Biochemical basis of resistance to pod borer (Helicoverpa armigera) in Australian wild relatives of pigeonpea

The domestication of pigeonpea has severely impacted the intrinsic host‐plant resistance (HPR) to pest and diseases, particularly pod borer (Helicoverpa armigera hubner). This study with 41 Australian wild Cajanus genotypes and interspecific hybrids demonstrated a high level of resistance to H. armigera in the accessions of Cajanus acutifolius, C. latisepalus, C. lanceolatus, C. pubescens, and C. reticulatus var. reticulatus. Significant variation in herbivory development and mortality (P < 0.001) was observed in the wild accessions and their hybrids in response to feeding on leaves. A strong positive relationship (R2 = 0.69, P < 0.001) between total phenolic compounds (TPC) and the HPR was observed. Australian wild genotypes demonstrated the role of TPC and the absence of certain flavonoids such as rutin and quercetin in resistant genotypes. The detached leaf bioassay technique separated the wild and domesticated accessions into wild resistant, with herbivory weight difference (HWD) (Day 7–Day 1) ranging between −27 ‐ 104 mg, wild susceptible, with HWD ranging between 124 ‐ 207 mg and domesticated susceptible, with HWD ranging from 208 ‐ 300 mg. Similarly, based on TPC, accessions were also categorised into wild high TPC, with TPC ranging between 32.3 ‐ 42.5 GAE mg/g DW, and wild low TPC had only 17.2–24.8 GAE mg/g DW. Low TPC concentrations were found in domesticated pigeonpea, with 10.7–17.6 GAE mg/g DW. The presence of very high concentrations of the flavone isoorientin, an important antioxidant implicated in the intracellular defence mechanism of cancer therapy, was identified for the first time in wild species of pigeonpea.

second position in Indian pulse production. However, the productivity of pigeonpea is severely constrained by biotic factors such as pod borer, Helicoverpa armigera (Hubner), causing crop losses of around two billion USD/year (Abigail et al., 2020;Shanower & Minja, 1999).
Screening of 10,000 accessions of the worlds' pigeonpea germplasm at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) did not reveal stable host plant resistance (HPR) to H. armigera (Srivastava et al., 1990;Upadhyaya et al., 2011). In recent years, wild species of pigeonpea have attracted attention as a source of resistance to biotic and abiotic stresses. Various biochemical markers, such as stilbenes, oxalic acid, and malic acid (Gadge et al., 2015) for resistance, high soluble sugars, low soluble protein, low condensed tannins, and low phenols for susceptibility were identified (Sharma & Manohar, 2009). For instance, high oxalic acid and malic acid concentrations were essential in H. armigera resistant btchickpea influencing cry1Ac expression. The anti-feedant activity of the acids might interact with the bt-endotoxins, consequently lowering consumption (Surekha & Arjuna, 2013) by the larvae. Macfoy et al. (1983) have reported amino acids and sugars in a susceptible cultivar of cowpea genotype, Vita-1, in higher concentrations than in the resistant variety, emphasizing that the susceptibility could be due to soluble sugars favouring the insect feeding. Likewise, the presence of non-glandular trichomes or low density glandular trichomes was identified as morphological markers for tolerance to H. armigera (Glas et al., 2012;Romeis et al., 1999;Sharma & Manohar, 2009;Sujana et al., 2012). The non-glandular trichome density was reported to be associated with the oviposition of H. armigera, whitefly and spider mites (Asif et al., 2019;Rakha & Ramasamy, 2017) in resistant potato accessions. Rashid et al. (2012) have identified the sticky exudates from the trichomes, such as flavonoids, terpenoids, and alkaloids (Rashid et al., 2012), which produce toxic chemicals that hinder the insect's growth, causing antibiosis. They may also act as a physical barrier preventing the insect-plant interaction (Abigail et al., 2020;Ranger & Hower, 2001). Enzymatic markers such as trypsin proteinase inhibitors (TPIs) (Swathi et al., 2016), amylase inhibitors (AIs) (Gadge et al., 2015;Rathinam et al., 2019) were reported in resistant wild Cajanus species. Increased expression of TPIs, AIs, H 2 O 2 , and polyphenol oxidase (Meitei et al., 2018;Rutwik et al., 2020) was observed in resistant genotypes, causing metamorphosis. Studies on C. scarabaeoides revealed protease inhibitors that may act on insect gut proteases, causing impaired digestion and altered amino acid absorption (Abigail et al., 2020;Swathi et al., 2015). Also, protease inhibitors were reported to act on mid gut proteases of Menduca sexta larvae in pigeonpea and black gram (Prasad & Padmasree, 2010), resulting in antifeedant activity, ultimately larval mortality. Despite the intensive research and identification of pod borer resistance traits in wild relatives of Cajanus species, the introgression of resistance genes from wild relatives into the cultivars has been limited (Sharma & Upadhyaya, 2016).
The role of secondary metabolites as defence molecules in plant-insect interactions is widely accepted. Among them, phenolic compounds such as flavonoids, lignin, and tannins (Sharma & Manohar, 2009) are vital in protecting the plant against herbivory (Rizwana & Ashok, 2007). An inheritance study on C. acutifolius showed a higher concentration of flavonols quercetin and rutin, leading to herbivory mortality  in resistant genotypes.
A similar study with groundnut noted the elevated levels of flavonoids during infestation by tobacco armyworm Spodoptera litura (Mallikarjuna et al., 2004). Endogenous flavonols in apple leaves were reported as an essential defence mechanism against fungus Venturia inaequalis (Mayr & Treutter, 1998;Picinelli & Mangas, 1995) while investigating scab symptoms. Condensed tannins often inhibit the digestion in larvae and denature proteins (Sharma & Manohar, 2009).
With an unpleasant bitter taste, they also act as repellents to larvae (Pagare et al., 2015). Therefore, tannins contribute to HPR (Kamila, 2016 Australia is one of the centres of diversity of pigeonpea, with 15 out of 32 wild Cajanus species with desirable agronomic traits (Khoury et al., 2015). Although traits such as heat and drought tolerance have been reported, there has been limited information on the mechanisms underpinning the resistance to major pest Helicoverpa armigera in the Australian wild pigeonpea accessions. This article reports on the screening of native Australian wild pigeonpea species for resistance to H. armigera and identifies major biochemical mechanisms underpinning the resistance. The study focuses on phenolic compounds in addition to total phenolic content (TPC).

| Plant material
The pigeonpea accessions included in this study were obtained from the Australian Grains Gene Bank, Horsham, Victoria. Two experiments were conducted using different sets of wild and domesticated accessions to explore the resistance components. Exp-1 conducted in 2018 used C. acutifolius and its interspecific hybrids (ISH) to screen for H. armigera resistance. The results of Exp-1 were verified with the second set with a broader range of wild and domesticated accessions in Exp-2 conducted in 2019.

| Leaf bioassay
The original culture of wild H. armigera was purchased from a commercial organization (AgBitech, Toowoomba, Queensland, Australia).
The insects were reared on a soya and wheat flour diet at room temperature (25-27 C), with humidity of 65 ± 5%, and 12 h photoperiod until the larvae reached the third instar.
The detached leaf bioassay protocol consisted of insect rearing jars with firmly holding ventilated lids ( Figure 1). A bed of wet sand (2-3 cm depth) was laid out, and two to three young trifoliate leaves with petioles were inserted into the damp sand vertically in the jars.
The third instar larvae's initial weight (WD1) was measured and placed on the leaf at one larva/jar ( Figure 1b). Each genotype was tested in three independent jars (replicates). Jars were randomised were kept closed using ventilated lids. Herbivory growth was measured every alternate day by weight, and the fresh leaf material was provided as needed. The jars were re-randomised once 2 days to minimise any local effects. This procedure was continued for 7 days. The herbivory weight difference (HWD) between the initial (WD1) and final weight (WD7) (Figure 1e) served as an indicator for antibiosis ( Figure 1f), which was used to rank the genotypes for their level of resistance ( Figure 1g). The leaf treatments (genotypes) on which the larvae could not survive when feeding on the leaves were identified as resistant. In contrast, the treatments on which the herbivory could survive by consuming the leaf but failed to pupate were considered medium resistant. Treatments in which herbivory survival was 100%, and the herbivory growth rates were high were reported as susceptible hosts (Brooks, 2008;Vawdrey & De Faveri, 2005).

| Sample preparation for phytochemical analysis
Flowers were collected from genotypes in Exp-1 and fully opened third and fourth leaves from the secondary branches of genotypes in Exp-2 for phytochemical analysis. All samples were freeze dried for 48 h and then pulverized in a retsch_MM400 ball mill (Retsch-Allee 1-5, 42,781 Haan, Germany) for 30 seconds at 25 oscillations speed.

| Extraction of phenolic compounds
Polyphenols were extracted as per the already reported method

| Total phenolic content (TPC)
TPC was estimated by employing a Folin_Ciocalteu assay as reported previously (Singleton & Rossi, 1965), using a microplate absorbance reader (Sunrise, Tecan, Maennedorf, Switzerland). The absorbance was measured at 700 nm. The TPC concentration was expressed as milligrams of gallic acid equivalents per gram of sample dry weight (GAE mg/g DW), based on a standard curve constructed from a serial dilution of gallic acid (from 0 to 105 mg/L).

| Statistical analysis
The data from leaf bioassays and the TPC measurements were subjected to Analysis of Variance (ANOVA), and the least significant difference at 95% confidence level using R software (R Core team 2017).
The correlation graphs between insect survival and the TPC for each genotype were produced using Microsoft excel (Praveen et al., 2013).
The reported data is the mean of the three replications in bioassay experiments and the nine replications in TPC analysis.

| RESULTS
Exp-1 demonstrated a significant variation (P < 0.001) in the HPR to H. armigera indicated by low HWD ranging between 1 and 94 mg in the native C. acutifolius and its interspecific hybrids (BC1W). UPLC analysis of these genotypes indicated the absence of already reported flavonoids such as chlorogenic acid, quercetin and rutin in genotypes expressing HPR ( Exp-2 had demonstrated significant (P > 0.005) variation in HPR between wild and domesticated accessions as well as among wild accessions. Based on the HWD, the accessions were separated into three groups. Six accessions were classified as wild resistant (WR), with a low rate of HWD (À27-124 mg) (Figure 2). The remaining six wild accessions were categorised as wild susceptible (WS), with HWD ranging between 132 and 207 mg. All seven domesticated accessions were (153-300 mg) categorized under domestic susceptible (DS). Among the WR accessions, C. pubescens (AGG309206WCAJ1) was found to be significantly more resistant, with reduced HWD (À27 mg) (Table 3), suggesting a higher level of resistance than C. acutifolious (AGG316914WCAJ1), which was proven to be more resistant in Exp-1.
The larvae could not survive for 36 h on C. pubescens (AGG309206WCAJ1) due to very little leaf consumption in all three replications. Even though two accessions, C. reticulatus (AGG300162WCAJ1) and C. latisepalus (AGG309208WCAJ1), demonstrated medium levels of antibiosis with an increased mean HWD up to104 mg and 124 mg, respectively (   Abbreviations: HWD = herbivory weight difference TPC = total phenolic content. *P value is significant. **P value is highly significant.

| Concentrations of flavonoids
In this study, UPLC-DAD analysis of leaf extracts could not detect apigenin, luteolin, vitexin, and iso-vitexin, the common flavonoids in domesticated pigeonpea. Instead, the spectra indicated the presence of phenolic compounds that could be near isomers. Three compounds eluted at different retention times (RT: 6.8, 7.3, and 9.4) in both wild and domesticated species ( Figure S1) had the same spectra. On the other hand, the phenolic compounds eluting at specific RT in different genotypes had different spectra, which adds complexity in understanding the biochemical basis of resistance. Quercetin and rutin were detected in all the samples except for all wild and one BC1W genotype. The concentration of rutin in the F1's was 0.58-1.43 mg/g, followed by 0.21-1.26 mg/g in domesticated and 0.67-2.14 mg/g in BC1D (Table 2). However, BC1W contained 0.00-1.09 mg/g rutin.

| TPC variation in genotypes
High variation in the TPC was observed between the genotypes of different progeny levels ( GAE mg/g DW) in C. acutifolius. A low TPC level was noted in the domesticated and BC1D genotypes (6.7-12.7 GAE mg/g DW).
The TPC estimated in the 19 accessions of Exp-2 categorised the samples into three groups (Figure 2). Eight wild accessions showed high concentrations of TPC ranging between 32.3 and 42.5 GAE mg/g DW (Table 3), which were grouped under W_htpc (wild high TPC).
The positive association of TPC with HPR enabled us to set the threshold levels of TPC (32.3-42.5 GAE mg/g DW) for resistance.
However, two wild accessions, one of each from C. reticulatus

F I G U R E 3
The role of total phenolic content (TPC) in the flower and leaf material of genotypic groups used concerning insect survival and their resistance (regression line was drawn using samples analysed in 2018 only, R 2 = 0.6937). The points in the circle were not considered for regression, as these are sampled from 2019. However, they still had high TPC and low larval growth The compound eluted at RT 9.2 in UPLC analysis had the same molecular mass and the fragmented ions, as shown in Figure S2. It was further confirmed by comparison with isoorientin sourced from Sigma Aldrich. Higher concentrations of isoorientin were recorded in all the wild species compared with the domesticated species (Figure 4). However, the role of isoorientin in HPR is yet to be established.

| Herbivory on a diet enriched with leaf extracts
Herbivory on a diet supplemented with 50 μl of leaf extracts had HWD ranged between 203 and 492 mg in all five treatments

| DISCUSSION
Exploiting wild species to improve genetic diversity and stress mitigation has emerged as a potential approach to crop protection (Mammadov et al., 2018). Despite identifying potential markers for  (Sharma & Manohar, 2009;Swathi et al., 2016), linkage drag has been the major limitation in transferring the resistance through conventional breeding (Venkata et al., 2019). Due to this limitation, attempts were made to develop transgenic pigeonpea with herbivory resistance through transformation using several different approaches, such as pricked embryo axes (cry1Ac) (Ajinder et al., 2016), in planta transformation (cry2Aa and cry1AcF) (Ramkumar et al., 2020) and embryo rescue in pigeonpea (Shivali et al., 2020). Being the second-largest source of Cajanus species, Australian wild species have been less exploited for their biotic and abiotic stress-resistance traits. The present investigation of the biochemical basis of resistance has proved the trait potential of Australian native Cajanus species. However, the expression of resistance to pod borer varied significantly in the WR group. C. pubescence was noted to be highly resistant as larval mortality was observed within 36 h. The remaining five accessions resulted in low HWD before the larval mortality. Lower HWD, increased mortality and prolonged larval development confirms the high (Kumari et al., 2006;Sujana et al., 2008) level of antibiosis existing in native species. A similar kind of resistance was reported in wild species with higher levels of antibiosis and reduced HWD compared with domesticated genotypes in tomatoes (Asif et al., 2019) and chickpea (Sivakumar et al., 2020). On the other hand, wild Cajanus species were reported to have traits resistant to Fusarium wilt (Mamta et al., 2012;Saxena et al., 2020) and sterility mosaic disease along with other agronomically essential traits like high seed weight and high protein content (Saxena & Rao, 2002;Upadhyaya et al., 2011). Pod borer resistance in BC1W genotypes showed the cross-compatibility and functional integrity of the trait. Hence the members of the WR group would be potential donors for pod borer resistance in pigeonpea.
The UPLC-DAD analysis of leaf extracts and detached leaf bioassay has revealed a significant negative relationship (R 2 = 0.69) ( Figure 3) between TPC and insect survival. The effect of TPC interfering with the metamorphosis of the insects (Tunaz & Uygun, 2004) was already reported. Phenolic compounds have been targeted as biochemical markers for biotic and abiotic stress resistance since 1986 (Rathi et al., 1986). For instance, the antioxidant activity of Nypa fruticans, a palm species, was positively correlated with the TPC concentration in leaves (Hermanto et al., 2020). Increased synthesis of phenylalanine in white cabbage infested by cabbage butterflies and flea beetles was shown to be an induced defence mechanism (Kovalikova et al., 2019). Anket et al. (2016) reported a high TPC concentration, increased expression of chalcone synthase (CHS), and phenylalanine ammonia-lyase (PAL) in stress induced mustard. However, in our study, C. lanuginosus (AGG316926WCAJ1) and C. reticulatus (AGG300161WCAJ1) were highly susceptible herbivory despite having high TPC. Several other factors, such as high soluble carbohydrates in leaves, might have favoured the herbivory feeding ability (Asif et al., 2019;Sharma & Manohar, 2009), leading to susceptibility.
Moreover, though TPC plays a significant role in plant defence, it has a minimal influence on individual compounds concentration (Pagare et al., 2015). The flower and leaf analysis of AGG316926WCAJ1 and ICPL14425 suggests the concentration of TPC in the leaf was higher than in flowers of the same genotype.
Similar results were also reported in Elaeagnus angustifolia (Saboonchian & Hosseini, 2014), commonly known as Russian olive had higher phenolic and flavonoid concentration in leaves than in flowers. The higher TPC in resistant accession than in the susceptible accession, suggesting that the TPC level is specific to the genotype.
Thus the TPC extracted from any plant tissue of the resistant genotype would be higher than the TPC extracted from the same plant tissue of the susceptible genotype. TPC extracted from different genotypes at the same time should be reliable ( Figure 2). Therefore, TPC could be used as a biochemical marker to categorise genotypes resistance to H. armigera. TPC and the related enzymes were reported as markers for biotic and abiotic stress in sorghum (Mamoudou et al., 2005). Although our studies indicate the high and low TPC concentration is specific to genotype, Khang and Liu have noted the individual flavonoids could vary depending on tissue and growth stage (Khang et al., 2016;. The synthesis of phenols in the chloroplast is a photosensitive phenomenon. Long days in autumn favour increased accumulation of phenolic compounds (Palavan-Unsal, 2011). However, the effect of the environment on TPC accumulation in wild Cajanus species is yet to be investigated.
The resistance phenomena observed in the WR group could be a synergistic association of two or more phenolic compounds. Higher expression of quercetin and rutin in pigeonpea , caffeic acid, dihydroxybenzoic acid and vanillic acid in groundnut (Rashid et al., 2016) were reported in H. armigera resistance genotypes. Similarly, phenolic compounds extracted from root hairs of tomato tested for H. armigera, survival on an artificial diet enriched with leaf extracts reported 53% of larval mortality. This study also showed a high level of antibiosis attributed to the presence of rutin, quercetin, kaempferol, gallic acid and caffeic acid (Harpal et al., 2014).
However, the reported resistance was the effect of individual flavonoid or the synergistic effect of TPC with one or more compounds is still unclear. The increased resistance, higher TPC, and isoorientin levels in Australian wild species could be due to the existing genetic difference between Asian and Australian Cajanus species (Kassa et al., 2012). However, the basis for the difference in phytochemistry is yet to be identified.
Isoorientin was reported as the most effective compound with great aphidicidal activity on mustard aphid affecting cruciferous vegetables (Gao et al., 2019). The concentration of isoorientin contributing to higher TPC and wild species resistance (Figure 4) could involve a dosage-dependent or a synergistic mechanism. For example, the presence of isoorientin at >0.2% (dry weight) was identified as the concentration required for resistance against Helicoverpa zea (Widstrom & Snook, 1998), in conjunction with maysin, in equal concentrations.
The hybrids with both compounds were identified as more resistant than those only with isoorientin or maysin (Widstrom & Snook, 1998).
Our analysis of herbivory on a diet enriched with leaf extracts (Figure 5) also demonstrated a similar type of mechanism. The herbivory feeding experiments on the WR accession (AGG309208WCAJI) showed a high antibiosis with less HWD than any other treatments.
Simultaneously, the feeding on DS spiked with WR caused continuously increased levels of antibiosis as the concentration of WR leaf extracts increased. In research on tomato, the trichomes extracts were found to contribute significantly to the antibiotic effect of the leaf against Heliothis zea. The result was attributed to rutin in synergy with other phenolic compounds (Duffey & Isman, 1981). Our study showed that isoorientin alone might not offer resistance but could act in synergy with other compounds such as maysin in maize (Mamoudou et al., 2005) to contribute to H. armigera resistance in the WR group.

| CONCLUSION
Our results demonstrated that secondary gene pool of pigeonpea collected in Australia C. acutifolius (AGG316925WCAJ1), C. latisepalus Australian wild species could help to identify the specific compounds acting synergistically. The expression of resistance could also be associated with environmental changes, which requires further investigation.

ETHICS STATEMENT
We declare that the work presented in this paper is the original research work conducted at UQ. It has neither manipulated nor submitted to any other journal before. This paper's authors are not engaged in any personal or financial relationships that can impact this publication.

AUTHOR CONTRIBUTIONS
Prameela Rani Vanambathina carried out the actual laboratory and bioassay experimental work and drafted the manuscript. Rao Rachaputi did the statistical assay assistance and fine changes in manuscript. Yasmina Sultanbawa provided guidance and supervision in phytochemical analysis. Anh Dao Thi Phan assisted in UPLC and LCMS/MS handling and analysis. Robert J. Henry did the manuscript proofreading, grammar, and spell-check. Hugh Brier provided guidance in leaf bioassay and larvae availability.

CONFLICT OF INTERESTS
The authors declare no conflict of interests.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.