Ostrinia nubilalis samples and extraction protocols

Fat body, midgut, head, and whole body samples were collected from 5^th instar O. nubilalis larvae. Additionally, fat body tissue was removed from 4^th and 5^th O. nubilalis larvae, and from larvae exposed to diapause conditions for one week (Beck 1982). All O. nubilalis were bivoltine Z-pheromone strain individuals from the USDA-ARS, Corn Insects and Crop Genetics Research Unit lab colony in Ames, Iowa. Dissected tissue samples were flash frozen in liquid nitrogen, ground to powder, and RNA extracted with RNeasy extraction kits (Qiagen,  www.qiagen.com) according to manufacturer instructions. All DNA extractions used adult thoracic tissue and were performed as described by Coates and Hellmich (2003). Extracts were quantified by absorption at 260 nm (A²⁶⁰), diluted to 50 ng/μl, and stored at −20°C.

PCR primer design, DNA amplification, and sequencing

Forward (DAP1-F: 5′-TTA GCG AGT GCT GCC TTG GT-3′) and reverse PCR primers (DAP1-R: 5′-GTC AGG CGC ATC ACA CTG TT-3′) were designed from an O. furnacalis diapause-associated protein gene sequence (GenBank accession no. AF169311) using Primer3 (Rozen and Skaletsky 1998). Primers spanned the coding sequence from amino acid positions 10 through 269 of the 291 amino acid long O. furnacalis gene. The homologous O. nubilalis gene region was PCR amplified using 5 pmol of each primer in 2.5 mM MgCl₂, 50 μM dNTPs, 100 ng of DNA template, and 0.9 U Taq DNA polymerase (Promega,  www.promega.com) in a 12.5 μl reaction. PTC-100 thermocycler conditions used 95° C for 2.5 m, followed by 40 cycles of 95° C for 30 s, 58° C for 30 s, and 72° C for 1 m. Amplification was confirmed and fragment size estimated by electrophoresis of 6 μl PCR product on a 10 cm 1.5 % agarose 1X tris-borate EDTA (1X TBE) gel containing 0.5 μg/ml ethidium bromide. PCR product was ligated using the pGEM-T Easy Cloning system (Promega) by overnight incubation at 4° C, and was used to transform 80 μl E. coli SURE (Stratagene,  www.stratagene.com). Cloned inserts were DNA sequenced in 10 μl DTCS Quickstart reactions (Beckman-Coulter,  www.beckman.com) using 1.6 pmol T7 primer according to manufacturer instructions. Sequence reaction products were purified by ethanol precipitation, suspended in 40 μl deionized formamide, and separated on a CEQ 8000 Genetic Analysis System (Beckman-Coulter) with method LFR-1 (denature: 90° C for 120 sec; inject: 2.0 kV for 15 sec; and separated: 4.2 kV for 85 min in a 50° C capillary). Sequence data was used to design transcript-specific reverse primers. Electropherogram output from duplicate runs was inspected visually for sequencing errors on the CEQ8000 Genetic Analysis Software.

Gene homology and orthology

The Ostrinia furnacalis diapause-associated protein sequence (DAP; GenBank accession no. AF169311; positions 246 to 762) was aligned with homologous regions from three O. nubilalis obp1 and 2 obp2 alleles, and M. sexta OBP gene sequence (GenBank accession no. L00975; positions 59 to 575). A 350 nt Onobp2 allele fragment (Onobp2-B4) that spanned O. furnacalis DAP positions 487 to 836 was omitted from the alignment. Additionally, expressed sequence tag sequences from Bombyx mori 5^th instar fat body (fbVm022; male; strain p50) and ovary (ovS318H0, and ovS10G0) were include in the alignment. The consensus alignment was generated using AlignX software (Informax; gap penalty = 10), and a shared 516-nucleotide region of the alignment was used to construct a parsimony-based phylogeny using programs from the PHYLIP package (Felsenstein 1989). Bombyx mori ovarian expressed sequence tag ovS10G0 was used as the outgroup. One thousand bootstrap resampling steps were produced by the SeqBoot program, parsimony trees were generated using DNAPars, a strict consensus tree was estimated from all possible phylogenies with CONSENSE, and was viewed using TreeView (Page 1996).

Expression analysis

Total RNA extracted from fat body, midgut, head, and whole larvae from 5^th instar larvae, and fat body from 4^th and 5^th, 5^th instar wandering, and 1 week old diapausing O. nubilalis larvae was subjected to reverse transcriptase (RT)-PCR analysis. Individual RT-PCR first strand cDNA synthesis reactions used 250 ng total RNA template, 10 pmol of primer DAP1-R or β-actin-R (5′-GAC AAC GGC TCC GGT ATG T-3′; controls only), 2.5 U Tth polymerase (Promega), 100 μ M dNTPs, 2.5 mM MnCl₂, and 2.0 μl of 10X Tth reverse transcriptase buffer (Promega) in a 10 μl reaction volume. A PTC-100 thermocycler (MJ Research,  www.mjr.com) performed a primer extension cycle of 85° C for 1 m, 56° C for 1 m, and 72° C for 20 m. A 4.0 μl aliquot of 1^st strand cDNA synthesis product was mixed with 1.6 μl chelate buffer (Promega), 10 pmol of primer OnOBP1-F, OnOBP2-F, or β-actin-F (5′-CCT TCG TAG ATA GGG ACG GT-3′; controls only), MgCl₂ concentration adjusted 2.0 mM in a 20 μl final volume. RT-PCR reactions were carried out on a PTC-100 thermocycler (MJ Research) using 40 cycles of 95° C for 30 s, 56° C for 30 s, and 72° C for 1 m. RT-PCR products (20 μl) were separated on a 10 cm 1.0 % agarose 1X tris-borate EDTA gel containing 0.5 μg/ml ethidium bromide, and digital images taken under UV illumination on a BioRad ChemiDoc System (BioRad,  www.bio-rad.com).

Ostrinia nubilalis ommochrome binding protein (OBP) polymorphism

Two putative O. nubilalis ommochrome binding protein genes, called Onobp1 and Onobp2, were co-amplified using primers DAP1-F and DAP1-R. Gene specific forward PCR primers OnOBP1-F (5′-GGG AGC GTG CTR AAG ACC AT-3′) and OnOBP2-F (5′-GGG ACT GTG CTG ATG AAG A-3′) were designed using Primer3 (Rozen and Skaletsky 1998) that could be used in combination with DAP1-R. The two PCR products were amplified separately, but with the same conditions (4 pmol of each primer, 2.5 mM MgCl₂, 50 μM dNTPs, 100 ng of DNA template, and 0.9 U Taq DNA polymerase (Promega) in a 10.0 μl reaction). PTC-100 thermocycler conditions used 95° C for 2.5 m, followed by 40 cycles of 95° C for 30 s, 58° C for 30 s, and 72° C for 1 m.

Polymorphic restriction endonuclease cleavage sites were identified in Onobp1 and Onobp2 gene fragment alignments, and used to screen pedigrees and genotype light trap samples. The Onobp1 fragment was digested with HinfI or Tsp509I, and the Onobp2 fragment was digested with HaeIII, MseI or TaqI. All endonucleases were purchased from New England BioLabs except TaqI (Promega). PCR-RFLP reactions included 6.0 μl of appropriate PCR product, 2.5 μl 10x Buffer, 0.1 mg/μl BSA, and 0.25 U of enzyme in 25 μl. Reactions were incubated at 37° C or 60° C (Tsp509I and TaqI) for 14 h. Entire PCR-RFLP reaction volumes were loaded onto 10 cm 2.0 % 1X TBE agarose gels that contained 0.5 μg/ml ethidium bromide. Samples were separated at 100 V for 1.5 h, and images captured under UV illumination on a ChemiDoc System (BioRad).

Mendelian inheritance of Onobp1 and Onobp2

Mendelian inheritance of Onobp1 and 2 was evaluated by allele segregation in pedigrees. Two paired matings of O. nubilalis adults from the USDA-ARS, CICGRU laboratory colony were used to establish F₁ families. The F₁ progeny were sib-mated (inbred) to produce F₂ larvae. F₂ larvae were reared on a semi-meridic diet (Guthrie 1987), DNA extracted from adults according to Coates and Hellmich (2003), and samples stored at – 20° C prior to use. Onobp1 and Onobp2 loci were screened by PCR-RFLP as described previously, and chi-square tests used to measure goodness-of-fit between observed allelic distributions in the F₂ generation and expectation Mendelian ratios.

Gene characterization and amplification

Two highly similar ommochrome-binding protein-like (obp) genes were characterized from the O. nubilalis genome. Both sequences were isolated from O. nubilalis by PCR, named Onobp1 and Onobp2, and the data were submitted under GenBank accession nos. AY819651 to AY819655. GenBank accession nos. AY819651 and AY819652 are partial Onobp1, and AY819653 to AY819655 and AY862870 are partial Onobp2 genes that lack N- and C-terminal peptide coding regions. Both O. nubilalis genes are transcribed in fat body tissue only, show derived structural characteristics similar to the M. sexta OBP, and likely encode glycoprotein products excreted into hemolymph. Structural identification of Ostrinia sequences as OBP encoding genes was based on peptide identity to the M. sexta OBP, similarity of the signal peptide, N-glycosylation, a conserved glycosyl hydrolase family 10 active site, and SwissProt classification of an O. furnicalis homolog within peptide Family PD125063 with the M. sexta OBP.

A PCR product amplified from O. nubilalis genomic DNA using primers DAP1-F and DAP1-R contained a heterogeneous mix of two approximately 775 bp products. The O. nubilalis sequences were gene fragments that omitted 9 and 22 codons from N- and C-terminal peptide regions, respectively. The O. nubilalis nucleotide sequences were 90–98% similar to an O. furnacalis diapause-associated peptide coding sequence (DAP; GenBank accession no. AF169311; Fig. 1; BLAST score ≥ 233 and E-score ≥ 6 × 10^-30). All O. nubilalis obp1 and 2 genes contained > 90% of the CDS length and shared ≥ 80% identity at the peptide-level compared to O. furnacalis. The unpublished O. furnacalis GenBank accession AF169311 was therefore used to establish similarity to known genes. The 291 amino acid O. furnacalis DAP1 peptide showed similarity (35 %) and identity (54 %) to a 274 residue (31 kDa) M. sexta ommochrome binding protein encoding gene (obp; Martel and Law 1991; Yepiz-Plascencia et al. 1993; GenBank accession no. L00975; BLASTp score = 134, E-value 10^-30). A BLASTp search of using the M. sexta peptide P31420 identified our released O. nubilalis obp1 and 2 sequences (E-values 4 × 10^-14 to E-value 3 × 10^-36; identities 34 to 38%; similarities 53 to 58%). Identities as low as 25% were present between putative chymotrypsin-like proteases from the Hessian fly (Mayetiola destructor; Zhu 2005) even though genes were derived from a single ancestor via gene duplication (Neurath et al. 1967). The peptide similarity of 46% was used to identify a 269-residue B. mori peptide fragment derived from fat body expressed sequence tag fbVm0227 as an ommochrome binding protein (SilkBase, Bombyx Genome Database Working Group; Toru Shimada, Kazuei Mita, and co-workers unpublished). This suggests relative homology between Ostrinia and M. sexta genes is sufficient to classify the former as ommochrome-binding protein-related genes.

The derived amino acid sequence from O. furnacalis DAP GenBank accession no. AF169311 had a predicted molecular weight ≅ 32.6 kDa and isoelectric point (pI) of 4.89 (compute pI/MW program; Bjellqvist et al. 1993) that was near 31 kDa and pI of 5.89 for M. sexta OBP (Martel and Law 1991). The O. furnacalis peptide was predicted to have extracellular localization with 66.7% assurance (PSORTII program; Paul Horton unpublished;  www.psort.org). Additionally, the O. furnacalis peptide may have an 18 amino acid signal peptide (residues 1 to 18) predicted by PSORTII program using methods described by von Heijne (1986), which is the exact length of the signal peptide from M. sexta. Although the two signal sequences appear to share little sequence identity, O. furnacalis and M. sexta OBP signal peptides were respectively composed of 12/18 and 11/18 residues with aliphatic side chains. Furthermore, the program SignalP 3.0 ( www.cbs.dtu.dk/services/SignalP/) used a hidden Markov models to predict presence of a peptide cleavage site following a VSS residue sequence of the O. furnicalis OBP peptide (probability 0.998). The partial O. nubilalis obp1 and obp2 sequences contained 50% of the entire signal sequence length and is the location of forward PCR primer (DAP1-F), showing the O. nubilalis obp1 and obp2 genes likely also encode a similar signal sequence. These properties suggest cell membrane transport out of the fat body, and proteins that are soluble in lepidopteran hemolymph (pH ≅ 7.0; see expression data below).

The derived peptide sequences from O. furnacalis or O. nubilalis gene fragments did not show similarity to Choristoneura fumiferana (Lepidoptera: Tortricidae) DAP1 (753 amino acids; 74 kDa) or DAP2 (749 amino acids; 72 kDa; Palli et al. 1998), indicating they constitute a different class of hemolymph proteins. Therefore, we propose that the O. furnacalis DAP gene and homologous O. nubilalis genomic regions may encode ommochrome-binding proteins (OBPs).

A PredictProtein search of the ProSite database indicated the presence of two N-glycosylation sites, NITE and NKTK, from both O. furnacalis and O. nubilalis obp genes (Rost 1996; Fig. 1). A mannose attachment was found on the M. sexta OBP (Martel and Law 1991), and a single corresponding N-glycosylation site found by Yepiz-Plascencia et al. (1993). These predictions indicate the putative Ostrinia obp1 and obp2 gene products are also glycosylated. Additionally, a ProDom search ( http://protein.toulouse.inra.fr/prodom/current/html/home.php) co-classified the Ostrinia and M. sexta OBPs within peptide Family PD125063 (Zdobnov and Apweiler 2001). All members of this protein family share a conserved glycosyl hydrolase family 10 active site that is present in O. furnacalis OBP residues 96 to 106 and O. nubilalis homologs (Fig. 1). Glycosyl hydrolases degrade cellulose and xylans, but the presence of this activity has not been previously described for OBPs. The function of this active site is unknown and it may be falsely identified by the Protein database. Nevertheless, the conserved domain provides evidence of co-ancestry between OBP encoding genes. Phylogenetic and expression level evidence for O. nubilalis obp1 and obp2 ommochrome binding protein classification also is shown (see below).

Gene homology and orthology

Phylogenetic methods can predict relationships and genealogical history among genes or alleles, and was used to explore common ancestry between lepidopteran obp genes. The parsimony-based phylogeny incorporated an alignment of 516 nucleotides shared by O. furnacalis DAP (OBP1) coding sequence (GenBank accession no. AF169311), M. sexta (Yepiz-Plascencia et al. 1993; GenBank accession no. L00975), O. nubilalis obp1 and obp2 allele fragments (Fig. 1), and B. mori expressed sequence tags fbVm022, ovS318H0, and ovS10G0. The resultant phylogeny showed that Ostrinia obp genes are nested within a clade with the M. sexta gene, indicating derivation from a single ancestral gene and common ancestry of contemporary genes. The tree also indicted ≥ 950 of 1000 bootstrap pseudoreplicates supported presence of a node between O. nubilalis obp1 and 2 alleles (Fig. 2). Lower interspecific compared to intraspecific similarity between Onobp1 and Onobp2 sequences and the phylogeny (Fig. 2) suggested duplication of Onobp1 and Onobp2 from a common ancestral gene. Gene duplications are maintained by gain of function or division of function (Force et al. 1999; Lynch and Force 2000), otherwise gene duplicates may accumulate mutations (“decay”) resulting in reduced or lost function and pseudogene formation.

Expression analysis

Semi-quantitative RT-PCR showed that transcription of Onobp1 and Onobp2 is differentially regulated in O. nubilalis fat body, and duplicate genes may be maintained in the genome due to their having different functions. The ommochrome-binding protein gene expressed in fat body of 5^th instar M. sexta larvae (Yepiz-Plascencia et al. 1993) corresponded to a peptide that increased in concentration in hemolymph during late instars (Martel and Law 1991; 1992). An OBP protein was also purified from 5^th instar A. yamamai hemolymph (Saito et al. 1998). Two co-expressed obp-like genes have not previously been described in Lepidoptera, except for the evidence presented here for larval O. nubilalis fat body. The tissue specificity of O. nubilalis obp1 and obp2 provides further evidence for correct classification of these genes as coding for ommochrome-binding proteins.

Expression patterns of two highly similar O. nubilalis genes (Onobp1 and Onobp2) were characterized by locus-specific RT-PCR. Results indicated Onobp1 and 2 transcripts are present in total RNA preparations from whole larvae and dissected fat body tissue, but not in midgut or head tissues of 5^th instar larvae (Fig. 3 A). Lack of obp1 and obp2 RT-PCR products suggests transcripts are not present or below the level of sensitivity in head and midgut tissues. Amplification of β-actin (b-actin) from total RNA obtained from each tissue indicated template integrity differences were not present between samples. Fat body expression of Onobp1 and Onobp2 transcripts was expected due to data from M. sexta (Yepiz-Plascencia et al. 1993) and further suggests correct classification as OBP-encoding genes.

The O. nubilalis obp1 gene has larval growth stage-specific expression. Onobp1 transcript was detected by RT-RCR in total RNA from fat body of 5^th instar larvae, whereas the transcript was not detected from 4^th instar, 5^th instar wandering, or 1 week old diapausing larval fat body using identical methods (Fig. 3 B). Similar larval instar-specific transcription was observed for a M. sexta obp but in 5^th instar wandering larvae (Yepiz-Plascencia et al. 1993). Also, an A. yamamai obp peptide product was identified in 5^th instar hemolymph (Saito et al. 1998). Data collected from O. nubilalis, M. sexta, and A. yamamai suggest an orthologous obp gene may be expressed in 5^th and late 5^th instar Lepidoptera.

In contrast to obp1, transcription of the O. nubilalis obp2 gene is not restricted to a particular developmental stage. The Onobp2 transcript was amplified in fat body total RNA samples by RT-PCR from all larval growth stages tested, suggesting constitutive fat body expression (Fig. 3 B). The obp2 transcript level in the fat body may vary between larval stages. It increased as larvae moved from 4^th instar to 5^th instar wandering larvae, and remained high among 1 week old diapausing larvae. Assuming primers (DAP1-R, OBP1-F and OBP2-F) anneal with equal efficiency, the level of Onobp2 transcription appears greater than Onobp1. Higher overall OBP2 expression may result, although eukaryotic transcriptional and translational efficiency are not always correlated. Transcription of Onobp2 across larval stages differed from the stage-specific expression of Onobp1 (Fig. 3 B) and M. sexta obp transcripts (Yepiz-Plascencia et al. 1993). Differential expression of Onobp1 and 2 among larval stages might suggest O. nubilalis paralogs might differ in developmental stage-specific enhancer elements (Ayer and Benyajati 1992).

Ostrinia nubilalis ommochrome binding protein polymorphism

Four peptide sequence conflicts were identified within the first 31 residues of M. sexta OBP N-terminal sequence by comparing evidence from Martel and Law (1991) and Yepez-Plascencia et al. (1993). These data suggest variability exists even within the same species, and may contribute to low peptide-level similarity and identity between species. Two unique sequence types with similarity to the O. furnacalis DAP gene (reclassified as OBP1) were isolated from O. nubilalis by PCR, (GenBank accession nos. AY819651 to AY819655 and AY862870). No introns were predicted within either gene fragment. Intraspecific alignment of 775 to 778 bp Onobp1 (alleles Onobp1-A1, and -A2) and 782 to 785 bp Onobp2 gene fragments (alleles Onobp2-B1, -B2, and -B3) identified 21 and 23 single nucleotide polymorphisms, respectively (Fig. 1). Onobp1 single nucleotide polymorphisms within HinfI or Tsp509I, and Onobp2 mutations within HaeIII, MseI or TaqI restriction endonuclease cleavage sites were observed (Fig. 1; Table 1).

The difference of three nucleotides was due to a Thr deletion near the N-terminus of Onobp1, and an Asp residue deletion among Onobp2 sequences. DNA sequence data indicated 97.0% and 97.3 to 98.3% similarity between Onobp1 and 2 alleles, respectively. Lower sequence similarity was observed between Onobp1 and 2 sequences (88.3 to 90.1%). The Onobp2-B4 sequencing fragment (GenBank accession no. AY862870) was not included in this analysis, but used to identify polymorphic nucleotides within pedigrees. Lower similarity between Onobp1 and Onobp2 might suggest sequences are from separate loci. Sequence differentiation can occur between alleles at the same locus, thus similarity comparisons between Onobp1 and Onobp2 do not prove they are unique genes with different genome positions. Fidelity of allelic inheritance was confirmed by pedigree analysis.

Mendelian inheritance

Two sets of divergent yet similar genome sequences (Onobp1 and Onobp2) isolated from O. nubilalis by PCR constitute separate loci. Gene specific primer fidelity was inferred by calculating degree of mismatch and estimated melting temperature. The Onobp1 specific OnOBP1-F primer showed 25% mismatch to the homologous region of Onobp2 template sequence with an estimated annealing temperature ≅ 48 °C. The Onobp2 locus specific OnOBP2-F primer had 30% mismatch with homologous Onobp1 template with an estimated annealing temperature of ≅ 46 °C. Template (gene or locus) specificity of primers was shown empirically by failure of OnOBP1-F to prime PCR that used a cloned Onobp2 gene fragment as template, and OnOBP2-F primer to prime PCR that used a cloned Onobp1 gene fragment as template (58° C annealing; data not shown).

Observed F₂ ratios indicated Onobp1 HinfI PCR-RFLP or Onobp2 HaeIII PCR-RFLP defined alleles did not deviate significantly from Mendelian expectation (Table 1). In combination with sequence specificity of forward primers, phylogeny, differential expression, Mendelian segregation of heterozygous allele states in Onobp1 and Onobp2 assays indicate PCR products are derived from unique loci, and show that null alleles were not present (Pemberton et al. 1995).