Characterization of factors involved in the coupling of 3' end processing and splicing and in the 3' end formation of mRNA precursors

Eukaryotic mRNA precursors are processed at their 5’ and 3’ ends and are spliced prior
to their export from the nucleus to the cytoplasm. Although all three processing reactions
can be studied separately in vitro, they are coupled in vivo.
3’ end processing of most mammalian pre-mRNAs involves endonucleolytic
cleavage followed by polyadenylation of the upstream cleavage product. Cleavage and
polyadenylation specificity factor (CPSF) is a multiprotein complex, which together with
cleavage factor Im and IIm (CF Im, CF IIm), cleavage stimulatory factor (CstF), poly(A)
polymerase (PAP) and nuclear poly(A) binding protein 1 (PABPN1) is required for
3’ end formation.
We have found that the U2 snRNA and subunits of the splicing factors 3a and 3b
(SF3a, SF3b), which are components of the U2 snRNP, were also present in highly
purified CPSF fractions. GST pull-down experiments indicated a direct interaction of
CPSF subunits with SF3b49 and SF3b130. Furthermore, antibodies directed against
CPSF100 co-immunoprecipitated subunits of SF3a and SF3b and the U2 snRNA Taken
together our results show that subunits of CPSF and the U2 snRNP directly interact with
each other.
In order to analyze whether this interaction plays a role in the coupling of 3’ end
processing and splicing, we depleted CPSF subunits from HeLa cell nuclear extract and
tested the extracts for splicing activity. CPSF100-depleted extract showed no detectable
cleavage activity and its splicing activity was significantly reduced in coupled assays but
not in un-coupled assays. Moreover, pre-mRNAs containing mutations in the binding site
of SF3b were not only less efficiently spliced but they also showed reduced cleavage
activity. Interestingly, efficient cleavage required the presence of the U2 snRNA in
coupled but not in un-coupled assays. Based on our studies, we propose that the
interactions between CPSF and U2 snRNP contribute to the coupling of splicing and 3’
end formation.
Furthermore, we depleted CPSF100 and the U2 snRNP subunits SF3b155 and
SF3b130 by means of RNAi. We observed that knock down of both SF3b proteins caused
high lethality of the cells indicating that these polypeptides are essential. However,
depletion of CPSF100 did not result in a significant increase in cell mortality, suggesting
that the protein is either not essential that the knock down was not efficient enough to
result in cell lethality or that CPSF100 shares redundant functions with another protein.
We were able to show that SF3b155 and SF3b130 are required for efficient splicing in
vivo but did not detect a splicing deficiency in CPSF100 depleted cells. Knock down of
neither of the proteins resulted in an observable 3’ end processing deficiency. Further
work is required to address the question if the U2 snRNP and CPSF couple splicing and
3’ end processing in vivo.
Splicing and 3’ end formation are highly conserved mechanisms from mammals
to yeast and the two organisms share homologues of most of the proteins involved in the
two reactions. To test whether the coupling mechanism mediated by CPSF and the
U2 snRNP is conserved between different organisms, we focused on the yeast system.
The essential protein Rse1p is the yeast homolog of SF3b130. We show that the rse1-1
strain is sensitive to cordycepin, which suggests that Rse1p might be involved in 3’ end
processing. Furthermore, Rse1p and 3’ end processing factors interacted genetically and
Northern blot analysis suggested that strains carrying mutations in Rse1p and subunits of
CPF had increased levels of unspliced pre-mRNA at restrictive temperature compared to
the single mutants. We therefore suggest that the coupling of 3’ end processing and
splicing mediated by CPSF and U2 snRNP is conserved between mammals and yeast.
Precursor tRNAs (pre-tRNAs) must undergo a number of processing steps before
they become mature tRNAs and some tRNAs contain introns. tRNA splicing is a three
step reaction and each step requires an individual set of proteins. In the first step the pretRNA
is cut at its two splice sites. This reaction is catalyzed by the so called tRNA
splicing endonuclease complex. Recently this complex was purified from mammalian
cells and interestingly hClp1 (a subunit of the 3’ end processing factor CF IIm) was
identified as one of its components. Furthermore, hSen2 a subunit of the endonuclease
complex was shown to be required for efficient 3’ end processing in vivo. A model was
proposed suggesting that the tRNA endonuclease complex is involved in 3’ end
processing. In collaboration with S. Paushkin and C. Trotta we continued to investigate if
this model is indeed correct. We found that biochemically purified CF IIm indeed carried
tRNA endonuclease activity. However, tRNA endonuclease complexes were not able to
reconstitute cleavage activity of CF IIm-depleted HeLa nuclear extract, unless they were
purified with His-Flag tagged hClp1. Taken all our results into account we cannot
exclude that the tRNA endonuclease complex is indeed involved in 3’ end processing.
However, we believe that the evidence supporting this model is rather weak. We think it
is more likely that hClp1 is part of the tRNA endonuclease complex as well as a subunit
of CF IIm, and that the two complexes are not functionally related.
As mentioned earlier, 3’ end processing is highly conserved form mammals to
yeast. In S. cerevisiae cleavage and polyadenylation factor (CPF) is a multiprotein
complex, which together with the cleavage factor IA (CF IA) and the cleavage factor IB
(CF IB) is required for both the cleavage and the polyadenylation step of the 3’ end
formation reaction.
Ydh1p/Cft2p is an essential component of CPF. Cleavage and polyadenylation
reactions revealed that the protein is required for both reactions to occur in vitro.
Previously, it was demonstrated that an important function of CPF lies in the recognition
of poly(A) site sequences and previous RNA binding analyses with recombinant
Ydh1p/Cft2p suggested that the protein may interact with the CYC1 poly(A) site region.
In accordance, we found that mutant ydh1 strains were deficient in recognition of the
ACT1 cleavage site in vivo.
Transcription by RNA polymerase II (RNAP II) and 3’ end processing reactions
are tightly linked. The C-terminal domain (CTD) of RNAP II plays a major role in
coupling the two events, as it tethers the factors involved in polyadenylation to the
polymerase. We provide evidence that Ydh1p/Cft2p interacts with the CTD, several
subunits of CPF and with Pcf11p, a component of CF IA. We propose that Ydh1p/Cft2p
contributes to the formation of important interaction surfaces that mediate the dynamic
association of CPF with RNAP II, the recognition of poly(A) site sequences and the
assembly of the polyadenylation machinery on the RNA substrate.