A Third Type of Resistance to Cydia pomonella Granulovirus in Codling Moths Shows a Mixed Z-Linked and Autosomal Inheritance Pattern

ABSTRACT

Different isolates of Cydia pomonella granulovirus (CpGV) are used worldwide to control codling moth larvae (Cydia pomonella) in pome fruit production. Two types of dominantly inherited field resistance of C. pomonella to CpGV have been recently identified: Z-chromosomal type I resistance and autosomal type II resistance. In the present study, a CpGV-resistant C. pomonella field population (termed SA-GO) from northeastern Germany was investigated. SA-GO individuals showed cross-resistance to CpGV isolates of genome group A (CpGV-M) and genome group E (CpGV-S), whereas genome group B (CpGV-E2) was still infective. Crossing experiments between individuals of SA-GO and the susceptible C. pomonella strain CpS indicated the presence of a dominant autosomal inheritance factor. By single-pair inbreeding of SA-GO individuals for two generations, the genetically more homogenous strain CpRGO was generated. Resistance testing of CpRGO neonates with different CpGV isolates revealed that isolate CpGV-E2 and isolates CpGV-I07 and -I12 were resistance breaking. When progeny of hybrid crosses and backcrosses between individuals of resistant strain CpRGO and susceptible strain CpS were infected with CpGV-M and CpGV-S, resistance to CpGV-S appeared to be autosomal and dominant for larval survivorship but recessive when success of pupation of the hybrids was considered. Inheritance of resistance to CpGV-M, however, is proposed to be both autosomal and Z linked, since Z linkage of resistance was needed for pupation. Hence, we propose a further type III resistance to CpGV in C. pomonella, which differs from type I and type II resistance in its mode of inheritance and response to CpGV isolates from different genome groups.

IMPORTANCE The baculovirus Cydia pomonella granulovirus (CpGV) is registered and applied as a biocontrol agent in nearly all pome fruit-growing countries worldwide to control codling moth caterpillars in an environmentally friendly manner. It is therefore the most widely used commercial baculovirus biocontrol agent. Since 2005, field resistance of codling moth to CpGV products has been observed in more than 40 field plantations in Europe, threatening organic and integrated apple production. Knowledge of the inheritance and mechanism(s) of resistance is indispensable for the understanding of host response to baculovirus infection on the population level and the coevolutionary arms race between virus and host, as well as for the development of appropriate resistance management strategies. Here, we report a codling moth field population with a new type of resistance, which appears to follow a highly complex inheritance in regard to different CpGV isolates.

INTRODUCTION

The codling moth (Cydia pomonella) is a lepidopteran (family Tortricidae) whose larvae cause serious damage in pome fruit production, mainly in apple and pear (1). C. pomonella is worldwide distributed mostly through the temperate fruit-growing regions (2). In conventional plant protection, this pest has been controlled with chemical insecticides entailing a variety of environmental and safety detriments (3). Thus, the demand for environmentally safe control agents for C. pomonella increased. Cydia pomonella granulovirus (CpGV), first described from diseased C. pomonella larvae found in Mexico (4), has a very narrow host range and is highly virulent for C. pomonella larvae. CpGV belongs to the genus Betabaculovirus of the Baculoviridae family. Its genome comprises double-stranded circular DNA of about 120.8 to 123.9 kb (5, 6). CpGV isolates are phylogenetically classified into five genome groups, A to E (6). Because CpGV is highly effective against early larval stages and harmless to nontarget insects, mammals, and the environment, biocontrol agents based on CpGV have been developed for C. pomonella control (1).

Since its first registration in 1987 in Switzerland, CpGV products have been used as biopesticides in virtually all apple-growing areas worldwide (7). The CpGV occlusion bodies (OB) can be applied in their natural form; C. pomonella larvae taking up the OB during feeding die within a few days after ingestion (1). Most of the commercial CpGV products of the first generation contained the Mexican isolate, termed CpGV-M (4, 7). In 2005, after nearly two decades of field use in Europe, several organic apple plantations where CpGV treatment had failed to control C. pomonella satisfactorily were identified in southwestern Germany (8). When laboratory bioassays were performed with progeny of these populations, a 1,000- to 100,000-fold-reduced susceptibility to CpGV-M was confirmed (9). This was the first time that resistance of insect populations to commercial baculovirus products had been reported. In the following years, field resistance of C. pomonella populations to CpGV products was observed in about 40 apple plantations in Austria, the Czech Republic, France, Germany, Italy, the Netherlands, and Switzerland (10–13). Acquired resistance of insects to baculoviruses has been described also for several insect-baculovirus associations, such as Phthorimaea operculella-Phthorimaea operculella granulovirus (PhopGV) (14), Anticarsia gemmatalis-Anticarsia gemmatalis multiple nucleopolyhedrovirus (AgMNPV) (15), Trichoplusia ni-Trichoplusia ni single nucleopolyhedrovirus (TnSNPV) (16, 17), and Adoxophyes honmai-Adoxophyes honmai nucleopolyhedrovirus (AdhoNPV) (18, 19). However, these incidental laboratory selections never reached the wide geographic distribution and the economic significance of CpGV field resistance in C. pomonella.

The codling moth exhibits a karyotype of 2n = 56 chromosomes, with 54 autosomes (A) and a WZ/ZZ (female/male) sex chromosome system (20). For the laboratory-selected strain CpRR1, which derived from the resistant field population CpR, a dominant, monogenic, and Z-linked inheritance of CpGV resistance was demonstrated (9, 21). Genetic analyses of the resistant C. pomonella strains RGV (France) and CpR-CZ (Czech Republic) revealed a Z-linked and dominant inheritance mode also for these geographically distant populations (11, 22). Hence, a single main inheritance type has been initially assumed in European C. pomonella populations. This so-called type I resistance appeared to be targeted only against CpGV-M (group A), whereas other CpGV isolates (genome groups B to E) were able to overcome resistance (6, 23). For CpRR1, it was demonstrated that resistance-breaking properties of the CpGV isolates depended on the lack of an insertion of 24 bp in the open reading frame (ORF) pe38 in CpGV-M (group A), which is missing in all resistance-breaking isolates of genome group B to E CpGVs (6). By registration of some of these isolates, a successful commercial control of the type I resistance in C. pomonella was eventually reestablished in Europe (23–28).

More recently, C. pomonella field populations which could not be controlled by the use of novel resistance-breaking CpGV products were identified. Resistance of one of these populations, termed NRW-WE, is targeted not only against CpGV-M but also against isolates belonging to the genome groups C to E (13), suggesting an alternative mode (=type II) of resistance in NRW-WE (13). Indeed, when the laboratory strains CpR5M and CpR5S were generated by selection experiments starting with NRW-WE individuals, it was demonstrated that type II resistance followed a dominant, monogenic but autosomal inheritance pattern. Furthermore, a cross-resistance to at least two CpGV isolates, CpGV-M and CpGV-S, was observed (29).

A further C. pomonella field population from northeastern Germany, called SA-GO, showed also a reduced susceptibility to both CpGV-M and CpGV-S. SA-GO was first reported in a study concerning the distribution of CpGV resistance in Europe (12). Its 50% lethal concentration (LC₅₀) was about 1,000,000-fold increased over that of susceptible C. pomonella, representing a C. pomonella population with one of the highest resistance levels ever observed in the field (12). The aim of the current study was to investigate the genetic basis and inheritance mode of CpGV resistance in SA-GO. Resistance tests were conducted to identify CpGV isolates with resistance-breaking properties for SA-GO. Single-pair crosses between field-collected individuals of SA-GO and susceptible C. pomonella (strain CpS) were carried out to obtain initial indications on the inheritance mode and the frequency of resistance in this field population. A two-step inbreeding of SA-GO individuals combined with simultaneous resistance testing of the offspring resulted in a genetically homogenous strain, CpRGO, with very low susceptibility to CpGV-M and CpGV-S. CpRGO was tested against further resistance-breaking CpGV genome groups to find suitable candidate CpGV isolates to be used as control agents for this type of resistance, whereas hybrid crosses and backcrosses between CpRGO and CpS unveiled the inheritance mechanism of this CpGV resistance. These investigations did not allow assigning resistance of CpRGO to either type I or type II but suggested a further resistance type with a highly complex mode of inheritance, which we termed type III resistance.

RESULTS

Resistance testing of SA-GO with different CpGV isolates.

Resistance of neonates of the field-derived C. pomonella population SA-GO to the isolates CpGV-M, -S, and -E2 was tested at a discriminating concentration of 5.8 × 10⁴ OB/ml, which would cause >95% virus-induced mortality in susceptible C. pomonella strains after 7 days (9); the susceptible laboratory strain CpS was included as a control (Fig. 1). Whereas virus-induced mortality of CpS larvae exposed to different CpGV isolates varied between 87% and 97% after 7 days and was more than 99% after 14 days, mortality of SA-GO neonates exposed to CpGV-M and CpGV-S was very low and did not exceed 9% for both viruses after 14 days. Significant mortality of SA-GO was observed only on exposure to CpGV-E2 with 28% mortality after 7 days and 81% mortality after 14 days. The differences of mortality were significant between SA-GO and CpS for all viruses after 7 and 14 days, except for CpGV-E2 (analysis of variance [ANOVA], pairwise t test, P < 0.05) (Fig. 1). These results indicated resistance of SA-GO to both CpGV-M and CpGV-S, whereas CpGV-E2 expressed full resistance-breaking characteristics in SA-GO only after 14 days.

FIG 1.

Mortality of neonate larvae of SA-GO and CpS tested for resistance on artificial diet containing either CpGV-M (M), CpGV-S (S), or CpGV-E2 (E2) at a discriminating concentration of 5.8 × 10⁴ OB/ml after 7 days (black bars) and 14 days (gray bars) of virus exposure. Given are the Abbott-corrected (38) mean mortality and the standard deviation (error bars) from three independent replicates. The total number of tested individuals per treatment is given below the chart. Columns marked by different letters differ significantly (ANOVA, pairwise t test, P < 0.05).

Resistance testing of F₁ progeny of the reciprocal single-pair crosses between field-derived SA-GO moths and susceptible CpS moths.

To obtain immediate information on the character of the resistance inheritance in SA-GO, reciprocal hybrid crosses between CpS moths and adults that had emerged from diapausing field larvae of SA-GO were performed. Neonates of the F₁ progeny were then subjected to resistance testing with a discriminating concentration of 5.8 × 10⁴ OB/ml. Virus-induced mortality of F₁ progeny of nine female crosses (SA-GO♀ × CpS♂) was between 0% and 34% after 7 days of exposure to CpGV-M, with a median of 4% (Fig. 2a). On CpGV-S, mortality ranged between 0% and 87%, with a median mortality of 20%. Similar mortality rates were observed for 24 single-pair crosses of the male crosses (SA-GO♂ × CpS♀) (Fig. 2a).

FIG 2.

Box plot analysis of the mortality of resistance testing of F₁ progeny of the female and male single-pair crosses between SA-GO and CpS. Mortality of neonate L1 larvae from single-pair hybrid crosses between individuals of field-collected SA-GO and the laboratory strain CpS tested for resistance after 7 days (a) and 14 days (b) on diet containing 5.8 × 10⁴ OB/ml of either CpGV-M or CpGV-S. Open boxes indicate the 25th to 75th percentile of the Abbott-corrected (38) mortality of all crosses, horizontal lines in the boxes give the medians, and vertical lines indicate the lower and upper ends of all observed mortality values. The numbers (n) of evaluated single-pair female crosses (SA-GO♀ × CpS♂) and male crosses (SA-GO♂ × CpS♀) are given on the top of the chart.

After 14 days, the F₁ progeny of the female crosses showed CpGV-M-induced mortality rates between 0% and 75%, with a median of 33%. On CpGV-S, mortality of female crosses ranged between 50% and 100% with a median of 62% (Fig. 2b). Mortality of the progeny of the male crosses was again similar to that of the female crosses. Full details of the crossing experiments are given in Fig. S1 in the supplemental material. The results of the single-pair crosses between CpS and SA-GO suggested a dominant resistance to both CpGV-M and CpGV-S in SA-GO; otherwise, a higher mortality close to 100% would have been expected after 7 and 14 days of virus exposure. Four female crosses showed 0% mortality, and no difference in mortality was observed between male and female crosses; both findings contradict the assumption of a Z linkage of resistance when mortality after 7 and 14 days was considered.

Establishment of CpRGO.

To further evaluate the inheritance of CpGV resistance in SA-GO, a genetically more homogeneous strain was established by two consecutive rounds of inbreeding using single-pair crosses (families) of SA-GO individuals combined with resistance testing of the offspring of the single families. Starting with 20 single-pair crosses, seven families produced sufficient F₁ offspring (20 to 60 neonates) to be further used for resistance testing and breeding. Mortality of the F₁ larvae exposed to CpGV-M or CpGV-S for 7 days was between 0% and 33%. The mean mortality of all tested siblings was 7% on CpGV-M and 9% on CpGV-S. After 14 days, the mean mortality increased to 19% for CpGV-M and to 11% for CpGV-S (Table S1).

All control groups of the different families were reared to pupae and separated by sex to undertake a second round of single-pair inbreeding. Only one family (family 12) produced enough F₁ moths to perform further four single-pair crosses (12.1, 12.2, 12.3, and 12.4). When offspring of these crosses were subjected to resistance testing, mortality ranged from 0 to 14% on CpGV-M and from 0 to 10% on CpGV-S at 7 days postinfection (Table 1); all offspring of the families 12.2 and 12.4 survived the treatment with CpGV-M, whereas mortality on exposure to CpGV-S was 0% and 6%, respectively, and did not increase to more than 18% for both viruses after 14 days. The offspring of the two families 12.2 and 12.4 was combined, further mass-reared, and termed CpRGO. Resistance testing of CpRGO progeny revealed mortality similar to that of the parental families 12.2 and 12.4 (Table 1). Eventually, CpRGO was reared in the laboratory without any selection pressure and further used for crossing and backcrossing experiments with CpS (see below). In addition, CpRGO individuals were frequently tested for their resistance level during 3 years of rearing; in 10 independent assays, mean mortality caused by CpGV-M and CpGV-S after 7 days was 11% and 20%, respectively, and increased further until 14 days (Table 1).

TABLE 1.

Mortality of first-instar larvae of F₁ family 12 after the single-pair inbreeding of SA-GO^a

Family or strain	CpGV-M			CpGV-S
	n	% mortality after:		n	% mortality after:
		7 days p.i.	14 days p.i.		7 days p.i.	14 days p.i.
12	22	5.3	19.5	20	32.3	47.9
12.1	29	13.8	20.7	29	10.3	20.7
12.2	28	0.0	13.5	34	6.1	17.6
12.3	18	11.1	11.1	17	0.0	15.4
12.4	7	0.0	10.0	4	0.0	0.0
Mean (SD) of values for 12.1 through 12.4		6.2 (6.3)	13.8 (4.2)		4.1 (4.4)	13.4 (8.0)
CpRGO	40	3.1	42.4	37	0.9	26.4
CpRGO (2013–2016)	273b			330
Mean (SD) for CpRGO (2013–2016)		11.3 (17.2)	28.0 (25.7)		19.6 (23.8)	33.9 (35.7)

^aLarvae were tested for resistance on artificial diet containing CpGV-M or CpGV-S at the discriminating concentration of 5.8 × 10⁴ OB/ml for 7 and 14 days postinfection (p.i.). Individuals of the untreated control cohort of family 12 were crossed to obtain F₂ offspring (families 12.1 to 12.4). Untreated control cohorts of families 12.2 and 12.4 (bold) were pooled to obtain strain CpRGO. n, number of tested larvae. Mortality data are corrected according to the Abbott formula with mortality of <6% (7 days) and <16% (14 days) in untreated control cohorts (38).

^bTen independent replicates.

Genetic inheritance studies.

Female and male single-pair crosses and backcrosses between CpRGO and CpS individuals were performed to determine (i) the mode of inheritance of the resistance to CpGV-M and CpGV-S and (ii) the sex ratio of the surviving pupae. Following an autosomal or Z-linked, dominant or recessive inheritance hypothesis, differences in the mortality rates in the crosses and backcrosses were expected (Fig. 3). Mortality was recorded after 7 days to assess larval resistance and after 21 days to measure pupation success.

FIG 3.

Inheritance schema for Z-linked and autosomal inheritance of resistance to CpGV. Hypothetically expected mortality after 7 and 21 days postinfection (p.i.) and ratio of surviving male pupae after 21 days of the offspring is given for neonate larvae with a dominant, monogenic inheritance of CpGV-M resistance, which were exposed to a discriminating concentration of CpGV-M. Red square, resistance gene on the Z chromosome; blue square, resistance gene on the autosome; A, autosome; Z and W, sex chromosomes; R, resistant; S, susceptible; ♂, male; ♀, female.

(i) Female crosses.

In total, 38 single female crosses (CpRGO♀ × CpS♂) were performed, of which nine crosses produced sufficient offspring for resistance testing. Seven of these female crosses (cluster A) showed similar responses; their mortality was below 50% (mean, 31%) after 7 days and increased to 47 to 76% (mean, 60%) after 21 days (Table 2). Two families did not follow this response pattern and were not considered further. It was found that 40% of the progeny exposed to CpGV-M successfully pupated until day 21; all of them were males (Table 2). This observation did not fully support autosomal inheritance of CpGV-M resistance, especially when pupation success is considered. Female crosses tested with CpGV-S drew a slightly different picture, though only three families produced enough individuals to complete these tests. Whereas mortality after 7 days was 43% and thus in the range of CpGV-M treatment, only 3 out of 60 tested individuals pupated; two of them were female. Hence, in contrast to the CpGV-M treatment, the rate of surviving pupae on CpGV-S was much lower; an influence of the sex on the pupation success was not determined.

TABLE 2.

Mortality of neonate larvae subjected to CpGV-M and CpGV-S at the discriminating concentration of 5.8 × 10⁴ OB/ml determined after 7 and 21 days p.i.^a

Cross type and cluster	CpGV-M				CpGV-S
	No. of crosses/cluster (no. of tested individuals)	Mean % mortality (SD) at:		% male pupae (no. of males:no. of females)	No. of crosses/(no. of tested individuals)	Mean % mortality (SD) at:		% male pupae (no. of males:no. of females)
		7 days	21 days			7 days	21 days
CpRGO♀ × CpS♂, A	7 (244)	31 (12)	60 (9)	100 (92:0)	3 (61)	43 (22)	95 (6)	33 (1:2)
CpRGO♂ × CpS♀
B	4 (179)	7 (3)	10 (7)	43 (33:44)	4 (151)	23 (10)	81 (17)	48 (10:11)
C	3 (143)	16 (23)	48 (3)	38 (27:45)	3 (138)	41 (45)	94 (7)	50 (4:4)
D	1 (32)	50 (—)	88 (—)	75 (3:1)	1 (33)	13 (—)	76 (—)	38 (3:5)

^aSingle-pair crosses were grouped in clusters (A to D) according to their mortality responses to CpGV-M. The rates of surviving male pupae were assessed after 21 days. —, not applicable.

(ii) Male crosses.

In total, 19 single-pair male crosses (CpRGO♂ × CpS♀) were performed. Eight single pairs produced sufficient offspring for resistance testing. Their mortality response to CpGV-M and CpGV-S was highly heterogenic and ranged from 2 to 50% and 4 to 91%, respectively, after 7 days. According to their response to CpGV-M, we grouped these crosses into three clusters (B to D). Cluster B exhibited a very low mortality of 7% and 10% after 7 and 21 days, respectively, whereas clusters C and D showed increased mortality but very high variation on both CpGV-M and CpGV-S. In contrast to the female crosses, the sex ratio of the surviving pupae on CpGV-M and CpGV-S was nearly equal (Table 2).

(iii) Backcross A.

In total, 37 families of the backcrosses A (BC A; F₁♂ × CpS♀) were set up, of which 19 single pairs with F₁♂ originating from clusters A, C, and D could be used for resistance testing (Table 3). Again, a considerable variation in the response to CpGV-M was observable, which could not be explained by either autosomal or Z-linked inheritance. All F₂ families derived from F₁♂ cluster A and half of cluster C (termed Ca in Table 3) showed similar mortality rates of 51% and 59%, respectively, after 7 days, and a slight increase of mortality until pupation. In contrast, the other three families derived from F₁♂ of cluster C (termed Cb in Table 3) and progeny of cluster D showed 100% mortality already after 7 days, without any surviving pupae. When cohorts of neonates of the same BC A backcrosses were subjected to CpGV-S, the 7-day mortality was 43% for progeny derived from F₁♂ cluster A and 66 to 75% for those from clusters Ca, Cb, and D (Table 3). Here, pupation was observed in only cluster A, which showed a balanced sex ratio of pupae.

TABLE 3.

Mortality of the offspring of the backcrosses BC A and BC B exposed to diet containing CpGV-M and CpGV-S at the discriminating concentration of 5.8 × 10⁴ OB/ml determined after 7 and 21 days p.i.^a

F1 from cluster	CpGV-M				CpGV-S
	No. of crosses/cluster (no. of tested individuals)	Mean % mortality (SD) at:		% male pupae	No. of crosses/cluster (no. of tested individuals)	Mean % mortality (SD) at:		% male pupae
		7 days	21 days			7 days	21 days
BC A; F1♂ × CpS♀
A	11 (406)	51 (13)	77 (10)	38	8 (242)	43 (24)	82 (21)	45
Ca	3 (111)	59 (14)	66 (9)	33	3 (119)	66 (17)	100 (0)	—
Cb	3 (126)	100 (0)	100 (0)	—	3 (138)	72 (12)	98 (2)	—
D	2 (93)	100 (0)	100 (0)	—	2 (90)	75 (2)	100 (0)	—
BC B; F1♂ × CpRSO♀
A	4 (141)	31 (21)	58 (17)	63	3 (48)	48 (28)	100 (0)	—
B	2 (32)	19 (15)	39 (13)	50	2 (29)	23 (21)	ND	—

^aThe F₁ male (♂) moth used for backcrosses originated from the control group of the female (♀) or male single-pair crosses between CpRGO and CpS (Table 1). Resulting F₁ males were reared to adulthood, and moths were backcrossed by single-pair crosses with female CpS moths (BC A; F₁♂ × CpS♀) or CpRGO female moths (BC B; F₁♂ × CpRGO♀). The rates of surviving male pupae were assessed after 21 days. —, not applicable; ND, not determined.

(iv) Backcross B.

Six out of 18 single-pair families set up in backcross B (BC B; F₁♂ × CpRSO♀) could be used for resistance testing (Table 3). The BC B families derived from clusters A and B showed similar responses, which were far below 50% mortality after 7 days of exposure to CpGV-M, a rate that would have been expected if resistance in CpRGO followed a classic monogenic Z-linked or autosomal inheritance. On CpGV-S, only progeny from cluster A showed mortality of 50%; however, all individuals died before pupation. Progeny from cluster B showed a low mortality rate of 23%.

Comparative resistance testing with different CpGV isolates in C. pomonella strains.

To obtain a more comprehensive picture of how CpRGO differs in its resistance from other resistant C. pomonella strains, neonate larvae of susceptible CpS and the four resistant strains CpRR1 (type I resistance), CpR5M and CpR5S (type II resistance), and CpRGO (proposed type III resistance) were systematically tested for their responses to five different CpGV isolates, representing the CpGV genome groups A to E. All tests were carried out at the discriminating concentration of 5.8 × 10⁴ OB/ml and were evaluated after 7 and 14 days (Fig. 4).

FIG 4.

Mortality of neonate larvae of the susceptible C. pomonella strains CpS, CpRR1 (type I resistance), CpR5M and CpR5S (type II resistance), and CpRGO tested for resistance on artificial diet containing either CpGV-M (M), CpGV-S (S), CpGV-E2 (E2), CpGV-I12 (I12), or CpGV-I07 (I07) at a discriminating concentration of 5.8 × 10⁴ OB/ml. Abbott-corrected mean mortality (38) and standard deviations (error bars) were determined 7 days (a) and 14 days (b) postinfection. Isolate names and corresponding genome groups, total number of tested individuals (n), and number of independent replicates (N) are given below the chart. Columns marked by different letters differ significantly (ANOVA, Tukey HSD test, P < 0.05).

In CpS, mortality ranged from 76% (CpGV-E2) to 97% (CpGV-I07) after 7 days and was higher than 97% for all CpGV isolates after 14 days (Fig. 4a and b). Thus, all tested CpGV isolates were infective for CpS. For CpRR1 (type I resistance), low mortality of 5% and 11% was observed only on CpGV-M after 7 and 14 days, respectively. In contrast, treatments with CpGV-S, -E2, -I12, and -I07 caused significantly higher larval mortality after 7 and after 14 days (ANOVA, post hoc Tukey honestly significant difference [HSD] test, P < 0.05), corroborating previous findings that type I resistance is targeted exclusively to CpGV-M (6). In CpR5M and CpR5S (type II resistance), mortality was low for CpGV-M, -S, and -I12 and did not exceed 23% after 7 days and 50% after 14 days. Only isolates CpGV-E2 and -I07 caused significant mortality of >80% after 7 days and of >95% after 14 days.

For CpRGO, the virus-induced mortality caused by CpGV-M and CpGV-S was 12% and 19%, respectively, after 7 days, and increased to a maximum of 37% (CpGV-S) after 14 days.

These mortality rates were very similar to the mean mortality observed for CpRGO during a 3-year survey (Table 1). Treatment with CpGV-E2, -I12, and -I07 resulted in mortality of CpRGO larvae between 64% and 91% after 7 days and increased to 87 to 99% until 14 days. The mortality in the treatments with CpGV-M and CpGV-S differed significantly from that in the treatments with CpGV-E2, -I12, and -I07 (ANOVA, pairwise t test, P < 0.05). In summary, CpRGO showed a response that was again different from C. pomonella strains representing type I and type II resistance.

DISCUSSION

In the present study, the resistance of the German field population SA-GO exhibiting an extreme level of resistance to CpGV-M (12) was analyzed. Infection experiments showed that SA-GO larvae were not susceptible to both CpGV-M and CpGV-S, concluding that the resistance in SA-GO is directed against at least two different genome groups, A and E, of CpGV (6, 23). Hence, resistance in SA-GO appears to be different from type I resistance reported previously for C. pomonella in Europe (9, 11, 12, 26) and seems more to resemble type II resistance, which is described as a cross-resistance against CpGV-M and CpGV-S (13, 29).

The results of the single-pair crosses between individuals of SA-GO and susceptible CpS provided initial evidence of a dominant and autosomal inheritance (Fig. 2; see also Fig. S1 in the supplemental material). Otherwise, higher mortality of the hybrid progeny should have occurred and reciprocal male and female hybrid crosses should have shown different mortality rates, which was not the case. In addition, zero mortality was observed in several SA-GO♀ × CpS♂ crosses (Fig. S1), which is also not compatible with a Z linkage of the resistance gene(s).

To gain more insight into the resistance of SA-GO, the genetically more homogenous strain CpRGO was selected from SA-GO by a two-step inbreeding process with single-pair crosses. At first, mortality of CpRGO and its parental families 12.2 and 12.4 was close to zero on both CpGV-M and CpGV-S, indicating strain homogeneity and success of the selection process; during 3 years of rearing of CpRGO without selection pressure, an increase of susceptibility was noted, suggesting some plasticity of resistance within CpRGO.

Hybrid crossing experiments and backcrossing experiments with CpRGO and CpS individuals confirmed some of our initial predictions of a dominant and autosomal inheritance that was already made with parental SA-GO. However, extended observation of infected animals for 21 days and thorough analyses of the mortality of siblings from single-pair crosses generated a much more complex picture of CpGV resistance in CpRGO (Tables 2 and 3). (i) Strikingly, female crosses resulted in significant mortality after 7 days on both CpGV-M and CpGV-S. This observation would be compatible with either Z linkage or an autosomal inheritance of heterozygous individuals. Z linkage was supported for resistance to CpGV-M because only males pupated. This finding could not be confirmed for the resistance to CpGV-S, because only a very few male and female pupae survived in these experiments. The outcome of the male crosses was even more puzzling because different response clusters (B to D) were observed, showing a highly heterogenic reaction among the different families. (ii) The backcrossing experiments again did not match with a Z-linked or an autosomal inheritance hypothesis. Even more confusing, the Z-linkage signal observed for CpGV-M resistance of the female cross in cluster A fully disappeared in BC A, and evidence for an autosomal inheritance was noted. Within BC A, the progeny of the male cross cluster C split into two response groups (Ca and Cb) for CpGV-M but not for CpGV-S, supporting different modes of inheritance for CpGV-M and CpGV-S resistances. (iii) In general, only a few males and female pupae developed on CpGV-S, indicating a variation in the efficacy of CpGV-M and CpGV-S in both infection process and sex selection. No correlation or a very low correlation in the mortality responses between treatments with CpGV-M and CpGV-S was detected in the hybrid crosses and backcrosses, also suggesting a somehow different and heterogeneous action of resistance in CpRGO for these two CpGV genome groups. To summarize, the results of the different hybrid crossing and backcrossing experiments did not support classic monogenic Z-linked or autosomal inheritance, as was clearly observed for type I and type II resistance (9, 13, 29). In addition, cross-resistance of CpRGO to CpGV-M and CpGV-S followed different inheritance patterns, which was also not the case for type II resistance (29).

Considering that CpRGO originated from only a single pair (pair 12) of C. pomonella individuals, the heterogeneity observed in the F₁ crosses and the backcrosses is unexpected, especially after mortality of early generations of CpRGO was very low. However, by assuming the combined action of a Z-linked and an autosomal genetic factor, most of the observed crossing results can be interpreted. We hypothesize, therefore, that resistance to CpGV-M is based on at least two loci on an autosome and the Z chromosome and depends on a sex and dose effect (two-gene hypothesis). By taking all results together, it is proposed that survivorship after 7 days of the crosses needs at least one resistance gene either on an autosome (A^R) or on the Z chromosome (Z^R) (Fig. 5). Individuals with only one resistance gene on the autosome but missing a second resistance gene on the Z chromosome are able to survive 7 days but do not develop to pupae. This is the case for F₁ females from the female crosses, which obtained their Z^S chromosome from a susceptible CpS male. Our hypothesis further includes a certain increase of the number of dead larvae until day 21 caused by long-term virus exposure (Fig. 5). An increase of larval mortality with extended virus exposure appears to be typical for CpGV resistance; it was also observed for type I and type II resistance (9, 13, 21).

FIG 5.

Hypothetical inheritance schema for the two-gene hypothesis for the resistance to CpGV-M in CpRGO. Hypothetically expected mortality of the offspring originating from female or male crosses (a) and backcrosses of CpRGO and CpS (b) exposed to the discriminating concentration of CpGV-M. Given are the expected mortality after 7 and 21 days postinfection (p.i.) and ratio of surviving male pupae. Parental and progeny chromosomal genotypes are shown for a dominant, polygenic, mixed (Z-linked and autosomal) inheritance for resistance to CpGV-M. A, autosome; Z and W, sex chromosomes; CL, cluster; F₀, parental moth; F₁, progeny of the crosses; F₂, progeny of crosses of the second generation; red square, resistance gene on the Z chromosome; blue square, resistance gene on the autosome. Expected mortality is estimated for the hypothesis that offspring with one resistance gene survive at least through the 7-day bioassay but die before 21 days p.i. Progeny with one resistance gene on the autosome and one resistance gene on the Z chromosome mostly survive 21 days p.i. and develop to pupae. Clusters represent crosses from Tables 2 and 3 related to the hypothetical chromosomal genotypes.

Our two-gene hypothesis is also compatible with the hypothetical mortality of the F₁ progeny of homogeneous resistant females crossed with susceptible CpS males (Z^RW A^RA^R × Z^SZ^S A^SA^S = Z^SW A^SA^R, Z^RZ^S A^RA^S), which fits all observed mortality of cluster A. For the male crosses, F₁ offspring were clustered in three groups (B to D). The results for cluster B can be explained by selecting a homogeneous male moth (Z^RZ^R A^RA^R) and C and D for two different genotypes of heterogeneous individuals (Z^RZ^S A^RA^R or Z^RZ^S A^RA^S). Because clear chromosomal genotypes were generated for the F₁ progeny of the crosses, most of the backcrossing results are explainable: the mortality of BC A, F₂ offspring originating from clusters A, B, and Ca, fits the expected mortality because the selected F₁ male moth had a homogeneous genotype (Z^RZ^S A^RA^S). If an F₁ male with one resistant gene (Z^SZ^S A^RA^S) from cluster Cb was randomly selected, mortality up to 100% was determined as shown in Fig. 5. The hypothesis also matched most of the observed results for BC B.

These results suggest that the original number 12 pairs used to establish CpRGO did not only consist of fully genetically homogeneous resistant individuals (Z^RW A^RA^R, Z^RZ^R A^RA^R) but were heterozygous in their Z-linked and/or autosomal component. Such heterogeneity also explains the observed decline of resistance during further breeding without selection pressure of CpRGO, which might have been caused by fitness costs or genetic drift. Therefore, the two-step inbreeding clearly helped to identify some genetic traits of the resistance in CpRGO, but it was not successful in generating a genetically fully homozygous line.

For the CpGV-S treatments, low mortality after 7 days but a strong increase in mortality until 21 days was detected for nearly all crosses and backcrosses. These findings were independent of the (number of) resistance allele(s), suggesting a more recessive nature of inheritance of the resistance in CpRGO to CpGV-S, when pupation success of the hybrid offspring is considered. No sex linkage was observable for CpGV-S. In heterozygous offspring, CpGV-S appeared to be infective although the death of the larvae was highly delayed compared to susceptible larvae.

Taking the results together, neither a Z-linked nor an autosomal inheritance alone can explain the highly complex mortality patterns of the crosses and backcrosses observed in this study. Indeed, statistical tests supported a polygenic inheritance pattern in the majority of the backcrosses for the resistance to both viruses (data not shown). These observations are in clear contrast to the monogenic mode of inheritance suggested for type I resistance to CpGV-M in CpRR1 (9, 11, 26) and for type II resistance in CpR5M/CpR5S (29). Interestingly, some circumstantial evidence for an autosomal and Z-linked inheritance was previously suggested for the French strain field population St-Andiol, though this trait was also lost during selection experiments, producing only Z linkage of the descendant strain RGV (22).

The observed combination of autosomal and Z-linked inheritance observed for CpRGO is unique. So far, reports on genetically based baculovirus resistance have referred mainly to laboratory selections (14–18). Inheritance patterns had been determined for baculovirus resistance in C. pomonella (9, 11, 22, 29) and P. operculella (14, 30); it was generally monogenic and dominant and either Z linked (C. pomonella) or autosomal (C. pomonella and P. operculella). Recently, a laboratory strain of A. honmai highly resistant to AdhoNPV was selected (18, 19), and a midgut-based mode of resistance to oral infection by AdhoNPV OB was identified. It is thus obvious that baculovirus resistance does not follow a common mode but varies among different baculovirus-host systems.

CpRGO shows a new and highly complex inheritance pattern that is different from type I and type II resistance in C. pomonella. Regarding its inheritance, it is not fully clear whether this new type of resistance, to which we refer as type III resistance, is functionally independent from type I and type II resistance or whether it is a combination of the two. A systematic comparison of all available resistant C. pomonella strains with CpGV isolates belonging to different genome groups revealed susceptibility of CpRGO to CpGV-E2 (genome group B), -I12 (D), and -I07 (C) (Fig. 4). Thus, in addition to the differing inheritance, also the response of CpRGO to different CpGV isolates does not comply with that of type I and type II resistant C. pomonella strains and lacks the strong correlation of cross-resistance to CpGV-M and CpGV-S that was found with type II resistance. These variations might be an effect of differences in the resistance mechanism(s) and/or a consequence of quantitative effects based on the numbers of expressed resistance genes in the hosts. Such quantitative effects could also explain why the level of resistance to CpGV-M observed in the original field population SA-GO (12) was much higher than in other tested C. pomonella populations. In any case, the startling complexity of CpGV resistance amounting to C. pomonella populations with highly different susceptibilities and inheritance of resistance provides a unique model to study the potential of host adaptation to baculovirus infections in the field. Understanding the underlying genetic factors of CpGV resistance will be crucial to sustaining the efficacy of CpGV isolates in the field.

MATERIALS AND METHODS

Insects.

The codling moth (C. pomonella) strain CpS is susceptible to all known CpGV isolates and has been reared at the Julius Kühn Institute (JKI), Institute for Biological Control, in Darmstadt (Germany), for many years (9). The C. pomonella field population SA-GO derived from an organic orchard in Saxony, Germany, where CpGV-M products had been applied in the past. In autumn 2008, cardboard strips were wrapped around tree trunks to collect diapausing larvae, which were reared in the laboratory subsequently. Resistance to CpGV-M in this orchard was already demonstrated by full-range bioassays (12). CpRR1 carries type I resistance to CpGV-M and arose from the resistant field population CpR (BW-FI-03, ‘Sudbaden’) by single-pair crosses (9). The two C. pomonella strains CpR5M and CpR5S originated from the field population NRW-WE, which was selected on either CpGV-M or CpGV-S for five generations; CpR5M and CpR5S show cross-resistance to both CpGV-M and CpGV-S and showed a second type of resistance (type II resistance) (13, 29). All C. pomonella strains were reared under laboratory conditions at 26°C with a 16-/8-h light-dark photoperiod and 60% relative humidity on a modified artificial diet (31).

Viruses.

Five different Cydia pomonella granulovirus (CpGV) isolates were used in this study. CpGV-M (4) belonged to genome group A, isolate CpGV-E2 (32) belonged to genome group B, and CpGV-S (6) belonged to genome group E. Both isolate CpGV-I07 (genome group C) and isolate CpGV-I12 (genome group D) originated from Iran (23, 33). CpGV occlusion bodies (OB) were purified as described previously (34, 35), and all samples were stored at −20°C. Quantitation of virus stocks was performed by OB counting with a light microscope (Leica DMRBE) using dark-field optics with the Petroff-Hauser counting chamber (depth, 0.02 mm).

Resistance testing.

To differentiate between resistant and susceptible individuals, first-instar larvae were tested for resistance as described elsewhere (36). For resistance testing of different CpGV isolates, purified OB were mixed with a diet modified from the work of Ivaldi-Sender (31) to obtain a final discriminating concentration of 5.8 × 10⁴ OB/ml, which causes >95% mortality in susceptible CpS neonates after 7 days (9). For untreated control, larvae were reared on a diet without virus. Mortality of larvae was determined 7, 14, and 21 days postinfection (p.i.).

For resistance testing in the crosses and backcrosses, all larval offspring were separated into three cohorts: (i) neonate larvae were tested for resistance on an artificial diet containing CpGV-M; (ii) larvae were exposed to a diet with CpGV-S; and (iii) for an untreated control, larvae were reared on a diet without virus. For determining the rate of surviving male individuals, surviving pupae were sexed after 21 days according to the number of their abdominal segments described previously (37).

Single-pair crosses.

For single-pair crosses (families), pupae of SA-GO were sexed as described before (37), one virgin female moth was mated with a male moth, and the obtained eggs were collected daily and stored at 4°C until the end of oviposition. Then, all eggs of a family were incubated at 26°C until hatching of the neonates. Single-pair families producing fewer than 20 neonates were excluded from the resistance testing. For details of single-pair crosses, see the work of Asser-Kaiser et al. (9).

Reciprocal single-pair crosses.

In order to obtain information about the mode of inheritance in both SA-GO and CpRGO, reciprocal single-pair crosses between one of the two resistant strains and the susceptible strain CpS were conducted. Two different crosses were performed; female crosses represent the mating of one virgin female (♀) moth of SA-GO or CpRGO with a male (♂) moth of CpS (SA-GO♀ or CpRGO♀ × CpS♂). Male crosses indicate the crosses of a male moth of SA-GO or CpRGO mated with a virgin female moth of CpS (SA-GO♂ or CpRSO♂ × CpS♀). The resulting offspring of the reciprocal male or female crosses (F₁) were divided into three cohorts and subjected to resistance tests on CpGV-M, CpGV-S, and untreated control as described above. If the number of F₁ neonates was fewer than 90 individuals, only one virus (CpGV-M or CpGV-S) treatment and the untreated control group were included.

Backcrossing experiments.

The F₁ control group of the male or female crosses were reared to pupae and separated by sex. The hatching male moths were individually backcrossed with female CpS moths (backcross A [BC A], CpRGO♀ × CpS♂) or with a virgin female moth of CpRGO (backcross B [BC B], CpRGO♂ × CpS♀). The resulting offspring (F₂) of BC A or BC B were tested for resistance as described above.

Statistical analysis.

The observed virus-induced mortality in the resistance testing was corrected by the mortality of the untreated control group according to Abbott's formula (38). Statistical differences in the resistance testing with the field resistance strain SA-GO and strain CpS were analyzed using ANOVA (pairwise t test). Differences between different C. pomonella strains and treatments in the resistance testing were statistically evaluated using ANOVA and post hoc Tukey HSD test of the means (RStudio edition 2.3.4.4).