Why does 4-thiouracil labelling to map RNA-binding proteins cause a T-C change?

Question

I am now reading a paper about the purification and identification of mRNAs and its RNA binding proteins by using UV crosslinking and immunoprecipitation. I came upon this sentence which puzzled me

Up to 28% of all uridines present in 3' UTRs showed diagnostic T-C changes in the protein occupancy profiling sequence reads. This number is reasonably high, considering observations that typically only one of a few uridines in RNA binding sites when substituted by 4SU, crosslinks to proteins.

What I do not understand is why does 4-thiouracil (4SU) labelling in the cell show T-C changes. Would the uracil-adenine base pairing not be affected?

How does the 4SU labelling really work?

Answer 1

The technique used in that paper is called PAR-CLIP (Photoactivatable Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation; see Ascano et al., 2012 and Spitzer et al, 2014). This technique involves substituting uridine (U) with 4-thiouridine (4SU) or substituting guanosine (G) with 6-thioguanosine (6SG). The rationale behind using these modified bases is that they improve the efficiency of RNA-protein crosslinking. Moreover, PAR-CLIP uses 365nm UV instead of the 254nm UV used in the conventional crosslinking. However, the crosslinking process causes the loss of the sulphur atom causing the U to be read as C during reverse transcription. Likewise, 6SG gets read as A.

_{Ascano et al., 2012}