We assessed mortality induced by 11 EPF candidates from Australia, New Zealand, and Malaysia that were in CSIRO’s fungal collection against the recently established lepidopteran pest S. frugiperda. Two Australian isolates of Beauveria sp. (B-0571 and B-1311) induced the highest mortality in 3rd instar S. frugiperda caterpillars, with mortality induced predominantly in the first day after infection. These two isolates also induced high mortality in S. frugiperda moths and 6th instar caterpillars, but not in pupae. Additionally, we examined mortality induced by B-0571 and B-1311 in an Australian native lepidopteran pest, H. armigera, for which both strains were less effective than in S. frugiperda.
4.1 Fungal Isolate Species identification
The candidate fungal species were identified via sequencing of the ribosomal 16S ITS region. Sequence results indicated that of the 11 fungal candidates, five Beauveria isolates matched the ITS of six species of Beauveria with high nucleotide identity (100-97%, B-0016, B-0077, B-0079, B-0571, B-0698, and B-1311), and five Metarhizium isolates matched the ITS regions of various Metarhizium species (M-0121, M-0122, M-0123, M-0999 and M-1000). The general consensus divergence for cut-off for delimitating across Fungi kingdom is approximately 3-5% for ITS sequences (intraspecific ITS variability in fungi is 2.51±4.57%) with ≥80% query coverage and based on 500-800bp (Schoch et al. 2012; Raja et al. 2017). In the phylum Ascomycota (which includes Beauveria and Metarhizium), the average variability in weight of infraspecific ITS has been reported to be 1.96±3.73% (Nilsson et al. 2008; Raja et al. 2017). Hence, although the genera of CSIRO fungal isolates were inferred by ITS sequences analysis (six isolates of Beauveria sp. and five isolates of Metarhizium sp.), the species of all candidates remain unknown. This is because the ITS regions of those isolates have ≥97% identical sites to more than one species from the reference sequences.
Based on the ITS sequence phylogeny, six of the CSIRO Beauveria fungal isolates clustered with high confidence (Fig. 2; node support value 88% (in red)) with B. bassiana and B. brongniarti, suggesting that resolving the taxonomic status of both B. bassiana and B. brongniarti as well as the CSIRO Beauveria isolates could not rely solely on the ITS region. Some isolates of B. bassiana and B. brongniarti as well as various CSIRO Beauveria isolates also have identical ITS sequences (e.g., OM131742, LN886699, B-0077, B-0571, B-1311), while many Beauveria species shared highly similar (>97%) ITS nucleotide identity. Similarly, the five CSIRO Metarhizium isolates were clustered in the same clade, albeit less confidently, with many species of Metarhizium (Fig. 3). This is due to high nucleotide similarity in Metarhizium ITS sequences both within the M. anisopliae complex (M. anisopliae, M. guizhouense, M. pingshaense, M. robertsii, and M. acridum) and other Metarhizium spp. (Metacordyceps indigotica M. gryllidicola, and Metacordyceps brittlebankisoides). There is currently insufficient data for molecular diagnostics of the various Beauveria and Metarhizium isolates based on the ITS regions, and multilocus sequence typing (MLST)/whole genome sequencing approach may offer greater resolution power to aid in the molecular taxonomy and assist with molecular diagnostics of these important fungal groups.
Phenotypic characters such as macro-and micro-morphology, production of certain chemicals, and host specificity are also used to help identify fungal species. These phenotypic characters, however, can be affected by external factors (e.g., growth conditions) and the strain of fungus (Francisco et al. 2006; Fernandes et al. 2010; Sepúlveda et al. 2016; Bridge et al.). For example, although B. brongniarti has been reported to mainly attack cockchafer (i.e., Melolontha spp.) and other coleopteran insects (Vestergaard et al. 2003; Shah and Pell 2003; Imoulan et al. 2017; Rohrlich et al. 2018), Wu and colleagues (Wu et al. 2019) demonstrate that B. brongniartii SB010 also infects lepidopteran species (Spodoptera litura). Molecular diagnostics and the molecular taxonomic status of CSIRO Beauveria and Metarhizium isolates used in this study will require further investigation, either through the whole genome sequencing approach or via a MLST approach involving other widely used DNA markers (i.e., TEF1α, RPB1, RPB2, β-tubulin, CaM, Bloc) (Bischoff and Rehner 2009; Hernández and Guzman-Franco 2017; Dizkirici and Kalmer 2019; Hoang et al. 2019; Bustamante et al. 2019; Castro-Vásquez et al. 2021).
4.2 Mortality of Spodoptera frugiperda 3rd instar caterpillars
The overall and daily mortality of 3rd instar caterpillars was investigated to determine mortality induced by each fungal candidate. Two isolates of Beauveria sp., B-0571 and B-1311, induced the highest mortality. The higher mortality levelsinduced by B-0571 and B-1311 could be because of the strength of spore adhesion, spore gemination rates, the production of enzymes and/or secondary metabolites, and growth conditions that support their development (see below).
The strength of spore adhesion is one of the crucial factors that indicates the virulence of EPF. Spore adhesion is the very first stage of fungal infection, and mostly relies on kinetic mechanisms (e.g., hydrophobic and/or electrostatic) (Boucias and Pendland 1991). Some strains of EPF may employ carbohydrate substances or specific receptors/ligands to strengthen the adhesion force (Boucias et al. 1988). A weakened adhesion strength could result in the spores being washed off from the surfaces of insects (Holder and Keyhani 2005) thus preventing them from infecting the host species (Wang and St Leger 2007). Herein, a diverse isolate of fungi may manifest different level of hydrophobicity and biochemistry. This may result in a fluctuation in virulence. The two most pathogenic isolates of Beauveria sp. (B-0571 and B-1311) may exhibit stronger spore adhesion strength than other tested fungal candidates.
Conidial germination rates are also frequently associated with the virulence of EPF (Heale et al. 1989; Yeo et al. 2003; Safavi et al. 2007; Tseng et al. 2014). After attaching to the insect cuticle, to infect the target insect the spores must germinate and form an appressorium to penetrate through the insect’s chitinous exoskeleton. A faster germination rate is speculated to not only quicken the killing process but also reduce the possibility of losing spores from the insect moving and moulting (Altre et al. 1999). Faria et al. (2015) found that fast-germination strains of B. bassiana exhibit higher virulence towards S. frugiperda than do slow-germination strains. In the present study, the two fungal isolates that induced the highest overall mortality (B-0571 and B-1311) also have the highest Day one mortality, which suggests that these fungal strains may have higher germination rates. At this time, germination rates of the candidate fungi have not been measured and so the contribution of germination rate to the observed mortality patterns remains unknown.
The production of hydrolytic enzymes has also been proposed as one determinant of EPF virulence. The arthropod exoskeleton is composed of various compounds including chitin, lipid, protein, and phenolic compounds that act as a barrier to protect the insect from desiccation and entomopathogens (Petrisor and Stoian 2017). EPF secrete a wide range of cuticle-degradation enzymes in order to infect the insect (e.g., chitinase, protease, and lipase) (Cheong et al. 2016). Consequently, fungal candidates that possess higher virulence may have higher enzyme activities. Further investigation is also required to verify this hypothesis for our high-performing Beauvaria candidates B-0571 and B-1311.
Growth conditions of the fungal candidates could play significant roles in fungal virulence. Culture media and growth temperature have been shown to greatly influence the strength of spore adhesion (Ibrahim et al. 2002; Rangel et al. 2008), production of cuticle-degrading enzyme protease (Butt et al. 1996; Safavi et al. 2007; Rangel et al. 2008), conidial gemination rates (Yeo et al. 2003; Safavi et al. 2007; Rangel et al. 2008), and production of secondary metabolites (Asai et al. 2012; VanderMolen et al. 2013). Therefore, some of those isolates which exhibit a low killing ability may simply require different culture media or/and growth conditions to improve virulence.
Some EPF have the ability to secrete secondary metabolites that could contribute to the outcome of fungal-insect interaction. This is because secondary metabolites could assist EPF to overcome insects’ immune systems and quicken mycosis (Zimmermann 2007). There is evidence that some strains of B. bassiana are able to produce host-specific secondary metabolites that can cause 50% mortality at very low concentrations (3.3µg/g body wt; Bassiacridin, infected Locusta migratoria) (Quesada-Moraga and Vey 2004). Regarding the high mortality induced by B-0571 and B-1311, these strains may produce bioactive compounds that have insecticidal activities toward S. frugiperda. Understanding and being able to produce these compounds could open new avenues for controlling invasive pest species, however further investigation is required to confirm these hypotheses.
4.3 Mortality of Spodoptera frugiperda 6th instar caterpillars, pupae, and moths
Isolates B-0571 and B-1311 which induced the highest mortality in S. frugiperda 3rd instar caterpillars also induced high morality in moths and 6th instar caterpillars, but not in pupae. Differences in the susceptibility of S. frugiperda at different developmental and life stages may be linked to the strength of the insect’s immune system, thickness, profile, and availability of susceptible locations on the exoskeleton, and pathogen defences (see below).
A more developed immune system could enhance the resistance of S. frugiperda toward fungal infection. In the present study, 6th instar caterpillars generally were less susceptible to pathogens than were 3rd instar caterpillars. Similarly, 5th instar S. frugiperda caterpillars are less susceptible to nematode (Steinernema feltiae) than 1st and 3rd instar caterpillars (Fuxa et al. 1988). The increasing of resistance in older caterpillars is believed to be strongly associated with the development of their immune system. Haemocytes tend to be much more abundant in older caterpillars than in younger caterpillars (Carper et al. 2019). This suggests that caterpillars increase their immunometabolism as they progress through their developmental stages. The stronger immune system would imbue older S. frugiperda with greater resistance against EPF (and/or insecticides) leading to reduced susceptibility to infection (or reduced mortality from exposure to insecticides).
The physical aspects of S. frugiperda at each life stage could also contribute to the outcome of fungal infection. Regarding the mode of action of Beauveria species, the spores infect the target host by penetrating through its external exoskeleton. Thus, the insect’s cuticular thickness and profile could be important factors in preventing penetration. According to Wrońska et al. (2018), the cuticles of caterpillars and the thorax of moths are the most susceptible areas to be digested by proteases and lipase (commonly produced by many EPF) (Erlacher et al. 2006; Hussein et al. 2012; Dhawan and Joshi 2017). This makes caterpillars and moths highly vulnerable to fungal infection. Additionally, the spores of EPF generally accumulate around the spiracles, hairs, pores of the wax glands, eyes, antennal segment, and legs articulating membranes and germinate from those areas (Toledo et al. 2010). Hence, thinner/weaker exoskeletons and/or higher availability of susceptible areas potentially provides the fungi with greater chances to successfully infect host species.
Spodoptera frugiperda commonly pupate in the soil and are therefore likely to encounter entomopathogenic microbes more than are caterpillars and moths. Thus, pupae may have evolved defence mechanisms, such as melanisation of the exoskeleton, to defend against soil-borne entomopathogens. Melanisation is a process that the insect employs to enhance cuticle pigmentation, cuticular sclerotization, wound healing, and its innate immune system (Sugumaran and Barek 2016). The darkness of the insect cuticle (higher concentrated melanin indicator) could, in some cases, indicate immune status (Nappi et al. 1995; Reeson et al. 1998; Barnes and Siva-Jothy 2000). Melanisation is likely to be a key mechanism that protects pupae from fungal infection. In addition, pupal cells of some insect species (e.g., Curculio caryae) have been shown to possess the antimicrobial properties which can suppress the growth and germination of B. bassiana (Shapiro-Ilan and Mizell 2015). Whether S. frugiperda also evolved a similar mechanism to protect themselves from B-0571 and B-1311 at the pupal stage is not known.
4.4. Mortality of 3rd instar Spodoptera frugiperda vs Helicoverpa armigera
The overall mortality of 3rd instar S. frugiperda caterpillars infected with isolates B-0571 or B-1311 are higher than that of H. armigera. This suggests that either H. armigera has a higher resistance to fungal infection or it is not a target species of B-0571 and B-1311. A similar scenario was reported by Gutierrez et al. (2015) who showed M. anisopliae to be highly pathogenic towards oriental cockroach (Blatta orientalis), but not to other cockroach species (e.g., German cockroach Blattella germanica). The authors reported at least 19 fatty acids on the outer layer surface of B. orientalis but not on B. germanica. The fatty acids were speculated to be a determinant of fungus virulence; however, the mechanism of how they facilitated M. anisopliae was not reported. Therefore, the lower susceptibility of H. armigera could be due to it lacking specific signals (e.g., fatty acids) on the cuticle such that the germination process is not triggered. Other factors that could contribute to the lower susceptibility of H. armigera may include the activation of antifungal mechanisms to increase protection against fungal infection, or production of secondary metabolites by B-0571 and B-1311 that better suited species within the Spodoptera genus, including S. frugiperda. This hypothesis could be tested by including endemic Spodoptera species e.g., S. litura and Australian S. exigua (a potentially cryptic species from African S. exigua; see (Agarwal et al. 2022)).