CLIP Sequencing for RNA–Protein Interactions: Methods, Comparisons, and Use Cases

Protein–RNA Interactions: Advances Through CLIP and Sequencing Technologies

Protein–RNA interactions are now recognised as essential regulators of gene expression. A major breakthrough came with the development of crosslinking and immunoprecipitation (CLIP), which uses ultraviolet irradiation to covalently link RNA-binding proteins (RBPs) to their RNA partners. Initially, the approach was limited because only a small number of RNA-binding sites could be identified.

The advent of next-generation sequencing changed the field. High-throughput sequencing of crosslinked protein–RNA fragments now allows researchers to profile these interactions comprehensively and without bias. Various CLIP-based methods have since been applied in cultured cells and animal models, generating important insights into how RBPs influence RNA processing and stability.

These advances have made it possible to systematically identify functional binding sites for a wide spectrum of RBPs. As experimental protocols and bioinformatics pipelines improve, the resolution and interpretability of interaction maps continue to increase. When combined with transcriptome-wide data on RNA stability, chemical modifications, structural conformations, and protein expression, high-resolution maps of protein–RNA contacts provide a powerful framework for dissecting post-transcriptional and translational control mechanisms.

clip sequencing overview

Introduction

Regulation of gene expression occurs throughout the life cycle of an mRNA, from transcription to protein translation. This process is guided by RNA-binding proteins (RBPs) and non-coding RNAs, particularly microRNAs. These molecules recognise cis-regulatory elements on both pre-mRNA and mature transcripts, assembling into ribonucleoprotein (RNP) complexes. The architecture and dynamics of RNPs determine when and where specific RNAs accumulate in cells.

RBPs further shape transcript diversity by controlling alternative splicing and 3′ end processing, thereby influencing which isoforms are produced. RNA was once viewed as little more than a template for protein synthesis. It is now clear, however, that RNP-mediated regulation directly impacts both the diversity and abundance of protein isoforms. Because of the vast range of RNA species and their widespread protein interactions, RNA itself represents a major source of biological variability.

To explore RNA–protein interactions, researchers often combine genetics, biochemical assays, and bioinformatics. Experimental methods include in vitro selection approaches and RNA immunoprecipitation (RIP) from cells or tissue extracts. In the RIP–CHIP method, microarrays are used to detect RNAs co-precipitated with specific RBPs. While useful, RIP–CHIP mainly captures stable protein–RNA complexes and provides limited resolution. Complementary methods are needed to detect transient interactions, and in some cases, RBP binding sites can only be inferred computationally.

Over the last decade, innovative experimental strategies have transformed this field. New technologies now allow mapping of RBP binding sites at near-nucleotide resolution, producing high-resolution, transcriptome-wide interaction maps. These approaches depend on next-generation sequencing (NGS), which has made it possible to sequence billions of nucleotides rapidly and at a manageable cost. In this review, we highlight how high-throughput sequencing is driving new applications in the study of protein–RNA interactions.

High-Throughput Sequencing in RNA Analysis

Over the past decade, sequencing technologies have advanced at an extraordinary pace. Next-generation sequencing (NGS) now enables millions of reads to be generated in parallel within a single run. Platforms such as 454 (Roche), HiSeq (Illumina), and SOLiD (Life Technologies) support transcriptome sequencing at a scale previously impossible. Today, a single experiment can yield close to one billion reads of around 100 nucleotides each.

RNA sequencing (RNA-seq), a specialised application of NGS, has become a standard tool for transcriptome profiling. Unlike microarrays, RNA-seq delivers higher resolution and a broader dynamic range, making it possible to quantify complex cellular RNA populations with accuracy. Importantly, RNA-seq is not restricted to known transcripts; it can be applied to any species of interest, enabling discovery beyond annotated genomes.

With RNA-seq, researchers can:

• Generate a complete map of the transcriptome.
• Assess alternative splicing, RNA editing, and differential 3′ end processing.
• Identify novel RNA molecules, including non-coding RNAs with regulatory functions.

Selective biochemical purification before sequencing further improves focus on particular RNA groups. This enrichment has been widely used to study RNAs bound to RNA-binding proteins (RBPs). In these cases, crosslinking and immunoprecipitation (CLIP) methods are employed to capture protein–RNA complexes. More recent CLIP adaptations—such as HITS-CLIP, PAR-CLIP, eCLIP and iCLIP—improve crosslinking efficiency and refine the mapping of binding sites, enabling researchers to probe RNA–protein interactions at near-nucleotide resolution.

Quick Comparison of CLIP-Derived Techniques

Use this table to pick the right assay for your RBP question.

Notes for readers

• "Single-nt" indicates the crosslinked nucleotide can be assigned directly.
• Choose PAR-CLIP for highest crosslink efficiency and diagnostic mutations.
• Choose iCLIP when input is limiting or thiolated nucleosides are impractical.
• Use CRAC for denaturing purification and very clean datasets, especially in yeast.
• Select eCLIP when reproducibility, scalability, and input controls are priorities, particularly for large-scale RBP profiling (e.g., ENCODE).

Please read our RNA CLIP-Seq: Principle, Advantages, Protocol, and Applications for more information about CLIP-Seq.

Cross-Linking and Immunoprecipitation (CLIP): A Turning Point in RNA Biology

Cross-linking and immunoprecipitation (CLIP) marked a major breakthrough in RNA biology when it was first introduced in 2003 by Robert Darnell's laboratory. While ultraviolet (UV) light had been used since the 1980s to induce protein–RNA crosslinks in live cells, CLIP provided the first robust framework to purify and characterise protein–RNA complexes inside tissues. Unlike chromatin immunoprecipitation (ChIP), which commonly relies on formaldehyde crosslinking for protein–DNA studies, CLIP employs 254 nm UV irradiation to generate covalent bonds between endogenous RNA-binding proteins (RBPs) and their RNA targets.

This covalent bond formation allows stringent immunoprecipitation of protein–RNA complexes with antibodies recognising native RBPs or epitope tags. The method was first applied to study Nova1, a splicing regulator. After UV treatment of brain homogenates, Nova1–RNA complexes were purified, trimmed with RNase to fragments of 20–100 nucleotides, radiolabelled, and size-fractionated on SDS-PAGE. Following protease digestion, the RNA was recovered, adaptor-ligated, reverse-transcribed, and amplified by PCR before sequencing. Sanger sequencing of cDNA libraries yielded 340 Nova1-associated tags, including 18 flanking exons. These findings were validated in Nova1 knockout mice, which showed mis-splicing of seven exons. Further analysis revealed the YCAY motif as Nova1's binding site, highlighting its role in coordinating splicing of transcripts encoding inhibitory synaptic components.

Since its inception, CLIP has been extended to multiple RBPs across diverse systems:

• SFRS1 (splicing factor): shown to regulate distinct transcripts depending on cellular compartment.
• HNRNPA1: demonstrated as essential for the biogenesis of microRNA miR-18a.
• Rrm4 (fungus Ustilago maydis): provided proof that CLIP could be applied beyond mammalian models.
• Cugbp1 (mouse brain): revealed RNA gain-of-function effects linked to spinocerebellar ataxia type 8.
• Msy2: identified interactions with piRNAs, localised to chromatin, cytoplasmic RNPs, and polysomes.
• DJ-1 (linked to Parkinson's disease) and RARα (retinoic acid receptor): both confirmed as RNA-binding proteins through CLIP.

Collectively, these applications demonstrate CLIP's versatility in mapping RBP targets and its central role in uncovering how RNA–protein interactions shape gene regulation in health and disease.

Why CLIP Stands Out

One of the key strengths of CLIP is its flexibility. Ultraviolet (UV) crosslinking can be applied across a wide spectrum of biological materials, from cultured cells to homogenised mouse tissues. This adaptability has made CLIP a valuable tool for studying RNA–protein interactions in diverse experimental systems.

The exact molecular mechanism of protein–RNA crosslinking remains under investigation. Current evidence suggests that UV light excites nucleic acid bases to a higher electronic state, which promotes the formation of new covalent bonds with nearby amino acid residues. This reaction locks proteins to their RNA targets, allowing rigorous purification and downstream sequencing.

By providing a covalent "snapshot" of RNA–protein contacts in living systems, CLIP enables researchers to capture dynamic regulatory interactions that would otherwise be lost during biochemical purification. This technical advantage has established CLIP as a cornerstone in modern studies of RNA biology.

High-Throughput Sequencing of RNA Isolated by CLIP (HITS-CLIP)

Early CLIP studies relied on Sanger sequencing, which generated only a limited number of cDNA reads. This low yield restricted the full potential of CLIP-based approaches. The advent of high-throughput sequencing transformed the field, enabling unbiased and comprehensive profiling of crosslinked RNA fragments bound to specific RNA-binding proteins (RBPs).

In the first demonstration of HITS-CLIP (also called CLIP-seq), Darnell and colleagues applied next-generation sequencing to study Nova2, a neuron-specific splicing factor. Unlike Nova1, which is more restricted in expression, Nova2 is broadly distributed but regulates many similar targets. Using the 454/Roche pyrosequencing platform, the team obtained ~412,000 sequence tags—over 1,000 times more than the earlier Nova1 Sanger CLIP dataset. Roughly 41% of these reads mapped uniquely to the mouse genome, largely within protein-coding genes. The tags clustered into 19,000 sites enriched for Nova2 consensus motifs, yielding a transcriptome-wide map of Nova2 interactions and confirming its role in splicing regulation. A notable finding was the enrichment of Nova2 tags in 3′ UTRs, suggesting an unexpected role in regulating alternative polyadenylation.

HITS-CLIP has also illuminated Argonaute (Ago)/microRNA complexes. In mouse brain experiments, Ago CLIP captured 454 distinct microRNAs and mapped Ago-binding sites across 829 transcripts. Many of these sites were complementary to brain-specific miR-124, validating their biological significance. Follow-up studies with an Ago2-specific antibody compared wild-type and Dicer-knockout embryonic stem cells, distinguishing between microRNA-dependent and independent Ago2 binding sites.

Applications of HITS-CLIP span a wide range of RBPs:

PTB (polypyrimidine tract binding protein): Revealed CU-rich hexamers and strategies for exon inclusion or skipping.

FOX2: Demonstrated position-dependent regulation of exon inclusion or exclusion.

SFRS1: Shown to modulate a diverse set of transcripts with broad effects on gene expression.

Khd1 in yeast: Identified targets using TAP-tag purification, uncovering regulation of FLO11 asymmetry and cell fate during filamentous growth.

A persistent challenge in CLIP experiments is pinpointing the exact crosslinked nucleotide. HITS-CLIP data have shown that mutations—insertions, deletions, or substitutions—often appear at crosslinking sites. More than two-thirds of reads contain such "hidden" mutations relative to reference genomes, providing a powerful signature for accurately identifying nucleotide-level crosslink sites.

Enhanced CLIP (eCLIP): Streamlined and Reproducible Protein–RNA Mapping

Enhanced CLIP (eCLIP) was developed to address some of the technical challenges of earlier CLIP protocols. By optimising RNA recovery and simplifying library preparation, eCLIP reduces experimental noise and increases reproducibility across biological replicates. One of its key innovations is the inclusion of size-matched input controls, which improve background correction and allow more confident identification of true binding sites. eCLIP retains the ability to map RBP binding at near-nucleotide resolution, but with higher throughput and better scalability, making it the current standard in several large consortium projects such as ENCODE.

Key advantages of eCLIP include:

• Simplified workflow with reduced library preparation bias.
• Incorporation of input controls for improved background subtraction.
• Increased reproducibility and scalability across many RBPs.
• Compatibility with large-scale RBP profiling efforts.

Cross-Linking and Analysis of cDNA (CRAC): Refining Protein–RNA Interaction Mapping

CRAC (Cross-linking and Analysis of cDNA) was developed by David Tollervey's group to overcome some limitations of conventional CLIP. Unlike antibody-based immunoprecipitation under native conditions, CRAC employs affinity resins that do not depend on peptide–protein interactions. This allows purification under denaturing conditions, enabling stringent washes to remove nonspecific material and yielding exceptionally clean protein–RNA datasets.

In yeast, CRAC was applied to study snoRNP components such as Nop1, Nop56, Nop58, and Rrp9. Tandem affinity purification (TAP)-tagged proteins were first recovered on IgG resin, eluted with TEV protease, denatured in guanidine hydrochloride, and rebound to nickel resin for further purification. After limited RNase digestion, recovered RNA fragments were ligated to adaptors, reverse-transcribed, and amplified as cDNA. This workflow produced a very high signal-to-noise ratio and pinpointed precise binding sites of snoRNP proteins on U3 RNA, offering new insights into U3 snoRNP architecture.

CRAC has since been extended to broader questions in RNA biology:

• Ribosome biogenesis: Binding sites of the helicase Prp43 on 18S and 25S rRNA precursors revealed how enzymatic RNA remodelers interact with pre-rRNA.
• Pre-40S assembly factors: Six factors (Dim2/Pno1, Enp1, Ltv1, Nob1, Rio2, and Tsr1) were localised on pre-18S rRNA, mapping their positions within pre-40S particles.
• Nuclear RNA surveillance: CRAC uncovered widespread targets of the TRAMP complex (Trf4/5–Air1/2–Mtr4) and the Nrd1–Nab3 dimer, unexpectedly revealing many novel mRNA substrates and suggesting that nuclear RNA turnover is far more active than previously thought.
• Exonuclease Rat1: Recent studies identified Rat1 binding sites at the 5′ ends of pre-rRNAs, clarifying its essential role in ribosome maturation.

By combining denaturing purification with high-throughput sequencing, CRAC provides both resolution and specificity, making it a powerful complement to other CLIP-derived approaches in mapping protein–RNA interactions.

Photoactivatable-Ribonucleoside-Enhanced CLIP (PAR-CLIP): Improving Crosslinking Efficiency and Resolution

Most CLIP methods rely on 254 nm UV irradiation to induce protein–RNA crosslinks. This process has relatively poor efficiency, with yields often estimated at only 1–5% using purified proteins and radiolabelled RNA. To address this limitation, Thomas Tuschl and colleagues developed PAR-CLIP, a technique that uses photoactivatable nucleoside analogues to enhance crosslinking.

Cells are metabolically labelled with 4-thiouridine (4-SU) or 6-thioguanosine (6-SG), which are readily incorporated into newly transcribed RNAs without detectable toxicity. Protein–RNA crosslinking is then induced in live cells with 365 nm UV light. For the RNA-binding protein IGF2BP1, for example, this approach produced far more efficient crosslinking than conventional 254 nm UV exposure.

The downstream workflow mirrors standard CLIP: immunoprecipitation of RBP–RNA complexes, partial RNase digestion, adaptor ligation, reverse transcription, PCR amplification, and sequencing on Illumina platforms. A distinctive feature of PAR-CLIP is the appearance of diagnostic nucleotide conversions during reverse transcription—T-to-C (from 4-SU) or G-to-A (from 6-SG). These mutations mark true crosslinked nucleotides, allowing precise identification of binding sites. Sequence clusters representing candidate sites can also be ranked by the number of diagnostic substitutions.

Applications of PAR-CLIP

• In the first PAR-CLIP studies, binding motifs and regulatory sequences were mapped for several well-characterised human RBPs, including PUM2, QKI, IGF2BP1–3, Argonaute/EIF2C1–4, and TNRC6A–C. The results confirmed known preferences for exonic, intronic, coding, and noncoding regions, while adding new binding motifs.
• Independent reports of HuR/ELAVL1 revealed more than 20,000 binding sites, most located in 3′ UTRs. HuR was shown to stabilise its target mRNAs, regulate splicing, and repress miR-7 activity.
• In yeast, PAR-CLIP was adapted for strains grown in media containing 4-SU. RBPs such as Nrd1, Nab3, and the helicase Sen1 were found to bind unexpected noncoding RNAs and tRNA transcripts, with stress-dependent changes in binding.
• PAR-CLIP also defined the sequence motif of PAPD5, a noncanonical poly(A) polymerase, demonstrating activity across diverse RNA substrates.

Comparison with HITS-CLIP

Quantitative comparisons of PAR-CLIP and HITS-CLIP showed only subtle differences in binding site recovery for human HuR and Ago2. However, extensive RNase digestion was found to bias site identification. Interestingly, U-rich HuR binding sites were not more enriched in PAR-CLIP compared with HITS-CLIP, despite the use of 4-SU. Later studies confirmed that 4-SU and 6-SG PAR-CLIP produce highly similar consensus motifs and reproducible binding profiles.

Key Advantages of PAR-CLIP

• Enables nucleotide-resolution mapping of binding sites.
• Uses diagnostic mutations (T→C, G→A) to discriminate true crosslinked positions.
• Provides quantitative ranking of candidate binding clusters.

These innovations make PAR-CLIP a powerful complement to HITS-CLIP, particularly when high crosslinking efficiency and precise binding site identification are essential.

Individual-Nucleotide Resolution UV Cross-Linking and Immunoprecipitation (iCLIP)

Individual-nucleotide resolution CLIP, or iCLIP, represents one of the most refined versions of CLIP technology. Developed by Jernej Ule's laboratory, iCLIP provides single-nucleotide resolution for identifying protein–RNA crosslinking sites. Like PAR-CLIP, this approach pinpoints precise binding events, but it relies on a different molecular strategy.

After UV irradiation, covalently crosslinked RNA–protein complexes are immunoprecipitated. RNA fragments are ligated to adaptors at the 3′ end. Digestion with proteinase K leaves a short peptide attached to the RNA, which causes reverse transcription (RT) to terminate prematurely at the crosslink site. RT is performed with an oligonucleotide primer carrying adaptors and a barcode sequence. The use of a splittable adaptor allows circularised cDNA to be re-linearised for PCR amplification and sequencing.

In sequencing data, crosslinking sites can be identified directly. The first nucleotide following the barcode corresponds to the position where RT stopped, thereby marking the crosslinked residue. The barcoding system further enables the distinction of unique RNA molecules, as tags derived from the same cDNA template share both an identical barcode and truncation point.

Applications of iCLIP

• HNRNPC mapping: iCLIP revealed that HNRNPC preferentially binds uridine-rich sequences and demonstrated how tetramer positioning influences exon inclusion.
• Splicing regulators TIA1 and TIAL1: High-resolution maps showed their binding near upstream 5′ splice sites, where they modulate distal 3′ splice site choice. These results highlighted the importance of kinetics in regulating alternative splicing.

Overall, iCLIP offers nucleotide-level precision similar to HITS-CLIP and PAR-CLIP. Its efficient cDNA circularisation method reduces the amount of starting material required, making it particularly valuable for low-abundance RBPs or limited biological samples.

Schematic summary of CLIP and iCLIP methods.

Future Perspectives

Next-generation sequencing (NGS) has dramatically accelerated the study of RNA–protein interactions. The development of diverse CLIP variants now enables researchers to choose the most suitable method for probing specific RBPs in different biological systems. Importantly, these approaches extend beyond stable interactions, allowing the investigation of transiently bound enzymatic proteins such as polymerases, helicases, nucleases, and RNA-modifying enzymes. This capacity will continue to broaden the scope of protein–RNA interaction research.

Among the available techniques, PAR-CLIP offers particular advantages. Although it requires the addition of thiolated nucleosides before crosslinking, it generates diagnostic mutations at crosslinked sites. These sequence changes not only pinpoint binding positions but also allow ranking of candidate sites by confidence. Tissue-specific expression of uracil phosphoribosyltransferase (UPRT) further enhances this method, enabling cell-type–specific analysis of protein–RNA interactions. In addition, metabolic incorporation of modified nucleosides allows labelling of newly transcribed RNAs, making it possible to monitor dynamic RNP assembly during stress responses, differentiation, self-renewal, and programmed cell death.

HITS-CLIP and iCLIP have also proven robust across large biological sample sets. All three methods—HITS-CLIP, iCLIP, and PAR-CLIP—can achieve near-nucleotide resolution. Although technically demanding, the efficient cDNA circularisation step in iCLIP reduces input requirements, making it attractive for low-yield samples.

Open Access Databases

To manage the flood of published CLIP data, several open-access databases have been established:

CLIPZ: provides experimentally defined RBP binding sites with an integrated analysis environment.

starBASE: allows inference of microRNA–mRNA interactions from Ago HITS-CLIP datasets.

doRiNA (http://dorina.mdc-berlin.de): stores transcriptome-wide binding site data for RBPs and microRNAs, offering combinatorial query searches.

Challenges and Opportunities Ahead

The key challenge for the coming years will be integrating high-resolution protein–RNA interaction maps with other large-scale datasets. This will clarify the functional and dynamic impact of protein–RNA networks at a systems level. Emerging high-throughput methods now enable researchers to evaluate how RNPs influence mRNA abundance and protein expression. NGS-based strategies also support transcriptome-wide probing of RNA secondary structures.

In summary, the fusion of CLIP technologies with NGS provides unprecedented depth for studying protein–RNA interactions. These advances are already yielding unexpected biological insights. Continued progress will enrich our understanding of RNPs in cellular processes, ultimately illuminating how post-transcriptional regulation controls growth, differentiation, and disease.

Q&As

Q: What are the main differences between HITS-CLIP, PAR-CLIP, and iCLIP?

HITS-CLIP uses 254 nm UV light and high-throughput sequencing, offering transcriptome-wide RBP maps but moderate resolution; PAR-CLIP adds photoreactive nucleosides (4-SU/6-SG) and 365 nm UV, improving crosslink efficiency and yielding T→C or G→A transitions for single-nucleotide accuracy; iCLIP captures cDNA truncations at crosslink sites, adds UMIs for quantification, and achieves nucleotide resolution even with low-input material.

Q: When should I choose PAR-CLIP over HITS-CLIP?

Choose PAR-CLIP when you need higher crosslinking efficiency and precise binding sites via diagnostic mutations (T→C/G→A), especially in cultured cells capable of incorporating 4-SU or 6-SG—even though these must be metabolically incorporated and may have cellular limitations.

Q: Can I use iCLIP when sample material is limited?

Yes. iCLIP's cDNA circularisation and use of UMIs enhance sensitivity and reduce required input material, making it well-suited for scarce or precious samples.

Q: How does CLIP-seq compare to RIP-seq for studying RNA-protein interactions?

CLIP-seq uses UV crosslinking before immunoprecipitation to capture direct, in vivo protein–RNA interactions with positional context, whereas RIP-seq lacks crosslinking and may include indirect or reassociated interactions. CLIP-seq is technically more demanding but yields higher specificity and resolution.

References:

1. Van Nostrand, E.L., Pratt, G.A., Yee, B.A. et al. Principles of RNA processing from analysis of enhanced CLIP maps for 150 RNA binding proteins. Genome Biol 21, 90 (2020).
2. Van Nostrand, Eric L., Gabriel A. Pratt, Alexander A. Shishkin, Chelsea Gelboin-Burkhart, Mark Y. Fang, Balaji Sundararaman, Steven M. Blue, et al. 2016. Robust Transcriptome-wide Discovery of RNA-binding Protein Binding Sites with Enhanced CLIP (eCLIP). Nature Methods 13 (6): 508–14.