Reviewers' comments: Reviewer #1 (Remarks to the Author): Cui et al identified four miRNA-like small RNAs from the entomopathogenic fungus, Beauveria bassiana using small RNA sequencing produced during infection of Anopheles stephensi

Cui et al identified four miRNA-like small RNAs from the entomopathogenic fungus, Beauveria bassiana using small RNA sequencing produced during infection of Anopheles stephensi mosquitoes. They showed that production of one of them, bba-miRI-1, was dependent on dicer2, but not the others. Knockout or overexpression of bba-miRI-1, but not the other three, affected fungal infection. The authors showed that bba-miRI-1 is highly expressed at 36 hours postinfection that coincides with cuticle penetration during which it targets the transcripts of Spazle, an inducer of the Toll pathway, therefore suppressing the expression of antifungal effector molecules. However, later in infection when the fungus gains entrance into the hemocoel, the production levels of bba-miRI-1 is significantly reduced to avoid induction of a clip domain protease (another target of bba-miRI-1) that activates the prophenoloxidase enzyme and melanisation. Therefore, the timing of expression of bba-miRI-1 aligns well with the two different stages of infection. This is a very interesting work and adds to our understanding of the role of small non-coding RNAs in interaction of pathogens and hosts. The investigators have done a thorough study and generally the manuscript is written well, although there are several issues with the usage of the article “the” which is either missing or not needed in many places. In addition, a couple of validations are missing listed below.

deriving from a B. bassiana isolated from Bombyx mori. The target site proposed, and convincingly shown with mutagenesis studies is indeed present in the plant TLR gene Xa21 as shown in this paper (see figure 3 nucleotides 176-182 here: http://www.plantcell.org/content/plantcell/9/8/1279.full.pdf). This is particularly relevant as B. bassiana is known to colonize plants symbiotically. Taking this broader phylogenetic perspective is not unreasonable and may reveal broader implications for this paper and the milRNAs under investigation.
Reviewer #3 (Remarks to the Author): The manuscript entitled "The fungal pathogen deploys a small silencing RNA to attenuate mosquito immunity and facilitate infection" by Cui et al. submitted to Nature Communications describes the role of one microRNA-like RNA (milRNA) of the fungus Beauveria bassiana in insect infection. The manuscript is well written and it follow a logical development trying to solve the questions that show up through the reading. It provides evidences that the fungus synthesizes milRNAs and the conclusions are attractive. The role of small RNAs in pathogenesis has been described in fungi that infect plants, but this is the first manuscript describing their possible role on the infection of insects. Therefore, the subject of the manuscript is very novel and it would be of interest of a wide research community. In addition, the authors have combined different techniques and the quality of the data and presentation is high. However, the conclusions about the role of a particular milRNA (milR1) in facilitating fungal infection are overstated and more experiments are needed to support them. In addition, the results of some routine experiments should be shown to verify the genotype of some mutants generated in this work. The main and minor issues concerning this manuscript are listed below.
Major concerns -The Identification of milRs expressed during mosquito infection. The shown information is scarce. It would be convenient to indicate some details about the types and features of the fungal small RNAs identified in the samples. What sample contained the studied 4 milRNAs? Are these milRNAs expressed by the fungus in axenic cultures? -Experiments that analyzed mutants (Figure 1, Figure 4 and Figure 5) should have been done with two independent mutants or complemented strains to avoid effects due to ectopic mutations that can occur in transformation or spontaneously. -One of the most critical result corresponds to figure 1. The difference between milR1-KO and WT strain are very small, although they are statically significant. This suggests that if milR1 has some effect in infection it is very small and all possible precautions should be considered in any experiment.
-Experiments to demonstrate binding of milR to mosquito Ago need a control experiment. Ago proteins are relatively well conserved, so the authors have to demonstrate that the used antibody is not able to immunoprecipitate B. bassiana Ago protein.
-People from plants (Fungal small RNAs suppress plant immunity by hijacking host RNA interference pathways. Weiberg A, Wang M, Lin FM, Zhao H, Zhang Z, Kaloshian I, Huang HD, Jin H. Science. 2013 342:118-23) created an Ago mutant to confirm that the effect of Botrytis cinerea small RNAs was mediated by the RNAi mechanism of the plant. They also analyzed the virulence of Dicer mutants of the fungus, which could be done in this manuscript because there are available.
-Explain why the interaction between bba-milR1 and the predicted target genes was assayed in HEK293T instead of insect cells. In addition, the results in figure 3 suggest that human Ago proteins behave in the same way as insect Ago proteins regarding the control of Spz4 and CLIPB9 mRNA levels, which is very surprising, particularly in the case of CLIPB9. Moreover, negative effects of RNAi in expression are well known but positive ones deserve some comments in the manuscript.
-PO activity in figure 5 should include milR1-OV and milR1-KO strains.
-Lines 223-225. "Here, we show that the fungal pathogen transfers a sRNA effector into host insect cells to attenuate host immunity and achieve its infection" This sentence is very blunt and should be rephrased. Insect cells in culture and in the insect are able to uptake naked milR1 but none experiment demonstrates that the fungus secretes milR1 when it is infecting insects. -Experiments (Southern or PCR) should be added to supplementary material confirming the deletion of dcl1, dcl2, milR1, milR2, milR3 and milR4.
-Consider change the title because it gives the idea that there is only one fungal species that infect insects.
Minor points: -Many figure legends repeat the conclusion of the experiments without give clues to understand the experiments, requiring surfing in the materials and methods section. Some missed information in legend is stated below, but not all, and the author should look at the legends in detail. We thank all reviewers for the evaluation of our manuscript and for the constructive comments. We have revised the manuscript taking into account the reviewers' comments. Our point-by-point responses follow. The reviewers' comments are quoted in bold and our responses follow in plain text. Response: These comments are much appreciated. We have carefully corrected the inappropriate usage of "the" in the manuscript.

1) Validation of bba-milR1 knock out hasn't been shown.
Response: Validation of bba-milR1 knock out and other mutants have been included in Supplementary Figure 2. Response: Thanks for the reviewer's insightful suggestion. We have showed that the host miRNAs ast-miR10-5p and ast-2940-3p could be detected in uninfected samples (Fig. 2c).

Line 1: The fungal pathogen should be A fungal pathogen
Response: Thank you for your correction. We have changed "the" to "A".

Line 186 and 213: the statistical significance hasn't been shown in the text or on the figure.
Response: The statistical significance has been added in Fig. 4f and Fig. 5g.

Line 367: … (OMEGA) and used as template to….
Response: Thank you for correction. We have changed "(OMEGA), were used as template to" to "(OMEGA) and used as the template to".

Line 379: correct the spelling of fungal
Response: "fugal" has been changed to "fungal".

Line 391 and 417: fatbody should be fat body
Response: "fatbody" and "fatbodies" in new Lines 394, 420 and Supplementary Fig. 8 have been changed to "fat body" or "fat bodies". Similarly, identification of targets of a specific sRNA based on sequence alignment search is also not accurate because we found many predicted targets of bba-milR1 are not the real targets (shown in Supplementary Fig. 7).
Although we have analyzed conservation of bba-milR1 target sites in Spz4  Response: We appreciate your comments. In the previous bioassays, we used very high concentration of fungal conidial suspension (10 8 conidia/mL), which resulted in small difference between milR1-KO and WT. We have conducted additional bioassays using lower concentration of fungal conidial suspension (10 7 conidia/mL), and it turns out that there is very big difference in the virulence between milR1-KO and WT (Fig 1b). We have also showed that B. bassiana Dicer mutants Dcl1/ Dcl2 resulted in decrease in fungal virulence against A. stephensi compared to WT (see figure below). One possible explanation could be that Dcl1/ Dcl2 mutants don't generate bba-milR1 ( as shown in Fig 1a in manuscript), thereby Dcl1/ Dcl2 could not repress mosquito Spz4 expression during infection.

-Experiments to demonstrate binding of milR to mosquito
However, we could not rule out another possibility that knock-out of Dicers affect other unknown milRNAs that might positively or negatively modulate fungal virulence. Based on the above considerations, it may not be suitable to evaluate a specific milRNA's function by analyzing the virulence of Dicer mutants of the fungus against insects.

Effect of Dicers on fungal virulence. Survival of adult female A. stephensi
mosquitoes infected with the wild-type (WT) ARSEF252 and two Dcl1/ Dcl2 mutant strains following topical application of a spore suspension (10 7 conidia/ml).
Mosquitoes sprayed with sterile 0.01% Triton X-100 were used as a negative control (Triton).  . 2011, 31;108(22):9250-5.). Our studies also showed that bba-milR1 activates the expression of CLIPB9, but the exact mechanism is unknown. Response: We have rephrased the sentence in new Line 226 to "Here, we

Explain why the interaction between bba
show that the fungal pathogen produces a sRNA effector that is translocated into host insect cells to attenuate host immunity and achieve its infection". In RIP assay, A. stephensi Ago1 antibody specifically pulled down bba-milR1 during infection progress, but not in samples mixed A. stephensi and B.
bassiana, which shows that the fungus produces bba-milR1 when it is infecting insects.
Response: Thanks for your suggestions. We have added the results confirming the deletion of dcl1, dcl2, milR1, milR2, milR3 and milR4 in Supplementary Figures 2d-j. 11. -Consider change the title because it gives the idea that there is only one fungal species that infect insects.
Response: We have changed "the" to "A". Response: Thanks for your suggestions. We have added the information to figure legends. For example, we added "RNA was extracted from fungal mycelium" to figure 1a legend.
The revised manuscript "The fungal pathogen deploys a small silencing RNA to attenuate mosquito immunity and facilitate infection" by Cui et al. has been improved according to the reviewers' criticisms. Moreover, I acknowledge that the authors have provided appropriate responses to all the issues raised by reviewers. However, there are some minor points that should be addressed.