agronomy A Wild Cajanus scarabaeoides (L.), Thouars, IBS 3471, for Improved Insect-Resistance in Cultivated Pigeonpea

: Cajanus scarabaeoides (L.), Thouars, is the closest wild crop relative of cultivated pigeonpea, Cajanus cajan (L.), Millspaugh. This wild pigeonpea has several insect-resistance mechanisms, particularly to Helicoverpa armigera (Hübner). Estimated economic losses in the semi-arid tropics from H. armigera damage in pigeonpea are approximately two billion USD / year. Therefore, it is imperative to improve pest resistance in this crop. In this study, we investigated insect-resistance components in IBS 3471, a C. scarabaeoides accession, and explored the possibility of transferring resistance mechanism / s to cultivated pigeonpea. A detached leaf bioassay revealed that IBS 3471 has more e ﬀ ective antibiosis and antixenosis resistance mechanisms against H. armigera compared to the susceptible C. cajan variety, ICPL 87. To further investigate the antibiosis resistance mechanism, we fed H. armigera larvae a heated and non-heated artiﬁcial diet supplemented with lyophilised IBS 3471 leaf powder. Incorporation of IBS 3471 leaf powder inhibited H. armigera larval weight and delayed larval development compared to larvae reared on diet supplemented with ICPL 87 leaf powder. The putative insect-resistance compounds in C. scarabaeoides were heat-labile. Proteomic analysis revealed higher levels of potential insecticidal proteins, namely lectin and cysteine proteinase inhibitor, in wild pigeonpea compared to the cultivated variety. Nutritional analysis and interspeciﬁc hybridisation experiments also indicated that IBS 3471 is a potential candidate for improvement of insect-resistance in pigeonpea. This study demonstrates that IBS 3471 has multiple resistance mechanisms against H. armigera, and they are transferable to cultivated pigeonpea.


Introduction
Pigeonpea (Cajanus cajan (L.) Millspaugh) is an important crop that sustains the livelihood of most people in the semi-arid tropics (SAT), where they depend on rain-fed agriculture [1]. It is grown worldwide and has an annual production approaching seven million tonnes per year [2]. Dry, mature seeds of pigeonpea are nutritionally rich in protein, carbohydrates, B-group vitamins, amino acids, calcium, fibre and iron [3]. Lack of high-yielding cultivars resistant to biotic stress is one of the major contributors for yield plateau and losses in pigeonpea, and the most important biotic stress is the insect pest, Helicoverpa armigera (Hübner), the pod borer [4]. H. armigera damage is a significant constraint in Agronomy 2020, 10, 517 3 of 16

Plant Material
IBS 3471 was selected out of 11 C. scarabaeoides accessions, based on a preliminary insect-resistance screening at the Queensland University of Technology (QUT), Centre for Agriculture and the Bioeconomy (CAB) laboratory (Supplementary Table S1 and Figure S1). ICPL 87, a high-yielding C. cajan variety, a universal susceptible check to H. armigera, was chosen as a control. All seeds used in this study were obtained from the Australian Grains Genebank (AGG) (Horsham, Victoria, Australia), except for the less susceptible ICPL 20338, the female parent in the interspecific hybridisation, which was obtained from the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), (Patancheru, Hyderabad, India). Mechanically scarified IBS 3471, ICPL 20338 and ICPL 87 seeds were germinated in Petri dishes (90 mm), with moist filter paper, in an incubator, at 36 • C. The germinated seeds were planted in Searles premium potting mix, in an environmentally controlled glasshouse (27 ± 1 • C, 13 h light and 11 h dark), at QUT Carseldine facilities, (Brisbane, Australia). Bamboo sticks were used to provide support for C. scarabaeoides accession, which has a climbing habit. No pesticide was applied to the plants during the experiments.

Insect Culture
H. armigera egg colonies were obtained from the Commonwealth Scientific and Industrial Research Organisation (CSIRO), (Narrabri, New South Wales, Australia). Eggs were hatched at 25 • C, and the neonate larvae were transferred to freshly prepared Stonefly Heliothis diet (Ward's scienceX), by mixing the powder diet with water-vinegar (19:1). Moths were fed on 10% honey (diluted with water) and placed in a Sanyo Versatile environmental test chamber (Model MLR-350H, Japan) at 27 ± 1 • C, with 16 h light and 8 h dark. Larval growth instar stages were categorised as described in the Queensland government department of primary industries and fisheries guideline [36].

Screening for Antibiosis and Antixenosis Defence Mechanisms in Wild and Cultivated Pigeonpea, Using Detached Leaf Bioassay
Five H. armigera eggs were placed on young, freshly harvested and fully expanded first trifoliate leaves of IBS 3471, ICPL 87, ICPL 20338 and an F 1 interspecific hybrid (IBS 3471×ICPL 20338), in Petri dishes with moist filter paper, and replicated ten times. The dishes were placed in an environmentally controlled growth room, at 25 ± 1 • C, with 16 h light and 8 h dark. The leaves were replaced with freshly harvested leaves every 24 h. First and second instar larvae of H. armigera were reared as groups, while larvae from the third instar and beyond were reared individually, to avoid cannibalism. The larval weights were recorded after eight days.

Trichome Identification
Scanning electron microscopy (SEM) was used to identify the types of trichomes on leaves and pods of wild and cultivated pigeonpea. Leaf samples were cut into three pieces, while for the pods, the epidermis was peeled. Both samples were flooded with 2.5% glutaraldehyde in 0.2 M cacodylate buffer solution and soaked overnight in a fume hood. After 24 h, the samples were rinsed three times with 0.2 M cacodylate buffer. After each rinse, the samples were soaked for ten minutes. Samples were dehydrated, using graded ethanol (EtOH) series, with each concentration step repeated twice, for 10 min, before moving onto the next concentration, as follows; 50%, 70%, 90% and 100% EtOH. The last step was repeated three times, each for 15 min. Dehydrated samples were mounted on SEM carbon tabs coated with gold, using Leica EM-SCD005 Sputter Coater, and placed on SEM stub holders and stored in a cool dry place, until the samples were analysed, using a Zeiss Sigma VP Field Emission SEM at 100× magnification. Each accession had three leaves and three pods, and each sample was scanned three times. The first fully expanded leaves of IBS 3471, ICPL 87, ICPL 20338 and an F 2 interspecific hybrid from self-pollinated interspecific F 1 (IBS 3471×ICPL 20338) were harvested from glasshouse-grown plants and transported to the laboratory on dry ice and stored at −80 • C. Due to a limitation in the number of F 1 interspecific hybrid plants, experiments for antixenosis and antibiosis bioassays using artificial diet supplemented with lyophilised leaf powder were carried out, using F 2 interspecific hybrids. Leaves were freeze-dried for 72 h, to minimize any changes in biochemical composition, using a Benchtop Pro with Omnitronics™ freeze dryer (VirTis SP Scientific, USA). The freeze-dried leaves were milled, using a Tube Mill Control (IKA Mills No. 0004180000, China) for 25 s, at 25,000 rpm. The artificial diet was prepared by mixing the Stonefly Heliothis diet (Ward's ScienceX) powder (21.7 g), lyophilised leaf powder (3.3 g) and 100 mL of water-vinegar (19:1). A plain artificial diet (PAD) without lyophilised leaf powder was used as a control. To investigate whether potential inhibitors were heat-labile, the effects of heated and non-heated diets were assessed. For heating, the mixed diets were microwaved for 90 s (750W, Sharp Carousel R-5A12).
A moist paintbrush was used to place neonate H. armigera larvae into each well of a 32-cell rearing tray (RT32W, Frontier Scientific Services Agriculture, USA). The tray was sealed with a clear rearing tray Lid-4 Cells (RTCV4, Frontier Scientific Services Agriculture, USA) and placed in a Sanyo Versatile environmental test chamber (Model MLR-350H, Japan), at 26 ± 1 • C, 16 h light and 8 h dark. The larval weights were recorded after five days, and each accession had 64 replications.

Protein Extraction and Digestion
Leaves of IBS 3471 and ICPL 87 were harvested at both vegetative and flowering/podding stages and immediately placed on dry ice, for transport to the laboratory. The leaves were snap-frozen with liquid nitrogen and ground with a mortar and pestle. Protein extraction and digestion were as described by Wu et al. [37], using the trichloroacetic acid-acetone method.
2.6.2. Data Acquisition Using Nanoflow Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry (NanoLC ESI MS/MS) Samples were tagged by tandem mass tag (TMT) reagent labelling of peptides (Thermo Fisher Scientific, USA), as per manufacturer's protocol. The labelled samples were prepared for MS analysis by reconstituting the high pH High-performance liquid chromatography (HPLC) fractions of each TMT set with 2% acetonitrile, 97.9% water and 0.1% formic acid (sample loading buffer). The samples were subjected to NanoLC-MS/MS system (Easy nLC 1000, Thermo Fisher Scientific, USA) and 1D Data-dependent acquisition (DDA) of peptides on Q-Exactive Quadrupole-Orbitrap (QE-Classic, Thermo Fisher Scientific, USA). The TMT labelled peptide samples were injected onto an in-house packed trap column (Halo C18, 160 Å, 2.7 µm, 100 µm × 3.5 cm) and desalted with loading buffer. Peptides were eluted from the trap into an in-house packed analytical column with the linear gradients of mobile phase A (99.9% water, 0.1% formic acid) and B (99.9% acetonitrile, 0.1% formic acid): mobile phase B (30%) over 110 min with a flow rate of 300 nL/min across the gradient. The eluent from the trap was separated over the analytical column (Halo C18, 160 Å, 2.7 µm, 75 µm × 15 cm). The column eluent was directed into the ionization source of the mass spectrometer. A 2.6 kV electrospray voltage was applied via a liquid junction upstream of the column. Peptide precursors from 350 to 1850 m/z were scanned at 70,000 resolution with an Automatic Gain Control (AGC) target value of 1 × 10 6 ions. The ten most intense ions from the preceding survey scan were fragmented by Higher-Energy Collisional Dissociation (HCD), using a normalized collision energy of 35% with an isolation width of 0.7 m/z. For MS/MS analysis, only precursors with a charge state of +2 to +4 were used. The MS method had a minimum signal required value of 2.5 × 10 4 for MS2 triggering, an AGC target value of Agronomy 2020, 10, 517 5 of 16 2 × 10 5 for MS2 and a maximum injection time of 250 ms for MS2. MS/MS scan resolution was set at 70,000 resolution, and the dynamic exclusion was 90 s.

Protein Identification and Quantification
The raw data files from NanoLC ESI MS/MS were processed, using Proteome Discoverer version 2.1 (Thermo Fisher Scientific, Waltham, MA, USA). Subsequently, SequestHT and Mascot (Matrix Science, London, UK) search engines were used to process the data against C. cajan and C. scarabaeoides sequences downloaded from the UniProt database (47,669 sequences accessed on 19th July 2019 [38]). Conditions chosen for analysis were trypsin with two missed cleavage points, and the MS precursor mass tolerance was ±20 ppm, with an MS/MS fragment mass tolerance of 0.02 Da. The dynamic modifications were oxidation of methionine, deamination of asparagine, glutamine and pyroglutamate, acetylation of protein N-terminus, Met-loss+Acetyl (Sequest), and TMT6plex tag on lysine residues and the peptide N-terminus. Static modification parameter was carbamido-methylation of cysteine. False discovery rate (FDR) was set at <1%, while the display filters were Protein, Peptide and PSM Master proteins only.

Generation of Interspecific Hybrids
ICPL 20338, a super-early maturing pigeonpea accession with the capability of escape mechanism for insects, was used as the female parent. To synchronise flowering, the male parent IBS 3471 was sown six weeks before the female parent (ICPL 20338). An artificial hybridisation procedure described by Singh and Oswalt [39] was used to generate the interspecific F 1 hybrids. The F 1 hybrids were self-pollinated to generate an F 2 generation.

Statistical Analysis
All the collected data, including larvae weights, larval lifecycle (days), number of pupa and trichomes, and protein levels, were analysed using Minitab ® Statistical Software (Version 18.1). Normal distribution and homogeneous variance were checked by using Normal probability plot of residuals. All the parameters comparing two means were analysed, using the independent t-test (the p-value of a Student's t-test < 0.05 and p < 0.01), while analysis of variance (one-way ANOVA) and post hoc mean comparisons with Tukey's HSD test at p < 0.05 and p < 0.01 was conducted for all parameters, comparing multiple means. Results were presented as the means ± SE.

IBS 3471 Leaves Inhibited Growth and Development of Helicoverpa Armigera Larvae
On the preliminary plant choice bioassay done at the beginning of this study, using leaves, flowers and pods from the different C. scarabaeoides and C. cajan accessions, we observed that the larvae prefer to feed on cultivated pigeonpea and would feed on the wild pigeonpea if they have exhausted the cultivated pigeonpea (data not showed). Therefore, the intact leaf bioassay without choice was carried out, using the most preferred cultivated pigeonpea (ICPL 87) and the least preferred wild pigeonpea (IBS 3471).
H. armigera larval weight after eight days fed on C. scarabaeoides accession, IBS 3471, displayed significant (t 18 = 8.42, p < 0.001) inhibition of growth when compared to cultivated pigeonpea, C. cajan ICPL 87. The larvae fed on IBS 3471 leaves were 100.5 mg, while the larvae on ICPL 87 weighed 333.6 mg ( Figure 1A). The development of H. armigera larvae was also delayed, leading to a prolonged lifecycle when fed on the leaves of IBS 3471 compared to ICPL 87 ( Figure 1B). Significant differences (t 18 = 5.76, p < 0.001) were observed between the larvae reared on ICPL 87 and IBS 3471 leaves. Larvae reared on ICPL 87 leaves took an average of 17 days to complete development and 28 days when reared on IBS 3471 leaves ( Figure 1B).

Trichome Identification Revealed a Strong Antixenosis Defence Mechanism in IBS 3471
In this study, the type (A, B, C and D), density, distribution, orientation and length of the trichomes differed considerably between IBS 3471 and ICPL 87 ( Figure 2). Type A trichomes, which are glandular, produce exudates that attract and act as phagostimulant to H. armigera larvae [11], were only observed in ICPL 87. Type A trichomes were not observed on IBS 3471 leaves or pods. Leaves contained fewer type B trichomes compared to pods for both species (Figure 2). Moreover, the overall density of type B trichomes was lower in both species, compared to other trichome types. Both leaves and pods of IBS 3471 contained a high density of type C and D trichomes. In contrast, the leaves of ICPL 87 contained very few type C and D trichomes (Figure 2).

Trichome Identification Revealed a Strong Antixenosis Defence Mechanism in IBS 3471
In this study, the type (A, B, C and D), density, distribution, orientation and length of the trichomes differed considerably between IBS 3471 and ICPL 87 (Figure 2). Type A trichomes, which are glandular, produce exudates that attract and act as phagostimulant to H. armigera larvae [11], were only observed in ICPL 87. Type A trichomes were not observed on IBS 3471 leaves or pods. Leaves contained fewer type B trichomes compared to pods for both species (Figure 2). Moreover, the overall density of type B trichomes was lower in both species, compared to other trichome types. Both leaves and pods of IBS 3471 contained a high density of type C and D trichomes. In contrast, the leaves of ICPL 87 contained very few type C and D trichomes (Figure 2).

IBS 3471 Possesses Both Antibiosis and Antixenosis Insect-Resistance Mechanisms
The results from intact-leaf bioassay and trichome characterisation indicate that IBS 3471 significantly (t 171 = 7.75, p < 0.001) inhibited the growth and development of H. armigera by antixenosis mechanism. To understand whether the inhibition was by antixenosis or whether antibiosis also played a role, H. armigera larvae were reared on an artificial diet supplemented with lyophilised leaf powder from the two genotypes. H. armigera larval weights were significantly (t 124 = 6.98, p < 0.001) lower when reared on an artificial diet supplemented with lyophilised leaf powder from 11 different C. scarabaeoides accessions, compared with larvae reared on the artificial diet with lyophilised leaf powder of susceptible cultivated pigeonpea variety, ICPL 87, (Supplementary Figure S1). Larvae fed on an artificial diet supplemented with ICPL 87 lyophilised leaf powder had the highest average weight of 304.6 mg, compared to those fed on an artificial diet supplemented with lyophilised leaf powder of IBS 3471, which had the least average weight of 182.3 mg by day five ( Figure 3A). Delayed development was also observed in larvae fed on an artificial diet supplemented with leaf powder of IBS 3471 compared to ICPL 87. Approximately 42% of larvae reared on an artificial diet supplemented with lyophilised leaf powder of IBS 3471 had pupated within 11 days, whereas 88% of the larvae fed on an artificial diet with leaf powder of ICPL 87 had already pupated by day 11 ( Figure 3B).

IBS 3471 Possesses Both Antibiosis and Antixenosis Insect-Resistance Mechanisms
The results from intact-leaf bioassay and trichome characterisation indicate that IBS 3471 significantly (t171 = 7.75, p < 0.001) inhibited the growth and development of H. armigera by antixenosis mechanism. To understand whether the inhibition was by antixenosis or whether antibiosis also played a role, H. armigera larvae were reared on an artificial diet supplemented with lyophilised leaf powder from the two genotypes. H. armigera larval weights were significantly (t124 = 6.98, p < 0.001) powder of IBS 3471, which had the least average weight of 182.3 mg by day five ( Figure 3A). Delayed development was also observed in larvae fed on an artificial diet supplemented with leaf powder of IBS 3471 compared to ICPL 87. Approximately 42% of larvae reared on an artificial diet supplemented with lyophilised leaf powder of IBS 3471 had pupated within 11 days, whereas 88% of the larvae fed on an artificial diet with leaf powder of ICPL 87 had already pupated by day 11 ( Figure 3B).

Insect-Resistance Factors in Pigeonpea Were Heat-Labile
To further investigate the component/s contributing to insect-resistance in pigeonpea, we heated the artificial diet supplemented with lyophilised leaf powder of C. scarabaeoides and C. cajan and assessed the growth of H. armigera for differences compared to non-heated media. A significant difference (F5,377 = 43.89, p < 0.001) in the weight of H. armigera larvae fed on the heated and non-heated artificial diets was observed. Average larval weight of H. armigera (day five) reared on heat-treated leaf powder was 244.8 and 384.27 mg for IBS 3471 and plain artificial diet (PAD), respectively ( Figure  4A). Meanwhile, the weights of larvae that were fed on a non-heat-treated artificial diet with leaf powder had less weight, with 182.3 mg for IBS 3471 and 354.4 mg for PAD on day five ( Figure 4A). Larvae fed on the artificial diet with ICPL 87 leaf powder for both heat and non-heat-treated diet had a higher weight when compared with larvae fed on the artificial diet with IBS 3471 leaf powder. Statistically, no significant (p < 0.01) differences were observed on heated or non-heated PAD, confirming that the heating process does not destroy any ingredient in the diet, contributing to the experimental differences ( Figure 4A).

Insect-Resistance Factors in Pigeonpea Were Heat-Labile
To further investigate the component/s contributing to insect-resistance in pigeonpea, we heated the artificial diet supplemented with lyophilised leaf powder of C. scarabaeoides and C. cajan and assessed the growth of H. armigera for differences compared to non-heated media. A significant difference (F 5,377 = 43.89, p < 0.001) in the weight of H. armigera larvae fed on the heated and non-heated artificial diets was observed. Average larval weight of H. armigera (day five) reared on heat-treated leaf powder was 244.8 and 384.27 mg for IBS 3471 and plain artificial diet (PAD), respectively ( Figure 4A). Meanwhile, the weights of larvae that were fed on a non-heat-treated artificial diet with leaf powder had less weight, with 182.3 mg for IBS 3471 and 354.4 mg for PAD on day five ( Figure 4A). Larvae fed on the artificial diet with ICPL 87 leaf powder for both heat and non-heat-treated diet had a higher weight when compared with larvae fed on the artificial diet with IBS 3471 leaf powder. Statistically, no significant (p < 0.01) differences were observed on heated or non-heated PAD, confirming that the heating process does not destroy any ingredient in the diet, contributing to the experimental differences ( Figure 4A).  There was less profound delayed development on the larvae fed on both heated and non-heated artificial diet supplemented with lyophilised leaf powder of ICPL 87 ( Figure 4B). However, no delay There was less profound delayed development on the larvae fed on both heated and non-heated artificial diet supplemented with lyophilised leaf powder of ICPL 87 ( Figure 4B). However, no delay in the development for larvae reared on both heated and non-heated PAD was observed ( Figure 4B). By day 11, 100% of the larvae fed on a heat-treated artificial diet supplemented with lyophilised leaf powder of ICPL 87 had pupated, consistent with those fed on PAD. In contrast, 88% of the larvae fed on a non-heated artificial diet with lyophilised powder of ICPL 87 had pupated ( Figure 4B). Seventy per cent of larvae fed on a heat-treated artificial diet supplemented with lyophilised leaf powder of IBS 3471 pupated by day 11, whereas only 42% pupation was observed on the larvae fed on a non-heated artificial diet supplemented with lyophilised leaf powder of IBS 3471 ( Figure 4B). These results indicated that the effects of insect-resistance component/s in leaves of pigeonpea were significantly (F 5, 378 = 33.0, p < 0.001) reduced by heat.

Proteomic Analysis Revealed High Presence of Two Potential Defence Proteins for Helicoverpa Armigera Resistance in IBS 3471
To investigate heat-labile insect-resistance components in pigeonpea, we performed a proteomic analysis on IBS 3471 and ICPL 87 genotypes. There were significant differences (t(2) = 6.92, p = 0.02, t(4) = 2.81, p = 0.048) in the levels of cysteine proteinase inhibitors ( Figure 5A,B, respectively, and Supplementary Table S3). Leaves at the vegetative stage of IBS 3471 contained a 5.1-times-higher level of cysteine protease inhibitor than leaves of ICPL 87 at the same stage ( Figure 5A,B). There was a two-fold reduction in cysteine proteinase inhibitor detected during the flowering and podding stage in IBS 3471 and a minimal increase in ICPL 87 ( Figure 5B). There were significant differences (t(2) = 14.41, p = 0.005, t(4) = 9.09, p = 0.001) in the levels of lectin-domain containing receptor kinase ( Figure 5C,D, respectively, and Supplementary Table S3). The lectin-domain containing receptor kinase was 2.5 times higher in the leaf of IBS 3471 than in leaf of ICPL 87, at the vegetative stage ( Figure 5C). For the flowering and podding stage, IBS 3471 accumulated two times more lectin-domain containing receptor kinase than ICPL 87 ( Figure 5D).

Insect-Resistance Components in IBS 3471 Are Transferrable to Cultivated Pigeonpea
Generation of interspecific hybrids is one possible strategy of improving the resistance of cultivated pigeonpea against pest damage, especially pod borers. In this study, F 1 and F 2 interspecific hybrids were generated and feeding bioassays conducted. The insect-resistance levels in F 1 and F 2 hybrids were comparable to their paternal parent IBS 3471 ( Figure 6A). The larvae fed on the maternal parent ICPL 20338 diet had more weight (161.1 mg) than larvae fed on the F 1 hybrid (58.5 mg) and IBS 3471 (54.3 mg) by day eight ( Figure 6A). Larvae fed on an artificial diet supplemented with lyophilised leaf powder of the paternal parent IBS 3471 were 167.8 mg, those on the hybrid F 2 generation diet were 182.4 mg and those on the maternal parent ICPL 20338 were 323.6 mg by day five ( Figure 6B). No significant differences (F 2,93 = 24.46, p < 0.001) were observed in larval weight between IBS 3471 and the F 2 hybrid ( Figure 6B). Agronomy 2020, 10, x FOR PEER REVIEW 10 of 16

Insect-Resistance Components in IBS 3471 Are Transferrable to Cultivated Pigeonpea
Generation of interspecific hybrids is one possible strategy of improving the resistance of cultivated pigeonpea against pest damage, especially pod borers. In this study, F1 and F2 interspecific hybrids were generated and feeding bioassays conducted. The insect-resistance levels in F1 and F2 hybrids were comparable to their paternal parent IBS 3471 ( Figure 6A). The larvae fed on the maternal parent ICPL 20338 diet had more weight (161.1 mg) than larvae fed on the F1 hybrid (58.5 mg) and IBS 3471 (54.3 mg) by day eight ( Figure 6A). Larvae fed on an artificial diet supplemented with lyophilised leaf powder of the paternal parent IBS 3471 were 167.8 mg, those on the hybrid F2 generation diet were 182.4 mg and those on the maternal parent ICPL 20338 were 323.6 mg by day five ( Figure 6B). No significant differences (F2,93 = 24.46, p < 0.001) were observed in larval weight between IBS 3471 and the F2 hybrid ( Figure 6B). Analysis of the trichomes present on the leaves and pods indicated that the F 2 hybrid contained similar trichome types, density and length to its highly insect-resistant paternal parent, IBS 3471 ( Figure 6C-I). These results confirm that insect-resistance components from IBS 3471 were transferred to the progeny.

Discussion
Crop improvement programs depend heavily on the availability of relevant germplasm. To utilise the resources available in a gene bank, the genetic diversity of the germplasm collection needs to be characterised [40]. Helicoverpa armigera is a major threat affecting pigeonpea production worldwide [5]. Wild pigeonpea has a high resistance to H. armigera and therefore has potential as a donor source for this trait to be incorporated into cultivated pigeonpea, provided there is no yield drag associated with the transfer [5]. In this study, we investigated the insect-resistance components in one of the closest relatives to cultivated pigeonpea, namely C. scarabaeoides, and explored the possibility of transferring one or more of the insect-resistant mechanisms to cultivated pigeonpea C. cajan.
Out of the 11 C. scarabaeoides accessions obtained from the Australian Grains Genebank (Supplementary Table S1 and Figure S1), IBS 3471 accession demonstrated the highest level of resistance to H. armigera. In addition, IBS 3471 accession had better nutritional profile (Supplementary  Table S2), making it an ideal candidate for use in an interspecific hybridisation program. This study proposes that IBS 3471 possesses both antixenosis (non-host preference) and antibiosis (harmful effect on insect biology) insect-resistance mechanisms. The antixenosis resistance mechanism present in IBS 3471 seemed to be related to the presence of dense, long and various types of trichomes on the leaves and pods (Figure 2). Trichomes provide insect-resistance by producing sticky exudates which hinder the insect, producing toxic chemicals which poison the insect through contact/ingestion/inhalation and acting as physical barriers preventing insect contact with the plant [41]. Trichome types, orientation, density and length influence insect damage in several crops [7]. Pigeonpea has five distinct types of trichomes (type A, B, C, D and E) that are identified by light or scanning electron microscopy [8]. Type A trichomes are glandular and secrete exudates, which act as attractants and phagostimulants toward H. armigera larvae [11,42]. Type B trichomes are globular and comprise an oily substance, which gives pigeonpea the characteristic fragrant smell [43]. The contents of type B trichomes are released upon the breakdown of the cell wall, caused either by chewing insects (H. armigera) or by abiotic factors (e.g., high temperatures or low air humidity). Type C and D are non-glandular trichomes. Type D trichomes are longer than type C [11,42]. Type E trichomes are multilobed, glandular and small, and therefore can only be identified by electron microscopy [11]. Results from this study showed that the types, density, distribution, orientation and length of the trichomes in leaf and pod varied considerably between IBS 3471 and ICPL 87, correlating with the different capacities of these genotypes to resist insect-feeding bioassay (Figure 2). Type A trichomes that act as H. armigera larvae attractants and phagostimulants were only present in the cultivated pigeonpea variety, ICPL 87 ( Figure 2C,D). Conversely, the non-glandular trichomes (type C and D), which may act as a physical barrier to feeding by H. armigera larvae [11,42,44], were present at much higher densities on the leaves and pods of IBS 3471, compared to ICPL 87 (Figure 2). Bioassays using intact detached leaves demonstrated that H. armigera larvae reared on C. scarabaeoides leaves, which have dense and long trichomes, had significantly lower weights compared to larvae reared on ICPL 87. This result coincides with the findings reported by Green et al. [42] and Sujana et al. [44].
To enable the improvement of insect-resistance in crops through breeding programs, donor parents should possess as many insect-resistant mechanisms as possible. In this study, we found that IBS 3471 possesses antixenosis, as well as antibiosis, resistance mechanisms. H. armigera larvae reared on an artificial diet supplemented with lyophilised IBS 3471 leaf powder had less weight and a prolonged developmental period, as compared to the larval reared on an artificial diet supplemented with lyophilised leaf powder of ICPL 87 (Figures 3 and 4). The differential of larval weight suggests the presence of more effective antibiosis defence systems in C. scarabaeoides, potentially involving proteinase inhibitors, phenolic or flavonoid compounds [45,46]. Delayed insect development is a good indicator of an antibiosis defence mechanism against pests [44]. Practically, delays in the pest growth cycle provide an opportunity to control H. armigera larvae through application of insecticides at the early stage of larval growth, before it reaches the destructive stage, and also insecticides are more effective on small larvae (below fifth instar) [36].
Plant protease inhibitors (PPIs) play crucial roles in plant defence against insects, pathogenic nematodes and fungi. They are also among the most abundant proteins in plants [3,47,48]. Several in vitro and in vivo studies showed that PPIs are toxic to larval gut proteases, causing digestion defects and disruption of amino acid absorption. Besides, PPIs result in retarded growth and development of the larvae, consequently affecting the fecundity and productivity of the moth [13]. Similarly, C. scarabaeoides accession IBS 3471 contained the largest quantities of cysteine proteinase inhibitors and lectin-domain containing receptor kinases ( Figure 5), resulting in the most significant inhibition to H. armigera larval growth (Figures 1, 3 and 4; Supplementary Table S3). Swathi et al. [49] had a similar observation, where pigeonpea protease inhibitors affected H. armigera gut proteinases. In addition, Prasad et al. [50] reported reduced activity of proteinases in the midgut of Manduca sexta larvae by proteinase inhibitors from pigeonpea and black gram.
Plant domestication has produced varieties containing low genetic variation. This is particularly true for self-pollinated crops such as pigeonpea. This is a result of using breeding methods which exclusively involved selection and mating of genetically related individuals, to generate elite varieties. This may have resulted in random loss of alleles, leading to genetic erosion, causing several genetic bottlenecks [51]. This is changing, as modern breeding programs are broadening and improving cultivated gene pools by using crop wild relatives [4]. Therefore, the use of wide hybridisation to introgress useful genes, such as insect-resistance genes, is one mechanism of broadening the genetic base of cultivated pigeonpea to combat pod borers.
The interspecific hybrid and the F 2 population derived from it inherited the insect-resistance mechanisms that inhibit the growth and development of the H. armigera larvae from the paternal parent C. scarabaeoides IBS 3471 ( Figure 6). Similar results were reported by Dodia et al. [52], in which they observed the adverse effects of C. scarabaeoides and their F 1 derivates on H. armigera larvae. Together with previous studies, this research confirms that wild pigeonpea, C. scarabaeoides, can be used as a donor for transferring pod-borer-resistance into cultivars through wide hybridisation, as suggested and observed by [3,[52][53][54].

Conclusions
The global population is growing at an unprecedented rate, and the demand for plant-based proteins is set to increase. Pigeonpea is an important pulse crop for the semi-arid tropics. It is one of the premium legumes and a major source of protein to over 20% of the world's population based largely on the Indian subcontinent. Despite its importance, pigeonpea productivity is substantially hindered by insect pressure such as Helicoverpa armigera (pod borer). This study has identified IBS 3471, a C. scarabaeoides accession, as a potential candidate for improvement of insect-resistance in pigeonpea. Further experiments are in progress to gain a better understanding of how the two-potential insect-resistant proteins, namely cysteine proteinase inhibitor and lectin domain-containing receptor kinase in IBS 3471, are involved in H. armigera resistance. This study has shown that the resistance traits from wild pigeonpea are transferable and stable at F 2 population; therefore, this can be used to improve insect resistance in a largescale breeding programs involving different cultivated pigeonpea with multiple resistance to H. armigera.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4395/10/4/517/s1. Table S1. Preliminary screening for insect-resistance, using non-heat-treated artificial diet supplemented with lyophilised leaf powder from 11 accessions of C. scarabaeoides and C. cajan received from Australian Grains Genebank. Figure S1. Average larval weight of H. armigera fed on different accessions of C. scarabaeoides and