Flies (Drosophila melanogaster) were raised at 25 °C with a 12 h of light/dark cycle and 65% humidity. Standard Drosophila medium (6.65% cornmeal, 7.15% dextrose, 5% yeast, 0.66% agar, 2.2% nipagin, and 3.4 mL/L propionic acid) was used to feed all flies. The w${}^{\mathrm{1118}}$ strain was used as the wild-type control and all other flies were back-crossed with w${}^{\mathrm{1118}}$ for isogenization (at least 5 generations). The detailed information of the flies used in this study are as follows: (1) key${}^{\mathit{\text{c02831}}}$ (#11044) and Uch-L5${}^{\mathit{\text{j2B8}}}$(#12078) were purchased from Bloomington Drosophila Stock Center; (2) Uch-L5 RNAi #1 (#103481) was obtained from Vienna Drosophila Resource Center; (3) Uch-L5 RNAi #2 (#04757.N) was acquired from Tsinghua Fly Center; and (4) c564-gal4 was described previously [32].

All dsRNAs were synthesized according to methods described previously [33]. In brief, DNA templates were amplified using specific primers (Supplementary Table 1). PCR products were then combined with 1/10 volume of 5 M NaCl and 2.5-fold volume of EtOH, followed by centrifugation for 10 min at high speed. The pellet was washed with 75% EtOH and dissolved in RNase/DNase-free water. The T7 in vitro transcription kit (Promega, Cat#P1300, Madison, Wisconson, USA) was used for dsRNA synthesis and purification according to the manufacture’s protocol. Purified dsRNAs were diluted in RNase/DNase-free water to a final concentration of 1 µg/µL. The quality of dsRNAs was confirmed by agarose gel electrophoresis (Supplementary Fig. 1A). For gene silencing in S2 cells, dsRNA at the amount of 3 µg was added directly into S2 cells for 48 h.

S2 cells were cultured in a 27 °C incubator using insect medium (Gibco) with 10% FBS (Hyclone). The MycoAlert${}^{\text{TM}}$ kit (Lonza, Cat#LT07-318, Basel, Halbkanton, Switzerland) was used for mycoplasma detection and validation of the authenticity of S2 cells. Lipofectamine${}^{\text{TM}}$ 2000 (Thermo Fisher, Cat#11668019, Waltham, Massachusetts, USA) was used for plasmid transfection into S2 cells according to previous protocols [34]. The Att-Luc [35] and Drs-Luc [33] reporter systems were utilized to detect the relative activities of the IMD and Toll pathways, respectively. For the IMD pathway, we transfected S2 cells with the plasmid that express Imd, in order to induce the Att-Luc activity. As a positive control, we used the dsRNA targeting kenny, which is one of the pivotal genes of the IMD pathway. For the Toll pathway, the Myd88 expressing plasmid and the pelle dsRNA (positive control) were used. The methods for detecting the Firefly and Renilla luciferase activities were as follows: S2 cells were adequately lysed with 100 µL of passive lysis buffer (Promega, Cat#PR-E1941, Madison, Wisconson, USA). After centrifugation, 30 µL of the supernatant from each sample was added into a 96-well plate containing detection reagents. Firefly and Renilla luciferase activities were detected and finally the ratio of Firefly to Renilla was calculated. Results were analyzed based on data from 3 independent biological replications.

S2 cells or fly samples were homogenized in Trizol Reagent (Invitrogen, Cat#15596026, Waltham, Massachusetts, USA). Chloroform solution (1/5 volume) was added into the sample, followed by intense votexing for 15 sec, and centrifugation at high speed for 15 min. The upper layer of the supernatant was transferred into a fresh tube and an equal volume of isopropanol was added. After centrifugation again at high speed for 10 min, the pellet was washed with 75% EtOH and diluted in RNase/DNase-free water. Samples were further incubated with DNase for 30 min, and the quality of RNA was assessed by examining the 260:280 ratio, and the pattern in the agarose gel electrophoresis assay. The first-strand cDNA synthesis kit (Transgen, Cat#AT341-01, Beijing, China) was used to reverse-transcribe RNA (1 µg) into cDNA. Quantitative PCR assays (in 3 technical repetitions) were performed using a SYBR Green Master Mix (Thermo Fisher, Cat#A46012, Waltham, Massachusetts, USA) in the Lightcycler 480 PCR platform (Roche, Cat#05015278001, Basel, Halbkanton, Switzerland). Rp49 was used as the internal control. Results were analyzed based on data from 7 independent biological replications. The primers used in RT-qPCR experiments are shown in Supplementary Table 1.

Infection experiments were performed as previously described [36]. In brief, overnight bacterial cultures were harvested and diluted in sterile PBS at a concentration of OD${}_{600}$ = 1. The Serratia marcescens strain (#1.1215) was obtained from China General Microbiological Culture Collection Center (CGMCC) and the Ecc15 was a kind gift from Dr. Dominique Ferrandon’s laboratory (Institut de Biologie Moleculaire et Cellulaire, France). For injection, male flies were collected and anesthetizsed on the fly pad with CO${}_2$. Diluted bacteria (4.6 nL), or the same volume of PBS, was injected into each fly with a nanoliter injector (Nanoject III, Drummond, Cat#3-000-207, Broomall, Pennsylvania, USA). Flies that died within 2 h after bacterial infection were not included in further studies. For survival analysis, flies were transferred to fresh vials and counted for death every day. For bacterial burden experiments, 10 flies were collected, dipped in 75% EtOH, and volatilized with EtOH on the fly pad for several minutes. The flies were then homogenized in 200 µL sterile PBS, followed by serial dilutions. Finally, 100 µL of each diluent was inoculated on an LB agar plate at 30 °C for 12 h before we counted the numbers of the colonies.

All statistical analyses were performed by using GraphPad Prism 9 (v. 9.4.0, Dotmatics, Boston, MA, USA). Data were shown as means plus standard errors. Statistical significance was determined by using the ANOVA or Mann-Whitney tests except for survival assays, in which the Log-Rank test (Kaplan-Meier method) was used. The p value of less than 0.05 was considered statistically significant. * p 0.05, ** p 0.01, *** p 0.001, ns (not significant) p $\mathrm{>}$ 0.05.

We examined the potential role of Uch-L5, a previously described Dub modulating Hh signaling [31], in affecting the innate immune reaction in Drosophila. To do this, we first designed 3 types of dsRNAs that targeted different regions of the coding sequence of Uch-L5 (referred to as Uch-L5 dsRNAs #1, #2, and #3, respectively, Supplementary Fig. 1A). Cultured Drosophila S2 cells were treated with these Uch-L5 dsRNAs (dsRNA that targeted GFP was used as the control), and the knockdown efficiency was monitored by RT-qPCR assays (Supplementary Fig. 1B). As illustrated in Fig. 1A, silencing of Uch-L5 resulted in drastic decreases (by ~75% to ~80%) of the Att-Luc activities upon Imd over-expression, indicating that Uch-L5 is a potentially positive regulator in the IMD signaling pathway in cultured S2 cells. Further, we performed RT-qPCR assays to detect the endogenous inductions of the AMPs that are downstream of the IMD pathway, including AttA, Dpt, and CecA1. We obtained consistent results: down-regulation of Uch-L5 significantly prevented the Imd-driven transcription of several AMP genes (Fig. 1B–D).

Fig. 1.

Silencing of ubiquitin carboxy-terminal hydrolase L5 (Uch-L5) reduces immune deficiency (IMD) signaling in Drosophila S2 cells. (A) S2 cells were pretreated with indicated dsRNAs. 48 h later, cells were transfected with various combinations of expressing plasmids, followed by dual-luciferase assays. (B–D) S2 cells were pretreated with dsRNAs as in (A) for 48 h. Cells were then transfected with indicated expressing plasmids for 36 h and harvested for RT-qPCR assays to detect the mRNA levels of AttA (B), Dpt (C), and CecA1 (D). In (A–D), each dot represents one independent replicate and data are shown as means plus standard errors. * p 0.05, ** p 0.01, *** p 0.001.

As mentioned in the Introduction, the Drosophila innate immune response is mostly governed by two signaling cascades, the Toll and the IMD pathways. We then conducted another luciferase reporter assay (Drs-Luc) in S2 cells, as previously reported, aiming to test whether Uch-L5 is also involved in affecting Toll signaling. However, we did not observe apparent alterations in the Myd88-driven Drs-Luc activities when Uch-L5 was knocked down, compared to that in the control (Supplementary Fig. 2A). Similar results were obtained when we looked at the transcript levels of Drs and Mtk, two well-known downstream AMP genes in the Toll pathway (Supplementary Fig. 2B,C). These data indicated that Drosophila Dub Uch-L5 is specifically involved in controlling the IMD other than the Toll innate immune signaling pathway in S2 cell cultures.

We collected more evidence on the functional role of Uch-L5 in controlling IMD innate immunity by constructing a plasmid expressing Myc-tagged Uch-L5 in S2 cells. By again using the Att-Luc reporter system, we observed that ectopic expression of Myc-Uch-L5 enhanced IMD signaling in a dose-dependent manner (Fig. 2A). Accordingly, co-transfection of Uch-L5 expressing plasmid resulted in increased inductions (by ~32% to ~136%) of the IMD-downstream AMPs upon Imd over-expression (Fig. 2B–D), which indicated a positive role of Uch-L5 in modulating IMD signaling.

Fig. 2.

Uch-L5 affects IMD signaling in a Uch domain-dependent manner. (A–D) S2 cells were transfected with expressing plasmids as indicated, followed by dual-luciferase assays (A) or RT-qPCR assays to examine the expression levels of AttA (B), Dpt (C), and CecA1 (D). (E) Domain analysis of Uch-L5. (F) S2 cells were transfected with indicated combinations of expressing plasmids, followed by dual-luciferase assays. In (A–D,F), each dot represents one independent replicate and data are shown as means plus standard errors. * p 0.05, ** p 0.01, *** p 0.001, ns p $\mathrm{>}$ 0.05.

A recent study that explored the biochemical characteristic of Uch-L5 demonstrated that Uch-L5 harbors a typical Uch domain at its N-terminus, which is the catalytical region for its Dub enzymatical activity [31]. We therefore constructed two kinds of plasmids that expressed truncated forms of Uch-L5, including Uch-L5${}^{\text{UD}}$ (Uch domain) and Uch-L5${}^{\text{CTD}}$ (C-terminal domain) (Fig. 2E), and transfected them into S2 cells for Att-Luc reporter assay. As illustrated in Fig. 2F, over-expression of Uch-L5${}^{\text{UD}}$, but not Uch-L5${}^{\text{CTD}}$, mimicked the positive contribution of Uch-L5${}^{\text{FL}}$ (full-length Uch-L5) to IMD signaling. Taken together, these data suggested that the N-terminal Uch domain is both required and sufficient for Uch-L5 to benifit Drosophila IMD signaling.

We next sought to decipher the immune function of Uch-L5 in vivo. To this end, we collected w${}^{\mathrm{1118}}$ flies (referred to as the wild-type control), key${}^{\mathit{\text{c02831}}}$ flies (key loss-of-function mutant as the positive control), and Uch-L5${}^{\mathit{\text{j2B8}}}$ flies (Uch-L5 loss-of-function mutant, isogenized with w${}^{\mathrm{1118}}$) for infection experiments using the Gram-negative bacteria S. marcescens (Serratia marcescens), to activate the host IMD innate immune defense. Six hours after infection, we detected decreases (by ~34% to ~56%) of AttA, Dpt, and CecA1 in the Uch-L5-defective flies relative to those in the wild-type control (Fig. 3A–C). To confirm our results, we subjected these flies to infection with another type of Gram-negative bacteria, Ecc15 (Erwinia carotovora carotovora 15). Similar reduction trends were observed when we compared the expression levels of AMPs in the Uch-L5 and key loss-of-function mutants to those in the wild-type flies (Fig. 3A–C). In addition, we noted comparable survival rates of these flies (w${}^{\mathrm{1118}}$, key${}^{\mathit{\text{c02831}}}$, and Uch-L5${}^{\mathit{\text{j2B8}}}$) after injection of sterile PBS buffer (Fig. 4A). However, the Uch-L5 loss-of-function mutant flies succumbed much faster than did the wild-type controls upon injection of either S. marcescens (Fig. 4B) or Ecc15 (Fig. 4C), indicating that Uch-L5 plays an essential role in the host defense against bacterial infections.

Fig. 3.

Loss of Uch-L5 antagonizes IMD downstream AMP expressions. (A–C) w${}^{\mathrm{1118}}$ (wild-type control), key${}^{\mathit{\text{c02831}}}$, and Uch-L5${}^{\mathit{\text{j2B8}}}$ males (6- to 7-day) were injected with PBS, S. marcescens, or Ecc15. 6 h later, flies were harvested for RT-qPCR assays to detect the expression levels of AttA (A), Dpt (B), and CecA1 (C). Each dot represents one independent replicate and data are shown as means plus standard errors. * p 0.05, ** p 0.01, *** p 0.001.

Fig. 4.

Uch-L5 plays a critical role in the fly anti-microbial defense. (A–C) w${}^{\mathrm{1118}}$ (wild-type control), key${}^{\mathit{\text{c02831}}}$, and Uch-L5${}^{\mathit{\text{j2B8}}}$ males (6- to 7-day) were infected with PBS (A), S. marcescens (B), or Ecc15 (C), followed by survival analyses. The numbers of flies in each figure are as follows. In (A), w${}^{\mathrm{1118}}$: 104, 103, 106; key${}^{\mathit{\text{c02831}}}$: 105, 105, 102; Uch-L5${}^{\mathit{\text{j2B8}}}$: 102, 103, 104. In (B), w${}^{\mathrm{1118}}$: 102, 104, 106; key${}^{\mathit{\text{c02831}}}$: 104, 103, 105; Uch-L5${}^{\mathit{\text{j2B8}}}$: 105, 103, 103. In (C), w${}^{\mathrm{1118}}$: 104, 107, 106; key${}^{\mathit{\text{c02831}}}$: 106, 104, 103; Uch-L5${}^{\mathit{\text{j2B8}}}$: 104, 103, 106. (D,E) w${}^{\mathrm{1118}}$ (wild-type control), key${}^{\mathit{\text{c02831}}}$, and Uch-L5${}^{\mathit{\text{j2B8}}}$ males were infected with S. marcescens (D) or Ecc15 (E), followed by bacterial burden assays at indicated time points (0, 6, and 12 h). Each dot represents one independent replicate and data are shown as means plus standard errors. For each replicate, 10 male flies were collected. In (B–E), * p 0.05, ** p 0.01, *** p 0.001.

To test the involvement of Uch-L5 in affecting the proliferation of injected pathogens (S. marcescens or Ecc15), we performed CFU (colony-forming-units) assays at different time points (0, 6, and 12 h) after injection. As shown in Fig. 4D,E, the burdens of S. marcescens and Ecc15 in Uch-L5${}^{\mathit{\text{j2B8}}}$ adults were significantly higher than those in the wild-type flies. In summary, our in vivo data strongly support the notion that Uch-L5 plays a critical role in the host defense against bacterial pathogens in Drosophila.

Fat body is one of the main responsible immune tissues/organs during systemic infection in Drosophila. As the Uch-L5 transcript is relatively abundant in the fat body cells, according to the high-throughput sequencing or array data in the FlyBase website (https://flybase.org/reports/FBgn0011327), we sought to investigate the functional role of Uch-L5 in Drosophila fat body. Two different Uch-L5 RNAi (RNA interference) flies (Uch-L5 RNAi #1 and #2, for detailed information, see Materials and Methods) were crossed with the fat body-specific driver c564-gal4. The Tub-gal80${}^{\text{𝑡𝑠}}$ strain was used in order to allow gene RNAi at adult stage (Supplementary Fig. 3A). The progenies, including c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #1 and c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #2, were collected and subjected to S. marcescens or Ecc15 infection (c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$+ was used as the wild-type control). Six hours after S. marcescens or Ecc15 infection, the inductions of AttA, Dpt, and CecA1 in Uch-L5 RNAi flies were prominently decreased compared to those in the control group (Fig. 5A–C). These results suggested that Uch-L5 functions in fat body to modulate IMD signaling in response to bacterial challenge. Consistently, we observed that the Uch-L5 RNAi flies were less resistant to S. marcescens or Ecc15 infection: they displayed higher mortality than did the wild-type control (Fig. 5D–F). Moreover, the bacterial burdens at various time points after infection were remarkably elevated by the loss of Uch-L5 in fat body (Fig. 5G,H).

Fig. 5.

Silencing of Uch-L5 in fat body results in immune defects. (A–H) c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$+ (wild-type control), c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #1, and c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #2 males (6- to 7-day) were injected with PBS, S. marcescens, or Ecc15. Flies were collected for RT-qPCR assays to examine the inductions of AttA (A), Dpt (B), and CecA1 (C) 6h after infection, or subjected to survival analyses (D–F), or bacterial burden assays (G,H). In (A–C, G,H), each dot represents one independent replicate and data are shown as means plus standard errors. In (D), the numbers of flies are as follows. c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$+: 104, 105, 104; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #1: 103, 102, 102; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #2: 103, 106, 105. In (E), c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$+: 102, 103, 106; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #1: 104, 105, 105; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #2: 103, 104, 102. In (F), c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$+: 105, 104, 102; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #1: 106, 103, 104; c564${}^{\text{𝑡𝑠}}$$\mathrm{>}$Uch-L5 RNAi #2: 105, 105, 104. In (A–H), * p 0.05, ** p 0.01, *** p 0.001.