iii

Solid-phase glycoconjugate synthesis

On-resin analysis with gel-phase

19F NMR spectroscopy

Mickael Mogemark

Akademisk avhandling

Som med vederbörligt tillstånd av rektorsämbetet vid Umeå universitet, för

avläggande av filosofie doktorsexamen vid tekniska-naturvetenskapliga

fakulteten, framläggs till offentligt försvar i Stora Hörsalen, KBC-huset,

fredagen den 21 januari 2005, kl 13.00. Avhandlingen kommer att försvaras

på engelska.

Fakultetsopponent: Prof. Dr. Peter H. Seeberger, Laboratorium für

Organische Chemie, ETH Hönggerberg, Zürich, Switzerland.

Department of Chemistry, Organic ChemistryUmeå University

Umeå 2005

iv

Title Solid-phase glycoconjugate synthesis – On-resin analysis with 19F

NMR spectroscopy

Author Mickael Mogemark, Department of Chemistry, Organic

Chemistry, Umeå University, S-901 87, Umeå, Sweden

Abstract An efficient and versatile non-destructive method to analyze

the progress of solid-phase glycoconjugate synthesis with

gel-phase 19F NMR spectroscopy is described. The method

relies on use of fluorinated linkers and building blocks

carrying fluorinated protective groups. Commercially

available fluorinated reagents have been utilized to attach

the protective groups.

The influence of resin structures for seven

commercial resins upon resolution of gel-phase 19F NMR

spectra was investigated. Two different linkers for

oligosaccharide synthesis were also developed and

successfully employed in preparation of α-Gal

trisaccharides and a n-pentenyl glycoside. Finally, reaction

conditions for solid-phase peptide glycosylations were

established.

Keywords: Solid-phase synthesis, protective groups, linkers,

monitoring, gel-phase, 19F NMR spectroscopy,

carbohydrates, oligosaccharides, glycopeptides.

Copyright © 2004 Mickael MogemarkISBN 91-7305-792-4

Printed in Sweden by Intellecta DocuSys, Göteborg 2004

v

To Lena

vi

Contents

List of papers ix

Abbreviations x

1 Solid-phase synthesis 1

1.1 Solid-phase oligosaccharide synthesis 2

1.2 Linkers 3

1.2.1 Linkers for solid-phase oligosaccharide synthesis 4

1.3 Analytical methods 6

1.3.1 High resolution magic angle spinning (HR-MAS) NMR

spectroscopy 6

1.3.2 Gel-phase NMR spectroscopy 7

1.3.3 Gel-phase 13C NMR spectroscopy 7

1.3.4 Miscellaneous analytical methods 8

1.4 Resins for solid-phase synthesis 8

2 Gel-phase 19F NMR spectroscopy 12

2.1 Monitoring polymer supported synthesis with 19F NMR

spectroscopy 13

2.2 Fluorinated linkers 14

2.3 Objectives of this thesis 18

vii

3 Comparison of resins used in solid-phase organic synthesis with

gel-phase 19F NMR spectroscopy 19

3.1 HR-MAS 1H NMR spectrum quality 19

3.2 Gel-phase 19F NMR spectrum quality 21

4 Fluorine labeled protective groups 25

4.1 Acetyl and benzoyl protective groups in glycoconjugate

synthesis 25

4.1.1 Reactivity of fluorinated benzoyl groups 26

4.2 Chemical 19F shifts for commercial protective groups and

reagents 28

4.3 Use of fluorinated protective groups in solid-phase

synthesis 29

5 Fluorine labeled linkers and their use in solid-phase

synthesis 32

5.1 Synthesis of an acid stable fluorinated linker 32

5.1.1 Establishing conditions for solid-phase glycosylation of

serine 33

5.2 Synthesis of an acid/base labile fluorinated Wang linker 36

5.3 Selenium containing resins 38

5.3.1 Synthesis of a fluorinated selenide for preparation of n-

pentenyl glycosides 39

viii

5.3.2 Establishing glycosylation conditions and cleavage from

the linker 41

5.4 Conclusions 47

6 Solid-phase synthesis of α-Gal trisaccharides 48

6.1 Xenotransplantation epitopes 48

6.2 Protective group strategy 49

6.3 Solid-phase assembly of α-Gal trisaccharides 50

6.4 Conclusions 57

7 Solid-phase peptide glycosylations 59

7.1 Reactivity in solid-phase peptide glycosylations 59

7.1.1 Solid-phase glycosylations of peptides containing serine

61

7.1.2 Conditions for cleavage from solid phase and removal of

protective groups 63

7.2 Conclusions 64

8 Future perspectives 65

9 References 67

10 Acknowledgements 81

ix

List of papers

This thesis summarizes the papers I-V, which are reprinted with kindpermission from the publishers.

I Mickael Mogemark, Frida Gårdmo, Tobias Tengel, Jan Kihlberg andMikael Elofsson.Gel-phase 19F NMR spectral quality for resins commonly used insolid-phase organic synthesis; a study of peptide solid-phaseglycosylations.Organic & Biomolecular Chemistry, 2004, 2, 1770–1776.

II Mickael Mogemark, Mikael Elofsson and Jan Kihlberg.Monitoring Solid-Phase Glycoside Synthesis with 19F NMRSpectroscopy.Organic Letters, 2001, 3, 1463–1466.

III Mickael Mogemark, Linda Gustafsson, Chrisoffer Bengtsson,Mikael Elofsson and Jan Kihlberg.A Fluorinated Selenide Linker for Solid-Phase Synthesis of n-Pentenyl Glycosides.Organic Letters (in press).

IV Mickael Mogemark, Mikael Elofsson and Jan Kihlberg.Fluorinated Protective Groups for On-Resin Quantification of Solid-Phase Oligosaccharide Synthesis with 19F NMR Spectroscopy.ChemBioChem, 2002, 3, 1266–1269.

V Mickael Mogemark, Mikael Elofsson and Jan Kihlberg.

Solid-Phase Synthesis of α-Gal Epitopes; On-Resin Analysis of

Solid-Phase Oligosaccharide Synthesis with 19F NMR Spectroscopy.

Journal of Organic Chemistry, 2003, 68, 7281-7288.

x

Abbreviations

Ac acetylAcOH acetic acidAla alanineArg arginineAsp Aspartic acidAq aqeuosBn benzylBz benzoylBzl benzylBoc tert-butoxycarbonylCbz benzyloxycarbonylcf. comparecHx cyclohexylCy cyclohexylCSA chemical shift anisotropyDAST (Diethylamino)sulfur trifluorideDBU 1,8-diazobicyclo[4,3,0]undec-7-eneDDQ 2,3-dichloro-5,6-dicyanobenzoquinoneDEAD diethyl azodicarboxylateDIC N,N’-diisopropylcarbodiimideDMAP N,N’-dimethylaminopyridineDMF N,N’-dimethylformamideDMS dimethylsulfideDMSO dimethylsulfoxideDMT 4,4´-dimethoxytritylDMTST dimethyl(methylthio)sulfonium triflateDVB divinylbenzenedppf 1,1´-bis(diphenylphosphanyl)ferrocenedppp 1,3-bis(diphenylphosphino)propane.g. for exampleEt ethylFG functional groupGal galactoseGln glutamineGly glycineFmoc 9-fluorenylmethoxycarbonylFT Fourier transformHEL hen-egg lysozymeHMBA 4-hydroxymethylbenzoic acidHMQC heteronuclear multiple quantum coherence

xi

HOAt N-hydroxy-9-azabenzotriazoleHOBt 1-hydroxy-1-benzotriazoleHPLC high performance liquid chromatographyHR high resolutioni.e. that isIgG Immunoglobulin GIle isoleucineIR infra redLeu leucineLev levulinoylLex Lewis xm metaMALDI matrix-assisted laser desorption/ionization

mass spectroscopyMAS magic angle spinningMBHA methylbenzhydrylaminemCPBA 3-chloroperoxybenzoic acidMe methylMeIm 1-methylimidazoleMOM methoxymethylMS mass spectroscopyMSNT 1-(mesitylene-2-sulfonyl)-3-nitro-1,2,4-

triazoleNBS N-bromosuccinimideNCLP non-cross-linked polystyreneNIS N-iodosuccinimideNMR nuclear magnetic resonanceo orthop paraPA p-methoxyphenyloxy carbonylPEG poly(ethylene glycol)PEGA (VERSABEAD-O) polyethylene glycol-polyacryl amidePfp pentafluorophenylPh phenylPhe phenylalaninePiv pivaloylPMB p-methoxybenzylPNBP (p-nitrobenzyl)-pyridinePOEPOP (VERSABEAD-O) polyoxyethylene-polyoxypropylenePS polystyreneR substituentRT room temperatureSer serineTBAF tetrabutyl ammonium fluoride

xii

tBu tert-butylTBDPS tert-butyldiphenylsilylTBS (TBDMS) tert-butyldimethylsilylTFA trifluoroacetic acidTfOH trifluormethanesulfonic acidTHF tetrahydrofuranThr threonineTLC thin layer chromatographyTMS trimethyl silylTMSOTf trimethylsilyl trifluoromethanesulfonateTOF time of flightTsOH toluensulfonic acidTyr tyrosineUv ultra violetVal valineX anomeric leaving group or functional

groupY functional group

1

1 Solid-phase synthesis

Solid-phase synthesis is a methodology in which synthetic transformationsare conducted with one of the reactants attached to an insoluble matrix. Thisapproach was developed by Merrifield in the beginning of the 1960´s1,2

when he described the preparation of peptides on a polystyrene resin, todayreferred to as Merrifield’s resin. It was envisioned that excess of reagentscould be used to drive reactions to completion, and that reagents and solubleby-products can then simply be removed by filtration and washing of thepolymeric support. These properties render solid-phase synthesis suitable forautomation,3-6 which is important for speeding up the synthetic process.Thus, solid-phase synthesis has been widely applied for preparation ofpeptides4,7,8 and oligonucleotides,9 on an individual basis or in the form oflibraries.10 This has been of major importance in our struggle to understandthe biological roles of proteins and nucleic acids. In recent years, increasedknowledge in organic synthesis and the introduction of combinatorialchemistry and parallel synthesis have renewed the interest in solid-phasesynthesis of, for instance, small ”drug-like” molecules,10-14 glycopeptides15-18

and oligosaccharides.19-23

2

1.1 Solid-phase oligosaccharide synthesis

Solid-phase synthesis has proven to be extremely efficient for preparation ofoligonucleotides and peptides, but preparation of complex oligosaccharidestructures by solid-phase synthesis is still in its infancy and constitutes aconsiderable challenge for all but a few laboratories.24-28 This can beattributed to the difficulties in the synthesis of the oligosaccharides, in whicheach glycosylation is unique due to the diverse reactivity of differentglycosyl donors and acceptors.29 In addition, the need to differentiatebetween the different hydroxyl groups requires use of an extensiveorthogonal protective group pattern.30-32 Thus, stereoselective glycosylationsand protective group manipulations in high yields are essential for asuccessful solid-phase oligosaccharide synthesis.

Numerous synthetic strategies have been developed andconsiderable progress in the field of solid-phase glycoconjugate synthesishas been made. Thus a wide variety of linkers,33-35 protective groups31,32,36,37

and glycosylation agents26,38-40 are available. In general, attachment of thefirst fully protected carbohydrate unit through its anomeric position to a solidsupport, via a linker, is the most efficient and frequently used strategy forsolid-phase oligosaccharide synthesis (Figure 1). After attachment the firstcarbohydrate moiety is subjected to protective group manipulation and theresulting glycosyl acceptor is subsequently glycosylated with a suitabledonor, which in turn is subjected to protective group manipulation to yield aglycosyl acceptor. Glycosylations and protective group manipulations in arepetitive cycle, and final cleavage from the linker releases theoligosaccharide from the solid support.

A major limitation when performing reactions on solid phase is thedifficulties in analyzing the outcome of complex reactions when the productis still attached to the polymeric support. Standard solution-phasechromatographic and spectroscopic methods are not compatible with solidsupports, and accumulation of resin-bound side-products may result inproblematic purifications of the final product. Hence, the lack of an efficientanalytical tool for optimization of solid-phase syntheses may result in wasteof both time and money.

3

Figure 1. General procedure for solid-phase oligosaccharide synthesis.

1.2 Linkers

The choice of linker for tethering the first synthon to the polymeric supportis a critical element in solid-phase synthesis. The field of protective groupchemistry has served as a basis for development of new linkers and reagents,and most of the linkers developed today are mimics of common protectivegroups used in solution-phase synthesis. It is necessary that (1) theattachment of the starting material to the resin can be achieved in high yield,(2) the linker is orthogonal to the protective groups employed and inert to the

X

Polymer

Linker FGY

Linker FG

OLinker

(OR)mO

Linker

(OR)m

Glycosyl donors

Glycosylation

Protective groupmanipulation

Oligosaccharide

Cleavage

HOn n

4

conditions used during synthesis, and (3) the target molecule can be releasedfrom the support in high yield using mild conditions that do not causedegradation. It is also desirable that the introduction of impurities duringcleavage is minimized and that the released products require minimal work-up and purification.

Usually a functional group in the starting material is attached to thelinker; this group is then unmasked during cleavage from the solid support.Introduction of “traceless linkers” in which the point of attachment to thesolid support is replaced by hydrogen and not apparent in the releasedproduct has recently emerged as an important tool in library synthesis.34

1.2.1 Linkers for solid-phase oligosaccharide synthesis

The carbohydrate scaffold displays several functionalities for attachment to alinker. However, anchoring through the anomeric position of thecarbohydrate is the most frequently employed strategy. During recent yearsnumerous linkers for solid-phase oligosaccharide synthesis have beendeveloped; a selection of linkers organized after the conditions used forcleavage is presented in Table 1.

5

Table 1. Structures of linkers, cleavage conditions, and products.

ORO O Si

Me Me

O

NHAc2O, BF3:Et2O O

ROOAc

OOSi

R

ROR O

OH

OR

ORO O

O ORO

TrBF4or

TFA/CH2Cl2

OH

RO RO

TBAF, AcOH

ORO

S X

NBS, DTBP, ROH ORO

OR

ORO S

O ORO

Hg(OCOF3), H2O

OH

NH4OHO

ORORO O

ORHORO

O

O

NH

OOR

ROHN

OO

HN DDQ O

ORRO

ORO O

O

ORO O

ORO O O

RO O

Grubb`s cat.Ethylene

Grubb`s cat.Ethylene

ORO O

O ORO

Raney-Ni W2

OH

ORO O

O ORO

hv

OH

O2N

Thioglycoside linker[41]

Thioglycoside linker[42,43]

Acid-labile Wang linker[44,45]

Acid-labile silyl linker[46]

Acid- and fluoride-labile silyl linkers[47–49]

Base-labile succinyl linker[50–52]

Wang linker cleaved by reduction[54]

Linker cleaved by oxidation[55]

Photo-labile linker[56]

Linker cleaved by olefin metathesis[57]

Linker cleaved by olefin metathesis[58]

O ORRO

OOB

Acetone, H2OO ORRO

HOOH

Linker cleaved by ester hydrolysis[53]

OOH

References41-58

6

1.3 Analytical methods

Monitoring solid-phase carbohydrate synthesis is not as straightforward ascompared to standard peptide synthesis, which is conveniently quantified bymeasuring residual amino groups with on-resin color tests, such as theKaiser test59 or with bromophenol blue (BFB).60 Destructive methods wherea small sample of resin is cleaved and subjected to ordinary solution-phaseanalysis, for instance LC-MS, TLC and NMR spectroscopy, are today incommon use in solid-phase oligosaccharide synthesis. The drawback withthese methods is that valuable product is consumed and thatintermediates/products may decompose during cleavage, which complicatesthe analysis. Even though MS methods are very sensitive and require smallquantities of compound,61,62 non-destructive methods are desirable andseveral innovative NMR spectroscopy methods suited for on-bead analysishave therefore been developed.63

1.3.1 High resolution magic angle spinning (HR-MAS)NMR spectroscopy

Magic angle spinning (MAS), i.e. spinning the sample at approximately 1–3kHz at an angle of 54°, is a well known method to reduce spin latticerelaxation and chemical shift anisotropy (CSA) that cause line-broadening.64

The method has been widely used for structural characterization of on-resinsolid-phase synthesis, mostly with 1H and 13C NMR spectroscopy.65-68

Development of high-resolution nano-probes has made it possible to performstructural determination on very small aliquots of beads or even on a singlebead.69-71 This constitutes a very valuable tool for validating hits fromcombinatorial libraries that have been screened directly on the resin.71-73

HR-MAS NMR spectroscopy has been applied to analyze theformation of a crude trisaccharide bound to a Merrifield’s resin,48 and a spin-echo technique was then employed to reduce the 1H NMR signals originatingfrom the resin. Furthermore, the carbohydrate was nicely assigned with 1H,13C and 1H–13C HMQC NMR experiments. However, a drawback is thatcomplicated spectra will be obtained due to the narrow spectral range of theproton if the reactions are incomplete and/or byproducts are formed. In

7

addition, the analysis is rather time consuming and requires specializedNMR equipment. Thus, for monitoring purposes this method is somewhatcumbersome and it is only available in a few specialized laboratories andtherefore not accessible for routine analyses.

1.3.2 Gel-phase NMR spectroscopy

Gel-phase NMR spectroscopy is essentially the transfer of solution-phaseNMR spectroscopy to the solid state. The resin is swollen in a deutueratedsolvent to give a solvent-swollen resin, i.e. a gel-phase, in order to obtainmaximum molecular mobility of the resin-bound molecules. The slurry isintroduced into a standard NMR tube and the spectrum is recorded usingordinary NMR equipment, i.e. MAS is not required, and after analysis theresin can be recovered. However, the resonances will suffer from line-broadening due to restricted motion and dipolar couplings. Depending on theresin used, this will give resonances that have line-widths up to severalhundreds of Hertz. As a consequence, this limits the use of gel-phase NMRspectroscopy to nucleic other than 1H that possess a wide spectral range, forinstance 13C, 32P, and 19F. Use of gel-phase 19F NMR spectroscopy for on-resin reaction analysis will be discussed in Chapter 2.

1.3.3 Gel-phase 13C NMR spectroscopy

13C NMR spectroscopy has not been extensively used for reactionmonitoring. This is due to the prolonged acquisition time required since thenatural abundance of the 13C isotope is low (1.1%). However, by utilizing13C-enriched building blocks, which greatly enhances the signal intensities,simple spectra with high resolution can be obtained, due to the wide spectralrange of the 13C nucleus.74 Thus, gel-phase gate decoupled 13C NMRspectroscopy was employed to monitor the preparation of the sialyl Lewisx

tetrasaccharide on a TentaGel resin,75,76 in which each step, i.e.glycosylations, deprotections and cleavage from the linker, could bequantified by comparison of the signals from the 13C-labeled linker with thetemporary 13C-labeled chloroacetyl protective group (ClAc).

8

1.3.4 Miscellaneous analytical methods

FT-IR spectroscopy has proved to be a valuable tool for monitoringformation and disappearance of certain functional groups.174 However,severe spectral overlap renders the analysis extremely difficult, which limitsthe general use of the method.

Real time monitoring using a color test in combination withMALDI-TOF MS analysis was recently introduced for solid-phaseoligosaccharide synthesis.62 The presence and deprotection of a temporarychloroacetyl group (ClAc) was quantified by dying with (p -nitrobenzyl)pyridine (PNBP), whereas the glycosylation was quantified bydetecting the hydroxyl group being glycosylated using cyanuricchloride–disperse red conjugation and/or detection of ClAc with PNBP.175

1.4 Resins for solid-phase synthesis

The structure and physiochemical properties of the resin may have a majorimpact on the outcome of a reaction performed on a polymeric support.Therefore, understanding the resin properties is crucial for an efficient solid-phase glycoconjugate synthesis. When planning the synthetic route theproperties of the resin, i.e. swelling, cross-linking, particle size, stability,functional groups, loading capacity, etc. have to be considered.77-81

Numerous resins with different synthetic and spectroscopic properties arecommercially available and the most frequently used resins are the gel-typeresins. These are typically spherical beads of cross-linked polystyrene (PS)or poly(styrene-oxyethylene) graft copolymers. In organic solvents suchresins furnish a solvent-swollen gel, which facilitates the diffusion ofreagents through the polymer network to the reactive sites.

The classical and the resin that is most commonly employed today isthe Merrifield resin (1 or 2% divinylbenzene cross-linked polystyrene)(Figure 2).1 It exhibits high loading capacity and excellent stability toleratinga wide range of reaction conditions. It has been shown that the reactive sitesare homogeneously distributed throughout the bead and that most of the

9

reactive sites are situated inside the swollen bead.82 Since the main part ofreactions take place inside the bead it has been anticipated that both thesolvent depending properties and steric hindrance of the Merrifield resinmay affect both the reactivity and the stereochemical outcome. In addition,the general use of the resin is somewhat restricted by its hydrophobiccharacter, which limits the swelling of the bead in polar solvents and also thediffusion of polar reagents into the polystyrene matrix.79,81

Realizing that the reaction rate is dependent on the swelling, cross-linking, and bead size, the swelling properties of the PS resin was furtherimproved by grafting the hydrophobic core with flexible poly(ethyleneglycol) spacers (PEG) to produce resins that swell in both polar and non-polar solvents. TentaGel83,84 and ArgoGel85 are PEG-grafted PS co-polymersthat have been extensively used because of their mechanical stability (due tothe stabilizing PS core) and swelling properties (as a consequence of theflexible PEG chains) (Figure 2). The high PEG content of the TentaGel andArgoGel resins (up to 70%) plays an important role to enhance both thehydrophilicity and the molecular mobility of the resins, which leads to thatthe resin creates a “solvent-like” environment.79,81 Thus, the rates ofreactions performed on such resins are significantly increased compared tothe standard PS resins. An another advantage is that the resins can beemployed in on-resin screening in biological assays.24,73 Furthermore, theNMR spectroscopic properties were improved as a result of the increasedmolecular mobility within the resin which decreases CSA.69 However,TentaGel and ArgoGel resins suffers from relatively low loading capacity(∼ 0.3 to ∼0.7 mmol/g), and are therefore generally insufficient forpreparation of solid-supported reagents and scavengers.86 The low loadingcapacity may however be an advantage in SPOS since a too high loadingcapacity may lead to incomplete reactions because of steric hindrance.87

NovaGel is a PEG grafted PS co-polymer, which has its functionalgroups situated directly on the PS polymer (Figure 2).77 Hence, the resinexhibits a loading capacity as high as the Merrifield resin and a swellingcapacity comparable with TentaGel and ArgoGel. Because the moleculesbound to the resin are fixed on the PS core, in contrast to the correspondingTentaGel and ArgoGel resins, the PEG chains will not increase their entropy.As a consequence of NovaGel´s “biphasic” composition the NMRspectroscopic properties remain rather poor (cf. section 3.2). To date nocomparison of reaction rates have been conducted with this resin.

10

VERSABEAD-A (PEGA)88,89 was developed in order to make aresin which was more compatible with aqueous solutions, and in particularfor biological screenings (Figure 2).73 This resin lacks the stabilizing PS coreand is a co-polymer that consist of polyacryl amide, which is cross-linkedwith PEG chains. Since the resin contains amide functionalities it is oftenincompatible with many of the harsh conditions used in organic synthesis,and has therefore mainly found applications for peptide and glycopeptidesynthesis.

The desire to develop a resin that exhibits physiochemical propertiesthat are compatible with conditions used in SPOS and with enzymaticchemistry led to the development of VERSABEAD-O (POEPOP) (Figure2).90 The VERSABEAD-O resin consists only of cross-linked flexible PEG-chains, and all the amide bonds of VERSABEAD-A are replaced by fairlystable ether linkages. The resin displays similar physical properties asVERSABED-A, and both of these resins display excellent swelling inaqueous solutions due to the high content of cross-linked PEG chains.91 Bothresins were shown to be highly permeable for enzymes and weresuccessfully used in direct on-resin biological screening.73 Furthermore, thepresence of flexible PEG-chains significantly increased the molecularmobility of the resin and it was shown that the HR-MAS 1H NMR spectralresolution of VERSABEAD-O was somewhat better in polar solvents whencompared to the PEG-grafted TentaGel and ArgoGel resins.91

11

Figure 2. Chemical structures of some commercial resins.

n mPolystyrene

O O1, 2

O Xn

Polystyrene

O O Xn

OOX n

n ≈ 30 - 50

HN O O OMen

**

Ph

NH2

n ≈ 30 - 50

Merrifield resin TentaGel

ArgoGel NovaGel

HO

O

OO O

O

O O OH

O

O O O

O

n

n

n

NH

O OHN

NH

O OHN

NH

O

O

O

O

O

R

R

OOH2NR

n

n

n

R

R

VERSABEAD-A (PEGA) VERSABEAD-O (POEPOP)

X

12

2 Gel-phase 19F NMR spectroscopy

Methods for analyzing the outcome of reactions directly on the resin plays apivotal role in development and optimization of novel solid-phase protocols.On-bead reaction monitoring with gel-phase 19F NMR spectroscopyconstitutes an excellent tool for evaluating reactions directly on the solidsupport and has found many applications. 19F NMR spectroscopy has severalfavorable properties, including high sensitivity (the natural abundance of the19F nucleus is 100%) and a broad distribution of chemical shifts as a result ofthe strong polarizability of the fluorine atom. Hence, chemical shift overlapis rather uncommon. Furthermore, since resins for solid-phase synthesis donot contain fluorine atoms interfering signals from the solid support are notencountered. Although the 19F resonances suffer from line broadeningoriginating from chemical shift anisotropy (CSA) and spin lattice relaxationthe resolution of spectra is usually high. Consequently, magic angle spinning(MAS) is not required and high quality spectra can be obtained in a coupleof minutes by measuring on a solvent-swollen resin using a standard NMRinstrument.

13

2.1 Monitoring polymer supported synthesis with19F NMR spectroscopy

Early in 1980, Manatt et al. reported the use of 19F NMR spectroscopy tomonitor peptide elongation on a Merrifield resin.92 Starting from thisobservation several studies has been carried out where gel-phase 19F NMRspectroscopy has been employed to control the advancement of solid-phasetransformations. The technique has been applied to SNAr reactions,93,94

amide synthesis,95-98 formation of tertiary amines,99 reductive aminations,97

ether bond formation,100,101 ester formation,102-104 Staudinger reactions,105

Michael additions,106,107 Knoevenagel condensations,97,108 glycosylations,109-

114 2-pyridinone formation,115 acidolysis,116 and for encoding ofcombinatorial libraries.106,117

In 1996 Shapiro et al. used the 19F NMR approach to investigate thereaction kinetics of a SNAr displacement of a fluorine atom on a Merrifieldresin (Scheme 1).93 Conversion of 1 to 2 was analyzed directly on the resinevery 15 minutes, by excluding an aliquot of resin and halting the reactionprior to analysis. In this study, the spectral quality was somewhat poor due tothe rigid nature of the Merrifield resin, but still allowed quantification.

Scheme 1. SNAr reaction on solid phase.

In the same year Svensson et al. established conditions for amidebond formation on a TentaGel resin in an initial study,95 in which it was alsofound that 19F signals in solution closely matched the resin-linkedcounterparts. In addition, it was concluded that TentaGel produced 19F NMRspectra of high resolution (due to the flexible poly(ethylene glycol) spacers),and it was anticipated that the narrow line widths should allow monitoring ofeven small chemical shift changes, without any need for MAS. The impact

HN

ONO2

F Cl•H3N CO2Et

DIEA, DMF

HN

ONO2

HN CO2Et

1 2

14

of the resin structure upon the 19F spectral quality is discussed further inchapter 3.

2.2 Fluorinated linkers

In view of the advantages that gel-phase 19F NMR spectroscopy was foundto provide for monitoring reactions on solid support, Svensson et al.introduced the use of fluorine labeled linkers that serve as internal analyticalstandards throughout the synthesis.95-97 In the study fluorinated analogs oflinkers commonly used in peptide synthesis were prepared and evaluated(Figure 1).

Figure 1. Fluorine labeled linkers used in solid-phase synthesis.

The linkers were expected to exhibit about the same reactivity astheir non-fluorinated variants. Hence, carboxylic acids attached to linker 3should be rather stable towards acids and therefore preferably cleaved bynucleophiles, whereas it may be possible to cleave linker 4 by strong acidicconditions, e.g. liquid hydrogen fluoride. The Wang linker is one of the mostfrequently used linkers in solid-phase synthesis mainly due to the ability toundergo cleavage under both acidic and nucleophilic conditions. Indeed, thefluorinated Wang-type linker 5 was also found to be cleaved by a weak acidsuch as trifluoroacetic acid (TFA), but a recent study revealed an incompletecleavage of a peptide linked to 5 at room temperature and it was shown thata temperature of 60 °C was required to achieve complete cleavage.116

However, the 19F resonance of linker-loaded resin 6 underwent a largechemical shift upon acylation (–115 → –117 ppm) and the well-separatedresonances allowed convenient quantification of coupling to and cleavagefrom the resin (Scheme 2).

FO

OH

CO2HFHO2C

OH

F

OH

HO2C

3 4 5

15

Scheme 2. Chemical shifts for linker 6 before and after acylation.

The concept was successfully applied in solid-phase synthesis of acoumarin- and a peptoid based library.97 Thus, use of the fluorinated linker-loaded resin 6 in combination with fluorinated building blocks, enabledoptimizations of a Knoevenagel condensation and a reductive amination,respectively, to allow development of a facile synthetic route to coumarincompounds 8 and peptoid compounds 9, which could be obtained in goodoverall yields.

Scheme 3. Solid-phase synthesis of coumarins and peptoids.

FO

OH

6O

HNFO

O

7O

HN

O

R

–115 ppm –117 ppm

Acylation

FO

COHN

OR1 O

N R3O R2

HO

ONR2

O O

OFO

O

COHN

ONR2

O O

O

FO

OH

COHN

R3

R3

HOR1 O

N R3O R2

8 (Coumarin library)

9 (Peptoid library)

R1

R1

LiOH (aq)

THF/MeOH

THF/MeOH

LiOH (aq)6

16

Recently, the group of Salvino prepared the fluorinated Wang linker12 on a Merrifield resin in high overall yield (92%) starting from 10(Scheme 4).107 The utility of the linker for on-resin analysis with MAS 19FNMR spectroscopy was demonstrated in a multistep synthesis, whichincluded an acylation, a Horner–Wadsworth–Emmons olefination, a Michaeladdition, and an oxidation to give resin 15. Cleavage from the resin wasaccomplished by acidolysis with TFA (30% in dichloromethane) at roomtemperature to give 16 in quantitative yield.

Scheme 4. Preparation of fluorinated Wang resin 12 and solid-phasesynthesis of 16.

In view of the previous studies, a set of fluorine containing Wanglinkers 17-19 was prepared on a non-cross-linked polystyrene (NCLP) resin(Figure 2).101 By exploiting the fluorine substitution pattern, it was foundthat the proximity of the fluorine atom to the hydroxyl group of the linker isessential in order to enable analysis of the loading of non-fluorine containing

O F

OH

12

O

HO

F O

OH

MOMO

F1) MOM-Cl, NaH, THF2) NaOCl, H2O, dioxane

1) MeOH, HCl2) PS-Cl, Cs2CO3, DMF3 LAH, THF

10 11HO P OEt

O O OEt

2,6-di-ClBzClPyridine, DMF

O F

O

13

P OEt

O O OEt

Ph

O, (Me3Si)2NLi

THF

O F

O

14 O

1) 3-FPhSH, DBU, DMF2) mCPBA, dioxane

HO

O O2S F

16

O F

O

15 O O2S

FTFA, CH2Cl2

17

substrates. The 19F resonances of resin 17 and 18 remained at the same shiftsafter attachment of the first building block and it was concluded that theselinkers are less suited as probes for monitoring with 19F NMR spectroscopy.However, the fluorine atoms in 19 appeared to be proximal enough to detectchemical changes in close distance to the reaction site. Unfortunately, theauthors did not investigate reactivity/stability towards acids. However, withrespect to the electron withdrawing fluorine substituents, the linker 19 maywell be stable towards acidic cleavage. Because the reactivity of thehydroxyl group of the linker is also decreased, loading of unreactivesubstrates may be problematic.

Figure 2. Fluorine containing Wang non-cross-linked polystyrene resins.

More recently, a perfluoroalkyl sulfonyl fluoride containing resinwas reported as a traceless linker for insertion of C-H or C-C bonds.103,104

The concept was based on attachment of aromatic phenols to 20, whichfurnishes an activated sulfonyl ester linkage, 21 , that is sensitive topalladium-mediated conditions. Furthermore, the loading of the substratecould be quantified by monitoring the displacement of the sulfonyl fluoridewith gel-phase 19F NMR spectroscopy. The linker was shown to be stable toconditions used in reductive aminations, and the products could be releasedfrom the solid support either by reductive deoxygenation to give 22, or bySuzuki cleavage/cross-coupling to yield 23.

OH

F

OHF F

O OH

FF

FF

17 18 19

18

Scheme 5. Cleavage/Cross-coupling of phenols linked to 20.

2.3 Objectives of this thesis

Methods for analyzing solid-phase synthesis are a fundamental requirementin order to determine the completion of reactions, especially for optimizationof the reaction conditions. In view of the advantages that gel-phase 19F NMRspectroscopy provides for monitoring reactions solid-phase, we anticipatedthat this technique would be particularly useful in solid-phaseglycoconjugate synthesis as the existing analytical techniques are somewhatrestricted by means of availability, user friendliness, and interpretation.

The main objectives of this thesis are:

♦ Introduction of fluorinated protective groups in order to facilitate analysisof solid-phase glycoconjugate synthesis.

♦ Preparation of fluorine labeled linkers for use in oligosaccharidesynthesis.

♦ Development of new methodologies for solid-phase glycoconjugatesynthesis.

HN O

FO F

F F F FS

F F

O O

F

20

= SO O

FF F

K2CO3, DMF

HO

SO O

F FR1

B(OH)2R2

R1

PdCl2(dppf), Et3N, DMF

Pd(OAc)2, dppp, Et3NDMF, HCO2H

R1 R1

R2

21

22 23

O

19

3 Comparison of resins used in solid-phase organic synthesis with gel-phase19F NMR spectroscopy

As emphasized in section 1.3.2 restricted molecular mobility of resin-boundmolecules will give rise to line-broadenings up to several hundreds Hertzwide. Thus, the resin structure has a major impact on the NMR spectraquality, in particular if no magic angle spinning is utilized. This fact restrictsthe use of gel-phase NMR spectroscopy to nuclei that exhibit a wide spectralrange, for instance 13C, 32P and 19F, in order to avoid severe spectral overlap.

3.1 HR-MAS 1H NMR spectrum quality

Keifer investigated the influence of resin structure, tether length, and solventupon high-resolution magic angel spinning (HR-MAS) 1H NMR spectrumquality.69 The study included nine commercially available resins loaded withdifferent linkers carrying an aspartic acid residue: Fmoc-Asp(OtBu)-Wang(24), Fmoc-Asp(OtBu)-NovaSyn KA 100 (25), Fmoc-Asp(OtBu)-NovaSynKA 500 (26 ), Fmoc-Asp(OtBu)-NovaSyn TGA (27), Fmoc-Asp(OtBu)-NovaSyn TGT (28), Boc-Asp(OcHx)-Merrifield (29), Boc-Asp(OBzl)-PAM(30), Boc-Asp(OcHx)-MBHA (31), Boc-Asp(OtBu)-Rink amide MBHA(32). Thus, four different polymers were studied: a 1% DVB cross-linkedpolystyrene resin (resins 24, 29 – 32), a composite of polyacrylamide andKiselguhr (KA 100, resin 25), a composite of polyacrylamide andpolystyrene (KA 500, resin 26), and a 1% DVB cross-linked polystyreneresin grafted with poly(ethylene glycol) (PEG) spacers (“TentaGelequivalents”, resins 27 and 28).

The spectra were recorded in seven different solvents, i.e. CD2Cl2,DMF-d7, DMSO-d6, acetone-d6, benzene-d6, CD3OD, and D2O, and it wasshown that both the choice of solvent and resin had a major impact on the

20

spectral quality. In general, the polar aprotic solvents (DMF-d7, CD2Cl2 andDMSO-d6) furnished the best spectrum for all resins, and it was concludedthat the PEG grafted resins 27 and 28 produced spectra having the narrowestline widths (Figure 1, spectra f – h), since they have the most mobilemoieties due to the long PEG tethers. These resins were closely followed bythe PA 500 resin 26 (Figure 1, spectrum e).

Figure 1. High-resolution magic angle spinning 1H NMR spectra forselected solvent-swollen resins recorded at a spin rate of 2 kHz. Reproducedwith permission from Keifer, P. A., J. Org. Chem. 1996, 61, 1558. Copyright1996 American Chemical Society.

21

In addition, it was pointed out that the Merrifield resins providedspectra of poor quality (Figure 1, spectra a - d), since they do not containflexible tethers, and also due to problems with presaturation of thebackground signals. A remarkable observation was that the resin signals ofthe PEG grafted resins 27 and 28 could be conveniently removed, sincepresaturation of the PEG signal at 3.5 ppm also quenched the polystyrenesignals.

A recent study performed by Bianco et al. compared HR-MAS 1HNMR spectra resolution of two 1% crosslinked DVB polystyrene resins,118

i.e. a Ellman polystyrene resin (carrying a dihydropyran linker) and a Wangpolystyrene resin (carrying a Wang linker), with three PEG containing resinsi . e . NovaSyn (TentaGel equivalent), Tentagel, and VERSABEAD-O(POEPOP, cf. section 1.4). In that study, the polystyrene based resins,including the TentaGel and NovaSyn resins, were found to give inadequatespectra due to problems with presaturation of the polymer signals and fromthat observation it was concluded that the VERSABEAD-O resin wassuperior to the polystyrene based resins.

3.2 Gel-phase 19F NMR spectrum quality (I)

In order to further provide insight into the NMR spectroscopic properties forresins commonly used in organic synthesis and to identify the best resin foranalyzing samples directly on the resin with gel-phase 19F NMRspectroscopy, we decided to investigate the effect of resin structure as wellas solvent upon quality of 19F NMR spectra (Figure 1). Seven differentcommercially available amino functionalized resins (cf. section 1.4 forstructure descriptions) carrying a N-acylated dipeptide with four fluorineatoms were analyzed: 1) polystyrene resin 33, cross-linked with 1% DVB, 2)ArgoPore 34, a highly DVB-cross-linked polystyrene resin, 3) ArgoGel 35,4) TentaGel 36, 5) NovaGel 37, 6) VERSABEAD-A 38, also referred to asPEGA, and 7) SPAR-50 39, which is an acrylamide resin. Each resin wasanalyzed in CDCl3, DMSO-d6, benzene-d6 and CD3OD.

Our results clearly show that the molecular mobility of moleculesattached to the resin and the resin structure is essential to afford spectra ofhigh resolution (Figure 2). The highly DVB-cross-linked ArgoPore resin 34have a very restricted molecular mobility, therefore no spectra could be

22

obtained. In sharp contrast, the resolution of the lightly DVB-cross-linkedpolystyrene 33 and NovaGel 37 resins were found to give adequate spectrain DMSO-d6, which illustrates that the degree of DVB-cross-linking isconnected with the spectral resolution. The PEG grafted/cross-linked resinsArgoGel 35, TentaGel 36, and VERSABEAD-A 38 displayed the narrowestline widths in all solvents due to the high content of flexible PEG chains.The ArgoGel resin 35 and TentaGel resin 36 produced high quality 19F NMRspectra in CDCl3, DMSO-d6 and benzene-d6, whereas use of CD3ODresulted in spectra of moderate quality (Figure 2). The influence of the PEGgrafts of these resins results in excellent swelling properties in a wide varietyof solvents and they swell well in both polar and non-polar solvents.Importantly, this also increases the molecular mobility of the organic moietyattached to the resins and as a consequence both spin lattice relaxation andthe chemical shift anisotropy is reduced to provide spectra with high qualityin a wide variety of solvents. Moreover, the 19F resonances from the two p-FPhe residues were completely separated in both CDCl3 and benzene-d6. Thebest spectra for both resins were obtained in CDCl3, closely followed byDMSO-d6, benzene-d6 and CD3OD. However, a closer inspection of the linewidths revealed that the TentaGel resin 36 generally gave rise to spectra ofslightly better resolution than the ArgoGel resin 35. VERSABEAD-A is apoly-acrylamide support that is cross-linked with PEG chains, which givesthe resin good spectroscopic and swelling properties in a wide range ofsolvents, and in particular in protic solvents. As expected, the resin 38indeed produced 19F NMR spectra of high resolution in both CDCl3 andDMSO (Figure 2). However, recording resin 38 in CDCl3 resulted infragmentation of one of the pFPhe residues into four resonances. Thisobservation was assumed to originate from inhomogeneous distribution ofthe functional sites and the presence of different stereoisomers of thesecondary amine. This may severely complicate analysis of the NMRspectra, which decreases the use of this resin for monitoring reactions onsolid phase with NMR spectroscopy. SPAR-50 is a poly-acrylamide polymerprovided with large pore size, which is particularly suitable for enzymaticapplications since the large pores facilitates the entrance of the enzyme intothe polymer matrix. In contrast to VERSABEAD-A, this polymer does notcontain PEG tethers, and consequently the molecular mobility of thispolymer is rather low compared to VERSABED-A. Hence, 19F NMR spectrawith low quality were obtained with resin 39.

23

Figure 2. Gel-phase 19F NMR spectra for resins 33 – 39.

HN

ONH

HN

O

O

F

F

F

F

= Polystyrene (33), ArgoPore (34), ArgoGel (35),TentaGel (36), NovaGel (37), VERSABEAD-A (38), SPAR -50 (39)

33–39

–103 → –105 ppm–108 → –109 ppm –116 ppm

–116 ppm

24

Unfortunately, at the time this study was conducted the promisingVERSABEAD-O (POEPOP) was not commercially available and wastherefore not included. However, an initial study of VERSABEAD-Oillustrated that the HR-MAS 1H NMR spectral quality was significantlyimproved when the length of the PEG tethers was increased.91 It was alsofound that VERSABEAD-O in protic solvents produced somewhat narrowerline widths than the TentaGel and ArgoGel resins. A major drawback withthis resin (cf. VERSABED-A 38) is that the resin contains both primary andsecondary hydroxyl functionalities, which gives duplication of the NMRresonances.

In conclusion, the PEG grafted/-crosslinked resins, i . e .VERSABEAD-A, TentaGel and ArgoGel, were by far superior to the otherresins used in this study giving the following general order of 19F NMRspectral quality: VERSABEAD-A ≥ TentaGel ≥ ArgoGel >> NovaGel ≥SPAR-50 ≥ Polystyrene >> ArgoPore. Our results also clearly demonstratethat the influence from the PEG tether and the rigidity of the resin structure,i.e. the degree of cross-linking, are the most important factors affecting thespectral quality.

25

4 Fluorine labeled protective groups

The objective of this chapter is to summarize the fluorinated protectivegroups employed in the thesis. At the beginning of this project, only littlewas known regarding the reactivity of fluorinated aromatic protectivegroups. Thus, this chapter also survey an investigation reported by Sjölin etal. in order to get a better picture of the reactivity of fluorinated benzoylgroups in deprotection and as well as participating group in glycosylation,

4.1 Acetyl and benzoyl protective groups inglycoconjugate synthesis

The acetyl group is a frequently used protective group in carbohydratesynthesis, in particular for preparation of glycosylated amino acids andglycopeptides, due to its ease of removal under mild basic conditions.Problems associated with the acetyl groups include insufficient neighboringgroup participation, orthoesther-formation, and acyl migration, which oftenresults in complex product mixtures that are difficult to separate.

The benzoyl group was introduced as an excellent alternative toovercome these problems,119,120 since the benzoyl group provide moreeffective neighboring group participation, is less prone to migrate and do notreadily form orthoesthers. However, a great disadvantage is the strong basicconditions required for deprotection, and as a consequence the removal ofthe O-benzoyl groups of glycopeptides is often hampered by base-catalysedepimerization of the α-stereo centers and by β-elimination carbohydratemoieties linked to serine and threonine.116,121,122

26

4.1.1 Reactivity of fluorinated benzoyl groups

Sjölin et al. systematically investigated the scope and limitations of theacetyl- and benzoyl groups in glycopeptide synthesis.116,123,124 In order totake advantage of the favorable properties of the benzoyl group, an electronwithdrawing fluorine atom was introduced to enhance the rate ofdeprotection of base-sensitive glycopeptides.124 An initial screening withsodium methoxide in methanol (6 mM) revealed that the fluorinated benzoylgroups indeed increased rate of the deprotection. The debenzoylation wasfound to be dependent on both the number of fluorine substituents and theirposition, giving the following rate of removal: 2,5-difluoro- > 3-fluoro- > 2-fluoro- > 4-fluorobenzoyl.

Previous attempts to synthesize an acetylated analog of 43 were notreadily accessible due to formation of inseparable α -/β-mixtures ororthoesters.122 Exchanging the acetyl protective groups to benzoyl groupscircumvented these problems and the lactosylated building block 43 could beobtained in excellent yield (Scheme 1).122 Adding to the difficult preparationof lactosylated peptide building blocks, it is a well-known fact that removalof benzoyl groups of the corresponding glycopeptides occurs at a substantialloss of material caused by β-elimination.124 In this perspective thefluorinated benzoyl protective groups constitute a valuable alternative if theydisplay normal reactivity in glycosylation. Indeed, the fluorinated benzoylgroups were found to perform equally well in glycosylations as thecorresponding benzoyl group, which allowed preparation of 43 – 45 in highyields (Scheme 1).124

Scheme 1. Glycosylation of lactose derivatives 40 – 42 with Fmoc-Ser-OPfp

O

RO

OR

ROO

O

OR

ORRO

RO

BrAgOTfCH2Cl2

O

RO

OR

ROO

O

OR

ORRO

RO O OPfp

O

NHFmocHO OPfp

O

NHFmoc

40 R = Bz41 R = 4-FBz42 R = 2,5-diFBz

43 R = Bz (72%)44 R = 4-FBz (59%)45 R = 2,5-diFBz (64%)

27

The building blocks 43 – 45 were subsequently subjected to solid-phase peptide synthesis to give glycopeptides 46 – 48 carrying an acylatedlactose O-linked to Ser56 in a hen-egg lysozyme [HEL(52-61)] peptidefragment (Scheme 2). Treatment of 46 - 48 with a catalytic concentration oflithium hydroxide in methanol (20 mM) produced a substantial amount of β-elimination product 50 (50%) for the BzO-protected glycopeptide 46 and40% of 50 for the 4-FBzO protected glycopetide 47. In contrast, no β-elimination was observed for the 2,5-diFBzO-protected analog 48 when itwas subjected to lithium hydroxide or ammonia in methanol and theglycopeptide 48 could be fully deprotected in less than two hours. Thus, byusing fluorinated benzoyl groups the previous problems with formation oforthoesters, α -/β -mixtures, and β -elimination could be elegantlycircumvented.

Compound R Base Reaction time Amount of 50

46 Bz LiOH 8 h ~50%

47 4-FBz LiOH 8 h ~40%

48 2,5-diFBz LiOH < 0.5 h 0%

48 2,5-diFBz NH3 2 h 0%

Scheme 2. De-O-benzoylation of HEL-glycopeptides 46 – 48.

Gln-Ile-Asn-Ser-Arg61-NH2

OAsp52-Tyr-Gly-Ile-NH

OO

OR

OR

ORO

O

ORRO

RO OR

Gln-Ile-Asn-Ser-Arg61-NH2

OAsp52-Tyr-Gly-Ile-NH

OO

OH

OH

OHO

O

OHHO

HO OH

Gln-Ile-Asn-Ser-Arg61-NH2

OAsp52-Tyr-Gly-Ile-NH

+

LiOH, MeOHor NH3, MeOH

49

50

46 R = Bz47 R = 4-FBz48 R = 2,5-diFBz

43 R = Bz44 R = 4-FBz45 R = 2,5-diFBz

Peptidesynthesis

28

4.2 19F chemical shifts for commercial protectivegroups and reagents (II, IV, V)

In order to evaluate reactions on solid phase with gel-phase 19F NMRspectroscopy, we decided to engage fluorinated variants of reagents thatcorresponds to classical protective groups used in carbohydrate chemistry,i.e. benzoyl chlorides, benzyl bromides and benzaldehydes. The 19F shifts forthese mono-fluorinated protective groups attached to monosaccharidesdisplay distinguishing resonances that are dispersed over a large spectralrange (Figure 1). Thus, the risk of encountering chemical shift overlap israther low. Furthermore, The fact that these reagents are commerciallyavailable and comes at a low price is a great advantage.

Figure 1. Approximate 19F chemical shifts for aromatic protecting groups.

OR OR OR

F

FF

OR OR OR

F

FF

RORORO

OR OR OR

F

FF

OOO

–119 ppm –114 ppm –115 ppm

–121 ppm –113 ppm –112 ppm

–109 ppm –111 ppm –105 ppm

Benzyl ethers:

Benzoates:

Benzylideneacetals:

Ortho Meta Para

29

In addition, Hochlowski et al. introduced alkyl- and benzoylic acidslabeled with trifluoromethyl- and difluoromethyl substituents as tags forencoding combinatorial libraries.117 The resonances of these reagentsattached to a resin via an amide linkage appeared between –40 to –80 ppm.Hence, they greatly expanding the chemical shift region of the acylprotective groups to constitute an interesting complement to the fluorinatedaromatic reagents outlined in Figure 1.

4.3 Use of fluorinated protective groups in solid-phase synthesis (I – V)

Besides acting as analytical markers, fluorinated aromatic protective groupsmust exhibit normal reactivity in glycosylations, hydrolysis, and reductionsin order to be useful in solid-phase synthesis. These certain requirements aresupported in several studies (cf. section 4.1 – 4.2 and chapter 5 – 7).Furthermore, it is important to point out that fluorinated variants of benzoylchlorides, benzyl bromides and benzaldehydes used in this thesis shows thesame reactivity in protection of carbohydrate building blocks as theircorresponding non-fluorinated variants. As summarized in Scheme 3, thesynthesis of glycosyl donors and building blocks that are carryingfluorinated protective groups could be prepared in adequate overall yields byfollowing established procedures.

30

Scheme 3. Synthesis of carbohydrate building blocks 52 – 73 carryingfluorinated protective groups and the linker building block 75.

31

Scheme 3. Cont.

O

ORoFBzO

OOpFPh

OH

OO

ORHOO

OpFPhO

RHO

HOHO

OH

oFBzCl, DMAPpyridine

RuCl3, NaIO4

68, R = NHAc (50%)69, R = OH (37%)

70, R = NHAc (84%)71, R = OoFBz (98%)

72, R = NHAc (67%)73, R = OoFBz (63%)

1) Allyl alcohol, HCl (g)2) pFPhCH(OMe)2, p-TsOH, MeCN

66, R = NHAc67, R = OH

CCl4/MeCN/H2O

O

OHHO

OO

SPhMe

mFPh1) NaOMe, MeOH2) mFPhCH(OMe)2, pTsOH, MeCN

51

60 (86%)

pFBzCl, Pyridine

O

OpFBzpFBzO

OO

SPhMe

mFPh

61 (98%)

O

OoFBnoFBnO

OO

SPhMe

mFPh

62 (90%)

oFBnBr, NaH, DMF

O

OHRO

OO

SPhMe

oFPh1) NaOMe, MeOH2) oFPhCH(OMe)2, pTsOH, MeCN

51

63, R = H (86%)FmocClPyridine

O

OpFBzFmocO

OO

SPhMe

oFPh

65 (98%)

pFBzCl

Pyridine

64, R = Fmoc (85%)

HO OH

74

pFBzClPyridine pFBzO OH

75 (73%)

32

5 Fluorine labeled linkers and their use insolid-phase synthesis (II, III)

As discussed in chapter 2, a linker carrying a fluorine label serves as aninternal analytical standard that allows analysis with gel-phase 19F NMRspectroscopy. By engaging fluorine containing building blocks or protectivegroups (cf. previous chapter), both stereochemical and quantitativeinformation may be obtained for each reaction step on solid phase. Thischapter mainly focuses on preparation of fluorinated linkers described inpaper II and III and some of their applications in glycoconjugate synthesis.

5.1 Synthesis of an acid stable fluorinated linker

Since glycosylations normally are carried out under acidic conditions, acidstability is of particular importance for glycoconjugate synthesis. Thus, in aneffort to develop a linker for oligosaccharide synthesis, an acid stable linkerfor anchoring carboxylic functionalities to the solid phase was effectivelyprepared directly on the solid support by a three steep sequence (Scheme 1).Attachment of the first building block 76 was achieved by acylation of theArgoGel-NH2 resin using N,N´-diisopropylcarbodiimide (DIC) and 1-hydroxy-1-benzotriazole (HOBt) for activation to give 77. Treatment of 77with methanolic sodium methoxide removed O-acylated byproducts andsubsequent alkylation of the phenolic hydroxyl group with mono-pFBz-protected pentadiol 75 under Mitsunobu conditions125 furnished resin 78.Subsequent deacylation with methanolic sodium methoxide gave the linker-loaded resin 79 in 97% overall yield.

Resin 79 have been successfully engaged in solid-phase synthesis ofα-Gal trisaccharides (cf. Chapter 6). It was also found to be useful inoptimization in solid-phase synthesis of bicyclic 2-pyridinones,115 which

33

resulted in an improved synthetic route for the preparation of a “pilicide”library, i.e. compounds that target pilus formation in pathogenic bacteria.

Scheme 1. Preparation of the resin bound linker 79.

5.1.1 Establishing conditions for solid-phase glycosylationof serine

In an effort to establish conditions for glycosylations of serine containingglycolipids and peptides (cf. section 5.2 and chapter 7), it was decided tofirst investigate glycosylation of the resin bound serine derivative 80.Linker-loaded resin 79 was esterified with Fmoc-serine using 1-(2-mesitylenesulfonyl)-3-nitro-1,2,4-triazole (MSNT)126 as coupling reagent(Scheme 2). Elimination of the Fmoc group followed by acylation of theamino group with p-fluorobenzoic acid introduced a second fluorine markerat –108 pp. Acidolysis of the tBu group furnished resin 80.

The alcohol functionality of serine has been recognized to beunreactive in glycosylations, which has been suggested to be a consequenceof intramolecular hydrogen bonding between the hydroxyl group and theamide hydrogen.127 With respect to the poor reactivity of the acceptor, wedecided to investigate thioglycosides as glycosylation agents, whoseversatility has been acknowledged due to their reactivity, ease of preparationand handling.128 Glycosylation of 80 with 52 (three equivalents) was first

H2N

NH

OF

HO

NH

OF

OHO4

ArgoGel-NH2 =

77 (quant)OH

OF

HO

76

NH

OF

OpFBzO4

+

78 (97%) 79 (quant)

DIC, HOBtOHpFBzO

4

DEAD, PPh3THF

75

NaOMe, MeOH

–134.4 ppm

DMF

34

carried out with freshly prepared dimethyl(methylthio)sulfonium triflate(DMTST)129 as Lewis acid catalyst. These conditions resulted in slowformation of 81 giving a low yield (∼30%) and modest stereoselectivity (α/β= 1:3) as a consequence of the inadequate reactivity (Scheme 2).

Scheme 2. Preparation of the serine resin 80 and glycosylations withthiogalactoside 52 using different promoters.

However, performing the reaction under activation by the promoter systemNIS and TfOH130,131 enhanced the reaction rate to give the β-anomer of 80 in∼70% with complete stereoselectivity. The amount of 52 was increased tofive equivalents to ensure a quantitative formation of 81. The efficiency of

35

these conditions was further confirmed by glycosylations of 80 with therelated glycosyl donors 54 and 56 (five equivalents, respectively) to give 82and 84 in quantitative yields (Scheme 3).

As determined with 19F NMR spectroscopy, the glycoconjugateswere quantitatively cleaved from the resin by mild saponification withsimultaneous hydrolysis of the fluoro benzoyl groups without causingepimerization or β-elimination of the serine moiety (Scheme 3). Thus,treatment of resin 81 and 84 with catalytic amount of aqueous lithiumhydroxide gave 83 and 85 in 39% and 37% yields, respectively, afterpurification with reversed-phase HPLC. Submitting resin 82 to aqueous TFAfollowed by hydrolysis with aqueous lithium hydroxide gave 83 in 19%yield after purification with reversed-phase HPLC.

Scheme 3. Glycosylation of 80 with different galactosyl donors andsubsequent resin cleavage.

O

pFBzOpFBzO

OO

OO

O

NHpFBzO

F

O

NH

pFPh

O

pFBzOTBDMSO

OTBDMSpFBzO

OO

O

NHpFBzO

FO

NH

4

4

O OOH

HO

OHHOOH

NHpFBz

O

O OOH

HO

OO

OHNHpFBz

O

pFPh LiOH/H2O, THF

83

82 (quant)

85

84 (Quant)

O

pFBzOTBDMSO

OTBDMSpFBzO

SPhMe

O

pFBzOpFBzO

OO

SPhMe

pFPh

80 +

80 +

NIS, TfOH

CH2Cl2

NIS, TfOH

CH2Cl2

1) TFA/H2O (9:1, v/v)2) LiOH/H2O, THF

LiOH/H2O, THF81

(39%)

(quant)

(quant)

(19%)

(37%)

54

56

36

5.2 Synthesis of an acid/base labile fluorinatedWang linker

An isomeric analog of the previous described acid labile fluorinated Wanglinker (cf. section 2.2) was synthesized in 51% total yield, over three steepsby Wallner et al. (Scheme 4).114

Scheme 4. Synthesis of fluorinated Wang linker 88.

The ArgoGel bound fluorinated Wang linker 88 was coupled to an ArgoGelresin, which was successfully used in solid-phase synthesis of a set of serine-based glycolipid analogs (Scheme 5).114 By taking advantage of being bothacid- and base labile linker, the glycolipids 90 and 91 could be cleaved byacidolysis, whereas 94 could be released from the resin by saponification.One of the most critical events in solid-phase synthesis is the purification ofthe products after cleavage. Therefore, it is a great advantage to have thepossibility to purify the glycolipids in both their protected unpolar anddeprotected polar form. This allowed the partially deprotected glycolipidderivatives 91, 92 and 94 to be purified by flash column purification prior tothe final deprotection steps.

OH

FO

HO

O

FHOOEt

O

O

FHOOH

O

1) BH3:DMS,B(OMe)3, THF2)DBU, BrCH2CO2Et,MeCN

LiOH (aq)

THF/MeOH

86 87 (57% over two steps)

88 (90%)

37

Scheme 5. Solid-phase synthesis of serine containing glycolipids.

O

FHO HN

OArgoGel

–134.5 ppm

=HO F

HN O

OO

O

HN

OO

oFBnOoFBnO

OO

mFPh

HN OH

OO

O

HN

OO

HOHO

OHHO

O

OHHO

OHHO

94F

HN O

n

OO

O

HN

O

m

90 n = m = 191, n = 7, m = 11

F

O

pFBzOpFBzO

OpFBzpFBzO

HN OH

n

OO

O

HN

O

m

92, n = m = 193, n = 7, m = 11

95

1) TFA, H2O, 60 °C2) NaOMe/MeOH

1) LiOH, H2O, THF2) H2, Pd/C, HOAc

89

38

5.3 Selenium containing resins

Selenium substituted resins was introduced by Mitchels et al. for use asoxidation agents.132 Later was this methodology introduced as an attractivemethod for traceless synthesis of aryl ethers, by utilizing radical mediatedcleavage.133 The benefits of using resin-linked selenium include theirodorless nature and reduced toxicity, since volatile organoseleniumcompounds are notorious to exhibit an unbearable odor and are in generalhighly toxic.

The method was further developed by the group of Nicolaou toinclude an oxidative cleavage protocol (Scheme 6).134 Lithiation of apolystyrene resin and subsequent treatment with dimethyl diselenide gaveresin 96. Exposure of 96 to bromine resulted in quantitative formation ofresin 97, as judged by consumption of bromine. Lithiation of 97, followed bysubstitution of iodo alkyl derivative 99 gave resin 100. Cleavage from theresin with radical conditions furnished the alkyl compound 101 in 89%yield, while oxidation with aqueous hydrogen peroxide gave the olefinicproduct 102 in 78% yield.

The selenium-based resin 102 was later successfully employed insynthesis of natural product libraries,40,135 in where the products werecleaved from the resin by oxidation in order to introduce conjugated doublebounds.

Scheme 6. Synthesis and cleavage of selenium-based resin 100.

Polystyrene

1) nBuLi2) (MeSe)2

SeMe SeBr

SeLi Se OTBDPS

I(CH2)5OTBDPS

Br2

LiBH4

OTBDPS OTBDPS

96 97

98 100

101 (89%) 102 (78%)

Bu3SnH / AlBNPPh3, 110 °C H2O2, THF, RT

99

39

5.3.1 Synthesis of a fluorinated selenide-linker forpreparation of n-pentenyl glycosides

Selenium containing resins provide tethers that are robust and resistant tomany chemical conditions in organic chemistry and in particular inoligosaccharide synthesis, for instance hard Lewis acids, alkylation agents,strong bases, nucleophiles and reductive conditions.136 The need for novellinkers for oligosaccharide synthesis that are compatible with protectivegroup and glycosylation protocols used in oligosaccharide prompted us toexploit the use of a selenide-linker for preparation of pentenyl glycosides.

Scheme 7. Initial route for preparation of linker 110.

An initial effort to prepare a fluorine-labeled selenide-linkerfollowing the procedure of Nicolaou et al.134 was not successful in our hands(Scheme 7). The seleno-ether linkage of 106 was however obtained inmoderate yield (45%) by quenching the lithiated compound of 103 withdimethyl diselenide. Having the phenol functionality of 103 protected with abenzyl group (104) or an ArgoGel resin (105) did not result in anyselenation. Unfortunately, the seleno-ether 106 did not react in the followingbromination step. However, by modifying the procedure a successful routewas achieved (Scheme 8). This route took advantage of the successfulselenation step in the previous route using dimethyl selenide. Thus thecomplicated bromination step was circumvented by a convergent approachin which the diselenide 111 was utilized. Subjecting the iodo alkyl derivative99134 to sodium hydride and elemental selenium powder gave diselenide 111

RO

Br

F

103 R = H104 R = Bn105 R = ArgoGel

1) tBuLi, –78 °C2) (MeSe)2, –78 °C → rt

RO

SeMe

F106 R = H (45%)107 R = Bn (0%)108 R = ArgoGel

Br2Et2O HO

SeBr

F

1) LiBH42) 99 TBDPSO Se

OH

F

110

109

40

in 51% yield. Finally, the linker building block 112 was obtained in 23%overall yield by quenching the dilithiated compound of 103 with diselenide111.

Scheme 8. Synthesis of linker 112.

Encouraged by these results, the synthetic route was furtherimproved by another linear strategy (Scheme 9). In this approach compound103 was treated with tBuLi followed by addition of elemental selenium toaccess the corresponding selenide anion. At this stage a significant amountof selenium was not consumed. However, upon addition of the iodo alkylderivative 99 the remaining selenium was consumed, probably due toformation of iodide. Thus, addition of a stoichiometric amount of 99furnished the linker building block 112 in excellent yield (97%) in a one-potreaction. Furthermore, the linker was extended with a carboxylicfunctionality, in order to simplify and ensure quantitative anchoring to thesolid phase via an amide linkage. Thus, alkylation of the phenol functionalityof 112 with ethyl bromoacetate using potassium carbonate as base in thepresence of tetrabutylammonium iodide gave compound 113 in 97% yield.Ester hydrolysis of 113 with aqueous LiOH was accomplished inquantitative yield and the desired linker building block 114 could beobtained in 94% overall yield starting from 103. Since the commercial resinwas contaminated with hydroxy functionalities, the linker building block 114was attached to the ArgoGel-NH2 resin by employing standard peptidecoupling conditions in order to avoid O-acylation. Capping of the remaininghydroxyl functionalities with acetic anhydride followed by cleavage of thesilyl protective group of 115 with TBAF liberated the alcohol functionalityto provide resin 116.

OTBDPSI

1) NaH 70 °C2) Se (s), rt

DMF

THF

1) tBuLi, –78 °C2) 111, –78 °C → rt TBDPSO Se

OH

F

111 (51%)

112 (45%)

99

OH

FBr

103

(TBDPSO(CH2)5Se)2

41

Scheme 9. Improved synthesis of linker building block 112 and attachmentto ArgoGel-NH2.

5.3.2 Establishing glycosylation conditions and cleavagefrom the linker

Glycosylation of the selenide-linker resin 116 was subsequently investigatedwith a glycosyl trichloroacetimidate 57, a glycosyl fluoride 58 , and aglycosyl sulfoxide 59a (Scheme 10). Submitting resin 1 1 6 to fiveequivalents of glycosyl trichloroacetimitate 57 in the presence TMSOTf (0.6equivalents) at room temperature using dichloromethane as solvent affordedthe β-galactoside 117 in quantitative yield, as determined by comparison ofthe 19F resonances of the protective groups (pFBz = 105.0 and 105.4 ppm;pFPhCHO2 = 113.3 ppm) with the linker signal (133.3 ppm).

TBDPSO Se

O OR

O

F

TBDPSO Se

OH

F

OH

FBri) tBuLi, THF, –78 °Cii) Se (s), –78 °C → rtiii) 99, –78 °C → 0 °C

K2CO3, Bu4NI,BrCH2CO2Et

LiOH (aq)MeOH/THF

103 112 (97%)

MeCN

THF

RO Se

OHN

O

F

113, R = Et (97%)

115, R = TBDPS

1) ArgoGel-NH2, HOAt, DIC, DMF2) Ac2O, Pyridine

114, R = H (quant)

TBAFTHF 116, R = H

42

Scheme 10. Glycosylations of linker-loaded resin 116.

Glycosyl fluoride 58 was significantly less reactive than thecorresponding trichloroacetimidate and a variety of reaction conditions wereemployed in attempts to optimize the glycosylation (Scheme 10). An initialeffort to glycosylate resin 116 with 58 under activation with BF3×Et2O didnot produce any glycosylation product. Glycosylation of 116 with 5 8activated by TMSOTf (0.6 equivalents) resulted in only ∼ 20% conversion ofresin 116, and it was found that an excess (four equivalents) of the promoterTMSOTf was required for the reaction to proceed. However, under thesesomewhat harsh conditions the reaction was hampered by partial cleavage ofone of the benzoyl groups. Employing a mixture of dichloromethane andether prevented the acidolysis and thus, glycosylation of resin 116 with 58(four equivalents) at room temperature resulted in complete formation ofresin 117 as determined by analysis of 19F NMR spectra.

An attempt to glycosylate resin 116 with glycosyl sulfoxide 59a atroom temperature was carried out (Scheme 10). Following the proceduredescribed by Kahne et al.137 the sulfoxide 59a was activated at –78 °C with

HO Se

OHN

O

F

116

O

pFBzOpFBzO

OO

pFPh

X57 X = OC(NH)CCl358 X = F

TMSOTf, rt

O

pFBzOpFBzO

OO

pFPh

SO

PhMe

O Se

OHN

O

O

pFBzOpFBzO

OO

pFPh

117

F

Tf2O, collidineCH2Cl2

No glycosylation59a

43

triflic anhydride in the presence of collidine in dichloromethane. Themixture was subsequently transferred to resin 116, which was allowed toreact at ambient temperature. 19F NMR spectroscopy of the resulting resinrevealed that no glycosylation had occurred. Since only degradation of thelinker was observed, this glycosyl donor was not further evaluated.

Scheme 11. Evaluation of cleavage with H2O2 and mCPBA.

Efforts to establish cleavage conditions for resin 117 and modelcompound 118138 were initially made with aqueous H2O2 or mCPBA asoxidizing agents (Scheme 11). When resin 117 was treated with aqueousH2O2 only stable by-products were formed,139 i.e. they did not undergo syn-elimination. Two major peaks appeared in the 19F NMR spectra (–130.0 and–133.2 ppm, respectively), that were suspected to be the correspondingselenone 120 and seleno hydrate 121 (Figure 1). Oxidation of 117 and 118with mCPBA resulted in formation of several by-products, as judged by 19FNMR spectroscopy, that indicated two major peaks at –130.0 and –133.2ppm. To our delight, washing the oxidized derivative of resin 118 withdichloromethane followed by refluxing in dichloromethane resulted inpartial cleavage of the resin. After repeated heating in a sealed tube at 120°C for one hour in CH2Cl2 using microwave assisted heating, the pentenylglycoside 119 could be isolated in 15% total yield after flash columnpurification. Furthermore, microwave assisted heating of the partially syn-eliminated resin at 100 °C for one hour in presence of mCPBA usingdichloromethane as solvent did not increase the yield of 119 and the two

O Se

OHN

O

O

pFBzOpFBzO

OO

pFPh

117

F

H2O2 (aq)or

mCPBANo syn-elimination

O Se

OHN

O

O

pFBzOpFBzO

OpFBzpFBzO

118

F

1) mCPBA2) Heat

CH2Cl2OO

pFBzOpFBzO

OpFBzpFBzO

119

44

major 19F resonances at –130.0 and –133.2 ppm remained intact. Refluxingthe resin oxidized counterpart of 118 in toluene using a Dean-Starke trapwas performed in an attempt to force the equilibrium of the assumedselenohydrate towards the corresponding selenoxide, but this did notproduce any detectable cleavage from the resin.

Figure 1. Structures of possible by-products formed from the oxidation ofresin 117.

From these results it could be concluded that 1) anhydrousconditions is required in order to avoid formation of the selenohydrate, 2) aweak oxidation reagent is needed to avoid formation of selenone, and 3) heatis necessary to induce decomposition of the selenoxide. Anhydrous tBuOOHin toluene was expected to fulfill these requirements and it was anticipatedthat the stability of the resin-linked selenoxide is an advantage in solid-phasesynthesis, since it would permit removal of the oxidizing reagents prior toreleasing the product from the solid support. Thus, treatment of selenide 117with anhydrous tBuOOH in toluene for four hours resulted in completeconversion of selenide functionality to yield the corresponding seleneoxide122 (Scheme 12). The 19F signal of resin 122 shifts to –130.2 ppm (Scheme12, spectrum b) and the linker signal at –133.3 ppm completely disappears(Scheme 12, spectrum a), confirming completion of the oxidation.Importantly, no decomposition of the selenoxide-resin 122 was detectedwhen the resin was subjected to oxidation at room temperature over night.The oxidation reagent was subsequently removed by washing the resinthoroughly with toluene. Finally, the selenoxide-resin underwent complete

O Se

OHN

O

O

pFBzOpFBzO

OO

pFPh

FO O

O Se

OHN

O

O

pFBzOpFBzO

OO

pFPh

FHO OH

Selenone 120

Seleno hydrate 121

45

syn-elimination when heated in toluene at 90 °C for four hours, and the puren-pentenyl glycoside 123 was isolated in 92% yield, based on the initialloading capacity of the resin. Inspection of the 19F NMR spectra of resin 124revealed complete cleavage, since the signal from the selenoxide resin 122 at–130.2 ppm had disappeared and only a signal derived from the presumedresin-bound selenic acid 124 at –132.4 ppm was detected (Scheme 12,spectrum c).

Scheme 12. Oxidation of resin 1 1 7 and thermal syn-elimination ofselenoxide 122.

Seeberger and co-workers developed a promising linker for solid-phase preparation of n-pentenyl glycosides,58 which is gently cleaved fromthe solid phase in high yield by olefin cross-metathesis using Grubb´scatalyst (Scheme 9). The linker was prepared from mono-DMT-protectedoctenediol 126 on a Merrifield resin 125. Acidolytic removal of the DMTgroup gave 127 and at this stage the resin loading could be estimated by acolorimetric assay. The linker proved to be compatible with hard Lewis acidactivation of both a glycosyl phosphate and trichloroacetimidate. The initialstudy included a highly efficient solid-phase synthesis of trisaccharide 132by a iterative TBS deprotection / glycosylation protocol employing a

46

glycosyl phosphate 128 as glycosyl donor under activation of TMSOTf. Inthe final step the trisaccharide was released from the solid phase bytreatment of the resin with Grubb´s catalyst under an atmosphere of ethyleneat room temperature to give the n-pentenyl trisaccharide 132 in 56% yield. n-Pentenyl glycosides are a well known as glycosylation reagent and havebeen used in construction of complex oligosaccharide structures.140 I naddition, the n-pentenyl functionality can be further oxidized to thecorresponding aldehyde or carboxylic acid to provide a handle forbioconjugation to carrier proteins or microchips for high-throughputscreening.141,142

The linker concept was successfully adapted to automation andseveral challenging natural occurring oligosaccharide structures have beenprepared using reactive glycosyl phosphates as donors in combination withan efficient and reliable protective group pattern in where the temporaryacetyl, and levulinoyl (Lev) protective groups were employed.5,6

Scheme 13. Preparation of linker 127 and solid-phase synthesis of n-pentenyl trisaccharide 132.

Cl

HODMTO

1) NaH,

2) NaH, MeOH3) 3% dichloroacetic acid

HOO

OPO

OBuOBu

BnOTBSO

BnOPivO

TMSOTf, –50 °C1h

OO

OBnO

ROBnO

PivO

129 R = TBS130 R = H

128, TMSOTf, –50 °C,1h

OOBnOO

BnO PivO

OBnO

TBSOBnO

PivO

OOBnOO

BnO PivO

OBnOOBnO PivO

1) TBAF2) 128, TMSOTf–50 °C, 1h.

3) Ethylene,

RuPhPCy3

PCy3

ClCl

OBnO

TBSOBnO

PivO

125 127

128

131132 (56%)

126

TBAF

47

5.4 Conclusions

We have shown that fluorine labeled linkers in combination with fluorinatedprotective groups is most useful for optimizing solid-phase glycoconjugatesynthesis. Two different linkers 79 and 112 were prepared and evaluated insolid-phase synthesis, which were concluded in glycosylations of serinecontaining substrates and in synthesis of a n-pentenyl glycoside. Theselenide linker 116 proved to be very efficient for installing terminal isolatedolefins. Since the selenoxide 122 did not decompose during oxidation inroom temperature, the resin could be washed with toluene before releasingthe n-pentenyl glycoside from the solid phase by thermal syn-elimination.Hence, the n-pentenyl glycoside 123 could be isolated in excellent purity andyield without purification.

48

6 Solid-phase synthesis of α-Galtrisaccharides (IV, V)

6.1 Xenotransplantation epitopes

The increased demand for organs and tissues in surgery has resulted inresearch to fill the need of organs by xenotransplantation i.e. transplantationof organs and tissues between animal species.143 The common pig has beenidentified as the most suitable donor candidate for humans. Initial attemptsto transplant porcine organs into apes were hampered by destruction andhyper acute rejection of the organ. This is a consequence of the largeamounts of circulating anti-Gal IgG antibodies that are naturally occurring inthe human body, as well as in other primates. α-Gal epitopes are abundant inlarge quantities on the surface of all mammalian endothelial cells other thanthose of humans and apes. The terminal Galα–3Gal moiety (132) is acommon structure element of these α-Gal epitopes (Figure 1). Approaches toavoid hyperacute rejection and achieve accommodation inxenotransplantation includes attachment of oligosaccharides containing theGalα–3Gal motif to affinity columns in order to remove the anti-Gal IgGantibodies by filtration before the new organ is installed.

Several research groups have previously prepared derivatives of α-Gal epitopes employing various glycosylation methods.87,110,111,144-149 Themost challenging step in the synthetic sequence includes installation of aterminal α-linkage, since the stereochemical outcome of this reaction isdifficult to predict due to the lack of neighbouring group participation at theC-2 position of the glycosyl donor. Thus, routes involving chemoezymaticglycosylations have been established in order to achieve the “difficult” α-glycosidic bond with absolute steroselectivity.150-155

49

Figure 1. α-Gal saccharide structures.

6.2 Protective group strategy

The often time-consuming preparation of carbohydrate building blocks is amajor hurdle in solid-phase oligosaccharide synthesis. This problem arisesfrom the structural complexity of naturally occurring oligosaccharides,which for their assembly may require several different glycosyl donorscarrying unique protective group patterns. Thus, major efforts have beendirected towards development of effective protective group protocols forsolid supported oligosaccharide synthesis.23,32 A feas ib leprotection–deprotection strategy was introduced by Wong et al. where thechloroacetyl (ClAc), p-methoxybenzyl (PMB), levulinoyl (Lev) and tert-butyldiphenyl silyl groups were demonstrated to be a promising orthogonalset of temporary protective groups.30 Recently the well known 9-fluorenymethoxycarbonyl (Fmoc) protective group, which is one of the mostused Nα-amino protecting group in peptide synthesis,156 was also establishedas a versatile hydroxyl protective group in solid-phase oligosaccharidesynthesis.31,37,56,110,157-160

OOHO

OH

OHO

OHO

HOOHO

HOHO

HOOH

OH

OOHO

NHAc

OHO

OHO

HOOHO

HOHO

HOOH

O

O

OHO

HOOHO

HOHO

HOOH

OOHO

OH

OHO

OH

HOOH

OH

OOHO

NHAc

OHO

OHO

HOOHO

HOHO

HOOH

OH

α-Gal epitope 132

α-Gal trisaccharide 133

α-Gal trisaccharide 134

α-Gal pentasaccharide 135

50

The protective groups described above are either deprotected bybase, acid, or nucleofiles and have been successfully employed for iterativeassembly of oligosaccharides. Regarding to combinatorial carbohydratesynthesis, solid supported preparation of complex branched oligosaccharideswould benefit from an extended temporary/permanent protective grouppattern, which is orthogonal to the existing protective group strategies. The4,6-O-benzylidene acetal is one protective group which fulfills theserequirements and appears to be promising for solid-phase synthesis.161,162

This classical protective group has proved to be most useful in solution-phase synthesis and has the advantage of protecting two hydroxyl groups atthe same time. The possibility to perform a mild reductive deprotection togive either of the corresponding 6-O-benzyl163 or 4-O-benzyl162,164 ethers isan advantage.

6.3 Solid-phase assembly of α-Gal trisaccharides

The solid-phase strategy for preparation of trisaccharides 136 and 137(Scheme 1), involved three key steps: 1) regioselective reductions of afluorine-labeled 4,6-O-benzylidene acetal, 2) β-galactosylations, and 3)installation of an α-galactosidic linkage.

The retrosynthetic analysis led to resins 138 and 139, in which thelinker 79 provided an acid stable attachment of building blocks 72 and 73 tothe resin. The fluorinated linker also served as an internal analytical 19F-marker throughout the synthesis. The thiogalactosyl donors 62 and 65carried fluorinated protective groups in order to allow analysis of theglycosylations and protective group manipulations on solid-phase with gel-phase 19F NMR spectroscopy.

51

Scheme 1. Retrosynthesis of target α-Gal trisaccharides 136 and 137.

Immobilization of the carboxyl functionality of glucoside 73 andglucosamine 75 (each three equivalents) on linker-resin 79 via an estercoupling afforded resin 140 and 1 4 1 (Scheme 2). Integration of theresonances from the fluorine atoms in the linker (−134.4 ppm), the 4,6-O-p-F-benzylidene group (−113.2 ppm) and the two o-F-benzoyl groups (−109.0and −110.1 ppm) revealed the yield to be quantitative for resin 140 (Scheme

O

ORoFBzO

OOpFPh

OH

O

O

OoFBnoFBnO

OO

SPhMe

mFPh

O

OpFBzFmocO

OO

SPhMe

oFPh

6562

OHO

FO

NH

HO

F

72, R = OoFBz73, R = NHAc 4

ArgoGel®

=

79

OOoFBzO RO

OpFBnO

pFBzOO

OO

oFPh

O

oFBnOoFBnO

OO

mFPh

F

Regioselective reduction of a benzylidene acetal

α-glycosylation

β-glycosylation

138 R = OoFBz 139 R = NHAc

136, R = NHAc 137, R = OH

OOHO

ROO

OH

OHO

OHO

HOOHO

HOHO

HOOH

52

2, Figure a). Resin 141 was also formed in quantitative yield as estimated by19F NMR spectroscopy. Resin-bound 140 and 141 was then converted intosecondary alcohols 142 and 143 via regioselective opening of the 4,6-O-p-F-benzylidene acetal under reductive conditions using NaCNBH3 in a solutionof HCl(g) in diethyl ether.162 After optimization using a model system it wasestablished that a large excess of NaCNBH3 (> 30 equivalents) was requiredin order to prevent acidolysis of the benzylidene acetal. Thus, performing thereaction with 50 equivalents NaCNBH3 led to the complete disappearance ofthe resonance from the 4,6-O-p-F-benzylidene group (-113.2 ppm) of 140,and the appearance of a single, new signal originating from the p-F-benzylether (−115.2 ppm) of resin 142 (Scheme 2, Figure b). Also reduction ofresin 141 proceeded equally well to give 143. Importantly, a washingsequence that includes water was required to remove salts from the resins.

53

Scheme 2. Solid-phase synthesis of trisaccharides 138 and 139.

54

The hydroxyl group of 142 and 143 was subsequently glycosylatedwith thiogalactoside 65 (four equivalents) using N-iodosuccinimid (NIS) andtrifluoromethane sulfonic acid (TfOH) as promotor system. The 19F NMRspectrum of resin 144 revealed that the glycosylation had not reachedcompletion since resonances from unreacted 142 were still visible, and theyield could be estimated to ~50%. As revealed by 19F NMR spectroscopy, theglycosylation had to be repeated twice to achieve complete conversion of 142to give resin bound disaccharide 144. The p-F-benzoyl group, which is inclose proximity to the newly formed glycosidic bond, displayed tworesonces; a major one at –105.1 ppm and a minor one at −105.9 ppm in a 2:1ratio. All other resonances from the product were also split into 2:1 doubletsand it was assumed that this reflected the formation of an anomeric mixture(Scheme 2, Figure c). To our surprise the corresponding glucosamine resin143 was found to be significantly more reactive than 142 on glycosylation.This observation was unexpected since glucosamines are known to be pooracceptors.165 Thus, using the same conditions and reaction time as for 142, a70% conversion of resin 143 was obtained after the first glycosylation.Repeating the reaction once furnished resin 145 in quantitative yield in a 2:1ratio. Next, the Fmoc group at HO-3´ in 144 and 145 was removed bytreatment with triethylamine in dichloromethane to afford resins 146 and147.

α-Glycosylation of resins 146 and 147 with thiogalactoside donor 62was carried out at reduced temperature under activation by NIS and TfOH togive resins 138 and 139. The quantitative formation of resin 138 as a singleanomer was evident from the appearance and integrals of the signals in the19F NMR spectrum of 138 (Scheme 2, Figure d). Resin 139 was formed in80% yield. Cleavage of 138 and 139 from the solid phase under basicconditions with LiOH/H2O/THF allowed isolation of the trisaccharides 148α(~14%) and 148β (37%) and the corresponding trisaccharides 149α (~8%)and 149β (27%) after purification with reversed phase HPLC (Scheme 3).The remaining fluorinated benzylidene acetals and benzyl groups of 148βand 149β were removed by hydrogenation to give the α-Gal trisaccharides136 and 137 in 36% and 25 % overall yield, respectively.

55

Scheme 3. Cleavage from the solid phase and deprotection of 138 and 139.

Recently, Wu and Schmidt reported a similar approach for solid-phase synthesis of α-Gal trisaccharide 157,160 in which an O-Fmoc protectedglycosyl donor was used (Scheme 4). Moreover, they also encounteredproblems in the β-glycosylation step, where the acceptor 152 was found tobe unreactive. This led to an incomplete reaction after two consecutiveglycosylations. However, only the desired β-linkage of 154 was obtainedand by capping the unconsumed acceptor with benzoyl isocyanate the finalpurification of the target trisaccharide was simplified. The final α -glycosylation with 155 resulted in an anomeric mixture of 156 (α/β = 8/1)and subsequent cleavage allowed the protected trisaccharide 157 to beisolated in 25% overall yield after acetylation and HPLC purification.

56

Scheme 4. Solid-phase synthesis of protected α-Gal derivative 157.

In a synthesis carried out by Drinnan et al.166 the inadequate neighbouringgroup participation of an acetyl or a p-chlorobenzoyl protecting group at theC-2 position of 158 and 159 also resulted in poor β-selectivity (Scheme 5).However, they found that exchanging the C-2 protecting group for a pivaloylor a benzoyl group significantly increased the β −selectivity to allowpreparation of the α-Gal epitopes 167 and 168 in high overall yields.

OFmocOBnO

NPhthO CCl3

OBn

NH

O

BzOFmocO

OO

Ph

HOO

O

PS 1) TMSOTf, –40 °CCH2Cl2

2) CH2Cl2/Et3N OHOBnO

NPhthO

OBn

O

CCl3

NH

1)TMSOTf, –40 °CCH2Cl22) BzNCO

OOBnO NPhth

O

OBnO

BzOFmocO

OO

Ph

OBnO

BnOO CCl3

OBn

NH

BnO2)

TMSOTf, –40 °CCH2Cl2

1) CH2Cl2/Et3NOO

BnO NPhthO

OBnO

BzOO

OO

Ph

OBnO

BnO

OBnBnO

1) NaOMe, CH2Cl2/MeOH2) Ac2O, Pyridine

OOBnO NPhth

O

OBnO

AcOO

OO

Ph

OBnO

BnO

OBnBnO

OAc

= HO150

151

152

153

154155

156

157 (25%)

+

57

Scheme 5. Evaluation of neighboring group participation of galactosyldonors 158 – 161 in β-glycosylation of 162.

6.4 Conclusions

A route for solid-phase synthesis of α -Gal trisaccharides has beenestablished. By using a fluorinated linker and fluorinated protective groupseach step could be analyzed directly on the resin with 19F NMRspectroscopy. This enabled us to establish conditions and confirm theoutcome of complex steps in the synthetic sequence. The 4,6-O-benzylideneacetal proved to be a robust and useful protective group also for solidsupported synthesis. The versatility of this protective group was illustrated inreduction of the acetals 140 and 141 to the corresponding 6-O-benzylderivatives, which were achieved in quantitative yields. Thus, glycosyldonors bearing a benzylidene acetal in combination with a temporary Fmocor Lev group constitutes an excellent orthogonal protective group pattern,which should allow selective protective group manipulations of eachposition of the glycoside. In view of these advantages, the 4,6-O-benzylideneacetal has potential to be adapted for the use in automated procedures.

The poor reactivity and stereoselectivity in the β-glycosylations, isprobably originating from a mismatch between the glycosyl acceptors 142

58

and 143 and donor 65. According to the results obtained by Drinnan et al.166

and Wu et al.,160 the β-selectivity could be enhanced by employing asterically hindered pivaloyl group at the C-2 position of the donor or byusing a more reactive acceptor. In view of these investigations exchangingthe p-fluorobenzoyl group at the C-2 position of 65 for a pivaloyl group, incombination with replacement of the electron withdrawing o-fluorobenzoylgroup at the C-3 position of 72 and 73 with a benzyl group, would perhapsenhance the reactivity and the stereoselectivity. These results clearlyillustrate that the well-known observation that the reactivity of both theacceptor and the donor is controlled by the nature of the protective groups toa large extent, and that every glycosylation is unique. This issue has emergedto a major burden in the struggle to find standardized conditions for use inautomated oligosaccharide synthesis. To tackle this problem, benzylatedglycosyltrichloroacetimidates and phosphates, in combination with atemporary levulinoyl or pivaloyl group as the only deactivating protectivegroups, have been used to maximize the reactivity of both the glycosyl donorand the acceptor.5,6

59

7 Solid-phase peptide glycosylations (I)

The classical building block approach, i.e. stepwise assembly of amino acidsand glycosylated amino acid building blocks, is at present the most reliableand widespread method for solid-phase synthesis of glycopeptides.15,16,29

However, a strategy based on glycosylation of hydroxyl groups of serine,threonine, tyrosine, hydroxyproline, or hydroxylysine residues in apreformed peptide is very appealing, since it offers possibilities to creatediversity using different glycosyl donors at a late stage in the synthesis. Inaddition, such a strategy may be of major importance towards futureresearch activities that are directed for construction of large glycopeptidesand glycoproteins. These compounds may be used as biotechnological toolsin efforts to understand the function of glycoproteins, as well as fordevelopment of novel biopharmaceuticals.

7.1 Reactivity in solid-phase peptide glycosylations

The origin of the poor reactivity of peptides in glycosylations is mainlyattributed to the low solubility of the peptide under the conditions used forglycosylation and/or adsorption of the Lewis acid by the peptide-amidefunctionalities.167 In addition, the reactivity of serine and threonine inglycosylations has been proposed to be particularly low due tointramolecular hydrogen bonding between the hydroxyl group and theacetamide hydrogen.127 Hence, attempts to find protocols towards direct O-glycosylation of resin bound peptides are fairly rare and initial attempts toglycosylate small peptide fragments linked to a polystyrene resin were verylow yielding (4 – 5% yields).168-170 However, a few successful glycosylationsof peptides linked to a VERSABEAD-O resin (POEPOP) by usingthrichloroacetimidates as glycosyl donor have been reported.72,171,172 Thus,pentapeptides 169 – 171 were attached to a VERSABEAD-O resin through aRink linker, an Uv-labile linker and a hydroxymethyl benzoic acid (HMBA)

60

linker (Scheme 1). Glycosylations of the peptides 169 – 170 withperbenzoylated galactosyl thrichloroacetimidate 172 (eight equivalents)using 0.1 equivalents TMSOTf at room temperature gave the glycosylatedresins 173 – 175 in quantitative yield, as estimated by cleavage from theresin followed by tlc analysis. Remarkably, the acid sensitive Rink linker of173 was stable under these rather harsh conditions and no problem withacidolytic cleavage was reported. Cleavage from the polymeric supports wasachieved by either acidolysis with TFA, photolysis, or treatment withsodium methoxide in methanol. Subsequent deprotection of the remainingbenzoyl groups with sodium methoxide in methanol gave the glycopeptides176 and 177 in high overall yields.

Scheme 1. Glycosylations of peptides attached to VERSABEAD-O resins.

61

7.1.1 Solid-phase glycosylations of pepetides containingserine

In an effort to find conditions for glycosylation of resin-boundpeptides, tripeptide 178 , pentapeptide 179, and heptapeptide 180 wereprepared on the linker-loaded ArgoGel resin 79 (Scheme 2). Our previouspositive results using thioglycosides as glycosyl donors prompted us tofurther evaluate their use also in solid-phase peptide glycosylations.Furthermore, since we previously found that N-iodosuccinimide (NIS) incombination with a catalytic amount of trifluoromethanesulphonic acid(TfOH) was superior to dimethyl(methylthio)sulfonium triflate (DMTST)(cf. section 5.1.1), we decided to use NIS and TfOH as promoter system.Hence, under the influence of NIS and TfOH peptides 178 – 180 wereglycosylated in a parallel manner with thioglycoside 61 (two equivalents) atroom temperature. Integration of the fluorine resonances revealed ~30%glycosylation of all the three peptides. The glycosidic linkages were assumedto have been formed as the desired β -anomers due to neighboringparticipation of the acyl functionality at the C-2 position of 61. Repeating theglycosylation once under identical conditions increased the yields to ~60%and after a third glycosylation the glycopeptides 181 – 183 were formed in~80% yields (Scheme 3, spectra a – c).

Scheme 2. Solid-phase peptide synthesis of resins 178 – 180.

The length of the resin bound peptides gave rise to a considerablevariation in 19F NMR spectral quality when recorded in CDCl3. Thus, highquality 19F NMR spectra of the resin bound tripeptide 178 and glycosylatedtripeptide 181 could be obtained in CDCl3 (Scheme 3, spectrum a), whereas

ArgoGel®HN

OO

HO

F

79

=

178

179

180

NH

HO

NH

HO

NH

HO

pFBz–Gly Val–O

O F

Val–Phe–O

O

pFBz–Ala–Gly

F

Val–Phe–Leu–O

O F

pFBz–Phe–Ala–Gly

HO F

62

the penta- and heptapeptides 179 and 1 8 0 and the correspondingglycopeptides 182 and 183 furnished inadequate spectra both in CDCl3 andbenzene-d6. This was assumed to originate from aggregation of resin boundpeptides and glycopeptides due to the low polarity of CDCl3 and benzene-d6.

167 Indeed, when the solvent was changed to DMSO-d6 the aggregationwas disrupted and well resolved 19F NMR spectra could be obtained for boththe resin bound peptides 179 and 180 and glycopeptides 182 and 183(Scheme 3, spectra b and c).

Scheme 3. Stereoselective glycosylations of resin bound peptides 181 – 183.

Interestingly, even though the glycosylations were performed in CH2Cl2,which is a less polar solvent than CDCl3, the yields of the glycopeptide 181– 183 were all the same after each consecutive glycosylation. Thus, ourresults indicate that the length of the resin bound peptides is not crucial for

63

the outcome of the glycosylation when the reactions is carried out with athioglycoside under activation by NIS and TfOH.

7.1.2 Conditions for cleavage from the solid phase andremoval of protective groups

Cleavage from the solid phase was initially carried out by hydrolysis withaqueous LiOH in THF. Careful monitoring of the cleavage of resin 181 with19F NMR spectroscopy and LC-MS analysis of the reaction mixture revealedthat the hydrolysis was hampered by β-elimination of the carbohydratemoiety from the peptide (Scheme 4). To our delight this could becircumvented by first releasing the glycopeptides from resins 181 – 183 byacidolysis with TFA at 60 °C,116 which resulted in simultaneous removal ofthe benzylidene group. Treatment with LiOH (20 mM) in methanol thenremoved the remaining p-fluorobenzoyl groups. Following this approach, noβ-elimination was observed and the glycopeptides 185, 186 and 187 couldbe isolated in 27, 26 and 29% overall yields after HPLC purification.Migration of the 2- and 3-O-fluorobenzoyl groups to the 6-OH position,where the saponification will occur at significantly faster rate, may explainthe increased reaction rates in hydrolysis and the lack of β-eliminations ofthe cleaved intermediates.173 Furthermore, once the carbohydrate moiety isdeprotected the glycopeptide is more resistant to β-elimination since thepartial negative charge on the deprotected carbohydrate moiety prevents theformation of a negative charge on the O-1 position, especially when thecharge is located at the O-2 position.

64

Scheme 4. Cleavage and deprotection of glycopeptides from solid phase.

7.2 Conclusions

By employing a thiogalactoside as glycosyl donor, conditions wereestablished allowing glycosylations of a tri-, a penta-, and a heptapeptide inhigh yields. Furthermore, it was found that neither the length nor aggregationof the peptides influenced the yields of the glycosylations. When comparingour results with a previous study,170 in which a similar peptide (Fmoc-Thr–Val–Val) linked to a polystyrene resin was glycosylated in 4% yieldusing a peracetylated allylcarbamoyl glucosyl donor, the choice of resinseems to be an important variable affecting the outcome of glycosylation.Further studies comparing both the influence from the resins and theglycosyl donors would provide valuable insight into the factors affecting theglycosylation of peptides.

181

182

183

O

pFBzO

OO

mFPh

pFBzO

NH

O

pFBz–Gly Val–O

O F

NH

O

Val–Phe–O

O

pFBz–Ala–Gly

F

O

pFBzO

OO

mFPh

pFBzO

NH

O

Val–Phe–Leu–O

O F

pFBz–Phe–Ala–Gly

O

pFBzO

OO

mFPh

pFBzO

1) TFA, H2O2) LiOH (aq), MeOH

185

186

187

O

OH

OHHO

HO

NH

O

pFBz–Gly Val–OH

O

NH

O

Val–Phe–OH

O

pFBz–Ala–Gly

O

OH

OHHO

HO

NH

O

Val–Phe–Leu–OH

O

pFBz–Phe–Ala–Gly

O

OH

OHHO

HO

1) TFA, H2O2) LiOH (aq), MeOH

1) TFA, H2O2) LiOH (aq), MeOH

184

NH

pFBz–Gly Val–OH

OLiOH (aq)

MeOH

65

8 Future perspectives

The work described in this thesis, with strong emphasis in “on-resin”monitoring with gel-phase 19F NMR spectroscopy, has provided new insightsinto solid-phase glycoconjugate synthesis. It was shown that fluorinated that19F signals of protective groups are spread over a wide spectral range. Thisallowed analysis of the reaction progress, for instance in the synthesis of α-Gal trisaccharides on solid phase without encountering severe spectraloverlap. I believe that by using the fluorinated protective groups presentedherein sparingly, this technique should allow convenient monitoring insynthesis of larger oligosaccharides. For this purpose additional fluorinatedtemporary protective groups should be most useful to expand the protectivegroup “tool box”. For instance, fluorinated variants of the para-methoxybenzyl (F-PMB), the benzyloxy carbonyl (F-Cbz), the para-methoxyphenyloxy carbonyl (F-PA), the pivaloyl (FPh-Piv), and the silyl(FPh-R2Si) group constitute an interesting set of future orthogonal protectivegroups (Figure 1).

Figure 2 Chemical structures of fluorinated variants of protective groups.

Furthermore, finding nearly neutral conditions for solid-phaseglycosylations of peptides that contain amino acid residues protected withacid labile protective groups constitutes a major challenge. Previous attempts

XR

OMe

F

OMe

F

O

RXO

ORO

F

SiOR1

RR

FF-PMB F-PA FPh-Piv FPh-R2Si

OMe

F

F-CBz

OO

XR

X = oxygen or nitrogen

66

to establish such conditions, using acetyl protected carbohydrate donors,failed either due to orthoesther formation or acidolytic cleavage of the sidechain protective groups, e.g. tert-butyl (tBu) protecting groups.72 Use ofglycosyl donors protected with fluorinated benzoyl protective groups maycircumvent these obstacles since they do not readily form orthoesters.Hence, it could be possible to use nearly neutral conditions for glycosylationand in the final deprotection step the fluorobenzoyl groups can be removedwithout causing β-elimination (cf. chapter 4).

67

9 References

(1) Merrifield, R. B. Solid phase peptide synthesis. I. The synthesis of atetrapeptide. J. Am. Chem. Soc. 1963, 85, 2149-2154.

(2) Merrifield, R. B. Solid-phase peptide synthesis. III. An improvedsynthesis of brandykinin. Biochemistry 1964, 3, 1385-1390.

(3) Caruthers, M. H. Gene synthesis machines: DNA chemistry and itsuses. Science 1985, 230, 281–285.

(4) Atherton, E.; Sheppard, R. C. Solid phase peptide synthesis: Apractical approach; IRL Press at Oxford University Press: Oxford,1989.

(5) Plante, O. J.; Palmacci, E. R.; Seeberger, P. H. Automated solid-phase synthesis of oligosaccharides. Science 2001, 291, 1523-1527.

(6) Seeberger, P. H. Automated carbohydrate synthesis to drivechemical glycomics. Chem. Commun. 2003, 1115–1121.

(7) Jung, G.; Beck-Sickinger, A. G. Multiple peptide synthesis methodsand their applications. Angew. Chem. Int. Ed. 1992, 31, 367–486.

(8) Eichler, J.; Hougthen, R. A. Identification of substrate-analogtrypsin inhibitors through the screening of synthetic peptidecombinatorial libraries. Biochemistry 1993, 32, 11035–11041.

(9) Beaucage, S. L.; Iyer, R. P. Advances in the synthesis ofoligonucleotides by the phosphoramidite approach. Tetrahedron1992, 48, 2223–2311.

(10) Gordon, E. M.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.;Gallop, M. A. Application of combinatorial technologies to drugdiscovery. 2. Combinatorial organic synthesis, library screeningstrategies, and future directions. J. Med. Chem. 1994, 37,1385–1401.

(11) Früchtel, J. S.; Jung, G. Organic chemistry on solid supports. Angew.Chem. Int. Ed. 1996, 35, 17–42.

(12) Ellman, J. A. Design, synthesis, and evaluation of small-moleculelibraries. Acc. Chem. Res. 1996, 29, 132–143.

(13) Thompson, L. A.; Ellman, J. A. Synthesis and application of smallmolecule libraries. Chem. Rev. 1996, 96, 555–600.

(14) Balkenhohl, F.; von dem Bussche-Hünnefeld, C.; Lansky, A.;Zechel, C. Combinatorial synthesis of small organic molecules.Angew. Chem. Int. Ed. 1996, 35, 2288–2337.

(15) Kihlberg, J.; Elofsson, M. Solid-phase synthesis of glycopeptides:Immunological studies with T cell stimulating glycopeptides. Curr.Med. Chem. 1997, 4, 79–110.

68

(16) Herzner, H.; Reipen, T.; Schultz, M.; Kunz, H. Synthesis ofglycopeptides containing carbohydrate and peptide recognitionmotifs. Chem. Rev. 2000, 100, 4495–4537.

(17) Seitz, O. Glycopeptide synthesis and the effects of glycosylation onprotein structure and activity. ChemBioChem 2000, 1, 214–246.

(18) Mogemark, M.; Kihlberg, J. The organic chemistry of sugars.Marcel Dekker Inc. In Press.

(19) Frechet, J. M.; Schuerch, C. Solid-phase synthesis ofoligosaccharides. I. Preparation of the solid support. poly(p-(1-propen-3-ol-1-yl)styren). J. Am. Chem. Soc. 1971, 93, 492-496.

(20) Frechet, J. M.; Schuerch, C. Solid-phase synthesis ofoligosaccharides. II. Steric control by C-6 substituents in glucosidesyntheses. J. Am. Chem. Soc. 1972, 94, 604-609.

(21) Osborn, H. M. I.; Khan, T. H. Recent developments in polymersupported syntheses of oligosaccharides and glycopeptides.Tetrahedron 1999, 55, 1807-1850.

(22) Hilaire, P. M. S.; Meldal, M. Glycopeptide and oligosaccharidelibraries. Angew. Chem. Int. Ed. 2000, 39, 1162-1179.

(23) Seeberger, P. H.; Haase, W.-C. Solid-phase oligosaccharidesynthesis and combinatorial carbohydrate libraries. Chem. Rev.2000, 100, 4349-4393.

(24) Liang, R.; Yan, L.; Loebach, J.; Ge, M.; Uozumi, Y.; Sekanina, K.;Horan, N.; Gildersleeve, J.; Thompson, C.; Smith, A.; Biswas, K.;Still, W. C.; Kahne, D. Parallel synthesis and screening of a solidphase carbohydrate library. Science 1996, 274, 1520-1522.

(25) Seeberger, P. H.; Danishefsky, S. J. Solid-phase Synthesis ofoligosaccharide and glycoconjugates by the glycal assembly method:A five year retrospective. Acc. Chem. Res. 1997, 31, 685–694.

(26) Bartolozzi, A.; Seeberger, P. H. New approaches to the chemicalsynthesis of bioactive oligosaccharides. Curr. Opin. Struct. Biol.2001, 11, 587–592.

(27) Nicolaou, K. C.; Pfefferkorn, J. A. Solid phase synthesis of complexnatural products and libraries thereof. Biopolymers (Pept Sci) 2001,60, 171-193.

(28) Kanemitsu, T.; Kanie, O. Recent developments in oligosaccharidesynthesis: Tactics, solid-phase synthesis and library synthesis.Comb. Chem. High Throughput Screen. 2002, 5, 339–360.

(29) Sears, P.; Wong, C.-H. Toward automated synthesis ofoligosaccharides and glycoproteins. Science 2001, 291, 2344-2350.

(30) Wong, C.-H.; Ye, X.-S.; Zhang, Z. Assembly of oligosaccharidelibraries with a designed building block and an efficient orthogonalprotection-deprotection strategy. J. Am. Chem. Soc. 1998, 120,7137-7138.

69

(31) Zhu, T.; Boons, G.-J. A new set of orthogonal-protecting groups foroligosaccharide synthesis on a polymeric support. Tetrahedron:Assymetry 2000, 11, 199-205.

(32) Orain, D.; Ellard, J.; Bradley, M. Protecting groups in solid-phaseorganic synthesis. J. Comb. Chem. 2002, 4, 1-16.

(33) James, I. W. Linkers for solid phase organic synthesis. Tetrahedron1999, 55, 4855–4946.

(34) Blaney, P.; Grigg, R.; Sridharan, V. Traceless solid-phase synthesis.Chem. Rev. 2002, 102, 2607–2624.

(35) Wills, A. J.; Balasubramanian, S. Recent developments in linkerdesign and application. Curr. Opin. Chem. Biol. 2003, 7, 246–352.

(36) Greene, T. W.; Wuts, P. G. M. Protective groups in organicsynthesis; 2nd ed.; John Wiley & Sons, Inc.: New York, 1991.

(37) Zhu, T.; Boons, G.-J. A novel and efficient synthesis of a dimericLex oligosaccharide on polymeric support. J. Am. Chem. Soc. 2000,122, 1022–1023.

(38) Toshima, K.; Tatsuta, K. Recent progress in O-glycosylationmethods and its application to natural products synthesis. Chem.Rev. 1993, 93, 1503-1531.

(39) Davis, B. G. Recent developments in oligosaccharide synthesis. J.Chem. Soc., Perkin Trans. 1 2000, 2137–2160.

(40) Nicolaou, K. C.; Mitchell, H. J. Adventures in carbohydratechemistry: New synthetic technologies, chemical synthesis,molecular design, and chemical biology. Angew. Chem. Int. Ed.2001, 40, 1576-1624.

(41) Yan, L.; Taylor, C. M.; Goodnow, R.; Kahne, D. Glycosylation onthe Merrifield resin using anomeric sulfoxides. J. Am. Chem. Soc.1994, 116, 6953–6954.

(42) Rademann, J.; Schmidt, R. R. A new method for the solid phasesynthesis of oligosaccharides. Tetrahedron Lett. 1996, 37,3989–3990.

(43) Rademann, J.; Schmidt, R. R. Repetitive solid phase glycosylationon an alkyl thiol polymer leading to sugar oligomers containing 1,2-trans- and 1,2-cis-glycosidic linkages. J. Org. Chem. 1997, 62,3650–3653.

(44) Shimizu, H.; Ito, Y.; Kanie, O.; Ogawa, T. Solid phase synthesis ofpolylactosamine oligosaccharide. Bioorg. Med. Chem. Lett. 1996, 6,2841–2846.

(45) Manabe, S.; Ito, Y.; Ogawa, T. Solvent effect in glycosylationreaction on polymer support. Synlett 1998, 628-630.

(46) Weigelt, D.; Magnusson, G. A linker for solid phase carbohydratesynthesis, permitting the introduction of variable anomeric

70

functionality in the release step. Tetrahedron Lett. 1998, 39,2839–2842.

(47) Danishefsky, S. J.; Bilodeau, M. T. Glycals in organic synthesis:The evolution of comprehensive strategies for the assembly ofoligosaccharides and glycoconjugates of biological significance.Angew. Chem. Int. Ed. 1996, 35, 13801419.

(48) Seeberger, P. H.; Beebe, X.; Sukenick, G. D.; Pochapsky, S.;Danishefsky, S. J. Monitoring the progress of solid-phaseoligosaccharide synthesis by high-resolution magic angle spinningNMR: Observation of enhanced selectivity for β-glycosideformation from α-1,2-anhydrosugar donors in solid-phaseCouplings. Angew. Chem. Int. Ed. 1997, 36, 491-493.

(49) Doi, T.; Sugiki, M.; Yamada, H.; Takahashi, T. Glycosidation ofsolid-supported glycosyl donors tethered by a trialkylsilane linker.Tetrahedron Lett. 1999, 40, 2141–2144.

(50) Douglas, S. P.; Withfield, D. M.; Krepinsky, J. J. Polymer-supportedsolution synthesis of oligosaccharides. J. Am. Chem. Soc. 1991, 113,5095–5097.

(51) Adinolfi, M.; Barone, G.; De Napoli, L.; Iadonisi, A.; Piccialli, G.Solid phase synthesis of oligosaccharides. Tetrahedron Lett. 1996,37, 5007–5014.

(52) Kononov, L. O.; Ito, Y.; Ogawa, T. Synthesis of a polymer-supported sialic acid glycosyl donor. Tetrahedron Lett. 1997, 38.

(53) Belogi, G.; Zhu, T.; Boons, G.-J. Polystyrylboronic acid as areusable polymeric support for oligosaccharide synthesis.Tetrahedron Lett. 2000, 41, 6965–6968.

(54) Fukase, K.; Nakai, Y.; Egusa, K.; Porco, J. A.; Kusumoto, S. Anovel oxidatively removable linker and its application to α-selectivesolid-phase oligosaccharide synthesis on a polymer support. Synlett1999, 1074-1078.

(55) Douglas, S. P.; Withfield, D. M.; Krepinsky, J. J. Polymer-supportedsolution synthesis of oligosaccharides using a novel versatile linkerfor the synthesis of D-mannopentaose, a structural unit of D-mannans of pathogenic yeasts. J. Am. Chem. Soc. 1995, 117,2116–2117.

(56) Nicolaou, K. C.; Winssinger, N.; Pastor, J.; DeRoose, F. A generaland highly efficient solid phase synthesis of oligosaccharides. Totalsynthesis of a heptasaccharide phytoalexin elicitor (HPE). J. Am.Chem. Soc. 1997, 119, 449-450.

(57) Knerr, L.; Schmidt, R. R. Solid-phase synthesis of a branchedhexasaccharide related to lacto-N-hexaose. Eur. J. Org. Chem. 2000,2803–2808.

71

(58) Andrade, R. B.; Plante, O. J.; Melean, L. G.; Seeberger, P. H. Solid-phase oligosaccharide synthesis: Preparation of complex structuresusing a novel linker and different glycosylating agents. Org. Lett.1999, 1, 1811–1814.

(59) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Color testfor detection of free terminal amino groups in the solid-phasesynthesis of peptides. Anal. Biochem. 1970, 34, 595-598.

(60) Flegel, M.; Sheppard, R. C. A sensitive general method forquantitative monitoring of continuos flow solid phase peptidesynthesis. J. Chem. Soc., Chem. Com. 1990, 536-538.

(61) Zambias, R. A.; Boulton, D. A.; Griffin, P. R. Microchemicalstructural determination of a peptoid covalently bound to apolymeric bead by matrix-assisted laser desorption ionization time-of-flight. Tetrahedron Lett. 1994, 35, 4283-4286.

(62) Ito, Y.; Manabe, S. Tag-reporter and resin capture - release strategyin oligosaccharide synthesis. Chem. Eur. J. 2002, 8, 3076-3084.

(63) Shapiro, M. J.; Gounarides, J. S. NMR methods in combinatorialchemistry research. Progr. NMR Spectr. 1999, 35, 153-200.

(64) Andrew, E. R.; Bradbury, A.; Eades, R. G. Nuclear magneticresonance spectra from a crystal rotated at high speed. Nature 1958,182, 1659.

(65) Anderson, R. C.; Stokes, J. P.; Shapiro, M. J. Structuredetermination in combinatorial chemistry: Utilization of MASHMQC and TOCSY NMR spectra in the structure determination ofWang-bound lysine. Tetrahedron Lett. 1995, 36, 5311-5314.

(66) Anderson, R. C.; Jarema, M. A.; Shapiro, M. J.; Stokes, J. P.; Ziliox,M. Analytical techniques in combinatorial chemistry: MAS CHcorrelation in solvent-swollen resin. J. Org. Chem. 1995, 60, 2650-2651.

(67) Sarkar, S. K.; Garigipati, R. S.; Adams, J. L.; Keifer, P. A. An NMRmethod to identify nondestructively chemical compounds bound to asingle solid-phase-synthesis bead for combinatorial chemistryapplications. J. Am. Chem. Soc. 1996, 118, 2305-2306.

(68) Wehler, T.; Westman, J. Magic angle spinning NMR: A valuabletool for monitoring the progress of reactions in solid phase synthesis.Tetrahedron Lett. 1996, 37, 4771-4774.

(69) Keifer, P. A. Influence of resin structure, tether length, and solventupon the high-resolution 1H NMR spectra of solid phase-synthesisresins. J. Org. Chem. 1996, 61, 1558–1559.

(70) Purch, M.; Schlottenbeck, G.; Tseng, L.-H.; Albert, K.; Rapp, W.Monitoring the reaction progress in combinatorial chemistry: HMAS NMR investigations on single macro beads in the suspendedstate. Angew. Chem. Int. Ed. 1996, 35, 2867-2869.

72

(71) Gottfredsen, C. H.; Grøtli, M.; Willert, M.; Meldal, M.; Duus, J. Ø.Single-bead structure elucidation. Requirements for analysis ofcombinatorial solid-phase libraries by nanoprobe MAS-NMRspectroscopy. J. Chem. Soc., Perkin Trans. 1 2000, 1167-1171.

(72) Halkes, K. M.; Gotfredsen, C. H.; Grøtli, M.; Miranda, L. P.; Duus,J. Ø.; Meldal, M. Solid-Phase Glycosylation of Peptide Templatesand On-Bead MAS-NMR Analysis: Perspectives for GlycopeptideLibraries. Chem. Eur. J. 2001, 7, 3584-3591.

(73) Groth, T.; Renil, M.; Meinjohanns, E. PEG based resins for proteasedrug discovery synthesis, screening and analysis of combinatorialon-bead libraries. Comb. Chem. High Troughput Screen. 2003, 6,589–610.

(74) Look, G. C.; Holmes, C. P.; Chinn, J. P.; Gallop, M. A. Methods forcombinatorial organic synthesis: The use of fast 13C NMR analysisfor gel phase reaction monitoring. J. Org. Chem. 1994, 59, 7588-7590.

(75) Kanemitsu, T.; Kanie, O.; Wong, C.-H. Quantitative monitoring ofsolid-phase synthesis using gated decoupling 13C-enriched protectinggroup and an internal standard in the synthesis of sialyl Lewisx

tetrasaccharide. Angew. Chem. Int. Ed. 1998, 37, 3415-3418.(76) Kanemitsu, T.; Wong, C.-H.; Kanie, O. Solid-phase synthesis of

oligosaccharides and on-resin quantitative monitoring using gateddecoupling 13C NMR. J. Am. Chem. Soc. 2002, 124, 3591-3599.

(77) Hudson, D. Matrix assisted synthetic transformations: A mosaic ofdiverse contributions. I the pattern emerges. J. Comb. Chem. 1999,1, 333–360.

(78) Hudson, D. Matrix assisted synthetic transformations: A mosaic ofdiverse contributions. II The pattern is completed. J. Comb. Chem.1999, 1, 403–457.

(79) Vaino, A. R.; Janda, K. D. Solid-phase organic synthesis: A criticalunderstanding of the resin. J. Comb. Chem. 2000, 2, 579–596.

(80) Gerritz, S. W. Quantitative techniques for the comparison of solidsupports. Curr. Opin. Chem. Biol. 2001, 5, 264–268.

(81) Walsh, D.; Wu, D.; Chang, Y.-T. Understanding the effects of thepolymer support on reaction rates and kinetics: Knowledge towardefficient synthetic design. Curr. Opin. Chem. Biol. 2003, 7,353–361.

(82) Rademann, J.; Barth, M.; Brock, R.; Egelhaaf, H.-J.; Jung, G.Spatially resolved single bead analysis: Homogeneity, diffusion, andadsorption in cross-linked polystyrene. Chem. Eur. J. 2001, 7,3884–3889.

73

(83) Bayer, E.; Dengler, M.; Hemmasi, B. Peptide-synthesis on the newpolyoxyethylene-polystyrene graft copolymer, synthesis of insulin-B12-30. Int. Pept. Protein Res. 1985, 25, 178–186.

(84) Bayer, E. Towards the chemical synthesis of proteins. Angew. Chem.Int. Ed. Engl. 1991, 30, 113–129.

(85) Gooding, O. W.; Baudart, S.; Deegan, T. L.; Heisler, K.; Labadie, J.W.; Newcomb, W. S.; Porco, J. A.; van Eikeren, P. On thedevelopment of new poly(styrene–oxyethylene) graft copolymerresin supports for solid-phase organic synthesis. J. Comb. Chem.1999, 1, 113–122.

(86) Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach, A.G.; Longbottom, D. A; Nesi, M.; Scott, J. S.; Storer, I.; Taylor, S. J.Multi-step organic synthesis using solid-supported reagents andscavengers: A new paradigm in chemical library generation. J.Chem. Soc., Perkin Trans. 1 2000, 3815–4195.

(87) Wu, X.; Schmidt, R. R. Solid-phase synthesis of complexoligosaccharides using a novel capping reagent. J. Org. Chem. 2004,69, 1853–1857.

(88) Meldal, M. Pega: A flow stable polyethylene glycol dimethylacrylamide copolymer for solid phase synthesis. Tetrahedron Lett.1992, 33, 3077–3080.

(89) Renil, M.; Meldal, M. Synthesis of a PEGA polymeric support forhigh capacity continuous flow solid-phase peptide synthesis.Tetrahedron Lett. 1995, 36, 4647–4650.

(90) Renil, M.; Meldal, M. POEPOP and POEPS: Inert polyethylenglycol crosslinked polymeric supports for solid synthesis.Tetrahedron Lett. 1996, 37, 6185–6188.

(91) Grøtli, M.; Gotfredsen, C. H.; Rademann, J.; Buchardt, J.; Clark, A.J.; Duus, J. Ø.; Meldal, M. Physical properties of poly(ethyleneglycol) (PEG)-based resins for combinatorial solid phase organicchemistry: A comparison of PEG-cross-linked and PEG-graftedresins. J. Comb. Chem. 2000, 2, 108–119.

(92) Manatt, S. L.; Amsden, C. F.; Bettison, C. A.; Frazer, W. T.;Gudman, J. T.; Lenk, B. E.; Lubetich, J. F.; McNelly, E. A.; Smith,S. C.; Templeton, D. J.; Pinnel, R. P. A fluorine-19 NMR approachfor studying Merrifield solid-phase peptide synthesis. TetrahedronLett. 1980, 21, 1397–1400.

(93) Shapiro, M. J.; Kumaravel, G.; Petter, R. C.; Beveridge, R. 19FNMR monitoring of a SNAr reaction on solid support. TetrahedronLett. 1996, 37, 4671–4674.

(94) Burgess, K.; Lim, D.; Bois-Choussy, M.; Zhu, J. Rapid and efficientsolid phase syntheses of cyclic peptides with endocyclic biaryl etherbonds. Tetrahedron Lett. 1997, 38, 3345–3348.

74

(95) Svensson, A.; Fex, T.; Kihlberg, J. Use of 19F NMR spectroscopy toevaluate reactions in solid phase organic synthesis. TetrahedronLett. 1996, 37, 7649–7652.

(96) Svensson, A.; Bergquist, K.-E.; Fex, T.; Kihlberg, J. Fluorinatedlinkers for monitoring solid-phase synthesis using gel-phase 19FNMR spectroscopy. Tetrahedron Lett. 1998, 39, 7193–7196.

(97) Svensson, A.; Fex, T.; Kihlberg, J. Preparation of fluorinatedlinkers: Use of 19F NMR spectroscopy to establish conditions forsolid-phase synthesis of pilicide libraries. J. Comb. Chem. 2000, 2,736–748.

(98) Salvino, J. M.; Kumar, N. V.; Orton, E.; Airey, J.; Kiesow, T.;Crawford, K.; Mathew, R.; Krolikowski, P.; Drew, M.; Engers, D.;Krolikowski, D.; Herpin, T.; Gardyan, M.; McGeehan, G.;Labaudiniere, R. Polymer-supported tetrafluorophenol: A newactivated resin for chemical library synthesis. J. Comb. Chem. 2000,2, 691–697.

(99) Gustafsson, M.; Olsson, R.; Andersson, C.-M. Generalcombinatorial synthesis of tertiary amines on solid support. A novelconditional release strategy based on traceless linking at nitrogen.Tetrahedron Lett. 2001, 42, 133–136.

(100) Stones, D.; Miller, D. J.; Beaton, M. W.; Rutherford, T. J.; Gani, D.A method for the quantification of resin loading using 19F NMRspectroscopy and a new method for benzyl ether linker cleavage insolid-phase chemistry. Tetrahedron Lett. 1998, 39, 4875–4878.

(101) Lakshmipathi, P.; Crévisy, C.; Grée, R. Reaction monitoring inLPOS by 19F NMR. Study of soluble polymer supports with fluorinein spacer or linker components of supports. J. Comb. Chem. 2002, 4,612–621.

(102) Hu, Y.; Porco, J. A. Ester enolates claisen rearrangement using apolymer-supported silyl triflate. Tetrahedron Lett. 1999, 40,3289–3292.

(103) Pan, Y.; Ruhland, B.; Holmes, C. P. Use of a perfluoroalkylsulfonyl(PFS) linker in a "traceless" synthesis of biaryls through Suzukicleavage. Angew. Chem. Int. Ed. 2001, 40, 4488–4491.

(104) Pan, Y.; Holmes, C. P. A traceless perfluoroalkylsulfonyl (PFS)linker for the deoxygenation of phenols. Org. Lett. 2001, 3,2769–2771.

(105) Le Roy, I.; Mouysset, D.; Mignani, S.; Vuilhorgne, M.; Stella, L.Solid phase β-lactams synthesis using the Staudinger reaction,monitored by 19F NMR spectroscopy. Tetrahedron 2003, 59,3719–3727.

75

(106) Drew, M.; Orton, E.; Krolikowski, P.; Salvino, J. M.; Kumar, N. V.A method for quantification of solid-phase synthesis using 19F NMRspectroscopy. J. Comb. Chem. 2000, 2, 8–9.

(107) Salvino, J. M.; Patel, S.; Drew, M.; Krowlikowski, P.; Orton, E.;Kumar, N. V.; Caulfield, T.; Labaudiniere, R. Synthesis of a newfluoro-Wang resin for solid-phase reaction monitoring by 19F NMRspectroscopy. J. Comb. Chem. 2001, 3, 177–180.

(108) Hamper, B.; Kolodziej, S. A.; Scates, A. M. Knoevenagelcondensation of unsymmetrical malonamic esters and malonates ona solid support. Tetrahedron Lett. 1998, 39, 2047-2050.

(109) Mogemark, M.; Elofsson, M.; Kihlberg, J. Monitoring solid-phaseglycoside synthesis with 19F NMR Spectroscopy. Org. Lett. 2001, 3,1463–1466.

(110) Mogemark, M.; Elofsson, M.; Kihlberg, J. Fluorinated protectivegroups for on-resin quantification of solid-phase oligosaccharidesynthesis with 19F NMR spectroscopy. ChemBioChem 2002, 3,1266–1269.

(111) Mogemark, M.; Elofsson, M.; Kihlberg, J. Solid-phase synthesis ofα-Gal epitopes: On-resin analysis of solid-phase oligosaccharidesynthesis with 19 F NMR spectroscopy. J. Org. Chem. 2003, 68,7281–7288.

(112) Mogemark, M.; Gårdmo, F.; Tengel, T.; Kihlberg, J.; Elofsson, M.Gel-phase 19F NMR spectral quality for resins commonly used insolid-phase organic synthesis; a study of peptide solid-phaseglycosylations. Org. Biomol. Chem. 2004, 2, 1770–1776.

(113) Mogemark, M.; Gustafsson, L.; Bengtsson, C.; Elofsson, M.;Kihlberg, J. A fluorinated selenide linker for solid-phase synthesis ofn-pentenyl glycosides. Org. Lett. 2004, In press.

(114) Wallner, F.; Norberg, H. A.; Johansson, A. I.; Mogemark, M.;Elofsson, M. Solid-phase synthesis of serine-basedglycosphingolipid analogues for preparation of glycoconjugatearrays. Org. Biomol. Chem. 2004, In press.

(115) Emtenäs, H.; Åhlin, K.; Pinkner, J. S.; Hultgren, S. J.; Almqvist, F.Design and parallel solid-phase synthesis of ring-fused 2-pyridinones that target pilus biogenesis in pathogenic bacteria. J.Comb. Chem. 2002, 4, 630–639.

(116) Sjölin, P.; George, S. K.; Bergquist, K.-E.; Roy, S.; Svensson, A.Kihlberg, J. Synthesis of deoxy and alanine-substituted derivativesof a T cell stimulating glycopeptide — An investigation ofconditions for cleavage from the solid phase and deprotection. J.Chem. Soc., Perkin Trans. 1 1999, 1731–1742.

76

(117) Hochlowski, J. E.; Whittern, D. N.; Sowin, T. J. Encoding ofcombinatorial chemistry libraries by fluorine-19 NMR. J. Comb.Chem. 1999, 1, 291–293.

(118) Bianco, A.; Furrer, J.; Limal, D.; Gulchard, G.; Elbayed, K.; Raya,J.; Piotto, M.; Briand, J.-P. Multistep synthesis of 2,5-diketopiperazines on different solid supports monitored by highresolution magic angle spinning NMR spectroscopy. J. Comb.Chem. 2000, 2, 681–690.

(119) Garegg, P. J.; Norberg, T. Observation on silver trifluoromethanesulfonate-promoted syntheses of 1,2-trans-glycosides from acylatedglycosyl bromides. Acta Chem. Scand. 1979, B 33, 116-118.

(120) Garegg, P. J.; Konradsson, P.; Kvarnström, I.; Norberg, T.;Svensson, S. C. T.; Wigilius, B. Studies on Koenigs-Knorrglycosidations. Acta Chem. Scand. 1985, B 39, 569-577.

(121) Reimer, K. B.; Meldal, M.; Kusumoto, S.; Fukase, K.; Bock, K.Small-scale solid-phase O-glycopeptide synthesis of linear andcyclized hexapeptides from blood-clotting factor IX containing O-(α-D-Xyl-1→3-α-D-Xyl-1→3-β-D-Glc)-L-Ser. J. Chem. Soc. PerkinTrans. 1 1993, 925-932.

(122) Elofsson, M.; Broddefalk, J.; Ekberg, T.; Kihlberg, J. Synthesis of awater-soluble serine-based neoglycolipid which can be covalentlylinked to solid phases. Carbohydr. Res. 1994, 258, 123-133.

(123) Sjölin, P.; Elofsson, M.; Kihlberg, J. Removal of acyl protectivegroups from glycopeptides: Base does not epimerize peptidestereocenters and β-elimination is slow. J. Org. Chem. 1996, 61,560-565.

(124) Sjölin, P.; Kihlberg, J. Use of fluorobenzoyl protective groups insynthesis of glycopeptides: β-Elimination of O-linked carbohydratesis suppressed. J. Org. Chem. 2001, 66, 2957–2965.

(125) Mitsunobu, O. The use of diethyl azodicarboxylate andtriphenylphosphine in synthesis and transformation of naturalproducts. Synthesis 1981, 1–28.

(126) Blankemeyer-Menge, B.; Nimtz, M.; Frank, R. An efficient methodfor anchoring Fmoc-amino acids to hydroxyl-functionalized solidsupports. Tetrahedron Lett. 1990, 31, 1701–1704.

(127) Polt, R.; Szabó, L.; Treiberg, J.; Li, Y.; Hruby, V. J. Generalmethods for α- and β-O-Ser/Thr glycosides and glycopeptides.Solid-phase synthesis of O-glycosyl cyclic enkephalin analouges. J.Am. Chem. Soc. 1992, 114, 10249-10258.

(128) Garegg, P. J. Thioglycosides as glycosyl donors in oligosaccharidesynthesis. Adv. Carbohydr. Chem. 1997, 52, 178-205.

(129) Fügedi, P.; Garegg, P. J. A novel promoter for the efficientconstruction of 1,2-trans linkages in glycoside synthesis, using

77

thioglycosides as glycosyl donors. Carbohydr. Res. 1986, 149, C9-C12.

(130) Veeneman, G. H.; van Leeuwen, S. H.; van Boom, J. H. Iodoniumion promoted reactions at the anomeric centre. II. An efficientthioglycoside mediated approach towards the formation of 1,2-translinked glycosides and glycosidic esters. Tetrahedron Lett. 1990, 31,1331-1334.

(131) Konradsson, P.; Udodong, U. E.; Fraser-Reid, B. Iodoniumpromoted reactions of disarmed thioglycosides. Tetrahedron Lett.1990, 31, 4313-4316.

(132) Mitchels, R.; Kato, M.; Heitz, W. Polymere selenreagenzien.Makromol. Chem. 1976, 177, 2311–2320.

(133) Ruhland, T.; Andersen, K.; Pedersen, H. Selenium-linking strategyfor traceless solid-phase synthesis: direct loading, aliphatic C-Hbond formation upon cleavage and reaction monitoring by gradientMAS NMR spectroscopy. J. Org. Chem. 1988, 63, 9204–9211.

(134) Nicolaou, K. C.; Pastor, J.; Barluenga, S.; Winssinger, N. Polymer-supported selenium reagents for organic synthesis. Chem. Commun.1998, 1947–1948.

(135) Nicolaou, K. C.; Roecker, A. J.; Hughes, R.; van Summeren, R.;Pfefferkorn, J. A.; Winssinger, N. Novel strategies for the solidphase synthesis of indolines and indoles. Bioorg. Med. Chem. 2003,11, 465–476 and references therein.

(136) Wirth, T.; Drabowicz, J. Organoselenium chemistry : moderndevelopments in organic synthesis; Springer, Berlin: New York,2000.

(137) Kahne, D.; Walker, S.; Cheng, Y.; Van Engen, D. Glycosylation ofunreactive substrates. J. Am. Chem. Soc. 1989, 111, 6881-6882.

(138) Bengtsson, C. Synthesis of n-pentenyl glycosides on a seleniumbased carbohydrate linker. Master thesis. Department of Chemistry,Organic Chemistry; Umeå University: Umeå, 2003; pp 56.

(139) Sharpless, K. B.; Young, M. W.; Lauer, R. F. Reactions ofselenoxides: Thermal syn-elimination and H2

19O exchange.Tetrahedron Lett. 1973, 14, 1979–1982.

(140) Fraser-Reid, B.; Udodong, U. E.; Wu, Z.; Ottoson, H.; Merritt, J. R.;Rao, C. S; Roberts, C.; Madsen, R. n-Pentenyl glycosides in organicchemistry: A contemporary example of serendipity. Synlett 1992,927–942.

(141) Flitsch, S. L.; Ulijn, R. V. Sugars tied to the spot. Nature 2003, 421,219-220.

(142) Ratner, D. M.; Adams, E. W.; Disney, M. D.; Seeberger, P. H. Toolsfor glycomics: Mapping interactions of carbohydrates in biologicalsystems. ChemBioChem 2004, 5, 1375–1383.

78

(143) Chen, X.; Andreana, P. R.; Wang, P. G. Carbohydrates intransplantation. Curr. Opin. Chem. Biol. 1999, 3, 650-658.

(144) Jacquinet, J. C.; Duchet, D.; Milat, M. L.; Sinay, P. Synthesis ofblood-group substances. Part 11. Synthesis of the trisaccharide O-α-D-galactopyranosyl-(1→3)-O-β-D-galactopyranosyl-(1→4)-2-acetamido-2-deoxy-D-glucopyranose. J. Chem. Soc., Perkin Trans.1 1981, 326-330.

(145) Garegg, P. J.; Oscarson, S. A synthesis of 8-methoxycarbonyloct-1-yl O-α-D-galactopyranosyl-(1→3)-O-β-D-galactopyranosyl-(1→4)-2-acetamido-2-deoxy-β-D-glucopyranoside. Carbohydr. Res. 1985,136, 207-213.

(146) Schaubach, R.; Hemberger, J.; Kinzy, W. Tumor-associated antigensynthesis. Synthesis of the Gal-α-(1→3)-Gal-β-(1→4)-GlcNAcepitope a specific determinant for metastatic progression? LiebigsAnn. Chem. 1991, 607-614.

(147) Reddy, G. V.; Jain, R. K.; Bhatti, B. S.; Matta, K. L. Synthesis of α-D-galactopyranosyl-linked oligosaccharides containing the α-Gal→β-Gal→GlcNAc sequence employing methyl-2,3,4,6-tetra-O-(4-methoxybenzyl)-1-thio-β-D-galagalactopyranoside as an efficientglycosyl donor. Carbohydr. Res. 1994, 263, 67-77.

(148) Gege, C.; Kinzy, W.; Schmidt, R. R. Total synthesis of the naturalantigen involved in the hyperacute rejection responce toxenotransplants. Carbohydr. Res. 2000, 328, 459-466.

(149) Hanessian, S.; Saavedra, O. M.; Mascitti, V.; Marterer, W.;Oehrlein, R.; Mak, C.-P. Practical syntheses of B disaccharide andlinear B type 2 trisaccharide-non-primate epitope markersrecognized by human anti-a-Gal antibodies causing hyperacuterejection of xenotransplants. Tetrahedron 2001, 57, 3267-3280.

(150) Nilsson, K. G. I. Glycosidase-catalyzed synthesis of di- andtrisaccharide derivatives related to antigens involved in thehyperacute rejection of xenotransplants. Tetrahedron Lett. 1997, 38,133-136.

(151) Vic, G.; Tran, C. T.; Scigelova, M.; Crout, D. H. G. Glycosidase-catalyzed synthesis of oligosaccharides: a one step synthesis oflactosamine and of the linear B type 2 trisaccharide α-D-Gal-(1→3)-β-D-Gal-(1→4)-β-D-GlcNAcSEt involved in the hyperacuterejection response in xenotransplantation from pigs to man and asthe specific receptor for toxin A from Clostridium difficile. Chem.Commun. 1997, 169-170.

(152) Fang, J.; Li, J.; Chen, X.; Zhang, Y.; Wang, J. et al. Highly EfficientChemoenzymic Synthesis of α-Galactosyl Epitopes with a

79

Recombinant α(1→3)-Galactosyltransferase. J. Am. Chem. Soc.1998, 120, 6635-6638.

(153) Sujino, K.; Malet, C.; Hindsgaul, O.; Palcic, M. M. Acceptorhydroxyl group mapping for calf thymus α-(1→ 3)-galactosyltransferase and enzymatic synthesis of α-D-Galp-(1→ 3)-β-D-Galp-(1 → 3)-β-D-GlcNAcp analogs. Carbohydr. Res. 1998,305, 483-489.

(154) Ramos, D.; Rollin, P.; Klaffke, W. Chemoenzymatic synthesis ofneoglycopeptides: Application to an alpha-Gal-terminatedneoglycopeptide. J. Org. Chem. 2001, 66, 2948-2956.

(155) Yan, F.; Gilbert, M.; Wakarchuk, W. W.; Brisson, J.-R.; Whithfield,D. M. Chemoenzymatic iterative synthesis of difficult linkages ofoligosaccharide on soluble polymeric supports. Org. Lett. 2001, 3,3265-3268.

(156) Carpino, L. A. The 9-Fluorenylmethyloxycarbonyl family of base-sensitve amino-protecting groups. Acc. Chem. Res. 1987, 20, 401-407.

(157) Freese, S. J.; Vann, W. F. Synthesis of 2-acetamido-3-O-acetyl-2-deoxy-D-mannose phosphoramidites. Carbohydr. Res. 1996, 281,313-319.

(158) Roussel, F.; Knerr, L.; Grathwohl, M.; Schmidt, R. R. O-Glycosylthrichloracetimidates bearing fmoc as temporary hydroxy protectinggroup: A new access to solid-phase oligosaccharide synthesis. Org.Lett. 2000, 2, 3043-3046.

(159) Grathwohl, M.; Schmidt, R. R. Solid phase syntheses ofoligomannosides and of a lactosamine containing milk trisaccharideusing a benzoate linker. Synthesis 2001, 2263-2272.

(160) Wu, X.; Grathwohl, M.; Schmidt, R. R. Efficient Solid-phasesynthesis of a complex, branched N-glycan hexasaccharide: use of anovel linker and temporary-protecting-group pattern. Angew. Chem.Int. Ed. 2002, 41, 4489-4493.

(161) Lipták, A.; Jodál, I.; Nánási, P. Stereoselective Ring-cleavage of 3-O-benzyl- and 2,3-di-O-benzyl-4,6-O-benzylidenehexopyranosidederivatives with the LiAlH4-AlCl3 reagent. Carbohydr. Res. 1975,44, 1-11.

(162) Garegg, P. J.; Hultberg, H. A novel, reductive ring-opening ofcarbohydrate benzylidene acetals, with unusual regioselectivity.Carbohydr. Res. 1981, 93, c10-c11.

(163) Lu, J.; Chan, T.-H. Borane/Bu2BOTf: A mild reagent for theregioselective reductive ring opening of benzylidene acetals incarbohydrates. Tetrahedron Lett. 1998, 39, 355-358.

80

(164) Debenham, S. D.; Toone, E. J. Regioselective reduction of 4,6-O-benzylidenes using triethylsilane and BF3*Et2O.Tetrahedron:Asymmetry 2000, 11, 385–387.

(165) Crich, D.; Dudkin, V. Why are the hydroxy group of partiallyprotected N-acetylglucoseamine derivatives such poor glycosylacceptors... J. Am. Chem. Soc. 2001, 123, 6819-6825.

(166) Drinnan, N. B.; Bornaghi, L.; West, M. L. Solid phase synthesis ofalpha-Gal containing oligosaccharides. Abstract of the XXIstInternational Carbohydrate Symposium, Cairns, Australia 2002.

(167) Fields, G. B.; Fields, C. G. Solvation effect in solid-phase peptidesynthesis. J. Am. Chem. Soc. 1991, 113, 4202–4207.

(168) Hollósi, M.; Kollát, E.; Laczkó, I.; Medzihradszky, K. F.; Thurin, J.et al. Solid-phase synthesis of glycopeptides: Glycosylation of resin-bound serine-peptides by 3,4,6,-tri-O-acetyl-D-glucose-oxazoline.Tetrahedron Lett. 1991, 32, 1531–1534.

(169) Andrews, D. M.; Seale, P. W. Solid-phase synthesis of O-mannosylated peptides: Two strategies compared. Int. J. PeptideProtein Res. 1993, 42, 165–170.

(170) Seitz, O.; Kunz, H. HYCRON, an allylic anchor for high-efficiencysolid phase synthesis of protected peptides and glycopeptides. J.Org. Chem. 1997, 62, 813–826.

(171) Paulsen, H.; Schleyer, A.; Mathieux, N.; Meldal, M.; Bock, K. Newsolid-phase oligosaccharide synthesis on glycopeptides bound to asolid phase. J. Chem. Soc., Perkin Trans. 1 1997, 281-293.

(172) Schleyer, A.; Meldal, M.; Manat, R.; Paulsen, H.; Bock, K. Directsolid-phase glycosylations of peptide templates on a novel PEG-based resin. Angew. Chem. Int. 1997, 36, 1976–1978.

(173) Sjölin, P. Synthesis of modified glycopeptides. Studies of β-elimination and epimerization; side reactions in glycopeptidesynthesis. PhD thesis. Department of Chemistry, OrganicChemistry; Umeå University: Umeå, 2000; pp 77.

*(174) Yan, B. Monitorin the progress and the yield of solid-phase organicreactions directly on resin supports. Acc. Chem. Res. 1998, 31,621–630.

*(175) Manabe, S.; Ito, Y. On-resin real-time reaction monitoring of solid-phase oligosaccharide synthesis. J. Am. Chem. Soc. 2002, 124,12638–12639.

* References appearing at page 8.

81

10 Acknowledgements

Förutom det stora nöjet att få jobba med organisk kemi så hardoktorandtiden på avdelningen för organisk kemi i Umeå varit enhärlig tid som är få förunnat. Detta beror mestadels på dem jag harhaft nöjet att arbeta med. Framförallt ett stort tack till…

Min huvudhandledare Jan Kihlberg som i ett tidigt skede igrundutbildningen välkomnade mig till gruppen och som på ettutmärkt sätt väglett och stöttat mig genom hela doktorandtiden.

Min biträdande handledare, vän och granne Mikael Elofsson somalltid funnits tillhands för att bolla idéer. De senaste åren har varitväldigt omtumlande – stort tack till dig och Cecilia för allt stöd.

Golfaren Petter Sjölin för handledning av mitt exjobb och för att dulivat upp stämmningen på labbet.

Fredrik Wallner och Henrik Nordberg för fluorerandet.

Mattias Hedenström och Tobias Tengel som alltid ställer upp då detär trubbel med nmr och framförallt då det är pokerkväll.

Hans Emtenäs för allt roligt på och utanför jobbet, din sköna stil, ochför att du alltid hjälper till med gott humör.

Hasse, Johan Broddefalk, Björn Holm för korrekturläsning

Frida Gårdmo, Christoffer Bengtsson, Linda Gustafsson och SaraHolmqvist för utmärkt utförda exarbeten/projektarbeten och allt annatskoj.

Carina för alla roliga skämt och för hjälp med det krångligaadministrativa.

Bert för all hjälp med datorn, trots att det är en mac…

82

Alla lab- och rumskamrater Sussi, Maciej, Anna, Blompa, Petter,Brodde, Mark, Zhongqing, Mattias, Tobias, Eric W, Lotta, Stina,Björn, Shaji, Baquer mfl. för allt stoj och skoj på lab.

Storfiskarna David Nilsson, Tomas Gustafsson, David Blomberg,Mark Rucker och Fredrik Almqvist för roliga fisketurer ochgripande fiskeberättelser.

Marcus Carlsson för allt mellan flått och Turbo Negro.

Innebandygänget för att ni stått ut med min spelstil.

De gamla rävarna Ulf Edlund, Dan Johnels, Ingemar Sethson,Roland Gustafsson, Åsa Nilsson-Lindgren, Bertil Eliasson ochGull-Britt Trogen för att de visar vart “skåpet ska stå”

Alla övriga på Organisk Kemi för en härlig stämmning.

Alla nya vänner jag och Lena fått under vår tid i Umeå.

Stora tjocka släkten för uppmuntrande nyfikenhet.

Min älskade fru Lena och underbara son Emil för att ni gör mig till enbättre människa.

Äntligen är nattskiften över…