Insect resistance management in Bacillus thuringiensis cotton by MGPS (multiple genes pyramiding and silencing)

The introduction of Bacillus thuringiensis (Bt) cotton has reduced the burden of pests without harming the environment and human health. However, the efficacy of Bt cotton has decreased due to field-evolved resistance in insect pests over time. In this review, we have discussed various factors that facilitate the evolution of resistance in cotton pests. Currently, different strategies like pyramided cotton expressing two or more distinct Bt toxin genes, refuge strategy, releasing of sterile insects, and gene silencing by RNAi are being used to control insect pests. Pyramided cotton has shown resistance against different cotton pests. The multiple genes pyramiding and silencing (MGPS) approach has been proposed for the management of cotton pests. The genome information of cotton pests is necessary for the development of MGPS-based cotton. The expression cassettes against various essential genes involved in defense, detoxification, digestion, and development of cotton pests will successfully obtain favorable agronomic characters for crop protection and production. The MGPS involves the construction of transformable artificial chromosomes, that can express multiple distinct Bt toxins and RNAi to knockdown various essential target genes to control pests. The evolution of resistance in cotton pests will be delayed or blocked by the synergistic action of high dose of Bt toxins and RNAi as well as compliance of refuge requirement.


Introduction
Cotton (Gossypium hirsutum) is an ultimate source of fiber for the textile industry and seed oil for biofuel. China, India, USA, Pakistan, Brazil and Australia are the leading cotton producers ( Figs. 1 and 2). The G. hirsutum and G. barbadense account for 90% and 8% of the global cotton production, respectively. Globally, 150 countries are involved in cotton industrial chain, providing income for more than 100 million families and employment for almost 7% of all labor in developing countries (Fig. 3). The global cotton production in 2018-2019 is 118.5 million bales which is 4.2% lower than the previous year (Dohlman et al. 2019). The insect pests and diseases cause 15%∼30% economic losses to cotton production and even up to 50% losses by direct damage or transmission of plant diseases (Cui et al. 2007;Chen et al. 2020;Tarazi et al. 2019). The major insect pests which cause substantial losses to cotton production are cotton jassid, cotton aphid, thrips, spotted bollworm, pink bollworm, American bollworm, cotton mealy bug, pink boll worm, fall armyworm and whitefly. The excessive use of pesticides leads to the insecticidal resistance, pest resurgence that exacerbates the situation. These chemicals kill indiscriminately beneficial and harmful insects, and also deteriorate the environment and human health (Sanahuja et al. 2011).
In the mid-1990s, the Monsanto (US company) developed Bacillus thuringiensis (Bt) cotton and then commercialized it to different cotton producing countries. The transgenic cotton kills some voracious insect pests without harming environment as well as human health and increases the yield and farmers' profit by reducing the application of insecticides (Flachs 2017;Wu et al. 2008). In eight cotton producing countries, the adoption of genetic modified (GM) insect resistant (IR) cotton causes reduction of over 331 million kg of insecticide active ingredient. The transgenic cotton is highly selective, efficient for numerous key lepidopteran pests, ecofriendly, and become an important part of intergrated pest management (IPM) (Naranjo 2011).
Globally, the transgenic cotton is grown on area of more than 33 million hectares (Tarazi et al. 2019). The adoption of Bt cotton has reduced the application of insecticides up to 305 million kilogram in China, India and USA during the period of 1996-2018. It has also minimized the farmers' exposure to hazardous insecticides, increased the yield and profits, and suppressed the insect pests population of cotton (Brookes and Barfoot 2020a, b). Globally, the adoption of transgenic cotton has grown up to 42% from 2017 to 2018 and its adoption by major cotton growing countries was remarkable, such as China (95%), India (93%), USA (96%), Brazil (84%) and Australia (100%) (ISAAA 2018). In India, Bt cotton has increased the yield by 100% (Kranthi and Stone 2020). The adoption of transgenic cotton in India has reduced the pest damage, increased 24% yield per acre and 50% profits (Fleming et al. 2018). In China, the Bt cotton has reduced the insecticide applications (47% to 79%) (Veettil et al. 2017).
The Bt crops provide significant economic and environmental advantages but these benefits can be eliminated by the evolution of resistance in insects and pests (Carpenter 2010;Tabashnik et al. 2013;Tabashnik et al. 2010). The growing of Bt crops on a large scale exerts more selection pressure on insect pests and results in resistance against insecticidal activity of Bt crops. The 21 cases of field-evolved resistance have been reported that decrease the efficacy of nine Cry proteins (Tabashnik and Carrière 2020;Calles-Torrez et al. 2019). To date, seven targeted pests of Bt crops have control problem due to the development of resistance (Tabashnik and Carrière 2017;Tabashnik et al. 2013). The efficacy of Bt crops has decreased due to field-evolved resistance. The field-evolved resistance is genetically controlled and reduced the susceptibility to Bt toxin caused by exposure of a pest population to the toxin in the field (Tabashnik et al. 2009). The proteins Cry1Ac belonged to Cry1A family is most widely used in Bt cotton to control some lepidopteran larvae. The open field resistance in Pectinophora gossypiella to Cry1Ac and Cry1Ab is reported in India (Naik et al. 2020). In USA, the Helicoverpa zea showed resistance to transgenic cotton exhibiting Cry1Ac and Cry2Ab (Tabashnik et al. 2013). The analysis of 51 field-obtained strains of P. gossypiella exhibits significant lower susceptibility to Cry1Ac during 2008 to 2010 than 2005 to 2007, that validated the evolution of resistance in pink bollworm (Wan et al. 2012). The evolution of resistance to Cry1Ac of H. armigera in open field condition (Tabashnik et al. 2013) has served as a warning because it has also gained the resistance against Bollgard-I cotton since 2012 (Tabashnik and Carrière 2017;Cui et al. 2007). Four cases of field-evolved resistance to transgenic crops expressing single toxin of Cry1A class are confirmed (Bemisia fusca, Cry1Ab maize, H. zea, Cry1Ac cotton, Spodoptera frugiperda, Cry1F maize, H. armigera, Cry1Ac cotton) (Tabashnik et al. 2009). Worldwide, the practical resistance against different endotoxins in seven major insect species has been reported (Naik et al. 2018;Tabashnik et al. 2013;Grimi et al. 2015). The third generation of transgenic technology is promising and is being tested in the fields against different insects and pests attacks. For example, different constructs containing Cry1Ac, Cry2A, ViP3A and EPSPS have been synthesized and developed. The 3 rd generation Bt cotton exhibits three genes (Cry1Ac + Cry2Ab + Vip3A), (Cry1Ab + Cry2Ac + Vip3Aa19) or (Cry1Ac + Cry1F + Vip3A). In Australia, the 3 rd generation cotton expressing Cry1Ac + Cry2Ab + Vip3Aa was planted on more than 90% area during 2016-2017 (Tabashnik and Carrière 2017). In case of USA, the 3 rd generation cotton expressing Cry and Vip3Aa covered 27% area of total cotton growing area in 2019 (USDA-FAS 2019). Most of the cotton pests like corn earworm, bollworms and fall armyworms get resistance to pyramided cotton expressing Cry1Ac + Cry1F and Cry1Ac + Cry2Ab and other combinations of different Cry genes Tabashnik and Carrière 2017). Various studies have confirmed the development of field-evolved resistance in major cotton pests against different Cry proteins used in the 3 rd generation Bt cotton, but only Vip3A is consistently effective against these pests (Tabashnik and Carrière 2017). The efficacy of Vip3A protein used in pyramided crops will be reduced due to resistance  . Reisig et al. (2018) also reported the increased damage to fruiting structure of pyramided cotton expressing two or multiple Bt genes by Cry1Ac resistant population of H. Zea. In another study, the population of H. Zea which have decreased susceptibility to Cry1Ac, showed increased survival & damage to Bollgard II and Bollgard III cotton (Little et al. 2019). The results of this study suggested that the addition of Vip3Aa gene in the 3 rd generation cotton is not sufficient for the management of pests which are resistant to Cry proteins. To preserve the efficacy of Vip3Aa proteins in pyramided cotton, the application of other resistance management tactics are necessary (Little et al. 2019). Yang et al. (2019) reported early warning of resistance in H. Zea against Vip3A protein. The cotton leaf bioassay of H. Zea larvae obtained from pyramided maize expressing Cry1Ab + Cry1F + Vip3A showed well survival on WideStrike3 cotton expressing Cry1Ac + Cry1F + Vip3A proteins . In another study, statistical significant, weak cross resistance is confirmed between Vip3 and Cry1 toxins with a mean of 1.5-fold cross-resistance in 21 cases (range: 0.30∼4.6-fold) (Tabashnik and Carrière 2020). Six cases of resistance in different insects against Vip3A have been reported, including S. litura (Barkhade and Thakare 2010), S. frugiperda (Bernardi et al. 2016;Yang et al. 2018), H. armigera (Chakroun et al. 2016), H. Zea (Little et al. 2019) and H. virescens (Pickett et al. 2017).

Factors contributing to resistance in insect pests of cotton
In Lepidopteran larvae, ATP-binding cassette (ABC) transporters, alkaline phosphatases (ALP), membrane bound cadherin (CAD) in midgut, and aminopeptidase N (APN) play a vital role in insecticidal activity. Gene expression modulation and mutations in receptor sites of insects are major reasons for Bt resistance decrease. The modification in Bt receptors, reduction of proteolytic activity in midgut and regeneration or replacement in midgut cell are responsible factors for resistance development to Bt toxins. The mutations disrupting a cadherin protein are tightly linked with recessive resistance to Cry1Ac in H. armigera, P. gossypiella and H. virescens (Gahan et al. 2001;Morin et al. 2003;Xu et al. 2005;Li et al. 2019). In field derived populations of H. armigera, recessive cadherin alleles accounted for 75% to 84% of resistance alleles detected. However, major resistance alleles were found in heterozygotes whereas at least one non-recessive resistance allele was observed in 59%∼94% of resistant individuals (Zhang et al. 2012). The cadherin allele r 1 and other cadherin resistance alleles accounted for 88% of the resistance alleles in field derived population of cotton bollworm (Zhang et al. 2012). The cellular trafficking is affected by CAD transmembrane mutation, causing resistance in P. gossypiella to Cry1Ac Bt toxin (Wang et al. 2018b). The mutations in promotor of trypsin gene (HaTryR) induced Cry1Ac resistance in H. armigera . The resistance to Cry1Ac in cotton bollworm occurred due to different mechanisms like qualitative changes or reduced levels of the confirmed and putative midgut receptors cadherin, alkaline phosphatase, aminopeptidase, and ABCC2 proteins (Zhang et al. 2012;Zhang et al. 2009;Jurat-Fuentes et al. 2011;Xiao et al. 2014). Three altered alleles of the cadherin encoding gene associated with Cry1Ac resistance in  (Brookes and Barfoot 2020a, b) pink bollworm were deletion (Morin et al. 2003). Mutations in ABC transporter proteins cause resistance to Cry1 and Cry2 toxins in eight Lepidopteran species (Heckel 2012;Wu et al 2019). The mutation in an ABC transporter gene (PgABCA2) is linked with resistance to Cry2Ab in pink bollworm (Mathew et al. 2018). The missplicing of ABCC2 gene results in loss of 150 amino acids and causes resistance in H. armigera to Cry1Ac. The mutation in ABCA2 gene conferred resistance to Cry2Ab in H. armigera. The CRISPR/Cas9 mediated downregulation of HaABCA2 results in resistance to Cry2Aa and Cry2Ab in cotton bollworm (Wang et al. 2020). Various ABC transporters bind to different Bt toxins; for example, ABCC2 is a receptor for Cry1Ab and Cry1Ac, and ABCA2 for Cry2Ab (Tay et al. 2015). The proteomic and genomic studies showed reduced expression of ALP in Bt resistant strains of H. armigera, H. virescens, and S. frugiperda than that in susceptible (Jurat-Fuentes et al. 2011). In H. armigera, a deletion mutation in the HaAPN1 gene is linked with the resistance to Cry1Ac ).
The concentration of toxin in Bt plants should be high enough to kill all or nearly all hybrid progeny, which is called "high dose" criterion. In India and China, the bollworm developed resistance to Cry genes because Bt cotton has not met the criterion of high dose. The field data of approved Bt cotton varieties expressing Cry1Ac from these countries showed significant survival of susceptible P. gossypiella larvae (Wan et al. 2004;Bambawale et al. 2010), indicating the non-fulfillment of the high dose criterion. In India, pink bollworm got resistance to Bt cotton after 7 years due to the use of illegal Bt cotton seeds with low doses of Bt protein and non-compliance with the refuge strategy ). The S. frugiperda and H. zea became resistant to Cry1Ac due to lower dose in Bt corn and cotton (Ali et al. 2006;Storer et al. 2012). Low expression level helps operate the natural selection in increasing the frequency of mutated resistant population.
Different studies suggested that the growing refuges with Bt crop reduced the selection pressure in susceptible insect pests of cotton and delayed the evolution of resistance. Lack of refuge caused high selection pressure which resulted in resistance in insect pests of cotton. The resistance in S. frugiperda (Storer et al. 2012) and P. gossypiella (Tabashnik et al. 2012) occurred owing to scarcity of refuges. The growing of non-Bt plants with Bt plants allowed the survival of susceptible individuals in Bt dominated environment. The random mating between dominant susceptible (SS) and recessive resistant (RR) individuals resulted in heterozygous (RS) progeny, which could be killed by Bt crops. The fulfilment of high dose criteria lessened the process of rapid resistance development, and in some instances, if this standard was not maintained throughout the growing season, the resistance could be delayed for more than 10 years with abundant refuges (Tabashnik et al. 2013).
From different geographical studies, it is evident that several other factors could also contribute to resistance development in pests against Bt cotton. These factors include lack of regulation and compliance withenvironment protection agency (EPA), multiple exposure to same Bt endotoxins, cross resistance to multiple Bt endotoxins and failure of producing high Bt endotoxin dose . For instance, Cry1Ab and Cry1Ac are still being practiced even after 20 years of its incorporation in most varieties of Bt cotton and corn, and a cross resistance in H. zea was detected between these two endotoxins (Brevault et al. 2013;Crespo et al. 2015). The simultaneous cultivation of a pyramid with a mono-toxin plants expressing a toxin, which is also part of pyramid toxins, can accelerate evolution of resistance to the pyramid. The resistance strain of S. frugiperda to Cry1Fa get rapid resistance to pyramid of corn producing Cry1A.105 + Cry2Ab, because the toxins Cry1Fa and Cry1A.105 are closely related (Santos-Amaya et al. 2015;Tabashnik et al. 2013).
Different studies suggested that the insecticidal efficacy of Bt cotton varied owing to variable expression of Bt protein during the cotton growing season (Chen et al. 2017a;Wan et al. 2005). The factors affecting the Bt concentration in transgenic cotton are discussed in Table 1.
The variance of insect resistance in cotton bollworm and armyworm is linked with differential expression of Cry1Ac in field, that effected by environment, varietal background as well as the age of plant (Chen et al. 2017a;Chen et al. 2018). Bt protein expression could be regulated by promotor, nucleotide sequence, insertion point, transgene amplification, natural conditions and cell's environmental factors (Wang et al. 2018c;Downes et al. 2016;Hobbs et al. 1993). Overwhelmingly, insecticidal ability was directly or indirectly influenced by intensity of pest and diseases, rain fall, soil characteristics and adequate and appropriate farm management. Taking all together, optimal environment is necessary for GM cotton production which ultimately leads to reinforce the expression of Bt gene.

Genome information of major insect pests of cotton
The host-plant diversification involves the expansion of chemosensory gene families used for recognition of volatile and nonvolatile molecules (Gouin et al. 2017). Chemical signals are detected by proteins which comprised of multigene families and moderately in size. These families encompasses: (i) ionotropic receptors (IR), (ii) chemoreceptor super family consisting of gustatory receptors (GRs) and olfactory receptors (OR), (iii) chemosensory proteins (CSPs) and odorant binding proteins (OBPs) (Sánchez-Gracia et al. 2001). The GRs are present on taste sensilla on tarsi, mouthparts and ovipositors where they recognize non-volatile molecules (e.g. bitter compounds salts and sugar) present on food substances and oviposition substrates (Isono and Morita 2010). OBPs and CSPs proteins exist on lymph antennae and dendrites of olfactory receptor neurons (ORNs). The OBPs and CSPs can catch and bind environmental chemical cues and then conveyed to ORs or IRs (Xu et al. 2009). The ORs are important especially for insect and host odor recognition (Carey et al. 2010). IRs are involved in detecting environmental chemical signals. Different antennal IRs perform various functions in the process of insects' recognition of external information (Chen et al. 2015). The evolution of resistance to insecticide occurs by interaction of multiple genes.
The increased metabolic detoxification and reduced target site sensitivity are major outcomes of insecticide resistance. In insects, the host toxic secondary metabolites and xenobiotics are normally detoxified by UDPglycosyltransferases, esterases (CCEs), glutathione-stransferases (GSTs) and cytochrome P450s (CYPs) (Gouin et al. 2017). Despite the various sizes of CYPomes in insects, many genes, often of CYP 3 and CYP 4 clans, are arrayed in tight clustersof tandemly duplicated genes, reflecting recent duplications. The presence of multiple, closely related CYP genes in the genomes of insect pests presents a challenge to the functional identification of the genes that are important in adaptation to plant chemicals and detoxification of insecticides (Wang et al. 2018a). Insecticide resistance is outcome of duplication of genes encoding detoxification enzymes (Bass et al. 2013). The GSTs are involved in various biological activities including xenobiotic detoxification and secondary metabolism (Sylvestre-Gonon et al. 2019). Esterases are entangled in neurogenesis, hormone and pheromone degradation, developmental regulation and xenobiotics detoxification. The UGTs played a vital role in endobiotic regulation and xenobiotics detoxification by catalyzing sugar with small hydrophobic compounds to produce glycosides (Teese et al. 2010). In herbivorous insects, the metabolism is governed by essential digestive proteases. The serine proteases (SPs) and serine protease homologs (SPHs) are involved in numerous physiological processes like digestion, development and immunity (Yang et al. 2017b).
In S. frugiperda the higher number of GRs (N = 231 genes), OBPs (50 genes), CSP repertoire (22 genes), OR (69 genes), and IR (42 genes) were found. Expansion occurs in GRs and OBPs,because tandem duplications and strong conservation in candidate antennal IRs were found (Gouin et al. 2017). The variety of evolutionary adaptive characteristics is a result of gene duplication (Conant and Wolfe 2008). A total of 117 CYP genes were annotated in S. frugiperda and strong gene expansion was observed in CYP6, CYP9, CYP321, CYP324 and CYP4 families. The S. frugiperda have 46 GST genes. These GST genes exhibit extraordinary diversification of epsilon class and expansion of epsilon and delta cytosolic classes. Its genome contained 96 carboxyl/cholinesterases (CCEs) with notable expansions. The expansion through tandem duplications is found in UGTs gene families which also revealed patterns of interspecific conservation in gene number. The S. frugiperda has conservative antioxidant defense system. There are 86 digestive SP genes and rapid gene duplication was found (Gouin et al. 2017). Chemosensory genes were found, showing almost non-significant variation in both C and R strains but significant variation in gene number of detoxification and digestion genes. The difference of detoxification and digestion genes between both strains results in differential adaptation to various ranges of hostplant (Gouin et al. 2017). The whitefly Bemisia tabaci (Gennadius) is a phloem feeding insect, recognized as a complex composed of 35 morphocryptic species. Within the species complex, the highly invasive and destructive species are Middle East-Asia Minor 1 (MEAM1/B) and Mediterranean (MED/Q) (Xie et al. 2018). Across the hemipteran insect genomes, B. tabaci has low genome synteny. The MEAM1/B and MED/Q are intricate in detoxification and metabolism reflecting similar gene families (Xie et al. 2018). The genome of MEAM1/B has 130 P450 genes, 81 UGT genes, 22 GST genes, 50 ABC genes and 51 COE genes. However, the MED/Q genome has 153 P450 genes, 63 UGT genes, 21 GST genes, 59 ABC genes and 51 COE genes. The MEAM1/B and MED/Q have significant expansion in P450 detoxification gene family, and CYP3 and CYP4 clades of MED/Q were expanded largely. This remarkable expansion in metabolic and detoxification genes results in insecticide resistance in B. tabaci (Xie et al. 2018). The significant expansion of cytochrome P450s, UDP-glucuronosyltransferases cathepsins, and phosphatidylethanolamine-binding proteins were found in genome of MEAM1/B (Chen et al. 2016). Eight OBP and 19 CSP genes were identified in MEAM1/B. The phylogenetic analyses revealed lineage-specific expansion in CSP genes (BtabBCSP1, BtabBCSP3, BtabBCSP13, BtabBCSP17, BtabBCSP18 and BtabBCSP19) (Zeng et al. 2019). The global invasion of pests and their resistance to resistance is a result of expansion of genes involved in chemo sensation, metabolism, detoxification and those related to pesticide resistance, as well as virus acquisition (Chen et al. 2016).
There are 12 OBPs, 23 IRs, 34 ORs and 50 GRs, all belong to chemosensory related genes, in genome of Aphis gossypii Glover. There are 62 P450s, 72 ABCs, 7 GSTs, 20 CCEs and 56 UGTs (all belong to detoxification related genes) in A. gossypii. As compared with other aphids, the A. gossypii has lower IRs, OR and GR (Quan et al. 2019). It is suggested that the size of chemosensory genes is closely associated with the host range in aphids (Nicholson et al. 2015). The A. gossypii genome also encodes extra ABC genes and fewer GSTs than other aphid species (Quan et al. 2019).
Pink bollworm is also regarded as model insect to understand the insect responses to Bt toxins. The molecular bases of its tolerance are not well documented. The 46 458 transcripts derived from 39 874 unigenes were used to construct a transcriptome assembly for the midgut of P. gossypiella.
The transcriptome data presented midgut proteins which are important for detoxification of xenobiotic, digestion of nutrients and their allocation, as well as for the discovery of protein receptors crucial for Bt intoxication (Tassone et al. 2016).
Final assembly of H. armigera contains 997 scaffolds with a total genome size of 337 Mb. The scaffold N50 size was 1.00 Mb. H. armigera detoxification gene families consist of 114 P450s, 97 CCEs, 42 GSTs, and 46 UGTs and 54 ABCs. The chemosensory gene families involved 84 ORs, 213 GRs, 29 CSP and 40 OBPs. The serine proteases, major digestive clades, consist of 45 trypsins and 49 chymotrypsins genes. Across the 300 genes, H. armigera detoxification gene families had over 70 genes, i.e., GSTs, CCEs and P450s, whereas over 90 gene families corresponding with digestion and more than 150 are chemosensory genes. The polyphagy behavior and insecticide resistance in H. armigera is due to extensive amplification, duplication and neofunctionalisation of genes involved in detoxification, chemosensation and digestion (Pearce et al. 2017).
The lack of detailed genome study on P. solenopsis is a barrier to understand the molecular bases of its growth, development as well as insecticide resistance. The RNA-Seq technology was applied to execute de novo transcriptome assembly and comparative transcriptome profiling of different developmental phases of P. solenopsis. About 182.67 million reads were assembled into 93 781 unigenes with an average length of 871.4 bp and an N50 length of 1 899 bp. The differentially expressed gene (DEG) libraries exhibited 29 415 DEGs upon comparison of gene expression profiles among different developmental stages. DEGs were associated with hormone biosynthesis, developmental processes, anti-microbial protection and functional protein synthesis. This study provides genomics resource covering all developmental stages of P. solenopsis and helps to identify vital RNAi target to control it (Arya et al. 2018). In P. solenopsis, present molecular sequence resources were amplified through de novo transcriptome assembly, and RNA sequencing generated 12 925 coding sequence (CDS) from 23 643 contigs with an average size of 1 077.5 bp per CDS. At all developmental stages, the expression of targeted genes (AQP, IAP,CAL, VATPase, SNF7, α-amylase, chitin synthase and bursicon) was studied and the silencing of these genes by injecting their respective dsRNA was achieved (Singh et al. 2019b). Fourteen candidate reference genes were assessed at five various stages as well as under starvation stress. The results revealed that GST, Actin, TFIID, SDHA, and 28s were identified as the several best reference genes for expression analysis studies in mealybug (Singh et al. 2019a).

Management strategies Gene pyramiding
The pyramided crops producing two or more distinct Bt toxins belongs to Cry or Vip insecticidal proteins are designed to impede the resistance evolution. In early 2000s, the pyramided Bt cotton was introduced for resistance management and control of insect pests like bollworm of cotton. Now in USA, Australia, India, and China, single Bt gene cotton has been replaced with pyramided Bt cotton that produces two Bt toxins, either Cry1Ac and Cry1F or Cry1Ac and Cry2Ab (Brevault et al. 2013). In 2004, the single Bt gene (Cry1Ac) is replaced by pyramided cotton expressing Cry1Ac + Cry2Ab in Australia for the management of H. punctigera and H. armigera (Downes and Mahon 2012). The development of resistance has been delayed for more than a decade after the introduction of this pyramid. The co-expression of Vip3A & Cry1Ac enhanced the insecticidal activity of transgenic cotton against lepidopteran pests. The pyramiding of vip3AcAa and cry1Ac genes into cotton has increased the larval mortality rates of S. litura, A. ipsilon and S. exigua as compared with single Bt-Cry1Ac cotton (Chen et al. 2017b). In S. frugiperda (Cry1F resistant), cross-resistant was observed against Cry1Ac and Cry1Ab. However, it was not crossresistant to Vip3A, Cry2Ae and Cry2Ab2. Because of this cross-resistance mechanism among Cry1 proteins, the pyramided cotton expressing two or more Bt genes to control Cry1F-resistant S. frugiperda has become necessary (Yang et al. 2017a). The recessive resistance was found in Cry1F-resistant S. frugiperda. The pyramiding of different Bt genes in corn was effective for managing the Cry1F resistance in S. frugiperda (Niu et al. 2014).

Factors effecting the durability of pyramids
Durability of pyramids crops is enhanced by following conditions: i) refuges are abundant; ii) frequency of resistance alleles is rare; iii) resistance is recessive; iv) fitness costs are linked with resistance; v) resistance is not complete; vi) the multiple toxins in pyramid can kill all susceptible pests; vii) there is no cross resistance among toxins used in pyramid; viii) there is no simultaneous cultivation of pyramids with mono-toxin plants that exhibits one of the toxins used in the pyramid. Retrospective analysis of all cases showed that significant deviations from the first three conditions causes practical resistance to single toxin crops (Brevault et al. 2013;Carrière et al. 2015;Zhao et al. 2005).
The cross-resistance has reduced the efficacy of pyramids, and it happens when selection of a pest population with one Bt toxin causes genetically controlled reduction in sensitivity to other toxins (Tabashnik et al. 2014). Theoretically, weak or strong cross-resistance will speed up resistance development in insect pests which are genetically less susceptible to Bt crops, while only strong crossresistance will speed up resistance in pests which are inherently more susceptible to Bt crops (Carrière et al. 2015). The cross resistance occurs in those Bt toxins which have more similar amino acid sequence and share more similar binding sites in midgut of insect. The similarity of the amino acid sequence in domain II of Bt toxins is associated with cross resistance between toxins in the pyramids. A recent assessment of cross resistance in 10 important pests against seven sets of Bt toxin in 80 cases verified this pattern and revealed that the resemblance of amino acid sequence of domain II is linked with crossresistance (Carrière et al. 2015). So, pyramiding of two or more toxins which are not cross resisted by the target pest is a better strategy for resistance management. It is supposed that the resistance to Bt toxins is autosomal and governed by single diallelic recessive genes. Redundant killing of pests is decreased by strong cross resistance among toxins, because if the target pest is resistant to one toxin of the pyramid, then it can survive to other toxins (Tabashnik et al. 2014). The stronger cross-resistance is likely to occur in Cry1, Cry2 and Cry3 toxins because they are more similar and share similar three-domain structure. If the pests have weak cross-resistance but more inherently susceptible to Bt toxins, the resistance development to pyramids will be accelerated. But if the pests have less inherent susceptibility with weak cross resistance, the evolution of resistance to Bt toxins of the pyramid can be accelerated (Carrière et al. 2015). The weaker cross resistance was observed among Cry3Bb or mCry3Aa & Cry34/35Ab because of different structural homology (Gassmann et al. 2014). The simultaneous cultivation of a pyramid with a monotoxin plants that express a toxin which is also part of pyramid toxins can accelerate evolution of resistance to the pyramid. The resistance strain of S. frugiperda to Cry1Fa get rapid resistance to dual gene Bt corn producing Cry1A.105 and Cry2Ab because the toxin Cry1Fa is closely related to Cry1A.105 toxin (Santos-Amaya et al. 2015). The cross-resistance and antagonism among toxins used in pyramids are common. The similarity of amino acid sequence in domains II and III is the major cause of cross-resistance and antagonism (Carrière et al. 2015).

High dose/refuge strategy
Transgenic crops can only be deemed successful if they have high dose/refuge strategy, which means that: 1) the Bt crops must express high dose; 2) the frequency of resistance alleles should be low; 3) abundant refuges (non Bt plants) are grown with Bt crops. The fieldevolved resistance in fall armyworm and pink bollworm is well documented in many countries. The factors which are involved in the field-evolved resistance of the insects against Bt-cotton are failure of the crop to express high dosage and lack of sufficient refuge plants . The concentration of each toxin in the pyramid must be high enough to kill 95% of the insect's susceptible populations. The pyramid will be more effective, if each toxin in a pyramid acts independently and kills 99.75% of susceptible pests (Carrière et al. 2015). Among 18 observations of nine test-pyramid combinations, only half met this criteria (Carrière et al. 2015). In case of pyramid cotton that expressed Cry1Ac and Cry2Ab, the mortality rate of H. zea and H. armigera was more than 99.75% (Carrière et al. 2015).
The refuge strategy plays an important role in regulating the evolution of resistance to Bt crops. The approach mainly depends on mating between resistant and susceptible individuals produced in Bt and non-Bt (refuge) host plants. The random mating between dominant susceptible (SS) and recessive resistant (RR) individuals results in heterozygous (RS) progeny, that can be killed by Bt crops (Tabashnik and Carrière 2017) so that the resistance against insects can be delayed against Btcrops (Jin et al. 2015;Carriere et al. 2012;Tabashnik et al. 2008). The susceptible insects live and grow on refuges (non-Bt plants). This is an effective way that can lead to the delay in the development of resistance in pests against Bt-crops that can assist in using the same genes for the longer period of time (Carrière et al. 2016).

Other strategies
To delay resistance, releasing sterile insects (Tabashnik et al. 2010) and seeds mixture strategy (Carrière et al. 2016) are also used. The development of modified Bt toxins is also used for the management of resistance. In this approach, the knowledge is involved inhow the insect pests acquire resistance against a toxin and then alter the formulating of that toxin so that resistance can occur in another manner that will lead to the delay in evolution of resistance to Bt-crops. Cry1AbMod and Cry1AcMod killed M. Sexta and P. gossypiella that had cadherin deletion mutations (Soberón et al. 2007). Modified toxins showed high insecticidal activity against the most resistant strains of H. virescens and H. armigera (Tabashnik et al. 2011). The conservation of natural enemies can be very effective in delaying the evolution of resistance in Bt-cotton. Nearly 500 species of natural enemies in cotton systems in China have been reported (Luo et al. 2014).

RNAi
RNA interference (RNAi) is a potential approach for effective insect control via downregulation of gene expression (Table 2). RNAi targets genomic sequences in the insect species to avoid those genomic regions that belong to the beneficial insects. The Bt-cotton gene pyramiding from Cry1Ac and Cry1Ab incorporates the resistance against bollworms coupled with RNAi, which assists in the dysfunctioning of those genes involved in the development of tolerance against Bt-genes, which results in long term resistance against bollworms in cotton (Ni et al. 2017). RNAi silencing involves the breakdown of dsRNA into short interfering RNA by the RNase II enzymes dicer and drosha, and these siRNA are loaded into another complex RNA-induced silencing complex (RISC). The siRNA is unwound during the RISC assembly and single stranded RNA hybridizes with mRNA target. Gene silencing results in two ways. The first is the nucleolytic degradation of the targeted mRNA by RNase enzyme Argonaute (Slicer), the second is that if there is mismatch between the mRNA and siRNA resultantly, mRNA cannot be cleaved, but resulting in a translational blockage. There are many RNAi methods, e.g., micro-injections, ds RNA spray and artificial diet-based feeding have been adopted. The efficacy of microinjection and feeding methods are variable depending on the type of genes and the organisms. Moreover, it has also been reported that neither of the two methods produce same results in the organisms (Watson 2018). These methods have been very successful in laboratory but they are not effective in the fields. At field level, in planta expression of dsRNAs to knockdown the specific genomic regions has been very economical for insect control (Younis et al. 2014). The dsRNA produced by transgenic plants against key gene of pests has been regarded as safeguard that endows transgenic resistant plants with new innovations (Mao et al. 2007) ( Table 2). The main benefits of employing RNAi are its high degree of specificity and efficiency. Therefore, RNAi is used in the functional analysis of genes to evaluate the inhibition of genes that lead to the loss of any specific phenotypic functions (Majumdar et al. 2017).

Multiple gene pyramiding and silencing
Insect pests of cotton can acquire resistance against single Bt toxins; therefore, pyramided Bt cotton and efficacy of refuge for regulating the evolution of resistance against Bt-crops were introduced to overcome this resistance (Carrière et al. 2019). Recently, studies have suggested that insect pests (i.e., P. gossypiella, H. zea, S. frugiperda) have developed tolerance against dual gene pyramided cotton, and refuge also lost its efficacy in case of non-recessive resistance, i.e., cotton bollworm (Jin et al. 2015). Presently, new strategies are needed to be developed to delay the evolution of resistance in cotton pests. Plant-mediated RNAi of essential pest genes involved in defense, detoxification, digestion and development is being utilized for enhancing tolerance against insects and pests. In recent years, new types of insect resistant transgenic cotton have been developed using RNAi technology or RNAi pyramided with Bt genes (Ni et al. 2017;Mao et al. 2011;Mao et al. 2013;Wu et al. 2016). Ni et al. (2017) developed a pyramid of cotton containing Bt and RNAi, and found excellent results against cotton bollworm, but also substantially delayed resistance as compare with using Bt alone.
Pyramiding of multiple RNAi expression cassettes against various essential genes involved in defense, detoxification, digestion and development of cotton pests will successfully obtain favorable agronomic characters for crop protection and production. The MGPS involves the construction of transformable synthetic chromosomes, that have multiple distinct Bt toxins and RNAi to knockdown various essential target genes of pest (Ren et al. 2019). The evolution of resistance in cotton pests will be delayed or blocked due to synergistic action of high dose of Bt toxins and RNAi(s) as well as compliance of ample refuge. The transgenic cotton based on MGPS coupled with refuge can be an effective and smart way to control pests.

Conclusion
The adoption of Bt cotton increase the yield, profit and reduced the application of pesticides as well as load of insect pests without harming the human health and environment. The development of resistance in insects and pests has reduced effectiveness of single and pyramided Bt cotton. The modification in midgut receptors, lack of high dose/refuge, cross resistance and fluctuation in expression of Bt protein during growing season are major factors that facilitate in resistance development. Besides, the resistance development in cotton pest and the drastic increment in population of secondary pest due to less application of insecticides have become a major concern for Bt cotton growers. Currently, different strategies like pyramided cotton expressing two or more distinct genes, refuge strategy, releasing of sterile insects, seed mixture, and genome editing by CRISPR/Cas9 and RNAi are being used to control insect pests. Recently, studies have suggested that insect pests (i.e., P. gossypiella, H. zea, S. frugiperda) have developed tolerance against dual gene pyramided cotton, and refuge also lost its efficacy in case of non-recessive resistance, i.e., cotton bollworm. The insects are remarkably adaptable and can develop resistance to any control tactics, including transgenic plants containing multiple Bt toxins and RNAi. The innovations like genetically modified Bt toxins and discovery of insecticidal proteins from bacteria other than Bt will continue to provide new tools for pest control. The MGPS-based cotton will be more durable with compliance of high refuges and other control tactics.