NF-κB factors cooperate with Su(Hw)/E4F1 to balance Drosophila/human immune responses via modulating dynamic expression of miR-210

Abstract

MicroRNAs (miRNAs) play crucial regulatory roles in controlling immune responses, but their dynamic expression mechanisms are poorly understood. Here, we firstly confirm that the conserved miRNA miR-210 negatively regulates innate immune responses of Drosophila and human via targeting Toll and TLR6, respectively. Secondly, our findings demonstrate that the expression of miR-210 is dynamically regulated by NF-κB factor Dorsal in immune response of Drosophila Toll pathway. Thirdly, we find that Dorsal-mediated transcriptional inhibition of miR-210 is dependent on the transcriptional repressor Su(Hw). Mechanistically, Dorsal interacts with Su(Hw) to modulate cooperatively the dynamic expression of miR-210 in a time- and dose-dependent manner, thereby controlling the strength of Drosophila Toll immune response and maintaining immune homeostasis. Fourthly, we reveal a similar mechanism in human cells, where NF-κB/RelA cooperates with E4F1 to regulate the dynamic expression of hsa-miR-210 in the TLR immune response. Overall, our study reveals a conservative regulatory mechanism that maintains animal innate immune homeostasis and provides new insights into the dynamic regulation of miRNA expression in immune response.

Graphical Abstract

Graphical Abstract.

Introduction

The Toll signaling pathway, initially discovered in Drosophila melanogaster, has crucial and evolutionarily conserved functions in innate immune responses across the animal kingdom (1,2). Drosophila Toll pathway is mainly responsible for resisting Gram-positive (G+) bacteria or fungi infections (3). Activation of this pathway induces the nuclear translocation of NF-κB family transcription factors Dorsal and/or Dif (Dorsal-related immunity factor) to activate the expression of specific antimicrobial peptide (AMP) genes such as Drosomycin (Drs) (3,4). However, imbalanced immune responses, either excessive or deficient, can be detrimental to the survival of Drosophila (5,6). Therefore, Drosophila needs to fine-tune the strength and duration of Toll signaling to maintain immune homeostasis (7). Notably, effective and adequate immune responses of Toll pathway can be intricately regulated by many positive and negative regulators in Drosophila (8).

MicroRNAs (miRNAs) as a class of small non-coding RNAs can negatively regulate gene expression to participate in controlling immune responses and maintaining immune homeostasis in animal (9). Currently, some miRNAs, such as miR-8, miR-964, miR-958, miR-317 and miR-959-962, have been demonstrated to negatively regulate Drosophila Toll innate immune responses by targeting various genes (10–14). However, how these expressions of immune-specific miRNAs themselves are governed remain unclear. Previous studies have indicated that miRNA expression is usually time-dependent and tissue-specific, which can be controlled by transcription factors (15). For instance, RelA as one human NF-κB factor can directly bind to the promoter regions of certain miRNAs to activate their transcriptions in the biliary epithelial cells (16). However, little is known about whether the Drosophila NF-κB factor Dorsal, a core component of Toll pathway controlling the expression of AMPs (17), can also regulate the expression of miRNA during Drosophila innate immune responses.

The Suppressor of Hairy-wing [Su(Hw)] is a protein containing 12 zinc finger architectural domains and is known as the best-characterized insulator protein in Drosophila (18). In addition to establishing the chromatin insulator of the gypsy retrotransposon (18–21), Su(Hw) may also function as a transcriptional suppressor to repress protein-coding gene transcription in Drosophila (22–24). Moreover, studies have shown that most transcription factors lack the effector activity necessary for inducing gene expression, thereby needing to interact with cofactors to regulate cooperatively gene transcription (25,26). For example, NF-κB/RelA can interact with YY1, a member of the Polycomb Group (PcG), to epigenetically suppress miR-29 expression (27). Interestingly, a recent study using yeast 2-hybrid (Y2H) assays has demonstrated that Drosophila NF-κB transcription factors can interact with Su(Hw) protein (28). These above studies led us to speculate that the NF-κB factor Dorsal may interact with transcriptional suppressor Su(Hw) to cooperatively regulate the miRNA expression in Drosophila Toll pathway immune responses.

To explore this hypothesis, in the present study, based on the bioinformatic data we firstly predicted the interaction of Su(Hw) and Dorsal with the promoter region of an immune-related miR-210 from our previous small RNA-seq study (29), and confirmed that miR-210 was involved in Drosophila Toll immune response via directly targeting Toll in vivo and in vitro. We next demonstrated that infection-induced Dorsal could bind to and repress the promoter activity of miR-210 in vivo and in vitro, and Dorsal-mediated inhibition required the cooperation with the repressor Su(Hw). Furthermore, the dynamic expression profiles of Dorsal, miR-210, Toll, Drs and su(Hw) in wild-type flies across 9 time points after M. luteus infection provided insights into their critical roles in maintaining strength and persistence of immune response as well as immune homeostasis. Importantly, we also validated that the human NF-κB factor RelA can interact with the zinc-finger protein E4F1 to regulate the dynamic expression of hsa-miR-210, thereby controlling the expression level of its target gene TLR6 to maintaining immune homeostasis in human cell. Overall, our study reveals an evolutionarily conserved mechanism by which NF-κB factors interact with co-repressors to regulate the dynamic expression of miRNA for governing innate immune responses and homeostasis in Drosophila and human.

Materials and methods

Acquisition, construction and maintenance of Drosophila strains

Flies were reared on the standard cornmeal/agar/yeast medium at 25°C and 60% humidity. To generate transient overexpressing flies, we performed crosses between UAS-gene transgenic strains and the Tub-Gal80ts; Tub-Gal4 strain. To minimize lethality during early developmental stages due to the gene requirements in development, all flies were raised in a light-dark (12 h cycle) incubator at a permissive temperature (18°C) until adult progeny of the appropriate genotype emerged. Subsequently, these 4–6 day-old progeny were shifted to restrictive temperature (29°C) for 24 h to induce the overexpression of genes and then infection.

These fly strains, including w¹¹¹⁸ (BDSC: #3605), Tub-Gal80^ts; TM2/TM6B (BDSC: #7019), Tub-Gal4/TM3, Sb1, Ser1 (BDSC: #5138), UAS-Dorsal (BDSC: #9319), UAS-miR-210 (BDSC: #41179), miR-210 KO (BDSC: #58899), Drs-GFP (BDSC: #55707), su(Hw) null mutant (Df(3R)su(Hw)⁷/TM6B, Tb^[1]) (BDSC: #1049), su(Hw) null mutant (w^[1118]; PBac{w[+mC]=RB}su(Hw)^[e04061]/TM6B, Tb^[1]) (BDSC:#18224), Dorsal-GFP (BDSC:#42677), su(Hw)-GFP (BDSC:#64794), su(Hw)-RNAi (BDSC:#33906), were obtained from the Bloomington Drosophila Stock Center, FlyORF and Tsinghua Fly Center.

To construct the w*; P{miR-210-EGFP} attP40 fly strain, 1968 bp of DNA immediately upstream of the transcription start site of miR-210 gene was cloned into Not I/Nde I sites of promoter-EGFP (from Qidong Fungene Biotechnology), which has the attB site and miniwhite marker was used as the backbone. Then 200 embryos were injected using the above plasmid. When the injected P0 embryos grew into adults, they were crossed with Sp/CyO. F1 flies were screened for candidates that carried miniwhite marker in eyes (orange eye). PCR was performed using primers for validation of miR-210-EGFP integration. w*; P{miR-210-Dorsal-motif-mut-EGFP} attP40 and w*; P{miR-210-su(Hw)-motif-mut-EGFP} attP40 were also constructed by the above method, but the motif sequences of Dorsal and Su(Hw) were mutated during the plasmid construction.

GFP fluorescence assay and quantification

To verify the AMP activity of Drs, UAS-miR-210 and miR-210 KO flies were crossed with the Drs-GFP strain, respectively. To evaluate the promoter activity of miR-210-promoter, miR-210-promoter-Dorsal-motif-mut and miR-210-promoter-su(Hw)-motif-mut in control and Dorsal overexpression flies in vivo post-infection, UAS-Dorsal and Tub-Gal4 flies were crossed with miR-210-EGFP, miR-210-Dorsal-motif-mut-EGFP and miR-210-su(Hw)-motif-mut-EGFP flies, respectively. The GFP fluorescence intensity of 6–10 flies per group was visualized and photographed using fluorescence microscope (Nikon, Japan). ImageJ and GraphPad 8.3 were used for analysis and quantification of the GFP fluorescence intensity.

Bioinformatics analysis

The mature sequences of miR-210 and hsa-miR-210 were downloaded from miRBase (http://www.mirbase.org/). These 3’UTRs of immune-related genes and the promoter sequences of miR-210 were acquired from the FlyBase (http://flybase.org/) (30). These target genes of miR-210 were predicted using TargetScan (www.targetscan.org/fly_12/) and miRanda v3.3 with default parameters (31,32). The promoter sequences of hsa-miR-210 were acquired from NCBI (https://www.ncbi.nlm.nih.gov/) and the target genes of hsa-mir-210 were predicted from the miRWalk database (http://mirwalk.umm.uni-heidelberg.de/). We predicted these transcription factors that may bind the upstream promoter region of miR-210 and hsa-miR-210 based on the PROMO database (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3/) (33), FIMO (https://meme-suite.org/meme/tools/fimo), Jarspar (http://jaspar.genereg.net/) and TransmiR 2.0 database (http://www.cuilab.cn/transmir). The Venn diagram was drawn on the jvenn website (http://jvenn.toulouse.inra.fr/app/index.html) (34). The regulatory network diagram was drawn by the Cytoscape_v3.7.2. The evolutionary tree was constructed using the IQ-TREE software and visualized using the FigTree v1.4.4. Furthermore, the miRNA promoter sequences were predicted through the BDGP (https://www.fruitfly.org/seq_tools/promoter.html). The human homologous gene is screened using DIOPT v8.0 (http://www.flyrnai.org/diopt). The ChIP-seq of Drosophila transcription factors were obtained from ENCODE and modENCODE project (https://www.encodeproject.org/) (35,36). The ChIP-seq peak analysis was conducted by utilizing the ChIPseeker package integrated with R version 4.2.1 (37). Visualization of Dorsal/Su(Hw) ChIP-seq peaks was performed using IGV version 2.9.4 (38).

Recombinant plasmid construction

The pAc5.1/V5-HisA expression vector (Thermo, USA) was used for constructing the recombinant plasmids. And the luciferase coding sequence was subcloned into pAc5.1/V5-HisA to generate pAc-luc as previous (39). To produce exogenous Dorsal, Su(Hw) in Drosophila S2 cells, the CDS sequences of Dorsal, Flag-su(Hw) were amplified and cloned into pAc5.1/V5-HisA vector to generate pAc-Dorsal-V5, pAc-Flag-su(Hw) plasmids using ClonExpress II One Step Cloning Kit (Vazyme, China). The 3’ UTR sequences of Toll with or without miR-210 binding sites were amplified by PCR and cloned into pAc-luc to generate pAc-luc-Toll-3’ UTR and pAc-luc-Toll-3’ UTR-mut. The promoter region of miR-210 was predicted by BDGP website (https://www.fruitfly.org/seq_tools/promoter.html). We constructed the promoter region, containing about 2 kb upstream from miR-210 TSS site, into the pGL3-Basic plasmid (Promega, USA). To exclude the influence of non-specific effects, we constructed the antisense strand sequence of the hsa-miR-210 promoter into the pGL3-Basic plasmid (pGL3-hsa-miR-210-promoter-AS). To explore the cooperative regulation of NF-κB and the zinc finger repressor Su(Hw)/E4F1 on the miR-210 promoter, we deleted the NF-κB and the zinc finger repressor Su(Hw)/E4F1 motifs in the miR-210 promoter region using overlap PCR. The 3' UTR sequence of TLR6 (approximate 4 kb) was extracted from NCBI (https://www.ncbi.nlm.nih.gov/) and the 3' UTR sequence of TLR6 with or without hsa-miR-210 binding sites were constructed into psiCHECK-2 plasmid. The CDS sequence of RelA and E4F1 were extracted from NCBI (https://www.ncbi.nlm.nih.gov/) and constructed into CMV-3xFlag and pcDNA3.0-HA plasmids, respectively. All construction ones were verified by Sanger sequencing. All primers used in this analysis were listed in Supplementary Table S2.

Septic injury and survival experiments

These 4–6 days adult male flies were used for septic injury experiments. The control flies and the flies with over-expressed or knockdown gene/miRNA were infected with Micrococcus luteus. Septic injury experiments were performed by pricking the thorax of the flies with a pulled glass capillary carrying M. luteus inoculant using a Nanoject apparatus (Nanoliter 2010, WPI). Next, the flies were collected at specified time-points for subsequent experiments. Survival of infection is the most comprehensive approach to evaluate these immune response deficiencies (40). For survival experiment, the flies were infected with a concentrated culture of Enterococcus faecalis as pricking above, and the fly survival rate of 24 h was then recorded in detail.

Mimics infection and rescue effect verification

For the mimics infection experiment, a pulled glass capillary carrying 2 μM mimics-miR-210 or negative control mimics-NC (the mimics-NC should be designed based on the sequence of the mimics-miRNA, modifying the mimics-miRNA sequence to contain at least 5-nt mismatches at the 5’-end ‘seed site (41)) was injected into the thorax of 4-day-old adult fruit flies. Prior to injection with M. luteus, the fruit flies were injected with mimics-NC/mimics-miR-210 at least 24 h in advance. Samples were then collected at corresponding time points thereafter. To ensure the rescue effect of mimics-miR-210, we examined the triglyceride (TG) content and the Nile red-stained lipid droplets in the retinas of 5-day-old w¹¹¹⁸, miR-210 KO and miR-210 rescued (miR-210 KO + mimics-miR-210) flies as previously described (42). For TG content analysis, 0.1g of retinas was weighed, and TG was extracted and measured following the instructions of TG Content Assay Kit (Solarbio Life Sciences, #: BC0625, China). To examine lipid droplets in the retinas, the heads of w¹¹¹⁸, miR-210 KO and miR-210 rescued flies were removed from adult flies and fixed in 4% formaldehyde for 1 h at room temperature. After removing the excess tissues, the retinas were fixed for an additional 30 minutes. Fixed retinas were washed three times (20 min each time) with PBS and then incubated with 1 mg/ml Nile red (Sigma 72485-100MG, 1:1000) for 20 min at room temperature. Subsequently, the retinas were washed three more times (20 min each time). Images were acquired using a Nikon AX confocal microscope.

RNA extraction and RT-qPCR

First, we collected these treated Drosophila adults in five per groups or the treated human cells in 6-well plates at specific time points in DNase/RNase free 1.5 ml EP tubes (Axygen, USA) and quickly frozen in liquid nitrogen. The total RNA was then extracted in RNA isolators Total RNA Extraction Reagent (Vazyme, China) using a tissue grinder (TIANGEN, China). Total RNA dissolved in DEPC-treated water was tested for concentration and purity with NanoDrop 2000 (Thermo, USA). The samples with A260/A280 of ∼2.0 and A260/A230 in the range of 1.8∼2.2 were used to reverse transcription. For RT-PCR, cDNA was prepared by 1 μg RNA each sample using HiScript II Q RT SuperMix for qPCR (Vazyme, China) with 50°C for 15 min and 85°C for 5 s. The stem-loop primers were synthesized for reverse transcription to generate the specific stem-loop cDNA of miRNA. The quantitative PCR reaction was performed on the Roche LightCycler® 96 real-time PCR system (Roche, Switzerland) using AceQ SYBR Green Master Mix (Vazyme, China). The RT-qPCR cycling conditions were: step 1: 95°C for 5 min; step 2: 95°C for 10 s; step 3: 60°C for 30 s, then steps 2 and 3 were cycled for 40 times. Both mRNA and miRNA expression levels were normalized to the control rp49/GAPDH and U6 snRNA, respectively. All experiments were performed with three biological replicates and three technical replicates. The relative 2^−△△CT method was used for data analysis (43). All primers used in this analysis were listed in Supplementary Table S3. All data collected from real-time PCR analysis were presented as means ± SD.

Cell culture, transfection and immune stimulation

Drosophila S2 cells were maintained at 28°C in SFX-insect medium (HyClone, USA) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, USA). The 293T cells were cultured under standard conditions in Dulbecco's modified Eagle's medium (Wisent, China), 10% fetal bovine serum, 1% penicillin/streptomycin. The THP1 cells were grown in suspension in RPMI/FBS media (RPMI supplemented with 1% of penicillin–streptomycin solution and 15% FBS). The 293T and THP1 cell lines were cultured at 37°C in a 5% CO₂ incubator and used within 6 months.

S2 cells, 293T cells and THP1 cells were transiently transfected with the transfection complex in 24-well plates for dual luciferase reporter assay and in 6-well plates for Co-IP or western blot, using X-treme Gene HP Transfection Reagent (Roche, Switzerland). The specific steps of transfection refer to the HP transfection reagent instructions.

THP1 is the monocyte using for studying human innate immunity and can be induced by phorbol-12-myristate-13-acetate (PMA) to become human monocyte-derived macrophages (MDMs) (44), as well as the heat-killed bacteria can induce the innate immune response (45). For immune stimulation, the THP-1 cells were treated with 10 ng/ml of 12-O-tetradecanoylphorbol-l3-acetate (PMA) for 36 h. Activated THP-1 cells were differentiated after 36 h in PMA, and then stimulated with PBS or heat-killed M. luteus [multiplicity of infection (MOI) = 50] (45).

Dual luciferase reporter assay

To investigate whether Toll is a target gene of miR-210, S2 cells were transfected with the mixed transfection complex contained either mimics-miR-210 or mimics-NC, either pAc-Luc-Toll-3′UTR or pAc-Luc-Toll-3′UTR-mut and Renilla luciferase plasmid (pRL) for a total volume of 50 μl. Promega pRL was used to normalize the transfection efficiency. To explore whether Dorsal and Su(Hw) could regulate the promoter activity of miR-210, we transfected pGL3-miR-210-promoter/pGL3-miR-210-promoter-Dorsal-motif-mut/ pGL3-miR-210-promoter-su(Hw)-motif-mut, pAc Vector encoding Dorsal and su(Hw) at different concentrations into S2 cells. For human cells, psiCHECK-2-TLR6-3’UTR/ psiCHECK-2-TLR6-3’UTR-mut, mimics-hsa-miR-210/mimics-NC or pGL3-hsa-miR-210-promoter/pGL3-hsa-miR-210-promoter-RelA-motif-mut/pGL3-hsa-miR-210-promoter-E4F1-motif-mut, CMV-Flag-RelA, pcDNA3.0-HA-E4F1 were transfected into 293T cells as described above. The luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, USA) according to manufacturer's instructions and the Renilla luciferase activity was used as a control for normalization.

Chromatin immunoprecipitation (ChIP)

S2 cells transfected with pAc-Dorsal-V5 were used for ChIP in vitro, while Dorsal-GFP, Su(Hw)-GFP, Su(Hw)-GFP with Dorsal overexpression, Su(Hw)-GFP with Su(Hw) knock down flies and THP1 cells transfected with CMV-Flag-RelA and pcDNA3.0-HA-E4F1 post-infection were prepared for ChIP in vivo. The protocol of ChIP is as described in the previous research (39). Briefly, for the ChIP experiment, the cells and ground GFP-tagged flies were fixed at room temperature by crosslinking with a final concentration of 1% formaldehyde solution for 10 min and then quenched with 125 mM glycine for 5 min. After washing twice with cold PBS containing the protease inhibitor cocktail and PMSF, the cells were lysed with cell lysis buffer and nuclear lysis buffer. The clarified lysates were sonicated. The chromatin was then sheared to 200–500 bp. The chromatin was used for ChIP incubating with Dynabeads protein G (Thermo Fisher Scientific, USA) coated with either anti-V5 antibody (ABclonal, #AE017), anti-GFP antibody (abcam, #ab290, USA), anti-Dorsal antibody (DSHB, #7A4-39, USA), anti-RelA Rabbit pAb (ABclonal #A2547, China), anti-E4F1 (D-12) monoclonal antibody (SANTA # sc-514718, USA), rabbit IgG control antibody (ABclonal, #AC005, China) or mouse IgG control antibody (ABclonal, #AC011, China) overnight at 4°C on a rotating platform. After repeated washing with a magnetic rack (Thermo Fisher Scientific, USA), genomic DNA bound to Dorsal/Su(Hw)/RelA/E4F1 was eluted from Dynabeads and then reversed cross-linked overnight at 65°C. The DNA fragments were then purified using AxyPrep PCR Cleanup Kit (Axygen, USA). ChIP-qPCR analysis was performed using DNA from the Input and ChIP experiments using the primers listed in Supplementary Table S4. At least three independent experiments and three technical replicates were performed on the rp49, miR-210 and Drs promoters. The rp49 promoter was used as a negative control and the Drs promoter as a positive control.

Western blot

Protein samples were prepared with RIPA (50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.1% SDS), and separated by 10% SDS-PAGE gel and then transferred to PVDF membranes. The membranes were blocked with 5% non-fat milk and incubated with the primary antibodies at 4°C overnight. GAPDH primary antibody (36 kDa) (ABclonal #AC001, China), Tubulin β polyclonal antibody (Biogot #AP0064, China), anti-V5 antibody (ABclonal, #AE017, China), anti-Flag antibody (ABclonal, #AE005, China), anti-HA antibody (ABclonal, #AE008, China), anti-RelA Rabbit mAb (ABclonal #A19653, China), anti-p-RelA (27. Ser 536) Rabbit mAb (ABclonal #AP1294, China), anti-GFP antibody (abcam, #ab290, USA), anti-Dorsal 7A4 (DSHB #AB_528204, USA) and E4F1 (D-12) monoclonal antibody (SANTA # sc-514718, USA) were used at suitable dilution ratio according to the instruction. Then the PVDF membranes were incubated for 1 h∼2 h at room temperature with 1:20,000 horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody and goat anti-mouse secondary antibody (Biogot #BS13278 #BS12478, China) diluent buffer. The antigen-antibody complex was visualized with an imaging instrument (Li-COR, USA and Clinx, China). The gray value of protein was analyzed using ImageJ.

Co-immunoprecipitation (Co-IP)

In vitro Co-IP experiments were performed using S2 cells transfected and co-transfected with pAc-Dorsal-V5 and pAc-Flag-su(Hw) plasmids, respectively. For in vivo Co-IP experiments, Su(Hw)-GFP flies, with or without infection, were used. Moreover, the 293T cells were transfected and co-transfected with CMV-3xFlag-RelA and pcDNA3.0-HA-E4F1 expression plasmids, respectively, for Co-IP experiments in vitro and THP1-induced macrophages, with or without infection, were used for Co-IP experiments in vivo. In the Co-IP procedure, 2 × 10⁷ cells or five flies were collected and suspended in RIPA buffer supplemented with protease inhibitor cocktail (Roche, USA) and PMSF on ice for 30 min. After centrifuging at 13 000 rpm for 10 min at 4°C, the lysates (equal allocation for each IP and Input experiment) were incubated with anti-V5 tag antibody (ABclonal, #AE017, China), anti-Flag tag antibody (ABclonal, #AE005, China), anti-HA tag antibody (ABclonal, #AE008), anti-RelA Rabbit mAb (ABclonal #A19653, China), anti-E4F1 (D-12) monoclonal antibody (SANTA # sc-514718, USA), anti-GFP antibody (abcam, #ab290, USA) and anti-Dorsal 7A4 (DSHB # AB_528204, USA) at 4°C overnight, followed by further incubation with Protein A/G PLUS-Agarose (SANTA,# sc-2003) for 2 h. The beads were washed five times with PBST buffer. The proteins were eluted by Co-IP Elution Buffers (Abmart #T10007, China) and analyzed by western blot with anti-V5 tag antibody (ABclonal, #AE017, China), anti-Flag tag antibody (ABclonal, #AE005, China), anti-HA tag antibody (ABclonal, #AE008), anti-RelA Rabbit mAb (ABclonal #A19653, China), anti-E4F1 (D-12) monoclonal antibody (SANTA # sc-514718, USA), anti-GFP antibody (abcam, #ab290, USA) and anti-Dorsal 7A4 (DSHB # AB_528204, USA).

In Drosophila, the in vitro Co-IP experiments were conducted using S2 cells that IP with anti-IgG or overexpressed Dorsal-V5 and Flag-Su(Hw), respectively, serving as a control. The total protein amount in the input sample was determined using anti-Tubulin. The output had been shown to evaluate the effectiveness of the IP. While input and all IP groups had been shown on the same gel for evaluation. WB gray scale analysis via ImageJ was used to quantify the percentage of input relative to the IP. For in vivo Co-IP experiments, we equally divided the lysates into four aliquots for each IP/Input experiment and also used anti-IgG as a control. The effectiveness of the IP was assessed using the output from these experiments. Similar to the in vitro experiments, the input and all IP groups had been shown on the same gel for evaluation. WB gray scale analysis via ImageJ was again used to quantify the percentage of input relative to the IP. The Co-IP assay performed in human cells were consistent with flies.

RNA-immunoprecipitation (RIP)

The main RIP experimental steps refer to this protocol (46,47). In brief, about 2 × 10⁷ 293T cells transfected with mimics-NC or mimics-hsa-miR-210 were lysed in RIPA buffer (Beyotime Biotechnology, China) containing the PMSF (Beyotime Biotechnology, China), protease inhibitor cocktail (Roche, Switzerland) and RNase inhibitor (Thermo Fisher Scientific, USA) for 30 min on ice. The supernatants after high-speed centrifugation were pre-cleared for 2 h using protein A agarose (Invitrogen, Carlsbad, CA, USA) at 4°C and take 10% as ‘Input’. After pre-clearing, anti-Ago2 antibody (Thermo, #AB_2576596, USA) were added in the supernatants and incubated at 4°C overnight. The next day, protein A agarose was joined and rotating incubated for 2 h. The agarose beads were washed using RIPA buffer for five times. The IP complexes were eluted using TE buffer with 1% (w/v) SDS. The eluted complexes of Ago2 and RNA were treated with protease K to separate the protein-bound RNA, and RNA was extracted and quantified using RT-qPCR.

Quantification and statistical analysis

All experimental data in this work were collected from three independent biological replicates. All statistical analyses were presented as means ± SD. Significant differences between the values under different experimental conditions were subjected to two-tailed Student's t-test. Statistical analysis of fly survival experiments was performed using the log-rank (Mantel–Cox) test. All graphs were drawn by the Graphpad prism 8.3. For all tests, P value <0.05 was considered as statistically significant. *P < 0.05; **P < 0.01; ***P < 0.001 and ns, no significance versus the control groups.

Results

miR-210 negatively regulates Drosophila Toll signaling via directly targeting Toll

We firstly used different fly strains to explore the roles of miR-210 in immune responses. Our results demonstrated that the expression level of the signature antimicrobial peptide Drosomycin (Drs) mRNA in the miR-210 overexpression (miR-210 OE) flies (Gal80^ts/+; Tub-Gal4/UAS-miR-210) decreased by 20%, 30%, 25% compared to the control ones (Gal80^ts/+; Tub-Gal4/+) at 6 h, 12 h, 24 h after infection with M. luteus, respectively, with no significant differences between the transgenic fly strain (Gal80^ts/+; UAS-miR-210/+) and the control Gal4 flies (Gal80^ts/+; Tub-Gal4/+) (Figure 1A and Supplementary Figure S1A). In contrast, the expression level of Drs in the miR-210 knockout (miR-210 KO) flies increased by 22%, 50% and 70% at 6, 12 and 24 h after infection compared to the control ones (w¹¹¹⁸), respectively (Figure 1B). Of note, the expression level of Drs in the miR-210 rescued flies (miR-210 KO + mimics-miR-210) was almost restored to the level in the control flies (w¹¹¹⁸) (Figure 1B and Supplementary Figure S1B–G). Especially, similar results were also observed for other AMPs Defensin (Def), Drosocin (Dro) and Metchnikowin (Mtk) (Supplementary Figures S2A and B). Therefore, Drs was chosen as a representative AMP to detect Toll immune response in the following studies. Subsequently, we used the Drs-GFP reporter fly strains to examine the expression level of Drs protein, and observed a significantly reduced Drs-GFP intensity in the miR-210 OE flies (Drs-GFP/+; Gal80^ts/+; Tub-Gal4/UAS-miR-210) compared to the control flies (Drs-GFP/+; Gal80^ts/+; Tub-Gal4/+), with no significant difference between the two control files (Drs-GFP/+; Gal80^ts/+; UAS-miR-210/+ and Drs-GFP/+; Gal80^ts/+; Tub-Gal4/+) (Figure 1C and Supplementary Figure S2C). On the contrary, the Drs-GFP intensity in the miR-210 KO flies (Drs-GFP/miR-210 KO) was significantly increased compared to the control flies (Drs-GFP/+), and the Drs-GFP intensity in the miR-210 rescued flies (Drs-GFP/miR-210 KO + mimics-miR-210) returned to near-normal levels (Drs-GFP/+) (Figure 1D and Supplementary Figure S2D). These above results demonstrated an important role of miR-210 in negatively regulating Drosophila Toll immune responses.

Figure 1.

miR-210 negatively regulates Drosophila Toll signal via targeting Toll (A) RT-qPCR analysis of Drs 'expression in the control flies (Gal80^ts/+; Tub-Gal4/+ and Gal80^ts/+; UAS-miR-210/+) and the miR-210 overexpressed flies (Gal80^ts/+; Tub-Gal4/UAS-miR-210) at different time points after M. luteus infection. (B) RT-qPCR analysis of Drs expression in the control flies (w¹¹¹⁸), the miR-210 KO flies and the miR-210 rescued flies (miR-210 KO + mimics-miR-210) at different time points after M. luteus infection. (C) Fluorescence microscopy images show the fluorescence intensity of Drosomycin-green fluorescent protein (Drs-GFP) in control flies (left and middle) and the miR-210 overexpressed flies (right) at 12 h after M. luteus infection. (D) Fluorescence microscopy images show the fluorescence intensity of Drs-GFP in control flies (left), the miR-210 KO flies (middle) and the miR-210 rescued flies (right) at 12 h after M. luteus infection. (E) Predicted binding sites of miR-210 in 3'UTR region of Toll were identified using TargetScan software. The 3'UTR target sequences were mutated base on the miR-210 seed sequence. (F) Dual luciferase assay was performed in Drosophila S2 cells to measure the luciferase activities of mimics-miR-210 and a reporter plasmid with or without a mutation in the target site of Toll. (G) RT-qPCR analysis of 12 Toll pathway genes in control flies and the miR-210 overexpressed flies. (H) RT-qPCR analysis of 12 Toll pathway genes in control flies, miR-210 KO flies, and miR-210 rescued flies.

We further predicted potential targets of miR-210 (Supplementary Figure S2E), and found that only target gene Toll was the key effector of Toll signaling pathway among 16 potential targets identified (Supplementary Figure S2F). We next used the dual luciferase reporter assays to validate whether miR-210 can target Toll in Drosophila S2 cells. Our results showed the luciferase activity of pAc-luc-Toll-3’UTR was repressed by 60% using mimics-miR-210 transfection compared to the negative control (mimics-NC), which was abolished when the binding site was mutated in the 3' UTR of Toll, indicating that miR-210 could bind and repress Toll gene in vitro (Figure 1E and F). To exclude other potential targets of miR-210, we further detected the expression levels of 12 known Toll pathway genes (GNBP1, SPE, SPE, PGRP-SA, PGRP-SD, Toll, Myc88, Tube, Pelle, Cactus, Dif, drosal) using the aforementioned fly strains, confirming that only the expression level of Toll exhibited significant decreases in the miR-210 OE flies, increased in the miR-210 KO, and restored in the miR-210 rescued flies (Figure 1G and H). Accordingly, the expression level of four AMPs in the Toll-RNAi flies (Gal80^ts/+; Tub-Gal4/UAS-Toll-RNAi) was significantly lower than those in the control flies (Gal80^ts/+; Tub-Gal4/+) (Supplementary Figure S2G). Collectively, our results supported that miR-210 could negatively regulate Toll pathway only via directly targeting Toll.

Dorsal inhibits miR-210 transcription via binding to its promoter region

We identified two potential transcription start sites (TSS) at –505 and –177 upstream of the left arm of miR-210 gene based on the BDGP website, and predicted the binding of the NF-κB transcription factor Dorsal to the promoter region of miR-210 (Figures 2A, Supplementary Figure S3A, C and E). Meanwhile, the ChIP-seq data analysis revealed a distinct peak of Dorsal on the miR-210 promoter (Figures 2B). We thus constructed the two promoter fragments, i.e. the pGL3-miR-210-TSS1 of 1531 bp and the pGL3-miR-210-TSS2 of 1842 bp covering both TSS1 and TSS2, into the pGL3-Basic plasmid, respectively (Figures 2A). Our results demonstrated that the promoter activity of TSS2 was stronger than that of TSS1 (Figures 2C). Remarkably, the promoter activity of miR-210 was significantly decreased when co-transfecting exogenously expressed Dorsal (pGL3-miR-210-TSS-1/2 + pAc-Dorsal) compared to the controls (pGL3-miR-210-TSS-1/2 + pAc-empty) (Figures 2D). Consistently, the expression level of miR-210 was significantly decreased to about 60-70% in the Dorsal overexpressed flies (Gal80^ts/UAS-Dorsal; Tub-Gal4/+) compared to the control flies (Gal80^ts/+; Tub-Gal4/+) in vivo (Figures 2G). The known positive control Drs, one AMP gene known to be regulated by Dorsal (48), was used to solidify the accuracy of the above results in vitro and in vivo (Figures 2A, E, F and H). Additionally, the results of in vitro ChIP-qPCR experiments showed that the enrichment signal of Dorsal was 2.2% of Input for the positive control Drs promoter region, 1.2% of Input for the miR-210 promoter, and only 0.15% for the negative control rp49, whereas the enrichment signal of the negative control anti-IgG for the promoter region of rp49, Drs and miR-210 was all about 0.15% (Figures 2I). Consistently, these results from in vivo ChIP-qPCR experiments also demonstrated that the enrichment signal of Dorsal was 0.8% of Input for the positive control Drs promoter, 0.7% of Input for the miR-210 promoter, but only 0.15% for the negative control rp49 promoter (Figures 2J). Together, these above in vitro and in vivo results revealed that Dorsal can bind to the promoter region of miR-210 and inhibit miR-210 transcription.

Figure 2.

Dorsal inhibits miR-210 transcription via binding to its promoter region (A) Schematic diagram depicts the promoter region of miR-210 and Drs. The predicted transcription start sites (TSS) and primer positions were indicated by a purple line. Predicted Dorsal motifs (blue line) and Su(Hw) motifs (green line) were shown at the promoter region of miR-210. (B) Visualization of ChIP-seq data for Dorsal and Su(Hw) using IGV. The red box indicates the peaks of Dorsal and Su(Hw) at the miR-210 promoter region. (C) Promoter activities of miR-210-TSS1 and miR-210-TSS2 were determined in Drosophila S2 cell using a luciferase assay. (D) Promoter activities were determined for miR-210-TSS1 and miR-210-TSS2 in the presence or absence of exogenous Dorsal in Drosophila S2 cell using a luciferase assay. (E) Promoter activities of Drs were determined in Drosophila S2 cell using a luciferase assay. (F) Promoter activities were determined for Drs in the presence or absence of exogenous Dorsal in Drosophila S2 cell using a luciferase assay. Remarkably, Fig. 2C and 2E were normalized against pGL3-Basic, Fig. 2D was normalized to pGL3-miR-210-TSS2 promoter + pAc-empty and pGL3-miR-210-TSS1 promoter + pAc-empty respectively, and Fig 2F was normalized against pGL3-Drs promoter + pAc-empty. (G) RT-qPCR analysis of miR-210 was detected in the Dorsal overexpressed flies (Gal80^ts/UAS-Dorsal; Tub-Gal4/+), normalized to the controls (Gal80^ts/+; Tub-Gal4/+). (H) RT-qPCR analysis of Drs was detected in the Dorsal overexpressed flies (Gal80^ts/UAS-Dorsal; Tub-Gal4/+), normalized to the controls (Gal80^ts/+; Tub-Gal4/+). (I) ChIP-qPCR was performed to detect the fold change in Dorsal binding on the promoters of miR-210, Drs and rp49 in vitro. (J) ChIP-qPCR analysis of Dorsal binding on the promoters of miR-210, Drs and rp49 in vivo.

Dorsal inhibits miR-210 transcription to de-suppress Toll expression

The mRNA expression levels of Drs and target gene Toll were also further tested in the Dorsal overexpressed flies (Gal80^ts; Tub-Gal4>Dorsal + mimics-NC), the Dorsal + miR-210 co-overexpressed flies (Gal80^ts; Tub-Gal4>Dorsal + mimics-miR-210) and the controls (Gal80^ts; Tub-Gal4/+ + mimics-NC) with or without M. luteus infection. Our results showed that the expression level of Drs in the Dorsal overexpressed flies was 80 folds higher than that of the controls under no infection condition, especially the expression level of Drs in the Dorsal + miR-210 co-overexpressed flies was returned to 50% of that in the Dorsal overexpressed flies (Figures 3A, Supplementary Figure S4A and B). Similarly, the expression level of Toll was increased in the Dorsal overexpressed flies and restored to that of the controls in the Dorsal + miR-210 co-overexpressed flies (Figures 3B). Meanwhile, at the early stage of Drosophila immune response to M. luteus infection (3 h post-infection), we observed a similar expression profile for both Toll and Drs but with less altered folds for Drs and Toll (Figures 3C and D). These results demonstrated that Dorsal can promote Drosophila immune response through repressing miR-210 transcription to de-suppress target gene Toll in the early stage of Drosophila immune response.

Figure 3.

Dorsal inhibits miR-210 transcription to de-suppress Toll expression (A) The expression levels of Drs were examined in the control flies (Gal80^ts/+; Tub-Gal4/+ + mimics-NC), the Dorsal overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-NC), the Dorsal and miR-210 co-overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210) before infection. (B) The expression levels of Toll were examined in the control flies (Gal80^ts/+; Tub-Gal4/+ + mimics-NC), the Dorsal overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-NC), the Dorsal and miR-210 co-overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210) before infection. (C) The expression levels of Drs were examined in the control flies (Gal80^ts/+; Tub-Gal4/+ + mimics-NC), the Dorsal overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-NC), the Dorsal and miR-210 co-overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210) at 3 h after M. luteus infection. (D) The expression levels of Toll were examined in the control flies (Gal80^ts/+; Tub-Gal4/+ + mimics-NC), the Dorsal overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-NC), the Dorsal and miR-210 co-overexpressed flies (Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210) at 3 h after M. luteus infection. (E) Survival rate changes were monitored for 24 h in the Dorsal over-expressed flies, the Dorsal and miR-210 co-overexpressed flies and the control flies with PBS as well as E. faecalis infection. Survival rates: Gal80^ts; Tub-Gal4/+ + mimics-NC—E. faecalis (n = 110); Gal80^ts; Tub>UAS-Dorsal + mimics-NC—E. faecalis (n = 103); Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210—E. faecalis (n = 101); Gal80^ts; Tub-Gal4/+ + mimics-NC—PBS (n= 103); Gal80^ts; Tub>UAS-Dorsal + mimics-NC—PBS (n = 102); Gal80^ts; Tub>UAS-Dorsal + mimics-miR-210—PBS (n = 104).

Herein, we further examined whether miR-210 could affect the survival of the above fly strains upon infected using the lethal G+ bacteria E. faecalis. Our results indicated that the survival rate of the Dorsal overexpressed flies was significantly higher than that of the controls. Whereas, the survival rate of the Dorsal + miR-210 co-overexpressed flies was significantly shortened than the Dorsal overexpressed flies (Figures 3E). Overall, these above findings revealed that the fine-tune regulation of miR-210 expression by Dorsal is essential for maintaining Drosophila immune homeostasis and survival.

Su(Hw) is required for Dorsal-mediated transcriptional repression of miR-210

Previous studies have indicated the Dorsal-mediated transcriptional repression was dependent on co-factors such as groucho and AP-1 (49,50). Here, we predicted the binding of a known transcriptional suppressor Su(Hw) to the promoter of miR-210 based on the ENCODE database and the TransmiR v2.0 database (Figures 2A, 4A, Supplementary Figure S3 and Supplementary Table S1) (51,52), implying that Su(Hw) may be involved in Dorsal-mediated transcriptional repression of miR-210. Of note, our results demonstrated the increased expression level of miR-210 by ∼ 50% in two su(Hw) null mutation flies (Df(3R)su(Hw)⁷ and PBac{RB}su(Hw)^e04061) compared to the control flies (w¹¹¹⁸) in vivo post-infection (Supplementary Figure S4C). We next examined the effect of different concentrations of Dorsal and Su(Hw) on miR-210 promoter activity (pGL3-miR-210-TSS2) by transfecting serial dosages of pAc-Dorsal-V5/pAc-Flag-su(Hw) plasmids (500, 1000 and 1500 ng) into S2 cells. Our dual luciferase reporter results showed that miR-210 promoter activity could be suppressed by exogenously overexpressed either Dorsal or Su(Hw) in a dose-dependent manner (Supplementary Figure S4E and F). Especially, we found that the co-overexpression of Dorsal and Su(Hw) could result in a stronger inhibition of miR-210 promoter activity compared to their separate overexpression (Figure 4B). When the Dorsal binding motif in the miR-210 promoter was mutated, the overexpression of Su(Hw) alone or the co-overexpression of Dorsal and Su(Hw) significantly repressed the miR-210 promoter activity, but their inhibitory effects were similar (Figure 4C). Conversely, once the Su(Hw) binding motif was mutated, the overexpression of Dorsal alone or the co-overexpression of Dorsal and Su(Hw) both significantly enhanced the promoter activity of miR-210 (Figure 4D). Remarkably, there seemed to be no genetic indirect interaction between Dorsal and Su(Hw) due to no difference in su(Hw) mRNA levels between the Dorsal-OE flies and the control flies (Supplementary Figure S4G), as well as no significant change in the Dorsal expression level between the su(Hw) null mutant flies and the controls (Supplementary Figure S4H). These above results suggested that Dorsal-mediated transcriptional repression of miR-210 might depend on Su(Hw).

Figure 4.

Dorsal-mediated transcriptional repression of miR-210 depends on Su(Hw). (A) Flowchart illustrates the process of screening the transcription co-repressor Su(Hw) from the ENCODE database and TransmiR database. (B–D) The reporter activities of pGL3-miR-210-TSS2-promoter, the pGL3-miR-210-TSS2- promoter-Dorsal-motif-mut and the pGL3-miR-210-TSS2-promoter–su(Hw)-motif-mut were determined in Drosophila S2 cell using the luciferase assay, with exogenous Dorsal and Su(Hw) transfected separately or together. “+” indicated transfection with 500 ng of the plasmid, “++” indicated transfection with 1000 ng of the plasmid, “–” indicated no transfection of the plasmid. (E, G, I) Observation of miR-210-EGFP, miR-210-Dorsal-mut-EGFP, and miR-210-su(Hw)-mut-EGFP reporter gene fluorescence in control flies (left) and the Dorsal overexpressed flies (right) using a fluorescent microscope. (F, H, J) Quantification of EGFP intensity in the miR-210-EGFP, miR-210-Dorsal-mut-EGFP and miR-210-su(Hw)-mut-EGFP using imageJ (10 flies per group).

We further generated three Drosophila strains: miR-210-promoter-EGFP (miR-210-EGFP), miR-210-promoter-Dorsal-motif-mut-EGFP (miR-210-Dorsal-motif-mut-EGFP) and miR-210-promoter-su(Hw)-motif-mut-EGFP (miR-210-su(Hw)-motif-mut-EGFP), which were crossed with the Dorsal-OE flies and the control flies. We compared the GFP fluorescence intensity of these strains at 3 h post-infection, and found that Dorsal overexpression considerably suppressed GFP fluorescence intensity in the miR-210-EGFP flies compared to the controls (Figure 4E and F), but had no significant effect on the miR-210-Dorsal-motif-mut-EGFP flies (Figure 4G and H). In contrast, Dorsal overexpression significantly enhanced GFP intensity in the miR-210-su(Hw)-motif-mut-EGFP flies (Figure 4I and J). Taken together, these in vitro and in vivo results indicated that Su(Hw) is required for Dorsal-mediated suppression of miR-210 during Drosophila Toll immune response.

Infection-induced Dorsal interacts with Su(Hw) to suppress miR-210 transcription

The previous study reported a potential direct interaction between Su(Hw) and NF-κB family transcription factors using yeast 2-hybrid assays (28). Interestingly, we also observed the overlapped ChIP-seq peak between Dorsal and Su(Hw) within the promoter region of miR-210 (Figure 2B). We thus hypothesized that Dorsal and Su(Hw) may cooperate to regulate miR-210 expression.

To prove this hypothesis, we firstly co-transfected pAc-Dorsal-V5 and pAc-Flag-su(Hw) plasmids or alone, and then performed the immunoprecipitation (IP) assay using antibodies anti-V5 for Dorsal and anti-Flag for Su(Hw) in S2 cells. Our results showed that after IP with anti-Flag antibodies, western blot could clearly detect Dorsal-V5 and vice versa (Figure 5A, quantified in Supplementary Figure S7A and B). We further conducted an in vivo co-IP assay using antibodies against Dorsal and GFP in Su(Hw)-GFP flies, and found that there was no interaction between Dorsal and Su(Hw) under non-infected conditions (Figure 5B), but a clear interaction between Dorsal and Su(Hw) was detected at 3 h post-infection (Figure 5B, quantified in Supplementary Figure S7C and D). Additionally, we examined the signals of Dorsal and Su(Hw) within the miR-210 promoter region using anti-Dorsal or anti-GFP antibodies at 3 h post-infection in the Dorsal-OE (su(Hw)-GFP/UAS-Dorsal; Tub-Gal4/+) flies and the su(Hw)-RNAi (+/su(Hw)-GFP; Tub-Gal4/UAS-su(Hw)-RNAi) flies, respectively. These results demonstrated that Dorsal overexpression significantly increased the signal of Dorsal within the miR-210 promoter region, especially this increase was accompanied by a significant increase of the Su(Hw) signal within the miR-210 promoter region, implying that Dorsal could interact with Su(Hw) to bind to the miR-210 promoter region (Figure 5C and D). To further reveal whether perturbation of Su(Hw) affects Dorsal signals in the miR-210 promoter, we knocked down Su(Hw) in Drosophila and examined the signals of Dorsal and Su(Hw) in the miR-210 promoter. When Su(Hw) was knocked down, no significant difference in the Dorsal signal was observed, indicating that Su(Hw) has no regulatory effect on the Dorsal signal in the miR-210 promoter region (Figure 5E and F). These above findings suggested that infection-induced Dorsal promotes interaction with Su(Hw) to bind to the miR-210 promoter and inhibit the expression of miR-210 during the early stage of Drosophila Toll immune response.

Figure 5.

Infection-induced Dorsal interacts with Su(Hw) to suppress miR-210 transcription. (A) In vitro Co-IP experiments were performed in S2 cells co-transfected with Dorsal-V5 and Flag-Su(Hw). Immunoprecipitation was performed using anti-V5 and anti-Flag antibodies, followed by western blot analysis with corresponding antibodies. (B) In vivo Co-IP experiments were conducted in Su(Hw)-GFP flies, both uninfected and at 3 h post-infection. Immunoprecipitation was performed using anti-Dorsal and anti-GFP antibodies, followed by western blot analysis with corresponding antibodies. (C) ChIP-qPCR analysis was performed in Su(Hw)-GFP flies with or without Dorsal overexpressed at 3 h post-infection to detect the fold change in Dorsal binding on the promoters of miR-210. (D) ChIP-qPCR analysis was performed in Su(Hw)-GFP flies with or without Dorsal overexpressed at 3 h post-infection to detect the fold change in Su(Hw) binding on the promoters of miR-210. (E) ChIP-qPCR analysis was performed in Su(Hw)-GFP flies with or without su(Hw) knocked down at 3 h post-infection to detect the fold change in Su(Hw) binding on the promoters of miR-210. (F) ChIP-qPCR analysis was performed in Su(Hw)-GFP flies with or without su(Hw) knocked down at 3 h post-infection to detect the fold change in Dorsal binding on the promoters of miR-210.

The Dorsal/Su(Hw)/miR-210/Toll regulatory loop contributes to Drosophila immune responses and homeostasis

To further explore the physiological role of the Dorsal/Su(Hw)/miR-210/Toll regulatory loop, we herein investigated the dynamic expression profiles of Dorsal, su(Hw), miR-210, Toll, Drs in the wild-type (w¹¹¹⁸) flies at different time points (0, 0.5, 1, 2, 3, 6, 12, 24 and 48 h) after M. luteus infection, with PBS injection as control. Our results demonstrated that the expression level of Drs did not significantly change at 0∼0.5 h post-infection compared to the control group, but significantly increased at 1 h and peaked at 6 h post infection, then gradually decreased and returned to near initial level at 48 h after infection (Figure 6A). The expression level of Dorsal was significantly up-regulated at the early stages of infection, such as 0.5, 1 and 3 h post-infection (Figure 6B), while the expression level of su(Hw) was only significantly up-regulated at 2 h after infection (Figure 6C). Interestingly, the expression level of miR-210 was significantly down-regulated at 2 and 3 h post-infection compared to the control group, but significantly up-regulated at 12 h post-infection, followed by a restoration to normal levels (Figure 6D). In contrast, the expression pattern of Toll was nearly opposite to that of miR-210 during the early 1∼12 h response stages, especially at 2, 3 and 12 h post-infection (Figure 6E). Moreover, we examined the changes of miR-210 and Toll expression in the miR-210KO flies at different times post-infection. Our results showed the expression level of Toll was remarkably enhanced in the miR-210KO flies compared to the WT flies at all time points post-infection, suggesting that miR-210 may be primarily responsible for the change of Toll expression level (Figure 6D and E). Taken together, it is reasonable to speculate that in the early stage of M. luteus infection, the NF-κB transcription factor Dorsal interacted with Su(Hw) to form a complex to cooperatively inhibit miR-210 transcription via binding to its promoter region, thereby up-regulating the expression level of target gene Toll as well as Drs to promote Toll pathway immune responses. Whilst in the late stage of M. luteus infection, these overexpressed AMPs may accompany the dissociation of the repressor complex of Dorsal and Su(Hw), resulting in the up-regulated expression of miR-210 to suppress the Toll expression for reducing AMP expressions, thereby preventing the over-activation of Toll pathway immune response and restoring a new homeostasis (Figure 6). Therefore, our results revealed that the Dorsal/Su(Hw)/miR-210/Toll regulatory loop could contribute to Drosophila immune responses and homeostasis in a time- and dose-dependent manner.

Figure 6.

The Dorsal/Su(Hw)/miR-210/Toll regulation loop contributes to Drosophila Toll immune responses and homeostasis (A-C) The dynamic expressions of Drs, Dorsal and su(Hw) were detected by RT-qPCR in the wild-type (w¹¹¹⁸) Drosophila infected with M. luteus at 0, 0.5, 1, 2, 3, 6, 12, 24 and 48 h post-injection. (D) The dynamic expressions of miR-210 in the wild-type (w¹¹¹⁸) and miR-210KO flies were detected by RT-qPCR at 0, 0.5, 1, 2, 3, 6, 12, 24 and 48 h post-infection. (E) The dynamic expressions of Toll in the wild-type (w¹¹¹⁸) and miR-210KO flies were detected by RT-qPCR at 0, 0.5, 1, 2, 3, 6, 12, 24 and 48 h post-infection.

The hsa-miR-210 targets and inhibits TLR6 mRNA expression

Since Drosophila miR-210 is homologous to human hsa-miR-210 and Toll is homologous to human Toll-like receptor 6 (TLR6) from the DIOPT v8.0 database (Supplementary Figure S5A) (53,54), we herein wondered whether the above regulatory loop identified in flies also exists in human. The TLR6 plays a crucial role in the immune response against G+ bacteria or yeast infections in human (55,56). We thus further validated the possible interaction between TLR6 and hsa-miR-210 using the miRWalk website (Supplementary Figure S5B), the dual luciferase reporter and RIP-qPCR assays. Our results demonstrated that the mimics-hsa-miR-210 significantly inhibited the luciferase activity of the reporter containing the 3' UTR of TLR6 (∼4 kb) compared to the control group (mimics-NC), while the point mutation in the 3' UTR of TLR6 disrupted the inhibition effect (Figures 7A and B). Interestingly, the RIP-qPCR experiments demonstrated that the TLR6 RNA immunoprecipitated by anti-Ago2 was increased nearly ten-fold after transfection with mimics-hsa-miR-210 compared to the control group (mimics-NC) in THP1-induced macrophages (Figures 7C). These results revealed that hsa-miR-210 could bind directly to the 3' UTR of TLR6 and inhibit its expression.

Figure 7.

hsa-miR-210 directly targets TLR6 and RelA-mediated transcriptional repression of hsa-miR-210 depends on E4F1. (A) Potential binding sites of hsa-miR-210 in 3'UTR region of TLR6 were predicted using the miRWalk website. Point mutations were introduced at the target site in the 3'UTR, which resulted in base-pairing with the seed sequence of hsa-miR-210. (B) Dual luciferase assay was performed in 293T cells to determine the luciferase activities of mimics-hsa-miR-210 and a reporter plasmid with or without the mutation site of target TLR6. (C) RIP-qPCR experiment was performed in THP1-induced macrophages transfected with mimics-NC or mimics-hsa-miR-210, using anti-Ago2 antibody. (D) Schematic diagram illustrates the promoter region of hsa-miR-210. The predicted TSS, position of the designed primer (purple line), RelA motifs (blue line) and E4F1 motifs (orange line) were marked. (E) Luciferase assay was performed in 293T cells to determine the activity of pGL3-hsa-miR-210-promoter and the pGL3-hsa-miR-210-promoter-AS reporters. (F) Luciferase assay was performed in 293T cells to determine the activity of the pGL3-hsa-miR-210-promoter reporters. Exogenous RelA and E4F1 were separately or jointly included in the assay. “+” indicated transfection with 500 ng of this plasmid, “–” indicated no transfection of the plasmid in 293T cells. (G) Luciferase assay was performed in 293T cells to determine the activity of the pGL3-hsa-miR-210-promoter-RelA-motif-mut reporters. Exogenous RelA and E4F1 were separately or jointly included in the assay. “+” indicated transfection with 500 ng of this plasmid, “–” indicated no transfection of the plasmid in 293T cells. (H) Luciferase assay was performed in 293T cells to determine the activity of the pGL3-hsa-miR-210-promoter-E4F1-motif-mut reporters. Exogenous RelA and E4F1 were separately or jointly included in the assay. “+” indicated transfection with 500 ng of this plasmid, “–” indicated no transfection of the plasmid in 293T cells.

RelA-mediaded transcriptional inhibition of hsa-miR-210 depends on E4F1

Phylogenetic analysis demonstrated that among the five human NF-κB family transcription factors, RelA, Rel and RelB have higher homology with Drosophila Dorsal (Supplementary Figure S6A). The further bioinformatics analysis found that the RelA has binding motif in the promoter region of hsa-miR-210 using TransmiR, PROMO and JASPAR database (Figures 7D and Supplementary Figure S5C–E). We thus examined the effect of RelA on the activity of the hsa-miR-210 promoter (hsa-miR-210-promoter) based on the TSS annotation of the hsa-miR-210 from miRGen v.3 database (57), with the reverse strand sequence of the hsa-miR-210 promoter (hsa-miR-210-promoter-AS) as a negative control. Our results showed that only the hsa-miR-210 promoter sequence had luciferase activity in human 293T cells, while the hsa-miR-210-promoter-AS had no promoter activity (Figures 7D and E). Especially, the inhibitory intensity of RelA on the hsa-miR-210 promoter activity increased with the increase of RelA plasmid concentrations and protein levels (Supplementary Figure S6B), which was consistent with these results of Drosophila in a dosage dependent manner (Supplementary Figure S4E).

We found that human E4F1 and ZNF358 as well as Drosophila Su(Hw) are all members of zinc finger protein family via the DIOPT v8.0 database, implying that human E4F1 and ZNF358 may have similar function with Drosophila Su(Hw) (Supplementary Figure S5F). Interestingly, we found that E4F1 has the binding motif in the promoter region of hsa-miR-210 (Figures 7D and Supplementary Figure S5G). Similar to Drosophila Su(Hw), E4F1 exhibited a more significant inhibitory effect on the promoter activity of hsa-miR-210 than RelA, and this inhibitory effect was also concentration-dependent (Supplementary Figure S6C). However, ZNF358 failed to inhibit the promoter activity of hsa-miR-210 and was served as a negative control (Supplementary Figure S6E). To further validate the regulatory pattern of RelA and E4F1 on the hsa-miR-210 promoter, we examined the activity of the hsa-miR-210 promoter after transfection with cmv-Flag-RelA and pcDNA3.0-HA-E4F1 alone or simultaneously, and found that the co-transfection of RelA and E4F1 exhibits stronger cooperative repression on the hsa-miR-210 promoter compared to their individual transfection (Figures 7F and Supplementary Figure S6F), which further supports the cooperative inhibitory function between these two factors. We further mutated the binding motifs of RelA and E4F1 in the promoter region of hsa-miR-210, and then transfected Flag-RelA and HA-E4F1 alone or simultaneously, finding that the mutation of the RelA motif did not affect the repression of E4F1 on the promoter activity of hsa-miR-210, but the regulation of RelA on hsa-miR-210 changed from inhibition to activation when the E4F1 motif was mutated (Figures 7G and H), which was similar to the regulation pattern of Drosophila Dorsal and Su(Hw) on the miR-210 promoter (Figure 4B-D). Our results revealed that RelA-mediated transcriptional repression of hsa-miR-210 is dependent on E4F1.

NF-κB/RelA interacts with E4F1 to suppress hsa-miR-210 transcription

Here, we further performed in vitro Co-IP experiments in 293T cells after transfecting CMV-3xFlag-RelA and pcDNA3.0-HA-E4F1 plasmids alone or simultaneously, and found that the Flag-RelA could be clearly detected after immunoprecipitation using anti-HA in 293T cells transfected with CMV-3xFlag-RelA and pcDNA3.0-HA-E4F1, and the same for the HA-E4F1 (Figure 8A, quantified in Supplementary Figure S7E and F). Especially, we found that under non-infected condition, there was no interaction between RelA and E4F1 in THP1-induced macrophages through the in vivo Co-IP assay using antibodies against RelA and E4F1 (Figure 8B). Interestingly, a clear interaction between RelA and E4F1 was observed at 0.5 h post-infection, suggesting that the interaction between RelA and E4F1 could be induced by infection (Figure 8B, quantified in Supplementary Figure S7G and H). These results are consistent with the interaction pattern between Drosophila Dorsal and Su(Hw), which suggests that the interaction of infection-induced NF-κB factors with zinc finger protein Su(Hw)/E4F1 is conserved in Drosophila and human.

Figure 8.

Interaction between RelA and E4F1 suppresses hsa-miR-210 transcription in the early stage of G+ bacterial stimulation. (A) In vitro Co-IP experiments were performed in 293T cells co-transfected with Flag-RelA and HA-E4F1 using anti-Flag and anti-HA antibodies for immunoprecipitation, followed by western blot analysis using corresponding antibodies. (B) In vivo Co-IP experiments were performed in THP1-induced macrophages, both uninfected and at 0.5 h post-infection, using anti-RelA, anti-E4F1 antibodies for immunoprecipitation, followed by western blot analysis using corresponding antibodies. (C) ChIP-qPCR was performed in THP1-induced macrophages, with and without overexpression of RelA at 0.5 h post-infection, to detect fold change on the promoters of hsa-miR-210 using anti-RelA antibodies for the immunoprecipitation in vivo. (D) ChIP-qPCR was performed in THP1-induced macrophages, with and without overexpression of RelA at 0.5 h post-infection, to detect fold change on the promoters of hsa-miR-210 using anti-E4F1 antibodies for the immunoprecipitation in vivo. (E) ChIP-qPCR was performed in THP1-induced macrophages, with and without overexpression of E4F1 at 0.5 h post-infection, to detect fold change on the promoters of hsa-miR-210 using anti-E4F1 antibodies for the immunoprecipitation in vivo. (F) ChIP-qPCR was performed in THP1-induced macrophages, with and without overexpression of E4F1 at 0.5 h post-infection, to detect fold change on the promoters of hsa-miR-210 using anti-RelA antibodies for the immunoprecipitation in vivo.

We further overexpressed RelA and E4F1 separately in THP1-induced macrophages at 0.5 h post-infection and examined the signals of RelA and E4F1 in the miR-210 promoter region using anti-RelA and anti-E4F1 antibodies. Our results demonstrated that the overexpression of RelA increased its signal in the hsa-miR-210 promoter region, which was accompanied by a significant increase of E4F1 signal in the hsa-miR-210 promoter region, suggesting that RelA could interact with E4F1 to bind to the hsa-miR-210 promoter region post-infection (Figure 8C and D). On the contrary, no difference was observed in the signal of RelA after the overexpression of E4F1, indicating that E4F1 has no regulatory effect on RelA signal in the promoter region of hsa-miR-210 (Figure 8E and F). Our findings revealed a highly conserved mechanism in which infection-induced NF-κB factors can promote the interaction with zinc finger protein Su(Hw)/E4F1 and bind to the miR-210 promoter, thereby leading to the inhibition of miR-210 expression in Drosophila and human in the early stage of G+ bacterial stimulation.

The role of the RelA/E4F1/hsa-miR-210/TLR6 regulatory loop in THP1-induced macrophages

We herein used THP1-induced macrophages to further detect the role of the RelA/E4F1/hsa-miR-210/TLR6 regulatory loop. Our results demonstrated that the transfection of mimics-hsa-miR-210 into the THP1-induced macrophages could significantly down-regulate the expression of downstream immune factors (IFNB, IL6 and TNFA) of human TLR signaling pathway compared to that of the mimics-NC control at 3 h after stimulation with the heat-killed G+ bacteria M. luteus (Figure 9A), indicating that hsa-miR-210 plays a critical role in human innate immune response. Of note, the levels of both RNA and protein of TLR6, the hsa-miR-210 target gene, in THP1-induced macrophages were significantly declined compared to the mimics-NC control (Figure 9B and C). Especially, we also examined the expression levels of RelA, E4F1, hsa-miR-210, and TLR6 in THP1-induced macrophages at various time points (0, 0.25, 0.5, 1, 2 and 3 h) after infection with heat-killed M. luteus, and demonstrated that the level of E4F1 protein was significantly increased by approximately 30% at 0.5 h post-infection compared to the control group, and subsequently decreased by about 50–60% at 2–3 h after immune stimulation (Figure 9D and E). Meanwhile, the expression level of phosphorylated-RelA protein was observably increased compared to the control groups at 0.5–3 h post-infection (Figure 9D and F). In contrast, the expression level of hsa-miR-210 was significantly decreased in the early stage of immune response (0.5 h) and increased significantly in the later stage (3 h) compared to the control group (Figure 9G). Moreover, the expression level of TLR6 exhibited an opposite pattern to hsa-miR-210 (Figure 9H). To further confirm the association between changes in hsa-miR-210 and TLR6 expression post-infection, we monitored the expression changes of hsa-miR-210 and TLR6 in THP1-induced macrophages transfected with inhibitor-hsa-miR-210 at different times post-infection. The results revealed that the expression level of TLR6 was increased significantly in THP1-induced macrophages transfected with inhibitor-hsa-miR-210 compared to that transfected with inhibitor-NC post-infection, suggesting that hsa-miR-210 may be primarily responsible for the change of TLR6 expression post-infection (Figure 9G and H). Taken together, our results revealed that the RelA/E4F1/hsa-miR-210/TLR6 regulatory loop can contribute to human immune responses and homeostasis in a time- and dose-dependent manner in the THP1-induced macrophages after infection, which is similar to that in Drosophila.

Figure 9.

The RelA/E4F1/hsa-miR-210/TLR6 regulation loop contributes to innate immune responses in THP1-induced macrophages. (A) RT-qPCR analysis of IFNB, IL6 and TNFA expression was measured in THP1-induced macrophages transfected with mimics-NC or mimics-hsa-miR-210 and stimulated with heat-killed M. luteus [multiplicity of infection (MOI) = 50]. (B) RT-qPCR analysis of TLR6 expression was measured in THP1-induced macrophages transfected with mimics-NC or mimics-hsa-miR-210 and stimulated with heat-killed M. luteus [multiplicity of infection (MOI) = 50]. (C) Changes in the protein levels of TLR6 were assessed in THP1-induced macrophages with mimics-NC or mimics-hsa-miR-210 transfection and heat-killed M. luteus [multiplicity of infection (MOI) = 50] stimulated. (D) Changes in the protein levels of E4F1, Phospho-RelA, RelA and GAPDH were analyzed in THP1-induced macrophages using western blot at different time points post-infection. (E) Western blot grayscale analysis of E4F1 in THP1-induced macrophages at different time points post-infection. (F) Western blot grayscale analysis of Phospho-RelA in THP1-induced macrophages at different time points post-infection. (G) The dynamic expressions of hsa-miR-210 in THP1-induced macrophages transfected with inhibitor-NC and inhibitor-hsa-miR-210 were detected by RT-qPCR at 0, 0.25, 0.5, 1, 2, 3 h post-infection. (H) The dynamic expressions of TLR6 in THP1-induced macrophages transfected with inhibitor-NC and inhibitor-hsa-miR-210 were detected by RT-qPCR at 0, 0.25, 0.5, 1, 2, 3 h post-infection.

Discussion

The tight control of immune response duration and strength is crucial for effective host defense. Although miRNAs have been demonstrated to play crucial roles in fine-tuning immune signals, the dynamic regulations of immune-specific miRNAs themselves remains poorly understood (58,59). In this study, we aimed to investigate how one conserved miR-210 regulates and is regulated in innate immune response to G+ bacterial stimulation. Our findings contribute to a better understanding of the regulation by and of immune-specific miRNAs.

The miR-210 is evolutionarily conserved between Drosophila and mammals (60), and has been involved in the regulation of photoreceptor neurons, circadian rhythms and lipid metabolism in Drosophila (42,61,62). However, its role in immune response remains unknown. In our work, we revealed that miR-210 can negatively regulate Drosophila Toll signaling via directly repressing the expression of target gene Toll (Figure 1). Interestingly, a similar regulatory relationship was also found in the human TLR signaling pathway (Figures 7 and 9). In contrast to Drosophila, the human hsa-miR-210 can extensively regulate a variety of biological processes, including inflammation, cell cycle, DNA damage repair, apoptosis, particularly the occurrence and development of cancer (63,64). However, little is known about the transcriptional activation and regulation of miR-210 itself in Drosophila and human. Mounting studies have focused on the regulatory roles of miRNAs, but little attention has been paid to how these miRNAs themselves are regulated, especially at transcriptional level. Interestingly, several studies have revealed that the dynamic and specific expressions of miRNAs can be precisely controlled by transcription factors (15,65,66). For instance, in mammals, the Hypoxia Inducible Factor-1 (HIF-1) acts as a principal transcription factor that binds to the miR-210 promoter and activates miR-210 transcription in response to hypoxia (67). Other transcription factors such as Oct-4, AP2, PPARc, and E2F have been reported to bind to the miR-210 promoter regions, although their regulatory roles remain unclear (68). These studies suggest that the multifunctional roles of miR-210 might be precisely regulated via different transcription factors. Similarly, the NF-κB family transcription factors are essential for innate immune response, whether they can also control the transcription of miR-210 to involve in innate immune response? Herein, we have provided compelling evidences that NF-κB counterparts from Drosophila Toll and human TLR signaling pathways regulate miR-210 expression. Specifically, we demonstrated that Drosophila NF-κB/Dorsal inhibits miR-210 transcriptional activity via binding to its promoter region (Figure 2). To ensure accuracy, the enhancement of positive control Drs expression was also detected when overexpressing Dorsal in vitro and in vivo, which is consistent with the previous report (69) (Figure 2). Of great importance, we also demonstrate that the human NF-κB family transcription factor RelA represses the transcription of hsa-miR-210 via binding to its promoter region (Figure 7), revealing an evolutionarily conserved mechanism between Drosophila and human. Overall, our study suggests that specific expression of miR-210 in Drosophila and human can be precisely regulated by NF-κB transcription factors, thereby performing critical immune regulation roles.

Currently, NF-κB factors are primarily considered to execute transcriptional activation functions on antimicrobial peptides in insects and downstream cytokines in human via its transactivation domain (TAD). For example, human RelA can activate miR-34a via directly binding to its promoter, thereby inhibiting Foxp3 expression involved in the regulatory of T cells (65). However, another study reported that RelA can inhibit miR-30a transcription to regulate Th17 differentiation (70), but its transcriptional inhibition mechanism has not been further investigated. We thus speculated that the inhibition of miR-210 by NF-κB factors may need to interact with transcriptional repressors. Interestingly, our work demonstrated that Drosophila zinc finger protein Su(Hw) can serve as a transcriptional repressor by binding to and inhibiting the promoter of miR-210 (Figure 4). Previous studies have indicated that Su(Hw) protein not only acts as a chromatin insulator (71), but also forms transcriptional inhibitory complexes with other proteins (22,72,73). Consistently, our present study has demonstrated that infection-induced Dorsal interacted with Su(Hw) to form a complex and further inhibit miR-210 transcription by binding to its promoter region during the early stage of M. luteus infection (Figure 4 and 5). Remarkably, the co-overexpression of Dorsal and Su(Hw) has a significantly stronger inhibitory effect on the promoter activity of miR-210 compared to their individual overexpression (Figure 4B). Overexpressing Su(Hw) can still significantly inhibit the promoter activity of miR-210 even with Dorsal motif mutation on miR-210 promoter (Figure 4C), whereas overexpressing Dorsal can promote transcriptional activation of miR-210 following Su(Hw) motif mutation on miR-210 promoter (Figure 4D, I and J). These results suggested that the Dorsal-mediated inhibition on miR-210 promoter is dependent on Su(Hw). Furthermore, our study demonstrated that infection-induced Dorsal can interact with more Su(Hw) to enhance the Su(Hw) signal on miR-210 promoter during the early stage (3 h) of Toll pathway immune response (Figure 5C and D). Our work reveals a novel immunomodulatory mechanism by which Dorsal suppresses miR-210 expression via interacting with Su(Hw) to enhance immune response during the early stage response to G+ bacterial stimulation in Drosophila. Meanwhile, another Drosophila NF-κB factor Dif can also inhibit the expression of miR-210 and interact with Su(Hw) in vitro, suggesting this immunomodulatory mechanism might be conserved among different NF-κB factors (Supplementary Figure S8).

However, it is worth noting that the interaction between transcription factors and their associated protein complexes is complex and often involves multiple partners and regulatory mechanisms. Although we have performed in vivo and in vitro Co-IP experiments between Dorsal and Su(Hw) to yield positive results in this study, this does not exclude the possibility of Dorsal interacting with the other compositions of the Su(Hw) complex to form the complex because Su(Hw) can also exert a positive effect on the transcription from nearby promoters (74,75) and its main partners Mod(mdg4) and CP190 are not known as suppressor proteins (76–79), implying that Dorsal may also need to stabilize/recruit other repressed complex at the Su(Hw) site in the miR-210 promoter to inhibit its expression. Therefore, further investigations on the other compositions of the complex besides Dorsal and Su(Hw) are needed to fully elucidate the dynamic expression mechanism of miR-210.

Strikingly, we have identified a human zinc-finger protein E4F1, which may perform a similar function to Drosophila Su(Hw). Although E4F1 as transcription repressor has been reported to be involved in stem cell homeostasis (80–82), ROS-mediated death in human myeloid leukemia cell lines (83,84) and DNA double-strand break repair (85), its regulatory role in innate immune responses has not yet been reported to date. In our study, we discovered that in the early stage of G+ bacterial stimulation, the protein levels of E4F1 are significantly up-regulated in the THP1-induced macrophages (Figure 9D and E). Especially, we found that RelA can interact with E4F1 to form a transcriptional inhibitory complex, thereby leading to the inhibition of hsa-miR-210 expression via binding to its promoter region (Figure 7 and 8). Notably, E4F1 has a more significant inhibitory effect on the activity of hsa-miR-210 promoter than RelA (Figure 7F and G). Similar to Drosophila Dorsal and Su(Hw), the binding motif of E4F1 is closer to the TSS site of hsa-miR-210 than RelA (Figure 7D). Of note, in this work, the level of inhibition on hsa-miR-210 promoter activity was very similar between the plasmid transfected with pcDNA3.0-HA-E4F1 (500 ng) + cmv-empty (500 ng) and the plasmid transfected with pcDNA3.0-HA-E4F1 (500 ng) + cmv-Flag-RelA (500 ng) (Figure 7F). This is probably because the levels of E4F1 plasmid being transfected in these experiments is too high. We thus repeated the experiment via reducing the plasmid levels of E4F1 and RelA by half (250 ng), and found that the co-transfection of RelA and E4F1 exhibited a stronger cooperative repression on the hsa-miR-210 promoter activity compared to their individual transfection, which further supported the cooperative inhibitory function between the two factors (Supplementary Figure S6F). Together, our work suggests that the regulation mechanism of NF-κB family transcription factors interacting with transcriptional repressors to inhibit miR-210 expression is evolutionarily conserved between Drosophila and human innate immune responses. In our work, the protein level of E4F1 is decreased, but the level of phosphorylated-RelA remains increased compared to the controls in the late stage of infection, indicating that the transition from transcriptional repression to reactivation of miR-210 may be due to the dissociation of the inhibitory complex or potential degradation of repressors.

Fascinatingly, our work demonstrated a tightly coordinated dynamic expression profile of Dorsal, su(Hw), miR-210, Toll, and Drs in the wild-type (w¹¹¹⁸) flies upon infection, as well as RelA, E4F1, hsa-miR-210 and TLR6 in the THP1-induced macrophage with infection, suggesting that animal innate immune responses can be finely regulated in a time- and dose-dependent manner (Figures 6 and 9). Therefore, we herein proposed an evolutionary conservative regulatory mechanism for maintaining innate immune homeostasis (Graphical Abstract). On the one hand, to prevent inadequate immune responses in the early stage of pathogen infection, the NF-κB family transcription factors (e.g. Drosophila Dorsal and human RelA) interact with transcription repressors (e.g. Drosophila Su(Hw) and human E4F1) at the appropriate time to form a complex and bind to the promoter region of miR-210, and inhibit miR-210 expression to de-suppress target gene Toll/TLR6, thereby enhancing immune response to eliminate invading pathogenic bacteria. On the other hand, with the continuous enhancement of immune response, the complexes containing Dorsal and Su(Hw) or RelA and E4F1 are dissociated with progressively decreased repressors to restore the miR-210 expression, thereby further inhibiting the expression of target gene Toll/TLR6 to prevent the excessive immune response and maintain a new immune homeostasis in the late stages of pathogen infections (Graphical Abstract). Mechanistically, our work sheds light on a novel conserved regulatory mechanism of immune responses between invertebrate and mammal.

Conclusion

In this work, we have identified miR-210 and Su(Hw) as new regulators in the Drosophila Toll pathway, along with E4F1 and hsa-miR-210 as new regulators in the human TLR pathway. We have also demonstrated that NF-κB factors (Dorsal in Drosophila and RelA in human) control miR-210 expression via interacting with transcription repressors (Su(Hw) in Drosophila and E4F1 in human) in a time- and dose-dependent manner. More importantly, we have elucidated a conserved regulatory mechanism for innate immune response and immune homeostasis in the Drosophila Toll and human TLR signaling pathways. Our study not only unveils a conserved TFs/miR-210/targets regulatory loop in both Drosophila and human, but also provides vital insights into the dynamic regulation of miRNA expression during animal innate immune responses.

Data availability

The data underlying this article are available in the article and in its online supplementary material. Further data underlying this article will be shared on reasonable request to the corresponding author.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

National Natural Science Foundation of China [32370516 to F.M., 31970477 to F.M., 32370515 to P.J., 32300401 to H.Z.]; Natural Science Foundation from Jiangsu Province [BK20231282 to P.J.]; Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions and Jiangsu Funding Program for Excellent Postdoctoral Talent (to H.Z.). Funding for open access charge: National Natural Science Foundation of China [32370516, 31970477, 32370515, 32300401].

Conflict of interest statement. None declared.