Degradation of Glyphosate by Pseudomonas sp. PG2982 via a ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 262, No.25, Issue of September 5, pp. 12164-12168,1987 Printed in U.S.A.

Degradation of Glyphosate by Pseudomonas sp. PG2982 via a Sarcosine Intermediate* (Received for publication, February 25, 1987)

Ganesh M. KishoreS and GaryS.Jacob4 From the Monsanto Company, Biological Sciences, AA31, St. Louis, Missouri 63198

Thebacterium Pseudomonas PG2982 metabolizes MATERIALS AND METHODS~ glyphosate (N-(phosphonomethy1)glycine)by convertRESULTS ing it to glycine, a one-carbon unit, and phosphate. Here we show that this conversion involves the interWhen Pseudomonas PG2982 was harvested during mid-log mediate formation of sarcosine. When cells are incu- phase and incubated with [3-’4C]glyphosate,3it metabolized bated with [“C]glyphosate, the I4C can be entrapped the radioactive substrate as evidenced by the loss of radioacin glycine or sarcosine. With addedsarcosine, 14C from tivity from the total suspension (Fig. 2, Miniprint). Since all three carbons of glyphosate is recovered solely in radioactivity was trapped into 0.1 N NaOH and was released sarcosine. In experiments with glycine, radioactivity by acidification, it was probably carbon dioxide. The radiofrom the carboxymethyl moiety of glyphosateis trap- activity remaining in the suspension after 60 min was solely ped in glycine as well as serine, whereas radioactivity from the products of [14C]glyphosatemetabolism. Conversion from the phosphonomethyl carbon is only incorporated of [14C]glyphosate to carbon dioxide also occurred in the into serine. These resultsare consistent witha pathway presence of chloramphenicol (datanot shown), suggesting involving the conversionof glyphosate to sarcosine by that new protein synthesis was not required for this reaction. cleavage of its carbon-phosphorus(C-P)bond, fol- The extentof metabolism of [ 14C]glyphosateto carbon dioxide lowed by theoxidationofsarcosine to glycine and was higher for phosphonomethyl carbon than the glycyl carformaldehyde. bons (Fig. 2). Within the glycyl moiety, a greater loss occurred for the carboxyl carbon with very little loss of methylene carbon. This is consistent with the glycine cleavage reaction in Pseudomonas PG2982 (9). Alternatively, conversion of glycine to serine and hence pyruvate followed by its oxidative Glyphosate (N-(phosphonomethy1)glycine) is the active in- metabolism would also account for a greater loss of the gredient of the commercial herbicide ROUNDUP.’ Glyphos- carboxyl carbon (9). Glyphosate metabolism in Pseudomonas PG2982 was folate is rapidly metabolized by soil bacteria to aminomethyl lowedby determining the amounts of radioactivity in the phosphonate (1-3), which does not have significant herbicidal medium and the total suspension. As shown in Fig. 3 (Miniproperties. Very little is known concerning the metabolism of print), glyphosate was rapidly taken up into cells and metabglyphosate in isolated cultures. Moore et al. (4) reported the olized but only when gluconate, the carbon source, was also isolation of Pseudomonas PG2982 which could use glyphosate present in the incubation medium. No metabolism of glyas a sole phosphorus source. Metabolic studies using solid- phosate could be demonstrated in cells grown with inorganic state NMR techniques revealed that glyphosate was converted phosphate (1 mM) alone or with inorganic phosphate plus to glycine, a one-carbon unit, and phosphate ( 5 ) . This result glyphosate (1 mM each; Fig. 3). The presence of inorganic was unexpected in view of the earlier reports that mixed soil phosphate (1 mM) in the incubation mixture also inhibited cultures degrade glyphosate to aminomethyl phosphonate (1- its metabolism, and the I4C wascompletely recovered as [“C] glyphosate. It can be concluded the metabolism of glyphosate 3). The metabolism of glyphosate was further investigated requires an energy source (gluconate) andis inhibited by here. Three mechanisms were envisaged for the conversion of inorganic phosphate. Bacterial cells actively metabolizing [‘4C]glyphosatewere glyphosate to glycine (6). In this report, we demonstrate that the metabolism of glyphosate involves the formation of sar- analyzed for radioactive metabolites by extracting cells with cosine which couldoccur by the direct cleavage of the carbon- 10% trichloroacetic acid. A significant portion ( S O % ) of the phosphorus bond of glyphosate (Fig. 1).Other products appear * Portions of this paper (including “Materials and Methods,” Tato be similarly made by the cleavage of the C-P bond during bles I and 11, and Figs. 2-5) are presented in miniprint at the end of the metabolism of phosphonates by bacteria. this paper. Miniprint is easily read with the aid of a standard * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed: Monsanto Co., Biological Sciences, AA31, 700 Chesterfield Village Pkwy., St. Louis, MO 63198. 8 Present address: Dept. of Biochemistry, University of Oxford, Oxford, OX1 3QU, United Kingdom. ROUNDUP isa Registered Trade Mark of the Monsanto Co.

magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 86M-4163, cite the authors, and include a check or money orderfor $4.80 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. For convenience, the carboxy, carboxymethyl, and phosphonomethyl carbons of glyphosate arereferred to as 1-, 2-, and 3-C glyphosate, respectively. Other abbreviations used are:Kcpm, lo3 cpm; MOPS, 4-morpholinepropanesulfonicacid; HPLC, high performance liquid chromatography.

12164

12165

Phosphonate Metabolism in Pseudomonas PG2982

FIG. 1. Mechanism for the conversion of glyphosate to glycine and a one-carbon unit. P corresponds to

OH

// J

the phosphorus moiety released from glyphosate during this conversion.

/ Purmes

I

=

CH2 OH

I A CHNH, I

-

Proteins

0 COOH

Nuclelcaclds

A CH,-

radioactivity was associated with proteins and nucleic acids with a trace amount of radioactivity in the trichloroacetic acid-soluble fraction. In the presence of chloramphenicol, nearly 30% of the radioactivity was recovered in serine and glycine. The remaining radioactivity was incorporated into unidentified compounds (datanotshown). We therefore tested the ability of the hypothetical intermediates of glyphosate metabolism (6) to function as trap reagents. In the presence of either sarcosine or glycine, a significant portion of radioactivity from ['4C]glyphosate was no longer metabolized to carbon dioxide and instead the radioactivity was recoverable from the medium (Fig. 4,Miniprint). These amino acids, therefore, interfere with the metabolism of the sarcosyl moiety of glyphosate (Fig. 4). Since glyphosate is metabolized to glycine ( 5 ) , the data with glycinecould be explained as either due to inhibition of metabolism of ['"C] glyphosate or dilution of the ['4C]glycine derived from glyphosate by the excess unlabeled glycine in the medium. The data with sarcosine could be rationalized in three ways: (i) sarcosine may be an intermediate in the conversion of glyphosate to glycine, (ii) sarcosine may be converted to glycine, and (iii) sarcosine may inhibit the catabolism of glyphosate. Portions of the incubation mixtures from the experiments in Fig. 4 were subjected to thin layer chromatography (TLC) on cellulose plates, and thedistribution of radioactivity in the individual components was determined using a radioactive scanner. In the presence of sarcosine, radioactivity from all three carbons of glyphosate was incorporated into a single compound which had the same RFas sarcosine (Fig. 5, Miniprint). Very little radioactivity was associated with glyphosate after 1 h of incubation. The kinetics of utilization of glyphosate in the presence of sarcosine was similar to that with glyphosate alone, ruling out the possibility of inhibition by sarcosine of metabolism of glyphosate by the bacterium. However, in the presence of glycine, asubstantialamount of unmetabolized ['"Clglyphosate was recovered, indicating that glycine inhibited the uptake ormetabolism of glyphosate (Fig. 5). Some incorporation of14C from [l-"Cc]- and [2-'4C]glyphosate to both glycine and serine was also detected under these conditions. As expected, with [3-14C]glyphosate,only serine was labeled and no radioactivity was detected in the region corresponding to glycine (Fig. 5 ) . These results indicate

0

CO- COOH

-

Genera' Metabolism

that the first product of glyphosate metabolism in Pseudomonas PG2982 is sarcosine. Consistent with this notion, an extract of Pseudomonas PG2982 grown on glyphosate as a phosphorus source had a &fold higher level of sarcosine dehydrogenase activity (8 nmol of sarcosine oxidized/min/mg of protein) than phosphate-grown cells. Samples of reaction mixtures of Fig. 4 were also resolved on an amino acid analyzer. The radioactivity was primarily associated with the sarcosine peak when sarcosine was used as thetrapping agent (TableI, Miniprint). In the presence of glycine, radioactivity wasrecovered in glyphosate, glycine, and serine when [l-'4C]glyphosate was used as a substrate (Table I).Identical results were obtained with glycine and [214C]glyphosate(data not shown). With [3-14C]glyphosateand glycine, radioactivity was only detected in glyphosate and serine. These data aresummarized in Table I and confirm the results of the TLC analyses. Similar results were obtained when the experiments were performed in the presence of the proteinsynthesisinhibitor chloramphenicol. These results indicate that sarcosine is acting primarily as a trap reagent. Additional evidence for the conversion of glyphosate to sarcosine was obtained using sarcosine dehydrogenase-inactivated bacterial cells. Repeated freezing and thawing of the bacterial suspension resulted in a complete loss of sarcosinemetabolizing activity of Pseudomonas PG2982. This suspension could convert ['4C]glyphosate to ['4C]sarcosine, which accumulated within the bacterial cells (data not shown). Attempts were also made to demonstrate the conversion of glyphosate to sarcosine using extracts of Pseudomonas PG2982. However, under a variety of conditions, no cell-free activity could be obtained. Experiments were therefore performed with whole cells to characterize partially the mechanistic features of this enzymatic reaction. Fate of 3H from N-(Ph~sphono[methyl-~H]glycim-Ifthe conversion of glyphosate to sarcosine by Pseudomonas PG2982 requires functionalization (ie. cleavage of the C-H bond) of the phosphonate carbon, then radioactivity from [3H]glyphosateshould be incorporated into the solvent. [3H] Glyphosate was therefore incubated with Pseudomonas PG2982 in the presence and absence of sarcosine. In the absence of sarcosine, a substantial incorporation of 3H from [3H]glyphosateinto water occurred (97%); in the presence of

12166

Phosphonate Metabolism in Pseudomonas PG2982

sarcosine, very little incorporation was detected (97%) was completely recovered in accord with the intermediate formation of sarcosine. Since from sarcosine. This finding indicates that the phosphonate the phosphonomethyl carbon is lost as a one-carbon unit at carbon of glyphosate is not functionalized during cleavage of the formaldehyde level of oxidation during theconversion of the C-P bond by Pseudomonas PG2982. sarcosine toglycine (13), no radioactivity from this carbon of Products of Metabolism of Other Phosphonates-Pseudomglyphosate is incorporated intoglycine. However, in the presonas PG2982 metabolizes a number of phosphonates besides ence of glycine, radioactivity from the phosphonomethyl carglyphosate (Table 11, Miniprint). The productsof metabolism bon of glyphosate is incorporated into serine. This is due to of these phosphonates have been determined and the results formation of serine from glycine and [14C]formaldehyde via are summarized in Table11. In each case, the organic moiety the serine trans-hydroxymethylase reaction (14). The formof the phosphonate is released as a product in which the C- aldehyde pool is expected to be labeled as a result of the H bond replaces the C-P bond. Thus, alkyl, aryl, andalkenyl sarcosine dehydrogenase reaction. In the light of these results, phosphonates were metabolized by Pseudomonas PG2982 itappearsthat mode of Fig. 1 accuratelyrepresentsthe leading totheformation of alkanes,arenes,andalkenes, conversion of glyphosate to glycine by cells of Pseudomonas respectively. In the presence of chloramphenicol, bacterial PG2982. cells grown on anyof these phosphonateswere able to metabConfirmation of sarcosine as the intermediate in the conolize glyphosate and likewise, glyphosate-grown cells rapidly version of glyphosate to glycine by Pseudomonas PG2982 metabolized other phosphonates (data not shown). illustrates that there are two pathways for the bacterial meMetabolism of Phosphonates by Escherichia coli B (ATCC tabolism of this herbicide. Glyphosate is converted to ami11303)"E. coli has been reported to utilize a variety of phos- nomethyl phosphonate ina Flauobacterium sp. (15) as well as phonates as a source of phosphorus in the growth medium mixed soil cultures of bacteria (1-3). However, metabolism of (10, 11). We find that E. coli B converts a number of alkyl, glyphosate in Pseudomonas PG2982 does not involve the aryl, andalkenyl phosphonates to the same products as those intermediate formation of aminomethyl phosphonate. In the reported here for Pseudomonas PG2982. Although E. coli Flauobacterium sps. (15) conversion of glyphosate to aminometabolizes a diverse range of phosphonates, it is unable to methyl phosphonateoccurs in the presence of inorganic phosutilize glyphosate as a phosphorus source. Furthermore, gly- phate, whereas in Pseudomonas PG2982 glyphosate metabophosate inhibits the growthof E. coli (12). Under conditions lism to sarcosine is inhibited by inorganic phosphate. These in which this inhibition is reversed by addition of aromatic results suggest that the aminomethyl phosphonate pathway amino acids to the growth medium, glyphosate still cannot may have evolved under conditions inwhich glyphosate must serve as a source of phosphorus for this bacterium. Interest- be metabolized in the presence of inorganic phosphate. An ingly, aminomethyl phosphonate, an analog of glyphosate, is Arthrobacter sp. GLP-1 also appears tometabolize glyphosate metabolized by E. coli. The product of thisreaction was via the sarcosine pathway(6), and inorganic phosphate inhibidentified as methylamine, which is also produced in Pseu- its glyphosate metabolism by this bacterium aswell. domonas PG2982 grown onaminomethylphosphonate as Of central importance is the natureof the enzyme catalyzphosphorus source. ing the cleavage of the C-P bond of phosphonates. The C-P E. coli was grown anaerobically in the presenceof nitrate, bond is extremely stable and resistant tohydrolysis by both as a terminal electron acceptor, to determine whether molec- acidsandalkali (11). Frostandhis co-workers (11) have ular oxygen was required for the cleavage of the C-P bond. recently reported a model system for the cleavage of organic As shown in Table11, methane production from methyl phos- phosphonates using lead tetraacetate. Despite extensive efphonate still occurred under anaerobic conditions, indicatingforts, we were unable to demonstrate theC-P bond cleavage that molecular oxygen was not involved in the conversion of reaction in a cell-free system. This enzyme must differ signifmethyl phosphonate to methane. Similar results were also icantly from phosphonatases which are unable to utilize sevobtained with other phosphonates tested with E. coli. Thus, eral phosphonates used here as substrates (16). Martell and the C-P bond cleavage in E. coli is a nonoxygenative reaction. Langlohr (17) have reported that pyridoxal phosphate catalyzes the dephosphonylation of phosphonoalanine. However, DISCUSSION this mechanism requires formation of the Schiff base between Our previous studies showed that the metabolism of gly- the aminoacid and pyridoxal phosphate, which is notpossible phosate involved the intermediate formation of glycine (5). with several phosphates used here. Several mechanismswere envisaged forthis conversion. HowT h e C-P bond cleavage enzyme in E. coli catalyzes an ever, a distinction between them could not be made by a n oxygen-independent cleavage reactionin which the C-H analysis of the intracellular metabolites of glyphosate because bonds of the phosphonate carbon are conserved. This elimiof the extremely rapid rates of their metabolism. Experiments nates the possibility of both oxygenative as well as functionwere therefore performed in the presenceof various interme- alized phosphonate-carbon for the C-P bond cleavage. In diates as potential trap reagents. Radioactivity from all three order to determine whether a reductive mechanism is opercarbons of glyphosate could be trapped into sarcosine. The ating in this reaction, it is of importance to characterize the identity of the radiolabeled product as sarcosine was con- phosphorus product released during the C-P bond cleavage. firmed by both thin layer chromatography and amino acid This should facilitate our understanding of the mechanismof analysis. In order to eliminate thepossible induction of new the C-P bond cleavage reaction in the absence of an i n vitro activities by sarcosine, resulting in a conversion of the ["C] enzyme activity. glycine derived from glyphosate t o sarcosine, trapping experVery recently, Wackett et al. (18) have reported thatAgroiments were also performed in the presenceof chloramphenbacterium radiobacter also metabolizes phosphonates by a icol. Identical resultswere obtained. In addition, the sarcosine pathwaysimilar tothat reported herefor Pseudomonas dehydrogenase activity of glyphosate-grown Pseudomonas PG2982 was inactivated without effect on the conversion of PG2982 and E. coli B. Sincephosphonate metabolism is tightly regulated by inorganic phosphate (19), these investiglyphosate to ~arcosine.~ gators reasoned that one or more of the phosphate-starvationinducible (Psi) proteinsmay actually represent theenzyme(s1 G . M. Kishore and G . S. Jacob, unpublished results.

Phosphonate Metabolism involved in phosphonate metabolism. Several E. coli K12 transposon Mudl-psi mutants were screened for their ability to grow on methyl phosphonate as phosphorus source. Mutants unable to adapt to growth on methyl phosphonate were found to have the Mud1 insertion in a locus designated Psi D (19). It appears, therefore, that Psi D may encode the er?zyme involved in C-P bond cleavage. Alternatively, it could encode a protein involved in either phosphonate transport or synthesis of a cofactor involved in C-P bond cleavage. It is clear that further progress inthis area requires detailed analysis of the gene productfs) of Psi D locus. REFERENCES 1. Rueppel, M. L., Brightwell, B. B., Schaefer, J., and Marvel, J. T.

(1977) J. Agric. Food Chem. 25, 517-528 2. Sprankle, P., Meggit, W. F., and Penner, D. (1975) Weed Sci. 2 3 , 229-234 3. Nomura, N. S., and Hilton, H.W. (1977) Weed Res. 17,113-121 4. Moore, J. K., Braymer, H. D.,and Larson, A. D. (1983) Appl. Enuiron. Microbiol. 46, 316-320 5. Jacob, G. S., Schaefer, J., Stejskal, E. O., and McKay, R. A.

i n Pseudomonas PG2982

12167

(1985) J. Biol. Chem. 260, 5899-5905 6. Pipke, R., Amrhein, N., Jacob, G. S., Schaefer, J., and Kishore, G. M.(1987) Eur. J. Biochem. 165.267-273 7. Niedhardt, F. C., Bloch, P. L., and Smith, D.F. (1974) J. Bacteriol. 119,736-747 8. Armstrong, J. McD. (1964) Biochim. Biophys. Acta 8 6 , 194-197 9. Jacob, G. S., Garbow, J. R., Schaefer, J., and Kishore, G. M. (1987) J. Biol. Chem. 262, 1552-1557 10. Harkness, D.R. (1966) J. Bacteriol. 9 2 , 623-627 11. Cordeiro, M. L.,Pompliano, D. L., and Frost, J. W. (1985) J.Am. C k m . SOC.108,332-334 12. Gresshoff, P. M. (1979) A u t . J. Plant Physiol. 6, 177-185 13. Frisell, W. R. (1971) Arch. Biochem. Biophys. 1 4 2 , 213-222 14. Schirch, L. (1982) Adu. Enzymol. Relat. Areas Mol. Biol. 53, 83 15. Balthazor, T. M., and Hallas, L. E. (1986) Appl. Enuiron. Microbiol. 5 1,432-434 16. La Nauze, J. M., Rosenberg, H., and Shaw, D. C. (1970) Biochim. Biophys. Acta 212, 332-350 17. Martell, A. E., and Langlohr, M. F. (1977) J. Chem. SOC.Chem. Commun. 10,342-345 18. Wackett, L. P., Shames, S. L., Vendetti, C. P., and Walsh, C. T. (1987) J. Bacteriol. 169, 710-717 19. Wackett, L. P., Wanner, B. L., Vendetti, C. P., and Walsh, C. T. (1987) J. Bucteriol. 1 6 9 , 1753-1756

Phosphonate Metabolism in PseudomomPG2982

12168 60

"Y

I

\ \ \

-.. .... . -0

0

10

I

10

20

30

40

Time (mind

50

,

60

" I

0

10

20

30

40

Time (mind

50

60