A UV-Based ChIRP Method For The Verifiable Identification of Proteins That Directly Interact With LncRNA: From Cultured Cells To Mouse Testis

Summary

A UV-based ChIRP method to identify direct lncRNA-protein interactions.

Abstract

Long non-coding RNAs (lncRNAs) have emerged as important regulators of diverse biological processes. However, the molecular mechanisms underlying the functions of most lncRNAs remain poorly understood. Identifying interacting proteins, particularly RNA-binding proteins (RBPs), is a key strategy for elucidating lncRNA function. Among available approaches, comprehensive identification of RNA-binding proteins (ChIRP) coupled with mass spectrometry is widely used because of its straightforward probe design and simple workflow. Conventional ChIRP relies on paraformaldehyde (PFA) crosslinking, which captures both direct RNA-protein contacts and indirect interactions mediated by bridging factors. In contrast, UV crosslinking combined with stringent washing conditions preferentially preserves direct RNA-protein interactions. Using the lncRNA Tug1 as a model, UV crosslinking and PFA crosslinking were compared within the ChIRP framework in cultured cells, demonstrating that UV-based ChIRP enriches for direct lncRNA-protein interactions. This UV-based workflow was subsequently extended to mouse testes, in which tissues were dissociated into single-cell suspensions prior to UV crosslinking, enabling the identification of Tug1-binding proteins in a native tissue context. Comparison with cell line-derived interactomes revealed both shared and tissue-specific interactors. Two shared direct interactions were independently validated by enhanced crosslinking immunoprecipitation (eCLIP). This UV-based ChIRP approach is referred to as cPDiRT (capturing proteins directly interacting with an RNA target). Together, cPDiRT enables the identification of lncRNA-binding proteins in both cultured cells and native tissues. The combination of tissue dissociation and UV crosslinking provides a broadly applicable strategy for studying direct lncRNA-protein interactions across diverse organ systems.

Introduction

Recent high-throughput transcriptomic studies have uncovered thousands of long non-coding RNAs (lncRNAs) that participate in diverse biological processes, ranging from chromatin organization to cellular signaling^1,2. Despite their widespread expression and functional importance, the molecular mechanisms of most lncRNAs remain poorly defined. A major obstacle is the identification of their physical protein partners, particularly RNA-binding proteins (RBPs)^3,4,5.

Current RNA-centric methods primarily rely on chemical crosslinking, which captures both direct and indirect RNA-protein interactions. As a result, these approaches are more suitable for global interactome mapping than for identifying bona fide direct contacts⁶. Comprehensive identification of RNA-binding proteins followed by mass spectrometry (ChIRP-MS) is a widely used technique for mapping lncRNA-RBP interactions because of its straightforward probe design and efficient workflow^7,8. Conventional ChIRP-MS uses paraformaldehyde (PFA) crosslinking, which captures both direct RNA-protein contacts and indirect associations mediated by bridging factors^9,10.

In contrast, UV crosslinking at 254 nm offers a critical advantage. It covalently links RNA and proteins only when they are in near-zero-distance contact (typically <1 Å)¹¹. This process converts hydrogen bonds into stable covalent bonds and enables stringent high-salt washing conditions that remove non-covalently bound indirect interactors. Consequently, UV-based approaches provide greater specificity for direct RNA-protein interactions than chemical crosslinking methods^12,13. Nevertheless, UV-based approaches remain underutilized in tissue-derived lncRNA interactome studies, largely due to concerns about crosslinking efficiency and limited UV penetration.

To investigate direct lncRNA-protein interactions in a physiologically relevant context, the lncRNA Tug1 was selected as a model system. Tug1 represents an ideal benchmark for validating a tissue-compatible direct-interaction workflow because of its severe knockout phenotype and high enrichment in the testis. Genetic ablation of Tug1 results in complete male sterility, impaired spermatogenesis, and abnormal sperm morphology¹⁴. In addition, Tug1 supports Sertoli cell function and blood-testis barrier integrity in high-fat diet mice¹⁵. Despite these well-established phenotypes, the direct protein partners that mediate Tug1 function in the testis—an organ expressing large numbers of RBPs that have recently been experimentally profiled¹⁶—remain unknown. This knowledge gap largely reflects the fact that most lncRNA-protein interaction mapping methods have been developed for cultured cell systems.

To address this limitation, the ChIRP framework was modified by replacing PFA crosslinking with UV crosslinking to capture proteins that directly interact with Tug1. Candidate interactions were independently validated by enhanced crosslinking immunoprecipitation followed by deep sequencing (eCLIP-seq)¹⁷. The workflow was further extended to tissue applications by dissociating fresh mouse testicular tissue into single cells prior to UV crosslinking. This optimized approach was designated cPDiRT (capturing proteins directly interacting with an RNA target).

Application of cPDiRT to Tug1 in the mouse testis enabled direct-interaction mapping in a native tissue context. Parallel experiments in cultured cells provided a side-by-side comparison of PFA-based and UV-based interaction profiles. Combined with independent eCLIP validation, this approach generated a high-confidence set of molecular partners that may contribute to Tug1 function during spermatogenesis. The protocol presented here enables the capture of proteins that directly interact with a target lncRNA in both cultured cells and native tissues. By defining the direct interactome of Tug1 in the testis, this method advances mechanistic studies of a fertility-essential lncRNA and provides a generalizable platform for investigating lncRNA-protein interactions in reproductive and other complex tissues.

Protocol

All solutions were prepared using nuclease-free, protease-free ultrapure water and reagents. All reagents and equipment are listed in the Table of Materials.

All animal experiments were approved by the Animal Ethics Committee of Nanjing Medical University (Approval No.: IACUC-1812003-4) and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 8th edition). Maintain male C57BL/6 mice (4 weeks old) under a 12 h light/12 h dark cycle with ad libitum access to food and water. Minimize animal suffering by using the fewest animals necessary to achieve statistical significance. Euthanize mice by CO₂ inhalation followed by cervical dislocation according to approved institutional protocols. 

1. Prepare buffers

1. Prepare cell lysis buffer
   1. Prepare cell lysis buffer containing 50 mM Tris-HCl (pH 7.0), 10 mM EDTA, and 1% SDS.
   2. Immediately before use, add 1 mM PMSF, protease inhibitor cocktail, and specialized RNase inhibitor.
      CAUTION: SDS is a skin and eye irritant. Avoid inhaling dust or aerosols. PMSF is highly toxic, acts as a cholinesterase inhibitor, and is unstable in aqueous solution. Prepare PMSF immediately before use. Wear gloves and safety glasses when handling these reagents. Handle PMSF in a chemical fume hood.
2. Prepare hybridization buffer
   1. Prepare hybridization buffer containing 750 mM NaCl, 1% SDS, 50 mM Tris-HCl (pH 7.0), 1 mM EDTA, and 15% formamide.
   2. Immediately before use, add 1 mM PMSF, protease inhibitor cocktail, and specialized RNase inhibitor.
      NOTE: Store formamide at 4 °C, protected from light.
      CAUTION: Formamide is a teratogen and reproductive hazard. Handle formamide in a chemical fume hood while wearing gloves, a lab coat, and safety glasses. Avoid inhalation and skin contact.
3. Prepare wash buffer
   1. Prepare wash buffer containing 50 mM Tris-HCl (pH 7.0), 1 M NaCl, 1 mM EDTA, 1% NP-40, 0.1% SDS, and 0.5% sodium deoxycholate.
   2. Protect the wash buffer from light during storage.
4. Prepare proteinase K (PK) buffer
   1. Prepare PK buffer containing 100 mM NaCl, 10 mM Tris-HCl (pH 7.0), 1 mM EDTA, and 0.5% SDS.
   2. Add proteinase K to a final concentration of 5% immediately before use.

2. Design and prepare probes

1. Design antisense probes
   1. Design antisense oligonucleotide probes using the online probe designer available at singlemoleculefish.com¹⁸.
2. Obtain biotinylated probes
   1. Order antisense DNA probes containing a 3'-BiotinTEG modification.
3. Prepare the probe pool
   1. Dilute each probe individually to 100 µM.
   2. Mix equal volumes of all probes to generate a probe pool. Maintain a final combined oligonucleotide concentration of 100 µM.
   3. Store the probe pool at -20 °C.
      NOTE: Recommended probe design parameters are as follows: (i) one probe per 100 bp of target RNA, (ii) target GC content of 45%, (iii) probe length of 20 nt, and (iv) spacing of 60–80 nt between probes.
      NOTE: If the RNA exceeds the size limit accepted by the probe designer, divide the sequence into smaller segments before design. Exclude repetitive regions and regions with extensive sequence homology.

3. Perform cell harvesting, tissue dissociation, and UV crosslinking

1. Prepare no-UV controls
   1. Prepare control samples in parallel with experimental samples.
   2. Omit the UV-crosslinking step from control samples.
   3. Perform all subsequent procedures, including sonication, RNA-Protein pull-down, RNA extraction, and protein extraction, identically for both groups.
      NOTE: Use the no-UV control to distinguish UV-dependent RNA-protein interactions from non-specific background binding.
2. Harvest cells and perform UV crosslinking
   1. Wash cultured cells
      1. Wash cells three times with PBS to remove residual culture medium.
   2. Crosslink cells
      1. Distribute 3 mL PBS evenly onto a 15 cm culture dish.
      2. Irradiate cells in a 254 nm UV crosslinker at a dose of 4000 mJ/cm².
   3. Collect crosslinked cells
      1. Collect cells in PBS following crosslinking.
      2. Centrifuge the suspension and discard the supernatant.
      3. Weigh the pellet and confirm a minimum pellet mass of 500 mg per sample.
3. Dissociate testicular tissue and perform UV crosslinking
   1. Prepare testes
      1. Decapsulate testes.
      2. Wash the tissue twice with DMEM.
   2. Perform collagenase digestion
      1. Incubate tissue in DMEM containing 1 mg/mL collagenase IV at 37 °C for 5–15 min.
      2. Continue digestion until seminiferous tubules are released.
         NOTE: Do not exceed 15 min of digestion.
   3. Wash the seminiferous tubules
      1. Centrifuge the suspension at 1000 × g for 1 min at 4 °C.
      2. Remove the supernatant.
      3. Wash the pellet with 5–10 mL DMEM.
      4. Centrifuge again under the same conditions.
   4. Perform trypsin digestion
      1. Resuspend the pellet in 0.25% trypsin containing 1 mg/mL DNase I.
      2. Incubate at 37 °C for 5–15 min.
         NOTE: Do not exceed 15 min of digestion.
   5. Stop enzymatic digestion
      1. Add an equal volume of DMEM supplemented with 10% FBS.
   6. Generate a single-cell suspension
      1. Filter the suspension through a 70 µm cell strainer.
   7. Wash cells
      1. Centrifuge cells at 2000 × g for 5 min at 4 °C.
      2. Discard the supernatant.
      3. Wash the pellet with 5–10 mL PBS.
      4. Centrifuge again under the same conditions.
   8. Crosslink testicular cells
      1. Resuspend the single-cell suspension in PBS.
      2. Distribute the suspension evenly onto a 15 cm culture dish.
      3. Irradiate cells in a 254 nm UV crosslinker at a dose of 4000 mJ/cm².
   9. Collect crosslinked cells
      1. For no-UV controls, omit irradiation.
      2. Collect cells in PBS.
      3. Centrifuge the suspension and discard the supernatant.
      4. Weigh the pellet and confirm a minimum pellet mass of 500 mg per sample.
         NOTE: The number of mice required to obtain approximately 500 mg of cell pellet depends on age and tissue yield. Approximately 10–12 four-week-old mice were used in this study.

4. Perform sonication

1. Prepare lysates
   1. Resuspend cell pellets in cell lysis buffer at a ratio of 100 mg pellet per 1 mL buffer.
   2. Mix thoroughly to obtain a homogeneous suspension.
2. Sonicate samples
   1. Sonicate lysates using a water bath sonicator or a focused ultrasonicator.
   2. Continue sonication until the lysate becomes visually clear.
      NOTE: Verify DNA fragmentation by agarose gel electrophoresis. Continue sonication until most DNA fragments are between 1 kb and 2 kb.

5. Perform RNA-protein pull-down

1. Collect input samples
   1. Remove aliquots of lysate for input controls.
   2. Reserve approximately 10 µL for RNA analysis and 10 µL for protein analysis.
   3. Maintain input samples at 37 °C during the hybridization and wash procedures.
2. Preclear lysates
   1. Add 30 µL of pre-washed magnetic beads to each sample.
   2. Incubate samples at 37 °C for 30 min with gentle mixing in a hybridization oven.
      NOTE: Wash magnetic beads before use. Place bead aliquots on a magnetic stand for 1 min or until the solution becomes clear. Remove the storage solution completely. Wash beads twice with 500 µL of cell lysis buffer.
3. Remove preclearing beads
   1. Place samples on a magnetic stand.
   2. Transfer the supernatant to a fresh tube.
   3. Repeat the transfer once to ensure complete removal of residual beads.
4. Hybridize probes
   1. Add two volumes of hybridization buffer to each sample.
   2. Add probes at a ratio of 1 µL of 100 µM probe stock per 1 mL of lysate (Supplementary Table 1).
5. Incubate hybridization reactions
   1. Incubate samples at 37 °C with end-over-end rotation for 12–16 h in a hybridization oven.
6. Prepare capture beads
   1. Wash 100 µL of magnetic beads three times with cell lysis buffer immediately before the end of hybridization.
   2. Remove all residual buffer after the final wash.
7. Capture probe-bound complexes
   1. Briefly centrifuge hybridization reactions.
   2. Resuspend the washed beads in 1 mL of hybridized sample.
8. Bind complexes to beads
   1. Return the bead suspension to the original tube.
   2. Incubate at 37 °C with mixing for 30–40 min.
9. Recover beads
   1. Briefly centrifuge samples.
   2. Place tubes on a magnetic stand for 1–2 min or until the solution becomes clear.
   3. Discard the supernatant.
   4. Remove residual liquid using a pipette tip.
      NOTE: Avoid disturbing the magnetic bead pellet during liquid removal.
10. Wash captured complexes
    1. Add 1 mL of wash buffer.
    2. Wash beads at 37 °C with continuous mixing for 5 min.
    3. Repeat the wash procedure five times.
       NOTE: The wash buffer contains 1 M NaCl, which disrupts non-covalent protein-protein and protein-RNA interactions while preserving covalent UV-induced crosslinks. These stringent conditions remove indirect interactors and non-specific background, thereby enriching direct protein partners¹⁹.
11. Reserve material for RNA analysis
    1. Transfer 1%–10% of the washed beads to a fresh tube for RNA analysis.
    2. Use the remaining beads immediately for protein extraction or store them at -80 °C.
       NOTE: Pause the protocol here if necessary. Store bead samples at -80 °C until protein extraction.
12. Elute proteins
    1. Add SDS sample buffer to the beads.
    2. Boil samples at 95 °C for 30 min to elute proteins and reverse crosslinks.
    3. Proceed to mass spectrometry analysis.

6. Extract RNA

1. Digest proteins
   1. Resuspend bead-bound RNA samples in 100 µL of PK buffer.
   2. Adjust RNA input samples to a final volume of 100 µL using PK buffer.
2. Incubate with proteinase K
   1. Incubate samples at 50 °C for 45 min with continuous mixing.
3. Inactivate proteinase K
   1. Heat samples at 95 °C for 10 min.
4. Add TRIzol reagent
   1. Add 500 µL of TRIzol reagent to each sample.
   2. Vortex thoroughly.
   3. Incubate for 5 min at room temperature.
      CAUTION: TRIzol reagent contains phenol and guanidine thiocyanate. Handle the reagent in a chemical fume hood while wearing gloves, a lab coat, and safety glasses. Avoid inhalation and skin contact. If skin contact occurs, wash immediately with soap and water.
5. Perform phase separation
   1. Add 100 µL of chloroform to each sample.
   2. Vortex thoroughly.
   3. Incubate for 5 min at room temperature.
6. Recover the aqueous phase
   1. Centrifuge samples at 4 °C for 15 min.
   2. Transfer the upper aqueous phase to a fresh tube.
7. Bind RNA to purification columns
   1. Add an equal volume of 100% ethanol to the aqueous phase.
   2. Mix thoroughly.
   3. Load samples onto RNA purification columns.
8. Purify RNA
   1. Perform RNA cleanup according to the manufacturer's instructions.
   2. Elute RNA in 30 µL of nuclease-free water.
9. Verify RNA enrichment by qRT-PCR
   1. Use 1 µL of eluted RNA per reaction for quantitative reverse transcription PCR (qRT-PCR).
   2. Assess target RNA enrichment relative to controls.

7. Perform LC-MS/MS analysis

1. Separate proteins by SDS-PAGE
   1. Mix proteins obtained in Step 5.12 with 4× LDS sample buffer.
   2. Separate proteins on a 4%–12% Bis-Tris gel.
   3. Stain the gel using a silver staining kit.
   4. Excise the entire protein lane as a single band.
2. Excise gel pieces
   1. Excise the protein band using a clean scalpel.
   2. Minimize the amount of excess gel.
   3. Cut the gel into approximately 1 mm³ pieces.
   4. Transfer gel pieces to a 0.5 mL tube.
3. Destain gel pieces
   1. Add 100 µL of destaining solution.
   2. Vortex for 10 min.
   3. Centrifuge at 14,000 × g for 1 min.
   4. Remove the supernatant.
   5. Repeat the destaining procedure until the gel pieces become transparent.
4. Reduce proteins
   1. Add 100 µL of reduction solution.
   2. Incubate at 56 °C for 30 min.
   3. Cool samples to room temperature.
   4. Centrifuge briefly and remove the supernatant.
5. Alkylate proteins
   1. Add 100 µL of alkylation solution.
   2. Incubate at room temperature for 20 min in the dark.
   3. Remove the supernatant.
   4. Wash gel pieces with 100 µL of digestion buffer for 5 min.
6. Digest proteins
   1. Add sufficient trypsin solution to completely cover the gel pieces.
   2. Incubate at 37 °C for 16–18 h.
7. Extract peptides
   1. Add 30 µL of extraction solution.
   2. Sonicate for 15 min or vortex for 30 min.
   3. Centrifuge at 14,000 × g for 1 min.
   4. Transfer the supernatant to a fresh tube.
   5. Repeat the extraction procedure twice.
   6. Pool all extracts.
   7. Dry pooled extracts in a vacuum concentrator for approximately 1 h.
8. Clean peptides using a StageTip
   1. Resuspend dried peptides in 20 µL of 0.1% formic acid (FA).
   2. Load samples onto a C18 StageTip.
   3. Centrifuge at 1,000 × g for 2 min.
   4. Wash the StageTip with 20 µL of 0.1% FA.
   5. Elute peptides with 20 µL of 60% acetonitrile/0.1% FA into an LC-MS vial.
   6. Dry the eluate.
   7. Resuspend peptides in 10 µL of 0.1% FA.
      NOTE: Pause the protocol here if necessary. Store purified peptides at -80 °C before LC-MS/MS analysis.
9. Acquire LC-MS/MS data
   1. Reconstitute peptides in 0.1% formic acid.
   2. Separate peptides on a 75 µm × 250 mm C18 analytical column at a flow rate of 300 nL/min.
   3. Apply the following solvent B gradient (0.1% formic acid in acetonitrile):
      2%–25% B over 0–40 min
      25%–40% B over 40–48 min
      40%–99% B over 50–57 min
      Hold at 99% B for 3 min
   4. Operate the mass spectrometer in data-dependent acquisition mode.
   5. Acquire full-scan MS spectra over an m/z range of 350–1200 at a resolution of 70,000.
   6. Acquire fifteen sequential higher-energy collisional dissociation (HCD) MS/MS scans at a resolution of 17,500.
   7. Record one microscan per scan.
   8. Set dynamic exclusion to 30 s.
10. Process mass spectrometry data
    1. Process raw data using MaxQuant (v2.1.0.0).
    2. Search spectra against mouse protein sequences downloaded from UniProt.
    3. Specify trypsin as the digestion enzyme.
    4. Allow a maximum of two missed cleavages.
    5. Specify methionine oxidation as a variable modification.
    6. Specify cysteine carbamidomethylation as a fixed modification.
    7. Set precursor and fragment mass tolerances to 15 ppm and 20 ppm, respectively.
    8. Control the false discovery rate at 1% at both the peptide and protein levels.
    9. Estimate protein abundance using iBAQ-based label-free quantification.

8. Perform Gene Ontology enrichment analysis

1. Analyze functional enrichment
   1. Perform Gene Ontology (GO) enrichment analysis using the clusterProfiler R package (v4.6.0).
   2. Use the GO database (v2024-09-20) as the reference annotation source.
   3. Apply Benjamini-Hochberg correction for multiple testing.
   4. Define statistical significance as an adjusted P value (FDR) < 0.05.

9. Perform eCLIP-seq validation

1. Prepare eCLIP samples
   1. Perform enhanced crosslinking immunoprecipitation (eCLIP) as previously described¹⁷.
   2. Prepare UV-crosslinked cell samples for immunoprecipitation of PABPC1 and CCT3.
   3. Lyse samples using eCLIP lysis buffer.
2. Immunoprecipitate RNA-protein complexes
   1. Add 10 µg of antibody to each sample.
   2. Capture RNA-protein complexes on magnetic beads.
   3. Collect input samples before immunoprecipitation.
3. Prepare RNA libraries
   1. Dephosphorylate protein-bound RNAs.
   2. Ligate a 3' RNA adapter to recovered RNA.
   3. Wash immunoprecipitated and input samples under stringent conditions.
   4. Elute complexes by boiling the beads.
   5. Resolve samples on a 4%–12% Bis-Tris gradient gel.
   6. Transfer separated complexes to a nitrocellulose membrane.
4. Recover RNA
   1. Excise a size-matched membrane region.
   2. Digest proteins using urea and proteinase K.
   3. Recover RNA from the membrane.
5. Generate sequencing libraries
   1. Reverse-transcribe recovered RNA into cDNA.
   2. Purify cDNA using an exonuclease I and shrimp alkaline phosphatase mixture.
   3. Ligate a DNA adapter to the 3' end of the cDNA using T4 RNA ligase.
   4. Amplify libraries using a high-fidelity DNA polymerase.
   5. Separate amplified libraries on a 3% low-melting-point agarose gel.
   6. Purify size-selected libraries using a gel extraction kit.
   7. Sequence libraries on an Illumina platform.
6. Analyze eCLIP-seq data
   1. Analyze sequencing data using the ENCODE eCLIP pipeline²⁰.
   2. Trim adapters and 3' adapter dimers using cutadapt (v1.14).
   3. Align reads to UCSC repeat elements using STAR (v2.7.6a).
   4. Remove reads mapping to repeat elements.
   5. Align remaining reads to the mm10 reference genome (GENCODE vM23).
   6. Remove PCR duplicates.
   7. Identify enriched binding sites using CLIPPER.
   8. Define significant peaks as those with fold enrichment ≥ 8 and adjusted P ≤ 0.001.
   9. Identify reproducible peaks using an irreproducible discovery rate (IDR) threshold of 0.01.

Results

To identify Tug1-binding proteins, UV crosslinking was incorporated into the ChIRP workflow, establishing a method designated cPDiRT (Figure 1). Conventional ChIRP was first performed in GC2 cells to enrich Tug1-interacting proteins (Figure 2A, B). Mass spectrometry analysis identified more than 190 proteins enriched relative to the negative control (Figure 2C; Supplementary Table 2). In contrast, cPDiRT-MS performed at the optimized UV dose of 4000 mJ/cm² identified a more selective set of proteins (Figure 2C; Supplementary Table 2), more than two-thirds of which overlapped with those detected by conventional ChIRP-MS (Figure 2C). As an example of a suboptimal outcome, insufficient UV crosslinking resulted in reduced RNA enrichment and decreased pull-down efficiency during preliminary optimization experiments. The proteins shared between conventional ChIRP-MS and cPDiRT-MS consistently ranked among the most enriched candidates, suggesting that they represent core Tug1 interactors.

Figure 1: Schematic overview of the cPDiRT workflow in cells and tissues. GC2 cells or testicular tissues are subjected to UV crosslinking and sonication to fragment cellular material. Biotin-labeled antisense probes are hybridized to the target RNA and captured using streptavidin-coated paramagnetic beads. Following stringent washing, bound proteins are recovered for downstream mass spectrometry analysis. Target RNA enrichment is assessed separately by qRT-PCR. The workflow enables identification of proteins that directly interact with the target lncRNA in both cultured cells and tissue-derived samples. Please click here to view a larger version of this figure.

Figure 2: Identification of Tug1-associated proteins by ChIRP-MS and cPDiRT-MS in GC2 cells and mouse testes. (A) RNA enrichment of Tug1 (blue bars) and the negative control 36b4 (green bars) in GC2-ChIRP, GC2-cPDiRT, and testis-cPDiRT assays. Data are presented as mean ± SEM from two biological replicates. (B) Representative silver-stained gel images showing proteins recovered under different crosslinking conditions. From left to right: GC2 cells without and with PFA crosslinking, GC2 cells without and with UV crosslinking, and mouse testes without and with UV crosslinking. Marker lanes are indicated. Experiments were performed twice with similar results. (C) Venn diagrams showing overlap of proteins identified by mass spectrometry (≥2 unique peptides) in two independent biological replicates of GC2-ChIRP-MS (left) and GC2-cPDiRT-MS (middle). The right panel shows overlap between the top 100 proteins identified by GC2-ChIRP-MS (light purple) and proteins reproducibly detected in two independent GC2-cPDiRT-MS experiments (dark blue). (D) Venn diagram showing overlap of proteins identified by mass spectrometry (≥2 unique peptides) in two independent biological replicates of testis cPDiRT-MS. (E) Gene Ontology (GO) enrichment analysis of biological process terms associated with proteins identified in GC2-ChIRP (blue), GC2-cPDiRT (red), and testis-cPDiRT (orange) datasets. Please click here to view a larger version of this figure.

To define the in vivo protein interaction network of Tug1, cPDiRT was applied to testes from 4-week-old mice. Because UV light exhibits limited tissue penetration, testes were first dissociated into single-cell suspensions prior to UV crosslinking. Two independent biological replicates were performed for both GC2 cells and mouse testes. Mass spectrometry analysis showed that a good number of proteins identified with at least two unique peptides were shared between replicates (Figure 2D; Supplementary Table 2), indicating high reproducibility of cPDiRT-MS in both cultured cells and tissue samples. Gene Ontology (GO) analysis of reproducibly identified proteins from each dataset revealed biological processes broadly consistent with known functions of Tug1 (Figure 2E), supporting the reliability of the workflow.

Comparison of three experimental conditions—conventional ChIRP-MS in GC2 cells, cPDiRT-MS in GC2 cells, and cPDiRT-MS in mouse testes—identified 15 proteins that were significantly enriched across all datasets. These proteins were designated high-confidence candidate direct interactors of Tug1 (Figure 3A,B; Supplementary Table 2).

Figure 3: Identification and validation of common Tug1-binding proteins across three experimental datasets. (A) Venn diagram showing overlap of proteins identified in GC2-ChIRP-MS (green), GC2-cPDiRT-MS (dark blue), and testis-cPDiRT-MS (light blue). The central intersection represents 15 common Tug1-binding proteins. (B) Gene Ontology (GO) enrichment analysis of biological process terms associated with the 15 common Tug1-binding proteins. Representative proteins associated with each enriched term are listed in the accompanying table. (C) Enhanced crosslinking immunoprecipitation sequencing (eCLIP-seq) validation of Tug1 binding by PABPC1 and CCT3. Integrative Genomics Viewer (IGV) browser tracks show eCLIP-seq peaks across the Tug1 locus for two biological replicates (eCLIP-1 and eCLIP-2) and corresponding input controls. Signal tracks are normalized by reads per million (RPM). Gray boxes indicate exons, and arrows indicate transcriptional direction. Please click here to view a larger version of this figure.

To validate the screening strategy and binding specificity of selected candidates, PABPC1 and CCT3 were analyzed by enhanced crosslinking immunoprecipitation followed by deep sequencing (eCLIP-seq). These proteins were selected because they represent two major functional categories enriched among the 15 common interactors: translation regulation (PABPC1)²¹ and protein folding/chaperone activity (CCT3)²², both of which are highly relevant to spermatogenesis. Although additional candidates, including HNRNPU and DDX5, were identified, PABPC1 and CCT3 were prioritized because of their central biological functions and the availability of validated antibodies suitable for eCLIP experiments. Visualization of eCLIP-seq data revealed clear and specific binding peaks on the Tug1 transcript for both proteins, confirming direct interaction with Tug1 (Figure 3C).

Together, these representative results demonstrate that cPDiRT-MS enables the identification of lncRNA-binding proteins in both cultured cells and native tissues. The reproducibility observed across biological replicates, the overlap with conventional ChIRP-MS datasets, and the independent validation by eCLIP support the utility of cPDiRT-MS for investigating lncRNA-protein interactions in physiologically relevant contexts.

Supplementary Table 1: Tug1 probe sequences used for cPDiRT and ChIRP experiments. Sequences of biotinylated antisense DNA probes designed against the mouse Tug1 transcript. All probes were pooled at equimolar concentrations and used for hybridization-based capture of Tug1-associated ribonucleoprotein complexes.Please click here to download this file.

Supplementary Table 2: Mass spectrometry analysis of proteins identified by ChIRP-MS and cPDiRT-MS. Complete mass spectrometry results from GC2-ChIRP-MS, GC2-cPDiRT-MS, and testis-cPDiRT-MS experiments. The table includes protein identifiers, gene names, numbers of unique peptides, sequence coverage, identification scores, q-values, and iBAQ-based abundance measurements for experimental and control samples. High-confidence common Tug1-binding proteins identified across datasets are indicated.Please click here to download this file.

Supplementary Table 3: qPCR primer sequences used for RNA enrichment analysis. Sequences of primers used for qRT-PCR to evaluate enrichment of target and control transcripts following ChIRP and cPDiRT procedures.Please click here to download this file.

Data Availability Statement: 
The eCLIP-seq data of PABPC1 and CCT3 have been deposited in the Gene Expression Omnibus (GEO) database under accession no. GSE334383.

Discussion

cPDiRT was developed as a modified ChIRP-based protocol that replaces paraformaldehyde (PFA) crosslinking with UV crosslinking to selectively capture direct lncRNA-protein interactions. By combining enzymatic dissociation of fresh testicular tissue into single cells with UV crosslinking, direct-interaction mapping was extended from cultured cell lines to native tissues. Parallel experiments in cultured cells enabled a side-by-side comparison of PFA-based and UV-based approaches, and UV-derived interactions were independently validated by eCLIP-seq. These results demonstrate that UV crosslinking can be successfully integrated into the ChIRP workflow for tissue-derived samples while maintaining high specificity.

Few studies have applied PFA-based RNA-centric approaches such as ChIRP to tissue samples because of challenges associated with tissue penetration, intermolecular crosslinking artifacts, and elevated background arising from indirect associations. These limitations have restricted the broader application of tissue-based lncRNA interactome mapping and motivated the development of cPDiRT as a more specific alternative for tissue-level studies.

A critical feature of cPDiRT is its ability to distinguish direct from indirect interactions under near-physiological conditions. Conventional PFA-based methods stabilize large macromolecular complexes and therefore capture both direct and indirect associations. In contrast, UV crosslinking covalently links proteins only when they are in direct physical contact with the target RNA²³. Specificity is further enhanced by stringent high-salt washing conditions, which efficiently remove non-covalently associated molecules. Application of this strategy to Tug1 in mouse testes revealed a more selective interactome than would be obtained using conventional approaches. Because many lncRNA functions depend on specific molecular interactions, including recruitment of chromatin modifiers^24,25 and stabilization of ribonucleoprotein complexes²⁶, the ability to resolve direct interactions within native tissues represents an important methodological advance⁴.

Tug1 provides an informative model for evaluating this workflow. Genetic deletion of Tug1 results in complete male sterility, although the molecular basis of this phenotype remains incompletely understood¹⁴. Identification of Tug1-binding proteins from testicular tissue provides a framework for investigating the molecular effectors involved in spermatogenesis, Sertoli cell function, and blood-testis barrier maintenance^14,15. More broadly, successful application of cPDiRT to Tug1 supports the utility of this approach for studying other tissue-enriched lncRNAs.

The amount of starting material required for cPDiRT (approximately 500 mg of cell pellet or an equivalent amount of testicular single-cell suspension) is comparable to that required for conventional PFA-based ChIRP-MS. Although the overall efficiencies of PFA and UV crosslinking are difficult to compare directly, UV crosslinking enriches direct interactions and thereby improves the signal-to-noise ratio by reducing background interference. The observed differences between ChIRP-MS and cPDiRT-MS are therefore more likely attributable to increased specificity than to improved crosslinking efficiency. Conventional ChIRP captures both direct and indirect interactions, whereas cPDiRT uses stringent washing conditions to enrich for direct interactors. As a result, low-abundance direct interactors that may be obscured by abundant indirect interactors in conventional ChIRP-MS can be more readily detected. This likely explains the identification of Tug1-associated proteins unique to cPDiRT.

Several steps are particularly important for successful implementation of the protocol. Low RNA recovery after pull-down often indicates RNA degradation or insufficient probe hybridization. Freshly prepared hybridization buffer can help improve RNA recovery. Probe hybridization failure is frequently associated with suboptimal probe design; therefore, probes should be designed across the full-length target transcript and evaluated carefully before use. Weak UV-crosslinking signals may be improved by increasing UV exposure or by performing a second round of crosslinking.

Several limitations should also be considered. First, enzymatic dissociation into single cells, while necessary for efficient UV penetration, partially disrupts seminiferous tubule architecture and cell-cell interactions²⁷. Consequently, interactions that depend on intact tissue organization may be underrepresented. Second, UV-crosslinking efficiency depends on RNA secondary structure and the proximity of protein-RNA contacts²⁸. Transient, low-affinity, or structurally shielded interactions may not crosslink efficiently and therefore may not be detected. Similarly, proteins that interact with highly structured or inaccessible RNA regions may be missed²⁹. Third, the current workflow was optimized for mouse testis and may require empirical optimization for application to other tissues or species. Single-cell dissociation may be particularly challenging in dense or fibrotic tissues and may require modified enzymatic or mechanical dissociation strategies.

Despite these limitations, cPDiRT provides a practical and broadly applicable platform for investigating direct lncRNA-protein interactions in complex tissues. Future applications integrating cPDiRT with quantitative mass spectrometry, single-cell proteomics, or functional validation could facilitate the construction of comprehensive lncRNA-protein networks in vivo. Such studies will contribute to understanding how lncRNAs regulate developmental processes and disease-associated pathways.

In summary, cPDiRT enables tissue-level identification of proteins that directly interact with lncRNAs and helps bridge the gap between transcriptomic discovery and mechanistic investigation. Application of cPDiRT to Tug1 demonstrates its utility in reproductive biology and establishes a methodological framework for studying lncRNA-protein interactions in native tissue environments.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.25% Trypsin-EDTA Solution	Gibco	25200056	for tissue dissociation and single-cell suspension preparation
100% Ethanol	Sigma-Aldrich	459844	for RNA cleanup
AffinityScript Reverse Transcriptase	Agilent	600107	for eCLIP cDNA synthesis
Anti-CCT3 Antibody	Abcam	Ab225878	antibody for eCLIP
Anti-PABPC1 Antibody	Abcam	Ab312314	antibody for eCLIP
Chloroform	Sigma-Aldrich	C2432	for RNA extraction
Collagenase Type IV	Sigma-Aldrich	C5138	for enzymatic dissociation of testicular tissue
Dephosphorylation reagents (T4 PNK)	New England Biolabs	M0201S	for eCLIP RNA dephosphorylation
DMEM High Glucose Medium	Gibco	11965092	for cell culture and tissue dissociation buffer
Dynabeads MyOne Streptavidin C1 Magnetic Beads	Thermo Fisher Scientific	65001	for capture of biotinylated probe-RNA-protein complexes
DynaMag-15 Magnet	Thermo Fisher Scientific	12301D	for magnetic bead separation
DynaMag-2 Magnet	Thermo Fisher Scientific	12321D	for magnetic bead separation
EDTA	Sigma-Aldrich	E9884	for chelating divalent cations in lysis and wash buffers
Exonuclease I	New England Biolabs	M0293S	for eCLIP cDNA purification
Fetal Bovine Serum (FBS)	Gibco	10099141C	for stopping trypsin digestion
Formamide	Thermo Fisher Scientific	AM9342	for hybridization buffer preparation
Hybridization Oven	TUOHE	LF-I	for probe hybridization and beads washing
Low-Melting-Temperature Agarose	Sigma-Aldrich	A9414	for eCLIP library size selection
MinElute Gel Extraction Kit	Qiagen	28604	for eCLIP library size selection
miRNeasy Mini Kit	Qiagen	217004	for RNA purification
NaCl	Sigma-Aldrich	S7653	for preparing high-salt wash buffer
Nitrocellulose Membrane	GE Healthcare / Amersham	10600002	for eCLIP western blot transfer
NP-40	Thermo Fisher Scientific	85125	for wash buffer detergent
Nuclease- and protease-free water	Thermo Fisher Scientific	10977035	for preparing nuclease-free solutions
NuPAGE 4-12% Bis-Tris Gel	Thermo Fisher Scientific	NP0322BOX	for eCLIP size selection
NuPAGE Bis-Tris Precast Gels	Thermo Fisher Scientific	NP0321BOX	for SDS-PAGE protein separation
PBS (pH 7.4)	Thermo Fisher Scientific	10010023	for cell washing and resuspension
Phenylmethylsulfonyl fluoride (PMSF)	Sigma-Aldrich	P7626	serine protease inhibitor for cell lysis buffer; added fresh before use
Protease Inhibitor Cocktail	Roche	4693132001	broad-spectrum protease inhibitor for cell lysis buffer; added fresh before use
Proteinase K	Sigma-Aldrich	P4850	for protein digestion in RNA extraction
qRT-PCR Master Mix	Thermo Fisher Scientific	11704044	for quantitative RT-PCR to confirm RNA enrichment
RNase/Protease-Free DNase I Solution	Thermo Fisher Scientific	EN0521	for tissue dissociation
RNeasy Mini Kit	Qiagen	74104	for RNA purification
SDS-PAGE Sample Loading Buffer	Beyotime	P0286	for protein sample preparation
Shrimp Alkaline Phosphatase	New England Biolabs	M0371S	for eCLIP cDNA purification
Silver Stain Kit	Beyotime	P00175	for visualizing proteins in SDS-PAGE gels
Sodium Deoxycholate	Sigma-Aldrich	D6750	for wash buffer detergent
Sodium dodecyl sulfate (SDS)	Sigma-Aldrich	L3771	for cell lysis and protein denaturation
Superase-in RNase Inhibitor	Thermo Fisher Scientific	AM2694	specialized RNase inhibitor; added fresh before use
T4 RNA Ligase	New England Biolabs	M0204	for eCLIP adapter ligation
Tris–HCl (pH 7.0)	Sigma-Aldrich	T5941	for preparing lysis, hybridization, and wash buffers
TRIzol Reagent	Thermo Fisher Scientific	15596026CN	for RNA extraction
Ultrasonic cell crusher	ATPIO	ATPIO-650D	for cell lysis and DNA fragmentation
Ultraviolet Crosslinker	Analytik Jena	UVP CL-1000	for UV crosslinking of RNA-protein interactions
urea	Sigma-Aldrich	U5128	for eCLIP protein degradation on membrane