Three helicases and a tethered enzyme dismantle the human spliceosome after each intron

The search for termination factors began with a clean comparison. Earlier work from the same group showed that in whole-worm extract of C. elegans, the intron-lariat spliceosome is the main spliceosome state and carries the helicase DHX15, and that adding ATP switches DHX15 on to start termination ^[9]. This time the team purified the ATP-treated complexes and compared them with the untreated intron-lariat spliceosome by mass spectrometry ^[10]. The starting complex was the same and only one ingredient was added. The treated complexes had lost all five known termination factors, TFIP11, PAXBP1, C19L1, C19L2 and DHX15 itself, and were significantly enriched in DHX35 and GPATCH1 ^[10].

Those two proteins mattered because nobody could explain the later steps. A spliceosome is built fresh on each intron, and after the joined mRNA leaves, the cut-out intron stays bound in the spliceosome's RNA active site ^[12]. The spliceosome then has to be taken apart so its proteins and snRNAs can be reused, and the intron has to be ejected, debranched at its branch-point adenosine and degraded ^[14]. Because the machine is assembled anew for every intron, that teardown happens every time an intron is removed ^[17]. Recent studies had described only the first step, recognition by four dedicated termination factors and DHX15 ^[13]. In those structures the branch point was buried in the spliceosome's RNA network ^[11]. DBR1, an essential enzyme, had been proposed to debranch the intron while it was still on the spliceosome, but that required some unknown way of pulling the branch point out first ^[11][15].

The pathway, as the authors report it, applies to human spliceosomes. The structures are of endogenous human complexes, and the team combined them with biochemistry, RNA sequencing and genetics ^[8]. The proteomic screen that flagged DHX35 and GPATCH1 ran in worm extract ^[10]. That split is reasonable, since the human structures carry the claim and the worm data pointed to where to look. It does mean that the human role of each factor rests on the structural and genetic evidence in human material.

Cryo-EM caught two sequential stages of termination ^[2], with three RNA helicases acting across them ^[16]. The two snapshots establish the order of states. They do not show how fast each step runs in a living cell, or which one limits how quickly snRNAs return for the next round of splicing; those rates need kinetic measurements.