Immobilized Cells for Preparing Phenylalanine

The present invention is primarily concerned with the production of phenylalanine from a precursor thereof, notably phenylpyruvate or phenylpyruvic acid, via transaminase. One embodiment of the invention utilizes immobilized whole cells having transaminase activity to produce phenylalanine from phenylpyruvate. However, according to a further embodiment of the invention, the desired enzyme activity may be obtained by using ruptured or permeabilized cells, as such or as purified fractions thereof, either in the free or immobilized state to obtain phenylalanine.

The invention also contemplates the possibility of preparing phenylalanine from cinnamic acid using immobilized whole cells having phenylalanine ammonia lyase activity.

The production of phenylalanine from phenylpyruvate has been attempted by numerous investigators. There are two possible routes to accomplish this transformation. One is by transamination with an appropriate amine donor while the other is direct reductive amination using a biological energy source such as NAD or NADP.

Sakurai (J. Biochemistry 43, 851, 1956) attempted the preparation of optically active amino acids via transamination. Sakurai used crude pig heart transaminase (freshly obtained) and found that after 20 hours the yield of phenylalanine reached a maximum of 78% when aspartic acid was used with a small amount of glutamic acid. When aspartic acid alone was used, the yield was only 70%. Sakurai concluded that both amino acids should be present for maximal yields. He explained this result as a coupled system in which glutamic acid was the amine donor for phenylalanine and the aspartic acid served to regenerate the glutamic acid.

Oishi ("The Microbial Production of Amino Acids", John Wiley & Sons, K. Yamada et al Eds. 1972, Chap. 16) reviewed the production of phenylalanine from precursor keto acids. He noted a maximum yield of 63.5% phenylalanine was obtained by Asai in screening a large number of microbes which had been dried. This yield was obtained from a strain of Alcaligenes faecalis. The two strains of E. coli surveyed showed a 38.5% and a 53% yield under the reaction conditions used. Asai obtained yields as high as 70.6% when the amine donor was a combination of L-aspartate, L-glutamate and L-leucine. Yields with aspartate in two-fold excess were only 54.5%.

It will be appreciated that the yields noted above with respect to the indicated prior procedures are not suitable for an economic industrial process. Yields in excess of 90% are generally considered essential for a commercially viable process.

Oishi also reported that, by using a coupled enzyme system, Kitai was able to reach 76.8% yield. The coupled system was a yeast alcohol dehydrogenase with beef liver glutamate dehydrogenase and the Serratia marscescens glutamate-phenylalanine amino transferase. The reaction was driven by the removal of acetaldehyde by semicarbazide. Additionally, Kitai was able to drive the reaction to the expected 100% yield of L-phenylalanine by use of a coupled system for reductive amination in which E. coli were used to provide NADP. Glutamate, which served as the amine donor, was the limiting reagent.

Wandrey et al (U.S. Pat. No. 4,304,858) describe a coupled system (with formate dehydrogenase) for the production of phenylalanine from phenylpyruvate while providing exogenous NAD or NADH. The system is also applicable when using alphahydroxycarboxylic acid for the precursor as illustrated in U.S. Pat. No. 4,326,031. In both of these systems, however, it is necessary to use the reagent NAD or NADH and to use a coupled system in order to regenerate this expensive and labile material.

The available literature reviewed above indicates that only when using coupled systems are high yields of phenylalanine obtained from phenylpyruvate. When uncoupled systems are used, yields no higher than 71% have been obtained using three different amine donors.

An important object of the invention is to provide a process for the production of phenylalanine in high yields from phenylpyruvic acid or phenylpyruvate via transaminase. A more specific object is to prepare phenylalanine by such a process which involves a single step, with a single amine donor and requiring neither a coupled system nor the addition of expensive cofactor reagents such as NADP or NAD. Other objects will also be hereinafter apparent.

According to the invention, phenylalanine is produced from phenylpyruvic acid or phenylpyruvate by transaminase using immobilized whole cells. In another embodiment of the invention, cells are used which have been ruptured or permeabilized so as to release their transaminase activity. These ruptured or permeabilized cells may be in the free or immobilized state.

In our above-mentioned earlier applications, we have described the preparation and use of compositions comprising whole cells having enzymatic activity wherein the cells are immobilized by means of an insoluble, crosslinked polymer obtained by curing a polyazetidine prepolymer, carboxymethyl cellulose, polyurethane hydrogel prepolymer or polymethylene isocyanate. Preferably the immobilizing polymer is a polyazetidine polymer although the other disclosed polymers may be used. Advantageously the immobilized cells are coated onto beads or other particulate material.

Thus, for example, in Ser. No. 465,551, we have described the immobilizing of cells such as E. coli, Pseudomonas Dacunhae and the like with polyazetidine by mixing together equal parts of aqueous polyazetidine solution (such as Hercules Polycup 172) and cell paste, stirring to homogenity, dispersing the mixture on beads, e.g. ion exchange beads, which have been dried. The coated beads are then air-dried for use.

The present invention contemplates the use of immobilized cell compositions as described in said earlier applications, as well as other forms of immobilized cells, provided the immobilized cells have transaminase activity. According to the invention, phenylalanine is produced by contacting a phenylalanine precursor, specifically phenylpyruvate or phenylpyruvic acid, with the immobilized cell composition having transaminase activity in the presence of an amine donor, so that the precursor is converted to phenylalanine. The literature indicates that small amounts of pyridoxal-5-phosphate (P5P) are required by the transaminase as a co-factor. This material (P5P) is also used in the present process in conventional co-factor amounts.

The precursor may be used in the form of the free acid or as a salt thereof, e.g. the ammonium or alkali metal salt.

A wide variety of amine donors may be used provided these are active with the transaminase. Preferably the donor is L-glutamic acid, L-aspartic acid or mixture thereof. However, other amine donors such as L-leucine or L-isoleucine may also give highly useful results. Preferably the donor is used in excess and it appears that higher yields are obtained as the excess is increased up to, for example, 30-50% excess or even more.

Any microorganism demonstrating transaminase activity may be used for present purposes. A wide variety of these are known (see Table 16-3, page 441, of Oishi publication mentioned above). These include the following:

Alcaligenes faecalis

Pseudomonas cruciviae

Pseudomonas aeruginosa MT

Aerobacter aerogenes

Escherichia coli

Achromobacter cycloclastes

Sarcina lutea

Kluyvera citrophila

Pseudomonas fluorescens

Micrococcus lysodeikticus.

The reaction conditions used for carrying out the transaminase reaction according to the invention can be widely varied, as will be understood by those in the art. For example, an aqueous solution of the precursor can be passed through a column containing the immobilized cells containing transaminase activity and the amine donor. Optimum ratios of precursor to donor and to cells, and other operating conditions, can be readily determined for any specific situation without undue experimentation. Typically, however, the ratio of the amine donor to the precursor will be at least 1:1 and preferably 1.1:1 or higher, e.g. 3:1. A preferred ratio is 1.5-2 parts doner per part precursor, parts being on an equivalent weight basis.

Acid or alkaline pHs may be used although there will generally be a readily determined optimum pH for any particular set of conditions. Usually it is desirable to use a pH above 4, and generally one in the range of 5-10, although pHs outside these ranges may also be used. Temperature of 30.degree. to 40.degree. C. normally will be used although any temperature below transaminase deactivation can be used.

The invention is illustrated by the following examples:

Saccharomyces cervisiae, E. coli, Alcaligenes faecalis and Pseudomonas dachunae cells were immobilized in separate batches with polyazetidine as described in Ser. No. 465,551 (see, for instance, Example 8 thereof) by mixing equal parts of cell paste and aqueous polyazetidine solution (Hercules Polycup 172), stirring to homogenity at 25.degree. C. by hand mixing with a wooden stick. This mixture was dispersed on Amberlite ion exchange beads which had been air-dried. The thin film of paste/prepolymer mixture on the beads was allowed to air dry at 25.degree. C. One ml of each group of beads containing 0.2 grams of a microbial cells per ml of beads was then placed into a 50 ml Erlenmeyer flask containing 25 ml of a 0.1 M aqueous solution of sodium pyruvate and either L-glutamic acid, L-aspartic acid or mixture thereof as amine donor and 0.1 mM of P5P. These were then compared with the use of free cells of Pseudomonas and Alcaligenes under otherwise similar conditions. The results in terms of phenylalanine (PHE) produced were determined by HPLC analysis of the supernatant after 17 hours of shaking and are presented below in Table I.

The foregoing example demonstrates that whole cells immobilized as described and having transaminase activity may be effectively used to produce phenylalanine from phenylpyruvate precursor when an appropriate amine donor is employed.

In the control using Pseudomonas and Alcaligenes in the fresh, wet free state (unruptured and unpermeabilized), negligible transaminase activity was noted. However, on rupturing the cells or by permeabilizing them, activity was substantially increased. This is surprising because the literature indicates that dried cells, which would normally be considered lysed or permeabilized, do not give commercially acceptable conversions.

Accordingly, the use of ruptured or permeabilized cells, whether immobilized or in the free state, to prepare phenylalanine constitutes a further aspect of the invention. Various techniques may be used to rupture or permeabilize the cells for use according to the invention. For example, the cells may be ruptured by sonication or grinding as known in the art. Alternatively the cells may be permeabilized by chemical treatment, e.g. by treatment with a permeabilizing surfactant such as Triton X100. These treatments apparently allow the phenylpyruvate or phenylpyruvic acid to more readily contact the enzyme and thus improve activity whether or not the microorganism is immobilized.

The use of ruptured cells, and the effect of pH and amine donor level on the results, are described in the following example:

2 grams of free E. coli cells were sonicated for 10 minutes in order to rupture the cells after which they were incubated for 23 hours with 25 ml aqueous solution containing different amounts of ASP (0.10 M, 0.15 M and 0.2 M), 0.1 mM P5P and 0.1 M PPA at 37.degree. C. on a Dubnoff H.sub.2 O shaker. H.sub.3 PO.sub.4 was used for pH adjustment.

The results obtained are shown below in Table II.

As shown, free cells, when ruptured, give the best and most useful conversions at higher ASP concentrations, the data given in Table II indicating that for a commercially acceptable yield level, the amount of ASP should exceed the amount of substrate on a molar equivalence basis.

The various aspects of the invention are further illustrated by the following additional examples:

Table III shows the effect of pH on transaminase activity using E. coli whole cells immobilized in bead form as in Example 1. Three experiments were conducted using 2 ml of beads incubated in 15 ml of 0.1 M PPA, 0.1 mM P5P and 0.15 M ASP at 37.degree. C. for 24 hours. The pH was adjusted with 1 N NaOH or N HCl.

The data in table III shows that high yields of PHE similar to those obtainable with ruptured free cells can be obtained using immobilized E. coli.

Table IV below provides the results in terms of yield of PHE obtained using immobilized E. coli in a continuous column operation (300 ml, 3.5.times.70 cm) with 0.1 M PPA, 0.15 M ASP and 0.1 mM P5P.

While polyazetidine polymer is preferred for immobilizing the microorganisms for use herein, the invention contemplates the possibility of using any other suitable immobilizing substrate. As representative of such alternatives, there may be mentioned such materials as polyacrylamide, Kappa-carrageenan, hollow fiber devices, Hypol or XAD coated beads. These materials have been shown to give excellent yields although the activity of the immobilized cells may vary from one immobilizing substrate to another. The results obtained in terms of yields and activities, using different systems involving immobilized E. coli, are shown below in Table V. The process used involved continuous flow onto a column of immobilized cells as described of an aqueous solution of 0.1 M PPA, 0.15 M ASP and 0.1 mM P5P at a pH 8.3-8.5 (adjusted with NH.sub.4 OH) at 37.degree. C. Flow varied according to column activity and space occupied. Equilibrium was reached at optimal flow prior to taking readings.

Of the materials referred to in Table V, XAD is a macroreticular styrene-divinylbenzene resin; IRA 938 is an ion exchange bead resin comprising styrene-divinylbenzene containing tertiary amine substituents; and Hypol is a polyurethane foam. The Kappa-carrageenan gum was cut into particles before use. The hollow fiber device was a commercially available item.

As an alternative to the procedures described above, phenylalanine may be made from cinnamic acid by using immobilized whole cells which are high in phenylalanine ammonia-lyase activity. This aspect of the invention represents an improvement in the process described by Yamada et al, Applied and Environmental Microbiology Nov. 1981, pages 773-778, incorporated herein by reference.

Yamada et al describe the preparation of L-phenylalanine from trans-cinnamic acid by an enzymatic method using Rhodotorula glutinis containing L-phenylalanine ammonia-lyase activity. According to the present invention, Rhodotorula glutinis ATCC 10788 was grown as described by Yamada et al and the harvested cells were immobilized with polyazetidine prepolymer. 14.9 grams of cells were mixed with 14.9 grams of polyazetidine prepolymer and coated onto 13.8 grams of IRA 938 ion exchange resin and assayed for phenylalanine ammonia lyase activity. Cinnamic acid was added to the beads by mixing 1 ml of beads produced as above with 5 ml of assay mixture which contained 740 mg transcinnamic acid, 45 ml 28% ammonium hydroxide, pH 10 diluted to 80 ml. After 24 hours the supernatant was spotted on a cellulose TLC plate and developed in a mixture of butanol, acetic acid, water (4:1:1) and the plates were sprayed with 0.2% ninhydrin and ethanol. Standards of phenylalanine were used for comparison and an estimation based on intensity and size of the spot indicated that 0.05 mg/ml of phenylalanine had been produced.

It will be appreciated that various modifications may be made in the invention described herein.