Insecticide resistance is achieved in a selected strain or population (i) by an alteration of the target site, (ii) by an alteration of the effective dose of insecticide that reaches the target or (iii) by a combination of the two. The resistance phenotypes have long been analyzed according to these useful biochemical and physiological criteria. At the molecular genetic level, several classes of mutations can account for these phenotypes (Taylor and Feyereisen, 1996; Feyereisen et al., 2015).

The precise molecular mutations responsible for P450-mediated insecticide resistance are only beginning to be explored. The cis-mutation (Accord insertion) in the 5‘UTR of the Cyp6g1 gene in Drosophila causing up-regulation of expression (Daborn et al., 2002) is currently the most detailed account and is discussed below. However, all classes of molecular mutations (structural, up- or down-regulation, Taylor and Feyereisen, 1996; Feyereisen et al., 2015) can be theoretically involved in P450-mediated resistance.

Traditionally, the first line of evidence for a role of a P450 enzyme in resistance has been the use of an insecticide synergist (e.g. piperonyl butoxide)(Feyereisen, 2015). A suppression or decrease in the level of resistance by treatment with the synergist being diagnostic. In cases too many to list here, this initial and indirect evidence is probably correct, however there are cases where piperonyl butoxide synergism has not been explained by increased detoxification (Kennaugh et al., 1993). Piperonyl butoxide may also be a poor inhibitor of the P450(s) responsible for resistance, or may inhibit some esterase activity so that the use of a second, unrelated synergist may be warranted (Brown et al., 1996, Zhang et al., 1997). In addition, the synergist as P450 inhibitor can decrease the activation of a proinsecticide, so that lack of resistance suppression can be misleading. Chlorpyriphos resistance in D. melanogaster from Israel vineyards maps to the right arm of chromosome 2 and is enhanced by piperonyl butoxide (Ringo et al., 1995). An independent and additional line of evidence is the measurement of total P450 levels or metabolism of selected model substrates. An increase in either or both being viewed as diagnostic. Such evidence is indirect, and the absence of change is not informative.

An increase in the metabolism of the insecticide itself in the resistant strain is more conclusive and has been demonstrated in many studies. For instance, permethrin metabolism to 4'-hydroxypermethrin was higher in a microsomes from Culex quinquefasciatus larvae that are highly resistant to permethrin (Kasai et al., 1998b) than in their susceptible counterparts. Total P450 and cytochrome b5 levels were 2.5 times higher in the resistant strain. Both permethrin toxicity and metabolism were inhibited by two unrelated synergists, TCPPE and piperonyl butoxide. A similarly convincing approach was taken to show P450 involvement in the resistance of house fly larvae of the YPPF strain to pyriproxifen. Gut and fat body microsomes were shown to metabolize the IGR to 4'-OH-pyriproxyfen and 5“-OH-pyriproxyfen at higher rates than microsomes of the susceptible strains and this metabolism was synergist-suppressible (Zhang et al., 1998). The major, dominant resistance factor was linked to chromosome 2 in that strain (Zhang et al., 1997).

Increased levels of transcripts for one or more P450 genes in insecticide resistant strains has now been reported in many cases. Transcriptomics approaches have been used in several species to accelerate the discovery of differentially expressed (mostly overexpressed) CYP genes in several species. These include Drosophila, mosquitoes, Tribolium castaneum and Myzus persicae. These studies suggest that overexpression of one or more P450 genes is a very common phenomenon of metabolic resistance but they do not by themselves establish a causal relationship with resistance. Instead, they merely provide lists of candidate genes (e.g. Brun-Barale et al., 2010).

Genetic linkage between increased mRNA or protein levels for a particular P450 and resistance has been obtained to the chromosome level or closer to marker genes (Feyereisen, 2005; Bogwitz et al., 2005; Wee et al., 2008; Hardstone et al., 2010). Linkage is just the first step in establishing a causal link between a P450 gene and resistance. QTL mapping of resistance in mosquitoes has also provided short lists of candidate genes for further functional studies.

Functional expression of the P450 enzymes in heterologous systems is required to establish the contribution to resistance of a CYP gene.

Transgenic expression in Drosophila or other insects followed by in vivo toxicity assays can provide convincing evidence, but the interpretation of such experiments is not always straightforward. For instance, while Bogwitz et al. (2005) demonstrated that Cyp12a4 overexpression in the midgut and Malpighian tubules confers lufenuron resistance in Drosophila, they also showed that high level ectopic expression of the gene was embryonic lethal (for an unknown reason).

CRISPR/Cas gene editing, such as gene knockouts or even knockout of gene clusters is an increasingly accessible and very effective technique to study the contribution of one or more P450 genes to xenobiotic metabolism.

RNAi suppression of CYP gene expression is very informative, but this technique is not equally effective in all insect species. Furthermore, the effects observed are often marginal.

The following is a discussion of selected cases of P450 genes associated with insecticide resistance. Evidence for mutations causing constitutive overexpression in cis and trans, as well as examples of P450 gene duplication/amplification are currently available. The variety of mechanisms, even in a single species in response to the same insecticide, is striking. The paucity of available data on the molecular definition of the resistant genotype and on its causal relationship to resistance is also striking when compared to the wealth of data on target site resistance (ffrench-Constant et al., 1999).

CYP6A1 is the first insect P450 cDNA to have been cloned. The gene is phenobarbital-inducible and constitutively overexpressed in the multiresistant Rutgers strain (Feyereisen et al., 1989). A survey of 15 house fly strains (Cariño et al., 1992) showed that CYP6A1 is constitutively overexpressed at various degrees in eight resistant strains, but not in all resistant strains - notably R-Fc known to possess a P450-based resistance mechanism. Thus the first survey with a P450 molecular probe confirmed the results of the first survey of house fly strains with marker P450 activities (aldrin epoxidation and naphthalene hydroxylation)(Schonbrod et al., 1968): there is no simple relationship between resistance and a molecular marker, here the level of expression of a single P450 gene. That different P450 genes would be involved in different cases of insecticide resistance was a sobering observation (Cariño et al., 1992), even before the total number of P450 genes in an insect genome was known.

The constitutive overexpression of CYP6A1 in larvae and in adults is linked to a semi-dominant factor on chromosome 2 (Cariño et al., 1994), but the CYP6A1 gene maps to chromosome 5 (Cohen et al., 1994), implying the existence of a chromosome 2 trans-acting factor(s) differentially regulating CYP6A1 expression in the two strains (Cariño et al., 1994). The gene copy number and the coding sequence of CYP6A1 are identical between Rutgers and a standard susceptible strain (sbo) (Cariño et al., 1994; Cohen et al., 1994). Competitive ELISA using purified recombinant CYP6A1 protein as standard showed that the elevated mRNA levels are indeed translated into elevated protein levels (Sabourault et al., 2001). Reconstitution of recombinant CYP6A1 expressed in E.coli with its redox partners (Sabourault et al., 2001) provided the conclusive evidence for its role in diazinon resistance, as CYP6A1 metabolizes the insecticide with a high turnover (18.7 pmol/pmol CYP6A1/min), and a favorable ratio (2.7) between oxidative ester cleavage and desulfuration.

The nature of the chromosome 2 trans-acting factor and of the mutation leading to resistance in the Rutgers strain has remained enigmatic despite considerable circumstantial evidence for a major resistance factor on chromosome 2 (Plapp, 1984). Diazinon resistance and high CYP6A1 protein levels could not be separated by recombination in the short distance between the ar and car genes (3.3-12.4 cM). This region carries an ali-esterase gene (MdαE7). A Gly137 to Asp mutation in this ali-esterase abolishes carboxylesterase activity towards model compounds such as methylthiobutyrate, and confers a low but measurable phosphotriester hydrolase activity towards an organophosphate (“P=O”), chlorfenvinphos, in both the sheep blowfly Lucilia cuprina LcαE7 and Musca domestica MdαE7 enzymes (Newcomb et al., 1997, Claudianos et al., 1999). Chromosome 2 of the Rutgers strain carries this MdαE7 Gly137 to Asp mutation, and low CYP6A1 protein levels are correlated with the presence of at least one wildtype (Gly137) allele of MdαE7. Recombination in the ar-car region could not dissociate diazinon susceptibility, low CYP6A1 protein level and the presence of a Gly137 allele of the ali-esterase (Sabourault et al., 2001). It was therefore hypothesized that the wildtype ali-esterase metabolizes an (unknown) endogenous substrate into a negative regulator of CYP6A1 transcription. Removal of this regulator (by loss-of-function of the ali-esterase) would increase CYP6A1 production and hence, diazinon metabolism. House fly strains that are susceptible or that are not known to overexpress CYP6A1 predictably carry at least one wild type Gly137 allele (Scott and Zhang, 2003). The LPR strain that overexpresses CYP6A1 (Cariño et al., 1992) and has increased OP metabolism (Hatano and Scott, 1993), as well as other resistant strains, carry other alleles of MdαE7 (Scott and Zhang, 2003; Gacar and Taskin, 2009). These alleles, Trp251 to Ser or Leu, also have impaired ali-esterase activity in Musca domestica (Claudianos et al., 1999; Gacar and Taskin, 2009). The hypothesis that the wild type allele of MdαE7 is a trans-acting negative regulator causing low expression of CYP6A1 in susceptible strains has implications. The pleiotropic effect of trans regulation is compatible with constitutive overexpression of CYP12A1 (whose product metabolizes diazinon as well, Guzov et al., 1998) and of GST-1 that are both controlled in the Rutgers strain by a chromosome 2 factor, possibly the same as the one controlling CYP6A1 expression.

There are alternative explanations, for instance very close linkage of the ali-esterase MdαE7 with an uncharacterized trans-acting factor, or interference of small inversions with genetic crosses and recombination in that region. The fragmented current assembly of the house fly genome is a hindrance to further research in this (historically important) case of resistance.

The diazinon resistance (Rop-1) and malathion resistance (Rmal) in Lucilia cuprina, are linked to the Gly137Asp and Trp251Leu mutations in LcαE7 (Newcomb et al., 1997). The Trp251Leu mutation enhances hydrolysis of dimethylorganophosphates and of the malathion carboxylesters, while the Gly137Asp mutation enhances preferentially the hydrolysis of diethylorganophosphates but virtually abolishes malathion carboxylesterase activity (Devonshire et al., 2003). In view of these data, could the Gly137Asp mutation alone be responsible for diazinon resistance in the blow fly ? The elegant calculations of Devonshire et al. (2003) show that despite the very low activity of the Gly137Asp mutant esterase to hydrolyze the oxon form of the pesticide (kcat ~ 0.05 min-1), the 10-20 fold resistance to diazinon and parathion could indeed be accounted by LcαE7. Those calculations, based on the oxon form do not take into account the necessary P450-dependent desulfuration of the parent OP to produce the oxon, that is invariably associated with ester cleavage (see above), so that P450 may still play a role in the blow fly. Indeed, the Q strain of the sheep blowfly is more resistant to parathion than to paraoxon (Hughes and Devonshire, 1982) and indirect evidence for a P450 involvement in Lucilia cuprina diazinon resistance has also been presented (Kotze, 1995, Kotze and Sales, 1995).

In the Rutgers strain of the house fly, the situation is different. The mutant ali-esterase cannot account for the P450-dependent carbamate and JHA resistance. Furthermore, the higher resistance to diazinon than in Lucilia cuprina (120x vs 10x) requires a more efficient clearance of the oxon. The contribution of CYP6A1 alone accounts for over 5 times more than the mutant ali-esterase to the timely removal of the toxic form. Lethality of the MdαE7 null (Sabourault et al., 2001) suggests that the Gly137Asp mutation is the optimal loss-of-function mutation as the Rutgers haplotype has swept through global populations of the house fly (Claudianos, 1999; Gacar and Taskin, 2009). The low phosphotriester hydrolase activity of the mutant ali-esterase probably helps clearing the activated form (P=O) of the insecticide (Sabourault et al., 2001). The endogenous function of the wild type MdαE7 gene remains to be elucidated.

The power of Drosophila genetics coupled with the tools made possible by the complete genome sequence has provided the most detailed, yet complex molecular genetic detail about a P450-based insecticide resistance mechanism. The resistance gene Rst(2)DDT has been genetically characterized for over 40 years (review in (Daborn et al., 2001, Wilson, 2001). The position of this gene around 64.5cM on the left arm of chromosome 2 has become almost mythical, as a number of phenotypes were linked to this locus, from dominant DDT resistance to phenylthiorurea susceptibility, from organophosphorus to carbamate resistance, from various P450-dependent activities to vinyl chloride activation. EMS mutagenesis of a wildtype stock and selection with imidacloprid led to two strains with moderate imidacloprid resistance and moderate cross-resistance to DDT (Daborn et al., 2001). Conversely, two DDT-resistant strains (Hikone-R and Wisconsin-1) were shown to be cross-resistant to imidacloprid. Fine scale mapping of this dominant resistance localized Rst(2)DDT to a region from 48D5-6 to 48F3-6 on the polytene chromosome map. Of three candidate P450 genes in this region, Cyp6g1, Cyp6g2 and Cyp6t3, only the first showed constitutive overxpression in the DDT and imidacloprid resistant strains tested (Daborn et al., 2001). A DNA microarray comprising probes for all the Drosophila P450 genes was addressed with target cDNAs from susceptible strains and from the DDT-resistant Hikone-R strain and the propoxur-resistant WC2 strain. In both cases, Cyp6g1 was the only P450 gene showing constitutive overexpression (Daborn et al., 2002). Overexpression of Cyp6g1 was confirmed by quantitative (RT)PCR in 20 strains, and DDT, imidacloprid, nitenpyram and lufenuron resistances were all independently mapped to the Cyp6g1 locus in the Hikone-R and WC2 strains. The insertion of a terminal direct repeat of the transposable element Accord was systematically found in the 5' UTR of 20 different resistant strains from across the globe. Phylogenetic analysis of the first intron sequence of the gene showed a unique haplotype in resistant strains vs a large diversity of susceptible haplotypes, suggesting a selective sweep had occured in global Drosophila populations (Daborn et al., 2002). This was further demonstrated in a survey of 673 lines from 34 populations collected around the world showing perfect correlation between the presence of the Accord insertion and resistance and a reduction in variability measured by microsatellite analysis in a 20kb region downstream of Cyp6g1 (Catania et al., 2004). In some resistant lines, the presence of a P-element insertion into the Accord element was reported in that study. In fact, the Cyp6g1 locus is even more complex, with at least six different alleles found in nature (Schmidt et al., 2010). Beyond the Cyp6g1 ± Accord insertion alleles, Schmidt et al. reported four alleles in which the Cyp6g1 with Accord insertion was duplicated, and in three of them additional insertions of P-element or HMS Beagle elements were found within the original Accord insertion. This allelic succession is adaptive, with higher resistance and Cyp6g1 transcription found for the most complex allele (Cyp6g1-[BP], Schmidt et al., 2010). Transgenic flies producing CYP6G1 under control of a variety of promoters in the GAL4/UAS system (heatshock, tubulin, or midgut/fat body/Malpighian tubules) showed increased survival to acetamiprid, imidacloprid and nitenpyram in larvae and to DDT in adults (Daborn et al., 2002; Le Goff et al., 2003; Chung et al., 2007; Daborn et al., 2007). Significantly, the 491 bp Accord sequence carries enhancer elements itself and can direct expression of reporter GFP in the tissues (gastric caeca, midgut, Malpighian tubules and fat body of larvae) in which Cyp6g1 expression is localized in resistant strains (Chung et al., 2007). Moreover, the Malpighian tubules of adults are critical, as overexpression of Cyp6g1, or its RNAi knockdown in just this tissue can significantly shift the toxicity of DDT to lower or higher levels, respectively (Yang et al., 2007). Definitive functional evidence that CYP6G1 metabolizes insecticides was provided by Joussen et al. (2008) who showed that the enzyme produced in tobacco cell suspensions cultures metabolizes imidacloprid and DDT as predicted by the experiments with transgenic flies resulting in resistance and by homology modeling (Jones et al., 2010). In most field-collected strains, DDT resistance is significant but low compared to that of strains further selected in the laboratory (e.g. 91-R, see below) suggesting that while Cyp6g1 may constitute a first line of defense seen in field populations, further insecticide pressure in the laboratory may select additional mechanisms. In addition, mechanisms other than Cyp6g1 overexpression (including target site resistance) can also be involved in DDT resistance. For instance, while Cyp6g1 overexpression is observed in the DDT-resistant Wisconsin and 91R strains, the Cyp12d1 (or Cyp12d2) gene is overexpressed as well in both strains (Brandt et al., 2002; Festucci-Buselli et al., 2005) and its transgenic overexpression can confer resistance to DDT (Daborn et al., 2007). Moreover, several other genes are overtranscribed in these strains as well (Pedra et al., 2004). Therefore it is hardly surprising that resistance, that has always been a relative term, is not restricted to Cyp6g1 overexpression, especially when compared to strains that are themselves resistant (Festucci-Buselli et al., 2005; Kuruganti et al., 2006). In a Brazilian strain of Drosophila simulans resistant to DDT, imidacloprid and malathion, only the Cyp6g1 ortholog is overexpressed (Le Goff et al., 2003). In a California population of D. simulans the 5'-flanking sequence of the Cyp6g1 ortholog is nearly fixed for a Doc transposable element insertion. This insertion is absent from African populations and is associated with increased transcript abundance of Cyp6g1 and resistance in a what appears to be a case of resistance analogous with the Accord case of D. melanogaster Cyp6g1 (Schlenke and Begun, 2004). UPDATE NEEDED

The LPR strain of the house fly is highly resistant to pyrethroids with a phenoxybenzyl moiety. This permethrin-selected strain has several resistance mechanisms with important contributions of the genes pen (for reduced penetration) and kdr (for target site resistance)(Liu and Scott, 1995) and with P450-based detoxification as a major contributor. An abundant form of P450 (P450Lpr) was purified from abdomens of adult LPR flies, and immunological data indicated that P450Lpr represents 67% of the P450 in microsomes from LPR flies, a 10-fold (Wheelock and Scott, 1990) increase over the reference susceptible strain. The P450Lpr gene, CYP6D1 (Tomita and Scott, 1995) is located on chromosome 1 of the house fly (Liu et al., 1995), and is constitutively overexpressed by about 10-fold in the LPR strain. This overexpression is not caused by gene amplification, but by increased transcription. It has been claimed that increased transcription of CYP6D1 causes insecticide resistance (Liu and Scott, 1998), but transgenic expression of CYP6D1 in Drosophila (Korytko et al., 2000a) has not been reported to confer resistance, and heterologously expressed CYP6D1 (Smith and Scott, 1997) has not been reported to metabolize pyrethroids. Instead, the evidence for the role of CYP6D1 in pyrethroid resistance is based on the inhibition of microsomal deltamethrin and cypermethrin metabolism by anti-P450Lpr antibodies (Wheelock and Scott, 1992b, Korytko and Scott, 1998). Deltamethrin is metabolized preferentially at the gem-dimethyl group on the acid moiety (Wheelock and Scott, 1992b) whereas cypermethrin is mainly hydroxylated at the 4' position on the alcohol moiety, at an extremely low rate (Zhang and Scott, 1996a). Whether overexpression or point mutations of CYP6D1 or both are involved in pyrethroid resistance in the LPR strain is yet unknown. Indeed, the CYP6D1 gene sequence from 5 strains shows a high polymorphism, with 57 variable sites in the coding region alone, of which 12 are non-silent (Tomita et al., 1995). Six amino acid changes are specific to the LPR strain (CYP6D1v1) when compared to pyrethroid-susceptible strains and several of these mutations appear to align with SRS3. The 170-fold piperonyl butoxide-suppressible resistance is conferred by a combination of the resistant chromosomes 1 and 2 from the parent LPR strain in the homozygous condition (Liu and Scott, 1995). There is no substantial resistance or CYP6D1 overproduction conferred by isolated LPR chromosomes 1 or 2, or by their subsequent combination (Liu and Scott, 1995, Liu and Scott, 1996). Thus, in the LPR strain, P450-mediated resistance requires both copies of the LPR chromosomes 1 and 2. The resistance and CYP6D1 overexpression linked to chromosome 1 are dominant, whereas the contributions of chromosome 2 are mostly recessive (Liu and Scott, 1996, Liu and Scott, 1997a). These data suggest a unique combination in the LPR strain of chromosome 2 trans-acting factor(s) with at least a matched cis-factor on chromosome 1. A key sequence difference between the 5'UTR of the CYP6D1v1 allele of LPR and of susceptible strains is the presence of a 15 bp insert that interrupts a binding site of the transcriptional repressor mdGfi-1, and reduces the amount of mdGfi-1 binding to the CYP6D1 promoter in electrophoretic mobility shift assays (Gao and Scott, 2006). This strongly suggests that the 15-bp insertion is the major cis-mutation on chromosome 1. This insert is also found in a permethrin-resistant strain NG98 from Georgia, USA that carries a virtually identical CYP6D1v1 haplotype (Seifert and Scott, 2002). However, in a strain from neighboring Alabama with high permethrin resistance (ALHF), chromosomes 1 and 2 play little role but chromosome 5 plays a major piperonyl butoxide-suppressible role (Liu and Yue, 2001). Two P450 genes located on chromosome 5, CYP6A36 and CYP6A5v2 (a probable recent duplicate of CYP6A5) are constitutively overexpressed in that strain (Zhu and Liu, 2008; Zhu et al., 2008a). A gene closely linked to CYP6D1 on chromosome 1 codes for a similar (78% identity) P450, CYP6D3 (Kasai and Scott, 2001a). CYP6D3 is 12-fold overexpressed in adult flies of the LPR strain, but it is also expressed in larvae (Kasai and Scott, 2001b), as opposed to CYP6D1 which has an adult-specific pattern of expression. CYP6D1 is overexpressed by about 2.4 fold in the Japanese strain YPER (Kamiya et al., 2001; Shono et al., 2002). This strain does not carry the CYP6D1v1 allele, and chromosome 2 has a major role in this permethrin-resistant strain. CYP6D3v2 and CYP6A24 are also overexpressed in Japanese strains (Kamiya et al., 2001). Pyrethroid resistance in the house fly appears thus to involve multiple mutations playing a role in the contribution of several P450s to the multifactorial resistance.