The interaction between cuticle free fatty acids (FFAs) of the cockroaches Blattella germanica and Blatta orientalis and hydrolases produced by the entomopathogenic fungus Conidiobolus coronatus

The interactions between entomopathogenic fungi and insects serve a classic example of a co-evolutionary arms race between pathogens and their target host. The cuticle, site of the first contact between insects and entomopathogenic fungus, is an important defensive barrier against pathogens. It is covered by a layer of lipids that appears to play a key role in these processes and cuticular free fatty acid (FFA) profiles are consider as a determinant of susceptibility, or resistance, to fungal infections. These profiles are species-specific. The cockroaches Blattella germanica (Blattodea: Blattidae) and Blatta orientalis (Blattodea: Ectobiidae) are unsusceptible to the soil fungus Conidiobolus coronatus (Entomophthorales: Ancylistaceae) infection, therefore we studied the profiles of FFAs in order to understand the defensive capabilities of the cockroaches. The fungus was cultivated for three weeks in minimal medium. Cell-free filtrate was obtained, assayed for elastase, N-acetylglucosaminidase, chitobiosidase and lipase activity, and then used for in vitro hydrolysis of the cuticle from wings and thoraces of adults and oothecae. The amounts of amino acids, N-glucosamine and FFAs released from the hydrolysed cuticle samples were measured after eight hours of incubation. The FFA profiles of the cuticle of adults, and the wings, thoraces and oothecae of both species were established using GC-MS and the results were correlated with the effectiveness of fungal proteases, chitinases and lipases in the hydrolyzation of cuticle samples. Positive correlations would suggest the existence of compounds used by the fungus as nutrients, whereas negative correlations may indicate that these compounds could be engaged in insect defence.

Introduction Insect populations are regulated in part by the activity of entomopathogens. Entomopathogenic fungi are proposed as an eco-friendly alternative to chemical insecticides and as model organisms to study insect infection [1][2][3][4]. Unlike bacteria or viruses, fungi infect insects by direct penetration of the cuticle, followed by multiplication in the hemocoel [5].
Infection by entomopathogenic fungi is a multi-stage process comprising adhesion of fungal spores to the insect cuticle, germination and the penetration of invasive hyphae into the host body, hyphae propagation inside the hemocoel and colonization of the host internal organs, followed by the release of toxic secondary metabolites, which might result in host death [6]. The fungus penetrates the insect cuticle by a combination of mechanical pressure from growing hyphae and the enzymatic degradation of the proteins, chitin and lipids comprising the cuticle: proteases are produced first, followed by chitinases and lipases [7,8].
Two key factors influencing the infection process are the structure and composition of the host exoskeleton, and the efficiency of the immune system. Since the cuticle is the first point of contact between the insect and fungus, it is the first and most decisive defence mechanism in insects, and its composition varies greatly according to the species and the developmental stage [9][10][11][12][13]. This complex structure is covered by a waxy layer rich in lipids which play a key role in resistance to entomopathogenic fungi [8,14]. However, although many cuticular lipids have antimicrobial properties, other stimulate the germination process, growth and virulence of fungi; and variations in lipid profiles between species are reflected in differential susceptibility to infection [15][16][17][18][19][20][21] The fungal proteases, chitinases and lipases used to degrade cuticle components play crucial roles in the infection process and are known to act in a coordinated fashion [5,[22][23][24]. Some cuticular proteins display protease inhibition, and protect the insect by suppressing conidial germination and penetration [25,26]. Although no lipase and chitinase inhibitors have been identified in the cuticle so far, several natural chitinase and lipase inhibitors, mostly of microbial origin, have been described [9,27]. Further studies might bring more information on the presence of substances tempering the activity of fungal chitinases and lipases in the insect cuticle.
Previous studies on four medically-important fly species (Lucilia sericata, Calliphora vicina, Calliphora vomitoria and Musca domestica) identified correlations between the efficiency of cuticle digestion by fungal enzymes and the content of cuticular free fatty acids (FFAs), free fatty acid methyl esters (FAME), fatty alcohols, n-alkanes, sterols and several non-typical compounds [28].
The German cockroach (B. germanica), and the oriental cockroach (B. orientalis) are two of the most common species of cockroaches worldwide. They usually reside in human habitats, where they act as hosts for parasites, viruses, bacteria and pathogenic fungi and can cause severe allergic reactions in humans [29][30][31][32]. As these insects are difficult to eradicate, due to their high rates of reproduction and resistance to commonly-used pesticides, biological control strategies based on the use of entomopathogenic fungi are becoming an increasingly desirable option [33][34][35].
The aim of the present work was to identify any relationships between the cuticular FFA profiles of two cockroach species, B. orientalis and B. germanica, and the efficiency of fungal enzymes in hydrolysing the insect cuticle.

Susceptibility of cockroaches to fungal infection
Exposure of B. orientalis and B. germanica imagines and oothecae to sporulating C. coronatus colonies showed high resistance of both cockroach species to fungal infection. No infection or mortality was observed in either the control or fungus-threated groups of B. orientalis. Mortality of fungus treated B. germanica was very low and comparable to the control groups (Table 1 and S1 Table).

Hydrolysis of cuticular proteins by C. coronatus enzymes
The effectiveness of fungal proteolytic enzymes in the culture medium during the 3 weeks of the C. coronatus in vitro cultivation was measured as the amounts of amino acids released from insect cuticle preparations. The greatest amounts of amino acids were produced during enzymatic digestion of B. germanica oothecae (978.29±45.49 μM/mg cuticle), and the least (127.86±52.69 μM/mg cuticle F (5, 12) = 14.37, p = 0.0001) for B. orientalis oothecae. Higher concentrations of amino acids were released from the thoraces, wings and imago of B. germanica than B. orientalis. Also 7.7-times more amino acids were released from B. germanica oothecae than B. orientalis. Results are given in Fig 1 and

Hydrolysis of cuticular chitin by C. coronatus enzymes
The effectiveness of hydrolysis by the C. coronatus chitinolytic enzymes was found to be similar in all samples, measured as the concentration of N-glucosamine (Fig 2 and S2 Table). The highest levels of N-glucosamine were observed for B. orientalis wings (66.70±0.80 μM/mg cuticle), and the lowest (43.49±3.80 μM/mg cuticle; F (5,12) = 4.98, p = 0.0106) for B. orientalis thoraces. The insects were exposed to sporulating C. coronatus colonies as described in Materials and methods. The susceptibility to fungal infection is expressed as percentage of mortality in tested populations.

GC-MS analyses of cuticular FFAs
In almost all cases, significantly higher cuticular FFAs were extracted from B. germanica than B. orientalis: whole body extracts: 2.6 vs.  (Table 3 and S3 Table).
Four FFAs predominated in all analysed internal extracts: C16:0, C18:2, C18:1 and C18:0 (Table 4)  Free fatty acids released during eight hours of incubation are presented as mean ± standard deviation μm/mg of cuticle from wings, thoraces and oothecae of the two cockroach species (for raw data see S2 Table: body. For the oothecae, B. germanica had a similar profile, while for B. orientalis, all FFAs, except C14:0 and C20:4, were more abundant in extracts I and II than extract III. Regarding the wings, higher concentrations of all FFAs were found in the cuticle for both species, apart from C20:3 for B. germanica. Regarding the thoraces, higher concentrations of FFAs were present in the combined extracts I and II for B. germanica (except for C7:0 and C24:0), while all FFAs were more abundant in extract III for B. orientalis (Tables 3 and 4).

Discussion
Although chemical pesticides are among the most popular methods of controlling cockroach infestations, their disadvantages have spurred the search for new strategies, including the use of entomopathogenic fungi [34,37,38]. C. coronatus is a cosmopolitan soil fungus that selectively attacks various insect species [39]. Our findings indicate that B. orientalis and B. germanica are not susceptible to infection by C. coronatus, but not to infection by other entomopathogenic fungi, such as Metarhizium anisopliae, Beauveria bassiana and Purpureocillium lilacinum [35,[40][41][42] The mycelia of C. coronatus cultivated in vitro secrete a plethora of enzymes, however, the activities of fungal enzymes measured in vitro are not necessarily correlated with their importance in the infection process occurring in nature [15,[43][44][45][46]. The enzymatic cocktail released by C. coronatus mycelia degrades cuticle samples from susceptible insects far more effectively than those from resistant species and/or developmental stages [11,28,47]. Similar differences were observed in the present study for B. orientalis and B. germanica, particularly regarding the digestion of oothecae proteins by fungal proteases; this might indicate higher levels of total protein in B. germanica oothecae than in B. orientalis, or of proteins susceptible to digestion by C. coronatus proteases. 13 C -NMR spectroscopy revealed higher levels of proteins in the cuticle of B. germanica oothecae than for B. orientalis [48]; this may be due to the higher protein requirement of developing nymphs [49] and/or differences in their physiology: B. orientalis females deposit oothecae as soon as they are formed while B. germanica females retain the oothecae until nymphs are ready to hatch.
In contrast to the oothecae, the two cockroach species released similar, low amounts of amino acids, suggesting their protein content was low in wings and had similar protein compositions. C. coronatus proteases have been found to be highly effective against the wings proteins of four fly species (L. sericata, C. vicina, C. vomitoria, M. domestica) and those of G. mellonella [28,47]. It could suggest a lower content of degradable proteins in the cockroach wings compared to other insects we have studied in terms of efficiency of cuticular protein digestion by C. coronatus proteases. Similar high concentrations of amino acids were released from the digested thoraces of B. orientalis and B. germanica, suggesting a high abundance of similar proteins. Taken together, our findings suggest that the protein composition of the cuticle varies considerably across the body of the insect.
In contrast, no significant differences were found in the effectiveness of chitin hydrolysis in all samples of both cockroach species, indicating no species-specific variation and similar spatial distribution of chitin in the bodies. However, the C. coronatus chitinolytic enzymes  Table 5 demonstrated greater efficiency against both cockroach species compared to four fly species [28] and the wax moth [47] suggesting higher levels of chitin in cockroach cuticles. N-glucosamine was released from oothecae incubated with C. coronatus enzymatic cocktail containing chitinases, thus confirming the presence of chitin. It has long been assumed that chitin was absent from oothecae [50], however, this belief has been challenged by recent studies [48,[51][52][53]. The lipases present in the C. coronatus enzyme cocktail demonstrated less hydrolytic activity against the cuticle samples than the proteases and chitinases; FFAs were only released from the oothecae of both species and thoraces of B. orientalis. Similar results have been noted against the previously described four fly species and G. mellonella [28,47]. It appears that in C. coronatus, lipases play a lesser role in the development of an infection to that of proteases and chitinases, in contrast with the pivotal role of lipolytic activity during M. anisopliae infection [27].
The cuticular and internal FFAs identified in this work are similar to those previously identified for B. orientalis and B. germanica [38,55,57]. Slight discrepancies in the presence and quantity of individual FFAs result from variation in the use of GC-MS instruments, extraction    Correlation coefficients (r) are presented in brackets. BO-B. orientalis; BG-B. germanica, NDC-not detected, NDT-not determined � data concerning compounds' effects on the in vitro growth, sporulation and virulence of C. coronatus are from [15] and derivatization procedures, and from the different starting materials: we used both males and females pooled together, while Paszkiewicz et al. examined B. orientalis females and B. germanica males only. Most previously examined insect species indicate higher abundance of FFAs in the internal lipids than in cuticular lipids [16,18,36]; however, Chorthippus brunneus appears to be an exception, as are the present results [58]. Higher amounts of FFAs were found in wings and thoraces (g -1 of tissue) than in the whole body of adults (g -1 insect body); this could be due to the high number of these light body parts (B. orientalis: 539 wings and 274 thoraces; B. germanica 377 wings and 186 thoraces) required to extract sufficient amounts of lipids for GC-MS analyses. B. orientalis display a clear wing dimorphism; the present study used equal amounts of reduced and leathery female wings and longer, membranous male wings.
Species specific differences in cuticular FFA profiles were found between cockroach species: C4:0 and C5:0 was present only in B. germanica, while C19:1 was found only in the thoraces of B. orientalis indicating an uneven spatial distribution. The physiological functions of these FFAs in cockroaches remain unknown. While C18:3 was found solely in the oothecae of both species, its exact role is unknown; however, it is likely to protect against fungal attack as C18:3 inhibits C. coronatus growth and the growth and germination of B. bassiana and Paecilomyces fumosoroseus [21,59]. The origin of C13:0, detected only in B. orientalis oothecae remains obscure. The same applies to C14:1 and C15:1, found only in B. germanica oothecae. C13:0, C14:1 and C15:1 have demonstrated antifungal activity against C. coronatus and several pathogenic fungi [15,60]. Eleven FFAs (C6:0, C7:0, C8:0, C9:0, C10:0, C12:0, C16:0, C18:1, C18:2, C18:3 and C20:0) known to inhibit key factors determining the ability of C. coronatus to infect insects, i.e. hyphal growth, sporulation and virulence [15], were found in the oothecae of both species, indicating multiple investments in protecting cockroach eggs and developing offspring. The cockroach ootheca is formed from the secretions of two colleterial glands containing proteins, enzymes and catechol derivatives [53]. The method of delivery of lipids to the ootheca is poorly understood [61], and the presence and amount of each cuticular FFA is the result of a number of poorly-understood processes of synthesis, degradation and distribution in the insect body and transportation to the target sites [28] The efficiency in degrading cockroach cuticle samples by C. coronatus proteases was found to be negatively correlated with concentrations of C6:0, C9:0, C10:0, C16:0 and C20:0. This suggests that these FFAs may play a protective role against fungal assault. However, this inference is weakened by the positive correlations found between fungal protease efficiency and concentrations of C6:0 in the wings, C12:0 and C20:3 in the thoraces, and C17:0 in the oothecae.
In the case of C. coronatus, the role of chitinases is even more complex, as both negative and positive correlations were found regarding the same FFAs, but these differed according to body part. Obviously, more experiments are necessary to demonstrate the impact of each FFA detected in cockroach cuticle on the activity of fungal enzymes engaged in the initial stage of fungal attack.
The present study partly elucidates the mechanisms underlying the non-susceptibility of two species of cockroaches, B. germanica and B. orientalis to fungal infection and highlights the role of FFAs in that process. Further studies on the role played by cuticular lipids in the interaction between the invading fungus and the insect host will shed greater light on the complexity of the infection process.

Insects
B. orientalis and B. germanica were cultured in the laboratory at 25˚C, 70% relative humidity (RH), and a 12:12-hour photoperiod. The insects were cultured on standard rodent food (Agropol, Poland). For cuticle preparations, both adults and oothecae were used.

Entomopathogenic fungus
The entomopathogenic fungus was C. coronatus (isolate no. 3491), originally isolated from Dendrolaelaps spp. (Mesostigmata: Digamasellidae), obtained from the collection of Professor Bałazy (Polish Academy of Sciences, Research Centre for Agricultural and Forest Environment, Poznań, Poland). The fungus was maintained in 90 mm Petri dishes at 20˚C in a 12:12-hour light/dark cycle to stimulate sporulation [62] on Sabouraud agar medium (SAM). The medium was supplemented by homogenized G. mellonella larvae to a final concentration of 10% wet weight. This addition enhances sporulation and virulence of the SAM cultures of C. coronatus. At seven days, conidia were harvested by flooding the plates with sterile water; 100μL portions of suspension, each containing approximately 50 conidia, were taken for inoculations.
To obtain the mixture of fungal enzymes to hydrolyze the insect cuticle, C. coronatus was cultivated at 20˚C in 500-ml Erlenmeyer flasks containing 250ml of minimal medium as described by Bania and co-workers but without shaking [43]. After three weeks, the mycelia were removed by filtration through Whatman no. 1 filter paper. The cell-free filtrates were assayed for their protein concentrations and protease, chitinase and lipase activities, and taken for in vitro hydrolysis of cockroach cuticle preparations.
The same C. coronatus cell-free filtrate was used in studies of cuticle hydrolysis in four medically-important fly species and Galleria mellonella [28,47].

Infection of insects with C. coronatus
B. orientalis and B. germanica adults were exposed for 24 hours at 20˚C to fully-grown and sporulating C. coronatus colonies, around 10 per Petri dish. Controls were exposed for 24 hours to sterile Sabouraud agar medium. After exposure, the insects were transferred to new, clean Petri dishes with appropriate food, and observed for seven days. Oothecae were exposed in the same way within 24 hours of being laid by the females. The effectiveness of fungus penetration into the oothecae and their impact on developing insects was measured as the percentage of larvae that were dead within three days of hatching.

Cuticle preparation
Frozen adults of B. orientalis and B. germanica were briefly (5-10 min) rinsed in tap water and then thoroughly dried with a paper towel. The wings were dissected, and the remnants of the muscles were removed. The cuticles were dissected from thoraces in 10 mM ice-cold Tris-HCl buffer (pH 7.0) and carefully cleaned of remnants of fat body, muscles and other tissues. Empty oothecae were cleaned inside to remove the remnants left by eggs and hatching larvae. All prepared cuticle pieces were washed three times in 10mM ice-cold Tris-HCl buffer (pH 7.0), allowed to dry on ice-cold towels and stored at −20˚C until use.

Enzymatic assays
Elastase, N-acetylglucosaminidase (NAGase), chitinase and lipase activity were measured in C. coronatus cell-free filtrates according to Boguś and co-workers [28]. Measurements were taken spectrophotometrically and spectrofluorimetrically (BioTek Synergy HT, USA) in 96-well polystyrene plates using suitable synthetic substrates (Merck, Germany). Elastolytic activity was measured using N-succinyl-alanine-alanine-proline-leucine-p-nitroanilide in 100mM Tris-HCl buffer containing 20mM CaCl 2 (pH 8.0). The reactions were performed in plate wells containing 2 μl of cell-free filtrate comprising fungal enzymes, 0.5mM final substrate concentration, and reaction buffer to a final volume of 200 μl. The reaction was started by the addition of the substrate, and readings were taken at A 410 to create a progress curve.

Determination of protein concentration
The protein concentration of the cell-free filtrate of C. coronatus was determined with the Bio-Rad Protein Assay (USA), according to Bradford. Briefly, an acidic dye (Coomassie Brilliant Blue) was added to the protein solution, and the absorbance was measured at 595 nm with a microplate reader. Absorbances were measured using BioTek Synergy HT. Bovine serum albumin (BSA) was used as the standard.

Hydrolysis of insect cuticle incubated with cell-free filtrate of C. coronatus
The insect cuticle samples were divided into 50 mg portions, ground in liquid nitrogen and then washed four times in 10 mM Tris-HCl buffer (pH 7.0); 10 mg of ground cuticle was suspended in 1ml of the 10mM Tris-HCl buffer (pH 7.0), 800 μl of which was mixed with 228 μl of the C. coronatus cell-free filtrate containing elastase, NAGase, chitobiosidase and lipase. The reaction mixture was incubated for eight hours at 30˚C. The reaction cocktail was divided into 20 μl portions and immediately frozen to stop further hydrolysis. Two negative controls were added, one consisting of reaction buffer with 1 mg of cuticle but without the cell-free C. coronatus filtrate (C1), and the other consisting of buffer with C. coronatus filtrate but without the insect cuticle (C2). The free amino acids produced by hydrolysis of the cuticle by proteases were measured according to Adler-Nissen, with some modifications [63]. The samples and the controls were mixed with 0.1% picrylsulfonic acid (Merck, Germany) and read at A 340 . The absorbance of the negative controls was subtracted from the samples. The amounts of N-glucosamine released by chitinase hydrolysis were measured using the D-glucosamine Assay Kit (Megazyme, Ireland) according to the producer's manual. The concentrations of free fatty acids (FFAs) released by lipases were determined with the use of the EnzymChrom TM Free Fatty Acid Assay Kit (BioAssay Systems, USA). Three independent replications of all procedures were performed. The hydrolytic efficiency of the fungal enzymes was calculated per mg of cuticle. No determination of cuticle protein, chitin and lipid content was not performed due to due to the amounts of insect-derived material being insufficient.

Extraction of free fatty acids (FFAs)
Cuticular and internal lipid components of insects were extracted, separated and analysed by GC-MS. Whole adults, oothecae, wings and thoraces isolated from adults (mass in Table 2) were extracted first in 20 ml of petroleum ether (Merck, Germany) for 5 min (extract I) and then again in 20 ml of dichloromethane (Merck, Germany) for 5 min (extract II) to yield cuticular lipids. The insects and cuticle preparations were sonicated with dichloromethane to produce Extract III containing internal lipids. The extracts were placed in glass flasks and evaporated under nitrogen.

Derivatization method
Trimethylsilyl esters (TMS) of FFAs were obtained by adding 100 μl of a BSTFA: TMCS mixture (99:1) (Merck, Germany) to 1 mg of sample and heating for 1h at 100˚C. The TMS of fatty acids were then analysed by GC-MS.