The molecular dissection of TRIM25‘s RNA-binding mechanism provides key insights into its antiviral activity

doi:10.21203/rs.3.rs-3692619/v1

Download PDF

Article

The molecular dissection of TRIM25‘s RNA-binding mechanism provides key insights into its antiviral activity

https://doi.org/10.21203/rs.3.rs-3692619/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Oct, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

TRIM25 is an RNA-binding ubiquitin E3 ligase with central but poorly understood roles in the innate immune response to RNA viruses. The link between TRIM25’s RNA binding and its role in innate immunity has not been established. Thus, we utilized a multitude of biophysical techniques to identify key RNA-binding residues of TRIM25 and developed an RNA-binding deficient mutant (TRIM25-m9). Using iCLIP2 in virus-infected and uninfected cells, we identified TRIM25’s RNA sequence and structure specificity and showed that its binding to specific viral RNA regions is critical for its antiviral activity.

Biological sciences/Biochemistry/RNA

Biological sciences/Biophysics/Molecular biophysics/Supramolecular assembly

Biological sciences/Structural biology/NMR spectroscopy/Solution-state NMR

Biological sciences/Molecular biology

Health sciences/Diseases

The innate immune system is the body's first line of defence against viral infection. When viruses are detected by the host cell, signalling pathways are activated to induce type I interferon response and the production of inflammatory cytokines¹. Post-translational modifications regulate these signalling pathways, allowing a dynamic and specific host response to infection². Ubiquitination plays a particularly important role and multiple E3 ubiquitin ligases have been identified to be involved in the regulation of immunity^3–6. The Tripartite Motif (TRIM) family of E3 ligases is characterized by their domain architecture, consisting of an N-terminal RING domain, which is responsible for the E3-ligase activity, followed by one or two B-box domains, a coiled-coil (CC) dimerisation domain and a C-terminal region containing non-catalytic domains that recognise E3 ligase substrates and other targets, which are variable across TRIM proteins with TRIM25 featuring a PRY/SPRY domain (Fig. 1a) ^7,89. TRIM25 has been identified as a key player in triggering the innate immune response to RNA viruses. It has been shown that TRIM25 plays a role in the RIG-I signalling pathway, where it is proposed to ubiquitinate the RIG-I caspase activation and recruitment domains (CARDs), which are exposed upon recognition of viral 5′-triphosphate-blunt-end double-stranded

RNA.^10,11 However, there are other E3 ubiquitin ligases that can polyubiquitinate and activate RIG-I, such as RING finger protein leading to RIG-I activation (Riplet)¹², mex-3 RNA binding family member C (MEX3C)¹³ and TRIM4¹⁴. In addition, there are studies suggesting that only RIPLET and not TRIM25 is sufficient to ubiquitinate and activate RIG-I^15–17. Despite these inconsistencies, TRIM25’s involvement in antiviral activity is undisputed. Apart from the RIG-I pathway TRIM25 is required for activating the antiviral function of zinc-finger antiviral protein (ZAP)^18,19, which inhibits the replication of a wide-range of viruses by binding to viral mRNA, repressing its translation and promoting its degradation^19–23.

TRIM25 was identified as an RNA-binding protein (RBP) by several RNA interactome capture studies, which reported peptides that crosslink to RNA^24–26. Several studies could identify different regions of the protein, which harboured RNA-binding residues (located in the CC, PRY/SPRY domains, and the linker in between^{17,25,27,28,29}). TRIM25 has also been shown to interact with a wide variety of RNAs, including 3’-UTRs and exons of mRNAs^25,28, lincRNAs²⁸, miRNAs²⁸, viral RNAs and their corresponding ribonucleoproteins (RBNPs)^30,31. Although no specific consensus motif for RNA binding was identified, G- and C-rich sequences were found to be overrepresented in the CLIP-seq data²⁸. In vitro assays showed no clear preference for single- or double-stranded RNA²⁹. The fact that different domains and regions of TRIM25 seem to be involved in RNA binding makes it difficult to identify specific sequences or conformational RNA motifs that TRIM25 prefers, and points towards a structurally intricate RNA-binding mechanism. Moreover, it was shown that TRIM25 required RNA binding for its E3-ligase function^28,29 and overall antiviral activity²⁹. However, the mutants employed in those studies still retained RNA-binding activity or were partially unfolded, as we demonstrate below. Thus, it remains poorly understood how TRIM25 binds RNA and how this affects its antiviral activity.

Here we combine biochemical, biophysical and cell biology approaches to characterise the TRIM25-RNA interaction and its role during antiviral activity. We have identified key residues in the PRY/SPRY and CC domains responsible for RNA binding. Furthermore, we developed a mutant, which retains the protein structure including an intact RING domain, responsible for E3-ligase activity, but completely lacks RNA-binding ability. Using iCLIP2³² with size matched inputs in virus-infected and uninfected cells, we could identify the host and viral RNA motifs that TRIM25 binds to and determined TRIM25’s RNA structure- and sequence specificity. In addition, we could confirm that RNA binding of TRIM25 is required for its antiviral activity as well as for its subcellular redistribution to the viral replication organelles (ROs) where it co-localises with viral RNA.

Two distinct sites of the TRIM25 PRY/SPRY domain confer binding to single- and double-stranded RNA

Previous work showed that the PRY/SPRY domain is responsible for TRIM25 RNA binding^25,28,29,33. Choudhury et al. identified a region in the PRY/SPRY domain encompassing residues 470–508 that bind to RNA by screening a series of deletion constructs²⁸. However, deletion of this region unfolds the PRY/SPRY domain and therefore it remained unclear whether impairment of RNA binding is due to removal of RNA-binding residues or unfolding of the domain (Extended Data Fig. 1a). To refine the RNA-binding interface of this domain at residue-level resolution, we performed NMR titrations with the reported RNA target of TRIM25, pre-let-7a-1@2³⁴ (hereafter referred to as pre-let-7, Extended Data Fig. 1b). Upon addition of pre-let-7, chemical shift perturbations (CSPs) are observed, which confirms binding (Fig. 1b). Specifically, we observed strong CSPs clustered around two regions of the PRY/SPRY domain. The first of the two regions (region 1, aa. 456–513) is located at the C-terminus of the PRY motif with the most affected residues H505 and K508 in the flexible loop connecting β-strands 3 and 4 (Fig. 1c, d). This region is located close to the previously reported interaction site between the PRY/SPRY and CC domains³⁵. The second region (region 2, amino acids 567–605) with the most affected residue K602 consists of β-strands 10 and 11, close to the N-terminal helix α1 (Fig. 1c, d). At this stage, we could not assess, whether the CSPs in this region were due to a direct RNA interaction or due to an allosteric effect.

The predicted structure of pre-let-7 consists of a stem-loop (Extended Data Fig. 1b). The formation of double-stranded regions indicative of a stem structure was confirmed by imino peaks observed in ¹H/¹H-NOESY experiments (Extended Data Fig. 1c). To obtain further insights into the RNA-binding mechanism of the PRY/SPRY, we designed shorter RNA constructs consisting of only the loop (pre-let-7-loop, single-stranded RNA) or stem of pre-let-7 (pre-let-7-stem) (Extended Data Fig. 1b). To ensure that the short stem was double-stranded, we fused the strands through a three-base long linker and confirmed its stem formation in solution by ¹H/¹H-NOESY NMR (Extended Data Fig. 1c). Pre-let-7-loop induced CSPs only for residues located in binding site 1, whereas titration of pre-let-7 stem affected only binding site 2 (Fig. 1d). To confirm that the binding sites are generally involved in RNA binding and not specific for pre-let-7 RNA, the NMR titration experiments were repeated with other known RNA binders of TRIM25 (such as Lnczc3h7a³⁶, DENV³⁰ and an arbitrary 28-mer²⁹, see Extended Data Fig. 1b), with similar outcome (Extended Data Fig. 1d). These results suggest that the PRY/SPRY domain interacts with RNA by two distinct binding sites with different structural specificities. The first binding site interacts with single-stranded RNA and the second with double-stranded RNA. The structural similarity between the RNAs reported to bind to TRIM25 suggests that TRIM25 specifies for stem-loop structures^29,30,36.

The PRY/SPRY domain interacts with different stem-loop RNAs with a dissociation constant (K_D) in the low micromolar range (1–4 µM) as assessed by isothermal titration calorimetry (ITC, see Fig. 1e and Supplementary Table 1). This is in the same range as canonical RNA-binding domains such as RNA recognition motifs (RRMs)³⁷. Our ITC data indicate that the PRY/SPRY domain preferentially binds to loops rich in A and G, as we could see binding to pre-let-7-loop but no binding to the Lnczc3h7-loop, which has three U and is shorter (Supplementary Table 1).

To further characterise the RNA-binding site of the PRY/SPRY, we mutated the three residues exhibiting the largest CSPs in NMR titrations. We created a triple mutant, termed PRY/SPRY-m3, in which we mutated residues H505 and K508 from binding site 1 and residue K602 from binding site 2 to glutamic acids. Importantly, the ¹H,¹⁵N-HSQC spectrum of the PRY/SPRY-m3 confirmed that the mutated PRY/SPRY domain is properly folded (Extended Data Fig. 1f). ITC showed that the PRY/SPRY-m3 lost its ability to interact with RNA, further confirming the importance of these residues for RNA binding (Fig. 1f and Supplementary Table 1). The co-occurrence of these RNA-binding residues seems to be a unique feature for TRIM25 amongst human TRIM family members from group IV⁹ (i.e. TRIM proteins carrying a PRY/SPRY domain) revealed by multiple sequence alignments. However, the RNA-binding residues are highly conserved in TRIM25 across different species (Extended Data Fig. 1g, h and Supplementary Table 2).

CC binds to RNA and the PRY/SPRY domain using adjacent interfaces

Having characterised RNA-binding activity of the PRY/SPRY domain, we followed up on previous work that has reported the CC domain of TRIM25 as RNA binder^25,28,38. The CC is not suitable for NMR studies due to its size and extended conformation, which causes slow molecular tumbling and NMR resonance line broadening beyond detection (Extended Data Fig. 2a). We therefore used ITC to test the putative RNA-binding capacity of the CC domain. We found that the CC interacts with pre-let-7 with low micromolar affinity (3.2 ± 0.7 µM) (Fig. 2a), in the same range as the PRY/SPRY domain and classical RNA-binding domains (e.g. RRMs). We performed ITC measurements with the other RNAs used in this study (Extended Data Fig. 1b and 2b). In contrast to the PRY/SPRY domain, the CC domain binds pre-let-7 or lnczc3h7a with equal affinities. However, as for the PRY/SPRY domain, pre-let-7-loop and lnczc7a-loop bind 10-fold weaker (Supplementary Table 1, and Extended Data Fig. 2b). Thus, we find that the CC domain is more promiscuous than the PRY/SPRY domain in its sequence preference.

Since we were unable to map the RNA-binding interface on the CC by NMR, we relied on indirect information to design CC RNA-binding deficient mutants. Based on the CC-PRY/SPRY structure³⁵, we deduced the RNA-binding residues from amino acids that are surface-exposed, close to the CC:PRY/SPRY interface and are commonly found to interact with RNA, such as lysines, arginines and aromatic residues. We found two candidates for mutational analysis (K283 and K285, Fig. 2b) on the basic surface close to the PRY/SPRY binding site 1. Mutation of these two residues into alanines (CC-m2 had a minor effect on affinity towards pre-let-7, but reduced the binding enthalpy 10-fold (Fig. 2c and Supplementary Table 1),

while retaining the ability to dimerize and to fold into a coiled coil (Extended Data Fig. 2c, d). This suggested that while these two amino acids are involved in energetically favourable interactions with RNA, other residues must contribute.

Using CLIR-MS³⁹ (Fig. 2d) with an equimolar mixture of unlabelled and uniformly ¹³C,¹⁵N-labelled 85-nucleotide-long RNA incubated in HeLa nuclear extract, we detected UV-crosslinks between nucleotides and amino acids of the peptide GISTKPVYIPEVELNHK of TRIM25 (Fig. 2e). The MS/MS spectrum allowed the identification of the crosslinked residue, which was either tyrosine 323 (Y323) or the neighbouring isoleucine 324 (I324). The peptide crosslinks to one uridine (mass shift 324) and could be identified by “light” searches (based on the expected mass of the RNA-peptide adduct) and by “light-heavy” searches (considering the characteristic peak doublet resulting from the isotope labelling), ensuring the specificity for the differentially isotope labelled RNA (Fig. 2e). According to a crystal structure (PDB: 6FLN³⁵) this region is devoid of secondary structure (Fig. 2b). Comparative analysis of TRIM25 sequences from different species (see Supplementary Table 2) revealed that the specific region responsible for RNA binding, as identified by CLIR-MS, is well conserved (Extended Data Fig. 2e). However, this region is not conserved in other group IV TRIM protein CC domains⁹ (Extended Data Fig. 2f). It is noteworthy that this region overlaps with the surface that interacts with the PRY/SPRY domain identified by Koliopoulos et al.³⁵ (Fig. 2b). To confirm that this region interacts with RNA, we mutated these four residues (K320, Y323, H331 and K332) to glutamic acids (CC-m4) and performed ITC experiments. These show that the CC-m4 has lost its RNA-binding activity (Fig. 2f), while still retaining its ability to fold into a coiled coil and dimerize (Extended Data Fig. 2c and d). We have thus identified the residues of the CC domain that are critical for RNA binding.

PRY/SPRY and CC domains bind to RNA cooperatively

As both domains, the CC and PRY/SPRY interact with RNA, we wanted to assess, whether they do so cooperatively. To test this, we performed ITC measurements of the CC-PRY/SPRY domain. The affinity of the CC-PRY/SPRY construct to pre-let-7 RNA is more than 30-50-fold higher than that of the isolated domains (Fig. 3a), which is a hallmark of chelating cooperativity, where multiple individually weaker binary interactions cooperate to form a stable multimeric complex^40,41. This cooperative effect can also be observed for binding to Lnczc3h7a RNA (Supplementary Table 1 and Extended Data Fig. 3a).

We have previously shown that the interaction between the CC and PRY/SPRY domains in solution is rather transient, despite being present in the crystal structure of the CC-PRY/SPRY tandem domain construct³⁵. Considering that the CC-PRY/SPRY interaction plays an important role in TRIM25's E3 ligase activity,³⁵ we investigated whether RNA binding modifies the properties of this interaction in solution. To achieve this, we monitored the changes in ¹H,¹⁵N-HSQC NMR spectra of CC-PRY/SPRY upon addition of pre-let-7. In the absence of RNA, the spectrum is dominated by sharp, high-intensity peaks in the centre corresponding to the disordered L2 linker region (Fig. 3b). Other well-dispersed peaks overlap with the spectrum of the isolated PRY/SPRY domain but with reduced intensity (Extended Data Fig. 3b). This is consistent with the previously described weak and transient interaction between CC and PRY/SPRY in solution³⁵, resulting in an independent tumbling of the CC and PRY/SPRY connected by a flexible linker. Upon addition of pre-let-7 RNA, the PRY/SPRY peaks disappear completely, while the L2 linker peaks remain visible (Fig. 3b). This suggests that the CC:PRY/SPRY interaction is stabilised by RNA (Fig. 3c), which in turn results in the structured regions of the protein being rigidly connected. To further confirm that this conformational change was due to RNA binding, we used solution small-angle X-ray scattering (SAXS). A comparison of the SAXS curves of free TRIM25 CC-PRY/SPRY and bound to pre-let-7 shows a significant decrease in the radius of gyration upon interaction with RNA (R_g = 6.83 ± 0.05 nm for the free protein compared to R_g = 5.7 ± 0.02 nm for the complex), which agrees with a more compact conformation than the free protein (Fig. 3d and Supplementary Table 3) in which the PRY/SPRY domain is mostly detached from the CC domain in solution³⁵. Similar effects were observed for the TRIM25-lnczc3h7a stem-loop complex (R_g = 5.8 ± 0.2 nm) at concentrations well above the K_D, confirming that this RNA stem-loop-induced conformation change is the general RNA-binding mechanism of TRIM25 (Extended Data Fig. S3c and Supplementary Table 3). The effect also occurs in SEC-SAXS with an excess of pre-let-7, clearly demonstrating that it is not an artefact caused by the

scattering contribution of the free RNA or aggregation (Extended Data Fig. S3d). We conclude that RNA is bound cooperatively by and enhances the interaction between both domains.

With the aim of generating a TRIM25 mutant with fully deficient RNA-binding activity, we combined the different mutations described in the previous sections (PRY/SPRY-m3, CC-m2 and CC-m4. Figure 3e), resulting in a mutant referred to as CC-PRY/SPRY-m9. ITC. ITC showed that this combined mutant does not bind RNA (Fig. 3f). Thus, with a full-length TRIM25-m9 RNA-binding deficient mutant, we established a powerful tool to test RNA binding by TRIM25 in cells and to understand the importance of RNA binding for TRIM25’s antiviral activity.

iCLIP2 data reveals cellular structure- and sequence specificity of TRIM25

To investigate the RNA-binding properties of TRIM25 within cells we used HEK293-TRIM25 knock-out (TRIM25 KO) cell lines,¹⁷ which contain a flippase recognition target (FRT) site to allow stable integration of a gene of choice into the genome using the Flp-In recombinase¹⁷. TRIM25-WT, TRIM25-m3 (lacking the RNA-binding properties of the PRY/SPRY domain) and TRIM25-m9 (no RNA binding by full-length TRIM25) were integrated into the genome, resulting in stable expression at levels close to endogenous TRIM25 in HEK293 WT cells (Extended Data Fig. 4a). To understand the relationship between TRIM25’s RNA binding and its antiviral activity, we chose the alphavirus Sindbis (SINV) as model system. SINV is a pathogenic virus transmitted from mosquitoes to vertebrates and is a broadly used alphavirus model system. It was chosen for several reasons: firstly, there is interactome data showing that TRIM25 is upregulated upon infection with SINV and that overexpression of TRIM25 potently reduces SINV infection⁴². In addition, there is no viral protein-dependent suppression of TRIM25 as it occurs with NS1 from influenza A virus³⁵.

To better understand TRIM25 specificity in cellulo, we applied iCLIP2⁴³ to SINV-infected and uninfected HA-FLAG-TRIM25 HEK293 cells as described in Garcia-Moreno et al. 2023⁴⁴. We added a size matched input (SMI) as previously described⁴⁵ to correct for background signal. Principal component analysis confirmed the high quality of the iCLIP2 data, showing separated clustering of samples based on sample type (SMI vs IP) and conditions (infected vs mock) (Extended Data Fig. 4b). Strikingly, iCLIP2 revealed that TRIM25 targets significantly more genes and more binding sites within targeted genes after SINV infection (Fig. 4a). In agreement with our biophysical data, TRIM25-m3 showed strong reduction in binding sites and targets, when compared to TRIM25 wild type (WT) protein. Almost no binding sites were observed for TRIM25-m9, suggesting that it also lacks RNA binding in cellulo. TRIM25 binding sites substantially differ in SINV infected and mock cells, with only ~ 22% of overlapping targets (Fig. 4b), suggesting that TRIM25 RNA-binding activity is regulated by virus infection. TRIM25-m3 binding sites barely overlapped with WT TRIM25, with only 130 (out of 990) being detected by both proteins. This suggests that while TRIM25-m3 is still able to interact with RNA to some extent, its specificity and affinity are substantially affected. The very few binding sites detected in TRIM25-m9 samples did not overlap with neither TRIM25-m3 nor TRIM25 WT (Fig. 4b), suggesting a full or near complete abolishment of its RNA-binding activity.

The iCLIP2 results show that TRIM25 interacts predominantly with mRNAs in both SINV-infected and uninfected (mock) conditions (Extended Data Fig. 4c). In uninfected cells, TRIM25 prefers 3’UTRs, followed by coding sequences (CDS). A similar preference has been observed in a previous study²⁸. In SINV-infected cells, this binding is reversed, leading to higher binding in CDS than in 3’UTRs. Interestingly, TRIM25 binding sites in SINV-infected cells peaks at the start and stop codon of CDS, which are key signatures for translational control. One potential explanation is the increased availability of these sequences in infected cells due to low ribosome occupancy consequently to the powerful translation initiation shut off occurring after phosphorylation of eIF2α⁴⁶. Interestingly, the TRIM25-m3 binding profile loses the peaks at the start and stop codons and 3’UTR, which indicates that the PRY/SPRY is necessary to define RNA-binding specificity. The very few reads detected in TRIM25-m9 samples were homogeneously distributed across the mRNA sequence, indicating that they are likely derived from experimental noise rather from real TRIM25 binding sites (Fig. 4c).

Since many of the RNAs reported for TRIM25 consist of multiple stem-loops of similar size, we tested whether TRIM25 recognizes specific structures by analysing the percentage of predicted base pairing around the crosslink site (Fig. 4d). Interestingly, we found a low incidence of base pairing at the crosslink site with base pairing frequently flanking it (considering 25 nucleotides on each side). This implies that TRIM25 recognises stem-loop structures, with UV crosslinking happening more frequently at the loop region. To rule out that this is an artefact of iCLIP2, we extended the analysis to other RBPs, observing that base pairing distributions compatible with stem-loops are not detected in other proteins (Extended Data Fig. 4d). To test whether TRIM25 recognises specific sequences within the stem-loop structure, we searched for enriched sequence motifs. We found two prominent classes, an AGAA motif and a UGG motif (Fig. 4e). Detailed analysis of the occurrence of these motifs revealed that the AGAA motif typically precedes the crosslinking site, and the UGG motif appears after it (Fig. 4f). Of note, this consensus is very similar to the model RNAs previously identified (pre-let-7 and DENV-SL) used in our biophysical investigation of TRIM25’s RNA-binding mechanism described above.

In addition to binding sites on cellular RNAs, we tested if TRIM25 also binds to the SINV RNA. iCLIP2 revealed several high-probability binding sites across the viral genome. SINV produces two positive sense RNAs, the genomic (g)RNA that encodes the non-structural proteins and is packaged into virions, and the subgenomic (sg)RNA that encodes for the structural proteins. The sgRNA overlaps with the last third of the gRNA and is expressed to a higher level. This explains why the peaks mapping to the sgRNA region are higher in magnitude than those mapping to the first two thirds of the gRNA. Despite this, several binding sites were significantly enriched over the background signal at gRNA and sgRNA regions (Fig. 4g), suggesting that TRIM25 engages with both viral transcripts. To determine whether these are specific binding sites, we searched for structure and sequence features that are enriched in the cellular binding sites of TRIM25. We used experimentally determined structural information previously generated by SHAPE reactivity data for the SINV genome.⁴⁷ SHAPE reactivity is closely related to nucleotide flexibility, with unconstrained nucleotides showing higher SHAPE reactivity than nucleotides involved in base pairing, stacking or other interactions. When we aligned the binding sites with the SHAPE reactivity data, we observed clear overlap of the crosslink sites with accessible nucleotides that are surrounded by low SHAPE reactivity regions compatible with base paring (Fig. 4g). Such architecture is reminiscent of the stem-loops used in our in vitro experiments and detected in TRIM25 binding sites on cellular mRNAs. Indeed, RNAfold⁴⁸, constrained by the SHAPE data, predicts hairpin loops in most of the binding sites (Fig. 4h). Interestingly, loop regions are rich in A/G bases, which are reminiscent of the TRIM25 consensus motifs identified above from our iCLIP2 data (Fig. 4h).

RNA binding is critical for TRIM25’s antiviral activity and relocation to viral replication organelles

TRIM25 is known to inhibit viral replication, as seen for Dengue, Influenza A and Rabies viruses^29,49. However, whether its RNA-binding activity is important for this antiviral role remains poorly understood. To test this, we infected SINV into TRIM25 KO cells rescued with either TRIM25 WT, an E3-ligase dead mutant (referred to as E3-dead, which contains a double mutation in the RING domain that abolishes the E3 ligase activity¹⁸), m3, or m9. Rescue with WT TRIM25 strongly reduced capsid levels when compared to the E3-dead mutant, suggesting that E3 ubiquitin ligase is important for the antiviral role (Fig. 5a). Importantly, rescue with TRIM25-m3 and particularly with TRIM25-m9 increased capsid abundance to similar levels as the E3-dead mutant. This demonstrates that RNA binding of TRIM25 is as important for its antiviral activity as the E3 ligase function of the protein. Noteworthy, the KO cell line rescued with WT TRIM25 restores the antiviral function to the levels of the parental cell line (Extended Data Fig. 5a).

To validate these results, we analysed the effect of our RNA-binding mutants on viral fitness using chimeric viruses expressing mScarlet (SINV_{NSP3−mScarlet}) or mCherry (SINV_mCherry) from the genomic or a duplicated subgenomic RNA, respectively. We observed that the RNA-binding deficient mutants increased viral fitness to a similar extent as the E3-dead mutant, with the effect being slightly stronger for TRIM25-m9 than for TRIM25-m3 (Fig. 5b), reinforcing the idea that the RNA binding is critical for TRIM25 antiviral activity. This experiment also confirms that RNA binding is as important as E3 ligase function for antiviral activity. Our chimeric viruses allow to test the expression of the gRNA and sgRNA (Extended Data Fig. 5b), and our data showed that TRIM25 suppresses the expression of both viral RNAs. This phenotype can be explained by the presence of TRIM25 binding sites in both the gRNA and sgRNA region. Again, we observed that the KO cell line rescued with WT TRIM25 decreases the viral fitness with similar efficiency to the parental cell line (Extended Data Fig. 5c).

SINV replicates in invaginations in endosome-derived membranes, referred to as replication organelles (Ros), where the replicating viral RNA is isolated from the host cytosol^49–51. TRIM25 is cytosolic in uninfected cells, but it accumulates in ROs upon SINV infection, together with viral RNA and capsid protein (Garcia-Moreno et al.⁴² and Fig. 5c). To determine whether the ability of TRIM25 to interact with SINV RNA promotes its redistribution to ROs we used immunofluorescence and RNA single molecule in situ hybridisation (smFISH). TRIM25-WT and TRIM25-m9 showed diffuse localization in the cytoplasm of uninfected cells. TRIM25-WT accumulated into cytoplasmic foci that co-localise with the viral RNA in SINV-infected cells, which are consistent with ROs (Fig. 5c). However, TRIM25-m9 remained diffuse after infection and did not co-localise with these foci (Fig. 5c). These results indicate that the RNA-binding activity is important for its localization to ROs. We hypothesise that interaction with viral RNA can both recruit TRIM25 to the ROs and potentially regulate its E3 ubiquitin ligase activity as previously proposed¹⁷.

Several attempts have been made to describe the RNA-binding mechanism of TRIM25^28,29, however a complete description of how TRIM25 recognises the RNA and its specific motif is still lacking. We have been able to identify all the RNA binding regions of TRIM25 at residue resolution, revealing the cooperative behaviour between the PRY/SPRY and CC domains, and describe a consensus RNA motif by combining a variety of biophysical methods (NMR, ITC, SAXS, SEC-MALS) and iCLIP2. Our results show that TRIM25 achieves very tight binding of RNA by using multiple sites on the CC and PRY/SPRY domains, with each binding site contributing cooperatively to affinity and RNA specificity. Not only were we able to precisely describe the residues involved in RNA binding, but we also discover that the PRY/SPRY domain has a dual binding mode with one site having single-stranded RNA specificity and the other site preferring double-stranded RNA. The proximity of these two sites suggests structural specificity, and indeed we found that the binding of the CC-PRY/SPRY domains to several stem-loops is within nanomolar affinity. This extra layer of complexity may explain the failure of previous studies at identifying a distinct target sequence motif for TRIM25. Here, we show that TRIM25 is indeed a bona fide and highly specific RBP.

The molecular dissection of TRIM25’s RNA-binding activity allowed us to generate a mutant that does not bind RNA neither in vitro nor in cells, without affecting its fold or dimerization property. With this tool in hand, we managed to elucidate the RNA-binding specificity of TRIM25 inside the cell.

Previous iCLIP data on TRIM25^17,28 showed that TRIM25 prefers G- and C-rich sequences, but no strong consensus motif was identified. Our iCLIP2 data allowed us to describe a structural preference of TRIM25 for stem-loop structures inside the cell, including both host and viral RNA. This observation can also explain the G- and C-rich sequences observed in the previous dataset, as G-C base pairing is more stable than A-U due to the higher number of hydrogen bonds formed. We also identified a preferred sequence motif within the hairpin loop of these stem-loop structures. Particularly, we found an AGAA motif that precedes the crosslinking site followed by a UGG motif. Interestingly, the sequence motif found is present in previously identified RNAs that bind TRIM25, such as pre-let-7 RNA and DENV-SL. Furthermore, this motif differs from the one present in the Lnczc3h7a RNA, which is consistent with the weaker binding we observed here. Importantly, the iCLIP2 data showed the lack of RNA-binding activity of TRIM25-m9, which is an important confirmation in cells of the mutations inferred by our biophysical data.

An open question is how TRIM25 exerts its antiviral activity and whether RNA binding is involved. There have been some efforts to answer these questions^28,29, but these studies used TRIM25 constructs that retained RNA-binding activity to some extent (Extended Data Fig. 3). Recently, it was shown to bind and destabilize IAV mRNAs¹⁷. Our data provide several lines of evidence supporting that RNA binding is an essential feature for the antiviral activity of TRIM25. We have shown that viral gene expression is strongly enhanced when RNA binding is completely abolished, phenocopying the mutant that lacks the E3 ligase function. In addition, our iCLIP2 data showed that TRIM25 binds to specific regions of viral RNA. Indeed, our microscopy analyses revealed that TRIM25 RNA-binding activity is a prerequisite to drive the localisation of TRIM25 to SINV ROs, which is likely a critical step for triggering its antiviral activity. This activity could be direct or through interaction with ZAP, which is known to localise to ROs upon SINV infection⁵². Understanding whether TRIM25 RNA-binding activity is important for ZAP localisation, or whether interaction of TRIM25 with ZAP is essential for the antiviral effect, as well as exploring how TRIM25 condensates within ROs, will contribute to a better understanding of the antiviral mechanism driven by TRIM25.

Protein expression and purification

TRIM25 CC (aa 189–379), PRY/SPRY (439–630) and CC-PRY/SPRY (189–630) were cloned using restriction free cloning into pETM22 featuring a 3C-protease cleavable hexa-histidine and thioredoxin tag (His-Trx). N-terminally extended PRY/SPRY construct (407–630) was cloned into pETM20 with a TEV-protease cleavable His-Trx tag. TRIM25 189–379 was expressed in BL21(DE3). TRIM25 439–630, 407–630 and 189–630 were co-expressed with chaperones KJE, ClpB and GroELS in E. coli BL21(DE3)⁵³. All proteins were induced at OD600 = 0.6 by 200 µM isopropyl-ß-D-1-thiogalactopyranoside (IPTG) and incubated at 18˚ C for 20h.

All proteins were purified by immobilized Nickel affinity chromatography (GE Histrap FF) in 50 mM Tris, pH 7.5, 300 mM NaCl and 0.2 mM TCEP and eluted with a gradient of imidazole (10–300 mM). In the case of co-expression with chaperones the column was washed with 50 mM Tris, pH 7.5, 350 mM KCl, 5 mM MgSO4 and 1 mM ATP prior to elution. The tag was removed by 3C- or TEV-protease digestion overnight and an additional passage over the HisTrap column (CC and extended PRY/SPRY) or GE HiTrap SP HP cation exchange column (PRY/SPRY, residues 439–630 and CC-PRY/SPRY, residues 189–630). As a final step, aggregates were removed by gel filtration on a GE Superdex S75 (CC, PRY/SPRY, extended PRY/SPRY) or S200 (CC-PRY/SPRY) in 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP. For stable isotope labelling, proteins were expressed in M9 medium using ¹⁵NH₄Cl (0.5 g/l) as sole nitrogen and carbon source.

RNA production

Pre-let-7a-1@2 (5'-GUA UAG UUU AAA AGG AGA UAA CUA UAC − 3'), Lnczc3h7a-SL (5’-UUUUAUCUGAGUUGGAGGUGAAG-3’) and DENV-SL (5’- GCA GGU CGG AUU AAG CCA UAG UAC GGG AAA AAC UAU GCU ACC UG-3’) were in vitro transcribed from DNA oligos using T7 RNA polymerase. Reaction mixtures containing 2 µM forward and reverse primer, 40 mM Tris, pH 8.0, 0.2 mM MgCl2, 10 mM NTPs, 10 mM spermidine, 15 mM DTT, 0.01% Triton X-100, 4 U/ml TIPP, 0.1 mg/ml T7 polymerase were incubated for 5 hours at 37˚ C and extracted by chloroform/phenol treatment. The aqueous phase was further purified by preparative gel electrophoresis via denaturing PAGE or HPLC using a Thermo DNA Pac PA100 22x250mm anion exchange column at 95˚ C. The purified RNA was dialyzed against 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP and refolded before use by heating up at 95˚ C for 5 minutes and snap-cooling on ice. The shorter pre-let-7 loop (5’-UAA AAG GAG AU-3’), stem-fusion (5'-G UAU AGU U C AAC UAU AC-3'), 28-mer duplex RNA (5’-AUG GCU AGC UGG AGC CAC CCG CAG UUC G-3’) and pre-let-7a-1@2 constructs were purchased from IBA or Biomers.

Nuclear magnetic resonance spectroscopy

NMR spectra were acquired on Bruker Avance III spectrometers operating at magnetic field strengths corresponding to proton Larmor frequencies of 600, 700, and 800 MHz, equipped with cryogenic triple resonance probes (600 and 800 MHz) and a room temperature triple resonance probe (700 MHz). Protein observed experiments were performed in 20 mM sodium phosphate, pH 6.5, 150 mM NaCl, 2 mM TCEP, 5% D₂O and 0.02% sodium azide at 293 K. For RNA observed experiments samples were prepared in 20 mM MES, pH 6.5, 75 mM NaCl, 5 mM MgCl₂, 0.02% sodium azide and measured at 278 K.

For titrations 100 µM ¹⁵N-labelled TRIM25-PRY/SPRY (439–630) was titrated with stock solutions of 10–20 mM natural abundance RNA to a molar excess of 2.5-3 (details indicated in the Fig. 1b). At each titration point a ¹H,¹⁵N-HSQC spectrum was acquired. Due to protein precipitation, a different number of scans were used for different titration points (256 scans for PRY/SPRY domain alone, 512 scans for 1:1 and 1:2 and 1024 for 1:3). Figure 1B was generated using the same number of contour and noise thresholds. Spectra were processed using NMRPipe⁵⁴, visualized and peak shifts tracked using SPARKY⁵⁵. Peak shifts for PRY/SPRY 439–630 were assigned based on a previously published assignment.³⁵ CSPs in histograms (Fig. 1d) were determined at a protein:RNA ratio of 1:3.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) data was collected on a Malvern MicroCal PEAQ-ITC at 20˚ C in 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP. Depending on the protein construct, RNA was titrated from the syringe at concentrations ranging from 20-1100 µM into the cell containing protein with concentrations between 2-150 µM while stirring at 750 rpm (see Figs. 1, 2 and 3 and Extended Data Figs. 1 and 3). For experiments with DENV-SL the RNA was kept in the cell at 15 µM and the protein titrated from the syringe at 110 µM. Experiments were typically done in triplicates and analysed using the MicroCal PEAQ-ITC analysis software. Data was fitted by assuming the simplest binding mode necessary to fully explain the data (one or two binding sites).

Sequence alignments

To reveal the conservation of TRIM25-PRYSPRY among TRIM proteins, sequences of the 44 human TRIM proteins belonging to group IV (containing the PRY/SPRY domain) and Riplet (RN135) were obtained from InterPro⁵⁶ using the PRY/SPRY domain as a seed. The alignment of the different PRY/SPRY domains was performed using ClustalW⁵⁷ and the alignment is provided in Extended Data Fig. 1g. Using the 44 sequences of human TRIM proteins containing the PRY/SPRY domain used previously, the structure of the proteins was modelled using Alphafold^58,59 to determine the boundaries of the CC domain for each protein. The alignment of the different CC domains was then performed using ClustalW ⁵⁷ and shown in Extended Data Fig. 2f.

To test sequence conservation of TRIM25 amongst different species, sequences of 18 TRIM25 proteins from different species were aligned using ClustalW⁵⁷ and the full alignment for the PRYSPRY domain is shown in Extended Data Fig. 1h. Sequences were obtained from ENSEMBL⁶⁰ using the human TRIM25 protein as seed (UniProt ID: Q14258). Details of the species are given in Supplementary Data Table 2. In addition, the sequence logo corresponding to the CC specific region responsible for RNA binding was generated by WebLogo (http://weblogo.berkeley.edu/logo.cgi)⁶¹ using the aligned protein sequences corresponding to the CC domain from the same 20 species (Supporting Table 2 and Extended Data Fig. 2e). The height of the stack of letters at each position reflects sequence conservation and reaches a maximum value of 4.32 for proteins. The height of each letter within a stack is proportional to its abundance.

Small-angle X-ray scattering

SAXS data were collected at the beamline P12, operated by EMBL Hamburg at the PETRA III storage ring (DESY, Hamburg, Germany)⁶². Measurements were done at 20°C in 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP in flow cell mode at 1.24 Å. For each sample and buffer 20–100 frames with 0.05–0.195 s exposure time were acquired and frames showing radiation damage manually removed. SEC-SAXS data was collected at the BM29 BioSAXS beamline using a wavelength of 0.99 Å and a Superdex 10/300 S200 Increase size exclusion column and analysed using Chromixs^63,64. Per run 2500 frames with 1s exposure time were collected. Data analysis was done using the ATSAS package version 2.8.3⁶⁵. PRIMUS was used for frame averaging and buffer subtraction⁶⁶. The radius of gyration, R_g, was estimated using the Guinier approximation in PRIMUS. Pair-wise distribution functions were calculated using GNOM⁶⁷. Collection statistics for SAXS measurements are summarized in Supplementary Table 2.

Circular dichroism spectroscopy

CD spectra were acquired using a Jasco 815 circular dichroism spectrometer. Samples were measured at 20 ˚C in quartz cuvettes with 1 mm path length in 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP at 0.5 mg/ml.

Protein-RNA crosslink identification by CLIR-MS/MS

TRIM25 crosslinks were identified by CLIR-MS/MS according to the procedure described in Dorn et al.³⁹, using a Thermo Orbitrap Fusion Lumos mass spectrometer. The cross-linked peptide is detected as a triple charged ion with a mass difference of approximately 11 Da, corresponding to the introduction of ¹³C and ¹⁵N stable isotopes in the heavy RNA. MS/MS spectra of the peptide crosslinked to light U (right, top) or heavy U (right, bottom). Assigned fragment ions with a relative intensity greater than 5% are labelled. All assigned fragment ions (b and y ions) are shown on the peptide sequence in the lower right-hand corner. The localisation of the crosslinking site to the tyrosine residue (highlighted in bold) is confirmed by observing the unmodified b7 ion and the b8 ion carrying the U modification. All fragment ions carrying the U modification (e.g. b₈, b₉, y₁₅) show a mass shift between the light and heavy form spectra.

Size exclusion chromatography – multi-angle light scattering (SEC-MALS)

50 µl of CC domain or the mutants were injected onto a Superdex 200 5/150 GL gel-filtration column (Cytiva) in 20 mM MES, pH 6.5, 75 mM NaCl and 0.5 mM TCEP, at a flow rate of 0.3 ml/min and at room temperature. The column was connected to a MiniDAWN MALS detector and Optilab differential refractive index detector (Wyatt Technology). Data were analyzed using the Astra 7 software s(Wyatt Technology).

Plasmids and recombinant DNA procedures

Synthetic genes for generation of inducible cell lines were obtain from Genscript. cloned into the pcDNA5/FRT with the HA-FLAG tag followed by a flexible linker encoding for GGSGGSGG (glycine and serine repeats) to facilitate the folding of the RBP of interest independently from the tag. The E3-dead mutant was generated by replacing Cys50 and Cys53 with serines which abolishes the E3 ligase activity of the RING domain as shown by Zheng et al¹⁸.

Cell lines generation

To generate the HEK293 TRIM25-WT, TRIM25-m3, TRIM25-m9 and TRIM25- E3-ligase dead cell lines, HEK293 TRIM25 KO cells were transfected with pOG44 and the corresponding pcDNA5/FRT plasmid containing the protein of interest using X-tremeGENE 360 transfection reagent following manufacturer’s recommendations (Roche, 8724121001). For the selection of inducible cell lines, 50 µg/ml Hygromycin B was added to the media. Cells were selected and checked TRIM25 variant expression by western blotting (Extended Data Fig. 4a).

Western blotting

Samples were prepared by mixing with NuPAGE LDS Sample Buffer 4X (Thermo Fisher Scientific, #NP0008) and incubated at 95ºC for 5 minutes. Proteins were resolved by SDS-PAGE for 50 min at 180 V and transferred to nitrocellulose membranes using the Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked for 1 hour at RT or overnight at 4◦C. 5% skimmed milk in 0.1% PBS-T was used for blocking and for the preparation of antibody dilutions. Membranes were incubated with primary antibodies either for 1 hour at RT or overnight at 4◦C. After three washes in 0.1% PBS-T, the membranes were incubated with the appropriate secondary antibodies for 1 hour at RT. The membranes were then washed three more times in 0.1% PBS-T prior to imaging. LI-COR Odyssey CLx imaging system for visualisation and Image Studio software for quantification.

Viruses

Sindbis virus was produced by in vitro transcription from the plasmid pT7svwt⁶⁸. SINV-mCherry and SINV-nsp3-scarlet was generated by duplication of subgenomic promoter and insertion of mCherry⁴² (Extended Data Fig. 5B). SINV-nsp3-scarlet was generated by amplifying mScarlet sequence from pUC19_CRISPR_CT_Scarlet_GSG_P2A_BSD (primer forward: GGGGACTAGTATGGTGAGCAAGG; reverse: ccccACTAGTCTTGTACAGCTCG) and cloned into pT7SINV-WT using SpeI restriction enzyme (Extended Data Fig. 5b).

Fluorescence-based viral fitness assay

6x10⁴ cells were seeded on each well of a 96-well microplate with flat µClear bottom (Greiner Bio-One, #655986) in DMEM lacking phenol-red (Gibco # 31053028) and supplemented with 2% FBS and 1 mM sodium pyruvate. The different cell lines (KO rescued WT, E3-ligase dead, m3, and m9) were infected with SINV-mCherry or SINV-nsp3-mScarlet (see Extended Data Fig. 5b) at 0.1 MOI in complete DMEM lacking phenol-red and supplemented with FBS 2%. Cells were incubated at 37°C and 5% CO₂ in a CLARIOstar fluorescence plate reader (BMG Labtech) for 24 h in FBS 2%; mCherry signal was monitored by measuring fluorescence (mCherry: excitation 570 nm, emission 620 nm; mScarlet: excitation 569 nm, emission 594 nm) every 15 minutes. Statistical significance of the difference in mCherry or mScarlet expression at 18 and 22 hpi was determined by t test (from n > 3).

iCLIP2

To identify TRIM25 binding sites for RNA on mock and SINV-infected cells at a high resolution, we employed iCLIP2-seq⁴³ with the modifications explained in Garcia-Moreno et al., 2023⁴⁴. 10x10⁶ HEK293 TRIM25 WT, TRIM25-m3, and TRIM25-m9 cells were seeded in 6 sets of 10 cm dishes and incubated for 24 h. Three cell sets per cell line were then infected with 3 MOI of SINV. At 9 hpi, cells were washed with PBS 1x, and UV irradiated with 300 mJ/cm² at 254 nm. To obtain RNA fragments of suitable length and to degrade DNA, 1 ml of thawed lysate of each replicate was incubated with 10 U Ambion RNase I (ThermoScientific, #AM2295) and 4 U Turbo DNase per ml of lysate for 3 minutes at 37°C, with 1100 rpm agitation. Subsequently, lysates were placed on ice and supplemented with 200 U RiboLock RNase Inhibitor (ThermoScientific, #EO0382). Then, 10 µl of lysate was taken from each sample for the size-matched input (SMI). 30 µl of ANTI-FLAG M2 Magnetic Beads (Millipore, #M8823) slurry per ml of lysate was pre-equilibrated in lysis buffer and resuspended in 50 µl of lysis buffer. Beads were added to the lysate and incubated for 2 hours at 4°C with gentle rotation. Next, we followed the protocol, explained in Garcia-Moreno et al., 2023. To decide the number of cycles, a qPCR was performed using EvaGreen ((Biotium, #31000), 2X Phusion HF PCR Master mix and P5/P3 Solexa primers (the optimal cycle number was considered as the cycle number at the centre at the “knee of exponential growth” -3). The number of cycles used for each replicate can be found in the supporting information (Supplementary Table 4). The final PCR products were purified by using ProNex beads (Promega, # NG2001) which allow for selecting the DNA of the correct size. The libraries were quantified using Qubit and the size was determined by using High Sensitivity D1000 TapeStation. Each group of samples was pooled equimolarly and then mixed considering the number of samples of the IP and the SMI and the ratio of reads. The final library was processed using single-end sequencing mode with a NextSeq 500/550 High Output v2 kit (75 cycles, Illumina).

Data processing for iCLIP2 and statistical analysis

Raw FASTQ files were demultiplexed using the Je Suite⁶⁹ and adapters were trimmed using Cutadapt⁷⁰. STAR was used to align reads to a concatenated human (GRCh38, ENSEMBL Release 106) and SINV (pT7-SVwt) genome in end-to-end alignment mode.⁷¹ Only uniquely aligned reads were retained for downstream analysis. PCR duplicates were collapsed using unique molecular identifiers (UMIs) with the Je Suite. The crosslink truncation site for each read (-1 from the 5’ start site of the read) was extracted using BEDTools⁷².Peak calling was performed with HTSeq-clip and the R/Bioconductor package, DEW-seq⁷³. HTSeq-clip was used to generate a sliding window annotation of the human and SINV genome (50nt window, 20nt step size) and calculate the frequency of crosslink truncation sites within each window. DEW-Seq was then used to calculate the differential enrichment of each window relative to size matched input control samples, with a cut-off of 2 log2 fold change and 0.01 adjusted p-value. Multiple hypothesis correction was performed using the Independent Hypothesis Weighting (IHW) method⁷⁴. Overlapping windows were merged to form binding regions. PCA analysis was performed using DESeq2. Following size correction and variance stabilisation, the 1000 most variable sliding windows were selected and used for PCA plotting. Binding site properties, including gene name, biotype, gene feature, were extracted from the ENSEMBL genome annotation using the GenomicRanges package. Metagene analyses were performed using functions from the cliProfiler package. Sequences for motif prediction and secondary structure prediction were defined for each binding site as a 50-nucleotide region, centred on the peak in BigWig signal. They were extracted from the CDS sequence of the longest isoform of the gene in which the binding site occurred. For motif prediction, a gene and gene region-matched background sequence was extracted for each binding site to allow for differential enrichment analysis. Enrichment analysis was performed using HOMER. MotifStack was used to cluster and plot motifs. Secondary structures were predicted using RNAfold from ViennaRNA⁴⁸ and specific structure features were extracted by forgi. Enrichment for paired sequence was calculated relative to 10 scrambled controls of each input sequence. For plotting of SINV viral genome structures, SHAPE data from Kutchko et al⁷⁵ was included as a constraint in the structure prediction. Coordinates for plotting specific structures were generated with the RRNA package. SINV genome coverage in reads per million was calculated using BEDTools. SMI signal was subtracted from IP signal for plotting.

Combined inmunofluorescence and RNA smFISH assays

1.8x10⁵ cells were seeded on High Precision Coverslips (Merienfeld, #0117520) and incubated for 24 hours in DMEM with 10% FBS. Cells were either mock-infected or infected with 1 MOI of SINV in DMEM without FBS at 37°C, followed by the replacement of the medium with DMEM supplemented with 2% FBS. At 18 hours post-infection, cells were washed with PBS and fixed in 4% paraformaldehyde for 10 minutes at room temperature. After three 5-minute washes in PBS, the cells were permeabilised with 1x PBS supplemented with 0.1% Triton X-100 for 10 minutes at room temperature. Cells were washed twice with 1X PBS for 5 minutes each, twice with 2x SSC for 5 minutes each, and twice with pre-warmed pre-hybridization buffer (2x SSC and 10% deionized formamide in DEPC-treated water) at 37°C for 20 minutes each. The cells were then incubated with 125 nM SINV RNAs specific Stellaris probes (LGC Biosearch Technologies) in pre-heated hybridization buffer (2x SSC, 10% deionized formamide and 10% dextran sulphate in DEPC water) at 37°C in a humidified chamber overnight. Cells were then washed twice with pre-warmed pre-hybridization buffer for 20 minutes at 37°C, twice with 2X SSC for 5 minutes each, and twice with 1X PBS 5 minutes each. All solutions used from now on were RNase free by adding RiboLock RNase Inhibitor (Thermo Fisher Scientific, # EO0381). Cells were incubated with 1X PBS-BSA 5% for 30 minutes followed by primary antibody (HA antibody, Biolegend 901502 1:250 dilution) in 1X PBS-Tween20 0.1% (PBST) supplemented with 5% BSA for 1 hour. Next, the cells were washed three times with PBST for 5 minutes each and were then incubated with secondary antibody (α-rabbit Alexa488 at 1:500 dilution) in 0.1% PBST supplemented with 5% BSA for 1 hour in the dark. The cells were washed three times in 0.1% PBST and once in 1X PBS for 5 minutes each. Cells were then incubated with DAPI (1:500 of 0.5 mg/mL stock) in 1X PBS for 10 minutes at room temperature, followed by two washes with 1x PBS for 5 minutes and one wash with DEPC-treated pure water for 1 minute. Finally, coverslips were collected and mounted on glass slides using ProLong Diamond Antifade Mounting Medium (Catalog number: P36961).

Image acquisition

Cells were imaged on a Zeiss LSM880 confocal system using a 63X oil objective (Plan-Apochromat 63x/1.4 Oil DIC M27). Images were generated using Zeiss Zen 3.4 Blue Edition software.

AUTHOR CONTRIBUTIONS

Conception/design: L.A., K.H., A.C., J.H; Data acquisition: L.A., K.H., V.R., L.G., S.A., I.H., B.S., P.W., F.G., A.L.; Material contribution: M.L., S.C., K.R., G.M., N.R.C.; Data analysis: L.A., K.H., L.I., L.G., K.L., F.G., A.L., A.C., J.H.; Data interpretation: L.A., K.H., L.I., A.L., F.A., A.C., J.H.; Writing, original draft: L.A., K.H., A.C., J.H.; Writing, editing, all authors; Funding acquisition: A.C., J.H.: Supervision, S.C., A.L., M.H., F.A., A.C., J.H..

ACKNOWLEDGEMENTS

L.A. was supported by a fellowship from the EMBL Interdisciplinary (EI4POD) program under Marie Skłodowska-Curie Actions COFUND (847543) and acknowledges support by the Joachim Herz Foundation through an Add-on Fellowship for Interdisciplinary Life Sciences. L.I. is funded by BBSRC DTP scholarship number BB/M011224/1 V.R. is funded by the European Union’s Horizon 2020 research and innovation programme under Marie-Sklodowska-Curie n. 892756. K.R. is supported by the Francis Crick Institute, which receives its core funding from Cancer Research UK (CC2075), the UK Medical Research Council (CC2075), and the Wellcome Trust (CC2075). A.L. acknowledges Paola Picotti (ETH Zurich) for access to instrumentation and laboratory infrastructure. G.M. is funded by project financed under Dioscuri, a programme initiated by the Max Planck Society, jointly managed with the National Science Centre in Poland, and mutually funded by Polish Ministry of Science and Higher Education and German Federal Ministry of Education and Research (2019/02/H/NZ1/00002). G.M. was also a recipient BBSRC project grant (BB/T002751/1). F.A. acknowledge support of the SNSF NCCR RNA and Disease (51NF40-182880). A.C. is funded by the European Research Council (ERC) Consolidator Grant ‘vRNP-capture’ 101001634 and the MRC grants MR/R021562/1 and MC_UU_00034/2. J.H. was funded by the Deutsche Forschungsgemeinschaft (DFG) via an Emmy-Noether Fellowship (HE7291/1–1) and acknowledges the EMBL for support.

Takeuchi, O. & Akira, S. Innate immunity to virus infection. Immunol Rev 227, 75–86 (2009).
Deng, Y., Wang, Y., Li, L., Miao, E. A. & Liu, P. Post-Translational Modifications of Proteins in Cytosolic Nucleic Acid Sensing Signaling Pathways. Front Immunol 13, 898724 (2022).
Zinngrebe, J., Montinaro, A., Peltzer, N. & Walczak, H. Ubiquitin in the immune system. EMBO Rep 15, 28–45 (2014).
Jiang, X. & Chen, Z. J. The role of ubiquitylation in immune defence and pathogen evasion. Nat Rev Immunol 12, 35–48 (2012).
Davis, M. E. & Gack, M. U. Ubiquitination in the antiviral immune response. Virology 479–480, 52–65 (2015).
Hu, H. & Sun, S.-C. Ubiquitin signaling in immune responses. Cell Res 26, 457–83 (2016).
Esposito, D., Koliopoulos, M. G. & Rittinger, K. Structural determinants of TRIM protein function. Biochem Soc Trans 45, 183–191 (2017).
Marín, I. Origin and diversification of TRIM ubiquitin ligases. PLoS One 7, e50030 (2012).
Short, K. M. & Cox, T. C. Subclassification of the RBCC/TRIM superfamily reveals a novel motif necessary for microtubule binding. J Biol Chem 281, 8970–80 (2006).
Gack, M. U. et al. TRIM25 RING-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity. Nature 446, 916–920 (2007).
Gack, M. U. et al. Roles of RIG-I N-terminal tandem CARD and splice variant in TRIM25-mediated antiviral signal transduction. Proc Natl Acad Sci U S A 105, 16743–8 (2008).
Oshiumi, H., Matsumoto, M., Hatakeyama, S. & Seya, T. Riplet/RNF135, a RING finger protein, ubiquitinates RIG-I to promote interferon-beta induction during the early phase of viral infection. J Biol Chem 284, 807–17 (2009).
Kuniyoshi, K. et al. Pivotal role of RNA-binding E3 ubiquitin ligase MEX3C in RIG-I-mediated antiviral innate immunity. Proc Natl Acad Sci U S A 111, 5646–51 (2014).
Yan, J., Li, Q., Mao, A.-P., Hu, M.-M. & Shu, H.-B. TRIM4 modulates type I interferon induction and cellular antiviral response by targeting RIG-I for K63-linked ubiquitination. J Mol Cell Biol 6, 154–63 (2014).
Hayman, T. J. et al. RIPLET, and not TRIM25, is required for endogenous RIG-I-dependent antiviral responses. Immunol Cell Biol 97, 840–852 (2019).
Cadena, C. et al. Ubiquitin-Dependent and -Independent Roles of E3 Ligase RIPLET in Innate Immunity. Cell 177, 1187-1200.e16 (2019).
Choudhury, N. R. et al. TRIM25 inhibits influenza A virus infection, destabilizes viral mRNA, but is redundant for activating the RIG-I pathway. Nucleic Acids Res 50, 7097–7114 (2022).
Zheng, X. et al. TRIM25 Is Required for the Antiviral Activity of Zinc Finger Antiviral Protein. J Virol 91, (2017).
Li, M. M. H. et al. TRIM25 Enhances the Antiviral Action of Zinc-Finger Antiviral Protein (ZAP). PLoS Pathog 13, e1006145 (2017).
Zhu, Y. et al. Zinc-finger antiviral protein inhibits HIV-1 infection by selectively targeting multiply spliced viral mRNAs for degradation. Proc Natl Acad Sci U S A 108, 15834–9 (2011).
Guo, X., Carroll, J.-W. N., Macdonald, M. R., Goff, S. P. & Gao, G. The zinc finger antiviral protein directly binds to specific viral mRNAs through the CCCH zinc finger motifs. J Virol 78, 12781–7 (2004).
Guo, X., Ma, J., Sun, J. & Gao, G. The zinc-finger antiviral protein recruits the RNA processing exosome to degrade the target mRNA. Proc Natl Acad Sci U S A 104, 151–6 (2007).
Zhu, Y., Wang, X., Goff, S. P. & Gao, G. Translational repression precedes and is required for ZAP-mediated mRNA decay. EMBO J 31, 4236–46 (2012).
Castello, A. et al. Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. Cell 149, 1393–406 (2012).
Kwon, S. C. et al. The RNA-binding protein repertoire of embryonic stem cells. Nat Struct Mol Biol 20, 1122–30 (2013).
Hentze, M. W., Castello, A., Schwarzl, T. & Preiss, T. A brave new world of RNA-binding proteins. Nat Rev Mol Cell Biol 19, 327–341 (2018).
Castello, A. et al. Comprehensive Identification of RNA-Binding Domains in Human Cells. Mol Cell 63, 696–710 (2016).
Choudhury, N. R. et al. RNA-binding activity of TRIM25 is mediated by its PRY/SPRY domain and is required for ubiquitination. BMC Biol (2017) doi:10.1186/s12915-017-0444-9.
Sanchez, J. G. et al. TRIM25 Binds RNA to Modulate Cellular Anti-viral Defense. J Mol Biol 430, 5280–5293 (2018).
Manokaran, G. et al. Dengue subgenomic RNA binds TRIM25 to inhibit interferon expression for epidemiological fitness. Science (1979) (2015) doi:10.1126/science.aab3369.
Meyerson, N. R. et al. Nuclear TRIM25 Specifically Targets Influenza Virus Ribonucleoproteins to Block the Onset of RNA Chain Elongation. Cell Host Microbe 22, 627-638.e7 (2017).
Wickenhagen, A. et al. A prenylated dsRNA sensor protects against severe COVID-19. Science (1979) 374, (2021).
Vaishali et al. Validation and classification of RNA binding proteins identified by mRNA interactome capture. RNA 27, 1173–1185 (2021).
Choudhury, N. R. et al. Trim25 Is an RNA-Specific Activator of Lin28a/TuT4-Mediated Uridylation. Cell Rep 9, 1265–1272 (2014).
Koliopoulos, M. G. et al. Molecular mechanism of influenza A NS1-mediated TRIM25 recognition and inhibition. Nat Commun (2018) doi:10.1038/s41467-018-04214-8.
Lin, H. et al. The long noncoding RNA Lnczc3h7a promotes a TRIM25-mediated RIG-I antiviral innate immune response. Nat Immunol (2019) doi:10.1038/s41590-019-0379-0.
Maris, C., Dominguez, C. & Allain, F. H.-T. The RNA recognition motif, a plastic RNA-binding platform to regulate post-transcriptional gene expression. FEBS J 272, 2118–31 (2005).
Lai, C. et al. Long noncoding RNA AVAN promotes antiviral innate immunity by interacting with TRIM25 and enhancing the transcription of FOXO3a. Cell Death Differ 28, 2900–2915 (2021).
Dorn, G. et al. Structural modeling of protein-RNA complexes using crosslinking of segmentally isotope-labeled RNA and MS/MS. Nat Methods 14, 487–490 (2017).
Whitty, A. Cooperativity and biological complexity. Nat Chem Biol 4, 435–9 (2008).
Ercolani, G. & Schiaffino, L. Allosteric, Chelate, and Interannular Cooperativity: A Mise au Point. Angewandte Chemie International Edition 50, 1762–1768 (2011).
Garcia-Moreno, M. et al. System-wide Profiling of RNA-Binding Proteins Uncovers Key Regulators of Virus Infection. Mol Cell (2019) doi:10.1016/j.molcel.2019.01.017.
Buchbender, A. et al. Improved library preparation with the new iCLIP2 protocol. Methods 178, 33–48 (2020).
Garcia-Moreno, M. et al. Incorporation of genome-bound cellular proteins into HIV-1 particles regulates viral infection. doi:10.1101/2023.06.14.544764.
Van Nostrand, E. L. et al. Robust transcriptome-wide discovery of RNA-binding protein binding sites with enhanced CLIP (eCLIP). Nat Methods 13, 508–14 (2016).
Carrasco, L., Sanz, M. & González-Almela, E. The Regulation of Translation in Alphavirus-Infected Cells. Viruses 10, 70 (2018).
Kutchko, K. M. et al. Structural divergence creates new functional features in alphavirus genomes. Nucleic Acids Res 46, 3657–3670 (2018).
Lorenz, R. et al. ViennaRNA Package 2.0. Algorithms Mol Biol 6, 26 (2011).
Yuan, Y. et al. Trim25 restricts rabies virus replication by destabilizing phosphoprotein. Cell insight 1, 100057 (2022).
Sánchez-Aparicio, M. T., Ayllón, J., Leo-Macias, A., Wolff, T. & García-Sastre, A. Subcellular Localizations of RIG-I, TRIM25, and MAVS Complexes. J Virol 91, (2017).
Yoo, J.-S. et al. DHX36 enhances RIG-I signaling by facilitating PKR-mediated antiviral stress granule formation. PLoS Pathog 10, e1004012 (2014).
Law, L. M. J. et al. ZAP’s stress granule localization is correlated with its antiviral activity and induced by virus replication. PLoS Pathog 15, e1007798 (2019).
de Marco, A., Deuerling, E., Mogk, A., Tomoyasu, T. & Bukau, B. Chaperone-based procedure to increase yields of soluble recombinant proteins produced in E. coli. BMC Biotechnol 7, 32 (2007).
Delaglio, F. et al. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR 6, 277–93 (1995).
Lee, W., Tonelli, M. & Markley, J. L. NMRFAM-SPARKY: enhanced software for biomolecular NMR spectroscopy. Bioinformatics 31, 1325–7 (2015).
Paysan-Lafosse, T. et al. InterPro in 2022. Nucleic Acids Res 51, D418–D427 (2023).
Larkin, M. A. et al. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947–8 (2007).
Varadi, M. et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res 50, D439–D444 (2022).
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Hubbard, T. J. P. et al. Ensembl 2009. Nucleic Acids Res 37, D690-7 (2009).
Schneider, T. D. & Stephens, R. M. Sequence logos: a new way to display consensus sequences. Nucleic Acids Res 18, 6097–100 (1990).
Blanchet, C. E. et al. Versatile sample environments and automation for biological solution X-ray scattering experiments at the P12 beamline (PETRA III, DESY). J Appl Crystallogr 48, 431–443 (2015).
Pernot, P. et al. Upgraded ESRF BM29 beamline for SAXS on macromolecules in solution. J Synchrotron Radiat 20, 660–4 (2013).
Panjkovich, A. & Svergun, D. I. CHROMIXS: automatic and interactive analysis of chromatography-coupled small-angle X-ray scattering data. Bioinformatics 34, 1944–1946 (2018).
Franke, D. et al. ATSAS 2.8: a comprehensive data analysis suite for small-angle scattering from macromolecular solutions. J Appl Crystallogr 50, 1212–1225 (2017).
Konarev, P. V., Volkov, V. V., Sokolova, A. V., Koch, M. H. J. & Svergun, D. I. PRIMUS : a Windows PC-based system for small-angle scattering data analysis. J Appl Crystallogr 36, 1277–1282 (2003).
Svergun, D. I. Determination of the regularization parameter in indirect-transform methods using perceptual criteria. J Appl Crystallogr 25, 495–503 (1992).
Castelló, A., Sanz, M. Á., Molina, S. & Carrasco, L. Translation of Sindbis Virus 26S mRNA Does Not Require Intact Eukariotic Initiation Factor 4G. J Mol Biol 355, 942–956 (2006).
Girardot, C., Scholtalbers, J., Sauer, S., Su, S.-Y. & Furlong, E. E. M. Je, a versatile suite to handle multiplexed NGS libraries with unique molecular identifiers. BMC Bioinformatics 17, 419 (2016).
Kechin, A., Boyarskikh, U., Kel, A. & Filipenko, M. cutPrimers: A New Tool for Accurate Cutting of Primers from Reads of Targeted Next Generation Sequencing. J Comput Biol 24, 1138–1143 (2017).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–2 (2010).
Sahadevan, S., Sekaran, T. & Schwarzl, T. A Pipeline for Analyzing eCLIP and iCLIP Data with Htseq-clip and DEWSeq. Methods Mol Biol 2404, 189–205 (2022).
Ignatiadis, N., Klaus, B., Zaugg, J. B. & Huber, W. Data-driven hypothesis weighting increases detection power in genome-scale multiple testing. Nat Methods 13, 577–80 (2016).
Kutchko, K. M. & Laederach, A. Transcending the prediction paradigm: novel applications of SHAPE to RNA function and evolution. Wiley Interdiscip Rev RNA 8, (2017).

There is NO Competing Interest.

Download PDF

Journal Publication

published 01 Oct, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

The molecular dissection of TRIM25‘s RNA-binding mechanism provides key insights into its antiviral activity

Status:

Journal Publication

Version 1

Abstract

Figures

INTRODUCTION

RESULTS

Two distinct sites of the TRIM25 PRY/SPRY domain confer binding to single- and double-stranded RNA

CC binds to RNA and the PRY/SPRY domain using adjacent interfaces

PRY/SPRY and CC domains bind to RNA cooperatively

iCLIP2 data reveals cellular structure- and sequence specificity of TRIM25

RNA binding is critical for TRIM25’s antiviral activity and relocation to viral replication organelles

DISCUSSION

MATERIAL AND METHODS

Protein expression and purification

RNA production

Nuclear magnetic resonance spectroscopy

Isothermal titration calorimetry

Sequence alignments

Small-angle X-ray scattering

Circular dichroism spectroscopy

Protein-RNA crosslink identification by CLIR-MS/MS

Size exclusion chromatography – multi-angle light scattering (SEC-MALS)

Plasmids and recombinant DNA procedures

Cell lines generation

Western blotting

Viruses

Fluorescence-based viral fitness assay

iCLIP2

Data processing for iCLIP2 and statistical analysis

Combined inmunofluorescence and RNA smFISH assays

Image acquisition

Declarations

AUTHOR CONTRIBUTIONS

ACKNOWLEDGEMENTS

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1