Conidia of the insect pathogenic fungus, Metarhizium anisopliae, fail to adhere to mosquito larval cuticle

Adhesion of conidia of the insect pathogenic fungus, Metarhizium anisopliae, to the arthropod host cuticle initially involves hydrophobic forces followed by consolidation facilitated by the action of extracellular enzymes and secretion of mucilage. Gene expression analysis and atomic force microscopy were used to directly quantify recognition and adhesion between single conidia of M. anisopliae and the cuticle of the aquatic larval stage of Aedes aegypti and a representative terrestrial host, Tenebrio molitor. Gene expression data indicated recognition by the pathogen of both hosts; however, the forces for adhesion to the mosquito were approximately five times lower than those observed for Tenebrio. Although weak forces were recorded in response to Aedes, Metarhizium was unable to consolidate firm attachment. An analysis of the cuticular composition revealed an absence of long-chain hydrocarbons in Aedes larvae which are thought to be required for fungal development on host cuticle. This study provides, to our knowledge, the first evidence that Metarhizium does not form firm attachment to Ae. aegypti larvae in situ, therefore preventing the normal route of invasion and pathogenesis from occuring.


Introduction
Metarhizium anisopliae is a widespread, soil-borne fungal pathogen of insects, ticks and mites [1][2][3] with much potential as an environmentally friendly alternative to conventional chemical pesticides for the control of pests of socio-economic importance [3]. assessed at 260 and 280 nm absorbance using a Nanodrop 2000 (Thermoscientific). Total RNA (1 µg) was reverse transcribed using the QuantiTect Reverse Transcription kit (Qiagen) with gDNA elimination reaction, for the experiment to quantify fungus-derived transcripts.

qPCR assay for expression of adhesin genes
Relative cDNA quantity was analysed by using the qPCR method previously described [29]. The same proctocol was followed, using 18S rRNA and elongation factor tEF as reference genes. Transcript levels were determined using the CFX96 Real-Time PCR detection system (Biorad). PCR reactions were performed in 10 µl volumes consisting of 1 µM of each primer, 2 µl of cDNA sample, 5 µl SYBR Green Fastmix (Quanta) and 1 µl diethylpyrocarbonate-treated water. All reactions were carried out in duplicate. 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 s, 60.3 • C for 10 s and 72 • C for 30 s. A dissociation step of 65-95 • C over 5 s was used for melt curve analysis for detection of non-specific products in the reaction. cDNA from each sample was pooled and diluted twofold, to serve as a standard curve and was included in each run to generate a standard curve for each primer set.
Bio Rad CFX Manager software was used to determine the cycle threshold of each sample (C t -Target) which was normalized to the geometric mean cycle threshold (C t ) of the appropriate endogenous qPCR products (C t -Control) for the same sample [31][32][33]. Relative gene expression was calculated using the comparative C t method (2-

Insect cuticle preparation
Aedes aegypti (strain AeAe) were reared at room temperature (22 • C ± 2 • C) in distilled water to late (L 3−4 ) instar and fed Tetramin fish food. Tenebrio molitor larvae were maintained at 27 • C (±2 • C) on wheat bran. Aedes aegypti larvae were removed from water and dried on Whatman No. 1 filter paper before flash freezing under liquid nitrogen. Tenebrio molitor larvae were washed in distilled water and flash frozen under liquid nitrogen before cutting the abdominal cuticle into 5 × 10 mm sections. Five mosquito and five T. molitor cuticles were subsequently fixed, with a small amount of expoxy resin adhesive, to a glass slide.

Spore probe preparation
Metarhizium anisopliae isolate ARSEF 4556 spores were harvested from cultures maintained on broken basmati rice [35]. Individual fungal spores were immobilized at the end of a V-shaped tipless AFM cantilever (Thermomicroscopes; electronic supplementary material, figure S1). Fungal spores were dusted onto a glass slide, previously cleaned with ethanol. Glue (Homebase glass glue) was applied to the AFM cantilever and a single conidium fixed to the AFM cantilever using the AFM itself [36][37][38]. To confirm attachment, the spore colloid probe was visualized under a VHX-600 Keyence microscope and later imaged with a scanning electron microscope.

Atomic force microscopy
A JPK nanowizard II AFM (JPK instruments, Berlin) was used to measure the adhesion forces between the spore probe and sample, by the vertical deflection of the cantilever. Cantilever deflection is measured as a change in the reflected laser path, resulting in a cantilever deflection-displacement curve (figure 1). Prior to force measurements, the cantilever was calibrated, to convert the deflection into a quantitative force measurement. The spring constant and sensitivity of cantilever were measured directly, by lowering the cantilever onto a clean glass surface submerged in deionized water, generating a spring constant of 0.26 Nm −1 and sensitivity of 53 nm V −1 , these values were used in subsequent force calculations following Hooke's Law. Experiments were carried out in a fluid cell containing deionized water. Measurements were made on 121 and 115 different areas of the cuticle of Tenebrio and Aedes, respectively. Each measurement was replicated three to five times before moving to a new location on the cuticle. Measurements were recorded across five cuticles from Aedes and five from Tenebrio. All forces curves were normalized resulting in tip deflection of 0 nm where there was no interaction. Once in contact the tip remains in contact as the separation between cantilever and sample decreases, causing a deflection of the cantilever (c). As the cantilever is retracted the tip remains in contact (d), owing to adhesion and the cantilever is deflected down (e). Eventually, the force on the cantilever is enough to overcome adhesion and the tip breaks free, returning to a neutral position (f).

Epicuticular wax extraction
All insects were washed in deionized water and dried at 30 • C (±2 • C) prior to solvent extraction. Six replicates of five Ae. aegypti or two T. molitor larvae were washed in hexane for 10 s before discarding, the hexane fraction was subsequently dried under nitrogen gas before resuspending in 1 ml transesterification solution: HCl/MeOH (5% v/v) containing 0.01% butylated hydroxyltoluene (w/v). Samples were heated to 90 • C for 90 min and allowed to cool to room temperature before incubating in 2 ml hexane with gentle agitation for 30 min. The hexane extract was collected and the sample extracted with hexane a further three times. All hexane fractions were pooled before drying under nitrogen gas. Heptadecanoic acid (C17:0) was added prior to extraction to act as an internal standard. Dried samples were stored at −80 • C until required.

Gas chromatography-mass spectrometry
Prior to gas chromatography-mass spectrometry (GC-MS) analysis, samples were resuspended in hexane and 1 µl of each sample injected onto a Thermo TR-5 ms SQC gas chromatography (GC) column (15 m × 0.25 mm × 0.25 µm) (Themo Scientific, UK) which was eluted by a Trace GC Ultra (Thermo Scientific, UK). There was a 0.5 s pre-injection dwell time and a single wash with 1 µl chloroform followed by one pre-injection rinse. The temperature gradient for elution used an initial temperature of 50 • C which was held for 1 min and then increased at 30 • C per minute up to a final temperature of 300 • C which was held for 5 min before returning to starting conditions. The eluent from the GC column was analysed by a DSQ II mass spectrometer which was used in positive ionization mode and a mass range within the XCALIBUR software (Thermo Scientific, UK) and identifications confirmed by probability score and manual inspection of the experimental and theoretical data.

Statistical analysis
Comparison of adhesion force measurements between samples was analysed using a Kruskal-Wallis H-test. Pairwise comparisons were performed using Dunn's post-hoc with a Bonferroni correction for multiple comparisons. GC-MS data was analysed using a Mann-Whitney U-test and molecular datasets were analysed using two-way analysis of variance (ANOVA) with Tukey HSD post-test. Prior to analysis, gene expression data was subjected to a Box-Cox transformation, conforming to ANOVA assumption of homogeneity of variance [33]. All statistical analyses were carried out using MATLAB R2013b and Rv. 3.0.2 [39].

Force measurements
Prior to force measurements the topography of a limited number of Aedes and Tenebrio cuticles was investigated using an AFM image scan (electronic supplementary material, figure S2). The surface of Aedes presented a rougher, more convoluted surface than Tenebrio, providing a greater surface area for attachment [40,41]. Contrary to previous research [42,43], the force curves generated were not reflective of a rougher surface topography. The force interactions between the abdominal cuticle segments and an immobilized spore were determined using AFM in order to assess whether the elevated expression of the adhesion genes allowed for effective attachment of the spore to the cuticle, compared with non-attachment to an inert glass surface. Figure 3a shows typical force-distance curves, with a single adhesion event, of M. anisopliae to glass, Aedes and Tenebrio.
Tenebrio approach curves show a large jump to contact peak, indicative of strong attractive forces, occuring at a distance of approximately 0.05 µm ( figure 3a). Typically, the cantilever snaps to the sample surface if the attractive force on the probe reaches or exceeds the value of the spring constant. In  contrast to Tenebrio, force-distance curves measured between glass and Aedes show no notable attractive interactions within the approach curve (figure 3a).
Similar forces were measured between the negative control glass (n = 100) and Aedes (n = 115) (figure 2a). Forces generated on Tenebrio (n = 121) cuticle were an order of magnitude greater than those of Aedes (χ 2 2 = 65.366, p ≤ 0.001). Indeed, the adhesion forces measured for Tenebrio were fivefold greater than the control sample, or Aedes (figure 3b; Mdn = 25.855, p ≤ 0.0001). There was no significant difference in forces measured between glass and Aedes (figure 3b; Mdn = 4.1505, 2.9980, respectively, p = 0.130) illustrating an absence of specific attachment to the cuticle of the mosquito.
Retraction curves demonstrated interesting differences; Tenebrio force-distance curves characteristically showed one adhesive peak over a large area with a sharp, de-adhesion to a neutral position, after a distance of approximately 1 µm. In contrast to this, retraction curves with multiple peaks and constant force plateaus at lower adhesion forces of approximately 0.5 nN were measured for Aedes, compared to weakest forces recorded in Tenebrio being between 15 and 20 nN. This indicates the possibility of other short-range molecular interactions, which are much weaker than those formed with Tenebrio (data not shown).

Epicuticular wax analysis
In order to determine whether the chemical make-up of the cuticle itself may-at least partially-be a contributing factor with regard to successful attachment or not of the spore to the different tested insects, the majority lipid composition of cuticle was analysed by GC-MS analysis (electronic supplementary material, figure S3). Seven hydrocarbons were identified whose abundance was similar across the two insects ( figure 4). The Tenebrio cuticle contained the long-chain hydrocarbons, tetramethylheptadecane and 9-methylnonadecane, whereas the Aedes cuticle did not (figure 4; U = 23.00, z = 2.263, p = 0.031 and U = 25.00, z = 2.785, p = 0.008, respectively).

Discussion
Adhesion of entomopathogenic fungi to host cuticle is considered to involve an initial binding of the conidia to the cuticle, largely through weaker hydrophobic forces, followed by consolidation and firm attachment [11,14]. This study found that there is an upregulation of Mad genes in response to exposure to Aedes cuticle which further corroborates work by Butt et al. [29] and is indicative of host-pathogen recognition. However, the inability of the fungus to establish firm attachment prevents the typical infection course from proceeding as the AFM data demonstrates that conidia of M. anisopliae fail to firmly adhere to the mosquito larvae cuticle, discrete force steps observed in both Tenebrio and Aedes force curves but absent in the glass force curves, may be attributed to the breaking force of a receptor-ligand pair [44]. While no firm attachment was determined for Aedes, the presence of weak short-range molecular interactions in Aedes further emphasizes recognition as suggested by the adhesion gene upregulation. Formation and strength of attachment is dependent upon the stability of molecular interactions, between the spore surface components and binding sites or receptors in the cuticle, under force [45], suggesting the initial recognition and binding to Aedes is extremely weak and not readily maintained. It has been shown that cells are able to regulate their receptor-ligand interactions, adapting them to the environments they are in [44].
Although preventing adhesion to the cuticle is a rare defence mechanism, it is well documented that hydrocarbon content of the waxy layer can influence fungal pathogenesis, with some compounds altering hydrophobicity or being fungistatic [12,46,47], in addition some insects actively secrete a variety of antimicrobial compounds on the cuticular surface. GC-MS analysis revealed no notable fungistatic compounds within the mosquito cuticle. There was, however, an absence of three long-chain hydrocarbons within the mosquito larval cuticle. For germination to occur, conidia require cuticular nutrients, namely long-chain fatty acids, lipids, sugars and amino acids [11,48], it is therefore feasible that the mosquito larval cuticle is not conducive for fungal development.
Metarhizium has demonstrated the ability to form firm attachment to a terrestrial host cuticle under water, indicating the ability of this fungus to adapt and form associations in an aquatic environment. In the mosquito larvae, the route of infection is more similar to the aquatic fungal pathogen Culicinomyces clavisporus, which is ingested rather than penetrating the cuticle [49], however, the upregulation of adhesin genes suggest there is recognition and an attempt to form attachment to the mosquito larvae. As a consequence, it can only be postulated that the cuticle itself is not providing an adequate substrate to further develop and, therefore, initiate full pathogenesis via traditional means.
Data accessibility. Supporting data are deposited with dryad doi:10.5061/dryad.bg48n and in the electronic supplementary material.