See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/229523252

Lonza: 20 Years of Biotransformations

Article in Advanced Synthesis & Catalysis · July 2003

DOI: 10.1002/adsc.200390049

CITATIONS READS

161 2,198

3 authors, including:

Nicholas M Shaw
Lonza Bioscience
31 PUBLICATIONS 918 CITATIONS

SEE PROFILE

All content following this page was uploaded by Nicholas M Shaw on 28 August 2019.

The user has requested enhancement of the downloaded file.

 REVIEWS
Lonza: 20 Years of Biotransformations
Nicholas M. Shaw,* Karen T. Robins, Andreas Kiener
Biotechnology Research and Development, Lonza AG, 3930 Visp, Switzerland
Phone: ( ‡ 41)-27-948-5937, Fax: ( ‡ 41)-27-947-5937, [email protected]

Received: December 6, 2002; Accepted: January 20, 2003

Abstract: Biotransformations are an important tool in 3 Reduction

organic synthesis, especially for the production of 3.1 (R)-Ethyl 4,4,4-Trifluoro-3-hydroxybutanoate
chiral molecules, and where chemical reactions are 4 Hydrolysis
not possible or inefficient. Lonza is a major custom 4.1 Nicotinamide
manufacturer of intermediates for the life science 4.2 (S)-Pipecolic Acid [(S)-Piperidine-2-carboxylic
industries, and uses biocatalysis in many of its Acid]
processes. A synthetic route may have one or more 4.3 CBZ-d-proline [(R)-N-CBZ-proline]
biotransformation steps. This article gives a number 4.4 (1R,4S)-1-Amino-4-hydroxymethylcyclopent-2-
of examples of syntheses that involve biotransforma- ene
tions for the production of both achiral and chiral 4.5 (S)-2,2-Dimethylcyclopropanecarboxamide
intermediates, at development and production scales. 4.6 (S)-3,3,3-Trifluoro-2-hydroxy-2-methylpropionic
For each example the reasons and advantages of using Acid
a biotransformation are given. 5 Multi-Enzyme Biotransformations
5.1 l-Alaninol
1 Introduction 5.2 l-Carnitine
2 Oxidation 6 Conclusion
2.1 Oxidation of Alkyl Groups on Aromatic Hetero-
cycles: 5-Methylpyrazine-2-carboxylic Acid and
Pyridyl-3-acetic Acid Keywords: biotransformations; hydrolysis; Lonza;
2.2 Regiospecific Hydroxylations of Aromatic N- multi-enzyme biotransformations; oxidation; reduc-
Heterocycles: 6-Hydroxynicotinic Acid tion

1 Introduction processes involving biocatalysis have been developed

for both chiral and non-chiral molecules.
Biotransformations are enzyme-catalysed conversions The possible synthetic routes for each target molecule
of non-natural substrates to products. They are an are evaluated in order to find the best overall route. The
important tool in organic synthesis, especially for the availability of starting materials, the number of steps
synthesis of chiral molecules, where the reactions involved, environmental considerations, scalability, de-
catalysed may be asymmetric syntheses or the resolution velopment time, product quality, and down-stream
of racemates. The main advantages associated with the processing are all taken into consideration. The route
use of single enantiomer compounds are increased of choice may be chemical, or a combination of
specificity and the avoidance of adverse side effects. chemistry and biocatalysis. For success, close coope-
Biotransformations are also used for reactions to achiral ration between chemistry and biotechnology is essen-
molecules where a chemical step would not be possible, tial.
or where the biotransformation has advantages. Enzymes for biocatalysis are used in a number of
Lonza×s biotechnology R&D group was founded in forms. They may be wild-type, or recombinant, or
1983 and consisted of three people. Since then the genetically modified to increase their specificity or
number of people engaged in biotechnology has risen to activity. One or more enzymes that carry out the
over 800. These people are employed in the production required steps may be present in whole cells, which
of pharmaceutical and agrochemical intermediates by may be growing, resting or immobilised. Alternatively,
fermentation and biocatalysis, and in the production of cell-free enzymes may be used in solution, in a
biopharmaceuticals with both microbial and mamma- membrane reactor, as a suspension, cross-linked, or
lian cells. Biotransformations have played a prominent immobilised. The medium for the reaction may be
role in the development of biotechnology in Lonza, and aqueous, organic or two-phase. For general reviews and

Adv. Synth. Catal. 2003, 345, No. 4 ¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/03/34504-425 ± 435 $ 17.50+.50/0 425
 REVIEWS Nicholas M. Shaw et al.

2 Oxidation
Nicholas Shaw (born 1955)
has been working in Lonza×s
2.1 Oxidation of Alkyl Groups on Aromatic
Biotechnology R&D Depart-
Heterocycles: 5-Methylpyrazine-2-carboxylic Acid
ment since 1987. He received
and Pyridyl-3-acetic Acid
his PhD in biochemistry from
the University of Wales,
Swansea in 1980, and was a Chemical oxidation reactions of heteroaromatic com-
post-doc at the Smithsonian pounds with one or more alkyl groups are generally
Institution in Washington unspecific and lead to the formation of by-products.
D.C, and the University of However, microorganisms grown on xylene or toluene
East Anglia. His interests are as their sole source of carbon and energy can selectively
in enzymology, biotransformations, and biopharma- oxidise single methyl groups on aromatic heterocycles,
ceutical production. using the enzymes of the xylene degradative pathway.[9]
5-Methylpyrazine-2-carboxylic acid (MPCA) is an in-
Karen Robins (born 1959) termediate for the production of Acipimox, which is an
studied science at the Univer- antilipolytic drug, and Glipicide, an antidiabetic drug.
sity of Sydney and has been The oxidation of 2,5-dimethylpyrazine (DMP) to
working in the biotechnology MPCA is carried out with Pseudomonas putida (Fig-
research department of Lonza ure 1). High product concentrations and yields are
since 1984. She completed her obtained by carrying out the biotransformation with
Masters Degree in Applied live cells growing on xylene.[10] DMP is added continu-
Science in Biopharmaceuti- ously to the fermentation broth up to 2 g/L. Higher
cals (University of New South concentrations of DMP are toxic for the biocatalyst. No
Wales) in 2001, and has just bacterial degradation products from xylene, which could
finished a 5 year stint at the interfere with product isolation, are detected in the
Lonza Guangzhou niacinamide plant in the P. R. fermentation broth. MPCA is precipitated by acidifying
China. Her interests are in biotransformations and the cell-free fermentation broth. Lonza produces
biopharmaceuticals. MPCA at the 15 m3 scale with a product concentration
of 24 g/L and an analytical yield of > 95%.
Andreas Kiener (born 1955) In an analogous way, a second group of microorgan-
received his PhD in 1986 in isms grown on n-octane can be used for the terminal
microbiology/biochemistry at oxidation of ethyl groups to the corresponding acetic
the ETH in Z¸rich. After- acid derivatives. For example, Pseudomonas oleovorans
wards he joined the group of catalyses the oxidation of 5-ethyl-2-methylpyridine to 6-
Christopher Wash at MIT for methylpyridyl-3-acetic acid, and illustrates the selectiv-
a postdoctoral fellowship un- ity of this reaction: no by-product with an oxidised
til 1988. Since then he has methyl group in position 2 was detected (Figure 2a).
been working at Lonza as a Lonza has used this reaction[9] for the conversion of 3-
Senior Scientific Researcher ethylpyridine to pyridyl-3-acetic acid (Figure 2b), which
in the Biotechnology R&D is an intermediate for the preparation of Risedronate, a
Department, where his main interest is in biotrans- drug for the treatment of osteoporosis.
formations.

monographs on biotransformations and their industrial

applications the following references are recommended
as a starting point.[1±8]
A number of examples are described below for
processes that include one or more biotransformation Figure 1. The oxidation of 2,5-dimethylpyrazine to 5-methyl-
steps. For each example the rationale for using a pyrazine-2-carboxylic acid by cells of Pseudomonas putida
biotransformation is given. growing on xylene.

426 Adv. Synth. Catal. 2003, 345, 425 ± 435

Lonza: 20 Years of Biotransformations REVIEWS

Achromobacter xylosoxydans LK1 strain. The optimum

temperature for the biotransformation is 30 8C and
aeration is necessary. The yield after the biotransforma-
tion is > 99%. The down-stream processing consists of
ultrafiltration of the biomass and adjustment of the
product solution pH to 1.8 with H2SO4. The 6-hydroxy-
nicotinic acid precipitates at this pH and the white
crystals are then filtered from the solution. The crystals
are washed and then dried under vacuum. The yield of
isolated 6-hydroxynicotinic acid is about 93% and the
purity > 99%.
The degradation of nicotinic acid is described in the
literature[16] and is shown in Figure 3. The second
enzyme of this pathway, 6-hydroxynicotinic acid hy-
droxylase is inhibited by > 1% nicotinic acid[13] while
Figure 2. Ethyl group oxidation by cells of Peudomonas the first enzyme, nicotinic acid hydroxylase is unaffect-
oleovorans. ed. This phenomenon is exploited in both steps of the
process. In the first step where biomass production is
critical the nicotinic acid concentration is kept low to
ensure that it is completely catabolised and no accumu-
2.2 Regiospecific Hydroxylations of Aromatic
lation of 6-hydroxynicotinic acid occurs. In the second
N-Heterocycles: 6-Hydroxynicotinic Acid
step the reverse situation, accumulation of 6-hydroxy-
nicotinic acid, is desired. In this step the nicotinic acid
The regiospecific hydroxylation of aromatic N-hetero- concentration is kept high ensuring the inhibition of the
cycles is difficult to carry out chemically, but progress second enzyme of the pathway which results in the
has been made in this field with biotransformations.[11] accumulation of 6-hydroxynicotinic acid.
Lonza have developed a number of biotransformations Lonza has successfully produced more than 10 tons of
for this type of reaction to produce 5-hydroxypyrazine- 6-hydroxynicotinic acid with this process.
carboxylic acid, 6-hydroxypicolinic acid and 6-hydroxy-
nicotinic acid.[9,12,13]
The starting product for the biosynthesis of 6-hydroxy-
nicotinic acid is the vitamin nicotinic acid. Nicotinic
acid is produced in large quantities by Lonza which
ensures both the availability and the competitive pricing
of the starting material. 6-Hydroxynicotinic acid is a
versatile building block. It is especially useful for the
production of insecticides.[13] 6-Hydroxynicotinic acid
can be synthesised chemically[14] but the formation of by-
products due to non-specific hydroxylation of the
pyridine ring and the complex down-stream processing
that is required for their separation from the product
increases costs, making the chemical process uneco-
nomic. Conversely the biohydroxylation is regiospecific
and therefore avoids the above-mentioned problems of
the chemical synthesis, making it an attractive alterna-
tive.
The enzymatic production of 6-hydroxynicotinic acid
occurs in two steps.[15] It is possible to carry out both
steps in the same fermenter. Firstly, Achromobacter
xylosoxydans LK1 (DSM 2783) biomass possessing a
highly active nicotinic acid hydroxylase is produced. To
achieve a high level of induction of this enzyme nicotinic
acid is used as the carbon and nitrogen source as well as
the enzyme inducer. In the second step the biomass is
used for the hydroxylation of the nicotinic acid. A
solution of nicotinic acid that has been neutralised with Figure 3. The degradation of nicotinic acid by Achromobacter
NaOH is added directly to the reactor containing the xylosoxidans.

Adv. Synth. Catal. 2003, 345, 425 ± 435 427

 REVIEWS Nicholas M. Shaw et al.

3 Reduction 4 Hydrolysis
Lonza uses a number of hydrolytic biotransformations
3.1 (R)-Ethyl 4,4,4-Trifluoro-3-hydroxybutanoate
to produce both achiral molecules, for example, nicoti-
namide, and a number of chiral molecules. Enzymes that
(R)-Ethyl 4,4,4-trifluoro-3-hydroxybutanoate is a build- metabolise nitriles and amides (nitrile hydrolases and
ing block for pharmaceuticals such as Befloxatone, an amidases) have found many applications.[21] Nitrile
antidepressant monoamine oxidase-A inhibitor from hydratases are not usually enantiospecific, whereas
Synthelabo.[17] The process utilises whole cells of Es- amidases can be (Figure 5). Nitrilases, which catalyse
cherichia coli that contain two plasmids.[18,19] One carries the conversion of nitriles directly to the corresponding
an aldehyde reductase gene from the yeast Sporobolo- carboxylic acids, were generally thought not to be
myces salmonicolor, which catalyses the reduction of enantiospecific, but recently several have been discov-
ethyl 4,4,4-trifluoroacetoacetate, and the second carries ered by modern cloning and screening techniques.[22]
a glucose dehydrogenase gene from Bacillus megateri-
um to generate NADPH from NADP‡ (Figure 4). The
strain was originally constructed to catalyse the stereo- 4.1 Nicotinamide
selective reduction of ethyl 4-chloro-3-oxobutanoate to
ethyl 4-chloro-3-hydroxybutanoate, a precursor for L- Nicotinamide is an essential nutrient in animal and
carnitine synthesis.[20] Productivities of up to 300 g/L, human nutrition. The common source of nicotinamide
and ee values of up to 92% were reported for the ethyl 4- in human and animal diets is the consumption of
chloro-3-hydroxybutanoate process. vegetables or plant matter. Nicotinamide was shown to
The Lonza process is carried out in a water/butyl cure the skin disease, pellagra[23] hence it is also known
acetate two-phase system to avoid inhibition of the as vitamin PP (pellagra preventing). Nicotinamide is a
reductase by the substrate and product. An advantage of component of vitamin premixes for many animal diets
the two phase system is that the cells are permeabilised, promoting health and growth of the animal. In the area
allowing the transfer of NADP‡ and NADPH through of human nutrition it is used as an additive for refined
the cell wall. Cells of Escherichia coli JM109 containing flour, breakfast cereals and in multi-vitamin prepara-
the two plasmids are grown at 22 8C to avoid inclusion tions.
body formation. The cells are then washed and stored There are several chemical processes for the produc-
frozen before use in the biotransformation. The product tion of nicotinamide.[24] One synthesis starts with the
has an ee value of > 99%, and the yield is about 50%. reaction of acetaldehyde, formaldehyde and ammonia
Lonza is the world×s leading manufacturer of ethyl 4,4,4- to produce a mixture of 3-picoline and pyridine. The 3-
trifluoroacetoacetate, so there is optimal backward picoline is converted to 3-cyanopyridine by ammoxida-
integration of the process. tion. The 3-cyanopyridine is then converted to nicoti-
NADP‡ is both essential and a major cost factor in the namide by alkaline hydrolysis. The disadvantages of this
biotransformation, so securing a supply at a reasonable process are low yields and the production of a significant
price was essential. The concentration of NADP‡ amount of nicotinic acid as a side-product which
required for maximum activity is about 0.5 g/L. Experi- complicates the down-stream processing. Until 1998
ments with a number of other esters showed that the Lonza also produced nicotinamide by chemical syn-
isopropyl ester of 4,4,4-trifluoroacetoacetate is at least thesis. The first step of the Lonza synthesis was the
as good a substrate as the ethyl ester. conversion of acetaldehyde and ammonia to 2-methyl-
5-ethylpyridine (MEP). MEP was then oxidised with
nitric acid to nicotinic acid. A batch amidation of the
nicotinic acid and several batch crystallisations pro-
duced pure nicotinamide. This Lonza process was no

Figure 4. The stereoselective reduction of (R)-ethyl 4,4,4,-

trifluoro-3-hydroxybutanoate by an aldehyde reductase from
Sporobolomyces salmonicolor in Escherichia coli. AR ˆ al-
dehyde reductase; GDH ˆ glucose dehydrogenase. Figure 5. Nitrile- and amide-metabolising enzymes.

428 Adv. Synth. Catal. 2003, 345, 425 ± 435

Lonza: 20 Years of Biotransformations REVIEWS

Figure 6. The new Lonza nicotinamide process.

longer competitive and full capacity had been reached. treatment of cancer multi-drug resistance[26] and the
For these reasons Lonza set about developing a new local anaesthetics Naropin (ropivacaine) from Astra-
process. Zeneca[27] and Chirocaine (levobupivacaine) from Chi-
The new Lonza process consists of 4 highly selective, roscience.[28] The Lonza process efficiently combines
continuous, catalytical reactions (Figure 6). The first chemistry and biotechnology to produce (S)-pipecolic
three steps are carried out using chemical catalysis. acid with an ee value of > 99%.[29,30,31]
These steps are carried out at temperatures above In the first step whole cells containing a nitrile
300 8C and in the gas phase. The last step, the biohy- hydratase convert 2-cyanopyridine to pyridine-2-car-
drolysis using a nitrile hydratase enzyme is the simplest boxamide, which is then chemically hydrogenated to
step. It is highly selective ( > 99%) and carried out under (R,S)-piperidine-2-carboxamide. This compound is the
very mild aqueous conditions. In this step the 3- substrate for an amidase-catalysed enantiomer resolu-
cyanopyridine is mixed with buffer and continuously tion with whole cells of Pseudomonas fluorescens
dosed into the reactor containing Rhodococcus rho- DSM9924. The product, (S)-pipecolic acid, is isolated
dochrous J1 cells immobilised in polyacrylamide. The by precipitation at acid pH (Figure 7).
crude solution of nicotinamide exiting the reactor is The production strain was selected by growth with the
decolourised, the bioburden is removed by nanofiltra- racemic substrate as the only source of nitrogen. Strains
tion and the solution is concentrated. In the final step that hydrolysed about 50% of the substrate, as analysed
nicotinamide is isolated as a granular, free-flowing, non- by thin layer chromatography, were chosen for further
caking, white solid. The advantages of this process are investigation to determine the stereospecificity of the
high yield, high energy efficiency, only one organic amidases.
solvent is used and the process water, ammonium and The pharmaceutical intermediates (S)- and (R)-piper-
hydrogen are recycled. The process is also environ- azine-2-carboxylic acid (Figure 8) are also produced by
mentally friendly and safe. Lonza using similar processes.[29,30,31]
The Rhodococcus rhodochrous J1 strain was isolated
by H. Yamada and T. Nagasawa.[25] This strain was then
immobilised in polyacrylamide by Nitto for commercial 4.3 CBZ-d-proline [(R)-N-CBZ-proline]
use. Lonza produces over 3,500 tonnes of nicotinamide
per year with this biocatalyst. A further example of a cyclic amino acid that is
produced by Lonza for use as a pharmaceutical inter-
mediate is CBZ-D-proline,[30,32] which is used for the
4.2 (S)-Pipecolic Acid [(S)-Piperidine-2-carboxylic synthesis of Eletriptan, a drug from Pfizer for the
Acid] treatment of migraine.[33]. l-Proline is chemically racem-
ised and derivatised to give racemic N-CBZ-proline,
(S)-Pipecolic acid is a building block for a number of which is then stereospecifically hydrolysed by a proline
pharmaceuticals, such as Incel from Vertex for the acylase in a strain of Arthrobacter sp. (Figure 9). This

Adv. Synth. Catal. 2003, 345, 425 ± 435 429

 REVIEWS Nicholas M. Shaw et al.

Figure 7. Process scheme for the production of (S)-pipecolic acid.

Figure 8. Process scheme for the production of (S)- and (R)-piperazine-2-carboxylic acid.

Figure 9. Process scheme for the production of CBZ-d-proline.

strain was newly isolated from soil samples using 4.4 (1R,4S)-1-Amino-4-hydroxymethylcyclopent-2-ene
enrichment methods. (R)-N-CBZ-proline is obtained
with a product concentration of up to 70 g/L, and with an (1R,4S)-1-Amino-4-hydroxymethylcyclopent-2-ene is
ee value of > 99.5%. The process has been scaled up to an intermediate for the Glaxo anti-HIV drug Abaca-
produce 100 kg amounts. The chemistry for the prepa- vir.[34] The racemic (cis) N-acetylamino alcohol was used
ration of the racemic N-CBZ-proline is carried out in as the substrate for the selection and screening of
water to avoid solvent changes, and the final isolation microorganisms that could release the acetyl group by
procedure is a simple extraction that yields (R)-N-CBZ- hydrolysis and then grow with acetate as the carbon
proline and an aqueous solution of (S)-proline (l- source. In this way a microorganism was isolated that
proline) that can be recycled as starting material. contained a stereospecific amidohydrolase for the
hydrolysis of the racemic N-acetylamino alcohol (Fig-
ure 10). The reaction yields the amino alcohol product
[(1R,4S)-1-amino-4-hydroxymethylcyclopent-2-ene] and
the non-hydrolysed (1S,4R)-N-acetylamino alcohol, which
it was not possible to recycle.[35]

430 Adv. Synth. Catal. 2003, 345, 425 ± 435

Lonza: 20 Years of Biotransformations REVIEWS

Table 1. Cost comparison of stereospecific amidase produc-

tion with a wild-type or a recombinant strain.
Strain C. acidovorans E. coli
A:18 DH5/pCAR6
(arbitrary units) (arbitrary units)
Inducer 17 0
Figure 10. Process scheme for the production of (1R,4S)-1- Medium components 7 2
amino-4-hydroxymethylcyclopent-2-ene. Fermentation 34 3
Total cost 58 5

4.5 (S)-2,2-Dimethylcyclopropanecarboxamide amide was required in the growth medium to induce the
amidase; and (4) the biotransformation was about 20
(S)-2,2-Dimethylcyclopropanecarboxamide is an inter- times faster. These improvements decreased production
mediate for the production Cilastatin (Merck), which is costs for the stereospecific amidase by about twelve
a renal dehydropeptidase inhibitor that is administered times (Table 1).
with penem and carbapenem antibiotics to prevent their In the first step of the one-pot, two-step biotransfor-
degradation in the kidney by this enzyme.[36] Chemical mation the (R,S)-nitrile is rapidly and quantitatively
and biotechnological processes were developed in hydrolysed to the (R,S)-amide. It is important that no
parallel, but the bioprocess was simpler, had less unit trace of the racemic nitrile remains in the reactor after
operations and resulted in higher quality product. this step because the amidase is inhibited by this
A microbial screening programme resulted in the compound. The amidase-containing biomass is then
isolation of several bacterial strains containing amidases added so that the (R)-amide is specifically hydrolysed to
that could specifically hydrolyse the (R)-amide. One of the (R)-acid, and the product, the (S)-amide accumu-
these strains, Comomonas acidivorans A:18 was par- lates. The down-stream processing steps are ultra-
ticularly effective.[37] filtration, electrodialysis, ion-exchange chromatogra-
The first step in the process is the hydrolysis of the phy, reverse osmosis, crystallisation, centrifugation and
racemic nitrile to the racemic amide (Figure 11) by drying. The product has been produced at 15 m3 scale.
nitrile hydratase-containing whole cells. This enzymic The yield was 35% and the product had an ee value of
step resulted in several improvements to the process: (1) > 98%. The (R)-acid can be recycled chemically into the
a 20% yield improvement compared to the chemical substrate for the amidase which leads to higher yields
hydrolysis; (2) it reduced the number of unit operations and minimises waste.
for the synthesis of the starting product; (3) it was carried
out by an in-house strain with very high productivity; (4)
it made it possible to carry out a one-pot, two-step bio- 4.6 (S)-3,3,3-Trifluoro-2-hydroxy-2-methylpropionic
transformation; and (5) the (R,S)-amide formed by this Acid
biotransformation had better miscibility properties than
the chemically synthesised material. For the second step (R)- and (S)-3,3,3-trifluoro-2-hydroxy-2-methylpro-
the gene for the amidase was cloned into Escherichia pionic acid are intermediates for the synthesis of a
coli, resulting in a number of improvements: (1) biomass number of potential pharmaceuticals, which include
production was faster than with the Comamonas strain; ATP sensitive potassium channel openers for the treat-
(2) only a minimal growth medium was required; (3) no ment of incontinence,[38] and inhibitors of pyruvate
dehydrogenase kinase for the treatment of diabetes.[39]
The aim was to develop efficient syntheses for these
intermediates that yielded pure products with high ee
values, and that were suitable for large-scale production.
Several possible synthetic routes were evaluated. Some
were purely chemical and others included biocatalytic
steps.
Of the routes tested, the biocatalytic routes were
the most promising. For example the resolution of
(R,S)-ethyl 3,3,3-trifluoro-2-hydroxy-2-methylpropionate
with the esterase from Candida lipolytica (Figure 12)
resulted in the (S)-ester with an ee value of 99%.[40]
However, this route was not developed further
Figure 11. Process scheme for the production of (S)-2,2- because of the amount and resulting cost of the enzyme
dimethylcyclopropane carboxamide. required to complete the reaction in a reasonable time.

Adv. Synth. Catal. 2003, 345, 425 ± 435 431

 REVIEWS Nicholas M. Shaw et al.

Figure 12. The resolution of (R,S)-ethyl 3,3,3-trifluoro-2-hy-

droxy-2-methylpropionate with the esterase from Candida
lipolytica.
Figure 15. Process scheme for the production of (R)- or (S)-
3,3,3-trifluoro-2-hydroxy-2-methylpropionic acid.
Other possible routes with one or more biocatalytic
steps included those involving an enantioselective oxy-
nitrilase reaction (Figure 13) and various routes starting The process can therefore be adapted to produce both
from the racemic cyanohydrin (Figure 14). Screening R- and S-enantiomers of 3,3,3-trifluoro-2-hydroxy-2-
for enantioselective oxynitrilases[41] and for enantiospe- methylpropionic acid, or (S)-3,3,3-trifluoro-2-hydroxy-
cific nitrilases[42] was started, but discontinued when the 2-methylpropionamide. The biocatalytic step is part of a
amidase route (below) was found to be successful. combined chemical and biocatalytic route that starts
For the successful route an enantiospecific amidase from the Lonza product ethyl 4,4,4-trifluoroacetoace-
from Klebsiella oxytoca was isolated, characterised and tate, so again backward integration of the process is
cloned and used to resolve (R,S)-3,3,3-trifluoro-2-hy- optimal. The products typically have a purity of greater
droxy-2-methylpropionamide, giving (R)-3,3,3-trifluo- than 98% and ee values of essentially 100% after
ro-2-hydroxy-2-methylpropionic acid and (S)-3,3,3-tri- isolation. The process has been used to produce 100 g
fluoro-2-hydroxy-2-methylpropionamide (Figure 15). amounts of the (S)-acid, and has been successfully scaled
The (S)-amide could then be hydrolysed chemically to up to produce 100 kg amounts of the (R)-acid, with the
(S)-3,3,3-trifluoro-2-hydroxy-2-methylpropionic acid. biotransformation carried out at the 1500 L scale.[43,44]
Cloning of the amidase gene into Escherichia coli was
carried out for a number of reasons: (1) to improve
safety; Klebsiella oxytoca is a risk class 2 microorganism.
Transfer of the amidase gene to a GRAS host such as an
Escherichia coli K12 derivative facilitated handling and
Figure 13. A possible route to 3,3,3-trifluoro-2-hydroxy-2- official registration procedures for the production
methylpropionic acid using an oxynitrilase. According to the process; (2) the productivity of the biotransformation
choice of enzyme, it should be possible to form either the was improved by cloning the amidase gene under the
(R)- or the (S)-enantiomer. The route to the (S)-cyanohydrin control of a strong promoter to improve its expression;
and (S)-acid is shown. (3) to avoid the slime-capsule problem encountered
when using Klebsiella oxytoca as the production strain;
(4) to have the possibility to use other microorganisms
with, for example, higher substrate or product toler-
ances, as hosts for the cloned gene.
The amidase from Klebsiella oxytoca is robust, stable,
and does not require cofactors. It can be heated to 70 8C
for 10 min without loss of activity, and therefore, heat
treatment of the biomass used in the biotransformation
was used to stabilise the enzyme, presumably by
inactivating proteases.
For the whole process attention was paid to safety,
efficiency and economics. The biotransformation was
Figure 14. 3,3,3-Trifluoro-2-hydroxy-2-methylpropionic acid: carried out in aqueous solution (which is often the case
possible routes starting from the racemic cyanohydrin. The with biotransformations). Large volumes of aqueous
routes to the (R)-acid are shown. The corresponding routes solution containing relatively low product concentra-
to the (S)-acid should also be possible. tions are difficult to handle in chemical plants (com-

432 Adv. Synth. Catal. 2003, 345, 425 ± 435

Lonza: 20 Years of Biotransformations REVIEWS

pared with ™normal∫ processes in organic solutions), so isopropylamine into the cell, a synthetase, a P450
particular attention was paid to the delivery of a oxygenase and a hydrolase (Figure 16). This product is
biotransformation product solution suitable for down- currently in development.[48]
stream processing. It was subjected to ultra-filtration
through a 70 kDa membrane to remove proteins that
could cause foaming during extraction, and contamina- 5.2 l-Carnitine
tion of the product. Concentration by thin-film evapo-
ration then removed as much water as possible before l-Carnitine or vitamin BT is an essential growth factor
transfer to the chemical plant for product isolation. for the meal worm, Tenebria molitor. Part of the daily
requirements of l-carnitine in humans is met by
endogenous synthesis in the liver from lysine and
5 Multi-Enzyme Biotransformations methionine. l-Carnitine functions in the transport of
activated fatty acids across the inner mitochondrial
membrane. It has pharmaceutical, food and feed
5.1 l-Alaninol
applications. Lonza produces l-carnitine in a whole-
cell biocatalytic process that utilises a naturally occur-
l-Alaninol has potential as an intermediate for both ring pathway of 4 enzymes (Figure 17). In this process 4-
pharmaceuticals, such as the quinolone antibacterial butyrobetaine is converted to l-carnitine, which is
Levofloxacin from Daiichi Pharmaceuticals,[45] and excreted into the medium.[13] The biotransformation
agrochemicals, such as the selective herbicide Metola- has a high energy requirement due to the necessity of
chlor from Novartis.[46] This biotransformation was cofactor regeneration and the transport of the 4-
studied in collaboration with Prof. Thomas Leisinger×s butyrobetaine and l-carnitine across the membrane.
group at the ETH in Z¸rich.[47] Biocatalysis is carried out For this reason the process must be carried out with
with whole cells of Pseudomonas sp., in which l-alaninol growing cells or cells that are in a maintenance state,
is produced as an intermediate of isopropylamine which has the advantage of low biomass production and
catabolism. The enzymes involved are a permease, high metabolic activity. The pathway is analogous to the
which is involved in transport of the starting material b-oxidation of fatty acids and occurs in a new genus of
microorganism that is related to Agrobacterium and
Rhizobium.[49] The introduction of the chiral centre is
performed by crotonobetainyl-CoA-hydrolase, which
belongs to the lyase class of enzymes. The natural
degradation of l-carnitine by carnitine dehydrogenase
is prevented by a classical mutation in the production
strain. l-Carnitine is produced at the 50 m3 scale with
a volumetric yield of greater than 80 g/L, a conversion
Figure 16. Asymmetric oxidation. Simplified pathway for l- of substrate to product of 99.5%, and an ee value of
alaninol production by Pseudomonas sp. 100%.

Figure 17. The pathway for l-carnitine production in Agrobacterium/Rhizobium HK13. 1 ˆ 4-butyrobetainyl-CoA synthetase,
2 ˆ 4-butyrobetainyl-CoA dehydrogenase, 3 ˆ crotonobetainyl-CoA hydrolase, 4 ˆ thioesterase (?), 5 ˆ carnitine dehydro-
genase.

Adv. Synth. Catal. 2003, 345, 425 ± 435 433

 REVIEWS Nicholas M. Shaw et al.

Other possibilities for the biotechnological produc- References

tion of l-carnitine involve either the production from
achiral precursors or the resolution of racemic carnitine [1] R. N. Patel, Enz. Microb. Tech. 2002, 31, 804 ± 826.
and its derivatives.[50,51,52] These alternate syntheses have [2] J. D. Rozzell, Bioorg. Med. Chem. 1999, 7, 2253 ± 2261.
the disadvantages of cofactor requirements, unstable [3] U. Bornscheuer, in Biotechnology, (Ed.: D. R. Kelly),
starting product, low productivity, maximum yields of Wiley-VCH, Weinheim, 2000, pp 278 ± 294.
50%, product ee values of less than 100% or complex [4] A. J. J. Straathof, S. Panke, A. Schmid, Curr. Opin.
down-stream processing steps. Biotech. 2002, 13, 548 ± 556.
l-Carnitine can also be produced chemically[53,54] [5] K. Faber, Biotransformations in Organic Chemistry,
starting from epichlorhydrin or diketene. The last step Springer, Berlin, 1997.
is the racemic resolution of d,l-carnitine using a [6] C.-H. Wong, G. M. Whitesides, Enzymes in Synthetic
Organic Chemistry, Pergamon, Oxford, 1994.
resolving agent. The yields are low and to reduce costs
[7] S. M. Roberts, G. Casy, M.-B. Nielsen, S. Phythian, C.
it is necessary to recover and recycle the resolving agent.
Todd, K. Wiggins, Biocatalysts for Fine Chemicals
A comparison of the waste stream from the bioprocess
Synthesis, Wiley, New York, 1999.
with that from the Lonza chemical synthesis[55] showed
[8] S. M. Roberts, J. Chem. Soc. Perkin Trans. 2001, 1475 ±
that the biotechnological process was also much more 1499.
environmentally friendly. The amounts of waste water, [9] A. Kiener, Chemtech 1995, 9, 31 ± 35.
total organic carbon, salts, and waste for incineration [10] A. Kiener, Angew. Chem. Int. Ed. Engl. 1992, 31, 774 ±
were all considerably lower for the bioprocess.[56] 775.
[11] T. Yoshida, T. Nagasawa, J. Biosci. Bioeng. 2000, 89,
111 ± 118.
6 Conclusion [12] A. Kiener, Synlett 1994, 814 ± 816.
[13] H. G. Kulla, Chimia 1991, 45, 81 ± 85.
There are many reasons for the success Lonza has had in [14] D. Quarroz, EP 0 084 118 B1.
the area of biotransformations. In 1983 the emphasis was [15] P. Lehky, H. Kulla, S. Mischler, EP 0 152 948 B1.
to create a strong in-house research and development [16] E. J. Behrman, R. Y. Stanier, J. Biol. Chem. 1957, 228,
group with specialists in the area of microbiology, 923 ± 945.
biochemistry, molecular biology and fermentation. [17] V. Rovei, D. Caille, O. Curet, D. Ego, F. -X. Jarreau, J.
The fledgling group profited from Lonza×s existing Neural. Transm. 1994, 41, 339 ± 347.
infrastructure, especially the analytical capabilities and [18] M. Kataoka, L. P. S. Rohani, M. Wada, K. Kita, H.
the synergy with the chemistry research and develop- Yanase, I. Urabe, S. Shimizu, Biosci. Biotechnol. Bio-
ment groups. Today the pharmaceutical industry is chem. 1998, 62, 167 ± 169.
under a lot of price and time pressure due to a number of [19] M. Petersen, O. Birch, S. Shimizu, A. Kiener, M.-L.
reasons, for example, patent expiry and a 20 year low in Hischier, S. Thoni, WO 99/42590.
the number of new molecular entities launched onto the [20] S. Shimizu, M. Kataoka, A. Morishita, M. Katoh, T.
market.[57] This has led to a general decrease in the Morikawa, T. Miyoshi, H. Yamada, Biotechnol. Lett.
number of products in the development pipelines but it 1990, 12, 593 ± 596.
has made the implementation of biosolutions to difficult [21] L. Martinkova, V. Kren, Biocatalysis and Biotransforma-
synthetic problems all the more important. This envi- tion 2002, 20, 73 ± 93.
[22] Diversa Corporation, San Diego, CA.
ronment has forced a change in our biotechnological
[23] P. J. Fouts, O. M. Helmer, S. Lepkovsky, T. J. Jukes, Proc.
activities. We have diversified so that we can now offer a
Soc. Exp. Biol. Med. 1937, 37, 405.
complete custom manufacturing service in biocatalysis,
[24] Ullman×s Encyclopedia of Industrial Chemistry, Vitamins,
microbial fermentation and production of biopharma-
Chapter 14: Niacin, VCH, Weinheim, 1996.
ceuticals with animal and microbial cells. The biotrans-
[25] T. Nagasawa, C. D. Mathew, J. Mauger, H. Yamada,
formation activities are moving to faster and more Appl. and Environ. Microbiol. 1988, 54, 1766 ± 1769.
efficient means of screening for the required enzymic [26] U. A. Germann, P. J. Ford, D. Shlyakhter, V. S. Mason,
activity. This means implementing higher throughput M. W. Harding, Anti-Cancer Drugs, 1997, 8, 141 ± 155.
screening methods, capitalising on our now extensive [27] M. D. Owen, L. S. Dean, Expert Opin. Pharmacother.
strain collection, networking and entering into selective 2000, 1, 325 ± 336.
collaborations with universities and companies. These [28] G. A. C. Frampton, H. S. Zavareh, WO 96/12700.
collaborations either complement our research efforts, [29] A. Kiener, J.-P. Roduit, J. Kohr, N. Shaw, EP 0 686 698
fill deficits in our technology, by-pass patent or licensing B1.
problems or gain us access to other extensive strain or [30] M. Petersen, M. Sauter, Chimia 1999, 53, 608 ± 612.
enzyme collections. [31] E. Eichhorn, J.-P. Roduit, N. Shaw, K. Heinzmann, A.
Kiener, Tetrahedron: Asymmetry 1997, 8, 2533 ± 2536.
[32] M. Sauter, D. Venetz, F. Henzen, D. Schmidhalter, G.
Pfaffen, O. Werbitsky WO 97/33987.

434 Adv. Synth. Catal. 2003, 345, 425 ± 435

Lonza: 20 Years of Biotransformations REVIEWS

[33] J. Ngo, X. Rabasseda, J. Castaner, Drugs of the Future [44] N. M. Shaw, A. Naughton, K. Robins, A. Tinschert, E.
1997, 22, 221 ± 224. Schmid, M.-L. Hischier, V. Venetz, J. Werlen, T. Zim-
[34] R. H. Foster, D. Faulds, Drugs 1998, 55, 729 ± 736. mermann, W. Brieden, P. de Riedmatten, J.-P. Roduit, B.
[35] C. Bernegger-Egli, O. M. Birch, P. Bossard, W. Brieden, Zimmermann, R. Neum¸ller, Org. Proc. Res. Dev. 2002,
F. Brux, K. Burgdorf, L. Duc, K.-S. Etter, Y. Guggisberg, 6, 497 ± 504.
M. Sauter, E. M. Urban, WO 97/45529. [45] I. Hayakawa, S. Atarashi, S. Yokohama, M. Imamura, K.
[36] J. Birnbaum, F. M. Kahan, H. Kropp, J. S. Macdonald, Sakano, M. Furukawa, Antimicrob. Agents Chemother.
Am. J. Med. 1985, 78, 3 ± 21. 1986, 29, 163 ± 164.
[37] T. Zimmermann, K. Robins, O. M. Birch, E. Boehlen, EP [46] C. Vogel, R. Aebi, DE 2328340 A1.
0 524 604 B1. [47] S. I. de Azavedo W‰sch, J. R. van der Ploeg, T. Maire, A.
[38] C. J. Ohnmacht, K. Russell, J. R. Empfield, C. A. Frank, Lebreton, A. Kiener, T. Leisinger, Appl. Env. Microbiol.
K. H. Gibson, D. R. Mayhugh, F. M. McLaren, H. S. 2002, 68, 2368 ± 2375.
Shapiro, F. J. Brown, D. A. Trainor, C. Ceccarelli, M. M. [48] S. I. de Azevedo-Wasch, J. R. van der Ploeg, T. Leisinger,
Lin, B. B. Masek, J. M. Forst, R. J. Harris, J. M. Hulsizer, A. Kiener, K. Heinzmann, T. Gilligan, WO 99/07199.
J. J. Lewis, S. M. Silverman, R. W. Smith, P. J. Warwick, [49] T. P. Zimmermann, K. T. Robins, J. Werlen,
S. T. Kau, A. L. Chun, T. L. Grant, B. B. Howe, J. H. Li, F. W. J. M. M. Hoeks, in: Chirality in Industry II, (Eds.:
S. Trivedi, T. J. Halterman, C. Yochim, M. C. Dyroff, M. A. N. Collins, G. N. Sheldrake, J. Crosby), John Wiley
Kirkland, K. L Neilson, J. Med. Chem. 1996, 39, 4592 ± and Sons Ltd, Chichester, 1997, pp. 287 ± 305.
4601. [50] H. Jung, K. Jungπ H. P. Kleber, Adv. Biochem. Eng.
[39] T. D. Aicher, R. C. Anderson, G. R. Bebernitz, G. M. Biotechnol. 1993, 50, 21 ± 44.
Coppola, C. F. Jewell, D. C. Knorr, C. Liu, D. M. [51] M.-R. Kula, U. Joeres, U. Stelkes-Ritter, Ann. N. Y.
Sperbeck, L. J. Brand, R. J. Strohschein, J. Gao, C. C. Acad. Sci. 1996, 799, 725 ± 728.
Vinluan, S. S. Shetty, C. Dragland, E. L. Kaplan, D. [52] H.-P. Kleber, FEMS Microbiol. Lett. 1997, 147, 1 ± 9.
DelGrande, A. Islam, X. Liu, R. J. Lozito, W. M. [53] L. Tenud, DE 2518813 A1, DE 2542196 A1, DE 2542227
Maniara, R. E. Walter, W. R. Mann, J. Med. Chem. A1.
1999, 42, 2741 ± 2746. [54] H. Loester, D. M. Mueller, Wiss. Z. Karl-Marx-Univ.
[40] W. Brieden, A. Naughton, K. Robins, N. Shaw, A. Leipzig Math.-Naturwiss. R. 1985, 34, 212 ± 223.
Tinschert, T. Zimmermann, DE 197 25 802 A1. [55] R. Voeffray, J.-C. Perlberger, L. Tenud, J. Gosteli, Helv.
[41] F. Effenburger, B. Hˆrsch, F. Weingart, T. Ziegler, S. Chim. Acta 1987, 70, 2058 ± 2064.
K¸hner, Tetrahedron Lett. 1991, 32, 2605 ± 2608. [56] H.-P. Meyer, A. Kiener, R. Imwinkelried, N. Shaw,
[42] M. Wieser, T. Nagasawa, in Stereoselective Biocatalysis, Chimia 1997, 51, 287 ± 289.
(Ed.: R. Patel), Marcel Dekker, New York, 2000. [57] www.cmr.org.
[43] W. Brieden, A. Naughton, K. Robins, N. Shaw, A.
Tinschert, T. Zimmermann, WO 98/01568.

Adv. Synth. Catal. 2003, 345, 425 ± 435 435

View publication stats