RIP answers a deceptively simple question: which RNAs are bound by a given protein at the moment you harvest? The naive version of the experiment — lyse, immunoprecipitate, sequence — is easy to run and easy to misinterpret, because abundant transcripts stick nonspecifically to beads and antibodies. The variants that have survived critical scrutiny (CLIP, CRAC, iCLIP, eCLIP, PAR-CLIP) all share one idea: force the interaction to become a covalent bond, then make that bond tell you where it formed. This post covers both the accessible RIP protocol and the higher-resolution crosslinking approaches, and where the boundary between them matters.
Crosslinking: UV versus formaldehyde
Crosslinking is the difference between an enrichment experiment and a binding-site experiment. Without it, the washing stringency needed to remove background also strips genuine interactions.
- UV (254 nm) forms zero-length covalent bonds between RNA bases and aromatic amino acids — essentially phenylalanine, tyrosine, tryptophan, and cysteine. Because the crosslink is zero-length, a UV crosslink nearly guarantees the protein was in direct contact, and the crosslinked peptide maps to the binding interface. Yield is low, and photochemistry depends on the geometry of the interacting residues.
- Formaldehyde is a longer-range crosslinker that bridges through protein–protein and protein–nucleic acid contacts. It works well for capturing indirect interactions and for RNA associated with chromatin, which is why RNA-ChIP and ChIRP-type methods lean on it, but it cannot distinguish direct from indirect binding.
The CLIP family is built on UV. If your goal is to claim a direct binding site, use UV. If your goal is to show an RNA is associated with a chromatin-associated complex, formaldehyde (often followed by sonication) is defensible, provided you say so.
Lysis conditions and the logic of limited RNase digestion
The lysis buffer must do three jobs: keep the protein–RNA complex soluble, inhibit phosphatases and proteases, and dissociate aggregates. For UV-CLIP work, denaturing buffers with SDS and heat inactivate endogenous nucleases and unfold the RNA, so the subsequent immunoprecipitation reports a covalently trapped, sequence-independent target. Standard RIP buffers (Tris pH 7.4–8, 100–150 mM salt, 0.5% NP-40 or Triton, with RNase inhibitor) preserve native complexes but require careful washing to control background.
The single most important parameter in CLIP is the RNase concentration. The RNA is partially digested so that each immunoprecipitated protein carries a short RNA tag — long enough to map, short enough to be a footprint. Under-digestion yields large RNA fragments that map to whole transcripts; over-digestion erases the footprint entirely. This is why the classic CLIP protocol calls for two-tiered RNase treatment and why the HITS-CLIP paper spent so much effort on the digestion ladder. The correct outcome is a smeared, distinctly lower-molecular-weight RNA band on a radiolabelled gel, not a clean high-molecular-weight signal.
From HITS-CLIP to eCLIP and PAR-CLIP
HITS-CLIP (originally CLIP-seq) combined UV crosslinking with deep sequencing to produce transcriptome-wide RNA maps of an RNA-binding protein. It set the field's standards. Three successors addressed its weaknesses:
- PAR-CLIP uses the photoactivatable nucleoside analogue 4-thiouridine (4SU) to boost crosslinking efficiency, with the bonus that the 4SU residue is read as a transition during sequencing, marking the exact crosslink site. It has very high resolution but requires 4SU incorporation into cells, which perturbs RNA metabolism and is impractical for tissue.
- iCLIP exploits the tendency of reverse transcriptase to stop at the crosslinked residue: after circular ligation of the cDNA, the truncation site is captured as the first nucleotide of the read. This gives base-precise binding sites, with a stronger quantitative signal for binding position than simple read density.
- eCLIP engineered the library chemistry to deliver comparable resolution with far less starting material, using random barcoded adapters so that control and immunoprecipitate libraries can be sequenced together. It is now the most reproducible and scalable of the family.
A regular RIP-seq experiment has no crosslink, so read density reflects RNA abundance, not binding. If you run RIP-seq, pool it with RNA-seq input and interpret relative to that input rather than against absolute transcript levels.
Antibody controls determine whether the result means anything
The immunoprecipitation step carries all the specificity risk. Three controls are non-negotiable:
- An isotype-matched, irrelevant antibody at the same concentration, processed identically — this absorbs bead and antibody background.
- Input and mock (no-antibody) lanes, to estimate the input complexity and the nonspecific RNA fraction.
- A known target and a known non-target transcript measured by RT-qPCR before sequencing, so that enrichment can be demonstrated by an independent readout.
For tagged proteins, an anti-tag antibody against a tagged but unbound control protein is stronger than an isotype control. For eCLIP, the field has largely converged on matched-size control libraries generated in parallel from the same lysate, because the difference in library complexity is itself a signal that needs a baseline.
Sequencing, peak calling, and motif analysis
Once libraries are sequenced, the standard pipeline is: trim adapters, align to the genome, call crosslink-centered peaks against input, annotate to transcript regions (5′UTR, CDS, 3′UTR, intron), and then run motif enrichment on the peaks. The resulting motif is the most informative and least appreciated output — an RRM protein typically shows a short, degenerate purine- or pyrimidine-rich motif, and its position within the peak reports where in the binding site the protein sits.
Two pitfalls recur. First, peak callers tuned for ChIP treat CLIP data as if reads were uniformly distributed rather than accumulating at a crosslink nucleotide, which inflates noise; use a CLIP-aware caller or an iCLIP-style crosslink-site extraction. Second, evolutionary conservation of the bound site is weaker than for transcription-factor binding, so conservation should be a supporting argument, not the primary evidence. Finally, functional follow-up — a reporter with the predicted site mutated, or an orthogonal binding assay — remains the only way to promote a predicted site to a real one.
References
- Ule J, Jensen KB, Ruggiu M, et al. CLIP identifies Nova-regulated RNA networks in the brain. Science. 2003;302(5648):1212–1215.
- Licatalosi DD, Mele A, Fak JJ, et al. HITS-CLIP yields genome-wide insights into brain alternative RNA processing. Nature. 2008;456(7221):464–469.
- Hafner M, Landthaler M, Burger L, et al. Transcriptome-wide identification of RNA-binding protein and microRNA target sites by PAR-CLIP. Cell. 2010;141(1):129–141.
- König J, Zarnack K, Rot G, et al. iCLIP reveals the function of hnRNP particles in splicing at individual nucleotide resolution. Nat Struct Mol Biol. 2010;17(7):909–915.
- Van Nostrand EL, Pratt GA, Shishkin AA, et al. Robust transcriptome-wide discovery of RNA-binding protein binding sites with enhanced CLIP (eCLIP). Nat Methods. 2016;13(6):508–514.
- Zarnegar BJ, Flynn RA, Shen Y, et al. irCLIP platform for efficient characterization of protein-RNA interactions. Nat Methods. 2016;13(6):489–492.
- Tenzer S, Moro A, Kuharev J, et al. Proteome-wide characterization of the RNA-binding protein RALY-interactome using the in vivo-Biotinylation-Pulldown-Quant (iBioPQ) approach. J Proteome Res. 2013;12(6):2869–2884.
- Milligan L, Huynh-Thu VA, Sergeant F, et al. Strand-specific, high-resolution mapping of modified RNA polymerase II. Mol Syst Biol. 2016;12(6):874.
- Hafner M, Katsantoni M, Köster T, et al. CLIP and complementary methods. Nat Rev Methods Primers. 2021;1:20.
- Darnell RB. HITS-CLIP, CLIP-seq and CLIP-based approaches: a window into RNA biology. Wiley Interdiscip Rev RNA. 2010;1(3):402–417.