SIDT1 Localizes to Endolysosomes and Mediates Double-Stranded RNA Transport into the Cytoplasm

Double-stranded RNA is a common by-product of replication during most viral infections (1), and its detection acts as a potent trigger of host antiviral immunity. The key cellular receptors responsible for viral dsRNA recognition consist of the endolysosomal protein TLR3 (2) and the cytoplasmic retinoic acid–inducible gene I (RIG-I)–like helicase receptors (RLR) and melanoma differentiation associated gene 5 (MDA-5) (3–5). Upon binding dsRNA, downstream signaling via each of these pathways culminates in the production of type I IFN and proinflammatory cytokines (2, 6).

During viral infections, infected cells can be lysed by the virus directly or via cell death, and this leads to the liberation not only of viral particles but also viral dsRNAs into the extracellular space (7). These extracellular dsRNAs can then be internalized by surrounding cells via clathrin-mediated endocytosis (8), where they are accessible for sensing by endosomal TLR3. Interestingly, dsRNA that has been internalized from the extracellular space also activates the RLR sensing pathway within the cytoplasm, and, surprisingly, it is this cytoplasmic sensing pathway that is largely responsible for facilitating IFN production following in vivo exposure to extracellular dsRNA (9). Specifically, it was observed that mice lacking MDA-5 produced significantly less type I IFN compared with wild type (WT) and TLR3-deficient mice following systemic injection of poly(I:C), which led the authors to postulate the existence of an efficient transport mechanism to deliver extracellular viral dsRNAs into the cytoplasm for RLR-dependent production of type I IFN (9).

We recently demonstrated that SIDT2, a mammalian ortholog of the Caenorhabditis elegans SID-1 dsRNA transporter (10–12), localizes within endolysosomes and transports internalized dsRNA into the cytosol for innate immune activation via RLRs (13). Mice deficient in SIDT2 produce less IFN-β and display impaired survival following HSV type 1 (HSV-1) and encephalomyocarditis virus (EMCV) infection (13).

In addition to SIDT2, mammals all share another SID-1 ortholog, SIDT1, but whether it too plays a role in the transport of viral dsRNA into the cytoplasm is unknown. SIDT1 has been previously reported to localize to the plasma membrane when overexpressed in a pancreatic cancer cell line (PANC1), and it was observed that this overexpression resulted in increased small interfering RNA (siRNA) uptake and enhanced gene knockdown (14). Consistent with this, knockdown of SIDT1 has been shown to result in a decrease in silencing by exogenous cholesterol-conjugated siRNAs in cultured liver cells (15). Moreover, biochemical studies of SIDT1 have suggested that its extracellular domain binds dsRNA (16). However, a more recent study found that SIDT1 was unable to facilitate uptake of longer dsRNA across the plasma membrane, and, instead, facilitated the transport of extracellular cholesterol into the cell (17).

It thus remains unclear from existing studies whether SIDT1 shares the dsRNA transport activity of SIDT2 and whether this has a physiological role. In this study, we demonstrate that SIDT1 can transport extracellular dsRNA into the cytoplasm following endocytosis in vitro and, moreover, is necessary for optimal production of type I IFN in vivo during infection with HSV-1. However, we also observed that SIDT1 is not essential for viral clearance or animal survival following infection in vivo, suggesting that its dsRNA transport activity is likely to be functionally redundant in the presence of SIDT2.

Sidt1^−/− mice were derived using embryonic stem cells carrying a Sidt1–targeting allele from Velocigene and provided via the Knockout Mouse Project (http//:www.komp.org). Mice were bred and maintained on a C57BL/6 background. The targeting allele leads to replacement of coding exons 1–3 with the Velocigene cassette ZENUb1, which includes a LacZ reporter and a floxed neomycin cassette. C57BL/6 mice were used as WT controls. This targeting event is expected to remove the signal peptide and result in a truncated, nonfunctional protein. Mice were bred and maintained in the animal facilities at the Walter and Eliza Hall Institute of Medical Research according to national and institutional guidelines for animal care. All experimental procedures were approved by the relevant animal ethics committees at the Walter and Eliza Hall Institute of Medical Research.

Primary bone marrow–derived dendritic cells (BMDCs) and bone marrow–derived macrophages (BMDMs) were obtained and cultured as previously described (18–20). DC2.4 cells were cultured in DMEM supplemented with 10% FBS. Cells stably expressing SIDT1-mCherry were generated using lentiviral transduction as described previously (13).

SIDT1-mCherry–transduced DC2.4 cells were treated with 1 μg/ml doxycycline for 32 h to induce SIDT1-mCherry expression and treated in ibidi eight-well chamber μ-Slides (Martinsried, Germany) with either 2 μg/ml poly(I:C)-fluorescein, or transfected with 100 nM dsDNA-fluorescein or siRNA-fluorescein complexed with FuGENE for 1 h at 37°C. Cells were then washed and fixed in 4% paraformaldehyde for 10 min on ice. Fluorescence-lifetime imaging microscopy (FLIM) data were recorded using an Olympus FV1000 microscope equipped with a PicoHarp 300 FLIM extension and a 485-nm pulsed laser diode from PicoQuant (Berlin, Germany). Cells were first imaged for green and red fluorescence by confocal microscopy using the Olympus FV1000 system with a 60× oil immersion objective to verify presence of both donor (fluorescein) and acceptor (mCherry) dyes. Subsequently, corresponding FLIM images of green fluorescence were recorded using the PicoHarp extension. Pixel integration time for FLIM images was kept at 40 ms per pixel, and fluorescence lifetime histograms were accumulated to at least 10,000 counts to ensure sufficient statistics for Förster resonance energy transfer (FRET)–FLIM analysis. Photon count rates were kept below 5% of the laser repetition rate to prevent pileup. FRET­–FLIM analysis was performed using the SymPhoTime 64 software (PicoQuant). Poly(I:C)-fluorescein– or dsDNA-fluorescein­–positive compartments of individual cells were chosen as regions of interest, then the fluorescence lifetime decay of each region of interest was deconvolved with the measured instrument response function and fitted with a biexponential decay. The amplitude weighted average lifetime was extracted from each fit and averaged over all values of one sample condition. The p values were determined to assess the significance of donor lifetime changes in the absence and presence of the SIDT1-mCherry acceptor. For FRET–FLIM analysis between SIDT1 and SIDT2, DC2.4 cells coexpressing SIDT1-mCherry and SIDT2-GFP were compared with cells expressing SIDT2-GFP alone (donor alone).

Sidt1 expression in mouse and human hematopoietic cells and tissue was obtained from publicly accessible microarray and RNA sequencing (RNAseq) expression data (21) accessed via Haemosphere (https://www.haemosphere.org). Average expression values obtained from the Haemopedia-Plus and Haemopedia-Human-RNAseq dataset was exported into GraphPad Prism 7 software to generate heat maps for visualization purposes.

Tissue specimens were fixed in 10% (w/v) neutral-buffered formalin for at least 12 h before being embedded in paraffin, sectioned (3 μm), and prepared for H&E staining. Specifically, slides were incubated with hematoxylin to stain nuclei (5 min), washed in dH₂O (1 min), then incubated in 0.3% (v/v) acid ethanol to destain (1 dip), rinsed in dH₂O followed by Scott’s tap water, then incubated in eosin (1 min). Slides were then rinsed in dH₂O, dehydrated, cleared, and mounted.

For poly(I:C) challenge, age-matched female mice were injected i.p. with 50 μg poly(I:C) per 25 g body weight, and serum was collected 3 and 6 h later for IFN analysis. For viral infection studies, 6-wk-old female mice were infected with 2 × 10⁷ PFU HSV-1 (KOS strain) or 50 PFU of EMCV via i.p. injection, and their survival was monitored for 8 d. For cytokine analysis, serum was harvested at 16 h postinfection (p.i.) with HSV-1. Assessment of serum IFN-β was performed via ELISA according to the manufacturer’s instructions (PBL Assay Science, Piscataway, NJ). Serum IL-6, RANTES, and IL-12 p40 levels were measured via a Bio-Plex cytokine assay (Bio-Rad Laboratories), according to manufacturer’s instructions.

WT BMDMs were seeded in six-well plates at a density of 2 × 10⁶ cells per well and stimulated with either 1000 U/ml mouse IFN-β (catalog no. 12400-1; PBL Assay Science) or 50 ng/ml mouse IFN-γ (catalog no. 485-MI; R&D Systems) for 4 h at 37°C, 5% CO₂. Primary keratinocytes were isolated from the tails of adult mice as previously described (22) and stimulated with 100 ng/ml IFN-λ (catalog no. 4635-ML; R&D Systems) for 4 h at 37°C, 5% CO₂. Cells were subsequently washed and lysed for RNA isolation. For Sidt1 RNA analysis, total RNA was isolated from WT BMDMs, BMDCs, or SIDT1-mCherry expressing mouse embryonic fibroblasts (MEFs) using the RNeasy Plus Mini Kit (QIAGEN), according to manufacturer’s instructions. Two micrograms of total RNA was reverse-transcribed using Superscript III Reverse Transcriptase enzyme (Thermo Fisher Scientific), and quantitative PCR was performed on a Viia 7 system using AccuPower 2× GreenStar qPCR MasterMix (Bioneer). The following primers were used: Sidt1 (forward, 5′-CCTCAGCACCGAGAACATCT-3′ and reverse, 5′-ACTTCTTTCTGCTGGCGAAC-3′); Rig-i (forward, 5′-AAAGACGGTTCACCGCATAC-3′ and reverse, 5′-TCTTGCACTTTCCACACAGC-3′); and Gapdh (forward, 5′-CAACTTTGTCAAGCTCATTTCCTG-3′ and reverse, 5′-CCTCTCTTGCTCAGTGTCCTT-3′).

For assessment of IFN regulatory factor (IRF) 3 phosphorylation, 2 × 10⁶ Sidt1^+/+ and Sidt2^−/− BMDMs were lysed in RIPA buffer supplemented with protease inhibitor and phosphatase inhibitor (Roche) and denatured in 4× SDS-PAGE sample buffer at 95°C. Proteins were separated on NuPAGE Novex 4–12% Bis-Tris Gels (Life Technologies), and transferred electrophoretically to Hybond ECL nitrocellulose membrane blocked in TBST supplemented with 5% BSA, incubated with the relevant primary Abs, and washed again. The primary Abs used were as follows: phospho–IRF3 (Ser396; catalog no. 29047; Cell Signaling Technology) and IRF3 (catalog no. sc-9082; Santa Cruz Biotechnology). Membranes were then incubated for 1 h with HRP-conjugated secondary Abs, washed, and treated with Luminata Forte Western HRP substrate (Millipore) and visualized on the ChemiDoc MP Imaging System (Bio-Rad Laboratories). β-Actin–HRP (catalog no. ab49900; Abcam) was used as a loading control.

To assess poly(I:C) internalization, cells were treated with 1 μg/ml of fluorescein-conjugated poly(I:C) (Invivogen) at 37°C for 1 h, washed three times with ice cold PBS, and analyzed on an LSR II flow cytometer (BD Biosciences). To characterize the subcellular localization of dsRNA, doxycycline-treated, SIDT1-mCherry–transduced DC2.4 cells were treated for 1 h with 1 μg/ml poly(I:C)-fluorescein at 37°C in eight-well chamber μ-Slides (ibidi), washed in PBS and fixed in 4% paraformaldehyde on ice, and then imaged on a Zeiss LSM 780 confocal microscope. To assess the proportion of poly(I:C) within the endocytic compartment, individual cells were delineated from brightfield images and the poly(I:C)-fluorescein signal was thresholded using the FIJI software package (23) to ensure that only those fluorescein-bright punctate regions within the cell were evident. For each individual cell, the total area occupied by these regions was then calculated as a proportion of the entire cell area; an average percentage was then derived across the entire cell population.

Statistical analyses were performed in GraphPad Prism 7 software using unpaired, two-tailed Student t tests where appropriate. For Bio-Plex cytokine analysis following HSV-1 infection, a two-way ANOVA was performed using Sidak multiple testing correction. For HSV-1 and EMCV survival analysis, a generalized Wilcoxon (Gehan–Breslow) test was used to compare survival curves. The p values <0.05 were considered statistically significant.

To determine the subcellular localization of mouse SIDT1, we generated MEFs that were stably transduced with a doxycycline-inducible mCherry-tagged SIDT1 overexpression construct. Unlike previous reports that human SIDT1 localizes to the plasma membrane (14, 15), we observed that mouse SIDT1 did not colocalize with the plasma membrane marker wheat germ agglutinin (Fig. 1A) but, rather, appeared within discrete punctate structures consistent with endosomal vesicles. To determine in which endosomal compartment SIDT1 localized, we transiently transfected cells with different endosomal markers and observed that, although SIDT1 did not colocalize with early endosomes (EEA1), it did colocalize with the late endosomal and lysosomal markers Rab7 (Fig. 1C) and LAMP1 (Fig. 1D), respectively. These results are reminiscent of previous observations that showed that SIDT2 also localizes to late endosomes and lysosomes (13, 24).

In a complementary approach, we generated MEFs that stably coexpress doxycycline-inducible SIDT1-mCherry and SIDT2-GFP (Fig. 2A, 2B). Consistent with their late endolysosomal localization, SIDT1-mCherry colocalized with SIDT2-GFP (Fig. 2C). Moreover, we also observed a significant and robust FRET–FLIM interaction between SIDT1-mCherry and SIDT2-GFP (Fig. 2D), suggesting that the two proteins interact and might share similar biological functions.

To begin to understand the possible biological role of SIDT1, we next sought to identify the tissues and cells in which it is normally expressed in vivo. Analysis of publicly available microarray data revealed that Sidt1 is strongly expressed in the brain, with low expression in other tissues (Fig. 3A). Sidt1 was also previously reported to be highly expressed in the thymus of adult mice (17). We, therefore, undertook a focused analysis of Sidt1 expression in hematopoietic cells. This revealed that Sidt1 is highly expressed in immune cells such as T and NK cells and moderately expressed in other cell types such as plasmacytoid dendritic cells and B cells in mice (Fig. 3B), and a similar expression pattern was observed in human immune cells (Fig. 3C). This expression pattern was in keeping with a potential role for SIDT1 during immune responses and was further supported by previous observations that Sidt1 is upregulated in the liver of mice following West Nile virus infection as well as in corneal epithelial cells following HSV-1 infection in vivo (25, 26).

The expression of many genes involved in the immune response to viral infection is stimulated by the production of type I IFNs, which can in turn be induced by exposure to dsRNA. To test whether Sidt1 might fall into this category, we treated WT BMDMs with various doses of extracellular poly(I:C) and measured Sidt1 mRNA levels using quantitative real-time PCR. Notably, Sidt1 was robustly upregulated in response to poly(I:C) in a dose-dependent manner (Fig. 3D). To determine whether this induction is specific to BMDMs or more broadly influenced immune cells, we also generated WT BMDCs and measured Sidt1 expression following extracellular poly(I:C) treatment over time. Similarly, Sidt1 expression increased upon poly(I:C) stimulation in a time-dependent manner (Fig. 3E). These data suggested that Sidt1 might be a type I IFN–stimulated gene (ISG), and to test this formally, we treated BMDMs and mouse dermal fibroblasts with either type I (IFN-β) or type II (IFN-γ) IFN and measured Sidt1 expression after 4 h. Notably, Sidt1 was induced in both cell types following stimulation with IFN-β but not IFN-γ, and this induction was comparable to that observed for the classic type I ISG Rig-i (Fig. 3F, 3G). Finally, because some type I ISGs are also induced by type III IFN (IFN-λ), we isolated and cultured primary keratinocytes from WT mice and, subsequently, measured Sidt1 expression following stimulation with IFN-λ. Interestingly, whereas we were able to detect induction of Rig-i in this way, we did not detect any change in Sidt1 expression (Fig. 3H). Taken together, these data, therefore, indicate that Sidt1 is induced in response to type I but not type II or type III IFN.

Given that SIDT1 interacts with SIDT2 and is upregulated following viral infection and exposure to dsRNA and type I IFN, we hypothesized that SIDT1 might be involved in endolysosomal trafficking of dsRNA, similar to SIDT2. To test this, we assessed whether SIDT1 is able to interact with dsRNA. We generated DC2.4 cells that stably express SIDT1-mCherry under the control of a doxycycline-inducible promoter (Supplemental Fig. 1) and, following incubation with doxycycline for 32 h, these cells were treated with either the long dsRNA analog poly(I:C), a short siRNA, or dsDNA (as a negative control) and visualized via confocal imaging. Consistent with its endolysosomal localization, SIDT1 colocalized with internalized poly(I:C)-fluorescein, siRNA-fluorescein, and dsDNA-fluorescein (Fig. 4A). To assess whether SIDT1 is able to interact with any of these nucleic acids, we performed FRET–FLIM analysis on SIDT1-mCherry⁺ or SIDT1-mCherry⁻ cells following exposure to either poly(I:C)-fluorescein, siRNA-fluorescein, or dsDNA-fluorescein. Similar to SIDT2 (13), we observed a significant reduction in fluorescence lifetime for poly(I:C)-fluorescein (Fig. 4B) but not siRNA-fluorescein or dsDNA-fluorescein (Fig. 4C, 4D) in the presence of SIDT1-mCherry. These data demonstrate that, like SIDT2, SIDT1 is able to interact with long dsRNA [poly(I:C)] but not short dsRNA (siRNA) or dsDNA that has been internalized from the extracellular space, and this finding further motivated us to determine whether SIDT1 might play a role in the innate immune response to long dsRNA.

Having established that SIDT1 is upregulated by and able to specifically interact with poly(I:C), we sought to determine whether SIDT1 can facilitate the trafficking of extracellular dsRNA from endolysosomes into the cytoplasm. However, previous studies have suggested that human SIDT1 directly promotes internalization of extracellular dsRNAs (specifically, siRNAs) into mammalian cells (14, 15). Given the similarities in sequence and expression of SIDT1 between mice and humans, we wanted to formally test whether mouse SIDT1 is able to promote uptake of extracellular dsRNA. To do so, we measured the amount of poly(I:C) internalization in SIDT1-mCherry⁺ (doxycycline treated) and SIDT1-mCherry⁻ (nondoxycycline treated) DC2.4 cells via flow cytometry. Consistent with our observation that SIDT1 localizes to endolysosomes and not the plasma membrane, we found no difference in the uptake of poly(I:C) between these two cell populations (Fig. 5A). This is consistent with a recent report that found that SIDT1 is unable to transport long (700 bp) dsRNA across the plasma membrane (17).

Given that SIDT1 did not play a role in the initial uptake of poly(I:C), we next assessed whether SIDT1 can facilitate the trafficking of extracellular dsRNA from endolysosomes into the cytoplasm by comparing the subcellular localization of poly(I:C)-fluorescein in SIDT1-mCherry⁺ and SIDT1-mCherry⁻ DC2.4 cells via confocal microscopy. Notably, we observed that SIDT1-mCherry⁺ cells displayed less punctate intracellular distribution of poly(I:C) than SIDT1-mCherry⁻ cells (Fig. 5B, 5C). Because both cell populations showed identical uptake of poly(I:C) (Fig. 5A), these data are, therefore, consistent with the hypothesis that SIDT1 promotes the escape of endocytosed dsRNA into the cytoplasm.

To investigate the role of SIDT1 in vivo, we generated Sidt1^−/− mice in which a SIDT1 gene–targeting construct had successfully disrupted exons 1–3 of the Sidt1 gene locus (Fig. 6A), leading to loss of Sidt1 expression (Fig. 6B). Sidt1^−/− mice were born at the expected Mendelian ratios and displayed normal viability and fertility. Unlike Sidt2^−/− mice (27), there was no observable difference between Sidt1^−/− mice and WT littermates in gross body weight, gross organ morphology, nor organ weights (data not shown). Although Sidt1 is highly expressed in lymphocytes, we did not observe any histological changes in lymphoid organs in the absence of SIDT1 (data not shown), nor did we observe any differences in the development of T lymphocytes (Supplemental Fig. 2) or B lymphocytes (Supplemental Fig. 3). Taken together, our data suggest that although SIDT1 is highly expressed in T and B cells, it does not appear to play an important role in the development or differentiation of these cells in naive mice.

Because SIDT1 was able to interact with poly(I:C) and its overexpression enhanced endosomal escape of poly(I:C), we decided to first assess whether loss of SIDT1 affected the response to poly(I:C) stimulation in vitro. Given that Sidt1 expression is relatively low in naive BMDMs, we primed cells with IFN-β prior to the addition of poly(I:C) to induce Sidt1 expression in WT but not Sidt1^−/− BMDMs. Similar to our previous results with Sidt2^−/− BMDM (13), we observed a reduction in phosphorylation of IRF3 protein in Sidt1^−/− BMDMs following poly(I:C) stimulation compared with WT cells (Fig. 6C), consistent with reduced stimulation of dsRNA-sensing pathways within the cytoplasm.

Next, to determine whether SIDT1 is also important for the response to poly(I:C) in vivo, we injected Sidt1^−/− mice with poly(I:C) and measured serum IFN-β levels 3 and 6 h later via ELISA. Serum IFN-β levels peaked at 3 h following poly(I:C) injection and were undetectable by 6 h, and Sidt1^−/− mice did not have an impaired type I IFN response at either time point (Fig. 6D).

To investigate whether loss of SIDT1 leads to impaired innate immunity during viral infection, we infected Sidt1^−/− and WT mice with 6.3 × 10⁷ PFU of HSV-1 via i.p. injection and assessed serum cytokine levels at 16 h p.i., a time point previously shown to be indicative of dsDNA-independent, MAVS-dependent type I IFN response in this model (28). Notably, compared with WT controls, Sidt1^−/− mice produced significantly less IFN-β and MCP-1 (CCL2) (Fig. 6E, 6F) as well as reduced amount of RANTES (Fig. 6G), although differences in the latter were not statistically significant. Unlike Sidt2^−/− mice (13), however, Sidt1^−/− mice produced similar levels of the proinflammatory serum cytokines IL-6 and IL-12 p40 compared with WT animals (Fig. 6H, 6I) and also showed an identical clinical course, with the vast majority of mice succumbing to lethal infection after 6 d (Fig. 6J). Similar results were observed when repeating this experiment using a lower infection dose of HSV-1 (1 × 10⁷ PFU) (Supplemental Fig. 4A–F). Taken together, these results suggest that, although SIDT1 may contribute to the production of type I IFN and different chemokines during HSV-1 infection, it is not essential for the survival of the animal.

To assess whether this was also true for other viral infections, Sidt1^−/− mice were infected with EMCV, in which survival post infection relies specifically on MDA-5–dependent type I IFN signaling (6, 9). Again, unlike Sidt2^−/− mice, which displayed significantly increased mortality following EMCV infection (13), Sidt1^−/− animals showed no difference in survival compared with controls (Supplemental Fig. 4G).

Our data implicate SIDT1 in the transport of long dsRNAs, similar to SIDT2 and SID-1 (9–12). This conclusion is consistent with previous biochemical studies showing that SIDT1 can bind long dsRNA (16). Superficially, it is also consistent with earlier studies reporting that SIDT1 can transport siRNAs across the plasma membrane (15). However, our results differ from these previous studies because we did not observe any SIDT1-mediated dsRNA transport across the plasma membrane. Instead, SIDT1-dependent dsRNA transport appears to occur across the endolysosomal membrane, consistent with our observation that SIDT1 is expressed in endolysosomes. Whether the failure to detect SIDT1-dependent transport of dsRNA across the plasma membrane is due to differences in the length of the dsRNA used in each study is unclear. Certainly, the length of poly(I:C), which in our experiments was several kB on average, and that of the siRNAs used in previous studies (21–22-nt long) are very different. Thus, it is conceivable that SIDT1 might preferentially traffic short dsRNA across the plasma membrane, and this might explain the discrepancy in results. However, it should be noted that previous studies indicate that, although SIDT1 has a high affinity for binding long dsRNA, it is unable to bind 20–50-bp dsRNA (16), which would argue against this being the case and, consistent with this, we failed to observe any interaction between SIDT1 and siRNAs.

Another discrepancy between our results and that of others is the subcellular localization of SIDT1. Three previous studies have reported on the subcellular localization of human SIDT1 (14, 17, 29). Although all three conclude that SIDT1 localizes to the plasma membrane, one study also found evidence of intracellular localization that overlapped with a marker of the endoplasmic reticulum (17). In our case, we studied mouse SIDT1, and it clearly showed strong intracellular expression in the absence of any detectable plasma membrane localization. Whether this difference is due to interspecies or cell type differences, mislocalization of the protein as a consequence of overexpression, or the use of different SIDT1 splice isoforms (of which six are annotated in humans and three in mice) is unclear, but further studies to explore this issue are warranted.

SIDT1 and SIDT2 share considerable homology at the sequence level, with 59% of amino acids identical and 74% similar (based on physiochemical properties) (data not shown). Taken together with our earlier work (13), our results in this study indicate that the two proteins also share many biological properties. For example, loss of both SIDT1 and SIDT2 in vivo is associated with a significant decrease in type I IFN production following HSV-1 infection, consistent with a shared role in the response to viral dsRNA. Moreover, both SIDT1 and SIDT2 localize to late endosomes and lysosomes, interact with dsRNA but not dsDNA, facilitate the endosomal escape of long dsRNA internalized by the cell, and facilitate activation of IRF3 by poly(I:C). Taken together, these data suggest that SIDT1-mediated transport facilitates cytoplasmic recognition of dsRNA and subsequent antiviral immunity, although it should be noted that, unlike our previous study with SIDT2 (13), in this study we did not attempt to explicitly assess whether loss of SIDT1 alters cytoplasmic RLR and/or endosomal TLR3 signaling.

However, we also find many important biological differences between SIDT1 and SIDT2. For instance, the tissue expression of Sidt1 is much more restricted compared with Sidt2 and, potentially because of this, we find that loss of SIDT1 in vivo has limited effects. Specifically, apart from the decreases in type I IFN and chemokine production following HSV-1 infection, Sidt1^−/− mice showed no differences following viral infection nor any overt changes in fertility, viability, body and organ weight, tissue morphology, or T and B cell development. Of course, more targeted phenotypic assessment of Sidt1^−/− mice in the future might reveal differences that we have not considered in this study; but for now, given that SIDT2 appears to share dsRNA transporter activity with SIDT1 and is expressed more broadly, it seems likely that the presence of SIDT2 in Sidt1^−/− mice is sufficient and can compensate for loss of Sidt1.

What might be the physiological role of SIDT1? Expression of Sidt1 is strongly upregulated following poly(I:C) and IFN-β treatment and during viral infection (25, 26), suggesting that once an infection is established the induction of SIDT1 might support SIDT2 to enhance endosomal escape of internalized dsRNA into the cytoplasm for downstream RLR activation. Moreover, our observation that SIDT1 interacts with SIDT2 suggests that the two proteins might synergize in this process. How they might do so is unclear, but it is interesting to note that both SID-1 and SIDT1 have been previously shown to form multimeric complexes (30, 31). More specifically, dominant negative studies with SID-1 indicate that it functions as a multimeric complex in C. elegans (31), whereas structural analyses of SIDT1 have previously shown that its ectodomain is likely to form a tetrameric structure with a central pore through which dsRNA might traffic (30). Whether the latter might incorporate SIDT2 and affect substrate specificity or alter transport activity is an open question, but our FRET–FLIM data indicate that SIDT1 and SIDT2 interact, raising this possibility. However, because our FRET–FLIM studies relied upon overexpression of SIDT1 and SIDT2 to supraphysiological levels, additional studies will be required to confirm the specificity of this interaction under endogenous conditions.

Recently, SIDT1 was reported to interact with cholesterol and mediate cholesterol uptake in mammalian cells (17). Interestingly, these authors also observed that mixing dsRNA with cholesterol resulted in increased transport of dsRNA by SIDT1, and that SIDT2 is similarly capable of interacting with cholesterol (17). Although we have not investigated what role, if any, cholesterol plays in SIDT1- or SIDT2-mediated endosomal escape of dsRNA, it is tempting to speculate that cholesterol may be involved in some way. Indeed, Sidt2^−/− mice were recently reported to have increased serum triglycerides and impaired lipid homeostasis in the liver, consistent with SIDT2 also being involved in lipid metabolism (32). Looking ahead, future investigations into lipid metabolism in Sidt1^−/− mice would clearly be of interest.

In conclusion, our study shows that SIDT1 shares many properties of SIDT2, including the ability to promote the endosomal escape of long dsRNA and to influence the production of type I IFN following HSV-1 infection, but it is not essential for survival following HSV-1 or EMCV infection. Moving forward, the novel Sidt1^−/− mouse model described in this study should, in conjunction with existing Sidt2^−/− mice, enable future studies to improve understanding of the complementarity and redundancy of SIDT1 and SIDT2.

We thank D. Huang and P. Gleeson for gifts of plasmids, as well as L. Mackiewicz, R. Crawley, L. Scott, C. Hay, and L. Inglis for technical assistance. We also thank all members of the Masters and Wicks laboratories for helpful discussions and comments.

The authors have no financial conflicts of interest.