An investigation into the role of ATP in the mammalian pre-mRNA 3' cleavage reaction

Abstract

RNA Polymerase II transcribes beyond what later becomes the 3' end of a mature messenger RNA (mRNA). The formation of most mRNA 3' ends results from pre-mRNA cleavage followed by polyadenylation. In vitro studies have shown that low concentrations of ATP stimulate the 3' cleavage reaction while high concentrations inhibit it, but the origin of these ATP effects is unknown. ATP might enable a cleavage factor kinase or activate a cleavage factor directly. To distinguish between these possibilities, we tested several ATP structural analogs in a pre-mRNA 3' cleavage reaction reconstituted from DEAE-fractionated cleavage factors. We found that adenosine 5'-(β,γ-methylene)triphosphate (AMP-PCP) is an effective in vitro 3' cleavage inhibitor with an IC50 of ∼300 μM, but that most other ATP analogs, including adenosine 5'-(β,γ-imido)triphosphate, which cannot serve as a protein kinase substrate, promoted 3' cleavage but less efficiently than ATP. In combination with previous literature data, our results do not support ATP stimulation of 3' cleavage through cleavage factor phosphorylation in vitro. Instead, the more likely mechanism is that ATP stimulates cleavage factor activity through direct cleavage factor binding. The mammalian 3' cleavage factors known to bind ATP include the cleavage factor II (CF IIm) Clp1 subunit, the CF Im25 subunit and poly(A) polymerase alpha (PAP). The yeast homolog of the CF IIm complex also binds ATP through yClp1. To investigate the mammalian complex, we used a cell-line expressing FLAG-tagged Clp1 to co-immunoprecipitate Pcf11 as a function of ATP concentration. FLAG-Clp1 co-precipitated Pcf11 with or without ATP and the complex was not affected by AMP-PCP. Diadenosine tetraphosphate (Ap4A), an ATP analog that binds the Nudix domain of the CF Im25 subunit with higher affinity than ATP, neither stimulated 3' cleavage in place of ATP nor antagonized ATP-stimulated 3' cleavage. The ATP-binding site of PAP was disrupted by site directed mutagenesis but a reconstituted 3' cleavage reaction containing a mutant PAP unable to bind ATP nevertheless underwent ATP-stimulated 3' cleavage. Fluctuating ATP levels might contribute to the regulation of pre-mRNA 3' cleavage, but the three subunits investigated here do not appear to be responsible for the ATP-stimulation of pre-mRNA cleavage.

Keywords: 3′ End formation; ATP; Cleavage and polyadenylation; Pre-mRNA processing.

Fig. 1. The effect of ATP and ATP structural analogs on the pre-mRNA 3′cleavage reaction reconstituted from HeLa DEAE-fractionated cleavage factors

A. In vitro reconstituted cleavage of the uniformly labeled, 5′ capped L3 pre-mRNA substrate. Nucleotide concentration, 0.5 mM H₂O, water added in place of nucleotide. Input, unprocessed RNA. 5′, upstream cleavage product; 3′, downstream cleavage fragments; R.C. Relative cleavage, normalized to ATP lane. Bar charts were made using the average of three independent identical experiments, ± one S.D. See text for nucleotide abbreviations. B. Same as in panel A except the SV40L pre-mRNA substrate was used. C. Structural differences of three of the nucleotides used. D. Same as in panel A, except 1 mM AMP-PNP was used. Spaces between gel lanes indicate rearrangement of lanes from a single representative experiment.

Fig. 2. The effect of AMP-PCP on ATP-stimulated L3 pre-mRNA cleavage

ATP (1 mM) and AMP-PCP at the indicated concentrations were included in a reconstituted in vitro cleavage reaction as in Fig. 1. The bar chart was made using the average of three independent identical experiments, ± one S.D.

Fig. 3. Excess ATP is not needed for CF II_m Clp1-Pcf11 complex stability

A. FLAG-tagged Clp1 was expressed in a stable HeLa cell line (Clp1-FLAG) and immunoprecipitated using FLAG-beads from whole cell extract with no added nucleotide or with 4 mM ATP or AMP-PCP. Normal HeLa cells were used as a control. The indicated fractions were resolved on SDS-PAGE and blotted. The top portion of the blot was probed with an anti-Pcf11 antibody. The lower portion was probed with an anti-Clp1 antibody. W.B., Western blot. B. The experiment shown in panel A was repeated (n) times with either ATP-depletion (as shown in A) or dialysis of the whole cell extract followed by ATP depletion. Bars indicate the Pcf11-Clp1 ratio normalized to the control where no nucleotide (nt) was included in the precipitation.

Fig. 4. Ap4A, a CF I_m25 Nudix domain ligand, does not affect in vitro pre-mRNA 3′ cleavage

A. Increasing concentrations of Ap4A do not inhibit ATP-stimulated cleavage of L3 pre-mRNA. B. Ap4A alone does not stimulate pre-mRNA cleavage in place of ATP. Cleavage reactions were done as described in Fig. 1.

Fig. 5. ATP does not stimulate pre-mRNA 3′ cleavage through the canonical PAP ATP-binding site

A. Bacterially expressed, N-term-His₆-tag affinity purified wt and mutant bovine PAP proteins (bPAP, ~3 μg) used in this study. SDS-PAGE with Coomassie staining. Asterisk (*) denotes bands arising through adventitious C-terminal proteolysis during expression or purification. BSA, bovine serum albumin standards. B. Evaluation of ATP binding using 1 mM Mn²⁺-mediated non-specific polyadenylation at 50 nM PAP concentration. wt(2×) denotes 100 nM PAP. RNA substrate/primer was the uniformly labeled L3 pre-mRNA. C. In vitro L3 RNA cleavage reconstituted from the HeLa nuclear extract CSF fraction supplemented with the indicated PAP protein and ATP. D. Evaluation of ATP binding as in panel B but with lower PAP (10 nM) and ATP concentrations. wt(2×), 20 nM PAP. wt(5×), 50 nM. E. Reconstituted in vitro cleavage using CSF and recombinant PAP as in panel C but using 10 nM PAP and 0.23 mM ATP. F. Reconstituted in vitro cleavage using CSF and recombinant PAP as in panels C and E but here only 5 nM PAP, with or without 0.23 mM ATP.