2.2.4.1 Natural variability
Comparing the intraspecific variation in susceptibility to Bt toxins between insect field populations is one approach to estimate the potential of insect field populations to evolve resistance to Bt toxins and giving an indication of baseline susceptibilities. It is common practice to compare the LC50 (concentration that kills 50 % of the
population) values. Low variability in susceptibility among populations, however, doesn’t necessarily mean a low potential for selection pressure to act on, as variability within populations can still be high (Ferré and Van Rie, 2002).
Studies measuring variation in susceptibility have reported both high and low natural variations among populations. Populations of O. nubilalis in Germany showed no differences in susceptibility to Cry1Ab (Saeglitz et al., 2006). There were small differences in susceptibilities to Cry1Ab between populations of Diatraea saccharalis F., sugar cane borer, in Louisiana and Texas, although all were as susceptible as a laboratory population (Huang et al., 2008). Susceptibility to Cry1Ac had a 50-fold range and Cry 1Ab had a 30-fold range in populations of Cnaphalocrocis medinalis Guenée (Lepidoptera: Pyralidae) an important lepidopteran rice pest, in China (Han et
al., 2008). Heliothis virescens susceptibilities to Cry2Ab2 in the USA varied up to
48-fold for nine laboratory, seven laboratory-cross and 28 field populations, although the means of these three groups varied only 2-fold (Ali and Luttrell, 2007). In the same study H. zea susceptibilities varied up to 37-fold but the means of laboratory, laboratory-cross and field populations varied only 3-fold (Ali and Luttrell, 2007). Populations of H. virescens in the USA varied 12-fold to Cry1Ac, while H. zea varied 130-fold (Ali et al., 2006).
Another method for estimating the variability of resistance genes is by measuring heritability (h2) in laboratory experiments (Tabashnik, 1994; Ferré and Van Rie, 2002). Tabashnik (1994) estimated heritability of resistance to Bt products and Cry1A toxins for 27 selection experiments and showed that compared with eight other insect species, P. interpunctella had a relatively high h2 value. Relatively high h2
Sayyed et al., 2000b). These high h2 values indicate high additive genetic variation for susceptibility to Cry proteins in those populations (Ferré and Van Rie, 2002).
2.2.4.2 Estimation of resistance allele frequency
The initial frequency of resistance alleles is important and influences the rate at which resistance may evolve (Gould et al., 1997). Therefore, it is a key element for predicting the rate of evolution in a population subjected to insecticide treatments (Ferré and Van Rie, 2002). Genetic models have assumed that initial allelic frequencies range from 10-2 to 10-6 based on theoretical assumptions regarding the balance between mutation and selection (Roush and McKenzie, 1987; Gould et al., 1997). A direct estimate with H. virescens, using homozygous recessive resistant females (YHD2) individually mated with field captured males, calculated that the frequency of resistance alleles that confer resistance to Cry1Ac in the field population was 1.5 x 10-3 (Gould et al., 1997). Tabashnik et al. (1997a) estimated the frequency of resistance alleles in a susceptible P. xylostella population to be 1.2 x10-1. Andow and Alstad (1998) used an F2 screening procedure to estimate the frequency of
Cry1Ab resistant alleles in O. nubilalis to be less than 1.3 x 10-2 for a Minnesota population (Andow et al., 1998) and less than 3.9 x 10-3 for an Iowa population (Andow et al., 2000). An F2 screen estimated that resistance allele frequency for field
derived H. virescens to Cry1Ac was 3.6 x 10-3 to 2.6 x 10-2 (Blanco et al., 2009). These direct estimates have high initial frequencies, as compared to the genetic model estimates, and if these estimates are typical, resistance may evolve more quickly than expected if precautionary measures are not in place.
2.2.4.3 Mode of inheritance of resistance
The dominance level of resistance is important as it influences the rate at which resistance may evolve. It is generally assumed that alleles for Bt resistance are rare initially (Gould et al., 1997) and individuals homozygous for resistance to Bt (RR) extremely rare initially. Therefore, the response of heterozygotes (RS) to Bt determines the initial course of evolution of resistance. If heterozygotes are killed by
Bt then the resistance is termed recessive, and if the heterozygotes survive exposure to
Bourguet et al. (2000) reported that dominance has been assessed in different ways in insecticide resistance studies. Values ranging from 0 for complete recessivity to 1 for complete dominance, have been obtained from single-dose mortality tests (DML,
dominance of survival at a given insecticide dose), from LC50 values of dose-
mortality curves (DLC, dominance of insecticide resistance) and from the fitness of the
3 geneotypes in insecticide-treated areas (DWT, dominance of relative fitness in the
treated area). The DWT, calculation is the most relevant to resistance management, but
at the same time is also the most difficult to estimate. The DLC value was modified
from the widely used Stone’s (1968) formula, for which the degree of dominance ranged from -1 to 1 (Ferré and Van Rie, 2002). It has been suggested that partial or completely recessive modes of inheritance are more closely associated with modification of binding sites, and dominant alleles seem to be associated with other mechanisms conferring broad spectrum resistance, for example, modification of midgut proteolytic activity (Gould et al., 1997; Liu and Tabashnik, 1997; Tabashnik
et al., 1998; Bourguet et al., 2000).
Recessive resistance to Bt was first reported in P. interpunctella (McGaughey, 1985). For H. virescens, resistance in the SEL and CP73-3 populations both showed incompletely dominant inheritance to Cry1Ac (Sims and Stone, 1991; Gould et al., 1992). Gould et al. (1995) reported that inheritance of resistance in the YHD2 population to Cry1Ac and Cry1Ab was partially recessive, but Cry2Aa was more dominant.
For P. xylostella populations from Hawaii (NO-QA) and Pennsylvania (PEN), resistance to Cry1Aa, Cry1Ab, Cry1Ac and Cry1F was partly to completely recessive, however a population from the Phillipines (PHI) showed recessive inheritance of resistance to Cry1Ab, but resistance to Cry1Aa and Cry1Ac was associated with partially dominant inheritance (Tabashnik et al., 1997b; Tabashnik et al., 1998). Another Hawaiin population (NO-95C) resistance to Cry1C was partially dominant (Liu and Tabashnik, 1997). A P. xylostella population (SERD4) from Malaysia showed incompletely dominant inheritance to Cry1Ac in the Cry1Ac-selected subpopulation and to Cry1Ab in the Cry1Ab-selected subpopulation (Sayyed et al.,
2.2.4.4 Number of resistance genes
The number of resistance genes involved may depend on the toxins that the population is resistant to and how many mechanisms of resistance are present, since having more than one resistance mechanism would suggest the presence of more than one resistance gene (polygenic resistance). Many of the studies, using backcross data, determining the number of loci involved in resistance have fitted the single locus model or monogenic resistance (Ferré and Van Rie, 2002).
For a H. virescens population (SEL) resistance to Cry1Ab was polygenic (Sims and Stone, 1991), whereas for a second population (YHD2) resistance to Cry1Ab was monogenic (Gould et al., 1995). Field-evolved resistance in P. xylostella populations NO-QA and Karak both exhibited monogenic resistance (Tabashnik et al., 1997b; Sayyed et al., 2004), whereas the field derived SERD4 population selected with Cry1Ac or Cry1Ab exhibited polygenic resistance with some parental sex influence (Sayyed et al., 2005). Two populations of O. nubilalis selected with Cry1Ab have shown polygenic resistance to Cry1Ab (Alves et al., 2006).
2.2.4.5 Fitness costs and stability of resistance
The evolution of resistance in a population may compromise the normal functions of individuals, such as survival, development time and fecundity, thereby causing a fitness cost associated with resistance. The nature of potential fitness costs are revealed in environments that lack the selecting agent. In the case of Bt crops, resistant individuals moving into refuges may exhibit fitness costs in comparison to susceptible individuals that have not been exposed to Bt toxin. Thus, a trade-off occurs with Bt resistance increasing fitness in the presence of Bt toxin, but causing a fitness cost when Bt toxin is absent (Tabashnik, 1994; Ferré and Van Rie, 2002; Gassmann et al., 2009). The stability of Bt resistance is linked to fitness costs associated with resistance. In many studies on resistance to Bt Cry toxins, resistance reverted when selection was stopped (Ferré and Van Rie, 2002; Gassmann et al., 2009).