Introduction

Amaranth is a common name for Amaranthus genus of the Amaranthaceae family1. It is inexpensive to grow, incredibly nutritious, and adapts to different environments2. Being highly nutritious and cheap, amaranth is being promoted to the vulnerable people in society, especially pregnant women, children3 and immunocompromised individuals, including people with HIV, to strengthen their immune systems4. Amaranth provides farmers with up to 7,117 USD/ha income5. Despite the importance of this crop, its production is affected by insect pests majorly lepidopteran leaf webber Spoladea recurvalis (Fabricius) (Lepidoptera: Crambidae). Spoladea recurvalis has been reported to be the major amaranth pest in America, Asia and Africa6. The larval stages feed on leaf lamina, leaving only the veins to make a skeleton and also webs the remaining parts to form a shelter7. In Kenya, S. recurvalis has been reported to cause up to 100% loss to farmers if not managed8. To manage this pest, farmers rely on synthetic pesticides use9 and no adequate biological measures are used in the management of amaranth pests10. However, the use of synthetic pesticides on such vegetables, which are quick-maturing and have a short pre-harvest interval, may leave residues that are hazardous to consumers, including humans and animals, and with negative repercussions to the environment and biodiversity11. In this instance, biological control agents like entomopathogens are promising alternatives that are safe to humans and eco-friendly12,13,14. Though this is a feasible alternative, a previous study by Opisa et al.12 reported moderate mortality of S. recurvalis by a Metarhizium anisopliae isolate ICIPE 30 of 58.3% while M. anisopliae ICIPE 18 could only induce 6.7% mortality to second instar larvae.

Efficacy of entomopathogens and host response to biological control agents (BCA) is affected by temperature variations15,16,17, which can limit the use of these entomopathogens for effective pest control18. Temperature affects the ecological suitability of insects by adjusting their physiology19, hence studies on the effects of temperature on BCA and host response are relevant for successful biological control programs. Reaction to cold and pathogen invasion/infection leads to variation in immune responses20,21. Wang et al.22 conducted immunological studies in Spodoptera frugiperda (J.E. Smith) (Lepidoptera: Noctuidae) exposed to M. rileyi and realized that there was impairment of cellular immunity, while Maingi et al.17 also got similar results with Phthorimaea absoluta (Meyrick) (Lepidoptera: Gelechiidae) after exposure to M. anisopliae ICIPE 18 and ICIPE 20 isolates. However, this has not been explored in S. recurvalis. Cooler temperatures may slow the growth of entomopathogens and increase insects’ ability to fight infections23. Insect immune systems and pathogens may behave differently as the temperature varies significantly, impacting biological control programs24. Insects use hemocytes to carry out mechanisms such as encapsulation and nodulation to fight pathogen invasion25; encapsulation involves hemocytes that activate and form multilayered capsules around the pathogens, while nodulation involves a number of cells that aggregate around a pathogen. Granulocytes and plasmatocytes are actively involved in defense against pathogen infection26,27, and these granulocytes are the first hemocytes that initiate the encapsulation process28,29. The gut microbiome also aids the insects in digestion, detoxification, stress tolerance, and immune defense by producing metabolites and signals that inhibit pathogens, activate systemic immunity, and create conditions unfavorable for fungal growth30,31,32. Therefore, this study was carried out to determine temperature-dependent immunological responses and gut bacterial community diversity in S. recurvalis following infection with M. anisopliae fungal isolates (ICIPE 30 and ICIPE 18) to provide information on the optimal conditions for integration of the biopesticide in the pest’s management system.

Results

Mortality of Spoladea recurvalis larvae

Cross-analysis of treatments and temperatures in relation to mortality revealed significant differences for both temperature (χ2 = 167.58, df = 2, p < 0.001) and treatment (χ2 = 871.27, df = 2, p < 0.001). The mortality rates significantly differed in EPF-treated larvae across days at 20 °C (χ2 = 106.227, df = 2, p < 0.001), 25 °C (χ2 = 340.19, df = 2, p < 0.001), and 30 °C (χ2 = 340.19, df = 2, p < 0.001) (Fig. 1). At 20 °C, M. anisopliae ICIPE 30 caused 21.5% mortality, while M. anisopliae ICIPE 18 caused only 6.5% (Fig. 1). As the temperature increased to 25 °C, mortality rates also rose, with M. anisopliae ICIPE 30 causing 57.5% and M. anisopliae ICIPE 18 causing 8.5%. At 30 °C, the highest mortality rates were observed, with M. anisopliae ICIPE 30 achieving 82.5% compared to 17.5% for M. anisopliae ICIPE 18 (Fig. 1). All mycosis tests conducted on the S. recurvalis larvae that died at 20 °C, 25 °C, and 30 °C showed over 95% positivity.

Fig. 1
Fig. 1
Full size image

Mean mortality (%) of S. recurvalis larvae post exposure to M. anisopliae isolates ICIPE 30 and ICIPE 18 at 20, 25 and 30 °C.

Hemocytes of Spoladea recurvalis

Total hemocyte counts (THC)

This study evaluated the effect of M. anisopliae ICIPE 30 and ICIPE 18 isolates on total hemocytes in the hemolymph of S. recurvalis at different temperature regimes. Cross analysis of treatments, temperatures and days in relation to hemocyte density was run and significant differences were observed between temperature regimes (χ2 = 460.49, df = 2, p < 0.001) and there was a strong interaction between temperature, treatments and days of the experiment (χ2 = 29.62, df = 4, p < 0.001). Further analysis was done at each regime and at 20 °C, a significant difference between treatments (χ2 = 208.382, df = 2, p < 0.001) and days (χ2 = 73.196, df = 3, p < 0.001) was exhibited as well as an interaction between days and treatments (χ2 = 16.061, df = 6, p < 0.05) (Fig. 2A). The application of M. anisopliae ICIPE 30 resulted in a significant rise in S. recurvalis’ total hemocytes, from 15.5 × 10⁴/ml on day 1 to 41.25 × 10⁴/ml on day 3, peaking at 48.25 × 10⁴/ml on day 5, before experiencing a slight decline to 37 × 10⁴/ml on day 7. Similarly, larvae treated with M. anisopliae ICIPE 18 exhibited a significant increase in total hemocyte counts, reaching 19 × 10⁴/ml by day 3. However, by days 5 and 7, the mean hemocyte counts in the M. anisopliae ICIPE 18-treated larvae were not significantly different from those in the control group (Fig. 2A).

Fig. 2
Fig. 2
Full size image

Mean Total Hemocyte Counts in hemolymph of Spoladea recurvalis after exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18 at 20 (A), 25 (B) and 30 °C (C).

For bioassays conducted at 25 °C (Fig. 2B), significant differences were observed in larval total hemocytes per treatment (χ² = 422.41, df = 2, p < 0.001), across the days (χ² = 201.63, df = 3, p < 0.001), and there was an interaction between days and treatments (χ² = 92.37, df = 6, p < 0.001). A rapid increase in total hemocytes up to day 3 (96.5 × 10⁴/ml) was noted in larvae treated with M. anisopliae ICIPE 30, followed by a drastic decline on day 5 (62 × 10⁴/ml) and day 7 (40.5 × 10⁴/ml). In larvae treated with M. anisopliae ICIPE 18, there was a statistically significant increase in the total hemocyte count on day 3 (42.75 × 10⁴/ml), day 5 (49.25 × 10⁴/ml), and day 7 (52.25 × 10⁴/ml). However, in control larvae, THC showed an insignificant increase on day 3 (13.5 × 10⁴/ml), day 5 (14.75 × 10⁴/ml), and day 7 (16.75 × 10⁴/ml) (Fig. 2B).

The analysis did not show significant differences at 30 °C (Fig. 2C) among the treatments (χ2 = 4.480, df = 2, p = 0.1065); but there were significant variations across the days (χ2 = 57.458, df = 3, p < 0.001) and in the interaction between days and treatments (χ2 = 35.336, df = 6, p < 0.001). Metarhizium anisopliae ICIPE 30 induced the highest number of hemocyte counts on day 3 (30.25 × 104/ml), but these counts declined drastically to 13.75 × 104/ml on day 5 and 10.75 × 104/ml on day 7. Similarly, an increase in total hemocytes was also observed in larvae treated with M. anisopliae ICIPE 18 on day 3 (19.5 × 104/ml), although the counts were lower compared to M. anisopliae ICIPE 30. The THC remained constant (19.25 × 104/ml) on day 5, but with a slight decline observed at day 7 (18 × 104/ml). However, in control larvae, THC gradually increased from day 3 (13.75 × 104/ml) to day 5 (16.5 × 104/ml) and then to day 7 (17 × 104/ml) (Fig. 2C).

Differential hemocyte count (DHC)

Microscopic observations revealed five types of cells, viz. granulocytes, prohemocytes, plasmatocytes, spherulocytes and oenocytes in the hemolymph of S. recurvalis (Fig. 3). We observed that at 20 °C, total number of granulocytes in S. recurvalis larvae did not show any significant differences among treatments (χ2 = 4.1530, df = 2, p = 0.1254) and days (χ2 = 2.1921, df = 3, p = 0.5335). However, the number of granulocytes in larvae treated with M. anisopliae ICIPE 30 insignificantly increased from day 1 to day 7, with percentages of 59.65%, 59.93%, 64.13%, and 67.42%, respectively (Fig. 4A1). Larvae treated with M. anisopliae ICIPE 18 showed an increased number of granulocytes compared to the control on day 5 (60.75%) and day 7 (61.25%). Granulocytes in the control groups ranged between 57.63% on day 1 to 58.25% on day 7 (Fig. 4A1). The percentage of plasmatocytes also did not differ significantly between treatments (χ2 = 3.3230, df = 2, p = 0.18985), although they varied across the days of the experiment (χ2 = 8.9290, df = 3, p = 0.03025). Plasmatocytes in both larvae treated with M. anisopliae ICIPE 30 and ICIPE 18, and controls, did not show significant changes except on day 7 (15.78%) in ICIPE 30-treated larvae (Fig. 4B1–B3).

Fig. 3
Fig. 3
Full size image

Images of hemocyte types, granulocytes (A), plasmatocytes (B), prohemocytes (C), spherulocytes (D) and oenocytes (E), observed in hemolymph of Spoladea recurvalis larvae post exposed to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18.

Fig. 4
Fig. 4
Full size image

Mean percent granulocytes at 20 (A1), 25 (A2) and 30 °C (A3) and plasmatocytes at 20 (B1), 25 (B2) and 30 °C (B3) observed in hemolymph of Spoladea recurvalis after exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18.

For S. recurvalis larvae which were treated at 25 °C, there was notable changes in total number of granulocytes per treatment (χ2 = 10.0390, df = 2, p < 0.006) and day (χ2 = 9.7400, df = 3, p < 0.02), with a strong interaction between days and treatments (χ2 = 27.7256, df = 6, p < 0.001). Total granulocytes (Fig. 4A1) in larvae exposed to M. anisopliae ICIPE 30 increased sharply from day 1 (74.21%) up to day 5 and 7, where a decline was observed (50.90% and 48.69%, respectively). In M. anisopliae ICIPE 18-treated larvae, granulocytes (Fig. 4A1) increased gradually, with higher values obtained on days 5 (61.48%) and 7 (64.82%). However, in the control groups, granulocytes declined slightly from 59.59% on day 1 to 57.57% on day 7 (Fig. 4A1). Plasmatocytes did not differ for both treatments (χ2 = 3.0554, df = 2, p = 0.2170) and days (χ2 = 5.0748, df = 3, p = 0.1664). The percent number of plasmatocytes did not show any significant changes across treatments and days, except for M. anisopliae ICIPE 30-treated larvae on day 7 (16.55%) (Fig. 4B2).

At 30 °C, the total number of granulocytes (Fig. 4A3) differed between treatments (χ2 = 29.743, df = 2, p < 0.001), days (χ2 = 25.003, df = 3, p < 0.001), and there was an interaction between days and treatments (χ2 = 13.438, df = 6, p = 0.03659). Total granulocytes in larvae exposed to M. anisopliae ICIPE 30 decreased in number from day 1 (53.89%) and continued to decline on days 3, 5, and 7 (48.33%, 40.83%, and 29.47%, respectively). Metarhizium anisopliae ICIPE 18-treated larvae had a gradual reduction in the number of granulocytes, with 58.50% on day 1, 55% on day 3, 51.34% on day 5, and 48.33% on day 7. However, the percent granulocytes in control larvae was also slightly reduced, from 59.24%, 57.27%, 54.79%, and 52.76% on days 1, 3, 5, and 7, respectively, without any significant variations.

The number of plasmatocytes also was significantly different across treatments (χ2 = 16.3538, df = 2, p < 0.001) and days (χ2 = 11.4586, df = 3, p < 0.01) but did not show any interaction between days and treatments (χ2 = 7.4801, df = 6, p = 0.28). Total number of plasmatocytes in M. anisopliae ICIPE 30-treated larvae declined from day 1 (22.5%) to day 7 (14.74%) (Fig. 4B3). The same trend was observed in M. anisopliae ICIPE 18-treated larvae, but the percentages were higher than in M. anisopliae ICIPE 30-treated larvae, with 25.94% on day 1, 25.59% on day 3, 18.06% on day 5, and 16.67% on day 7. Control groups changed in the proportion of plasmatocytes from day 1 (24.95%) to day 5 (26.53%) and a slight decrease on day 7 (24.69%), but without any significant variations.

Microbiome of Spoladea recurvalis

The 16 S rRNA sequencing produced a total of 2,330,328 reads and the analysis revealed 1,119 bacteria genera in S. recurvalis gut. Their abundance per sample depended on the treatment, temperature and days post-infection. The most abundant bacterial genera were Enterobacter, Enterococcus, and Klebsiella. All bacterial genera with abundance below 0.5% were merged into others (Fig. 5).

Fig. 5
Fig. 5
Full size image

Relative abundance of bacteria genera in Spoladea recurvalis gut after exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18 at 20 °C (A), 25 °C (B) and 30 °C (C). All bacteria genera below 0.5% abundance were grouped into others.

Enterococcus was abundant in larvae treated with M. anisopliae isolates ICIPE 30 and ICIPE 18 compared to controls. However, the abundance of Enterococcus was observed to decrease over time in fungal-isolate-exposed larvae. On the other hand, the abundance of Enterococcus in control larvae increased over time for all days and at various temperatures. The highest abundance of Enterococcus spp. (34.85%) was observed in larvae treated with M. anisopliae ICIPE 30 at 20 °C and 30 °C on day 1 (Fig. 5A and C).

Klebsiella abundance increased with time in the control larvae for all temperatures. The percentage abundance was observed rising from day 1 to 7 at 20 °C (18.30% to 21.99%), 25 °C (11.02% to 27.09%) and 30 °C (14.79% to 25.59%) (Fig. 5A–C). Larvae treated with M. anisopliae isolates ICIPE 30 and ICIPE 18 also showed an increase in Klebsiella abundance though their trend irregularly varied; for example, larvae treated with M. anisopliae ICIPE 30 showed a significant decline in Klebsiella abundance at 20 °C (17.22% on day 1 to 15.45 on day 7) and 25 °C (24.07% on day 1 to 19.29% on day 7). However, there was a rapid rise of Klebsiella spp. abundance in larvae treated with M. anisopliae ICIPE 30 at 30 °C from 6.59% on day 1 to 26.85% on day 7 (Fig. 5C).

Enterobacter spp. was abundant with time throughout all temperatures, regardless of the treatments, but there was a rapid rise in its abundance in larvae treated with M. anisopliae ICIPE 30 at 30 °C from 3.88% on day 1 to 16.93% by day 7. The same trend was observed in M. anisopliae ICIPE 18-treated larvae at 30 °C, where Enterobacter spp. abundance increased from 5.70% on day 1 to 13.16% by day 7 (Fig. 5C). The highest and rapid change of Enterobacter spp. abundance was observed in the larvae treated with M. anisopliae ICIPE 30 at 30 °C from 3.88% on day 1 to 16.93% on day 7 (Fig. 5C).

We also observed that Serratia spp. population/community had significant changes where its abundance increased with time, though this increase was slight but significant in larvae treated with M. anisopliae ICIPE 30 at 30 °C from 0.69% on day 1 to 2.16 on day 7 (Fig. 5C). The highest abundance was observed in larvae treated with M. anisopliae ICIPE 18 at 20 °C on day 7 (17.49%) (Fig. 5C), while the larvae that were exposed to M. anisopliae ICIPE 30 at 30 °C had their Serratia spp. abundance reduced to as low as 2.16% by day 7.

Alpha diversity of the bacteria varied depending on treatments, temperatures and days post-infection. It was observed that bacterial diversity reduced with time and temperature, influencing the alpha diversity. True Shannon (Fig. 6) ranged between 1.4 and 3.7 and the highest diversity (Index 3.6) was observed in the control larvae at 25 °C on day 1 (Fig. 6B), while the lowest (Index 2.7) was observed in M. anisopliae ICIPE 30 at 30 °C on day 7 (Fig. 6C). Metarhizium anisopliae ICIPE 30 at 30 °C led to the reduction in bacterial diversity from Index 3.2 on day 1 to 2.7 by day 7 (Fig. 6C).

Fig. 6
Fig. 6
Full size image

Alpha diversity of bacteria (Shannon-Wiener Index) in Spoladea recurvalis gut after exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18 at 20 °C (A), 25 °C (B) and 30 °C (C).

The bacterial richness (Fig. 7A–C) ranged between 420 and 645 species across all treatments, temperatures and days post-infection. Larvae treated with M. anisopliae ICIPE 18 at 30 °C on day 7 showed the highest average species richness (442.3), while the lowest species richness was observed in the control larvae at 25 °C on day 7 (271). Bacterial evenness index (Fig. 7A–F) ranged between 0.4 and 0.6 across all samples. Bacteria distribution was even more (0.57) in control larvae at 25 °C on day 1. The lowest distribution was observed in larvae treated with M. anisopliae ICIPE 30 at 25 °C on day 7 (0.43).

Fig. 7
Fig. 7
Full size image

Bacterial richness at 20 °C (A), 25 °C (B), 30 °C (C) and evenness at 20 °C (D), 25 °C (E), 30 °C (F) in gut of Spoladea recurvalis larvae post exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18.

Beta diversity (Fig. 8) was analyzed using the Jaccard dissimilarity index and the PCoA clustered samples differently per temperature. At 20 °C (Fig. 8A), the highest microbial diversity was observed between M. anisopliae ICIPE 30-treated larvae on day 1 and those treated with M. anisopliae ICIPE 18 on day 1 (68.43%), while the lowest was obtained between M. anisopliae ICIPE 30-treated larvae on day 7 and those in the control on day 1 (21.57%). For 25 °C (Fig. 8B), the highest diversity was recorded between M. anisopliae ICIPE 30-treated larvae on day 1 and day 7 (50.00%), while the lowest was observed between M. anisopliae ICIPE 30-treated larvae on day 7 and the control larvae on day 7 (16.30%). Finally, for larvae treated at 30 °C (Fig. 8C), the highest microbial diversity was observed between M. anisopliae ICIPE 30-treated larvae on day 1 and those treated with M. anisopliae ICIPE 18 on day 7 (68.44%), while the lowest was recorded between M. anisopliae ICIPE 30-treated larvae on day 7 and those in the control on day 7 (13.60%).

Fig. 8
Fig. 8
Full size image

Beta diversity of bacteria in the gut of Spoladea recurvalis larvae post exposure to Metarhizium anisopliae isolates ICIPE 30 and ICIPE 18 at 20 °C (A), 25 °C (B) and 30 °C (C).

Discussion

Regardless of the number of studies conducted on the efficacy of entomopathogens on various insect pests, a number of avenues have remained unexplored18,33,34. In our study, we demonstrate the impact of entomopathogenic fungal isolates M. anisopliae ICIPE 30 and ICIPE 18 on both cellular and gut microbiome immunity of S. recurvalis at varying temperatures.

Cellular-wise, we noted differences in hemocyte counts for each treatment, with the highest hemocyte numbers recorded in larvae treated with M. anisopliae ICIPE 30 at 25 °C. It is likely that larvae were stable at the optimum temperature and initiated strong defense mechanisms upon infection, thus an increase in hemocyte counts12,17. While both isolates triggered an increase in hemocyte numbers across all temperatures and days, a decline was observed starting from day 5 going forward, with the highest decline observed in M. anisopliae ICIPE 30-treated larvae. These differences are relatable to the germination and mycelial growth rates of these isolates since effective germination and mycelial growth are important factors in terms of the pathogenicity and virulence of the entomopathogen after host pest infection35,36. The two fungal isolates have been studied, and it was established that M. anisopliae ICIPE 30 has higher germination and mycelial growth rates than M. anisopliae ICIPE 18 at 20, 25 and 30 °C37,38.

Overall, across all fungal-treated larvae, a rapid rise in hemocyte numbers was recorded between day 1 and day 3 before a decline that continued up to day 7. It is possible that in the first days, larvae imposed a quick response to fight the fungal infection as it has been a defense mechanism of most lepidopteran insects, as reported by Gillespie et al.39. According to this study, circulating hemocytes in hemolymph can change rapidly in response to infection, and this translates to an initial increase in THC followed by a decline after the pathogen gains momentum39. This kind of decline has also been reported by Hung and Boucias40 in Spodoptera exigua (Hübner) (Lepidoptera: Noctuidae) after exposure to Beauveria bassiana. Furthermore, although we observed variations in hemocyte counts among the fungal-treated larvae, the control groups also exhibited slight, though statistically insignificant, changes under the different temperature regimes. However, hemocyte numbers in all control groups consistently remained lower than those in the treatment groups. This phenomenon could be attributed to the normal hemocyte changes that can happen during larval growth and instar transformations, as well as life changes41,42. Our results demonstrated that S. recurvalis immune responses differed depending on the incubation temperatures and the fungal virulence, which is in line with previous studies17,43. Metarhizium anisopliae fungal isolates have been reported to be more effective at 30 °C37,38 and this may also explain the highest decline of hemocytes at 30 °C in this study.

In this study, we considered the identification, classification and counting of different cell types to ascertain the effect of the tested fungal isolates on cell types in the hemolymph of S. recurvalis. Notably, granulocytes and plasmatocytes were the most observed cell types across all the treatments though their total counts varied depending on the fungal isolate, temperature and days after infection. However, we noted that both hemocytes in fungal isolate-treated larvae at 20 °C showed a gradual increase up to day 7. This trend differs from that seen at other temperatures, where hemocyte levels drastically rose but later declined. We attribute this difference to the relatively unfavorable conditions at 20 °C, which may have hindered the fungal isolates from fully establishing early on. Nonetheless, they still triggered an immune response in the larvae, leading to a sustained increase in hemocyte levels38.

On the other hand, the number of granulocytes in M. anisopliae ICIPE 30-treated larvae at 25 °C and 30 °C increased significantly right from day 1 and later started declining from day 3 however, the highest decline was noted at 30 °C by day 7. Insects use hemocytes to carry out mechanisms such as encapsulation and nodulation to fight pathogen invasion25; encapsulation involves hemocytes that activate and form multilayered capsules around the pathogens, while nodulation involves a number of cells that aggregate around a pathogen. Granulocytes and plasmatocytes are actively involved in defense against pathogen infection26,27, and these granulocytes are the first hemocytes that initiate the encapsulation process28,29. It is possible that the increase in percentage granulocytes and plasmatocytes in the larvae was a response to infection with fungal isolates, while a decline might have been due to other processes, such as apoptosis, where cells can be damaged in the process of defense and hemolymph completely colonized44.

The observed changes in total and differential hemocytes prompted us to explore the effect of M. anisopliae isolates ICIPE 18 and ICIPE 30 on the gut microbiome of S. recurvalis, where we identified over 1,000 bacterial genera, indicating a rich and diverse microbial community; consistent with previous findings that lepidopteran guts can harbor a vast number of bacteria, both pathogenic and non-pathogenic45. Overall, we observed low susceptibility to the tested fungal isolates in S. recurvalis larvae that exhibited higher bacterial abundance and diversity. Notably, larvae treated with fungal isolates at 20 °C and 25 °C on days 1 and 7, as well as at 30 °C on day 1, showed both higher bacterial abundance and diversity, which coincided with reduced susceptibility to the fungal isolates. Meanwhile, M. anisopliae ICIPE 30-treated larvae had the lowest bacterial abundance and diversity by day 7 and that might be why we observed the highest mortality. Several studies have reported similar results that are in line with ours, where gut bacteria are implicated to offer strong immunological responses, protecting insects from infections and biological control agents46,47. A study by Dillon and Charnley48 reported a strong correlation between gut bacteria of the desert locust Schistocerca gregaria Forsskål (Orthoptera: Acrididae) and its defense against pathogens. Additionally49, reported that lepidopteran insects use their gut microbiome to protect themselves from infection. Furthermore, for insect survival, it is critical for them to mount immunological responses against fungal infections, as also observed in our current study33.

Some bacterial genera have been reported to have effects on the insect immune responses50; for example, Enterococcus can reduce entomopathogenic fungal spore germination, making the host insects resistant to infection51. In addition, Enterococcus mundtii has been reported to protect Spodoptera littoralis Boisduval (Lepidoptera: Noctuidae) from pathogens through resisting colonization of the gut by these external microbes52,53. Xia et al.54 also reported that some species of Enterobacter can reduce larval mortality in high abundance of these bacteria. Furthermore, most gut bacteria can boost the immunity of their host; for example, Snyman et al.55 noted that Klebsiella could produce specific enzymes that aid in digestion and facilitate good nutrition in Busseola fusca Fuller (Lepidoptera: Noctuidae), and it is known that insects can resist infections if their nutritional state is stable56,57. All these highlighted mechanisms could have contributed to larval resistance to fungal isolates before the gut bacterial community was disrupted by the tested fungal isolates.

Bacteria of the genus Serratia have also been reported to induce the immunity of lepidopteran insects such as Hyphantria cunea Drury (Lepidoptera: Erebidae)58. However, when the insect gut is destroyed, Serratia may shift and develop outside the gut59. This may explain why we observed a reduction in the abundance of Serratia in M. anisopliae ICIPE 30-treated larvae on day 7, indicating the most significant shift of microbiome structures and the highest larval mortality. Our results showed the lowest bacterial richness and evenness in S. recurvalis larvae treated with M. anisopliae ICIPE 30 at 30 °C by day 7. This finding presents evidence of dysbiosis compared to Maingi et al.17, who observed similar results in P. absoluta when challenged with entomopathogenic fungal isolates.

Beta diversity, clustered differently the larvae treated with M. anisopliae ICIPE 30, ICIPE 18 and those in the controls. This result shows that treatments had a significant effect on the gut microbiome structure. The lowest microbial diversity was observed in larvae treated with M. anisopliae ICIPE 30 at 30 °C on day 7, and this suggests that M. anisopliae ICIPE 30 caused the highest shift in S. recurvalis gut microbiome structure. Generally, the gut microbiome was impacted by the entomopathogenic fungal isolates according to our cumulative abundance, beta and alpha diversity. Our results are in line with Maingi et al.17, who also observed gut community disruption in P. absoluta post-infection with entomopathogenic fungi M. anisopliae ICIPE 20 and ICIPE 18.

Conclusion

Our study demonstrated that, at 30 °C, M. anisopliae ICIPE 30 could weaken the microbiome of S. recurvalis and lower cellular immunity. We also noticed that temperature influenced the immune response of the insect as well as the activity of the tested fungal isolates. Furthermore, S. recurvalis has a robust bacterial community that provides immunity to it when it gets infected but this could be degraded at favorable temperatures. An abundant and diverse gut microbiome was responsible for immunity in the S. recurvalis larvae, while dysbiosis did the opposite at 30 °C. Since M. anisopliae ICIPE 30 worked well at higher temperatures, it might be a good biocontrol option in the field and therefore, more research studies are warranted for its development as a biopesticide for sustainable management of S. recurvalis in vegetable cropping systems.

Materials and methods

Spoladea recurvalis colony

Spoladea recurvalis adults and larvae were obtained from amaranth fields in Transmara, Narok county (0°35′32.892″N 3°0′49.14″E) and Yatta, Machakos County (01°08.295′S 037°25.892′E) in May and June 2014 and a colony was established. At the start of this experiment, field collections were frequently done to infuse the colony to avoid inbreeding and here, the moths were maintained in ventilated sleeved Perspex cages (40 × 40 × 45 cm) and fed on 10% honey solution. Potted amaranth plants were placed in the cages for oviposition. After 24 h, plants were removed and relocated to a table and transferred into wooden cages (50 × 50 × 60 cm) until the eggs hatched. Three days later, leaves with larvae were removed from the plants and placed in clear plastic dishes (20 × 15 × 7 cm) lined with paper towel to absorb excess moisture and fine netting infused lid for ventilation. The larvae were daily supplied with fresh amaranth leaves to obtain the second instar larvae required for use in the bioassays. To maintain the colony, the remaining larvae were fed on new amaranth leaves every day until pupation. Using a fine camel hair-brush, the pupae were removed from the clear plastic dishes (20 × 15 × 7 cm) and placed into other clean clear plastic dishes (20 × 15 × 7 cm) and put into Perspex cages (40 × 40 × 45 cm) for adult emergence. The colony was kept at the temperature of 25 ± 2 °C, 50–70% Relative Humidity and 12:12 L: D photoperiod.

Fungal isolates

The entomopathogenic fungi (EPF) isolates, Metarhizium anisopliae ICIPE 30 and ICIPE 18, were obtained from icipe’s Arthropod Germplasm Centre. The isolates were sub cultured on SDA media using streaking method and maintained at 25 ± 2 °C in complete darkness. The method involves use of a fine streaking loop to evenly spread fungal spores on media to make a four quadrant streak lines. To obtain fungal suspensions, conidia were harvested from two-to-three-week-old fully mature plates using a sterile spatula and suspended in 10 ml of distilled water containing 0.05% Triton X- 100 (MERCK KGaA, Darmstadt, Germany) solution. The suspension was vortexed for five minutes at about 700 rpm to obtain a uniform/homogenous spore suspension. The conidial concentration was determined using a new improved Neubauer chamber and a light microscope where a serial dilution was conducted to adjust the suspension to 1 × 108 conidia/ml. Prior to the bioassays, a viability test was conducted where fungal suspensions were adjusted to a concentration of 3 × 106 conidia/ml and 0.1 ml of the suspension was equally spread over SDA plates. Three sterile microscope cover slips were placed on the surface of each plate and incubated at 25 ± 2 °C. After 18 h, conidia germination was examined by counting 100 randomly selected germinated conidia beneath each coverslip under a light microscope. Conidia were considered to have germinated when the length of the germ tube was at least twice the diameter of the conidium60,61. Four replicate plates were used per isolate, and viability of each isolate was determined, where more than 95% of the conidia germinated.

Laboratory experiment

Second instar larvae were exposed to M. anisopliae ICIPE 30 and ICIPE 18 at the temperatures of 15 20, 25 30 and 35 °C. Petri dishes were lined with two pieces of filter paper and 50 s instar larvae per EPF treatment, per temperature regime, were placed on the filter paper using a soft camel brush. These larvae were sprayed with 10 ml at the concentration of 1 × 108 conidia/ml of each of the fungal isolates ICIPE 30 and ICIPE 18 at room temperature using potter precision laboratory spray tower (Burkard Manufacturing Co Ltd, UK). Control groups were sprayed with 0.05% Triton X-100 (MERCK KGaA, Darmstadt, Germany) solution.

Sprayed larvae were transferred into clear plastic dishes (10 cm diameter), well-ventilated by a fine net on the lid lined with sterile paper towel and containing pre-sterilized fresh amaranth leaves. Larvae were kept under the various defined temperatures in MIR-554 cooled incubator (PHC Europe B.V, Netherlands), monitored daily and fed on fresh surface-sterilised amaranth leaves until the end of the experiment. Each set up was replicated four times. Surviving larvae were randomly selected at days 1, 3, 5 and 7 for the studies on immunological responses and the effects of the EPF on the gut microbiota of S. recurvalis. In addition, dead larvae were also recorded at days 1, 3, 5 and 7 to determine the mortality rates at the different temperature regimes. For all mortality observed, mycosis tests were conducted to confirm whether death was due to infection by the fungus used in the bioassays. The insect cadavers were surface sterilized with 70% ethanol, rinsed three times in distilled water, and then kept separately in 9 cm Petri dishes lined with sterile moistened filter paper to allow for fungal outgrowth and verify if mortality could be attributed to the respective fungal isolates they were treated with. Mortality due to fungal infection was confirmed by the presence of hyphae and conidia on the surface of the cadavers.

Hemocyte study

Hemocyte studies included total hemocyte count (THC) and differential hemocyte count (DHC). For THC, selected surviving larvae were put in sterile falcon tubes and placed in ice bucket for 10 min to inactivate them. Inactive larvae were then surface sterilized with 70% ethanol, 1.5% sodium hypochlorite followed by three washes of distilled water. Larvae were placed on sterile paper tower at room temperature for 10 min to allow draining of excess distilled water and returning to normal physiological processes. Larvae were carefully pierced using blood lancet and hemolymph collected using sterile micro-capillary glass tubes. Hemolymph was quantified to have 1 µl for THC and another 1 µl for DHC17. For THC, 1 µl of hemolymph was added in Eppendorf tube containing 9 µl of ice cold sterile anticoagulant buffer (186 mM NaCl, 41 mM citric acid, 98 mM NaOH and 17 mM EDTA), mixed well and placed in ice. The 10 µl mixture was loaded on hemocytometer and total cells counted under light microscope. THC was reported as “Number of cells/ml×104”.

To identify and count different cell types, hemolymph was spread evenly on sterile glass slides and allowed to air dry at room temperature. The cells were fixed by dipping the slide in a fixative made by mixing methanol and glacial acetic acid in 3:1 ratio for about 5 s followed by air drying at room temperature. Staining was done by dipping the dried slides in staining glasses containing 5% Giemsa/PBS (v/v) for 35–45 min. Slides were then gently rinsed with distilled water to remove excess stain and air dried at room temperature. Dry slides were viewed under light microscope (Leica DM 2500 LED, Leica Microsystems Wetzlar Germany) to identify and count the different hemocytes present. For each slide, 100 cells were counted and grouped into different cell types according to62. Images were captured using the Leica LASZ Software, Leica Microsystems Wetzlar Germany.

Gut microbiome study

To establish the effect of the tested fungal isolates on gut microbiota, the insect larvae were aseptically surface sterilized as described above and dissected to remove the gut. DNA was extracted from the gut using ISOLATE II Genomic DNA Kit, (Bioline, London, UK) according to the manufacturer’s procedure. DNA quantity and purity was assessed using Nanodrop 2000 Spectrophotometer (Thermo Fischer Scientific, Wilmington, US), from where samples within A260nm/A280nm ratio of 1.8-2 were barcoded. Library preparation was done using the 16 S Barcoding Kit 1–24 (SQK-16S024, Oxford Nanopore Technologies, UK) following the manufacturer’s protocol. The library was prepared using a 50 µl mix containing; PCR Water (20.7 µl) (New England Biolabs, US), MyTaq buffer (8 µl) (Meridian Biosciences, UK), MyTaq enzyme (0.5 µl) (Meridian Biosciences, UK), 25mM.MgCl2 (0.8 µl) (New England Biolabs, US), 16 S Barcode (10 µl) (Oxford Nanopore Technologies) and DNA template (10 µl).

Amplification was done using Master Cycler Nexus gradient thermal cycler (Eppendorf, Germany) by running 1 cycle for 2 min at 95 °C for initial denaturation, 40 cycles for 30 s at 95 °C for denaturation, 40 cycles for 45 s at 55 °C for annealing, 40 cycles for 1 min at 72 °C for extension, 1 cycle for 10 min at 72 °C for final extension and then was held at 10 °C. Amplicons were then purified using AMPure beads following the ONT protocol and checked for quality and quantity using Nanodrop 2000 Spectrophotometer, sampled and pooled to a 1.5 ml Eppendorf DNA LoBind tube and 1 µl of RAP adapters (Oxford Nanopore Technologies, UK) added. This was mixed gently by flicking and down spinning, incubated at room temperature for 20 min before loading for sequencing.

The prepared library was loaded into a FLO-MIN106 flow cell (Oxford Nanopore Technologies, UK) and sequenced on an ONT MK1C device (Oxford Nanopore Technologies, UK) with live base-calling for 4 h.

Data analysis

Data from THC and larval mortality were subjected to a normality test using the Shapiro-Wilk test63 and then analyzed through the Generalized Linear model with negative binomial regression in the MASS package64. DHC data were analyzed using beta regression implemented in the Betareg package65. GLM served as a unifying framework that generated a structured model object, which was then used in the subsequent ANOVA step to break down the variance and test the significance of specific factors. Post hoc was performed using the estimated marginal means with a Tukey adjustment using the R packages multcomp66 and emmeans67 to compare treatment means at p ≤ 0.05. The analysis examined the effects of explanatory variables: temperature, fungal isolates, and the interaction with time on response variables, including total hemocyte density, hemocyte types, and larval mortality percentage. Model significance was assessed through analysis of deviance using Chi-square tests.

Sequencing data/reads were uploaded to the EPI2ME Desktop agent software (3.7.3)68 to identify classified bacteria. Genera with cumulative abundance below 0.5% were classified as others. Alpha diversity69 statistics, such as the Shannon-Wiener index, richness, abundance, and evenness, were used to evaluate bacterial diversity in each sample. The Jaccard70 dissimilarity index was used to calculate beta diversity among bacterial genera in samples. Finally, using the “vegan"71 package, the inter-sample microbiome relationship was calculated using principal coordinate analysis (PCoA). All the analyses were run in R. Software (4.3.1)72.