AU7316194A

AU7316194A - Glyoxylic acid/aminomethylphosphonic acid mixtures prepared using a microbial transformant

Info

Publication number: AU7316194A
Application number: AU73161/94A
Authority: AU
Inventors: David Leroy Anton; Robert Di Cosimo
Original assignee: EI Du Pont de Nemours and Co
Current assignee: EIDP Inc
Priority date: 1993-03-03
Filing date: 1994-06-29
Publication date: 1996-01-25
Also published as: JPH10501984A; WO1996000793A1; BR9408601A

Description

TITLE

GLYOXYLIC ACID/AMINOMETHYLPHOSPHONIC ACID MLXTURES

PREPARED USING A MICROBIAL TRANSFORMANT

BACKGROUND OF THE INVENTION 1. Field of the Invention:

This invention relates to a process for the production of mixtures of glyoxylic acid and aminomethylphosphonic acid (AMP A), where glycolic acid and oxygen are reacted in an aqueous solution in the presence of AMPA and catalysts consisting of a genetically-engineered microbial transformant which expresses the enzyme glycolate oxidase from spinach ((S)-2-hydroxy-acid oxidase, EC 1.1.3.15), and catalase (EC 1.11.1.6). The glyoxylic acid/aminomethylphosphonic acid mixtures prepared in this manner are useful intermediates in the production of N-(phosphonomethyl)glycine, a broad-spectrum, post-emergent herbicide. 2. Description of the Related Art: Glycolate oxidase, an enzyme commonly found in leafy green plants and mammalian cells, catalyzes the oxidation of glycolic acid to glyoxylic acid, with the concomitant production of hydrogen peroxide:

HOCH₂CO₂H + O₂ → OCHCO₂H + H₂O₂

N. E. Tolbert et al., J. Biol. Chem.. Vol. 181, 905-914 (1949) first reported an enzyme, extracted from tobacco leaves, which catalyzed the oxidation of glycolic acid to formic acid and CO via the intermediate formation of glyoxylic acid. The addition of certain compounds, such as ethylenediamine, limited the further oxidation of the intermediate glyoxylic acid. The oxidations were carried out at a pH of about 8, typically using glycolic acid concentrations of about 3-40 mM (millimolar). The optimum pH for the glycolate oxidation was reported to be 8.9. Oxalic acid (100 mM) was reported to inhibit the catalytic action of the glycolate oxidase. Similarly, K. E. Richardson and N. E. Tolbert (J. Biol. Chem.. Vol. 236, 1280- 1284 ( 1961 )) showed that buffers containing tris(hydroxymethyl)amino- methane (TRIS) inhibited the formation of oxalic acid in the glycolate oxidase catalyzed oxidation of glycolic acid. C. O. Clagett, N. E. Tolbert and R. H. Burris (J. Biol. Chem.. Vol. 178, 977-987 (1949)) reported that the optimum pH for the glycolate oxidase catalyzed oxidation of glycolic acid with oxygen was about 7.8-8.6, and the optimum temperature was 35-40°C. I. Zelitch and S. Ochoa (7. Biol. Chem. (Vol. 201, 707-718 (1953)) and J. C. Robinson et al. (J. Biol. Chem.. Vol. 237, 2001-2009 (1962)) reported that the formation of formic acid and CO₂ in the spinach glycolate oxidase-catalyzed oxidation of glycolic acid resulted from the nonenzymatic reaction of H₂O₂ with glyoxylic acid. They observed that addition of catalase, an enzyme that catalyzes the decomposition of H₂O₂, greatly improved the yields of glyoxylic acid by suppressing the formation of formic acid and CO₂. The addition of FMN (flavin mononucleotide) was also found to greatly increase the stability of the glycolate oxidase. N. A. Frigerio and H. A. Harbury (J. Biol. Chem.. Vol. 231, 135-157

(1958)) have reported on the preparation and properties of glycolic acid oxidase isolated from spinach. The purified enzyme was found to be very unstable in solution; this instability was ascribed to the relatively weak binding of flavin mononucleotide (FMN) to the enzyme active site, and to the dissociation of enzymaticaUy active tetramers and/or octamers of the enzyme to enzymaticaUy- inactive monomers and dimers, which irreversibly aggregate and precipitate. The addition of FMN (flavin mononucleotide) to solutions of die enzyme greaύy increased its stability, and high protein concentrations or high ionic strength maintained the enzyme as octamers or tetramers. There are numerous other references to the oxidation of glycolic acid catalyzed by glycolate oxidase. The isolation of the enzyme (and an assay method) are described in the following references: I. Zelitch, Methods of Enzvmologv. Vol. 1, Academic Press, New York, 1955, p. 528-532 (from spinach and tobacco leaves), M. Nishimura et al., Arch. Biochem. Biophys.. Vol. 222, 397-402 (1983) (from pumpkin cotyledons), H. Asker and D. Davies, Biochim. Biophys. Acta. Vol. 761, 103-108 (1983) (from rat liver), and M. J. Ernes and K. H. Erismann, Int. J. Biochem.. Vol. 16, 1373-1378 (1984) (from Lemna Minor L). The structure of the enzyme has also been reported: E. Cederlund et al., Eur. J. Biochem.. Vol. 173, 523-530 (1988) and Y. Lindquist and C. Branden, J. Biol. Chem.. Vol. 264, 3624-3628, (1989).

SUMMARY OF THE INVENTION This invention relates to the preparation of mixtures of glyoxylic acid (or a salt thereof) and aminomethylphosphonic acid (AMPA)(or a salt thereof), by oxidizing glycolic acid with oxygen in aqueous solution and in the presence of AMPA and two catalysts: a genetically-engineered microbial transformant which expresses the enzyme glycolate oxidase from spinach ((S)-2-hydroxy-acid oxidase, EC 1.1.3.15), and catalase (EC 1.11.1.6). Such mixtures of glyoxylic acid and AMPA are useful for the preparation of N-(phosphonomethyl)glycine, a post- emergent herbicide. BRIEF DESCRIPTION OF THE BIOLOGICAL DEPOSITS

Applicants have made the following biological deposits under the terms of the Budapest Treaty:

Depositor Identification International Depository Reference Designation Date of Deposit

Aspergillus nidulans T17 NRRL Y-21000 24 September 1992

Pichia pastoris, GS115-MSP10 NRRL Y-21001 24 September 1992

Hansenula polymorplia, G01 NRRL Y-21065 30 March 1993

As used herein, "NRRL" refers to the Northern Regional Research

Laboratory, Agricultural Research Service Culture Collection international depository located at 11815 N. University Street, Peioria, IL 61604 U.S.A. The "NRRL No." is the accession number to cultures on deposit at the NRRL. DESCRIPTION OF THE PREFERRED EMBODIMENTS A related patent, U.S. 5,135,860 (August 4, 1992), "Process for Preparing Glyoxylic Acid/Aminomethylphosphonic Acid Mixtures" (PCT/US92/09419), describes a process for the enzymatic conversion of glycolic acid to glyoxylic acid in the presence of oxygen, aminomethylphosphonic acid (AMPA), and the soluble enzymes glycolate oxidase and catalase. It was unexpected that the addition of AMPA to enzymatic oxidations of glycolic acid within the pH range of 7 to 9 would result in the high yields of glyoxylic acid obtained, since an unprotonated amine was believed to be necessary to form an oxidation-resistant N-substituted hemiaminal and/or imine complex with glyoxylate. The pKa of the protonated amine of AMPA is reported to be 10.8 (Lange's Handbook of Chemistry. J. A. Dean, ed., McGraw-Hill, New York, 1979, 12th edition), therefore AMPA would be present in the reaction mixture predominantly as the protonated ammonium ion and not be expected to form such protective complexes with glyoxylate. The present invention provides an improvement to the above process in the form of a whole microbial cell as a catalyst for this process.

The previously reported use of soluble enzymes as catalysts for the production of glyoxylic acid/aminomethylphosphonic acid mixtures poses several problems: catalyst recovery for reuse is not easily performed, enzyme activity is not as stable as with immobilized enzyme or whole cell catalyst systems, and soluble glycolate oxidase is not stable to the sparging of the reaction mixture with oxygen (required to increase the rate of oxygen dissolution and, thus, reaction rate). A second related patent application, U.S.S.N. 08/085,488, filed 1 July 1993,

"Glycolate Oxidase Production" (PCT US94/ ), described the construction of several transformants of Aspergillus nidulans, using genetic engineering techniques commonly known to those skilled in the art, which express the glycolate oxidase from spinach as well as an endogenous catalase. Several advantages for the use of these whole-cell catalysts over the use of soluble enzymes for the production of glyoxylic acid/aminomethylphosphonic acid mixtures in the present invention are:

(1) the whole-cell catalysts are easily recovered from the reaction mixture at the conclusion of the reaction for reuse, whereas the soluble enzyme is only recovered with great difficulty and loss of activity;

(2) they are more stable than the soluble enzyme, both for the number of catalyst turnovers obtained versus the soluble enzyme, as well as for recovered enzyme activity at the conclusion of a reaction; and

(3) most importantly, they are stable to reaction conditions where oxygen or an oxygen-containing gas is sparged into the reaction mixture to increase the rate of oxygen dissolution and reaction rate, where under similar reaction conditions the soluble glycolate oxidase is rapidly denatured. The Aspergillus nidulans transformants were prepared by first cloning the spinach gene which codes for glycolate oxidase, then introducing this gene into a strain of Aspergillus nidulans which already produced an endogenous catalase. The resulting transformants were cultured in either minimal media or SYG rich media in shaker flasks or fermenters, and additionally, different agents such as oleic acid (OL), hydroxyacetic acid (HA), or corn steep liquor (CSL) were added to the media to increase levels of expression of glycolate oxidase and/or catalase. The different transformants were then screened by assaying the whole cells (untreated) for catalase and glycolate oxidase activity, and by running reactions with the cells as catalysts for the oxidation of glycolic acid to glyoxylic acid.

When used as catalyst for the oxidation of glycolic acid to glyoxylic acid, Aspergillus nidulans cells were not pre-treated or permeabilized to increase accessibility of the reaction mixture to the enzymes in the interior of the cells; some permeabilization of the cells may take place, possibly from exposure to the reaction mixture or any of its components, or by freezing and thawing, which was used to store the whole cell catalysts until needed.

The use of Aspergillus nidulans transformants as a whole cell catalyst for the production of glyoxylic acid has been previously demonstrated (PCT/US93/00077, "Oxidation of Glycolic Acid to Glyoxylic Acid using a

Microbial Cell Transformant as Catalyst"), where employing an amine buffer (e.g., ethylenediamine or TRIS) capable of forming a chemical adduct with glyoxylic acid resulted in yields of glyoxylic acid as high as 98%; in that case, the concentration of endogenous catalase within the cell was sufficient to limit the oxidation of glyoxylate to formate by byproduct hydrogen peroxide. In the present case, where AMPA is less effective at protecting glyoxylate from oxidation by hydrogen peroxide, the concentration of endogenous A. nidulans catalase within the cells was insufficient to produce the desired high yields of glyoxylate. It has been discovered that adding additional soluble catalase (e.g., from A. niger or S. cerevisiae) to the AMPA-containing reaction mixture which uses these whole cells as a source of glycolate oxidase activity can result in yields of glyoxylate similar to those achieved when using soluble or immobilized enzymes.

When using an A. nidulans transformant as catalyst for the oxidation of glycolate/AMPA mixtures, both glyoxylate and hydrogen peroxide are produced within the cells containing the glycolate oxidase, and in the absence of added soluble catalase, glyoxylate is rapidly oxidized to formate (for example: 87% formate, 1.7% glyoxylate as in Example 2). It was unanticipated that high yields of glyoxylate could be obtained by simply adding soluble catalase to the reaction, since (1) the relative concentration of glyoxylate and hydrogen peroxide produced throughout the course of the reaction was greater within the cells than in the surrounding aqueous solution, (2) the concentration of endogenous catalase within A. nidulans cells was typically 2% to 15% of the concentration of soluble catalase usually added to a reaction mixture, and (3) the hydrogen peroxide would have to diffuse from inside the cell into the surrounding aqueous solution in order to be decomposed to water and oxygen by the soluble catalase. However, yields of glyoxylate as high as 94% have been obtained by simply adding soluble catalase to the reaction mixture at the same concentrations as have been previously employed when using only soluble catalase and soluble glycolate oxidase as catalysts (U.S. Patent No. 5,135,860, PCT/US92/09419). A second microbial cell catalyst which has been utilized in the present invention is a transformant of Hansenula polymorpha (a methylotrophic yeast) which expresses the glycolate oxidase enzyme from spinach, as well as an endogenous catalase. Several transformants of H. polymorpha having sufficient glycolate oxidase activity have been prepared by inserting the DNA for glycolate oxidase into an expression vector under the control of the formate dehydrogenase (FMD) promoter. H. polymorpha was transformed with this vector and a strain producing high levels of glycolate oxidase was selected and designated H. polymorpha G01 (NRRL Y-21065). H. polymorpha cell catalysts were typically prepared by first growing an inoculum of an H. polymorpha transformant in 500 ml of YPD (Difco), pΗ 4.4. This culture was then inoculated into a fermenter containing 10 L of Yeast Nitrogen Base (YBN, Difco) w/o amino acids (14 g), ammonium sulfate (50 g) and methanol (100 g), at pΗ 5.0. The fermenter was operated for 42.5 h at 37°C, an agitation rate of 400 rpm, constant pΗ of 5.0, 40% dissolved oxygen (controlled), and 14 psig of air. At the conclusion of the fermentation, 1.0 kg of glycerol was added and the cells harvested by centrifugation, frozen in liquid nitrogen, and stored at -80°C.

A third microbial cell catalyst which has been utilized in the present invention is a transformant of Pichia pastoris (a methylotrophic yeast) which expresses the glycolate oxidase enzyme from spinach, as well as an endogenous catalase. Several transformants of P. pastoris having sufficient glycolate oxidase activity have been prepared by inserting a DNA fragment containing the spinach glycolate oxidase gene into a P. pastoris expression vector (pΗIL-D4) such as to be under control of the methanol inducible alcohol oxidase I promoter, generating the plasmidpMPl. P. pastoris strain GTS115 (NRRL Y-15851) was transformed by plasmid pMPl and a selection was done such as to allow integration of the linearized plasmid pMPl into the chromosomal alcohol oxidase I locus and replacement of alcohol oxidase gene with glycolate oxidase gene. A pool of such transformants were next selected for maximal number of integrated copies of the expression cassette. A high copy number transformant designate P. pastoris strain GS115-MSP10 was isolated and deposited in the NRRL, Peoria, Illinois.

P. pastoris cells were typically prepared by growing an inoculum in 100 mL of YNB containing 1% glycerol. After 48 hours growth at 30°C, the cells were transferred into a fermenter containing 10 L of media composed of yeast nitrogen base (YBN) w/o amino acids (134 g), glycerol (100 g), and biotin (20 mg). The fermentation was operated at pH 5.0 (controlled with NH4OH), 30°C, agitation rate of 200 rpm, aeration of 5 slpm, 5 psig of air, and dissolved oxygen maintained at no lower than 50% saturation. When glycerol was depleted, the cells were induced to express glycolate oxidase by growth in the same media except that methanol (50 g) was substituted for glycerol. Glycolate oxidase activity during induction was followed by enzyme assay. After 24 hours of induction the cells were harvested following treatment with glycerol (1 kg). Following harvest the cells were frozen in liquid nitrogen and stored at -80°C. Unlike A. nidulans, H. polymorpha and P. pastoris cell transformants required permeabilization prior to use as catalyst for the oxidation of glycolic acid to glyoxylic acid. A variety of known methods of permeabilization were useful for preparing cells with sufficient glycolate oxidase activity (see Felix, H. Anal. Biochemistry. Vol. 120, 211-234, (1982)); typically, a suspension of 10 wt % wet cells was suspended in a 0.1% (v/v) "TRITON" X-100/20 mM phosphate buffer (pH 7.0) for 15 minutes, then frozen in liquid nitrogen, thawed, and washed with 20 mM phosphate/0.1 mM FMN buffer (pH 7.0).

When using either H. polymorpha and P. pastoris cell transformants as catalyst for the oxidation of glycolic acid/AMPA mixtures, the addition of soluble A. niger catalase was again found to be necessary for the production of high yields of glyoxylic acid. Although the accessible catalase activity of permeabilized cells of H. polymorpha or P. pastoris was ca. 10-fold greater than that of A. nidulans, the endogenous catalase of either methylotrophic yeast was less effective at decomposing byproduct hydrogen peroxide in reaction mixtures containing AMPA than catalase from A. nidulans or A. niger. Addition of soluble A. niger catalase, or permeabilized whole cells of Saccharomyces cerevisiae as a supplemental catalase source, resulted in marked improvements in glyoxylic acid production when compared to reactions run in the absence of added catalase.

A fourth microbial cell catalyst which has been utilized in the present invention is a transformant of Escherichia coli (a bacteria) which expresses the glycolate oxidase enzyme from spinach, as well as an endogenous catalase. Such an E. coli transformant was prepared as described in Macheroux et al., Biochem. Biophys. Acta. Vol. 1132, 11-16 (1992). E. coli transformants expressing glycolate oxidase activity were not permeabilized prior to use as catalyst in the present invention. Many of the deficiencies of using soluble soluble enzymes as catalysts in the present application have been eliminated by employing whole microbial cell transformants (either unpermeabilized or permeabilized) as catalyst. Recovery and reuse of the whole-cell catalyst was easily performed by centrifugation or by filtering the catalyst away from the reaction mixture and recycling it to fresh reaction mixture; in this manner, turnover numbers for glycolate oxidase of as high as 10⁶ have been obtained. The ability to bubble oxygen through the reaction mixture without rapidly denaturing the glycolate oxidase (as is observed when using the soluble enzyme) resulted in increases in the reaction rate of at least ten- fold over reactions where the reaction mixture is not bubbled, and this increase in rate significantly reduces the cost of manufacture for this process.

The glycolate oxidase activity (added as whole microbial cell transformant) used in the reaction should be present in an effective concentration, usually a concentration of 0.01 to about 100 IU/mL, preferably about 0.1 to about 10 IU/mL. An IU (International Unit) is defined as the amount of enzyme that will catalyze the transformation of one micromole of substrate per minute. A procedure for the assay of this enzyme is found in I. Zelitch and S. Ochoa, J. Biol. Chem.. Vol. 201, 707-718 (1953). This method is also used to assay the activity of recovered or recycled glycolate oxidase. The concentration of catalase should be 50 to 100,000 IU/mL of reaction mixture, preferably 500 to 14,000 IU/mL. It is preferred that both the glycolate oxidase and catalase enzymes be present within the same microbial cell (in the accompanying examples, a transformant of A. nidulans, H. polymorpha, P. Pastoris, or E. coli). If the concentration of endogenous catalase within the cells is insufficient to efficiently decompose the hydrogen peroxide produced (as in the accompanying examples), an additional source of catalase (e.g., soluble Aspergillus niger catalase, or whole cells of Saccharomyces cerevisiae) may be added to supplement the endogenous catalase present. Additionally, the catalase and glycolate oxidase concentrations should be adjusted within the above ranges so that the ratio (measured in IU for each) of catalase to glycolate oxidase is at least about 250: 1. Flavin mononucleotide

(FMN) is an optional added ingredient, used at a concentration of 0.0 to 2.0 mM, preferably 0.01 to 0.2 mM.

In view of the above effective concentration ranges, it should be appreciated that the improved process according to the instant invention encompasses and is intended to cover the use of all microbial cell catalysts, i.e., whole cell catalysts, that express glycolate oxidase and/or catalase in these operative ranges. Thus, for purposes of this invention the term "whole microbial cell catalyst" includes, by way of example but not limited thereto, genetically- engineered microbial transformants of the preferred embodiments and selected strains of either a natural or mutated microbe of enhanced expression capacity and/or mixtures thereof that actually result in effective concentration ranges as set out above.

Glycolic acid (2-hydroxyacetic acid) is available commercially. In the present reaction its initial concentration is in the range of 0.10 M to 2.0 M, preferably between 0.25 M and 1.0 M. It can be used as such or as a compatible salt thereof, that is, a salt that is water-soluble and whose cation does not interfere with the desired conversion of glycolic acid to glyoxylic acid, or the subsequent reaction of the glyoxylic acid product with the aminomethylphosphonic acid to form N-(phosphonomethyl)glycine. Suitable and compatible salt-forming cationic groups are readily determined by trial. Representative of such salts are the alkali metal, alkaline earth metal, ammonium, substituted ammonium, phosphonium, and substituted phosphonium salts.

The conversion of glycolic acid to glyoxylic acid is conveniently and preferably conducted in aqueous media. Aminomethylphosphonic acid (AMPA), or a suitable salt thereof, is added to produce a molar ratio of AMPA/glycolic acid (starting amount) in the range of from 0.01/1.0 to 3.0/1.0, preferably from 0.25/1.0 to 1.05/1.0. After combining AMPA and glycolic acid in an aqueous solution, the pH of the resulting mixture is adjusted to a value between 6 and 10, preferably between 7.0 and 8.5. Within this pH range, the exact value may be adjusted to obtain the desired pH by adding any compatible, non-interfering base, including alkali metal hydroxides, carbonates, bicarbonates and phosphates. The pH of the reaction mixture decreases slightly as the reaction proceeds, so it is often useful to start the reaction near the high end of the maximum enzyme activity pH range, about 9.0-8.5, and allow it to drop during the reaction. The pH can optionally be maintained by the separate addition of a non-interfering inorganic or organic buffer, since enzyme activity varies with pH.

It is understood that glycolic and glyoxylic acid are highly dissociated in water, and at pH of between 7 and 10 are largely if not substantially entirely present as glycolate and glyoxylate ions. It will also be appreciated by those skilled in the art that glyoxylic acid (and its conjugate base, the glyoxylate anion) may also be present as the hydrate, e.g. (HO)₂CHCOOH and/or as the hemiacetal, HOOCCH(OH)OCH(OH)COOH, which compositions and their anionic counterparts are equivalent to glyoxylic acid and its anion for the present purpose of being suitable reactants for N-(phosphonomethyl)glycine formation. Oxygen (O₂), the oxidant for the conversion of the glycolic acid to glyoxylic acid, may be added as a gas to the reaction by agitation of the liquid at the gas-liquid interface, through a membrane permeable to oxygen, or by sparging (bubbling) oxygen through the reaction mixture. It is believed that under most conditions, the reaction rate is at least partially controlled by the rate at which oxygen can be dissolved into the aqueous medium. Thus, although oxygen can be added to the reaction as air, it is preferred to use a relatively pure form of oxygen, and even use elevated pressures. Although no upper limit of oxygen pressure is known, oxygen pressures up to 50 atmospheres may be used, and an upper limit of 15 atmospheres is preferred. Agitation is important to maintaining a high oxygen dissolution (hence reaction) rate. Any convenient form of agitation is useful, such as stirring.

The reaction temperature is an important variable, in that it affects reaction rate and the stability of the enzymes. A reaction temperature of about 0°C to 40°C may be used, but the preferred reaction temperature range is from 5°C to 15°C. Operating in the preferred temperature range maximizes recovered enzyme activity at the end of the reaction. The temperature should not be so low that the aqueous solution starts to freeze. Temperature can be controlled by ordinary methods, such as, but not limited to, by using a jacketed reaction vessel and passing liquid of the appropriate temperature through the jacket. The reaction vessel may be constructed of any material that is inert to the reaction ingredients.

Following the cessation of contacting the reaction solution with O₂, the microbial cell catalyst may be removed by decantation, filtration or centrifugation and reused. Flavin mononucleotide (FMN) may optionally be removed by contacting the solution with activated carbon. The solution containing glyoxylic acid and aminomethylphosphonic acid (which are believed to be in equilibrium with the corresponding imine), is treated in accordance with any of the processes known to the art for producing N-(phosphonomethyl)glycine.

Catalytic hydrogenation is a preferred method for preparing N-(phosphono- methyl)glycine from a mixture containing glyoxylic acid and aminomethyl- phosphonic acid. Hydrogenation catalysts suitable for this purpose include (but are not limited to) the various platinum metals, such as iridium, osmium, rhodium, ruthenium, platinum, and palladium; also various other transition metals such as cobalt, copper, nickel and zinc. The catalyst may be unsupported, for example as Raney nickel or platinum oxide; or it may be supported, for example as platinum on carbon, palladium on alumina, or nickel on kieselguhr. Palladium on carbon, nickel on kieselguhr and Raney nickel are preferred. The hydrogenation can be performed at a pH of from 4 to 11, preferably from 5 to 10. The hydrogenation temperature and pressure can vary widely. The temperature is generally in the range of 0°C to 150°C, preferably from 20°C to 90°C, while the H₂ pressure is generally in the range of from about atmospheric to about 100 atmospheres, preferably from 1 to 10 atmospheres. N-(phosphonomethyl)glycine, useful as a post-emergent herbicide, may be recovered from the reduced solution, whatever the reducing method employed, by any of the recovery methods known to the art. In the following examples, which serve to further illustrate the invention, the yields of glyoxylate, formate and oxalate, and the recovered yield of glycolate, are percentages based on the total amount of glycolic acid present at the beginning of the reaction. Analyses of reaction mixtures were performed using high pressure liquid chromatography: organic acid analyses were performed using a Bio-Rad HPX-87H column, and AMPA and N-(phosphonomethyl)glycine were analyzed using a Bio-Rad Aminex Glyphosate Analysis column.

Microbial cell transformants were assayed for glycolate oxidase activity by accurately weighing ca. 5-10 mg of the wet cells into a 3-mL quartz cuvette containing a magnetic stirring bar and 2.0 mL of a solution which was 0.12 mM in 2,6-dichlorophenol-indophenol (DCIP) and 80 mM in TRIS buffer (pH 8.3). The cuvette was capped with a rubber septum and the solution deoxygenated by bubbling with nitrogen for 5 min. To the cuvette was then added by syringe 40 μL of 1.0 M glycolic acid/1.0 M TRIS (pH 8.3), and the mixture stirred while measuring the change in absorption with time at 605 nm (ε = 22,000).

Catalase activity was assayed by accurately weighing ca. 2-5 mg of the wet cells into a 3-mL quartz cuvette containing a magnetic stirring bar and 2.0 mL of a distilled water, then adding 1.0 mL of 59 mM hydrogen peroxide in 50 mM phosphate buffer (pH 7.0) and measuring the change in absorption with time at 240 nm (ε = 39.4). Glycolate oxidase and catalase activities of the A. nidulans wet cells (unpermeabilized) cultured in different media ranged from 0.5-4.0 DCIP IU/gram wet cells for glycolate oxidase and 500-7000 IU/gram wet cells for endogenous catalase. Glycolate oxidase and catalase activities of the H. polymorpha or P. pastoris wet cells (permeabilized) cultured in different media ranged from 20-60 DCIP IG/gram wet cells for glycolate oxidase and 30,000-80,000 IU/gram wet cells for endogenous catalase. EXAMPLE 1

A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino¬ methylphosphonic acid (0.375 M), isobutyric acid (0.100 M, ΗPLC internal standard), and flavin mononucleotide (0.01 mM) at pΗ 8.3 (adjusted with 50% NaOΗ), and the solution cooled to 5°C. To the reactor was then added 26 g of frozen (-80°C) Aspergillus nidulans FT17SYCSL/OL (124 IU glycolate oxidase and 57,800 IU catalase) and 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma), the cells allowed to thaw at 5°C, and the pΗ of the resulting mixture re¬ adjusted to 8.3 with 50% NaOΗ. This mixture was stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at 50 mL/min. The reaction was monitored by taking a 0.40 mL aliquot of the reaction mixture at regular intervals, filtering the aliquot using a Millipore "ULTRAFREE"-MC 10,000 NMWL Filter Unit, and analyzing the filtrate by ΗPLC. After 10 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 91%, 0%, and 7.9%, respectively, with 2.5% recovery of glycolic acid. The final activities of glycolate oxidase and Aspergillus niger catalase were 11% and 87% of their initial values.

EXAMPLE 2 (COMPARATIVES The reaction described in Example 1 was repeated, except that the addition of Aspergillus niger soluble catalase (Sigma) was omitted. After 22 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 1.7%, 0%, and 87.4%, respectively, with complete conversion of glycolic acid. There was no detectable glycolate oxidase or catalase activity in the Aspergillus nidulans. FT17SYCSL/OL cells at the end of the reaction. EXAMPLE 3

A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino¬ methylphosphonic acid (0.375 M), isobutyric acid (0.100 M, ΗPLC internal standard), and flavin mononucleotide (0.01 mM) at pΗ 8.3 (adjusted with 50% NaOΗ), and the solution cooled to 5°C. To the reactor was then added 26 g of frozen (-80°C) Aspergillus nidulans FT17SYCSL/OL (124 IU glycolate oxidase and 57,800 IU catalase) and 5.6 x 10⁵ units of Aspergillus niger soluble catalase (Sigma), the cells allowed to thaw at 5°C, and the pH of the resulting mixture re¬ adjusted to 8.3 with 50% NaOH. This mixture was stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at

50 mL/min. After 10 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 83%, 0%, and 9.4%, respectively, with 4.4% recovery of glycolic acid. The final activities of glycolate oxidase and Aspergillus niger catalase were 17% and 88% of their initial values. EXAMPLE 4

A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino¬ methylphosphonic acid (0.375 M), isobutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3 (adjusted with 50% NaOH), and the solution cooled to 5°C. To the reactor was then added 15 g of frozen (-80°C) Aspergillus nidulans FT17SYCSL/OL (72 IU glycolate oxidase and 33,300 IU catalase) and 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma), the cells allowed to thaw at 5°C, and the pH of the resulting mixture re¬ adjusted to 8.3 with 50% NaOH. This mixture was stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at

50 mL/min. After 17 hours, the yields of glyoxylic acid, oxalic acid, arid formic acid were 76%, 0%, and 4.1%, respectively, with 16.8% recovery of glycolic acid. The final activities of glycolate oxidase and Aspergillus niger catalase were 7% and 49% of their initial values. EXAMPLE 5

A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino¬ methylphosphonic acid (0.375 M), isobutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3 (adjusted with 50% NaOH), and the solution cooled to 5°C. Frozen (-80°C) Aspergillus nidulans FT17SYCSL/OL (26 g, 124 IU glycolate oxidase and 57,800 IU catalase) were allowed to thaw at 5°C, then washed with 2 x 100 mL of KH₂PO₄ (50 mM, pH 7.0)/FMN (0.01 mM) buffer at 5°C and the washed cells added to the reactor with 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). The pH of the resulting mixture was re-adjusted to 8.3 with 50% NaOH, then stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at 50 mL/min. After 11.5 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 94%, 3.5%, and 2.7%, respectively, with 1.3% recovery of glycolic acid. The final activities of glycolate oxidase and Aspergillus niger catalase were 7% and 77% of their initial values.

EXAMPLE 6 A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino¬ methylphosphonic acid (0.375 M), isobutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3 (adjusted with 50% NaOH), and the solution cooled to 5°C. Freshly-harvested Aspergillus nidulans FT17SYCSL/OL (37 g, 44 IU glycolate oxidase and 57,800 IU catalase) were washed with 4 x 100 mL of KH₂PO₄ (50 mM, pH 7.0)/FMN (0.01 mM) buffer at 5°C and the washed cells added to the reactor with 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). The pH of the resulting mixture was re-adjusted to 8.3 with 50% NaOH, then stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at 50 mL/min. After 15 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 84%, 0%, and 9.4%, respectively, with 6.4% recovery of glycolic acid. At the completion of the reaction, the reaction mixture was centrifuged at

5°C and the supernatant decanted. The resulting pellet of Aspergillus nidulans cells was resuspended in 100 mL of fresh reaction mixture at 5°C, and the reaction repeated under conditions identical to those described above. After 25 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 59%, 1.1%, and 1.6%, respectively, with 43% recovery of glycolic acid. An assay of glycolate oxidase in the cells at this time indicated no remaining activity.

EXAMPLE 7 A 300-mL EZE-Seal stirred autoclave reactor (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), amino- methylphosphonic acid (0.375 M), flavin mononucleotide (0.01 mM), and no added HPLC internal standard (pH 8.3, adjusted with 50% NaOH), and the solution cooled to 5°C. Frozen (-80°C) Aspergillus nidulans FT17SYCSL/OL (26 g, 124 IU glycolate oxidase and 57,800 IU catalase) were allowed to thaw at 5°C, then washed with 4 x 100 mL of KH₂PO₄ (50 mM, pH 7.0)/FMN (0.01 mM) buffer at 5°C and the washed cells added to the reactor with 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). The resulting mixture was stirred at 400 rpm and 5°C under 120 psig of oxygen while bubbling oxygen through the mixture at 50 mL/min. After 11 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 89%, 4.5%, and 2.0%, respectively, with complete conversion of glycolic acid.

EXAMPLE 8 Into a 3 oz. Fischer-Porter glass aerosol reaction vessel was placed a magnetic stirring bar and 10 mL of an aqueous solution containing glycolic acid (0.500 M), aminomethylphosphonic acid (0.375 M), isobutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3

(adjusted with 50% NaOH), and the solution cooled to 5°C. To the vessel was then added 0.47 g of Hansenula polymorpha transformant GO1 (10 IU glycolate oxidase and 22,100 IU catalase) which had been permeabilized by treatment with 0.1% "TRITON" X-100/1 freeze-thaw, and 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). The pH of the resulting mixture was re-adjusted to 8.3 with 50% NaOH, then the reaction vessel was sealed and the reaction mixture was cooled to 5°C. The vessel was flushed with oxygen by pressurizing to 70 psig and venting to atmospheric pressure five times with stirring, then the vessel was pressurized to 70 psig of oxygen and the mixture stirred at 5°C. Aliquots (0.10 mL) were removed by syringe through a sampling port (without loss of pressure in the vessel) at regular intervals for analysis by HPLC to monitor the progress of the reaction. After 16 hours, the HPLC yields of glyoxylate, formate, and oxalate were 90.1%, 1.3%, and 5.9%, respectively, and 3.0% glycolate remained. The remaining glycolate oxidase activity and total catalase activity were 86% and 136%, respectively, of their initial values.

EXAMPLE 9 (COMPARATIVE) The reaction in Example 8 was repeated, except that the addition of 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma) was omitted. After 16 hours, the HPLC yields of glyoxylate, formate, and oxalate were 57.6%, 32.5%, and 2.6%, respectively, and 8.9% glycolate remained. The remaining permeabilized cell activity of glycolate oxidase and catalase were 60% and 378%, respectively, of their initial values.

EXAMPLE 10 (COMPARATIVE) The reaction in Example 8 was repeated, except that 1.67 g of permeabilized (0.1% "TRITON" X-100/1 freeze-thaw) Sacchraomyces cerevisiae cells having 1.4 x 10⁶ units of catalase activity were substituted for the addition of 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). After 16 h, the HPLC yields of glyoxylate, formate, and oxalate were 67.2%, 19.7%, and 6.0%, respectively, and no glycolate remained. EXAMPLE 11

Into a 3 oz Fisher-Porter glass aerosol rection vessel was placed a magnetic stirring bar and 10 mL of an aqueous solution containing glycolic acid (0.500 M), aminomethylphosphonic acid (0.375 M), iosbutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3 (adjusted with 50% NaOH), and the solution cooled to 5°C. To the vessel was then added 0.75 g of a Pichia pastoris transformant GS115-MSP10 (13.2 IU glycolate oxidase and 21,200 IU catalase) which had been permeabilized by treatment with 0.1% "TRITON" X-100/1 freeze thaw. The pH of the resulting mixture was re-adjusted to 8.3 with 50% NaOH, then the reaction vessel was sealed and the reaction mixture was cooled to 5°C. The vessel was flushed with oxygen by pressurizing to 70 psig and venting to atmospheric pressure five times with stirring, then the vessel was pressurized to 70 psig of oxygen and the mixture stirred at 5°C. Aliquots (0.10 mL) were removed by syringe through a sampling port (without loss of pressure in the vessel) at regular intervals for analysis by HPLC to monitor the progress of the reaction. After 16 hours, the HPLC yields of glyoxylate, formate, and oxalate were 30.5%, 59.2%, and 10.7%, respectively, and 0.8% glycolate remained.

EXAMPLE 12 A 300 mL EZE-Seal stirred autoclave reactor equipped with Dispersimax Impeller (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.500 M), aminomethylphosphonic acid (0.375 M), isobutyric acid (0.100 M, HPLC internal standard), and flavin mononucleotide (0.01 m) at pH 8.3 (adjusted with 50% NaOH), and the solution cooled to 5°C. To the reactor was then added 10 g of Pichia pastoris transformant NRRL Y-21001 (391 IU glycolate oxidase and 457,000 IU catalase) which had been permeabilized by treatment with 0.1% "TRITON" X-100/1 freeze-thaw, and 1.4 x 10⁶ units of Aspergillus niger soluble catalase (Sigma). The pH of the resulting mixture re-adjusted to 8.3 with 50% NaOH. This mixture was stirred at 1000 rpm, which bubbled oxygen through the mixture via the action of the turbine impeller, and at 5°C under 120 psig of oxygen. The reaction was monitored by taking a 0.40 mL aliquot of the reaction mixture at regular intervals, filtering the aliquot using a Millipore "ULTRAFREE"-MC 10,000 NMWL Filter Unit, and analyzing the filtrate by HPLC. After 1.0 hour, the yields of glyoxylic acid, oxalic acid, and formic acid were 89.8%, 6.0%, and 2.9%, respectively, with 1.7% recovery of glycolic acid. The final activities of permeabilized-cell glycolate oxidase and catalase were 117% and 78% of their initial values.

The microbial cell catalyst was recovered from the reaction mixture described above by centrifugation. Without further treatment the cell pellet was mixed with 100 mL of fresh reaction mixture and an additional 1.4 x 10⁶ units of Apsergillus niger soluble catalase, and the reaction repeated. After 1.0 hour, the yields of glyoxylic acid, oxalic acid, and formic acid were 87.8%, 5.0%, and 5.3%, respectively, with 2.9% recovery of glycolic acid. The final activities of permeabilized-cell glycolate oxidase and catalase were 172% and 61% of their initial values. EXAMPLE 13

The mixture of glyoxylic acid and AMPA produced via the reaction described in Example 2 was centrifuged to remove the Aspergillus nidulans whole cell catalyst, then filtered using an Amicon "CENTRJPR ?" 10 concentrator (10,000 molecular weight cut-off) to remove the soluble Aspergillus niger catalase. The resulting solution was stirred with activated carbon (0.50 g) to remove FMN, filtered, and placed in the 300 mL, 316 SS cup of an Autoclave Engineers EZE-Seal reactor, along with 0.50 g of 10% Pd on activated carbon. The reactor was flushed with nitrogen, then charged with hydrogen at 300 psig and stirred at 1000 rpm and 27°C. After 22 hours, the yield of N-(phosphonomethyl)- glycine (based on AMPA) was 80%.

EXAMPLE 14 The mixture of glyoxylic acid and AMPA produced via the reaction described in Example 4 was centrifuged to remove the Aspergillus nidulans whole cell catalyst, then filtered using an Amicon "CENTRIPREP" 10 concentrator (10,000 molecular weight cut-off) to remove the soluble Aspergillus niger catalase. The resulting solution was stirred with activated carbon (0.50 g) to remove FMN, filtered, and placed in the 300 mL, 316 SS cup of an Autoclave Engineers EZE-Seal reactor, along with 0.50 g of 10% Pd on activated carbon. The reactor was flushed with nitrogen, then charged with hydrogen at 300 psig and stirred at 1000 rpm and 27°C. After 22 hours, the yield of N-(phosphonomethyl)- glycine (based on AMPA) was 89%.

EXAMPLE 15 The mixture of glyoxylic acid and AMPA produced via the reaction described in Example 6 was centrifuged to remove the Aspergillus nidulans whole cell catalyst, then filtered using an Amicon "CENTRIPREP" 10 concentrator (10,000 molecular weight cut-off) to remove the soluble Aspergillus niger catalase. The resulting solution was stirred with activated carbon (0.50 g) to remove FMN, filtered, and placed in a 300 mL glass liner, along with 0.50 g of 10% Pd on activated carbon. The glass liner was sealed in an autoclave and flushed with nitrogen, then charged with hydrogen at 1000 psig and mixed by rocking at 27°C. After 11 hours, the yield of N-(phosphonomethyl)glycine (based on AMPA) was 76%.

EXAMPLE 16 A 300 ML EZE-Seal stirred autoclave reactor equipped with Dispersimax

Impeller (Autoclave Engineers) was charged with 100 mL of a solution containing glycolic acid (0.50 M), aminomethylphosphonic acid (0.375 M), isobutyric acid (0.100 M), HPLC internal standard), and flavin mononucleotide (0.01 mM) at pH 8.3, and the solution cooled to 5°C. To the reactor was then added 30 g of E. coli transformant (25.2 IU glycolate oxidase and 39,900 IU catalase), and the mixture stirred at 1000 rpm, which bubbled oxygen through the mixture via the action of the turbine impeller, and at 5°C under 120 psig of oxygen. The reaction was monitored by taking a 0.40 mL aliquot of the reaction mixture at regular intervals, filtering the aliquot using a Millipore Ultrafree-MC 10,000 NMWL Filter Unit, and analyzing the filtrate by HPLC. After 19 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 8.6%, 8.3%, and 1.1%, respectively, with 57.3% glycolic acid remaining. The recovered activities of microbial glycolate oxidase and catalase were 10% and 68% of their intial values, respectively.

EXAMPLE 17 The reaction described in Example 16 was repeated using a second growth of 30 g of E. coli transformant having a higher glycolate oxidase activity (72 IU glycolate oxidase and 29,600 IU catalase). After 23 hours, the yields of glyoxylic acid, oxalic acid, and formic acid were 18.3%, 1.9%, and 2.9%, respectively, with 42.6% glycolic acid remaining. The recovered activities of microbial glycolate oxidase and catalase were 18% and 71% of their initial values, respectively. Having thus described and exemplified the invention with a certain degree of particularity, it should be appreciated that the following claims are not to be so limited but are to be afforded a scope commensurate with the wording of each element of the claim and equivalents thereof.

Claims

WHAT IS CLAIMED IS:

1. In a process for preparing a mixture of glyoxylic acid and amino¬ methylphosphonic acid comprising the step of oxidizing glycolic acid with oxygen in an aqueous solution, in the presence of aminomethylphosphonic acid and the enzymes glycolate oxidase and catalase wherein the improvement comprises using a microbial cell catalyst that expresses glycolate oxidase.

2. A process of Claim 1 wherein said microbial cell catalyst also expresses endogenous catalase.

3. A process of Claim 2 wherein soluble catalase is also used.

4. A process of Claim 1 wherein said microbial cell catalyst is selected from the group consisting of Aspergillus, Hansenula, and Pichia.

5. A process of Claim 1 wherein said microbial cell catalyst is Pichia pastoris.

6. A process of Claim 1 wherein said microbial cell catalyst is Hansenula polymorpha.

7. A process of Claim 1 wherein said microbial cell catalyst is Aspergillus nidulans.

8. A process of Claim 1 wherein said microbial cell catalyst is Escherichia coli.