (brd, J_6a,6b=11.0, 2H; H-6_a

H-6_a^G), 4.34 (dd, J_6a,6b=12.0, J_6a,5=1.5 Hz, 1H; H-6_a^C(E)), 4.25 (dd, J_6b,6a

12.0, J_6b,5=2.5 Hz, 1H; H-6_b^A), 4.20 (ddt, J_gem=13.0, J_a,b=5.0, J_a,cc=J_a,ct

1.5 Hz, 1H; H-a), 4.22 4.16 (m, 2H; H-6_b^E(C), H-6_b^E), 4.17 (dd, J_6b,6a

12.0, J_6b,5=3.0 Hz, 1H; H-6_b^C(E)), 4.05 (ddt, J_gem=13.0, J_a’,b=6.0, J_a’,cc

,

=

E(C)

J

_6b,5=2.5 Hz, 1H; H-6_b^A), 4.20 (ddt,

J_gem=13.0, J_a,b=5.0, J_a,cc=J_a,ct=

1.5 Hz, 1H; H-a), 4.19 (dd, J_6b,6a=12.0, J_6b,5=2.5 Hz, 1H; H-6_b^E), 4.16

(dd, J_6b,6a=12.5, J_6b,5=2.5 Hz, 1H; H-6_b^C), 4.05 (ddt, J_gem=13.0, J_a’,b=6.0,

J

_a’,cc=J_a’,ct=1.5 Hz, 1H; H-a’), 3.99 (t, J_4,3=J_4,5=4.0 Hz, 1H; H-4^F), 3.98

(dd, J_4,3=5.5, J_4,5=4.5 Hz, 1H; H-4^D), 3.95 (t, J_4,5=J_4,3=9.0 Hz, 1H; H-

J

_a’,ct=1.5 Hz, 1H; H-a’), 4.20 3.97 (m, 3H; H-4^D, H-4^F, H-4^H), 3.96 (dd,

4^C), 3.92 (dd, J_3,4=5.5, J_3,2=6.0 Hz, 1H; H-3^D), 3.90 (dd, J_3,4=4.0, J_3,2

3.5 Hz, 1H; H-3^F), 3.89 3.85 (m, 3H; H-3^A, H-4^A, H-5^A), 3.87 (dd, J_4,5

=

_4,5=10.0, J_4,3=9.0 Hz, 1H; H-4^C(E)), 3.95 3.90 (m, 3H; H-3^D, H-3^F, H-

=

3^H), 3.90 3.87 (m, 3H; H-3^A, H-4^A, H-5^A), 3.88 (dd, J_4,5=10.0, J_4,3=9.0,

10.0, J_4,3=9.0 Hz, 1H; H-4^E), 3.83 (brt, J_5,4=9.0 Hz, 1H; H-5^C), 3.82 3.78

2H; H-4^E(C), H-4^G), 3.87 3.79 (m with s at d=3.80, 7H; H-4^B, H-5^C(E), H-

(m with s at d=3.80, 5H; H-4^B, H-5^E, PhOMe), 3.77 (t, J_3,2=J_3,4=5.0 Hz,

5^E(C), H-5^G, PhOMe), 3.77 (t, J_3,2=J_3,4=6.0 Hz, 1H; H-3^B), 3.67 (dd, J_3,2

=

1H; H-3^B), 3.66 (dd, J_3,2=10.0, J_3,4=9.0 Hz, 1H; H-3^E), 3.62 (dd, J_3,2

=

10.0, J_3,4=9.0 Hz, 1H; H-3^G), 3.64 (dd, J_3,2=10.0, J_3,4=9.0 Hz, 1H; H-

3^E(C)), 3.63 (dd, _3,4=9.0 Hz, 1H; H-3^C(E)), 3.57 (s, 3H;

_3,2=10.0,

COOMe), 3.54 (s, 3H; COOMe), 3.53 (s, 3H; COOMe), 3.48 (s, 3H;

COOMe), 3.39 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-2^A), 3.30 (dd, J_2,3

10.0,

J

_3,4=9.0 Hz, 1H; H-3^C), 3.57 (s, 3H; COOMe), 3.53 (s, 3H;

J

COOMe), 3.48 (s, 3H; COOMe), 3.39 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H;

H-2^A), 3.29 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-2^C), 3.27 (dd, J_2,3=10.0,

=

J

_2,1=3.5 Hz, 1H; H-2^E), 2.12 (s, 3H; CH₃OAc), 2.09 (s, 3H; CH₃OAc),

10.0, J_2,1=3.5 Hz, 1H; H-2^C(E)), 3.28 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-

2^E(C)), 3.27 (dd, J_2,3=10.0, J_2,1=3.5 Hz, 1H; H-2^G), 2.12 (s, 3H; CH₃

OAc), 2.10 (s, 3H; CH₃OAc), 2.09 (s, 3H; CH₃OAc), 2.06 (s, 3H; CH₃

OAc), 2.03 (s, 6H; CH₃OAc), 2.02 (s, 6H; CH₃OAc) ppm; ¹³C NMR

(100.6 MHz, CDCl₃): d=170.6, 170.5 (3C), 169.9 (2C), 169.7, 169.6 (2C),

2.06 (s, 3H; CH₃OAc), 2.03 (s, 3H; CH₃OAc), 2.02 (s, 3H; CH₃OAc),

2.01(s, 3H; CH OAc) ppm; ¹³C NMR (62.5 MHz, CDCl₃): d=170.6,

3

ꢀ

170.5 (2C), 169.9 (2C), 169.8, 169.7, 169.5, 169.3 (C=O), 159.5 (C OMe

pMBn), 137.8, 137.7, 137.6, 137.4, 137.3 (C_{quat arom}), 133.1 (C-b), 129.6,

129.3, 128.5, 128.4, 128.2, 128.0, 127.9, 127.8, 127.7 (C_arom), 118.3 (C-c),

113.7 (C_mpMBn), 98.0 (C-1^F), 97.8 (C-1^B, C-1^C, C-1^D), 97.4 (C-1^E), 96.5

(C-1^A), 78.3 (C-3^A), 77.8 (C-3^C), 77.7 (C-3^E), 75.7 (C-4^A), 75.6 (C-4^B), 75.5

(C-4^C), 75.4 (C-3^B), 75.3 (C-4^D), 75.2 (C-3^F), 74.9 (CH₂Ph), 74.7 (CH₂Ph),

74.0 (CH₂Ph), 73.6 (C-3^D, CH₂Ph), 73.4 (CH₂Ph), 73.3 (CH₂Ph), 72.6 (C-

4^F, CH₂pMBn), 71.3 (C-5^B), 71.1 (C-2^B), 69.7 (C-5^E), 69.6 (C-5^C), 69.3 (C-

5^F), 69.2 (C-5^A), 68.8 (C-2^F), 68.7 (C-a), 63.3 (C-2^A), 62.9 (C-2^E), 62.7 (C-

2^C), 62.1(C-6 ^A), 61.6 (C-6^C, C-6^E), 55.2 (CH₃pMBn), 51.9 (2îCH₃

COOMe), 51.7 (CH₃COOMe), 20.8, 20.7 (CH₃OAc) ppm; IR (thin

ꢀ

169.5, 169.4, 169.2 (C=O), 158.9 (C OMe pMBn), 137.8, 137.6, 137.5,

137.4, 137.3, 137.2 (C_{quat arom}), 132.9 (C-b), 129.6, 129.2, 129.9, 128.4, 128.3,

128.2, 128.1, 128.0, 127.9, 127.8, 127.7, 127.6, 127.5, 127.4 (C_arom), 118.3

(C-c), 113.7 (C_mpMBn), 97.9 (C-1^H), 97.8 (C-1^B, C-1^C(E), C-1^D, C-1^F), 97.4

(C-1^E(C)), 97.1(C-1 ^G), 96.5 (C-1^A), 78.3 (C-3^A), 77.8, 77.7 (C-3^C, C-3^E, C-

3^E), 75.6, 75.5, 75.4, 75.3, 75.2 (C-4^A, C-4^B, C-4^C, C-4^Dor C-4^F, C-4^E, C-4^G,

C-3^B), 74.8 (C-3^H, CH₂Ph), 74.7 (2CH₂Ph), 74.6 (CH₂Ph), 74.0 (CH₂Ph),

73.9 (CH₂Ph), 73.5 (C-3^D, C-3^F, CH₂Ph), 73.3 (C-4^Dor C-4^F, CH₂Ph), 72.5

(C-4^H, CH₂pMBn), 71.2 (C-5^B), 71.1 (C-2^B), 70.5 (C-5^Dor C-5^F), 70.2 (C-

2^Dor C-2^F, C-5^Dor C-5^F), 70.0 (C-2^Dor C-2^F), 69.6 69.5 (3C, C-5^C, C-5^E,

C-5^G), 69.2 (C-5^H), 69.1(C-5 ^A), 68.7 (C-2^H), 68.6 (C-a), 63.3 (C-2^A), 62.8,

62.7, 62.6 (C-2^C, C-2^E, C-2^G), 62.0 (C-6^A), 61.5 (C-6^C, C-6^E, C-6^G), 55.2

(CH₃pMBn), 51.9 (3îCH₃COOMe), 51.7 (CH₃COOMe), 20.8, 20.7,

ꢀ

film): n˜ =3100, 3066, 3033, 2999 (C H_arom), 2953, 2928, 2870 (C H_aliph),

2109 (N₃), 1742 (C=O), 1613, 1515, 1497, 1455, 1438, 1371, 1304,

1235 cm^ꢀ1

;

elemental analysis: calcd (%) for C₁₀₄H₁₁₉N₉O₃₈

(2103.2 gmol^ꢀ1): C 59.39, H 5.70, N 5.99, O 28.91; found: C 59.41, H 5.88,

N 5.97.

ꢀ

20.6 (CH₃OAc) ppm; IR (thin film): n˜ =3097, 3065, 3034 (C H_arom),

ꢀ

2955, 2932, 2877 (C H_aliph), 2115 (N₃), 1749 (C=O), 1613, 1515, 1497,