The Splicing Factor RNA-Binding Fox 1 Mediates the Cellular Immune Response in Drosophila melanogaster

This information is current as Ashley E. Nazario-Toole, Javier Robalino, Kwame Okrah, of September 27, 2021. Hector Corrada-Bravo, Stephen M. Mount and Louisa P. Wu J Immunol published online 11 July 2018 http://www.jimmunol.org/content/early/2018/07/10/jimmun ol.1800496 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2018/07/10/jimmunol.180049 Material 6.DCSupplemental http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 27, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published July 11, 2018, doi:10.4049/jimmunol.1800496 The Journal of Immunology

The Splicing Factor RNA-Binding Fox Protein 1 Mediates the Cellular Immune Response in Drosophila melanogaster

Ashley E. Nazario-Toole,*,† Javier Robalino,* Kwame Okrah,‡ Hector Corrada-Bravo,‡ Stephen M. Mount,*,‡ and Louisa P. Wu*,†

The uptake and destruction of bacteria by phagocytic cells is an essential defense mechanism in metazoans. To identify novel involved in the phagocytosis of Staphylococcus aureus, a major human pathogen, we assessed the phagocytic capacity of adult blood cells (hemocytes) of the fruit fly, Drosophila melanogaster, by testing several lines of the Drosophila Genetic Reference Panel. Natural genetic variation in the RNA-binding Fox protein 1 (Rbfox1) correlated with low phagocytic capacity in hemocytes, pointing to Rbfox1 as a candidate regulator of phagocytosis. Loss of Rbfox1 resulted in increased expression of the Ig superfamily member Down syndrome adhesion molecule 4 (Dscam4). Silencing of Dscam4 in Rbfox1-depleted blood cells rescued the fly’s

cellular immune response to S. aureus, indicating that downregulation of Dscam4 by Rbfox1 is critical for S. aureus phagocytosis Downloaded from in Drosophila. To our knowledge, this study is the first to demonstrate a link between Rbfox1, Dscam4, and host defense against S. aureus. The Journal of Immunology, 2018, 201: 000–000.

he fruit fly, Drosophila melanogaster, is a powerful ge- In Drosophila, hemocytes play a critical role in host defense netic model for studying host–pathogen interactions. To against bacteria. Adult flies lacking blood cells are unable to T combat infection, the fruit fly solely relies on a robust clear Staphylococcus aureus, a Gram-positive extracellular bac- http://www.jimmunol.org/ innate immune system. The absence of the adaptive immune re- teria (6–8). Similarly, phagocytic blood cells are critical to host sponse in Drosophila facilitates direct in vivo studies of innate defense against S. aureus in humans (8–10). Several hemocyte- immunity. Consequently, studies characterizing the innate defense associated receptors have been shown to be important for reactions of Drosophila have yielded valuable insight into path- S. aureus recognition in the fruit fly: Croquemort (11), scavenger ogen recognition and host defense in higher organisms (1). receptor C1 (12), Eater (13), Nimrod C1 (14), the peptidoglycan The defense reactions of Drosophila include potent humoral and recognition PGRP-SC1A and PGRP-SA (15), Draper cellular responses. Humoral immunity is characterized by the (16), and the integrin heterodimer aPS3/bn (16, 17). With this systemic production of antimicrobial peptides after immune cells large number of receptors, the loss of any given receptor only of the blood and fat body detect bacteria or fungi in the hemo- partially impairs S. aureus uptake. Therefore, the study of phago- by guest on September 27, 2021 lymph. The cellular immune response is carried out by specialized cytosis is complicated by the redundancy of ligand-binding speci- blood cells known as hemocytes, which engulf and eliminate ficities and the overlapping functions of components of the pathogens via phagocytosis (2). During pupal development, phagocytic machinery. phagocytic hemocytes, called plasmatocytes, ingest apoptotic cells In an effort to identify novel genes and signaling pathways and larval tissues (3). After metamorphosis, plasmatocytes adhere involved in the phagocytosis of S. aureus in Drosophila, we un- to heart tissue in the dorsal vessel and serve as the primary effector dertook an approach distinct from traditional forward or reverse of cellular immunity in adult flies (4, 5). genetic screens. We carried out a genome-wide association study using the Drosophila Genetic Reference Panel (DGRP; a panel of D. melanogaster strains with fully sequenced genomes) to find *Department of Cell Biology and Molecular Genetics, University of Maryland, College Park, MD 20742; †Institute for Bioscience and Biotechnology Research, single-nucleotide polymorphisms (SNPs) (18–20) associated with University of Maryland, College Park, MD 20742; and ‡Center for Bioinformatics S. aureus phagocytosis in adult flies. We found that a SNP in RNA- and Computational Biology, University of Maryland, College Park, MD 20742 binding Fox protein 1 (Rbfox1) was present in lines with low ORCID: 0000-0003-2748-8205 (S.M.M.). S. aureus phagocytosis. Received for publication April 4, 2018. Accepted for publication June 14, 2018. Rbfox1 is a member of the highly conserved Fox-1 (Rbfox) This work was supported by a National Science Foundation University of Maryland family of RNA-binding proteins (21–23). Individual Rbfox protein ADVANCE Seed Grant (to S.M.M. and L.P.W.), two National Institutes of Health grants (5T32AI089621 and 5T32GM080201 to A.E.N.-T.), and National Science Foundation isoforms localize either to the nucleus or the cytoplasm of cells Division of Biological Infrastructure Award 1564785 (to co-principal investigator S.M.M). (24–26). Nuclear isoforms regulate tissue-specific alternative The sequences presented in this article have been submitted to the splicing by binding to intronic (U)GCAUG elements in target Omnibus (http://ongen.us/A2BP1-GEOdata) under accession number GSE79488. RNAs, leading to the exclusion of downstream exons or the in- Address correspondence and reprint requests to Dr. Louisa P. Wu, University of clusion of upstream exons (27–31). Recent evidence from studies Maryland, 4112 Plant Sciences Building, College Park, MD 20742-4454. E-mail in Drosophila and humans indicates that cytoplasmic Rbfox1 address: [email protected] isoforms bind to (U)GCAUG elements in the 39 untranslated re- The online version of this article contains supplemental material. gions (UTRs) of specific mRNAs, regulating their stability and Abbreviations used in this article: BH, Benjamini and Hochberg; DE, differential expression; DGRP, Drosophila Genetic Reference Panel; Dscam4, Down syndrome translation (26, 32). In humans, there are three Rbfox family adhesion molecule 4; FDR, false discovery rate; Rbfox1, RNA-binding Fox protein 1; genes: RBFOX1, RBFOX2, and RBFOX3 (23). Disruptions in RNAi, RNA interference; RNAseq, RNA sequencing; RRM, RNA recognition motif; RBFOX1 are associated with neurologic diseases such as mental SNP, single-nucleotide polymorphism; UTR, untranslated region; WT, wild type. retardation, epilepsy, and autism (33, 34). RBFOX2 has been Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 implicated in cancer development and progression as well as

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1800496 2 RBFOX1 MEDIATES THE CELLULAR IMMUNE RESPONSE IN DROSOPHILA cardiac and neurologic development (35, 36). In Drosophila, of the SNP in Rbfox1 to reflect the most current genome release Rbfox1 is the only Rbfox family member. Rbfox1 has been (FlyBase 6.07). shown to play critical roles in oogenesis and in wing imaginal Larval phagocytosis assay disc development (26, 37, 38). Additionally, Rbfox proteins regulate and transcriptional networks in For each experiment, 10 third-instar wandering larvae were injected with equal volumes of fluorescein-labeled S. aureus (resuspended to 1.6 mg/ml in mouse, human, and Drosophila brains (24, 39–42), indicating sterile PBS with 5% green dye). After 30 min, larvae were bled directly onto that the biological functions of these proteins are conserved polylysine coverslips. After 2 min, carcasses were removed, and the cells across species. were fixed with cold 4% paraformaldehyde in PBS. The fluorescence of In this study, we report a novel role for Rbfox1 in the Dro- bioparticles not taken up by the cells was quenched by briefly washing fixed sophila cellular immune response to Gram-positive bacteria. hemocytes with 5% trypan blue in PBS. The cells were then washed in PBS and the slides were mounted with ProLong with DAPI. Hemocytes were Rbfox1 negatively regulates the expression of the Ig superfamily imaged using a 633 oil immersion lens on a Zeiss LSM 700 confocal mi- member Down syndrome adhesion molecule 4 (Dscam4)inadult croscope, and differential interference contrast was used to visualize the cell hemocytes. Importantly, silencing of Dscam4 was sufficient to boundary. Approximately 10 cells per larva were imaged, and individual restore a normal cellular immune response to S. aureus in bioparticles were counted. Experiments were performed in triplicate. Rbfox1-depleted blood cells. These results reveal that Rbfox1 Survival after S. aureus infection affects S. aureus phagocytosis, in part, by suppressing Dscam4 expression in phagocytes. Thus, our study reveals new pathogen- Groups of 24–30 adult flies, 5–7 d old, were injected with equal quantities (80 hl) of logarithmic phase culture of S. aureus (final OD 0.1 or 0.5). Flies specific roles for Rbfox1 and Dscam4 in the innate immune injected with PBS served as a wounding control. Flies were kept at 25˚C response of the fly. and transferred regularly to new food, and death was assessed every 24 h. Downloaded from The experiments were repeated at least two more times. Log-rank tests were used to determine if survival curves were significantly different, and Materials and Methods p values ,0.05 were deemed significant. Flies and fly husbandry Bacteria load assays DGRP stocks were generated by Dr. T. Mackay’s laboratory at North Carolina State University. The core 40 DGRP stocks (43) were provided by Thirty to fifty adult flies per genotype were injected with equal quantities of Dr. J. Leips’s laboratory at the University of Maryland, Baltimore County. logarithmic phase culture of S. aureus (final OD 0.5). At 0 and 24 h http://www.jimmunol.org/ The UAS-Rbfox1-RE flies were provided by Dr. L.S. Shashidhara from the postinjection, six to eight flies from each group were immediately washed Indian Institute of Science Education and Research in Pune, India (38). in 70% EtOH, rinsed in PBS, and homogenized in Luria–Bertani media The following stocks were obtained from the Bloomington Stock Center: containing 1% Triton X-100. Homogenates were serially diluted and plated w1118, the blood cell–specific driver w1118; hmlDGAL4, y1w*; w1118;P on Luria–Bertani agar plates. Plates were stored at 37˚C overnight and {HemolectinDGAL4}2, P{UAS-2xEGFP}AH2, P{UAS-mCD8::GFP.L} CFUs per fly were calculated. One-tailed t tests were performed and , LL5 (stock no. 5137), y1w*; Mi{MIC}Rbfox1M101918 (no. 44669), w1118; p values 0.05 were determined to be statistically significant. Experiments Df(3L)ED4457 (no. 9355), y1sc*v1; P{TRiP.HMS00478}attP2 (no. 32476) were done at least three times. and y1v1; P{TRiP.JF02600}attP2 (no. 27286), y1sc*v1; and P{TRiP. For S. aureus–GFP bacteria load experiments, flies were infected with HMC03277}attP2 (no. 51508). The argus mutant and isogenic parental S. aureus expressing GFP (OD 1.0). Flies were then kept at 25˚C, and images were taken at 0, 24, and 48 h postinjection using a Zeiss stereomicroscope strain cn bw were identified from the EMS collection of Zuker lines (44). by guest on September 27, 2021 UAS-Rbfox1-RE and Rbfox1 RNA interference (RNAi) lines were crossed (Discovery V8) with an AxioCam HC camera. For each fly, the fluorescence to hmlDGAL4 flies or to w1118 as a control. Flies were reared at 25˚C with intensity of the first two segments of the dorsal side of the abdomen was 60% humidity under a 12-h light–dark cycle and were fed a standard measured. At least 16 flies per genotype per time point were used in each molasses/cornmeal/agar medium. Experiments were conducted at the same experiment. The experiment was performed three times. time each day. FACS of adult hemocytes. We used a protocol modified from Clark et al. (45). One hundred w1118;P{HemolectinDGAL4}2, P{UAS-2xEGFP} In vivo phagocytosis AH2 flies (50 male and 50 female) were ground for 45 s in ice-cold PBS with 2 mM EDTA. Cells were filtered through a 70-mmcell To assay S. aureus and Escherichia coli phagocytosis, approximately six to sorter and centrifuged at low speed to pellet intact cells. Cells were then eight adult flies (3–5 d old) per genotype per experiment were injected resuspended in ice-cold PBS/EDTA and filtered through a 40-mm with either fluorescein-conjugated S. aureus resuspended in PBS (S2851, strainer, pelleted, and resuspended in PBS/EDTA. GFP-positive cells 1.6 mg/ml; Invitrogen) or fluorescein-conjugated E. coli resuspended in were sorted once using a FACSAria cell sorter with a 100-mM nozzle and PBS (E2861, 1.6 mg/ml; Invitrogen). Injections were performed using a 20 psi of pressure. Cells were then resorted into the RLT buffer of the Pneumatic PicoPump PV820 (World Precision Instruments). After 30 min, Qiagen RNeasy Mini Kit plus 3.75% 2-ME. RNA was isolated according flies were injected again with trypan blue to quench extracellular fluo- to the RNeasy Mini Kit protocol. rescence and mounted ventral side down, and images of the dorsal vessel RNA isolation and quantitative PCR. were taken using a Zeiss stereomicroscope (Discovery V8) with an Adult RNA samples were obtained by AxioCam HC camera. Fluorescence intensity of the area around the dorsal homogenizing six flies (three males and three females) in TRIzol. To obtain vessel was quantified using AxioVision 4.7, and the background fluores- RNA from larval hemocytes and carcasses, 10 or 40 third-instar wandering cence of an adjacent area was also quantified. The ratio of dorsal vessel larvae were washed in PBS and bled by gently nicking the cuticle and fluorescence intensity was calculated as the ratio of fluorescence in the bleeding hemolymph directly into PBS, and TRIzol was immediately added. dorsal vessel area to fluorescence in an adjacent area of equal size. Two- RNA was extracted from the adult and larval samples using chloroform and tailed t tests were used to calculate p values. To assay phagocytosis of ethanol precipitation and loaded onto RNeasy Mini Columns. RNA was beads, flies were injected with ∼1.0 mm of red fluorescent carboxylate- subsequently purified using the Qiagen RNeasy Mini Kit manufacturer’s modified FluoSpheres diluted 1:20 in PBS. After 30 min, flies were protocol. The RNA was digested with DNase (DNase I, RNase-free, injected with trypan blue and then mounted and visualized as described EN05021; Thermo Fisher Scientific), and cDNA was synthesized using above. reverse transcription (RevertAid First Strand cDNA Synthesis Kit, K1621; Thermo Fisher Scientific). Quantitative real-time PCR was carried out Genome-wide association analysis using iQ SYBR Green supermix (Bio-Rad Laboratories) on an Applied Biosystems 7300 following the manufacturer’s protocols. At least three To identify candidate SNPs that contribute to differences in S. aureus biological replicates were tested per genotype. Primer sequences are as phagocytosis, we tested 30 lines of the DGRP using the in vivo S. aureus follows: Rbfox1-RL, -RH, -RE, -RF, -RJ, -RI, -RK, and -RM:59-GTCTC- phagocytosis assay. We submitted the median fluorescence intensity phe- CAACATACCGTTCCG-39,59-CATCGTTGCTGTTAGCGAATG-39; notypes to the DGRP Freeze 1 release 5.49 analysis pipeline (http://dgrp. Rbfox1-RL and -RH:59-TTCACAAGAGCACGTCGATC-39,59-GTGGG- gnets.ncsu.edu/), and genome-wide association analysis was performed as CGTTTATAGAGTGGG-39; Hemolectin:59-GAGGACTAACAGCTTGG- described in Mackay et al. (18). SNPs were previously identified by whole- CAG-39,59-CGGCATGAGACGTCTTTATC-39; Dscam4:59-GGCAT genome sequencing of the DGRP lines (18), and SNP positions were TTCTGGCTCTGATTTG-39,59-CGATTTATGGGCAGCGTTTG-39; RP49:59-GC annotated according to FlyBase release 5.49. We updated the position AAGCCCAAGGGTATCGA-39,59-TAACCGATGTTGGGCATCAG-39. The Journal of Immunology 3

RNA sequencing reduced phagocytosis of S. aureus bioparticles (p =6.673 1026; Control flies (w1118; P{UAS-mCD8::GFP.L}LL5, hmlDGAL4; attP2) Fig. 1A, 1F). Rbfox1 encodes nine isoforms generated by alternative expressed the mCD8:GFP fusion protein on the surface of hemocytes and splicing, alternative promoter usage, and alternative polyadenylation. expressed normal levels of Rbfox1 in hemocytes. Rbfox1 RNAi flies were Of these, the isoforms Rbfox1-RL and -RH share a distal promoter, 1118 generated with the following genotype: w ; P{UAS-mCD8::GFP.L} whereas the seven remaining isoforms share a common promoter D LL5, hml GAL4; P{HMS0048}attP2. Rbfox1 RNAi flies expressed both 35 kb downstream of the first. The SNP at 3L:10538501 is within the mCD8:GFP fusion protein and the 21-bp short RNAi hairpin against Rbfox1 in hemocytes. Groups of 60 female flies were injected with equal a long in all isoforms, 22.5 kb downstream of the second volumes of S. aureus (OD 5.0) in sterile PBS or PBS (wounding control). promoter. Uninjected flies served as a control. Flies were incubated at 25˚C for 3 h We verified that Rbfox1 is expressed in adult hemocytes and then homogenized in Cell Dissociation Solution (C1544-100; Sigma) by quantifying the mRNA levels of Rbfox1 in phagocytic in an RNase/DNase-free 1.5-ml tube with an RNase-free mini pestle for 45 s. Supplemented Schneider’s Drosophila medium (0.1% BSA and (GFP-positive) adult hemocytes sorted by FACS to the levels in 2 mM EDTA) was added to the homogenate, and cells were filtered whole flies (Supplemental Fig. 1B). We then compared the ex- through a 70-mm cell sorter and centrifuged at low speed to pellet intact pression levels of Rbfox1 in DGRP lines with the major allele (T) cells. The cell pellet was resuspended in Supplemented Schneider’s at 3L:10538501 to lines with the minor allele (A) at position 3L: Drosophila medium and incubated with Dynabeads Mouse CD8 (Lyt-2) 10538501. Rbfox1 mRNA levels were significantly reduced in (11447D; Invitrogen) for 30 min at 4˚C. mCD8-positive cells were isolated using a magnetic stand and resuspended in RLT buffer from the QIAGEN larval hemocytes from line 307 compared with hemocytes from RNeasy Mini Kit. Total RNA was then isolated according to the manufac- line 365 (Fig. 1C). In adult flies, expression of two isoforms of turers’ protocols. Following cDNA library preparation, Illumina adapters and Rbfox1 (Rbfox1-RL and Rbfox1-RH, both of which are nuclear) indexes were added. Sequencing was carried out using 50-bp single-end was significantly reduced in lines carrying the (A) at 3L:10538501 Downloaded from reads on an Illumina HiSeq 1000 system. (Supplemental Fig. 1C). These results indicate that (A) at Differential expression analyses 3L:10538501 correlates with decreased expression of specific Alignment and counting. Raw reads (51 nt each) were trimmed 13 nt from isoforms of Rbfox1 in the affected DGRP lines. the 5-prime end using Trimmomatic 3 (46). Trimmed reads were aligned to To see if (A) at 3L:10538501 exerts its effect through a loss of the FlyBase D. melanogaster genome (dmel-all-aligned-r6.03.fasta.gz) function, we performed deficiency and transposon complementa- using Rsubread version 1.18.0 (47), allowing a total of three mis- tion tests. Progeny from a cross between a deficiency [Df(3L) http://www.jimmunol.org/ matches. The portion of total reads mapped for each sample is found in Supplemental Fig. 3. Quantification was performed based on the Fly- ED4457] uncovering the genomic region of Rbfox1 (chromosomal Base annotation file (dmel-all-no-analysis-r6.03.gff.gz) using the fea- deletion from 67E2 to 68A7) and DGRP line 307 exhibited im- tureCount (48) utility of Rsubread with default parameters to obtain a paired S. aureus phagocytosis compared with control flies, DGRP matrix of raw counts. line 365/Df(3L)ED4457 (Fig. 1D). This failure to complement Normalization and differential expression. Genes with at least one count per implies that the SNP is a loss of function with respect to the million in three (minimum number of biological replicates within a group) S. aureus phagocytosis phenotype. Additionally, a transposon in- or more samples were kept and quantile normalized (49) for library size. M101918 Differential expression (DE) analysis was performed using the voom sertion line Mi[ET]Rbfox1 (M101918) also failed to function of the limma (50) R package. In particular, the linear model gene complement line 307 when compared with control flies (Fig. 1E). exprs = factor 1 + factor 2 was fit. Factor 1 has two levels (mutant and These results indicate that the SNP in Rbfox1 causes a by guest on September 27, 2021 wild type [WT]), and factor 2 has three levels (PBS, S. aureus, and loss-of-function defect. Furthermore, the M101918 insertion uninjected). Contrasts were used to assess comparisons of interest. All p values were adjusted using the Benjamini and Hochberg (BH) (51) affected the expression of Rbfox1 transcripts, and homozygous M101918 method. Rbfox1 mutants showed S. aureus–specific phagocytosis defects (Supplemental Fig. 1D–G). Rbfox1M101918 flies were also Accession number more susceptible to S. aureus infection than the background The accession number for the raw sequencing data reported in this paper is control w1118 (Supplemental Fig. 1H). It was evident that the GSE79488 and can be accessed at http://ongen.us/A2BP1-GEOdata. transposon insertion resulted in loss of function of Rbfox1 because flies carrying M101918 over a deletion that uncover the genomic Results region of Rbfox1 also showed significantly decreased S. aureus The RNA-binding protein Rbfox1 regulates phagocytosis uptake. of S. aureus In blood cells, the M101918 insertion caused an increase in the To characterize natural genetic variation in the innate immune expression of two Rbfox1 isoforms that have alternative, upstream response of Drosophila to Gram-positive bacteria (S. aureus), we start sites (RL and RH) (Fig. 1F, Supplemental Fig. 1D). To determine screened 30 lines of the core 40 subset of the DGRP using an whether increased or decreased expression of Rbfox1 ledtoimpaired in vivo adult phagocytosis assay (Fig. 1A, Supplemental Fig. 1A) S. aureus uptake, we directly manipulated Rbfox1 levels in hemoc- (5). We injected individual flies with fluorescein-labeled S. aureus ytes. Using two Rbfox1 RNAi constructs that target the 39UTR of all and quantified phagocytosis by measuring the fluorescence of Rbfox1 isoforms, the expression of Rbfox1 was silenced specif- dorsal vessel–associated hemocytes. To control for experimental ically in blood cells using the hmlDGAL4 driver (52). Phago- variability, the data from every assay were normalized relative to cytosis of S. aureus bioparticles was significantly decreased in a laboratory strain, cn bw. A previously characterized line (argus, hmlD.Rbfox1 RNAi hemocytes as compared with hemoc- A.E. Nazario-Toole and L.P. Wu, unpublished observations) with ytes from control flies (Fig. 2A, 2B). Phagocytosis of live, defects in S. aureus phagocytosis served as an additional control GFP-expressing S. aureus was also significantly decreased in (0.747, relative to 1.0 for cn bw). For all 30 DGRP lines, the range hmlD.Rbfox1 RNAi flies (Supplemental Fig. 2A). was 0.070–1.37, and the median phagocytosis was 0.99. Expression of a transgenic Rbfox1-RE in the Rbfox1 RNAi flies Fig. 1A shows phagocytosis images for line 365 (an example of was sufficient to rescue the S. aureus phagocytosis phenotype median phagocytosis), line 307 (with significantly reduced phago- (Fig. 2A, 2B), suggesting that Rbfox1 expression above a certain cytosis), and line 786 (with significantly higher phagocytosis). threshold is required to mediate S. aureus phagocytosis by Genotype–phenotype association analyses (18) revealed that an SNP hemocytes. However, overexpression of the same Rbfox1-RE in Rbfox1 [minor allele (A) versus major allele (T) at position transgene in a WT background resulted in dramatically reduced 3L:10538501] was a natural polymorphism found in lines with phagocytosis of S. aureus (Fig. 2C). Of note, hmlD.Rbfox1 RNAi 4 RBFOX1 MEDIATES THE CELLULAR IMMUNE RESPONSE IN DROSOPHILA Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. An in vivo phagocytosis screen identifies an SNP in Rbfox1 that affects uptake of S. aureus.(A) Phagocytosis of fluorescein-labeled S. aureus in 30 lines of the DGRP. The screen was performed over the course of several days, and mean phagocytosis of each line tested on a given day was normalized relative to the daily control, cn bw. Representative images (original magnification 332) of S. aureus phagocytosis in line 365 (median phagocytosis), 307 (significantly reduced phagocytosis), and 786 (significantly higher phagocytosis). (B) Box shows the area of the fly imaged during the adult phagocytosis assay, and the circled area is the dorsal vessel, where a large number of sessile hemocytes localize. (C) Expression of Rbfox1 in larval hemocytes and carcasses from lines 365 and 307. Hemocytes were collected from 40 larvae. Transcript levels were measured via quantitative PCR (qPCR), and rp49 was used as an endogenous control gene. (D) Phagocytosis of fluorescein-labeled S. aureus in F1 progeny of crosses between lines 365 or 307 and a chromosomal deletion line [3L(ED4457)] that uncovers the genomic region of Rbfox1. (E) Phagocytosis of fluorescein-labeled S. aureus in F1 progeny of crosses between lines 365 or 307 and a transposon insertion line that disrupts expression of Rbfox1, Mi[ET]Rbfox1M101918. (Figure legend continues) The Journal of Immunology 5 and hmlD.Rbfox1-RE flies do not have fewer blood cells than After S. aureus infection, CR45605 was the most strongly down- WT flies, as measured by larval hemocyte count (hmlD.Rbfox1 regulated transcript in Rbfox1 RNAi hemocytes (19.4-fold), and it RNAi) and adult phagocytosis of latex beads (hmlD.Rbfox1 encodes a computationally predicted long noncoding RNA of un- RNAi and hmlD.Rbfox1-RE) (Supplemental Fig. 2). Addition- known function. As a long noncoding RNA, CR45605 may play a role ally, hmlD.Rbfox1 RNAi and hmlD.Rbfox1-RE adult flies show in blood cells by regulating the expression of genes important for no changes in the phagocytosis of E. coli (Supplemental Fig. 2). phagocytosis of S. aureus. Tollo/Toll-8, peptidoglycan recognition This indicates that the blood cells of these flies develop normally protein LE (PGRP-LE), and tamo are immune-responsive genes that and possess functional phagocytic machinery but are specifically were among those that were significantly down in Rbfox1 mutant unable to phagocytose S. aureus. Based on these results, we hemocytes after S. aureus infection (53–56). Tamo regulates the nu- conclude that the expression of Rbfox1 must be tightly regulated clear import of Dorsal in fat body cells postinfection (56). Tollo/Toll-8 in hemocytes because too little or too much expression specifically and PGRP-LE are pathogen recognition receptors that both play a role impairs phagocytosis of S. aureus. during the humoral immune response to Gram-negative bacteria. Blood cell–specific silencing or overexpression of Rbfox1 sig- Thirty genes were up in Rbfox1 RNAi hemocytes after wounding nificantly increased fly mortality following S. aureus infection (PBS injection). Within this group were six genes involved in the (Fig. 2D, 2E). Both lower expression (hmlD.Rbfox1 RNAi) and long-chain fatty acyl-CoA pathway: members of the IPR002076 overexpression (hmlD.Rbfox1-RE) flies showed an increased ELO family (eloF, CG16904, CG9458,andCG9459), CG1444,and bacterial load following S. aureus infection (Fig. 2F, 2G). Thus, Sc2 (Holm–Bonferroni-corrected p = 6.76 3 1029). These results altered Rbfox1 expression affects the fly’s ability to limit the suggest that Rbfox1 acts to limit fatty acid elongation in hemocytes growth of bacteria, indicating that these flies are less resistant to after wounding. Mammalian macrophages metabolize fatty acids to Downloaded from S. aureus than WT flies. fulfill the energy needs necessary for activation and phagocytosis (57). It is possible that Rbfox1 regulates hemocyte metabolic ac- Transcriptome DE analysis reveals Rbfox1 targets in tivity in response to wounding by diverting energy and resources to adult hemocytes lipid metabolism. Furthermore, seven genes were up in Rbfox1 As a splicing factor, Rbfox1 is likely regulating the posttranscriptional RNAi hemocyte samples in all conditions: CG6432, CR12628,

processing of mRNAs that are important for managing the immune diazepam-binding inhibitor (Dbi), peritrophin A, tumorous testis http://www.jimmunol.org/ response against S. aureus. To identify transcripts that are affected in (tut), Tektin-C,andDscam4. Based on standard adult hemocytes after loss of Rbfox1, an RNA sequencing (RNAseq) annotations, two of these genes, CG6432 and Dbi, are also involved approach was taken. Hemocytes that expressed WT levels of Rbfox1 in synthesis and oxidation of fatty acids (58). Peritrophin A is a were used as a control (WT hemocytes), and hemocytes expressing a constituent of the midgut peritrophic membrane, tut is an RNA- single Rbfox1 RNAi construct were used to test for the effects of loss binding protein, CR12628 is a non–protein-coding RNA, and of Rbfox1 (Rbfox1 RNAi hemocytes). Groups of 60 adult female flies Dscam4 is an Ig-like cell surface protein. These diverse genes are were either uninfected, mock infected with PBS, or infected with live likely to be targets of Rbfox1 in WT adult blood cells, and Rbfox1 S. aureus (OD 5.0) for 3 h. Tagged hemocytes were then isolated negatively regulates their expression. using immunoselection, and their cDNA was sequenced using 50-nt by guest on September 27, 2021 single-end reads on an Illumina platform (Supplemental Fig. 3A). Suppression of Dscam4 rescues the immune response defects DE analysis was carried out to identify genes that were sig- caused by Rbfox1 RNAi nificantly (BH 2log10-adjusted p value ,0.05) up- or downreg- Phagocytosis of microbes by immune surveillance cells is an im- ulated in Rbfox1 RNAi hemocytes compared with control mediate response that occurs within minutes of an infection. We hemocytes (Fig. 3, Supplemental Fig. 3A). Thus, if a gene’s ex- hypothesized that Rbfox1 regulates the expression of cell surface pression was down in Rbfox1 RNAi hemocytes, Rbfox1 function receptors or coreceptors that may be important for S. aureus rec- is required to maintain normal levels of mRNA expressed from ognition and uptake. Thus, we focused our remaining studies on that gene. Conversely, if Rbfox1 suppresses a target mRNA by any Dscam4, which encodes a plasma membrane transmembrane protein means (including direct destabilization), then mRNA from that with Ig-like and fibronectin type III (FN3)domains. There are four gene would be expressed at higher levels in Rbfox1 RNAi Dscam-like proteins in the Drosophila genome, and the most ex- hemocytes compared with controls. tensively characterized of these is Dscam1 (59, 60). Dscam1 has the To examine how Rbfox1 regulates gene expression and splicing potential to express over 18,000 alternative splice isoforms and is after S. aureus infection, the lists of genes that were differentially essential for axon guidance and the formation of neural connections expressed in uninfected, PBS-infected (wounding control), and in Drosophila (61–63). An RNAi screen in the S2 cell line identified S. aureus–infected WT hemocytes were compared with Rbfox1- 36 RNA-binding proteins that regulate alternative splicing of silenced hemocytes exposed to the same treatments (Fig. 3, Dscam1 but found that Rbfox1 had no effect on alternative splicing Supplemental Fig. 3B). Twenty-two genes were down in Rbfox1 of Dscam1 (64), consistent with our finding that Rbfox1 RNAi in RNAi hemocytes, as compared with WT hemocytes, after S. aureus hemocytes does not significantly alter Dscam1 expression. Dscam1 infection. In contrast, only eight genes were down in Rbfox1 RNAi also has a role in the innate immune response and is important for after PBS treatment. Five genes (cpo, cyc, Obp99d, CR45605, phagocytosis of E. coli in Drosophila and for phagocytosis of E. coli and CG32039) were found significantly down in Rbfox1 RNAi and S. aureus in the Anopheles gambiae immune-competent cell line hemocytes in all conditions, indicating that these five genes are Sua5B (65, 66). Additionally, RNAi depletion of AgDscam de- primary targets for Rbfox1. creased survival and increased bacterial loads in mosquitos

(F) Genomic region of Rbfox1 and its eight isoforms. Location of SNP 3L:10538501 (arrow) and the Minos transposon M101918 (black, inverted triangle) within Rbfox1 are shown. The Drosophila Transgenic RNAi Project short hairpin RNA construct HMS00478 targeting a 21-bp sequence in the 39UTR of all isoforms is depicted by the gray triangle. All phagocytosis assays were performed in triplicate; n = 8–10 flies per genotype in each experiment. qPCR assays were performed three times. Error bars, 6 SEM. *p , 0.05, **p , 0.01. ns, not significant. 6 RBFOX1 MEDIATES THE CELLULAR IMMUNE RESPONSE IN DROSOPHILA Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 2. S. aureus phagocytosis and survival are affected when Rbfox1 expression is either increased or decreased in hemocytes. (A) Quantification and representative images of phagocytosis of fluorescein-labeled S. aureus by larval hemocytes from WT/Rbfox1 RNAi (control), hmlD.Rbfox1 RNAi, and hmlD. Rbfox1-RE/Rbfox1 RNAi flies. For each experiment, 10 larvae were injected with equal amounts of fluorescein-labeled S. aureus. Approximately 10 cells per larva were imaged, and individual bioparticles per cell were counted. Fluorescein-labeled S. aureus bioparticles appear as white dots. DIC was used to visualize cell boundary. (B) Adult phagocytosis of fluorescein-labeled S. aureus in WT/Rbfox1 RNAi, hmlD.Rbfox1 RNAi, and hmlD.Rbfox1-RE/Rbfox1 RNAi flies and DAPI was used to stain nucleic acids. (C) Adult phagocytosis of fluorescein-labeled S. aureus in WT/Rbfox1-RE and hmlD.Rbfox1-RE flies. Adult in vivo phagocytosis experiments were performed in triplicate with 8–10 flies per genotype in each experiment. (D) Representative survival curves of WT/Rbfox1 RNAi, hmlD.Rbfox1 RNAi, and hmlD.Rbfox1-RE/Rbfox1 RNAi flies after injection with S. aureus (OD 0.1) n = 24–30 flies. (E) Representative survival curves of WT/Rbfox1-RE and hmlD.Rbfox1-RE flies after injection with S. aureus (OD 0.1) n = 24–30 flies. Dotted lines represent mock-infected groups. (F) S. aureus (OD 0.5) bacteria load in WT/Rbfox1 RNAi and hmlD.Rbfox1 RNAi at 0, 24, and 48 h postinfection. (G)Comparisonof (Figure legend continues) The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 3. RNAseq and DE analysis reveal Rbfox1 targets in adult hemocytes. (A) The DE between WT and Rbfox1 RNAi uninfected hemocytes. The volcano plot depicts the magnitude of DE between WT and Rbfox1 hemocytes. Each dot represents a single gene (n = 8858) that is expressed in hemocytes, and those with a BH adjusted p value , 0.1 (indicated by the horizontal line) are differentially expressed. Genes with decreased expression in Rbfox1 RNAi hemocytes have positive log2 fold changes (top right quadrant), and genes with increased expression in Rbfox1 RNAi samples have negative log2 fold changes (top left quadrant). (B) DE between WT and Rbfox1 hemocytes 3 h after PBS infection. (C) DE between WT and Rbfox1 hemocytes 3 h after S. aureus infection. (D) Overlap of differentially expressed genes in each condition. See also Supplemental Fig. 3. (E) Bar graph of 20 Rbfox1 target genes and the magnitude of DE after Rbfox1 RNAi. n = 60 female flies per sample and three biological replicates per genotype and treatment. 2log10-adjusted p value ,0.05. infected with S. aureus and E. coli. Our RNAseq results show do not rule out less significant or consistent effects on the other that Dscam4 is the Dscam family member on which Rbfox1 has three Dscam family members. the greatest effect in Drosophila hemocytes and that this regu- Dscam4 expresses five transcripts with a pair of alternative first lation is important for phagocytosis of S. aureus. However, they exons and alternatively spliced exons 21, 29, 31, and 32 (FlyBase

S. aureus (OD 0.5) bacteria load recovered in WT/Rbfox1-RE and hmlD.Rbfox1-RE at 0, 24, and 48 h postinfection. Bacteria load experiments were performed three times, with six to eight flies collected per time point. Error bars, 6 SEM. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. DIC, differential interference contrast; ns, not significant. 8 RBFOX1 MEDIATES THE CELLULAR IMMUNE RESPONSE IN DROSOPHILA Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 4. The Ig superfamily member Dscam4 affects S. aureus phagocytosis. (A)ExpressionofDscam4 in larval hemocytes from WT/Rbfox1 RNAi and hmlD.Rbfox1 RNAi via quantitative PCR (qPCR). Hemocytes from 40 larvae were pooled for each experiment; n =3.(B) Dscam4 mRNA levels in adult hemocytes from WT/Rbfox1 RNAi and hmlD.Rbfox1 RNAi measured using qPCR. Relative expression was calculated using rp49 as an endogenous control. All RNAseq samples were assessed: n =12WTandn =12Rbfox1 RNAi. (C)ExpressionofRbfox1 (solid symbols) and Dscam4 (open symbols) in larval hemocytes. Hemocytes from 10 larvae were pooled for each experiment; n =3.(D) Phagocytosis of fluorescein-labeled S. aureus in WT/Rbfox1 RNAi, hmlD.Rbfox1 RNAi, and hmlD.Dscam4 RNAi/Rbfox1 RNAi flies. (E) Phagocytosis of fluorescein-labeled S. aureus in WT/Dscam4 RNAi and hmlD.Dscam4 RNAi flies. Six to eight flies were used in each experiment. (F) Representative survival curve of WT/Dscam4 RNAi and hmlD.Dscam4 RNAi flies after infection with S. aureus (OD 0.5); n = 28–30 flies. (G) Representative survival curve of hmlD.Rbfox1 RNAi and hmlD.Dscam4 RNAi/Rbfox1 RNAi flies (Figure legend continues) The Journal of Immunology 9 release FB2016_01). Each isoform contains nine extracellular Ig- to the infection at the same time (Fig. 4G). Similarly, blood cell– like domains and six FN3 (IPR003961) domains. The extracellular specific expression of Dscam4 RNAi restored normal suppression regions of the Dscam4 proteins are nearly identical (98–100% of bacterial growth at 24 h as compared with Rbfox1 RNAi files, identity among isoforms). The main source of variability be- indicating that Dscam4 is important to control bacteria growth at tween Dscam4 isoforms is found in the cytoplasmic tail, this time point. However, by 48 h, bacteria levels were similar in which contains no distinguishable protein domains. flies of both genotypes, indicating that Dscam4 RNAi was not DE analysis comparing PBS-injected WT hemocytes to S. aureus– sufficient to control the growth of bacteria at the later stages of infected WT hemocytes revealed nonsignificant changes in Rbfox1 infection (Fig. 4H). It appears that reducing the levels of Dscam4 and Dscam4 transcript levels: 0.132-fold change (false discovery rate in the Rbfox1 RNAi background is sufficient to protect the fly only [FDR]–adjusted p value 0.713) in Rbfox1 levels and a 0.22-fold at early time points. Upon exposure to S. aureus, Rbfox1 also change (FDR-adjusted p value 0.773) in Dscam4 levels, indicating regulates the expression of several other genes, including some that neither gene is transcriptionally regulated after S. aureus in- immune-related transcripts known to play a role in the humoral fection. In contrast, Dscam4 transcripts showed a 5.3-fold increase immune response (Supplemental Fig. 4). Together, these results (FDR-adjusted p value 5.88 3 1025) in uninfected Rbfox1 RNAi indicate that Dscam4 is important early in the response for hemocytes. This upregulation was also observed in Rbfox1 phagocytosis and control of bacteria loads, but as the infection RNAi hemocytes after wounding (2.4-fold increase; FDR-adjusted progresses, other target genes of Rbfox1 may play protective roles. p value 6.38 3 1025) and after S. aureus infection (2.7-fold in- crease; FDR-adjusted p value 2.0 3 1025). Quantitative PCR Discussion analyses of larval and adult hemocytes confirmed that loss of Previous studies have demonstrated that Rbfox1 is crucial for Downloaded from Rbfox1 led to significantly increased levels of Dscam4 mRNA neuronal and germline development in Drosophila. In this study, (Fig. 4A, 4B). Additionally, in larval hemocytes overexpressing we have presented a novel role for Rbfox1 in the maintenance of Rbfox1, Dscam4 expression is 1.7-fold (SD 0.32) higher than WT phagocyte function and immunocompetence by providing func- hemocytes. tional evidence that Rbfox1 is required for regulation of immune The fact that Rbfox1 is a known splicing factor suggests an responses in S. aureus–infected hemocytes.

effect through splicing. However, analysis of the splicing of The current annotation of the Drosophila genome reports eight http://www.jimmunol.org/ Dscam4 in our RNAseq data [using Salmon (67) and SUPPA2 isoforms of Rbfox1, and a recent study noted an additional ovary- (68)] detected no significant change in the pattern of alternative specific isoform, Rbfox1-RN (26). Seven of the reported Rbfox1 splicing, and the major Dscam4 isoform is the same in all treat- isoforms contain nuclear localization signals in their C-terminal ments. Furthermore, the effect is consistent and unaffected by domain, whereas the remaining two, -RN and -RF, do not. Nuclear other treatments; Dscam4 RNA abundance is higher in all 12 isoforms regulate mRNA alternative splicing by binding to Rbfox1 knockdown samples than in any of the 12 control samples. intronic (U)GCAUG elements, whereas cytoplasmic isoforms These findings indicate that Dscam4 is not a direct target of control mRNA stability and translation by binding to GCAUG Rbfox1, but additional analyses are necessary to identify the elements in target mRNA 39UTRs. The alternative splicing mechanism by which Rbfox1 knockdown or overexpression leads function of the nuclear localized isoforms is consistent with that of by guest on September 27, 2021 to increased Dscam4 transcript levels. Rbfox1 orthologs in higher organisms (21–23, 27–31). We then tested whether decreasing the levels of Dscam4 mRNA In this study, we undertook a broad examination of the function of would rescue the phenotypes seen in Rbfox1 RNAi flies. Indeed, Rbfox1 in Drosophila hemocytes using genome-wide association coexpression of RNAi constructs against Rbfox1 and Dscam4 in studies, RNAseq, and RNAi. By targeting an exon common to all hemocytes was sufficient to normalize Dscam4 mRNA levels to that Rbfox1 isoforms, we characterized how total loss of Rbfox1 function of control hemocytes (Fig. 4C). There was a strong rescue of the affects hemocytes. Furthermore, we find that only two nuclear local- phagocytosis phenotype in hmlD.Rbfox1 RNAi/Dscam4 RNAi flies ized isoforms, Rbfox1-RL and Rbfox1-RH, show reduced expression (Fig. 4D). In contrast, hemocyte Dscam4 RNAi in a WT background in DGRP lines carrying the minor allele (A) at position 3L:10538501 hadnoeffectonadultS. aureus phagocytosis or survival (Fig. 4E, and in the Minos insertion line Mi[ET]Rbfox1M101918.Thus,the 4F). These findings indicate that cellular immune phenotype of phenotypes observed in DGRP lines carrying the minor allele (A) Rbfox1 mutants is caused by the increased expression of Dscam4 3L:10538501 and in Rbfox1M101918 mutants are likely due to alter- and that reduction in Dscam4 expression in Rbfox1 RNAi hemocytes native splicing changes caused by reduced expression Rbfox1-RL and can restore phagocytosis of S. aureus. The increased level of Dscam4 Rbfox1-RH, indicating that these two specific isoforms may be re- transcripts in Rbfox1 RNAi hemocytes may cause Dscam4 to be sponsible for the immune defects observed in our studies. overly abundant on the surface of hemocytes. If Dscam4 is func- To identify Rbfox1 targets, we performed transcriptome analyses tioning as a negative regulator of phagocytosis, then an excess of the in hemocytes with silenced expression of Rbfox1, with or without receptor could inhibit uptake of S. aureus. S. aureus infection. Expression of the Ig superfamily member Dscam4 Rbfox1 RNAi flies show increased susceptibility to infection was increased over 5-fold when Rbfox1 was downregulated, indicat- and increased microbial growth after S. aureus infection, and these ing that Dscam4 is negatively regulated by Rbfox1 at the level of phenotypes were partially rescued by coexpression of Dscam4 transcription. Strikingly, silencing of Dscam4 in Rbfox1-depleted RNAi (Fig. 4G, 4H). During early time points, significantly more blood cells was sufficient to rescue the fly’s cellular immune re- hmlD.Rbfox1 RNAi flies died than hmlD.Rbfox1 RNAi/Dscam4 sponse to S. aureus, indicating that negative regulation of Dscam4 RNAi flies, but the longest-living flies in both groups succumbed expression is critical for Rbfox1 regulation of phagocytosis. Thus, our

after injection with S. aureus (OD 0.5); n = 28–30 flies. (H) Bacteria load in adult WT/Rbfox1 RNAi, hmlD.Rbfox1 RNAi, and hmlD.Dscam4 RNAi/ Rbfox1 RNAi flies 0, 24, and 48 h after S. aureus–expressing GFP (OD 1.0) injections. For each fly, the fluorescence intensity of the first two segments of the dorsal side of the abdomen was measured. Approximately 16 flies per genotype per time point were used in each experiment. Error bars, 6 SEM. *p , 0.05, **p , 0.01. ns, not significant. 10 RBFOX1 MEDIATES THE CELLULAR IMMUNE RESPONSE IN DROSOPHILA results have exploited natural variation in immune competency in flies 4. Lanot, R., D. Zachary, F. Holder, and M. Meister. 2001. Postembryonic hema- topoiesis in Drosophila. Dev. Biol. 230: 243–257. to reveal a role for Dscam4 as a negative regulator of phagocytosis. 5. Elrod-Erickson, M., S. Mishra, and D. Schneider. 2000. Interactions between the Additionally, we found that silencing all Rbfox1 isoforms and cellular and humoral immune responses in Drosophila. Curr. Biol. 10: 781–784. overexpressing one specific isoform, Rbfox1-RE, impaired the cellular 6. Defaye, A., I. Evans, M. Crozatier, W. Wood, B. Lemaitre, and F. Leulier. 2009. Genetic ablation of Drosophila phagocytes reveals their contribution to both immune response to S. aureus in the same way. A recent publication in development and resistance to bacterial infection. J. Innate Immun. 1: 322–334. which Rbfox1 levels were manipulated in Drosophila oocytes reported 7. Charroux, B., and J. Royet. 2009. Elimination of plasmatocytes by targeted similar findings (69). Upon starvation, Rbfox1-RE–overexpressing cells apoptosis reveals their role in multiple aspects of the Drosophila immune re- sponse. Proc. Natl. Acad. Sci. USA 106: 9797–9802. exhibited reduced apoptosis and decreased cell death. In contrast, 8. Nehme, N. T., J. Quintin, J. H. Cho, J. Lee, M. C. Lafarge, C. Kocks, and Rbfox1 RNAi–expressing cells showed increased cell death, indicating D. Ferrandon. 2011. Relative roles of the cellular and humoral responses in the that Rbfox1-RE is protective under starvation, and loss of Rbfox1-RE Drosophila host defense against three gram-positive bacterial infections. PLoS One 6: e14743. (and all isoforms of Rbfox1) is detrimental to the cell. However, both 9. DeLeo, F. R., B. A. Diep, and M. Otto. 2009. Host defense and pathogenesis in overexpression (Rbfox1-RE) and RNAi-mediated knockdown (of all Staphylococcus aureus infections. Infect. Dis. Clin. North Am. 23: 17–34. isoforms) lead to the same cellular proliferation defects and decreased 10. Haine, E. R., Y. Moret, M. T. Siva-Jothy, and J. Rolff. 2008. Antimicrobial defense and persistent infection in insects. Science 322: 1257–1259. expression of the transcription factor cut. This finding is analogous to 11. Stuart, L. M., J. Deng, J. M. Silver, K. Takahashi, A. A. Tseng, E. J. Hennessy, the work reported in this study in that Rbfox1 overexpression or si- R. A. Ezekowitz, and K. J. Moore. 2005. Response to Staphylococcus aureus lencing can lead to similar effects on gene expression and cellular requires CD36-mediated phagocytosis triggered by the COOH-terminal cyto- plasmic domain. J. Cell Biol. 170: 477–485. phenotype. In our overexpression experiments, Rbfox1-RE may behave 12. Ra¨met, M., A. Pearson, P. Manfruelli, X. Li, H. Koziel, V. Go¨bel, E. Chung, in a dominant negative fashion to impede the functions of -RL and M. Krieger, and R. A. Ezekowitz. 2001. Drosophila scavenger receptor CI is a

-RH, which could lead to increased expression of Dscam4 observed in pattern recognition receptor for bacteria. Immunity 15: 1027–1038. Downloaded from 13.Kocks,C.,J.H.Cho,N.Nehme,J.Ulvila,A.M.Pearson,M.Meister, Rbfox-RE overexpression hemocytes. C. Strom, S. L. Conto, C. Hetru, L. M. Stuart, et al. 2005. Eater, a trans- Our results indicate that a delicate balance of Rbfox1 expression membrane protein mediating phagocytosis of bacterial pathogens in Dro- must be maintained to support hemocyte functions. We also find ev- sophila. Cell 123: 335–346. 14. Kurucz, E., R. Ma´rkus, J. Zsa´mboki, K. Folkl-Medzihradszky, Z. Darula, idence for all isoforms of Rbfox1 in hemocytes. Previous biochemical P. Vilmos, A. Udvardy, I. Krausz, T. Lukacsovich, E. Gateff, et al. 2007. Nimrod, analysis indicates that Rbfox1 isoforms associate with themselves and a putative phagocytosis receptor with EGF repeats in Drosophila plasmatocytes. Curr. Biol. 17: 649–654. with each other (37). Future research elucidating the roles of specific http://www.jimmunol.org/ 15. Garver, L. S., J. Wu, and L. P. Wu. 2006. The peptidoglycan recognition protein Rbfox1 isoforms will shed light on the relative contribution and PGRP-SC1a is essential for Toll signaling and phagocytosis of Staphylococcus functions of specific isoforms during the cellular immune response. aureus in Drosophila. Proc. Natl. Acad. Sci. USA 103: 660–665. Finally, RNA-binding proteins have been shown to regulate im- 16. Shiratsuchi, A., T. Mori, K. Sakurai, K. Nagaosa, K. Sekimizu, B. L. Lee, and Y. Nakanishi. 2012. Independent recognition of Staphylococcus aureus by two munological responses by altering the posttranscriptional processes receptors for phagocytosis in Drosophila. J. Biol. Chem. 287: 21663–21672. of splicing, editing, decay, and translation (reviewed in Ref. 70). The 17. Nonaka, S., K. Nagaosa, T. Mori, A. Shiratsuchi, and Y. Nakanishi. 2013. Integrin aPS3/bn-mediated phagocytosis of apoptotic cells and bacteria in RNA recognition motif (RRM) of Drosophila Rbfox1 and human Drosophila. J. Biol. Chem. 288: 10374–10380. RBFOX proteins are 90% identical, and the splicing function of 18. Mackay, T. F., S. Richards, E. A. Stone, A. Barbadilla, J. F. Ayroles, D. Zhu, Rbfox family proteins is mediated exclusively by the presence of an S. Casillas, Y. Han, M. M. Magwire, J. M. Cridland, et al. 2012. The Drosophila

melanogaster genetic reference panel. Nature 482: 173–178. by guest on September 27, 2021 intact RRM (23). Given the similarities in the RRM domain be- 19. Miller, L. S., and J. S. Cho. 2011. Immunity against Staphylococcus aureus tween of Rbfox proteins in both species, it is possible that RBFOX1 cutaneous infections. Nat. Rev. Immunol. 11: 505–518. may function in a manner analogous to Drosophila Rbfox1 in pe- 20. van Kessel, K. P., J. Bestebroer, and J. A. van Strijp. 2014. Neutrophil-mediated phagocytosis of Staphylococcus aureus. Front. Immunol. 5: 467. ripheral blood cells by controlling the expression of cell surface 21. Shibata, H., D. P. Huynh, and S. M. Pulst. 2000. A novel protein with RNA- receptors important for phagocytosis. RBFOX1 mRNA is expressed binding motifs interacts with ataxin-2. Hum. Mol. Genet. 9: 1303–1313. in human peripheral blood cells (33, 34), therefore identifying 22. Bajpai, R., N. Sambrani, B. Stadelmayer, and L. S. Shashidhara. 2004. Identi- fication of a novel target of D/V signaling in Drosophila wing disc: Wg- in vivo mRNA targets of RBFOX1 in polymorphonuclear leuko- independent function of the organizer. Gene Expr. Patterns 5: 113–121. cytes, such as neutrophils, may serve to elucidate a new function for 23. Kuroyanagi, H. 2009. Fox-1 family of RNA-binding proteins. Cell. Mol. Life Sci. RBFOX1 in the human innate immune response. 66: 3895–3907. 24. Lee, J. A., Z. Z. Tang, and D. L. Black. 2009. An inducible change in Fox-1/ A2BP1 splicing modulates the alternative splicing of downstream neuronal target exons. Genes Dev. 23: 2284–2293. Acknowledgments 25. Nakahata, S., and S. Kawamoto. 2005. Tissue-dependent isoforms of mamma- We thank Wen Huang and Michael Magwire from the laboratory of lian Fox-1 homologs are associated with tissue-specific splicing activities. Dr. Trudy Mackay at North Carolina State University for providing sup- Nucleic Acids Res. 33: 2078–2089. port with using the DGRP bioinformatics online tool. We also thank 26. Carreira-Rosario, A., V. Bhargava, J. Hillebrand, R. K. Kollipara, M. Ramaswami, and M. Buszczak. 2016. Repression of pumilio protein ex- Dr. Jeff Leips at the University of Maryland, Baltimore County for pro- pression by Rbfox1 promotes germ cell differentiation. Dev. Cell 36: 562–571. viding the core 40 DGRP lines and Dr. L.S. Shashidhara from the Indian 27. Jin, Y., H. Suzuki, S. Maegawa, H. Endo, S. Sugano, K. Hashimoto, K. Yasuda, Institute of Science Education and Research in Pune, India for providing and K. Inoue. 2003. A vertebrate RNA-binding protein Fox-1 regulates tissue- the UAS-Rbfox1-RE flies. Finally, we thank Yifei Shi for the principal specific splicing via the pentanucleotide GCAUG. EMBO J. 22: 905–912. component analyses of the RNAseq data and for generating the principal 28. Underwood, J. G., P. L. Boutz, J. D. Dougherty, P. Stoilov, and D. L. Black. 2005. Homologues of the Caenorhabditis elegans Fox-1 protein are neuronal component analysis plots. splicing regulators in mammals. Mol. Cell. Biol. 25: 10005–10016. 29. Ponthier, J. L., C. Schluepen, W. Chen, R. A. Lersch, S. L. Gee, V. C. Hou, A. J. Lo, S. A. Short, J. A. Chasis, J. C. Winkelmann, and J. G. Conboy. 2006. Disclosures Fox-2 splicing factor binds to a conserved intron motif to promote inclusion of The authors have no financial conflicts of interest. protein 4.1R alternative exon 16. J. Biol. Chem. 281: 12468–12474. 30. Auweter, S. D., R. Fasan, L. Reymond, J. G. Underwood, D. L. Black, S. Pitsch, and F. H. Allain. 2006. Molecular basis of RNA recognition by the human al- References ternative splicing factor Fox-1. EMBO J. 25: 163–173. 31. Fukumura, K., A. Kato, Y. Jin, T. Ideue, T. Hirose, N. Kataoka, T. Fujiwara, 1. Igboin, C. O., A. L. Griffen, and E. J. Leys. 2012. The Drosophila melanogaster H. Sakamoto, and K. Inoue. 2007. Tissue-specific splicing regulator Fox-1 in- host model. J. Oral Microbiol. 4: 10368. duces exon skipping by interfering E complex formation on the downstream 2. Williams, M. J. 2007. Drosophila hemopoiesis and cellular immunity. intron of human F1gamma gene. Nucleic Acids Res. 35: 5303–5311. J. Immunol. 178: 4711–4716. 32.Lee,J.A.,A.Damianov,C.H.Lin,M.Fontes,N.N.Parikshak, 3. Tepass, U., L. I. Fessler, A. Aziz, and V. Hartenstein. 1994. Embryonic origin of E.S.Anderson,D.H.Geschwind,D.L.Black,andK.C.Martin.2016. hemocytes and their relationship to cell death in Drosophila. Development 120: Cytoplasmic Rbfox1 regulates the expression of synaptic and autism-related 1829–1837. genes. Neuron 89: 113–128. The Journal of Immunology 11

33.Bhalla,K.,H.A.Phillips,J.Crawford,O.L.McKenzie,J.C.Mulley,H.Eyre, 50. Law, C. W., Y. Chen, W. Shi, and G. K. Smyth. 2014. voom: precision weights A. E. Gardner, G. Kremmidiotis, and D. F. Callen. 2004. The de novo chro- unlock linear model analysis tools for RNA-seq read counts. Genome Biol. 15: mosome 16 translocations of two patients with abnormal phenotypes (mental R29. retardation and epilepsy) disrupt the A2BP1 gene. J. Hum. Genet. 49: 308– 51. Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: a 311. practical and powerful approach to multiple testing. J. Roy. Stat. Soc. B 57: 289– 34. Martin, C. L., J. A. Duvall, Y. Ilkin, J. S. Simon, M. G. Arreaza, K. Wilkes, 300. A. Alvarez-Retuerto, A. Whichello, C. M. Powell, K. Rao, et al. 2007. Cyto- 52. Goto, A., T. Kadowaki, and Y. Kitagawa. 2003. Drosophila hemolectin gene is genetic and molecular characterization of A2BP1/FOX1 as a candidate gene for expressed in embryonic and larval hemocytes and its knock down causes autism. Am. J. Med. Genet. B. Neuropsychiatr. Genet. 144B: 869–876. bleeding defects. Dev. Biol. 264: 582–591. 35. Venables, J. P., J. P. Brosseau, G. Gadea, R. Klinck, P. Prinos, J. F. Beaulieu, 53. Leulier, F., and B. Lemaitre. 2008. Toll-like receptors–taking an evolutionary E. Lapointe, M. Durand, P. Thibault, K. Tremblay, et al. 2013. RBFOX2 is an approach. Nat. Rev. Genet. 9: 165–178. important regulator of mesenchymal tissue-specific splicing in both normal and 54. Takehana, A., T. Katsuyama, T. Yano, Y. Oshima, H. Takada, T. Aigaki, and S. Kurata. 2002. Overexpression of a pattern-recognition receptor, cancer tissues. Mol. Cell. Biol. 33: 396–405. peptidoglycan-recognition protein-LE, activates imd/relish-mediated antibacte- 36. Homsy, J., S. Zaidi, Y. Shen, J. S. Ware, K. E. Samocha, K. J. Karczewski, rial defense and the prophenoloxidase cascade in Drosophila larvae. Proc. Natl. S. R. DePalma, D. McKean, H. Wakimoto, J. Gorham, et al. 2015. De novo Acad. Sci. USA 99: 13705–13710. mutations in congenital heart disease with neurodevelopmental and other con- 55. Akhouayri, I., C. Turc, J. Royet, and B. Charroux. 2011. Toll-8/Tollo negatively genital anomalies. Science 350: 1262–1266. regulates antimicrobial response in the Drosophila respiratory epithelium. PLoS 37. Tastan, O. Y., J. Z. Maines, Y. Li, D. M. McKearin, and M. Buszczak. 2010. Pathog. 7: e1002319. Drosophila ataxin 2-binding protein 1 marks an intermediate step in the mo- 56. Minakhina, S., J. Yang, and R. Steward. 2003. Tamo selectively modulates nu- lecular differentiation of female germline cysts. Development 137: 3167– clear import in Drosophila. Genes Cells 8: 299–310. 3176. 57. Biswas, S. K., and A. Mantovani. 2012. Orchestration of metabolism by mac- 38. Usha, N., and L. S. Shashidhara. 2010. Interaction between Ataxin-2 binding rophages. Cell Metab. 15: 432–437. protein 1 and cubitus-interruptus during wing development in Drosophila. Dev. 58. Attrill, H., K. Falls, J. L. Goodman, G. H. Millburn, G. Antonazzo, A. J. Rey, and Biol. 341: 389–399. S. J. Marygold, FlyBase Consortium. 2016. FlyBase: establishing a gene group 39. Fogel, B. L., E. Wexler, A. Wahnich, T. Friedrich, C. Vijayendran, F. Gao, resource for Drosophila melanogaster. Nucleic Acids Res. 44(D1): D786–D792. Downloaded from N. Parikshak, G. Konopka, and D. H. Geschwind. 2012. RBFOX1 regulates both 59. Vogel, C., S. A. Teichmann, and C. Chothia. 2003. The immunoglobulin su- splicing and transcriptional networks in human neuronal development. Hum. perfamily in Drosophila melanogaster and Caenorhabditis elegans and the Mol. Genet. 21: 4171–4186. evolution of complexity. Development 130: 6317–6328. 40. Koizumi, K., H. Higashida, S. Yoo, M. S. Islam, A. I. Ivanov, V. Guo, P. Pozzi, 60. Armitage, S. A., R. Y. Freiburg, J. Kurtz, and I. G. Bravo. 2012. The evolution of S. H. Yu, A. C. Rovescalli, D. Tang, and M. Nirenberg. 2007. RNA interference Dscam genes across the arthropods. BMC Evol. Biol. 12: 53. screen to identify genes required for Drosophila embryonic nervous system 61. Schmucker, D., J. C. Clemens, H. Shu, C. A. Worby, J. Xiao, M. Muda, development. Proc. Natl. Acad. Sci. USA 104: 5626–5631. J. E. Dixon, and S. L. Zipursky. 2000. Drosophila Dscam is an axon guidance

41. Guven-Ozkan, T., G. U. Busto, S. S. Schutte, I. Cervantes-Sandoval, receptor exhibiting extraordinary molecular diversity. Cell 101: 671–684. http://www.jimmunol.org/ D. K. O’Dowd, and R. L. Davis. 2016. MiR-980 is a memory suppressor 62. Wojtowicz, W. M., J. J. Flanagan, S. S. Millard, S. L. Zipursky, and microRNA that regulates the autism-susceptibility gene A2bp1. Cell Reports 14: J. C. Clemens. 2004. Alternative splicing of Drosophila Dscam generates axon 1698–1709. guidance receptors that exhibit isoform-specific homophilic binding. Cell 118: 42. Gehman, L. T., P. Stoilov, J. Maguire, A. Damianov, C. H. Lin, L. Shiue, 619–633. M. Ares, Jr., I. Mody, and D. L. Black. 2011. The splicing regulator Rbfox1 63. Graveley, B. R. 2005. Mutually exclusive splicing of the insect Dscam pre- (A2BP1) controls neuronal excitation in the mammalian brain. Nat. Genet. 43: mRNA directed by competing intronic RNA secondary structures. Cell 123: 706–711. 65–73. 43. Ayroles, J. F., M. A. Carbone, E. A. Stone, K. W. Jordan, R. F. Lyman, 64. Park, J. W., K. Parisky, A. M. Celotto, R. A. Reenan, and B. R. Graveley. 2004. Identification of alternative splicing regulators by RNA interference in Dro- M. M. Magwire, S. M. Rollmann, L. H. Duncan, F. Lawrence, R. R. Anholt, and sophila. Proc. Natl. Acad. Sci. USA 101: 15974–15979. T. F. Mackay. 2009. Systems genetics of complex traits in Drosophila 65. Watson, F. L., R. Pu¨ttmann-Holgado, F. Thomas, D. L. Lamar, M. Hughes, melanogaster. Nat. Genet. 41: 299–307.

M. Kondo, V. I. Rebel, and D. Schmucker. 2005. Extensive diversity of Ig- by guest on September 27, 2021 44. Koundakjian, E. J., D. M. Cowan, R. W. Hardy, and A. H. Becker. 2004. The superfamily proteins in the immune system of insects. Science 309: 1874–1878. Zuker collection: a resource for the analysis of autosomal gene function in 66. Dong, Y., H. E. Taylor, and G. Dimopoulos. 2006. AgDscam, a hypervariable Drosophila melanogaster. Genetics 167: 203–206. immunoglobulin domain-containing receptor of the Anopheles gambiae innate 45. Clark, R. I., K. J. Woodcock, F. Geissmann, C. Trouillet, and M. S. Dionne. immune system. PLoS Biol. 4: e229. b 2011. Multiple TGF- superfamily signals modulate the adult Drosophila im- 67. Patro, R., G. Duggal, M. I. Love, R. A. Irizarry, and C. Kingsford. 2017. Salmon mune response. Curr. Biol. 21: 1672–1677. provides fast and bias-aware quantification of transcript expression. Nat. Methods 46. Bolger, A. M., M. Lohse, and B. Usadel. 2014. Trimmomatic: a flexible trimmer 14: 417–419. for Illumina sequence data. Bioinformatics 30: 2114–2120. 68. Trincado, J. L., J. C. Entizne, G. Hysenaj, B. Singh, M. Skalic, D. J. Elliott, and 47. Liao, Y., G. K. Smyth, and W. Shi. 2013. The subread aligner: fast, accurate and E. Eyras. 2018. SUPPA2: fast, accurate, and uncertainty-aware differential scalable read mapping by seed-and-vote. Nucleic Acids Res. 41: e108. splicing analysis across multiple conditions. Genome Biol. 19: 40. 48. Liao, Y., G. K. Smyth, and W. Shi. 2014. featureCounts: an efficient general 69. Kucherenko, M. M., and H. R. Shcherbata, 2018. Stress-dependent miR-980 purpose program for assigning sequence reads to genomic features. Bioinformatics regulation of Rbfox1/A2bp1 promotes ribonucleoprotein granule formation 30: 923–930. and cell survival. Nat. Commun. 9: 312. 49. Bolstad, B. M., R. A. Irizarry, M. Astrand, and T. P. Speed. 2003. A comparison 70. Kafasla, P., A. Skliris, and D. L. Kontoyiannis. 2014. Post-transcriptional co- of normalization methods for high density oligonucleotide array data based on ordination of immunological responses by RNA-binding proteins. Nat. Immunol. variance and bias. Bioinformatics 19: 185–193. 15: 492–502.