Biocatalysis For Industrial Green Chemistry

Zhi Li, Martin Held, Sven Panke, Andrew Schmid, Renata Mathys, and Bernard Witholt

Eidgenössische Technische Hochschule Zürich, Switzerland

INTRODUCTION

Biocatalysis can make an important contribution to green chemistry due to several distinctive features, such as mild reaction conditions, nontoxicity, high chemo-, regio-, and stereoselectivity, and high turnover frequency. The potential of biocatalysis for the synthesis of chemicals is evident1-3 and examples of several industrial processes that are operational at BASF (Ludwigshafen, Germany), DSM (Geleen, the Netherlands), and Lonza (Visp, Switzerland) have recently been described.4 These industries use enzymes for the production of medium- to high-priced compounds that cannot be produced equally well using chemical approaches.5

The key challenge in the development of a bioprocess is the discovery and development of an appropriate biocatalyst. The discovery of enantioselective bio-catalysts for a given transformation can be achieved by many methods, such as the screening of collections of wild-type microorganisms6 or clonal libraries,7 in vitro evolution,8 and site-directed mutagenesis.9 While the latter requires knowledge of the structure and preferably also of the catalytic mechanism of the enzyme, the first two methods need fast and specific detection systems. In vivo screening of microorganism is often the method of choice for the discovery of cofactor dependent multi-component enzymes. The range of reactions that can be

Methods and Reagents for Green Chemistry: An Introduction, Edited by Pietro Tundo, Alvise Perosa, and Fulvio Zecchini

Copyright © 2007 John Wiley & Sons, Inc.

carried out with microorganisms and the range of microorganisms that have already been isolated or remain to be discovered is enormous. As a result, much energy goes into the selection of new enzymatic activities by using natural isolates and/or their mutants.

15.1 BIOCATALYSTS FUNCTION IN THE PRESENCE OF ORGANIC SOLVENTS

A critical consideration in the development of biocatalytic systems is the form in which the enzyme or enzyme system is going to be used. There are two general approaches. One is to use isolated enzymes. If these are inexpensive, they can be used as disposable biocatalysts, as is the case for glucose isomerase,10 which is the key biocatalyst in the production of high-fructose corn syrups from starch, or the lipases and proteases that are present in detergents. Alternatively, if enzymes are expensive to produce, they can be immobilized and used repeatedly by recovering the enzyme particles after each use.

The second general approach is to use whole cells that contain the enzyme or enzymes used in the biocatalytic process.11 The use of whole cells has the added advantage that coenzyme-dependent enzymes can be used because it is possible to regenerate the relevant coenzyme, through metabolism of the whole cells. This, of course, requires that the whole cells are not only physically intact but also meta-bolically active. Since coenzymes are often involved in building new molecules, industrial biocatalysis typically uses whole-cell systems.

Much of industrial chemistry takes place in organic solvents, or involves apolar compounds. Biocatalysis, in contrast, typically involves aqueous environments. Nevertheless, enzymes and microorganisms do in fact encounter apolar environments in Nature. Every cell is surrounded by at least one cell membrane, and more complex eukaryotic cells contain large amounts of intracellular membrane systems. These membranes consist of lipid bilayers into which many proteins are inserted: present estimates, based on genomic information, are that about one-third of all proteins are membrane proteins, many of which are so-called intrinsic proteins that are intimately threaded through the apolar bilayer. These proteins are essentially dissolved in, and function partly within, an apolar phase.

The notion that enzymes might well function in apolar solvents has been explored in detail and confirmed by Klibanov and his followers during the past two decades.12,13 Similarly, many microorganisms have been found to grow well in the presence of bulk solvents.14-16 This has permitted the development of biotransformation systems in which water-insoluble compounds are dissolved in apolar phases, to be converted to products by whole cells present in an aqueous phase, the products generally then dissolving again in the apolar phase.16-20

In short, microbial cells can be employed as very effective reactors for the conversion of substrates to products, operating in mixed aqueous-apolar systems, optimized for the best space-time yields attainable at lowest cost.

15.2 CHEMISTRY WITH MONOOXYGENASES

The regio- and stereoselective hydroxylation of specific nonactivated carbon atoms is a very useful reaction. For example, it can be used for the functionalization of alkanes and for the preparation of chiral alcohols that are useful pharmaceutical intermediates. This type of transformation is, however, a significant challenge in classic chemistry. On the other hand, Nature has found a general solution to this challenge via the development of a large number of monooxygenases, some of which, such as the soluble cytochrome P450 monooxygenases (P450 cam, P450 BM-3), soluble methane monooxygenase (sMMO), and membrane-bound alkane hydroxylase (AlkB) of Pseudomonas putida GPo1, have been well investigated. Their synthetic applications are limited, however, due to narrow substrate ranges or poor selectivity. The filamentous fungus Beauveria bassiana ATCC 7159 is often used for laboratorial synthesis,21 because it contains one or more unknown oxygenase systems, but these generally show low activities and selectivities. Thus, for specific hydroxylations, it is generally necessary to find appropriate biocatalysts.

To search for an appropriate whole-cell biocatalyst, it is necessary to identify an organism that contains large amounts of the desired enzyme. Equally important, the organism should not contain related pathway enzymes that modify or destroy the product synthesized by the desired enzyme.

In addition, the substrate and product should be transported through the cell membrane, either passively or actively, and necessary cofactors should be regenerated. Finally, the specific organism used should function well in an optimized bioreactor system.22,23 All of these requirements can be met by using strains that contain the desired enzyme in question.

15.3 SCREENING FOR THE REGIO- AND STEREOSPECIFIC HYDROXYLATION OF PYRROLIDINE

We were interested in the discovery and development of biocatalysts for the regio- and stereoselective hydroxylation of pyrrolidine to 3-hydroxypyrrolidine, since both (S)- and (R)-3-hydroxypyrrolidines are useful intermediates for the preparation of several pharmaceuticals (Figure 15.1), and they are difficult to prepare by chemical syntheses. To avoid random screening of a large number of microorganisms, we concentrated on microorganisms with selected biodegradation abilities. Microorganisms that degrade n-alkanes often contain alkane hydroxy-lases that catalyze the hydroxylation of alkanes at a terminal position as the first degradation step. These alkane hydroxylases might also catalyze the desired hydroxylation of pyrrolidines. Moreover, the alcohol dehydrogenases in alkane-degrading microorganisms that catalyze the oxidation of terminal alcohols to the corresponding aldehydes during biodegradation may not be able to oxidize the desired product 3-hydroxypyrrolidine, which is a cyclic secondary alcohol. Therefore, we selected 70 n-alkane-degrading strains as the source of catalyst for n02 O

H3COOC

H3COOC

Barnidipine, calcium antagonist

DX-9065a, anticoagulant h2n/,,

DX-9065a, anticoagulant h2n/,, hci h2o

Darifenacin, flor Irritable bowel syndrome

RS-533, carbapenem antibiotics hci h2o

ABT-719, antibacterial drug

Figure 15.1 (S)- and (R)-3-hydroxypyrrolidines as pharmaceutical intermediates.

screening. By use of a miniaturized system, which allows for parallel inoculation, growth, and bioconversion on a microtiter plate,24 coupled with a fast and sensitive Liquid Chromatography-Mass Spectroscopy (LC-MS) detection, we found that 25 of the 70 strains catalyze the 3-hydroxylation of N-benzylpyrrolidine.25 Further investigation with 12 more active strains demonstrated complementary enantioselectivities (Table 15.1). Hydroxylation of N-benzylpyrrolidine with Pseudomonas oleovorans GPol, the prototype alkane hydroxylating strain that contains a well-known membrane-bound alkane hydroxylase (alkB) afforded 62% of (R)-N-benzyl-3-hydroxypyrrolidine in 52% enantiomeric excess (ee), whereas

TABLE 15.1 Enantioselectivity and Activity of the Hydroxylation of N-Benzylpyrrolidine to N-Benzyl-3-hydroxypyrrolidine with Several Alkane-Degrading Strains

Entry

Microorganisms

Product ee (%)

Relative Activity

1

HXN-11CC

VC (R)

4

2

HXN-4CC

65 (R)

0.5

3

P.putida P1

62 (R)

1

4

P. oleovorans GPo1

52 (R)

1

5

BC2C

4C (R)

1

6

HXN-15CC

25 (R)

3

V

HXN-5CC

10 (R)

11

S

HXN-2CC

53 (S)

6

9

HXN-1CC

10 (S)

3

1C

HXN-19CC

<10 (S)

10

11

HXN-1000

<10 (S)

1

12

HXN-6CC

C

3

"Relative activity was based on the activity with P. oleovorans GPol.

"Relative activity was based on the activity with P. oleovorans GPol.

I HXN-200 |

R R

Sphingomonas sp.

Sphingomonas sp.

Activity (U/g cdw)

ee ee ee after crystalization (%)

R = C02CH2Ph R = C02f-Bu

Figure 15.2 Improvement of the hydroxylation activity and enantioselectivity with Sphingomonas sp. HXN-200 by substrate modification.

hydroxylation with Sphingomonas sp. HXN-200 gave 62% of (S)-enantiomer in 53% ee with six times higher activity.

It was found that Sphingomonas sp. HXN-200 contains a soluble alkane mono-oxygenase and accepts a broader range of substrates. The enantioselectivity and activity were further improved by introducing a "docking/protecting" group into the pyrrolidine substrates. As shown in Figure 15.2, changing the N-substitution from a benzyl to a benzyloxycarbonyl group resulted in a three-fold increase of the activity and an improvement of product ee from 53% (S) to 75% (R).26 We found that the ee can be further increased by simple crystallization of the biopro-duct from a mixture of n-hexane and ethyl acetate. Thus, it is possible to produce (R)- and (S)-3-hydroxypyrrolidine in 98% and 96% ee, respectively.

We optimized the growth of strain HXN-200 on n-octane and produced the cells in a large amount. It was found that the cells can be stored at — 80°C for two years without significant loss of activity. The frozen/thawed cells, that are easy to handle for the organic chemist, can be used for routine hydroxylation in an organic chemistry laboratory.

15.4 SYNTHESIS OF HYDROXYPYRROLIDINES, HYDROXYPYRROLIDINONES, AND HYDROXYPIPERIDINES

3-Hydroxypyrrolidines were easily prepared in 62-94% yield and in 1-2 g/L by hydroxylation with the frozen/thawed cells. Even higher productivity was obtained by hydroxylation with growing cells of HXN-200,26 which is obviously the best choice for an industrial hydroxylation. It was found that Sphingomonas sp. HXN-200 can also catalyze the hydroxylation of 2-pyrrolidinone.27 As shown in Figure 15.3, hydroxylation of N-benzylpyrrolidinone gave the corresponding (S)-4-hydroxypyrrolidine in >99.9% ee, demonstrating the best enantioselectivity ever reported for biohydroxylations. Changing the "docking/protecting" group to a tert-butoxycarbonyl group increased the activity to 11 U/g cell dry weight (cdw). Although the enantioselectivity was slightly lower, simple crystallization

I I

R R

R = CH2Ph R = CH2Ph >99.9%ee Act. 4.6 U/g cdw

R = C02f-Bu R = C02f-Bu 92.0% ee Act. 11 U/g cdw

99.9% ee (after crystalization)

Carbapenems 2

O (Phase II) Jj

Figure 15.3 Biohydroxylation of 2-pyrrolidinones with Sphingomonas sp. HXN-200.

from n-hexane/ethyl acetate improved the ee from 92% to 99.9%. Here again, (S)-4-hydroxypyrrolidines are useful intermediates for the preparation of several pharmaceuticals.

Sphingomonas sp. HXN-200 was also able to accept a six-member ring substrate. Hydroxylation of N-benzyl- and N-teri-butoxycarbonyl-2-piperidinone gave the corresponding (R)-4-hydroxy-piperidin-2-ones in 31% and 68% ee, respectively28 (Figure 15.4). This provides a simple synthesis of such types of useful synthons.

Further investigation demonstrated that Sphingomonas sp. HXN-200 is also the best biocatalyst known thus far for the hydroxylation of N-substituted piperidines.29 Even small cyclic compounds such as azetidines, which are difficult to be hydroxyl-ated by other system, are also good substrates for strain HXN-200.29 Excellent regioselectivity and high activity were obtained in all hydroxylations shown in Figure 15.5. This provides simple and efficient synthesis of 4-hydroxypiperidines and 3-hydroxyazetidines, which are useful pharmaceutical intermediates.

Sphingomonas sp. -

O HXN-200

Sphingomonas sp. -

O HXN-200

Activity (U/g cdw)

31 68

Figure 15.4 Biohydroxylation of 2-piperidinones with Sphingomonas sp. HXN-200.

Purines And Pyrimidines

Ebastine, drug for allergic rhinilis

Dezinamide Antiepllephllc (Phase II) American Home Products

Ebastine, drug for allergic rhinilis

^LH p COOR

Oral carbapenem antibiotics

Dezinamide Antiepllephllc (Phase II) American Home Products

Nadlfloxacin, antibacterial drugs

Figure 15.5 Biohydroxylation of piperidines and azetidines with Sphingomonas sp. HXN-200.

Nadlfloxacin, antibacterial drugs

Azelnidipine, antihypertensive Sankvo: Libe

Figure 15.5 Biohydroxylation of piperidines and azetidines with Sphingomonas sp. HXN-200.

15.5 BIOCATALYST FORMULATION AND TUNING

To facilitate its application in organic synthesis, we developed a lyophilized cell powder of Sphingomonas sp. HXN-200 as a biohydroxylation catalyst.29 Hydro-xylation of N-benzyl-piperidine with such catalyst powder showed 85% of the activity of a similar hydroxylation with frozen/thawed cells, shown in Figure 15.6. The fact that rehydrated lyophilized cells are able to carry out such a reduced nicotinamide adenine dinucleotide (NADH)-dependent hydroxylation indicates that these cells are capable of retaining and regenerating NADH at rates equal to or exceeding the rate of hydroxylation. To our knowledge, this is the first example of the use of lyophilized cells for a cofactor-dependent hydroxylation.

For an industrial biotransformation, it is often necessary to further optimize an appropriate biocatalyst. This includes the elimination of the follow-up enzymes in the wild-type strain by mutations and the improvement of other characteristics by additional mutations and the selection of improved strains. Alternatively, the genetic information for a desired enzyme might be introduced in a host that has many of the preceding characteristics, and that has no enzymes that could modify or degrade the desired product.

Figure 15.6 Hydroxylation of N-benzylpiperidine (5 mM) with lyophilized cell powder and frozen/thawed cells of Sphingomonas sp. HXN-200 at cell concentration of 4.0 g cdw/L.

Sphingomonas sp.

CI-LPh

HXN-200

P% frozen/thawed cells S% frozen/thawed cells P% Lyophilized cells S% Lyophilized cells

CH,Ph

Figure 15.6 Hydroxylation of N-benzylpiperidine (5 mM) with lyophilized cell powder and frozen/thawed cells of Sphingomonas sp. HXN-200 at cell concentration of 4.0 g cdw/L.

15.6 AN IDEAL BIOCATALYST BASIS: E. coli

Perhaps surprisingly, we have found in our work that E. coli is an excellent

23 30_32

whole-cell biocatalytic host. , It is relatively easy to introduce new desired enzymes into various E. coli strains. Generally, for many of the hydroxylation reactions that we have studied, degradative enzymes or downstream pathway enzymes that could modify or eliminate desired hydroxylation products, are not present in E. coli. In working with two liquid-phase systems, E. coli is more sensitive to apolar solvents than Pseudomonas strains. However, we have developed mixed apolar phase systems, based on highly apolar solvents such as hexadecane or substituted phthalates,19 which are highly compatible with E. coli. Thus, it is possible to use very toxic substrates, and produce equally toxic products, which dissolve in the hexadecane or phthalate phase and have very little effect on the host organism present in the aqueous phase.33

The beauty of working with only a few host strains is that these can be optimized for growth and growth medium, expression system for a wide range of bio-catalysts, behavior in the presence of solvents, regeneration of coenzymes, downstream processing, recovery of cells for further biocatalysis use, and waste treatment. Thus, function can be optimized generally, cost can be minimized generally, and different enzymes can be introduced from a wide range of sources. In addition, each enzyme to be used can be optimized further for top biocatalytic performance by mutagenesis, directed evolution,34 or gene shuffling.35

15.7 SYNTHESIS OF CATECHOLS BY A RECOMBINANT WHOLE CELL BIOCATALYST

An interesting example of the application of recombinant whole-cell biocatalysis is the conversion of 2-hydroxybiphenyl (2-phenylphenol) to 2,3-dihydroxybiphenyl

2-Hydroxybiphenyl 3-monooxygenase (E.C. 1.14.13.44)

NADH NAD

NADH NAD

Figure 15.7 Production of 3-substituted catechols using a designer biocatalyst.

or 3-phenylcatechol (Figure 15.7). Catechols are important building blocks for the chemical and pharmaceutical industries. However, their chemical synthesis is cumbersome. Especially the synthesis of 3-substituted catechols requires numerous agents with low environmental compatibility (organometallic reagents, HBr) and energy intensive reaction conditions (low temperature). At the same time, catechol and some of its derivatives are central metabolites in the microbial catabolism of aromatic compounds.36 Besides naturally occurring aromatics such as tyrosine and phenylalanine, nonbiological aromatic solvents and numerous polycyclic aromatic hydrocarbons are also readily degraded or modified by bacteria, with the genus Pseudomonas playing an important role in such turnover of aromatics in nature.

Pseudomonas azelaica HBP1 is a prominent example of interesting aromatic compound degradation.37 The strain readily degrades 2-phenylphenol—a man-made compound that has been widely used as a food-protecting agent and as a germicide. The initial step of 2-phenylphenol transformation by the strain is formation of 3-phenylcatechol by a 2-hydroxybiphenyl 3-monooxygenase (Figure 15.7).38 This enzyme has a broad substrate range and oxidizes numerous other 2-substituted phenols to corresponding 3-substituted catechols in a highly regioselective cofactor-dependent reaction. However, the strain cannot be used for catechol synthesis, because reaction products are instantly broken down. Hence, the 2-hydroxybiphenyl 3-monooxygenase gene was cloned and expressed in E. coli JM101.39 The resulting biocatalyst E. coli JM101 [pHBP461] efficiently overproduces the monooxygenase but does not degrade the products formed, which makes this strain a promising candidate for synthesis of 3-substituted catechols.

A major challenge that had to be met arose from the extreme bactericidal properties of phenols and catechols. Most microorganisms are poisoned at phenol or catechol concentrations in the 0.1-1 g/L range. The biocatalyst E. coli JM101 [pHBP461] is no exception, and was inactivated by 200 mg/L of both 2-phenylphenol and 3-phenylcatechol.40 Furthermore, 3-substituted catechols are of limited stability in aerated aqueous solutions and form multimeric humic-acid-like structures as unwanted side products. 3-Phenylcatechol, for instance, has a half-life time of only 14 h at pH 7.2.

Last but not least, catechols are highly water-soluble (the water solubility of catechol is approximately 1 g per 2.3 mL of water), which makes it difficult to directly extract them in situ from reaction media with organic, water immiscible solvents. Nevertheless, extraction of catechols from aqueous systems with hydro-phobic polymers such as the polystyrene-based resin Amberlite XAD-4 is

straightforward. We have therefore employed XAD-4 to combine biocatalytic synthesis with simultaneous product extraction. The system (Figure 15.8) comprises a continuously stirred tank reactor, a starting material feed pump, a product recovery loop with a (semi-) fluidized bed of XAD-4, and a pump to circulate the entire reaction mixture through the loop.40 Preliminary studies indicated that XAD-4 had no detrimental effects on E. coli JM101 (pHBP461), hence, separation of biomass and reaction liquid prior to catechol extraction was not required. The biocatalytic reaction was carried out at very low concentrations of the toxic substrate and product. This was achieved by feeding the substrate at a rate lower than the potential bioconversion rate in the reactor.

This assured that all substrate fed to the bioreactor was instantly converted to product: no substrate accumulated in the bioreactor. To prevent accumulation of product in the reactor, the reactor contents were circulated through the external fluidized-bed module, which contained Amberlite XAD-4 resin. All product adsorbed to the resin, while cells and medium components passed through the bed and back into the bioreactor, ready to convert more of the substrate newly fed into the bioreactor.

Figure 15.9 shows that this approach worked quite nicely: the substrate was added to a total concentration of 2.5 g/L, but neither substrate nor product accumulated in the bioreactor medium.30 Without a product recovery loop the product concentration (3-phenylcatechol) did not exceed 0.4 g/L, because of biocatalyst deactivation (results not shown). With the loop, 2-phenylphenol and 3-phenylcatechol concentrations remained below 0.1 g/L. Therefore, cell viability and biocatalytic activity were maintained, as indicated by the constantly low dissolved oxygen tension in the aerated reactor. As a result product yields (based on the 3-phenylcatechol eluted from the product sink) increased by one order of magnitude.40

The system was used for the preparative scale synthesis of numerous catechols. Table 15.2 shows that a half dozen other phenols could be converted to the corresponding catechols. In each case, 1 to 2 g of substrate was converted with yields

15.7 SYNTHESIS OF CATECHOLS

15.7 SYNTHESIS OF CATECHOLS

Figure 15.9 Synthesis and in situ recovery of 3-phenylcatechol.

of 80 to 90%, and after desorption and isolation of product, overall yields of 60 to 80% were estimated.41 The entire process was scaled up to the 300 L scale, resulting in the formation of a total of 1.1 kg 3-phenylcatechol, at a purity of 77%. One crystallization was sufficient to bring 25 g of the material to a purity of >98%; this compound is now for sale by Fluka, in Switzerland.

Though a detailed life-cycle analysis has not yet been made, the reagents that go into these biocatalytic reactions already indicate a certain trend. Besides starting materials the route requires only salts, glucose, glycerol and air, XAD-4 for product extraction, acidified methanol for product elution, and a solvent for product polishing (hexane or methanol/acid mixtures). With the exception of the XAD-4 resin, all reagents are from renewable or almost inexhaustible stocks. With sufficient XAD-4 reuse—exceeding a hundred or so cycles—waste streams are minor and comparatively green since they consist mainly of biomass and water.

This bioconversion illustrates several points. First, E. coli is a highly adequate host for the bioconversion of toxic compounds. Second, it is possible to attain adequate fluxes at low concentrations of substrate. Third, it was relatively easy to scale up the system from laboratory (several liters) to pilot scale (several hundred liters).

TABLE 15.2 Production Data for the Bioconversion of Substituted Phenols to the Corresponding Catechols

Starting Material

Yield (Total)

Yield (Adsorbed and

Compound

Added (g)

(mole%)°

Eluted) (mole%)°

3-Et-catecholb

2.0

90

82

3-Pr-catechol

1.2

81

69

3-i-Pr-catechol

2.6

91

76

3-sec-Bu-catechol

1.8

84

78

3-Cl-catechol

1.1

91

71

3-Br-catechol

2.6

84

73

"Mole % of total starting-material added. bNo standard available.

"Mole % of total starting-material added. bNo standard available.

15.8 SYNTHESIS OF EPOXYSTYRENE WITH VERY HIGH ENANTIOSELECTIVITY

Another useful bioconversion is the epoxidation of styrene to epoxystyrene. Styrene oxide is a valuable building block, because the epoxide function allows versatile synthetic chemistry, and the benzene ring is part of the majority of today's drugs (Figure 15.10). It is used, for example, in the production of the anti-helmintic drug Levamisole.42 However, to be an attractive building block for drug synthesis, the styrene oxide needs to be enantiopure. There are some chemical asymmetric synthesis routes described, but they usually deliver only moderate to good enantiomeric excesses. Kinetic resolution of styrene oxide employing the Jacobsen catalyst affords enantiopure styrene oxide, but this method has the inherent drawback of being limited to a maximum chemical yield of 50%.43 In order to circumvent these various limitations, we attempted to design an asymmetric synthesis route from styrene to styrene oxide using biocatalysis. Asymmetric synthesis in principle allows a 100% ee/100% yield process, and the enantioselectivity of enzymes should aid in overcoming the problem just mentioned in achieving satisfactory ee's.

We have used two enzyme systems for this purpose. The first is the xylene monooxygenase system of P. putida mt2, which is capable of utilizing xylene and toluene derivatives for growth. In the first enzymatic step, xylene monooxygenase introduces an oxygen atom in a toluene or xylene methyl substituent group. This monooxygenase can also introduce an epoxide in the vinyl double bond of styrene and substituted styrenes.44 The reaction was carried out using E. coli recombinants carrying only the xylene monooxygenase system, encoded by xylMA, resulting in efficient formation of epoxystyrene with an enantiomeric excess of 92%45 (see Figure 15.10).

Still, a biocatalyst that performs this reaction with an even higher ee was desirable, which led us to investigate styrene degradation in various bacterial soil strains. One of these, Pseudomonas sp. strain VLB120 was selected for further a o.

Styrene a o.

Styrene

Figure 15.10 Biological oxidation of styrene to (S)-styrene oxide.

ScPEXhC

stySc t>

lacZp

Pg Nc

HSc HSnPg

styR

styA

^>styAp

NrPNcStSm

styD

XhXbASt

pSPW 1

StyAB

StyAB

StyC

StyC

StyD

StyD

Figure 15.11 Styrene degradation in Pseudomonas sp. strain VLB120.

study, because it appeared to degrade styrene via styrene oxide and phenylacetal-dehyde (Figure 15.11). After preparation of a genomic library in E. coli, clones that could convert indole to indigo were selected and analyzed for their ability to transform styrene to styrene oxide. One such clone contained a 5.7-kb DNA fragment that encoded the major part of a styrene degradation pathway, the first step of which consists of the oxidation of styrene to epoxystyrene with a cytoplasmic two-component styrene monooxygenase (Figure 15.11). The genes encoding the styrene monooxygenase were used to construct a recombinant biocatalyst in E. coli JM101, which achieved the conversion of styrene into (S)-styrene oxide with an ee of more than 99% (Figure 15.10). In other words, natural biodiversity was sufficient to increase the ee to more than satisfactory levels.33,46

In designing a production process for (S)-styrene oxide, several issues must be considered. First, styrene monooxygenase activity depends on the availability of NADH, making the use of an enzyme reactor with (partially) purified enzyme complex and expensive because the cofactor needs to be regenerated while product is produced. Second, the substrate as well as the product of the reaction are not very soluble in water and, even at such low concentrations, toxic to living cells, complicating the development of a whole-cell biocatalytic system. One way to still create a potentially economically attractive process is to use growing cells in a two-liquid phase culture. The partition coefficients of substrate and product dictate that both compounds remain preferentially in the organic phase and the aqueous concentrations remain below toxic levels, while the overall concentrations of substrate and product in the reactor can be increased far beyond what would be possible if only the aqueous phase was present. In the present example, the overall styrene concentration could be increased from 2 mM for a one-liquid phase culture to 135 mM for a two-liquid phase culture that consisted of 50 vol % of aqueous medium with the biocatalyst and 50 vol % of the organic phase with 2 vol % styrene. During the reaction, a small amount of the styrene partitions out of the organic phase, is oxidized, and is then reextracted into the organic phase. The organic phase serves as substrate pool and product sink at the same time. We investigated this mode of production on a 2-L scale with recombinant E. coli JM101 cells that carried the styrene monooxygenase genes on the expression vector pSPZ10 under control of the alk regulatory system. This vector carries the styrene monooxygenase genes under control of the alkBp promoter, which is induced by octane and is not repressed by glucose in E. coli. The pBR322-based vector has been optimized by introducing transcriptional terminators to transla-tionally shield regions important for plasmid propagation and by replacing the tetracycline and ampicillin resistance genes with a kanamycin resistance gene.33 Such a genetic system allows easy induction in a two-liquid-phase culture and the use of a cheap carbon source for the cultivation.

We optimized the cultivation protocol with respect to aqueous medium composition, organic phase, and phase ratio (the ratio of the volume of the organic phase to the total liquid volume). The best system consists of a defined mineral medium, with glucose as the carbon source and diethylhexylphthalate as the organic phase at a phase ratio of 0.5. The organic phase contained 2 vol % styrene and 1 vol % of octane, which we added as an inducer for gene expression.

Loading For Peform
Figure 15.12 Two-liquid-phase-based biooxidation of styrene to styrene oxide with a recombinant whole-cell biocatalyst.

With this system we converted 135 mM styrene (relative to the total liquid volume) to styrene oxide in 10 h at a cell dry weight of around 10g/L aqueous phase, with an average activity of 152 U/L total liquid volume. This corresponds to a space-time yield of 1.1 g (S)-styrene oxide per liter and hour. These are the highest specific activities reported thus far for a microbial epoxidation process.33

The bioconversion was carried out in a two-liquid phase system (Figure 15.12), which was developed at the 2-L level, and scaled up to the 30-L level to produce almost 400 g of product. Several apolar phases were used, of which bis(2-ethylhexyl)phthalate (BEHP) was preferred because it showed a better partitioning of epoxystyrene toward the apolar phase and away from the aqueous phase than did hexadecane. This was important because the product was quite toxic to the recombinant biocatalyst when it appeared in the aqueous phase. This bioconversion illustrates that apolar compounds like styrene and its epoxide, which are quite toxic to microorganisms, can be handled successfully in two-liquid-phase cultures. The toxicity of the substrate and product are not significant issues here.

Again, a recombinant E. coli strain performs quite well in this system. Subsequent phase separation and distillation permit the simple purification of the product. The final product had an ee >99%, which is significantly better than that seen for the XylMA based system (Figure 15.10).

15.9 LIKELY COSTS OF LARGE SCALE HYDROXYLATION OR EPOXIDATION PROCESSES WITH WHOLE CELL BIOCATALYSTS

To get some idea of the prices to be expected for compounds produced with these approaches, we have estimated the total cost of producing 10,000 tons per annum of 1-octanol from «-octane, based on data collected for this conversion by P. oleovorans, during growth in a two-liquid-phase system containing 15% (v/v) hexadecane as a carrier phase.47 n-Octane is dissolved in the carrier phase to a concentration of 5 -10% (v/v), converted by the P. oleovorans cells in the aqueous phase, and the product 1-octanol dissolves in the hexadecane phase once more. Downstream processing consists of a phase separation, followed by two distillation steps. In the first step, the C8 alkane/alkanol are separated from the hexadecane, which is recycled into the bioreactor. In the second step, the n-octane is distilled off the n-octanol: the octane is recycled to the bioreactor, and the octanol is collected as the desired product. This approach leads to a very clean product stream of >98% pure 1-octanol.48

The total cost of the process was estimated to be about 8 US$ per kg product (Figure 15.13). The most significant cost item (ca. 40% of the total) was due to the glucose and salts necessary to support cell growth; that is, these are part of the biocatalyst formation costs. Biocatalyst costs can be reduced by increasing specific activities per g cells, and by extending the useful lifetime of whole-cell biocatalysts. We estimate that reductions in biocatalyst cost and associated process costs might lower the preceding estimate to 5 US$ per kg product. Although the estimate described here was carried out for a specific

Figure 15.13 Biocatalysis for the oxidation of n-octane to 1-octanol at 10,000 tons/year.

alkane-to-alkanol conversion, it applies to similar hydroxylations or oxidations of other apolar compounds. A key parameter in all such estimates is the specific activity per g cells that can be achieved. Typical activities are in the order of 5 to 50 U/g cell dry mass, where U is the international enzyme activity unit, expressed as mmol of product formed or mmol substrate utilized per minute.

The preceding estimate suggests that products valued at more than 10 US$ per kg are potential targets for biocatalytic production, provided the market is sufficiently large. Clearly, compounds valued at 50-100 US$ or more per kg are interesting biocatalysis targets. However, high tonnage targets of 5-7 US$ per kg are still worth investigating.

15.10 CONCLUSION

It is interesting to speculate on the development of such a focused biocatalysis-based chemical industry. It is likely that at least one more decade will pass before a significant biocatalysis-driven company emerges. A very important attribute of such a company will be the ability of management not to be side-tracked by nonissues, examples of which are:

• Toxic compounds: educts, products

• Polar, apolar, mixed-phase biotransformation media

• Reactor systems: batch or continuous

• Product recovery: batch or integrated

• Waste handling

• Geopolitics

• Biosafety regulation

What will matter is the extent to which a biocatalysis company manages to focus on the market, on the development of biocatalysts to address market needs, and on the ability to compare the potential of biocatalysts and chemical alternatives. Finally, it is possible to list some of the properties of that company that is likely to emerge as the premier biocatalysis company in the next decade:

• Understands the chemicals market

• Solid experience in synthetic chemistry: deep understanding of its potentials and limitations

• Biocatalyst and process development: done by organic chemists, chemical engineers, as well as molecular biologists

REFERENCES

1. Drauz, K.; Waldmann, H. Enzyme Catalysis in Organic Synthesis: A Comprehensive Handbook, VCH, Weinheim, New York, 1995.

2. Faber, K. Biotransformations in Organic Chemistry: A Textbook, 3rd, completely rev. ed., Springer-Verlag, Berlin, New York, 1997.

3. Wubbolts, M. G.; Witholt, B. Selected Industrial Biotransformations, Montie, T. C. (Ed.) Plenum Publishing Co. New York, 1998, pp. 271-329.

4. Schmid, A.; Dordick, J. S.; Hauer, B.; et al. Nature, 2001, 409, 258-268.

5. Tramper, J. Biotechnol. Bioeng., 1996, 52, 290-295.

6. Duetz,W. A.; Minas, W.; Kuhner, M.; et al. Bioworld, 2001, 2, 8-10.

7. DeSantis, G.; Zhu, Z.; Greenberg, W. A.; et al. J. Am. Chem. Soc., 2002, 124, 9024.

8. Reetz, M. T.; Zonta, A.; Schimossek, K.; Lebeton, K.; et al. Angew. Chem. Int. Ed. Engl., 1997, 36, 2830.

9. Peters, M. W.; Meinhold, P.; Glieder, A.; Arnold, F. H. J. Am. Chem. Soc., 2003, 125, 13442.

10. Jensen, V. J.; Rugh, S. Meth. Enzymol., 1987, 136, 356-370.

11. Wubbolts, M. G.; Favre-Bulle, O.; Witholt, B. Biotechnol. Bioeng., 1996, 52, 301 -308.

12. Klibanov, A. M. Acc. Chem. Res., 1990, 23, 114-120.

13. Laane, C.; Boeren, S.; Vos, K.; Veeger, C. Biotechnol. Bioeng., 1987, 30, 81-87.

14. De Smet, M. J.; Witholt, B.; Wijnberg, H. J. Org. Chem., 1981, 46, 3128-3131.

15. De Smet, M. J.; Kingma, J.; Wijnberg, H.; Witholt, B. Enzyme Microbiol Technol., 1983, 5, 352-360.

16. Witholt, B.; Van Beilen, J.; Wubbolts, M. G.; et al. Bioforum, 1997, 1, 32-36.

17. Witholt, B.; de Smet, M. J.; Kingma, J.; et al. Trends Biotechnol, 1990, 8, 46-52.

18. Witholt, B.; Wubbolts, M. G. Ind. Bioprocess, 1996, 18, 6-7.

19. Schmid, A.; Kollmer, A.; Mathys, R. G.; Witholt, B. Extremophiles, 1998, 2, 249-245.

20. Schmid, A.; Sonnleitner, B.; Witholt, B. Biotechnol. Bioeng., 1998, 60, 10-23.

21. Grogan, G. J.; Holland, H. L. J. Mol. Cat. B: Enzymatic, 2000, 9, 1-32.

22. Rothen, S. A.; Sauer, M.; Sonnleitner, B.; Witholt, B. Biotechnol. Bioeng., 1998, 58, 356-65.

23. Favre-Bulle, O.; Schouten, T.; Kingma, J.; Witholt, B. Biotechnology (NY), 1991, 9, 367-371.

24. Duetz, W. A.; Ruedi, L.; Hermann, R.; O'Connor, K.; et al. Appl. Environ. Microbiol., 2000, 66, 2641-2646.

25. Li, Z.; Feiten, H.-J.; van Beilen, J. B.; Duetz, W.; Witholt, B. Tetrahedron: Asymmetry, 1999, 10, 1323-1333.

26. Li, Z.; Feiten, H.-J.; Chang, D.; Duetz, W. A.; et al. J. Org. Chem., 2001, 66, 8424-8430.

27. Chang, D.; Witholt, B.; Li, Z. Org. Lett., 2000, 2, 3949-3952.

28. Chang, D.; Feiten, H.-J.; Witholt, B.; Li, Z. Tetrahedron:Asymmetry, 2002, 13, 2141-2147.

29. Chang, D.; Feiten, H.-J.; Engesser, K.-H.; et al. Org. Lett., 2002, 4, 1859-1862.

30. Favre-Bulle, O.; Witholt, B. Enzyme Microbial Technol., 1992, 14, 931-937.

31. Favre-Bulle, O.; Weenink, E.; Vos, T.; et al. Biotechnol. Bioeng., 1993, 41, 263-272.

32. Schneider, S.; Wubbolts, M. G.; Sanglard, D.; Witholt, B. Tetrahedron Asymmetry, 1998, 9, 2833-2844.

33. Panke, S.; Wubbolts, M. G.; Schmid, A.; Witholt, B. Biotechnol. Bioeng., 2000, 69, 91-100.

34. Arnold, F. H.; Volkov, A. A. Curr. Opin. Chem. Biol., 1999, 3, 54-59.

35. Stemmer, W. P. C. Nature, 1994, 370, 389-391.

36. Harayama, S.; Timmis, K. N. Catabolism of Aromatic Hydrocarbons by Pseudomonas, Hopwood, D.; Chater, K. (Eds.) Academic Press, London, 1987, pp. 151-174.

37. Kohler, H.-P. E.; Kohler Staub, D.; Focht, D. D. Appl. Environ. Microbiol., 1988, 54, 2683-2688.

38. Suske, W. A.; Held, M.; Schmid, A.; et al. J. Biol. Chem., 1997, 272, 24257-24265.

39. Schmid, A.; Kohler, H.-P. E.; Engesser, K.-H. J. Mol. Cat. B.: Enzymatic, 1998, 5, 311-316.

40. Held, M.; Schmid, A.; Kohler, H. P.; et al. Biotechnol. Bioeng., 1999, 62, 641-648.

41. Held, M.; Suske,W.; Schmid, A.; et al. J. Mol. Cat. B.: Enzymatic, 1998, 5, 87-93.

42. Hasegawa, J.; Ohashi, T. Chemistry Today, 1996, 14, 44-48.

43. Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science, 1997, 277, 936-938.

44. Wubbolts, M. G.; Reuvekamp, P.; Witholt, B. Enzyme Microbial Technol., 1994, 16, 608-615.

45. Wubbolts, M. G.; Hoven, J.; Melgert, B.; Witholt, B. Enzyme Microbial Technol., 1994, 16, 887-894.

46. Panke, S.; Witholt, B.; Schmid, A.; Wubbolts, M. G. Appl. Environ. Microbiol., 1998, 64, 2032-2043.

47. Mathys, R. G.; Schmid, A.; Witholt, B. Biotechnol. Bioeng., 1999, 64, 459-477.

48. Mathys, R. G.; Kut, O. M.; Witholt, B. J. Chem. Technol. Biotechnol., 1998, 71, 315-325.

Was this article helpful?

0 0