The binary toxin is the major active component of Bacillus sphaericus, a microbial larvicide used for controlling some vector mosquito-borne diseases. B. sphaericus resistance has been reported in many part of the world, leading to a growing concern for the usefulness of this environmental friendly insecticide. Here we characterize a novel mechanism of resistance to the binary toxin in a natural population of the West Nile virus vector, Culex pipiens. We show that the insertion of a transposable element-like DNA into the coding sequence of the midgut toxin receptor induces a new mRNA splicing event, unmasking cryptic donor and acceptor sites located in the host gene. The creation of the new intron causes the expression of an altered membrane protein, which is incapable of interacting with the toxin, thus providing the host mosquito with an advantageous phenotype. As a large portion of insect genomes is composed of transposable elements or transposable elements-related sequences, this new mechanism may be of general importance to appreciate their significance as potent agents for insect resistance to the microbial insecticides.
Mosquitoes transmit serious human diseases, such as malaria, yellow fever, West Nile fever and dengue. Resistance to all the major classes of chemical insecticides has now been recorded in several mosquito species, threatening to undermine the mosquito vector control programmes (Zaim and Guillet, 2002). To obviate resistance and the risks for human health and the environment caused by broad-spectrum conventional pesticides, new strategies for insect control have been developed based on the environmentally friendly biopesticides derived from the bacteria Bacillus sphaericus and Bacillus thuringiensis. Highly toxic strains of B. sphaericus Neide have been proven to effectively control mosquito-borne diseases of the genera Culex and Anopheles (Karch et al., 1992; Mulla, 1994; Becker, 2000). The toxic properties of B. sphaericus are mainly attributed to parasporal protein crystals produced under unfavourable conditions. Following ingestion by the susceptible larvae, crystalline inclusions are solubilized in the midgut, and two protoxins, BinA and BinB, are released. Both proteins are activated by larval proteases and act together as a binary toxin (Bin). Bin interacts with high affinity with Cpm1, an alpha-glucosidase present on the brush border membrane surface of the midgut epithelium of the West Nile virus vector larvae, Culex pipiens (Silva-Filha et al., 1999; Darboux et al., 2001). The toxicity of Bin correlates directly with its interaction with the cell membrane. We demonstrated previously that the resistant phenotype of the laboratory-selected Culex strain GEO was due to a nonsense mutation causing the premature termination of translation, eliminating the glycosylphosphatidylinositol (GPI) anchorage of the receptor to the membrane (Darboux et al., 2002).
High levels of B. sphaericus resistance have been reported in several field populations of Culex isolated from Brazil, France, India and China after intensive exposure (Rao et al., 1995; Yuan et al., 2000; Chevillon et al., 2001). Two different mechanisms have been observed in resistant populations from the French Mediterranean Coast (Chevillon et al., 2001). In the SPHAE strain, identified in 1994, resistance does not involve the toxin-receptor binding step (Nielsen-LeRoux et al., 1997) and might be rather correlated with intracellular activity of the Bin toxin after its binding to the midgut epithelium of the larvae. In contrast, the high level of resistance displayed by the BP strain collected in 1997 was related to the lack of toxin binding on the epithelial cell membrane (Nielsen-LeRoux et al., 2002). Hence the absence of a functional receptor for the Bin toxin may be a resistance mechanism developed by laboratory-selected strains, such as GEO and field-collected strains, such as BP.
Here we report the mechanism of resistance for the strain BP. We show that insertion of a transposable element (TE)-like sequence in the coding region of cpm1 promotes an unexpected form of genetic variation: it changes the splicing pattern of the host mRNA, leading to the creation of a new intron in the host gene. This molecular event leads to the production of a shortened receptor, which is incapable of interacting with the microbial toxin, allowing the insect to survive.
Expression of Cpm1 in midgut epithelial cells from BP larvae
We did not detect the receptor Cpm1 in BBMF from BP larvae by Western blot analysis (Fig. 1A). This agrees with previous results showing that the toxin does not bind to BBMF (Nielsen-LeRoux et al., 2002). Expression of cpm1 transcripts was measured by reverse transcription polymerase chain reactions (RT-PCR). Amplification of the complete coding sequence of cpm1 yielded the expected fragment of 1740 bp and an additional fragment around 1540 bp, specific to the resistant sample (Fig. 1B). Sequence analysis revealed that the 1740 bp fragment, we called cpm1BP, differs from the cpm1S-LAB cDNA by several replacements that lead to five amino acid substitutions and a nonsense mutation Gln396Stop (Fig. 1B). All these mutations are distinct from those found in cpm1GEO (Darboux et al., 2002). The 1540 bp cDNA fragment (named cpm1BP-del) has a 198 bp internal deletion that does not cause frameshift in the open reading frame (ORF). If translated, this mRNA would encode a protein 66 amino acids smaller than the Cpm1S-LAB protein, but retaining the predicted GPI-anchor sequence in the C-terminus. The deduced amino acid sequence of Cpm1BP-del also contains two substitutions, Arg307Lys and Ala503Val, as compared with the Cpm1S-LAB protein sequence.
Characterization of two resistant cpm1 alleles in the BP larvae
We amplified the complete coding sequence of cpm1 from the genomic DNA of susceptible and resistant mosquitoes. A fragment of 1850 bp was amplified from both strains. Sequence analysis revealed the presence of two short introns. Intron 1 was 50 bp long for both strains, whereas the length of intron 2 was 57 bp in S-LAB mosquitoes and 56 bp in BP mosquitoes. The coding sequence of the 1850 bp fragment produced from BP genomic DNA was identical to the cpm1BP cDNA. An additional PCR product of ∼3.3 kb was generated from the BP sample. Sequence analysis revealed a coding sequence displaying two single-nucleotide substitutions resulting in Arg307Lys and Ala503Val replacements and the presence of a 1451 bp insert within exon 2. The foreign fragment is flanked by a target site duplication (Fig. 2A) characteristic of the insertion of a TE. Others structural features reminiscent of class II transposons (Tu and Coates, 2004) are also observed: the presence of 41 and 38 bp imperfect terminal inverted repeats (TIRs) at each end (Fig. 2B), high A+T richness (70.8%) and a potential to form stable secondary structures. Several subterminal repeats are also present (Fig. 2B). The foreign element, we called TE-BP, lacks any ORF for transposase, so its characteristics strongly suggest that it is a non-autonomous TE-related sequence introduced into the cpm1BP-del allele. We did not detect significant sequence similarity with eukaryotic TEs described previously, suggesting that it may constitute a new endogenous TE in C. pipiens.
Segregation analysis of the cpm1BP-del allele in the BP population
The presence of cpm1BP and cpm1BP-del suggests that either the BP strain is genetically heterogeneous or that the cpm1 gene has been duplicated. Thus, we used primers flanking the insertion of TE-BP to amplify genomic DNA from individual BP mosquitoes. The amplified product from cpm1BP-del was larger than that from cpm1BP, facilitating the genotyping of each individual (Fig. 2C). In 108 mosquitoes, we show that the BP strain was heterogeneous: 50% contained both cpm1BP and cpm1BP-del, whereas 25% were homozygous for one allele or the other. A previous report showed that dose–response regression lines data on the backcross offspring are consistent with a single-locus, two-alleles (susceptible, resistant) model (Chevillon et al., 2001). If the genotypic structure of the BP strain remained stable over time, this would suggest that both cpm1BP and cpm1BP-del alleles provide BP mosquitoes with a similar resistant phenotype under these bioassay conditions.
Functional analysis of cpm1BP and cpm1BP-del alleles
We expressed resistant Cpm1 proteins in Sf9 cells to analyse the functional consequences of the mutations found in both alleles. With the nonsense mutation, cpm1BP must produce a receptor truncated by 184 amino acids. The missing portion includes GPI-anchoring signals. Therefore, Cpm1BP would not be a functional receptor because it would be secreted into the extracellular space, similar to what was observed for the GEO strain (Darboux et al., 2002).
Unlike Cpm1BP, we detected Cpm1BP-del in the membrane-enriched fractions (Fig. 3A). However, homologous competition experiments did not reveal an interaction between radiolabelled Bin and membrane fractions from Sf9 producing Cpm1BP-del or Cpm1BP (Fig. 3B). Even after we reintroduced the wild-type Lys-307 and Val-503 codons by site-directed mutagenesis, the protein (named del-KV) did not interact with BinB, the binding component of Bin, as demonstrated by a GST pull-down assay (Fig. 3C).
We have shown previously that a secreted form of Cpm1S-LAB can be expressed in a enzymatically active form in Sf9 cells (Darboux et al., 2002). We thus used the expression system to assess the in vitro maltase activity of the resistant Cpm1 proteins by measuring the release of p-nitrophenol by p-nitrophenyl α-d-glucopyranoside. We did not detect significant alpha-glucosidase activity on Cpm1BP, Cpm1BP-del or del-KV (Fig. 4).
In this study, we have shown that two B. sphaericus-resistant alleles coexist in a field-evolved population of C. pipiens. Both alleles encoded Cpm1 receptors, which are incapable of interacting with the Bin toxin, resulting in the lack of toxicity of B. sphaericus towards the BP mosquitoes.
In the cpm1BP allele, a nonsense mutation causes the loss of the C-terminal domain required for a proper anchoring of the receptor to the cell surface, and thus disrupts a crucial step in the toxic properties of B. sphaericus. The cpm1BP-del allele, characterized by a transposon element-like sequence inserted in the exon 2, produced a shortened cpm1 mRNA. The boundaries of the deletion found in the corresponding cDNA agree with the classical 5′[(C/A)AGGT(A/G)AGT] and 3′ (TCnNCTAGGT) splice site junctions (Fig. 2A). A World Wide Web service (http://www.fruitfly.org/seq_tools/spliceAbst.html) predicted scores of 0.41 and 0.65 for the 5′- and 3′ cryptic splice sites, respectively, in the cpm1BP-del genomic sequence. This observation suggests that the shortened mRNA is generated by aberrant splicing, unmasking putative cryptic donor and acceptor splice sites. Alterations in the splicing process induced by TE insertion have already been described (Wessler et al., 1987; Cox et al., 2000), but the molecular mechanism is unknown. The conserved consensus sequences for the 5′-and 3′ splice sites are important for splice site selection, however, other factors are also involved. These include intron and exon length and pre-mRNA structure. Possibly, insertion of the TE modifies the environment of these factors, interfering with normal recognition of the splice sites.
Functional analysis of Cpm1BP-del and del-KV showed that the in-frame deletion in exon 2 was sufficient to prevent the binding of Bin to Cpm1BP-del. It did this either by directly affecting the binding site or indirectly causing conformational changes. In contrast to the mutant del-KV, we only weakly detected Cpm1BP-del in the high-level expression system of Sf9 cells. We can thus presume that Arg307Lys and Ala503Val substitutions have additional deleterious effects on the structure and/or stability of the protein. This would explain why Cpm1BP-del is not observed in BBMF from BP larvae (Fig. 1A).
The association of a TE-like sequence with the cpm1BP allele clearly illustrates how TEs can provide beneficial genetic responses to insects so that they become resistant to pesticides (ffrench-Constant et al., 2006). Recent studies report an Accord TE insertion in the 5′ regulatory region of the P450 gene, Cyp6g1, associated with transcriptional upregulation of the gene and/or resistance to the organochlorine pesticide DDT [1,1-Bis(4-chlorophenyl]-2,2,2-trichloroethane) (Daborn et al., 2002; Schlenke and Begun, 2004). This adaptive phenotypic change has apparently been fixed in unrelated Drosophila populations. Another study showed a correlation between the insertion of a Doc1420 retrotransposon and resistance to organophosphates in Drosophila (Aminetzach et al., 2005). In that case, the presence of Doc1420 in the coding sequence of a gene disrupts the normal RNA processing and lead to the expression of multiple hybrid [host gene-Doc1420] transcripts, which are probably non-functional. The potential adaptive significance of TEs in biopesticide resistance has been proven in one case so far: an in-frame stop codon within a retrotransposon sequence resulted in the premature termination of translation of the B. thuringiensis toxin receptor. This mutational event potentially generates a protein, which lacks both its membrane anchor and toxin-binding domain, leading to B. thuringiensis resistance in a laboratory-selected strain of a major cotton pest, Heliothis virescens (Gahan et al., 2001). The data presented in this study are a novel example of insecticide resistance associated to a transposon insertion. A further characterization of the TE-BP insertion strain is needed to better understand the biological significance of this insertion on the natural Culex population. For example, it will be interesting to compare the B. sphaericus-resistance levels between Culex strains homozygous for cpm1BP or cpm1BP-del, and measure the correlation of frequency with resistance phenotype. Nevertheless, our data show that the remodelling of the pre-mRNA splicing pattern by a TE-like sequence insertion can contribute to host genome adaptation by generating a new protein, which cannot bind the microbial toxin, conferring biopesticide resistance. A substantial portion of insect genomes is composed of TEs or TE-related sequences (e.g. 16% of the euchromatin in Anopheles gambiae) (Holt et al., 2002). We may thus expect that other beneficial adaptive changes mediated by transposons will be revealed.
Both Cpm1BP and Cpm1BP-del proteins had no detectable alpha-glucosidase activity. This indicates that the mutations, including the splicing of the new intron from cpm1BP-del pre-mRNA, abolish the physiological function of the maltase. BP mosquitoes can be maintained in the laboratory without obvious defects in growth and development. Under laboratory conditions, B. sphaericus resistance due to the loss of Cpm1 activity in a Culex quinquefasciatus strain entails a fitness cost (de Oliveira et al., 2003; Romão et al., 2006). It is unknown whether the loss of digestive enzyme activity of Cpm1 is disadvantageous to the fitness of mosquitoes in the field. Cpm1 is a member of the alpha-amylase multigene family. Mechanisms involving the overlapping functions of the maltases may be activated to compensate for the absence of Cpm1 in BP mosquitoes. This phenomenon has already been observed in herbivorous insects fed an artificial diet with high endogenous proteinase inhibitor levels (Bown et al., 1997). We have preliminary data indicating that there are several alpha-glucosidases highly similar to Cpm1 in the midgut of Culex larvae. Their characterization should reveal redundancies of the enzymes and possible compensatory mechanisms in the resistant populations, an aspect which has important implications for resistance management.
Resistance to the B. sphaericus binary toxin has been reported in many parts of the world leading to a growing concern for the usefulness of this environmental friendly microbial insecticide. B. sphaericus resistance in treated mosquito populations in Southern France was recently shown to involve other factors than cpm1BP and cpm1BP-del alleles. Such a factor is present in the SPHAE population, where no change in the binding step as compared with the susceptible strain, has been detected (Nielsen-LeRoux et al., 1995). The existence of various resistance mechanisms represents a serious threat to the use of B. sphaericus. However, different strategies have been suggested to counteract the development of such field resistance. According to the lack of cross-resistance between B. sphaericus and Bacillus thuringiensis ssp. israelensis (Bti) toxins, the utilization of both microbial pesticides in rotation or mixture have been shown to present a high potential to restore B. sphaericus susceptibility in B. sphaericus-resistant mosquitoes (Rao et al., 1995; Rodcharoen and Mulla, 1996). Another approach for B. sphaericus resistance management consists in associating toxins with different mode of action to the binary toxin in recombinant strains of B. sphaericus or Bti using genetic engineering (Servant et al., 1999; Wirth et al., 2000a,b). Altogether, these measures should extend the usefulness life of B. sphaericus as an environmentally safe and effective microbial agent to control Culex mosquitoes.
Culex pipiens eggs were kindly provided by Dr Michel Raymond (ISEM, Montpellier, France). Two strains were used: S-LAB, a susceptible laboratory strain (Georghiou et al., 1966), and BP, a resistant field population (i.e. more resistant to Bin toxin than the S-LAB strain by a factor of 6000) collected in southern France in 1997 (Chevillon et al., 2001). Larvae were raised on a diet of ground cat food.
Western blot analysis
BBMF extracts from the midgut of fourth-instar larvae and Sf9 cell membrane-enriched fractions were prepared as described previously (Silva-Filha et al., 1999; Darboux et al., 2002). The samples were separated by 9% SDS-PAGE and blotted onto a PVDF membrane with a Multiphor II semidry electroblotting apparatus (GE Healthcare). The membranes were probed with anti-Cpm1 rat polyclonal antibodies (Darboux et al., 2002) or horseradish peroxidase-conjugated anti-V5 antibody (Invitrogen). An ECL Western blotting detection kit (GE Healthcare) was used to visualize the signal.
Reverse transcription PCR analysis
Total RNA was isolated from the midgut of C. pipiens fourth-instar larvae with the TRIzol reagent (Invitrogen). First-strand cDNA was synthesized from 3 μg of total RNA with Superscript II reverse transcriptase (Life Technologies). The entire ORF of cpm1 was amplified by PCR with the primers CPC-K (KpnI site underlined): (5′-CGGGGTACCCCGATGCGACCGCTGGGAGC-3′; nucleotides 1–17) and CPT-X (XbaI site underlined): (5′-CTAGTCTAGATTCACGAAGATATACCTGGC-3′; nucleotides 1720–1740). The RT-PCR programme consisted of 2 min at 94°C and then 30 cycles of (30 s at 94°C, 1 min at 53°C and 1 min at 72°C), followed by a 5 min extension. The amplification products were electrophoresed on an agarose gel, gel-purified individually using the Gel Extraction kit (Qiagen) and directly sequenced.
Polymerase chain reaction amplification of genomic DNA
Genomic DNA isolated from mosquitoes as described previously (Martinez-Torres et al., 1998) was used as a template for the primers CPC-K and CPT-X. PCR was carried out with the proofreading Advantage KlenTaq LA polymerase mix (Clontech) according to the manufacturer's recommendations. The cycling parameters consisted of an initial denaturation at 94°C for 2 min, then 94°C for 30 s, 50°C for 30 s, 70°C for 3 min for 26 cycles followed by a 10 min final extension at 70°C. The amplified fragments were cloned into the pCR2.1 vector (Invitrogen). Four independent clones corresponding to each resistant and susceptible allele were sequenced. Genomic DNA was prepared from 108 BP individuals for genotyping. PCR was performed with a pair of primers flanking the insertion: BP1 (forward), 5′-CCGTTGGGCTGAAGAGTTCAATC-3′ and BP2 (reverse), 5′-CTCGCATCTTCTGCGGCAGATTC-3′. The amplification conditions were as follows: 3 min at 94°C, 30 cycles of 94°C for 30 s, 30 s at 56°C and 3 min at 72°C, followed by a 10 min final extension at 72°C. PCR products were separated on 1% agarose gels and visualized under ultra violet light after staining with ethidium bromide.
Determination of the 5′- and 3′-ends of the cDNA by rapid amplification of cDNA ends (RACE)
The 5′- and 3′-ends of the cpm1BP and cpm1BP-del cDNA were obtained with the Marathon cDNA Amplification kit (CLONTECH). PCR products were cloned directly into the pCR2.1 vector (Invitrogen). Two independent clones for each RACE were sequenced.
All constructs were generated by standard cloning procedures and verified by restriction enzyme analysis and DNA sequencing. Expression vector for Cpm1BP was produced by amplifying the corresponding cDNA with the primers PS2 (5′-CGGGGTACCCCGATGCGAACGCTGGGAGC-3′, nucleotides 1–17, KpnI site underlined) and CPT-X and subcloning the cDNA into the KpnI–XbaI site of the pIZ vector (Invitrogen). Expression vector for Cpm1BP-del was produced by amplifying the corresponding cDNA with the primers CPC-K and CPT-X and subcloning into pIZ. For GST pull-down assays, cpm1BP and cpm1BP-del were expressed as secreted recombinant proteins. For that purpose, cpm1BP cDNA was amplified with the primer pairs PS2 and Gln396Xba (5′-CTCCTCTAGATGCATGGCTACCTCTTCACC-3′, reverse, nucleotides 1168–1197, XbaI site underlined) and cpm1BP-del cDNA was amplified with CPC-K and Leu-X (Darboux et al., 2002). Reverse primers were designed so that the amplified fragment could be inserted in frame into the upstream region of the V5 epitope and 6xHis tag of pIZ for detection and purification. Expression vectors for the susceptible receptor Cpm1 and its secreted form were described previously (Darboux et al., 2002). The wild-type Lys-307 and Val-503 codons were reintroduced into the cpm1BP-del cDNA by site-directed mutagenesis with the QuickChange II site-directed mutagenesis kit (Stratagene).
Cell culture and transfection
Sf9 cells maintained in complete Grace's insect media (Invitrogen) at 70% density in 25-cm2 plates were transfected with 5 μg of each plasmid with the jetPEI (QBiogen) reagent according to the supplier's instructions. Forty-eight hours after transfection, 400 μg ml−1 of Zeocin was added to the culture medium to select cells expressing the Zeocin resistance gene. Three weeks later, Zeocin-resistant cells were progressively transferred into TNM-FH protein-free medium (Invitrogen) for maintenance.
GST pull-down assays
The plasmid construct encoding GST-BinB (Oei et al., 1992) was kindly provided by C. Berry (Cardiff University, Wales, UK). The GST-BinB protein was produced in Escherichia coli BL21(DE3)pLysS and induced by 0.4 mM isopropyl-β-d-thiogalactopyranoside at 37°C for 4 h. The cell pellets were resuspended in PBST buffer (PBS with 1% Triton X-100, 2 mM EDTA, 0.1% β-mercaptoethanol) submitted to multiple freeze/thaw cycles and then sonicated. After centrifugation, the GST-fusion protein was isolated from the supernatant by glutathione–Sepharose affinity chromatography. Sf9 cell culture media containing recombinant proteins were collected after 7 days of culture and centrifuged at low speed to remove cells. An aliquot (1 ml for the Cpm1BP-del and del-KV samples, and 0.2 ml for the Cpm1S-LAB sample) was incubated for 2 h at 4°C under continuous agitation with 20 μl of GST-BinB immobilized on glutathione-Sepharose 4B beads (GE Healthcare). The beads were then collected by low speed centrifugation and washed six times with 1 ml of 20 mM Tris-HCl containing 0.5% NP40, 300 mM NaCl and 0.1 mM EDTA. Bound proteins were released by the addition of 2 × Laemmli sample buffer followed by incubation for 5 min at 95°C. The proteins were detected by Western blot analysis with anti-V5 antibody (Invitrogen) and an enhanced chemiluminescence system (ECL, GE Healthcare).
Homologous competition was assessed as described previously (Darboux et al., 2002), with 10 nm 125I-Bin, 1–1000 nm unlabelled toxin and 25 μg of cell membrane proteins at 25°C for 16 h.
Nucleotide sequence accession numbers
Sequences for genomic cpm1S-LAB, cpm1BP, cpm1BP-del alleles and TE-BP have been submitted to GenBankTM under Accession numbers EF061759, EF061760, EF061761 and EF054859 respectively.
This work was supported by grants from the MENRT (Impact des OGM) and the INRA/CNRS program (Impact des biotechnologies dans les agroécosystèmes). We thank René Feyereisen and Thomas Guillemaud for helpful discussions and suggestions. We are grateful to Zhijian Tu for his homology searches for the foreign DNA. We thank Colin Berry for his generous gift of plasmid encoding GST-BinB and Elodie Marchi for the maintenance of the Sf9 cells.