Introduction

Neurotoxic organophosphates (OPs) were introduced as insecticides about 60 years ago to replace more persistent organochloride insecticides like DDT, HCH etc. Mainly due to their indiscriminate use in agriculture, the residues of OP compounds are found in various environments, including groundwater (Singh et al. 2014). Due to the evolution of novel degradative enzymes, certain soil bacteria use OP insecticide residues as sole source of carbon and phosphate. Initially these hydrolytic enzymes were designated as parathion hydrolases, methyl parathion hydrolases and paraoxonases etc. depending on the substrate (OP insecticide) used to assay the enzyme activity. Subsequently they were divided into three major subgroups namely organophosphate hydrolases (OPH), methyl parathion hydrolases (MPH) and organophosphate acid anhydrolases (OPAA). Based on the similarities in structure and catalytic mechanisms, the OP-degrading enzymes are assigned to one of these three subgroups. Among these OP-hydrolyzing enzymes, the physiological substrate is known only for OPAA. Dipeptides ending with a prolyl residue serve as substrates for OPAA. Therefore, the OPAA has been renamed as prolidase (Cheng et al. 1999). Since physiological substrate is known for OPAA, they are not taken as part of genes evolved for degradation of organophosphates.

Authentic physiological substrates are not known for either OPH or MPH. Despite structural differences, these two enzymes have identical active sites and follow similar catalytic mechanisms while hydrolyzing OPs (Afriat et al. 2006; Afriat-Jurnou et al. 2012; Parthasarathy et al. 2017c). They hydrolyze certain OP insecticides with rates close to their substrate limits (Chaudhry et al. 1988). Therefore, OPs are considered as cognate substrates for both MPH and OPH (Pandeeti et al. 2012; Purg et al. 2016; Parthasarathy et al. 2017a). It is also proposed that these two structurally different enzymes have converged functionally to provide selective advantage to the bacteria. They minimize toxic effects of OP residues and generate a phosphate pool to be used as a phosphate source (Dong et al. 2005; Afriat et al. 2006; Tawfik 2006; Afriat-Jurnou et al. 2012). This proposition gained strength due to the existence of identical mobile pte-elements in geographically and taxonomically well separated bacteria. However, this review critically examines this point of view in the light of the short half-life of OP residues and proposes a possible physiological role for pte genes by taking into account the existing knowledge on pte genes and sequence information from genome and meta-genome databases.

Studies on the biodegradation of organophosphates gained momentum with the isolation of parathion-degrading Flavobacterium sp. from soil samples collected by the International Rice Research Institute, Manila (Sethunathan and Yoshida 1973). Since then a number of reports appeared on biodegradation of OP compounds. Most of them, either described degradation pathways or the characterization of the corresponding enzymes. Only a few studies have focussed on the genetics of OP degradation in bacteria (Serdar et al. 1982; Mulbry et al. 1987; Zhongli et al. 2001; Horne et al. 2002a, b; Siddavattam et al. 2003; Yang et al. 2003; Zhang et al. 2006; Pandeeti et al. 2011, 2012; Parthasarathy et al. 2017b). Unlike their eukaryotic homologues discovered accidentally in patients with injured kidneys, the bacterial pte genes were identified and cloned while specifically trying to understand the genetic basis of organophosphate degradation (Ali et al. 2012; Mulbry et al. 1987; Harper et al. 1988; McDaniel and Wild 1988; Mulbry and Karns 1989; Somara and Siddavattam 1995; Zhongli et al. 2001; Horne et al. 2002a, b). Initial studies reported the existence of identical pte genes on indigenous plasmids isolated from soil bacteria strains. However, the survey of genome and meta-genome sequences has revealed the existence of pte homologues both on plasmids and chromosomes of archaea and bacteria, including human pathogen like Mycobacterium tuberculosis.

The opd plasmids

The opd genes are found both on plasmids and on chromosomes (Harper et al. 1988; McDaniel and Wild 1988; Horne et al. 2002b; Ali et al. 2012; Parthasarathy et al. 2017b) and detailed studies have been conducted on plasmid-borne opd sequences cloned from Flavobacterium sp. ATCC 27551 and B. diminuta. These two soil isolates, reclassified as Sphingobium fuliginis ATCC 27551 (Kawahara et al. 2010) and Sphyngophyxis wildii (Parthasarathy et al. 2017c), respectively, were isolated on different continents. Sphingobium fuliginis ATCC 27551 was isolated from rice fields of the International Rice Research Institute, Manila (Sethunathan and Yoshida 1973), whereas the Sphyngophyxis wildii species was isolated from sewage samples collected in California, USA (Munnecke and Hsieh 1974). In both cases, the opd sequences were localized on large indigenous plasmids designated as pPDL2 and pCMS1. The plasmid pPDL2 isolated from S. fuliginis ATCC 27551 is a 40-kb plasmid and the 65-kb pCMS1 is an indigenous plasmid of S. wildii. Except for the region (5.1 kb) containing opd gene, no obvious similarity was seen between these two plasmids (Mulbry et al. 1987; Pandeeti et al. 2011, 2012; Parthasarathy et al. 2017c). Sequence analyses of both plasmids has provided significant insights into the lateral transfer of opd genes among soil bacteria.

Multiple strategies to maintain plasmidome

Mobilizable plasmids do not always encounter favourable situations to replicate in their recipient cells. If they fail to replicate they can be lost, particularly if they do not integrate into the genome. Plasmid pPDL2 follows unique strategies to maintain its genome in recipient cells. A toxin–antitoxin module ensures survival of the cells that retain the plasmid. The replication and site-specific integration modules facilitate plasmid maintenance in the host either as an episome or as a plasmid. The mobilization module consisting of oriT and relA gene enables its horizontal mobility in the presence of a genetic repertoire coding for type 4 secretion system (T4SS). However, no T4SS-coding sequences are found on pPDL2. Consistent with the sequence information, the plasmid pPDL2 derivative, pPDL2-K has shown lateral mobility only in the presence of helper plasmid (Pandeeti et al. 2012).

To survive in recipient cells, the plasmid should have a broad host-range replicative origin (oriV). Analyses of the plasmid sequence deposited in GenBank (NC_019376.1) revealed the existence of a replicative origin typically seen in theta-replicating plasmids, with a well conserved DnaA binding box. A 776-bp long oriV sequence was identified immediately upstream of repA (17238 to 18114). The RepA protein of pPDL2 shows 99% amino acid sequence identity with the RepA of plasmid pUT1 isolated from Sphingobium japonicum UT26S. Likewise, considerable similarities were also found between the oriVs of pPDL2 and pUT1 (Pandeeti et al. 2012). The special arrangements of direct repeats of DnaA boxes resemble the structure of oriV sequences found in pUT1. In general, the replicative origins of plasmids show configuration conservation (spatial arrangement of repeats) rather than showing strict sequence conservation (Puyet et al. 1988). The oriV of plasmid pPDL2 has shown similarities in spatial arrangements of repeats and palindromic sequences with several plasmids isolated from Sphingobium and Pseudomonas strains (figure 1). The existence of such a conserved oriV facilitates its replication in a wide range of soil bacteria. In addition to the well conserved replication origin, a par locus and toxin–antitoxin modules exist among the 42 ORFs predicted in the sequence of pPDL2 to ensure its segregation and stability in recipient cells.

Fig. 1
figure 1

(a) Replicative origin (oriV) of plasmid pPDL2. Comparison of the particular arrangements of dnaA boxes and intron sequences in plasmids pUT1 and pPS10 isolated from S. japonicum and P. savastanoi is shown. (b) Secondary structure of predicted oriT sequence. The dotted box indicates the relaxase target sequence. The figure is reproduced from the article published in G3: Genes, Genomics and Genetics (Pandeeti et al. 2012).

Full size image
Fig. 2
figure 2

(a) Organization of plasmid pPDL2-borne xerD sequences in Sphingobium fuliginis ATCC 27551. Panel b-I shows the existence of the attB site at the \(3^\prime \) end of the tRNA coding gene. Comparison of attB and attP sequences is shown in panel b-II. The figure is reproduced from the article published in G3: Genes, Genomics and Genetics (Pandeeti et al. 2012).

Full size image
Fig. 3
figure 3

(a) Genetic organization of the plasmid pPDL2-borne opd element in Sphingobium fuliginis ATCC 27551. Panel b indicates the existence of an identical opd region found between plasmids pPDL2 and pCMS1. Creation of orf306 due to insertion of Tn3 in plasmid pPDL2 is shown with vertical lines. The figure is reproduced from the article published in G3: Genes, Genomics and Genetics (Pandeeti et al. 2012).

Full size image

Integration module

In the sequence of plasmid pPDL2, two ORFs, int1 (28584-29570c) and int2 (31182-32028) code for phage integrase (IntP) family proteins forming two integration modules CIP-I (24734-28063) and CIP-II (28913-30866) along with the other ORFs (figure 2). These two proteins show high sequence similarity to the site-specific recombinase XerD (integrase/recombinase) of Pseudomonas syringae pv. tomato str. DC3000. Phage integrases are recombinases that mediate unidirectional site-specific recombination between two recognition sequences namely attP (the phage attachment site) and attB (the bacterial attachment site). The phage integrases are tyrosine recombinases that are characterized by the presence of a signature sequence H-X-X-R in their C-terminal domain. The integrase of pPDL2 aligns well with the integrases present in other bacteria suggesting the possibility of site-specific integration of pPDL2 in the recipient chromosome in case if it fails to replicate in the recipient bacterium. The attP attachment site (figure 2b, I&II), shows sequence identity to the \(3^\prime \) end of the seryl-tRNA gene of S. japonicum, and is found on the sequence of pPDL2 (Pandeeti et al. 2012). The plasmid pPDL2 has also been shown to integrate site-specifically at an artificially generated attB site on a compatible plasmid (Pandeeti et al. 2012). Considering its ability to site-specifically integrate and its lateral mobility, the plasmid pPDL2 is designated as an integrative mobilizable element.

Fig. 4
figure 4

Organization of opdA element in Agrobacterium radiobacter P230: The opdA gene is shown in red colour arrow. The ORFs, orfA and orfB flanking the opdA gene are shown with blue and yellow colour arrows, respectively. The orfA codes for a truncated transposase and orfB codes for an ATP-binding protein. The transposase gene tnpA is coloured in pink. Small black arrows represent left (LIR) and right (RIR) inverted repeats.

Full size image

The Tnopd element

As stated in earlier sections, the DNA region containing the opd gene is highly conserved in plasmids pPDL2 and pCMS1. Sequence analyses of the conserved opd region has revealed a transposon-like organization on plasmid pPDL2 of Flavobacterium sp ATCC 27551 (Siddavattam et al. 2003). The presence of an insertion sequence (IS) element (IS21 class) and a transposon (Tn3) upstream and downstream of the opd gene, respectively, resembled that of a typical complex catabolic transposons reported in a number of bacterial species involved in the degradation of recalcitrant aromatic compounds (Tan 1999; Nojiri et al. 2004). However, in a standard transposition assay performed in E. coli, this conserved region alone failed to show transposition (Siddavattam et al. 2003). Moreover, this conserved region did not show the existence of inverted and direct repeat sequences typically seen in complex transposons (Siddavattam et al. 2003). Nevertheless, such sequences were identified in flanking mobile elements Tn3 and y4qE (figure 3a). Transposition was only observed when the plasmid pPDL2 derivative pPDL2-K was used as donor plasmid while performing the transposition assay. The left and right border sequences of the opd cluster at the target sequence revealed the repeat sequences indicated flanking the mobile elements Tn3 and y4qE, suggesting that the opd region found between these two mobile elements is an active transposon (figure 3a). In addition to the opd gene and orf306, the opd element contains genes that code for an outer membrane transporter (MFS permease), and protocatechuate dioxygenase, which is required for degradation of organophosphate insecticides and the degradation products like catechols and substituted catechols generated during biodegradation of OP insecticides (figure 3a). The mobilizable nature of plasmid pPDL2 and transposability of the opd gene cluster, if seen together with the existence of identical opd genes among taxonomically distant bacterial strains provides strong evidence to suggest the existence of horizontal mobility of opd genes among soil bacteria.

pCMS1 is a self-transmissible plasmid

The complete genome sequence for S. wildii is available. It contains a chromosome and a 65, 908-bp plasmid designated as pCMS1 (Parthasarathy et al. 2017c). The annotated sequence of plasmid pCMS1 is available in the genome database (NZ_CP026382.1). It is a self-transmissible plasmid and contains all necessary genes for coding functional T4SS. In a typical conjugation experiment, the pCMS1-selectable derivative pCMS1::tet, constructed by replacing the opd gene with opd::tet, is horizontally transferred into P. putida (Pandeeti et al. 2011). In addition to the opd gene, the genes involved in degradation of aromatic compounds, especially those encoding lower pathway enzymes are identified on plasmid pCMS1. Nevertheless, the opd gene does not show a typical transposon-like organization on plasmid pCMS1 (Pandeeti et al. 2011). The conserved opd regions found in plasmids pCMS1 and pPDL2 show sequence identity only in the upstream region of the opd gene (figure 3b). There is a marked difference in sequence in the regions downstream of the opd gene. In contrast to pPDL2, pCMS1 does not contain Tn3 downstream of the opd gene. Instead, it contains a larger ORF, orf345, that codes for a carboxyesterase. Interestingly, the sequence of orf306, later identified as mfhA (Khajamohiddin et al. 2006) on pPDL2 is identical to the \(3^\prime \) end of the orf345 sequence. Insertion of the Tn3 sequence in the coding region of orf345 created a new ORF, called orf306 in plasmid pPDL2. Our unpublished results indicate that orf306 is an independent transcriptional unit. It appears to be transcribed from the distally located promoter element generated due to insertion of Tn3 in the coding region of orf345. The newly created orf306 codes for an active esterase/lipase and hydrolyzes meta-fission products generated from aromatic compounds (Khajamohiddin et al. 2006). Further, the lipase activity of Orf306 caused accumulation of propionate in E. coli (Chakka et al. 2015). The propionate-dependent metabolic diversion in E. coli enabled the cells to survive on less preferred carbon compounds like p-nitrophenol, which are generated during biodegradation of OP insecticides such as parathion and methyl parathion (Chakka et al. 2015). Since the orf306 is part of the plasmid pPDL2-borne opd element, its existence probably stimulates the innate ability of the recipient strain, and promotes, to mineralization of OP insecticide residues completely, which have accumulated in the soil.

Fig. 5
figure 5

Genetic organization of mpd elements: The mpd gene is shown in red colour arrow. The IS elements flanking the mpd gene of Pseudomonas sp. strain WBC-3 are shown using purple colour arrows (panel a). Genetic organization of the 4.7-kb mpd element identified in seven bacterial strains is shown in panel b. The mpd gene is shown with red colour arrow. The orf463, orf232 and orf259, which code for sigma factor 70, permease and an energy-transducing component, ExeB, are shown with blue, yellow and green colour arrows, respectively.

Full size image

The TnopdA element

The opd gene homologues are also found in a number of soil bacteria (Iyer et al. 2013). The most notable among them is opdA, which has been isolated from Agrobacterium radiobacter P230 (Horne et al. 2002b). The chromosomally located opdA gene is organized as a transposable element (figure 4). In addition to opdA, the TnopdA element contains three ORFs and inverted repeats typically seen in transposon Tn610 of Mycobacterium fortuitum (Horne et al. 2003). The transposase (TnpA) of TnopdA is identical to the TnpA sequence of Tn610. The two ORFs found between tnpA and opdA are predicted to code for a truncated transposase (orfA) and an ATP-binding protein (orfB). The event of transposition was successfully demonstrated in E. coli, demonstrating the ability of horizontal mobility of opdA among soil bacteria (Horne et al. 2003)

Tn mpd elements

The methyl parathion hydrolase (MPH) encoding mpd genes reported to date have been isolated from agricultural soils or activated sludge collected from pesticide manufacturing plants in China. The first mpd gene was cloned from Plesiomonas sp. strain M6 (Zhongli et al. 2001). This chromosomally located mpd gene surprisingly showed no homology to any of the known OP-degrading genes. Following this discovery, a number of other mpd sequences were also cloned from different soil bacteria (Liu et al. 2005; Zhang et al. 2006; Iyer 2016). Isolation of a plasmid-borne mpd gene from Pseudomonas sp. strain WBC-3 provided sufficient indication for lateral gene transfer of mpd. In this soil isolate, an indigenous plasmid, pZWL0, contained both a mpd gene and genes responsible for degradation of p-nitrophenol (Liu et al. 2005). Interestingly, the plasmid-borne mpd gene has shown genetic organization typically seen in class I transposons (Wei et al. 2009). The IS elements, IS6100 are found flanking the mpd gene (figure 5a). The mobile nature of the mpd element has also been demonstrated in a typical transposition assay performed in Pseudomonas sp. strain WBC-3 (Wei et al. 2009). The lateral transfer of mpd genes became evident with the discovery of a highly conserved mpd element in seven bacterial strains (Zhang et al. 2006). In all of them, a 4.7 kb region containing a mpd gene is highly conserved. Five ORFs were conserved in this region. One of them is similar to the TnpA-coding sequence of an IS element, IS6100. A perfect 14 bp inverted repeat also exists in its flanking sequences. The second ORF, designated as orf463, found immediately upstream of the IS element, codes for a protein that shows considerable similarity to a house-keeping sigma factor. The other ORFs, orf232 and orf259, found upstream of the mpd sequence code for a permease and an energy-transducing component ExeB, respectively. Both ORFs are involved in the membrane transport process (figure 5b). The significance of their roles as part of the mpd element is currently unclear and are speculated to have a role in processing and membrane-targeting of MPH, which has a 35 amino acid-long signal peptide (Zhang et al. 2006). Although the functions of MPH-associated proteins are speculative, the existence of identical mph elements in seven different bacterial strains having weak taxonomic relationship to each other, strongly supports horizontal mobility of the mph genes among disparate soil microbes.

Fig. 6
figure 6

Organization of opd genes in meta-genome: (SEED, http://pubseed.theseed.org). The opd-linked genes encoding membrane transport proteins are shown with coloured arrows. The orf1388 codes for a probable conserved transmembrane protein and orf534 codes for a MFS permease. The ORFs orf397, orf414, orf430, orf395, orf532, orf566, orf562 and orf546- code for a possible membrane proteins and rfaB- codes for glycosyltransferase. The black arrows represent nonmembrane proteins.

Full size image

Unanswered questions

Evolution of novel genes and their lateral transfer play a very crucial role in adaptation of bacterial cells to various stress conditions. Enzyme promiscuity also plays a critical role in the evolution of new catalytic functions (Toscano et al. 2007; Khersonsky and Tawfik 2010; Purg et al. 2016). An ancillary activity can be changed into main activity of an enzyme in presence of substrate analogues due to few mutations. For example, the caboxylesterase (CE) of Mesorhizobium loti also shows promiscuous phosphodiesterase (PDE), phosphotriesterase (PTE) and lactonase activities. A single mutation (E183K) turns CE to PDE and substitution of E183 to glutamine abolishes all the activities (Mandrich and Manco 2009). The organisms retain such conversions if gained activity contributes to organismal fitness (Russell et al. 2011). The phenomenon described as ‘substrate-assisted gain-of-function’, the promiscuous enzymes can attain high catalytic activity towards substrates that were hitherto unknown to the native environment, in particular xenobiotics (Mandrich and Manco 2009; Russell et al. 2011; Davidi et al. 2018). A number of studies are available on evolution of PTEs, and combined structural and functional analyses have clearly established an evolutionary link between quorum-quenching lactonases and organophosphate hydrolases (OPH) (Afriat-Jurnou et al. 2012; Elias and Tawfik 2012; Bergonzi et al. 2018; Rhoads et al. 2018). Further, there also exists a strong structural similarity between \(\upbeta \)-lactamase and methyl parathion hydrolase (Dong et al. 2005). As described earlier in this review, the organization of these newly evolved genes (opd and mpd), as mobile elements and their localization on self-transmissible plasmids contribute to the spread of these genes among soil bacteria.

The accumulated experimental evidence obtained on the recent and rapid evolution of PTEs is indisputable. The primary question remaining, however, concerns the persistence of OP residues in agricultural soils and their contribution to the evolution of PTEs. Due to their short half-life, the recalcitrant organochlorides like HCH, DDT etc., are replaced with OPs to minimize the damage to the environment. Their half-life in most of the environments is less than 15 days. The impact of OP residues with such a short half-life to the evolution of PTEs is thus questionable. If OPs are physiological substrates, why are OPH homologues found in pathogenic bacteria like M. tuberculosis? These are pathogens that rarely come in contact with OPs and are known for both their genome reduction and microbial minimalism (Moran 2002). If the primary role of PTEs is indeed to metabolize OP residues, the opd homologues should have been lost during the adaptation of these bacteria to their pathogenic life style. Therefore, their presence and maintenance strongly suggests an unknown physiological function for PTEs. Structural insights of PTEs support their evolution through substrate-assisted gain-of-function and the lateral mobility of the corresponding genes among soil bacteria strongly suggests the existence of selection pressure necessary for retention and dispersion of the evolved genes. What is this selection pressure? If promiscuity and primary activities of PTEs are taken into consideration then they are likely to be either phosphotriesterases or lactonases. Identification of such substrates and establishing their physiological role in bacteria and archaea should be a key future goal of research in this area.

Fig. 7
figure 7

Organization of mpd genes in meta-genome: The mpd-linked genes coding for membrane transport proteins are shown using coloured arrows. The orf324 codes for a possible conserved transmembrane protein; cimH- codes for L-malate or citrate/H+ symporter CimH; dipB- codes for a sensor kinase, DpiB; orf332- codes for a nitrate ABC transporter-permease protein; orf418- codes for a nitrate ABC transporter-nitrate-binding protein. The black arrows represent nonmembrane proteins.

Full size image

Future perspectives

Most of genomes and meta-genomes have pte-like genes. Interestingly, the genes linked to opd and mpd sequences are to a large extent conserved. As stated already in this review, no structural homology exists between opd and mpd genes. Their progenitors and evolutionary paths are also quite different (Parthasarathy et al. 2017a). Interestingly, despite following independent evolutionary paths, the genes linked to these structurally distinct pte genes are conserved (figures 6&7). Frequently, genes encoding MFS-permeases and proteins associated with the ABC transporter family are also linked to pte genes. These are interesting observations and they correlate well with the recent reports on OPH biology (Parthasarathy et al. 2016). OPH has recently been shown to be a periplasmically located lipoprotein, which is anchored to the periplasmic face of the inner membrane via a diacyl glycerol moiety, and which has also been shown to exist in a 293-kDa multiprotein complex (Parthasarathy et al. 2016). Membrane transport proteins such as outer membrane permeases, components of ABC transporters, are potential interaction partners of OPH. There are also striking similarities between OPH-interacting partners and the proteins coded by opd/mph linked genes (figure 6&7). What is the physiological relevance for the link between pte genes and genes coding for membrane transporters? For example, does OPH have a role in membrane transport? Since opd null mutants failed to grow in a medium having OPs as sole phosphate source, can we implicate OPH in phosphate acquisition? Phosphate is a macro-nutrient and it is used only if available in inorganic form. However, because most of the soil phosphate sources are in organic form and provide the major phosphate source available to soil bacteria, it is conceivable that PTEs in association with phosphodiesterases (PDEs) can generate an inorganic phosphate pool from these compounds in periplasmic space. The role of MFS permeases in transport of nutrients from the outer membrane to the periplasmic space is a well-established phenomenon (Pao et al. 1998). Thus, the MFS-permease together with ABC transporters has the potential to play an important role in the acquisition of phosphate from OPs.

The OPs need not always have to be insecticides or nerve agents. They also include a number of natural phosphate-triesters (PT) that exist in soils. All of them have the potential to serve as a phosphate source if mineralized in the presence of appropriate PTEs and PDEs. A recent study on the functional meta-genomics of these systems suggested the existence of novel PTEs that cannot be detected through homology search (Colin et al. 2015). Structural diversity of PTEs and their broad substrate range strongly suggest a role in the generation of inorganic phosphate from complex natural and man-made PTs. Due to the fact that phosphate is a major nutrient, its limitation can create a strong selection pressure for the evolution of novel PTEs from a variety of progenitors that show promiscuous activity towards a different PTs. The hypothesis implicating PTEs in phosphate acquisition appears to be promising, as an adequate supply of inorganic phosphate is essential to support microbial life in the soil.