Kinetic and Molecular Differences in the Amplified and Non-amplified Esterases from Insecticide-resistant and Susceptible Culex quinquefasciatus Mosquitoes*

Two non-amplified esterases were purified from the insecticide-susceptible Pel SS strain of Culex quinquefasciatus . These were the two major esterase activity peaks in this strain. The two corresponding amplified carboxylesterases, Est (cid:97) 2 and Est (cid:98) 2, involved in organophosphate sequestration were purified from two resist- ant C. quinquefasciatus strains. The Pel SS esterases were significantly less reactive with the organophos- phates than those from the resistant strains. One of the Pel SS esterases was electrophoretically identical to am- plified Culex Est (cid:98) 1. However, it differed kinetically, and in its nucleotide and predicted amino acid sequences from the two characterized amplified Est (cid:98) 1s, it is classified as Est (cid:98) 1 3 . Restriction fragment analysis sug-gested Pel SS has only one Est (cid:97) and one Est (cid:98) gene, while the resistant Pel RR has both amplified and non-ampli- fied forms of Est (cid:97) and Est (cid:98) . The Eco RI fragments for both Pel SS esterases were distinct from those of the amplified Est (cid:97) 2 1 , Est (cid:98) 2 1 , or Est (cid:98) 1 1&2 . An esterase with the same size Eco RI fragment as Est (cid:98) 1 3 was also present in Pel RR. This and restriction

Classification of these esterases is based on their preferences for ␣or ␤-naphthyl acetate, their mobility on native polyacrylamide gel electrophoresis (PAGE), 1 and their nucleotide sequence (1). The overproduction of the Est␣2 1 and a series of Est␤ esterases is due to gene amplification (1,5,7,8). Identical EcoRI restriction fragment sizes of the amplified Est␤2 from resistant C. quinquefasciatus worldwide has been reported, in contrast to a high level of variability in Est␤ from insecticidesusceptible mosquitoes (9). A cDNA with 97% identity to Est␤2 1 has been cloned from an insecticide susceptible strain (Pel SS) of C. quinquefasciatus from Sri Lanka (8); however, the protein has not been characterized. After native starch or PAGE of homogenates of individual resistant Culex larvae of the amplified Est␣2/Est␤2 esterase phenotype, two electromorphs can be visualized. Under the same conditions, no esterase bands are visible in homogenates of susceptible insects. This has lead to the suggestion that the susceptible insects have null alleles for these loci (10).
Here, we report the purification and characterization of both an ␣and a ␤-naphthyl acetate-specific esterase from the Pel SS strain of mosquito and compare these at a biochemical and basic molecular level to the elevated esterases Est␣2 1 , Est␤1 2 , and Est␤2 1 . The role of the amplified esterases in resistance is sequestration, which is rapid binding and slow turnover of insecticides (11)(12)(13). 2 The current study elucidates the relative efficiencies of the purified esterases from a susceptible and two further resistant strains at binding the carbamates and biologically active oxon analogues of the organophosphorus insecticides. Characterization of the amplified and non-amplified esterases will facilitate future site-directed mutagenesis studies on the essential amino acid residues involved in the enzymeinsecticide interaction.

EXPERIMENTAL PROCEDURES
Four mosquito strains of C. quinquefasciatus were used. Pel was established from a large (Ͼ1,000) sample of larvae collected from Peliyagoda (Sri Lanka) in 1984. The population was heterogeneous for insecticide resistance, and both Pel SS and Pel RR were derived from this strain. The insecticide-susceptible Pel SS strain was obtained by selection and pooling of multiple single families for low esterase activity over three generations. The resistant Pel RR strain was selected from the Pel strain by mass larval selection with temephos (2, 3). The Muheza strain was collected from Tanzania in 1987 and maintained under intermittent chlorpyrifos selection. SPerm was collected from Jeddah (Saudi Arabia) in 1989. It was selected for 20 generations with permethrin and then intermittently with malathion and temephos (15). The three resistant strains both have the * This work was funded by a project grant from the Wellcome Trust. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  Field collections, each of Ͼ1000 larvae, were made in six suburbs of Colombo (Sri Lanka) in 1994. The areas included Peliyagoda, where the 1984 collection had been made. The frequency of the amplified esterases in these field populations was analyzed by nonspecific esterase assays of 100 -200 larvae (3). Batches of 100 larvae of the remaining insects were used for DNA extraction.
Esterases were copurified from the resistant and susceptible strains to ensure that any differences in the enzymes seen between this and previous studies were not due to minor variations in our purification and kinetic determination procedures over time.
Purification of Carboxylesterases-Batches of 4th instar larvae were homogenized and centrifuged at 15,000 ϫ g for 5 min, and the supernatant was taken. Carboxylesterases Est␣ and Est␤ were purified several times from numerous batches of larvae of each strain by sequential column chromatography and preparative electrophoresis as described previously (12,17). Final enzyme preparations were homogeneous as determined by SDS-PAGE. Crude homogenates for the insecticide interaction experiments were prepared in ice-cold 50 mM phosphate buffer (pH 7.4) with 5% (v/v) glycerol and 10 mM DL-dithiothreitol.
All specific activities are given in units/mg of protein. A unit corresponds to the hydrolysis of 1 mol of substrate in 1 min under the assay conditions used. Kinetic constants were determined from at least three experiments for each substrate or inhibitor using enzymes from several different purifications. For the inhibition kinetics, stopped time inhibition assays were performed using pNPA or p-nitrophenyl hexanoate (pNPH) as the substrate. Insecticide stock solutions were prepared in acetonitrile and diluted in phosphate buffer (pH 7.4). The purified enzyme was incubated with a series of concentrations of the test insecticide (acetonitrile concentration of the medium never exceeded 1% (v/v)) for fixed time intervals. Residual activity was determined from the rate of substrate hydrolysis.
Inhibitor concentrations were usually in large excess so that linear pseudo-first order kinetics were obtained. The bimolecular rate constants for the formation of acylated enzyme (k a ) were derived as described previously (18). If inhibitor concentration could not be maintained in large excess, k a values were determined in the presence of substrate (19). To minimize the effect of the reversible enzyme-substrate complex on the rate of acylation, the substrate concentration was maintained at a very low concentration so that the [S]/K m ratio was always less than 0.5 (18).
Genomic DNA Studies-DNA was extracted (8), precipitated with ethanol, resuspended in a small volume of TE, and stored at 4°C until used for Southern blotting. Pel RR Est␤2 1 and Est␣2 1 esterase cDNA fragments (1,8) were used as probes to determine the haplotype of the esterases in each laboratory strain or field collection of mosquitoes. Genomic DNA (10 g) was digested to completion with EcoRI, HindIII, or BamHI and separated on 0.8% (w/v) agarose gels. The DNA was transferred to charged nylon membranes (Amersham) and hybridized with 32 P-labeled probe (specific activity Ͼ 2 ϫ 10 6 cpm/g) at 65°C for 16 h in hybridization buffer (5 ϫ Denhardt's solution, 6 ϫ SSC, 0.1% (w/v) SDS, 0.1% (w/v) sodium pyrophosphate, 5% (w/v) polyethylene glycol 8000, 100 g/ml boiled sheared herring sperm DNA). The final membrane washes were at 65°C in 0.1 ϫ SSC and 0.1% (w/v) SDS for 20 min. Membranes were probed first with the Est␤2 1 cDNA then stripped and probed with Est␣2 1 cDNA.

RESULTS
On native PAGE gels, the susceptible Pel SS strain had no visible esterase bands, whereas all the resistant strains had the amplified esterase bands Est␣2 and Est␤2 (Fig. 1). The specific activity of Pel SS crude homogenate for the substrate pNPA was 0.02 units/mg, approximately 50-fold less than that routinely observed for resistant crude homogenates. For pNPH, the specific activities were much higher (0.14 units/mg for Pel SS crude homogenate), and this was used as the assaying substrate in subsequent esterase purifications from Pel SS, as the higher specific activity made peak detection simpler. However, the first purification was followed with both pNPA and pNPH, and the major peaks detected were the same with both substrates. Since hydrolysis of pNPH is linear for less than 1 min, pNPA was still the substrate of choice for all purifications from the resistant strains. After purification, only 5-10 g of each of the Est␣ and Est␤ enzymes were obtained from 10 -15 g, wet weight, of Pel SS larvae.
SDS-PAGE with purified esterases demonstrated that the molecular masses of the Est␣ and Est␤ enzymes (67 and 62 kDa, respectively) from the susceptible and both resistant strains were similar to those previously reported for Est␣2 1 and Est␤2 1 (11,12). The electrophoretic mobility on native PAGE of the Pel SS-purified Est␣ was reproducibly faster than that of the resistant Est␣2 1 (Fig. 2). In contrast, the Pel SS Est␤ had a slower mobility than that of the resistant Est␤2 1 (Fig. 3) but exactly the same mobility as the elevated C. quinquefasciatus Est␤1 (14).
For the substrates pNPA and pNPH, the K m values of the Pel SS Est␤ enzyme were 247.2 Ϯ 23. 3  Previous studies showed k a to be the most important constant in the interaction between the Culex esterases and the insecticides (12). 10-fold differences in k a between the Pel SS and the resistant Muheza and SPerm strains for chlorpyrifosoxon were seen for Est␣ (Table I). Efficiencies in binding chlorpyrifos-oxon and paraoxon of the Pel SS Est␤ enzyme were, respectively, 1000-and 100-fold less than those of the resistant strains (Table II). Thus, the non-amplified enzymes from Pel SS are less able to bind the insecticides than their respective amplified counterparts from various resistant strains. The ratios of the reaction rates with the insecticides for the crude homogenates of Pel RR and Pel SS were similar to those observed for the purified enzymes (Table III). As with the purified enzymes, the greatest differences were for chlorpyrifos-oxon and paraoxon, which suggests that this method may be valid as a crude means of detecting the level of interaction between these enzymes and insecticides.
The genomic EcoRI digests of the Pel SS and Pel RR strains and the field-collected insects sequentially probed for the Est␣ and Est␤ esterases are shown in Fig. 4. The amplified Est␣2 1 and Est␤2 1 bands in Pel RR are clearly distinguishable. The elevated Est␣2 1 esterase is seen as a 3.5-kb band in the Pel RR strain, along with a fainter (non-amplified) 4-kb band. In Pel SS, a single non-amplified 4.1-kb Est␣ band is present. In both the Pel RR and Pel SS strains, there is a fainter non-amplified 3.3-kb Est␤ band. The six field collections of insects had elevated esterase frequencies ranging from 0.12 to 0.84 on the basis of nonspecific esterase assays. When DNA digests from field-collected insects were probed, the amplified Est␣2 1 and Est␤2 1 bands had amplification levels broadly in agreement with the different frequencies of elevated esterase individuals in each of the populations. There were three distinct nonamplified bands of each of Est␣ and Est␤ bands in all six field collections analyzed with each of the restriction enzymes (Fig.  4). These bands were identical in all field collections. Blots of DNA digests from individual Pel SS and field-caught insects showed that the amplified bands were easily visible in these samples, but the non-amplified bands were not visible under our experimental conditions. The non-amplified bands from Pel SS were, however, visible in the pooled DNA from 4 to 5 insects. Hence, the non-amplified restriction bands seen in the field material must be present in more than one insect from the pool of 100.

DISCUSSION
Both Est␣ and Est␤ esterases are expressed in the insecticide-susceptible Pel SS strain of C. quinquefasciatus, although at a lower level than in resistant strains where these esterase genes are amplified (1,8). The two esterases from Pel SS now need classifying in line with the system adopted for mosquito esterases (1). Recently, two Est␤1 esterases were found to be different from each other in their nucleotide sequence and inferred amino acid sequence, demonstrating that multiple alleles of the Est␤ locus with identical electrophoretic mobility occur (8). The TEM-R strain of C. quinquefasciatus has an amplified Est␤1 1 EcoRI band of 2.1 kb (9), while the MRES Est␤1 2 has a doublet of bands of 3 and 3.2 kb (8). We have now shown that the Est␤ esterase in Pel SS electrophoretically would be classified as Est␤1, but it has an EcoRI band of 3.3 kb, which is distinct from those in both the MRES and TEM-R strains, and has 98% identity at the amino acid level with the amplified Est␤ (8), hence the Pel SS esterase should be classified as Est␤1 3 . The high stringency washing conditions for the Southern blots show that the non-amplified and amplified Est␣ esterases must be closely related. The Est␣ esterase in Pel SS has a different electrophoretic mobility to Est␣2 and Est␣1 (formerly A 1 ), hence it should be classified as Est␣3.
The Pel SS Est␤1 3 K m for pNPA was significantly higher than those of the resistant Est␤2 previously observed for Pel RR (140.8 Ϯ 5.24 M), Dar91 (85.41 Ϯ 3.94 M), and Tanga85 (90.11 Ϯ 6.49 M) (12,17). In contrast, the Pel SS Est␣3 and Est␤1 3 esterase pNPH K m values are not significantly different from those previously reported for Est␣2 1 and Est␤2 1 (17). The differences (up to 1000-fold) between the inhibition kinetic FIG. 3. A native PAGE of purified Est␤ type carboxylesterases from the susceptible Pel SS and five resistant strains of C. quinquefasciatus stained for esterase activity. Crude homogenate of Pel RR, which has elevated Est␣2 1 and Est␤2 1 , is shown for reference.

TABLE I
The kinetic constant k a (M Ϫ1 min Ϫ1 ) for insecticide interactions with the susceptible (Pel SS) Est␣3 carboxylesterases and the elevated Est␣2 1 esterase from five resistant strains of C. quinquefasciatus The data are means Ϯ S.D. In the same row, different superscript letters indicate a significant difference (p Ͻ 0.05). Two letters indicate values that are not significantly different from either single letter alone. k a values for the enzymes purified from Pel RR, Dar91, and Tanga85 were reported previously (17) and given here for comparison. 10 constants for the enzymes from the susceptible and resistant strains for the oxon analogues of the organophosphorus insecticides are far greater than those observed between resistant strains. The k a values for the purified esterases from the resistant strains used in this study were only slightly different from those determined previously for other resistant strains (see Tables II and III) (17). Thus, the high level of variability between the k a values for the susceptible and all the resistant strains cannot be accounted for by minor variations in conditions and experimental techniques between the studies, and we conclude that all the amplified esterases are more efficient at binding the insecticides than their non-amplified Pel SS counterparts. This superiority of insecticide binding suggests that there has been a positive insecticide selection pressure to maintain amplification of favorable resistant alleles. The kinetic differences between the purified amplified Est␣2 and Est␤2 enzymes from different strains were previously suggested to be due to either allelic mixtures of the esterases or different single allelic forms of both Est␣2 and Est␤2 in each of the resistant strains (12,13,17). However, the Est␤2 EcoRI digest pattern, unlike that for Est␤1, does not appear to vary (9), hence the observed kinetic differences may not be reflected at the DNA level. Our current data show that Pel RR has the invariant Est␤2 1 EcoRI band and that it apparently contains an allelic mixture of both Est␣ and Est␤, with a minor nonamplified band of each of the esterase types being present along with the prominent amplified band. Thus, in this and previous biochemical studies, all the purified Est␣ and Est␤ esterases from the resistant strains may have been mixtures of amplified and non-amplified alleles, as the purification conditions used may not have separated these minor variants from each other. Such allelic mixtures would not have been detected, as both the Pel SS Est␤1 3 and the Pel RR Est␤2 1 esterase genes code for proteins of 540 amino acids, whose predicted molecular weights would not allow their separation by SDS-PAGE (8). This may explain the reported differences between purified esterases from different resistant strains, as variability could arise from different proportions of the amplified and non-amplified alleles, given the big differences in k a values between the amplified and non-amplified alleles of both the Est␣ and Est␤ esterases. Alternatively, different non-amplified alleles could be mixed with an identical amplified allele to give a similar result. Our results also suggest that there is limited variability in the non-amplified Est␣ and Est␤ alleles of the Sri Lankan field population, as only single non-amplified Est␣ and Est␤ EcoRI bands were apparent from DNA obtained from mass homogenates of Pel SS, suggesting that this colony contains only a single Est␣ and a single Est␤ esterase, even though the strain originated from numerous pooled single families selected from a large field collection of insects. Several years of laboratory colonization could have resulted in the loss of much of the variability. However, if the latter is true, it is surprising that the non-amplified Est␤ allele of Pel RR has an identical EcoRI fragment to the Est␤1 3 allele of Pel SS, given that the two strains are maintained in separate insectaries. Similarly, restriction digest analysis of recent field collections of C. quinquefasciatus also suggest that the variability of the non-amplified Est␣ and Est␤ alleles in the Sri Lankan field population is limited, with three common restriction fragments, as well as the amplified EcoRI fragment, occurring with each of three different restriction enzymes in six independent field collections from the Colombo area for each esterase.   1 esterases from five resistant strains of C. quinquefasciatus The data are means Ϯ S.D. In the same row, different superscript letters indicate a significant difference (p Ͻ 0.05). k a values for the enzymes purified from Pel RR, Dar91, and Tanga85 were reported previously (17) and are given here for comparison.