Lack of resistance development in Bemisia tabaci to Isaria fumosorosea after multiple generations of selection

The emergence of insecticide resistant insect pests is of significant concern worldwide. The whitefly, Bemisia tabaci, is an important agricultural pest and has shown incredible resilience developing resistance to a number of chemical pesticides. Entomopathogenic fungi such as Isaria fumosorosea offer an attractive alternative to chemical pesticides for insect control, and this fungus has been shown to be an effective pathogen of B. tabaci. Little is known concerning the potential for the development of resistance to I. fumosorosea by B. tabaci. Five generations of successive survivors of B. tabaci infected by I. fumosorosea were assayed with I. fumosorosea. No significant differences in susceptibility to I. fumosorosea, number of ovarioles, or ovipostioning were seen between any of the generations tested. Effects of I. fumosorosea and cell-free ethyl acetate fractions derived from the fungus on the B. tabaci fat body, ovary, and vitellogenin were also investigated. These data revealed significant deformation and degradation of ovary tissues and associated vitellogenin by the fungal mycelium as well as by cell-free ethyl acetate fungal extracts. These data indicate the lack of the emergence of resistance to I. fumosorosea under the conditions tested and demonstrate invasion of the insect reproductive tissues during fungal infection.


Results
Susceptibility of B. tacabi selected under I. fumosorosea pressure to I. fumosorosea. B. tabaci was selected under I. fumosorosea pressure for five generations, labeled F1 to F5, via iterative exposure of surviving insects to the fungal agent as detailed in the Methods section. The susceptibility of the F1, F3, and F5 generations to various concentrations of I. fumosorosea conidial suspensions was examined as described in the Methods section (Fig. 1). No significant differences were seen in the time course or mean cumulative mortalities of the different B. tabaci generations to either a low (10 4 conidia/ml) or high dose (10 7 conidia/ml) of the fungal agent ( Table 1). The mean lethal time to kill (LT 50 ) values for the high/low dose was 5.5 ± 0.5 d/28.1 ± 8.0 d, 6.3 ± 0.5 d/29.7 ± 7.8 d, and 6.4 ± 0.5 d/26.9 ± 6.8 d, for the F1, F3, and F5 B. tabaci generations, respectively, and were not significantly different.  (1 × 10 4 and 1 × 10 7 conidia/ml) were dissected and cultured on PDA. Fungal colonies could readily and continuously be isolate from dissected ovaries of adult at different ages cultured on PDA (Fig. 3A,B and C, images for 1 × 10 7 conidia/ml infection shown). Microscopic observations revealed distortion and deformation of infected ovaries that was not seen in control B. tabaci ovaries ( Fig. 3E and F). Degradation and almost complete loss of vitellogenin was also noted (Fig. 3D). Quantification of the number of ovarioles produced in the F1, F3, and F5 fungal-selected B. tabaci generations after infection with low (10 4 conidia/ml) and high doses (10 7 conidia/ml) of I. fumosorosea revealed (1) no significant effects with respect to infection, i.e. infection by either low or high doses of I. fumosorosea did not change the number of ovarioles seen as compared to control, and (2) no significant differences between the F1, F3, and F5 B. tabaci generations, the concentration of fungal inoculate used, or untreated controls (Table 3).
In order to further examine the effect of fungal growth on the host reproductive system, dissected adult reproductive tissues were treated with fungal conidial suspensions. Dissected fat bodies and ovaries from healthy B. tabaci adults could readily support growth of I. fumosorosea in vitro. In the presence of dissected B. tabaci ovaries, I. fumorosea rapidly grew forming extended hyphae and emerging mycelia within 12 h of co-incubation (Fig. 4a,b). In contrast, little to no growth was seen in fungi kept in the buffer solution alone (Fig. 4c), and no microbial growth was evident in untreated dissected ovaries (Fig. 4i,l). Similarly, dissected fat bodies provided a good growth substrate for the fungus (Fig. 4d,e). Degradation of vitellogenin was also evident (Fig. 4g,h) in in vitro incubations of dissected tissues incubated with I. fumosorosea, as well as infection of dissected eggs (Fig. 4j,k,m,n).   54 . In order to examine whether these extracts can directly affect B. tabaci immune and reproductive tissues, healthy dissected B. tabaci fat bodies and ovaries were incubated in the presence of the I. fumosorosea cell-free EthOAc extracts and examined visually over time. Degradation of dissected B. tabaci fat bodies was evident as early as 6 h post treatment with clear disruption of the tissue seen by 24 h (Fig. 5A,B), whereas control untreated dissected fat bodies remained intact. Similarly, eggs treated with the cell-free I. fumosorosea extracts were significantly damaged and loss of vitellogenin was evident at 24 h post treatment, whereas control eggs were unaffected (Fig. 5C,D).

Discussion
An important consideration is the development of insect resistance to the agents employed for their control. To date, there are no reports of the development of resistance to insect pathogenic fungi used in biological control efforts in the field, although it is well known that some insects (even between closely related species) display     higher intrinsic resistance to fungal infection than others 45 . Greenhouse and field trials using B. bassiana to target the emerald ash borer, Agrilus planipennis indicated sublethal effects that included decreased longevity and egg laying, and increased larval development periods 55 . However, in several instances although mortality by these fungal agents can occur, little to no sublethal effects have been observed 56 , or else field conditions have been shown to have the potential to mask or eliminate any significant sublethal effects 57 .
A number of reports have examined the sublethal effects of entomopathogenic fungi on various insects under laboratory conditions. These can impact the insect behavior, i.e. result in reduced feeding 58,59 , behavioral fever 60 , or can affect developmental programs including molting and egg laying 24,61,62 . Exposure to entomopathogenic fungi can sometimes impact the subsequent generation, i.e. progeny that survive fungal exposure. Reduced reproductive fitness of the progeny of Frankliniella occidentalis exposed to B. bassiana has been noted 63 , and Aedes aegypti larvae surviving infection by the narrow host range fungal pathogen Leptolegnia chapmanii were disrupted in reproductive success, laying fewer eggs 64 . Reports concerning selection for resistance to entomopathogenic fungi have been mixed. Galleria mellonella larvae appeared to develop only minimal resistance to B. bassiana even after 25 generations of selection 48 . However, although significant sublethal effects were seen in surviving generations of B. tabaci exposed to B. bassiana, a gradual reduction in mortality rates was seen in 1 st to 3 rd generations 41 . Here we observed little to no changes in susceptibility of B. tabaci to either low or high doses of I fumosorosea after up to 5 generations of selection. In addition, similar numbers of eggs were laid by each respective generation, and they remained equally affected by fungal exposure, i.e. a significant decrease in egg laying was noted after I. fumosorosea exposure, however, the effect was similar across the various B. tabaci generations examined. These data suggest that in our experimental set-up, no significant resistance development was detected in B. tabaci to I. fumosorosea infection after several generations under fungal selection. These observations may also be linked to the observed lack of resistance development in that although these tissues were affected during fungal infection, no significant differences were seen across the fungal-selected whitefly generations, suggesting no carry-over effects on fitness with respect to reproduction that might lead to the development of resistance.
Although subtle developmental effects on resistance were not examined in this study, fungal infection was shown to reach the fat body and ovaries of insect adult, resulting in loss of vitellogenin and damage to eggs, an effect readily apparent in in vitro assays where these tissues provided a readily utilizable nutrient source supporting fungal growth. In B. tabaci, the number of ovarioles per adult varies between 8 and 18 depending on the time (days) after eclosion, with an average of 13-15 within 12 days 65 . Examination of dissected ovaries revealed no significant differences in the total number of ovarioles between uninfected and I. fumosorosea treated insects, as well as no differences were seen across the B. tabaci generations (1 st to 5 th ) whether or not they were infected by I. fumosorosea.
As cell-free culture supernatants of I. fumosorosea have been shown to be insecticidal 54 , we also investigated any effects these extracts may have on dissected B. tabaci fat body and ovaries. Our data showed that these cell-free extracts have toxigenic activity on immune and reproductive tissues, suggesting a means by which both lethal and sub-lethal effects can be exerted in biological control. A number of lethal and sub-lethal effects of cell-free culture supernatants derived from entomopathogenic fungi have been noted. Complex outcomes were reported in the application of cell-fee M. anisopliae culture supernatant extracts towards the Mediteranean fruit fly, Ceratitis capitata 66,67 . Aside from direct toxicity (death), female fly fecundity was reduced upon initial exposure, but little to no effects were seen with respect to the egg fertility or mortality of larvae, although pupae were affected. Here we show that the extracts can result in direct damage to tissues in the absence of the presence of the fungal agent itself. Damage to the fat body, ovaries, and degradation of vitellogenin was observed in vitro, indicating the presence of toxic compounds in the culture supernatant. Although further characterization of the extract is warranted, a number of potential candidate compounds exist. Dipicolinic acid (2,6-pyridine dicarboxylic acid, DPA) is known to be an inhibitor of the prophenoloxidase activation 38 and I. fumosorosea produces abundant amounts of DPA 68 . A wide range of potential insecticidal toxic metabolites have been reported in M. anisopliae and B. bassiana [69][70][71] however work in I. fumosorosea has lagged behind. Our data confirm the potential for I. fumosorosea as an agent for B. tabaci control as part of Integrated Pest Management practices and indicate that subsequent generations are likely to continue to be susceptible to the fungus.

Methods
Fungal strain and insect maintenance. Strain PF01-N10 of I. fumosorosea (CCTCC No. M207088) was originally isolated from a B. tabaci nymph 27 . For routine use, I. fumosorosea was grown on potato dextrose agar (PDA) and conidia were prepared as described 24 . Conidia of I. fumosorosea were counted in a Fuchs-Rosenthal hemocytometer using a compound microscope and adjusted to indicate spore suspensions (10 4 -10 7 conidia/ml) in 0.05% Tween-80 water. Spore viability was examined by spreading 0.2 ml of the 1 × 10 4 conidia/ml suspension on PDA and counting the number of germinated cells after 24 h of incubation at room temperature. Cells were considered germinated/viable if the germ tube length was as long as the width of the conidia. Conidial viability was assessed for each batch of cells and only batches estimated to be > 95% viable were considered used in experiments. B. tabaci was originally collected from plant of Brassica campestris L. in Guangzhou and then maintained on the same plant in a greenhouse. Identity of the insect was confirmed by PCR-restriction fragment-length polymorphism analysis and mtCOI sequencing as described 72 . The identified mtCOI sequence was identical to the GenBank sequence accession no. GQ332577. Second instar B. tabaci were reared and prepared as described 24 for use in fungal virulence bioassays. Plants of B. campestris L. were grown in plastic pots and incubated in an artificial climate room at 26 ± 2 °C. Sufficient slow release fertilizer (N/P/K = 13:7:15) was added as required to maintain normal plant growth. Intact plants were maintained in greenhouse and used in this experiment at the six to eight leaf stages with 12 to 18 cm tall. Preparation of fungal cell-free culture supernatant ethyl acetate extracts. Fungal conidia (10 ml, 1 × 10 7 conidia ml −1 ) of I. fumosorosea were inoculated into shake cultures in a 1 L flask containing 300 ml of Czapek-Dox broth supplemented with 1% peptone (CZP) and incubated with aeration (180 rpm) at 26 ± 1 °C for 3 d for the production of seed inoculum. The seed inoculum was added to fresh CZP at a 1:9 ratio (v/v, 3 L total volume) and the mixture was incubated with aeration (200 rpm) at 26 ± 1 °C for an additional 6 d after which fungal cells were removed by centrifugation (12000 × g, 15 min) and the cell-free culture supernatant was stored at 4 °C. Metabolites were extracted from the cell-free culture supernatant using ethyl acetate (EthOAc). The cell-free supernatant was mixed with an equal volume of ethyl acetate (1:1, 6 L total final volume) and mixed vigorously for 30 min. The organic phase was collected and concentrated by rotary evaporation (RE − 52A, Shanghai Ya Rong Biochemical Instrument Factory, Shanghai, China) under reduced pressure and then stored at − 20 °C for use.
Insect bioassays. Isaria fumosorosea insect bioassays were performed using standard methods as described 24 . Newly molted 2 nd instars of B. tabaci were treated by dipping infested leaves (not excised leaves) into indicated concentrations of I. fumosorosea (0, 10 4 and 10 7 conidia/ml) for 10 seconds. Each treatment (each concentration) involved at least four leaves with > 50 whitefly nymphs per leaf and the entire experiment repeated three times with new batches of insects and new conidial suspensions. Within experiment treatments were performed at the same time, using randomized groups of insects from a single batch. Plants with treated insects/ leaves were placed in an air-conditioned room at 26 ± 2 °C, R.H.> 85%, L:D = 14:10 h. Treatments were monitored daily until death or new adult emergence, with B. tabaci mortality recorded every 24 h after treatment. Dead insects were removed immediately upon detection and placed separately in a clear Petri dish to allow for fungal sporulation on the cadavers. If the sporulation of I. fumosorosea was observed, the insect was considered to have been killed as a result of infection by I. fumosorosea.

Selection of B. tabaci under I. fumosorosea pressure.
Newly emerged individual adults (zero generation) were maintained for 1 d and allowed to mate. Mated adults were allowed to lay eggs on B. campestris plant leaves (not excised leaves). Eggs were allowed to hatch and the nymph was considered as the 1 st generation after selective pressure (only second instar nymph treated with 10 4 and 10 7 conidia/ml of I. fumosorosea). Insects from the 1 st generation were used in experiments re-treated as above to yield the second generation, and the experiment was continued for 5 generations. The nymph mortality of the F1, F3 and F5 generations to various concentrations of I. fumosorosea conidial suspensions was examined.
Measurement effect of I. fumosorosea on B. tabaci ovipositioning. B. tabaci newly emergence adults, derived from nymphs topically treated with indicated concentrations (0, 1 × 10 4 , 1 × 10 7 conidia/ml) of I. fumosorosea conidia as described in the insect bioassays section, were placed in cages (60 × 60 × 60 cm) and allowed to mate for ~1 day. Mated B. tabaci pairs were placed separately in a plastic Petri dishes (Ø 9 cm diameter), containing a leaf disk (8 cm diameter) of B. campestris L. placed on sterile 2% water agar. Insects (cages and Petri dishes) were kept in an air-conditioned room at 26 ± 2 °C, R.H. > 85%, L:D = 14:10 h. Egg production was recorded at 2 d intervals till death of the adult. Fresh leaf discs and bee honeydew were provided every 2 d to maximize ovipositioning and as a food supplement. Each treatment was repeated three times, for each repetition there were total 12-20 pairs of adult (selected at random from the mating adult pairs emergence at different day).
Microscopic visualization of the B. tabaci reproductive system. Newly emergence B. tabaci adults, derived from nymph treated as described above, i.e. topically with 0, 1 × 10 4 , 1 × 10 7 conidia/ml I. fumosorosea, were placed in cages (60 × 60 × 60 cm), and allowed to mate. The cages were kept in an air-conditioned room at 26 ± 2 °C, R.H. > 85%, L : D = 14 : 10 h. Mated insects were removed at 2, 5 and 8 d post-treatments and their ovaries dissected in PBS buffer solution. The number of ovarioles were observed under a microscope and recorded. Ovary morphology and vitellogenin were observed microscopically. Isolated ovaries were then washed 3 times with sterilize distilled water (ddH 2 O) and placed on Petri dishes containing PDA in order to observe any fungal outgrowth from infected tissues. Each treatment was repeated three times, and 10-15 ovaries were examined per replicate.
In separate experiments, the fat body and ovaries were dissected from un-infected B. tabaci adults and placed in a small Petri dish (Ø 3 cm) containing 15 ml of 5 × 10 4 conidia/ml of freshly harvested I. fumorososea conidia in sterilize 0.05% Tween-80 water. Ovaries placed in 15 ml sterilize 0.05% Tween-80 water were used as one control, and conidia (5.0 × 10 4 conidia/ml in sterile 0.05% Tween-80) placed in a Petri dish were used as another control. The morphology of the fat bodies, ovaries, and associated vitellogenin were observed every 6 h with a microscope. To observe growing hyphae and mycelia, select samples were stained with cotton blue dye (cotton blue 0.05 g, lactic acid 20 ml, phenol 20 g, glycerol 40 ml, ddH 2 O water 20 ml). Each treatment was repeated three times, for each repetition there were total 20 ovaries of adult selected at random from the different treatments to observe.
Fat bodies, ovaries, and associated vitellogenin dissected from un-infected B. tabaci were also treated with I. fumosorosea ethylacetate fractions (as isolated above). The evaporated I. fumosorosea EthOAc fraction was re-suspended in acetone at a concentration of 200 mg/mL and subsequently diluted to a working solution of 10 mg/mL sterilize 0.05% Tween-80. Dissected tissues (ovary), washed with ddH 2 O were placed in a small Petri dish (Ø 3 cm) containing 15 ml of the EthOAc extract working solution. Control samples included isolated tissues placed in 0.05% Tween-80:5% acetone. The morphology of the fat bodies, ovaries, and vitellogenin were observed every 6 h with a microscope. Data analyses. Mortality and ovipositing data were analyzed by using one-way analysis of variance (ANOVA). Mean values were compared by Turkey's student range test (Tukey's HSD, a = 0.05) 73 .