Genomic innovations, transcriptional plasticity and gene loss underlying the evolution and divergence of two highly polyphagous and invasive Helicoverpa pest species.

BACKGROUND: Helicoverpa armigera and Helicoverpa zea are major caterpillar pests of Old and New World agriculture, respectively. Both, particularly H. armigera, are extremely polyphagous, and H. armigera has developed resistance to many insecticides. Here we use comparative genomics, transcriptomics and resequencing to elucidate the genetic basis for their properties as pests. RESULTS: We find that, prior to their divergence about 1.5 Mya, the H. armigera/H. zea lineage had accumulated up to more than 100 more members of specific detoxification and digestion gene families and more than 100 extra gustatory receptor genes, compared to other lepidopterans with narrower host ranges. The two genomes remain very similar in gene content and order, but H. armigera is more polymorphic overall, and H. zea has lost several detoxification genes, as well as about 50 gustatory receptor genes. It also lacks certain genes and alleles conferring insecticide resistance found in H. armigera. Non-synonymous sites in the expanded gene families above are rapidly diverging, both between paralogues and between orthologues in the two species. Whole genome transcriptomic analyses of H. armigera larvae show widely divergent responses to different host plants, including responses among many of the duplicated detoxification and digestion genes. CONCLUSIONS: The extreme polyphagy of the two heliothines is associated with extensive amplification and neofunctionalisation of genes involved in host finding and use, coupled with versatile transcriptional responses on different hosts. H. armigera's invasion of the Americas in recent years means that hybridisation could generate populations that are both locally adapted and insecticide resistant.

The effect of diet on the expression of lipase genes in the midgut of the lightbrown apple moth (Epiphyas postvittana Walker; Tortricidae).

We have identified lipase-like genes from an Epiphyas postvittana larval midgut EST library. Of the 10 pancreatic lipase family genes, six appear to encode active lipases and four encode inactive lipases, based on the presence/absence of essential catalytic residues. The four gastric lipase family genes appear to encode active proteins. Phylogenetic analysis of 54 lepidopteran pancreatic lipase proteins resolved the clade into five groups of midgut origin and a sixth of non-midgut lipases. The inactive proteins formed two separate groups with highly conserved mutations. The lepidopteran midgut lipases formed a ninth subfamily of pancreatic lipases. Eighteen insect and human gastric lipases were analysed phylogenetically with only very weak support for any groupings. Gene expression was measured in the larval midgut following feeding on five artificial diets and on apple leaves. The artificial diets contained different levels of triacylglycerol, linoleic acid and cholesterol. Significant changes in gene expression (more than 100-fold for active pancreatic lipases) were observed. All the inactive lipases were also highly expressed. The gastric lipase genes were expressed at lower levels and suppressed in larvae feeding on leaves. Together, protein motif analysis and the gene expression data suggest that, in phytophagous lepidopteran larvae, the pancreatic lipases may function in vivo as galactolipases and phospholipases whereas the gastric lipases may function as triacylglycerol hydrolases.

Changes in gene expression in the permissive larval host lightbrown apple moth (Epiphyas postvittana, Tortricidae) in response to EppoNPV (Baculoviridae) infection.

Host cell and virus gene expression were measured five days after per os inoculation of 3rd instar lightbrown apple moth (LBAM) larvae with the Epiphyas postvittana nucleopolyhedrovirus (EppoNPV). Microarray analysis identified 84 insect genes that were up-regulated and 18 genes that were down-regulated in virus-infected larvae compared with uninfected larvae. From the 134 viral open reading frames represented on the microarray, 81 genes showed strong expression. Of the 38 functionally identifiable regulated insect genes, 23 coded for proteins that have roles in one of five processes; regulation of transcription and translation, induction of apoptosis, and maintenance of both juvenility and actin cytoskeletal integrity. Of the 34 functionally identifiable viral genes that were most strongly expressed, 12 had functions associated with these five processes, as did a further seven viral genes which were expressed at slightly lower levels. A survey of the LBAM-expressed sequence tag library identified further genes involved in these processes. In total, 135 insect genes and 38 viral genes were analysed by quantitative polymerase chain reaction. Twenty-one insect genes were strongly up-regulated and 31 genes strongly down-regulated. All 38 viral genes examined were highly expressed. These data suggest that induction of apoptosis and regulation of juvenility are the major 'battlegrounds' between virus and insect, with the majority of changes observed representing viral control of insect gene expression. Transcription and translational effects seem to be exerted largely through modulation of mRNA and protein degradation. Examples of attempts by the insect to repel the infection via changes in gene expression within these same processes were, however, also noted. The data also showed the extent to which viral transcription dominated in the infected insects at five days post inoculation.

Phenotypic changes and the fate of digestive enzymes during induction of amber disease in larvae of the New Zealand grass grub (Costelytra zealandica).

Amber disease of the New Zealand grass grub Costelytra zealandica (Coleoptera: Scarabaeidae) is caused by ingestion of pADAP plasmid carrying isolates of Serratia entomophila or Serratia proteamaculans (Enterobacteriaceae) and causes infected larvae to cease feeding and clear their midgut to a pale amber colour where midgut serine protease activities are virtually eliminated. Using bacterial strains and mutants expressing combinations of the anti-feeding (afp) and gut clearance (sep) gene clusters from pADAP, we manipulated the disease phenotype and demonstrated directly the relationship between gene clusters, phenotype and loss of enzyme activity. Treatment with afp-expressing strains caused cessation of feeding without gut clearance where midgut protease activity was maintained at levels similar to that of healthy larvae. Treatment with strains expressing sep-genes caused gut clearance followed by a virtual elimination of trypsin and chymotrypsin titre in the midgut indicating both the loss of pre-existing enzyme from the lumen and a failure to replenish enzyme levels in this region by secretion from the epithelium. Monitoring of enzymatic activity through the alimentary tract during expression of disease showed that loss of serine protease activity in the midgut was matched by a surge of protease activity in the hindgut and frass pellets, indicating a flushing and elimination of the midgut contents. The blocking of enzyme secretion through amber disease appears to be selective as leucine aminopeptidase and alpha-amylase were still detected in the midgut of diseased larvae.

Serratia entomophila inoculation causes a defect in exocytosis in Costelytra zealandica larvae.

Rapid elimination of midgut luminal proteinase activity and gut clearance are the two major symptoms of amber disease in Costelytra zealandica larvae because of the three-subunit protein toxin complex produced in Serratia entomophila and Serratia proteamaculans. Quantitative PCR analysis of mRNA from the major serine proteinase gene families showed that loss of proteinase activity did not result from transcriptional downregulation. Unexpectedly, protein levels and rates of protein synthesis increased, rather than decreased, in the midgut of diseased insects. Proteomic analysis of midgut tissues showed marked differences between healthy and diseased midguts. Large increases in soluble forms of both actin and tubulin were identified from 2D-gels, together with concurrent decreases in the levels of polymeric actin-associated proteins: actin depolymerizing factor and cyclophilin. These results suggest that the Serratia toxin acts to cause degradation of the cytoskeletal network and prevent secretion of midgut gut digestive proteinases as both the actin cytoskeleton and microtubules are involved in exocytosis. Proteinases synthesized in the diseased midgut must be rapidly degraded because they do not accumulate in an inactive form.

Serine proteases identified from a Costelytra zealandica (White) (Coleoptera: Scarabaeidae) midgut EST library and their expression through insect development.

Costelytra zealandica larvae are pests of New Zealand pastures causing damage by feeding on the roots of grasses and clovers. The major larval protein digestive enzymes are serine proteases (SPs), which are targets for disruption in pest control. An expressed sequence tag (EST) library from healthy, third instar larval midgut tissue was constructed and analysed to determine the composition and regulation of proteases in the C. zealandica larval midgut. Gene mining identified three trypsin-like and 11 chymotrypsin-like SPs spread among four major subgroups. Representative SPs were examined by quantitative PCR and enzyme activity assayed across developmental stages. The serine protease genes examined were expressed throughout feeding stages and downregulated in nonfeeding stages. The study will improve targeting of protease inhibitors and bacterial disruptors of SP synthesis.

Tri-trophic effects of transgenic insect-resistant tobacco expressing a protease inhibitor or a biotin-binding protein on adults of the predatory carabid beetle Ctenognathus novaezelandiae.

Tri-trophic impacts on adult predatory carabid beetles, Ctenognathus novaezelandiae, of insect-resistant transgenic tobacco plants expressing a serine protease inhibitor, bovine spleen trypsin inhibitor (BSTI), or a biotin-binding protein, avidin, were investigated. Both proteins could potentially affect this beetle, since avidin is known to be insecticidal to many beetle species and C. novaezelandiae midguts were shown to contain high levels of trypsin, a protease powerfully inhibited by bovine pancreatic trypsin inhibitor (a BSTI homologue) in vitro. Newly emerged field-collected adult C. novaezelandiae were fed exclusively for 280 days on Spodoptera litura larvae raised either on non-transgenic control, transgenic avidin (55 ppm) or transgenic BSTI (68 ppm) tobacco. Despite this long-term exclusive diet, there was no treatment effect on survival or fecundity and only minor and transient effects on beetles were observed. Data pooled across time and genders showed control-prey-fed beetles weighed 3% more than BSTI-prey-fed beetles and avidin-prey-fed beetles consumed 3-4% fewer prey than control- or BSTI-prey-fed individuals. Females in all treatments gained more mass and survived longer than males. Low exposure to the proteins because of dilution and deactivation within the prey is the most likely explanation for the lack of tri-trophic effects observed. Aditionally, the presence of a digestive chymotrypsin only partially inhibited by BSTI may provide an alternative path for proteolysis.

A Plant Chloroplast Glutamyl Proteinase.

A glutamyl proteinase was partially purified from Percoll gradient-purified spinach (Spinacia oleracea) chloroplast preparations and appeared to be predominantly localized in the chloroplast stroma. The enzyme degraded casein, but of the 11 synthetic endopeptidase substrates tested, only benzyloxycarbonyl-leucine-leucine-glutamic acid-[beta]-napthylamide was hydrolyzed at measurable rates. In addition, the enzyme cleaved the oxidized [beta]-chain of insulin after a glutamic acid residue. There was no evidence that native ribulose-1,5-bisphosphate carboxylase/oxygenase was cleaved by this proteinase. The apparent Km for benzyloxycarbonyl-leucine-leucine-glutamic acid-[beta]NA at the pH optimum of 8.0 was about 1 mM. Cl-ions were required for both activity and stability. Of the proteinase inhibitors covering all four classes of the endopeptidases, only 4-(2-aminoethyl)-benzenesulfonyl-fluoride HCl and L-1-chloro-3-[4-tosylamido]-4-phenyl-2-butanone significantly inhibited the proteinase. The partially purified enzyme had a molecular weight of about 350,000 to 380,000, based on size-exclusion chromatography. The enzyme has both similar and distinctive properties to those of the bacterial glutamyl proteinases. To our knowledge, this is the first description of a plant glutamyl proteinase found predominantly or exclusively in the chloroplast.

Characterization of major midgut proteinase cDNAs from Helicoverpa armigera larvae and changes in gene expression in response to four proteinase inhibitors in the diet.

A Helicoverpa armigera larval midgut cDNA library from larvae raised on an artificial, protein-rich, inhibitor-free diet contained very large numbers of serine proteinase positive clones. DNA sequencing of six random positive cDNAs and 12 PCR derived products identified trypsin genes classifiable into three families, and chymotrypsin and elastase genes classifiable into a single family each. Genomic blots established that the most highly expressed of the trypsin families contained about 18 genes, and that the chymotrypsin and elastase families contained about 14 and 2 genes respectively. The levels of mRNA corresponding to the highly expressed trypsin and chymotrypsin families were determined following chronic ingestion of four proteinase inhibitors. Compared to insects on an inhibitor-free diet, chymotrypsin mRNA was increased by all inhibitors, and trypsin mRNA levels decreased. This occurred independent of whether the inhibitor was solely a trypsin inhibitor (aprotinin), an inhibitor of both trypsin and chymotrypsin (proteinase inhibitor II, soybean trypsin inhibitor) or predominantly a chymotrypsin inhibitor (proteinase inhibitor I). Changing the protein level of the diet did not affect trypsin mRNA levels, but chymotrypsin mRNA levels decreased with increasing dietary protein.

Wounding induces a series of closely related trypsin/chymotrypsin inhibitory peptides in leaves of tobacco.

Wounding of tobacco (Nicotiana tabacum) leaves induced the expression of acid-stable trypsin/chymotrypsin inhibitory activity. Analysis by gel filtration determined that the inhibitory activity was contained within a fraction with a native M(r) of ca 5-7 x 10(3). Using ion-exchange column chromatography, this was resolved further into two major fractions, each of which inhibited both trypsin and chymotrypsin. Reverse-phase HPLC identified a total of six peptides from both fractions and each was purified to homogeneity. Four of these peptides inhibited both trypsin and chymotrypsin, a fifth inhibited trypsin only, while the sixth inhibited chymotrypsin almost exclusively. Sequencing of the N-terminal revealed that each peptide had an identical amino acid sequence and that these proteins are similar to a series of trypsin/chymotrypsin inhibitory peptides that are expressed predominantly in the stigmas of Nicotiana alata flowers.

A method to distinguish between chemical shift and susceptibility effects in NMR microscopy and its application to insect larvae.

We propose a simple method of distinguishing Zeeman broadening arising from susceptibility inhomogeneity and chemical shift variation, applicable to NMR microscopy. The method is based on the use of a specially built probe-head in which orthogonal sample alignment is possible using the same radiofrequency (RF) coil. This allows the investigation of alignment effects in image distortion and relies on the fact that the isotropic chemical shift is invariant under reorientation, whereas the susceptibility-related local field will depend strongly on relative orientation of bounding surfaces with the external polarizing field. We apply this approach to the study of a simple phantom, and an insect larva (Spodoptera litura Fabricius), demonstrating in the latter case that susceptibility variations are sufficiently small to allow chemical shift imaging on a scale greater than 1 ppm.

AtSUC2 has a role for sucrose retrieval along the phloem pathway: evidence from carbon-11 tracer studies.

The location of the phloem within a plant, and its vulnerability to disruption, make it a difficult tissue to study and therefore non-invasive studies of phloem functionality are important. Here we compare, phloem transport, measured non-invasively, in wild type Arabidopsis thaliana, and transposon-insertion mutants for AtSUC1 or AtSUC2, giving in vivo information on the importance of these sucrose transporters for phloem transport. The suc2 mutant showed an increase in both phloem leakage and transport time, consistent with reduced sucrose uptake into both transport and collection phloem. The results are consistent with the AtSUC2 transporter being important for retrieval of leaked sucrose in the transport phloem of Arabidopsis. There was no difference in phloem transport properties between the wild type and the suc1 mutants, implying that the AtSUC1 transporter does not play a significant role within the transport phloem of Arabidopsis under the conditions of our study.

Expressed sequence tags from the midgut of Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae).

The midgut is a key tissue in insect science. Physiological roles include digestion and peritrophic membrane function, as well as being an important target for insecticides. We used an expressed sequence tag (EST) approach to identify candidate genes and gene families involved in these processes in the light brown apple moth, Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae). Two cDNA libraries were constructed from dissected midgut of third to fifth instar larvae. Clustering analysis of 6416 expressed sequence tags produced 1178 tentative unique genes comprising 725 tentative contigs and 453 singletons. The sequences show similar codon usage to sequences from other lepidopterans, a Kozak consensus sequence similar to Drosophila and single nucleotide polymorphisms (SNPs) were detected at a frequency of 1.35/kb. The identity of the most common Interpro families correlates well with major known functions of the midgut. Phylogenetic analysis was conducted on representative sequences from selected multigene families. Gene families include a broad range of digestive proteases, lipases and carbohydrases that appear to have degradative capacity against the major food components found in leaves, the diet of these larvae; and carboxylesterases, glutathione-S-transferases and cytochrome P450 monooxygenases, potentially involved in xenobiotic degradation. Two of the larger multigene families, serine proteases and lipases, expressed a high proportion of genes that are likely to be catalytically inactive.

The effects of bivalent cations on ribulose bisphosphate carboxylase/oxygenase.

The half-saturation constants for binding of the bivalent cations (Mg2+, Ni2+, Co2+, Fe2+ and Mn2+) to ribulose bisphosphate carboxylase/oxygenase from Glycine max and Rhodospirillum rubrum were measured. The values obtained were dependent on the enzyme and the cation present, but were the same for both oxygenase and carboxylase activities. Ribulose bisphosphate rather than its cation complex was the true substrate. The kinetic parameters Vmax.(CO2), Vmax.(O2), Km(CO2), Km(O2), and K1(O2) were determined for both enzymes and each cation activator. The evolutionary and mechanistic implications of these data are discussed.

Purification of a trypsin inhibitor (PFTI) from pumpkin fruit phloem exudate and isolation of putative trypsin and chymotrypsin inhibitor cDNA clones.

The major trypsin inhibitor from pumpkin (Cucurbita maxima cv Supermarket Hybrid) fruit phloem exudate was purified by affinity and reverse phase chromatography. The protein has a molecular weight of approximately 8100 by SDS-PAGE and is blocked at the N-terminal serine. Following sequencing of a CNBr fragment, 3'- and 5'-RACE were used to isolate full length cDNAs corresponding to a trypsin inhibitor and to two chymotrypsin inhibitors. The three genes are similar, both in their translated and non-translated regions. Comparison of the full length translated proteins show that they are members of the proteinase inhibitor I family and almost identical apart from the P1 site in the proteinase binding loop. The genes encode proteins of 67 amino acids and appear to lack not only both pre- and prepro-peptide sequences but also the single disulphide present in most proteinase inhibitor I family members.

Effects of manganese ions and magnesium ions on the activity of soya-bean ribulose bisphosphate carboxylase/oxygenase.

The Michaelis constants of soya-bean ribulose bisphosphate carboxylase for CO2 in the carboxylation reaction and for O2 in the oxygenation reaction depend on the nature of the bivalent cation present. In the presence of Mg2+ the Km for bicarbonate is 2.48 mM, and the Km for O2 is 37% (gas-phase concentration). With Mn2+ the values decrease to 0.85 mM and 1.7% respectively. For the carboxylation reaction Vmax. was 1.7 mumol/min per mg of protein with Mg2+ but only 0.29 mumol/min per mg of protein with Mn2+. For the oxygenation reaction, Vmax. values were 0.61 and 0.29 mumol/min per mg of protein respectively with Mg2+ and Mn2+.

A kinetic study of ribulose bisphosphate carboxylase from the photosynthetic bacterium Rhodospirillum rubrum.

The activation kinetics of purified Rhodospirillum rubrum ribulose bisphosphate carboxylase were analysed. The equilibrium constant for activation by CO(2) was 600 micron and that for activation by Mg2+ was 90 micron, and the second-order activation constant for the reaction of CO(2) with inactive enzyme (k+1) was 0.25 X 10(-3)min-1 . micron-1. The latter value was considerably lower than the k+1 for higher-plant enzyme (7 X 10(-3)-10 X 10(-3)min-1 . micron-1). 6-Phosphogluconate had little effect on the active enzyme, and increased the extent of activation of inactive enzyme. Ribulose bisphosphate also increased the extent of activation and did not inhibit the rate of activation. This effect might have been mediated through a reaction product, 2-phosphoglycolic acid, which also stimulated the extent of activation of the enzyme. The active enzyme had a Km (CO2) of 300 micron-CO2, a Km (ribulose bisphosphate) of 11--18 micron-ribulose bisphosphate and a Vmax. of up to 3 mumol/min per mg of protein. These data are discussed in relation to the proposed model for activation and catalysis of ribulose bisphosphate carboxylase.

Phosphoglycolate phosphatase. Effect of cation and pH on activity.

P-glycolate phosphatase requires divalent cations for activity. Activity-pH curves identified 2 active site residues with pK values at pH 5.7 and pH 9.1 in the presence of magnesium and at pH 5.7 and pH 7.5 in the presence of manganese or cobalt. Saturation velocity kinetics enabled the identification of two distinct divalent cation binding sites. The first, nonspecific site has a K0.5 of 2 to 7 x 10(-5) M, depending on the cation and the pH. The second site, which is specific for magnesium, binds this cation in a negatively cooperative fashion. The affinity at pH 8.1 varies approximately 100-fold from the first magnesium bound to the fourth. The negative cooperativity is greatest at high pH. Because the pH range of activity is very broad, both the phosphate monoanion and dianion of P-glycolate must be bound as the substrate. The concentration of these two species at the apparent Km is independent of magnesium concentration. The P-glycolate.magnesium complex is kinetically inactive.

Phosphoglycolate phosphatase. Purification and properties.

Phosphoglycolate phosphatase (EC 3.1.3.18) was purified 1500-fold from field-grown tobacco leaves by acetone fractionation, DEAE-cellulose and molecular sieve chromatography, and preparative polyacrylamide gel electrophoresis. Preparations were judged 90 to 95% homogeneous by chromatography on DEAE-cellulose, polyacrylamide gel electrophoresis, and by isoelectric focusing. The highest specific activity obtained was 468 mumol of phosphate released/min/mg of protein. The native protein has a molecular weight of 80,500 by Ferguson plot analysis and 86,300 by sedimentation velocity on sucrose density gradients. Sodium dodecyl sulfate-polyacrylamide gels gave a molecular weight of 20,700, indicating the P-glycolate phosphatase is a tetramer with identical or near identical subunits. The enzyme, freshly purified or in crude homogenates, had a pI of 3.8 to 3.9 pH units by isoelectric focusing. Phosphosphoglycolate phosphatase from spinach leaves has a molecular weight of 93,000 and, unlike the enzyme from tobacco leaves, it is extremely unstable after DEAE-cellulose chromatography and is inactivated by lipase (EC 3.1.1.3). The phosphatase from both plants was stabilized by the addition of citrate or isocitrate in the buffers. Ribose 5-phosphate is a competitive inhibitor of phosphoglycolate phosphatase at physiological concentration, while other phosphate esters of the photosynthetic carbon cycle were without effect.