WO2023020789A1

WO2023020789A1 - Methods for preparing oxazaborolidines

Info

Publication number: WO2023020789A1
Application number: PCT/EP2022/070814
Authority: WO
Inventors: Mike Collins; Lloyd Cooper
Original assignee: Pathway Intermediates Limited
Priority date: 2021-08-17
Filing date: 2022-07-25
Publication date: 2023-02-23
Also published as: GB202111761D0

Abstract

A method for the preparation of oxazaborolidines having a formylation step of an amino acid using formic-acetic anhydride (FAM) with formic acid solvent to produce a formyl amino acid; a reduction step of the formyl amino acid to the corresponding amino alcohol by reacting the formyl amino acid with lithium aluminium hydride; and a condensation step to condense the amino alcohol using a boronic acid to provide the desired oxazaborolidine.

Description

Methods for Preparing Oxazaborolidines

Field of the Invention.

The present invention relates generally to improved methods of synthesizing biologically active oxazaborolidines.

Background

Quorum sensing is the ability to detect and respond to cell population density by gene regulation. It is the chemical signalling between bacteria that enables them to communicate. Bacteria export chemical signalling molecules into their environment (Bassler, B & Lossick, R [2006], Bacterially Speaking, Cell, pp 237-246) and the information supplied by these molecules is critical for synchronizing their activity. This chemical communication by bacteria involves producing, releasing, detecting and responding to the accumulation of small hormone-like molecules, called ‘autoinducers’. The cell-cell communication mechanism known as quorum sensing allows bacteria to monitor their environment for other bacteria and to alter behaviour on a population-wide scale in response to changes in the number and/or species present in a community (Waters, C. & Bassler, B.I., [2001], The Languages of Bacteria. Genes & Development. Volume 15, pp 1469-1480). For example, quorum sensing enables bacteria to restrict the expression of specific genes to the high cell densities at which the resulting phenotypes will be most beneficial. Many species of bacteria use quorum sensing to coordinate gene expression according to the density of their local population. Quorum sensing has been found to be an integral regulatory component of cellular processes such as bioluminescence production, virulence gene expression, biofilm formation, cell division, motility, metabolism and recombinant protein production.

Bacteria require three characteristics in order to use quorum sensing. The bacteria must (i) secrete a signalling molecule (the autoinducer), (ii) detect the change in concentration of the signalling molecule and (iii) regulate gene expression in response. Different types of bacteria have different autoinducers. Acylated homoserine lactone (AHL) is used by gram-negative bacteria. These molecules are synthesized by Luxl and are detected by the LuxR protein. The LuxR-AHL complex acts as a transcription factor to regulate gene expression (Schauder, S. & Bassler, B., [2001] The languages of bacteria. Genes and Development, Vol 15, pp 1469- 1480). Gram-positive bacteria secrete modified oligopeptides, or autoinducing peptides, into the media. The detectors for the oligopeptides are two-component adaptive response proteins. Binding of the oligopeptide autoinducer to these response proteins at a critical concentration triggers an intracellular phosphorylation cascade that cumulates in the phosphorylation, and subsequent activation, of the cognate response regulator (Ng, WL. & Bassler, BL., [2009] Bacterial Quorum- Sensing Network Architectures. Annu Rev Genet, Vol 43, pp 197-222). Upon activation, the response regulator acts as a transcription factor, resulting in modulation of gene expression (Kleerebezem, M et al., [1997] Quorum Sensing by peptide pheromones and two-component signal-transduction systems in grampositive bacteria. Molecular Microbiology, Vol. 24, pp. 895-904).

A third autoinducer is autoinducer 2 (AI-2), a furanosyl borate diester and one of the only few known biomolecules incorporating boron. It is a signal molecule produced by the protein LuxS, an enzyme found in many bacterial species and considered to be responsible for switching on a large number of metabolic and catabolic pathways. Critically, it may be used by both gram-positive and gram-negative bacteria. Therefore, it is considered to be an important signal for inter-species communication and as such may be exploited in alternative strategies to antibiotic growth promotors as the AI-2 system may be used by both gram-positive and gram-negative bacteria.

The ongoing emergence of antibiotic-resistant bacterial strains means alternative therapeutic strategies are required. The ability of autoinducers to influence population-wide activity of bacteria, such as their growth and pathogenicity, indicates that they are feasible candidates for use as an alternative to antibiotic treatments. Quorum sensing can be influenced through the use of external autoinducers to inhibit the growth of pathogenic bacteria while promoting or not changing the growth of beneficial bacterium. This variation in effect is possible due to (i) the variable threshold concentrations between bacteria, and (ii) the differing pathways induced by the same autoinducer across different populations of bacteria. Oxazaborolidines have been used to control biofilm production because they are structurally similar to the autoducer-2 (AI-2). Oxazaborolidines are five-membered heterocyclic boron compounds containing both oxygen and nitrogen atoms and which can be synthesized to have a chemical structure that resembles that of AI-2 (Aharoni, R. et al., [2008], Oxazaborolidine derivatives inducing autoinducer-2 signal transduction in Vibrio harveyi. Bioorganic & Medicinal Chemistry, Vol. 16, pp. 1596- 1604). This structural similarity suggests that the oxazaborolidines may provide a similar biological effect and, indeed, several oxazaborolidines have been found to have physiological/metabolic and enzymatic effects on bacterial adhesion to the substrate. Oxazaborolidines have also been found to have modulatory effects on anti-enzymatic bacterial activity, act as anti-bacterial agents, and, pertinent to this application, have been shown to influence quorum sensing between bacteria leading to a decrease in communication and a resultant damage to the integrity of bacterial biofilm.

WO 2005/021559 A2 (Yissum Research Development Company of the Hebrew University of Jerusalem) discloses oxazaborolidines as bacteria effectors for modulating at least one bacteria-related parameter selected from adhesion of the bacteria to its substrate; enzymatic activity of the bacterial enzymes; viability of the bacteria; effect on quorum sensing and biofilm formation by the bacteria. This document describes a number of different oxazaborolidine structures but fails to disclose a process that produces a pure oxazaborolidine compound. Figure 1 of the accompanying drawings illustrates Gas Chromatography Mass Spectrometry (GCMS) spectra for the end product produced by the method disclosed in WO 2005/021559 A2. Three significant peaks are shown, the starting materials (A) boronic acid and (B) ephedrine and (C) oxazaborolidine, with the end product (C) making up only around 10% of the end product.

Significant challenges exist in the preparation of oxazaborolidines, in particular because existing routes to these compounds involve the use of ephedrine, an expensive and controlled substance due to its use in narcotic production. It is the aim of the present invention to provide improved methods for the preparation of oxazaborolidines that overcome, or at least alleviate, the abovementioned problems.

Summary of the Invention

The present invention provides a method for the preparation of oxazaborolidines, the method comprising the steps of:

(i) formylation of an amino acid using formic-acetic anhydride (FAM) with formic acid solvent to produce a formyl amino acid;

(ii) reducing the formyl amino acid to the corresponding amino alcohol by reacting the formyl amino acid with lithium aluminium hydride; and

(iii) condensing the amino alcohol using a boronic acid to provide the desired oxazaborolidine end product.

It is to be appreciated that the particular starting amino acid and the boronic acid used for the condensation reaction should be selected dependent upon the desired oxazaborolidine required for the end product. Preferably, the amino acid is phenylalanine or phenylglycine. The boronic acid is preferably phenyl boronic acid.

The racemic formyl amino acid or either of the enantiomers (L or D) of the formyl amino acid may be selected from the product of step (i) for further reaction to produce an oxazaborolidine with a selected stereochemistry.

It is preferable for the formic-acetic anhydride (FAM) to be prepared in situ for use in step (i) of the method. Preferably, formic acid is mixed with acetic anhydride with cooling, optionally under an atmosphere of nitrogen, then heated to up to 60°C followed by cooling to provide FAM. The amino acid is preferably dissolved in formic acid, optionally with cooling, priorto the introduction of the FAM to produce the formyl amino acid.

Preferably, excess formic acid is used.

It is to be appreciated that formyl amino acid may be further concentrated using standard techniques, such as fractional distillation and re-crystallization steps. Preferably, a desired racemate or enantiomer of the formyl amino acid produced in step (i) is selected for the reduction and condensation steps (ii) and (iii) for forming an oxazaborolidine with desired stereochemistry.

The reduction step (ii) is preferably carried out in an atmosphere of nitrogen. Preferably, a tetrahydrofuran, especially 2-methyltetrahydrofuran (2MeTHF), preferably dried, is boiled under a vacuum and batches of lithium aluminium hydride are added to the 2MeTHF. Preferably excess lithium aluminium hydride is used. The formyl amino acid produced in step (i), preferably dried, is then added, preferably using a sealed solids addition tube. Cooling may be applied, if necessary.

Washing may then be carried out, for example using one or more washings of dry 2MeTHF, preferably with the washings being added to the mixture.

Again, additional standard techniques may be applied to purify and extract the amino alcohol produced by the reduction step, such as filtration, distillation and recrystallization techniques.

The condensation step (iii) is preferably carried out using the amino alcohol and boronic acid in a ratio 1 :1-1 :1.5, preferably 1 :1-1 :1.2, more preferably 1 :1.1. Preferably, the amino alcohol and boronic acid are mixed with a solvent, preferably toluene and fractional distillation is carried out on the solution. Once all the volatiles are removed, preferably under reduced pressure, the residue is dissolved in a solvent, such as ether, for transfer to a fresh bulb for further distillation.

The solvent is preferably removed under vacuum and further distillation step is carried out, preferably at around 50°C and at a reduced pressure. The product fraction is collected at a temperature of greater than 150°C, preferably 160-180°C.

The method of the present invention may be adapted to provide a desired oxazaborolidine of a given formulae, depending upon the constituents (R¹) of the amino acid and the constituents (R²) of the boronic acid, as illustrated below:

Thus, the process of the invention may be used to prepare a wide variety of oxazaborolidines of the general formula (I) below:

(I)

wherein: n = 0, 1 , 2, 3;

R1 is selected from hydrogen, C1-C8 alkyl, C2-C8 alkenyl, aryl, and C3-C7 cycloalkyl; R2 is selected from hydrogen, C1-C8 alkyl, C2-C8 alkenyl, aryl, C3-C7 cycloalkyl, - (C=O)R, and -S(=O)2R, where R is selected from C1-C8 alkyl, aryl, and C3-C7 cycloalkyl;

R3 and R4 are each independently selected from hydrogen, C1-C8 alkyl, C2-C8 alkenyl, aryl, C3-C7 cycloalkyl and Benzyl (C6H5CH2-)

R5 and R6 are each independently selected from hydrogen, C1-C8 alkyl, C2-C8 alkenyl, aryl, and C3-C7 cycloalkyl; or one of R3 and R4 together with one of R5 and R6 form a substituted or unsubstituted aromatic ring, a substituted or unsubstituted cycloalkyl ring, or a substituted or unsubstituted heterocyclic ring, fused to the oxazaborolidine ring;

R1’ is selected from null, hydrogen, hydroxyl, C1-C8 alkyl, C2-C8 alkenyl and aryl;

R2’ is selected from null, hydrogen, C1-C8 alkyl, C2-C8 alkenyl and aryl; or R1’ together with R2’ are a group selected from -OR10 -, -O-(C=O)R10-, and -O- R10(C=O)-, wherein R10 is selected from a substituted or unsubstituted C1-C3 alkyl, a substituted or unsubstituted aryl and a substituted or unsubstituted heteroaryl, thereby forming a ring fused to the oxazaborolidine ring.

Preferably, n is null and R1 is H and R1 ’ is a cyclo alkyl, preferably a phenyl group. R2 is preferably H and R2’ is a C1-C6 alkyl group, preferably C1-C4 alkyl, especially methyl. One of R3 and R4 is preferably a benzyl group with the other being a hydrogen.

Detailed Description The present invention provides synthetic processes for preparing oxazaborolidines that removes the need to use ephedrine from the synthesis process. The ephedrine starting material that is used in prior art methods of producing oxazaborolidines is a precursor for a number of controlled substances, including amphetamines, and therefore it is beneficial to be provide synthetic methods that do not utilise this substance.

The method according to the present invention has been devised to remove ephedrine from the synthesis process. The oxazaborolidines are synthesised according to the invention by a three-step process as detailed below, using formylation, reduction and condensation steps, and with controlled chirality.

Example 1: Preparation of 4-Benzyl-3-methyl-2-phenyl-1,3,2-oxazaborolidine according to an embodiment of the invention.

This synthesis route uses the readily available safe starting material phenylalanine (or other amino acid) and works for both enantiomers of Phenylalanine (D & L) and also the racemic mixture (DL). Step 1 can be performed identically for both enantiomers and the racemic mix, whereas step 2 preferably requires some small changes to be made, such as the omission of the EtOAc quench for the D and L isomers in the reduction step.

Step l : Phenylalanine is formylated using a Formic-Acetic Anhydride

Mixture (FAM) in Formic Acid to give N-Formyl Phenylalanine.

1a,b,c - N-Formylation of Phenylalanine (D, or L or DL) with Formic-Acetic Anhydride

Mixture. Formic acid solvent

8

SUBSTITUTE SHEET (RULE 26) 14g of Acetic Anhydride was weighed into an oven dried 50ml round bottomed (RB) flask fitted with a magnetic stir bar. The flask was then fitted with a nitrogen (N₂) feed and a dropping funnel and cooled in an ice-water bath with stirring for 30 minutes. 8g of Formic acid (97%) was then placed in the dropping funnel and added drop-wise over 15 minutes whilst maintaining cooling. The ice-water bath was removed after 30 minutes and the mixture allowed to warm to room temperature (RT). The mixture was then heated in an oil bath at 60°C for 2hrs to equilibrate, followed by cooling. This produced Formic-Acetic Anhydride Mixture (FAM).

8.0414g of Phenylalanine (48.6mmol) was then weighed into an oven dried 100ml round-bottomed (RB) flask fitted with a dried magnetic stir bar. 16ml Formic acid (97%) was added and the mixture was cooled in an ice-water bath and stirred to give a clear white solution.

The Formic-Acetic Anhydride Mixture prepared above was then transferred to a dropping funnel fitted to the flask and, with ice cooling and stirring, the mixture was added to the Phenylalanine solution in Formic acid. Addition took approximately 15 minutes giving a clear solution. After 10 minutes, solids began to appear and the mixture was stirred with cooling for an additional 15 minutes. The ice bath was then removed and the mixture allowed to warm to RT. At 1 hr the reaction mixture had completely solidified with white solids. A sample taken at 1 hr30 total was then checked with thin layer chromatography (TLC), sample applied in 10:1 Acetone Water concentration approx 5mg/ml. The reaction was then cooled using an ice-water bath and quenched by cautiously adding 5ml water, then 50ml of water. There resulted a mass of white fine powder. These solids were filtered but were very slow and difficult to filter. The filtrate was stripped of volatiles under vacuum 50mbar approx to give wet white solids. These were dissolved in 80ml boiling water. On cooling; small brilliant white crystals deposited. These were easily filtered free and were washed with water and dried overnight.

TLC 12:3:5 n-Butanol, Acetic Acid Water of the crude solids and the recrystallised material showed 1 spot. No Phenylalanine.

The retention factor (rF) matches the earlier positive IR ID reference material. N-Formyl-DL-phenylalanine rF 0.71 Strong UV, no colour Ninhydrin DL-Phenylalanine rF 0.53 Weak UV, purple red fading to brown

Ninhydrin

The mother liquor from the recrystallisation was concentrated to 1/4 volume and deposited a small amount of solids on standing. These showed 2 spots and were discarded.

The combined solids were then vacuum dried at O.l mbar and 50°C for 12hrs.

Solids from initial filtration, 2.17g (11.2mmol), recrystallised solids from mother liquor=5.41g (28.0mmol) combined yield 7.58g (39.2mmol) 73% yield of N-Formyl-DL- Phenylalanine MW=193.2.

The synthesis of N-Formyl amino acids using Formic-Acetic Anhydride Mixture and Formic acid as a solvent is clean and rapid, avoiding the poor solubility of amino acids in inert solvents. The excess Formic acid causes faster decomposition of the mixed anhydride and greater CO generation but the reaction is complete before this is a problem. This is the method of choice.

Step 2: N-Formyl Phenylalanine undergoes reduction using Lithium

Aluminium Hydride (LAH) in 2MeTHF in an atmosphere of N₂. This gives N-Methyl Phenylalaninol.

2a: Lithium Aluminium Hydride (LAH) reduction of N-Formyl-DL-

Phenylalanine to N-Methyl-DL-Phenylalaninol (Racemic). 150ml of 2MeTHF, previously dried by standing for 48hrs over freshly activated 3a sieves, was transferred into an oven dried 500ml 3 neck flask, equipped with a mechanical stirrer, reflux condenser, dip temperature probe, sealed solids addition tube and inert gas (N2). The flask was vacuum purged until the 2MeTHF boiled at RT and then the vacuum broken with inert gas. This was repeated 2 times.

3.3g of powder LAH (87mmol approx) was weighed out into a dry 25 ml Erlenmeyer flask. 100mg of LAH was then added to the reaction flask, the slurry stirred and the hydrogen evolution allowed to slow before adding another 100mg portion of LAH. On the 3rd addition of LAH there was no additional hydrogen evolution seen in the bubbler, and the main portion of 3.3g LAH was added against a gentle nitrogen stream. The internal temperature was seen to rise slightly (overall deltaT 5°C). Once addition was complete the mixture was stirred giving a grey suspension.

4.9g (25mmol) of N-Formyl-DL-Phenylalanine prepared in step 1 that had been previously dried at 50°C and 10 mbar for 1 hr, was then added portion-wise using the sealed solids addition tube. There was a strong evolution of hydrogen as the acid reacted with the LAH. The addition rate was moderated and any frothing and foaming allowed to calm between additions. Addition took approximately 20 minutes and the internal temperature rose to 46°C, no cooling was applied. (On any larger scale cooling would be advisable). The solids addition setup was then washed with 50 ml dry 2MeTHF and the washings added to the reaction mixture. The resulting grey slurry was then stirred and heated using an external oil heating maintained at 70°C; internal temperature 65°C. A slight positive nitrogen pressure was maintained throughout using a line with oil bubbler.

The light grey suspension was stirred and maintained at 65°C overnight. At 20hrs run time, the reaction mixture was sampled. Approximately 0.8ml of reaction mixture was sampled and 0.2ml EtOAc was added to the sample to quench excess LAH. The sample was cooled in ice-water and approximately 0.5ml 5% NaOH solution was added to the sample. White solids drop out. The liquid was transferred and diluted to 2ml with 2MeTHF and analysed using gas chromatography-mass spectrometry (GC- MS). GCMS showed poor chromatographic performance. Adding BSTFA and incubating the sample produced better chromatography but a lot of activity was still present. TLC was carried out instead. Using the previous eluent and method 13-3-5 n-Butanol- Acetic acid-Water, this system was not ideal as it is relatively viscous and slow to develop.

The reaction sample showed no trace of N-Formyl-DL-Phenylalanine at rF 0.65 but showed a new single spot at rF 0.39.

The reaction mixture was allowed to cool to RT maintaining an inert atmosphere then cooled to 5°C internal using an external ice-water bath. An addition funnel was fitted and 100 ml dry 2MeTHF was added to dilute the slurry. With external cooling, 4 ml EtOAc was added drop-wise, there was a 13°C exotherm and Hydrogen evolution. The mixture was stirred, then 3.3ml H2O was added drop-wise, there was continued Hydrogen evolution. The mixture was stirred and then 3.3ml of 15% NaOH (1g NaOH dissolved in 3ml H2O cooled and made up to 6.6ml) was added drop-wise over 10 minutes. There was a temporary solidification and thickening of the mixture, which broke up slowly on stirring. Then 10ml H2O was added and the mixture vigorously stirred whilst cooling in with an ice-water bath. The mixture becomes easily stirrable and was stirred with cooling for a further 30 minutes. Stirring was stopped and the white solids allowed to settle. The mixture was then filtered, the initial filtrate was refiltered through the cake to remove carried fines. The cake was then rinsed with 2MeTHF 150ml, the washings and filtrate were combined; giving a clear, colourless solution. The 2MeTHF was quickly dried over MgSO4 filtered. The volatiles were then removed by distillation at reduced pressure, 220mbar and the residual light-yellow oil which was pumped at 30mbar and allowed to cool where it crystallised into yellow tinted white crystals. These weighed 3.4628g 20.9mmol (83.3% crude yield).

These were transferred to a Kugelrohr flask for distillation. The evaporation flask residue was rinsed with warm ether and on cooling this deposited white crystals which were added to the Kugelrohr bulb.

3.4620g of crude solids were distilled at 0.1 to 0.08mbar. After a small fore run at 50°C, the main fraction distilled as a white oil at 90-100°C 0.08mbar which crystallised in the receiving bulb.

The main fraction weighed 3.1311g. 81 mg remained as distillation pot residue. Distilled yield 3.13g (18.9mmol, 75.8% yield) of N-Methyl-DL-Phenylalaninol. 2- (Methylamino)-3-Phenylpropan-1-ol as bright white waxy solids.

Melting point was determined in duplicate, temperatures not corrected, shrinkage at 68°C melting complete 69°C. Duplicate 2 shrinkage at 69°C melting complete at 70°C slow heating.

Literature reported mp 68-70°C Karim, Montreux, Petit, Buono, Peiffer J Organomet Chem 1986 317 93

IR KBr NH sharp 3290 cm ¹

TLC, IPA NEt3 H2O 10 0.1 2 1 spot rF 0.40 tailing, TLC method not optimised.

TLC Hexane EtOAc NEt3 10 10 0.1 rF 0.31

GC-MS BSTFA derivatised poor chromatography MS spectrum matches expected mass spectrum for N/O monoTMS derivative, reaction incomplete 1 hr in AcN at RT.

2b: LAH reduction of N-Formyl-D-Phenylalanine to N-Methyl-D-

Phenylalaninol.

5.0633g (26.2mmol) of vacuum dried N-Formyl-D-Phenylalanine was weighed out. 3.4g of LAH was weighed out. 3 equiv. LAH = 78.6mmol. 3.3 equiv. = 3.3g LAH.

Following the previous method 150ml of sieve dried 2MeTHF was placed in a 500ml 3 neck RB flask, fitted with reflux condenser, mechanical stirrer temperature probe and solids addition funnel. The flask was inerted and degassed with nitrogen and 3.4g of LAH added portion-wise. Initial addition was very cautious there was again some initial H2 generation and a 10°C exotherm. Once the addition of further small amounts of LAH stopped generating gas, the remainder was added in one portion against gentle nitrogen flow. The slurry was then stirred in a cooled with an ice-water bath under N2.

With vigorous stirring the solid N-Formyl-D-Phenylalanine was added, portion-wise to moderate foaming, over approximately 15 minutes. Once the addition was complete and the slurry uniform, the cooling bath was removed and the reaction mixture heated and stirred maintaining positive inert gas pressure. Stirring and heating was continued for 23hrs. A sample was taken at this point quenched with H2O and subjected to TLC, the sample produced H2 on quenching. TLC showed no starting material and a single spot rF 0.44.

The reaction mixture was quenched and worked up in a different way to previously. The reaction mixture was cooled using an ice water bath with stirring. The EtOAc quench was omitted and instead 3.4ml H2O was added very cautiously over 20 minutes and there was vigorous Hydrogen evolution. 3.4 ml of 15% NaOH solution was then added, again cautiously drop-wise, then 11 ml H2O was added and the mixture stirred and allowed to warm to RT. After 30 minutes, the white solids were separated granular and easy to stir.

The mixture was filtered, the cake washed with 2MeTHF, the filtrate and washing combined and dried with MgSO4 filtered and the volatiles removed under vacuum. After pumping and heating at 30 mbar, the crude brown oil crystallised in a similar way to the earlier racemate.

These crude brown solids weighed 5.1g (33.7mmol) 130% crude yield.

With the aid of ether, the solids were transferred to a Kugelrohr bulb and the ether removed under vacuum. The contents of the bulb were pumped at <30mbar to give 4.3395g of off-white solids with yellow-brown staining. These were then distilled, following a fore-run. The main fraction, a white oil which crystallised in the receiving bulb, came over at an oven temperature of 80°C and 0.08mbar. This main fraction weighed 3.9163g (23.6mmol) giving N-Methyl-D-Phenylalaninol as slightly waxy white solids in 90.4% yield.

68.8 mg of residue remained in the still pot.

TLC of the distillate showed 1 spot rF 0.44 IR was identical to racemate.

Material appears optically active, levorotatory 1M in IPA, exact rotation subject to determination. 2c: LAH reduction of N-Formyl-L-Phenylalanine to N-Methyl-L-

Phenylalaninol.

Following the general method, 150ml sieve dried 2MeTHF was placed in an oven dried 3 neck flask arranged as before and the flask degassed and inerted with nitrogen. 5.0655g (26.2mmol) of dried N-Formyl-L-Phenylalanine was weighed into a 25ml conical flask arranged to act as a sealed solids addition device.

3.6g (93mmol, 3.6 equiv) of LAH was weighed out and added cautiously to the 2MeTHF, portion-wise initially to avoid excessive gas generation from reaction with residual water and the BHT stabiliser in the 2MeTHF.

The LAH slurry was stirred vigorously and was then cooled to 10°C using an external ice water bath. Using the solids addition device, the N-Formyl-L-Phenylalanine was added portion-wise over 20 minutes to minimise foaming due to gas generation.

The cooling bath was then removed and the mixture stirred and heated to 65°C internal, whilst maintaining a slight positive pressure of nitrogen. Heating and stirring was continued overnight. At 17hrs total time the reaction was sampled and the sample analysed by TLC. No starting material was seen. The reaction mixture was then allowed to cool to RT and then cooled with an ice water bath maintaining a nitrogen atmosphere throughout.

3.6ml of H2O was then added drop-wise over 15 minutes, there was gas evolution and a 7°C exotherm. 3.6ml of 15% NaOH solution was then added drop-wise, there was a continuing exotherm. This was followed by 50ml 2MeTHF to thin the slurry. Then 12ml of H2O was added and the mixture stirred and allowed to warm to RT. The mixture was stirred at RT for 30 minutes after which it had set up as white solids. The mixture was then vacuum filtered as before, the filter cake was washed with 3x 100ml 2MeTHF. The filtrate and washing were combined and dried with MgSO4 and filtered. The filtrate was then stripped of volatiles at 230-240 mbar and the residue pumped at 4mbar. The light brown oil residue crystallised and weighed 4.42g (102% crude yield)

These solids were transferred to the Kugelrohr and distilled. A fore run of white liquid weighed 100mg and came over at 60°C oven temperature and 0.1 mbar. The main fraction came over at 95°C at O.l mbar as a clear colourless oil which crystallised in the receiving bulbs. This main fraction weighed 3.8089g (23.04mmol) of N-Methyl-L- Phenylalaninol as white solids in 88% yield. TLC of this material against the racemate showed identical rF and one spot. The material is optically active and is dextrorotatory, opposite to the D material, exact rotation to be determined but 1M in IPA is approximately +180.

Step 3: N-Methyl Phenylalaninol is condensed using Phenylboronic acid to yield the product compound BNO-PI (4-Benzyl-3-methyl-2- phenyl-1,3,2-oxazaborolidine).

3a: 4-Benzyl-3-methyl-2-phenyl-1 ,3,2-oxazaborolidine.

Condensation of N-Methyl-DL-Phenylalaninol with Phenyboronic acid.

Phenylalaninol = 5 mmol = 0.8262g

1.1 eq Phenylboronic acid = 5.5 mmol = 0.6706g

20 ml Toluene. Reaction produces approx 180ul H2O on complete consumption of limiting reagent, (5mmol x 2 x 18). Excess Phenylboronic acid will condense to Phenylboroxine.

In a 50ml oven dried flask was fitted a short fractionating column with a thermocouple placed at the top followed by a dean-stark trap and reflux condenser. There was also placed a magnetic stir bar, followed by 0.8283g (5mmol) of distilled N-Methyl-DL- Phenylalaninol and 0.6807g (5.6mmol 1.1 eq) 98% Phenylboronic acid followed by 20 ml of sieve dried (1 week over activated 3a sieves) Toluene. The mixture was stirred and became cloudy with most of the solids dissolving. The mixture was then heated to reflux, the fractionating column was lagged to reduce heat loss. Shortly before the solvent began to boil all the solids had dissolved. An initial fraction with a boiling point of 104-106°C distilled and was collected in the dean-stark trap, after refluxing for 40 minutes the column temperature had risen to 110°C. Refluxing was continued for 1 further hour. At the 2hr point the reaction mixture was allowed to cool.

TLC using Hexane EtOAc NEt3 showed almost complete consumption of the starting aminoalcohol, phenylalaninol.

The volatiles were then removed under reduced pressure giving a pink-brown viscous oil which became a glass after continued pumping at 30 mbar. This was fully dissolved in 5ml warm Et20 and this solution was transferred to a Kugelrohr bulb and the ether removed under vacuum.

Kugelrohr distillation 50°C at 0.2mbar gave a small fore run of what appeared to be water. Then as heating was increased there was seen foaming and bubbling of the glassy solid which became liquid at approximately 100°C. At an oven temperature of 160-180°C and 0.15mbar, a viscous colourless oil distilled which set up as a sticky colourless transparent glass on cooling. At the same temperature the still pot contents solidified and remained solid. The glass distillate 500mg approx was very sticky and difficult to handle requiring transfer using diethyl ether into a beaker.

The ether was removed by evaporation at RT and the glass scratched. No sign of crystallisation was seen and repeated attempts to obtain seed crystals from a variety of solvents failed. Considerable material was lost due to the sticky nature of the glass. Eventually 257.9 mg of racemic 4-Benzyl-3-methyl-2-phenyl-1 ,3,2-oxazaborolidine was transferred by heating into a vial and retained as for further examination.

Distilled yield 40% but reduced by severe mechanical losses. A 2mg/ml sample in Et20 was subjected to GC-MS. H2 Carrier gas. The chromatography was less than perfect. The material produced a tailing peak with a mass spectrum showing 78 m/z, 160 m/z as the major fragment ions. No molecular ion was seen. Interpretation of the spectrum follows. Significant liner contamination was evident on repeated blank runs. Trace Phenylboroxine was seen but it was not clear whether this was due to sample decomposition in the inlet liner or was present in the injected sample.

HPLC analysis was also carried out using MeOH/water gradient elution C18 RP column UV detection 254nm. This showed one major peak. Purity by uncorrected peak area 89.1%.

IR thin film showed no sharp absorbance at 3260 cm-1 which was seen in the starting amino alcohol.

Melting point of the glass is unclear, the material becomes fully liquid at around 105°C On standing at 5°C under nitrogen for 3 weeks the colourless semi liquid glass was unchanged.

3b: 4-Benzyl-3-methyl-2-phenyl-1,3,2-oxazaborolidine. Condensation of N-

Methyl-D-Phenylalaninol with Phenyboronic acid.

Following the general method above, 0.8245g (5.0mmol) of N-Methyl-D- Phenylalaninol and 0.6712g (5.5mmol, 1.1 eq) was placed in an oven dried 50ml pear shaped flask and a magnetic stir bar added. 25ml sieve dried Toluene was added and a short column reflux condenser and a dean stark trap fitted. The mixture was heated in an oil bath until boiling. The solids all dissolved just before this point to make a cloudy mixture. The column head temperature slowly rose from 104°C to 110°C and the mixture cleared. The mixture was heated with stirring for 1.5hrs then cooled and the solvent removed under vacuum giving a foamy clear semi liquid glass weighing 2.08g. This glass was then heated to 60°C at 20mbar vacuum for 30 minutes, then 10mbar for a further 30 minutes. The glass became more solid. These sticky solids weighed 1.78g. 1.0501g of these sticky solids was transferred by scraping to a Kugelrohr bulb, the solids had become a white electrostatic powder and sticky flakes.

HPLC of the crude solids showed 1 major peak 94.5% by peak area. GC-MS showed one tailing peak with trace of Phenylboroxine, the Trimeric Phenylboronic Acid Anhydride.

1 .0501 g of these crude solids were distilled. Initial heating at 50°C and 0.08mbar which gave a very small initial fore run. The temperature was then increased to 100°C at 0.15mbar and the solids melted around 100°C. The melting point is not sharp. What appeared to be water appeared. The vapour pressure increased markedly causing an increase in the observed pressure. The clear colourless liquid was then slowly heated to 120°C and 0.1 Ombar and a small amount of condensation appeared in the receiving flask. This appeared to be water (CoCI2 paper). Continued heating at this temperature and pressure resulted in the still pot contents becoming increasingly solid. No further condensation was seen in the receiving bulbs. After 30 minutes, the temperature was increased to 150°C causing the still pot contents to melt once more and a clear oil to distil over after 20 minutes. The still pot contents once again solidified and remained solid even when the temperature was increased to 200°C.

It was suspected that there was chemistry occurring in the still pot at higher temperatures producing water. Absorbed water in the initial crude material should have distilled away at 50°C and 0.08mbar but water was seen at 100°C plus and 0.1 Ombar. The main fraction was a clear water white liquid which formed a skin and slowly solidified to a clear glass on cooling. This set up slowly as sticky white solids and was able to be mechanically removed to give 744mg (2.96mmol) of 4-Benzyl-3-methyl-2- phenyl-1 ,3,2-oxazaborolidine as white solids, 59% yield. Approximately 50mg was lost in transfer from the bulb.

The still pot residue weighed 117.7mg and consisted of brown tinted, crunchy solids. GC-MS of the distilled material, 2mg/ml in IPA showed one tailing peak with 78.0 m/z (100%) base peak 52.0 (23%) and 160.0 (20%) and no sign of phenylboroxine.

HPLC analysis of the distilled material also showed 1 peak >95% by peak area Melting point approximately 105°C-108°C but very unclear.

The optical purity of the material has not been accurately determined but the material is optically active.

Thus, the D enantiomer is produced, as shown below:

3c: 4-Benzyl-3-methyl-2-phenyl-1,3,2-oxazaborolidine. Condensation of N-

Methyl-L-Phenylalaninol with Phenyboronic acid.

Following the general method described previously, 0.8283g (5.0mmol) of N-Methyl-L- Phenylalaninol and 0.6726g (5.5mmol, 1.1 eq) was placed in a 50 ml flask. 25 ml dried Toluene was added and the mixture heated to reflux. The initial column head temperature was 100°C which rose to 110°C after 30 minutes. The mixture was initially cloudy but cleared as reflux was reached. The mixture was heated at reflux with stirring for 1 hr and then allowed to cool. The volatiles were removed at 70mbar giving glassy solids that were pumped at 40°C and 10mbar to give 1 ,43g, 113% crude yield of sticky white semi-solid which gradually set up as white slightly sticky solids. 250 mg of these solids was retained for examination and 795.4 mg of the crude solids was transferred to a Kugelrohr bulb for distillation.

The white solids were heated at 50°C and 0.06mbar for 30 minutes. A very small amount of liquid condensed in the receiver bulbs. The temperature was increased to 100°C at O.IOmbar. The solids melted to give a clear liquid. A vapour or gas was simultaneously produced which increased the pressure in the vacuum system. Condensation appeared in the receiver bulbs and the pressure of the system fell. The bath temperature was then increased to 120°C and a colourless oil began to distil and condense in the still pot. This oil formed a surface skin in the receiver bulbs, the bath temperature was increased slowly to 150°C at 0.06mbar by which time the oil had all distilled over. The still pot residue had solidified and consisted of 108.4mg of pinkbrown crunchy hard solids. The still pot residue was insoluble in Acetone in contrast to the crude material and the distillate.

The main fraction weighed 394.0 mg (some mechanical losses). The obtained 4- Benzyl-3-methyl-2-phenyl-1 ,3,2-oxazaborolidine as white soft solids in 31% distilled yield (49% overall) was analysed using GC-MS.

The mass spectrum of this material derived from the L enantiomer, N-Methyl-L- Phenylalaninol was identical to that previously obtained from the D enantiomer and the racemic N-Methyl-DL-Phenylalaninol, HPLC showed 1 peak. The material was optically active with the opposite rotation to the material derived from the D enantiomer N-Methyl-D-Phenylalaninol. Absolute rotation was not determined. Chiral HPLC Chiracel AD-H Normal phase was inconclusive due to poor peak shape and lack of separation. Absolute rotation and configuration to be determined.

Thus, the L enantiomer is produced, as shown below:

Example 2: Preparation of 3-methyl-2,4-diphenyl-1,3,2 oxazaborolidine according to another embodiment of the present invention.

Step l : Phenylglycine is formylated using a Formic-Acetic Anhydride

Mixture (FAM) in Formic Acid to give N-Formyl Phenylglycine.

N-Formyl-DL-Phenylglycine is produced by formylation of DL-Phenylglycine with Formic-Acetic Anhydride Mixture with Formic acid solvent.

A check of the solubility of in 97% Formic acid showed the solubility of DL- Phenylglycine in Formic acid to be slightly less than DL-Phenylalanine.

6.105g (40.3mmol) of DL-Phenylglycine was weighed into an oven dried 100ml RB flask and a dried stir bar added. 25ml Formic acid 97% was added and the mixture stirred whilst cooling in ice-water. A clear yellow brown solution resulted.

Formic-Acetic Anhydride Mixture was prepared as follows. To 13.4g ice cold acetic anhydride (Ac2O) there was added over 15 minutes 7.5g 97% Formic acid. The mixture was stirred at 0°C for 30 minutes. Cooling was removed and the mixture was allowed to warm to RT and then stirred whilst heating with an oil bath at 60°C (bath temp) for 2hrs to equilibrate the mixture. The mixture was then cooled to RT.

The Formic-acetic anhydride mixture was then added dropwise over 20 minutes to the Formic acid solution of DL-Phenylglycine with ice-water cooling. The clear mixture was stirred at 0°C for 30 minutes, then cooling was removed and the mixture stirred and allowed to warm to RT. Solids began to form 10 minutes after cooling was removed and the mixture solidified and became almost un-stirrable at 1 hr. A sample was taken at 1hr30 and examined by TLC. One spot rF 0.45 UV, no colour with Ninhydrin. No Phenylglycine.

At the 2hr point the reaction was quenched by cooling in ice-water and adding 4ml H2O (220mmol) and stirred with cooling for half an hour. The volatiles were then stripped by evaporation under vacuum 100 to 60 mbar to give yellow-white damp solids weighing 17.78g.

A small test sample was taken 10mg approx and recrystallised from 0.5 ml water giving fine white solids.

A second small test sample was recrystallised from boiling isopropyl alcohol (IPA) to give small white crystals.

The bulk material was dissolved in 50ml hot IPA giving an orange yellow solution, this was filtered and transferred to a beaker and cooled with stirring to give a slurry of fine white crystals.

These were left to age and then filtered to give a yellow filtrate and fine slightly off white solids which were washed with 10ml cold IPA and left to dry on the filter. These solids were sticky and weighed 8.16g >100% yield.

After drying overnight the solids remained sticky and weighed 7.18g. These solids were dissolved in 70ml of boiling water, which was cooled with stirring to give white solids. These were dried on the filter to give 5.63g of bright white crystals (31.4mmol) 78% crude yield.

After air drying for 48hrs there was obtained, 4.3905g of N-Formyl-DL-Phenylglycine as bright crystals 61% yield. TLC of these showed one spot rF 0.65, UV detection. No colour with Ninhydrin. Reference DL-Phenylglycine rF 0.46 tailing spot weak UV. yellow then orange red with Ninhydrin .

4.3530g of N-Formyl-DL-Phenylglycine was dried under a hard vacuum <0.1 mbar for 1 hr at 70°C to give 4.3463g of solids. 183.7mg Loss on drying 4.2%.

A TLC purity check after drying with N-Formyl-L-Phenylglycine as the reference, using nBuOH:HoAc:H20 12:3:5 as the eluent was carried out.

N-Formyl-DL-Phenylglycine one major spot rF 0.59 no colour with Ninhydrin faint spot at 0.39 no colour with Ninhydrin N-Formyl-L-Phenylglycine one spot rF 0.59. The reaction of step 1 is summarised below:

Step 2: LAH reduction of N-Formyl-L-Phenylglycine to N-Methyl-L- Phenylglycinol, (2S)-2-Amino-2-Phenylacetic acid.

Scaling, 3 eq LAH = 3x28.5mmol 86mmol= 3.24g >3.5g to ensure excess.

5.1108g (28.5mmol) of dried N-Formyl-L-Phenylglycine prepared in step (1 ) was weighed into a conical flask fitted with a length of flexible large bore thin PTFE tubing and a ground glass adapter, to allow simple controllable sealed solids addition through a neck on a 500ml reaction flask.

150 ml of 2MeTHF (dried over activated 3a sieves for 48hrs under N2) was added to a 500ml oven dried RB flask was fitted with a solids addition tube similar to above, a reflux condenser internal temperature probe mechanical stirrer and a nitrogen inlet.

4.2g of LAH (111 mmol 3.9 equiv) was weighed out and covered. A small amount of LAH was added cautiously in 100mg portions to the 2MeTHF maintaining an inert atmosphere. With good stirring, the remainder was added in portions against a gentle nitrogen stream. The internal temperature climbed to 35°C and external ice-water bath cooling applied.

Once the internal temperature was 5°C, the N-Formyl-L-Phenylglycine was added portion-wise as a solid with good stirring, pausing to allow the hydrogen generation to subside between additions. Ice-water cooling was applied throughout addition. The cooling was removed and then the mixture was heated with an oil bath at 70°C. The internal temperature was then held at 65-75°C and stirred under Nitrogen overnight.

24

SUBSTITUTE SHEET (RULE 26) At 23hrs a sample was taken hydrolysed and subjected to TLC nBuOH, HOAc, H2O which showed a new single spot, no SM.

The reaction mixture was then cooled maintaining a Nitrogen atmosphere to RT, then cooled to 10°C using external ice-water bath. Some of the excess hydride was then destroyed by the cautious addition of 10ml EtOAc diluted in 50 ml 2MeTHF. Moderate exotherm to 20°C and vigorous hydrogen evolution was seen during this addition. 4.4 ml of water was then added drop-wise followed by 4.4ml 15% NaOH solution and finally 13ml water. The mixture almost completely solidified after the final addition of water but broke up on continued stirring to give a white suspension.

The solids were filtered off and washed with 2MeTHF. The filtrate is a single phase. The washings were combined with the filtrate and quickly dried with MgSO4, filtered and the volatiles removed in vacuo to give 3.696g of faintly yellow oil, crude yield 85.6%.

Dilution of the reaction mixture using another volume of dry 2MeTHF before quenching the reaction appeared to assist in the formation of free filtering solids.

The crude product was then distilled a small fore run at 50°C at O.l mbar and a main fraction a viscous white clear oil at 80-100°C at O.lmbar. The main fraction weighed 3.2639g, 21.6 mmol, 75.7%.

The oil was triturated under warm ether and cooled and scratched. Unfortunately it did not show any sign of crystallisation. The oil was then fully dissolved in 1 ml warm ether and diluted with 10ml warm hexane giving a clear solution, on cooling the mixture clouded and product oiled out but no sign of crystallisation. Literature melting point 74- 75°C (Et20-petroleum ether) J. Org Chem 49 22 pp 4107 (1984).

The solvent was removed and the oil refrigerated for several weeks. TLC showed 1 spot rF 0.50

The reaction of step 2 is summarised below:

Step 3: Condensation of N-Methyl-DL-Phenylglycinol with Phenyboronic acid to produce 3-methyl-2,4-diphenyl-1,3,2 oxazaborolidine.

Thus, the present invention provides a three-step process for the cheaper production of oxazaborolidines with controlled chirality using improved starting materials.

Claims

CLAIMS: A method for the preparation of oxazaborolidines, the method comprising the steps of:

(iii) condensing the amino alcohol using a boronic acid to provide the desired oxazaborolidine end product. The method according to claim 1 wherein the amino acid is selected from phenylalanine and phenylglycine. The method according to claim 1 or claim 2, wherein the boronic acid is phenyl boronic acid. The method according to any one of the preceding claims wherein the formic-acetic anhydride (FAM) is prepared in situ for use in step (i) of the method. The method according to claim 4, wherein the preparation of FAM comprises mixing formic acid with acetic anhydride with cooling, optionally under an atmosphere of nitrogen, then heating followed by cooling to provide FAM. The method according to any one of the preceding claims wherein the amino acid is dissolved in formic acid, optionally with cooling, prior to the introduction of the formic- acetic anhydride (FAM) to produce the formyl amino acid. The method according to claim 6 wherein excess formic acid is used. The method according to any one of the preceding claims wherein a desired racemate or enantiomer of the formyl amino acid produced in step (i) is selected for

27 the reduction and condensation steps (ii) and (iii) for forming an oxazaborolidine with desired stereochemistry. The method according to any one of the preceding claims, wherein the reduction step (ii) is carried out in an atmosphere of nitrogen. The method according to any one of the preceding claims wherein the reduction step (ii) is carried out by boiling a tetrahydrofuran, preferably dried, under a vacuum and adding batches of lithium aluminium hydride thereto. The method according to claim 10 wherein excess lithium aluminium hydride is used. The method according to claim 10 or claim 11 , wherein the tetra hydrofuran is 2- methyltetrahydrofuran (2MeTHF). The method according to any one of the preceding claims wherein the condensation step (iii) is carried out using the amino alcohol and boronic acid in a ratio 1 :1-1 :1 .5, preferably 1 :1-1 : 1.2, more preferably 1 :1.1. The method according to any one of the preceding claims wherein in step (iii) the amino alcohol and boronic acid are mixed with a solvent and fractional distillation is carried out on the solution. The method according to claim 14 wherein the solvent is toluene. The method according to claim 14 or claim 15 further comprising removing all volatiles under reduced pressure, dissolving the residue in a solvent for further distillation. The method according to claim 16, further comprising removing the solvent under vacuum and carrying out further distillation, preferably at around 50°C and at a reduced pressure. The method according to claim 17 further comprising collecting the product fraction at a temperature of greater than 150°C, preferably 160-180°C.

. The method according to any one of the preceding claims for the preparation of oxazaborolidines of the general formula (I) below:

(I)

wherein: n = 0, 1 , 2, 3;

R2’ is selected from null, hydrogen, C1-C8 alkyl, C2-C8 alkenyl and aryl; or R1’ together with R2’ are a group selected from -OR10 -, -O-(C=O)R10-, and -O- R10(C=O)-, wherein R10 is selected from a substituted or unsubstituted C1-C3 alkyl, a substituted or unsubstituted aryl and a substituted or unsubstituted heteroaryl, thereby forming a ring fused to the oxazaborolidine ring. The method according to claim 19, wherein n is null and R1 is H and R1’ is a cyclo alkyl, preferably a phenyl group. The method according to claim 19 or claim 20, wherein R2 is H and R2’ is a C1-C6 alkyl group The method according to claim 21 wherein R2’ is methyl. The method according to any one of claims 19 to 22, wherein one of R3 and R4 is a benzyl group with the other being a hydrogen.