RNA-protein interactions regulate all cellular processes, with dysregulation driving numerous diseases. UV crosslinking and immunoprecipitation methods (CLIP-seq) revolutionised our understanding of these interactions but currently suffer critical limitations. Only fragmented coverage is used, losing isoform-specific information; PCR amplification is employed, introducing quantitative biases; binding sites cannot be resolved within their native full-length transcript contexts and at true single-molecule level. These constraints fundamentally limit our ability to understand how proteins recognise and regulate specific RNA isoforms - a critical gap given widespread alternative splicing and processing involvement in cellular regulation.
Here we developed DIR-CLIP (Direct Identification of RNA binding sites by Crosslinking and Immunoprecipitation), which combines selective protein-RNA complex isolation with Oxford Nanopore Technologies long-read sequencing. This amplification-free approach preserves full-length RNAs, enabling unprecedented isoform-specific and single-molecule binding site mapping.
Our method employs two complementary sequencing strategies. Direct RNA sequencing provided proof-of-principle, detecting binding sites through both signal deviation analysis using Nanocompore and premature read termination profiling - a novel signature where protein adducts obstruct nanopore translocation. While demonstrating single-nucleotide resolution, premature termination reduced discovery capacity in 5' regions. To address this, we implemented amplification-free cDNA sequencing, where reverse transcription errors at crosslink sites create diagnostic mutation signatures detectable across entire transcripts.
We validated DIR-CLIP by mapping serine-arginine rich splicing factor 3 (SRSF3) binding sites in P19 teratocarcinoma cells. Sites identified through both approaches showed strong correlation with established iCLIP data, confirming specificity and sensitivity. Notably, our method revealed isoform-specific binding patterns invisible to conventional approaches. Using complementary computational pipelines, we achieved comprehensive binding site detection: Nanocompore identified 787 sites via signal deviation, read termination analysis detected 417 sites, with minimal overlap suggesting distinct crosslink signatures captured by each approach.
DIR-CLIP represents a paradigm shift in studying RNA-protein interactions, providing unbiased, quantitative, single-nucleotide resolution mapping within native full-length transcripts. By preserving RNA integrity and eliminating amplification bias, our method enables investigation of how proteins differentially recognise and regulate distinct RNA isoforms - critical for understanding post-transcriptional regulation in health and disease. Our technology opens new avenues for discovering isoform-specific therapeutic targets and understanding the true complexity of RNA-protein regulatory networks.