Using host molecules to increase fungal virulence for biological control of insects

Unlike the situation for many fungal pathogens, i.e., those of most plants and animals, where efforts at understanding the pathogenic process are aimed toward controlling the disease-causing agent, research on entomopathogenic fungi seeks to develop new strategies for improving the ability of the fungus to target its insect hosts. Previous work has shown that fungal virulence can be increased via overexpression of insect cuticle-degrading enzymes, e.g., proteases and chitinases, and by expression of insect toxins derived from various organisms. A new approach in which insect derived molecules are used against the insects themselves within the context of the microbial pathogenic process has recently been described. Depending upon the host and molecule chosen it is theorized that target specific enhancement of virulence is achievable, a critical step in developing safer and more effective biopesticides. Exploiting host molecules also places a higher burden on potential resistance development as compared with use of chemical pesticides and/or insect toxins.

As an alternative to chemical pesticides, entomopathogenic fungi have long been considered as potential candidates for insect pest control. The history of this research and any of its applications appears to have undergone periodic cycles of alternating intense and waning interest. In part, these cycles have been as a result of the development and introduction of cheaper and more effective chemical pesticides and the often less than satisfying performance of the fungal microbial agents in the field. The current status of the field appears to be one of cautious renewal of interest. The economic and industrial aspects of these efforts will not be considered here, but have been recently reviewed (Glare et al., Trends Biotechnol 2012). The major impetus driving this interest is the gradual banning of the use of chemical pesticides coupled to the lack of few new such chemicals in the near-term use and/or developmental stage. The two most widely studied entomopathogenic fungi are Metarhizium anisopliae and Beauveria bassiana, both Environmental Protection Agency (EPA, US, as well as by the respective EU regulatory agency)-approved insect biological control agents. Both organisms are currently used in various field applications and are available from a number of commercial sources worldwide. However, while there are several notable instances of their successful use, a number of factors continue to limit practical field applications of mycoinsecticides. These include relatively low resistance to abiotic factors that can affect field use, i.e., low tolerance to high (> 37°C) and/or fluctuating temperatures, reduced efficacy at low humidity and low resistance to sunlight-derived UV irradiation. In addition, issues concerning the length of time to kill target insects remain since it can take from 3 to 15 d for the fungus to kill its insect targets. Finally, the need for high spore concentrations for effective suppression of target insect populations and (low) persistence issues place significant constraints on the ability to use spores and/or other fungal infectious propagules to cover large topographical areas, e.g., agricultural fields, in a cost-effective manner.

Recombinant DNA techniques have afforded an opportunity for strain improvement currently impossible. Attempts at targeted strain isolation and/or selection, parasexual crossing or protoplast fusion have resulted in only incremental increases in virulence and/or a desired trait. Genetic engineering via addition of genes whether for pathogenicity determinants, stress resistance or some other factor that could increase the applicability of the fungal agents represents the most versatile approach at strain manipulation. The first such application targeted the (over) expression of an endogenous fungal protease that had been implicated in the degradation of insect cuticle substrates, i.e., facilitated the penetration of the host integument (St. Leger et al., Proc Natl Acad Sci U S A 1996). In this early work, although the median lethal concentration (LD₅₀) was unaffected, increased expression of the protease (Pr1) decreased the survival time (LT₅₀) by ~25%. Since then a variety of proteases, chitinases, the latter facilitating the degradation of the chitin polymer that constitutes the major carbohydrate constituent of the insect cuticle, and protease-chitinase hybrid proteins have been expressed in either M. anisopliae and/or B. bassiana, each showing some improvement in targeting of insect hosts.

A second, major development in the field occurred with the expression of an insect-specific scorpion neurotoxin in M. anisopliae which increased fungal virulence (i.e., decreased the LD₅₀) a dramatic 22-fold against the tobacco hornworm (Manduca sexta) and 9-fold against adult yellow fever mosquitoes (Aedes aegypti) (Wang and St. Leger, Nat Biotechnol 2007). This strain also resulted in a ~25–30% decrease in LT₅₀, thus killing insects faster and with fewer spores. Since then a number of other insect toxins have been expressed in either M. anisopliae or B. bassiana with similar results. This approach, however, has not been without controversy. Concerns regarding the use of strains expressing neurotoxins (even insect-specific) as well as the nature of the relatively non-specific enhancement of virulence that could potentially affect non-target or beneficial insects have been raised. More recently, the development of transgenic M. anisopliae that target the Plasmodium malarial parasite has been reported (Fang et al., Science 2011). In this approach, rather than increasing the virulence of the biological control agent, the fungus was engineered to express a single-chain antibody that agglutinates the sporozoite, thus reducing the parasite load in the mosquito and decreasing transmission of the disease causing agent. This development may be transformative in that it is likely to be the best current candidate for approval in field trials.

Given this context, we sought to expand the toolbox available for increasing the virulence of these fungi within a different conceptual framework. All insects (as well as most other organisms) use proteins and peptides to regulate critical physiological processes. These molecules are often regulated by strict developmental and tissue-specific mechanisms. We reasoned that disruption of the expression pattern of specific proteins/peptides via exogenous addition of the molecule could result in physiological consequences that would make the insect more susceptible to attack by microbial pathogens. Given the broad spectrum of molecules that would meet this description, several criteria were used to guide the selection of candidate molecules. First, evidence (in the literature) that the candidate molecule was essential for proper development or critical in key physiological and/or immune related processes. Second, evidence that the protein/peptide displayed some form of toxicity when administered to the insect. A third criteria, that significantly narrowed the field of potential candidates was evidence of field tests and/or agency approval of the molecule for actual use against insects. Finally, as targets for these efforts, we chose mosquitoes, which represent globally important vectors of disease causing agents, and fire ants, an invasive pest species that results in significant agricultural, ecological and hence economic losses, which also have important human health-related impacts.

One of the most promising candidates identified was the trypsin-modulating oostatic factor (TMOF) derived from the mosquito, Aedes aegypti (Fig. 1). TMOFs are hexa- and deca-peptides that terminate trypsin biosynthesis in the insect gut and are required for normal development and ovipositing. TMOF circulates in the hemolymph, binding to receptors on the hemolymph side of the gut and inhibits trypsin biosynthesis via translational control of trypsin mRNA. TMOF resists proteolysis in the gut and can traverse the gut epithelium to enter into the hemolymph. Due to this property, it was shown that exogenous feeding of Aea-TMOF to mosquitoes blocks their ability to digest food, bloodmeal or otherwise, subsequently resulting in starvation and death in both adults and larvae. A TMOF ortholog was also identified from the gray flesh fly, Sarcophaga bullata. The two TMOF sequences are highly divergent with the A. aegypti sequence corresponding to 10 amino acids: YDPAPPPPPP, and the S. bullata TMOF to 6 amino acids: NPTNLH. TMOFs therefore appear to be good candidates regarding specificity with minimal non-target effects, a key factor with respect to the desire to develop target specific agents. Commercial production and use of Aea-TMOF is currently under development, however, the most significant obstacle to practical application of the peptide is the lack of a mechanism by which the peptide can be delivered to the mosquito. Since B. bassiana is already a pathogen of mosquitoes, we hypothesized that expression of Aea-TMOF in the fungus would represent an approach at delivering the peptide to the target, while increasing the virulence of the biological control agent in a host specific manner.

Figure 1. Using host molecules to improve the efficacy of fungal biological control agents. Various host-derived peptides, including the mosquito (A. aegypti) trypsin modulating oostatic factor (TMOF), the Manduca sexta diuretic hormone (MSDH) and the fire ant (S. invicta) β-neuropeptide (b-NP) were selected as candidate molecules for expression in a fungal-insect pathogen (B. bassiana) in order to determine whether these factors can be used against the insects themselves using the fungus as a means of delivering the compound. Images courtesy of C. Zettel, L. Buss, and Castner (University of Florida, Entomology and Nematology).

As further proof-of-concept two additional insect-derived candidate molecules were selected for expression in B. bassiana: the tobacco hornworm, Manduca sexta, diuretic hormone (MSDH) and the fire ant, Solenopsis invicta, β-neuropeptide (β-NP). Insect diuretic hormones regulate fluid homeostasis and MSDH is a member of the corticotropin-releasing factor-related family of peptides. These peptides act upon insect Malpighian tubules, excretory organs that lie in the abdominal cavity and empty into the junction between the midgut and hindgut, affecting their secretion of water and solutes. Administration of synthetic MSDH to insects results in fluid loss through the gut and epidermis, decreased feeding, and ultimately leads to the death of the organism. β-NP is a member of the pyrokinin/pheromone biosynthesis activating neuropeptide (PBAN) family that is characterized by the presence of a C-terminal FXPRL-NH₂ sequence. β-NP is often (co-transcribed and) co-translated into a larger protein that is subsequently post-translationally cleaved to yield smaller peptides. These smaller peptides can include diapause hormone (DH), three peptides termed α, β and γ-neuropeptides and PBAN itself, the latter of which is often found between the β- and γ-neuropeptides. All of the peptides contain the C-terminal FXPRL-NH₂ motif, with the amidation reaction representing a subsequent (after cleavage) post-translational modification of the peptides. Depending upon the insect species, these peptides are thought to function in a wide range of physiological processes that includes stimulation of pheromone biosynthesis, induction and/or termination of diapause, pupariartion, melanization and potentially mediation of defense responses. In S. invicta, the α-peptide is lacking.

In all, we assessed the impact of expressing Aea-TMOF, Sb-TMOF, MSDH and Si-β-NP in the fungal insect pathogen B. bassiana (Fan et al., Nat Biotechnol 2012; Fan et al., PLoS One 2012). Each peptide was independently expressed in constructed fungal strains via transformation using vectors containing a constitutive B. bassiana derived gpd promoter and the nucleotide sequences of each peptide fused to a signal sequence to drive extracellular secretion of the molecule produced. Production of the peptides in culture media was confirmed by mass spectrometry and the concentrations ranged from 0.2 to 1.0 μM. Expression of MSDH increased the virulence (i.e., decreased the lethal dose, LD₅₀) from 5- to 10-fold toward Lepidopteran hosts including M. sexta and the greater waxmoth, Galleria mellonella, as well as toward A. aegypti. In addition, the time to death (LT₅₀) decreased by ~25–35% as compared with the wild-type parental strain. These data suggest that MSDH would act as a broad host-range targeting molecule.

Similarly, expression of Aea-TMOF in B. bassiana decreased the LD₅₀ 6- to 7-fold and decreased the LT₅₀ by ~30% against blood-fed female mosquitoes. However, Aea-TMOF expressing strains were no more effective against flesh flies (S. bullata) than the wild-type parent, and expression of Sb-TMOF did not increase virulence toward mosquitoes. Interestingly, expression of Sb-TMOF also did not increase virulence toward S. bullata, suggesting that the physiological requirement and/or role of TMOF in the flesh fly may differ significantly from that in mosquitoes. In addition to increasing virulence toward both the adult and larval stages of the mosquito, several important “side effect” impacts were noted for the B. bassiana strain expressing Aea-TMOF. As expected, gut trypsin activity was inhibited, but larval development was stunted and fecundity was dramatically decreased, with female mosquitoes laying ~55% fewer eggs than uninfected controls. Since one of the major obstacles regarding field use of fungi against mosquitoes is that the relative time frame it takes to kill the insect would still allow for a reproductive cycle, the decreased fecundity effect could have a significant impact on biological control efforts that seek to decrease vector populations and hence disease transmission.

Expression of Si-β-NP in B. bassiana increased fungal virulence approximately 6-fold toward fire ants (decreased LD₅₀) and decreased the mean survival time (LT₅₀) by ~30%. In this instance no significant differences were seen between the wild-type strain and the Si-β-NP expressing strain toward the Lepidopteran hosts Galleria mellonella and Manduca sexta. However, (unexpectedly) alternations in social behavior were noted in ants infected with the Si-β-NP expressing strain that did not occur during infections using the wild-type B. bassiana parental strain. Whereas under normal conditions as well as during infection by the wild-type B. bassiana strain, the fire ants would remove their dead from the surrounding areas of their nests, forming discrete “bone piles” or “cemeteries,” a phenomenon termed necrophoretic behavior, ants infected with Si-β-NP appeared to leave their dead conspecifics randomly distributed in the assay chambers. Necrophoretic behavior along with grooming represent two important behavioral adaptations that have limited the use of biological control agents in targeting fire ants and other social insect pests, since these actions help minimize the spread of infections throughout a colony. It remains to be seen whether the observations made in the laboratory would have any real impact in the field; however, these results do illustrate that insect pest behaviors can potentially be modulated in a target specific manner.

Although we have used fungi as the vehicle to deliver host molecules to an intended insect target, the use of insect derived proteins and peptides could potentially be expanded to other insect biological control agents such as viruses, bacteria and nematodes. There are several potential advantages to the approach outlined here that although require further investigation and verification, should be noted. First, appropriate selection of the host molecule can lead to target specific increases in virulence. Second, additional beneficial (in the sense of increasing the impact of the biological control agent) can be incorporated in the design selection, i.e., via selection of molecules that disrupt fecundity, feeding or other processes and/or behaviors. Finally, we theorize that the development of resistance to our approach is minimized as the host molecules (peptides) selected participate in critical or ideally essential processes that are species and tissue specific. Contrary to the use of chemical pesticides and/or toxins, where often a single mutation could result in resistance, mutations that might arise and could compensate for the dose given by the fungal-expressed product during infection would result in severe developmental defects, particular since the fungus remains a pathogen even in the absence of the expressed molecule. We speculate that this would represent a steep fitness cost greater than the potential development of resistance to a pesticide or toxin. Furthermore, multiple host molecules and/or combinations with other pathogenicity/targeting factors can be expressed in the same fungal strain to further improve the efficacy and safety of biologically based pesticides.

Fan Y, Borovsky D, Hawkings C, Ortiz-Urquiza A, Keyhani NO. Exploiting host molecules to augment mycoinsecticide virulence. Nat Biotechnol. 2012;30:35–7. doi: 10.1038/nbt.2080.

Fan Y, Pereira RM, Kilic E, Casella G, Keyhani NO. Pyrokinin β-neuropeptide affects necrophoretic behavior in fire ants (S. invicta), and expression of β-NP in a mycoinsecticide increases its virulence. PLoS One. 2012;7:e26924. doi: 10.1371/journal.pone.0026924.