Down syndrome cell adhesion molecule (Dscam) generates tens of thousands of isoforms by alternative splicing, thereby providing crucial functions during immune responses. In this study, a novel Dscam signaling pathway was investigated in crab, which remains poorly characterized in invertebrates. Bacterial infection induced the cytoplasmic cleavage of Dscam intracellular domains (ICDs) by γ-secretase, and then the released ICDs carrying specific alternatively spliced exons could directly interact with IPO5 to facilitate nuclear translocation. Nuclear imported ICDs thus promoted hemocyte proliferation and protect the host from bacterial infection. Protein-interaction studies revealed that the ectodomain of Dscam bound to a disintegrin and metalloprotease domain 10 (ADAM10) rather than ADAM17. Inhibition or overexpression of ADAM10 impaired or accelerated Dscam shedding activity post–bacterial stimulation, respectively. Moreover, the shedding signal then mediated Dscam with an intact cytoplasmic domain to promote the cleavage of ICDs by γ-secretase. Furthermore, the transcription of ADAM10 was regulated by Dscam-induced canonical signaling, but not nuclear imported ICDs, to serve as a feedback regulation between two different Dscam pathways. Thus, membrane-to-nuclear signaling of Dscam regulated hemocyte proliferation in response to bacterial infection.
Alternative splicing serves to enhance transcriptomic and proteomic diversity in numerous species from invertebrates to vertebrates (1). Splicing in general, and alternative splicing in particular, is important for regulation of the quantity and specificity of gene expression and, if disrupted, can lead to the malfunction of genes and proteins, ultimately resulting in disease (2, 3). In perhaps the most striking example of the complexity of alternative splicing, the single pre-mRNA of the Drosophila melanogaster receptor gene Down syndrome cell adhesion molecule 1 (Dscam1) can be processed to generate potentially 38,016 different mature transcripts (4), unlike the vertebrate homolog, which is not alternatively spliced. Even if only a portion of the Dscam isoforms is expressed in vivo, this combinatorial usage of spliced exons represents an incredible source of diversity, especially given that the entire Drosophila genome codes for only 14,000 genes (5).
Alternative splicing can provide biological options for a determinative biological response in combination with many other mechanisms (6). As integral components of the innate immune system, hemocytes must be capable of expressing most, if not all, of the potential Dscam isoforms to mount an effective defense against invading pathogens (4, 7), suggesting that Dscam may mediate distinct functions by the preferential use of alternatively spliced variants, which differs from the roles of vertebrate homologs. Recent studies have shown that mammalian DSCAM intracellular domains (ICDs) are cleaved by γ-secretase, and the liberated ICDs are localized to the nucleus via a nuclear localization signal (NLS) via interaction between ICDs and importin 5, subsequently regulating the expression of genes capable of inhibiting synapse formation (8), as a previously unknown mechanism of these important molecules. However, this process has not been reported in invertebrates. So, it remains unclear whether invertebrate Dscam has similar functions of intracellular cleavage and nuclear import. Moreover, the factors that induce Dscam cleavage remain unknown in both vertebrates and invertebrates. Furthermore, whether and how specific alternatively spliced Dscam ICDs are translocated from the cytoplasm into the nucleus remain elusive, as with the biological outcomes of nuclear imported Dscam ICDs, especially the effects on the innate immune response of invertebrates.
Blood cells (hemocytes) play important roles in tissue remodeling during embryogenesis and metamorphosis as well as the immune response of invertebrates. Drosophila contains three distinct hemocyte types (i.e., plasmatocytes, lamellocytes, and crystal cells), which collectively play crucial roles in phagocytosis, encapsulation, lysis of foreign cells, and antimicrobial peptide expression (9). Significant progress has recently been made in understanding the genetic control of hemocyte differentiation during development (10). Although hemocytes are essential in immunity, the mechanisms underlying the regulation of hematopoiesis in invertebrates remain largely unknown. Interestingly, previous studies have not found that mature circulating hemocytes have the ability to proliferate in crustaceans, although proliferated hemocytes play critical roles in the early phase of infection. However, the single-cell RNA deep sequencing (RNA-seq) data of Anopheles gambiae have shown the potential ability of mature hemocyte proliferation (11). In addition, the molecular mechanisms by which hemocytes trigger proliferation upon pathogen infection are still elusive during the immune response.
In this study, the immunological roles of Dscam in hemocytes were investigated. Bacterial stimulation induced cleavage of Dscam ICDs in the cytoplasm, and binding of IPO5 with Dscam was found to mediate the translocation of specific Dscam ICDs from the cytoplasm into the nucleus. RNA-seq data suggested that nuclear imported Dscam ICDs may regulate cell proliferation, which was also confirmed in crab hemocytes. Moreover, γ-secretase and a disintegrin and metalloprotease domain 10 (ADAM10) were found to play critical roles in Dscam cleavage. The high expression of ADAM10 induced by bacteria can bind to the extracellular domain of Dscam to mediate its shedding and then activate the cleavage of Dscam-ICD by γ-secretase. Furthermore, nuclear translocation of Dscam ICDs had no role in ADAM10 transcription, but the canonical signaling activated by Dscam-Dock-ERK-Dorsal was found to significantly regulate ADAM10 expression upon bacterial stimulation. Thus, in response to bacterial infection, Dscam promoted ADAM10 expression and other antibacterial immunological outcomes in the early phase via canonical signaling pathways, then regulated hemocyte proliferation via the ADAM10–γ-secretase–IPO5 axis controlled by Dscam ICD cleavage and nuclear translocation in the later phase. These findings provided new insights into how Dscam membrane-to-nucleus signaling is tightly regulated and its immunological outcome impact on hemocyte proliferation in invertebrates.
Materials and Methods
Animal and cells
The study protocol was approved by the Animal Care and Use Committee of East China Normal University (Shanghai, China; approval number AR2012/12017) and conducted in accordance with the animal care guidelines of the Ministry of Science and Technology of the People’s Republic of China.
Healthy adult Chinese mitten crabs (Eriocheir sinensis; mean body weight, 100 ± 10 g) were obtained from the Songjiang Aquatic Farm (Shanghai, China). After quick transfer to our laboratory, the crabs were maintained in filtered aerated freshwater and fed daily with a commercially formulated antibiotic-free diet.
E. sinensis hemolymph was collected from the nonsclerotized membrane of the posterior walking leg using a 10-mL sterile syringe preloaded with 5 ml precooled sterile anticoagulant (0.14 M NaCl, 0.1 M glucose, 30 mM trisodium citrate, 26 mM citric acid, and 10 mM EDTA [pH 4.6]) at a ratio of 1:1. The collected hemolymph was immediately centrifuged at 300 × g for 10 min at 4°C. Then, the serum was removed and washed with PBS. The isolated hemocytes were gently resuspended in Leibovitz’s L-15 medium (Sigma-Aldrich, St. Louis, MO) supplemented with 1% antibiotics (10,000 U/ml penicillin and 10,000 μg/ml streptomycin; Life Technologies, Thermo Fisher Scientific, Waltham, MA) and 0.2 mM NaCl (676 ± 5.22 mOsm/kg) at pH 7.2–7.4, then counted using an automated cell counter (Countess; Thermo Fisher Scientific) before using 4 ml (1 × 106 cells/ml) to seed 60-mm dishes.
S2 cells were cultured in Drosophila Schneider’s medium (Life Technologies) supplemented with 10% FBS and antibiotics. Human embryonic kidney (HEK)–293T cells were cultured in DMEM (Life Technologies) containing 10% FBS (Invitrogen, Carlsbad, CA), 1% penicillin, and streptomycin. Cultures were maintained in 5% CO2 at 37°C.
Immune stimulation and sample collection
Staphylococcus aureus, which is commonly used for bacterial infection of vertebrates and invertebrates, was obtained from the National Pathogen Collection Center for Aquatic Animals (stock number BYK0113; Shanghai Ocean University, Shanghai, China), cultured, collected, and resuspended in sterile PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 2 mM KH2PO4 [pH 7.4]). Bacterial counts were determined using the agar plating method.
In vitro and in vivo bacterial challenges were performed in accordance with previously described methods (12). For in vitro bacterial stimulation, cultured bacteria (1 × 107 cells/dish, 50 μl) in 50 μl sterile PBS were added separately to a hemocyte-cultured dish serving as the control. Total RNA was collected from hemocytes at specific time points poststimulation. Each sample consisted of hemocytes pooled from at least three crabs. First-strand cDNA was synthesized with a Reverse Transcriptase kit (Takara Bio, Shiga, Japan) in accordance with the manufacturer’s instructions. For in vivo bacterial infection, bacteria (1 × 108 CFU/crab, 200 μl) were injected into the hemolymph from the nonsclerotized membrane of the posterior walking leg but at a different site from that used for hemolymph collection (Supplemental Fig. 1), with sterile PBS (200 μl) used as a control. Total RNA was collected from hemocyte samples pooled from at least three crabs that were collected at specific time points postinfection.
RNA interference assay
For the in vitro RNA interference (RNAi) experiments, specific small interfering RNA (siRNA) against the relevant genes and targeting GFP (control) were synthesized by GenePharma. The primer sequences are listed in Supplemental Table II. E. sinensis primary hemocytes isolated from healthy crabs were transfected with siRNA using RNAi-mate transfection reagent (GenePharma) at a final concentration of 10 nM in accordance with the manufacturer's instructions. At 24 h post–siRNA transfection, total RNA was extracted to evaluate RNAi efficiency by quantitative real-time PCR (qRT-PCR) and semiquantitative RT-PCR.
For the in vivo RNAi experiments, the final concentration of siRNA was adjusted to 2 μg/μl, each sample was injected with a Microliter syringe (25 μl and 50 μl) through the arthrodial membrane of the fifth walking leg, and it is designated on the ipsilateral side of crab in this study. The amount of siRNA infected into each crab was proportional to its body weight. The siRNA injections were repeated 24 h after the first siRNA injection, and the efficiency of RNAi was determined via real-time RT-PCR, with siGFP RNA as the control.
Gene expression profile analysis
The relative gene expression levels in differentially treated hemocytes were determined by real-time RT-PCR using the CFX96 Real-Time System (Bio-Rad Laboratories, Hercules, CA) and SYBR Premix Ex Taq polymerase (Tli RNaseH Plus; Takara Bio) with gene-specific primers (Supplemental Table II) and the following reaction conditions: 94°C for 3 min, followed by 40 cycles at 94°C for 10 s and 60°C for 1 min, and then melting from 65°C to 95°C. The obtained data were normalized to the control samples using the 2−ΔΔCT method. The results are expressed as the mean ± SD of three independent experiments.
Plasmids and transfection procedures
HEK-293T cells were transfected with plasmids carrying the PcDNA 3.0 vector, whereas S2 cells were transfected with plasmids carrying the pAc 5.1b vector. The PCR primers used to produce the plasmids are listed in Supplemental Table II. For transient transfection, HEK-293T cells seeded in 60-mm dishes were grown to ∼60% confluence and transfected with 2 μg PcDNA-construct using Lipofectamine 2000 transfection reagent (Invitrogen). S2 cells seeded in 60-mm dishes were grown to ∼80% confluence and transfected using Effectene Transfection Reagent (Qiagen, Hilden, Germany) in accordance with the manufacturer’s instructions. The crab hemocytes were transfected with plasmids carrying the pCDH vector, whereas the HEK-293T cells were transfected with the target plasmid and plasmids pspAX2 and pMD2G at a ratio of 2:2:1. After 48 h, the virus was collected from the infected crab hemocytes. The transfected cells were continuously cultured in complete medium for 48 h and then collected for subsequent experiments.
Assessment of crab survival and bacterial clearance assay
Crabs were randomly divided into two groups, each with 30 animals. Each crab in the first or second group was injected with siRNA of target gene or siRNA of GFP, respectively. After EsDscam was knocked down by the siRNA injection, all crabs were injected with S. aureus or Vibrio parahaemolyticus (1 × 109 CFU/crab, 200 μl). Dead crabs were counted daily in each group, and survival was tabulated over 6 d. Three days after the bacterial injection, each crab’s hemolymph was collected, diluted, and cultured. The number of CFUs was determined by manual counting after treatment.
Western blot analysis
Protein samples obtained from crab hemocytes, S2 cells, or HEK-293T cells were quantified with Pierce Coomassie Plus (Bradford) Assay Reagent (Thermo Fisher Scientific). Crab hemocytes, S2 cells, and HEK-293T cells were lysed using radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors (Roche Applied Science, Penzburg, Germany). Cytoplasmic or nuclear proteins were extracted using a Nuclear Protein Extracting Kit (Beyotime Institute of Biotechnology, Haimen, China) in accordance with the manufacturer’s instructions. Protein concentrations were quantified using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). The proteins were then separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane, which was blocked for 1 to 2 h with 3% nonfat milk in TBS (10 mM Tris-HCl [pH 8] and 150 mM NaCl), and then incubated with antiserum against the proteins of interest or with commercial Abs against hemagglutinin (HA), GAPDH, histone H3, or actin after confirming that the Abs generated nonspecific signals in TBS with 3% nonfat milk for 3 h. After washing three times with TBS, the membrane was incubated with alkaline phosphatase–conjugated goat anti-rabbit/mouse IgG (dilution 1:10,000 in TBS) for 3 h and then washed again to remove the unbound IgG. Afterward, the membrane was visualized by incubation with 4-chloro-1-naphthol oxidation for 5 min in the dark. Abs against HA, GAPDH, H3, and actin were purchased from Abcam (Cambridge, U.K.), whereas those against Dscam and IPO5 in E. sinensis were obtained from WuXi Apptec (Shanghai, China). All images were captured using an Odyssey CLx Imaging System (LI-COR Biosciences, Lincoln, NE).
Immunocytochemical staining was used to examine the subcellular location of Dscam in S2 and HEK-293T cells. Briefly, pretreated cells were blocked with 3% BSA for 30 min at 37°C and then incubated overnight at 4°C with a Dscam-tagged Ab against HA (dilution, 1:100 in blocking buffer). After washing with PBS, the cells were incubated with 3% BSA for 10 min followed by a secondary goat anti-mouse Alexa Fluor 488 Ab (dilution 1:1000 in 3% BSA) for 1 h at 37°C in the dark. Afterward, the cells were washed with PBS and then stained with DAPI (AnaSpec, San Jose, CA) for 10 min at room temperature, washed again, and observed under a Revolve Hybrid Microscope (Echo, San Diego, CA).
Coimmunoprecipitation (Co-IP) analysis was conducted as described in a previous report (12). In brief, 1 × 107 D. melanogaster S2 cells were seeded in 60-mm dishes overnight and transfected with 1 μg empty plasmid as a control or various expression plasmids. At 48 h posttransfection, the medium was carefully removed, and the cell monolayer was washed twice with ice-cold PBS. The cells were then lysed with 500 μl radioimmunoprecipitation assay buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, sodium orthovanadate, sodium fluoride, EDTA, and leupeptin) (Beyotime Institute of Biotechnology) containing a protease mixture (Yeasen Biotechnology, Shanghai, China). Lysates were centrifuged at 14,000 × g for 15 min, and the supernatant was transferred to a fresh tube and precipitated for 2 h at 4°C with 30 μl of either an anti-HA or anti-GFP affinity gel (Biotool, Houston, TX). The affinity gel was washed with cold TBS four times and eluted by boiling for 10 min in TBS and 6× SDS loading buffer (2% SDS, 60 mM Tris-HCl [pH 6.8], 10% glycerol, 0.001% bromophenol blue, and 0.33% 2-ME). The cell lysates were also eluted by boiling in the same 6× SDS loading buffer. Proteins isolated from the beads and cell lysates were separated by SDS-PAGE and subjected to Western blot analysis with the indicated Abs. All images were captured with an Odyssey CLx Imaging System.
Laser-scanning confocal analysis
Confocal analysis was used to test the possible interaction between ICDs and IPO5 according to our previous report (13). HEK-293T cells were seeded onto 24 × 24-mm glass cover slips in 35-mm dishes overnight and transfected with 1 μg each plasmid, followed by removal of the medium, cells washed, and fixed with 4% paraformaldehyde. The cells were then permeabilized with 0.5% Triton X-100 in PBS for 10 min, washed with PBS three times, blocked with 3% BSA for 2 h, and then incubated with anti-Flag Ab (1:2000) and anti-HA Ab (1:2000) overnight at 4°C. Unbound Ab was removed by washing with PBS plus Tween, and goat anti-mouse IgG H (Rhodamine Red-X Conjugated) (1:100) was used to stain the cells for 2 h at 37°C. After PBS plus Tween washing, cells were stained with DAPI for 5 min, and cover slips were sealed with nail polish on the edges for observation under a laser-scanning confocal microscope (Leica, Wetzlar, Germany).
The proliferation of crab hemocytes was detected using the CFSE and 5-ethynyl-2′-deoxyuridine (EdU) methods according to our previous report (13).
For the CFSE method, a CFSE stock solution was prepared (10 mM in DMSO; Invitrogen, Merelbeke, Belgium) and stored at −20°C. The stock solution was thawed and diluted in PBS to the desired working concentrations. Pilot experiments were conducted to test different CFSE-labeling conditions (final concentrations, 0.2, 0.5, 1, 2, 5, and 10 μM) to obtain high cell viability and broad CFSE signal measurements. The crab hemocytes were resuspended in PBS (with 0.1% BSA) at 2 × 106 cells/ml, incubated with 1 μM CFSE for 20 min at 37°C, and then washed and resuspended in culture medium for 5 min to stabilize the CFSE dye. Subsequently, the cells were plated and infected with the virus. After 48 h, the crab hemocytes were harvested, and the CFSE signal of gated cells was analyzed using flow cytometry.
For the EdU method, EdU incorporation was analyzed using the Click-iT Plus EdU Imaging Kit (Life Technologies) in accordance with the manufacturer’s instructions. Briefly, half of the crab hemocyte culture media was replaced with fresh media containing 20 μM EdU and incubated for another 24 h. Subsequently, the cells were plated and infected with the virus. After 48 h, the cells were fixed with formaldehyde, permeabilized with 0.5% Triton X-100, and incubated with a reaction mixture and Hoechst stain in accordance with the manufacturer’s instructions. After imaging with a fluorescence microscope, the percentage of EdU-positive cells was evaluated by fluorescence microscopy.
Recombinant expression, purification, and antiserum production
PCR fragments representing crab Dscam and IPO5 were amplified using specific primers (Dscam-F and Dscam-R; and IPO5-F and IPO5-R; Supplemental Table I). Subsequently, the PCR products were purified and digested with restriction enzymes (EcoRI and XhoI) for ligation of the final DNA fragments into the pET-28a (+) vector (Novagen, Madison, WI). Escherichia coli Rosetta (DE3) cells (TransGen Biotech, Beijing, China) were transformed with the recombinant plasmid pET-28a (+)-Dscam or pET-28a (+)-IPO5 to express the recombinant Dscam or IPO5 protein. Fusion protein expression was induced under four different conditions: 1 mM isopropyl-β-d-thiogalactoside (IPTG) for 3 h at 37°C; 0.25 mM IPTG for 3 h at 37°C; 1 mM IPTG for 3 h at 30°C; and 0.25 mM IPTG for 3 h at 30°C. The fusion protein was purified with Ni-NTA resin (Transgene) in accordance with the manufacturer’s instructions. Subsequently, rabbit antiserum against Dscam or IPO5 was prepared according to a previous report (14).
Following transfection with or without HA-tagged Dscam, stimulation with or without S. aureus, and ADAM10 inhibition or transfection, crab hemocytes or D. melanogaster S2 cells were washed twice with PBS, trypsinized, fixed with 4% paraformaldehyde for 10 min, washed twice with PBS, and stored at 4°C. After thawing, the cells were stained for 1 h with HA Ab (Abcam) at 4°C in PBS/1% BSA and washed twice with PBS-BSA prior to FACS analysis. Data were acquired using a CytoFLEX flow cytometer (Beckman Coulter, Irving, TX) and analyzed using FlowJo software version 10 (Beckman Coulter). A minimum of 10,000 cells were acquired for each experiment.
Panning of the T7 phage display library
Panning was performed according to our published method (13). Briefly, 40 μg purified Dscam protein was added to the wells of a 96-well plate and incubated at 4°C overnight. After removal of the unbound protein by PBS washing, 10 μl T7 phage display library derived from crab hemocytes was added to the wells for incubation. Dscam-bound phages were eluted using 200 μl 1% SDS, centrifuged at 6000 × g for 5 min, and then the supernatant (10 μl) was inoculated into 1 ml BLT5403 cell and cultured at 37°C for 2 h. After centrifugation, the supernatant was plated on an agarose plate containing BLT5403 cells and cultured at 37°C for 3 h. Extraction buffer (100 mM NaCl, 6 mM MgSO4, and 20 mM Tris-HCl [pH 8]) was added to the wells, incubated at 4°C overnight, and then collected for the next panning. After three repeats, single plaques appeared, and the eight longest fragments were sequenced.
Libraries were prepared from RNA acquired from three independent experiments for each group. Sequencing libraries were generated using the TruSeq RNA Library Preparation Kit (Illumina, San Diego, CA). Briefly, mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Fragmentation was conducted using divalent cations under elevated temperature in Illumina proprietary fragmentation buffer. First-strand cDNA was synthesized using random oligonucleotides and SuperScript II Reverse Transcriptase (Thermo Fisher Scientific). Second-strand cDNA synthesis was subsequently performed using DNA polymerase I and RNase H. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities, and the enzymes were removed. After adenylation of the 3' ends of the DNA fragments, Illumina PE adapter oligonucleotides were ligated prior to hybridization. To select cDNA fragments of the target length of 200 bp, the library fragments were purified using the AMPure XP system (Beckman Coulter). DNA fragments with ligated adaptor molecules on both ends were selectively enriched using an Illumina PCR Primer Cocktail in a 15-cycle PCR reaction. The PCR products were purified with the Agencourt AMPure XP PCR purification system (Beckman Coulter) and quantified using a high-sensitivity DNA assay (Agilent Technologies, Santa Clara, CA) with a Bioanalyzer 2100 instrument (Agilent Technologies). The library was then sequenced with the HiSeq platform (Illumina), and differentially expressed genes (DEGs) were profiled using the Hisat2-Stringtie-Ballgown pipeline. Transcript-level expression analysis of RNA-seq experiments was conducted with HISAT (hierarchical indexing for spliced alignment of transcripts) with a cutoff p value <0.05 and log2foldchange >1. The Cytoscape plugins ClueGO (version 2.5.7) and CluePedia were used to perform gene ontology enrichment analyses using the Kyoto Encyclopedia of Genes and Genome database.
Detection of Dscam cytoplasmic fragments and correlations with mediated nuclear translocation
Recent studies revealed that human DSCAM ICD was released by γ-secretase–dependent cleavage and then efficiently translocated to the nucleus by IPO5 (8). However, only arthropod Dscam has the ability to undergo alternative splicing, suggesting that only specific alternatively spliced cytoplasmic isoforms of Dscam can be imported into the nucleus. To confirm this possibility, the exon structure of the crab Dscam cytoplasmic tail was investigated, which showed that Dscam ICD was composed of nine exons (exons 32–40) (Fig. 1A) that encoded the Dscam ICD. The Dscam ICD generated a total of six distinct isoforms, which included the full-length ICD and five truncated isoforms produced by alternative splicing of exons 33, 36, and 35/36: ICD-Δ35, ICD-Δ35 + 36, ICD-Δ33, ICD-Δ33 + 35, and ICD-Δ33 + 35+36 (12). The highly conserved NLS motif for IPO5 binding spanned the junction between constant exon 32 and alternatively spliced exon 33, which also acted as a dock binding site in constitutive exon 32 that mediated the canonical Dscam pathway (Fig. 1B), suggesting the importance of exon 33 in mediating nuclear translocation of the Dscam ICD.
To test this hypothesis, a primary polyclonal Ab target was constructed that comprised the constant exon 38, which was shared by the Dscam proteome. The Western blot results showed that the Dscam protein of untreated hemocytes comprised a prominent ∼220-kDa fragment as well as a smaller 45-kDa fragment and that the amount of the 45-kDa fragment (the enzymatic cleavage product of Dscam) was largely induced (Fig. 1C) and significantly enhanced (Fig. 1D) post-coculturing with S. aureus, suggesting that the Dscam underwent enzymatic digestion.
Several transmembrane proteins, including Notch, undergo ectodomain cleavage (15–17), directly followed by γ-secretase–mediated intramembrane cleavage, leading to the release of the ICDs, indicating that the ICD of human DSCAM could also be released by γ-secretase (8). To verify that γ-secretase cleaved the crab Dscam ICD, hemocytes were treated with or without the proteasome inhibitor lactacystin in the presence or absence of two different γ-secretase inhibitors: N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester and (5S)-(t-butoxycarbonylamino)-6-phenyl-(4R)hydroxy-(2R)benzylhexanoyl)-l-leu-l-phe-amide (L-685, 458). Lactacystin was added to stabilize the intracellular fragments generated by γ-secretase, which are rapidly degraded by the proteasome (8). Western blot analyses of lysates in the presence of lactacystin and the absence of either γ-secretase inhibitor revealed the presence of a stabilized fragment of ∼45 kDa (Fig. 1E). Taken together, the results suggested that Dscam was also a γ-secretase substrate in invertebrates and that its cleavage released an apparent 45-kDa γ-secretase product.
Next, immunofluorescence staining was performed to investigate possible correlations between NLS-carrying Dscam ICDs and nuclear translocation. Six C-terminally HA-tagged Dscam PcDNA 3.0/pAc-5.1b plasmids with or without NLS and alternatively spliced exons 33, 35, and 36 were constructed (Fig. 1F) and expressed in HEK-293T cells to analyze the subcellular localizations and quantify the protein contents in the cytoplasm and nucleus (Fig. 1G, 1I). Both full-length and NLS-containing ICDs were found to be predominantly localized in the nucleus, whereas alternative splicing of exon 35 or 36 had no effect on the nuclear localization of Dscam ICD (Fig. 1G). However, Dscam ICD without an NLS or alternatively spliced exon 33 showed very weak localization and protein levels in the nucleus (Fig. 1G, 1I). Furthermore, HA-tagged DSCAM without alternatively spliced exons 33, 35, and 35 + 36 was expressed in Drosophila S2 cells (Fig. 1H, 1J). The results showed that exon 35 or/and 36 had no effect on Dscam localization and protein levels in the nucleus, whereas the lack of exon 33 dramatically suppressed the localization and protein levels in the nucleus (Fig. 1H, 1J), consistent with the results in HEK-293T cells. Collectively, these findings indicated that the cytoplasmic fragment of crustacean Dscam ICDs and its nuclear translocation resembled the DSCAM of mammalian vertebrates, highlighting the indispensable role of alternative splicing of exon 33 in mediating nuclear import of Dscam ICD.
Nuclear translocation of Dscam ICDs carrying exon 33 is mediated by direct interaction with IPO5
IPO5 functions as a nuclear transport receptor for the import of nuclear proteins, serves as a receptor for the NLS of cargo substrates, and plays a crucial role in mediating nuclear import of human DSCAM (8). However, the expression pattern of IPO5 in invertebrates remains largely unknown. IPO5 was widely expressed in different tissues of the crab (Fig. 2A), and expression was significantly induced during the early stages of infection with S. aureus (Fig. 2B), suggesting a potential role in immunity.
To determine whether IPO5 mediated nuclear translocation of Dscam ICD, siRNA was used to knockdown IPO5 expression. This RNAi system suppressed the expression of IPO5 by >50% at both the mRNA (Fig. 2C) and protein (Fig. 2D) levels. Furthermore, Western blot analysis (Fig. 2E) and immunostaining (Fig. 2F) showed that IPO5 positively regulated Dscam ICD localization and protein levels in the nucleus, respectively.
To determine the role of alternatively spliced exons in binding to IPO5, Co-IP assays were conducted with polyclonal Abs that against Dscam or IPO5 in crab hemocytes, which showed Dscam binds to IPO5 in hemocytes (Fig. 2G). Moreover, we used different ICD plasmids to analyze the role of alternative spliced exons on protein interaction. The results showed that only the isoform lacking exon 33 significantly reduced binding to IPO5, whereas the lack of exon 35 or 35 + 36 had no effect on these protein interactions (Fig. 2H). Subsequently, laser confocal scanning in HEK-293T cells also demonstrated the critical role of exon 33 on ICDs-IPO5 interaction (Fig. 2I). Taken together, these results showed that in arthropods, alternative splicing of exon 33 has a critical role in the interaction with IPO5 and subsequent nuclear translocation.
RNA-seq indicates nuclear localization of Dscam ICDs may regulate cell proliferation
Studies with primary neurons and cell lines have suggested that transcriptional regulation via membrane-to-nucleus signaling plays a critical role in repressing the functions of DSCAM during neuronal circuit formation (8), yet this mechanism remains largely unclear in arthropods. Because it is very difficult to perform gene overexpression experiment in primary cultured crab hemocytes, we studied if nuclear imported ICDs regulated gene expression in S2 cells. S2 cells were stably expressing the full-length HA-tagged Dscam ICD (i.e., ICD-FL), ICD without alternatively spliced exon 35 (i.e., ICD-Δ35), and ICD without alternatively spliced exons 35 and 36 (i.e., ICD-Δ35 + 36), as well as HA-tagged NLS controls were generated for RNA-seq. Then, changes in the global transcriptome of cells expressing HA-tagged Dscam ICDs versus cells expressing HA-tagged NLS as controls were determined (Fig. 3A). Statistical analysis (p < 0.05; log2 FoldChange >1) of DEGs that were compared with different ICD-expressing cells and controls revealed a total of 1401 DEGs using Dscam ICD-FL expression: 413 using Dscam ICD-Δ35 expression and 22 using Dscam ICD-Δ35 + 36 expression (Fig. 3B). Moreover, no DEG was common among the three Dscam ICDs (Fig. 3C).
To understand the function of DEGs regulated by Dscam ICDs, gene ontology enrichment analysis with S2 cell DEGs was used. The results showed that the signaling pathways and related genes involved for cell proliferation were significantly enriched, and we narrowed the ICD-FL, ICD-Δ35, and ICD-Δ35 + 36 data sets down to ∼22 candidate genes based on their well-known functions in cell proliferation, cell cycle process, and hemocyte proliferation (Fig. 3D). To verify gene expression level after different Dscam ICD nuclear translocations, we overexpressed different ICDs into the hemocytes of crab and then tested the genes covering different terms, including Tl, CycE, Cdc2, STAT, and platelet-derived growth factor (PDGF)/vascular endothelial growth factor (VEGF), and CSN5. The results showed that these genes were induced by different Dscam ICDs, which was consistent with the results of RNA-seq (Fig. 3E).
Dscam ICDs with obvious nuclear transport can promote cell proliferation
The RNA-seq data revealed that ICDs regulated cell proliferation. Therefore, as shown in (Fig. 4A, the crab hemocytes were labeled with CFSE and then infected with three types of Dscam ICDs with obvious nuclear localization through lentivirus packaging to determine whether these three ICDs have the ability to regulate the proliferation of hemocytes. The flow cytometry results showed that all three Dscam ICDs had the ability to promote cell proliferation, with ICD-FL having the strongest effect, followed by ICD-Δ35 and ICD-Δ35 + 36 (Fig. 4B, 4C). To further confirm these results, crab hemocytes were labeled with EDU (Fig. 4D), which showed the strongest fluorescence in the ICD-FL group, followed by ICD-Δ35 and ICD-Δ35 + 36 (Fig. 4E), consistent with the flow cytometry results.
Nuclear import of specific Dscam ICDs controlled the antibacterial activities in hemocytes
The above experiment confirmed that the specific Dscam ICDs can promote cell proliferation, but it is not clear whether the proliferating cells have bactericidal function. To test these possibilities, we overexpressed each ICD with nuclear import activity into crab hemocytes in vitro and found that all three ICDs promote bacterial clearance, among which ICD-FL was the strongest, followed by ICD-Δ35 and ICD-Δ35 + 36 (Fig. 5A). This result also corresponds to their ability to promote cell proliferation. Considering the critical role of exon 33 in Dscam ICDs nuclear translocation, we injected siRNA into the crabs in vivo to knock down the expression of alternatively spliced exon 33 containing Dscam isoforms and then infected with S. aureus (Fig. 5B) or V. parahaemolyticus (Fig. 5C); results have shown exon 33 positively regulates crab survival rate, which demonstrated the important function of Dscam ICD nuclear translocation on antibacterial immunity.
Bacterial-induced Dscam ectodomain shedding may be caused by ADAM10
The shedding and cleavage of the Dscam ectodomain has been confirmed in both mice (18) and fruit flies (4). Moreover, the cleavage of type I transmembrane receptors that include notch, amyloid precursor protein, and receptor protein tyrosine phosphatases κ were mediated by ectodomain shedding, then followed by the cleavage of ICDs (19–21). Therefore, we speculate that the shedding of the Dscam ectodomain may activate cytoplasmic ICD cleavage by γ-secretase. However, the protease responsible for the shedding of Dscam receptors has not yet been identified. To investigate this process, we determined whether Dscam was constitutively shed in vitro after being first regulated by bacterial stimulation. A plasmid with a GFP tag at the N-terminal and an HA tag at the C-terminal (GFP-Ig8-HA) (Fig. 6A) was transfected into crab hemocytes for stable expression with or without S. aureus stimulation. FACS showed that GFP-tagged Dscam was not detected in untreated crab hemocytes, plasmid-transfected crab hemocytes showed Dscam shedding, and S. aureus stimulation enhanced the shedding of Dscam (Fig. 6B).
Then, proteases with the potential to facilitate Dscam shedding were screened. After constructing a transmembrane-recombinant Dscam protein that covered the region from Ig8 to the last domain of the outer membrane shared by all Dscam isoforms (Fig. 6A, 6C), a T7 phage display library expressing crab genes was used to screen potential Dscam-binding proteins. The results showed some possible binding proteins (Fig. 6D), with ADAM10 being of particular interest because of its role in shedding cell membrane receptors, like Notch and others.
ADAM10 binds with Dscam to mediate ectodomain shedding and intracellular cleavage
Both protease ADAM10 and ADAM17 have critical roles in shedding of type I cell-membrane receptors, followed by the cleavage of ICDs (22, 23). Therefore, the potential role of ADAM10 and ADAM17 in shedding of Dscam was tested by using S2 cells as a research model, which is an available cell line to overexpress genes. In this study, Dscam-expressing S2 cells were treated with or without GI254023X and GW280264X, which specifically inhibit ADAM10 and ADAM17 to make them noncatalytically active, respectively (22, 24). The FACS results showed that GFP-tagged Dscam was not detected in untreated S2 cells and that plasmid-transfected S2 cells showed Dscam shedding at 0 h after stimulation with S. aureus, whereas overexpression of ADAM10 in S2 cells showed enhanced Dscam shedding, as compared with the suppressed Dscam shedding in ADAM10-inhibited S2 cells at 12 h poststimulation with S. aureus (Fig. 7A). However, ADAM17 had no role in Dscam ectodomain shedding (Fig. 7B), which demonstrated that ADAM10 might bind to Dscam during ectodomain shedding.
To confirm the interactions among Dscam, ADAM10, and ADAM17, Co-IP assays were conducted with the use of S2 cells. The results showed that Dscam bound to ADAM10 rather than ADAM17 (Fig. 7C), which was consistent with the results of the T7 library screenings. Furthermore, HA-tagged Dscam ICD cleavage fragment (∼45 kDa) was enhanced in S2 cells overexpressing ADAM10, but suppressed in S2 cells where ADAM10 was inhibited (Fig. 7D, 7E), which showed the ectodomain shedding of Dscam by ADAM10 is important for Dscam ICDs cleavage.
Canonical Dscam signaling regulates the expression of ADAM10
The real-time RT-PCR results showed that S. aureus infection (Fig. 8A) or stimulation (Fig. 8B) induced high mRNA expression of ADAM10 in hemocytes in the later stages, indicating a possible relationship with immune reactions. The importance of the Toll and immune deficiency signaling pathways in the regulation of crustacean immunity has been widely confirmed. To determine whether these two signaling pathways regulated the expression of ADAM10, RNAi was used to silence crab Dorsal (Fig. 8C) and Relish (Fig. 8D) to observe changes in ADAM10 expression after bacterial infection. The results clearly showed that ADAM10 expression was significantly decreased after silencing of Dorsal (Fig. 8E), suggesting that ADAM10 expression might be regulated through the Toll pathway (Fig. 9). The results of our previous study showed that the cytoplasmic tail of Dscam can bind to the Src homology 3 (SH3) domain of the Dock protein, which then promotes ERK phosphorylation via indirect binding and regulates Dorsal phosphorylation and translocation from the hemocyte cytoplasm to nucleus (12). To determine whether ADAM10 expression was regulated by this canonical pathway, crab Dscam (Fig. 8F) and ERK (Fig. 8G) were also silenced, and the siDscam primers target its constitutive exon 32. The results showed that silencing of Dscam or ERK significantly decreased ADAM10 expression (Fig. 8H). The EsDscam ICD generated a total of six distinct isoforms (Fig. 1A), three of which had the ability to regulate nuclear transcription. To determine whether nuclear translocation of ICDs regulated ADAM10 expression, these three ICDs with nuclear import activity were overexpressed. Subsequent qRT-PCR analysis showed that there was no significant difference in ADMA10 expression between the overexpression and control groups regardless of bacterial stimulation (Fig. 8I–K).
Because alternative spliced Dscam isoforms are capable of specifically binding bacteria and promote bacterial phagocytosis in hemocytes, they are considered to be an “antibody-like” protein in invertebrates (25). In this study, novel membrane-to-nucleus signaling of Dscam to regulate hemocyte proliferation was found in the crab to defend against bacterial infection. This pathway was tightly regulated by canonical Dscam signaling (12) regulation of ADAM10 expression, ADAM10-mediated ectodomain shedding, γ-secretase control of ICD cleavage, and IPO5-mediated nuclear translocation of specific ICDs.
It has been shown that invertebrate hemocytes play crucial roles in the immune defenses of arthropods (26). However, the mechanisms underlying the triggering of hemocyte proliferation and differentiation during an immune response have remained elusive. Pioneering studies on hematopoiesis in insects and shrimp demonstrated that maintenance of hemocyte populations during the larval stage was also dependent on both the proliferation of cells already in circulation and the release of hemocytes from hematopoietic organs (27). However, maintenance of hemocyte populations via proliferation of cells already in circulation appears to be more important during the early stage of infection, whereas the production and release of hemocytes from the hematopoietic organs are delayed. Several studies have implicated the JAK/STAT (28) and Pvr (29) signaling pathways in the proliferation of prohemocytes. PDGF/VEGF-like growth factor induces hemocyte proliferation in Drosophila larvae, and differentiation of prohemocytes into plasmatocytes, crystal cells, or lamellocytes requires downregulation of Dome in insects (30). In the current study, Dscam ICDs were found to be critical for regulation of hemocyte proliferation, which might be dependent on the expression of proliferation-related genes, including positive regulate genes that include Cyclins (31) and PDGF/VEGF (29) that promote hemocyte proliferation and negative genes that regulate the expression of CSN5 that has been proved to suppress cell proliferation (32). However, further studies are needed to determine whether a single cell expresses all alternatively spliced Dscam ICD isoforms and to elucidate the detailed mechanisms of ICD-regulated hemocyte proliferation.
Dscam promotion of hemocyte proliferation is tightly regulated by membrane-to-nucleus signaling via ADAM10, γ-secretase, and IPO5. The potential phenomenon of Dscam ectodomain cleavage and shedding has been shown in Drosophila (4); however, the proteases responsible for Dscam shedding have not yet been identified. In the current study, ADAM10, but not ADAM17, specifically interacted with the Dscam receptor, leading to the cleavage and shedding of the Dscam ectodomain, thus strongly enhancing the cleavage of cytoplasmic Dscam ICDs (Fig. 7). In the case of Notch, metalloprotease cleavage is dependent on ligand binding, whereas subsequent intramembrane cleavage by γ-secretase is constitutive (15, 33). In this study, Dscam cleavage released cytoplasmic ICDs with different alternatively spliced exons, and only the ICDs carrying alternative spliced exon 33 were efficiently translocated from the cytoplasm into the nucleus via direct binding with IPO5, which functioned in nuclear protein import as a nuclear transport receptor and served as a receptor for NLS in cargo substrates.
Our previous study showed that bacteria-activated crab membrane-bound Dscam initiated canonical antibacterial responses via the Dock-ERK-Dorsal axis, resulting in significant upregulation of antimicrobial peptides (12). The results of the current study revealed novel membrane-to-nucleus Dscam signaling that promoted ectodomain shedding, cytoplasmic enzymatic cleavage, and nuclear import to promote hemocyte proliferation during bacterial stimulation. The latter signaling was activated by ADAM10-mediated shedding of the Dscam ectodomain, which suppressed the expression of antimicrobial peptides by canonical Dscam signaling because the bacteria-binding alternative spliced Ig domains (4, 7) were released from the Dscam receptor. Thus, understanding the relationship between these two Dscam signaling pathways could help to clarify the function of Dscam in immune reactions. Furthermore, these results demonstrate that ADAM10 expression was largely induced at the latter stage of bacterial stimulation, and the transcription of ADAM10 was significantly regulated by canonical Dscam signaling rather than the nuclear translocation of Dscam ICDs. This is consistent with the results of our RNA-seq, whereas the expression of ADAM family genes was not regulated by ICDs nuclear import. Collectively, these data suggested that in response to bacterial infection, Dscam regulated the expression of antimicrobial peptides and ADAM10 via canonical signaling in the early phase, enrichment of ADAM10 at the cell membrane activated another Dscam pathway via ectodomain shedding, and Dscam promoted hemocyte proliferation through membrane-to-nucleus signaling in the later phase. Thus, the two signaling pathways induced by Dscam were linked by ADAM10, which protected the host from bacterial infection by promoting the expression of antimicrobial peptides and hemocyte proliferation in the early and latter phase, respectively.
In summary, our study demonstrates that alternatively spliced Dscam possesses nuclear imported receptor properties that allow it to mediate host innate immunity through a novel membrane-to-nuclear pathway in addition to the canonical signaling cascades (Fig. 9). Importantly, we established the feedback relationship and time axis between the canonical and membrane-to-nucleus signaling of Dscam. We also identified conclusive evidence that this novel membrane-to-nuclear signaling mechanism regulates cell proliferation in invertebrates, and the proliferation of hemocytes have a significant bacterial clearance capacity. This study advances our knowledge about the novel signaling regulatory mechanisms of Dscam signaling on invertebrate innate immunity and reveals the significance of Dscam in invertebrate antimicrobial responses. The comprehensive identification of the Dscam signaling pathway also provides insights into innate immunity in invertebrates.
We thank the Experimental Platform for Molecular Zoology (East China Normal University, Shanghai, China) for providing instruments essential for conducting this study.
This work was supported by the National Natural Science Foundation of China (31972820 to W.-W.L. and 31970490 to Q.W.), Shanghai Rising-Star Program (20QA1403000 to W.-W.L.), and Fundamental Research Funds for the Central Universities (B200202141 to X.-K.J.).
H.L., Q.W., and W.-W.L. designed research; H.L., K.-M.Z., and H.Z. performed research; X.-K.J. performed bioinformatic analyses; Y.-H.Z. contributed new reagent/analytic tools; H.L., X.-K.J., and W.-W.L. analyzed data; W.-W.L. wrote the paper; W.-W.L. supervised the study.
The RNA-sequencing data in this article have been submitted to the National Center for Biotechnology Information Sequence Read Archive (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA649344) under BioProject accession code PRJNA649344.
The online version of this article contains supplemental material.
Abbreviations used in this article
a disintegrin and metalloprotease domain
differentially expressed gene
Down syndrome cell adhesion molecule
human embryonic kidney
nuclear localization signal
platelet-derived growth factor
quantitative real-time PCR
RNA deep sequencing
Src homology 3
small interfering RNA
vascular endothelial growth factor
The authors have no financial conflicts of interest.