Utilization of glyphosate as phosphate source: biochemistry and genetics of bacterial carbon-phosphorus lyase

Abstract

After several decades of use of glyphosate, the active ingredient in weed killers such as Roundup, in fields, forests, and gardens, the biochemical pathway of transformation of glyphosate phosphorus to a useful phosphorus source for microorganisms has been disclosed. Glyphosate is a member of a large group of chemicals, phosphonic acids or phosphonates, which are characterized by a carbon-phosphorus bond. This is in contrast to the general phosphorus compounds utilized and metabolized by microorganisms. Here phosphorus is found as phosphoric acid or phosphate ion, phosphoric acid esters, or phosphoric acid anhydrides. The latter compounds contain phosphorus that is bound only to oxygen. Hydrolytic, oxidative, and radical-based mechanisms for carbon-phosphorus bond cleavage have been described. This review deals with the radical-based mechanism employed by the carbon-phosphorus lyase of the carbon-phosphorus lyase pathway, which involves reactions for activation of phosphonate, carbon-phosphorus bond cleavage, and further chemical transformation before a useful phosphate ion is generated in a series of seven or eight enzyme-catalyzed reactions. The phn genes, encoding the enzymes for this pathway, are widespread among bacterial species. The processes are described with emphasis on glyphosate as a substrate. Additionally, the catabolism of glyphosate is intimately connected with that of aminomethylphosphonate, which is also treated in this review. Results of physiological and genetic analyses are combined with those of bioinformatics analyses.

FIG 1

Pathways for biological extraction of glyphosate-phosphorus. Glyphosate may be cleaved through the phn-specified C-P lyase pathway with the formation of PRPP and N-methylglycine or by the oxidation of C₂ by glyphosate oxidase with the formation of AMPA and glyoxylate. AMPA-phosphorus in turn may be converted to PRPP through the phn-specified C-P lyase pathway, which also results in the formation of N-methylacetamide (see the text for details).

FIG 2

Organization of C-P lyase pathway-encoding operons in various microbial species and strains. Each open reading frame is shown as a rectangle, with the phn gene designation shown above in italics. Displacement of an open reading frame relative to the previous open reading frame indicates overlapping reading frames. Genes for transport proteins are labeled in red. A nomenclature has been adopted so that genes encoding periplasmic binding proteins are labeled D, and genes encoding a nucleotide-binding protein are labeled C, whereas genes encoding a membrane-spanning protein are labeled E. In some cases, this nomenclature may deviate from the originally published gene designations. Genes for products with known enzymatic activity in phosphonate catabolism are labeled in blue, and genes specifying putative regulatory proteins are labeled in green, whereas genes specifying auxiliary polypeptides involved in phosphonate catabolism are labeled in black. Genes apparently not involved in phosphonate catabolism are unfilled and unlabeled. In some cases, more than one cistron appears to encode homologous proteins, e.g., phnM of O. anthropi. 1045 indicates the domain of unknown function 1045 gene (encoding DUF1045). It is included as phosphonate degradation relevant, as it is usually located in proximity to phn genes (129). Similarly, 1868 indicates the gene encoding DUF1868, whereas 2HP designates a gene that may specify a member of the two-histidine phosphodiesterase superfamily (see the text for details). The designations of the P. stutzeri htx operon cistrons have been altered (except for htxA) to underline the similarity with other phosphonate catabolism operons. The htx operon was originally designated htxABCDEFGHIJKLMN (30). Similarly, the S. meliloti and B. japonicum phnCDEE genes were originally designated phoCDET (119, 130, 131). Numbers expressed in kilobases indicate interoperon distances. Arrows beneath the genes indicate the direction of transcription and attempts to estimate the boundaries of individual transcripts. Contiguous arrows do not necessarily indicate operon expression, except for the E. coli phnC-P, A. ferrooxidans phnG-M, and P. stutzeri htxA-P genes, where operon expression has been experimentally confirmed (29, 30, 76). An asterisk under an arrow indicates that a Pho box has been identified in the DNA sequence and, thus, that transcription may be regulated by the P_i supply. Unless otherwise indicated, operons are located in the chromosomes. The GenBank accession numbers of the nucleotide sequences employed are U00096.2 for E. coli (132), AF061267 (htx) and AY505177 (phn) for P. stutzeri (117), CP000758.1 for O. anthropi ATCC 49188 (133), AL591985 (plasmid pSymB) and AL591688 (chromosome) for S. meliloti 1021 (57, 58), CP000628.1 for A. radiobacter K84 (63), BA000012.4 (chromosome) and BA000013.4 (plasmid pMLa) for M. loti MAFF303099 (32, 134), BA000040.2 for B. japonicum USDA 110 (130, 135), CP002833.1 (chromosome 1) for B. pseudomallei 1026b (136), BA000019.2 for Nostoc sp. PCC7120 (70), CP001219.1 for A. ferrooxidans ATCC 23270 (77), CP000393.1 for T. erythraeum IMS101, and CP000031.2 for R. pomeroyi DSS-3 (137).

FIG 3

Pathways for the conversion of phosphorus of glyphosate and AMPA to P_i. (A) Conversion of glyphosate. Compounds: 1, glyphosate; 2, 5′-triphospho-α-d-ribosyl 1′-(N-phosphonomethylglycine); 3, 5′-phospho-α-d-ribosyl 1′-(N-phosphonomethylglycine); 4a, N-methylglycine; 5, 5-phospho-α-d-ribosyl 1,2-cyclic phosphate; 6, α-d-ribosyl 1,5-bisphosphate; 7, 5-phospho-α-d-ribosyl 1-diphosphate (PRPP); 8, diphosphate ion (PP_i); 9, phosphate ion (P_i). Reactions are indicated by their enzymes as follows. PhnI* indicates purine ribonucleoside 5′-tri/diphosphate phosphonylase, an enzyme complex where PhnI plays a crucial catalytic role and which may also involve PhnG, PhnH, PhnJ, PhnK, and/or PhnL. PhnM is 5′-triphospho-α-d-ribosyl 1′-phosphonate diphosphohydrolase. PhnJ* indicates S-adenosylmethionine-dependent carbon-phosphorus lyase. PhnJ* may constitute a protein complex also containing PhnG, PhnH, PhnI, PhnK, and/or PhnL. PhnI* and PhnJ* may be the same protein complex. PhnP is the phnP-specified phosphoribosyl cyclic phosphodiesterase. PhnN is the phnN-specified ribosyl bisphosphate phosphokinase. APRTase is the apt-specified adenine phosphoribosyltransferase. Ppa is the ppa-specified inorganic diphosphate hydrolase (138). The enzymes catalyzing the latter two reactions are not specified by the phn operon. APRTase was arbitrarily chosen among the 10 phosphoribosyltransferases of E. coli (53). Any of these 10 enzymes may participate in the process. (B) Conversion of AMPA. Compounds: 10, AMPA; 11, N-acetamidomethylphosphonate; 4b, N-methylacetamide; compound 5 is described above for panel A. Compound 5 is further catabolized as described above for panel A. PhnO is aminoalkylphosphonate N-acetyltransferase, and PhnI*, PhnM, and PhnJ* are described above for panel A. The pathways were established on the basis of previously reported data (39–41, 52, 139, 140).

FIG 4

Metabolic pathways of ribosyl 1,5-bisphosphate of E. coli. α-d-Ribosyl 1,5-bisphosphate (Rib1,5bP) is shown at the center in red. The compound is generated by the activity of phosphoribosyl cyclic phosphodiesterase (PhnP) (42) and presumably by the activity of ribosyl 1-phosphate kinase (RPK) (52). Ribosyl 1,5-bisphosphate may be further converted to PRPP by ribosyl 1,5-bisphosphate phosphokinase (PhnN) and presumably to some unknown metabolite, metabolite X (52). The possibility of hydrolysis with the formation of P_i by a phn operon-specified gene product (PhnG, PhnH, PhnK, or PhnL or a combination thereof) is indicated by PhnG/H/K/L?. Presently, there is no evidence for this activity of any phn-specified polypeptide.

FIG 5

Putative substrates for C-P lyase of cells grown with various phosphonates. (A) Cells grown with glyphosate; (B) cells grown with AMPA; (C) cells grown with 2-aminoethylphosphonate; (D) cells grown with 3-aminopropylphosphonate; (E) cells grown with methylphosphonate; (F) cells grown with phenylphosphonate.

FIG 6

Hydrolytic and oxidative cleavage of C-P bonds. (A) Transformation of phosphonoalanine (compound 1) to P_i and pyruvate (compound 3) via phosphonopyruvate (compound 2). Enzymes: I, phosphonoalanine:α-ketoglutarate transaminase; II, phosphonopyruvate phosphohydrolase. (B) Transformation of phosphonoacetate (compound 4) to P_i and acetate (compound 5) catalyzed by phosphonoacetate phosphohydrolase (enzyme III). (C) Transformation of 2-aminoethylphosphonate (compound 6) to P_i and acetaldehyde (compound 8) via phosphonoacetaldehyde (compound 7). Enzymes: IV, 2-aminoethylphosphonate:pyruvate transaminase; V, phosphonoacetaldehyde phosphohydrolase; VI, phosphonoacetaldehyde dehydrogenase. (D) Oxidative transformation of 2-aminoethylphosphonate (compound 6) to P_i and glycine (compound 10) via 2-amino-1-hydroxyethylphosphonate (compound 9). Enzymes: VII, α-ketoglutarate-dependent 2-aminoethylphosphonate dioxygenase (PhnY); VIII, 2-amino-1-hydroxyethylphosphonate oxygenase (PhnZ).

FIG 7

Alignment of amino acid sequences of E. coli (Ec) and S. enterica (Se) aminoalkylphosphonate N-acetyltransferases and putative S. meliloti (Sm) and R. pomeroyi (Rp) aminoalkylphosphonate N-acetyltransferases. Asterisks above the sequences indicate identical amino acid residues in all four sequences, whereas asterisks below the sequences indicate identical amino acids in S. meliloti and R. pomeroyi aminoalkylphosphonate N-acetyltransferase sequences. Amino acid residues highlighted in white on a red background are believed to be involved in the binding of the substrate acetyl coenzyme A (98, 99).

FIG 8

Proposed phosphonate proteome and transport metabolon of E. coli. (A) The 14-cistron phn operon. The color code is the same as that described in the legend of Fig. 2. (B) Proposed location of phn-specified polypeptides. Individual polypeptides are identified by their cistron designation and a color code similar to that of the encoding cistrons in panel A; i.e., red spheres indicate the ABC transport system, blue spheres indicate polypeptides with assigned enzymatic function, the green sphere indicates the putative repressor of phn operon expression, and black spheres indicate auxiliary polypeptides without an assigned biochemical function.