A tale of two tails

The fate of RNAs in the nucleus is determined by the polymerase that made them. For example, the same primary transcript that will give rise to a messenger RNA if made by RNA polymerase II (pol II), will never mature if is made by pol I, pol III or a bacteriophage RNA polymerase¹,². How can we explain this link between the machines that make and process mRNAs? On page 73 of this issue, Hirose and Manley³ bring us one step closer to an answer. They report that the carboxy-terminal domain (CTD) — a protein domain unique to pol II — acts as part of the protein complex that fashions the 3′ end of the mature mRNA by cleavage and polyadenylation.

The formation of 3′ ends is a two-step process in which the RNA transcript is cleaved downstream of the sequence AAUAAA, then a poly(A) tail is added at the cut end. RNAs made in vivo by pol II lacking the CTD — a conserved repeat of the sequence YSPTSPS, found at the carboxy terminus of the pol II large subunit — are not efficiently capped, spliced or cleaved at the poly(A) site⁴,⁵. This suggests that the CTD helps to target mRNA processing factors to pol II transcripts⁶,⁷. Hirose and Manley have now discovered that the CTD facilitates the 3′ RNA cleavage reaction, even in the absence of transcription. They suggest that, in effect, the CTD is a cofactor for 3′ processing.

Production of a mature mRNA 3′ end can be reconstituted in vitro with cleavage/polyadenylatn specificity factor (CPSF), cleavage-stimulation factor (CstF), cleavage factors CFI and CFII, and poly(A)polymerase (PAP)⁸. The association of CPSF and CstF with the CTD in vitro⁴, and the colocalization of CstF and phosphorylated pol II in vivo (Fig. 1), indicate that a stable ‘mRNA factory’ complex of pol II with processing factors carries out synthesis and maturation of the primary transcript⁹.

Figure 1: Colocalization of the cleavage-stimulation factor CstF and phosphorylated RNA polymerase II (pol II).

Immunofluorescence of Drosophila polytene chromosomes with anti-pol II monoclonal antibody (left) and rabbit antibody against Suppressor of forked, the Drosophila homologue of CstF (right). (Courtesy of M. Sikes and A. Beyer, University of Virginia.)

Hirose and Manley provide an insight into how the mRNA factory operates. They followed up the curious observation that, in a purified system, cleavage at the poly(A) site is stimulated by creatine phosphate and by phosphoamino acids. They reasoned that these compounds may mimic a phosphoprotein that allosterically activates the cleavage reaction. An attractive candidate phosphoprotein is pol II — the CTD undergoes a cycle of hyperphosphorylation and dephosphorylation (mainly at residues two and five of the YSPTSPS repeat¹⁰) as pol II cycles through transcriptional initiation, elongation and termination. Sure enough, the authors found that phosphorylated pol II and recombinant CTD stimulate cleavage in vitro. Unexpectedly, however, the unphosphorylated CTD was also effective.

These experiments show that the CTD stimulates 3′ processing in the absence of transcription, although it is not clear whether the CTD is acting in the same way as creatine phosphate. Hirose and Manley also found that the cleavage reaction was inhibited when they immunodepleted pol II from a crude nuclear extract, after dissociating poly(A) factors from the CTD with high salt. Moreover, they could restore cleavage activity by adding back pure pol II. In other words, the CTD stimulates cleavage in a crude extract, as well as in a purified system.

But how does the CTD stimulate cleavage in the absence of transcription? Does it promote the formation of a stable complex (within which cleavage can then occur), or does it contact polyadenylation factors only fleetingly, triggering cleavage? Assembly of a complex — which might comprise CstF, CPSF, CFI, CFII and PAP, together with the RNA (Fig. 2) — may be rate limiting both in vivo and in vitro. The CTD could accelerate assembly of the complex by acting as a scaffold that positions the proteins optimally, all in the same place at the same time. A scaffold might even impose an order of assembly of the polyadenylation machine.

Figure 2: Model for activation of 3′ processing by the pol II carboxy-terminal domain (CTD).

Hirose and Manley³ have shown that the CTD is part of the protein complex that carries out cleavage and polyadenylation to produce a mature messenger RNA transcript. CPSF, cleavage/polyadenylation specificity factor; CstF, cleavage-stimulation factor; CFI, cleavage factor I; CFII, cleavage factor II; PAP, poly(A)polymerase.

Hirose and Manley suggest that the CTD is not just a passive scaffold but an active participant in the cleavage reaction. According to this model, which is more consistent with the effect of creatine phosphate, the CTD behaves as a cofactor or allosteric activator. One possible target for allosteric activation is CstF, which binds the CTD in vitro⁴. The importance of this interaction for scaffolding or allosteric activation could be tested using CTD mutants that do not bind CstF.

It is now clear that pol II stimulates 3′ processing, both in vivo and in vitro, through the CTD. Conversely, poly(A) factors may affect pol II by influencing the decision to terminate or elongate the RNA chain. Termination depends on transcription of the poly(A) signal¹¹, poly(A) factors¹² and the CTD (ref. 4). Signalling from the polyadenylation machine to the polymerase may ensure that the chain is terminated only after pol II has reached the end of a gene. So, a two-way line of communication between pol II and the 3′ processing factors may be built into the mRNA factory.

Working out how the machines in the mRNA factory interact with one another is a big challenge. Hirose and Manley³ have provided the first in vitro system to study one aspect of the complex communication system that integrates different nuclear processes through protein-protein interactions. We can look forward to the development of more in vitro systems, which will reveal the details of how the CTD affects splicing and capping as well as polyadenylation.