Bispyridylamides as Ligands in Asymmetric Catalysis

Oscar Belda de Lama

Doctoral Thesis

Stockholm 2004

Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan iStockholm framlägges till offentlig granskning för avläggande av filosofiedoktorsexamen i kemi med inriktning mot organisk kemi, fredagen den28:e maj, kl 10.00 i Kollegiesalen, KTH, Valhallavägen 79, Stockholm. Avhandligen försvaras på engelska. Opponent är Professor Dr. AndreasPfaltz, Universität Basel.

ISBN-91-7283-724-1 ISRN KTH/IOK/FR--04/88--SE ISSN 1100-7974 TRITA-IOK Forskningsrapport 2004:88 © Oscar Belda de Lama 2004

Belda de Lama, O., 2004 ”Bispyridylamides as ligands in asymmetric catalysis”. Kungliga Tekniska Högskolan, Stockholm.

Abstract This thesis deals with the preparation and use of chiral bispyridylamides as ligands in metal-catalyzed asymmetric reactions.

The compounds were prepared by amide formation using different coupling reagents. Bispyridylamides having substituents in the 4- or 6- positions of the pyridine rings were prepared by functional group interconversion of the 4- or 6-halopyridine derivatives. These synthetic approaches proved to be useful for various types of chiral backbones. Pseudo C2-symmetric bispyridylamides were also synthesized by means of stepwise amide formation.

The compounds were used as ligands in the microwave-accelerated Mo-catalyzed asymmetric allylic alkylation reaction. Ligands having π-donating substituents in the 4-positions of the pyridine rings gave rise to products with higher branched to linear ratio. The catalytic reaction, which proved to be rather general for allylic carbonates with an aromatic substituent, was used as the key step in the preparation of (R)-baclofen. The Mo-bispyridylamide catalyst precursor was studied by NMR spectroscopy.

Bispyridylamide complexes of metal alkoxides were also evaluated in the asymmetric addition of cyanide to aldehydes and the metal complexes involved were studied by NMR spectroscopy and X-ray crystallography.

Chiral diamines were used as additives to study the ring opening of cyclohexene oxide with azide, catalyzed by Zr(IV)-bispyridylamide complexes.

Various bispyridylamides were attached to solid supports of organic or inorganic nature. The solid-supported ligands were used in Mo-catalyzed asymmetric allylic alkylation reactions and in the asymmetric addition of cyanide to benzaldehyde. Keywords: asymmetric catalysis, chiral ligand, pyridine, amide, allylic alkylation, enantioselective, cyanation, ring-opening, chiral Lewis acid.

A mis padres que me dieron todo

Table of contents

Abstract. List of publications. 1. Introduction 1

1.1 Aim of the thesis 2 References 3

2. Bispyridylamides 5 2.1 Introduction 5 2.2 Synthesis 6 2.2.1 Amide formation 7 2.2.2 Functional group interconversion 10 2.2.3 Solid-supported pyridylamides 12 2.3 Coordination chemistry 15 2.4 Use in catalysis 19 References 20

3. Application of bispyridylamides as ligands in Mo- catalyzed asymmetric allylic alkylations 23

3.1 Introduction 23 3.2 Electronic and steric effects of the ligand 25 3.3 Substrates and nucleophiles 29 3.4 Heterogeneous catalysts 32 3.5 Application of the reaction in enantioselective synthesis 33 References 36

4. Application of bispyridylamides as ligands in asymmetric Lewis acid catalyzed processes 41

4.1 Introduction 41 4.2 Addition of cyanide to carbonyl groups 42

4.2.1 Use of a heterogeneous bispyridylamide as ligand 49 4.3 Ring opening of cyclohexene oxide 50 References 53

5. Concluding remarks 55 Appendix 57 Acknowledgements 61

List of publications This thesis is based on the following papers, referred to in the text by their Roman numerals:

I. “Highly Stereo- and Regioselective Allylations Catalyzed by Mo-Pyridylamide Complexes: Electronic and Steric Effects of the Ligand”. Belda, O.; Kaiser, N.-F.; Bremberg, U.; Larhed, M.; Hallberg, A.; Moberg, C. J. Org. Chem. 2000, 65, 5870.

II. “Substituted Pyridylamide Ligands in Microwave-Accelerated Mo(0)-Catalysed Allylic Alkylations”. Belda, O.; Moberg, C. Synthesis 2002, 1601.

III. “Recoverable Resin-Supported Pyridylamide Ligand for Microwave-Accelerated Molybdenum-Catalyzed Asymmetric Allylic Alkylations: Enantioselective Synthesis of Baclofen”. Belda, O.; Lundgren, S.; Moberg, C. Org. Lett. 2003, 5, 2275.

IV. “Chiral Bispyridylamide Metal Complexes as Catalysts for the Enantioselective Addition of TMSCN to Aldehydes”. Belda, O.; Duquesne, S.; Fischer, A.; Moberg, C. (Manuscript).

1

1 Introduction

Since Pasteur’s discovery of chirality in organic matter, tremendous advances have been made in this field, even though fundamental questions such as the origin of chirality have not been answered yet.

Nevertheless, a clear trend towards single-enantiomer new chemical entities (NCEs) appeared in the late 1980s. The world market for chiral fine chemicals sold as single-enantiomers was $6.63 billion in the year 2000 with an expected growth of 13% annually to $16 billion in 2007.1 Single-enantiomer compounds made for pharmaceutical formulations represented 36% of the total worldwide sales of pharmaceutical products in 2001 and here there was also a growing tendency.2 The development of methods for producing enantiomerically pure compounds, e.g. stereoselective synthesis, became necessary and a lot of re-search was devoted to this subject.

In general, stereoselective synthesis falls into three broad categories:3

– Diastereoselective syntheses. – Enantioselective syntheses. – Double stereodifferentiating reactions.

In industry, enantiomerically pure compounds are traditionally made from

enantiomerically pure naturally occurring compounds or by resolution of race-mates. Enantioselective reaction with a chiral nonracemic metal catalyst, i.e. asymmetric catalysis, keeps attracting the attention of industry since the pio-neering work on asymmetric hydrogenation reactions.4 The field experiences a constant growth ever since and today synthetic chemists find a number of trans-formations using chiral nonracemic metal catalysts within their box of every-

2

day enantioselective synthesis tools. The use of substoichiometric amounts of the chiral catalyst is one of the major advantages of asymmetric catalysis, in some cases making these reactions atom economical.5 However, chiral cata-lysts are often substrate-specific, which leads to a lack of generality. Ideally, a chiral synthetic catalyst should be enantioselective over a wide range of reac-tions and, indeed, some examples of these privileged catalysts are known.6

Chiral nonracemic metal catalysts are usually prepared through complexation of enantiomerically pure organic ligands with the appropriate metal salts. In the design of a new catalyst, the steric and electronic properties of the ligand as well as the nature of the metal are important. The ligand design plays a signifi-cant role in the development of new catalysts, and when making new success-ful ligands, one must take some structural trends into account. For instance, rigid structures with multiple oxygen-, nitrogen-, or phosphorus-containing functional groups that allow them to bind strongly to the metal, and a two-fold axis of symmetry that reduces the number of transition state geometries avail-able in the reaction, are common features in many highly selective ligands. However, there are no general rules and a structure that possesses these fea-tures does not necessarily display high selectivity. In the same manner, not all highly selective ligands have these properties, so that today the development of new catalysts is still to a large extent based on trial and error.

For a given ligand, it is also interesting whether both enantiomers are easily accessible or not and whether their electronic and steric properties can be changed in a simple manner, allowing a fine tuning of the reaction and a degree of substrate and reaction generality.

1.1 Aim of the thesis This thesis is a continuation of earlier work in our group, dealing with the use of bispyridylamides in asymmetric catalysis.7 The aim of this thesis has been to develop new bispyridylamides as ligands for asymmetric catalysis. We have studied the preparation of picolinic acid based bispyridylamides with substitu-ents in either one or both pyridine rings and examined the versatility of these compounds as ligands in asymmetric catalysis by using them in Mo-catalyzed asymmetric allylic alkylation reactions and asymmetric Lewis acid catalyzed processes. We have also studied the coordination chemistry of the metal com-plexes involved in the catalytic reactions, the preparation and utilization of heterogeneous pyridylamide ligands and the application of the catalytic reac-

3

tions studied in the enantioselective syntheses of chiral products of industrial interest.

References 1. Stinson, S. C. Chiral Chemistry. Chem. Eng. News 2001, 79, 45-56. 2. Rouhi, A. M. Chiral Roundup. Chem. Eng. News 2002, 80, 43-50. 3. Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic Compounds. Wiley: New York, 1994; pp 839-843. 4. Knowles, W. S. Asymmetric hydrogenation. Acc. Chem. Res. 1983, 16, 106-112. 5. Trost, B. M. The atom economy: a search for synthetic efficiency. Science 1991, 254, 1471-1477. 6. Tehshik, P. Y.; Jacobsen, E. N. Privileged Chiral Catalysts. Science 2003, 299, 1691-1693. 7. Adolfsson, H. Ligand Design for Selective Catalysis. Doctoral Thesis. KTH: Stockholm, 1995.

4

5

2 Bispyridylamides

2.1 Introduction Compounds with the general structure shown in Figure 2.1 will be given the general name of bispyridylamides in this thesis.

Figure 2.1. General structure of a picolinic acid based bispyridylamide.

Bispyridylamide ligands were first prepared by Ojima in 1967.1 Since then, a

number of achiral and chiral bispyridylamides were prepared and their coordi-nation chemistry was extensively studied.2 However, only a few studies of their use in catalysis, especially in asymmetric catalysis, have been described in the literature.

This chapter deals with the synthesis of bispyridylamides, the chemistry of their coordination with transition metals and selected applications as ligands in catalysis.

6

2.2 Synthesis A bispyridylamide is usually prepared through the formation of an amide bond from an appropriate diamine and an activated picolinic acid (Figure 2.2A). A number of diamines, including chiral nonracemic ones, are commercially avail-able. The most commonly used achiral diamines are 1,2-diaminobenzene, 1,2-diaminoethane and 1,3-diaminopropane. As for chiral diamines, C2-symmetric R*,R*-1,2-diaminocyclohexane, R*,R*-1,2-diphenyl-1,2-diaminoethane and R*-2,2’-diamino-1,1’-binaphthyl are the most employed ones.

Few substituted picolinic acids are commercially available. Therefore, they must be synthesized from different pyridine derivatives in multistep proce-dures. In some cases, it can be tedious to isolate the neutral substituted pi-colinic acids due to their amphoteric nature.

In a first report, three bispyridylamides were prepared by direct reactions at high temperatures between methyl picolinate and three diamines.1 Later, Biniecki and Herold reported the synthesis of a bispyridylamide derived from meso-1,2-diphenyl-1,2-diaminoethane and ethyl picoloyl carbonate.3

Triphenyl phosphite has been the most commonly employed reagent among the general coupling methods available for the reaction between picolinic acid or its derivatives and a diamine. The method was originally used in 1978, by Vagg and coworkers, to prepare ligands derived from picolinic acid and 1,2-diaminoethane, 1,3-diaminopropane, rac-1,2-diaminocyclohexane, 1,2-diami-nobenzene and piperazine in varying yields from 47 to 96%.4 One of the draw-backs of this method was the formation of coproducts and, thus, the necessity of sometimes tedious separations.

The transformation of the carboxylic acid into the acid chloride by treatment with either thionyl chloride or oxalyl chloride is usually not a viable method due to the formation of byproducts and low reproducibility. Nevertheless, a few examples exist.5

For bispyridylamides with substituents in the pyridine rings, there is an alter-native retrosynthetic analysis (Figure 2.2B), in which a functional group inter-conversion (FGI) of a suitable bispyridylamide precursor is involved.

7

Figure 2.2. Retrosynthetic analyses of bispyridylamides. A: amide formation. B: functional group interconversion.

2.2.1 Amide formationI,II,III,IV We prepared ligand 1a (Figure 2.3) from picolinic acid and (1R,2R)-1,2-diami-nocyclohexane, using different coupling methods (Table 2.1), to examine their advantages and disadvantages.

Table 2.1. Preparation of 1a by Means of Different Coupling Reagents.

entry coupling reagent % yield 1a P(OPh)3 47 2 P(OPh)3 62 3 oxalyl chloride 32 4 Mukaiyama’s reagent 62

5b CDI 87 a Ref. 4a. b Ref. 6.

The coupling of the reagents with triphenyl phosphite required a separation of the coproducts (vide supra). In the original report by Vagg (entry 1), that sepa-ration was achieved by crystallization from chloroform to give the desired product, rac-1a, as brownish crystals, in 47% yield. However, we could in-

8

crease the yield to 62% of a white crystalline product, if the purification was instead done by column chromatography followed by crystallization from EtOAc (entry 2).

When substituted picolinic acids were coupled using this method, the yields were much lower, since the column chromatography was tedious and the crys-tallization was not easy.

Formation of the acid chloride with oxalyl chloride was an attractive method, since the coproducts were volatile and, therefore, easily removed. In practice we experienced that the reaction was sluggish and the very dark-colored product was obtained in low and variable yields (entry 3). Mukaiyama’s re-agent (2-chloro-1-methylpyridinium iodide) also afforded 1a in 62% yield, but the separation of the coproducts was somewhat easier when substituted pi-colinic acids were used as reactants.

Figure 2.3. Some bispyridylamides prepared by amide formation from (1R,2R)-1,2-diamino-cyclohexane and various substituted picolinic acids.

During the course of this investigation Conlon and Yasuda reported the syn-

thesis of 1a on the kilo-scale, using Staab’s reagent, 1,1’-carbonyldiimidazole (CDI), in 87% yield.6 The main advantage of this method was that the copro-duct imidazol was easily removed.

9

We prepared bispyridylamides derived from substituted picolinic acids (Figure 2.3) by using a combination of the methods described above (Table 2.2). As opposite to 1a, all of these derivatives were difficult to crystallize. Column chromatography was the purification method used in most cases.

Table 2.2. Preparation of Bispyridylamides 1b-e, 2a-b and 3a.

bispyridylamide coupling reagent % yield 1b oxalyl chloride 23 1c Mukaiyama’s reagent 45 1d Mukaiyama’s reagent 62 1e Mukaiyama’s reagent 10 1e CDI 54 2a Mukaiyama’s reagent 53 2b CDI 54 3a oxalyl chloride 31

The reaction with a second equivalent of the carboxylic acid derivative oc-

curred more slowly than the first amide bond formation. This opened up the possibility to form unsymmetrical pyridylamides containing pyridine rings with different substituents. With careful control of the reaction conditions, espe-cially the concentration and the rate and order of addition, we synthesized compounds 4a and 4b (Figure 2.3) in 46 and 26% yield, respectively, using CDI as the coupling reagent.

We also prepared bispyridylamides based on (1R,2R)-1,2-diphenyl-1,2-dia-minoethane (Figure 2.4) in similar manners in variable yields (Table 2.3).

Figure 2.4. Bispyridylamides prepared by amide formation from (1R,2R)-1,2-diphenyl-1,2-diaminoethane and different substituted picolinic acids.

10

Table 2.3. Preparation of Bispyridylamides 5a-c and 6a.

bispyridylamide coupling reagent % yield 5aa P(OPh)3 73 5a Mukaiyama’s reagent 58 5a CDI 84 5b CDI 34 5c oxalyl chloride 39 6a CDI 35

a Ref. 4b.

2.2.2 Functional group interconversionII,III One of the aims of this study was to find a simple way of preparing bis-pyridylamides with substituents on the pyridine rings. The linear syntheses described above resulted in low overall yields and, in some cases, in high num-bers of steps. Therefore, we decided to try whether functional group intercon-version could be a more direct approach towards these kinds of compounds.

Scheme 2.1. Preparation of substituted bispyridylamide derivatives of 1a by nucleophilic

aromatic substitution reactions and Ni-catalyzed cyanation.

11

We identified bispyridylamides 1e and 2b (Figure 2.3), with halo-substituted

pyridines, as suitable building blocks for a number of derivatives, since it was well-known that α- and γ-halosubstituted pyridines undergo nucleophilic aro-matic substitution with several types of nucleophiles.7 Moreover, the syntheses of 1e and 2b were quite straightforward, since 4-chloropicolinic acid was pre-pared in multigram quantities in a one pot two steps procedure from pyridine, and 6-bromopicolinic acid was commercially available or otherwise prepared in multigram quantities by oxidation of 6-bromopicoline.8 Nucleophilic aro-matic substitution reactions of pyridine derivatives were usually performed at high temperatures and with excess of the nucleophile.9 We decided to run the reactions of 1e and 2b in a microwave cavity10 at 150-165 °C for 15-20 min in the presence of an excess of the nucleophile, which also acted as solvent (Scheme 2.1).

Scheme 2.2. Preparation of pseudo C2-symmetric pyridylamides by nucleophilic aromatic substitution reactions.

After complete reaction, work-up and evaporation of the excess of nucleo-

phile gave the products desired in almost quantitative yields and in high purity. In this way, bispyridylamides with pyridine rings carrying –OMe and –NC4H8

substituents could easily be prepared. In order to introduce electron-withdraw-ing substituents, Ni-catalyzed cyanation of 1e was performed, resulting in 42% yield of 1g.

The same approach was used for preparing pyridylamides with only one sub-stituted pyridine ring, starting from monochloro-substituted 4b (Figure 2.3, Scheme 2.2).

These reactions could be performed in a similar manner for preparing bis-pyridylamides based on (1R,2R)-1,2-diphenyl-1,2-diaminoethane (Figure 2.5).

12

Figure 2.5. Ligands based on (1R,2R)-1,2-diphenyl-1,2-diaminoethane prepared by nucleo-philic aromatic substitution of 5b and 6a (see Fig. 2.4).

2.2.3 Solid-supported pyridylamidesIII,IV Heterogeneous asymmetric catalysis has been a field of interest for many re-search laboratories due to the increased demand for robust, reliable and eco-nomical synthetic methods of preparing chiral compounds. A number of solid-supported chiral ligands have been prepared for nearly every class of asymmetric catalytic organic transformation including oxidation, reduction and C-C bond formation among others.11 The main advantages of solid-supported catalysts are the easy separation of the catalyst from the reaction mixture and the possibility to reuse the catalyst separated. In some cases, the solid-supported catalyst has reduced the leaching of metals into the product, which has been of especial importance for complying with the severe regulations on the preparation of pharmaceuticals.

Different strategies have been used for the immobilization of a catalyst to an organic or inorganic solid support. The attachment to the solid support has generally been made by chemical bond formation between the catalyst and the solid support, by adsorption or ion-pair formation, by entrapment or by micro-encapsulation.

Crosslinked polystyrene polymers or co-polymers are among the most com-monly used solid organic supports. The swelling of the polymer during the catalytic reaction is very important, because if the polymer does not swell in the solvent, the catalysts will be less available for the reactants. A number of resins exist to overcome this problem. Among them, Tenta Gel and Argo Gel are the most common ones. These resins are widely used and their swelling properties are outstanding, allowing a “homogeneous” behavior of the hetero-geneous catalyst. These resins also allow an easier analysis of the compounds in the swollen state, for example by 13C NMR spectroscopy.

13

Inorganic supports are usually thermally, chemically as well as mechanically more stable than organic ones. Silica, zeolites, alumina, zirconia, ZnO and clay are frequently used as inorganic supports for organic catalysts.

Bispyridylamides could be attached to solid supports via either the diamine or the pyridine part of the molecule. The former strategy, which in principle could allow to maintain the C2-symmetry of the ligand, was used in the synthesis of a bispyridylamide derivative of 1,2-diaminobenzene.12 In that way, picolinic acid was coupled to a polymer-supported 1,2-diaminobenzene derivative and the resulting product was used in the Mn-catalyzed oxidation of alkanes and al-kenes.

We were interested in the preparation of chiral solid-supported bispyridyla-mides for their application as ligands in asymmetric catalysis. Some syntheses of enantiomerically pure diamines with suitable functional groups that allowed their attachment to solid supports have been described.13,14 However, the attachment via the pyridine ring allowed for a higher flexibility, and the func-tionalization could also be employed for tuning of the electronic properties of the ligand. We envisioned that 4e (Scheme 2.2) and 6b (Figure 2.5), each having a free primary amino group in one of the pyridine rings, were suitable for the reaction with a solid support functionalized with an electrophilic group. We attached compound 4e to an organic as well as an inorganic solid support in order to compare their stability, activity and selectivity in catalytic reactions. A carboxylic acid functionalized Tenta Gel resin (TG HL-COOH, particle size 110 µm, loading 0.40 mmol/g) was chosen as the organic support. In this case, we observed that the amide bond formation occurred smoothly using DCC as the coupling reagent, giving 84% yield of 4f according to elemental analysis (Scheme 2.3).

14

Scheme 2.3. Attachment of 4e and 6b to a Tenta Gel HL-COOH by amide coupling.

As for the inorganic solid support, propionyl chloride functionalized silica gel

(Silicycle-60Å, particle size 40-63 µm, loading 0.93 mmol/g) was used at first (Scheme 2.4). However, the yield of the coupling was only 22% and no in-crease was observed by changing the reaction conditions. Using a carboxylic acid functionalized Kromasil-100Å silica gel (EKA, particle size 10 µm, loading 1.11 mmol/g) and performing the amide bond formation with DCC af-forded a maximum of 56% yield of attached 4e (Scheme 2.4). The reason for the lower yields observed in these cases might be the fact that some reactive sites in the solid support were not available due to its heterogeneous nature, i.e. the diffusion of the reagents through the pores was very slow.

15

Scheme 2.4. Immobilization of 4e to two silica-based supports.

We could also attach 6b to a Tenta Gel HL-COOH with DCC in 74% yield

according to elemental analysis (Scheme 2.3).

2.3 Coordination chemistryIV The coordination chemistry of bispyridylamides has been extensively studied since the first report by Ojima in 1967.1 As a result, numerous metal complexes with different bispyridylamides have been prepared and characterized.2 Coor-dination through the pyridine nitrogen atoms, the carbonyl oxygen atoms and the amide nitrogen atoms is possible. Generally, bispyridylamides bind to the metal through the pyridine nitrogen atoms and either by the deprotonated amide nitrogens, the oxygen atoms of the neutral amide groups or a combina-tion of both, but some exceptional cases exist. The deprotonated dianionic ligands are strong σ-donors, able to stabilize metal ions in high oxidation states. Square planar complexes of Ni(II) and Pd(II) with deprotonated bis-pyridylamides acting as N4 ligands are known.1,15 Cu(II) complexes favor square pyramidal geometries with the deprotonated ligand coordinating in a planar N4 fashion and a molecule of water in the fifth position.16 The analogous Ni(II) complexes do not favor coordination of a fifth ligand. With sterically hindered ligands, e.g. having methyl groups in the 6 position of the pyridine rings, tetragonal distortion was observed.17 Complexes with the neutral ligand

16

show a wide array of geometries, such as octahedral with Cu(II)15 and poly-meric quasi-linear chains with trans bis-bidentate binding coordination to Ag(I) involving the pyridine nitrogen and the carbonyl oxygen.18

Enantiomerically pure complexes of the deprotonated ligand 5a (Figure 2.4) have been prepared with Cu(II), Ni(II) and Pd(II).19 The geometry of the Cu(II) complex in the crystal structure has been shown to be square pyramidal with the deprotonated ligand coordinating in a planar N4 fashion and an amide oxy-gen atom of a neighbouring molecule in the apical position. The phenyl groups of the ligand display a pseudo-axial conformation which is interesting for applications in asymmetric catalysis.

We have studied the coordination chemistry of 5a with diverse early transi-tion metals in high oxidation states, in relation to Lewis acid catalyzed asym-metric processes. We prepared a complex by mixing 5a and Zr(OtBu)4. The complex was crystallized from benzene and its structure was determined by X-ray analysis (Figure 2.6, Table 2.4). The metal had a distorted octahedral ge-ometry with the deprotonated ligand coordinated to Zr in a planar N4 fashion and two tert-butoxide groups in the apical positions. In this case, the phenyl groups of the ligand were also in a pseudo-axial conformation. Attempts to form similar complexes with Ti(OiPr)4, Sc(OiPr)3 and Yb(OiPr)3 failed. In the case of Yb(OiPr)3 a gel-like solution was obtained and broad signals were ob-served in the 1H-NMR spectrum, an indication of aggregate formation.20 Addi-tion of an excess of trimethylsilyl cyanide to a 1:1 mixture of 5a and Ti(OiPr)4 afforded a mixture of complexes which could not be separated by crystalliza-tion.

Figure 2.6. X-ray structure of Zr complex with ligand 5a. Ellipsoids are drawn at the 50% probability level.

17

Table 2.4. Selected Bond Lenght (pm) and Bond Angles (deg) for Zr Complex with Ligand 5a.

bond length bonds angle Zr1-O3 192.4(8) O3-Zr1-O4 124.1(3) Zr1-O4 192.2(7) O4-Zr1-N3 105.7(4) Zr1-N1 234(1) O3-Zr1-N3 118.8(3) Zr1-N2 220.7(8) O4-Zr1-N2 120.6(3) Zr1-N3 220.6(9) O3-Zr1-N2 70.6(3) Zr1-N4 236.5(8) O3-Zr1-N1 84.3(3)

N3-Zr1-N1 138.1(3) N2-Zr1-N1 69.5(3) O4-Zr1-N4 82.9(3) O3-Zr1-N4 82.9(3) N3-Zr1-N4 69.2(3) N2-Zr1-N4 137.7(3) N1-Zr1-N4 152.4(3)

In relation to applications of bispyridylamide 7 as ligand for Mo-catalyzed

asymmetric allylic alkylations, an allyl complex derived from 8 was prepared by Krska et al. and the structure of the complex was determined by X-ray studies (Scheme 2.5).21 The ligand was coordinated to the metal with the nitro-gen atom of the pyridine ring, the adjacent deprotonated amide nitrogen atom and the carbonyl oxygen atom of the other amide group. The allyl moiety was bound to Mo in a η3 fashion and was placed trans to the deprotonated amide group.

Scheme 2.5. Preparation of a Mo-allyl complex with ligand 7 by oxidative addition.

In contrast to this, a complex made by reaction of 1a (Figure 2.3) with

[MoCl(η3-C3H5)(CO)2(NCMe)2] contained a 1:2 ligand to metal ratio in which

18

the bispyridylamide was bound to Mo via the pyridine nitrogen atoms as well as the carbonyl oxygen atoms (Figure 2.7).22

Figure 2.7. Structure of a Mo-allyl complex of ligand 1a.

We prepared a Mo complex of 1a by heating an excess of Mo(CO)6 together

with the ligand in C6D6. The dark red reaction product contained unreacted ligand together with a metal complex which, according to NMR experiments (Figure 2.8) had lost the C2-symmetry.

Figure 2.8. Above: 1H NMR spectrum of a mixture of Mo complex and 1a resulting from reaction with Mo(CO)6. Stars indicate signals coming from 1a. Below: 1H NMR spectrum of

1a.

In order to elucidate whether the complex contained two molecules of the

ligand, we prepared a mixture of complexes using rac-1a. We did not observe any new signals in the 1H NMR spectrum due to diastereomeric complex for-mation, which strongly indicated that the complex was not composed of two or more ligands. Assuming an 1:1 ratio of ligand to metal, it seemed reasonable,

19

that the structure of this complex was similar to that found for related ligands by Pfaltz (Figure 2.9),23 and in which the ligand would coordinate via the two pyridine nitrogen atoms and one of the carbonyl oxygens. No conclusive evi-dence for this was obtained by us for the Mo complex with 1a, as the isolation of the pure complex was difficult.

Figure 2.9. Structure of a Mo complex with related ligands.

2.4 Use in catalysis Even though the coordination chemistry of bispyridylamides has been studied intensively, there have not been many reports on the use of the metal com-plexes in catalysis. For enantiomerically pure complexes, the first report of their use in asymmetric catalysis appeared in 1995.24

Catalytic oxidations were the reactions most commonly studied because of the analogy of these ligands to porphyrins. The oxidation of alkenes employing metal complexes of Os(IV),25 Mn(III)26 or Fe(III)27 using PhIO as reoxidant, re-sulted in low to moderate yields of the products. Hydroxylation of alkenes with Mn(III) using PhIO as reoxidant yielded mixtures of alcohols and ketones in low to moderate yields.28 Oxygen activation, achieved by means of Co(II), allowed the oxidation of phenols to benzoquinones.29

As for chiral enantiomerically pure catalysts, a complex made in situ by mixing 5a (Figure 2.4) and Zr(OtBu)4 catalyzed the ring opening of cyclohexene oxide with trimethylsilyl azide in 45% yield and 56% ee after 4 days.24 Addition of diethylamine increased the reactivity and enantioselectivity of the catalyst and the product was obtained in a maximum yield of 60% and 71% ee.

Complexes of 5a with Co(II), Ir(II) and Rh(II) were used in the asymmetric hydride transfer reduction of acetophenone resulting in low conversion and poor enantioselectivities.30

20

The most successful use of these ligands in asymmetric catalysis has so far been found in Mo-catalyzed asymmetric allylic alkylation reactions (see next Chapter).

References 1. Ojima, H. Synthesis of N,N'-bis(picolinoyl)ethylenediaminato-, N,N'-bis(picolinoyl)-1,2-propanediaminato-, and N,N'-bis(picolinoyl)-1,3-propanediaminatocopper(II) and -nickel(II) and their properties. Nippon Kagaku Kaishi 1967, 88, 333-339. CAN 67:7591. 2. (a) Belda, O.; Moberg, C. Bispyridylamides-Coordination Chemistry and Applications in Catalytic Reactions. Submited to Coord. Chem. Rev. (b) Moberg, C; Adolfsson, H.; Wärnmark, K. Pyridinamides in Asymmetric Catalysis. Acta Chem. Scand. 1996, 50, 195-202. 3. Biniecki, S.; Herold, F. Synthesis of meso-1,2-diphenylethylenediamine Derivatives N,N’-diacylated with Pyridinecarboxylic Acids. Acta Pol. Pharm. 1972, 29, 117-123. CAN 77:101346. 4. (a) Barnes, D. J.; Chapman, R. L.; Vagg, R. S.; Watton, E. C. Synthesis of Novel Bis(amides) by Means of Triphenyl Phosphite Intermediates. J. Chem. Eng. Data 1978, 23, 349-350. (b) Fenton, R. R.; Stephens, F. S.; Vagg, R. S. Studies on the Metal Amide Bond 21. Metal Complexes of Chiral Phenyl-Substituted Bis-Picolinamide Tetradentates. J. Coord. Chem. 1991, 23, 291-311. 5. Lin, J-H.; Che, C-M.; Lai, T-F.; Poon, C-K.; Cui, Y. X. An Approach to Chiral Metal Catalysts; the Synthesis and X-Ray Crystal Structures of a Chiral Tetradentate Binaphthyl Bis-amide Ligand and its Osmium(VI)-oxo Derivative. J. Chem. Soc., Chem. Commun. 1991, 468-470. 6. Conlon, D. A.; Yasuda, N. Practical Synthesis of Chiral N,N’-Bis(2’-pyridinecarbox-amide)-1,2-cyclohexane Ligands. Adv. Synth. Catal. 2001, 1, 137-138. 7. Joule, M.; Mills, K. Heterocyclic Chemistry; Blackwell Science: Oxford, 2000. 8. Funeriu, D. P.; Lehn, J-M.; Baum, G.; Fenske, D. Double Subroutine Self-Assembly; Spontaneous Generation of a Nanocyclic Dodecanuclear CuI Inorganic Architecture. Chem. Eur. J. 1997, 3, 99-104. 9. Sammakia, T.; Hurley, T. B. Enhanced Selectivities for the Hydroxyl-Directed Methanoly-sis of Esters Using the 2-Acyl-4-aminopyridine Class of Acyl Transfer Catalysts: Ketones as Binding Sites. J. Org. Chem. 2000, 65, 974-978. 10. Throughout this work, microwave heating has been employed in many cases as a heating method alternative to conventional heating in an oil bath. Performing reactions in microwave

21

cavities has become a general practice in many laboratories due to a general increase in reac-tion rates and yields. Since the aim of this thesis was not to study the effect of microwave heating in the reactions, no discussion on how microwave heating functions or on the possible acceleration mechanism has been included here. For the interested reader, I recommend the following reviews of the field: 1) Perreux, L.; Loupy, A. A Tentative Rationalization of Mi-crowave Effects in Organic Synthesis According to the Reaction Medium, and Mechanistic Considerations. Tetrahedron 2001, 57, 9199-9223. 2) Lidström, P.; Tierney, J.; Wathey, B.; Westman, J. Microwave Assisted Organic Synthesis - A Review. Tetrahedron 2001, 57, 9225-9283. 11. For excellent reviews see a special issue devoted to Recoverable Catalysts and Reagents: Gladysz, J. A. Ed. Chem. Rev. 2002, 102, 3215-3892. 12. Havranek, M.; Sames, D. Evolution and Study of Polymer-Supported Metal Catalysts for Oxygen Atom Transfer: Oxidation of Alkanes and Alkenes by Diamide Manganese Com-plexes. J. Am. Chem. Soc. 1999, 121, 8965-8966. 13. Song, C. E.; Yang, J. W.; Roh, E. J.; Lee, S.; Hyeon, A.; Han, H. Heterogeneous Pd-Catalyzed Asymmetric Allylic Substitution Using Resin-Supported Trost-Type Bisphosphane Ligands. Angew. Chem., Int. Ed. 2002, 41, 3852-3854. 14. Trost, B. M.; Pan, Z.; Zambrano, J.; Kujat, C. Polymer-Supported C2-Symmetric Ligands for Palladium-Catalyzed Asymmetric Allylic Alkylation Reactions. Angew. Chem., Int. Ed. 2002, 41, 4691-4693. 15. Chapman, R. L.; Vagg, R. S. Studies on the Metal-Amide Bond. I. Metal Complexes of the Bis-amide Tetradentate Ligand N,N’-Bis(2’-pyridine-carboxamide)-1,2-benzene. Inorg. Chim. Acta 1979, 33, 227-234. 16. Chapman, R. L.; Stephens, F. S., Vagg, R. S. Studies on the Metal-Amide Bond. II. The Crystal Structure of the Deprotonated Copper(II) Complex of N,N’-Bis-(2’-pyridinecarbox-amide)-1,2-benzene. Inorg. Chim. Acta 1980, 43, 29-33. 17. Stephens, F. S.; Vagg, R. S. Studies on the Metal-Amide Bond. XIX. A Comparison of Molecular Distortions in the Crystal Structures of [N,N’-Bis(2’-pyridinecarboxamido)-1,2-benzene]nickel(II) with its 6’-Methyl-substituted Analogue. Inorg. Chim. Acta 1986, 120, 165-171. 18. Muthu, S.; Yip, J. H. K.; Vittal, J. J. Coordination Polymers of d10 Metals and N.N’-Bis(3-pyridinecarboxamide)-1,2-ethane. J. Chem. Soc., Dalton Trans. 2001, 24, 3577-3584 19. Fenton, R. R.; Stephens, F. S.; Vagg, R. S. Studies on the Metal Amide Bond 21. Metal Complexes of Chiral Phenyl-Substituted Bis-Picolinamide Tetradentates. J. Coord. Chem. 1991, 23, 291-311. 20. Zafiropoulos, T. F.; Perlepes, S. P.; Tsangaris, J. M. Complexes of Lanthanide(III) Salts with N, N’-Bis(2’-Pyridinecarboxamide)-1,8-Naphthalene. J. Coord. Chem. 1985, 14, 87-90.

22

21. Krska, S. W.; Hughes, D. L.; Reamer, R. A.; Mathre, D. J.; Sun, Y.; Trost, B. M. The Un-usual Role of CO Transfer in Molybdenum-Catalyzed Asymmetric Alkylations. J. Am. Chem. Soc. 2002, 124, 12656-12657. 22. Morales, D.; Pérez, J.; Riera, L.; Riera, V.; Corzo-Suárez, R.; García-Granda, S.; Miguel, D. An Easily Accessed Molybdenum Lewis Acid as a Catalyst for Imine Aziridination. Organometallics 2002, 21, 1540-1545. 23. Glorius, F.; Neuburger, M.; Pfaltz, A. Highly Enantio- and Regioselective Allylic Alkylations Catalyzed by Chiral [Bis(dihydrooxazole)]molybdenum Complexes. Helv. Chim. Acta 2001, 84, 3178-3196. 24. Adolfsson, H.; Moberg, C. Chiral Lewis Acid Catalysed Asymmetric Nucleophilic Ring Opening of Cyclohexane Oxide. Tetrahedron: Asymmetry 1995, 6, 2023-2031. 25.Che, C.-M.; Cheng, W.-K. Manganese(III) Amide Complexes as a New Class for Efficient Alkene Epoxidation. J. Chem. Soc., Chem. Commun. 1986, 1443-1444. 26. Knops-Gerrits, P.-P.; De Vos, D. E.; Jacobs, P. A. Oxidation Catalysis with Semi-Inor-ganic Zeolite-Based Mn Catalysts. J. Mol. Catal. A. 1997, 117, 57-70. 27. Yang, Y.; Diederich, F.; Valentine, J. S. Lewis Acidic Catalysts for Olefin Epoxidation by Iodosylbenzene. J. Am. Chem. Soc. 1991, 113, 7195-7205. 28. Leung, W.-H.; Ma, J.-X.; Yam, V. W.-W.; Che, C.-M.; Poon, C.-K. Syntheses, Electro-chemistry and Reactivities of Pyridine Amide Complexes of Chromium(III) and Manga-nese(III). J. Chem. Soc., Dalton Trans. 1991, 1071-1076. 29. Ganeshpure, P. A.; Sudalai, A.; Satish, S. Oxidation of Phenols with Molecular Oxygen Catalysed by [N,N’-Bis(2’-pyridinecarboxamido)-1,2-benzene]cobalt(II), Chelate. Tetrahedron Lett. 1989, 30, 5929-5932. 30. Halle, R.; Bréhéret, A.; Schulz, E.; Pinel, C.; Lemaire, M. Chiral Nitrogen-Metal Com-plexes for the Asymmetric Reduction of Ketones. Tetrahedron: Asymmetry 1997, 8, 2101-2108.

23

3 Application of bispyridylamides as ligands

in Mo-catalyzed asymmetric allylic alkylations

3.1 Introduction1

The Pd-catalyzed asymmetric allylic alkylation reaction has been the subject of numerous studies and a large variety of ligands have been developed, allowing the highly enantioselective formation of carbon-carbon and carbon-heteroatom bonds.2 The analogous Mo-catalyzed alkylation reaction has been less studied, but it has already developed into a powerful synthetic procedure, complemen-tary to the Pd-catalyzed process in that, in contrast to the situation with Pd, allylic alkylation takes place mostly at the more substituted carbon atom when unsymmetrical substrates are used.3 Other metals, such as W,4 Rh,5 Ir,6 Ni,7 Pt8 and Cu9, have been used in order to catalyze these reactions and present the same preferential regioselectivity as Mo. The choice of the metal has therefore been determined by the reactivity of the metal (the amount of catalyst needed), its selectivity for a given class of allylic substrates, its reactivity towards different types of nucleophiles (stabilized or nonstabilized carbon nucleophiles or heteroatomic nucleophiles) and various practical aspects, e.g. stability and cost.

Molybdenum has been commonly used in catalysis partly because of the vast oxidation number states (from –4 to +6) and the flexible coordination number (from 4 to 8) exhibited in the compounds which it forms.10 Molybdenum π-allyl complexes and their reactivity were widely studied, and this fact made the use of this metal as a catalyst in allylic alkylation reactions very interesting.11

24

Chiral, highly selective ligands for Mo-catalyzed allylic alkylations have emerged only quite recently and different classes of chiral ligands have been developed for these reactions.

The first report on asymmetric Mo-catalyzed allylic alkylation was published by Trost and Hachiya in 1998 (Scheme 3.1)12, but some preliminary results were mentioned earlier by Kočovský.13 The reaction was performed with a catalyst produced by heating bispyridylamide 1a (Figure 2.3) together with (EtCN)3Mo(CO)3. Addition of the allylic substrate and the nucleophile to the reaction mixture afforded the product in high yield as well as regio- and enan-tioselectivity. In this way, several aryl-substituted allylic carbonates were al-kylated with sodium dimethyl malonate. Bulkier 2-substituted malonates were also used and the products were still obtained in high yields and with high se-lectivities.

Scheme 3.1. Mo-catalyzed asymmetric allylic substitution of aromatic allylic carbonates with sodium

dimethyl malonate.

The excellent temperature stability of this catalyst prompted Hallberg and co-workers, together with us, to develop a more convenient experimental method for this reaction. As previously shown for Pd,14 microwave heating could be used instead of conventional heating to accelerate the reaction without any sig-nificant loss in selectivity (5 minutes compared to 4 hours). In addition, a more stable and cheap Mo source, Mo(CO)6, could be used instead of (EtCN)3Mo(CO)3 by using microwave heating, (Table 3.1).15 The yield as well as the regioselectivity were somewhat higher when microwave heating was used (86% yield and 19:1 compared to 59% yield and 11:1). It was shown later, that the reaction could be performed on a large scale with conventional heating, using Mo(CO)6 and the appropriate activation temperature and time.16

25

Table 3.1. Allylic Alkylation of 10a (Eq. 1) with Malonate, with 1a as Ligand.

activation reaction Mo salt % cat.

(mmol) time (h)

T (°C)

time (h)

T (°C)

% yield

b:l % ee

(EtCN)3Mo(CO)3 10 1 60 3 rt 70 49:1 99b

(EtCN)3Mo(CO)3 10 1 60 3 70 88 32:1 99b

Mo(CO)6 4 - - 0.083 165a 86 19:1 98c

Mo(CO)6 4 - - 0.1 165 59 11:1 98c

Mo(CO)6 10 4 85 8 85 80 19:1 99d

a Heating performed in a microwave cavity. b Ref. 12. c Ref. 14. d Ref. 15.

Molybdenum-catalyzed asymmetric allylic alkylations have been performed with other types of chiral ligands, including bis(dihydrooxazoles)17 and bipyri-dines.18

3.2 Electronic and steric effects of the ligandII

Electronic and steric factors play important roles in asymmetric induction with metal catalysts. For example, in Pd-catalyzed asymmetric allylic alkylations, the different trans influence of the donor atoms in P,N ligands is assumed to be the major enantiodifferentiating factor, whereas for ligands having donor atoms with similar trans influence, such as N,N ligands, the enantioselectivity is assumed to originate from the different steric environment in the allylic ter-mini, resulting in a different C-Pd bond distances and electron densities in the allyl ligand. Changes in the electronic and steric properties can also affect the reactivity of the system. Therefore, the variation of these properties can be used to fine-tune the catalyst.

The electronic properties of the catalyst are varied not only by using ligands with atoms with different trans influence, but also by changing the σ donor π acceptor properties of the atoms in the ligand, e.g. by introducing different electron-donating or electron-withdrawing substituents in a pyridine ring.

26

We used bispyridylamides 1a-f, 2a-c, 3a and 4a-e (Figure 2.3, Schemes 2.1 and 2.2) with differently substituted pyridine rings as ligands, in order to study the influence of electronic and steric factors in Mo-catalyzed asymmetric ally-lic alkylations (Table 3.2).

Table 3.2. Differently Substituted Bispyridylamides as Ligands for Mo-catalyzed Asymmetric Allyla-

tion of Substrate 10a in THF using Mo(CO)6 (Eq. 1).

entry ligand T (°C) time (min) % yield b:l % ee 1 1a 160 6 80 19:1 98 2 1b 165 5 46 13:1 98 3 1c 165 8 32 16:1 97 4 1c 150 15 37 16:1 97 5 1d 165 4 88 41:1 >99 6 1e 165 6 89 74:1 96 7 1f 170 12 91 88:1 96 8 2a 165 5 30 13:1 79 9 2b 160 6 traces - - 10 2c 160 6 traces - - 11 3a 160 6 traces - - 12 4a 160 6 89 74:1 97 13 4b 160 6 traces - - 14 4c 160 6 90 98:1 97 15 4d 160 6 89 75:1 99 16 4e 160 12 traces - -

The results showed that electronic and steric effects had a strong influence on

the outcome of this reaction. Thus, ligands containing π-donor substituents in the 4-positions of the pyridine rings favored the formation of the branched product. For the bispyridylamide 1d, having methoxy groups, an increased re-action rate was also observed (entry 5). In contrast to this, the bispyridylamide 1f, having pyrrolidyl groups instead of the methoxy groups, made the catalyst work more slowly but afforded the product with higher regioselectivity. Bis-pyridylamide 1c, with nitro groups in the 4-positions of the pyridine rings, gave a less active as well as less regio- and enantioselective catalyst (entries 3 and 4). Ligands with substituents in the 6-positions of the pyridine rings (entries 8, 9 and 10) gave rise to both less active and less selective catalysts. Low activity was also observed with pseudo C2-symmetric 4b, that had one pyridine ring substituted by a nitro- group in the 4-postion and the other pyridine ring

27

substituted by a chlorine atom in the 4-position. All other pseudo C2-symmetric ligands displayed a behavior similar to that of the corresponding C2-symmetric analogs, even though the ligand that induced the highest regioselectivity was that with a 4-MeO-substituted pyridine ring (4c, entry 14). Ligands with free hydroxyl groups (3a) or primary amino groups (4e) formed rather inactive catalysts.

It is difficult to rationalize these results with the scarce mechanistic informa-tion available. However, it is reasonable to suppose that electronic effects are more important than steric ones for ligands with substituents in the 4-positions. For ligands with substituents in the 6-positions, steric effects are predominant over electronic ones.

Trost and co-workers found that higher regioselectivity (46:1 compared with 19:1) was obtained when one picolinamide group was replaced by a nicotina-mide group.19 An even higher regioselectivity (60:1) was obtained when one picolinamide group was replaced by a benzoylamide group. This result showed that coordination of both pyridine nitrogen atoms was not a necessary condition for high regioselectivity. In the same investigation it was also reported that analogue ligands, in which the pyridine rings were replaced by quinoline rings, were not as efficient as ligands in the reaction.

The operating mechanism of the molybdenum-catalyzed asymmetric allylic alkylation has been less studied than that of the palladium-catalyzed process. However, a catalytic cycle has been proposed after isolation and characteriza-tion of several intermediate metal complexes with ligand 7.20 Thus, the Mo π-allyl complex 9 (Scheme 2.5) reacted with sodium dimethyl malonate in the presence of CO to give the product and intermediate (11), which then reacted with the allylic carbonate to regenerate 9 (Scheme 3.2). However, this model does not account for the origin of the regio- and enantioselectivity observed.

28

Scheme 3.2. Proposed mechanism of the Mo-catalyzed asymmetric allylic alkylation of substrate 10a

with ligand 7.

Although the stereochemistry of Mo-catalyzed allylic alkylation was known

to proceed with overall retention of the stereochemistry, the stereochemistry of each step, the oxidative addition and the nucleophilic attack, was unknown un-til recently, even though some evidence was obtained.13

Conclusive evidence has now been provided for a retention-retention pathway in Mo-catalyzed asymmetric allylic alkylations.21 In order to explain the reten-tion mechanism of the oxidative addition, a model has been proposed in which precoordination of the metal to the leaving group occurs.22 As for the syn attack of the nucleophile, precoordination of malonate to the metal and subsequent attack on the π-allyl ligand23 or nucleophilic attack in a SN2´ mode on a η1 complex have both been proposed.22 However, the isolation of π-allyl complex 9 has made the former hypothesis more solid.

29

3.3 Substrates and nucleophiles A number of allylic substrates and carbon nucleophiles (vide infra) have been employed in Mo-catalyzed asymmetric allylic alkylation reactions. The most widely used leaving group is carbonate because of its high reactivity in the re-action, but acetate17 and phosphate24 have also been successfully employed. Catalysts made with bispyridylamides as ligands have been more efficient for aryl-substituted allylic substrates in general, whereas for aliphatic allylic sub-strates, bis(dihydrooxazoles) ligands have afforded the products with higher regio- and enantioselectivity. Polyallylic substrates have also been successfully alkylated and, in those cases, substitution has mostly taken place at the 4-posi-tion. In contrast to Pd-catalyzed allylation of these kind of polyallylic substrates, no attack on the benzylic position has been observed.25 Diverse sta-bilized carbon nucleophiles including malonates, the lithium enolate of glycine ester24 and lithium salts of azlactones,24 have been used in the reaction.

Figure 3.1. Structures of allylic carbonates.

We tried a number of derivatives of 10a and of its branched isomer 12a,

containing substituents in the para position of the aromatic ring (Figure 3.1), in the Mo-catalyzed asymmetric allylic alkylation using the parent bispyridyla-mide 1a and 4-chloro-substituted 1e (Figure 2.3) as ligands, in order to study the generality of the results obtained (Scheme 3.3, Table 3.3).

Scheme 3.3. Reaction of 10a and related compounds (Figure 3.1) with sodium dimethyl malonate in

THF.

30

Table 3.3. Mo-Catalyzed Asymmetric Allylic Alkylation of 10a-d and 12a-c.

entry substrate ligand % yield b:l % ee 1 10a 1a 80 19:1 98 2 10a 1e 89 74:1 96 3 10b 1a 51 32:1 96 4 10b 1e 86 47:1 90 5 10c 1a 59 51:1 98 6 10c 1e 79 57:1 94 7 10d 1a 50 11:1 99 8 10d 1e 66 22:1 96 9 12a 1a 76 13:1 96 10 12a 1e 89 69:1 86 11 12b 1a 78 26:1 96 12 12b 1e 70 34:1 74 13 12c 1a 48 10:1 98 14 12c 1e 52 20:1 90

Most products were obtained in good yields and with high regio- and

enantioselectivity. When the catalyst was formed with bispyridylamide 1e, the products were consistently obtained with higher b:l (branched to linear) ratios and in higher yields. The enantioselectivity on the other hand was generally somewhat lower. These results also showed that π-donating groups in the para position of the aromatic ring of the allylic carbonate afforded the corresponding products with higher b:l (branched to linear) ratios.

Isomeric branched carbonates 12a-c, gave lower regio- and enantioselectivity than substrates 10a-d, in all cases. This is in agreement with the results found by Hughes et al. They observed that even though the two enantiomers of 12a gave the same enantiomer of the product, they reacted with different rates.26 Thus, it was shown that the allylic alkylation of these substrates, using ligand 1a, proceeded via a kinetic resolution.

When the reaction is conducted to full conversion, the process should be re-ferred as a dynamic kinetic transformation (DYKAT),2 as distinguished from a kinetic resolution, in which the reaction is stopped at less than 50% conversion and enantioenriched starting material and product are obtained. Furthermore, the rate of diastereomeric equilibration of the π-allyl intermediates relative to the subsequent steps in the reaction was shown to be dependent on the solvent and the nucleophile resulting in different ee:s (Scheme 3.4).

31

Scheme 3.4. Memory effect in Mo-catalyzed asymmetric allylic alkylations.26

Thus, in solvents such as THF, tetrahydropyran, iso-PrOAc, and MeCN, a

significant stereochemical memory effect was operative, with the slow-reacting enantiomer providing the product in lower ee than the fast reacting one. There-fore, when the reaction was run to completion in these solvents, branched car-bonates afforded the product in lower ee than linear ones. In toluene and 1,2-dichloroethane, the product was obtained with similar ee:s from the two enantiomers of the branched carbonate, probably due to the lower solubility of sodium dimethyl malonate, making the nucleophilic attack step slower, allowing the equilibration of the π-allyl intermediates before the nucleophilic attack and thus a minimized memory effect.

Figure 3.2. Allylic substrates used for allylic alkylation with dimethyl malonate.

32

We also tried other aryl-substituted (13, 14) and aliphatic substrates (15-17)

(Figure 3.2) in the reaction. We could not observe any product under the reac-tion conditions. This was in agreement with previous observations regarding aliphatic allylic carbonates and ligand 1a.22 However, by increasing the catalyst loading to 10% and using a more concentrated solution of substrate 17 and a stoichiometric amount of sodium dimethyl malonate, i.e. no BSA, the reaction was completed after 5 min at 160 °C, and the branched product could be iso-lated in 48% yield. Unfortunately, the product obtained in this way was race-mic, because it was formed via an achiral pathway, more favorable than the chiral one under these reaction conditions.

Figure 3.3. Heteroatomic nucleophiles.

We also used N and O nucleophiles in this study (Figure 3.3). No reaction

was observed when using tosylamide as nucleophile, neither using its sodium salt nor the BSA method. As for sodium phenoxide, we observed a complete conversion of the starting material into two products in about a 1:1 ratio. Nei-ther the structural characterization of the isomeric products (a subsequent Claisen rearrangment was possible), nor a separation of the enantiomers could be made.

3.4 Heterogeneous catalystsIII

Several solid-supported ligands had been developed and used for Pd-catalyzed asymmetric allylic alkylation.27 However, no report on solid-supported ligands for the Mo-catalyzed reaction had been published when this work started.

We used the solid-supported bispyridylamides 4f (Scheme 2.3) and 4h (Scheme 2.4) in the allylic alkylation of 10a (Figure 3.3) with dimethyl malo-nate (Table 3.4).

33

Table 3.4. Allylic Alkylation of 10a with Solid-Supported Bispyridylamides as Ligands.

ligand T (°C) t (min) % yield b:l % ee 4f 160 20 90a 35:1 97 4f 160 25 95a 35:1 97 4f 160 30 82b 35:1 97 4h 160 30 tracesa - -

a Determined by GC. b Isolated yield.

When using solid-supported bispyridylamide 4f, the product was obtained in

similar yield as well as similar regio- and enantioselectivity as when the homo-geneous system was used. However, a decreased reactivity was observed with this system, probably due to the heterogeneous nature of the catalyst. Note-worthy, the product was easily isolated by filtration and the solid-supported bispyridylamide could be reused up to seven times without change in the reac-tion outcome. When the bispyridylamide attached to an inorganic solid support 4h, was used, the reaction was very slow. Whether the reason was the hetero-geneous nature of catalysts or some interaction of Mo with the silica employed was not clear.

3.5 Application of the reaction in enantioselective synthesisIII

The Mo-catalyzed asymmetric allylic alkylation has been employed as the key step in several enantioselective syntheses, demonstrating its high synthetic value. Reaction with lithium salts of azlactones afforded protected unnatural quaternary α-amino acids.24 Palucki et al. used the allylation reaction in the de-velopment of a synthetic method towards a 3,4-disubstituted cyclopentanone drug intermediate (Figure 3.4).28 Trost and Andersen utilized both Pd and Mo DYKAT for the synthesis of Tipranavir (Figure 3.4).29

34

Figure 3.4. Structure of an advance drug intermediate and Tipranavir.

We have used the Mo-catalyzed asymmetric allylic alkylation reaction for the

synthesis of (R)-baclofen, a γ-aminobutyric acid (GABA) derivative which is used as a muscle relaxant. Due to the high substrate generality of the allylation reaction, this synthesis could also be used, in principle, for preparing other GABA derivatives, such as (R)-rolipram (Figure 3.5).

Figure 3.5. Structures of commercial drugs based on GABA.

As seen before (Table 3.3), substrate 12b could be alkylated with dimethyl

malonate in the presence of a Mo catalyst formed in situ from Mo(CO)6 and ligand 1a in 78% yield, 26:1 branched to linear ratio and 96% ee (Figure 3.6).

Figure 3.6. Allylic alkylation of 12b to give (R)-baclofen intermediate 18.

35

We could also prepare 18 by using the solid-supported bispyridylamide 4f as

ligand for the reaction (Figure 3.6, Table 3.5). In this case, the solvent should be carefully chosen in order to minimize the memory effect. Thus, by using a 9:1 mixture of toluene and THF the ee of the product could be rised from 48% (entry 1) to 89% (entry 3). The observation of a lower memory effect in toluene than in THF is in agreement with the results shown by Hughes for ligand 1a.26

Table 3.5. Allylic Alkylation of 12b with Solid-supported Bispyridylamide (S,S)-4f (Figure 3.6).

entry THF (mL)

PhMe (mL)

t (min)

% yield b:l % ee

1 1 0 30 95a 27:1 48 2 0.5 0.5 30 98a 24:1 76 3 0.1 0.9 30 76b 25:1 89

a Determined by GC. b Isolated yield.

Decarboxylation of the geminal diester was achieved by heating a solution of

18 and NaCl in a mixture of DMSO and H2O to 200 °C for 20 min in a micro-wave cavity. Oxidation of the double bond in 19 with ozone at -78 °C for 10 min and cleavage of the ozonide obtained with Me2S, afforded the corresponding aldehyde30 that was immediately treated with NH4OAc and NaBH3CN31 at room temperature for 12h. After addition of 2 N NaOH to the reaction mixture at room temperature, work-up and formation of the hydrochlo-ric salt, the product was obtained in 22% overall yield from 18 and 96% ee according to optical rotation (Scheme 3.5).

Scheme 3.5. Synthesis of (R)-baclofen from intermediate 18.

36

References 1. For a review of the field see: Belda, O.; Moberg, C. Molybdenum-Catalyzed Asymmetric Allylic Alkylations. Acc. Chem. Res. 2004, 37, 159-167. 2. Trost, B. M. Pd Asymmetric Allylic Alkylation (AAA). A Powerful Synthetic Tool. Chem. Pharm. Bull. 2002, 50, 1-14. 3. There are few examples where this regioselectivity is also observed for Pd, see: (a) You, S-L.; Zhu, X-Z; Luo, Y-M.; Hou, X-L.; Dai, L-X. Highly Regio- and Enantioselective Pd-Catalyzed Allylic Alkylation and Amination of Monosubstituted Allylic Acetates with Novel Ferrocene P,N-Ligands. J. Am. Chem. Soc. 2001, 123, 7471-7472, and references therein. (b) Prétôt, R.; Pfaltz, A. New Ligands for Regio- and Enantiocontrol in Pd-Catalyzed Allylic Al-kylations. Angew. Chem., Int. Ed. 1998, 37, 323-325. (c) For a reversed regioselectivity with Mo see: Co, T. T.; Paek, S. W.; Shim, S. C.; Cho, C. S.; Kim T-J.; Choi, D. W.; Kang, S. O.; Jeong, J. H. 1,2-Ferrocenediylazaphosphinines: An Unusual Coordination Behavior and Application to Allylic Alkylation. Organometallics 2003, 22, 1475-1482. 4. Lloyd-Jones, G. C.; Pfaltz, A. Chiral Phosphanodihydrooxazoles in Asymmetric Catalysis: Tungsten-Catalyzed Allylic Substitution. Angew. Chem., Int. Ed. 1995, 34, 462-464. 5. Hayashi, T.; Okada, A.; Suzuka, T.; Kawatsura, M. High Enantioselectivity in Rhodium-Catalyzed Allylic Alkylation of 1-Substituted 2-Propenyl Acetates. Org. Lett. 2003, 5, 1713-1715. 6. (a) Kanayama, T.; Yoshida, K.; Miyabe, H.; Takemoto, Y. Enantio- and Diastereoselective Ir-Catalyzed Allylic Substitutions for Asymmetric Synthesis of Amino Acid Derivatives. Angew. Chem., Int. Ed. 2003, 42, 2054-2056. (b) Takeuchi, R.; Kashio, M. Iridium Complex-Catalyzed Allylic Alkylation of Allylic Esters and Allylic Alcohols: Unique Regio- and Stereoselectivity. J. Am. Chem. Soc. 1998, 120, 8647-8655. (c) Janssen, J. P.; Helmchen, G. First Enantioselective Alkylations of Monosubstituted Allylic Acetates Catalyzed by Chiral Iridium Complexes. Tetrahedron Lett. 1997, 38, 8025-8026. 7. Bricout, H.; Carpentier, J-F.; Mortreux, A. Bis(aminophosphine)-Nickel Complexes as Efficient Catalysts for Alkylation of Allylic Acetates with Stabilized Nucleophiles. Tetrahedron Lett. 1996, 37, 6105-6108. 8. Brown, J. M.; MacIntyre, J. E. Allylic Alkylation catalysed by Platinum Complexes: Structure and Reactivity of Intermediates, and the Overall Stereoselectivity. J. Chem. Soc., Perkin Trans. 2 1985, 961-970. 9. van Klaveren, M.; Persson, E. S. M.; del Villar, A.; Grove, D. M.; Bäckvall, J-E.; van Koten G. Chiral Arenethiolatocopper(I) Catalyzed Substitution Reactions of Acyclic Allylic Substrates with Grignard Reagents. Tetrahedron Lett. 1995, 36, 3059-3062. 10. Curtis, M. D. In Encyclopedia of Inorganic Chemistry; King, R. B., Ed.; John Wiley & Sons Ltd., 1994, pp 2346-2361.

37

11. Trost, B. M.; Lautens, M. Molybdenum catalysts for allylic alkylation. J. Am. Chem. Soc. 1982, 104, 5543-5545. 12. Trost, B. M.; Hachiya, I. Asymmetric Molybdenum-Catalyzed Allylic Alkylations. J. Am. Chem. Soc. 1998, 120, 1104-1105. 13. Dvořák, D.; Starý, I.; Kočovský, P. Stereochemistry of Molybdenum(0)-Catalyzed Allylic Substitution: The First Observation of a Syn-Syn Mechanism. J. Am. Chem. Soc. 1995, 117, 6130-6131. 14. (a) Kaiser, N-F.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. Microwave-mediated palladium-catalyzed asymmetric allylic alkylation; an example of highly selective fast chemistry. J. Organomet. Chem. 2000, 603 (1), 2-5. (b) Bremberg, U.; Lutsenko, S.; Kaiser, N-F.; Larhed, M.; Hallberg, A.; Moberg, C. Rapid and Stereoselective C-C, C-O, C-N and C-S Couplings via Microwave-Accelerated Palladium-Catalyzed Allylic Substitutions. Synthesis 2000, 7, 1004-1008. (c) Bremberg, U.; Larhed M.; Moberg, C.; Hallberg, A. Rapid Microwave-Induced Palladium Catalyzed Asymmetric Allylic Alkylation. J. Org. Chem. 1999, 64, 1082-1083. 15. Kaiser, N-F.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg, A. Fast, Convenient, and Efficient Molybdenum-Catalyzed Asymmetric Allylic Alkylation under Noninert Conditions: An Example of Microwave-Promoted Fast Chemistry. Angew. Chem., Int. Ed. 2000, 39, 3596-3598. 16. Palucki, M.; Um, J. M.; Conlon, D. A.; Yasuda, N.; Hughes, D. L.; Mao, B.; Wang, J.; Reider, P. J. Molybdenum-Catalyzed Asymmetric Allylic Alkylation Reactions Using Mo(CO)6 as Precatalyst. Adv. Synth. Catal. 2001, 343, 46-50. 17. (a) Glorius, F.; Neuburger, M.; Pfaltz, A. Highly Enantio- and Regioselective Allylic Al-kylations Catalyzed by Chiral [Bis(dihydrooxazole)]molybdenum Complexes. Helv. Chim. Acta 2001, 84, 3178-3196. (b) Glorius, F.; Pfaltz, A. Enantioselective Molybdenum-Cata-lyzed Allylic Alkylation Using Chiral Bisoxazoline Ligands. Org. Lett. 1999, 1, 141-144. 18. Malkov, A. V.; Baxendale, I. R.; Bella, M.; Langer, V.; Fawcett, J.; Russell, D. R.; Mansfield, D. J.; Valko, M.; Kočovský, P. Synthesis of New Chiral 2,2’-Bipyridyl-Type Ligands, Their Coordination to Molybdenum(0), Copper(II), and Palladium(II), and Applica-tion in Asymmetric Allylic Substitution, Allylic Oxidation, and Cyclopropanation. Organometallics 2001, 20, 673-690. 19. Trost, B. M.; Dogra, K.; Hachiya, I.; Emura, T.; Hughes, D. L.; Krska, S.; Reamer, R. A.; Palucki, M.; Yasuda, N.; Reider, P. J. Designed Ligands as Probes for the Catalytic Binding Mode in Mo-Catalyzed Asymmetric Allylic Alkylation. Angew. Chem., Int. Ed. 2002, 41, 1929-1932. 20. Krska, S. W.; Hughes, D. L.; Reamer, R. A.; Mathre, D. J.; Sun, Y.; Trost, B. M. The Un-usual Role of CO Transfer in Molybdenum-Catalyzed Asymmetric Alkylations. J. Am. Chem. Soc. 2002, 124, 12656-12657.

38

21. Lloyd-Jones, G. C.; Krska, S.; Hughes, D. L.; Gouriou, L.; Bonnet, V. D.; Jack, K.; Sun, Y.; Reamer, R. A. Conclusive Evidence for a Retention-Retention Pathway for Molybdenum-catalyzed Asymmetric Alkylation. J. Am. Chem. Soc. 2004, 126, 702-703. 22. Kočovský, P.; Malkov, A. V.; Vyskočil, Š.; Lloyd-Jones, G. C. Transition metal catalysis in organic synthesis: reflections, chirality and new vistas. Pure Appl. Chem. 1999, 71, 1425-1433. 23. (a) Trost, B. M.; Lautens, M. Chemoselectivity and Stereocontrol in Molybdenum-Cata-lyzed Allylic Alkylations. J. Am. Chem. Soc. 1987, 109, 1469-1478. (b) Krafft, M. E.; Procter, M. J.; Abboud, K. A. Synthesis of Molybdenum Dicarbonyl Complexes Bearing Tethered Homoallylic Amines and Sulfides. Organometallics 1999, 18, 1122-1124. 24. Trost, B. M.; Dogra, K. Synthesis of Novel Quaternary Amino Acids Using Molybdenum-Catalyzed Asymmetric Allylic Alkylation. J. Am. Chem. Soc. 2002, 124, 7256-7257. 25. Trost, B. M.; Hildbrand, S.; Dogra, K. Regio- and Enantioselective Molybdenum-Cata-lyzed Alkylations of Polyenyl Esters. J. Am. Chem. Soc. 1999, 121, 10416-10417. 26. Hughes, D. L.; Palucki, M.; Yasuda, N.; Reamer, R. A.; Reider, P. J. Solvent-Dependent Dynamic Kinetic Asymmetric Transformation/Kinetic Resolution in Molybdenum-Catalyzed Asymmetric Allylic Alkylations. J. Org. Chem. 2002, 67, 2762-2768. 27. (a) Song, C. E.; Yang, J. W.; Roh, E. J.; Lee, S.; Hyeon, A.; Han, H. Heterogeneous Pd-Catalyzed Asymmetric Allylic Substitution Using Resin-Supported Trost-Type Bisphosphane Ligands. Angew. Chem., Int. Ed. 2002, 41, 3852-3854. (b) Trost, B. M.; Pan, Z.; Zambrano, J.; Kujat, C. Polymer-Supported C2-Symmetric Ligands for Palladium-Catalyzed Asymmetric Allylic Alkylation Reactions. Angew. Chem., Int. Ed. 2002, 41, 4691-4693. (c) Hallman, K.; Macedo, E.; Nordström, K.; Moberg, C. Enantioselective Allylic Alkylation Using Polymer-supported Palladium Catalysts. Tetrahedron: Asymmetry 1999, 10, 4037-4046. 28. Palucki, M.; Um, J. M.; Nobuyoshi, Y.; Conlon, D. A.; Tsay, F-R.; Hartner, F. W.; Hsiao, Y.; Marcune, B.; Karady, S.; Hughes, D. L.; Dormer, P. G.; Reider, P. J. Development of a New and Practical Route Chiral 3,4-Disubstituted Cyclopentanones: Asymmetric Alkylation and Intramolecular Cyclopropanation as Key C-C Bond-Forming Steps. J. Org. Chem. 2002, 67, 5508-5516. 29. Trost, B. M.; Andersen, N. G. Utilization of Molybdenum- and Palladium-Catalyzed Dy-namic Kinetic Asymmetric Transformations for the Preparation of Tertiary and Quaternary Stereogenic Centers: a Concise Synthesis of Tipranavir. J. Am. Chem. Soc. 2002, 124, 14320-14321. 30. Resende, P.; Almeida, W. P.; Coelho, F. An efficient synthesis of (R)-(-)-baclofen. Tetrahedron: Asymmetry 1999, 10, 2113-2118. 31. Brenna, E.; Caraccia, N.; Fuganti, C.; Fuganti, D.; Graselli, P. Enantioselective synthesis of β-substituted butyric acid derivatives via orthoester Claisen rearrangement of enzymati-

39

cally resolved allylic alcohols: application to the synthesis of (R)-(-)-baclofen Tetrahedron: Asymmetry 1997, 8, 3801-3805.

40

41

4 Application of bispyridylamides as ligands

in asymmetric Lewis acid catalyzed processes

4.1 Introduction

Lewis acid and base catalysts have been widely used to accelerate organic re-actions such as Diels-Alder and other cycloaddition reactions, aldol reactions and additions of nucleophiles to carbonyl groups and epoxides.1 The develop-ment of chiral Lewis acids able to promote equivalent asymmetric reactions has been a growing field during the last decades. Such catalysts have usually been prepared by in situ formation of complexes of an electron deficient metal and a chiral ligand. The metal has commonly been chosen by experimental trial and error, since several fundamental aspects of Lewis acids have not been fully un-derstood yet. Lewis acids have recently been classified on the basis of activity and selectivity in a given reaction.2 However, the most generally used classifi-cation has been that introduced by Pearson in 1963,3 which refers to Lewis acids as being “hard” or “soft” according to their polarizability.

The chiral ligand has a direct influence on the rate and selectivity of the resulting catalyst and, therefore, ligand design is also an important part in the development of chiral Lewis acids for catalytic applications to asymmetric re-actions. The use of bispyridylamides as ligands in Lewis acid catalyzed addi-tion of cyanide to carbonyl groups as well as in ring opening of meso epoxides is discussed in this chapter.

42

4.2 Addition of cyanide to carbonyl groupsIV

The asymmetric addition of cyanide to carbonyl groups (Figure 4.1) is a very useful reaction in asymmetric synthesis, because the resulting enantiomerically enriched cyanohydrins can be transformed into synthetically versatile building blocks without loss of the optical purity.4 Enzymatic methods as well as chiral Lewis acids and bases are all able to induce asymmetry in the reaction.5 Metal-salen complexes are particular examples of catalysts containing tetradentate ligands that are used for this reaction.6

Figure 4.1. Asymmetric Lewis acid catalyzed addition of cyanide to a carbonyl group.

Trimethylsilyl cyanide (TMSCN) has been a common source of cyanide

widely employed in reactions affording silylated cyanohydrins. Uses of more convenient cyanide sources, such as KCN together with Ac2O, have also been reported.7

The reactions have been performed with different kinds of aromatic and ali-phatic carbonyl groups, but the catalysts have been very substrate-specific and, basically, each of the substrates has had to be handled as an individual case for optimization.

Aromatic aldehydes are usually more reactive than aliphatic ones, and ke-tones, especially aliphatic ketones, need even longer reaction times and larger catalyst loadings.

We studied the reaction between benzaldehyde and TMSCN8 at room tem-perature, using a catalyst formed from bispyridylamide 5a (Figure 2.4) and Ti(IV). In order to find the optimal conditions for the reaction we examined the effect of the solvent on a constant amount of a catalyst (5 mol %) formed in situ from 5a and Ti(OiPr)4 (Figure 4.2, Table 4.1). We found that solvents such as THF, EtOAc and toluene afforded the product in both low yield and low ee. When the solvent was dichloromethane or acetonitrile, the product was ob-tained in good yield and with 60% ee. The reaction was faster when acetonitrile was employed than when dichloromethane was used as solvent (1 hour reaction time compared to 4 hours). Addition of small amounts of dichloromethane in order to dissolve the ligand was necessary when acetonitrile was used as sol-vent.

43

Figure 4.2. Optimization of the addition of TMSCN to benzaldehyde: various solvents.

Table 4.1. Effect of the Solvent on the Reaction (Figure 4.2).

entry solvent t (h) % yield of 20a % ee

1 CH2Cl2 4 >95 60 (S) 2 THF: CH2Cl2 (4:1) 1 15 4 (R) 3 EtOAc 1 16 28 (S) 4 CH3CN: CH2Cl2 (5:1) 1 93 60 (S) 5 toluene 1 11 2 (S)

a Determined by GC.

We studied the influence exerted on the reaction by (1) the amount of catalyst in relation to the amount of benzaldehyde, (2) the metal to ligand ratio and (3) the concentration of benzaldehyde at the start of the reaction, when using di-chloromethane as solvent to keep the system as anhydrous as possible (Table 4.2). The results showed that, while the metal to ligand ratio had no major in-fluence on the reaction (entries 3 and 4, Table 4.2), both the relative amount of catalyst and the concentration of benzaldehyde had important effects on the outcome of the reaction. When the reaction started with 1-2.5 mol % of catalyst with respect to benzaldehyde and a concentration of benzaldehyde of 5-10 M, the product was obtained in high yields and with 70% ee (entries 5 and 6). Di-lution of the reaction mixture had a negative impact on the enantioselectivity, but the yield was still high (entries 7-9).

44

Table 4.2. Effect of the Relative Amounts of Ti(OiPr)4 and 5a and of the Concentration of Benzalde-

hyde in the Reaction.

entry 5a (mol %)

Ti(OiPr)4 (mol %)

[PhCHO] (M)

% yield of 20a

% ee

1 5 5 1 >95 60 (S) 2 10 10 1 >95 60 (S) 3 5 10 1 90 60 (S) 4 10 5 1 >95 60 (S) 5 1 1 10 >95 (86)b 70 (S) 6 2.5 2.5 5 >95 70 (S) 7 2.5 2.5 0.5 >95 66 (S) 8 2.5 2.5 0.33 >95 63 (S) 9 2.5 2.5 0.25 >95 57 (S)

a Determined by GC. b Isolated yield.

When we ran the reaction with 1 mol % of catalyst in a microwave cavity at 100 °C, instead of at room temperature, the product was obtained in 91% yield and with 34% ee after 5 minutes of reaction time. We also performed the reac-tion at 0 °C but then the ee of the product was the same as when the reaction was run at room temperature.

We also employed bispyridylamides 1a, 2b (Figure 2.3) and 5b (Figure 2.4) as ligands in the reaction in order to see the influence of different chiral back-bones and of different substituents in the pyridine rings of the ligand (Figure 4.3, Table 4.3). Ligands based on the (1R,2R)-1,2-diaminocyclohexane chiral backbone afforded the product with lower ee than ligand 5a, based on (1R,2R)-1,2-diphenyl-1,2-diaminoethane. A derivative of 1a that contained 6-bromo-substituted pyridine rings, afforded the racemic product (entry 3). The product was also obtained in low ee, if the derivative 5b, having 4-chloro- substituted pyridine rings, was used instead of 5a (entry 4).

Figure 4.3. Asymmetric addition of TMSCN to benzaldehyde with bispyridylamides 1a, 2b and 5b as

ligands.

45

Table 4.3. Effects of Various Bispyridylamides as Ligands in the Reaction (Figure 4.3).

entry ligand ligand (mol %)

Ti(OiPr)4 (mol %)

% yield of 20a

% ee

1 1a 10 10 >95 36 (S) 2 1a 1 1 >95 46 (S) 3 2b 1 1 >95 0 4 5b 1 1 69 20 (S)

a Determined by GC.

In the course of this optimization we observed that the enantiomeric excess of the product 20 increased with time (Figure 4.4).

Figure 4.4. Increase of the ee of 20 with the reaction time.

A possible explanation of this effect may be that a catalyst exhibiting higher enantioselectivity was formed from the enantiomerically enriched product. In order to test this hypothesis, small amounts of trimethylsilyl-protected or un-protected mandelonitrile 20 (Figure 4.2) with 70% ee were added together with the catalyst (5 mol %) and the reagents at the beginning of the reaction. How-ever, this had no noticeable effect on the enantioselectivity (Table 4.4), as the ee obtained was the same as when no enantioenriched substrate was present at the beginning of the reaction (compare with entry 1 in Table 4.2).

Table 4.4. Influence of Addition of (S)-20 70% ee to the Reaction Mixture (Figure 4.2).a

entry added 20 (mol %) % yield of 20b % ee 1c 2.16 >95 60 (S) 2d 1.64 >95 60 (S)

a The initial concentration of benzaldehyde was 1 M. b Determined by GC. c Protected 20. d Unprotected

20.

46

A second hypothesis was formulated on the supposition that a more

enantioselective catalyst might form as a result of the change of concentration of benzaldehyde during the reaction. To test this hypothesis we performed the reaction by slowly adding the substrate, benzaldehyde, to the reaction mixture. Once more, the ee of the resulting product was identical with that of an ex-periment at which the substrate was added at once.

It was observed in similar systems that slow diffusion of water into the reac-tion mixture produced an increase of the ee of the product with time, due to the formation of a dimeric Ti complex with O bridges, that was more active in the reaction than the initial monomeric complex.9 We added small amounts of wa-ter to the reaction system, but the product 20 was then obtained with lower ee (30% ee with 0.5% water and 0% ee with 1.2% water).

At this point, we decided to study the complexes, which were involved in the reaction, by NMR spectroscopy to see whether the increase of ee with time was caused by slow formation of various catalytic species. As already discussed in chapter 2.3, no complex was formed upon addition of Ti(OiPr)4 to a solution of bispyridylamide 5a, as evidenced by NMR. On the other hand, addition of an excess of TMSCN to a solution of Ti(OiPr)4 and 5a afforded a mixture of com-plexes. These complexes were probably different aggregates with iso-propoxide ligands bridging different Ti centers. Indeed we found that Ti(OiPr)4 and TMSCN reacted to give a single Ti(IV)-CN complex, which was able to react with 5a to give the same mixture of complexes as when mixing all three reagents together. A Ti(III)-CN was prepared and characterized by Girolami et al10 and several groups proposed chiral cyanide sources in reactions forming asymmetric cyanohydrins.11

47

Scheme 4.1. Proposed mechanism for the asymmetric addition of TSMCN to benzaldehyde, catalyzed

by Ti-salen complexes.12

Belokon’ and North proposed, that a Ti-salen catalyst used in the addition of

TMSCN to benzaldehyde formed a monometallic Ti(IV)-CN complex, which later reacted with a metalloacetal, formed by reaction of the Ti-salen complex with benzaldehyde, to give a bimetallic Ti complex that activated both the nu-cleophile and the carbonyl substrate (Scheme 4.1).12 The NMR data, observed for the Ti(IV)-bispyridylamide 5a system, suggested the formation of similar bimetallic systems. We carried out various attempts to crystallize these bi-metallic complexes, in order to characterize them, but without success.

We also studied the reaction with other aldehydes than benzaldehyde (Figure 4.5) and found that longer reaction times were needed and the products ob-tained had lower ee:s than found with benzaldehyde (Table 4.5).

Figure 4.5. Addition of TMSCN to several aldehydes.

48

Table 4.5. Addition of TMSCN to Several Aromatic and Aliphatic Aldehydes.

entry R t (h) % yielda % ee 1 Ph 6 >95 70 2 4-MeOPh 16 >95 47 3 4-F3CPh 66 >95 11 4 2-BrPh 25 >95 24 5 C4H9 5 >95 41 6 tBu 5 >95 12

a Determined by GC.

The para-substituted benzaldehyde containing an electron-donating group showed higher reactivity and enantioselectivity in the reaction than the one having an electron-withdrawing group (compare entries 2 and 3). This was in agreement with previous results found for the asymmetric cyanation of these kinds of substrates catalyzed by Ti-salen complexes, but the reason for this was not clear.9 As for the ortho-substituted benzaldehyde (entry 4), the enantiose-lectivity was also lower than for benzaldehyde. Aliphatic aldehydes were cyanated slightly faster than benzaldehyde but the enantioselectivities of these reactions were low, especially if bulky substrates were used (entries 5 and 6). We employed acetophenone as a substrate in the reaction and obtained the product after 119 hours in >95% yield and with 8% ee.

We also employed other metal alkoxides than Ti(OiPr)4 together with bis-pyridylamide 5a in the asymmetric addition of cyanide to benzaldehyde (Figure 4.6, Table 4.6).

Figure 4.6. Asymmetric cyanation of benzaldehyde with other metal alkoxides than Ti(OiPr)4.

Table 4.6. Use of Other Metal Alkoxides than Ti(OiPr)4 in the Reaction.

M(OR)x mol % t (h) % yield of 20a % ee Zr(OtBu)4 1 23 78 3 Zr(OtBu)4 10 4 91 26 Sc(OiPr)3 1 3 >95 2 Yb(OiPr)3 1 16 >95 8 Cu(OAc)2 1 3 10 0

a Determined by GC.

49

All metals except Cu(II) were able to catalyze the reaction, but low or no asymmetric induction was observed. Despite these disappointing results, we decided to optimize the reaction conditions for the Zr complex (Figure 2.6). As reported previously, the addition of small amounts of a secondary amine to a catalyst formed from Zr(OtBu)4 and bispyridylamide 5a, afforded a more active catalytic system in the ring opening of cyclohexene oxide by means of trimethylsilyl azide.13 When we added pyrrolidine to the reaction mixture (en-try 3), no significant change in the reaction outcome was observed. However, addition of small amounts of water had a great effect on the outcome of the re-action (entries 4-6).

Table 4.7. Optimization of the Zr-catalyzed Asymmetric Cyanation of Benzaldehyde (Figure 4.6).

entry 5a (mol %)

Zr(OtBu)4 (mol %)

remarks t (h) % yield of 20a

% ee

1 1 1 - 23 78 3 2 10 10 - 4 91 26 3 10 10 1 mol % C4H8NH 5 >95 26 4 10 10 1 mol % H2O 3 >95 29 5 10 10 5 mol % H2O 2 >95 56 6 10 10 10 mol % H2O 2 >95 7

a Determined by GC.

When a small amount of water (5 mol %) was added to the reaction mixture, the catalyst formed became both more active and more selective than that without water (compare entries 1 and 2 with entry 5). If more water was added, the catalyst was still rather active but the product was nearly racemic (entry 6). We could observe by NMR that addition of water to the reaction mixture afforded a dimeric or oligomeric complex, probably containing O-bridges. These observations were similar to those reported by Kobayashi et al. for a Zr(IV) catalyst for asymmetric aldol reactions.14 No change of ee with time was observed in these reactions.

4.2.1 Use of a heterogeneous bispyridylamide as ligand

We used the solid-supported bispyridylamide 6c (Scheme 2.3) as a ligand in the asymmetric addition of TMSCN to benzaldehyde. When the metal used was Ti(OiPr)4, the reaction was slower than when the homogeneous ligand 5a was

50

used (16 hours compared to 4 hours), and the product was obtained in 90% yield with 12% ee. When Zr(OtBu)4 was used instead of Ti(OiPr)4, the reaction was even slower (24 hours) and the product obtained was racemic.

The reason for these disappointing results may be the fact that more active and selective dimeric complexes are more difficult to form when the ligand is attached to a solid support than when it is free in solution.

4.3 Ring opening of cyclohexene oxide The asymmetric ring opening of meso-epoxides has the potential to generate two adjacent stereogenic centers from an achiral starting material. Epoxides are readily available by oxidation of alkenes and it is well-known that their ring strain-induced reactivity can be enhanced by coordination to a Lewis acid. Chiral Lewis acids can be used to induce asymmetry in the reaction in this way. Various nucleophiles can be employed to ring-open epoxides, making these reactions quite versatile.15

Highly enantioselective catalytic ring opening of meso-epoxides was first re-ported by Nugent. He used a Zr(IV) trialkanolamine complex for the ring opening of meso-epoxides with iPrEt2SiN3.16

Jacobsen reported the use of chiral metal-salen complexes in the ring opening of epoxides and proposed a mechanism involving activation of both the nu-cleophile and the electrophile by two metal centers.17

Our group reported that a complex made in situ by mixing 5a (Figure 2.4) and Zr(OtBu)4 catalyzed the ring opening of cyclohexene oxide with trimethyl-silyl azide (TMSN3) in 45% yield and 56% ee after 4 days.13 Addition of di-ethylamine increased the reactivity and enantioselectivity of the catalyst and the product was obtained in a maximum yield of 60% and 71% ee.

It was shown some years ago that achiral18 and chiral additives19 could im-prove the rate and enantioselectivity of asymmetric catalytic reactions. We were interested in studying the effect that the addition of small amounts of chiral amines would have on the enantioselectivity of the ring opening of cyclohexene oxide with TMSN3, catalyzed by the 5a-Zr(OtBu)4 complex (Figure 4.7).

51

Figure 4.7. Effect of the additive in the ring opening of cyclohexene oxide with trimethylsilyl azide

catalyzed by the 5a-Zr(OtBu)4 complex.

We performed the reaction using standard conditions (0.5 mmol cyclohexene

oxide, 0.05 mmol Zr(OtBu)4, 0.05 mmol 5a, 0.53 mmol TMSN3 in 0.5 mL di-chloromethane, stirring at room temperature for 3 days) and seven chiral amines and one chiral alcohol (Table 4.8). The results suggested that for chiral diamines different diastereomeric complexes could be formed, as the product was obtained with different ee:s depending on the enantiomer of the chiral diamine that was used in the reaction. In the case of the chiral primary amine, the ee of the product was approximately the same for both enantiomers. When the chiral alcohol was used as an additive, the reaction afforded the racemic product in a very low yield.

Table 4.8. Screening of Chiral Additives in the Ring Opening Reaction (Figure 4.7).

additivea % yield of 21b % ee None 45 56

diiso-propylamine 82 62 (1R,2R)-1,2-diphenyl-1,2-diaminoethane 90 63 (1S,2S)-1,2-diphenyl-1,2-diaminoethane 50 35

(1R,2R)-1,2-diaminocyclohexane 30 22 (1S,2S)-1,2-diaminocyclohexane 30 48 (2R,5R)-2,5-dimethylpyrrolidine 90 46

(R)-phenylethylamine 90 54 (S)-phenylethylamine 90 56

(S)-2-butanol 10 0 a 0.01 mol % of the additive was used. b Determined by GC.

In view of the results obtained with (1R,2R)-1,2-diphenyl-1,2-diaminoethane,

we decided to vary the amount of the additive in order to see its effect on the enantioselectivity of the reaction (Table 4.9).

52

Table 4.9. Variation of the Amount of (1R,2R)-1,2-Diphenyl-1,2-diaminoethane used in the Ring

Opening Reaction (Figure 4.7).

(1R,2R)-1,2-diphenyl-1,2-diaminoethane (mol %)

% yield of 21a % ee

0.018 >90 56 0.01 >90 63

0.005 >95 64 0.0033 >90 58 0.0025 >90 64

a Determined by GC.

The product was obtained in good yields and with 56-64% ee after 4 days of

reaction at room temperature. No clear trend could be seen in the variation of the enantiomeric excess of the product.

We evaluated the electronic effects of the ligand by using bispyridylamide 5c (Figure 2.4) with MeO-substituted pyridine rings and 0.01 eq. of diiso-pro-pylamine. The product was obtained in 80% yield and 64% ee after 3 days of reaction time at room temperature. This result indicated that electron-donating groups in the 4-position of the pyridine ring of the ligand had very little effect on the outcome of the reaction.

The ring opening of cyclohexene oxide with TMSCN was also studied (Figure 4.8).

Figure 4.8. Ring opening of cyclohexene oxide with TMSCN catalyzed by 5a and metal alkoxides.

The complex 5a-Zr(OtBu)4 afforded the racemic product in 83% yield after 2

days at room temperature. When a small amount of water was added (0.5 eq. with respect to the ligand), the reaction was completed after 15 hours, but the product was nearly racemic (2% ee). The reaction was rather slow when 5a-Ti(OiPr)4 was used instead (2% conversion after 18 hours). Addition of small amounts of FeCl3 to the reaction mixture increased the rate of the reaction (46% yield after 48 hours) and the ee of the product (12% ee after 48 hours).

53

References 1. Lewis Acids in Organic Synthesis. Yamamoto, H. Ed. Wiley-VCH 2000. 2. Kobayashi, S.; Busujima, T.; Nagayama, S. A Novel Classification of Lewis Acids on the Basis of Activity and Selectivity. Chem. Eur. J. 2000, 6, 3491-3494. 3. Pearson, R. G. Hard and Soft Acids and Bases. J. Am. Chem. Soc. 1963, 85, 3533-3539. 4. Mori, A. and Inoue, S. in Comprehensive Asymmetric Catalysis; Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H., Eds.; Springer: Berlin, 1999; pp 983-994. 5. For a recent review see: North, M. Synthesis and Applications of Non-racemic Cyanohydrins. Tetrahedron: Asymmetry 2003, 14, 147-176. 6. Belokon’, Y.; Ikonnikov, N.; Moscalenko, M.; North, M.; Orlova, S.; Tararov, V.; Yashkina, L.; Asymmetric Trimethylsilylcyanation of Aldehydes Catalyzed by Chiral (salen)Ti(IV) Complexes Tetrahedron: Asymmetry 1996, 7, 851-855. 7. Belokon’, Y. N.; Gutnov, A. V.; Maleev, V. I.; Moskalenko, M. A.; Yashkina, L. V.; Lesovoy, D. E.; Ikonnikov, N. S.; Larichev, V. S.; North, M. Catalytic Asymmetric Synthesis of O-acetyl Cyanohydrins from KCN, Ac2O and Aldehydes. Chem. Commun. 2002, 244-245. 8. We realized that the quality of the TMSCN used was very important for the outcome of the reaction. When dark-coloured TMSCN from Lancaster was used the reaction times were longer, the products were obtained in low ee:s (10-46% ee) and the reproducibility was low. On the other hand, when we used TMSCN from Acros or Aldrich, the reproducibility was high whereas the ee:s of the products varied slightly between two different batches of the reagent supplied by Acros (batch 1 afforded the product in 70% ee and batch 2 in 60% ee un-der the best conditions found). No significant differences were observed in the 1H NMR of the reagent from the different suppliers. The reagent from Lancaster and batch 1 from Acros were both dark-colored liquids, whereas the reagent from batch 2 Acros and Aldrich were colorless liquids. Since Fe-CN complexes are dark-colored, we decided to analyze by ICP whether any Fe was present in the various reagents and found that the reagents did not contain any Fe above 0.08 ppm. We decided to use TMSCN from Aldrich kept under N2, and the experi-ments were run several times in order to verify their reproducibility. 9. Belokon', Y. N.; Caveda-Cepas, S.; Green, B; Ikonnikov, N. S.; Khrustalev, V. N.; Larichev, V. S.; Moscalenko, M. A.; North, M.; Orizu, C.; Tararov, V. I.; Tasinazzo, M.; Timofeeva, G. I.; Yashkina, L.V. The Asymmetric Addition of Trimethylsilyl Cyanide to Al-dehydes Catalyzed by Chiral (Salen)Titanium Complexes. J. Am. Chem. Soc. 1999, 121, 3968-3973. 10. Entley, W. R.; Treadway, C. R.; Wilson, S. R.; Girolami, G. S. The Hexacyanotitanate Ion: Synthesis and Crystal Structure of [NEt4]3[TiIII(CN)6].4MeCN. J. Am. Chem. Soc. 1997, 119, 6251-6258.

54

11. (a) Minamikawa, H.; Hayakawa, S.; Yamada, T.; Iwasawa, N.; Narasaka, K. Asymmetric Hydrocyanation of Aldehydes Using Chiral Titanium Reagents. Bull. Chem. Soc. Jpn. 1988, 61, 4379-4383. (b) Narasaka, K.; Yamada, T.; Minamikawa, H. The Asymmetric Hydrocyanation of Aldehydes with Cyanotrimethylsilane Promoted by a Chiral Titanium Reagent. Chem. Lett. 1987, 2073-2076. 12. Belokon , Y. N.; Green, B.; Ikonnikov, N. S.; Larichev, V. S.; Lokshin, B. V.; Moscalenko, M. A.; North, M.; Orizu, C.; Peregudov, A. S.; Timofeeva, G. I. Mechanistic Investigation of the Asymmetric Addition of Trimethylsilyl Cyanide to Aldehydes Catalysed by Dinuclear Chiral (Salen)titanium Complexes. Eur. J. Org. Chem. 2000, 2655-2661. 13. Adolfsson, H.; Moberg, C. Chiral Lewis Acid Catalysed Asymmetric Nucleophilic Ring Opening of Cyclohexene Oxide. Tetrahedron: Asymmetry 1995, 6, 2023-2031. 14. Yamashita, Y.; Ishitani, H.; Shimizu, H.; Kobayashi, S. Highly anti-Selective Asymmetric Aldol Reactions Using Chiral Zirconium Catalysts. Improvement of Activities, Structure of the Novel Zirconium Complexes, and Effect of a Small Amount of Water for the Preparation of the Catalysts. J. Am. Chem. Soc. 2002, 124, 3292-3302. 15. Jacobsen, E. N.; and Wu, M. H. in Comprehensive Asymmetric Catalysis; Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H., Eds.; Springer: Berlin, 1999; pp 1309-1326. 16. Nugent, W. A. Chiral Lewis Acid Catalysis. Enantioselective Addition of Azide to Meso Epoxides. J. Am. Chem. Soc. 1992, 114, 2768-2769. 17. Jacobsen, E. N. Asymmetric Catalysis of Epoxide Ring-Opening Reactions. Acc. Chem. Res. 2000, 33, 421-431. 18. Vogl, E. M.; Gröger, H.; Shibasaki, M. Towards Perfect Asymmetric Catalysis: Additives and Cocatalysts. Angew. Chem., Int. Ed. 1999, 38, 1570-1577. 19. Ohkuma, T.; Doucet, H.; Pham, T.; Mikami, K.; Korenaga, T.; Terada, M.; Noyori, R. Asymmetric Activation of Racemic Ruthenium(II) Complexes for Enantioselective Hydro-genation. J. Am. Chem. Soc. 1998, 120, 1086-1087.

55

5 Concluding remarks

Asymmetric catalysis is a powerful synthetic tool and in the future a number of asymmetric transformations, based on metal catalysis, will become the standard for industrial production of fine and specialty chemicals. Due to the specificity of a catalyst towards a substrate or substrate family, research on new ligand families, that allow the preparation of selective catalysts for other kinds of substrates, becomes necessary. Today, the discovery of highly selective homo-geneous catalysts is still too much based on trial and error. However, a deeper understanding of the catalytic intermediates leads to semi-rational design of catalysts for specific substrates in a reaction. Ligands which allow an easy variation of their electronic and steric properties are, therefore, interesting. We have demonstrated that bispyridylamides can easily be constructed with different chiral backbones, and substituents are simply introduced by functional group interconversion of halo-substituted derivatives. Furthermore, we have also shown that the attachment of bispyridylamide to solid supports is possible, offering the possibility of easy separation from the reaction mixture and reuse with a significant degree of robustness.

Bispyridylamides were excellent as ligands in the microwave-accelerated Mo-catalyzed asymmetric allylic alkylation reaction, and by introducing elec-tron-donating groups in the pyridine rings, catalysts displaying higher regiose-lectivity for the branched product were obtained. That reaction was used to prepare (R)-baclofen, a commercial GABA derivative used as a muscle relaxant. A solid-supported bispyridylamide could be employed in this reaction and the product was still obtained in high yield and with high regio- and enan-tioselectivity.

Chiral Lewis acid complexes, especially those of Ti(IV) and Zr(IV), were employed in the addition of cyanide to aldehydes. The asymmetry induced by

56

these catalysts was generally lower for aldehydes other than benzaldehyde. The electronic effects of the ligand were not as important in this reaction as in Mo-catalyzed asymmetric allylic alkylations. However, the chiral backbone used was decisive, and a more selective catalyst was obtained when bispyridyla-mides derived from (1R,2R)-1,2-diamino-1,2-diphenylethane were used as ligands in the reaction. Moreover, interesting mechanistic insight into this reaction was obtained by studying the complexes involved by NMR spectros-copy. These studies suggested the formation of bimetallic dimeric structures activating both the electrophile, benzaldehyde, and the nucleophile, cyanide, when Ti(IV) was used. For Zr(IV), an initial monomeric complex with a bis-pyridylamide ligand could be isolated and its structure was determined by X-ray diffraction. Addition of a small amount of water promoted the formation of a more reactive and selective catalyst than the monomeric species. When a solid-supported ligand was used in the addition of cyanide to benzaldehyde, the reaction was slow and the product obtained displayed very low enantiomeric excess. This observation constituted an additional piece of evidence for the fact that a dimeric complex is a more selective catalyst than a monomeric one in these reactions. Further studies in order to elucidate the structure of the metal complexes involved in the reaction should be made.

The asymmetric ring opening of cyclohexene oxide with azide, catalyzed by Zr(IV) bispyridylamide complexes was further investigated. Addition of chiral additives showed that diamines gave rise to the formation of diastereomers, displaying different reactivities and inducing different degrees of asymmetry. When cyanide was used instead of azide as a nucleophile in the reaction, the enantiomeric excess of the product was low.

Because of the easy variation of the chiral structure and the electronic properties of bispyridyamides, their use as ligands in other catalytic reactions is very interesting. These easy modification of the ligands should also be used in order to find a catalyst able to perform Mo-catalyzed allylic alkylations using other nucleophiles than dimethylmalonate.

57

Appendix This appendix contains experimental details and spectroscopical data of com-pounds 1g, 4b, 4h, 5b-d, 6a-c. General: all solvents were dried using standard techniques. Unless otherwise noted, all 1H and 13C NMR spectra were recorded in CDCl3 on a 500 or 400 MHz instrument. Chemical shifts are reported in ppm relative to the solvent as internal standard. (1R,2R)-1,2-Bis[(4-cyanopyridine)-2-carboxyamido]cyclohexane (1g). A suspension of (1R,2R)-1,2-Bis[(4-chloropyridine)-2-carboxyamido]cyclohex-ane (1e) (128 mg, 0.33 mmol), sodium cyanide (65 mg, 1.33 mmol) and anhydrous nickel bromide (145 mg, 0.66 mmol) in 1-methyl-2-pyrrolidinone (1 mL) was heated to 200 ºC in a microwave cavity for 40 min. The resulting dark green mixture was poured into water and extracted with CH2Cl2. The combined organic extracts were dried (Na2SO4) and the solvent was evaporated to give crude 1g that was purified by flash cromatography on silica gel (eluent: hexanes:EtOAc, 1:1) to give 52 mg (42%) of the pure product; 1H NMR δ 8.73 (1H, d, J = 4.9 Hz), 8.27 (1H, s), 8.13 (1H, d, J = 6.6 Hz), 7.60 (1H, d, J = 4.9 Hz), 4.04, (1H, b), 2.19 (1H, b), 1.86 (1H, b), 1.46 (2H, b); 13C NMR δ 163.0, 151.4, 149.6, 127.9, 124.5, 122.3, 116.3, 54.0, 32.9, 25.1. (1R,2R)-1-[(4-chloropyridine)-2-carboxyamido]-2-[(4-nitropyridine)-2-carboxyamido]cyclohexane (4b). A suspension of 4-chloropicolinic acid (519 mg, 2.03 mmol), K2CO3 (282 mg, 2.04 mmol) and N,N’-carbonyldiimidazole (329 mg, 2.03 mmol) in THF (5 mL) was stirred for 2 h at 50 oC under nitro-gen. The mixture was diluted with THF (35 mL) and added dropwise to a solu-tion of (1R,2R)-1,2-diaminocyclohexane (232 mg, 2.03 mmol) in THF (40 mL). The reaction mixture was stirred for an additional 1 h and the solvent was removed by rotatory evaporation. The solid monoamide resulting was dissolved in CH2Cl2 (10 mL) and added to a suspension of 4-nitropicolinic acid (341 mg, 2.03 mmol), N,N’-carbonyl-diimidazole (329 mg, 2.03 mmol), which had been stirred in THF:CH2Cl2 (5:1, 6 mL) in a separate flask for 6 h at 50 oC under ni-trogen. The suspension was stirred for 18 h at room temperature and an aque-

58

ous solution of saturated K2CO3 (20 mL) was then added. The mixture result-ing was extracted with CH2Cl2 (3 × 10 mL). The combined organic extracts were dried (Na2SO4) and the solvent evaporated. The crude product was puri-fied by column chromatography on silica gel (eluent: hexane:EtOAc 1:1). Yield: 215 mg (26%) of 4b; 1H NMR δ 8.84 (1H, d, J = 5.29 Hz), 8.73 (1H, d, J = 2.0 Hz), 8.41 (1H, d, J = 5.0 Hz), 8.20 (1H, d, J = 7.3 Hz), 8.12 (1H, d, J = 7.3), 8.08 (1H, dd, J = 5.3, 2.0 Hz), 8.03 (1H, d, J = 2.0 Hz), 7.35 (1H, dd, J = 5.3, 2.0 Hz), 4.05 (2H, b), 2.18 (2H, b), 1.86 (2H, b), 1.45 (4H, b); 13C NMR δ 163.9, 162.9, 162.8, 155.4, 153.5, 153.3, 151.4, 151.0, 150.9, 149.5, 146.1, 126.7, 126.6, 123.2, 118.9, 118.8, 115.6, 54.5, 54.2, 53.9, 53.6, 32.9, 25.2, 25.1. Silica-supported ligand (4h). Carboxylic acid functionalized Kromasil silica gel was prepared by shaking a suspension of NH2-Kromasil silica phase (1g, 1.29 mmol NH2) and succinic anhydride (1.29 g, 12.9 mmol) in DMF (3 mL) at room temperature for 64 h. After washing (CH2Cl2, MeOH, acetone) the re-sulting silica was dried under vacuum to give 1.09 g (97% yield, according to elemental analysis: 1.11 mmol carboxyl groups/g). Then a suspension of the carboxylic acid functionalized Kromasil silica gel (216 mg, 0.24 mmol), (1R,2R)-1-{[N-(1,2-diaminoethyl)pyridine]-2-caboxamido}-2-(pyridine-2-car-boxamido)cyclohexane (4e), DCC (58 mg, 0.281 mmol) and 4-dimethylamino-pyridine (3.4 mg, 0.028 mmol) in CH2Cl2 (2 mL) was refluxed under nitrogen for 4 days. The resulting solid was filtered, washed (CH2Cl2, MeOH, CH2Cl2) and dried under vacuum to give 137 mg (56% yield, according to elemental analysis: 0.043 mmol ligand/g). (1R,2R)-Bis-[(4-chloropyridine)-2-carboxyamido]-1,2-diphenylethane (5b). A suspension of 4-chloropyridine-2-carboxylic acid mixed salt (1.0 g, 5.15 mmol), 1,1’-carbonyldiimidazol (845 mg, 5.2 mmol) and potassium carbonate (720 mg, 5.2 mmol) in THF (5 mL) was heated at 50 ºC under N2 for 1 h. Then (1R,2R)-1,2-diphenyl-1,2-diaminoethane (547 mg, 2.6 mmol) was added at once and the reaction mixture was stirred at 50 ºC for a further 1 h. Water (20 mL) was added to the cold reaction mixture and the resulting suspension was extracted with CH2Cl2 (3x10 mL). The combined organic extracts were dried (Na2SO4) and the solvent was evaporated. The crude product was purified by recrystallization from EtOH to yield 445 mg (34%) of the pure product; 1H NMR δ 8.84 (1H, d, J = 5.9 Hz), 8.46 (1H, d, J = 5.1 Hz), 8.10 (1H, d, J = 2.2

59

Hz), 7.40 (1H, dd, J = 5.1, 2.2 Hz), 7.30-7.26 (5H, m), 5.61 (1H, d, J = 5.9 Hz); 13C NMR δ 163.8, 151.4, 149.6, 146.0, 138.8, 129.0, 128.4, 128.1, 126.8, 123.2, 59.5. (1R,2R)-Bis-[(4-methoxypyridine)-2-carboxyamido]-1,2-diphenylethane (5c). A suspension of 5b (100 mg, 0.20 mmol), and solid sodium methoxide (54 mg, 1 mmol) in MeOH (1 mL) was heated to 140 ºC for 30 min in a mi-crowave cavity. The suspension resulting was poured into water and extracted with CH2Cl2. The combined organic extracts were dried (Na2SO4) and the sol-vent was evaporated to give 98 mg (>99%) of the title compound; 1H NMR δ 8.97 (1H, d, J = 5.5 Hz ), 8.38 (1H, d, J = 5.5 Hz), 7.69 (1H, d, J = 2.7 Hz), 7.37-7.21 (5H, m), 6.90 (1H, dd, J = 5.5, 2.7 Hz), 5.64 (1H, d, J = 5.5 Hz), 3.91 (3H, s). (1R,2R)-Bis-[(4-N-pyrrolidylpyridine)-2-carboxyamido]-1,2-diphenylethane (5d). A solution of 5b (132 mg, 0.27 mmol) in pyrrolidine (2 mL) was heated to 160 ºC for 15 min in a microwave cavity. The resulting mixture was concentrated in vacuo and the brownish solid was extracted with CH2Cl2 and water. The combined organic extracts were dried (Na2SO4) and the solvent was evaporated to give 151 mg (>99%) of the title compound; 1H NMR δ 8.96 (1H, d, J = 6.0 Hz), 8.10 (1H, d, J = 6.0 Hz), 7.25-7.12 (5H, m), 6.35 (1H, dd, J = 6.0, 2.7 Hz), 5.58 (1H, d, J = 6.0 Hz), 3.36-3.25 (4H, m), 2.04-1.95 (4H, m); 13C NMR δ 165.8, 152.9, 150.2, 148.6, 139.6, 128.7, 128.2, 127.8, 108.6, 106.0, 58.8, 47.5, 25.7. (1R,2R)-1-[(4-chloropyridine)-2-carboxyamido]-2-(pyridine-2-carboxyamido)-1,2-diphenylethane (6a). A solution of picolinic acid (250 mg, 2.03 mmol) and N,N’-carbonyldiimidazole (329 mg, 2.03 mmol) in THF (5 mL) was stirred for 30 min at 50 oC under nitrogen. The mixture was diluted with THF (15 mL) and CH2Cl2 (20 mL) and added dropwise to a solution of (1R,2R)-1,2-diphenyl-1,2-diaminoethane (431 mg, 2.03 mmol) in THF (40 mL). The reaction mixture was stirred for an additional 1 h and the solvent was removed by rotatory evaporation. The resulting solid monoamide was dissolved in CH2Cl2 (10 mL) and added to a suspension of 4-chloropicolinic acid (519 mg, 2.03 mmol), K2CO3 (282 mg, 2.03 mmol) and N,N’-carbonyldiimidazole (329 mg, 2.03 mmol), which had been stirred in THF (5 mL) in a separate flask for 2 h at 50 oC under nitrogen. The suspension was then stirred for 16 h at room temperature and then an aqueous solution of saturated K2CO3 (20 mL)

60

was added. The resulting mixture was extracted with CHCl3 (3 × 10 mL). The combined organic extracts were dried (Na2SO4) and the solvent evaporated. The crude product was purified by column chromatography on silica gel (eluent: hexane:EtOAc 1:1). Yield: 321 mg (35%) of 6a; 1H NMR δ 8.93 (2H, b), 8.59 (1H, d, J = 4.5 Hz), 8.48 (1H, d, J = 4.5 Hz), 8.19-8.09 (2H, m), 7.84-7.77 (1H, m), 7.47-7.38 (12H, m), 5.67-5.64 (2H, m); 13C NMR δ 165.0, 163.8, 151.5, 149.9, 149.5, 148.6, 146.0, 139.0, 138.9, 137.5, 129.0, 128.9, 128.3, 128.2, 128.1, 126.7, 126.6, 123.2, 122.6, 59.6, 59.2. (1R,2R)-1-{[N-(1,2-Diaminoethyl)pyridine]-2-carboxyamido}-2-(pyridine-2-carboxyamido)-1,2-diphenylethane (6b). A solution of 6a (64.5 mg, 0.141 mmol) in 1,2-diaminoethane (2 mL) was heated to 180 ºC for 20 min in a mi-crowave cavity. The mixture resulting was washed with saturated K2CO3 and extracted with CH2Cl2. The combined organic extracts were dried (Na2SO4) and the solvent was evaporated to give 64 mg (94%) of the title compound; 1H NMR δ 8.99 (1H, d, J = 4.4 Hz), 8.93 (1H, d, J = 4.4 Hz), 8.56 (1H, d, J = 4.0 Hz), 8.12 (1H, d, J = 7.7 Hz), 8.09 (1H, d, J = 5.9 Hz), 7.79-7.73 (1H, m), 7.40-7.08 (12H, m), 6.47 (1H, dd, J = 5.5, 2.2 Hz), 5.61 (2H, d, J = 4.4), 4.90 (1H, b), 3.27-3.15 (2H, m), 2.99-2.91 (2H, m); 13C NMR δ 165.7, 164.8, 154.8, 150.5, 150.1, 148.9, 148.6, 139.4, 139.2, 137.5, 128.9, 128.8, 128.2, 128.2, 128.1, 128.0, 126.4, 122.7, 109.7, 106.5, 59.2, 58.9, 45.2, 41.0.

61

Acknowledgements I would like to express my gratitude to the following: Prof. Christina Moberg; for taking me as a PhD student at KTH. For your guidance, enthusiasm and trust. All my coworkers, especially my diploma workers Sophie Duquesne and Stina Lundgren; for doing a great job. All present and former members of Ki’s group; for the friendly environment and fruitful discussions. All present and former members of the Organic Chemistry Department at KTH; for the stimulating working atmosphere and valuable help. All members of the SELCHEM program; for all the inspiring meetings. Gunhild Aulin-Erdtman; for teaching me proper English while preparing this thesis. The Swedish Foundation for Strategic Research (SSF); for financial support. The Aulin-Erdtman Foundation; for making possible to present this work into several conferences. Personal Chemistry; for the microwave cavity.