Enrichment of an Endosulfan-Degrading Mixed Bacterial Culture

Abstract

An endosulfan-degrading mixed bacterial culture was enriched from soil with a history of endosulfan exposure. Enrichment was obtained by using the insecticide as the sole source of sulfur. Chemical hydrolysis was minimized by using strongly buffered culture medium (pH 6.6), and the detergent Tween 80 was included to emulsify the insecticide, thereby increasing the amount of endosulfan in contact with the bacteria. No growth occurred in control cultures in the absence of endosulfan. Degradation of the insecticide occurred concomitant with bacterial growth. The compound was both oxidized and hydrolyzed. The oxidation reaction favored the alpha isomer and produced endosulfate, a terminal pathway product. Hydrolysis involved a novel intermediate, tentatively identified as endosulfan monoaldehyde on the basis of gas chromatography-mass spectrometry and chemical derivatization results. The accumulation and decline of metabolites suggest that the parent compound was hydrolyzed to the putative monoaldehyde, thereby releasing the sulfite moiety required for growth. The monoaldehyde was then oxidized to endosulfan hydroxyether and further metabolized to (a) polar product(s). The cytochrome P450 inhibitor, piperonyl butoxide, did not prevent endosulfan oxidation or the formation of other metabolites. These results suggest that this mixed culture is worth investigating as a source of endosulfan-hydrolyzing enzymes for use in enzymatic bioremediation of endosulfan residues.

The chlorinated cyclic sulfite diester endosulfan (Thiodan, bicyclo-[2.2.1]-2-heptene-5,6-bisoxymethylene sulfite) is a broad-spectrum insecticide that has been used extensively for over 30 years on a variety of crops. Endosulfan is often classified as a cyclodiene and has the same primary action and target site as other cyclodienes (3). However, it has chemical and physical properties significantly different from other cyclodiene insecticides that affect both its environmental and biological fates. In particular, endosulfan has a relatively reactive cyclic sulfite diester group (32) and, as a consequence, its environmental persistence is lower than that of other cyclodienes, albeit still higher than that of many other insecticides. Since the deregistration in many countries of most cyclodiene insecticides, the ongoing availability of endosulfan has become important as an alternative option in resistance management strategies of pest species. Additionally, compared to many other available insecticides, it has low toxicity to many species of beneficial insects, mites, and spiders (10). However, endosulfan is extremely toxic to fish and aquatic invertebrates and it has been implicated increasingly in mammalian gonadal toxicity (28–31), genotoxicity (4), and neurotoxicity (24). These environmental and health concerns have led to an interest in postapplication detoxification of the insecticide.

The aim of this research is the investigation of an enzymatic method for endosulfan detoxification. Enzymatic detoxification of pesticides is receiving serious attention as an alternative to existing methods, such as incineration and landfill. In particular, enzymatic insecticide bioremediation is the focus of extensive study after the isolation of a phosphotriesterase capable of detoxifying a range of organophosphate compounds from several bacterial species (for review, see reference 7 and references within). An essential initial step in the investigation of an enzymatic method for endosulfan detoxification is the definitive identification of a biological source of endosulfan-degrading activity. Numerous studies have described the degradation of endosulfan in soils (14, 26), soil inocula (11, 21, 27), mixed microbial cultures (1), and isolated microorganisms (9, 13, 17, 20, 22). The compound is degraded by attack at the sulfite group via both oxidation and hydrolysis to form the toxic endosulfate (endosulfan sulfate) and the nontoxic endodiol (endosulfan diol), respectively. The formation of endosulfate is thought to occur only through biological transformation, whereas hydrolysis to the diol occurs readily at alkaline pH (20). Many studies describing degradation of endosulfan in microbial cultures do not differentiate between chemical and biological hydrolysis, as culturing often leads to an increase in pH of the growth medium (20, 21). In addition to stringent pH controls, the detection of metabolites is important for the confirmation of degradation, as losses of endosulfan from culture media or soils occur readily by volatilization and adsorption to surfaces (12).

Microorganisms have increasingly been investigated as a source of xenobiotic-degrading enzymes (5). We are interested in the isolation of an endosulfan-degrading bacterium for further investigation into enzymatic endosulfan bioremediation. Using endosulfan as the only available sulfur source, we enriched soil inocula for microorganisms capable of releasing the sulfur from endosulfan, thereby providing a source of sulfur for growth. Since removal of the sulfur moiety dramatically decreases vertebrate toxicity of endosulfan (8, 10), this results in concurrent detoxification of the insecticide. We report here on the resultant mixed bacterial culture that, to our knowledge, is the first reported enrichment of an endosulfan-degrading microbial culture. The culture degrades endosulfan to produce a novel metabolite not reported to occur as a result of chemical hydrolysis. These results suggest this mixed culture is a potential source of an enzymatic bioremediating agent.

MATERIALS AND METHODS

Soil.

The soil used in this study was collected from a cotton field near Narrabri, New South Wales, Australia, at the end of the growing season. The field had generally received several applications of endosulfan in the summer months for at least the previous 5 years. The soil was fertile grey clay at pH 7.5. Topsoil was collected from the first 15 cm, air dried, and stored at 4°C for up to 1 month prior to enrichment.

Isolation of soil bacteria.

Soil (approximately 15 g) was first enriched for endosulfan-degrading organisms by the addition of 2 mg of technical-grade endosulfan in 100 μl of acetone to remoistened soil, followed by incubation in the dark at room temperature for 1 month. Further enrichment was then achieved by initiating shake flask enrichment cultures from these samples by using endosulfan as the only added source of sulfur. The enrichment medium (pH 6.6 to 6.8) consisted of 20 mg of technical-grade endosulfan (99% pure), 0.05% Tween 80, 2.0 g of KH₂PO₄, 7.5 g of K₂HPO₄, 1.0 g of NH₄Cl, 0.5 g of NaCl, 1.0 g of glucose, 0.1 g of MgCl₂, 0.86 mg of p-amino benzoic acid, 0.86 mg of nicotinic acid, and 10 ml of a trace element solution per liter. The stock trace element solution contained 20 mg of (NH₄)₆Mo₇O₂₄ · 4H₂O, 50 mg of H₃BO₃, 30 mg of ZnCl₂, 3 mg of CoCl₂ · 6H₂O, 10 mg of (CH₃COO)₂Cu · H₂O, and 20 mg of FeCl₂ · 6H₂O per liter. According to impurity data provided by Sigma Chemical Co. Castle Hill, New South Wales, Australia) the maximum limit of sulfite/sulfate contamination in the enrichment medium was less than 0.4 × 10⁻³ g liter⁻¹. Approximately 1 g of endosulfan-enriched soil was inoculated into 50 ml of enrichment media and cultured in a 400-ml Erlenmeyer flask on a rotary shaker (200 rpm) at 28°C for up to 14 days. Substrate levels were measured using thin-layer chromatography (TLC), and when approximately 50% of the endosulfan had degraded relative to sterile controls, 5 ml of the culture was transferred into 50 ml of fresh enrichment medium. Ten different soil samples were enriched for endosulfan-degrading activity. An endosulfan-degrading culture was obtained from only one of these samples. After approximately six transfers into enrichment media, cultures were transferred into sulfur-free media (see below) for further enrichment.

A sulfur-free medium was also designed because contaminating sulfur in the enrichment medium could promote culture growth, resulting in increases in optical density at 595 nm of the culture from 0.05 to 0.3. A second soil culture was initiated for the sole purpose of preparing medium free of contaminating sulfur. Sulfur-free medium was prepared by growing the second soil culture overnight in enrichment medium without endosulfan and then removing cells by centrifugation and filtering the supernatant through a 0.22-μm-pore-size filter. After inoculation of this medium with either the endosulfan-degrading culture or Escherichia coli TG1, no growth was observed until the addition of a source of sulfur. After the addition of either 50 μM sodium sulfite or magnesium sulfate, both the endosulfan-degrading culture and the E. coli TG1 culture were able to grow to an optical density at 595 nm of at least 0.8. The sterility of the sulfur-free medium was confirmed by the absence of growth when aliquots were incubated on rich medium agar plates. After initial enrichment, the endosulfan-degrading culture was maintained in sulfur-free medium with endosulfan as the only sulfur source. Rich medium agar used in this study included low-salt Luria broth (LB; 10 g of tryptone liter⁻¹, 5 g of yeast extract liter⁻¹, 0.5 g of NaCl liter⁻¹, 15 g of Noble agar liter⁻¹) and BUGM (Oxoid, Melbourne, Victoria, Australia).

Chemicals.

Technical-grade endosulfan (99% pure) for bacterial growth was a gift from Hoechst Schering AgrEvo Pty Ltd. Technical-grade endosulfan (used commercially) is a mixture of two diastereoisomers: alpha-endosulfan and beta-endosulfan in a ratio of 7:3. With the exception of endosulfan diacetate, insecticide and metabolite standards (at least 99% pure) were purchased from Chem. Services Inc. (West Chester, Pa.). Endosulfan diacetate was synthesized by peracetylation of endosulfan diol with acetic anhydride in dry pyridine at 80°C for 1 h and purified by silica chromatography. The O-benzyl oxime of endosulfan monoaldehyde was prepared by the reaction of the putative aldehyde (recovered by thin-layer chromatography [TLC] on alumina) with a fivefold excess of benzylhydroxylamine hydrochloride (Alltech, Baulkham Hills, New South Wales, Australia) in dry pyridine at room temperature for 8 h. All other chemicals used were of at least reagent grade.

TLC.

Cultures were extracted with equal volumes of ethyl acetate. The organic phase was passed through a 6-cm MgSO₄ column in a Pasteur pipette stoppered with glass wool to remove any residual water, gently evaporated under a dry nitrogen stream, dissolved in acetone, and then applied to neutral aluminium oxide F₂₅₄ TLC plates (Alltech). The plates were developed in either petroleum ether-acetone (4:1) or chloroform-ethyl acetate (3:1). R_f values for endosulfan and its metabolites in these solvent systems are given in Table 1. The aqueous phase was reduced to dryness by rotary evaporation, and the resultant residue was extracted with dichloromethane (DCM) to recover any hydrophilic metabolites. The DCM-soluble products were spotted onto TLC plates as described above and developed in methanol. Chlorine-containing constituents were visualized by spraying plates with silver nitrate-saturated methanol and then exposing them to UV light. The lower limit of detection of this method for endosulfan and metabolites containing the hexachlorinated ring structure was 0.1 μg (data not shown). As detection is based on formation of silver chloride, dechlorinated metabolites will have a detection limit relative to the level of dechlorination.

TABLE 1.

R_f values for TLC with different solvent systems and retention times of FID GC for endosulfan isomers and metabolites

Endosulfan isomers and metabolites	Rf values in various TLC solvent systemsa			Retention time (min)b
	Petroleum ether-acetone (4:1)	Chloroform-ethyl acetate (3:1)	Methanol
alpha-endosulfan	0.90	0.88		15.65
Endosulfan ether	0.83	0.86		11.92
beta-endosulfan	0.70	0.82		18.27
Endosulfate	0.60	0.78		21.02
Putative monoaldehyde product	0.60	0.57		12.37
Endosulfan lactone	0.60	0.45		14.36
Endosulfan hydroxyether	0.47	0.30		13.31*
Endodiol	0.35	0.13		16.33*
Endosulfan diacetate (internal standard)				19.46
Polar unknown	NAc	NA	0.22	NA

^aTLC was performed on aluminium oxide plates, and product detection was performed with silver nitrate and UV exposure. 

^bRetention time was determined using FID GC conditions described in Materials and Methods. *, Silylated with bis(trimethylsilyl)trifluoroacetamide. 

^cNA, not applicable. 

GC and GC-mass spectrometry (GC-MS) analysis.

As endosulfan and its chlorine-containing metabolites are strongly electronegative, previous studies have employed electron-capture GC for detection of these compounds. As we demonstrate in the present study, flame ionization detection (FID) can replace this if preliminary steps are used to recover the metabolites selectively. In addition, FID enables the use of DCM for clean and efficient solvent extraction.

Cultures (15 ml) were extracted with DCM (10 ml), and the organic phase was dried with MgSO₄, as described above. The solution of endosulfan and its lipophilic metabolites was diluted with hexane to yield a 20% hexane–DCM solution, which was applied to a 5-cm silica column (DCC silica gel, 63-200; Aldrich) within a Pasteur pipette. The column was flushed with a further 3 ml of 20% hexane–DCM. Control experiments demonstrated that endosulfan hydroxyether and endodiol were the only metabolites retained by the silica under these conditions. Endosulfan diacetate (50 μg) was added as an internal standard to the combined eluate and washings, which were then concentrated to 25 μl under a gentle stream of nitrogen before storage at −20°C and subsequent GC analysis using FID.

The more polar metabolites, endosulfan hydroxyether and endodiol, were subsequently eluted from the silica column with 10% methanol–DCM (8 ml). Endosulfate (40 μg) was added as the internal standard, and the recovered solution was evaporated to near dryness under a gentle stream of nitrogen. The residue was taken up in DCM (10 μl), and bis(trimethylsilyl)trifluoroacetamide (BSTFA; 25 μl) was added with initial vortex mixing to silylate the metabolites (6 h, room temperature) before storage at −20°C and GC analysis.

The addition of the internal standards to the fractions enabled both qualitative assessment of the metabolites from their relative retention times by GC and quantitative evaluation of the metabolic pathways. Losses from volatilization and extraction efficiencies ranged from 15 to 40% (depending on length of time incubated, media composition, and compound) and were calculated by comparison with stocks of known concentration. GC was performed using a Varian model 3300 with a cool on-column injector, an FID, and a computer with data acquisition and processing software. The capillary column was 5% phenyl methylsilicone (SE54, Alltech Econocap, 30 m by 0.32 mm [inside diameter], 0.25-μm film thickness) with a helium flow rate of 2 ml min⁻¹. The column was preceded by a retention gap of deactivated silica (2 m) to preserve the integrity of the column. A typical temperature program for analysis of the endosulfan metabolites comprised an initial period after injection of 2 min at 40°C and a temperature gradient of 20°C min⁻¹ to 200°C for 10 min, followed by a temperature gradient of 10°C min⁻¹ to 300°C.

The identities of the known metabolites in the fractions were confirmed by GC-MS using a VG Trio 2000 mass spectrometer interfaced to a Hewlett-Packard 5890 gas chromatograph (cool on-column injector), with VG MassLynx software for control and data acquisition. The GC column was 5% phenyl methylsilicone (SE54, Alltech Econocap, 30 m by 0.32 mm [inside diameter], 0.5-μm film thickness), with a helium flow rate of 1 ml min⁻¹. Ionization modes used for MS of the metabolites were either electron ionization (EI; 70 eV) or positive-ion chemical ionization (PCI; ammonia reagent gas; source pressure, 60 Pa). The molecular and fragment ions were generally represented by peak distributions over several masses because their respective chlorine compositions included the additional natural isotope ³⁷Cl.

RESULTS

Enrichment of microorganisms.

After approximately six rounds of successive subculturing in enrichment media followed by four rounds in sulfur-free media, TLC analysis and optical density measurements of the enrichment culture confirmed substantial disappearance of endosulfan with a simultaneous increase in bacterial mass. The culture was incapable of growth in sulfur-free medium without the addition of a sulfur source (Fig. 1). Addition of either alpha-, beta-, or technical-grade endosulfan promoted growth to various degrees. No significant growth was seen in the absence of endosulfan, and endosulfate could not substitute for endosulfan as a utilizable source of sulfur (Fig. 1).

FIG. 1.

Growth of a bacterial culture after enrichment for endosulfan degradative ability in sulfur-free medium containing endosulfan isomers (50 μM). OD₅₉₅, optical density at 595 nm.

When dilutions of the enrichment culture were incubated on agar plates for 48 h, two distinct colony types became apparent on enrichment medium agar and five were revealed on rich medium agar (LB and BUGM). Inoculation of sulfur-free medium broth containing endosulfan with isolates or combinations of isolates, after purification on agar plates, failed to produce a culture capable of growth. However, growth was observed when “scrapes” from the initial dilution plating on either medium type were used to inoculate sulfur-free medium broth containing endosulfan. Many separate components of the culture may be required for metabolism of endosulfan, including enzymes involved in hydrolysis, sulfite oxidation, and the provision of nutrients not supplied in the minimal medium. The inability of combinations of isolates to utilize endosulfan as a sulfur source suggested that a key species in the degradative process either did not form obvious colonies on agar plates or lost its ability to contribute to endosulfan metabolism after passage on rich media.

Continuous subculturing in sulfur-free medium led to an increase in rates of endosulfan disappearance, with no detectable levels of endosulfan remaining after 4 days by the twentieth subculture as compared to 8 days after the tenth subculturing. Degradative ability was retained in cultures initiated from frozen 20% glycerol stocks after several months at −80°C.

Characterization of endosulfan metabolites.

TLC and GC analysis indicated the disappearance of both diastereomers of endosulfan and the concomitant formation of endosulfan metabolites. The known metabolites of endosulfan are not diastereomeric. Three metabolites were identified as endosulfan hydroxyether, endosulfate, and endodiol on the basis of comigration with authentic standards on TLC plates developed in different solvent systems, coincident retention times on GC (Table 1), and structural confirmation by GC-MS (data not shown). A single additional metabolite, with mobility on TLC plates similar to that of endosulfate, was also detected. MS analysis (70-eV EI) of the compound after purification by TLC indicated a molecular ion with an m/z of 342 (³⁵C₆), isomeric with that of endosulfan ether (Fig. 2). The fragmentation pattern was also similar to that obtained with endosulfan ether, except for the absence of a prominent fragment ion with an m/z of 69 derived from the pentacyclic ether moiety. An analogous ion with an m/z of 85 was observed in the 70-eV EI mass spectrum of endosulfan hydroxyether (data not shown). Thus, the molecular structure of the novel isomer does not include a pentacyclic ether ring.

FIG. 2.

Mass spectra (70 eV EI) of endosulfan ether and putative endosulfan monoaldehyde.

The PCI mass spectrum [PCI(NH₃)] of the novel metabolite displayed the molecular parent ions [M+H]⁺ and [M+NH₄]⁺ with m/z of 341 and 358, respectively (data not shown), confirming the molecular mass (³⁵Cl₆) indicated previously in the EI mass spectrum (Fig. 2). The preliminary evidence indicated that the molecular structure of the novel isomer was that of endosulfan monoaldehyde (Fig. 3). The PCI mass spectrum of the metabolite also displayed fragment ions indicating consecutive losses of two molecules of HCl from [M+H]⁺ ions. Since the most probable site for gas phase proton attachment in the putative structure would be the carbonyl oxygen atom, the initial HCl loss may be rationalized as elimination of the reagent proton together with the vicinal bridgehead chlorine atom via a favored six-centered transition structure.

FIG. 3.

Proposed pathway for metabolism of endosulfan by the microbial culture described in this paper.

Support for the structure of the novel metabolite is provided by the observation that it forms an O-benzyl oxime derivative. Although the expected molecular ion is absent in its 70-eV EI mass spectrum (Fig. 2), an [M−CH₃]⁺ ion with an m/z of 430 is present (data not shown), indicating a relative molecular mass of 445 (³⁵Cl₆) for the derivative and substantiating a monoaldehyde structure for the metabolite.

Formation of endosulfan metabolites by the culture.

TLC analysis of the culture allowed qualitative comparison of the amounts of each product during the growth cycle after the tenth subculture in sulfur-free medium (Table 2). Endosulfate appeared rapidly in the growing culture, followed by the putative monoaldehyde, hydroxyether, and then small amounts of the diol. The endosulfate continued to accumulate until both isomers of the parent compound had disappeared and then remained at constant levels in the medium. Of the other metabolites, levels of the putative monoaldehyde decreased first, followed by the hydroxyether and then the diol.

TABLE 2.

Qualitative TLC analysis of endosulfan isomers and metabolites present in organic extracts from different stages of the culture growth cycle^a

	Growth stage	Days of growth	Presence of endosulfan metabolite
alpha-endosulfan			beta-endosulfan	Endosulfate	Putative monoaldehyde product	Endosulfan hydroxyether	Endodiol
Initial	0	+++++	+++	—	—	—	—
Early log	2	+++	++	+++	+	—	—
Mid log	3.5	+	+	++++	+++	++	+
Late log	4.5	—	—	++++	++	+++	+
Stationary	7	—	—	++++	—	+	+

^aAnalysis was of 2-ml culture samples grown in sulfur-free media with 50 μM technical-grade endosulfan. Extracts were developed in petroleum ether-acetone (4:1) and in chloroform-ethyl acetate (3:1). +++++, major product; ++++, +++, and ++, intermediate product; +, minor product; —, no product detected. 

Table 3 shows quantitation by GC analysis of various chlorinated products in DCM extracts from cultures at one time point in log phase growth during the twelfth subculture in sulfur-free medium as a percentage of the added sulfur-containing substrate (corrected for extraction efficiency and volatilization from the medium). Four sulfur-containing substrates were tested separately, each added to an initial concentration of 50 μM: technical-grade endosulfan, alpha-endosulfan, beta-endosulfan, and endosulfate. The final column of Table 3 shows the amount of substrate that is not accounted for by products in the organic extract. Significant amounts of chlorinated product(s) were detected by TLC in the DCM extract of the dried residue from the aqueous fraction. While we were unable to analyze this by GC or GC-MS, it corresponded semi-quantitatively to the amount of product unaccounted for in the organic fraction.

TABLE 3.

GC quantitation of endosulfan-related product appearance in log phase cells^a

Sulfur source	Endosulfan metabolite (% of added substrate)
	alpha- endosulfan	beta- endosulfan	Endosulfate	Putative monoaldehyde product	Endodiol	Endosulfan hydroxyether	Further metabolized productb
Technical-grade endosulfan	2.5	10.4	55.0	10.0	3.2	6.4	12.5
alpha-endosulfan	58.3	NDc	16.2	<0.5	<0.5	<0.5	25.1
beta-endosulfan	ND	9.5	4.2	1.4	1.4	3.5	80.0
Endosulfate	ND	ND	100.0	ND	ND	ND	0.0

^aThe amounts of the metabolites were determined after approximately six rounds of successive subculturing in enrichment media followed by six rounds in sulfur-free media supplied with 50 μM concentrations of different sulfur sources. 

^bPercentages were derived by subtraction from total (corrected for volatilization and extraction efficiency as described in Materials and Methods) and were semi-quantitatively accounted for as unidentified polar product in DCM residue from the aqueous fraction after removal of the metabolites named above by organic extraction. 

^cND, not detected. 

The predominant metabolite detected in cultures grown with alpha-endosulfan as the only sulfur source was endosulfate (16.2%) (Table 3). The putative endosulfan monoaldehyde, endosulfan hydroxyether, and endodiol were detected at much lower levels (<0.5%). Conversely, less endosulfate was detected in cultures grown with the beta isomer (4.2%) and the other metabolites were all detected at comparatively higher levels (1.4 to 3.5%). In cultures grown in the presence of beta-endosulfan, the majority of metabolites was found in the DCM extract of residue from the aqueous fraction after organic extraction. This was attributed to further metabolism of the described products. No growth was observed in cultures grown with endosulfate (Fig. 1), and the added endosulfate was recovered from the media after 7 days (data not shown).

Despite their initial accumulation and then disappearance in cultures grown in the presence of endosulfan, endodiol and endosulfan hydroxyether were not degraded when added to cultures grown in sulfur-free media with 50 μM added sulfite (data not shown). This may be due to an inability of these more polar compounds to traverse the cell membrane and therefore an inability to interact with their respective degradative enzymes. However, endodiol and endosulfan hydroxyether have been degraded by inocula from soil (27). Endosulfan strongly adsorbs to microorganisms, with the majority of the insecticide being associated with the cell membrane rather than the growth medium (22, 25). Hence, degradation of endosulfan presumably leads to an accumulation of products within the cell, facilitating their further degradation.

Effect of piperonyl butoxide.

Inclusion of 1 and 10 μM piperonyl butoxide (PBO, a known cytochrome P450 inhibitor) in sulfur-free medium with endosulfan did not prevent the formation of any metabolites, including endosulfate, nor did it significantly inhibit growth (data not shown). While culture growth was approximately halved by the addition of 100 μM PBO, the formation of metabolites was not qualitatively affected (data not shown).

DISCUSSION

Enzymatic bioremediation of insecticides is receiving considerable attention, particularly since the extensive characterization of a phosphotriesterase enzyme capable of detoxifying a range of organophosphate compounds (7). The basis of similar investigations for enzymes capable of detoxifying other classes of insecticides requires a source of enzymes for catalytic detoxification. This study describes the enrichment of a culture of soil bacteria capable of degrading endosulfan. Enrichment was achieved and maintained by providing endosulfan as the only sulfur source. Endosulfan is a poor biological energy source, as it contains only six potential reducing electrons and previous attempts to enrich for endosulfan-degrading microorganisms using the insecticide as a carbon source have been unsuccessful (11). However, endosulfan has a relatively reactive cyclic sulfite diester group (32). In this study, microorganisms were selected for their ability to release the sulfite group from endosulfan and to use this as a source of sulfur for growth. This selection procedure enriches for a culture capable of either the direct hydrolysis of endosulfan or the oxidation of the insecticide followed by its hydrolysis. The degradation products in our culture indicate that both hydrolysis and oxidation reactions are occurring. However, the accumulation of endosulfate and the inability of the culture to grow when this is provided as the sole sulfur source indicate that we have selected for the direct hydrolysis of endosulfan.

The strategy for enrichment also addressed the issues that endosulfan is virtually insoluble in water and spontaneously hydrolyzes at alkaline pH. A study into the distribution of the compound in sterile microbial broth showed that it concentrated at the glass-medium interface and that the inclusion of Tween 80 (a mixture of oleic, linoleic, palmitic, and stearic acids) resulted in dispersion of the pesticide (12). Therefore, we included Tween 80 in the enrichment broth to emulsify endosulfan, thereby increasing the amount of insecticide in contact with the soil bacteria. This detergent has previously been used to solubilize pyrethroids during the isolation of microorganisms capable of metabolizing permethrin (19).

Endosulfan is susceptible to alkaline hydrolysis (20), with approximately 10-fold increases in hydrolysis occurring with each increase in pH unit. Many previous studies have been unable to differentiate between chemical and biological hydrolysis of endosulfan because microbial growth has led to increases in the alkalinity of the culture medium (20, 21). To minimize nonbiological hydrolysis, the enrichment medium was buffered to pH 6.6 and cultures were monitored constantly to ensure that growth did not decrease hydrogen ion concentrations. We observed detectable levels of endodiol (>0.1 ppm) in sterile media inoculated with 50 μM endosulfan at pH 7.2 after 4 days (data not shown) and recommend that the pH of the medium for biodegradation studies of endosulfan be maintained below pH 7.0.

The formation and decay of metabolites led us to propose a pathway of endosulfan metabolism by the culture (Fig. 3). According to this pathway, the parent compound is either oxidized or hydrolyzed. The oxidation reaction is favored for the alpha isomer and produces endosulfate. Preferential oxidation of this isomer has been previously reported, and it is thought that it contributes a disproportionate amount of endosulfate found in the environment (2, 6, 21). Cytochrome P450 monooxygenases are known to contribute to the oxidation of many sulfur-containing pesticides. In contrast to other studies investigating endosulfan oxidation (13, 16, 17), the cytochrome P450 inhibitor PBO does not prevent the formation of endosulfate by this culture. While this does not exclude the possibility that oxidation is catalyzed by a cytochrome P450 monooxygenase, as PBO is not a universal inhibitor of these enzymes, it suggests that the reaction is being catalyzed by a type of oxidase that is different to that in the previous studies.

The culture is unable to utilize endosulfate as a sulfur source, and it accumulates as a terminal pathway product as a result of endosulfan oxidation. While an inability to transport the more polar compound into the cell may contribute to this, the accumulation of endosulfate in cultures provided with endosulfan suggests the absence of an enzyme capable of hydrolyzing the oxidized compound. The different oxidation states of the sulfur in endosulfan and endosulfate make it unlikely that the same enzyme will be capable of releasing the sulfur-containing moiety from both.

According to our proposed pathway, enzymatic hydrolysis of endosulfan forms the monoaldehyde and releases sulfite. Both isomers are substrates for this reaction, although the rapid oxidation of the alpha-endosulfan makes it difficult to estimate the rate of hydrolysis of this isomer. Because of the strong selection pressure imposed on the culture, release of sulfur from endosulfan does not have to be energetically favorable. The subsequent metabolism of non-sulfur-containing metabolites is presumably driven by energy demands. We propose that oxidative cyclization of the monoaldehyde leads to endosulfan hydroxyether, which is further metabolized to polar products. According to this pathway, the low levels of endodiol we detected in the later stages of culture growth are a result of chemical hydrolysis of the parent compound despite incubation in the slightly acidic medium. The pathway of metabolism we propose for our culture is substantially different from the degradation pathway in inocula from sandy-loam soil (21), tobacco leaf (6), and white-rot fungi (17), and the detoxification pathway of the Indian honey bee (23). The pathway proposed in these other studies involves a double hydration to produce endodiol and then a dehydration to produce endosulfan ether. While insects are under pressure to detoxify the insecticide, endosulfan is not toxic to plants (18), bacteria, or soil fungi (20); hence, the degradation observed in the other studies is most likely the result of cometabolism.

Although the monoaldehyde has not been reported previously, a putative dialdehyde metabolite of endosulfan has been characterized in white-rot fungi (17). In that case, however, it was inferred that the dialdehyde was derived by oxidation of initially formed endosulfan diol, with endosulfan hydroxyether as the intermediate.

From this study, we cannot predict the prevalence or relevance of this pathway in the soil environment. We were successful in enriching endosulfan-degrading organisms from only 1 out of 10 soil samples, but our method relies on bacteria being able to grow in the minimal media; hence, the number of culturable bacteria is severely restricted. We are also unable to predict if the endosulfan-degrading organism would utilize endosulfan as a sulfur source in the soil environment. The majority of the sulfur content of soils is found in an organic form, with over 95% present as sulfonates and sulfate esters. As a result, it is expected that soil bacteria will have numerous enzymes capable of releasing sulfur from organic compounds. Studies to date confirm this (for review, see reference 15).

The formation of the novel monoaldehyde product further supports our conclusion that the degradation we are observing is biological, as this compound has not been described as a product of chemical degradation (10). After approximately 25 rounds of subculturing, the culture metabolizes 50 μM endosulfan to undetectable levels in less than 4 days. This rate of degradation is significantly higher than those previously measured for bacteria (1, 11, 20, 21, 27), especially considering that chemical hydrolysis is not thought to be a significant contributing factor. Other studies have provided sufficient nutrient sources, and the biological transformation observed is presumably cometabolism. Our study differs from previous studies by the application of strong selection pressure on the culture to release the sulfur moiety from the insecticide, allowing us to enrich for the degradative activity. We are currently further characterizing the hydrolytic ability of this culture as a potential enzymatic bioremediating agent for endosulfan.