Seven P450 genes distributed in the CYP6B, CYP4M, and CYP321A subfamilies have been cloned from H. zea 212223. Previous studies have demonstrated that CYP321A1 is highly inducible by xanthotoxin 23 and the four CYP6B transcripts accumulate to varying degrees in response to a range of allelochemicals naturally encountered in hostplants (xanthotoxin, indole-3-carbinol, chlorogenic acid, flavone) as well as in response to synthetic chemicals not naturally encountered (cypermethrin, phenobarbital) 2122. The CYP6B transcripts are also inducible by plant defense signalling molecules jasmonate and salicylate, allowing this species to "eavesdrop" on plant defense signals for activating detoxification systems in advance of induced biosynthesis of host plant toxins 41. Baculovirus-mediated expression of the CYP6B8 and CYP321A1 proteins has directly demonstrated that they are capable of metabolizing a range of allelochemicals (xanthotoxin, chlorogenic acid, quercetin, flavone) and insecticides (diazinon, cypermethrin, and aldrin) 2324.

Here we cloned the 5'-flanking sequences of the four xenobiotic-metabolizing H. zea CYP6B genes by genome walking approach [see Additional file 1] and then searched for the presence of transposon insertions by sequence analysis. We successfully recovered the genomic sequences and the 5'-flanking sequences of CYP6B8 and CYP6B27 but failed to recover the 5'-flanking sequences of CYP6B9 and CYP6B28 from our current laboratory population and from the H. zea midgut cell line RP-HzGUT-AW1 42. It is possible that CYP6B28 is just an allele of CYP6B8 since it is essentially identical to CYP6B8 except for a 227-bp insertion in its intron (Figure 3a) 43. Re-examination of this 227-bp insertion sequence reveals that it has a 60-bp perfect terminal inverted repeats (TIRs) flanked by 2-bp target site duplications (TSDs), lacks coding potential and is AT rich (59%). These are the common structural characteristics of MITEs 44, thus we designate it as HzMITE1 (Figure 1a; Figure 2c; Table 1).

In the laboratory population and the cell line, CYP6B8 and CYP6B27 are adjacent to each other in a head-to-tail arrangement, with CYP6B8 located 2030-bp upstream of CYP6B27 (Figure 1a). By comparing the sequence of the CYP6B8-CYP6B27 cluster in the laboratory colony and that in the cell line, three transposon insertions are characterized. One transposon insertion is found in the 5'-flanking promoter region of CYP6B8 in the midgut cell line CYP6B8-CYP6B27 cluster (Figure 1a; Table 1). This 3518-bp DNA transposon insertion has long TIRs and is flanked by 5-bp (GCTCG) TSDs (Figure 2a). The deduced 489 amino acids sequence has 34% identity with transib5 from Drosophila melanogaster 45. Based on these features, it is designated as Hztransib. Another transposon is found in the exon1 of CYP6B8 in the midgut cell line (Figure 1a; Table 1). This 1558-bp transposon has typical features of Tc1 family transposon, such as 27-bp perfect TIRs flanked by 2-bp (TA) TSDs (Figure 2b). Its putative open reading frame (ORF) encodes 375 amino acids and share 23% identity with Tc1 transposon from Caenorhabditis elegans 46. Thus it is designated as HzTc1.

The third one is inserted in the intergenic region of the CYP6B8-CYP6B27 cluster, i.e. the 5' flanking region of CYP6B27 or the 3'-flanking sequence of CYP6B8 (Figure 1a). This insertion is present in the CYP6B8-CYP6B27 cluster of the laboratory population but not in the midgut cell line (Table 1). This transposon is flanked by 6-bp TSDs (Figure 2d) and has an ORF sharing 29.1% amino acids identity with Maui transposon, a member of CR-1 (chicken repeat 1) family of non-LTR retrotransposons, from the pufferfish Fugu rubripes 47. This retrotransposon insertion also contains microsatellite (TA dinucleotides) repeats at the 3' end but lacks a poly-A tail, which usually happened in the CR1 family of non-LTR retrotransposons 47. Thus, we designate it as HzCR-1. Amino acids sequence alignment with other CR1 clade retrotransposons demonstrates that HzCR1 is truncated at the 5' end, losing apurinic-apyrimidic endonuclease (AP ENDO) domains and most of reverse transcriptase (RT) subdomains. Among the seven RT subdomains defined by Xiong and Eickbush 48, only the seventh subdomain and part of the sixth subdomain are retained in the HzCR1 (see Additional file 2).

Two cDNA sequences representing CYP9A12 and CYP9A14 were isolated from a pyrethroid-resistant strain of Helicoverpa armigera 25, the old world sibling species of H. zea. The fact that the CYP9A12 and CYP9A14 transcripts in the pyrethroid-selected resistant strain are overexpressed 433- and 59-fold in the fatbody and 19- and 4.3-fold in the midgut, respectively, in comparison with the unselected parental strain, indicates that these two P450s are associated with pyrethroid resistance 25. To examine if any transposon is inserted in the homologs of these two insecticide-metabolizing H. armigera CYP9A genes in H. zea, gene-specific primers (see Additional file 1) are designed to PCR-amply the genomic sequences of the H. zea homologs from the laboratory strain and the midgut cell line. The orthologous CYP9A genes obtained from the H. zea laboratory strain are designated as CYP9A12v3 (5.7 kb) and CYP9A14 (8.4 kb) (Nelson, personal communication). Sequence comparison with the cDNA sequences of the H. armigera CYP9A12 and CYP9A14 reveals that the two paralogous H. zea CYP9A genes share an identical genomic structure, each having 10 exons and 9 introns of variable sizes (Figure 1b, 1c). Their intron/exon locations and boundary sequences (all of the 9 introns follow the GT-AG rule in both CYP9A12v3 and CYP9A14) are identical.

One transposon is characterized from the H. zea CYP9A12v3 by using blastx against a nonredundant database. This transposon insertion, present in both the laboratory strain and the midgut cell line, is 1764-bp long, flanked by 10-bp target site duplications (TSDs), and inserted in the third intron of CYP9A12v3 (Figure 1b; Figure 2e; Table 1). It has one ORF that is orientated in an opposite direction with CYP9A12v3 (Figure 1b; Figure 2e) and encodes a reverse transcriptase with 40% amino acids sequence identity with the RTE-1 retrotransposon in Caenorhabditis elegans 49. Its 3' untranslated region is unusually short and is predominantly composed of TGA trinucleotide repeats, the typical feature found in the members of the RTE clade 50 (Figure 2e). The typical feature and high amino acids sequence identity it shares with the members of the RTE-1 clade indicates that it is a RTE-1 like non-LTR retrotransposon element and thus is designated as HzRTE-1. Amino acids sequence alignment with several RTE-1 retrotransposons demonstrates that HzRTE-1 is a 5'-truncated non-LTR retrotransposon that retains most of RT domain but loses AP ENDO domains. Among the seven RT subdomains, only subdomains 3rd-7th are retained in the HzRTE-1 (see Additional file 3).

Surprisingly, four transposons are characterized from the CYP9A14 locus (Figure 1c; Table 1). They are inserted into intron1, 7, and 8 of CYP9A14, respectively. The two elements inserted into intron 1 and 7 are 99% identical to each other, suggesting they are two copies of the same transposon. These two copies, however, have different TSDs and are oriented opposite to each other (Figure 1c; Figure 4a). Interestingly, this element is also inserted into the CYP321A2-CYP321A1 gene cluster (see below), namely, the 3'-flanking sequence of CYP321A2 and the 5'-flanking sequence of CYP321A1 (Figure 1d; Table 1). This copy also has different TSDs and is shorter in the 3' end than the two copies in the intron 1 and 7 of CYP9A14A1 (Figure 2f, 2g and 2l; Figure 4a). Because this element has no sequence similarity to any previously published insertion sequence and has no features that belong to Class II or Class I transposon, it is thus referred to as a TE-like element. We designate this element as HzIS1 (H. zea Insertion Sequence 1), with HzIS1-1, HzIS1-2 and HzIS1-3 representing its insertions in the first and seventh introns of CYP9A14 and the CYP321A2-CYP321A1 cluster, respectively. The laboratory strain has all three copies, whereas the midgut cell line has HzIS1-1 and HzIS1-3 only (Table 1).

In the eighth intron of CYP9A14 from the laboratory strain, two transposons are characterized, with one nested into another (Figure 1c; Table 1). BLAST search shows that the flanking transposon shares 96% sequence identity with the middle sequence (in an opposite orientation) of the 1^st intron of the H. zea acyl-CoA desaturase gene (HzPGDs3) (Figure 3b) 51. In addition, these two sequences also share identical 12-bp TIRs and identical TSDs, indicating they are two copies of the same transposon (Figure 2h; Figure 3b). Because they have the common structural characteristics of MITEs 44, including small size, AT rich (67.5–67.7%), TIRs and lack of coding potential, we designate them as HzMITE2-1 (the flanking transposon in the 8^th intron of CYP9A14) and HzMITE2-2 (the copy in the intron of the HzPGDs3), respectively. The nested transposon within HzMITE2-1 (Figure 1c; Figure 3b) is 946-bp long, AT rich (67.8%), lacks coding potential, and is flanked by 9-bp TSDs (Figure 2i). These features suggest that it is also a MITE-like element although without flanking TIRs 44. This nested transposon is designated as HzMITE3. The CYP9A14 allele from the midgut cell line has the flanking HzMITE2-1 only (Table 1).

To examine if there are transposons around the CYP321A1 locus, another well-defined xenobiotic-metabolizing P450 in H. zea 23, we cloned the 5'-flanking sequence of CYP321A1 from both the laboratory strain and the midgut cell line. A 5.4-kb fragment upstream of CYP321A1 was obtained from the laboratory population by genome walking. From this fragment, we identified another P450 gene, a paralog of CYP321A1, is designated as CYP321A2 (Nelson, personal communication). Like the CYP6B8-CYP6B27 cluster, the CYP321A2-CYP321A1 paralogs are oriented in a head-to-tail arrangement, with CYP321A2 located 1467-bp upstream of CYP321A1 (Figure 1d). Compared with CYP321A1 and CYP321A2 alleles isolated from the midgut cells, the CYP321A2 allele from the laboratory population contains a 556-bp TE insertion located 126-bp upstream of its heme-binding signature motif in its ORF (Figure 1d; Table 1), resulting in a 3'-truncated transcript (Chen & Li, unpublished data). The insertion sequence is flanked by 5-bp (AAAAA) TSDs and ends with tetramer repeats (GACR) at its 3'end (Figure 2k), which is a feature found in some SINEs such as Artw and Pst elements in cows 52 and CR1 element in chickens 53. The fact that it has no sequence similarity to any previously published transposable element suggests that it is a novel SINE element. We therefore named it as HzSINE1.

In addition, there are two other transposable elements in the CYP321A2-CYP321A1 cluster, one located in the intergenic region, another inserted in the 5'-flanking region of CYP321A2 (Figure 1d; Table 1). The transposon in the intergenic region is 99% identical to HzIS1-1 and HzIS1-2 found in the intron 1 and 7 of CYP9A14 (Figure 4a) and thus is designated as HzIS1-3 (see above). The transposon inserted in the 5'-flanking region of CYP321A2 is an extremely 5'-truncated copy of the HzRTE-1 element found in the 3^rd intron of CYP9A12v3 (see above). This HzRTE-1 copy completely loses its ORF and retains only the 62-bp 3'UTR and the flanking 7-bp TSD (Figure 2j; Figure 4b). To differentiate the two HzRTE-1 copies, we name them as HzRTE-1-1 (the one found in the 3^rd intron of CYP9A12v3) and HzRTE-1-2 (the extremely-truncated copy in the 5'-flanked promoter of CYP321A2).

An extensive literature search shows that thirteen out of the 90 P450 genes in the D. melanogaster genome have been functionally defined (Table 2) 54. Eight of them, including CYP4E2 27, CYP6A2 282930, CYP6A8 28, CYP6A9 29, CYP6G1 313233, CYP6W1 30, CYP12A4 34, CYP12D1 3536, are implicated in insecticide resistance and xenobiotic metabolism. The other five P450 genes (CYP302A1, CYP306A1, CYP307A1, CYP314A1 and CYP315A1) are involved in ecdysone biosynthesis and developmental regulation 1837383940. To test if transposons are more enriched in the xenobiotic-metabolizing Drosophila P450 genes as opposed to the ecdysone-synthesizing Drosophila P450s, we surveyed the insertions of TE within or in close proximity to the thirteen P450 loci and GSTD1, a DDT-metabolizing gene, by in silico mapping at the flybase 26. Seven out of the eight insecticide resistance responsible P450 genes and the DDT-metabolizing GSTD1 are associated with a variety of TEs (Table 2). In contrast, TEs are excluded from all of the five ecdysone-synthesizing P450 genes (Table 2).

A laboratory strain of H. zea, generously provided by Dr. May R. Berenbaum (Department of Entomology, University of Illinois at Urbana-Champaigen), was maintained in an insectary kept at 28°C with a photoperiod of 16 h light:8 h dark on a semi-synthetic control diet containing wheat germ 68. Several 5^th instar larvae randomly picked from the colony were individually flash-frozen in liquid nitrogen and stored in -80°C for subsequent genomic DNA isolation.

H. zea midgut cell line RP-HzGUT-AW142, generously provided by Dr. Cynthia L. Goodman (BCIRL, USDA, ARS), was cultured in ExCell 401 serum-free medium (JR BioSciences, Lenexa, KS) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 50 μg/ml streptomycin, and 50 units/ml penicillin (Sigma). The cells were harvested after 72 hour culture at 27°C by a brief centrifugation at 1,000 g for 2 min. The cell pellets obtained were flash-frozen in liquid nitrogen and stored in -80°C until use for DNA isolation.

Genomic DNA was isolated from the 5^th instar larva and cell pellets, using the procedure described in Li et al 43. The 5'-flanking sequences of the four xenobiotic-metabolizing P450 genes, CYP6B8, CYP6B27, CYP321A1, and CYP321A2, were obtained by genome walking. In brief, genomic DNA was digested by several restriction enzymes supplied in the Universal GenomeWalker kit (Clontech, CA, USA) and then ligated to the genome walking adapters according to the manufacturer's manual. The resulting DNA fragments were used as templates to PCR-amply the 5'-flanking sequences of the four P450 genes using the two general forward primers complementary to the adapter sequences and the two corresponding gene-specific reverse primers (see Additional file 1) for each P450 gene. The nested PCR reactions began with the primary PCR consisting of 25 cycles of 94°C denaturation for 2 min, 68°C annealing/extension for 4 min, followed by the secondary PCR consisting of 35 cycles of 94°C denaturation for 2 min and 68°C annealing/extension for 4 min. PCR products were run on a 1% agarose gel in 1×TAE buffer. The longest band for each gene was eluted from the gel using the QIAquick Gel Extraction Kit (Qiagen), and then directly cloned into the PGEM-T easy vector (Promega, MI). One white clone for each band was sequenced on Applied Biosystems 3730 DNA Analyzer twice in both directions using M13 forward and M13 reverse primers as well as internal primers designed on the basis of the determined sequences at the Genomic Analysis & Technology Core Facility of the University of Arizona.

For cloning the CYP9A12v3 and CYP9A14 genomic sequences, two pairs of primers, CYP9A12F/CYP9A12R and CYP9A14F/CYP9A14R (see Additional file 1), were designed based on the cDNA sequences of their orthologs CYP9A12 [GenBank: AY371318] and CYP9A14 [GenBank: AY487948] in H. armigera, the Old World sibling species of H. zea. Their full length DNA sequences were then PCR-amplified with 35 cycles of 94°C denaturation for 2 min and 68°C annealing/extension for 4 min (CYP9A12v3) or 6min (CYP9A14). As described above, the PCR products obtained were cloned into the PGEM-T easy vector and sequenced.

Sequences analyses were performed first by using blastn and blastx against a nonredundant database 69 and by using CENSOR against the Repbase 70 to scan the obtained H. zea P450 genomic sequences for TE insertions. The optimal global sequence alignment program was then used to further identify the sites of TE insertions, their TSDs, TIRs and other features by comparing a TE-free sequence and a TE-containing sequence 71. DNA and protein sequence alignments were conducted using Genedoc software 72. Sequences reported in this paper were deposited in the GenBank under the following accession numbers: Hztransib [GenBank: DQ788836], HzTc1 [GenBank: DQ788837], HzCR1 [GenBank: DQ788838], genomic DNA of CYP9A12v3 containing HzRTE-1-1 [GenBank: DQ788839], genomic DNA of CYP9A14 containing HzHzIS1-1, HzIS1-2, HzMITE2-1 and HzMITE3 [GenBank: DQ788840] and 5'-flanking sequence of CYP321A1 containing HzRTE-1-2, HzSINE1 and HzIS1-3 [GenBank: DQ788841].

An extensive literature search was conducted to identify D. melanogaster P450s whose functions have been relatively defined. This led to characterization of eight insecticide resistance associated P450 genes (a DDT-resistant GST gene also included) and five ecdysone-synthesizing P450 genes (Table 2). A survey of TE insertions within or in close proximity (until encounter other genes or beyond 10 kb in both the 5' and 3' directions) to the thirteen P450 genes and the one GST gene was then conducted by in silico mapping at the flybase 26.