Wide substrate range for a candidate bioremediation enzyme isolated from Nocardioides sp. strain SG-4 G

Abstract Narrow substrate ranges can impact heavily on the range of applications and hence commercial viability of candidate bioremediation enzymes. Here we show that an ester hydrolase from Nocardioides strain SG-4 G has potential as a bioremediation agent against various pollutants that can be detoxified by hydrolytic cleavage of some carboxylester, carbamate, or amide linkages. Previously we showed that a radiation-killed, freeze-dried preparation (ZimA) of this strain can rapidly degrade the benzimidazole fungicide carbendazim due to the activity of a specific ester hydrolase, MheI. Here, we report that ZimA also has substantial hydrolytic activity against phthalate diesters (dimethyl, dibutyl, and dioctyl phthalate), anilide (propanil and monalide), and carbamate ester (chlorpropham) herbicides under laboratory conditions. The reaction products are substantially less toxic, or inactive as herbicides, than the parent compounds. Tests of strain SG-4 G and Escherichia coli expressing MheI found they were also able to hydrolyse dimethyl phthalate, propanil, and chlorpropham, indicating that MheI is principally responsible for the above activities.


Introduction
Enzymatic bioremediation is attracting increasing attention as an effectiv e, 'gr een' method for pollution abatement (Sutherland et al. 2004, Scott et al. 2008, Sharma et al. 2018, Thakur et al. 2019, Wackett and Robinson 2020, Mousavi et al. 2021, Saravanan et al. 2021 ).It gener all y involv es the use of cell-fr ee pr epar ations of r edo x cofactor-inde pendent enzymes as catalysts for r a pid r emov al of organic and inorganic pollutants from contaminated environments (Scott et al. 2008, Thatoi et al. 2014, Sharma et al. 2018, Mousavi et al. 2021 ).Such enzymes have been developed for agricultural and industrial pollutants such as pesticides (Sutherland et al. 2004, Pandey et al. 2010, Coppin et al. 2012, Hennessy et al. 2018 ), plastics/plasticizers (Boll et al. 2020 ), and explosives (Fida et al. 2014, Karthik e yan et al. 2016 ).
Ho w e v er, v ery fe w of these enzymes have yet seen widespread commercial use, despite demonstrated efficacy in specific field trials, and in many cases this has been because their substrate range has been too narrow to allow versatile use against a range of pollutants (Sutherland et al. 2004, Scott et al. 2008, Scott et al. 2010 ).Many pollutants can be degraded in cofactor-independent r eactions catal ysed by ester hydr olases, but the div ersity of ester bonds and substituent moieties has so far hindered widespread deployment of such enzymes as bioremediation agents (Scott et al. 2008, Thakur et al. 2019 ).Modern protein engineering technology now often enables a broadening of substrate range in respect of the substituent moieties, but enzymes able to attack different types of ester bonds have generally proven elusive.For example, se v er al carboxylester ases can hydr ol yticall y detoxify pyr ethr oid insecticides, but none are yet known to also hydrol-yse organophosphates , carbamates , or amides (Hatfield and Potter 2011 ).Conv ersel y man y amidases and carbamate hydr olases hav e br oad substr ate specificity within amides and carbamates, r espectiv el y, but negligible activity for other substrate classes (Wu et al. 2020, Mishra et al. 2021 ).In a similar v ein, the ester ases so far known to degrade phthalate diesters do not degrade carbamates, amides , organophosphates , and other classes of ester pollutants (Boll et al. 2020 ).
One pr omising bior emediation enzyme in this r espect is the ester hydr olase MheI fr om Nocardioides sp.str ain SG-4 G.This ca. 25 kDa enzyme (awaiting an E.C. number from the Enzyme Commission) can be produced at scale, formulated in active form in a radiation-killed freeze-dried preparation of the wild type bacterium (hereafter designated ZimA, the term MheI being r eserv ed for the enzyme itself), and used to decontaminate water from commercial fruit-dipping operations containing high levels of the benzimidazole fungicide carbendazim (methyl-1H-benzimidazol-2-ylcarbamate; MBC) (Pandey et al. 2010 ).Furthermore, carbendazim is a carbamate ester, but Mhe1 also hydr ol yses model carboxylesters such as naphthyl acetate , p-nitrophenyl acetate , and methyl salicylate (Pandey et al. 2010 ).
Tests of the ability of strain SG-4 G to grow by utilizing some compounds as carbon sources used a pr e viousl y described carbon-free minimal salt medium (MSM; Pandey et al. 2010 ), as did tests of the ability of pr e-gr own cells of this strain and E. coli pDEST17-mheI to degrade these compounds in the absence of cell gro wth.Po wdered forms of the compounds were dissolved in MSM, which was then filtered through 0.22 μm sterilized filters (Millipore, MA).Nutrient agar (NA) and nutrient (NB) or Luria-Bertani (LB) broth were also used as rich media to support bacterial growth when required.

Types of degr ada tion assay
Four different types of assays were used to characterize the ability of ZimA to degr ade the c hemicals of inter est.One simpl y used the radiation-killed freeze-dried ZimA preparation.The second involv ed gr owth assays to determine whether strain SG-4 G could use the test chemicals as a sole source of carbon and energy.The third tested pr e-gr own str ain SG-4 G cells to see whether their ability to hydr ol yse the test chemicals was constitutively expr essed.Finall y, pr e-gr own E. coli cells expressing the mheI gene were tested to see if the MheI enzyme was responsible for the hydr ol ysis of the test chemicals.Except where noted, all assays were carried out in triplicate.

Degr ada tion assays with ZimA
Following Pandey et al. ( 2010 ), various doses of ZimA were added to 50 ml MSM in 100 ml Erlenmeyer flasks containing the test c hemicals (at concentr ations below their aqueous solubility limits), and the reactions allo w ed to proceed under aseptic conditions (28 • C, 100 r pm).One-millilitr e samples wer e collected at v arious time points and filtered through 0.45 μm sterile filters (Millipor e, MA), befor e formic acid was added to a final concentration of 1% v/v.Samples were then stored at 4 • C until analysed as below.Flasks with all r ea gents except ZimA wer e used as negativ e controls.

Degr ada tion assays with growing cells of strain SG-4 G
An aer obicall y gr own (28 • C, 100 r pm) ov ernight seed cultur e of strain SG-4 G in NB was used to inoculate (1%, v/v) 50 ml of the supplemented MSM, which was then incubated at 28 • C with shaking at 100 r pm.One-millilitr e samples wer e collected at v arious time points, filtered through 0.45 μm filters (Millipore, MA), formic acid added to a final concentration of 1% v/v, and the resultant mixtur e stor ed at 4 • C until anal ysed.Flasks without SG-4 G cells but with the r espectiv e c hemicals of inter est wer e used as negativ e contr ols.

Degr ada tion assays with pre-grown cells of strain SG-4 G and E. coli expressing mheI
Following Pandey et al. ( 2010 ), a single colony of SG-4 G or E. coli BL21-AI™ pDEST17-mheI was picked from a plate and inoculated into 1 l of nutrient broth (NB) for SG-4 G or 250 ml of LB plus 100 mg/ml ampicillin for E. coli pDEST17-mheI.These cultures were then incubated at 28 • C with shaking at 100 rpm until they reached mid exponential phase (OD 600 ∼0.6 and 1.0, r espectiv el y).Fifty millilitr e of NB-gr own cells and 5 ml of LB-grown cells were then harvested, washed with ice-cold MM, and resuspended in 10 ml of MM containing 10 mM glucose and the r equisite concentr ation of the compound in question.The resuspended cultures were then incubated at 28 • C with shaking at 100 rpm.Samples were collected at various time points, filtered through 0.45 μm filters (Millipore), formic acid added to a final concentration of 1% v/v, and the r esultant mixtur e stor ed at 4 • C until anal ysed.Samples containing all the r ea gents except the pr e-gr own cells were used as negative controls.

Analytical methods
A high-performance liquid c hr omatogr a phy system (HPLC, Agilent Technologies , C A) coupled in series to either a diode array detector (DAD; Agilent Technologies) or a time-of-flight mass spectrometer (TOF/MS; Agilent Technologies) was used for quantitative and qualitative analysis, respectively.
Compounds (20 μl injection per sample) were separated and eluted at 25 • C on an Aqua C18 column (5 μm particle size, 250 × 4.60 mm; Phenomenex, CA) under v arious isocr atic conditions of acetonitrile in water (containing formic acid at a final concentration of 0.1%) as a mobile phase at various flow rates ( Supplementary Table S1 ).Dimethyl phthalate, propanil, monalide, and c hlor pr opham and their degr adation pr oducts (monomethyl phthalate, 3,4-dic hlor oaniline, 4-c hlor oaniline, and 3-c hlor oaniline, r espectiv el y) wer e quantified using authentic standards with the DAD operating at appropriate wavelengths ( Supplementary Table S1 ).The identities of all test chemicals and, wher e possible, their tr ansformation pr oducts wer e confirmed with the TOF/MS operating as described pr e viousl y (P andey et al. 2010 ).

Phthalate esters
ZimA at 400 ppm completely transformed 130 μM dimethyl phthalate within 48 h.The reaction was a ppr oximatel y linear over the first 24 h, after which time over 90% of substrate was consumed (Fig. 1 A and Supplementary Fig. S1 ).Assays starting with higher substrate concentrations (maximum 1.29 mM) did not increase the absolute amount of substrate consumed, so the percentage consumed decreased (Figs 1 B and 2 A).This indicates that the K M of the enzyme(s) catalysing the first cleavage reaction was < 130 μM.Near-stoichiometric amounts of the product, monomethyl phthalate (Rt.7.53 min, m / z 181) were observed (Fig. 1 C), so only one of the two ester bonds of the phthalate was efficientl y hydr ol ysed.Howe v er, small amounts of phthalic acid in its protonated anhydride form ( m / z 149) were detected, suggesting some cleav a ge of the r emaining ester bond (data not shown).
Tests for growth of strain SG-4 G on MM supplemented with dimethyl phthalate found that it was not used as a sole source of carbon and energy for growth (data not shown).Howe v er, pr egr own liv e SG-4 G cells harv ested fr om 50 ml cultur es (OD 600 ∼ 0.6) completely transformed 1.29 mM dimethyl phthalate in 24 h (Fig. 1 C), while pr e-gr own E. coli cells harv ested fr om 5 ml cultur es (OD 600 ∼ 1) expressing a his-tagged version of MheI transformed just over half (68%) the 130 μM dimethyl phthalate provided in the same timeframe (Fig. 1 D).As with ZimA, both these cell types pr oduced near-stoic hiometric amounts of monomethyl phthalate (Fig. 1 C and D).We do not know the r elativ e concentr ations of the MheI in the two cell types, or whether the his-tag influenced activity.Ho w e v er, the low yield of Mhe1 obtained from the E. coli cells (Pandey et al. 2010 ) suggests MheI was at least a major contributor to the phthalate degradation activity of ZimA.
The m uc h lo w er aqueous solubilities of dibutyl and dioctyl phthalate ( ∼40.2 and 0.56 μM, r espectiv el y) meant equiv alent time course analyses of their degradation by 400 ppm ZimA had to use m uc h lo w er starting substr ate concentr ations.Ne v ertheless, these assays sho w ed complete degradation of 54 and 2.56 μM of those substr ates, r espectiv el y, within 4 h (Fig. 3 ).The r espectiv e products, monobutyl phthalate and monooctyl phthalate, were detected at their expected m / z values (223 and 279, respectively; Supplementary Figs S2 and S3 ) but could not be quantified because authentic standards were not a vailable .Only trace amounts of phthalic acid were seen (data not shown).

Anilide herbicides
Assays of pr opanil ( N -(3,4-dic hlor ophen yl)pr opanamide) at starting concentrations between 0.23 and 0.57 mM with 800 ppm ZimA sho w ed 50%-70% of the substrate, depending on its starting concentr ation, wer e degr aded within 24 h (Figs 2 B, 4 A, and Supplementary Fig. S4 ).As with dimethyl phthalate, the percenta ge degr aded w as lo w er with the higher starting concentration but, unlike the dimethyl phthalate situation, the absolute amount was higher.This suggests the K M for propanil could be as high or higher than the highest starting concentration (0.57 mM).As with some of the other substrates tested, the low aqueous solubility limit of propanil (0.6-1.0 mM) constrained our ability to further inv estigate the substr ate concentr ation dependence of the r eaction.The incomplete degradation of the propanil also enabled us to compare the amounts degraded in the first and second 12 h of the assa ys .Only ∼20% less was degraded in the second interval than the first ( Supplementary Fig. S5 ).T hus , under the conditions of the assay at least, the ZimA pr epar ation should be sufficiently stable to continue significant degr adativ e activity for se v er al da ys .
The metabolite expected from propanil hydrolysis, 3,4dic hlor oaniline, was detected in these assays (Rt.6.39 min, m / z 162), but in less than stoichiometric amounts, just 0.08-0.15mM at 24 h (data not shown).This was pr obabl y not due to further metabolism because the transformations known for this compound r equir e cofactors, whic h the cell-fr ee ZimA pr epar ation would have limited ability to recycle (Scott et al. 2008 ).Rather we expect that the higher hydrophobicity of the 3,4-dic hlor oaniline ma y ha v e compr omised its r ecov ery during sample pr ocessing.
As with dimethyl phthalate, tests for growth of strain SG-4 G on MM supplemented with propanil found it was not used as a sole source of carbon and energy for growth (data not shown).Ho w e v er, the pr e-gr own SG-4 G cells degr aded 0.272 mM pr opanil completely after 18 h, with 0.10 mM 3,4-dichloroaniline recovered, Our monalide ( N-(4-c hlor ophen yl)-2,2-dimethylpentanamide) pr epar ation contained 10% (based on peak area) of an impurity that the supplier termed an isomer.We suspect this was a structural isomer of monalide, possibly N -(2c hlor ophen yl)-2,2-dimethylpentanamide or N -(3-c hlor ophen yl)-2,2-dimethylpentanamide, since no geometric or stereoisomers of it are possible.Reactions of 800 ppm ZimA with four starting monalide concentrations between 46 and 89 μM sho w ed pr ogr essiv e degr adation ov er 24 h (Figs 2 C, 4 B, and Supplementary Fig. S6 ).Up to 35%-40% of both the nominate monalide and the isomer were degraded in this time, although this was only 10%-15% above bac kgr ound degr adation in the no-enzyme negativ e contr ols ( Supplementary Fig. S7 ).There was little difference between starting concentrations in the amount of monalide degraded in percentage terms, but more in terms of absolute amounts was seen at the highest starting concentration, suggesting K M could be as high or higher than 89 μM.As with pr opanil, ther e was less than stoic hiometric r ecov ery ( < 5%) of the hydr ol ytic product, in this case 4-chloroaniline (Rt.4.33 min, m / z 128; Supplementary Fig. S7 ), again, we suspect, reflecting losses due to its high hydrophobicity during sample processing.
Consistent with the results for propanil, strain SG-4 G was unable to grow on MM supplemented with monalide, but pre-grown SG-4 G cells degraded over 90% of 19 μM of it in 18 h, forming 3 μM 4-c hlor oaniline (data not shown).

Chlorpropham
Five doses of ZimA varying from 1 to 1000 ppm were tested for their ability to degrade 42 μM chlorpropham (Fig. 5 and Supp lementary Fig. S8 ).There was no detectable degradation, even at 96 h, with 1 or 10 ppm ZimA, but complete degradation at 18 h with 500 and 1000 ppm, and lar gel y complete degradation at 42 h with 100 ppm.Inter estingl y, in the latter case, significant degradation continued between 24 and 42 h, confirming the stability of the enzyme evidenced in the propanil results abo ve .One product expected, 3-chloroaniline (Rt.5.5 min, m / z 128) (the other being CO 2 ), was detected at the higher enzyme doses but, as with the c hlor o-and dic hlor oanilines pr oduced fr om monalide and pr opanil, r espectiv el y, at less than stoic hiometric amounts and, in this case, there was too little to quantitate accur atel y ( Supplementary Fig. S8 ).
Assays at various substrate concentrations sho w ed the activity/substr ate concentr ation curv e had not plateaued at the high- est substrate concentration (37 μM), suggesting the K M may exceed the aqueous solubility limit (42 μM; ICSC 2004 ) of the chlorpropham (Fig. 2 D).
As with the other compounds abo ve , strain SG-4 G did not grow on MM supplemented with 40 μM c hlor pr opham but pr e-gr own SG-4 G cells and E. coli cells expressing MheI could transform it to the same product as ZimA, in this case 3-c hlor oaniline (data not shown).

Discussion
We have found that a r adiation-killed fr eeze-dried pr epar ation of a growth culture of strain SG-4 G can degrade three phthalate di-Figure 6.Heat map-coded correlation matrix of amino acid similarity percentages for enzymes with activity against any of the phthalates, carbamates, amides found to be substrates for MheI here or in Pandey et al. ( 2010 ).Note that only one of the four MheI isolates other than the one her ein fr om SG-4 G is shown because the r est wer e not bioc hemicall y c har acterized in detail.Abbr e viations for substr ates and accession numbers for enzymes are as per Supplementary Table S2 .Clustal Omega 1.2.2 (mBed algorithm) was used for multiple protein sequence alignment and generating the identity matrix (Sie v ers and Higgins 2018 ).esters , two anilide herbicides , and a carbamate ester herbicide in addition to its pr e viousl y demonstr ated activity for the benzimidazole fungicide carbendazim.Furthermore, we have shown that substantial amounts of these activities are due to the cofactorindependent MheI enzyme .T hus , the strain and enzyme ha ve acti vity against carbo xylester, carbamate , and amide linkages .Importantl y, the hydr ol ysis of the phthalates gr eatl y r educes their toxicity, and the hydr ol ysis of the carbendazim, anilides, and c hlor pr opham eliminates their r espectiv e fungicidal and herbicidal activities (Jonsson and Baun 2003, Pandey et al. 2009, Carvalho et al. 2010, Pujar et al. 2018 ).Given the ubiquity of the phthalates in various industries and the large numbers of benzimidazole fungicides and anilide and carbamate herbicides used in a gricultur e, str ain SG-4 G and MheI are potentially versatile bioremediant resources.
Se v er al other bacteria are known to degrade phthalate esters, and others are known to degrade benzimidazole, or other carbamate, anilide or other amide herbicides/pesticides (Xu et al. 2007, Liang et al. 2008, Benjamin et al. 2015, Boll et al. 2020, Wu et al. 2020 ; Supplementary Table S2 ).In a few cases, the enzyme responsible for the degradation has also been characterized ( Supplementary Table S2 ) (Zhang et al. 2012, 2019, Li et al. 2013, Sun et al. 2013, Chen et al. 2016 ).
Notwithstanding the w ell-kno wn promiscuity of esterases for a range of substrates within particular chemotypes of pesticides (Kourist et al. 2008, Martinez-Martinez et al. 2018 ), ho w e v er, none of these other bacteria/enzymes have yet been shown to possess both the carboxylesterase activity against the phthalates and carbamate hydrolase and/or amidase [and in one case also urease (Reichel et al. 1991 )] activities against the herbicides/pesticides ( Supplementary Table S2 ).On current knowledge at least, SG-4 G/MheI is thus uniquely versatile in showing both sets of activities.Notabl y thr ee gene/enzyme systems (GenBank accession numbers CCI05716, ARQ80492, and AEA07594) closely related to mheI /MheI and two others (AEB78730 and A O W37542) that are identical to it have also been isolated from various soil bacteria and, if tested, these w ould likely also sho w a similar substrate range (Fig. 6 ) (Zhang et al. 2013(Zhang et al. , 2017 ) ).Some of the other enzymes shown to have either the phthalate esterase or the herbicide/pesticide carbamate hydrolase or amidase/urease activities might also pr ov e to hav e similarl y wide substr ate r anges if their individual substr ate r anges wer e tested.Significantly, one, CemH, has shown amidase activity against anilide herbicides plus esterase activity against a phenoxyester insecticide ( Supplementary Table S2 ) (Li et al. 2013 ).All these other enzymes, including CemH, show < 20% similarity to ZimA and most also show < 20% similarity to each other (Fig. 6 ).The exceptions in respect of the latter are the amidases, PrpH and DmhA, with 72% similarity to one another, and three enzymes isolated from metagenomes with phthalate esterase activity, which are up to 48% similar to one another and 31%-37% similar to Cemh.There are thus several largely independent lineages of enzyme to assess as potential br oad-substr ate-r ange bior emediation enzymes.
Significantly also, some of these lineages likely utilize different catal ytic mec hanisms, whic h could in turn impact on their ester or amide substr ate r anges .For example , MheI and some other hydrolases utilize a charge relay mechanism that operates most effectiv el y at neutr al-basic pHs (Lei et al. 2017, Du et al. 2021 ), but the negativ el y c har ged carboxylic gr oup gener ated by hydr olysis of the first ester bond in phthalate diesters is detrimental to hydr ol ysis of the second, for which more acidic pH is optimum 5.5-6.5 (Fojan 2000 , Chahinian andSarda 2009 ).This likely explains why ZimA/MheI is better able to hydr ol yse the first than the second ester bond in these compounds.Ho w e v er, other ester ases hav e been described that can effectiv el y hydr ol yse both the diesters and monoesters, and other individual bacteria and consortia are known that utilize two different esterases for the two reactions (Hara et al. 2010, Wu et al. 2013, Prasad and Suresh 2015, Lu et al. 2020, Qiu et al. 2020 ).
The limited data pr esentl y av ailable also suggest that, steric constraints aside, the Mhe1 enzyme might prefer carboxylester and carbamate over amide (and urea) substrates.Combined across our study and Pandey et al. ( 2010), the highest rates calculated for ZimA with carbamates involved essentially complete degradation of 2.6 mM carbendazim in 24 h by 800 ppm enzyme and 47 μM c hlor pr opham in 18 h by 500 ppm, while that for a car-boxylester was 130 μM dimethyl phthalate in 48 h by 500 ppm.By comparison, the highest value calculated for an amide involved about half of 23-57 μM propanil in 12 h by 800 ppm.Depending on the specifics of the application, substantial increases in the latter rates, in particular, would be needed for cost-effectiv e bior emediation (see belo w).Ho w e v er, suc h incr eases should be ac hie v able with modern enzyme engineering technology (Turner 2003, Lutz and Iamurri 2018, Sharma et al. 2018, Mousavi et al. 2021 ).Notabl y, while ZimA is alr eady r elativ el y stable in both laboratory and field en vironments , enzyme stability can also be enhanced with this technology (Pongsupasa et al. 2022, Teufl et al. 2022 ).
The two major a ppr oac hes to bioremediation in which ZimA/MheI may be useful are the clean-up of contaminated water by fr ee-enzyme form ulations, as per ZimA, and the clean-up of contaminated soils by transgenic phytoremediation crops expressing MheI.The fact that SG-4 G expressed MheI constitutively (Pandey et al. 2010 ) is adv anta geous for the free enzyme applications because it facilitates its production.The fact that it had little activity for phthalate monoesters is disadv anta geous for fr ee enzyme remediation of the phthalates, although hydr ol ysis of the first ester bond still r epr esents substantial detoxification (Jonsson and Baun 2003, Xu et al. 2016, Tian et al. 2019 ).Similarly, while the ZimA assays and SG-4 G growth assays suggest they do not completely mineralize the anilide herbicides, this need not be problematic as the hydrolysis destroys their herbicidal activity (Kanawai et al. 2016).Some of the herbicide-deactiv ating pr operties of MheI may also hav e v alue in the de v elopment of herbicideresistant food or fibre crops.
In respect of free-enzyme contamination of w astew ater, w e note that ZimA has already been trialled successfully in decontaminating millimolar concentrations of carbendazim in rinsates from commercial potato processing operations (Pandey et al. 2010 ).High concentrations of other fungicides or insecticides a gainst whic h ZimA/Mhe1 has activity could also occur in the post-harv est pr ocessing wastes of other horticultur al oper ations (Felsot et al. 2003, Damalas et al. 2008, Scott et al. 2008, Galuszka et al. 2011, Delgado-Moreno et al. 2017 ), although, as noted, some enzyme engineering might be needed for cost-effective removal of some of those compounds from such waste streams.On the other hand, the phthalates and pesticides/herbicides encountered in many other w aste w aters of concern, such as from chemical production facilities and irrigated a gricultur e, ar e in the low nMlow μM r ange (Ritter 1990, Sta ples et al. 1997, Chang et al. 2007, Liang et al. 2008, Agarwal et al. 2015, Benjamin et al. 2015, Boll et al. 2020 ), and ZimA could be a ppr opriate as is for many such situations.
The le v els of activity r equir ed for enzymes deplo y ed in phytor emediation ar e substantiall y lo w er than those needed in freeenzyme bioremediation because the enzymes can be produced semi-continuously by the host plant and have much longer time frames in which to act (Macek et al. 2008 ).The mheI gene may ther efor e be a ppr opriate as is for phytor emediation trials a gainst industrial chemicals without herbicidal activity, e.g. the phthalates.Ho w e v er, in planta a pplications dir ected at herbicides, including in herbicide-r esistant cr ops, could clearl y be mor e pr oblematic.Notably MheI is secreted naturally by SG-4 G and is effectiv e extr acellularl y in our assa ys , but differ ent secr etion mac hinery would be needed for in planta applications.
In conclusion, we find that ZimA/MheI has potential as a versatile resource for the bioremediation of a range of problematic compounds in various contamination situations.We have only tested a small number of the many carbamate and anilide pesticides and herbicides that are used in a gricultur e, so its sub-str ate r ange is likel y to be significantl y wider than so far demonstrated.It will likely need optimization, including by modern enzyme biotechnology methods, for some substrates and applications, but for others, it may already be a relatively cost-effective option for bioremediation.

Figure 1 .
Figure 1.Degradation of dimethyl phthalate by ZimA, strain SG-4 G, and E. coli cells expressing the mhe I gene: (A) Degradation of 0.130 mM dimethyl phthalate and production of monomethyl phthalate by 400 ppm ZimA; (B) Degradation of different concentrations of dimethyl phthalate by 400 ppm ZimA; (C) Degradation of 1.28 mM dimethyl phthalate by live cells of the strain SG-4 G; and (D) Degradation of 0.13 mM dimethyl phthalate by E. coli cells expressing mhe I.