Efficient Synthesis by Intra- and Supramolecular Control

Hai Dong

Doctoral Thesis Stockholm 2008

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av teknologie doktorsexamen i kemi, med inriktning mot organisk kemi, torsdagen den 5 Feb 2009, kl 10.00 i sal F3, KTH, Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. Opponent är Ulf Nilsson, Lunds Tekniska Högskola/Lunds Universitet.

ISBN 978-91-7415-207-4 ISSN 1654-1081 TRITA-CHE-Report 2009:2

© Hai Dong, 2008 Universitetsservice US AB, Stockholm

献给晓溪, 东东和爱玲 .

Till Emilia, Dongdong och Ailing.

The road ahead is hard and long, but nothing will stop me as I go searching up and down. ………Qu Yuan (B.C. 340 - 278) Translated by Hai

Hai Dong, 2008: “Efficient Carbohydrate Synthesis by Intra- and Supramolecular Control” Organic Chemistry, KTH Chemistry, Royal Institute of Technology, S- 10044 Stockholm, Sweden.

Abstract

The Lattrell-Dax method of nitrite-mediated substitution of carbohydrate triflates is an efficient method to generate structures of inverse configuration. In this study, the effects of the neighboring group on the Lattrell-Dax inversion were explored. A new carbohydrate/anion host-guest system was discovered and the ambident reactivity of the nitrite anion was found to cause a complicated behavior of the reaction. It has been demonstrated that a neighboring equatorial ester group plays a highly important role in this carbohydrate epimerization reaction, restricting the nitrite N-attack, thus resulting in O-attack only and inducing the formation of inversion compounds in good yields. Based on this effect, efficient synthetic routes to a range of carbohydrate structures, notably β-D-mannosides and β-D-talosides, were designed by use of double parallel and double serial inversion. A supramolecularly activated, triggered cascade reaction was also developed. This cascade reaction is triggered by a deprotonation process that is activated by anions. It was found that the anions can activate this reaction following their hydrogen bonding tendencies to the hydroxyl group in aprotic solvents.

Keywords: , Carbohydrate Protection, Epimerization, Inversion, Neighboring Group Participation, Supramolecular Control, Anion Activation, Ambident Reactivity, Cascade Reaction, Hydrogen Bonding, Basicity.

Abbreviations

A Anion Ac Acetyl group AcCl Acetyl chloride Ac2O Acetic anhydride aq aqueous Bn Benzyl group BnBr Benzyl bromide Bz Benzoyl group BzCl Benzoyl chloride Bu Butyl Bu2SnO Dibutyltin oxide Conv Conversion DCM Dichloromethane DMF Dimethylformamide DMSO Dimethylsulfoxide eq./equiv. equivalent Et Ethyl EDA Ethylenediamine Gal Galactoside Glc Glucoside h hour HSAB Hard-Soft Acid-Base theory IM Imidazole Man Mannoside Manp Mannopyranoside Me Methyl NGP Neighboring group participation NMR Nuclear magnetic resonance rt/r.t. room temperature S Solvent T Temperature Tal Taloside TBA Tetrabutylammonium TEA Triethylamine Tf2O Trifluoromethylsulfonic anhydride THF Tetrahydrofuran P/PG Protecting group py. Pyridine PMO Perturbation Molecular Orbital theory

List of publications

This thesis is based on the following papers, referred to in the text by their Roman numerals.

I. Reagent-Dependent Regioselective Control in Multiple Carbohydrate Esterifications Hai Dong, Zhichao Pei, Styrbjörn Byström and Olof Ramström J. Org. Chem. 2007, 72, 1499-1502.

II. Stereospecific Ester Activation in Nitrite-Mediated Carbohydrate Epimerization Hai Dong, Zhichao Pei and Olof Ramström J. Org. Chem. 2006, 71, 3306-3309.

III. Supramolecular Control in Carbohydrate Epimerization: Discovery of a New Anion Host-Guest System Hai Dong, Martin Rahm, Tore Brinck, and Olof Ramström J. Am. Chem. Soc. 2008, 130, 15270-15271.

IV. Control of the Ambident Reactivity of the Nitrite Ion in Carbohydrate Epimerization Hai Dong, Lingquan Deng, and Olof Ramström Preliminary manuscript.

V. Efficient Synthesis of β-D-Mannosides and β-D- Talosides by Double Parallel or Double Serial Inversion Hai Dong, Zhichao Pei, Marcus Angelin, Styrbjörn Byström and Olof Ramström J. Org. Chem. 2007, 72, 3694-3701.

VI. Synthesis of Positional Thiol Analogs of β-D- Galactopyranose Zhichao Pei, Hai Dong, Rémi Caraballo and Olof Ramström Eur. J. Org. Chem. 2007, 29, 4927-4934. VII. Supramolecular Activation in Triggered Cascade Inversion Hai Dong, Zhichao Pei and Olof Ramström Chem. Commun. 2008, 11, 1359-1361.

VIII. Enhanced Basicity by Supramolecular Anion Activation Hai Dong and Olof Ramström Preliminary manuscript.

Papers not included in the thesis.

IX. Solvent Dependent, Kinetically Controlled Stereoselective Synthesis of Thioglycosides Zhichao Pei, Hai Dong and Olof Ramström J. Org. Chem. 2005, 70, 6952-6955.

X. Direct, Mild, and Selective Synthesis of Unprotected Dialdo- Marcus Angelin, Magnus Hermansson, Hai Dong and Olof Ramström Eur. J. Org. Chem. 2006, 19, 4323-4326.

Table of Contents

ABSTRACT ABBREVIATIONS LIST OF PUBLICATIONS

1 Introduction ...... 1 1.1 – A General Introduction ...... 1 1.2 Carbohydrates – Challenging Synthetic Targets...... 2 1.3 Regioselective Protection/Deprotection...... 4 1.4 Epimerization ...... 4 1.5 Neighboring Group Participation...... 6 1.6 Aim of Study ...... 7 2 Regioselective Carbohydrate Protection...... 9 2.1 Common Protection Strategies...... 9 2.1.1 Acylation ...... 9 2.1.2 Alkylation...... 9 2.1.3 Organotin Protection ...... 10 2.1.4 Integrated Protection Strategies ...... 10 2.2 Organotin Mutiple Esterification ...... 11 3 Lattrell-Dax Epimerization ...... 15 3.1 Effects of Protecting Groups...... 15 3.1.1 Effects of Protection Patterns...... 15 3.1.2 Effects of Neighboring Group Configurations...... 17 3.2 Neighboring and Remote Group Participation...... 18 3.2.1 Neighboring Group Participation Effects...... 18 3.2.2 Remote Group Participation...... 19 3.3 Supramolecular Control ...... 21 3.3.1 Unusual Solvent Effect...... 21 3.3.2 Carbohydrate-Anion Complex ...... 22 3.3.3 Binding Model...... 24 3.4 Ambident Reactivity of Nitrite Anions ...... 25 3.4.1 An O/N Selectivity Case in Carbohydrate Epimerization...... 26 3.4.2 Carbohydrate Epimerization with Non-Ambident Reagents .....27 3.4.3 Nitration Products ...... 27 3.4.4 Solvent Effect...... 29 3.4.5 Neighboring Equatorial Ester Group Activation...... 30 3.5 Conclusions ...... 31 4 Applications of the Lattrell-Dax Epimerization...... 33 4.1 Application in Synthesis of β-D-Mannosides and Talosides ...... 33 4.1.1 Introduction...... 33 4.1.2 Double Parallel Inversion...... 33 4.1.3 Double Serial Inversion...... 35 4.2 Application in Synthesis of Thio-β-D-Galactosides ...... 37 4.2.1 Introduction...... 37 4.2.2 Synthesis of Methyl 3-Thio-β-D-Galactoside ...... 37 4.2.3 Synthesis of Methyl 4-Thio-β-D-Galactoside...... 38 4.3 Conclusions ...... 39 5 Enhanced Basicity by Supramolecular Anion Activation ...... 41 5.1 Supramolecular Activation in Cascade Inversion ...... 41 5.1.1 Triggered Cascade Inversion...... 41 5.1.2 Anion Activation...... 42 5.2 Enhanced Basicity by Supramolecular Effects ...... 44 5.2.1 Supramolecular Effects of Anions and Solvents...... 44 5.2.2 Basicity Controlled Cascade Reaction ...... 46 5.2.3 Enhanced Basicity by Supramolecular effects...... 48 5.3 Conclusions ...... 50 6 General Conclusions ...... 51

ACKNOWLEDGEMENTS APPENDIX REFERENCES

1 Introduction

1.1 Carbohydrates – A General Introduction

Carbohydrates are the most abundant natural products, including monosaccharides, , , and (Figure 1). They also include substances derived from monosaccharides, such as when the carbonyl group is reduced to form an alditol, one or more terminal groups are oxidized to form carboxylic acids, or a hydroxyl group is replaced by a hydrogen, amino-, or thiogroup. All members of the carbohydrates consist of monosaccharides as the basic units. There are a large number of monosaccharides, classified in several ways. For example, they are classified as furanoses (five-membered rings) and pyranoses (six-membered rings), etc, by the size of the ring. The stereocenter at C-1 is called the anomeric center, where the hydroxyl group connected to C-1 can alternate between pointing up and down, referred to as β and α, respectively. When this hydroxyl group is replaced with an aglycone moiety, the configuration is locked into either α or β and the term is used to classify this class of carbohydrates.

Monosaccharides

HO OH HO HO HO OH 6 OH 5 O O O 4 O HO 2 OH OMe HO 1 HO HO HO 3 HO HO NH2 OH OMe α-D-galactopyranose β-D-fucose methyl α-D-mannoside methyl 2-amino-β-D-galactoside

Disaccharides HO O OH OH OH O HO HO O O O O OH OH OH HO HO HO OH OH HO HO Sucrose (cane sugar) Lactose

Oligosaccharides OH HO OH O HO O HO O O O NH AcHN O O OH O HO OH Blood group antigen H glycoprotein (O-glycosidic)

Polysaccharides

OH OH OH NHAc O HO O HO HO O O HO O O OH NHAc OH OH Cellulose Chitin Figure 1 Natural carbohydrates.

1 with other large biomolecules such as lipids or proteins, playing essential roles in diverse biological processes.[1] Polysaccharides constitute the major volume of carbohydrates, biosynthesized by plants, algae, animals and microbes. For example, cellulose is found in all plants as the major structural component of the cell walls, and chitin is the principal structural component of the exoskeleton of invertebrates such as crustaceans and insects.

1.2 Carbohydrates – Challenging Synthetic Targets

Carbohydrates have attracted an increasing amount of attention up to today, on account of their diverse biological function. For example, specific protein-carbohydrate interactions are involved in cell differentiation, cell adhesion, immune response, trafficking and tumor cell metastasis.[2-4] These important processes occur between carbohydrates (glycoproteins, glycolipids, and entities at cell surfaces) and lectins, proteins with carbohydrate-binding domains. Carbohydrates are also widely used in medicine, for example as anticoagulants, antibiotics and vaccines.[5, 6]

Uncovering the contributions of carbohydrates in cell biology would greatly promote advancements in the biological and medical areas. However, the functions of carbohydrates in biology have not been extensively studied due to the high complexity of oligosaccharides and to a lack of general methods for synthesizing and analyzing these molecules. One important case is β-mannoside synthesis. The β-mannopyranosidic linkage is a common structural element in a wide range of natural products.[7-10] This biologically important and widespread class of structures contains β-D-Manp units as the relevant component. For example, the β-D-Manp unit is present as a central component in the ubiquitous N-glycan core structure of glycoproteins,[7] and makes part of a range of fungal and bacterial entities (Figure 2).[11, 12]

OH OH O HO HO HO OH O O HO OH HO OH NHAc O HO O HO O O O O HO OH NHAc N-linked pentasaccharide core structure

HO OH HO O OH OH OH HO O OR HO O HO OR HO OH O HO β-D-Manp Fungal metabolite deacetyl-caloporoside Figure 2 Natural entities containing β-mannopyranosidic linkages.[11, 12]

The chemical synthesis of this 1,2-cis-mannosidic linkage is, however, especially difficult. The α-mannosidic linkage is strongly favored because of the concomitant occurrence of both the anomeric effect and the repulsion between the axial C-2 substituent and the

2 approaching nucleophile (Figure 3). Moreover, neighboring group participation of a 2-acyl substituent leads to α-mannosides only.[1]

R

ROCO POH O O O O O O OP

OP OP Favored by the Favored by the Partial neighboring group participation anomeric effect dipole moments

α β HO O O 1,2-trans OP

OP OH easiest easy

HO O O 1,2-cis OP

HO OP harder hardest Figure 3 Anomeric effect and neighboring group participation.

Another important case is thiosaccharide synthesis. Thiosaccharides, where an exocyclic oxygen is replaced by a sulfur atom, constitute an increasingly important group of compounds in glycochemistry, possessing unique characteristics compared to their oxygen analogs (Figure 4).

HO O HO HO SH OH HO 1-thio-β-D-glucose OH HO OH O HO HO HO S S OH HO OH O HO O O HO HO O S HO S O HO O O HO O OH HO S HO HO O OH HO HO OH OH partial structure of the cell cyclic thiooligosaccharides wall phosphomannan antigen modified by a sulfur atom Figure 4 Thiosaccharides.[13-15]

These compounds are often used as efficient glycoside donors and acceptors in and neoglycoconjugate synthesis,[16-22] because the thiolate is a potent nucleophile and a weak base that reacts easily and selectively with soft electrophiles. Furthermore, the resulting thioglycosides and S-linked conjugates possess increased

3 resistance to degradation by glycosidases potentiating their use as efficient building blocks in drug design and therapeutics.[23] Generally, in order to obtain those carbohydrate targets, economic and efficient synthetic strategies have to be designed.[14, 15, 24, 25] Two of the most important strategies in carbohydrate synthesis are epimerization and regioselective protection/deprotection.

1.3 Regioselective Protection/Deprotection

Regioselectivity is a prominent challenge in carbohydrate chemistry since carbohydrates contain several hydroxyl groups of similar reactivity. Selective protecting groups and efficient protecting group strategies are therefore of crucial importance to efficiently obtain desired carbohydrate structures. The most common protecting groups for hydroxyl functions are esters, ethers, and acetals. Carbohydrate hydroxyl groups differ somewhat in reactivity depending on whether they are anomeric, primary or secondary, and also depending on their configurations. For example, in order to obtain 3- or 4-thio-β-D- galactosides (Scheme 1), reasonable epimerization and regioselective protection /deprotection strategies have to be used.[26]

HO OH AcO OAc HO OH O protection/ O epimerization O OR epimerization /deprotection HO OR HS OR OH OAc OH OH

OH OBz HS OH HO O regioselective HO O epimerization O /protection /deprotection HO OR BzO OR HO OR OH OBz OH Scheme 1 Synthesis of methyl 3- and 4-thio-β-D-galactopyranosides.

1.4 Epimerization

Epimerization of carbohydrate structures to the corresponding epi-hydroxy stereoisomers is an efficient means to generate compounds with inverse configuration that may otherwise be cumbersome to prepare. Several different synthetic methods have been developed, including protocols based on the Mitsunobu reaction,[27] sequential oxidation/reduction routes[28] as well as enzymatic methods,[29] all of which having their respective advantages and shortcomings. A common route to stereocenter inversion in carbohydrate chemistry involves the triflation of a given hydroxyl group, followed by substitution using a variety of nucleophilic reagents (Scheme 2). This method was used by Dax and co-workers who first reported that glycoside triflate displacement by nitrite ion,[30] a reaction first found by Lattrell and Lohaus,[31] produced carbohydrates with inversed hydroxyl configuration under very mild conditions. Despite its reported efficiency,[16, 32-34] the Lattrell-Dax method has unfortunately not been extensively adopted, likely because of difficulties in predicting the product outcome.

4 RO OR RO OR 1. py, Tf O, O 2 O CH2Cl2 OMe OMe HO 2. KNO , OR 2 OR DMF OH Scheme 2 Lattrell-Dax epimerization.

Binkley[35] reported a simple technique for converting methyl 2,6-dideoxy-β-D-arabino/ galacto-hexopyranosides into the corresponding ribo- and lyxo-isomers through internal triflate displacement by a neighboring benzoyl group and a direct inversion method through triflate displacement by nitrite ion when neighboring participation could not take place (Scheme 3). He further reported that the inversion reaction appeared to be related to the configuration, but no explanation was given.

O O BzO H2O HO TfO OMe CHCl3 OMe neighboring group participation OBz

O O BzO nitrite BzO OMe OMe Fast TfO toluene OH

BzO BzO O nitrite O Slow TfO OMe toluene OMe OH Scheme 3 Effect of carbohydrate configuration on inversion reaction.[35]

More recently, von Itzstein and co-workers[19] needed to perform a 3-position glycoside inversion reaction when they developed a new approach toward the synthesis of lactose- based S-linked sialylmimetics of α-(2,3)-linked sialosides. Their strategy however failed when they chose a glycoside where one hydroxyl group in 3-position was free and the other positions protected with benzyl groups (Scheme 4). Interestingly, they obtained a satisfactory result when the 2-position benzyl group was replaced by a benzoyl group. It clearly showed that the choice of protecting group was crucial for the inversion of the configuration at the 3-position of the galactose moiety. In light of these studies, it seems that the type and configuration of the neighboring protecting group is crucial for the reactivity in the Lattrell-Dax inversion. An equatorial trans-configuration is favored for the inversion. However, the trans-configuration is also favored for the neighboring group participation. Thus, a question can be put forward: can the neighboring ester group activate the nitrite inversion process via a neighboring group participation mechanism?

5 Ph Ph

O O O O O nitrite O TfO OR DMF OR OBz OBz Ph OH

O O O nitrite failed TfO OR DMF OBn Scheme 4 Effect of the neighboring groups on the inversion reaction.[19]

1.5 Neighboring Group Participation

The neighboring group participation (NGP) mechanism requires two conditions: a neighboring ester group and trans-configuration.[26, 36] In the polar solvent DMF, the neighboring group participation reaction takes place immediately. However, in the nonpolar solvent toluene the neighboring group participation is restrained. This indicates that neighboring group participation is favored in polar solvents. Further analysis showed that the products of the neighboring group participation were always compounds where the ester group is axial and the hydroxyl group equatorial. The explanation was given by King[37] and Binkley[36] (a in Scheme 5) according to Deslongchamps´ stereoelectronic theory[38, 39]. a)

HO O R O O H2O O O R major R O TfO O O O O O HO O O H O R OH major R 2 O O minor b) HO

TfO O nitrite O O toluene O O O R R

Scheme 5 Comparison of the products formed by NGP reaction and the nitrite inversion. a): Mechanism of major product formation by neighboring group participation given by Binkley.[36] b): Nitrite-mediated inversion.[35]

6 In addition, for the Lattrell-Dax nitrite-mediated inversion, it is clear that the ester group always remains in the same position and the hydroxyl group is inversed on the carbon atom directly connected to the triflate group (b in Scheme 5). This indicates that the effect of the neighboring ester group on the Lattrell-Dax inversion is not via a neighboring group participation process.

1.6 Aim of Study

The main aims of this thesis has been to investigate the effects of the protecting group pattern on the Lattrell-Dax nitrite-mediated inversion reaction, and by making use of this very important reaction to synthesize thio-β-D-galactoside derivatives and develop efficient methods for the synthesis of β-D-mannosides and talosides. In order to investigate the Lattrell-Dax reaction, a series of galacto- and gluco-type derivatives, where one hydroxyl group in the 2, 3, or 4-position is free and the other positions are protected with acetyl, benzoyl, or benzyl/benzylidene groups were chosen for further evaluation (Figure 5). These compounds would be tested in the Lattrell-Dax nitrite-mediated reaction.

Ph O Ph O O O O O BzO OMe BnO OMe Ph OH 12OH O AcO OAc BzO OBz O O O O HO OMe HO OMe HO OMe OAc 3 OBz 45OBn

OBz HO OBz HO OBn OBn HO O O O HO O BzO OMe BzO OMe BnO OMe BnO OMe OBz 6 OBz 7 OBn 8 OBn 9 Figure 5 Galacto- and gluco-type derivatives with different protecting group patterns.

To further analyze and explore the effect of the neighboring ester group configuration on the reactivity, other systems were designed. To avoid effects from the 2- and 6-positions and to isolate the effects arising from ester groups in the 3- and 4-positions, the 2- and 6- positions were protected with benzyl ether groups (Figure 6). Thus, a range of compounds where one of the hydroxyl groups in the 3- or 4-position is protected with an acetyl group had to be prepared and subsequently tested in the Lattrell-Dax epimerization reaction. However, all compounds mentioned above first had to be synthesized. Therefore we had to make use of or develop efficient regioselective protection methods before these investigations.

AcO OBn OBn OBn HO OBn OBn OBn O AcO O AcO O O HO O HO O HO OMe HO OMe OMe AcO OMe AcO OMe OMe OBn 10 OBn 11 OBn 12 OBn 13 OBn 14 OBn 15 OH OAc Figure 6 Methyl glycoside derivatives where the 2- and 6-positions are protected with benzyl ether groups.

7

8 2 Regioselective Carbohydrate Protection

2.1 Common Protection Strategies

In order to efficiently obtain the desired carbohydrate compounds 1-15, selective protecting groups and efficient regioselective protection strategies are of crucial importance. The most common protection strategies for carbohydrates are acylation and alkylation. In some case, they can be directly used to protect certain specific hydroxyl groups by making use of reactivity differences. Furthermore, organotin acylation and alkylation are often used to increase the selectivity. However, in most cases, it requires the integration of several protection strategies to efficiently synthesize the desired protected carbohydrates.

2.1.1 Acylation

Acylation can be used to protect hydroxyl groups with an acyl group. The most common acylating reagents include benzoyl chloride, acetyl chloride and acetate anhydride. Usually, these reagents are used in pyridine at room temperature to protect all free hydroxyl groups. However, at low temperature regioselectively acylated structures may also be obtained. For example, methyl 2,3,6-tri-O-benzoyl galactoside 7 could be efficiently synthesized in 60% yield by a one-step esterification process at -40 oC, starting from galactoside 16 (Scheme 6).

HO OH HO OBz BzCl O O OMe py, CH2Cl2 OMe HO o BzO OH 16 -40 C OBz 7 Scheme 6 Synthesis of compound 7.

2.1.2 Alkylation

An often used alkylation method in carbohydrate chemistry is benzylation. Especially, the 4- and 6-positions of a pyranoside can be regioselectively protected by a benzylidene group, which can be reductively opened from the 4- or 6-position to obtain a free hydroxyl group and a benzyl group. The glycoside derivatives 5 and 9 could be synthesized by this benzylation method. Starting from galactoside 16 and glucoside 17, the 4,6-O-benzylidene 18 and 19 respectively were synthesized first. Compounds 18 and 19 were then allowed to react with benzyl bromide in the presence of sodium hydride, producing 4,6-O- benzylidene-2-O-benzyl galactoside 5 (in 30% total yield) and 4,6-O-benzylidene-2,3-di- O-benzyl glucoside 20 (Scheme 7). After the benzylidene ring of 20 was opened by reduction, the 2,3,6-tri-O-benzyl glucoside 9 was finally obtained in 72% overall yield.

9 Ph Ph

O O HO OH O O

O PhCH(OMe)2 O BnBr, NaH O HO OMe DMF, H HO OMe DMF HO OMe OH 16 OH 18 OBn 5

OH OBn Ph O Ph O Et3SiH, O O O O HO PhCH(OMe)2 O BnBr, NaH O CF3CO2H HO OMe HO OMe DMF, H HO DMF BnO OMe CH Cl BnO OMe OH 17 OH 19 OBn 20 2 2 OBn 9 Scheme 7 Synthesis of compound 5 and 9.

2.1.3 Organotin Protection

For obtaining mono-substituted compounds in one or a few steps, the use of organotin reagents such as tributyltin oxide or dibutyltin oxide[40], provide useful means to efficient regioselective acylations[41-44], alkylations[41, 45-47], silylations[48], sulfonylations[41, 49, 50] and [51-53]. Stannylene acetals are easily prepared by treatment of carbohydrates with organotin in methanol at reflux condition, and generally lead to intermediate structures with predictable reactivities. In these reactions, stoichiometric amounts of organotin reagent are normally used. 4,6-O-Benzylidene-3-O-benzoyl glucoside 1 and 4,6- O-benzylidene-3-O-benzyl glucoside 2 can for example easily be obtained through regioselective organotin-mediated protection (Scheme 8). Starting from the 4,6-O- benzylidene 19 with 1.1 equivalent of dibutyltin oxide in methanol at 70 oC, a stannylene intermediate can be obtained after removing the methanol. When the stannylene intermediate was treated with benzoyl chloride or benzyl bromide in toluene, glucoside 1 (in 55% yield) and 2 (50% yield) can be obtained respectively. The relatively low yield was in these cases caused by the similar reactivity of the hydroxyl groups in the 2- and 3- positions of 19.

Ph O 1.Bu2SnO, Ph O 1.Bu2SnO, Ph O O O O O MeOH O MeOH O BnO OMe 2.BnBr, TBAI HO OMe 2.BzCl, BzO OMe OH 2 toluene OH 19 toluene OH 1 Scheme 8 Synthesis of compounds 1 and 2.

2.1.4 Integrated Protection Strategies

Most of the glycoside derivatives were synthesized using a combination of esterification, benzylation and organotin protection strategies. Some required only a few steps, whereas others were more cumbersome. The synthesis of methyl 2,4,6-tri-O-acetyl galactoside 3 and methyl 2,4,6-tri-O-benzoyl galactoside 4 was somewhat more complex than the synthesis of compounds 1 and 2. The hydroxyl group in the 3-position of galactoside 16 was first protected with a benzyl group by regioselective tin oxide benzylation, and then the obtained compound 21 was acylated in the presence of pyridine in methanol to form compounds 22 and 23. Finally, after removing the benzyl group in the 3-position by

10 catalytic hydrogenation, both methyl galactosides 3 and 4 were acquired in more than 70% total yield (Scheme 9).

HO OH HO OH AcO OAc 1.Bu2SnO O O MeOH Ac2O, py O OMe OMe HO 2.BnBr, TBAI, BnO MeOH BnO OMe OH 16 toluene, 90 oC OH 21 OAc 22 BzCl, py Pd, MeOH H2 BzO OBz BzO OBz AcO OAc

O Pd, H2 O O HO OMe BnO OMe HO OMe OBz 4 OBz 23 OAc 3 Scheme 9 Synthesis of compound 3 and 4.

Generally, several steps are required to synthesize glycoside derivatives where one of the hydroxyl groups in the 3- or 4-position is protected with an acetyl group and the 2- and 6- position are blocked with benzyl groups. However, the methyl 4-O-acetyl-2,6-di-O-benzyl galactoside 10 could be relatively easily obtained in more than 70% total yield via a one- pot reaction (Scheme 10).[54] Removal of the acetyl group in compound 10 resulted in 2,6- di-O-benzyl galactoside 26, which could easily be converted into 2,3,6-tri-O-benzyl galactoside 8 or 3-O-acetyl-2,6-di-O-benzyl galactoside 13 by organotin methods (Scheme 10). Furthermore, the 2,6-di-O-benzyl compounds 11, 12, 14 and 15 can be synthesized by epimerization and migration starting from compounds 10 and 13.

HO OH EtO O OH EtO O OBn

O CH3(OEt)3 O BnBr, NaH O HO OMe THF, H O OMe THF O OMe OH 16 OH 24 OBn 25

H

HO OBn HO OBn AcO OBn 1.Bu2SnO O MeOH O MeOH O MeONa AcO OMe 2.Ac2O HO OMe HO OMe OBn 13 toluene OBn 26 OBn 10

1.Bu2SnO MeOH 2.BnBr, TBAI toluene HO OBn O BnO OMe OBn 8 Scheme 10 Synthesis of compounds 8, 10, and 13.

2.2 Organotin Multiple Esterification (Paper I)

Of particular interest in carbohydrate protection is the possibility of acquiring multiple protections in single step processes. The organotin method provided a good method for

11 such multiple protection strategies.[55] The unprotected glycoside was first treated with excess (2-3 equivalents) of dibutyltin oxide in methanol at reflux condition, producing a stannylene intermediate that was not isolated. This intermediate was subsequently treated with the acylation reagent to yield the protected products in a one-pot process. For example, 2,3,6-tri-O-benzoyl galactoside (or glucoside) 7 (or 6) can easily be obtained by treatment of the stannylene intermediates, formed by the reaction of unprotected galactoside 16 (or glucoside 17) and 3.3 equivalents of dibutyltin oxide, with 3.3 equivalents of benzoyl chloride (Scheme 11).

OH HO OBz HO 1. Bu2SnO O MeOH O OMe HO OMe 2. BzCl, rt, BzO OH 16 toluene OBz 7 85%

OH OBz 1. Bu2SnO MeOH O HO O HO OMe HO OMe 2. BzCl, rt BzO OH 17 toluene OBz 6 85% Scheme 11 Synthesis of compound 6 and 7 by organotin multiple benzoylation.

Further studies indicated that the multiple esterification processes were highly dependent on the acylation reagents and the polarity of the solvents. Different protection patterns could be acquired from the same starting material by control of temperature, acylation reagents, reagent mole ratio, and solvent polarity. In the course of these studies, it was found that the benzoyl group can migrate to 3- and 4-position from 2- and 3-position at high temperature (Scheme 12). Thus, the temperature could be used for dynamic migration control.

OBz HO OH 1. Bu2SnO HO O MeOH O OMe HO OMe 2. BzCl, rt, BzO OH 16 toluene OH 27 90%

OH BzO OBz HO 1. Bu2SnO O MeOH O o OMe HO OMe 2. BzCl, 90 C HO OH 16 toluene OH 28 85%

HO OH BzO OBz 1. Bu2SnO O MeOH O o OMe HO OMe 2. BzCl, 90 C BzO OH 16 toluene OH 29 90% Scheme 12 Multiple benzoylation controlled by temperature and reagent mole ratio.

According to the proposed organotin acyl group migration mechanism (Figure 7),[50, 56-58] the resulting tin alkoxide intermediate is able to attack the acyl carbonyl group. However, it is reasonable to assume that acylation reagents in general are able to migrate under the same conditions. And yet, different from benzoyl chloride, it was found that migration could be observed with acetyl chloride at room temperature, whereas acetic anhydride proved inefficient under these conditions (Scheme 13). On the other hand, product 33,

12 protected in the 3,6-positions, was obtained with acetic anhydride whereas product 34, protected in the 2,6-positions, was obtained with benzoyl chloride at room temperature. Since no migration resulted with either acetic anhydride or benzoyl chloride at room temperature, it is apparent that, in this case, the results controlled by the acylation reagents were not brought about by this organotin acyl group migration (Scheme 13). Good selectivity was always obtained when the esterification reactions were done in a more polar solvent. The reason is likely due to decreased reactivity of the esterification reagent from solvent-induced destabilization of the stannylene intermediates.[59]

Bu Bu Cl Cl O Sn Bu Sn Bu R O OP O O OP O OP O OP O Bu O O O O O R Sn R O OMe O OMe O OMe O OMe OP OP Bu Sn OP Bu Cl OP Bu Cl Figure 7 Proposed organotin acyl group migration mechanism.

HO OAc O AcO OMe OH 32

Bu2SnO, MeOH Ac2O, DMF, rt 90% AcO OAc HO OH HO OAc Bu2SnO Bu2SnO O MeOH O MeOH O OMe OMe AcO AcCl, rt HO Ac2O, rt AcO OMe OH 31 toluene OH 16 toluene OAc 30 57% 70%

OBz OH OAc Bu2SnO Bu2SnO O O O HO MeOH HO MeOH HO OMe OMe OMe HO BzCl, rt HO Ac2O, rt AcO OBz 34 OH 17 OH CH3Cl DMF 33 51% 70% Bu2SnO, MeOH Ac2O, CH3CN, rt 85% OAc O HO AcO OMe OAc 35 Scheme 13 Multiple esterifications controlled by acylation reagents and solvents.

13

14 3 Lattrell-Dax Epimerization

3.1 Effects of Protecting Groups (Paper II)

All the glycoside derivatives, designed to explore the effect of the neighboring group on the Lattrell-Dax epimerization, were synthesized via the use of esterification, benzylation or organotin methods. It was hypothesized that, whenever the triflate group is in 2-, 3-, or 4-position of these pyranosides, good inversion yields would be obtained with neighboring ester groups, whereas poor inversion yields or complex mixtures would be obtained with neighboring benzyl groups. Furthermore, good inversion yields would be obtained with only neighboring equatorial ester groups, whereas neighboring axial ester groups would be inefficient. Our first approach was to investigate the effect of the protecting group pattern on the inversion reaction.

3.1.1 Effects of Protection Patterns

Initially, glycoside derivatives carrying a triflate group in the 3-position were subjected to the examination. In order to compare the effects of different ester groups, two types of ester-protected galactopyranosides (3, 4) were synthesized.

AcO OAc AcO OAc 1. py, Tf2O, O O CH2Cl2, 2h OMe OMe HO 2. KNO2, 3h OAc 3 o OAc 36 73% DMF, 50 C OH

BzO OBz BzO OBz 1. py, Tf2O, O CH2Cl2, 2h O OMe OMe HO 2. KNO2, 6h OBz 4 DMF, 50 oC OBz 37 77% Ph OH

O O 1. py, Tf O, O 2 CH2Cl2, 2h Complex HO OMe mixture OBn 5 2. KNO2, 3h DMF, 50 oC Scheme 14 Epimerization of glycosides where the 3-OH is unprotected.

As can be seen (Scheme 14), good yields were in these cases obtained only on the condition that esters were chosen as protecting groups, benzoyl groups being slightly less activating than the acetyl counterparts. When the ester protecting groups were replaced by benzyl/benzylidene groups, a mixture of different products was instead obtained.

Similar results were obtained from the epimerization of glycopyranosides where the hydroxyl group in the 4-position was unprotected, and all other positions were protected with either benzoyl or benzyl groups (Scheme 15). Only when an ester group was present at the carbon adjacent to the carbon atom carrying the leaving triflate group did the

15 reaction proceed smoothly, the axially oriented triflate being less reactive than the equatorial leaving group.

OBz HO OBz 1. py, Tf O, O 2 O HO CH2Cl2, 2h BzO OMe BzO OMe OBz 6 2. KNO2, 2h OBz 7 70% DMF, 50 oC

HO OBz OBz 1. py, Tf O, O 2 O CH2Cl2, 2h HO BzO OMe BzO OMe OBz 7 2. KNO2, 5h OBz 6 75% DMF, 50 oC

HO OBn 1. py, Tf2O, O CH2Cl2, 2h Complex OMe BnO 2. KNO ,0.5h mixture OBn 8 2 DMF, 50 oC

OBn 1. py, Tf2O, O HO CH2Cl2, 2h Complex OMe BnO 2. KNO , 0.5h mixture OBn 9 2 DMF, 50 oC Scheme 15 Epimerization of glycosides where the 4-OH is unprotected.

In contrast to this effect, no efficient reaction occurred when benzyl groups were employed where compound mixtures were instead rapidly obtained. These results suggest that a neighboring ester group is able to induce or activate the inversion reaction, whereas an ether derivative is unable to produce this effect. The results also showed that the inversion reaction proceeded smoothly regardless of the triflate configuration.

Ph O 1. py, Tf O, Ph O OH O 2 O O CH2Cl2, 2h O OMe OMe BzO 2. KNO , 6h BzO OH 1 2 38 74% DMF, 50 oC

Ph O 1. py, Tf2O, O O CH2Cl2, 2h Complex OMe BnO 2. KNO , 3h mixture OH 2 2 DMF, 50 oC Scheme 16 Epimerization of glycosides where the 2-OH is unprotected.

Further tests were performed for glucopyranosides where the hydroxyl groups in the 2- position were free (Scheme 16). After observing the inversion behavior in the 3- and 4- position of the hexopyranosides, the 2-position was probed. The ester-protected glucopyranoside compound 1 afforded the inversion mannopyranoside product 38 in good yields, whereas the ether-protected compound 2 proved inefficient. In this case, slightly longer reaction times were however necessary due to the lower reactivity of the 2-OTf derivative.

16 3.1.2 Effects of Neighboring Group Configurations

It was demonstrated that a neighboring ester group was essential for the reactivity of the nitrite-mediated triflate inversion from the above experiments. To further analyze these findings and explore the effects of the neighboring ester group configurations on the reactivity, glycoside derivatives 10 to 15 were tested in the nitrite-mediated inversion reactions. To avoid the effects from the 2- and 6-positions and to isolate the effects arising from ester groups in the 3- and 4-positions, the 2- and 6-positions were protected with benzyl ether groups. The experimental results presented in Table 1 clearly indicated that the configuration of the neighboring ester group directed the reactivity of the epimerization reaction. Good inversion yields depended mainly on the relative configurations of the two groups, and only with the ester group in the equatorial position did the reaction proceed smoothly, regardless of the configuration of the triflate, whereas a neighboring axial ester group proved inefficient.

Table 1 Epimerization reactions studied.

Entry Reactant Time /h Product Yield /% AcO OBn _ 1 O 3 mixture HO OMe OBn 10 OBn OBn AcO O AcO O 2 HO OMe 1.5 OMe 69 OBn OBn 11 OH 12

OBn OBn O O AcO AcO 3 OMe 4 HO OMe 73 OBn OBn OH 12 11

HO OBn OBn O O 4 HO 75 AcO OMe 4 AcO OMe OBn 13 OBn 14 OBn HO OBn HO O O 5 AcO OMe 0.5 AcO OMe 72 OBn 14 OBn 13

OBn HO O _ 6 OMe 3 mixture OBn OAc 15 o o Reaction conditions: i: Tf2O , py, CH2Cl2, -20 C-10 C, 2h; ii: KNO2, 50 oC, DMF, 0.5-4h. All starting materials were consumed.

Rapid internal triflate displacements by neighboring acetyl or benzoyl groups will occur if the ester group and the leaving group have trans-diaxial relationships. This leads to products where the configuration is retained, thus excluding these combinations from the present investigation. This internal displacement is indicative of the fast formation of an intermediate acyloxonium carbocation, stabilized by polar solvent. In our cases, compounds 11 and 14 hold 3,4-trans configurations in diequatorial relationships, where the internal triflate displacement by the neighboring ester group is considerably less

17 efficient. Contrary to this situation, compounds 12 and 13 hold 3,4-cis configurations where the ester groups are in the equatorial positions. This structural situation largely excluded the conventional neighboring group participation.[60, 61]

Further study indicated that the nitrite-mediated reaction produced the same results also in acetonitrile, a mixture being produced without neighboring equatorial ester group. The reaction rates in the less polar solvent acetonitrile were always lower than in the more polar solvent DMF. Normally, Lattrell-Dax epimerization reactions are performed in polar aprotic solvents such as acetonitrile or DMF,[16, 19, 30, 32, 34, 62, 63] and nonpolar solvents are mainly chosen to avoid neighboring group participation.[16, 26, 64]

3.2 Neighboring and Remote Group Participation

3.2.1 Neighboring Group Participation Effects

The results obtained seem to point to the importance of a neighboring group (acyloxonium) effect. Compounds 11 and 14 (3,4-trans) expressed higher reactivity compared to compounds 12 and 13 (3,4-cis) as a result of the activation from the neighboring ester group inducing the inversion reaction. This is reflected in the longer reaction times for the 3,4-cis compounds, as displayed in Table 1. However, acyloxonium formation is still unlikely to be the sole explanation of the results for two reasons: first, starting compounds 12 and 13 both have a cis relationship between the ester and the leaving group, which largely disqualifies acyloxonium formation;[60, 61] and second, formation of a carbocation intermediate would result in a nucleophilic displacement from the triflate face of the compound leading to retention (double inversion) of configuration rather than single inversion (Figure 8).

OBn OBn OBn O Tf2O/py AcO AcO O KNO2 AcO O HO OMe CH Cl TfO OMe DMF OMe OBn 11 2 2 OBn OBn 12 OH DMF

OBn OBn O H2O O O HO OMe OMe O OBn OBn 15 OAc

OBn OBn HO OBn Tf O/py KNO HO O 2 TfO O 2 O OMe OMe AcO CH2Cl2 AcO DMF AcO OMe OBn 14 OBn OBn 13

DMF

O OBn AcO OBn O H2O O O OMe HO OMe OBn OBn 10 Figure 8 Comparison of nitrite-mediated inversion with neighboring group participation.

18 The importance of the acyloxonium formation in the trans-configuration cases was further supported by studies with addition of water. Compounds 11 and 14, both with 3,4-trans- diequatorial relationships, mainly yielded compounds 12 and 13 from reaction with potassium nitrite in dry DMF (Table 2). If on the other hand wet DMF without nitrite was used, compounds 15 and 10 were instead obtained as the main products. This suggests acyloxonium formation to the five-membered-ring intermediate, which rapidly collapses in the presence of water to produce the axial ester and the equatorial hydroxyl group. These results are indicative of (partial) acyloxonium formation in the trans-configuration cases, but that the nitrite ion is unable to open the five-membered ring from either the triflate face or from attacking the carbonyl cation, as has been suggested for water.[35] More importantly, the ester group is, therefore, likely to induce or stabilize the attacking nitrite ion regardless of the trans- or cis-configurational relationships. The effects observed for the ether-protected carbohydrates are likely a result of their lower degree of positive charge destabilization than the corresponding ester groups, leading to side reactions such as ring contraction and elimination.[65, 66]

Table 2 Water effects in studied nitrite-mediated inversion reactions.

Reactant Nucleophile Product Yield /% OBn OBn O O AcO KNO AcO HO OMe 2 OMe 69 OBn OBn 11 OH 12

OBn OBn O AcO O H O HO OMe 2 OMe 70 HO OBn OBn 11 OAc 15 OBn HO OBn O O HO KNO 72 AcO OMe 2 AcO OMe OBn 14 OBn 13 OBn AcO OBn O O HO H O OMe 70 AcO OMe 2 HO OBn 14 OBn 10 o o Reaction conditions: i: Tf2O , py, CH2Cl2, -20 C-10 C, 2h; ii: KNO2, o 50 C, DMF, 0.5-1.5h, or H2O, r.t., DMF, 6h.

3.2.2 Remote Group Participation

When the inversion of the triflate intermediates of compounds 29 and 31 was performed with nitrite in DMF, it was expected to acquire inversed compounds 39 and 41 in high yields since both compounds 29 and 31 contain a neighboring equatorial ester group. However, a mixture of two compounds was obtained in both cases. Further experiments in acetonitrile showed the same results, in which compounds 40 and 42 were also formed simultaneously besides the expected compounds 39 and 41. 1H-NMR experiments indicated that the formation of methyl talosides 39 and 41, where the hydroxyl group in the 2-position is unprotected, were more favored in less polar solvent acetonitrile (50%, 80%) and less favored in polar solvent DMF (45%, 40%), whereas the formation of methyl talosides 40 and 42, where the hydroxyl group in 4-position is free, were more favored in

19 polar solvent DMF (55%, 60%) and less favored in less polar solvent acetonitrile (50%, 20%). As a comparison, starting from the triflate intermediate of methyl 3,4,6-tri-O-acetyl galactoside 31, it was expected that the fully protected methyl taloside would be produced via the use of five equivalents of tetrabutylammonium acetate. However, the same mixture of methyl talosides 41 (52%) and 42 (48%) was produced (Scheme 17).

OBz OBz BzO OBz BzO HO OH OBz 39(%) 40(%) O O O a or b a 80 20 BzO OMe BzO OMe + BzO OMe OH b 40 60 29 39 40 OAc OAc AcO OAc AcO HO 41(%) 42(%) OH OAc O a, b or c O O a 50 50 AcO OMe AcO OMe + AcO OMe b 45 55 OH 31 41 42 c 52 48

o (a) i: Tf2O, py, CH2Cl2, ii: TBANO2, CH3CN, 50 C, 30h. o NMR-yields. (b) i: Tf2O, py, CH2Cl2, ii: TBANO2, DMF, 50 C, 20h. o (c) i: Tf2O, py, CH2Cl2, ii: TBAOAc, CH3CN, 50 C, 30h. Scheme 17 Epimerization by neighboring and remote group participation.

All of these results support a remote group (4-position) participation mechanism, where a six-membered ring is generated first, and then opened by trace water to produce either a free 4-hydroxyl group or a free 2-hydroxyl group in a reaction that is favored by polar solvents (Figure 9). The direct nitrite competition reaction resulted in that the 2-hydroxyl group products (39, 41) were favored in less polar solvents. In combination with the steric effects of the nucleophilic reagent, this also explains why a mixture of methyl talosides 41 and 42 were primarily obtained when tetrabutylammonium acetate was employed as a nucleophilic reagent.

AcO AcO OH AcO OAc O O O O O O H2O O AcO OMe AcO OMe AcO OMe OTf 43 OTf

OAc OAc HO AcO OAc OH O + O AcO OMe AcO OMe 42 41 Figure 9 Remote group participation.

To further support this mechanism, the triflate of methyl taloside 31 was directly tested in wet acetonitrile at 50 oC for 20 hours. As a result, a mixture including methyl talosides 41 and 42 was also obtained. However, addition of the nucleophilic reagents tetrabutylammonium nitrite/acetate can increase the reactivity of the remote group participation. The test for neighboring group participation also supported this result. It

20 seems that not only the neighboring ester group can activate the nitrite-mediated epimerization but also the nitrite ion can activate the neighboring or remote group participation.

3.3 Supramolecular Control (Paper III)

Supramolecular control is an important advancement in modern synthetic chemistry,[67-69] enabling for example improved selectivities and enhanced reaction rates. Recognition- based proximity effects of participating reactants are generally involved in these systems, positioning the components by templating prior to the reaction sequence.

3.3.1 Unusual Solvent Effect

For the inversion of 3-OTf β-D-galactopyranoside derivatives, an unusual solvent effect was found. For example, when the β-D-galactopyranoside derivatives 3 and 4 were tested in the reaction, the reaction rates proved generally higher in more polar solvents (Table 3). Thus, the rate increased in the order: CHCl3 < CH2Cl2 < CH3CN < DMF. However, the results in toluene and benzene broke this trend. Although these two solvents have the lowest polarity, the reactions proceeded at a faster rate. When other glycoside triflate derivatives were tested in toluene or benzene, the reaction rates were always lower in less polar solvents. For example, inversion of the triflate intermediates of methyl β-D- galactopyranosides 29 and 31 was successful in DMF and acetonitrile (Scheme 17), whereas it completely failed in toluene. Furthermore, in contrast to the rapid reaction times for the inversion of the 3-position of β-D-galactopyranoside 3 in benzene or in toluene, no reaction occurred for its α-.

Table 3 Comparison of the reactivity of 3-position inversion of β-galactoside.

AcO OAc AcO OAc 1. py, Tf2O, O O 2. TBANO2, HO OMe 50 oC OMe a OAc 3 OAc 36 OH

BzO OBz BzO OBz 1. py, Tf2O, O O 2. TBANO2, HO OMe 50 oC OMe b OBz OBz 37 4 OH

Solvent DMF CH3CN CH2Cl2 CHCl3 Benzene Toluene Time/h 1 3 6 20 1 1 a Conv/% 87 91 80 60 90 89 Time/h 2 5 12 9 2 2 b Conv/% 92 89 80 35 91 91

21 3.3.2 Carbohydrate-Anion Complex

During 1H-NMR studies of these reactions, it was surprisingly found that certain signals were dramatically shifted when nitrite anion was added. More detailed studies were performed, and the 3-OTf intermediate 44 was analyzed in deuterated DMF, acetonitrile, chloroform, and benzene, respectively. Interestingly, it was found that the relative shift differences in absence and presence of nitrite were close to negligible in DMF, acetonitrile and chloroform, whereas pronounced differences were recorded in benzene (Table 4). The signals for the protons in the 1-, 3- and 5-positions were in this case shifted downfield by 0.82, 1.03 and 1.14 ppm, respectively, while the shifts of the 2-, 4-, and 6-protons remained largely constant.

Table 4 Comparison of 1H-chemical shifts of intermediate 44 with and without nitrite anion in various d-solvents.

BzO OBz O TfO OMe OBz 44

d-solvent H1 H2 H3 H4 H5 H6a H6b w/o nitrite benzene 4.11 6.16 5.10 6.07 3.42 4.22 4.58 w nitrite benzene 4.93 6.24 6.13 6.34 4.56 4.41 4.69 Δδ 0.82 0.08 1.03 0.27 1.14 0.19 0.11 w/o nitrite CDCl3 4.67 5.74 5.23 6.05 4.22 4.41 4.66 w nitrite CDCl3 4.68 5.63 5.30 5.96 4.26 4.32 4.54 Δδ 0.01 -0.09 0.07 -0.09 0.04 -0.09 -0.12 w/o nitrite CD3CN 4.85 5.60 5.60 6.04 4.40 4.42 4.55 w nitrite CD3CN 4.94 5.57 5.78 6.03 4.52 4.41 4.52 Δδ 0.09 -0.03 0.18 -0.01 0.12 -0.01 -0.03 w/o nitrite DMF 5.20 5.82 6.21 6.29 4.85 4.70 4.56 w nitrite DMF 5.25 5.85 6.28 6.32 4.91 4.72 4.60 Δδ 0.05 0.03 0.07 0.03 0.06 0.02 0.04

Similar effects were also recorded for the 2-OTf intermediate 45 (Table 5), where the 1-, 3-, and 5-protons were deshielded by 0.78, 0.44 and 0.88 ppm, respectively, in benzene, and the relative shifts of the 2-, 4-, and 6-protons were close to zero. In DMF, no deshielding could be seen. Similar results were obtained when the benzoyl group was replaced with an acetyl group for intermediates 44 and 45.

22 Table 5 Comparison of 1H-chemical shifts of intermediate 45 before and after addition of nitrite anion in DMF and benzene.

BzO OBz O BzO OMe OTf 45

d-solvent H1 H2 H3 H4 H5 H6a H6b w/o nitrite benzene 3.96 5.42 5.53 6.07 3.47 4.14 4.58 w nitrite benzene 4.74 5.46 5.97 6.20 4.35 4.26 4.67 Δδ 0.78 0.04 0.44 0.13 0.88 0.12 0.09 w/o nitrite DMF 5.30 5.13 6.10 6.09 4.85 4.53 4.66 w nitrite DMF 5.30 5.15 6.10 6.10 4.85 4.53 4.65 Δδ 0.00 0.02 0.00 0.01 0.00 0.00 0.01

Furthermore, in contrast to the results for the β-anomer, no effects were observed for the α- form of the 3-OTf intermediates 46 and 47 (Table 6). All these results point to a supramolecular control effect. The carbohydrate structures present polar binding regions in their favored conformations, accentuated by electron-withdrawing protecting/leaving groups. Negatively charged species can thus interact with these structures, forming relatively strong molecular complexes. The nitrite ion can in this case be accommodated at the center of the pyranoside B-face to produce a carbohydrate complex, apparently reinforced by weak CH-O bonds, resulting in a 1-, 3-, 5-hydrogen deshielding effect. This also explains why the effect was only found in non-polar solvents, since competition in better solvating media hampers the binding effect.

Table 6 Comparison of 1H-chemical shifts of intermediates 46 and 47 with and without nitrite anion in d-benzene.

AcO OAc BzO OBz O O TfO TfO AcO BzO 46 OMe 47 OMe

Compound d-solvent H1 H2 H3 H4 H5 H6a H6b 46 w/o nitrite benzene 4.95 5.44 5.35 5.60 3.54 4.08 4.01 w nitrite benzene 4.95 5.42 5.35 5.60 3.60 4.08 4.01 Δδ 0.00 -0.02 0.00 0.00 0.06 0.00 0.00 47 w/o nitrite benzene 5.27 5.85 5.70 6.03 3.84 4.55 4.17 w nitrite benzene 5.27 5.85 5.71 6.04 3.86 4.55 4.18 Δδ 0.00 0.00 0.01 0.01 0.02 0.00 0.01

23 This model is further corroborated by the experimental results obtained from the Lattrell- Dax reaction. Especially for the inversion of the β-galacto-type 3-position in nonpolar solvents, an improved formation of this anion-carbohydrate complex controls the overall rate leading to accelerated reaction compared to more polar solvents (Figure 10). Although the host-guest complex is also formed between nitrite and derivatives 43 and 45, the outcome is unproductive since the inversion path of the reaction originates from the A- face.

RO OR RO OR O - O NO2 O Activated TfO OMe OMe OMe OR OR H 5 OH H3 H1 O- OR RO OR RO - OH O NO2 N Deactivated O O X RO OMe RO OMe OTf Figure 10 Supramolecular control from carbohydrate-anion recognition.

3.3.3 Binding Model

In order to further support the host-guest model, a quantum chemical study was performed. Compound 43 and tetramethylammonium nitrite were chosen as the model, and the binding mode and the association constant of the host-guest system were calculated (Figure 11). The binding constant could be in this case estimated to 1.0 x 102/M, and the guest nitrite anion was found to be situated slightly closer to the 1- and 3-hydrogens than to the 5-hydrogen at the carbohydrate B-face.

OAc AcO O AcO OMe H OTf 43 H H O-

N N O

Figure 11 Quantum chemical model of the nitrite-compound 43 complex.

The calculation predictions were subsequently confirmed by 1H-NMR titration experiments using compound 43 and tetrabutylammonium nitrite in d-toluene (Figure 12).

24 The association constant amounted to 7.4 x 102/M, and Job's analysis indicated a 1:1 ratio of the binding partners. The 1-, and 3- hydrogens were notably more deshielded than the 5- proton, suggesting that the nitrite anion is located closer to these in the complex. In principle, these results are compliant with the well known carbohydrate-aromatic interactions often found in carbohydrate-binding proteins,[70-74] where recent results suggest a “three point landing surface” via a CH 1-, 3-, 5-π interaction.[73, 74] Consequently, the anion recognition effect presented here may have unknown implications in biological recognition. To explore whether this supramolecular effect between β-glycosides which contain axial 1-, 3-, 5-hydrogen and anions in nonpolar solvents is general, these glycosides have been tested with different anions such as acetate, chloride anions and so on. For β-glycosides which have at least one unprotected hydroxyl group, it seems that a similar supramolecular effect can be observed, however the hydrogen bonding effect between the hydroxyl group and anions make the results more complicated. Experiments with fully protected glycosides indicated that the interaction between anions and the 1-, 3-, 5-hydrogen “bowl” from the carbohydrate B-face is related to the electron-withdrawing ability of the protecting groups.

OAc AcO OAc δ = 4.298 AcO O δ = 4.773 O AcO OMe H AcO OMe 5 OTf K H H H a 5 OTf 3 + 1 H H δ = 4.056 δ = 4.788 3 O- 1 O- Kd δ = 4.804 δ = 5.281 N N O O 5.6 0.20 H1 5.2 0.15 x

δ 4.8 H 0.10 5 Δδ

4.4 0.05 H3 4.0 0.00 0 1 2 3 4 5 0.0 0.2 0.4 0.6 0.8 1.0 - [NO2 ]/[43]0 mole fraction x of 43

Figure 12 Job´s plot for compound 43 with tetrabutylammonium nitrite in benzene. K = 7.4 x 102/M, R2 = 0.9996.

3.4 Ambident Reactivity of Nitrite Anions (Paper IV)

The ambident reactivity of the nitrite ion has been debated for a very long time.[75-82] The Hard-Soft Acid-Base theory (HSAB) or the Perturbation Molecular Orbital theory (PMO) has been used to explain how the O-attack or the N-attack generates mainly nitrite or nitro- products.[78, 79, 83] Recently, Mayr´s group suggested that nitrite anions are a third type of ambident anions (besides SCN- and CN-) which do not fit any previous theories.[84]

25 According to their experiments, the effect of the leaving groups on the O/N selectivity is small, and instead it seems that solvents and reagents play more important roles (Figure 13). It was also shown that when tetrabutylammonium nitrite was allowed to react with methyl triflate in chloroform, a mixture of nitrite and nitro products was obtained.

Solvent MeONO/MeNO2 EtOH 30/70

KNO2/[18]crown-6 MeCN 50/50 Me-OSO Me Me-ONO + Me-NO 2 2 THF 92/8

C6H6 85/15

TBANO2 CHCl Me-OSO2CF3 Me-ONO + Me-NO2 3 59/41 Figure 13 O/N selectivities for methylations of nitrite ions.[84]

3.4.1 An O/N Selectivity Case in Carbohydrate Epimerization

The phenomenon of O-attack or N-attack of benzoylcarbamate was reported in carbohydrate epimerization by Knapp´s group when using a benzoylcarbamate cyclization method for the synthesis of amino glycoside from precursors bearing a hydroxyl group and a nearby electrophilic carbon center (Figure 14).[85]

O O NCOPh OH Ph O O NHCOPh O OH H2N N TfO O HO O N-attack O-attack O O O O O

Ph O Ph O O O O O O TfO TfO O O OMe NHCOPh PhCOHN OMe O 48 51

Ph Ph

O O O O Ph O O O O O OMe TfO OMe TfO O OMe O O O TfO PhCOHN PhCOHN NHCOPh O O 49 50 52 Figure 14 O/N selectivities controlled by glycoside configuration.[85] It can be seen that glycosides 48-50 undergo N-attack to produce amino products and that glycosides 51 and 52 undergo O-attack to produce diol derivatives under the same conditions. The key step in the reaction is the proton abstraction at nitrogen creating a negative species that may undergo either N- or O-cyclization. However, it was difficult to predict the outcome of the reaction by the HSAB theory, whereas steric reasons could instead be used to explain the selectivity. The N-attack requires more free space than the

26 O-attack. For glycosides 48-50, the face which will be attacked by benzoylcarbamate anion has enough space to accommodate the whole N-attacking group. However, for glycoside 51 and 52, the methoxy and benzylidene groups are both on the A-face/B-face, blocking the attack of the more hindered side of the benzoylcarbamate anion (N-attack). As a result, only O-attack can occur. The above results show that O- or N- attack can be controlled by the glycoside configuration and protecting group pattern.

3.4.2 Carbohydrate Epimerization with Non-Ambident Reagents

Similar effects could be the cause leading to the complex mixture in the Lattrell-Dax inversion. In this case, the nitrite ester would be transformed to a hydroxyl group whereas the nitro products could lead to side reactions such as decomposition, elimination and ring contraction. In contrast, if carbohydrate triflate intermediates react with non-ambident nucleophilic reagents, a single product will mainly be formed, independent of any neighboring group to the triflate group. The experiments shown in Scheme 18 supported this point. Even though triflate intermediates of 2 and 10 lead to complex mixture with nitrite in acetonitrile, single products 53 and 54 were obtained in near quantitative yields with acetate. The experiments further support that it is the ambident reactivity of nitrite that leads to a mixture of products in the Lattrell-Dax epimerization.

1. Tf2O, py, CH2Cl2 Ph O Ph O OAc O 2. TBAOAc, O O MeCN O BnO OMe BnO OMe OH 253

1. Tf2O, py, AcO OBn CH2Cl2 AcO OBn O 2. TBAOAc, O MeCN HO OMe OMe OBn 10 OBn 54 AcO Scheme 18 Carbohydrate epimerization with single reactivity reagents.

3.4.3 Nitration Products

To obtain nitro products, the triflate intermediates of compounds 2, 9, and 10 were chosen to perform a nitrite-mediated epimerization in acetonitrile at room temperature. The experiments were followed by 1H-NMR. Compound 2 led to a very complex mixture (Scheme 19), but the nitro compound 56, where the nitro group is in equatorial position, could be separated from this mixture by column chromatography, instead of the expected nitro compound 55 where the nitro group is in the axial position. IR analysis of compound 56 indicated a strong absorption at 1550 cm-1, typical for nitro groups. Separation of compound 57 by column chromatography, however, always resulted in a mixture of compounds 57 and 58. Within 24 hours, compound 57 was totally converted into compound 58, supporting the presence of compound 57 in Scheme 19.

27 1. Tf2O, py NO Ph O 2. TBANO2, Ph O 2 Ph O ONO O O O O MeCN O O OMe BnO BnO OMe + BnO OMe OH 2 55 57 24h, 24h, r.t. r.t.

Ph O Ph O Ph O OH O O O O O O OMe + BnO BnO OMe BnO OMe 59 56 58 ONO NO2

1. Tf2O, py Ph O OH 2. TBANO2, Ph O Ph O O O O O MeCN O O BnO OMe BnO OMe + BnO OMe 56 59 58 NO2 ONO

1. Tf O, py OBn 2 O2N OBn ONO OBn 2. TBANO2, O O O HO MeCN OMe BnO OMe BnO OMe + BnO OBn 9 OBn 60 OBn 62

24h, 24h, r.t. r.t. OBn OBn HO OBn O O O ONO + O2N BnO OMe BnO OMe BnO OMe OBn 63 OBn 61 OBn 8

1. Tf O, py HO OBn 2 OBn OBn 2. TBANO2, O O O MeCN O2N ONO OMe + OMe BnO OMe BnO BnO OBn 8 OBn 61 OBn 63 Scheme 19. Carbohydrate nitro compounds forming from nitrite-mediate reactions.

Compound 9 also led to a complex mixture. Similar to the reaction with compound 2, instead of the expected inversed nitro product 60, nitro product 61, where the nitro group retains the same configuration as the triflate group, was separated from the mixture. According to the 1H-NMR experiments, it can be seen that the expected nitro compound 60 and the nitrite compound 62 were formed immediately upon addition of nitrite to the solution of compound 9. Then, the nitrite compound 62 was slowly cleaved to compound 8. The nitro compound 60 also disappeared. To further explore this, the triflate intermediates of compounds 58 and 8 were tested in the same way. The 1H-NMR spectra clearly indicated that compound 58 lead to 41% of the expected nitro compound 56 and 59% of nitrite compound 59, and that compound 8 led to 36% of the expected nitro compound 61 and 64% of nitrite compound 63. Both the nitrite compounds 59 and 63 were cleaved to the related compounds 2 and 9 in 24 hours at room temperature. All these results indicated that the compounds where the nitro group is in axial position are very unstable and undergo a second attack by the nitrite anion to produce nitro products where the nitro group is in the more stable equatorial position and nitrite products that are cleaved to the starting material. That starting material 2 was also separated from the above reaction mixture supports this conclusion. Further analyses indicated that the reaction mixture

28 produced with compound 2 includes compounds 2, 55-59 and an unknown compound resulting from the elimination of the nitro group. A trace of nitrite compound 63 could also be seen during the reaction of compound 9, which suggests that the nitro compound 60 should lead to nitro compound 61 and nitrite compound 63.

Further experiments were carried out to test the product distribution for the triflate intermediate of compound 64, a side product when preparing compound 2, where the triflate group is in the equatorial 3-position (Scheme 20). According to the 1H-NMR analysis, 55% of nitrite compound 65, where the nitrite group is in axial position, and 45% of nitro compound 66, where the nitro group is in equatorial position were formed. For the triflate intermediate of compound 10, due to its low reactivity in acetonitrile at room temperature, it proved difficult to find any nitrite- or nitro products from the 1H-NMR spectra. However, compound 10 was also separated from this reaction suggesting a neighboring group participation mechanism of nitro compound 67 (Scheme 20), where an axial nitro group in 3-position was formed first, followed by a neighboring group participation from the acetate group in 4-position.

1. Tf2O, py Ph O 2. TBANO2, Ph O Ph O O O O O MeCN O O OMe + HO OMe O N OMe 65 2 66 OBn 64 ONO OBn OBn

1. Tf2O, py AcO OBn AcO OBn AcO OBn 2. TBANO2, O MeCN O + O HO OMe 2 d OMe OMe OBn 10 OBn 67 OBn 68 NO2 ONO

O OBn AcO OBn O O O OMe OMe OBn OBn 69 HO Scheme 20 Nitrite-mediated epimerization with compounds 64 and 10.

3.4.4 Solvent Effect

When the triflate intermediate of compound 10 was tested in d-benzene (Scheme 21), due to the supramolecular control effect increasing the reactivity, nitrite compound 68 could be clearly seen before it decomposed. However, almost no nitro compounds could be identified. This is likely due to the solvent effect mentioned above, decreasing the extent of N-attack. This effect was also found in Binkley´s experiments.[35] As can be seen (Scheme 21), compounds 70 and 71 led to very good inversion yields in the nonpolar solvent toluene in the nitrite-mediated epimerization reaction. These compounds should otherwise lead to mixtures in polar solvents due to the absence of neighboring equatorial ester groups.

29

1. Tf2O, py AcO OBn AcO OBn AcO OBn 2. TBANO2, O benzene O 24h, r.t. O HO OMe 7h OMe OMe OBn 10 OBn 68 OBn 69 ONO HO

1. Tf2O, py BzO BzO 2. TBANO2, O toluene O HO OMe 18h, r.t. OMe 70 72 88% HO

1. Tf2O, py HO 2. TBANO2, O toluene O HO OMe 2d, r.t. OMe 71 73 96% BzO BzO Scheme 21 Nonpolar solvents decreasing the extent of N-attack.

3.4.5 Neighboring Equatorial Ester Group Activation

During the nitrite-mediated epimerization of carbohydrates with a neighboring equatorial ester group, neither nitro products nor nitrite products were observed when the same tests were performed at room temperature. The inversed hydroxyl group products were always formed in near quantitative yields. These results indicate that only O-attack occurred and that the nitrite products were cleaved simultaneously due to the effect of the neighboring ester group. Then the question is why the neighboring equatorial ester group can decrease the extent of N-attack and promote the cleavage of the nitrite group.

δ NO ONO O OH O H O ONO N 2 ONO 2 O 2 ONO ONO O ONO 2 2 δ 2 Scheme 22 A neighboring nitrate group activation mechanism.[86] During the of nitrite esters, Thatcher and coworkers suggested an intramolecular Lewis acid or charge transfer catalysis by the β-nitrate group, in which the β-nitrate substituent assists the departure of the leaving group (Scheme 22).[86] Their hypothesis was further supported by ab initio and semi-empirical molecular orbital - calculations, on the model compound O2NO[CH2]2O . This model can be applied also in the Lattrell-Dax reaction, where a neighboring equatorial ester group can play the same role as the neighboring nitrate group, whereas a neighboring axial ester group cannot (a in Scheme 23).

30

a) ON δ+ O O O ONO O H2O HO OMe OMe O OMe R O R O R O O δ− O b)

O O ONO - O O N O NO2 N O O O TfO TfO R O R O R O R O OMe OTf O O O O A B Scheme 23 A neighboring ester group assistant catalysis mechanism. However, it is more difficult to explain how the neighboring equatorial ester group can restrain the occurrence of N-attack. First of all, we tried to apply the HSAB theory to understand the experimental results. With neighboring ester protecting groups, the carbon atom carrying the triflate group becomes more positively charged and thus more hard. With ether protecting group, this carbon atom becomes less positively charged and behaves as more soft. The harder nitrite oxygen sites will thus be favored in attack at the harder electrophile with ester protecting groups, and the softer nitrogen site will be favored in attack at the softer electrophile with ether protecting groups. However, this model can not explain why the neighboring axial ester group does not have this effect, for example, in reactions with compound 10.

Another explanation is instead proposed in Scheme 23b. In this case, secondary interactions between the incoming nitrite and the neighboring ester group guide the reaction towards O-attack. A six-membered transition state is here formed and the O-attack preference is explained due to the interaction of the nitrogen center with the ester group. The axial ester group is in this case unable to form an efficient transition state. In principle, the incoming nucleophile may adopt the opposite angle where the oxygen interacts with the ester group rather than the nitrogen, and this may be more favorable following the HSAB theory, but this pathway appears to be unproductive. Ongoing molecular orbital calculations aim at delineating these effects. A model where the nitrite activates the neighboring group participation, and the carbonyl oxygen acts as the nucleophile, can also be envisaged. This model is justified from the sometimes observed rate acceleration of acyl migration in presence of nitrite, but can be largely ruled out in this case. Contrary to acyl migration, the reaction proceeds well with equatorial ester groups, independent of the configuration of the triflate moiety.

3.5 Conclusions

In conclusion, it has been demonstrated that esters play highly important roles in the Lattrell-Dax reaction, facilitating nitrite-mediated carbohydrate epimerizations. Despite the

31 higher reactivity of carbohydrate triflates protected with ether functionalities, these compounds proved inefficient in these reactions, where mixtures of compounds caused by the ambident reactivity of nitrite ion were rapidly obtained. Neighboring ester groups, on the other hand, could induce the formation of inversion compounds in good yields by restraining the occurrence of the N-attack. The reactions further demonstrated stereospecificity, inasmuch as axially oriented neighboring ester groups were unproductive and only equatorial ester groups induced the nitrite-mediated reaction. Equatorial esters thus favored O-attack, restrained the occurrence of N-attack, and activated the leaving of triflate group and the cleavage of the nitrite group. A supramolecular control effect was also found in this Lattrell-Dax reaction. This effect also proved sensitive to the carbohydrate structure, requiring a H1-, H3-, H5-cis pattern for efficient complexation. These findings expand the utility of this highly useful reaction in carbohydrate synthesis as well as for other compound classes.

32 4 Applications of the Lattrell-Dax Epimerization

4.1 Application in Synthesis of β-D-Μannosides and Talosides (Paper V)

4.1.1 Introduction

It has been demonstrated that a neighboring equatorial ester group plays a highly important role in the Lattrell-Dax (nitrite-mediated) carbohydrate epimerization reaction, inducing the formation of inversion compounds in good yields. These studies suggested that new, efficient synthetic methods to complex glycosides would be feasible under the guidance of this principle, where the activating ester groups should be able to control the inversion of two neighboring positions simultaneously. Thus, we next attempted to meet the synthetic challenges of β-D-mannoside synthesis.

In consequence to these synthetic challenges, several different synthetic methods have been developed for β-mannoside synthesis. These include Koenigs-Knorr coupling methods using insoluble silver salt promoters blocking the α-face of mannosyl halides,[87- 89] sequential oxidation/reduction routes,[24, 90, 91] use of 2-oxo and 2-oximinoglycosyl [92, 93] [94-97] [25, 98, 99] halides, use of intermolecular, or intramolecular, SN2 reactions and intramolecular aglycone delivery method, [100-107] inversion of configuration of α-mannosyl triflate donors,[108-110] epimerization of β-glucopyranosides to β-mannopyranosides through [33, 51, 52, 111, 112] [113-115] SN2 reactions, as well as enzymatic methods, all of which with their respective advantages and short-comings. The 1,2-cis-glycosidic linkage is present also in β-D-talopyranosides. However interesting, recently evaluated for their intriguing H- bonding motifs, these structures have been less investigated in part due to their cumbersome synthesis.[85, 116, 117] However, via the application of the Lattrell-Dax epimerization, novel and efficient methods to synthesize β-D-mannoside and β-D-taloside can be designed.

4.1.2 Double Parallel Inversion

The glycoside derivatives 27, 29 and 31-33, which were synthesized by the one-pot organotin multiple esterification strategy, were chosen as starting materials (Figure 15).

HO OBz BzO OBz AcO OAc HO OAc OAc O O O O O HO BzO OMe BzO OMe AcO OMe AcO OMe AcO OMe OH 27 OH 29 OH 31 OH 32 OH 33 Figure 15 Glycosides obtained by organotin multiple esterification.

The taloside derivatives can be acquired starting from 29 and 31 via the inversion of the 2- position, or starting from 33 via the double parallel inversion of 2- and 4-positions, if the equatorial ester group in the 3-position is able to activate the epimerization of the neighboring 2- and 4-positions at the same time (Figure 16). On the other hand, the

33 mannoside derivatives can be acquired starting from 27 and 32 via the same double parallel inversion strategy.

OPG HO OPG OH O O i) Tf2O, Py, CH2Cl2 HO OMe OMe R O ii) KNO2, DMF R O OH O O Figure 16 Double parallel inversion.

In order to evaluate whether an equatorial ester group in the 3-position would be able to activate the epimerization of the neighboring 2- and 4-positions at the same time, a series of inversion reactions was probed (Scheme 24). Galacto- and gluco-type derivatives 27-33 where the 3- and 6-positions were protected with acetyl groups and the other two positions left unprotected were subjected to conventional triflation by triflic anhydride followed by treatment with tetrabutylammonium nitrite in acetonitrile or toluene at 50 oC. In acetonitrile, when methyl 3,6-di-O-acetyl glucopyranoside 33 was used as reactant, methyl 3,6-di-O-acetyl talopyranoside 76 was obtained in 85% yield. In contrast, the double inversion of methyl 3,6-di-O-acetyl galactopyranoside 32 was not successful and a very complex mixture was produced. It was hypothesized that this effect is likely due to acetyl group migration and neighboring group participation from the 3-O-acetyl group. If this explanation would be valid, the products produced would constitute an inversed-type mixture, that is to say, only the free methyl β-D-talopyranoside would be obtained if the inversed mixture was not isolated but directly deprotected under basic conditions. The experimental results showed that only one compound was obtained following deprotection of the complex mixture, indicating that this hypothesis was indeed valid.

OBz HO OBz 1. py, Tf2O, CH Cl , 2h OH O 2 2 O HO BzO OMe 2. TBANO2, OMe o BzO OH 27 CH3CN, 50 C 74 70% 5h

OAc HO OAc 1. py, Tf2O, OH O CH Cl , 2h O 2 2 HO OMe AcO 2. TBANO2, AcO OMe OH 32 toluene, rt 75 76% 5h OAc OAc 1. py, Tf O, HO 2 OH CH2Cl2, 2h O HO O AcO OMe 2. TBANO2, AcO OMe o OH 33 CH3CN, 50 C 76 85% 5h OAc HO OAc 1. py, Tf2O, CH Cl , 2h OAc O 2 2 O AcO OMe 2. TBAOAc, AcO AcO OMe OH 32 CH3CN, rt 77 90% 5h Scheme 24 Double parallel inversion reagent and conditions.

34 It is however well known that benzoyl groups are less reactive than acetyl counterparts to migration, as well as for neighboring group participation. In addition, neighboring group participation is disfavored in non-polar solvents.[26, 118] Thus, in order to avoid these side reactions, both these approaches were tested. On the one hand, reactions with methyl glucoside 33 and methyl galactoside 32 were performed in the non-polar solvent toluene; on the other hand, the inversion of methyl 3,6-di-O-benzoyl galactopyranoside 27 was attempted in acetonitrile. For comparison, the triflate of methyl galactoside 32 reacting with tetrabutylammonium acetate in acetonitrile was also tested. When methyl glucoside 33 was inversed at 50 oC in toluene, the reaction time had to be prolonged to twelve hours to obtain product 76 in 85% yield. This result indicates, as expected, that the reactivity was decreased in non-polar solvent. In addition, both these approaches proved successful for the double inversion of the methyl galactosides, efficiently reducing the neighboring group participation.

4.1.3 Double Serial Inversion

During this epimerization process, it was found that the reactivity in the 4-position was much higher than in the 2-position. At room temperature, the epimerization reaction in the 4-position occurred instantaneously, completed within ten to twenty minutes, whereas in the 2-position the epimerization reaction proceeded very slowly under these conditions. This result incited us to make use of the reactivity difference between the different positions to develop a new method, stepwise inversion of the hydroxyl groups amounting to a double serial inversion protocol, by which carbohydrate structures where one position is a hydroxyl group and the other positions were protected with ester groups could be obtained (Figure 17).

i) Tf O, Py, CH Cl i) Tf O, Py, CH Cl OPG HO OPG 2 2 2 OPG 2 2 2 ii)TBAOAc, CH3CN ii) TBANO2, CH3CN OH O O O AcO AcO OMe OMe R O R O R O OMe OH OTf O O O

i) Tf O, Py, CH Cl OPG HO OPG i) Tf2O, Py, CH2Cl2 OPG 2 2 2 ii)TBANO , CH CN ii) TBAOAc, CH3CN OAc O 2 3 O O HO HO OMe OMe R O R O R O OMe OH OTf O O O Figure 17 Double serial inversion.

Using the same initial step for the double serial inversion strategy, from methyl glucoside 33, the 2,4-triflate intermediates 78 could be produced via a triflation process (Scheme 25). The 4-triflates of these intermediates were subsequently inversed to the corresponding 4- O-acetyl intermediates 79 by substitution with tetrabutylammonium acetate, followed by inversion of the 2-position by tetrabutylammonium nitrite, to yield a mixture of methyl 3,4,6-tri-O-acetyl taloside 81 and methyl 2,3,6-tri-O-acetyl taloside 82. Conversely, When

35 the 2,4-triflates of intermediates 78 were first inversed to the corresponding 4-hydroxyl groups intermediates 80 via the use of tetrabutylammonium nitrite, directly followed by inversion of the 2-position by tetrabutylammonium acetate, in this case, however, instead of the formation of product 82, the compound 83 was formed in 90% yield, due to the strong deprotonation effect of the acetate anion.

TBANO2, OAc OAc AcO OAc AcO HO CH3CN, 24h OH OAc O O + O OMe AcO OMe 90% AcO OMe AcO OTf 45% 45% Tf2O, py, 79 81 82 OAc OAc 90% CH2Cl2, 2h TBAOAc, O O toluene HO TfO OMe OMe AcO AcO TBANO , OH OTf 90% 2 33 78 CH3CN HO OAc TBAOAc, AcO OAc O CH3CN, 4h O O AcO OMe 90% OMe OTf 80 83 Scheme 25 Double serial inversion from 33.

Starting from methyl glucoside 32 the 2,4-triflate intermediate 84 could be produced as well (Scheme 26). The 4-triflate of this intermediate was subsequently inversed to the corresponding 4-O-acetyl intermediates 85 by substitution with tetrabutylammonium acetate, followed by inversion of the 2-position by tetrabutylammonium nitrite, to yield methyl 3,4,6-tri-O-acetyl mannoside 87. When the intermediates 84 were first inversed to the corresponding 4-hydroxyl groups intermediates 86 via the use of tetrabutylammonium nitrite, directly followed by inversion of the 2-position by two equivalents of tetrabutylammonium acetate, methyl 2,3,6-tri-O-acetyl mannoside 88 was efficiently produced.

TBANO , OAc 2 AcO toluene, 6h OH O O AcO AcO AcO OMe 90% AcO OMe OTf 85 87 Tf2O, py, HO OAc TfO OAc 90% CH2Cl2, 2h TBAOAc, O O toluene

AcO OMe AcO OMe TBANO2, OH OTf 90% CH3CN 32 84 OAc TBAOAc, AcO CH CN, 6h OAc O 3 O HO HO AcO OMe 75% AcO OMe OTf 86 88 Scheme 26 Double serial inversion from 32.

In addition, due to the fact that methyl glucoside 17 was produced in a lower yield (70%) than methyl galactoside 16 (90%), following the double serial inversion strategy, an alternative, more high-yielding, synthetic route to methyl taloside could be devised starting

36 from methyl galactoside 16 instead of methyl glucoside 17. Thus, compound 32 acquired from methyl galactoside 16 by the organotin method, could be inversed to the intermediate 78 via intermediate 84 (Scheme 27). As a result, the use of methyl glucoside 17 could be avoided and the overall yield increased.

HO OAc TfO OAc 1. TBANO2, OAc Tf2O, py, CH CN, 1h O O 3 O CH2Cl2 TfO AcO OMe AcO OMe AcO OMe OH OTf OTf 2. Tf O, py, 32 84 2 78 CH2Cl2 Scheme 27 Alternative double serial inversion strategy to intermediate 78.

4.2 Application in Synthesis of Thio-β-D-Galactosides (Paper VI)

4.2.1 Introduction

For sulfur-containing carbohydrates, the benzyl ether group is less attractive because of deprotection difficulties by common reduction protocols. Organic sulfur compounds are very poisonous to metal hydrogenation catalysts, consequently hampering the deprotection. Thus, the ester and benzylidene protecting groups remain the more prevalently used for the synthesis of sulfur-containing carbohydrates. Selective protection protocols can furthermore be used to generate a variety of structures by hydroxyl epimerization strategies. The Lattrell-Dax inversion reaction is an efficient means to generate compounds with inverse configuration, where neighboring ester groups are important for the inversion reactivity. Consequently, the ester protecting strategy is essential for the synthesis of sulfur-containing carbohydrates when Lattrell-Dax epimerization protocols are employed. In addition, due to neighboring group participation of ester functionalities in triflate-activated carbohydrates, where 5- or 6-membered acyloxonium intermediates may form during thiolation, the solvent also plays important roles.

4.2.2 Synthesis of Methyl 3-Thio-β-D-Galactoside

The methyl guloside derivative 36, where the hydroxyl group in the 3-position was unprotected and the other positions were protected with acetyl groups, was essential for the synthesis of the 3-thiogalactose derivative. Compound 36 was easily obtained by Lattrell- Dax epimerization from the methyl galactoside derivative 3. However, treatment of 36 with triflic anhydride, and subsequent thiolation with KSAc in DMF instead afforded side product 3, which was generated via the neighboring ester group participation followed by hydrolysis. Our study indicated that the efficient stereoselective synthesis of 3- thiogalactose derivative from the corresponding 3-OTf gulose derivatives, where ester protecting groups were used, were highly dependent on the solvent and the nucleophile concentration. It could be proposed that the choice of the solvent and the nucleophile concentration controlled the product formation. Neighboring group participation could be attenuated by performing the reactions at high nucleophile concentration in toluene.

37 Consequently, introduction of a triflate group at C-3 in 36 followed by the SN2 reaction with 40 equiv. of TBASAc in toluene at room temperature provided the desired compound 89 (Scheme 28). With this strategy, the byproduct 3-thiogulose derivative 91 was competitively formed in only 4 % yield. Final deprotection of compound 89 under Zemplén conditions afforded the methyl 3-thio-β-D-galactoside 90.

AcO OAc AcO OAc AcO OAc py, Tf2O TBANO2, 3h o O CH2Cl2, 2h O 50 C, MeCN O HO OMe TfO OMe 80% OMe OAc OAc OAc 3 OH 36

py, Tf2O CH2Cl2, 2h

OAc OAc AcO OAc AcO AcO TBASAc, O O toluene, r.t. O OMe + AcS OMe 78% OMe OAc OAc OAc SAc 91 89 OTf

NaOMe, 86% MeOH, XTBASAc, r.t., 2h DMF, r.t.

HO OH 89 O HS OMe OH 90

Scheme 28 Synthesis of methyl 3-thio-β-D-galactoside.

4.2.3 Synthesis of Methyl 4-Thio-β-D-Galactoside

In our preliminary study, the key intermediate methyl β-D-glucoside derivative 6, where the hydroxyl group in the 4-position was unprotected and the other positions were protected with benzoyl groups, was obtained in three steps from methyl β-D-glucoside 17 by ester group migration. A more convenient route to obtain this key intermediate was based on an one-pot organotin-mediated multiprotection procedure starting from methyl β- D-glucoside 17 (Scheme 29) or the Lattrell-Dax epimerization of methyl 4-OH galactoside derivative 7 which can be obtained by direct benzoylation. Our initial strategy included the introduction of a triflate at C-4 in compound 6, followed by inversion with potassium thioacetate in DMF to yield the desired C-4 inverted compound 92. Unfortunately, attempted thiolation of the triflate intermediate with potassium thioacetate in DMF afforded a reaction mixture. It was assumed that the problem was also caused by neighboring group participation from the 3- or 6-OBz group in the polar solvents. To suppress that, thiolation of the triflate intermediate with tetrabutylammonium thioacetate in toluene was instead used, in this case, affording the C-4 inverted compound 92. Subsequent deprotection of galactoside 92 yielded compound 93 in good yield.

38 OH OBz HO OBz O Organotin O Lattrell-Dax O HO esterification HO epimerization HO OMe BzO OMe BzO OMe OH 17 OBz 6 OBz 7

Tf2O, py, CH2Cl2, -20 - 0 oC OBz AcS OBz OBz AcS KSAc, TBASAc, O DMF, r.t. O toluene, r.t. O x TfO OMe BzO OMe OMe 76% BzO BzO OBz 92 OBz 92 OBz

NaOMe, MeOH, r.t. 86%

HS OH O HO OMe OH 93

Scheme 29 Synthesis of methyl 4-thio-β-D-galactoside.

4.3 Conclusions

In conclusion, by the application of the Lattrell-Dax epimerization, novel and convenient double parallel, and double serial inversion methods have been developed, by which methyl β-D-mannoside, methyl β-D-taloside, have been efficiently synthesized in very high yields at very mild conditions in few steps. By use of the reactivity difference of the hydroxyl groups in 2- and 4-positions, a range of methyl β-D-mannoside and methyl β-D- taloside derivatives could be easily synthesized. On the other hand, the methyl 3- and 4- thio-β-D-galactosides were also conveniently synthesized by the choice of the appropriate synthetic strategies, including the Lattrell-Dax epimerization.

39

40 5 Enhanced Basicity by Supramolecular Anion Activation

5.1 Supramolecular Activation in Cascade Inversion (Paper VII)

5.1.1 Triggered Cascade Inversion

During the double serial inversion, an unexpected behavior of the nucleophilic reagent was found and prompted us to study these reactions further (Scheme 25). When the purified intermediates 86 and 80 were subjected to five equivalents of tetrabutylammonium acetate in toluene or acetonitrile, expecting to obtain methyl β-D-mannoside 88 or taloside 82, methyl 2,3-anhydro-4,6-di-O-acetyl-β-D-mannopyranoside 94 and talopyranoside 83 were however obtained in near quantitative yields instead (Scheme 30).

OAc OAc OH OH O H2O O O HO OMe OMe OAc OAc 95 TBAOAc, O OAc O r.t. 2h O HO AcO O H OMe OAc OAc AcO OMe toluene OTf 86 94 OH H O H2O O AcO O AcO OMe OMe OH 96

OAc OAc OAc AcO AcO HO OAc AcO OH TBAOAc, H H O O r.t. 2h O H O 2 O O O OMe OMe OMe AcO OMe toluene OTf 80 83 OH 97 Scheme 30 Carbohydrate cascade epimerization controlled by acetate.

With acid, the 2,3-anhydro mannoside 94 was hydrolyzed to a mixture of 3,6-di-O-acetyl- β-D-altropyranoside 95 and 4,6-di-O-acetyl-β-D-altropyranoside 96, resulting from neighboring group participation and direct hydrolysis, whereas 2,3-anhydro taloside 83 was hydrolyzed to near quantitative yield of 4,6-di-O-acetyl-β-D-idopyranoside 97 (Scheme 30). For mannoside 94, when the reaction was performed in the polar solvent acetonitrile, the neighboring acetyl group can participate in the hydrolysis process, where attack from the acetyl group in the 4-position on C-3 opened the epoxide ring and formed a five-membered acetoxonium intermediate. This intermediate was subsequently opened by water, yielding the acetyl group in the axial position as the main product. Thus a mixture of 3,6-di-O-acetyl-β-D-altropyranoside 95 (75%) and 4,6-di-O-acetyl-β-D-altropyranoside 96 (25%) were obtained. When the reaction was performed in the non-polar solvent toluene, direct hydrolysis and neighboring group participation occurred competitively, and a mixture of altroside 95 (50%) and altroside 96 (50%) were obtained.

In order to evaluate how the intermediates 80 and 86 were transformed into 2,3-anhydro taloside 83 and the 2,3-anhydro mannoside 94, 1H-NMR-analyses were carried out (a, b in Figure 18), indicating that only starting materials and the final products co-exist in the

41 reaction mixture, and no build-up of intermediates could be recorded. These results suggest a base-dependent cascade reaction (Scheme 31), where the acetate anion initially acts as a base in deprotonating the 4-OH group. Then due to a dynamic acetyl group migration between the 4- and 3-positions, the 3-position alkoxide was generated that instantly attacked the 2-position. As results of this nucleophilic substitution, 2,3-anhydro taloside 93 and 2,3-anhydro mannoside 94 were produced. Base-promoted deprotonation being the trigger for the reaction cascade, similar results would also be obtained if the acetate anion is replaced with a different base. The NMR-analyses with ethylenediamine (EDA) further proved the base-dependent mechanism (c, d, e in Figure 25), where besides the starting materials and the final products, a large amount of intermediate accumulated.

Figure 18 Cascade reaction with 5 eq. of TBAOAc (a-b) or EDA (c-e). * and ¤ indicate resonances from compounds 83 and 80, respectively.

HO OAc O OAc O O OAc k1 Fast O O O k AcO OMe -1 O OMe O OMe OTf O OTf OTf 80 Fast k-2 k2

AcO OAc AcO OAc AcO OAc

O k1 O k3 O O HO OMe O OMe OMe OTf k-1 OTf 98 83 Scheme 31 Proposed cascade reaction mechanism for methyl galactoside 80.

5.1.2 Anion Activation

In order to further explore why the large amount of intermediate accumulated with EDA, more tests with compound 80 was performed (Table 7). With imidazole no reaction occurred, but with triethylamine (TEA) the cascade reaction was also initiated. The reaction proceeded however in this case very slowly and a lower amount of migration intermediate 98 accumulated compare to EDA (1-4 in Table 7). However, although nitrite

42 anion alone was unable to induce the cascade reaction (entries 5 in Table 7), the combination of nitrite or acetate with amine resulted in large rate accelerations (entries 6- 10 in Table 7). For example, one equivalent of acetate and ten equivalents of ethylenediamine yielded full conversion in one hour (entry 10 in Table 7). In contrast, acetate or EDA alone resulted in only 50% conversion after eight and forty hours, respectively (entries 4, 9 in Table 7).

Table 7 Cascade reaction of compound 80 with anionic reagents and/or amine base.

HO OAc AcO OAc O Nu / base O O AcO OMe Benzene, r.t. OMe OTf 80 83 Entry Reagent (eq.) Base (eq.) Time /h Yield /% 1 - IM (50) 72 - 2a - TEA (15) 96 -b 3a - EDA (30) 7 -b 4a - EDA (10) 40 -b c 5 TBANO2 (10) - 120 -

6 TBANO2 (10) EDA (30) 0.1 quant

7 TBANO2 (1) EDA (10) 4 quant 8 TBAOAc (5) - 2 quant 9 TBAOAc (1) - 8 -b 10 TBAOAc (1) EDA (10) 1 quant a Intermediate 98 formed. b 50% conversion. c Inversion obtained.

Interestingly, the large amount of intermediate 98 that accumulated with amine alone, rapidly disappeared when adding acetate or nitrite. This suggests that the anions are able to activate the cascade reaction. In this case, the acetate anion acts not only as a base, but also as an activator of the whole cascade reaction. The nitrite anion, on the other hand, exclusively acts as a cascade activator. According to these results, it could also be anticipated that amino acids should display strong activation abilities for the cascade reaction, carrying both an amine and an acetate group. Indeed, this proposition proved to be valid. Starting from compound 80, and adding only two equivalents of the TBA salt of either α-L-alanine or β-alanine in benzene, the 2,3-anhydro product 83 was obtained in quantitative yield within one hour (Scheme 32). Based on these results, improved triggered cascade reactions directly starting from intermediates 78, 84 could be designed (Scheme 32), where the cascade sequence involved two inversions, migration and epoxidation. By combination of tetrabutylammonium nitrite with EDA, the cascade reaction proceeded smoothly and compounds 83 and 94 were produced directly in up to 90% yield. In these cases, EDA could not attack the 4-position and only maintained a basic condition. The nitrite ion not only triggered the entire cascade reactions by substitution and inversion of the 4-position, but furthermore activated the cascade reaction.

43 O O

O TBA H2N O TBA NH2 AB

HO OAc AcO OAc O O A or B O OMe OMe AcO toluene, r.t.1h OTf 80 83

OAc AcO OAc O TBANO /EDA O TfO 2 O OMe AcO OMe toluene, r.t. 4h, OTf 78 83

TfO OAc OAc O TBANO2/EDA O AcO O AcO OMe toluene, r.t. 4h, OMe OTf 84 94 Scheme 32 Carbohydrate cascade epimerization activated by anions.

Under mild acidic work-up conditions, compounds 83 and 94 could subsequently be transformed to the corresponding β-D-altrosides,[119, 120] and β-D-idosides,[121, 122] respectively, in near quantitative yields. This provides a very efficient route to these unusual carbohydrate structures,[123, 124] accessible in high overall yields (up to 80%) in very few steps from the parent unprotected glucosides/galactosides. The 2,3-anhydro compounds are furthermore potentially useful building blocks for alternative carbohydrate substitution patterns, using a variety of suitable reagents.[125-127]

5.2 Enhanced Basicity by Supramolecular Effects (VIII)

5.2.1 Supramolecular Effects of Anions and Solvents

To further explore how the anionic reagents activate the cascade reaction, a range of anions and solvents were further explored. Thus, hydroxide, fluoride, benzoate, chloride, bromide, nitrate, iodide, hydrogensulfate, thiocyanate as well as acetate and nitrite ions were tested together with compound 80 in the aprotic solvents benzene, acetonitrile and DMSO, respectively. 1H-NMR analyses indicated the formation of hydrogen bonds between the anions or solvents and 4-OH of compound 80.

When the tests were performed in d-benzene (Figure 19), the chemical shifts of the 4-OH - proton changed from 1.8 ppm for compound 80 alone to up to 9.0 ppm for NO2 . With AcO- and BzO-, the 4-OH resonances were indiscernible, but the downfield change in chemical shift of the 4-H protons indicates formation of hydrogen bonds. The 4-H - - resonances thus changed from 3.7 ppm (compound 80 alone) up to 4.8 ppm (NO2 , AcO and BzO-).

44

Figure 19 1H-NMR spectra of compound 80, H-bonding forms between hydroxyl group and anions. a: with TBANO2; b: with TBACl; c: with TBABr; d: with TBASCN.

When the tests were performed in d-acetonitrile (Figure 20), it can be seen that the chemical shifts of the 4-OH proton changed from 3.7 ppm for compound 80 alone to up to 3.8 ppm for TBAI, 4.1 ppm for TBASCN, 4.5 ppm for TBABr and 5.5 ppm for TBACl. - - - With NO3 , NO2 and BzO , the 4-OH resonances were indiscernible. Similarly, the downfield change in chemical shift of the 4-H protons also indicates formation of hydrogen bonds. The 4-H resonances thus changed from 4.1 ppm (compound 80 alone) up - - - to 4.2 ppm (NO3 and NO2 ) and 4.4 ppm (BzO ).

Figure 20 1H-NMR spectra of compound 80 with various anions in d-acetonitrile.

Similarly, when the tests were performed in d-DMSO (Figure 21), except for TBAOAc, TBAF and TBAOH, all other reagents induced downfield chemical shifts of the 4-OH proton, which changed from 5.6 ppm (compound 80 alone or with TBAI, TBASCN,

45 TBAHSO4, TBANO3 and TBABr separately) up to 5.7 ppm (with TBACl), 6.3 ppm (with TBANO2) and 6.4 ppm (with TBAOBz).

Figure 21. 1H-NMR spectra of compound 80 with various anions in d-DMSO.

5.2.2 Basicity Controlled Cascade Reaction

Further tests were performed to explore the solvent effects of the reaction and the origin of the anion activation for the cascade reaction. In order to analyze and compare the reactions, compound 80 was used as starting material for the cascade reaction in d- benzene, d-acetonitrile and d-DMSO, respectively. 1H-NMR analysis was used to follow the reaction over time. Three systems were tested in all cases: a) Anions alone were used as base; b) Neutral amines alone were used as base; and c) Combinations of anions and neutral amines were used as base. Using kinetic analysis, the reaction rates could be compared, and the reaction half lives of each reaction determined.

It seems that a tentative conclusion can be obtained from Table 8; the cascade reaction was controlled by the basicity of anions or neutral amines. It can be seen that, in all solvents, the hydroxide, fluoride, and acetate anions showed strong basicity, and the cascade reaction proceeded smoothly and no any migration intermediate was recorded in the entire process (entries 1-3 in Table 8). With chloride, bromide, nitrate, hydrogensulfate, thiocyanate and nitrite, showing low basicity, the cascade reaction could not be triggered (entries 7-13 in Table 8). However, some unexpected results were also observed when using EDA, TEA, acetate or benzoate alone. For example, when using EDA alone, it seems that more polar solvents promote the cascade reaction (entry 6 in Table 8), which is expected because more polar solvents possess stronger hydrogen bonding properties. Thus,

46 the reaction half life was 7.4 hours in benzene, 4.3 hours in acetonitrile and 1.7 hours in DMSO, respectively. However, when using acetate, benzoate or TEA alone, although the more polar acetonitrile promoted the cascade reaction much more than the nonpolar benzene, the reactions were suppressed or did not occur at all in the most polar solvent DMSO (entries 3-5 in Table 8).

Table 8 Cascade reaction half lives in the presence of anions or amines alone in various aprotic solvents.

+ + Entry Reagent pKa(AH ) pKa(AH ) t½ (h) t½ (h) t½ (h) (eq.) (DMSO) (CH3CN) (benzene) (acetonitrile) (DMSO) 1 TBAOH (5) 31.4 < 0.1 < 0.1 < 0.1 2 TBAF (5) 15 < 0.1 < 0.1 < 0.1 3 TBAOAc (5) 12.6 22.3 0.5 0.2 1.0 4 TBAOBz (5) 11.1 20.7 5.9 3.7 -a 5 TEA (15) 9.0 18.5 78.6 39.0 -a 6 EDA (10) - 18.46 7.4 4.3 1.7

b 7 TBANO2 (5) 7.5 - - - 8b TBACl (5) 1.8 - - - 9b TBABr (5) 0.9 - - - 10b TBAI (5) - - -

b 11 TBANO3 (5) - - -

b 12 TBAHSO4 (5) - - - 13b TBASCN (5) - - - a no or very slow reaction. b no or very slow inversion reaction.

When combining these anionic reagents with ten equivalents of ethylenediamine, the cascade reactions were accelerated (entries 1-9 in Table 9), compared to the neutral amine alone (entries 7 in Table 8). Furthermore, since these anions can activate the base-triggered cascade reaction, according to the activation ability, either a very small amount of migration intermediate or no any intermediate at all was accumulated. All the anions showed activation abilities in relation to their hydrogen bonding tendencies, in the order F-, ------OH > OAc > OBz , Cl , NO2 > Br > NO3 > I , SCN . An exceptional case was the hydrogensulfate anion. Although the hydrogensulfate anion possesses low hydrogen bonding tendencies, it showed strong activation ability in the reaction (entry 3 in Table 9). The reason is likely due to the formation of the non-soluble salt RNH3HSO4. Mixing compound 80 with five equivalents of TBAHSO4 in d-acetonitrile or d-DMSO, and adding ten equivalents of EDA, a white precipitate appeared immediately and the cascade reaction was simultaneously completed. Unexpected results were also observed. For example, when using combinations of anions and EDA, the anions showed the highest activation ability in

47 the nonpolar solvent benzene, which seems to indicate that polar solvents suppress the activation ability of anions (Table 9). The anions thus showed lower activation ability in acetonitrile and DMSO than in benzene. Although it is expected that the anions show stronger activation ability in acetonitrile than in DMSO due to the lower polarity of acetonitrile, the anions showed the lowest activation ability in acetonitrile. In particular, some anions which have low hydrogen bonding tendencies, such as nitrate, iodide and thiocyanate, totally lost their activation ability in DMSO (comparing entry 7-9 in Table 9 with entry 6 in Table 8). Although the benzoate anion displays much stronger hydrogen bonding tendency than the nitrite anion, it only showed a comparable activation ability. Similarly, when comparing the nitrite and the chloride anions, it can be seen (entry 4, 5 in Table 9) that the chloride anion showed stronger activation ability in benzene and acetonitrile, whereas it had a weaker activation ability in DMSO.

Table 9 Cascade reaction half lives in the presence of combinations of anions and EDA in aprotic solvents.

Entry Reagent t½ (h) t½ (h) t½ (h) (eq.) (benzene) (acetonitrile) (DMSO) 1 TBAOAc (5) < 0.1 < 0.1 0.4 2 TBAOBz (5) 0.5 1.5 0.7

3 TBAHSO4 (5) - < 0.1 < 0.1

4 TBANO2 (5) 0.5 1.8 0.7 5 TBACl (5) 0.4 1.3 1.2 6 TBABr (5) 1.0 3.1 1.5

a 7 TBANO3 (5) - 3.4 1.7 8a TBAI (5) - 3.9 1.7 9a TBASCN (5) 1.5 3.9 1.7 a The migration intermediate 98 were recorded in these NMR spectra.

5.2.3 Enhanced Basicity by Supramolecular Effects

These results indicate anion-assisted deprotonation through hydrogen bonding as a rationale for the activation effect.[128] Reactions with amine base alone led to accumulation of the migration product intermediate 98, but when the anionic reagent is present from the start, or added after build-up of the intermediate, this is rapidly consumed and the 2,3- anhydro product formed. The kinetic analysis indicates that the cascade reaction is mainly controlled by basicity and the rate of the deprotonation. Fast deprotonation and medium basicity led to a large amount of intermediate and a medium reaction rate, such as EDA. Slow deprotonation and medium basicity lead to trace amounts of intermediate and a slow reaction rate, such as TEA. Strong basicity led to no accumulation of intermediate and a fast reaction rate, such as acetate. Too slow deprotonation and weak basicity could not

48 trigger this cascade reaction, such as imidazole and nitrite. Thus, how the anions can activate the cascade reaction is likely to be due to either increasing the rate of the deprotonation or the enhancing the basicity.

O OAc R R O 1 1 or AHA SHS AcO OMe 2 NR H NR2 OTf CA R3 R3 a c CC

S A or R1R2R3N

H H R1 O OAc A O OAc e 2 NR H A O O R2 R AHS AcO OMe or AcO OMe 3 OTf OTf or R1 N R3 CD CS CN R1 b d R1 2 NR H S 2 NR H S R3 A: Anions; S: Solvent or AHS CB R3 CB Figure 22 Enhanced deprotonation through supramolecular interaction.

It is clear that a hydrogen bond complex CS is formed between compound 80 and the solvents. With amines or anions alone, a new hydrogen bond complex CN is competitively formed between compound 80 and the amines or anions (Figure 22). These two complexes will be further attacked by solvents or amines (or anions), leading to the deprotonation of compound 80 and two homoconjugate complexes CA and CC as well as a heteroconjugate complex CB by approaches a, b, c and d (usually by approaches c and d).[129, 130] For EDA and TEA, the heteroconjugate complex CB is relative stable due to the lower steric congestion compared to the homoconjugate complex CC. Especially for EDA, a more stable complex CB can be formed, due to three possible hydrogen bonds. If the complex CC is enough stable, such as with fluoride, benzoate, or acetate (likely forming a [RCOO….H….OOR]- complex), these anions can activate the deprotonation process. Support for this conclusion was also seen when one equivalent of acetate was used in the reaction, resulting in final 50% conversion. On the other hand, with nitrite and the other anions alone, the hydrogen bond complex is weaker, and the possible [A….H….A]- complex is equally weak. As a result, deprotonation is less efficient. Since the more polar solvents lead to more stable complexes CB, the more polar solvents will promote the cascade reaction. However, the more polar solvents also lead to more stable complexes CS, which result in fewer complexes CN due to competition. Consequently, in the polar solvent DMSO, only very low amounts of the complex CN will be formed, and the deprotonation process will be hampered for pathways c and d. As a result, the reactions were much suppressed or did not occur at all in DMSO.

Dramatic rate enhancements were furthermore recorded with combinations of anions and amines. In analogy with the formation of [RCOO….H….OOR]-, this effect suggests possible …. …. complex formation between the amine and the anion [RNH2 H A], leading to more efficient formation of product 83, and no accumulation of intermediate 98. In the nonpolar

49 solvent benzene, the hydrogen bond complex CN is the major species due to the very weak interaction between compound 80 and benzene (Figure 22). After addition of EDA, complex CN was deprotonated following pathway e to form complex CD. Thus, the cascade reaction rates will mainly depend on the stability of complex CD. In acetonitrile and DMSO, the lower degree of complex CN formed leads to a decrease in the deprotonation rate, and the stable complexes CA and CB formed lead to enhanced basicity. Overall, the reaction was usually more suppressed in acetonitrile than in DMSO. However, when the anion is acetate, due to its very strong hydrogen bonding tendency, the hydrogen bonding effect caused by acetonitrile can be omitted, and, thus, only DMSO suppress the proton transfer process (entry 1 in Table 9). In spite of the fact that the benzoate anion shows much stronger hydrogen bonding tendency than the nitrite anion, the steric effect reduces the deprotonation rate compare to nitrite, and, as a result, benzoate anion shows almost the same activation ability as nitrite in all the solvents (entry 2, 4 in Table 9). Similarly, although the nitrite anion displays stronger hydrogen bonding tendency than chloride anion, the steric effect also reduces its deprotonation rate compared to chloride, and, as a result, the nitrite anion shows weaker activation ability than the chloride anion in benzene and acetonitrile (entry 4, 5 in Table 9). In DMSO, for both the nitrite and the chloride ions, the solvent from the CS complex cannot be displaced. If the pathway b (Figure 22) is the main deprotonation process, there will be no steric effect from nitrite anion, and the hydrogen bonding tendency of the anions will govern the reaction so that nitrite anion shows stronger activation ability than the chloride anion in DMSO. Especially, since the solvent DMSO shows almost the same hydrogen bonding effect as iodide, thiocyanate and nitrate, comparing the lower amount of ions with the large quantity of DMSO, five equivalents of these anions are inefficient in DMSO, and, as a result, they do not show any activation ability.

5.3 Conclusions

In conclusion, a convenient and highly efficient method for multiple carbohydrate epimerization through triggered cascade reactions has been introduced. It was found that reactions that normally involve many steps could be completed in one step in quantitative yields. An intriguing activation effect was furthermore discovered, where combinations of anionic reagents and amines resulted in dramatic rate enhancements. The mechanism by which the anionic reagents activate the cascade reaction was initially explored, suggesting a supramolecular process emanating from the enhanced deprotonation due to plausible amine-anion complexation. The solvents also show important effects on this reaction due to the supramolecular interactions between the solvents and the solutes.

50 6 General Conclusions

The effects of the neighboring group in the Lattrell-Dax epimerization have been explored. During this research, a new carbohydrate/anion host-guest system was discovered and the ambident reactivity of nitrite anion was found to cause a complicated behavior of the reaction. Based on this effect, efficient synthetic routes to β-D-mannosides, β-D-talosides, β-D-altrosides, and β-D-idosides from the corresponding β-D-galactosides and β-D- glucosides, have been designed. The supramolecular effect in these reactions was also explored.

• It has been demonstrated that a neighboring ester group is essential for the reactivity of the Lattrell-Dax nitrite-mediated triflate inversion. Furthermore, a good inversion yield also depended on the relative configuration of the neighboring ester group to the triflate. Only with the ester group in the equatorial position, whatever the configuration of the triflate, did the reaction proceed smoothly, whereas a neighboring axial ester group proved largely inefficient. • A new carbohydrate/anion host-guest system has been discovered in the Lattrell- Dax epimerization. This host-guest system exerts pronounced effects in the nitrite–mediated inversion. The interaction between pyranosides and nitrite thus resulted in dramatic rate control in the inversion reaction, leading to either activation or deactivation effects. • It has been demonstrated that the origin of the importance of the neighboring equatorial ester group in the Lattrell-Dax epimerization is its restriction of the nitrite N-attack, thus resulting in O-attack only. • Based on the efficient multiple carbohydrate esterifications and the Lattrell-Dax carbohydrate epimerization, novel and convenient double parallel- and double serial inversion methods have been developed, by which methyl β-D-mannosides and methyl β-D-talosides have been efficiently synthesized in very high yields at very mild conditions in few steps. The results also indicate that an ester group can, either in parallel or serially, induce its two neighboring groups in the epimerization reaction. • A supramolecularly activated, triggered cascade reaction was developed. This cascade reaction is triggered by a deprotonation process that is activated by anions. It was found that the anions can activate this reaction following their hydrogen bonding tendencies to the hydroxyl group in aprotic solvents.

51

52 Acknowledgements

I would like to express my sincere gratitude to:

My supervisor Olof Ramström for accepting me as a PhD-student, for sharing your vast knowledge in chemistry, for your inspiring guidance and for your patience whenever I want to discuss with you.

Prof. Tobias Rein, Prof. Krister Zetterberg for taking time and efforts of proof-reading my licentiate and doctor thesis. Prof. Torbjörn Norin, Prof. Mingdi Yan for constructive discussion in my licentiate defense.

My co-authors: Prof. Tore Brinck, Dr. Zhichao Pei, Dr. Styrbjörn Byström, Martin, Marcus, Remi and Lingquan “Vince”.

My colleagues/friends in the Ramström group: A special thanks to Zhichao and Rikard for helping me during my early time at the department. Marcus, Pornrapee “Jom” Vongvilai, Remi, Gunnar, Oscar, Morakot, Lingquan and Dr. Luis for constructive criticism and comments on this thesis, for being good friends, for pleasant times in the lab and for spreading a nice atmosphere throughout the group, and all past and present group-members.

Prof. Christina Moberg, Prof. Peter Somfai, Prof. Torbjörn Norin, Prof. Licheng Sun, Prof. Anna-Karin Borg Karlson, Dr. Zoltan Szabo for interesting graduate courses.

Dr. Ulla Jacobsson and Dr. Zoltan Szabo for support with the NMR instruments.

Lena Skowron, Ilona Mozsi, Henry Challis, Ingvor Larsson, Ann Ekqvist for all kinds of help.

Everyone at the chemistry deparment for a friendly and pleasant working atmosphere.

All former and present colleagues at KTH for your friendship.

The Aulin-Erdtman foundation, Knut och Alice Wallenbergs Stiftelse, and Ragnar och Astrid Signeuls for conferences and traveling support.

The Swedish Research Council (VR), the Swedish Foundation for International Cooperation in Research and Higher Education, the Carl Trygger Foundation, for financial support.

My children Emilia, Dongdong and my wife Ailing for their love and support during all these years.

My parents and sister for their endless support and encouragement.

53 Appendix

The following is a description of my contributions to papers I-VIII.

Paper I: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

Paper II: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

Paper III: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript, excluding the computational chemistry section.

Paper IV: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

Paper V: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

Paper VI: I contributed to the formulation of the research problems and performed some of the experimental work.

Paper VII: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

Paper VIII: I contributed to the formulation of the research problems, performed the experimental work and wrote the manuscript.

54 References

[1] T. K. Lindhorst, Essentials of Carbohydrate Chemistry and Biochemistry, Wiley- VCH, Weinheim, 2000. [2] C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357-2364. [3] H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637-674. [4] E. A. Smith, W. D. Thomas, L. L. Kiessling, R. M. Corn, J. Am. Chem. Soc. 2003, 125, 6140-6148. [5] Q. Vicens, E. Westhof, Biopolymers 2003, 70, 42-57. [6] L. Schofield, M. C. Hewitt, K. Evans, M. A. Siomos, P. H. Seeberger, Nature 2002, 418, 785-789. [7] R. A. Dwek, Chem. Rev. 1996, 96, 683-720. [8] R. Kornfeld, S. Kornfeld, Annu. Rev. Biochem. 1985, 54, 631-664. [9] D. Crich, A. Banerjee, Q. J. Yao, J. Am. Chem. Soc. 2004, 126, 14930-14934. [10] D. R. Bundle, J. R. Rich, S. Jacques, H. N. Yu, M. Nitz, C. C. Ling, Angew. Chem. Int. Ed. 2005, 44, 7725-7729. [11] K. Tatsuta, S. Yasuda, Tetrahedron Lett. 1996, 37, 2453-2456. [12] X. Y. Wu, D. R. Bundle, J. Org. Chem. 2005, 70, 7381-7388. [13] Z. J. Witczak, R. Chhabra, H. Chen, X. Q. Xie, Carbohydr. Res. 1997, 301, 167- 175. [14] L. F. Fan, O. Hindsgaul, Org. Lett. 2002, 4, 4503-4506. [15] D. R. Bundle, J. R. Rich, S. Jacques, H. N. Yu, M. Nitz, C. C. Ling, Angew. Chem. Int. Ed. 2005, 44, 7725-7729. [16] J. R. Rich, D. R. Bundle, Org. Lett. 2004, 6, 897-900. [17] C. S. Rye, S. G. Withers, Carbohydr. Res. 2004, 339, 699-703. [18] X. M. Zhu, F. Stolz, R. R. Schmidt, J. Org. Chem. 2004, 69, 7367-7370. [19] A. Liakatos, M. J. Kiefel, M. von Itzstein, Org. Lett. 2003, 5, 4365-4368. [20] S. Knapp, D. S. Myers, J. Org. Chem. 2002, 67, 2995-2999. [21] L. A. Marcaurelle, C. R. Bertozzi, J. Am. Chem. Soc. 2001, 123, 1587-1595. [22] D. Crich, H. M. Li, J. Org. Chem. 2000, 65, 801-805. [23] H. Driguez, ChemBioChem 2001, 2, 311-318. [24] K. K. C. Liu, S. J. Danishefsky, J. Org. Chem. 1994, 59, 1892-1894. [25] C. Unverzagt, Angew. Chem. Int. Ed. 1996, 35, 2350-2353. [26] Z. C. Pei, H. Dong, O. Ramström, J. Org. Chem. 2005, 70, 6952-6955. [27] K. Weinges, S. Haremsa, W. Maurer, Carbohydr. Res. 1987, 164, 453-458. [28] C. W. T. Chang, Y. Hui, B. Elchert, Tetrahedron Lett. 2001, 42, 7019-7023. [29] J. Samuel, M. E. Tanner, Nat. Prod. Rep. 2002, 19, 261-277. [30] R. Albert, K. Dax, R. W. Link, A. E. Stutz, Carbohydr. Res. 1983, 118, C5-C6. [31] R. L. Lattrell, G., Justus Liebigs Ann. Chem. 1974, 901-920. [32] B. M. Trost, H. B. Yang, G. D. Probst, J. Am. Chem. Soc. 2004, 126, 48-49. [33] A. A. H. Abdel-Rahman, S. Jonke, E. S. H. El Ashry, R. R. Schmidt, Angew. Chem. Int. Ed. 2002, 41, 2972-2974. [34] H. M. R. Hoffmann, R. Dunkel, M. Mentzel, H. Reuter, C. B. W. Stark, Chem. Eur. J. 2001, 7, 4771-4789. [35] R. W. Binkley, J. Org. Chem. 1991, 56, 3892-3896.

55 [36] R. W. Binkley, G. S. Goewey, J. C. Johnston, J. Org. Chem. 1984, 49, 992-996. [37] J. F. King, A. D. Allbutt, Can. J. Chem. 1970, 48, 1754. [38] P. Deslongchamps, P. Atlani, D. Frehel, A. Malaval, Can. J. Chem. 1972, 50, 1204. [39] P. Deslongchamps, Tetrahedron 1975, 31, 2463-2490. [40] S. David, S. Hanessian, Tetrahedron 1985, 41, 643-663. [41] K. Takeo, K. Shibata, Carbohydr. Res. 1984, 133, 147-151. [42] F. Iwasaki, T. Maki, O. Onomura, W. Nakashima, Y. Matsumura, J. Org. Chem. 2000, 65, 996-1002. [43] G. Reginato, A. Ricci, S. Roelens, S. Scapecchi, J. Org. Chem. 1990, 55, 5132- 5139. [44] F. Peri, L. Cipolla, F. Nicotra, Tetrahedron Lett. 2000, 41, 8587-8590. [45] V. K. Srivastava, C. Schuerch, Tetrahedron Lett. 1979, 3269-3272. [46] D. J. Jenkins, B. V. L. Potter, Carbohydr. Res. 1994, 265, 145-149. [47] L. Ballell, J. A. F. Joosten, F. A. el Maate, R. M. J. Liskamp, R. J. Pieters, Tetrahedron Lett. 2004, 45, 6685-6687. [48] A. Glen, D. A. Leigh, R. P. Martin, J. P. Smart, A. M. Truscello, Carbohydr. Res. 1993, 248, 365-369. [49] M. J. Martinelli, R. Vaidyanathan, V. Van Khau, Tetrahedron Lett. 2000, 41, 3773-3776. [50] M. J. Martinelli, R. Vaidyanathan, J. M. Pawlak, N. K. Nayyar, U. P. Dhokte, C. W. Doecke, L. M. H. Zollars, E. D. Moher, V. Van Khau, B. Kosmrlj, J. Am. Chem. Soc. 2002, 124, 3578-3585. [51] G. Hodosi, P. Kovac, J. Am. Chem. Soc. 1997, 119, 2335-2336. [52] G. Hodosi, P. Kovac, Carbohydr. Res. 1997, 303, 239-243. [53] G. Hodosi, P. Kovac, Carbohydr. Res. 1998, 308, 63-75. [54] B. Mukhopadhyay, R. A. Field, Carbohydr. Res. 2003, 338, 2149-2152. [55] Z. Y. Zhang, C. H. Wong, Tetrahedron 2002, 58, 6513-6519. [56] M. W. Bredenkamp, H. S. C. Spies, Tetrahedron Lett. 2000, 41, 543-546. [57] S. Roelens, J. Org. Chem. 1996, 61, 5257-5263. [58] M. W. Bredenkamp, H. S. C. Spies, M. J. van der Merwe, Tetrahedron Lett. 2000, 41, 547-550. [59] S. David, A. Thieffry, A. Forchioni, Tetrahedron Lett. 1981, 22, 2647-2650. [60] E. V. Anslyn, D. A. Dougherty, Modern Physical Organic Chemistry, University Science Books, Sausalito, CA, 2006, 659-660. [61] S. Winstein, E. Grunwald, L. L. Ingraham, J. Am. Chem. Soc. 1948, 70, 821-828. [62] H. Dong, Z. C. Pei, O. Ramström, J. Org. Chem. 2006, 71, 3306-3309. [63] H. Dong, Z. C. Pei, M. Angelin, S. Byström, O. Ramström, J. Org. Chem. 2007, 72, 3694-3701. [64] J. R. Rich, A. Szpacenko, M. M. Palcic, D. R. Bundle, Angew. Chem. Int. Ed. 2004, 43, 613-615. [65] M. Kassou, S. Castillon, J. Org. Chem. 1995, 60, 4353-4358. [66] A. ElNemr, T. Tsuchiya, Carbohydr. Res. 1997, 301, 77-87. [67] J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives, Wiley-VCH, Weinheim, 1995. [68] J. M. Lehn, Science 2002, 295, 2400-2403.

56 [69] L. J. Prins, D. N. Reinhoudt, P. Timmerman, Angew. Chem. Int. Ed. 2001, 40, 2383-2426. [70] M. Mazik, H. Cavga, P. G. Jones, J. Am. Chem. Soc. 2005, 127, 9045-9052. [71] E. Klein, M. P. Crump, A. P. Davis, Angew. Chem. Int. Ed. 2005, 44, 298-302. [72] G. Terraneo, D. Potenza, A. Canales, J. Jimenez-Barbero, K. K. Baldridge, A. Bernardi, J. Am. Chem. Soc. 2007, 129, 2890-2900. [73] J. Screen, E. C. Stanca-Kaposta, D. P. Gamblin, B. Liu, N. A. Macleod, L. C. Snoek, B. G. Davis, J. P. Simons, Angew. Chem. Int. Ed. 2007, 46, 3644-3648. [74] E. C. Stanca-Kaposta, D. P. Gamblin, J. Screen, B. Liu, L. C. Snoek, B. G. Davis, J. P. Simons, Phys. Chem. Chem. Phys. 2007, 9, 4444-4451. [75] N. Kornblum, R. A. Smiley, R. K. Blackwood, D. C. Iffland, J. Am. Chem. Soc. 1955, 77, 6269-6280. [76] N. Kornblum, L. Fishbein, R. A. Smiley, J. Am. Chem. Soc. 1955, 77, 6261-6266. [77] N. Kornblum, Org. React. 1962, 12, 101-156. [78] R. G. Pearson, J. Songstad, J. Am. Chem. Soc. 1967, 89, 1827-1836. [79] G. Klopman, J. Am. Chem. Soc. 1968, 90, 223-234. [80] X. F. Huang, L. J. Huang, H. S. Wang, X. S. Ye, Angew. Chem. Int. Ed. 2004, 43, 5221-5224. [81] I. Fleming, Frontier Orbitals and Organic Chemical Reactions, 1976, Wiley, Chichester, 41-42. [82] A. Streitwieser, C. H. Heathcock, E. M. Kosower, Introduction to Organic Chemistry 1992, Macmillan, New York, 779. [83] R. F. Hudson, Angew. Chem. Int. Ed. 1973, 12, 36-57. [84] A. A. Tishkov, U. Schmidhammer, S. Roth, E. Riedle, H. Mayr, Angew. Chem. Int. Ed. 2005, 44, 4623-4626. [85] S. Knapp, P. J. Kukkola, S. Sharma, T. G. M. Dhar, A. B. J. Naughton, J. Org. Chem. 1990, 55, 5700-5710. [86] F. Buckell, J. D. Hartry, U. Rajalingam, B. M. Bennett, R. A. Whitney, G. R. J. Thatcher, J. Chem. Soc. Perkin Trans. 2 1994, 401-403. [87] P. Grice, S. V. Ley, J. Pietruszka, H. M. I. Osborn, H. W. M. Priepke, S. L. Warriner, Chem. Eur. J. 1997, 3, 431-440. [88] I. Matsuo, Y. Nakahara, Y. Ito, T. Nukada, Y. Nakahara, T. Ogawa, Bioorg. Med. Chem. 1995, 3, 1455-1463. [89] R. K. Jain, K. L. Matta, Carbohydr. Res. 1996, 282, 101-111. [90] J. Kerekgyarto, J. G. M. Vanderven, J. P. Kamerling, A. Liptak, J. F. G. Vliegenthart, Carbohydr. Res. 1993, 238, 135-145. [91] S. J. Danishefsky, S. Hu, P. F. Cirillo, M. Eckhardt, P. H. Seeberger, Chem. Eur. J. 1997, 3, 1617-1628. [92] F. W. Lichtenthaler, T. Schneideradams, J. Org. Chem. 1994, 59, 6728-6734. [93] L. Cipolla, L. Lay, F. Nicotra, J. Org. Chem. 1997, 62, 6678-6681. [94] K. Sato, A. Yoshitomo, Chem. Lett. 1995, 39-40. [95] I. Matsuo, M. Isomura, R. Walton, K. Ajisaka, Tetrahedron Lett. 1996, 37, 8795- 8798. [96] A. Furstner, I. Konetzki, Tetrahedron Lett. 1998, 39, 5721-5724. [97] S. Weiler, R. R. Schmidt, Tetrahedron Lett. 1998, 39, 2299-2302. [98] C. Unverzagt, Angew. Chem. Int. Ed. 1994, 33, 1102-1104.

57 [99] J. Seifert, C. Unverzagt, Tetrahedron Lett. 1996, 37, 6527-6530. [100] G. Stork, J. J. LaClair, J. Am. Chem. Soc. 1996, 118, 247-248. [101] A. Dan, Y. Ito, T. Ogawa, Tetrahedron Lett. 1995, 36, 7487-7490. [102] A. Dan, Y. Ito, T. Ogawa, J. Org. Chem. 1995, 60, 4680-4681. [103] A. Dan, M. Lergenmuller, M. Amano, Y. Nakahara, T. Ogawa, Y. Ito, Chem. Eur. J. 1998, 4, 2182-2190. [104] Y. Ito, T. Ogawa, J. Am. Chem. Soc. 1997, 119, 5562-5566. [105] M. Lergenmuller, T. Nukada, K. Kuramochi, A. Dan, T. Ogawa, Y. Ito, Eur. J. Org. Chem. 1999, 1367-1376. [106] F. Barresi, O. Hindsgaul, J. Am. Chem. Soc. 1991, 113, 9376-9377. [107] T. Ziegler, G. Lemanski, Angew. Chem. Int. Ed. 1998, 37, 3129-3132. [108] D. Crich, S. X. Sun, J. Org. Chem. 1996, 61, 4506-4507. [109] D. Crich, S. X. Sun, J. Am. Chem. Soc. 1997, 119, 11217-11223. [110] D. Crich, S. X. Sun, J. Am. Chem. Soc. 1998, 120, 435-436. [111] K. C. Nicolaou, F. L. vanDelft, S. R. Conley, H. J. Mitchell, Z. Jin, R. M. Rodriguez, J. Am. Chem. Soc. 1997, 119, 9057-9058. [112] A. A. H. Abdel-Rahman, S. Jonke, E. S. H. El Ashry, R. R. Schmidt, Angew. Chem. Int. Ed. 2004, 43, 4389-4389. [113] C. H. Wong, R. L. Halcomb, Y. Ichikawa, T. Kajimoto, Angew. Chem. Int. Ed. 1995, 34, 412-432. [114] S. Singh, M. Scigelova, D. H. G. Crout, Chem. Commun. 1996, 993-994. [115] G. M. Watt, L. Revers, M. C. Webberley, I. B. H. Wilson, S. L. Flitsch, Angew. Chem. Int. Ed. 1997, 36, 2354-2356. [116] V. Vicente, J. Martin, J. Jimenez-Barbero, J. L. Chiara, C. Vicent, Chem. Eur. J. 2004, 10, 4240-4251. [117] I. A. Ivanova, A. V. Nikolaev, J. Chem. Soc. Perkin Trans. 1 1998, 3093-3099. [118] H. Dong, Z. C. Pei, O. Ramström, J. Org. Chem. 2006, 71, 3306-3309. [119] K. Fujita, K. Ohta, Y. Ikegami, H. Shimada, T. Tahara, Y. Nogami, T. Koga, K. Saito, T. Nakajima, Tetrahedron Lett. 1994, 35, 9577-9580. [120] N. K. Richtmyer, Methods Carbohydr. Chem. 1962, 1, 107-113. [121] A. S. Perlin, B. Casu, G. R. Sanderson, J. Tse, Carbohydr. Res. 1972, 21, 123-132. [122] E. Sorkin, R. Reichstein, Helv. Chim. Acta 1945, 28, 1-17. [123] R. W. S. Weber, A. Meffert, H. Anke, O. Sterner, Mycol. Res. 2005, 109, 619-626. [124] S. S. Afiyatullov, T. A. Kuznetsova, V. V. Isakov, M. V. Pivkin, N. G. Prokof'eva, G. B. Elyakov, J. Nat. Prod. 2000, 63, 848-850. [125] V. L. Maxwell, E. L. Evinson, D. P. G. Emmerson, P. R. Jenkins, Org. Biomol. Chem. 2006, 4, 2724-2732. [126] Y. Wang, Q. Li, S. H. Cheng, Y. F. Wu, D. J. Guo, Q. H. Fan, X. F. Wang, L. H. Zhang, X. S. Ye, Org. Lett. 2005, 7, 5577-5579. [127] A. Procopio, R. Dalpozzo, A. De Nino, L. Maiuolo, M. Nardi, B. Russo, Adv. Synth. Catal. 2005, 347, 1447-1450. [128] V. Amendola, D. Esteban-Gomez, L. Fabbrizzi, M. Licchelli, Acc. Chem. Res. 2006, 39, 343-353. [129] V. Amendola, M. Boiocchi, L. Fabbrizzi, A. Palchetti, Chem. Eur. J. 2005, 11, 5648-5660.

58 [130] M. Boiocchi, L. Del Boca, D. Esteban-Gomez, L. Fabbrizzi, M. Licchelli, E. Monzani, Chem. Eur. J. 2005, 11, 3097-3104.

59