Metarhizium brunneum Blastospore Pathogenesis in Aedes aegypti Larvae: Attack on Several Fronts Accelerates Mortality

Abstract

Aedes aegypti is the vector of a wide range of diseases (e.g. yellow fever, dengue, Chikungunya and Zika) which impact on over half the world's population. Entomopathogenic fungi such as Metarhizium anisopliae and Beauveria bassiana have been found to be highly efficacious in killing mosquito larvae but only now are the underlying mechanisms for pathogenesis being elucidated. Recently it was shown that conidia of M. anisopliae caused stress induced mortality in Ae. aegypti larvae, a different mode of pathogenicity to that normally seen in terrestrial hosts. Blastospores constitute a different form of inoculum produced by this fungus when cultured in liquid media and although blastospores are generally considered to be more virulent than conidia no evidence has been presented to explain why. In our study, using a range of biochemical, molecular and microscopy methods, the infection process of Metarhizium brunneum (formerly M. anisopliae) ARSEF 4556 blastospores was investigated. It appears that the blastospores, unlike conidia, readily adhere to and penetrate mosquito larval cuticle. The blastospores are readily ingested by the larvae but unlike the conidia are able infect the insect through the gut and rapidly invade the haemocoel. The fact that pathogenicity related genes were upregulated in blastospores exposed to larvae prior to invasion, suggests the fungus was detecting host derived cues. Similarly, immune and defence genes were upregulated in the host prior to infection suggesting mosquitoes were also able to detect pathogen-derived cues. The hydrophilic blastospores produce copious mucilage, which probably facilitates adhesion to the host but do not appear to depend on production of Pr1, a cuticle degrading subtilisin protease, for penetration since protease inhibitors did not significantly alter blastospore virulence. The fact the blastospores have multiple routes of entry (cuticle and gut) may explain why this form of the inoculum killed Ae. aegypti larvae in a relatively short time (12-24hrs), significantly quicker than when larvae were exposed to conidia. This study shows that selecting the appropriate form of inoculum is important for efficacious control of disease vectors such as Ae. aegypti.

Aedes aegypti is the vector of a wide range of diseases (e.g. yellow fever, dengue, Chikungunya and Zika) which impact on over half the world's population. Entomopathogenic fungi such as Metarhizium anisopliae and Beauveria bassiana have been found to be highly efficacious in killing mosquito larvae but only now are the underlying mechanisms for pathogenesis being elucidated. Recently it was shown that conidia of M. anisopliae caused stress induced mortality in Ae. aegypti larvae, a different mode of pathogenicity to that normally seen in terrestrial hosts. Blastospores constitute a different form of inoculum produced by this fungus when cultured in liquid media and although blastospores are generally considered to be more virulent than conidia no evidence has been presented to explain why. In our study, using a range of biochemical, molecular and microscopy methods, the infection process of Metarhizium brunneum (formerly M. anisopliae) ARSEF 4556 blastospores was investigated. It appears that the blastospores, unlike conidia, readily adhere to and penetrate mosquito larval cuticle. The blastospores are readily ingested by the larvae but unlike the conidia are able infect the insect through the gut and rapidly invade the haemocoel. The fact that pathogenicity related genes were upregulated in blastospores exposed to larvae prior to invasion, suggests the fungus was detecting host derived cues. Similarly, immune and defence genes were upregulated in the host prior to infection suggesting mosquitoes were also able to detect pathogen-derived cues. The hydrophilic blastospores produce copious mucilage, which probably facilitates adhesion to the host but do not appear to depend on production of Pr1, a cuticle degrading subtilisin protease, for penetration since protease inhibitors did not significantly alter blastospore virulence. The fact the blastospores have multiple routes of entry (cuticle and gut) may explain why this form of the inoculum killed Ae. aegypti larvae in a relatively short time (12-24hrs), significantly quicker than when larvae were exposed to conidia. This study shows that selecting the appropriate form of inoculum is important for efficacious control of disease vectors such as Ae. aegypti.

Author Summary

Mosquitoes transmit a range of diseases which have a profound impact on human health. Aedes aegypti vectors dengue, one of the fastest emerging diseases and, more recently, the Zika virus, which has been linked to thousands of birth defects over the last two years in Brazil. Insect pathogenic fungi such as Metarhizium brunneum are effective in killing mosquito adults and larvae. They exhibit much plasticity, producing aerial conidia on solid substrates and blastospores in liquid media. We not only show that blastospores are more virulent than conidia but present evidence explaining why they are more aggressive. The blastospore mode of pathogenesis differs from that of conidia in several ways. Firstly, blastospores appear to be more dependent on entry using mechanical force than by secretion of cuticle degrading proteases such as Pr1. Blastospores produce copious mucilage which ensures that many spores attach to the cuticle. They are also readily ingested and able to penetrate the gut wall rapidly and colonize the haemocoel. Multiple entry points and gross damage to the cuticle and gut results in rapid larval death. Conidia neither adhere to the cuticle nor germinate in the gut but cause Pr1 stress induced mortality, which takes a slightly longer time. Blastopores, therefore, have greater potential for the control of Ae. aegypti larvae in mosquito control programmes

Citation: Alkhaibari AM, Carolino AT, Yavasoglu SI, Maffeis T, Mattoso TC, Bull JC, et al. (2016) Metarhizium brunneum Blastospore Pathogenesis in Aedes aegypti Larvae: Attack on Several Fronts Accelerates Mortality. PLoS Pathog 12(7): e1005715. doi:10.1371/journal.ppat.1005715

Editor: Elizabeth Ann McGraw, Monash University, AUSTRALIA

Received: February 23, 2016; Accepted: June 1, 2016; Published: July 7, 2016

Copyright: © 2016 Alkhaibari et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: Funding for this research was obtained from ERASMUS (http://www.erasmusprogramme.com) for SIY, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior www.capes.gov.br for RIS ATC TCM, Conselho Nacional de Desenvolvimento Científico e Tecnológico www.cnpq.br for RIS, Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro www.faperj.br for RIS, European Regional Development Fund, Ireland-Wales Programme (INTERREG 4A) http://wefo.wales.gov.uk/programmes/territorial/irelandwales/?lang=en) for TMB, Saudi Arabian Cultural Bureau, London http://uksacb.org/uk-en1313/ for AMA, and Tabuk University, Kingdom of Saudi Arabia https://www.ut.edu.sa/web/university for AMA. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Aedes aegypti is the vector of a wide range of viral diseases (e.g. yellow fever, dengue, Chikungunya and Zika) [1-5]. Dengue fever annually affects 284 to 528 million people around the world [6]. The range of this pest appears to be expanding due to global warming [7]. Of major concern is the establishment of Ae. aegypti and Aedes albopictus throughout Europe with the latter now firmly established in Southern Europe [7]. The success of these two species is partly due to their ability to readily adapt to urban environments and the tolerance of the eggs to desiccation [8]. Current control is still heavily dependent upon the use of chemical pesticides, which should be discouraged because of the risks they pose to human health and the environment [9, 10]. Moreover, mosquitoes are also rapidly developing resistance to chemical insecticides as well as to the biological larvicide Bacillus thuringiensis [11-14]. Much attention is currently being focussed on the use of entomopathogenic fungi (EPF) such as Beauveria bassiana and Metarhizium anisopliae for the control of mosquito adults and larvae [15-24] as they are considered to be environmentally friendly and highly versatile [25].

Both aerial conidia and blastospores are highly efficacious in killing mosquito larvae [26-28]. Blastospores differ from conidia in several ways. The former are thin-walled, pleomorphic, hydrophilic spores produced relatively inexpensively due to short fermentation times within 2-3 days in liquid media, whereas conidia are uniform shaped, hydrophobic spores produced within 12-20 days on solid substrates such as rice [28, 29]. Although aerial conidia have a comparatively longer shelf life, blastospores are normally considered more virulent against susceptible hosts [28-38]. Exactly why blastospores are more aggressive is unclear. Blastospores generally germinate faster than conidia (2-8hrs versus 12-24 hrs) and this attribute is considered to be a virulence determinant [29, 39]. Slower germination means longer exposure of propagules to deleterious biotic (e.g. antagonistic microbes) and abiotic (e.g. humidity, UV, temperature) factors that negatively affect propagule viability [40, 41]. Furthermore, it gives the host more time to mobilise its defences and resist infection [42, 43]. In the aquatic environment, blastospores of B. bassiana, Tolypocladium cylindrosporum and M. anisopliae were found to be more virulent against mosquito larvae than aerial conidia [27, 28, 44]. According to Miranpuri and Khachatourians [28] the primary infection sites of B. bassiana blastospores were the head and the anal region, although the most preferred site for invasion was the larval gut. However, none of these studies provided an explanation as to why the blastospores were more virulent than conidia. Both conidia and blastospores adhere to the surface of terrestrial arthropod hosts and penetrate the cuticle using a combination of enzymes and mechanical force [45, 46]. Recent studies have shown that the mode of M. anisopliae pathogenesis against Ae. aegypti larvae was radically different from that observed when attacking terrestrial hosts in that the conidia failed to adhere to the cuticle surface and that death was due to stress induced in the insect gut by the spore bound proteases on the surface of ingested conidia [47, 48]. Furthermore, the ingested conidia did not germinate and colonise the haemocoel but remained confined to the gut lumen [47]. It was postulated that M. anisopliae, which is normally found in the soil, has not evolved to infect aquatic invertebrates, hence the atypical mode of pathogenesis. To date, there has been no detailed study of the infection processes of EPF blastospores. Some fluorescence and ultrastructural studies of EPF blastospores have been conducted but mostly of those infecting terrestrial insects [49, 50]. The current multidisciplinary study investigates various aspects of Metarhizium brunneum blastospore pathogenesis in Ae. aegypti larvae with the goal of understanding the infection process when compared to that observed in terrestrial hosts. The current study establishes why blastospores were more virulent than conidia and are highly interesting candidates for vector control, given that mortality was observed in hours rather than days. The significance of these findings as regards the use of Metarhizium for mosquito control is discussed.

Results

Comparison of blastospore and conidial virulence

Significant differences in survival were observed among treatments ([chi]2 = 163.7, df = 3, overall P<0.001, Fig 1). Blastospores of M. brunneum ARSEF 4556 were significantly more virulent against Ae. aegypti larvae than either the wet ([chi]2 = 49.13, pairwise P<0.001) or dry ([chi]2 = 55.32, P<0.001) conidial formulations (Fig 1). There was no significant difference in survival between the wet and dry conidia ([chi]2 = 0.568, P = 0.451, Fig 1). Blastospores caused 100% mortality 2 days post inoculation (pi) (LT50 = 0.92 days), while wet and dry conidia only caused 100% mortality at 5 days pi with LT50 values of 2.52 and 2.76 days respectively (Fig 1).

Fig 1. Susceptibility of Aedes aegypti larvae to blastospores, wet and dry conidia of Metarhizium brunneum ARSEF 4556.

Survival curves of Ae. aegypti (3rd-4th) instar larvae (n = 30 per treatment) inoculated with blastospores or conidia (wet and dry) of M. brunneum ARSEF 4556 applied at final dose of 107 spores ml-1. Blastospores significantly decreased survival compared to all other treatments (p < 0.001), but there were no statistical differences in the survival when comparing wet and dry conidia. The negative controls were either distilled water or 0.05% aqueous Tween. Error is represented as SE.

http://dx.doi.org/10.1371/journal.ppat.1005715.g001

Protease activity during blastospore pathogenesis

Three lines of evidence are presented which show that proteases are not critical virulence determinants during blastospore infection. Firstly, the level of Pr1 associated with blastospore pellets was significantly lower than for wet (diff = 41.97 [95% c.i.: lower = 17.55, upper = 66.38], P = 0.003) or dry conidia (50.49 [26.08, 74.90], P = 0.001) as shown in Fig 2. Secondly, conducting assays at increased temperatures did not accelerate larval mortality with blastospores ([chi]2 = 1.000, df = 3, P = 0.317), where the respective morality rates from blastospore infection were 97% and 100% at 20°C or 27°C 24 hrs pi (Fig 3). These values are similar to those observed at 25°C (Fig 1). In contrast, conidia (wet/dry) bound proteolytic activity increased with temperature and corresponding mortality rates of larvae inoculated with dry or wet conidia were significantly higher at 27°C than 20°C as shown in Fig 3 (Dry: [chi]2 = 5.214, df = 3, P = 0.022; Wet: [chi]2 = 6.513, df = 3, P = 0.011). Finally, protease inhibitors did not influence blastospore virulence (Fig 4) whereas they greatly affected conidial virulence [47]. Although mortality appeared to be slightly lower in suspensions containing protease inhibitors 24 hr pi, there were no statistically significant differences in virulence between M. brunneum untreated blastospores when compared to blastospores treated with [alpha] 2-macroglobulin ([chi]2 2.778, df = 6, P = 0.096) or chicken egg white inhibitor ([chi]2 = 1.100, df = 6, P = 0.294) as seen in Fig 4. No differences in larval survival were observed when using [alpha] 2-macroglobulin and chicken egg white inhibitors (Fig 4: [chi]2 = 0.406, df = 6, P = 0.524,). Furthermore, there were no significant differences in survival between larvae exposed to heat killed blastospores and the untreated control (Fig 4: [chi]2 = 1.000, df = 6, P = 0.317).

Fig 2. Comparison of Pr1 enzyme activity between blastospores, wet and dry conidia of Metarhizium brunneum ARSEF 4556.

Activity of Pr1 bound to the cell wall of conidia (dry and wet) and blastospores in culture media. Letters denote statistical differences. Error is represented as 95% ci.

http://dx.doi.org/10.1371/journal.ppat.1005715.g002

Fig 3. Survival of Aedes aegypti larvae exposed to conidia or blastospores at 20°C and 27°C.

Survival curves of Ae. aegypti larvae (n = 30) exposed to conidia (wet & dry) and blastospores of M. brunneum at a concentration of 107 spores ml-1 incubated at 20°C and 27°C. Survival was similar for larvae inoculated with blastospores at 20°C or 27°C (P = 0.317) but significant differences were observed for larvae exposed to either dry (P = 0.022) or wet (P = 0.011) conidia when comparing these two temperatures. Controls consisted of distilled water or 0.05% aqueous Tween. Error is represented as SE.

http://dx.doi.org/10.1371/journal.ppat.1005715.g003

Fig 4. Survival of Aedes aegypti larvae exposed to blastospores in the presence of different protease inhibitors.

Ae. aegypti larvae (n = 72) were exposed to M. brunneum blastospores with and without the addition of the protease inhibitors chicken egg white (CEW, a Pr1 specific inhibitor) and [alpha] 2-macroglobulin (global protease inhibitor). The Kaplan-Meier method was used to plot survival curves of larvae; Log-rank test was used to assess difference in survival between treatments. Controls consisted of either distilled water or distilled water with protease inhibitors. Error is represented as SE.

http://dx.doi.org/10.1371/journal.ppat.1005715.g004

Interaction of blastospores with the integument and intestine

Blastospores of M. brunneum adhered to almost any part of the mosquito larval cuticle (Fig 5A and 5B). There appeared to be no blockage of the mouthparts or siphons (Fig 5A). Blastospores often formed clumps but they were also present as individual propagules (Fig 5C). Low temperature-scanning electron microscopy (SEM) showed that the blastospores were often covered with copious mucilage which was present in sheet, reticulate and strand form (Fig 5C). The mucilage appeared to be water insoluble since it was present when larvae were recovered from water. Mucilage strands were extruded at the fungus-cuticle interface and were particularly abundant at blastospore apices (Fig 5C). Mucilage strands were strong as they resisted destruction when samples were plunged in the preparatory pre-cooled nitrogen slush and their structure was not affected by the solvents used during sample preparation for transmission electron microscopy (TEM). In thin sections, mucilage appeared as an amorphous, non-uniform, matrix of fibrils that coated the relatively thin cell wall but also extended beyond the blastospores (Fig 6A).

Fig 5. Scanning electron microscopy of Aedes aegypti larvae infected with Metarhizium brunneum blastospores.

Larvae were inoculated with 1X107 blastospores ml-1 and prepared for SEM 20 hrs post inoculation. (A): Head of Ae. aegypti larva showing blastospores (BS) attached to the surface of the cuticle. (B): Blastospores at different stages of germination attached to surface of abdomen. (C): Germinating and non-germinating blastospores surrounded by a mucilaginous matrix (M). (D). Cross section of infected larva showing blastospores of M. brunneum occluding the gut lumen (GL).

http://dx.doi.org/10.1371/journal.ppat.1005715.g005

Fig 6. Transmission electron microscopy of Aedes aegypti larvae infected with Metarhizium brunneum blastospores 24 hr post-inoculation.

(A) Blastospore (BS) firmly adhering to surface of the cuticle (CU). The blastospore has dense cytoplasm, a relatively thin wall and a coating of mucilage (MU) which extends beyond the fungus. The mucilage consists of heterogeneous electron opaque material. (B) Penetration of the larval cuticle. Vacuoles (V) containing electron dense material evident in the penetration hyphae. Cuticle readily distorted by penetration hypha. (C) Blastospores in gut lumen penetrating the peritrophic membrane (PM). (D) One blastospore has penetrated the midgut epithelium and has entered the haemocoel (H).

http://dx.doi.org/10.1371/journal.ppat.1005715.g006

Most blastospores appeared to be turgid, cylindrical cells with a smooth surface (Fig 5C). They grew over the surface of the cuticle but not extensively with little evidence of branching. Blastospores rarely produced appressoria, with penetration pegs being produced sub-apically (Fig 6B). The relatively short, narrow penetration pegs quickly expanded within the cuticle or soon after breaching the cuticle (Fig 6B). The peg retained a relatively thin cell wall which was covered with an amorphous matrix (Fig 6A). There was no obvious clearing around the peg or gross distortion of the cuticle at the penetration site.

Both SEM and light microscopy confirmed that blastospores were ingested by the larvae. Numerous blastospores were present in the gut lumen (Figs 5D and 7). Many cells adhered to the peritrophic membrane, these became swollen and ultimately gave rise to penetration hyphae which penetrated the peritrophic membrane and eventually the midgut epithelium (Fig 6C). The blastospores colonizing the haemocoel after breaching the gut were similar in phenotype to those produced in liquid media (Fig 7).

Fig 7. Cross section of Aedes aegypti larvae 24 hrs post infection with Metarhizium brunneum blastospores.

(A) Blastospores of M. brunneum mostly confined to gut lumen. (B) Blastospores adjacent to the peritrophic membrane are swollen with some having penetrated the peritrophic membrane and midgut epithelium. Cells colonizing the haemocoel consisted of short filaments or hyphal bodies as well as yeast like cells. There was no evidence of branched filamentous hyphae. EP: Epithelium, PM: peritrophic membrane, MV: Microvilli, N: Nuclei, BS: Blastospores, GL: Gut lumen.

http://dx.doi.org/10.1371/journal.ppat.1005715.g007

Blastospore structure

TEM revealed that each blastospore contained dense cytoplasm, several nuclei and mitochondria but very few vacuoles. Most of the vacuoles were relatively small, containing electron opaque material. This material appeared to be extruded through the cell wall and deposited between the fungus and host cuticle (Fig 6B). The electron opaque material was associated with individual blastospores as well as blastospores in groups at the cuticle surface. Electron opaque material was also observed associated with ingested blastospores in the gut lumen (Fig 6C). The distribution of this material was not extensive or uniform.

The cytoplasm retained a dense appearance irrespective of whether the blastospores grew on the host surface, penetrating the cuticle or midgut epithelium or when colonizing the haemocoel. Blastospores appeared to divide by budding but the size of the daughter cells varied before breakage from the mother cell. In some instances the daughter cells appeared as short filaments (Fig 7).

A striking feature of blastospores was the poor staining of the cell wall with calcofluor (Fig 8). Both incipient and fully formed septa fluoresced intensely. Intense fluorescent spots were observed at the apices of some blastospores and occasionally along the length of the cell (Fig 8). The pattern of staining was similar whether the blastospores were produced in vitro or in vivo. The plasma membrane stained weakly with filipin, fluorescence being intense at septa and cell apices (Fig 8). Fluorescent spots were also observed in a number of cells (Fig 8).

Fig 8. Fluorescence microscopy of Metarhizium brunneum blastospores.

(A) Calcofluor White staining of blastospores. Cell walls fluoresced weakly except at apices and septa. (B) Rhodamine 123 was used to visualise mitochondria in blastospores. (C) Filipin staining of ergosterol present in the plasma membrane. (D) Fluorescent staining of nuclei with DAPI. (E) FITC staining of proteins (F) Blastospores as seen in bright-field (F). Scale bar = 5 Mm.

http://dx.doi.org/10.1371/journal.ppat.1005715.g008

Expression of Metarhizium pathogenicity related genes

Expression of genes directly and indirectly linked with pathogenesis namely Pr1, Pr2, Mad1, Mad2, Mos1, Cag8, and nrr1 was generally higher in blastospores in the aquatic setting (i.e. in infected larvae or in the presence or absence of the larvae) than in Tenebrio, the terrestrial positive control host (Fig 9). Expression of the cuticle degrading enzyme Pr1 was highest in blastospores in the presence of larvae (i.e. blastospores that were still in suspension), which was statistically similar to blastospores in infected but live larvae (diff = -0.198 [95% c.i.: lower = -1.837, upper = 1.440], Tukey's HSD: P = 0.993). Expression was slightly lower in infected dead insects but much lower in blastospores suspended in water in the absence of Aedes larvae (Fig 9). Expression of the protease Pr2 was highest in dead infected larvae but with similar levels of expression when comparing blastospores in the presence and absence of larvae (-2.320 [-7.699, 3.060], P = 0.595). In infected live larvae, Pr2 expression was the lowest among all treatments excluding Tenebrio (Fig 9). Most notably there was an elevated expression of the adhesion genes Mad1 and Mad2. Mad 1 expression was highest in M. brunneum 4556 blastospores in the presence of larvae (Fig 9). However, this was not significantly different to infected live larvae (-5.735 [-12.39, 0.924], P = 0.096), dead infected larvae (-2.706 [-10.15, 4.739], P = 0.723), and blastospores in absence of larvae (4.833 [-1.826, 11.49], P = 0.181, Fig 9). Expression of Mad2 was equally high in infected dead larvae and blastospores in presence of live larvae (0.590 [-3.066, 4.245], P = 0.980). Similar expression levels of Mad2 were observed in infected live larvae, blastospores in the absence of larvae and Tenebrio (Fig 9).

Fig 9. Expression of pathogenicity related genes in Metarhizium brunneum blastospores 20.3 hours post-infection.

Gene expression analysed by quantitative PCR included: proteases (Pr1A, Pr2), adhesins (Mad1, Mad2), an osmosensor (Mos1),regulators of G-protein signalling (Cag8) and nitrogen (nrr1). X axis shows: 1: infected living larvae, 2: infected dead larvae, 3: blastospores in the presence of Ae. aegypti larvae 4: blastospores in absence of the larvae, 5: Tenebrio molitor (terrestrial host) positive control. Boxes denote interquartile range, bisected horizontally by median values; whiskers extend to 1.5× interquartile range beyond boxes; outliers are marked as dots beyond whiskers. Expression is shown as the inverse of the number of amplification cycles to reach Critical Threshold values (CT-1).

http://dx.doi.org/10.1371/journal.ppat.1005715.g009

Expression of Cag8 and Mos1 was highest in live infected larvae followed by dead infected larvae. Cag8 was higher in blastospores in the presence of live larvae than in their absence. Mos1 was lowest in blastospores whether in the presence or absence of live larvae but similar to blastospores infecting Tenebrio (Fig 9). Expression of the nitrogen regulator gene nrr1 was high in blastospores in all treatments except for Tenebrio (Fig 9).

Expression of Ae. aegypti immune and stress management genes during infection

The immune defence response of Ae. aegypti larvae was rapid following exposure to M. brunneum AFSEF 4556 blastospores; expression of all five AMPs (AeDA, AeDB1, Ada-defD, AeCA2, Ada-ccg) was elevated immediately after exposure to the pathogen (Fig 10). However, at 12 hrs pi only the expression of AeDA, AeDB1 remained high. However by 20.5hrs pi, all the AMP genes were down regulated (Fig 10). The expression level of AeDA, AeDB1 at 20.5 pi was significantly different in comparison with the control (AeDA: -4.468 [-6.428, -2.508], P = 0.002), (AeDB1: -5.375 [-6.803, -3.947], P < 0.001), or 12 hrs pi (AeDA: -5.428 [-7.619, -3.236], P = 0.001), (AeDB1: -5.985 [-7.581, -4.388], P < 0.001), where the gene expression was down regulated, with no difference between the control and 12hrs pi (AeDA: 0.960 [-1.232, 3.151], P = 0.398), (AeDB1: 0.610 [-0.987, 2.206], P = 0.481). Expression of Ada-defD, AeCA2 and Ada-ccg was significantly lower than the controls at 12 hrs (Ada-defD: -9.078 [-11.79, -6.371], P < 0.001), (AeCA2: -10.67 [-13.79, -7.548], P < 0.001), (Ada-ccg: -10.54 [-12.59, -8.496], P < 0.001) and 20.5 hrs pi (Ada-defD: -8.708 [-11.13, -6.286], P < 0.001), (AeCA2: -10.215 [-13.01, -7.424], P < 0.001), (Ada-ccg: -10.43 [-12.26, -8.600], P < 0.001). However, between 12hrs and 20.5 hrs pi there was no significant difference (Ada-defD: 0.370 [-2.337, 3.078], P = 0.899), (AeCA2: 0.454 [-2.667, 3.574], P = 0.887), (Ada-ccg: 0.112 [-1.935, 2.159], P = 0.983).

Fig 10. Expression of Aedes aegypti antimicrobial peptides (AMP) and stress management genes during infection with Metarhizium brunneum blastospores.

Quantitative real time PCR used to analyse expression of stress management and AMP genes in Ae. aegypti larvae inoculated with M. brunneum blastospores at 1 = 0hr, 2 = 12hr and 3 = 20:30 hrs post-inoculation. AMP genes included; AeDA (Defensin A), AeDB1 (Defensin B), Ada-defD (Defensin D), AeCA2 (Cecropin A), and Ada-ccg (Cecropin G). Stress management genes include; HSP70 (Heatshock protein 70), HSP83 (Heatshock protein 83), GPX (Glutathione peroxidase), Cyp6Z6 (Cytochrome P450), and TPX10 (Thiol peroxidase 10). Boxes denote interquartile range, bisected horizontally by median values; whiskers extend to 1.5× interquartile range beyond boxes; outliers are marked as dots beyond whiskers. Expression is shown as the inverse of number of amplification cycles to reach Critical Threshold values (CT-1).

http://dx.doi.org/10.1371/journal.ppat.1005715.g010

Expression of five genes associated with stress management (HSP70, HSP83, GPX, Cyp6Z6, TPX10) showed an unusual disjointed pattern of expression (Fig 10). Expression of the heat shock genes HSP70 and HSP83 was similar at all three time points (HSP70: F2,5 = 1.701, P = 0.273), (HSP83: F2,5 = 0.811, P = 0.495). Expression of GPX, Cyp6Z6, and TPX10 was high at 0hrs pi but was low at 12 hrs pi. All three genes appeared to be up-regulated at 20.5hrs pi. The level of GPX at 20.5 hr pi was significantly higher than at 12 hrs pi (diff = 3.754 [95% c.i.: lower = 1.844, upper = 5.663], Tukey's HSD: P = 0.003), but it was significantly lower than in non-infected larvae (-1.1858 [-3.566,-0.150], P = 0.037). Expression of Cyp6Z6 and TPX 10 was significantly lower at 12 hrs pi than in the controls (Cyp6Z6: -7.178 [-10.46, -3.893], P = 0.002), (TPX 10: -2.543 [-4.330, -0.757], P = 0.013).

Discussion

Blastospores of M. brunneum strain ARSEF 4556 were significantly more virulent than conidia in killing Ae. aegypti larvae. This study shows for the first time that blastospores have specific characteristics which explain why they are more aggressive to mosquito larvae than conidia. These attributes include the ability to readily adhere to the cuticle surface, production of copious mucilage and the ability to infect without the differentiation of appressoria or excessive production of the proteases Pr1 and Pr2. Blastospores infected larvae through integument and gut, retaining the dynamic blastospore form both on the host surface and during invasion of the haemocoel. This study also shows that the host immune and stress management responses rapidly kick in but are inadequate as they fail to prevent infection.

One of the most striking features of M. brunneum blastospores is their ability to adhere to the surface of the mosquito larval cuticle whereas conidia of the same fungus are unable to do so [47, 48]. Blastospores are hydrophilic, they readily suspend in water and have high affinity for hydrophilic surfaces [51]. This study shows for the first time that M. brunneum blastospores produce copious, robust, water insoluble mucilage, which may not only explain why these cells tended to clump but also firmly adhere to the mosquito larval surface in an aquatic environment. Exactly why the blastospores expend resources and energy in mucilage production is unclear. Fungi produce an extracellular matrix for a variety of reasons including adhesion of conidia, hyphae, appressoria and blastospores to host surfaces, protection against harmful UV radiation and creation of environments conducive for cuticle degrading enzyme activity [52]. There appears to be a relationship between mucilage production and hydrophilicity since several EPF species that produce wettable conidia often possess a mucilaginous coat [53].

Expression of the adhesin Mad1 gene was elevated in blastospores exposed to Ae. aegypti larvae. This corroborates the findings of Barelli et al. [54] who found that Mad1 was upregulated in the presence of insect cuticle. Disruption of Mad1 in M. anisopliae delays germination, suppresses blastospore formation, and greatly reduces virulence [55]. Sustained expression of Mad1 may explain why M. brunneum retained the blastospore form during invasion and colonisation of Ae. aegypti larvae. In contrast, Mad2 expression was high in dead infected Ae. aegypti larvae and blastospores in the presence of larvae. Mad2 is associated with adhesion to plant surfaces; disruption of this gene, which is under stress control, blocks adhesion of M. anisopliae to plant surfaces but has no effect on fungal differentiation or virulence [55]. Expression of Mad2 as seen here, suggests that the gene was under starvation stress control as previously reported by Barelli and colleagues [54].

It was not possible to determine the chemical composition of the mucilage but it was clear from the SEM and TEM studies that the mucilage was highly resilient and that it contained feint electron opaque fibrils. Similar material has been reported associated with appressoria of several plant pathogenic fungi [56]. None of the fluorochromes tested stained the mucilage. Calcofluor only weakly stained the cell wall of M. brunneum blastospores suggesting that its composition was substantially different to that of the septa, bud scars and growing points which fluoresced intensely. Presumably the blastospore cell wall lacked [Beta]-glucans or that these sugars were masked by mucilage. Several researchers have reported that blastospores growing in vivo possess fewer carbohydrate epitopes than conidia or blastopores incubated in artificial culture media [57, 58]. [Beta]-glucans are powerful elicitors of the insect immune system [59]. Absence or masking of [Beta]-glucans may be an adaptation to avoid recognition by the host's immune defences and may explain why no significant cellular or humoral defence responses were observed in Aedes larvae. Expression of AMPs was elevated immediately after exposure to the inoculum but not later suggesting that water soluble factors were being detected by the host, eliciting an immune response.

Another novel finding of this study was the failure of protease inhibitors to prevent blastospores from causing larval mortality even though the Pr1 and Pr2 cuticle degrading protease genes were expressed in blastospores. Pr1 is a virulence determinant of terrestrial arthropod hosts [60] and can cause stress induced mortality in Ae. aegypti larvae following ingestion of viable conidia [47]. It is tempting to speculate that blastospores are therefore more dependent upon mechanical forces or enzymes other than Pr1 to penetrate the larval cuticle and peritrophic membrane. The current observations and the fact that EPF are able to penetrate non-host plant cells and inert substrates, suggest that mechanical force may be more important than previously realised [61, 62].

The highest level of expression of Cag8, nrr1, and Mos1 was detected in infected living larvae, presumably their concerted activity enables M. brunneum blastospores to adapt to the host and cope with disparate stresses encountered particularly at the cuticle surface and gut lumen [63]. Mos1 plays a role in fungal differentiation (appressoria, hyphal bodies) and ability to cope with stress (osmotic, oxidative). Inactivation of this gene reduces virulence [55]. Although M. brunneum blastospores failed to readily produce appressoria, Mos1 probably plays a role in the infection process such as maintaining cell turgor and stress management since the level of expression declines in dead infected insects. Cag8 plays a role in the regulation of conidiation, virulence and hydrophobin synthesis [64]. Disruption of this gene can result in the production of irregular shaped blastospores and a decline in virulence [63]. In the current study this gene may be involved in the retention of the blastospores phenotype and contribute to its virulence, hence the elevated activity during infection in the live insects but a decline in dead infected insects.

Pr1 and Pr2 are subject to nitrogen derepression by nrr1. The nrr1 protein binds to GATA sites from the promoter region of the Pr1 gene in M. anisopliae during nitrogen regulation [65]. The highest levels of Pr1 were observed in blastospores exposed to live Aedes larvae or during host colonisation suggesting that the fungus was responding to host cues before and during infection, with the level of expression remaining constant. Interestingly, Pr1 levels were depressed in dead hosts whereas Pr2 levels were highly elevated suggesting that Pr2 may have functions other than cuticle degradation. The high level of expression of all the virulence related genes examined in blastospores incubated with live Ae. aegypti larvae suggests that the blastospores can sense the presence of the insect host even before adhering to the cuticle.

Just as the fungus was able to sense the insect before initiating infection, mosquito larvae could also sense the pathogen since AMP genes were elevated 0 hrs pi with only defensin A and B remaining elevated at 12 hrs pi when many insects were highly infected and even dying. In contrast, expression of genes involved with stress management showed an unusual pattern; expression was high at 0 and 22.5hrs pi but declined at 12 hrs pi. The only exception was Hsp70 which peaked at 12 hr pi. Investment in stress management was a high priority and remained high even at the point of death. Presumably, this was a better investment of resources than the AMPs since most of these are known to exhibit antibacterial rather than antifungal activity [63].

This study shows that the physiological and morphological adaptations of blastospores enable them to adhere to and infect mosquito larvae via the integument as well as invade through the gut following ingestion. These multiple routes of entry result in Ae. aegypti larvae being killed within hours. Blastospores unlike conidia produce copious mucilage which ensures strong adhesion to the host surface. Furthermore, blastospores are not dependent upon differentiation of appressoria or production of Pr1 for infection. These adaptations may explain why blastospores are more virulent than conidia at least when infecting mosquito larvae. However, there are other attributes which make this form of inoculum ideal for control of Ae. aegypti larvae. They are inexpensive and rapidly produced in liquid media. Their hydrophilic nature means that surfactants are not required. Aerial conidia not only take longer to produce but need surfactants to suspend them in water. Thus, M. brunneum blastospores meet most of the commercial criteria for selecting pest control products i.e. being relatively inexpensive, safe to humans and the environment, easy to use and highly efficacious (fast acting) against the target pest. These criteria are particularly important since mosquito control often requires treatment of large areas and most affected countries have limited resources.

Materials and Methods

Mosquito source and maintenance

Aedes aegypti (strain AeAe) eggs, obtained from the London School of Hygiene and Tropical Medicine (UK), were hatched in tap water. Larvae were fed rabbit food (Burgess) and fish food (Tetra pro) and kept at room temperature (25±2°C) in a 16L:8D photoperiod.

Fungal strains and production

Metarhizium brunneum (Formerly M. anisopliae) isolate ARSEF 4556, identified as highly pathogenic to mosquitoes [24, 47], was maintained on Sabouraud dextrose agar (SDA). Conidia used in assays had over 95% viability. Conidia were harvested from 15 day old sporulating cultures and suspended in 0.03% (v/v) aqueous Tween 80. Blastospores were produced in Adamek's medium [66]. The medium was inoculated with 107 conidia ml-1 and incubated at 27°C in a rotary shaker at 130 rpm for 3 days. Blastospores were harvested by filtering through 2 layers of lens cleansing tissue (Whatman No.105), washed twice with sterile distilled water then centrifuged at 5000 rpm for 10 min (IEC-Centra-3E) and the pellet was suspended in sterile distilled water. Conidia and blastospore concentrations were determined using an improved Neubauer haemocytometer.

Virulence of M. brunneum blastospores and conidia

Assays were conducted to compare the virulence of M. brunneum ARSEF 4556 blastospores and conidia against Ae. aegypti (L3-4) larvae. Cohorts of ten larvae (n = 30) per incubated in 280 ml plastic beakers containing 100 ml of distilled water + fungus used at a final concentration of 107 propagules ml-1. Conidia were applied either as a dry powder dusted over the water surface (dry conidia) or as a suspension in 0.03% aqueous Tween 80 (wet conidia). Blastospores were applied as a suspension in distilled water. Larval mortality was recorded every 24 hours for 7 days. Dead larvae were transferred to Petri dishes lined with moist filter paper and incubated at 27°C to encourage fungal emergence and sporulation. Controls consisted of either distilled water or 0.03% aqueous Tween 80. Each treatment was replicated 3 times and the complete experiment was repeated three times. Assays were performed at 25 ± 2°C.

Role of proteases in pathogenesis

Since conidial proteases play a key role in mosquito larval pathogenesis [47], several complementary studies were conducted to determine if blastospore proteases also played a role in pathogenesis. The first study consisted of virulence bioassays as outlined above being conducted at 20°C and 27°C, since proteolytic activity increases with temperature. The assay was repeated 3 times.

Another study was conducted using protease inhibitors in 24 well plates (Nunc, Roskide, Denmark, n = 72) with one larva per well. Treatments included larvae exposed to 1 ml of 107 blastospores under the following conditions: (1) live blastospores, (2) heat killed blastospores (autoclaved for 15 min at 121°C), (3) live blastospores incubated with chicken egg white Pr1 protease inhibitor (0.1 mg/ml), (4) live blastospores incubated with [alpha] 2-macroglobulin (1 Mg/ml). The latter is an inhibitor of serine, cysteine and metallo-proteases. The inhibitors were purchased from Sigma-Aldrich. Control larvae were exposed to 1ml of distilled water or the inhibitors in distilled water. Mortality was recorded daily over 4 days and the assays repeated 3 times. For more details of the assays see Text S1 in S1 File.

Infection processes of M. brunneum blastospores

Larvae were inoculated with blastospores as described above and examined at 12 and 24 hrs post inoculation (pi). Infected larvae were examined by light microscopy to determine if there were preferential sites for blastospores adhesion and to observe the blastospores in the gut. Larvae were also prepared for examination by transmission electron microscopy (TEM). Full details on the preparation and staining of material for TEM are provided in Text S2 in S1 File. Thick sections of the resin embedded larvae were also examined using a light microscope. Full details on the preparation and staining of the material for light microscopy are provided in Text S3 in S1 File. Larvae (n = 20) were also examined by cryo-scanning electron microscopy (SEM) using a Hitachi S4800 field emission microscope equipped with a Quorum PPT2000 cryogenic stage and preparation chamber. Full details on the cryo-SEM are provided in Text S4 in S1 File.

Blastospore structure

In addition to SEM and TEM, certain cell attributes of blastospores were investigated using a range of fluorochromes. Calcofluor White, Filipin, Rhodamine 123 and DAPI (4', 6-diamidino-2-phenylindole) were used to visualize [Beta]-glucans (in cell walls, septa, mucilage), ergosterol (in cell membranes), mitochondria and nucleic acids, respectively. The fluorochromes were purchased from Sigma. Stained cells were examined in a Zeiss AxioCam MR3 fluorescence microscope. Details on the preparation, staining and visualization of the cell components are provided in Text S5 in S1 File.

Transcript quantification of insect and fungus-derived genes

Samples, RNA extraction and cDNA synthesis.

Studies focussed on specific genes which are known to play a major role in fungal pathogenesis or insect defence against infection [63]. Full details of the transcript quantification are given in Text S6 in S1 File. Briefly, Ae. aegypti larvae (L3-4) (n = 10 larvae per replicate with 3 replicates per treatment) were exposed to blastospores of M. brunneum ARSEF 4556, controls included larvae not exposed to fungus (negative control), as well as a terrestrial insect, Tenebrio molitor, that was exposed to blastospores (positive control). Biological samples were frozen in liquid nitrogen and stored at -80°C until required. All samples were ground with a micropestle and total RNA extractions carried out using the RNeasy Micro kit (Qiagen) following the manufacturer's instructions. RNA concentration and purity was assessed at 260 and 280 nm absorbance using a Nanophotometer (Implen). Total RNA (1 Mg) was treated with RNase-Free DNase Set (Qiagen) and reverse transcribed using a QuantiTect Reverse Transcription kit (Qiagen) with gDNA elimination reaction, for the experiment to quantify insect-derived transcripts and fungus-derived transcripts respectively. Relative quantification of cDNA was performed by PCR using two reference genes for insect or fungal cDNA samples to ensure consistency between values. For Ae. aegypti: ribosomal S7 (accession number: AAEL009496) and ribosomal protein 49/L32 (AAEL003396). For M. brunneum: 18S rRNA and elongation factor tEF (Table S1 in S1 File).

Quantitative real time PCR (qRT-PCR).

PCR reactions were performed in 10 Ml volumes consisting of 1MM of each primer, 2 Ml of the cDNA sample, 5 Ml SYBR Green Fastmix (Quanta) and 1 Ml ultrapure water. All reactions were carried out in triplicate. PCR cycling conditions were as follows: one cycle of 45°C for 5 min and 95°C for 3 min followed by 39 cycles of 95°C for 10 sec, 60°C for 10 sec and 72°C for 30 sec. A dissociation step of 65-95°C over 5 sec was used for melt curve analysis to detect non-specific products in the reaction. The reaction was performed using a Rotor-Gene 6000 (Corbett Research). The comparative Ct method (2-[increment][increment]Ct) was used to calculate relative gene expression [67].

Statistics

Mosquito larval survival was investigated in three treatment groups; 1) conidia, 2) blastospores and 3) blastospores incubated in protease inhibitors. Each experimental treatment was replicated three times. The whole study was repeated three times. Cumulative survival was quantified using Kaplan-Meier plots. Pairwise comparison were made between; 1) conidia vs blastospores and 2) blastospores vs blastospores with inhibitor using log-rank tests. Pr1 activity and molecular data sets were analysed using one-way Analysis of Variance (ANOVA) with Tukey´s HSD post-hoc test to assess pairwise comparisons.

Prior to analysis, gene expression data was logarithm transformed, conforming to ANOVA assumption of homogeneity of variance [68]. All statistical analyses were carried out using SPSS v22.0 [69], Rstudio Version 0.99.482 [70]and GraphPad Prism v5.0 (GraphPad Software, USA).

Genes and corresponding accession numbers (see Text S6 in S1 File)

Ae. aegypti genes; Ribosomal S7 (AAEL009496), Ribosomal protein 49/L32 (AAEL003396), Defensin A (AAEL003841), Defensin B (AF156090.2), Defensin D (AAEL003857), Cecropin A (AAEL000627), Cecropin G (AAEL015515), heat shock protein 70 (AAEL016995), heat shock protein 83 (AAEL011704), Thiol peroxidase (AAEL004112), Cytochrome P450 (AAEL009123), and Glutathione peroxidase (AAEL008397)

M. brunneum; PR1a (MBR_01491), PR2 (MBR_06579), MAD1 (MBR_08250), MAD2 (DQ338438.1), MOS1 (MBR_07375), Cag 8 (MBR_00569), Nrr1 (MBR_08301), 18s (DQ288247.1),Translation elongation factor (MBR_08275)

Supporting Information

S1 File.

Text S1: Enzyme and enzyme inhibitor assays. Text S2: Transmission electron microscopy. Text S3: Light microscopy of resin embedded sections. Text S4: Cryo-SEM. Text S5: Fluorescence Microscopy. Text S6: Transcript quantification of insect and fungus-derived genes.

(DOC)

Acknowledgments

We are grateful to Ursula Potter at the University of Bath for help with TEM.

Author Contributions

Conceived and designed the experiments: AMA ATC SIY TM TCM RIS TMB. Performed the experiments: AMA ATC SIY TM TCM RIS TMB. Analyzed the data: AMA JCB. Contributed reagents/materials/analysis tools: TMB RIS. Wrote the paper: AMA JCB RIS TMB.

References

WHO. Yellow fever fact sheet N°100. Weekly Epidemiological Record. 2010;85:33-6. pmid:20151492

WHO. Global strategy for dengue prevention and control 2012-2020: World Health Organization; 2012

Tolle MA. Mosquito-borne diseases. Current Problems in Pediatric and Adolescent Health Care. 2009;39(4):97-140. doi: 10.1016/j.cppeds.2009.01.001. pmid:19327647

Marcondes CB, Ximenes MdFFd. Zika virus in Brazil and the danger of infestation by Aedes (Stegomyia) mosquitoes. Revista da Sociedade Brasileira de Medicina Tropical. 2015;(AHEAD):0-. doi: 10.1590/0037-8682-0220-2015

WHO. Epidemiological update on Zika virus infection--16 October 2015 2015 [cited 2015 14-12-2015]. Available from: http://www.paho.org/hq/index.php?option=com_docman&task=doc_view&Itemid=270&gid=32021&lang=en

WHO. Dengue and severe dengue 2015 [cited 2015 30/12]. Available from: http://www.who.int/mediacentre/factsheets/fs117/en/

Medlock JM, Hansford KM, Schaffner F, Versteirt V, Hendrickx G, Zeller H, et al. A review of the invasive mosquitoes in Europe: ecology, public health risks, and control options. Vector-borne and Zoonotic Diseases. 2012;12(6):435-47. doi: 10.1089/vbz.2011.0814. pmid:22448724

Juliano SA, Philip Lounibos L. Ecology of invasive mosquitoes: effects on resident species and on human health. Ecology Letters. 2005;8(5):558-74. pmid:17637849 doi: 10.1111/j.1461-0248.2005.00755.x

Rose RI. Pesticides and public health: integrated methods of mosquito management. Emerging Infectious Diseases. 2001;7(1):17. pmid:11266290 doi: 10.3201/eid0701.010103

Zaim M, Guillet P. Alternative insecticides: an urgent need. Trends in Parasitology. 2002;18(4):161-3. pmid:11998703 doi: 10.1016/s1471-4922(01)02220-6

Hemingway J, Ranson H. Insecticide resistance in insect vectors of human disease. Annual Review of Entomology. 2000;45(1):371-91. doi: 10.1146/annurev.ento.45.1.371

Ranson H, N'Guessan R, Lines J, Moiroux N, Nkuni Z, Corbel V. Pyrethroid resistance in African anopheline mosquitoes: what are the implications for malaria control? Trends in Parasitology. 2011;27(2):91-8. doi: 10.1016/j.pt.2010.08.004. pmid:20843745

Paris M, Tetreau G, Laurent F, Lelu M, Despres L, David JP. Persistence of Bacillus thuringiensis israelensis (Bti) in the environment induces resistance to multiple Bti toxins in mosquitoes. Pest Management Science. 2011;67(1):122-8. doi: 10.1002/ps.2046. pmid:21162152

Tilquin M, Paris M, Reynaud S, Despres L, Ravanel P, Geremia RA, et al. Long lasting persistence of Bacillus thuringiensis subsp. israelensis (Bti) in mosquito natural habitats. PLoS One. 2008;3(10):e3432. doi: 10.1371/journal.pone.0003432. pmid:18941501

Bukhari T, Takken W, Koenraadt CJ. Development of Metarhizium anisopliae and Beauveria bassiana formulations for control of malaria mosquito larvae. Parasit Vectors. 2012;4(23):1-14. doi: 10.1186/1756-3305-4-23

Bukhari T, Middelman A, Koenraadt CJ, Takken W, Knols BG. Factors affecting fungus-induced larval mortality in Anopheles gambiae and Anopheles stephensi. Malaria Journal. 2010;9(1):22. doi: 10.1186/1475-2875-9-22

Kannan SK, Murugan K, Kumar NA, Ramasubramanian N, Mathiyazhagan P. Adulticidal effect of fungal pathogen, Metarhizium anisopliae on malarial vector Anopheles stephensi (Diptera: Culicidae). Journal of Biotechnology. 7(6)

Mohant SS, Raghavendra K, Dash AP. Induction of chymoelastase (Pr1) of Metarhizium anisopliae and its role in causing mortality to mosquito larvae. World Journal of Microbiology and Biotechnology. 2008;24(10):2283-8. doi: 10.1007/s11274-008-9742-2

Scholte EJ, Njiru BN, Smallegange RC, Takken W, Knols BG. Infection of malaria (Anopheles gambiae ss) and filariasis (Culex quinquefasciatus) vectors with the entomopathogenic fungus Metarhizium anisopliae. Malaria Journal. 2003;2(1):29. doi: 10.1016/j.jip.2005.10.006

Scholte EJ, Knols BG, Takken W. Autodissemination of the entomopathogenic fungus Metarhizium anisopliae amongst adults of the malaria vector Anopheles gambiae ss. Malaria Journal. 2004;3(1):45. doi: 10.1016/j.jip.2005.10.006

Scholte EJ, Ng'habi K, Kihonda J, Takken W, Paaijmans K, Abdulla S, et al. An entomopathogenic fungus for control of adult African malaria mosquitoes. Science. 2005;308(5728):308(5728), 1641-2. pmid:15947190 doi: 10.1126/science.1108639

Scholte E-J, Takken W, Knols BG. Infection of adult Aedes aegypti and Ae. albopictus mosquitoes with the entomopathogenic fungus Metarhizium anisopliae. Acta Tropica. 2007;102(3):151-8. pmid:17544354 doi: 10.1016/j.actatropica.2007.04.011

Pereira CR, de Paula AR, Gomes SA, Pedra PCO Jr, Samuels RI. The potential of Metarhizium anisopliae and Beauveria bassiana isolates for the control of Aedes aegypti (Diptera: Culicidae) larvae. Biocontrol Science and Technology. 2009;19(8):881-6. doi: 10.1080/09583150903147659

Greenfield BP, Peace A, Evans H, Dudley E, Ansari MA, Butt TM. Identification of Metarhizium strains highly efficacious against Aedes, Anopheles and Culex larvae. Biocontrol Science and Technology. 2015;25(5):487-502. doi: 10.1080/09583157.2014.989813

Butt TM, Jackson C, Magan N. Fungi as biocontrol agents: progress problems and potential: CABI; 2001

Nadeau M, Boisvert J. Larvicidal activity of the entomopathogenic fungus Tolypocladium cylindrosporum (Deuteromycotina: Hyphomycetes) on the mosquito Aedes triseriatus and the black fly Simulium vittatum (Diptera: Simuliidae). Journal of the American Mosquito Control Association. 1994;10(4):487-91. pmid:7707051

Riba G, Keita A, Soares G Jr, Ferron P. Comparative studies of Metarhizium anisopliae and Tolypocladium cylindrosporum as pathogens of mosquito larvae. Journal of the American Mosquito Control Association. 1986;2(4):469-73. pmid:2906985

Miranpuri G, Khachatourians G. Larvicidal activity of blastospores and conidiospores of Beauveria bassiana (strain GK 2016) against age groups of Aedes aegypti. Veterinary Parasitology. 1990;37(2):155-62. pmid:2251749 doi: 10.1016/0304-4017(90)90070-r

Jackson M. Optimizing nutritional conditions for the liquid culture production of effective fungal biological control agents. Journal of Industrial Microbiology and Biotechnology. 1997;19(3):180-7. doi: 10.1038/sj.jim.2900426

Shockley M, DOSSEY A. Mass production of beneficial organisms invertebrates and entomopathogens. USA: Academic Press; 2014

Jackson MA, Mcguire MR, Lacey LA, Wraight SP. Liquid culture production of desiccation tolerant blastospores of the bioinsecticidal fungus Paecilomyces fumosoroseus. Mycological Research. 1997;101(01):35-41. doi: 10.1017/s0953756296002067

Wang Y-B, Yang Z-H, Yu J-J, Zhang Y-A, Xue J-J, Li Z, et al. Comparison between conidia and blastospores of Esteya vermicola, an endoparasitic fungus of the pinewood nematode, Bursaphelenchus xylophilus. World Journal of Microbiology and Biotechnology. 2013;29(12):2429-36. doi: 10.1007/s11274-013-1433-y. pmid:23918632

Hall R. Pathogenicity of Verticillium lecani conidia and blastospores against the aphid, Macrosiphoniella sanborni. Entomophaga. 1979;24(2):191-8. doi: 10.1007/bf02375133

Hegedus DD, Bidochka MJ, Miranpuri GS, Khachatourians GG. A comparison of the virulence, stability and cell-wall-surface characteristics of three spore types produced by the entomopathogenic fungus Beauveria bassiana. Applied Microbiology and Biotechnology. 1992;36(6):785-9. doi: 10.1007/bf00172195

Holder DJ, Kirkland BH, Lewis MW, Keyhani NO. Surface characteristics of the entomopathogenic fungus Beauveria (Cordyceps) bassiana. Microbiology. 2007;153(10):3448-57. doi: 10.1099/mic.0.2007/008524-0

Poprawski T, Jackson M. Laboratory activity of blastospores of Paecilomyces fumosoroseus on Bemisia argentifolii nymphs, 1997. Arthropod Management Tests. 1999;24(1):L8. doi: 10.1093/amt/24.1.l8

Behle R, Garcia-Gutierrez C, Tamez-Guerra P, McGuire M, Jackson M. Pathogenicity of blastospores and conidia of Paecilomyces fumosoroseus against larvae of the Mexican bean beetle, Epilachna varivestis Mulsant. Southwestern Entomologist. 2006;31(4):289

Shapiro-Ilan DI, Cottrell TE, Jackson MA, Wood BW. Virulence of Hypocreales fungi to pecan aphids (Hemiptera: Aphididae) in the laboratory. Journal of Invertebrate Pathology. 2008;99(3):312-7. doi: 10.1016/j.jip.2008.07.001. pmid:18675272

Vega FE, Jackson MA, McGuire MR. Germination of conidia and blastospores of Paecilomyces fumosoroseus on the cuticle of the silverleaf whitefly, Bemisia argentifolii. Mycopathologia. 1999;147(1):33-5. pmid:16308757

Rangel DE, Braga GU, Anderson AJ, Roberts DW. Variability in conidial thermotolerance of Metarhizium anisopliae isolates from different geographic origins. Journal of Invertebrate Pathology. 2005;88(2):116-25. pmid:15766928 doi: 10.1016/j.jip.2004.11.007

Lacey L. The effect of selected allelochemicals on germination of conidia and blastospores and mycelial growth of the entomopathogenic fungus, Paecilomyces fumosoroseus (Deuteromycotina: Hyphomycetes). Mycopathologia. 1998;142(1):17-25. pmid:16284855

Ortiz-Urquiza A, Keyhani NO. Action on the surface: entomopathogenic fungi versus the insect cuticle. Insects. 2013;4(3):357-74. doi: 10.3390/insects4030357. pmid:26462424

Vega FE, Dowd PF, McGuire MR, Jackson MA, Nelsen TC. In VitroEffects of Secondary Plant Compounds on Germination of Blastospores of the Entomopathogenic Fungus Paecilomyces fumosoroseus (Deuteromycotina: Hyphomycetes). Journal of Invertebrate Pathology. 1997;70(3):209-13. pmid:9367728 doi: 10.1006/jipa.1997.4693

Soarés GG Jr. Pathogenesis of infection by the hyphomycetous fungus, Tolypocladium cylindrosporum in Aedes sierrensis and Culex tarsalis [Dip.: Culicidae]. Entomophaga. 1982;27(3):283-99. doi: 10.1007/bf02374812

Butt T, Ibrahim L, Clark S, Beckett A. The germination behaviour of Metarhizium anisopliae on the surface of aphid and flea beetle cuticles. Mycological Research. 1995;99(8):945-50. doi: 10.1016/s0953-7562(09)80754-5

Vestergaard S, Butt T, Bresciani J, Gillespie A, Eilenberg J. Light and Electron Microscopy Studies of the Infection of the Western Flower Thrips Frankliniella occidentalis (Thysanoptera: Thripidae) by the Entomopathogenic Fungus Metarhizium anisopliae. Journal of Invertebrate Pathology. 1999;73(1):25-33. pmid:9878286 doi: 10.1006/jipa.1998.4802

Butt TM, Greenfield BP, Greig C, Maffeis TG, Taylor JW, Piasecka J, et al. Metarhizium anisopliae Pathogenesis of Mosquito Larvae: A Verdict of Accidental Death. PloS One. 2013;8(12):e81686. doi: 10.1371/journal.pone.0081686. pmid:24349111

Greenfield BP, Lord AM, Dudley E, Butt TM. Conidia of the insect pathogenic fungus, Metarhizium anisopliae, fail to adhere to mosquito larval cuticle. Royal Society Open Science. 2014;1(2):140193. doi: 10.1098/rsos.140193. pmid:26064542

Boucias D, Pendland J. Ultrastructural studies on the fungus, Nomuraea rileyi, infecting the velvetbean caterpillar, Anticarsia gemmatalis. Journal of Invertebrate Pathology. 1982;39(3):338-45. doi: 10.1016/0022-2011(82)90058-1

Pendland JC, Boucias DG. Phagocytosis of lectin-opsonized fungal cells and endocytosis of the ligand by insect Spodoptera exigua granular hemocytes: an ultrastructural and immunocytochemical study. Cell and Tissue Research. 1996;285(1):57-67. doi: 10.1007/s004410050620

Holder DJ, Keyhani NO. Adhesion of the entomopathogenic fungus Beauveria (Cordyceps) bassiana to substrata. Applied and Environmental Microbiology. 2005;71(9):5260-6. pmid:16151112 doi: 10.1128/aem.71.9.5260-5266.2005

Doss RP. Composition and enzymatic activity of the extracellular matrix secreted by germlings of Botrytis cinerea. Applied and Environmental Microbiology. 1999;65(2):404-8. pmid:9925560

Boucias DG, Pendland JC. Attachment of mycopathogens to cuticle. The fungal spore and disease initiation in plants and animals: Springer; 1991. p. 101-27

Barelli L, Padilla-Guerrero IE, Bidochka MJ. Differential expression of insect and plant specific adhesin genes, Mad1 and Mad2, in Metarhizium robertsii. Fungal Biology. 2011;115(11):1174-85. doi: 10.1016/j.funbio.2011.08.003. pmid:22036295

Wang C, Leger RJS. The Metarhizium anisopliae perilipin homolog MPL1 regulates lipid metabolism, appressorial turgor pressure, and virulence. Journal of Biological Chemistry. 2007;282(29):21110-5. pmid:17526497 doi: 10.1074/jbc.m609592200

Ikeda K, Kitagawa H, Meguro H, Inoue K, Park P, Shimoi S. The Role of the Extracellular Matrix (ECM) in Phytopathogenic Fungi: A Potential Target for Disease Control: INTECH Open Access Publisher; 2012

Pendland J, Hung S, Boucias D. Evasion of host defense by in vivo-produced protoplast-like cells of the insect mycopathogen Beauveria bassiana. Journal of Bacteriology. 1993;175(18):5962-9. pmid:8376342

Wanchoo A, Lewis MW, Keyhani NO. Lectin mapping reveals stage-specific display of surface carbohydrates in in vitro and haemolymph-derived cells of the entomopathogenic fungus Beauveria bassiana. Microbiology. 2009;155(9):3121-33. doi: 10.1099/mic.0.029157-0

Gottar M, Gobert V, Matskevich AA, Reichhart J-M, Wang C, Butt TM, et al. Dual detection of fungal infections in Drosophila via recognition of glucans and sensing of virulence factors. Cell. 2006;127(7):1425-37. pmid:17190605 doi: 10.1016/j.cell.2006.10.046

Leger RJS, Butt TM, Goettel MS, Staples RC, Roberts DW. Productionin vitro of appressoria by the entomopathogenic fungus Metarhizium anisopliae. Experimental Mycology. 1989;13(3):274-88. doi: 10.1016/0147-5975(89)90049-2

Robinson R. Studies on penetration of insect integument by fungi. International Journal of Pest Management B. 1966;12(2-3):131-42. doi: 10.1080/05331846609438426

Magalhães BP, Butt TM, Humber RA, Shields EJ, Roberts DW. Formation of appressoria in vitro by the entomopathogenic fungus Zoophthora radicans (Zygomycetes: Entomophthorales). Journal of Invertebrate pathology. 1990;55(2):284-8. doi: 10.1016/0022-2011(90)90065-e

Butt T, Coates C, Dubovskiy I, Ratcliffe N. Entomopathogenic Fungi: New Insights into Host-Pathogen Interactions. Advances in Genetics. 2016. doi: 10.1016/bs.adgen.2016.01.006

Fang W, Bidochka MJ. Expression of genes involved in germination, conidiogenesis and pathogenesis in Metarhizium anisopliae using quantitative real-time RT-PCR. Mycological research. 2006;110(10):1165-71. doi: 10.1016/j.mycres.2006.04.014

Screen S, Bailey A, Charnley K, Cooper R, Clarkson J. Isolation of a nitrogen response regulator gene (nrr1) from Metarhizium anisopliae. Gene. 1998;221(1):17-24. pmid:9852945 doi: 10.1016/s0378-1119(98)00430-2

Adamek L. Submerse cultivation of the fungus Metarhizium anisopliae (Metsch.). Folia Microbiologica (Prah). 1963;10:255-7. doi: 10.1007/bf02875956

Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2- [delta][delta]CT method. Methods. 2001;25(4):402-8. pmid:11846609 doi: 10.1006/meth.2001.1262

Eastwood DC, Mead A, Sergeant MJ, Burton KS. Statistical modelling of transcript profiles of differentially regulated genes. BMC Molecular Biology. 2008;9(1):66. doi: 10.1186/1471-2199-9-66

Morgan GA, Leech NL, Gloeckner GW, Barrett KC. IBM SPSS for introductory statistics: Use and interpretation: Routledge; 2012

RCoreTeam. R: A language and environment for statistical computing. R Foundation for Statistical Computing Vienna, Austria2015. Available from: http://www.R-project.org/