Knockdown resistance to DDT and pyrethroids: from target-site mutations to molecular modelling



Naturally derived insecticides such as pyrethrum and man-made insecticides such as DDT and the synthetic pyrethroids act on the voltage-gated sodium channel proteins found in insect nerve-cell membranes. The correct functioning of these channels is essential for the normal transmission of nerve impulses, and this process is disrupted by binding of the insecticides, leading to paralysis and eventual death. Some insect pest populations have evolved modifications of the sodium channel protein that inhibit the binding of the insecticide and result in the insect developing resistance. This perspective outlines the current understanding of the molecular processes underlying target-site resistance to these insecticides (termed kdr and super-kdr), and how this knowledge may in future contribute to the design of novel insecticidal compounds. Copyright © 2008 Society of Chemical Industry


The use of chemical pest control is thought to date back at least 4500 years. However, it was only with the industrialisation and mechanisation of agriculture in the eighteenth and nineteenth centuries, and the large-scale production of naturally derived insecticides such as pyrethrum1 (extracted from Chrysanthemum cinerariifolium Vis. flowers), that chemical pest control became the method of choice. In the twentieth century the production of synthetic insecticides such as DDT2 (introduced in the 1940s) and the synthetic pyrethroids3 (first commercialised in the 1970s) further boosted this trend.

Research on the basis of the toxicity of DDT (dichloro-diphenyl-trichloroethane) and the pyrethroid insecticides has shown that both are potent neurotoxicants that interfere with nerve membrane function by interacting with a common target within nerve membranes.4, 5 Consequently, it can be seen in retrospect (although it was not obvious at the time when the pyrethroids were first commercialised) that the development of resistance among insect pests to DDT was also likely to confer resistance to pyrethroids. The common resistance mechanism that did emerge, referred to as ‘knockdown resistance’ (or kdr), is due to a 10–20-fold reduction in target-site sensitivity to these compounds, and is probably the most important of several resistance mechanisms reported in insect species.6 The genes that confer kdr are ‘recessive’, which means that the trait will only be expressed in homozygous individuals (i.e. two identical copies of the same recessive gene affecting the trait need to be present). Such recessive genes can persist at low levels in a population without being detected, even if no selection pressure is present. Thus, although DDT had not been used for a number of years, the subsequent introduction and intensive use of synthetic pyrethroids was able to select for kdr and reduce their efficacy against several pest species within a very short time period. The first cases of such cross-resistance were documented in the mid-1970s in houseflies in Europe,7 and by 1980 over 200 insect pests were documented as having resistance to DDT,8 many of which have since also been reported as showing cross-resistance to the synthetic pyrethroids. In certain species, kdr is also accompanied with additional, enhanced resistance trait(s), designated super-kdr, capable of conferring much greater (100-fold) resistance to pyrethroids.9


DDT, the pyrethrins and the synthetic pyrethroids all act on the voltage-gated sodium channel (VGSC) proteins10 found in nerve-cell membranes. The VGSCs are integral membrane proteins (Fig. 1) responsible for the conduction of sodium ions, and the channels open and close in response to changes in membrane potential. The correct functioning of these channels is essential for normal transmission of nerve impulses, and, when this process is disrupted by binding of insecticides, paralysis and eventual death occur. Various inherited point mutations in the VGSC genes, which alter the amino acid sequence of the VGSC protein, have been shown to be responsible for causing DDT and pyrethroid resistance in a wide range of agricultural pests and disease vectors (recently reviewed elsewhere11, 12). Two of these amino acid substitutions, L1014F (in domain IIS6 of the channel) and M918T (in the IIS4–S5 channel linker), were originally identified in pyrethroid-resistant housefly strains as being associated with high levels of resistance.13 Subsequent in vitro expression studies of mutated VGSCs confirmed these changes as being responsible for kdr and super-kdr phenotypes respectively.14 Since the initial classification of the L1014F and M918T substitutions, other resistance-associated mutations have also been characterised. One of the most effective single nucleotide substitutions, T929I (housefly numbering) in domain IIS5, was found along with the leucine to phenylalanine (kdr) mutation in a highly resistant diamondback moth (Plutella xylostella L.) strain.15 At the same position, a T to C change and/or a T to V change (both requiring a double nucleotide change) have been detected in highly resistant thrips (Frankliniella occidentalis Perg.)16 and cat fleas (Ctenocephalides felis Bche.).17 A T to I mutation is also found in pyrethroid-resistant head lice (Pediculus humanus capitis Deg.), but in this case it is accompanied with a second mutation, L932F.18 L925I (housefly numbering) has recently been described in whitefly19 (Bemisia tabaci Genn.) and is associated with a decrease in channel affinity intermediate between kdr and super-kdr.20 F1538I, identified in cattle ticks (Boophilus microplus Can.), has also been demonstrated to be a highly effective resistance locus for the synthetic pyrethroids.21 Although multiple amino acid residue changes have been identified, how the majority of these mutations reduce pyrethroid sensitivity remains largely speculative.

Figure 1.

A schematic representation of a voltage-gated sodium channel. The pore-forming α-subunit consists of a single polypeptide chain with four internally homologous domains (I–IV), each having six transmembrane helices S1–S6. The domains assemble to form a central aqueous pore, lined by the S5 and S6 helices and the S5–S6 linkers (P-loops). The S1–S4 helices of each domain assemble to form four physically discrete voltage sensors. The voltage dependence of channel activation is thought to derive from the movements of the four positively charged (+) S4 segments. The identity and location of some of the mutations associated with resistance (kdr and super-kdr) are shown (●), with residues numbered according to the sequence of the housefly (Musca domestica) voltage-gated sodium channel (EMBL accession: X96668).

The authors' recent research has focused on the development of a three-dimensional model of the pore region of the housefly sodium channel22 (Fig. 2), and has facilitated the prediction (through ligand docking simulations) of a putative pyrethroid-binding pocket on the VGSC. As can be seen in Fig. 3, the potential binding ‘crevice’ that transiently exists in the channel's open state encompasses the resistance-associated residues M918, L925, T929 and L932 in the IIS4–S5 linker and IIS5 helix region of the channel and F1538 on the IIIS6 helix. Binding of DDT or pyrethroids to this open channel conformation is predicted to restrict further movement of the S5 and S6 pore-forming helices, the insecticides acting as a molecular ‘wedge’ between IIS5 and IIIS6 that prevents movement towards the channel's closed state. Interaction of pyrethroids (but not DDT23) with M918 would enhance this effect by stabilising the IIS4–S5 linker in the open state, hence making the channel's voltage sensor(s) much less responsive to membrane repolarisation. Binding of insecticide thus locks the channel pore in the open configuration, leading to an induction of repetitive activity owing to a prolongation of the transient increase in sodium permeability of the nerve membrane associated with excitation. The model predicts that selective mutations that disrupt molecular interactions of the insecticide with M918 or T929 would give the highest levels of resistance, and this is consistent with experimental findings.24 Furthermore, T929 has also been demonstrated to be a key resistance-associated locus for both DDT and pyrethroids, with the T929I mutation rendering the channel highly insensitive to all compounds (DDT, type-I and type-II pyrethroids) that have been tested so far.20 This may be due to the natural proclivity of this threonine residue to hydrogen bond (both as donor and acceptor) with the ester linkage of pyrethroids. Introduction of an α-cyano group (as in type-II pyrethroids) is likely to extend the hydrogen-bonding interactions of T929 with the pyrethroid ester linkage, and this would tightly ‘lock’ the central portion of type-II compounds into the binding cavity, hence increasing the insecticide's neurotoxicity. The VGSC model22 (and previous electrophysiological studies of the M918T substitution24) further suggests that interaction with the M918 residue on the domain IIS4–S5 linker is critical for the high-level toxicity of pyrethroid insecticides, especially for the highly potent type-II (cyano-phenoxy-benzyl) compounds such as deltamethrin (Fig. 4). These observations would explain why type-II pyrethroids inhibit the deactivation of sodium channels to a greater extent than do type-I pyrethroids.24 Other resistance loci, L925I, L932F and F1538I, are also found in close proximity to bound pyrethroids in the model. However, it is possible that some of these residues do not physically interact with the pyrethroids, but may subtly alter the binding site conformation in order to exclude them. The overall stereochemical structure of pyrethroid insecticides is also known to influence their toxicity. Small changes in substituents and stereochemistry are sufficient to produce compounds differing in their insecticidal potency, spectrum of activity and mammalian toxicology.25 The availability of a three-dimensional model of the binding site will facilitate interpretation of why these structural variations are so important; for example, halogenated compounds (e.g. fenfluthrin and flumethrin) are exceptionally active against many species of insect. The VGSC model suggests that this may be due to enhanced interaction of halogenated compounds with aromatic residues on the domain IIIS6 helix.20

Figure 2.

Model of the activated-state voltage-gated sodium channel from a housefly. The four voltage sensor domains are shown in surface representation (grey). The S4–S5 linkers, the S5 helices, the pore helices and the S6 helices are shown in cartoon (yellow, cyan, brown and blue respectively). Residues implicated in pyrethroid binding in various pest species (M918, L925, T929 and L932) are shown in space fill (green). Reproduced from reference 22, © The Biochemical Society, London, UK. The model is based on the crystal structures27 of the potassium channels KvAP and Kv1.2.

Figure 3.

Docking predictions for DDT (left-hand panel), deltamethrin (middle panel) and fenfluthrin (right-hand panel) with the voltage-gated sodium channel (insecticide structures shown in stick format). The S4–S5 linkers, the S5 helices, the pore helices and the S6 helices are shown in cartoon (yellow, cyan, brown and blue respectively). Residues implicated in pyrethroid binding in various pest species (M918, L925, T929 and L932) are shown in green stick format. Predicted distances of H-bonding interactions from the T929 threonine oxygen (donor) to each pyrethroid's carbonyl oxygen (acceptor) are 3.03 Å (deltamethrin) and 2.73 Å (fenfluthrin). Reproduced with permission from reference 20, © Elsevier BV.

Figure 4.

Structures of DDT, pyrethrum compounds (R1 = H, R2 = CH = CH2 = pyrethrin I, R2 = CH3 = cinerin I, R2 = CH2CH3 = jasmolin I, R1 = C02, R2 = CH = CH2 = pyrethrin II, R2 = CH3 = cinerin II, R2 = CH2CH3 = jasmolin II) and the synthetic pyrethroids permethrin (type I), fenfluthrin (fluorinated type I), deltamethrin (type II), fenvalerate (non-cyclopropane ring containing type II) and flumethrin (fluorinated type II).

The location of the pyrethroid-binding site in the domain IIS4-5 region rather than in association with the IIS6 segment (the kdr locus) is consistent with the interpretation of structure–activity relationships for interaction of pyrethroids with kdr and super-kdr channels.26 A conformational effect is therefore most likely to account for kdr, as, according to the VGSC model, the L1014 residue does not make physical contact with the insecticide. One possible scenario (based on recent potassium channel crystallographic data27) is that the IIS6 inner helix, by curving parallel to the inner membrane plane, makes a platform or ‘receptor’ for the IIS4-S5 helix; this interaction is seen as being essential for the coupling of voltage-sensor movements to channel pore opening and closing.27 Thus, when the VGSC is closed, the IIS4–S5 linker packs tightly against the bottom half of IIS6, closing the pore (Fig. 5). Upward movement of the IIS4 voltage sensor in response to membrane depolarisation moves the IIS4–S5 linker, allowing bending of IIS6 and channel opening.27 The position of the ‘kdr’ locus next to the ‘hinge’ glycine (residue 1012—housefly numbering) on IIS6 may influence the propensity of this helix to bend and thereby have an impact on positioning of the IIS4–S5 linker and IIS4 voltage sensor, thus altering channel activation kinetics (in this case the channel is less prone to open). This is again in line with previous electrophysiological studies showing that the kdr substitution L1014F on domain IIS6 promotes closed-state inactivation, whereby 70–80% of the sodium channels never open.14 The latter reduces the action of pyrethroids, because these insecticides bind preferentially to the open state of the channel. Interestingly, the super-kdr substitutions M918T and T929I also promote closed-state inactivation, but in addition speed the rate of dissociation of pyrethroids up to 100-fold.14, 24

Figure 5.

Proposed mechanism of kdr resistance. In the closed-channel state, the IIS4–S5 linker packs tightly against the bottom half of IIS6, closing the channel. Upward movement of the IIS4 voltage sensor in response to a membrane depolarisation moves the IIS4–S5 linker, allowing bending of IIS6 and channel opening. Positioning of kdr next to the ‘hinge’ glycine on IIS6 influences the propensity of IIS6 to bend and thereby impacts on positioning of the IIS4–S5 linker and IIS4 voltage sensor, thus altering the channel's activation kinetics (i.e. the channel becomes less prone to open).

Since the original identification of the L1014F (kdr) mutation, two variants (L1014H and L1014S) have also been reported.11 Bioassays of houseflies homozygous for the L1014H mutation have indicated a lower level of resistance to permethrin, compared with that conferred by L1014F.28 The L1014S mutation, which is common in malarial mosquito (Anopheles gambiae Giles) populations from East Africa,29 appears to be involved in a slightly higher level of resistance to DDT combined with a lower level of resistance to permethrin, while their West African equivalents (commonly carrying L1014F) show moderate resistance to both pyrethroids and DDT.29 Therefore, it is possible (if metabolic detoxification is discounted) that these differences are a reflection of differing degrees of influence of different kdr substitutions on the positioning of the IIS4–S5 linker. Different insecticide selection pressures (e.g. exposure to either type-I or type-II pyrethroids or DDT) or environmental constraints may also influence which kdr mutation is present (i.e. a particular kdr residue may be more effective at reducing susceptibility to one chemical structure than another, or may impose a fitness cost on the insect). However, it is necessary to wait for the results of future electrophysiological studies on these substitutions in a common VGSC background before drawing firm conclusions on the significance of differential bioassay responses to insecticides between the different kdr variants. In insects, kdr is also further complicated by secondary mutations in the cytoplasmic portion of the channel that appear to enhance the level of resistance associated with the primary L1014 polymorphism but are not themselves primary resistance mutations.30


Both DDT and pyrethroid insecticides kill insects by binding to the VGSCs in nerve membranes and interfering with the functioning of the nervous system. Point mutations within the genes encoding these channels are responsible for pyrethroid resistance in a wide range of agricultural pests and insects that vector human diseases. Laboratory-based studies have both confirmed the role of the mutations in causing resistance and provided a fundamental understanding of the mechanism of action and selectivity of pyrethroids. In silico modelling studies have great potential to enhance knowledge of these underlying mechanism(s) by predicting some of the key insecticide–target interactions and thereby provide an explanation as to how the different resistance mutations exert their effect at the molecular level. The authors anticipate that this will also lead to new prospects for targeting specific pyrethroid (or next-generation pyrethroid-like) compounds to particular insect pests while at the same time avoiding toxicity to non-target organisms such as beneficial insects.


Rothamsted Research is supported by the Biotechnology and Biological Sciences Research Council. This work was also supported in part by a project grant from the BBSRC to BAW.