Introduction

Genetic transformation of insects may lead to control or reduced pest status of species destructive of food and fiber (DeVault et al., 1996). A transgenic system for Heliothis virescens is needed to explore genetic control or to unravel the complexities of insecticide resistance, a phenomenon for which this pest is among the most notorious in agriculture (Brown, 1996). Organophosphorus insecticide resistance in this species has been associated with the gene AceIn (Brown and Bryson, 1992) which was genetically linked to methyl parathion resistance (Gilbert et al., 1996) and is located on chromosome 2 (Heckel et al., 1998). Pyrethroid insecticide resistance appear to be associated with the gene hscp (Taylor et al., 1993) in which there are several point mutations (Park et al., 1998; Lee et al., 1999).

While genetic transformation of dipterans has become routine, it was not until very recently that technology was developed to express a foreign gene in non-dipteran insects (Handler, 2000). Lepidopterans were transformed using piggyBac transposon as a vector for the transgenic marker green fluorescent protein (Peloquin et al., 2000; Tamura et al., 2000). This promising vector has not been tested in H. virescens or in other noctuids. Transformation of dipterans was developed by exploiting recessive eye color mutants beginning with rosy of Drosophila melanogaster (Rubin and Spradling, 1982). Transformants were identified by expression of the transgenic wild type alleles such as w+ (Komori et al., 1993), v+ (White et al., 1996) or cn+ (Jasinskiene, 1998) that restored color in the mutant background. Recessive eye color mutants for lepidopterans have been reported (Dittrich and Luetkemeier, 1980; Marec and Shvedov, 1990), but are not understood biochemically. Recently, the gene Bmwh3, similar in sequence and inferred protein structure to D. melanogaster w, was cloned from Bombyx mori and its expression was linked to w3 and w3ol mutantions (Abraham et al., 2000).

We describe a new visible marker in H. virescens, yes, for yellow eye and yellow scale that arose spontaneously in our strain that also carried a visible autosomal recessive marker, y, for yellow scale (Mitchell and Leach, 1994). We report the independent assortment of y, yes, AceIn and hscp, the latter two of which confer resistance to insecticides. We present a hypothesis to explain the control of pigmentation by y and yes with a view toward future transgenic technology. Our results characterize the most convenient and visible genetic landmarks, those of visible mutants which are rare in the species. We intend to find molecular markers linked to each visible marker in order to locate the genes controlling pigment deficiencies in a library of bacterial artificial chromosomes.

Materials and Methods

Results

All crosses between double mutant Yyes and Dalzell98 produced all wild type progeny indicating recessive inheritance of the trait yellow eye and confirming the recessive inheritance of yellow scale (Mitchell and Leach, 1994). We assumed a single gene controlled each trait. Average emergence of adults from pupae was 89 ±12%. There was no significant difference in results when separated by sex, therefore, pooled results are presented.

Independent assortment between y and yellow eye was proven by the presence of progeny of the recombinant class with yellow scales and grey eyes among backcross progeny (Table 1), however, neither the expected model for independent assortment, nor the expected model for linkage fit our observations (Tables 1 & 2). The recombinant class with yellow scales and grey eyes was found in the expected proportion of 0.25 in addition to the two parental classes. Had y been linked to a single gene for yellow eye, this recombination would not be expected (Experimental design and analysis). Confirmation of independent assortment was found in the observation of this recombinant class in the expected proportion in the dihybrid cross (Table 2).

Absent in both the test cross and dihybrid cross was the second expected recombinant class with green scales and yellow eyes (Tables 1 & 2). Also, the proportion of the parental class with yellow scales and yellow eyes was increased beyond the expected proportion by an amount equal to the missing class with green scales and yellow eyes. Observations did not fit a model in which y+ y, yellow eyed was lethal.

The observed data fit a model in which a single gene, independent of y, conferred both yellow eye and yellow scale, suggesting that two unlinked genes conferred yellow scale while one of those genes also controlled yellow eye (Tables 1 & 2). According to this model, y+ y, yellow eyed moths would express yellow scales due to the same gene that conferred yellow eye. The expected recombinant y+ y, yellow eyed moths would actually occupy the parental class with yellow scales and yellow eyes.

We conclude that a single gene, yes, confers not only yellow eye, but also yellow scale. It does so via a mechanism that cannot be overcome by one allele of y+ background; i. e., both genes y and yes confer yellow scales. While this result could also be due to two linked genes in the backcrosses, there were no recombinants with green scales and yellow eyes observed among 516 progeny in the intercross experiments in which recombination was possible in heterozygous fathers (Table 2).

When considered independently from yellow eye, yellow scale was found at nearly 3:1 over green scale (Table 3), which is predicted by our model in which yes confers not only yellow eye, but also yellow scale. Considered independently from y, yellow eye was found at a ratio of 0.98: 1 with the wild type eye indicating that eye color was controlled by the single, recessive gene, yes (Table 3). This was confirmed in dihybrid crosses in which the observed count of yes yes to yes+ yes was 122 to 394 (expected ratio 129: 378).

No genetic linkage was observed between y and AceIn, as indicated by the presence of 23 parental phenotypes and 19 recombinant phenotypes in two backcross experiments involving strain Yel (Figure 3). A model for independent assortment of two genes predicted a 1:1:1:1 ratio in the four possible phenotypes: yellow scale, AceIn+ AceIn+; green scale, AceIn AceIn+; yellow scale, AceIn AceIn+; and green scale, AceIn+ AceIn+. In the backcross (Ace-R × Yel) × Yel the observed count of 6; 8; 7; 5 moths among the phenotypes, respectively, did not differ from the model of independent assortment (?2 = 1.5768, p<.05, df = 3). In the backcross (Pyr-R × Yel) × Yel the observed count of 4; 5; 4; 3 moths among the phenotypes, respectively, did not differ from the model of independent assortment (?2 = 0.5, p<.05, df = 3).

The ratio of green to yellow was 13:13 in the backcross (Ace-R × Yel) × Yel and 8:8 in the backcross (Pyr-R × Yel) × Yel confirming the hypothesis that a single gene encodes the yellow trait (Mitchell and Leach, 1994). Also, the ratios of 15 AceIn AceIn+ : 11 AceIn+ AceIn+ in the backcross (Ace-R × Yel) × Yel and 9 AceIn AceIn+ : 7 AceIn+ AceIn+ in the backcross (Pyr-R × Yel) × Yel confirmed that a single gene AceIn controlled acetylcholinesterase insensitivity to propoxur (Brown and Bryson, 1992).

Of the progeny from the control group, Ace-R × Ace-R/YEL, 27 were green and one was yellow. The yellow moth in the control group was unexplained, but it was likely due to human error when pupae were sorted into containers for emergence. The other explanation was that there was a yellow allele in the Ace-R strain at a frequency of 0.04 which was very unlikely in that no yellow mutants had been observed in this lineage for over ten years. These observations confirmed the recessive inheritance of yellow scale as y.

AceIn was genetically linked to isocitrate dehydrogenase-2 (IDH-2) placing its locus on chromosome two (Heckel et al., 1998). AceIn is linked to increased survivorship (resistance) of larvae exposed to methyl parathion and resistance is completely dominant, being fully expressed in heterozygotes that are AceIn AceIn+ (Gilbert et al., 1996). This linkage is consistent with the model of a structural mutation giving a resistant enzyme that confers complete resistance from one allele. That is, a half dose of the AceIn enzyme can function in the presence of methyl parathion to restore synaptic responsiveness. However, evidence for a structural gene mutation model is lacking until the acetylcholinesterase gene from tobacco budworm can be cloned and sequenced. Independent assortment of yes from AceIn and from hscp was observed in one family of the test cross Dalzell98/Yyes × Yyes (Table 4). This family was informative for genes y, yes, AceIn and hscp as seen in the pedigree.

The AceIn allele from the heterozygous mother, absent in the father, was found in five grey-eyed progeny of the parental type, but it was observed also in four yellow-eyed progeny which were of the recombinant type (Table 4). Because recombination was prohibited in this backcross (see Experimental Design), this result demonstrated independent assortment. Likewise, the hscp allele CAT present in the heterozygous mother, but absent in the father, was found in two grey-eyed progeny, but it was observed also in three yellow-eyed progeny proving the independent assortment of yes from hscp (Table 3). The hscp allele CAT was found to segregate independently of scale color; one greenscaled progeny inherited this allele from the mother, but three green-scaled progeny inherited the CTT allele from the mother and were thereby recombinant type (Table 4). The hscp allele CAT was found to segregate independently of AceIn as there were four recombinant progeny exhibiting the genotype AceIn AceIn+, CTT/CTT (Table 4). In this family, one recombinant type for pigmentation was observed (yellow scale with grey eye), but the other recombinant type was absent (green scale with yellow eye) (Table 4). These data were typical of the model described above for two genes controlling yellow scale, one of which also controlled yellow eye (Table 1 and 2).

Discussion

In D. melanogaster, some eye color mutations result from deficiencies in specific enzymes of biosynthesis (Sarkar and Collins, 2000) while other mutations are due to deficient ATPbinding cassette (ABC) transporters of precursors of pigments, or in the transport of pigment granules (Lloyd et al., 1998). In H. virescens, our double mutant adult was not devoid of color, but retained a brilliant yellow scale color and a greenish-yellow eye color versus the wild type in which scales were green with black bands and eyes are grey (Figure 1). The source of the yellow color could be the synthesized compound xanthopterin, a guanosine derivative known to impart yellow color to lepidopteran scales. Pterins are also found in compound eyes and they are cofactors in synthesis of another group of pigments, the ommochromes. Synthesized from tryptophan through kynurenine, xanthommatin is a very common red or brown pigment of insect compound eyes and deficiencies in this synthetic pathway are the basis of several eye color mutants of D. melanogaster, of several other dipterans, and of B. mori.

A black pigment synthesized in insects is melanin, a polymer of indole-5,6-quinone derived from tyrosine. This black pigment, when deposited with the background yellow pigment in scales, could result in the green scale color and black bands on wings observed in the wild type.

In D. melanogaster, some eye color mutations result from the loss of ATP-binding cassette (ABC) transporter proteins which appear to be responsible for the active transport and availability of precursors such as tryptophan or guanine (Ewart and Howells, 1998). These transporters are heterodimers so that the transporter of tryptophan is composed of the products of w+ and st+, while the transporter of guanine is composed of the products of w+ and bw+. White eye was displayed in homozygous w fruit flies or in homozygous st bw fruit flies (Sarkar and Collins, 2000).

There are several ways to interpret these results including that a single gene coordinately regulates transcription of genes in both the melanin and ommochrome pathways, or that a single gene produces an inhibitor of enzymes in both the melanin and ommochrome pathways. We favor the simplest interpretation based on known mechanisms which is that an ABC transporter limits the precursors to both melanin and ommochrome. It is known that point mutations in the white gene of D. melanogaster can decrease both red and brown pigment putatively impairing both tryptophan and guanine transport (Mackenzie et al., 1999).

In H. virescens, assuming that the background yellow color is due to xanthopterin, our yes mutant might be analogous to the yellow eye mutant of Ephestia kuehniella which complemented white eye and red eye (Marec and Shvedov, 1990), or to the st mutant of D. melanogaster, that has lost transport of tryptophan for ommochrome synthesis while retaining transport of guanine for pterin synthesis. The observed interaction in H. virescens in which green and black scales are not found in yellow-eyed progeny as should be expected from segregation of these genes, could be explained by the simultaneous loss of tyrosine transport and tryptophan transport so that both ommochrome and melanin synthesis were prevented, even in the presence of y+ alleles, leaving only xanthopterin yellow synthesis to proceed. On the other hand, presence of yellow scales does not prevent expression of the wild type eye color and this suggests that the mutation for yellow scale is specific for melanin and does not affect ommochrome synthesis.

Although it has been conventional to reject linkage and accept independent assortment upon the first observation of recombination in test cross progeny from diheterozygous females (Shimada et al., 1994), this unusual case in which one recombinant class was observed in full proportion, but the other recombinant class was absent requires that the alternate interpretation be explored as follows: the lack of the recombinant class with green scales and yellow eyes could have been due to linkage of y to yes and half the expected parental wild type class y+ y, yes+ yes somehow expressed yellow scale to produce the then unexpected recombinant class with yellow scales and grey eyes. We cannot provide a logical biochemical mechanism for this alternative explanation of the test cross results. We conclude that there was no linkage between y and yes based on the presence of recombinant progeny.

We can test this hypothesis by repeating the test cross and mating progeny of the parental double marker class to y y, yes+ yes+ moths. The hypothesis predicts that one -half parental double marker class progeny are actually y+ y, yes yes rather than y y, yes yes as they appear; therefore, one-fourth the proposed matings would produce all wild type progeny. If the alternate hypothesis is correct, then no matings of this type would produce wild type progeny. We have purified a line of y y, yes+ yes+ to make this test possible.

Another yellow eye mutant has been observed independently in a strain cultured in Georgia (Hasty and Payne, 1999). This mutant was controlled by a single, autosomal, recessive gene and was found in moths having wild type scales. In Yyes and in the crosses described herein, we have not observed yellow eye in a moth with wild type scales. Our hypothesis predicts that this is a separate locus from yes or a different allele that does not simultaneously control yellow scale. Test crosses for complementation revealed complementation of yes and this second gene for yellow eye, ye (Cho et al., unpublished observations). We have purified inbred lines of yellow scale, grey eye, constructed another strain of yellow scale with yellow eye from this mutant and are testing the segregation of the genes y from ye.

Development of selectable markers for transgenesis in this species should be directed to finding the wild type allele of an enzyme involved in pigment synthesis. This gene could then be exploited as a marker in vectors when transforming a visible mutant strain. Our results also suggest possible pitfalls in such research when visible marker genes interact to yield unexpected results. In particular, we suggest the possibility that restoration of an enzyme activity as likely conferred by an allele such as y+ might not yield the expected transgenic phenotype for visible selection in the presence of a deficiency in a transporter. To our knowledge, this is the first report of a single gene controlling two traits in this species and it is the first report of inheritance relationships of both visible mutations and biochemical markers in this species.

In conclusion, two genes conferring pigmentation and two genes conferring insecticide resistance assorted independently indicating that the genes y, yes, AceIn and hscp occupied four separate autosomes among the 31 chromosomes of H. virescens. These fo7ur genes now become convenient markers for future linkage mapping, y and yes because they are visible as homozygotes. AceIn and hscp are convenient due to their clear codominant expression in a simple biochemical assays as described herein. Although these visible mutants are not known from wild populations at this time, AceIn was present at an average frequency of 14% (Brown et al., 1996b) and hscp1029H was observed at approximately 20% (Park, 1998) in collections from agricultural fields. It should be easy to capture these alleles for future genetic linkage studies.