Desymmetrisation of Prochiral Diols for Mechanically Chiral Rotaxane Synthesis

(1)

MSc Chemistry

Molecular Science

Literature Thesis

Desymmetrisation of Prochiral Diols for Mechanically

Chiral Rotaxane Synthesis

by

Koen van de Vrande

11022787

November 2019

12 EC

5 September 2019 – 19 November 2019

Supervisor/Examiner:

S. Pilon MSc

Examiner:

Prof. Dr. J.H. van Maarseveen

Prof. Dr. J.N.H. Reek

Van ‘t Hoff Institute for Molecular Science

Synthetic Organic Chemistry

(2)

(3)

3

Abstract

Due to the recent discovery of lasso peptides, the interest in the synthesis of mechanically chiral rotaxanes has risen. A new synthetic strategy has been proposed by the group of Van Maarseveen, where a prochiral rotaxane ring can be synthesised with the use of the copper(I)-catalysed alkyne-azide cycloaddition reaction. In order to complete the synthesis of a mechanically chiral rotaxane, desymmetrisation of a diol is a key step. This report describes several different ways to desymmetrise diols, using non-catalytic reactions, biocatalysis, metal catalysis and organocatalysis. In order to investigate which method was the most selective and effective, all methods were compared based on yield, ee, substrate scope, availability of multiple enantiomers and the overall applicability in the synthesis of mechanically chiral rotaxanes. It was found that the substrate scope of both non-catalytic and metal catalysed methods was limited to 1,2- and 1,3-meso-diols. The biocatalytic hydrolysis had the highest yield and ee, but due to the limited size of the active site of the enzyme, an organocatalytic method by Kündig et al. was concluded to be the most applicable for the synthesis of mechanically chiral rotaxanes.

(4)

4

1. Introduction

The interest in mechanically interlocked molecules (MIMs), such as rotaxanes (Figure 1, a) and catenanes (Figure 1, b) has risen drastically over the past few years.1_{A rotaxane is a molecule consisting}

of two parts: a macrocyclic ring, which loops around a linear axle. Sterically bulky stoppers are present on the ends of the axle, preventing the ring from slipping off. A catenane is made up by two cyclic molecules which loop through each other. The interesting structures and properties of these molecules as well as the synthetic challenge has led to the Nobel prize in Chemistry of 2016 being awarded for the application of these kinds of systems.2_{MIMs have promising properties for the use in catalysis}3

and as host for molecule and ion storage.1,4_{Partially because of the discovery of natural asymmetric}

MIMs, the so-called lasso peptides, in 2003 by Ebright et al. the focus in the field of MIM synthesis has shifted towards making asymmetric MIMs as model compounds for these peptides.5_{The advantage of}

chiral MIMs is that they could be used for asymmetric catalysis or storage of specific enantiomers of chiral molecules.4_{The main difference between MIMs and simpler, not interlocked molecules is that}

MIMs can have two different categories of chirality: classical chirality and mechanical chirality, which will be discussed in the following paragraph.

Figure 1: Schematic representation of a rotaxane (a) and a catenane (b).

1.1 Mechanical Chirality

The term ‘Chirality’ was first mentioned in a lecture by Lord Kelvin in 1893.6_{In this lecture, the term}

referred to the situation in which the mirror image of an object could not be superimposed on the original. The term originates from the Greek word for hand, χειρ (Cheir), as the most well-known example of chirality are your hands. Examples of chirality are found in every sub-discipline of chemistry and are most important in biologically active systems.6–8_{Whilst classical chirality has been known for}

over 200 years, mechanical chirality can only be found in MIMs and has thus far only recently been getting interest.4_{Mechanical chirality can appear in both rotaxanes and catenanes and it can originate}

from multiple sources.4,9_{For rotaxanes with only one ring, there are two possible forms of mechanical}

chirality: oriental chirality which arises from the use of an asymmetric ring, or mechanically planar chirality, which originates from directionality in the ring. An interesting property of mechanical chirality is that it can arise in MIMs whose components do not have to be chiral by themselves.

A mechanically planar chiral rotaxane is obtained when there is some form of direction in the ring, such as amide or ester bonds, and two different stoppers are used, as is shown in Figure 2. When these rotaxanes are mirrored, the mirror image cannot be superimposed onto the original because the direction in the ring will be inverted. Enantioselective synthesis of mechanically chiral rotaxanes has only been published recently by the group of S.M. Goldup in 2018.10_{Before this, mechanically chiral}

MIMs were mostly synthesised as racemates or at a very low ee value of less than 10%. The discovery of natural mechanically chiral MIMs, the lasso peptides, has led to increased interest in the synthesis of chiral MIMs as a model for these fascinating natural compounds.5_{Because these chiral MIMs give}

an interesting synthetic challenge and they have the potential for asymmetric catalytical properties, new synthetic strategies are being developed.4

(6)

6

Figure 2: Schematic representation of a mechanically chiral rotaxane and its mirror image. The arrows represent the direction in the ring, which normally consists of direction in chemical bonds such as amides.

1.2 Desymmetrisation of Prochiral Diols

One example of a new strategy to synthesise mechanically chiral rotaxanes has been proposed by Van Maarseveen et al. In their proposal, which is shown in Scheme 1, they proposed to first synthesise prochiral rotaxanes with a direction in the ring arising from the copper(I)-catalysed alkyne-azide cycloaddition (CuAAC) used for the ring closing reaction.11_{This reaction introduces direction in the}

rotaxane ring, as one side of the ring will be connected to the nitrogen atom and the other side to carbon. After the ring closure, pro-chiral pre[2]rotaxane 1 is obtained, which can be further functionalised to give a rotaxane by adding the stoppers onto the hydroxy groups of the central aromatic system and cleaving the ester bonds connecting this aromatic centre part to the ring. If two different stoppers would be added to 1, a mechanically chiral rotaxane is formed. The challenging part is, however, that the two aromatic hydroxy groups are chemically equivalent due to symmetry of 1. Therefore this aromatic diol has to be desymmetrised for complete enantioselective synthesis of a mechanically chiral rotaxane. The exact reaction step to obtain the desymmetrised diol has yet to be determined.

Scheme 1: Proposed synthetic pathway for the synthesis of a mechanically chiral rotaxane. Courtesy of Prof. Dr. J.H. van Maarseveen for providing this scheme.

(7)

7 The aim of this report is to give an overview of the different known methods to desymmetrise prochiral diols. These reactions will be compared by overall yield, enantiomeric excess (ee) and substrate scope to find out what the most selective and effective method to enantioselectively desymmetrise meso-diols is. The different methods will be ordered firstly by the type of catalyst used and secondly by the functional group that the hydroxy is reacted to. First, an overview of older, non-catalytic methods will be given, followed by bionon-catalytic approaches. The section after that will discuss the different metal catalysed desymmetrisations, and finally the organocatalytic pathways will be reviewed. In the last section, all of the discussed reactions will be compared and an overall conclusion will be given. Eventually this report should give an overview of which methods Van Maarseveen et al. could use to finish their synthesis of mechanically chiral rotaxanes.

2. Non-Catalytic Desymmetrisation

2.1 Non-Catalytic Acylation

In the 1990s it was already concluded that the desymmetrisation of meso-diols was a useful method to obtain chiral building blocks.12,13_{At this time, the most common way to produce enantiomerically}

pure compounds was kinetic resolution.14_{To obtain kinetic resolution, a racemic mixture is reacted}

with a chiral reagent. If one of the enantiomers of the starting product has a faster reaction rate in this reaction, the mixture can be kinetically resolved, as this enantiomer will be completely converted before the other enantiomer, and two separable substances will be formed. The main disadvantage of kinetic resolution is that the maximum theoretical yield is 50%, as half of the starting material is effectively reacted away. Desymmetrisation of prochiral compounds, on the other hand, has a maximum theoretical yield of 100%, since all of the starting material can be converted to the preferred enantiomer. In the early days of desymmetrisation, there were two main methods: biocatalysis or chemical methods. The advantage of chemical methods over biocatalysis is that enzymes have a very specific substrate scope and development of a biocatalytic pathway required a long time of enzyme screening, and enzymes are commonly able to catalyse the formation of only one enantiomer.12,15_In

order to avoid these disadvantages, research was performed onto chemical desymmetrisation. The first chemical methods, which are described in this chapter, did not use any catalyst in their reaction, but instead relied upon cleavage of acetals.

The first of these methods to be discussed was published by the group of Iwata in 1995.15_This

reactive pathway is shown in Scheme 2. In their desymmetrisation of 1,2-cyclopentadiol (2) the diol is first reacted with the sulphur ketone 3 to form chiral acetal 4. Acetal 4 is then oxidised using the Davis’ reagent DR1 to obtain 5 in both stereoisomers, 5a and 5b, which were separated. After treatment with lithium diisopropylamide (LDA), a strong base that kinetically deprotonates organic compounds,16_the

acetal bond would break, and an olefin was formed, giving 6a and 6b. The hydroxy groups of 6 were then protected with (+)-Mosher’s acid chloride so the absolute stereochemistry could later be determined using NMR,17_{then the substance underwent ozonolysis and finally it was desulphurised}

with Raney nickel, a well-known desulphurisation catalyst.18_{This reaction would give the}

desymmetrised diol in both stereoisomers 7a and 7b, with an overall yield of 35% for 7a and 48 % for

7b over six steps. The Mosher’s acid protection step can be left out when the reaction is used, as this

was purely done for determination of the absolute stereochemistry. The advantage of this method is that it is possible to isolate both enantiomers of 7, however six steps are needed, using many expensive and aggressive chemicals and it is not possible to selectively synthesise only one enantiomer. In a later publication, Iwata et al. were able to synthesise only 7a with an ee of 78% by using a non-cyclic sulphur compound instead of 3.19

(8)

8

Scheme 2: The desymmetrisation of 1,2-meso-cyclopentadiol following the method of Iwata et al. a) TMSOTf (0.1eq), CH2Cl2,

rt.; b) DR1, CH2Cl2, rt.; c) LDA (3eq), THF, -78°C; d) (+)-MTPACl, DMAP, Et3N, CH2Cl2, -50°C, then O3, MeOH, -78°C, then Raney

Ni(W2), EtOH, rt.

2.2 Non-Catalytic Etherification

Two other acetal cleavage desymmetrisation pathways were published by the group of Harada a few years later.20,21_{In both of their pathways, the hydroxy groups were not acylated as in the method by}

Iwata et al.15 _{but were arylated}20_{or alkylated.}21 _{Since both arylation and alkylation of a hydroxy group}

lead to an ether being formed, both reactions will be put under the umbrella term of ‘etherification’ in this report.

The first of the two methods by Harada et al. was published in 1997.20_{The reaction, which is shown}

in Scheme 3, starts with the formation of acetal 10 from diol 8 and diether 9. Chiral Lewis acid 14 is then added, and because it is chiral, the Lewis acid will preferentially bind to the pro-R hydroxy group, activating this side for the reaction with olefin 15. The hydroxy group of 11 is then deprotonated using the strong, non-nucleophilic base potassium bis(trimethylsilyl)amide (KHMDS) and benzylated with benzyl bromide, which leads to 12.20,22_{In the final step, cleaving of the carbon-oxygen bond with TFA}

yields desymmetrised diol ether 13 in 74% yield and 85% ee. Whilst only cyclohexadiol 8 is shown in Scheme 3, the same reaction can be done on a variety of cyclic and acyclic 1,2-meso-diols, with the highest yield of 81% being obtained for 2,3-butadiol. The key step in this reaction is the use of chiral Lewis acid 14, which coordinates to only one of the oxygen atoms in acetal 10, activating this side for the reaction and forming 11 in 85% ee. The main disadvantage of this reaction is that only syn-10 is reactive, whilst the anti-enantiomer of 10 is also formed. Another disadvantage was the use of a strong acid in the final step, which makes this reaction not compatible with substrates that have acid-sensitive protective groups.

(9)

9

Scheme 3: First acetal cleavage desymmetrisation of Harada et al. a) TsOH (cat), THF; b) Cat. 14, Me2C=C(OEt)(OTMS) (15),

CH2Cl2, -78°C, 15h; c) KHMDS, BnBr, THF, 0°C; d) TFA, 0°C, then rt. 3h, then NaOH (aq), MeOH, 60°C, 30 min.

In order to make their reaction more applicable, Harada et al. published a revised strategy a year later, in which no strong acids are required (Scheme 4).21 _{In this reaction, acetal 10Me is formed from}

1,2-meso-diol 8Me with 9 in the same way as how 10 was formed in the previous reaction.20,21_The

pro-R hydroxy group is once again activated by coordination of chiral Lewis acid 14, and after reaction with

18, 16 is formed in 97% yield with 96% ee. The free hydroxy group on 16 is then protected by formation

of a MEM ether, after which the other hydroxy group is restored by addition of the kinetic base LDA.16,21

The desymmetrised ether 17 was obtained in 74% isolated yield and 96% ee. This reaction was also performed on other diols where the methyl groups were substituted by different R groups, however no cyclic diols were used in the publication. However, since the reaction is quite similar to the reaction given in Scheme 3, it would most likely work for cyclic 1,2-meso-diols as well. For the substrate with benzyl ethers instead of methyl groups, Harada et al. were also able to convert it to a chiral hydroxyamine (Scheme 5). In this side reaction, the chiral, opened acetal 16* is first converted into its triflate and then azide 19 is formed after reaction with tetramethyl guanidium azide. Azide 19 can then be hydrogenated with palladium black as catalyst to yield amine 20 in 75% isolated yield and 96% ee.

Scheme 4: Second acetal cleavage desymmetrisation by Harada et al. a) 9, TsOH (cat), CH2Cl2; b) Cat. 14, 18, CH2Cl2, -78°C; c)

(10)

10

Scheme 5: Amination side reaction of 16* by Harada et al. a) Tf2O, pyridine, then tetramethyl guanidium azide; b) Pd black, H2.

The ee value of the second reaction of Harada et al. (Scheme 4) is higher than for their first reaction (Scheme 3), with the second reaction having an ee of 96% and the first one 89% for the same substrate, while the isolated yields of both reactions are similar.20,21_{Both reactions also require a similar amount}

of reaction steps, with five and four steps for the first and second reaction, respectively. The MEM ether in 17 can be exchanged for the benzyl ether in two additional steps, leading to both reaction pathways being able to yield the same product. The main difference between the two pathways is the removal of the alkyne ether group: the first reaction uses the strong acid TFA, while in the second reaction the strong base LDA is used. In order to decide which of these reactions one would use in their own desymmetrisation therefore depends on whenever their substrate is sensitive to acids or bases: the first reaction should be used with base sensitive substrates and the second reactions for acid sensitive substrates.

2.3 Conclusions on Non-Catalytic Desymmetrisation

All three non-catalytic reactions, the Iwata reaction15_{(Scheme 2), the first Harada reaction}20_(Scheme

3) and the second Harada reaction21_{(Scheme 4) were able to desymmetrise 1,2-meso-diols over six,}

five and four steps respectively. While the Iwata and first Harada reactions were proven to be able to handle cyclic substrates, the second Harada reaction should also be able to handle these due to its similarity to the first Harada reaction. A severe disadvantage is that these reactions are limited to only 1,2-meso-diols: whilst acetals with larger ring systems are known to exist,23_{9 is not likely to react with}

any diol larger than 1,2 because of the limitation imposed by it being a 1,1 diether and five- and six-membered rings being the most stable ring structures.24_{The low flexibility of benzene rings}25_{as well}

as the presence of the rotaxane ring in 1 (Scheme 1) makes these reactive pathways not suitable for application in the proposed synthesis of mechanically chiral rotaxanes.

(11)

11

3. Biocatalytic Desymmetrisation

Historically the first use of asymmetric catalysis was the use of enzymes.12,13_{Due to the importance of}

chirality in biologically active molecules the use of enzymes in asymmetric synthesis has become common practice.6,7,12,13_{The main advantage of biocatalysis is the ability to obtain chiral products in}

very high ee, contrary to most early chemical methods. The main disadvantages are that enzymes have a specific substrate scope and are not always able to synthesise both enantiomers selectively. While asymmetric chemical catalysts can produce the different enantiomer of the product if a different enantiomer of the catalyst is used, in biochemical methods a new enzyme has to be discovered to synthesise the opposite enantiomer of a product from the same starting materials in most cases.12,13,15

The investigation and optimisation of enzymes for application in chemical reactions is also a time-consuming process, as many different enzymes have to be found, investigated and compared to find the ideal catalyst for a reaction. Nevertheless, biocatalysis is still a major part of asymmetric synthesis and no review about desymmetrisation would be complete without addressing its use and high enantioselectivity. In general, biocatalytic desymmetrisation of diols follows one of two possible pathways: 1) enantioselective acylation of one of the hydroxy groups or 2) enantioselective hydrolysis of a diester to an ester alcohol.26_{Both pathways will be discussed in this report. The most common}

type of enzyme used for these reactions are lipases.13,26_{In their natural environment, these enzymes}

catalyse the hydrolysis of fats and oils to glycerol and fatty acids.27_{In organic solvents, however, the}

equilibrium is reversed and shifted towards esterification.28

3.1 Biocatalytic Acylation

The first of the two main pathways to be discussed is biocatalytic enantioselective monoacylation of

meso-diols. 29_{In an early study by Nicolosi et al. in 1994, diphenyldiol 21 was selectively acylated}

(Scheme 6). In their study, two different enzymes were found, each able to catalyse the formation of a different enantiomer. To synthesise 22a Rhizopus Javanicus lipase (RJL) was used and 22a was obtained with 70% yield and 90% ee. For the synthesis of 22b Candida Cylindracea lipase (CCL) was used. The reaction took place for 3h, after which the enzyme was washed and recycled for a total of five times. Product 22b was obtained with 28% yield and 90% ee. After the reaction there was still 65% of unconverted 21 present, which could be recycled to obtain higher yields. Both of these reactions were able to give 90% ee, however the yield of 22b was significantly lower than for 22a. Whilst these reactions were developed before the non-catalytic reactions discussed in Section 2, their yield is significantly higher than those reactions, and only one step is needed where the chemical methods needed at least four.15,20,21,29

Scheme 6: Biocatalytic desymmetrisation by Nicolosi et al. a) vinyl acetate, RJL, cyclohexane/t-amyl alcohol (9:1), rt., 24h; b) vinyl acetate, CCL, toluene, 45°C, 15h.

Another biocatalytic acylation was published in 2001 by Hirama et al.30_{As part of their total}

synthesis of trans-4,5-dihydroxy-2-cyclopenten-1-one, they desymmetrised cyclopentenediol 23 to both enantiomers of 24 (Scheme 7). Enantiomer 24 could be obtained in one biocatalytic acylation step using lipase Amano AK as catalyst and vinyl acetate as acetate donor. The product 24 was obtained

(12)

12 with 99% yield and >99% ee. Enantiomer 24* could not be synthesised as easily as 24 but required one additional step. Acylation of 23 with acetic anhydride gave diester 25, which could be hydrolysed to

24* using the same Amano AK lipase catalyst, only this time in an aqueous phosphate buffer instead

of an organic solvent. After these two steps, 24* was obtained with 91% isolated yield and also >99% ee. These reactions are an excellent example of the influence of solvents on enzymatic reactions. In its natural environment, a lipase catalyses the hydrolysis of fats to fatty acids and glycerol.27_{This is the}

case in the formation of 24*, where the reaction was performed in an aqueous solution. When the solvent is changed to an organic solvent, as in the formation of 24, the enzyme will instead catalyse the inverse reaction, in this case the esterification of an alcohol.13,26,28,30

Scheme 7: Desymmetrisation used in the total synthesis by Hirama et al. a) Amano AK, vinyl acetate, benzene, 30°C, 13h; b) Ac2O, DMAP, pyridine, 1h; c) Amano AK, phosphate buffer 0.1M (pH 7.0), 33°C, 62h.

In a more recent study by Deska et al. in 2011, they were also able to desymmetrise a diol and obtain both enantiomers using the same enzyme to catalyse different reactions.31_{Allenediol 26 could}

be acylated to 27 with the use of porcine pancreatic lipase (PPL) as the catalyst and vinyl butyrate as the butyrate donor in dioxane (Scheme 8). Product 27 was obtained with 95% yield and 98% ee after this reaction. The other enantiomer, 27*, could be obtained by solvolysis of 28 using the same PPL biocatalyst, but this time with a 9:1 mixture of heptane and t-butanol as the solvent. The t-butanol would be used for the solvolysis, and 27* was obtained with 79% yield and 97% ee. This method uses a very specific allenediol, but the concept of changing the solvent to reverse the direction of the reaction from esterification to hydrolysis and obtain the other enantiomer has been proven again by this study.30,31_{This concept makes it easier for biocatalysis to be applied in organic chemistry when}

both enantiomers of a desymmetrised diol are needed in high enantiopurity.

Scheme 8: Desymmetrisation of allenediol by Deska et al. a) PPL, vinyl butyrate, dioxane, 40°C, 24h; b) PPL, heptane/t-butanol (9:1), 40°C, 48h.

(13)

13 Another enzymatic desymmetrisation of cyclic 1,2-diols was developed by Wirz et al in 2000.32_They

developed a reaction to desymmetrise 1,3,5-cyclohexatriol to be used as a key step in the synthesis of vitamin D derivates. The protected triol 29, see Scheme 9, was desymmetrised in vinyl acetate and ethyl acetate (9:81) which functioned both as acetate donor and as solvent. Many different lipases were used, and the enzyme that gave the highest yield, selectivity and reaction rate was lipase QL, which gave 30 with quantitative yield and >99% ee. In a later article by Hilpert and Wirz, this method was used in the synthesis of enantiopure vitamin D derivates.33 _{Among the enzymes tested were also}

CCL, which gave 87% yield and >99% ee, and AK, which gave 95% yield and >99% ee as well.32_When

comparing these results to the results of the use of these enzymes in the earlier discussed reactions, it can be concluded that CCL is much better at the desymmetrisation of 1,3-diols than 1,2-diols as was performed by Nicolosi et al.29,32_{and that AK can selectively desymmetrise both five- and six-membered}

cyclic 1,3-diols.30,32

Scheme 9: Desymmetrisation of cyclohexadiol by Wirz et al. a) Lipase QL, vinyl acetate, ethyl acetate, 22°C, 46h.

3.2 Biocatalytic Hydrolysis

As was discussed before, both esterification of diols and hydrolysis of diesters can be used to obtain desymmetrised diols.26_{An enzyme can catalyse both reactions, with the equilibrium depending on the}

use of solvents: water shift the equilibrium towards hydrolysis whilst organic solvents shift to esterification.26,28,30,31_{In most cases, usage of hydrolysis gives the other enantiomer than}

esterification.30,31_{In this section, the focus will be on enzymes that are primarily used to desymmetrise}

diols using hydrolysis and for which no acylation pathway was published.

The first of these hydrolysis methods to be discussed was published by Trauner et al. in 2003.34_In

their reaction, which is shown in Scheme 10, they wanted to desymmetrise 1,4-cyclohexadiol 31 using pig liver esterase (PLE). Cyclic diol 31, which was obtained after Birch reduction, hydroboration and oxidative work-up of para-xylene, was acylated with acetic anhydride to obtain 32 in 98% yield. Diacetate 32 could then be hydrolysed with PLE as the catalyst in an aqueous solution of ammonium sulphate until it had been completely converted to 33, which was obtained with an ee of >99.5%. The isolated yield of 33 was 98%. As of now, there has been no known publication in which a different enantiomer of 33 has been synthesised. Since this reaction was able to desymmetrise a cyclohexadiol derivate, this reaction could possibly be applied in the desymmetrisation of 1,4-benzenediol derivates, as is needed for the synthesis of mechanically chiral rotaxanes.

Scheme 10: Desymmetrisation of cyclohexadiol by Trauner et al. a) Ac2O, Et3N, DMAP, CH2Cl2, rt., 3h; b) PLE, (NH4)2SO4 (aq),

(14)

14 A recent paper by Zhang et al. described the desymmetrisation of an aromatic 1,4-diol (Scheme 11).35_{In their study, they optimised the desymmetrisation of 34 using biocatalytic hydrolysis. Diol 34}

first underwent esterification using propionyl chloride to give 35. Propionyl groups were chosen as usage of these groups gave a high conversion without giving an unpleasantly smelling by-product, as was the case for butyl groups, which led to the formation of butyric acid during hydrolysis. Many different lipases were tested for the formation of 36, and AK was found to be the most selective and efficient. Finally, the pH and buffer were optimised, and the most efficient way to convert 35 into 36 was to use AK as the biocatalyst within an aqueous sodium dihydrogen phosphate – sodium hydroxide buffer of pH 7.2. With these optimised conditions, 36 was obtained with 81% isolated yield and 98% ee. As the paper describing this reaction was only published several months before this report, no synthesis of the other enantiomer of 36 was known.

Scheme 11: Desymmetrisation of an aromatic diol by Zhang et al. a) Propionyl Chloride, Et3N, DMAP, Ethyl Acetate, 0-25°C,

5h; b) Lipase AK, NaH2PO4-NaOH buffer (pH 7.2), THF, 15°C.

3.3 Conclusion on Biocatalytic Desymmetrisation

When compared to the chemical methods in section 2, the biocatalytic desymmetrisations have a significant higher yield and ee, and a reduced number of chemical steps.15,20,21,29–31,33–35_{Where the}

chemical methods needed at least four steps, the biochemical desymmetrisation only require one or two steps for esterification and hydrolysis, respectively. Of the biochemical methods, the highest yields were obtained for the direct esterification of the diol, with 99% yield obtained by Hirama et al.30 _and

Wirz et al.32 _{The lowest yield of 28% and ee of 90% was obtained by the method of Nicolosi et al.}29

which is not unexpected, since this is the oldest of all the methods discussed. The other methods all had an ee value of at least 97%. In general, the biochemical methods are able to obtain high yields and almost enantiopure products. The main disadvantage is that most enzymes have a smaller substrate scope than most chemical catalysts, and it is not always possible to obtain both enantiomers.34,35_The

methods of Trauner et al.34_{and Zhang et al.}35_{both use six-membered cyclic 1,4-diols, and Zhangs}

substrate even contains an aromatic system. Therefore, usage of either PLE or AK to hydrolyse a acylated 1 (Scheme 1) would be the most likely biocatalytic pathway .

(15)

15

4. Metal Catalysed Desymmetrisation

The use of transition metal complexes as catalysts in organic chemistry goes back to the early 1970s.36

Due to the unique binding modes between transition metal d-orbitals and the organic s- and p-orbitals, many different types of reactions can be catalysed by organometallic complexes. These reactions include, but are not limited to: hydrogenation of olefins,37_{partial oxidation of organic molecules}38_and

carbon-carbon bond formation reactions such as the Heck39_{and Suzuki}40_{reactions. Many of these}

reactions have also been optimised for asymmetric reaction, such as asymmetric epoxidation38_or

carbon-carbon bond formation.39,40_{The use of asymmetric transition metal catalysis, whilst being a}

more recent method, has not passed over the field of desymmetrisation of meso-diols. In this section of the report, different transition metal catalysed desymmetrisation reactions will be discussed. Due to the large number of known reactions and catalysts, the reactions will not only be sorted on the type of reaction, but also on the type of catalyst used.

4.1 Copper-BOX Catalysed Acylation

One of the most popular ligand classes for asymmetric catalysis are the C2-symmetric bis(oxazoline),

or BOX ligands (Figure 3, 37).41_{The two nitrogen atoms of this molecules are coordinated to the metal}

centre of the complex, and the use of bulky R groups effectively block one of the sides of the substrate from attack by the reactant and therefore, one enantiomer of the product has an energetically preferred transition state, leading to enhanced enantioselectivity. The first known metal catalysed desymmetrisation of meso-diols was developed by the group of Matsumura in 2003, and they used a copper(II)-BOX derivate as the catalyst.42

Figure 3: General structure of the C2-symmetrical BOX ligands.

In their article, Matsumura et al. first developed a catalyst for the kinetic resolution of chiral diols.42

Their CuCl2-BOX catalyst, Scheme 12, compound 39, was able to recognise only one enantiomer within

the racemic mixture and acylate this enantiomer with 99% ee. When the reaction conditions were optimised, they were applied to meso-diol 21, and desymmetrised diol 38 was obtained with 79% yield and 94% ee. The mechanism of this desymmetrisation is given in the catalytic cycle in Scheme 12. At the start of this cycle, 21 coordinates to the copper atom of 39 with the two hydroxy groups, replacing the chloride anions coordinated to 39 in the process. Complex 39a was then deprotonated with the bulky, non-nucleophilic base DIPEA, which is well known for its application in metal catalysed coupling reactions.42,43_{Because of the chiral BOX ligand, one of the two protons is surrounded by three phenyl}

groups. This causes that only the less hindered proton is easily deprotonated by the bulky DIPEA base, inducing preference for one enantiomer of the product. The deprotonated complex 39b is acylated with benzoyl chloride, giving 38 in 79% yield and 94% ee. Whilst Matsumura et al. did only desymmetrise 21, a later study by Arai et al. did desymmetrise 1,2-cyclohexadiol with 32% yield and 56% ee using the same catalyst and reaction conditions.44

(16)

16

Scheme 12: The Copper(II) catalysed desymmetrisation of Matsumura et al. and the proposed catalytic cycle. a) 39, BzCl, DIPEA, CH2Cl2, 0°C.

The reactivity of a transition metal catalyst depends on many factors. When developing new BOX-type ligands, the geometry of the bridging atom and the charge of the ligand are some of the most important among these factors.45_{Neutral ligands with a tetrahedral bridging atom, such as 39, have}

been known for a long time, as have neutral ligands with planar bridging atoms46_{and anionic planar}

ligands.47_{In order to investigate a new type of BOX ligands, the group of Pfaltz thus developed anionic}

tetrahedral ligands, using a negatively charged boron atom as the tetrahedral bridging atom (Scheme 13, 41).45_{Their new catalyst was tested and optimised for cyclopropanation, and inspired by the}

desymmetrisation reaction by Matsumura et al. they decided to study their catalyst’s use in desymmetrisation of 1,2-meso-diols.42,45_{The same procedure as in Scheme 12 was used, where 8 was}

desymmetrised with catalyst 41 and the bulky base DIPEA, to obtain 40 with 83% yield and 90% ee. Since this reaction was similar to the one of Matsumura et al., the catalytic cycle in Scheme 12 also applies to the reaction in Scheme 13, with the bulky benzyl groups of 41 blocking one of the hydroxy groups from being deprotonated by DIPEA which causes the stereoselectivity of the reaction. Pfaltz et

al. did not only use 8 as the substrate, but also tested cyclopentadiol (73% yield, 76% ee) and

1,2-butadiol (65% yield, 94% ee) and different ligands, where the aromatic groups on the central boron atom and at the dihydrooxazole rings were substituted, however 41 had the highest yield and ee of all ligands tested. The main difference with the catalyst of Matsumura et al. is that the ee and the yield were higher for the group of Pfaltz then when Arai used the Matsumura catalyst to desymmetrise 8, with 32% yield and 56% ee for the Matsumura catalyst versus 83% yield and 90% ee for the Pfaltz catalyst.42,44,45

(17)

17

Scheme 13: Desymmetrisation of 1,2-cyclohexadiol using the catalyst developed by Pfaltz et al. a) 41, DIPEA, BzCl, CH2Cl2,

0°C – rt.

A third BOX-derivate copper catalysed desymmetrisation was discovered by Arai et al. whilst investigating a new method to optimise reaction conditions for catalytic reactions.48_{In their method,}

they used circular dichroism spectroscopy (CD spectroscopy) to visualise the formation of chiral products. CD spectroscopy is a form of absorption spectroscopy, where the difference in absorption between left and right circular polarised light is measured.49_{Since chiral systems interact with}

polarised light while achiral systems do not, CD spectroscopy can be used to track the formation of chiral molecules during a desymmetrisation reaction, and it can also be used to determine the absolute stereochemistry. In their experiment, Arai et al. were able to immobilise their BOX derivate 43 (Scheme 14) by connecting one of the NTs groups to a polymer support.48_{They were then able to try many}

different reaction conditions, and the CD spectroscopy showed which conditions gave the highest reaction rate. The optimised conditions were then used in the reaction depicted in Scheme 14. Some further optimisation of the reaction conditions lead to 8 being desymmetrised using para-bromo benzoyl chloride as the acyl donor, with 42 being obtained with 72% yield and 80% ee. Whilst the yield and ee are lower than the values of Matsumura42_{and Pfaltz}45_{, these values were the highest of all}

methods used by Arai et al., thus proving their optimalisation.48_{A notable difference is the use of}

copper(I), where the other two methods both used copper(II) as the metal centre.

(18)

18

4.2 Copper-Diamine Catalysed Acylation

A different type of copper catalysts for desymmetrisation reactions were developed by Shirai et al. in 2006.44_{In their study, they did not use the BOX-type ligands, but instead focused on cyclic diamines.}

They developed catalysts 44 and its methylated derivate 44Me (Scheme 15) and compared their yields and ee with the natural compound (-)-Sparteine, a well-known chiral ligand,50_{and Matsumura catalyst}

39.42,44_{A small selection of both cyclic and acyclic 1,2-meso-diol substrates were tested. The reaction}

conditions were the same as used by Matsumura, with the ligand and CuCl2 being used as the catalyst,

and the bulky, non-nucleophilic base DIPEA43_{being used to deprotonate substrate 8, which is then}

attacked by BzCl, leading to the desymmetrised product 40. The Matsumura catalyst 39 had the lowest potential, as 40 was obtained with only 32% yield and 56% ee. Shirai catalyst 44 gave 40 in 80% yield and 91% ee, (-)-sparteine gave 80% yield and 94% ee, and the methylated 44Me gave 100% yield and 91% ee. In a later study by Kaluza et al. in 2014, however, 44Me only yielded 40 in 94% yield and 90% ee under the same reaction conditions.44,51_{The methylated 44Me was, however, still the most potent}

catalyst for desymmetrisation of 8 in their research. When this is taken into account, the most potent ligand for desymmetrisation of 8 using a copper catalyst is (-)-sparteine, for this gives the best ee.

Scheme 15: The desymmetrisation reaction used by Shirai et al. and the four different ligands used. a) CuCl2, ligand, DIPEA,

BzCl, CH2Cl2, -78°C, 24h.

4.3 Zinc Catalysed Acylation

Whilst the last sections have focussed on enantioselective acylation of meso-diols using a copper complex, the second metal that has been used for desymmetrisation catalysts is zinc. Where copper catalysis is mostly used to desymmetrise 1,2-meso-diols, zinc catalysis is used mostly for desymmetrisation of 1,3-meso-diols.52_{The first of these methods to be discussed was published Trost}

and Mino in 2003.53_{For their desymmetrisation of 1,3-meso-diols and cyclic 1,4-meso-diol 45 (Scheme}

16) they developed chiral ligand 48. After deprotonation of all three hydroxy groups of 48, it would form a binuclear complex with two zinc(II) cations. An etheneolate anion, coming from the reaction with 46, was also coordinated to one of the zinc ions, to bring the total charge to zero (48a). As seen in the catalytic cycle in Scheme 16, the 1,3-diol would coordinate whit each oxygen atom to a different zinc ion, giving complex 48b. The ethenolate fragment would then deprotonate one of the hydroxy groups of the substrate, and the protonated enolate would leave in its more stable keto form as acetaldehyde.53,54_{Deprotonated complex 48c would then react with 46, where the carbonyl oxygen of}

46 would coordinate to the zinc atom where acetaldehyde had left a vacant coordination site. The

benzyl fragment on 48d would then move to the deprotonated hydroxy, releasing the product from the catalyst complex, and the ethenolate fragment would stay coordinated to zinc, returning the cycle to 48a. Desymmetrised product 47 was formed in 93% yield and with 91% ee. The catalytic cycle shown in Scheme 16 was published for the desymmetrisation of 1,3-meso-diols, but since the same catalyst could be used for desymmetrisation of 45, the same catalytic cycle applies to this reaction.

(19)

19

Scheme 16: The Zn catalysed desymmetrisation by Trost and Mino and the proposed catalytic cycle. Ar = -C6H4Ph. a) 48,

ZnEt2, Toluene, -15°C, 2h.

A different zinc catalyst for desymmetrisation of 1,3-meso-diols was developed by Nagao et al.55

Their catalyst was formed by the addition of diethyl zinc to two equivalents of the chiral ligand of 51 (Scheme 17), which would coordinate to give catalyst 51. They wanted to desymmetrise 1,3-dihydroxy propane-2-amines (49) and were inspired by the work of Trost and Mino,53_{but their catalyst was}

unable to desymmetrise 49. Therefore, a new zinc(II) catalyst was developed (51), which was able to desymmetrise 49 to 50 with 92% yield and 88% ee.55_{In their proposed catalytic cycle, which is shown}

in Scheme 17, the substrate 49 would first coordinate to catalyst 51, giving complex 51a. Deprotonation of 51a yielded 51b, which could be acylated by acetic anhydride, forming acetate as the base used in the deprotonation step as well as complex 51c. Dissociation of 50 from 51c then yielded the product as well as closing the catalytic cycle, regenerating 51. To expand the substrate scope of their catalyst, Nagao et al. used substrates where the methyl group was substituted for different R groups, which gave similar ee values and slightly lower yields, except for Bn, where both yield and ee dropped to 70%.

(20)

20

Scheme 17: Desymmetrisation reaction and proposed catalytic cycle by Nagao et al. a) 51, Ac2O, tBuOMe, 0°C, 20h.

Another publication by the group of Nagao in 2010 described a different zinc catalyst for the desymmetrisation of 1,3-meso-diols.56_{This time they did not desymmetrise protected amine derivates}

such as 49, but protected glycerol derivates (Scheme 18, 52). The ligand used in this reaction is a cinchona alkaloid containing the same aromatic sulfonyl moiety as catalyst 51 in Scheme 17.55,56_Since

the substrate and the conditions in Scheme 18 are similar to the one in Scheme 17, the catalytic cycle will most likely be similar. The zinc atom will probably coordinate to the nitrogen atom connected to the sulfonyl aromatic moiety and the nitrogen in the bridged bicyclic group. Diol 52 was desymmetrised with 54 and diethyl zinc, and acetic anhydride was used as acyl donor and base, similar to Scheme 17. Desymmetrisation product 53 was obtained with 78% yield and 86% ee, slightly lower than catalyst 51.

(21)

21 The final zinc catalysed enantioselective acylation to be discussed in this report was published by Subramanian et al. in 2019.57_{They used the binuclear zinc(II) complex 55 (Scheme 19) to desymmetrise}

21 into 38. No catalytic cycle was published, but as this is a binuclear zinc complex, the catalytic cycle

would be similar to the one in Scheme 16.53,57 _{The diol would coordinate to both zinc atoms with its}

two hydroxy groups, then one would be deprotonated by lutidine, a steric base of natural origin.57,58

The deprotonated diol can then be attacked by benzoyl chloride and would dissociate from the catalyst complex, regenerating 55 and expelling 38 in 80% yield and 75% ee. This catalyst is thus far the only zinc acylation catalyst used for 1,2-meso-diols.

Scheme 19: Desymmetrisation by Subramanian et al. a) 55, lutidine, BzCl, CH3CHCl2, rt.

4.4 Copper-Catalysed Desymmetrisation using Different Protective Groups

In this final section about metal catalysed desymmetrisation of diols, three different methods will be discussed. Each of these methods introduces a different group after desymmetrisation and all methods make use of the BOX type ligands. The first of these methods to be described is the carbamoylation of cyclic 1,2-meso-diols developed by Matsumura et al. in 2006.59_{After their development of the copper}

catalysed acylation shown in Scheme 1242_{it was found that the basic conditions used to deprotonate}

the diol also led to acyl transfer, which decreased the enantioselectivity of the reaction.50,59_{In order to}

avoid this side reaction, Matsumura and co-workers experimented with base free desymmetrisation. Their first reaction is shown in Scheme 20, where diol 8 is desymmetrised using chiral ligand 39 in combination with copper(II) triflate as catalyst. This is the same catalyst as they used in their acylation.42,59_{The advantage of carbamoylation over acylation is that this reaction does not need any}

other reactants such as an acid or base, but takes place spontaneously upon mixing of an alcohol with an isocyanate.60_{In their reaction, Matsumura et al. desymmetrised 8 by addition of chiral ligand 39,}

which coordinated to copper(II) triflate.59_{Product 56 was obtained with 69% yield and 86% ee at -40°C}

and 92% yield and 76% ee at room temperature. The enantioselectivity was introduced due to one of the hydroxy groups being sterically more shielded by the phenyl groups of 39, in the same manner as for Scheme 12. Whilst no mechanism for the reaction was given, it is most likely that the diol would coordinate to the catalyst with both hydroxy groups, followed by carbamoylation of the least hindered position.42,59,60

(22)

22 Whilst the above method by Matsumura et al. (Scheme 20) was able to desymmetrise diols without the use of a base, the enantioselectivity did not satisfy the research group.59_{Therefore, a new}

desymmetrisation reaction was developed using the same copper(II)-BOX catalyst, however, this time a sulfonylation reaction was performed instead of a carbamoylation.59,61_{In this reaction, which is}

shown in Scheme 21, cyclic diol 8 was desymmetrised with chiral ligand 39 and copper(II) triflate as the catalyst, and p-toluenesulfonyl chloride was used as the sulfonylation reagent. Since a base was needed in this reaction, the catalytic cycle should be similar to the cycle for acylation shown in Scheme 12. Diol 8 coordinated to the catalyst, and due to the steric hindrance of the phenyl groups on 39, only one of the hydroxy groups can be deprotonated. The deprotonated hydroxy group is then sulfonylated, and dissociation of the product 57 from the catalyst would close the catalytic cycle. Desymmetrised diol 57 was obtained with 94% yield and 97% ee. Since the product was sulfonylated and not acylated, the use of a base could not lead to acyl transfer reactions.42,60–62_{This reaction was done on a large}

substrate scope containing many cyclic and acyclic 1,2-diols and the values of the yield and ee were compared to carbamoylation and benzoylation, with sulfonylation giving the highest ee and good yield for all substrates investigated, with 93% ee being the lowest.

Scheme 21: Sulfonylation reaction by Matsumura et al. a)TsCl, Cu(OTf)2, 39, K2CO3, CH2Cl2, rt., 12h.

The most recent advance in copper(II)-BOX catalysed desymmetrisation of diols was published by Kuriyama et al. in 2016.63_{They developed a benzylation reaction, where the hydroxy group would be}

transformed into a phenyl ether (Scheme 22). This reaction was developed since the only other known asymmetric benzylation method was discovered in 1986 by Brunner et al. and therefore a more modern approach was needed.63,64_{Kuriyama and co-workers used diphenyl iodonium triflate as the}

arylating agent, as diaryl iodonium salts are known as strong arylation reagents.63,65_{The mechanism of}

this reaction, which is based on the mechanism of Scheme 12,42_{starts with the diol coordinating to the}

catalyst complex, forming 39a, which is deprotonated enantioselectively by the phosphate66_{to give}

39b. Complex 39b can then be arylated with the diaryliodonium salt to obtain complex 39c, which after

dissociation of the product 58 regenerates the catalyst.63_{Phenyl ether 58 was obtained with 47% yield}

and 24% ee, but the method was better suited for larger cyclic 1,2-meso-diols, such as cyclooctane (94% yield, 76% ee) or cyclododecane (84% yield, 85% ee).

(23)

23

Scheme 22: Etherification of meso-diol by Kuriyama et al. a) CuCl2, 39, Na3PO4, toluene, 80°C, 15h.

4.5 Conclusion on Metal Catalysed Desymmetrisation

In this section, a total of eleven different methods for metal catalysed desymmetrisation of meso-diols have been discussed. Several different ligands used with both copper and zinc as metal centre were used as catalyst. In general, the copper catalysed acylation reactions were used for 1,2-meso-diols,42,44,45,48,51_{whilst the zinc catalysed acylations were used for larger diols,}53,55,56_{mainly 1,3 and 1,4,}

with the notable exception being the zinc catalysed desymmetrisation of 1,2-diphenylethane-1,2-diol (Scheme 19, 21) by Subramanian et al.57_{The advantage of metal catalysis is that it is easy to synthesise}

the other enantiomer of a chiral ligand to obtain the other enantiomer of the product.67_{The most}

selective and efficient desymmetrisation acylation catalyst was the copper(II)- (-)-sparteine complex investigated by Shirai et al.44_{Three other copper-BOX catalysed reactions were discussed, namely}

carbamoylation59_{, sulfonylation}61_{and etherification.}63_{Of these three reactions, the sulfonylation}

showed the highest potential, with high yields and ee and no unwanted acyl transfer.61,62

The main disadvantage of the metal catalysed desymmetrisation reactions is that the yields and ee are in general lower than for the biocatalytic pathways.34,57,61_{Another major disadvantage is that due}

to the coordination needed between the metal catalyst and the substrate, the two hydroxy groups of the diols have to be relatively close to each other.68,69_{Since the coordination modes between}

1,4-benzenediol and transition metals does not include a mode in which both hydroxy groups are coordinated to the metal, these methods will be unable to catalyse the desymmetrisation reaction needed for the synthesis of mechanically chiral rotaxanes.70

(24)

24

5. Organocatalytic Desymmetrisation

By far the most studied and diverse category of enantioselective desymmetrisation of diols is the category of organocatalytic methods.12,52,71_{Many different types of organocatalysts have been}

developed, and many different kinds of desymmetrisation reactions have been published. In order to give a structured overview of all different organocatalytic methods, this section will be organised similar to section 4, where the methods will be sorted not only on the type of reaction used but also the kind of catalyst used. This section will start with the different types of organic catalysed acylations and the types of catalysts used, and this will be followed by examples of etherification and silyl etherification.

5.1 Diamine Catalysed Acylation

The first group of reactions to be discussed are the diamine catalysed acylation reactions. One of the oldest and most famous organic desymmetrisation catalysts was developed by the group of Oriyama in 1998.72_{Since they already knew that proline derived diamines were able to catalyse kinetic}

resolution of racemic alcohols,73_{they wanted to apply these kinds of catalysts for desymmetrisation,}

as desymmetrisation has a higher theoretical yield than kinetic resolution.14,74_{Oriyama and co-workers}

developed proline derived diamine 59 (Scheme 23) and tested this on model compound 8. According to their proposed mechanism, which is shown in Scheme 23, catalyst 59 would form complex 59a upon reaction with benzyl chloride. The chirality of 59a would sterically protect one side of the oxomethylium ion from attack by the diol, so only the unprotected side could be attacked, introducing the stereoselectivity. After attack by the diol, protonated catalyst 59b would be deprotonated by triethyl amine, and could once again form complex 59a upon reaction with benzyl chloride. With this catalyst, 40 was obtained with 87% yield and 97% ee, which was quite high for the time of this paper.20,21,29,74_{Different 1,2-meso-diols were tried as substrate, with all having high yields and ee.}

Scheme 23: Diamine catalysed desymmetrisation by Oriyama et al. and the proposed mechanism. a) 59, BzCl, Et3N, MS4Å,

CH2Cl2, -73°C, 3h.

After their success with 1,2-meso-diols, Oriyama and co-workers decided to expand the substrate scope of their proline derived diamines to 1,3-meso-diols.72,74_{For this purpose, the same catalyst as}

was used for their initial kinetic resolution, catalyst 62 (Scheme 24) was used.72,73_{In this reaction, diol}

60 was desymmetrised to ester 61 using 4-(tert-butyl)-benzoyl chloride as the acylating agent and

(25)

25 the same reaction mechanism as in Scheme 23, where the catalyst would form a complex with the benzoyl cation, then be attacked by and deprotonate the substrate, and be deprotonated by DIPEA and once again form a complex with a benzoyl cation.72,74_{This would give 61 with 22% yield and 93%}

ee. Whilst the ee value for this reaction with phenyl and different substituents is quite high, the yields were low for the complete substrate scope investigated by Oriyama and co-workers.

Scheme 24: Second diamine catalysed desymmetrisation by Oriyama et al. a) 62, p-tBuC6H4COCl, DIPEA, nPrCN, MS4Å,

-78°C, 3h.

Inspired by the desymmetrisation reactions by Oriyama et al., the group of Kündig decided to investigate if they could desymmetrise organometallic π complexes in the same way.72,74–76_They

synthesised diol 63 (Scheme 25) by reduction of a naphthalene diol and formation of its chromium tricarbonate complex. Then, they first tried the known Oriyama catalyst 59 to catalyse the reaction shown in Scheme 25 with benzoyl chloride as acylating agent and triethyl amine as the base. The (-) enantiomer of 64 was obtained with 74% yield and 98% ee.75,76_{Since these results were quite good,}

Kündig et al. then developed their own diamine catalysts 65 and its enantiomer 65*. Usage of catalyst

65 gave the same product as 59 but with 89% yield and 99% ee, which is higher than for 59. When the

enantiomer of 65, 65* was used, the (+) enantiomer of 64 was obtained in 80% yield and 97% ee. With their research, Kündig and co-workers did not only develop a more effective catalyst for the desymmetrisation of 63, but were also able to prove that both 65, 65* and Oriyama catalyst 59 were able to catalyse the desymmetrisation of cyclic 1,4-meso-diols as well. Since usage of catalyst 65 has high yield and ee and an easy way to obtain the other enantiomer of the product, this catalyst could have the potential for the desymmetrisation step needed for the synthesis of mechanically chiral rotaxanes.

Scheme 25: Desymmetrisation of cyclic 1,4-meso-diol by Kündig et al. a) Cat 59, 65 or 65*, BzCl, Et3N, MS4Å, CH2Cl2, -60°C,

(26)

26

5.2 DMAP-Derivate Catalysed Acylation

The second group of organocatalytic acylation catalysts are the derivates of 4-dimethylamine pyridine (DMAP). By itself, DMAP is known as a good catalyst for many functional group transfer reactions, including but not limited to acyl transfer reactions.77_{Because of these known catalytic properties, it is}

not unusual that chiral DMAP derivates were among the first chiral organocatalysts used for desymmetrisation of diols. The first of these methods was developed by Fu et al. in 1998.78_They

developed the ferrocene-DMAP derivate 68 (Scheme 26) initially as a catalyst for kinetic resolution of a racemic mixture of chiral alcohols. Further investigation of the catalytic properties of 68 in the presence of meso-diols yielded good desymmetrisation results. The reaction in Scheme 26 was catalysed by 68, where 1,5-diol 66 was acylated to 67 using acetic anhydride as the acylation agent. The product was obtained with 91% yield and 99.7% ee. The mechanism shown in Scheme 26 is the general mechanism for DMAP catalysed acylation, as proposed by Zipse et al.77_{In this mechanism, the}

acetic anhydride reacts, coordinating the acetyl group to DMAP, forming DMAP-1. DMAP-1 can then interact with alcohols, forming complex DMAP-2, where the carbonyl carbon of the acyl fragment coordinates to the oxygen atom of the alcohol, and the alcohol proton is coordinated to the acetate fragment. In the final step of the mechanism, the acylated alcohol and acetic acid are released, and the DMAP catalyst is regenerated. Usage of an external base, as was done by Fu et al. can accelerate the reaction, as the deprotonation of the alcohol will be faster.78,79_{The chiral ferrocene group used by}

Fu and co-workers blocked one side of DMAP-2 and therefore the stereoselectivity was introduced. The catalyst can be used for aromatic systems, as was the case by the group of Fu, but it can also be used for other six membered rings with 1-hydroxyethyl groups at the 1,3 position.80

Scheme 26: Desymmetrisation described by Fu et al. (top) and the DMAP catalysed acylation mechanism described by Zipse et al. (bottom). a) 68, Ac2O, Et3N, t-amyl alcohol, 0°C.

A more recent DMAP derived desymmetrisation catalyst was developed in 2006 by Yamada et al.81

Their catalyst, 71 in Scheme 27, was also developed as a catalyst for kinetic resolution at first, and afterwards used on meso-diols to investigate its effects. The 1,4-diol 69 was desymmetrised using isobutyric anhydride as the acyl donor and triethyl amine as the base. Product 70 was obtained with

(27)

27 87% yield and 97% ee. Many different substrates were used, ranging from 1,2- to 1,6-diols. Of these substrates, 69 gave the highest yield and ee. A cyclic derivate of 69 was also tried, where the two terminal methyl groups were connected to form a cyclohexane ring, but the ee of this substrate was only 18% and the yield 44%. Since catalyst 71 is also a DMAP derivate, the general mechanism depicted in Scheme 26 applies to this reaction.77,81_{The low yield and ee for a cyclohexane derivate of 69 leads}

to the conclusion that some flexibility of the alcohol groups is needed to obtain high yields and ee’s of desymmetrised product when using catalyst 71.

Scheme 27: Desymmetrisation reaction developed by Yamada et al. a) 71, isobutyric anhydride, Et3N, 0°C.

A third DMAP derivate desymmetrisation catalyst was developed by Mandai and co-workers in 2017.82_{Their catalyst, 73 in Scheme 28, contains the binaphthyl chiral fragment. The chirality of this}

fragment originates from the high rotation barrier of the two naphthalene fragments.83_{The steric}

hindrance makes the racemisation kinetically unfavourable and therefore binaphthyl is a well-known fragment in chiral catalysts. The chiral DMAP catalyst 73 of Mandai et al. was used for the desymmetrisation of many different 1,2-meso-diols.82_{Of these substrates, the highest yield was}

obtained when substrate 8 was desymmetrised to 72 (Scheme 28). In this reaction, 72 was obtained with 79% yield and 94% ee. In a later study by the same group, catalyst 73 could also be used for the desymmetrisation of 1,3-meso-diols, with the highest yield being 90%, with a corresponding ee of 88%.84_{Since the catalyst is also a DMAP derivate, the mechanism is the same as the mechanism shown}

in Scheme 26.77,82,84

Scheme 28: Desymmetrisation reaction described by Mandai et al. a) 73, isobutyric anhydride, Et3N, TBME, -20°C, 3h.

5.3 Phosphine Catalysed Acylation

Due to phosphorus and nitrogen atoms being isoelectronic, phosphines and amines are able to catalyse similar reactions.85_{Both amines and phosphines are known for their application in organocatalysis.}

Whilst the last two sections has focussed on the use of amine catalysts for enantioselective desymmetrisation, this section will focus on the use of phosphines as catalysts. The first ever phosphine catalysed desymmetrisation was developed by Vedejs et al. in 1996.86_{This reaction, which}

is shown in Scheme 29, made use of catalyst 74, and desymmetrised 21 to 38, using benzoyl anhydride as benzoyl donor, with 84% yield and 68% ee. Due to the low yield of this reaction, another phosphine catalyst was developed by the same group, which was published in 2004.87_{Catalyst 75 catalysed the}

same desymmetrisation reaction, using the same benzoyl donor, but with 97% yield and 93.7% ee. The main difference between catalyst 74 and 75 is that the phosphorus atom in 75 is a chiral centre, which brings the chirality of the catalyst closer to the substrate than for 74.

(28)

28

Scheme 29: Phosphine catalysed desymmetrisation reactions of Vedejs et al. a) 74, (PhCO)2O, CH2Cl2, 0-20°C; b) 75,

(PhCO)2O, CH2Cl2, -30°C, 22h.

Another group famous for their development of phosphine based acylation catalysts is the group of Fujimoto. They are mainly known for their work with phosphinite derivates of cinchona alkaloids.12,71

While cinchona alkaloids were already discussed in this report as a ligand for zinc catalysed desymmetrisation,56_{they are on their own also known to catalyse many asymmetric reactions,}

including asymmetric acylation of diols.88_{The group of Fujimoto developed phosphinite derivates of}

these natural compounds, in order to combine the Lewis basic properties of the phosphorus atom with the Brønsted basic properties of the nitrogen and so create a potent desymmetrisation catalyst.89_The

first of their catalysts, which is shown in Scheme 30, was synthesised in situ by addition of chlorodiphenylphosphane to the cinchona alkaloid, giving catalyst 76. This catalyst was then used for the desymmetrisation of many different 1,2-meso-diols using benzoyl chloride as benzoyl donor and DIPEA as the base. Of all substrates, desymmetrisation of 8 to 40* gave the highest ee. Product 40* was obtained with 85% yield and 94% ee.

Scheme 30: Phosphinite-Cinchona alkaloid catalysed desymmetrisation by Fujimoto et al. a) 76, BzCl, DIPEA, EtCN, -78°C.

After developing catalyst 76 (Scheme 30) for the desymmetrisation of 1,2-meso-diols, the group of Fujimoto wanted to increase their substrate scope to include larger diols.89_{Using a slightly different}

cinchona alkaloid, they obtained catalyst 79 (Scheme 31), with the only difference between catalyst

76 and 79 being the methoxy group of 79.90_{This catalyst was able to catalyse the desymmetrisation of}

six-membered cyclic 1,4-diol 77 using the same benzoylation agent and base as catalyst 76. Desymmetrisation of 77 to 78 was done with 55% yield and 82% ee. Whilst the yield of this reaction is moderate, catalyst 79 could desymmetrise many different 1,3-meso-diols in much higher yields, up to 82%.

Scheme 31: Second Phosphinite-Cinchona alkaloid catalysed desymmetrisation by Fujimoto et al. a) 79, BzCl, DIPEA, CH2Cl2,

(29)

29 A third phosphinite catalyst was also developed by Fujimoto et al., which was however not based on cinchona alkaloids.91_{Catalyst 81, which is shown in Scheme 32, does still have the phosphorus Lewis}

base and nitrogen Brønsted base groups that the cinchona alkaloid derivates had. The catalyst was able to desymmetrise 8 into 80, as depicted in Scheme 32. With this reaction, 80 was obtained with 83% yield and 93% ee. The substrate scope of this catalyst includes many cyclic and acyclic 1,2-meso-diols and acyclic 1,3-meso-1,2-meso-diols. The yields for the desymmetrisation of the 1,2-1,2-meso-diols was on average higher than for 1,3-diols. The mechanism of the catalytic reaction was not explored, but it was suggested that it would go via intermediate 81a, when the acylating agent reacts with the catalyst to form this pentacoordinate phosphorus complex.

Scheme 32: Third phosphinite catalyst developed by Fujimoto. a) p-tBuPhCOCl, 81, DIPEA, MS4Å, toluene, 0°C, 12h.

5.4 NHC Catalysed Acylation

In the last section, it was shown that phosphines could be used to catalyse the desymmetrisation of

meso-diols.86,87,89–91_{Since the catalytic activity came mostly from the Lewis basicity of the phosphorus}

atom, it makes sense that an isoelectronic compound would be able to catalyse the same reactions. A type of compounds that is isoelectronic with phosphines are carbenes, of which N-heterocyclic carbenes (NHCs) are the most stable subgroup.92_{Two different NHCs able to catalyse}

desymmetrisation of diols were synthesised: catalyst 83 (Scheme 33) by Scheidt et al.93,94_{and catalyst}

84 by Yashima et al.95_{The only difference between both catalysts is the aromatic group on the nitrogen}

rings. Both catalysts were initially developed for oxidative esterification, but also showed activity for kinetic resolution and desymmetrisation of diols. Catalyst 83 was able to catalyse the desymmetrisation of 8 to 82, with the use of cinnamaldehyde as protective group. In order to deprotonate 8, the organic base known as proton sponge, or 1,8-bis(dimethylamino)naphthalene was used, which due to conjugation is a particularly strong base.93,94,96_{Catalyst 84, developed by Yashima}

et al., catalysed the desymmetrisation of 8 into 40.95_{For this reaction, benzaldehyde was used as}

benzoyl donor. Co-catalyst 85 was also used in the publication, however the exact use of this substance in the reaction is unknown, as it did not influence the ee, and Scheidt and co-workers could use 84 without the co-catalyst to catalyse desymmetrisation of 8.93–95_{Product 40 was obtained with 30% yield}

and 64% ee, and when the other enantiomer of 84 would be used, the other enantiomer of 40 would be obtained with similar yield and ee. The mechanism of this transformation included intermediate

83a, which due to the similarities of the catalysts would also apply for catalyst 83. The main advantage

of the NHC catalysts is that they are able to use aldehydes instead of acid chlorides or anhydrides as substrate, leading to lower waste production and less dangerous chemicals being used. The yield and ee values, on the other hand, are on the lower end of the spectrum. But since the use of NHCs as catalysts is still fairly new, it could become possible in the future to obtain a green NHC catalyst able to desymmetrise diols in high yields and ee.