The last E-Newsletter of 2003 contains highlights from the Second International Conference on Industrial Applications of Biocatalysis and the Large Scale Chromatography Symposium and Exhibition, both held in Boston, Ma (USA) in September / October 2003.
Will Watson
This lively meeting (the second in the current series) was attended by 49 people, with 12 nationalities represented. There were 15 presentations including a Keynote presentation by Professor Alex Klibanov from MIT. Copies of the full proceedings can be purchased from Scientific Update. The conference was proceeded by a short course entitled given by Professor Andreas Liese on September 29th.
Copies of the full conference proceedings can be purchased from Scientific Update upon request.
Professor Liese’s talk was in three parts, the development of a lipase catalysed resolution of ethyl 2-hydroxy-4-phenylbutyrate, development of a reaction cascade to produce each of the 4 possible stereo-isomers of 1-phenyl-1,2-propanediol and finally a membrane based aeration process for regeneration of NADPH with hydrogenase, which is required in the reaction cascade. The main problem in resolving ethyl 2-hydroxy-4-phenylbutyrate is that the reaction is subject to product inhibition. This problem was solved by continuously removing the product as it is formed using a diafiltration reactor. This enables the productivity to be increased more than 5-fold.
The first intermediate in the cascade reaction, 2-hydroxy-1-phenylpropanone can be made from benzaldehyde and acetaldehyde using either benzaldehyde lyase or benzoylformate dehydrogenase depending on which enantiomer is required. In both cases thiamine diphosphate is required as a co-factor. An improvement to the overall process was realised by changing the reaction solvent from DMSO to MTBE / buffer which greatly simplified the downstream processing. Reduction of the keto-group is achieved using alcohol dehydrogenase and NADPH.
This presentation outlined work carried out at BMS on the synthesis of chiral intermediates for various developmental drugs. This included microbial reduction of 2-bromo-4-fluoroacetophenone, and the ketone group in 4-(2’-acetyl-5’-fluorophenyl)butyrate esters. The preparation of some unnatural amino acids was also described. Various biochemical routes to C6 amino acid intermediates (1, R2 = CH2OH, CH2NH2, CH(OR)2) such as (S)-6-hydroxynorleucine were described. The intermediates were coupled with the appropriate homocysteine derivative to provide the key intermediate (2) for cyclisation to generate the thiazepine core (3) of omapatrilat.
L-Phenylalanine was produced by a selective enzymatic deprotection of racemic N-Cbz-DL-phe alanine. This deprotection also worked as a very mild method of removing Cbz groups where standard hydrogenation approaches might be problematic. The synthesis of (S)-tert-leucine (by enzymatic reductive amination) and of amino-alcohol (4) (by a diastereoselective enzymatic reduction) was discussed, these compounds being intermediates for the production of atazanavir, which has recently been approved by the FDA.
Various biocatalytic routes for the making 4-vinyl-2,3-dihydrobenzofuran were investigated (the chemical route has recently been published (Org. Process Res. Dev., 2003, 7, 821). The final part concerned the enzymatic resolution of cis-3-acetyloxy-4-(1,1-dimethylethyl)-2-azetidinone, which forms the side chain of paclitaxel.
Mr Otey, from the group of Frances Arnold, discussed work on making unnatural P450 by combinatorial techniques for use in the oxidation of unactivated C-H bonds. There are various approaches to generating P450 chimeras, with a huge potential protein space to explore, but by breaking existing enzymes into smaller sections and recombining these the chances of finding active enzymes is increased and this is coupled with selective mutations that are less likely to disrupt the protein structure. The number of contacts broken by recombination is called the E value and the prediction is that chimeric sequences with low E values will be more likely to fold and so produce active enzymes. A rapid screening technique for activity on a variety of substrates was designed based on colorimetric differentiation. 8-(4-Nitrophenoxy)octanoic acid is one such substrate which is oxidised at C-8 to generate a hemi-acetal which breaks down under the test conditions to generate 4-nitrophenol, which reacts with 4-aminoantipyrine to produce a red colour. Chimeric P450s with specific activity for oxidation of 2-phenoxyethanol were found.
Dr Mink outlined DSM’s approach to using biocatalysis and integrating it with chemical and / or fermentation capabilities leading to efficient green processing routes. DSM have pioneered the PlugBug® concept, which seeks to maximise research and production efficiency by limiting the number of microorganisms used. This was exemplified by examples using DERA (2-deoxyribose-5-phosphate aldolase), threonine aldolases, and hydroxynitrile lyases. 5-Deoxy-sugars can be produced using DERA and this has been refined to produce STATON (5), which can be elaborated in to lipitor intermediate (6) or (S)-CDHP (7), a versatile C6 building block.
Threonine aldolases (TA’s) were used for coupling glycine with a variety of aldehydes to give –hydroxy--amino acids. The stereospecificity at the -carbon was described as “relaxed”! There are a large number of natural L-type TA’s, but only 3 D-type TA’s limiting the number of compounds with this configuration which can be made. Some phenylserine derivatives can be resolved with TA’s. Hydroxynitrile lyases were used to convert aldehydes to cyanohydrins enantiospecifically with HCN.
Professor Flitsch presented work on the use of enzymes in solid supported synthesis. Several aspects were covered including the choice of solid support, using enzymes to catalyse reactions between 2 substrates, one of which is supported, using enzymes to cleave compounds off solid supports, and analysis of enzyme reactions on immobilised substrates. Examples included the use of penicillin amidases to cleave linkers, thermolysin catalysed peptide synthesis and the use of fluorescence microscopy to follow enzyme reactions on immobilised substrates. This analytical technique correlated well with hplc data for the same reaction.
Dr Trauthwein described the various chiral technologies that have been developed at Degussa. Biotechnologies discussed included acylases (-amino acids) and an extension of this technology to the synthesis of –amino acids, alternative “ester” hydrolases for dynamic kinetic resolution such as sulphatases, hydantoinases, the use of lipases in conjunction with organometallic complexes to improve the yield by recycling the off-enantiomer, oxidoreductases for reductive amination of keto-acids. This latter case included modifications to formate dehydrogenase enzyme (required for co-factor regeneration) and improvements to the reaction conditions such as using biphasic media.
This presentation outlined Direvo’s approach to highlighting the usefulness of the Recombination Chain Reaction (RCRTM), which in combination with directed evolution allows the exchange of positive mutations. The key to fast development of biocatalysts by this method is to screen the catalysts under application-relevant conditions in order to precisely hit the intended application conditions. This requires a highly flexible platform with all relevant steps being fully automated and with a high degree of miniaturisation. The best results are obtained by working at the optimum of evolutionary progress by means of error-threshold mutagenesis (ETM), RCR and other DNA variation technologies.
The keynote presentation given by Professor Klibanov covered major developments in applied biocatalysis from 1980 to the present and some predictions of the future. Methods for enhancing the enantioselectivity of a given enzyme in a given reaction such as varying the solvent, modifying the history of the enzyme sample or temporarily enlarging the chiral substrate were described. So for example the activity and the stereoselectivity of the acylation of the pro-R or the pro-S hydroxyl of 2-(3,5-dimethoxybenzyl)propane-1,3-diol using chymotrypsin can be varied simply by changing the solvent. These differences arise because of the differential free energy of solvation in the transition states. The solvent dependence of the prochiral selectivity also depends on the method of preparation of the chymotrypsin (results with cross linked crystalline material, lyophilised or acetone precipitated chymotrypsin were presented). The enantioselectivity of the kinetic transesterification or selective acylation of racemic amino acid or derivatives can also be modified by forming a salt with a bulky counter ion. The selectivity of the acylation of 1-(hydroxymethyl)phenylacetic acid with vinyl acetate with P. cepacia lipase can be increased by a factor of 7 by using an adamantamine or quinuclidine salt as the substrate. In the final section of the talk, some examples of unnatural reactions, such as asymmetric sulfoxidation were presented including an example where changing the oxidant from hydrogen peroxide to tert-butyl hydroperoxide led to an inversion of the stereoselectivity.
Enzyme or, more generally, protein coated microcrystals (PCMCs) are formed by dissolving the enzyme in a concentrated solution of a crystalline material such as potassium sulphate and then co-precipitating by addition of a water miscible solvent such as propanol. The enzyme-activated sites can be titrated with benzylsulponyl fluoride and then quantified using Electrospray Ionisation Mass Spectroscopy. PCMCs in general show much enhanced reactivity over enzyme powders and retain their activity for long periods of time. In one case a PCMC that had been stored in propanol for 18 months still retained 85% of it’s original activity (this was discovered by accident!). PCMCs show higher activity than other enzyme preparations in almost all cases provided the enzyme is soluble in the salt solution to allow formation of a good PCMC. A comparison of a PCMC derived from Candida Antarctica lipase B was compared with industrially available preparations and found to be comparable with immobilised material and significantly better than a CLEC or lyophilised material.
Dr Dordick’s lecture concerned the use of enzymes on a microfluidic biochip for a variety of chemical transformations. This allows biocatalysis to be performed on small scale (nanomolar). In addition the use of microfluidic platforms to generate metabolic enzymes through in vitro reconstruction and manipulation can give unique products. This allows array-based biocatalytic systems to be used for the rapid synthesis and screening of bioactive compounds. This was exemplified by the oxidation of polyketide derived and / or polyketide like guaiacols.
This presentation concerned the development of an industrial process based on work carried out by Professor J-E. Bäckvall in Stockholm on the dynamic kinetic resolution of secondary alcohols. Typically this involves acylating one enantiomer of a racemic mixture in the presence of an enzyme and an organometallic catalyst (to racemise and recycle the off-enantiomer by oxidation / reduction). In this way yields up to 100% are theoretically possible as opposed to a maximum 50% in the absence of the organometallic catalyst. There were 3 issues to tackle from an industrial perspective – the acyl donor (4-chlorophenyl acetate was used by Bäckvall), the racemisation catalyst, and the amount of biocatalyst required. In the optimised process immobilised Candida Antarctica lipase (Novozyme 435 ®) is used as biocatalyst, ruthenium complex (7) is used, in place of the "SHVO" catalyst (8), for racemisation in toluene as solvent. Isopropenyl acetate, isopropyl acetate or isopropyl butyrate can be used as the acyl donor depending on the ester required.
The by-product from isopropenyl acetate (acetone) has to be removed as it is formed by distillation to avoid “transfer oxidation” of the product and potassium carbonate (0.2mol%) is also required to neutralise any acetic acid formed, as this forms an inactive ruthenium acetate complex. Under these conditions yields over 90% and ee’s >98% are achievable using either aromatic or aliphatic alcohols as substrates and this technology has been demonstrated on scale.
Dr Matcham discussed work carried out at Celgro to produce L-glufosinate a single enantiomer form of glufosinate, which is an established, systemic herbicide sold as a racemate (>1000Tpa). The target was to develop an efficient (maximum performance, minimum cost) biocatalytic transamination of keto-acid (9) to L-glufosinate (L-GA, 10).
Although existing Celgro biocatalysts showed excellent isopropylamine conversion activity, they showed no activity with PPO, so the biocatalyst had to be modified to accept new keto-acids such as PPO as substrates. After 5 rounds of mutagenesis that started with a thermally stable isopropylamine transaminase, an active catalyst was generated. The catalyst was formulated as a stable, spray-dried powder for ease of handling. The optimised process produces 250-300g/L product in 8 hours, with 99.9% conversion of the raw material (PPO) and less than 0.1% impurities. The biocatalyst is removed by ultrafiltration and the product L-GA is produced as a technical concentrate ready for formulation. In addition this technology has also been extended so that it can be applied to the synthesis of a wide variety of both (R)- and (S)-amino acids.
Professor Bornscheuer described some variations on standard uses of enzymes for dynamic kinetic resolution. For example better results can often be achieved in the resolution of carboxylic acid esters if vinyl esters rather than ethyl esters are used as substrates. The talk also included some work on recombinant pig liver esterases, which showed greatly enhanced activity, compared to commercially available enzymes. Directed evolution was used to modify various esterases in an attempt to produce an enzyme that would give good activity for the resolution of ester (11) to produce (12) a building block for the synthesis of epothilones.
A number of assays were described to allow rapid determination of synthetic activity in mutant organisms in high throughput screens. For example if the hydrolysis of an acetate derivative of an alcohol is being studied, using a hydrolase, one mole of acetic acid is produced for every mole of substrate converted. The acetic acid is then involved in the ATP / AMP cycle leading to the transformation of NAD+ in to NADH which can be quantified spectrophotometrically at 340nm. In the hydrolysis of vinyl esters, acetaldehyde is produced and this can be quantified by derivatisation with 4-hydrazino-7-nitro-2,1,3-benzoxadiazole (13), which is non-fluorescent, but reacts with acetaldehyde to produce hydrazone (14) which is fluorescent.
These assays work efficiently in a variety of organic solvents (except of course those that dissolve the micro titre plate!) and show high sensitivity with the detection limit being in the nanomolar range. The final part of the talk concerned work in progress on the rational design of hydrolases that exhibit activity on tertiary esters such as 3-phenylbut-1-yn-3-yl acetate, 3-methylpent-1-yn-3-yl acetate, and linalyl acetate.
This presentation focused on the technology that Proteus uses to develop novel biocatalysis, working at the interface of two worlds – biotechnology and chemistry. Proteus has access to a wide variety of natural enzymes and organisms from very different environments, e.g. deep sea hydrothermal vents, (hyper)saline environments such as the Dead Sea, oil fields, hot volcanic environments and hypersaline deep sea basins. These organisms provide the starting point for controlled mutagenesis to produce a pool of parental genes, which are then recombined using L-Shuffling™. The principle of L-Shuffling™ is that parent genes, are fragmented and the fragments are then subjected to a series of cycles of denaturation / hybridisation to templates/ ligation before finally removing the templates to generate new gene diversity. An HTS screen called CLIPS-O™ was also described and this has been published (Org. Process Res. Dev., 2002, 7, 441).
Dr Basch presented work carried out at BMS to improve the overall yield of Cephalosporin C (Ceph-C) in the fermentation of Acremonium chrysogenum. Significant amounts of Ceph-C are lost by non-enzymatic degradation under the fermentation conditions (pH 4-7) to compound X. This can be minimised by stabilising the Cephalosporin nucleus by converting Ceph-C to deacetylCeph-C. This conversion can be achieved by adding Rhodosporidium toruloides whole cells to the fermentation and results in the accumulation of deacetylCeph-C. However this also requires additional fermentation to produce R. toruloides and extra downstream processing. But this problem was overcome by cloning the gene for the active ester (Ceph-C esterase) and then using it to modify A. chrysogenum. The final hurdle is to convert deacetylCeph-C in to Ceph-C and this can be achieved either chemically using isopropenyl acetate / dicyclohexylamine or biochemically using Cephalosporin C esterase with isopropenyl acetate. Overall the direct fermentation of Deacetylcephalosporin C results in a 35-40% increase in the Cephalosporin nucleus recovered from the fermentation broth of A. chrysogenum.
Peter Spargo
The second Large Scale Chromatography Symposium was attended by some 70 people from 8 different countries. There were 15 different speakers, and a number of companies manned stands in the adjoining exhibition room. A brief summary of the presentations is given below.
Copies of the full conference proceedings can be purchased from Scientific Update upon request.
Dr Hsu provided a general overview of how GSK manages preparative chromatography projects, putting what is technically achievable (e.g. using simulated moving bed chromatography) into the context of pharmaceutical development programmes, and emphasising that chromatography can be the key to speed in early development.
Counter Current Chromatography (CCC) has been known and used as a laboratory tool for many years. “Dynamic Extraction” is a recent development of CCC which provides a robust, rapid, scalable and simple liquid-liquid separation process. It consists of a sample, tubing and two immiscible solvents. The tubing (coil) is wound onto a drum (bobbin). The bobbin is centrifugally rotated in planetary motion. Coils are filled with stationary (liquid) phase, and mobile (liquid) phase is pumped over it. The planetary motion ensures volume retention of the stationary phase against the flow of the mobile phase and sets up alternating zones of mixing and settling providing high-resolution separations based upon the distribution ratio between the two phases. Separations take minutes rather than hours, and is more efficient on scale-up. A pilot scale facility is currently being constructed.
Five preparative chromatography case studies in development were presented, including applications of simulated moving bed and reverse phase chromatography. Of note was a reverse phase batch chromatography study in which substitution of an HCl salt with the corresponding TFA salt prevented “breakthrough” on the column, when running in a TFA-containing eluent. Stability studies on a chiral stationary phase were also described, indicating a lifetime of at least two years. (Some manufacturers claim this to be a conservative estimate.)
Dr Suteu gave a detailed description of Chiral Technologies’ approach to chiral separations. Key aspects were selection of chiral stationary phase, mobile phase selection, and optimisation of parameters. She reported that although they had invested more than 10 man years into efforts to understand the mechanism of chiral recognition, using X-Ray, NMR and molecular mechanics, this had met with limited success. This is not entirely surprising, and clearly leaves us with the fact that identification of a successful chiral separation is still heavily reliant upon empirical screening!
Mr Lumley-Wood outlined the specific challenge of scale which faces discovery/medicinal chemists, specifically that discovery is “large scale” in terms of the numbers of different compounds (as opposed to larger quantities of a single compound). A “platform approach” to chromatographic purification was described encompassing normal and reverse phase chromatography, supercritical fluid chromatography, flash chromatography and counter current chromatography. Also presented was a parallel purification system capable of purifying over 15,000 compounds per month.
Dr Ma described a proprietary development of simulated moving bed (SMB) chromatography, which combines the advantages of SMB (continuous separation process, high productivity, low production cost) with the advantages of batch chromatography (simultaneous separation of multiple components from complex mixtures). The specific workings of this technology were not fully disclosed for intellectual property reasons, but a key aspect is the “focussing” (i.e. enrichment as opposed to dilution) of minor components during separation.
Solvent management is, of course, a vitally important aspect of large scale chromatography, without which this technology would be completely impractical. Dr Giberti reviewed the application of thin film evaporation technology including rising film, falling film, disc and tube, wiped/scraped film, both horizontally and vertically.
Mr Blehaut outlined the Varicol® continuous chromatography system and how it differs from standard SMB - the key difference being that simulation of a counter-current steady state process is NOT a requirement. Simultaneous switching of inlet and outlet lines is not needed, and therefore the number of average column/zone configurations is essentially unlimited. This leads to higher productivity and lower dilution, and requires less chiral stationary phase (i.e. less columns). The largest Varicol system uses 1 metre diameter columns and processes up to 200 Mtons per year. (He also showed a photograph of the world’s largest HPLC column, which is 1.6m diameter, 4m long and contains 3,500kg packing material.) A numerical optimiser simulation software was described which combines physical data (isotherms, pressure drop etc) with constraints (purity, recovery yield etc) to accurately predict the performance of a large scale separation.
Mr Blehaut also described the integration of chiral chromatography with other tools in the “chiral toolbox” (i.e. chiral pool, asymmetric synthesis, and diastereomeric crystallisation), the most obvious one being the use of crystallisation to enhance the enantiomeric purity of material separated by chiral chromatography.
Dr Dapremont took a stepwise diagrammatic approach to describing the principle of multi-column chromatography/(SMB), then reviewed the scale-up issues and finally discussed the fate of impurities, demonstrating that SMB can be more than simply a binary separation tool, and that in certain cases, impurities can be separated out, albeit at the cost of higher eluent consumption and lower throughput.
Dr Chan described some case histories of preparative chromatographic separations from BMS development programmes, including that of a chiral alcohol in which the costs of chiral chromatography vs enzymic resolution vs enzymatic hydroxylation were compared. The chromatographic option was the cheapest.
Dr Dutta presented a couple of case studies (a decapeptide and a diastereomeric oxirane), and briefly mentioned “KromaguideTM 3.0”, a computerised support system, which basically takes information from laboratory scale separations and combines this with equipment restrictions, cost factors and physical properties, and predicts optimised running conditions for best economy on large scale. A couple of notable conclusions were: (i) Repetitive mass-overloaded injections should be seriously considered as an industrial high throughput method. (ii) For small molecules, capacity is maximised with a narrow pore size distribution in the solid support.
Dr Bruno presented his cost calculations relating to the hypothetical production of homochiral naproxen by simulated moving bed chiral chromatography, concluding that a cost of less than $50/kg should be achievable within a couple of years.
This was a nice overview of Finorga’s successful application of simulated moving bed chromatography for the separation of enantiomers, which included one of the clearest depictions (in a single slide) of the principle of multi-column chromatography. Finorga are now capable of running an SMB enantiomer separation at up to 200 tonnes per annum.
In an interesting review of silica gel properties and manufacture, the importance of pH, surface area, particle flow and dry packing properties to the effectiveness of chromatographic separations was highlighted, and new and efficient particulate and monolithic silica sorbents were described. Full disclosure was not possible, owing to intellectual property issues, but a new silica (known as “Pharmprep®”), which is made from alkoxysilanes, has superior physical parameters which can be easily tuned.
Protein-A chromatography is widely used in the production of monoclonal antibodies. Dr McCue described how the dynamic capacity of a Protein-A column was significantly increased by increasing the surface area of the porous glass backbone. It was also shown that columns can be successfully scaled in both diameter and height (conventionally, increasing the diameter is the focus for chromatography scale-up.)