The conserved AAUAAA hexamer of the poly(A) signal can act alone to trigger a stable decrease in RNA polymerase II transcription velocity

Abstract

In vivo the poly(A) signal not only directs 3'-end processing but also controls the rate and extent of transcription. Thus, upon crossing the poly(A) signal RNA polymerase II first pauses and then terminates. We show that the G/U-rich region of the poly(A) signal, although required for termination in vivo, is not required for poly(A)-dependent pausing either in vivo or in vitro. Consistent with this, neither CstF, which recognizes the G/U-rich element, nor the polymerase CTD, which binds CstF, is required for pausing. The only part of the poly(A) signal required to direct the polymerase to pause is the AAUAAA hexamer. The effect of the hexamer on the polymerase is long lasting--in many situations polymerases over 1 kb downstream of the hexamer continue to exhibit delayed progress down the template in vivo. The hexamer is the first part of the poly(A) signal to emerge from the polymerase and may play a role independent of the rest of the poly(A) signal in paving the way for subsequent events such as 3'-end processing and termination of transcription.

FIGURE 1.

The poly(A) signal core and G/U-less variants thereof. (A) The core of the polyadenylation apparatus (Zhao et al. 1999; Proudfoot 2004). (B) Sequence elements used as inserts in the reporter constructs of this work. There is no G/U-rich region downstream of the insert.

FIGURE 2.

The G/U-rich region of the poly(A) signal is not required for pausing. (A) Pausing occurs in the absence of a G/U-rich region. The reporters used are identical to the corresponding A₃〈EC_n〉 constructs of Orozco et al. (2002) except that they contained inserts iii and v of Figure 1B instead of a poly(A) signal (also an AATAAA hexamer centered 65 bp upstream of the 140-bp cassette was removed by converting it to CCCGGG by site-directed mutagenesis). The region upstream of the 140-bp cassette contains the SV40 small intron and is mostly identical to the first two-thirds of the A₃ region described in Orozco et al. (2002). The upstream cassette here is 140 bp rather than the 136 bp of Orozco et al. (2002) because of the different cloning junctions. The results plotted are the average and range from two different experiments. (B) Processing does not occur in the absence of a G/U-rich region. RNase protection was carried out on nuclear RNA. The probe (lane 3) for the G/U-less constructs in lanes 1 and 2 was generated by RNA polymerase T3 transcription of a PCR fragment generated from construct C₁ (A, top) containing insert iii (Fig. 1B). The RNase protection in lane 4 for a poly(A) signal was as in Park et al. (2004). In both cases only T1 RNase was used.

FIGURE 3.

Pausing and termination with the penta-cassette reporter. (A) Intact poly(A) signal. The gel lanes show results of representative G-less cassette run-on assays of nuclei isolated from COS cells transfected with reporter plasmids containing five G-less cassettes and either insert i or insert ii (centered at the position of the arrow). The reporter (roughly to scale) was derived from pAP〈cat〉 of Figure 6A in Tran et al. (2001), to which it is identical downstream of the HindIII site. The region upstream of the 104-bp G-less cassette contains the SV40 small intron and is mostly identical to the first two-thirds of the A₃ region described in Orozco et al. (2002). The graph summarizes the results of three experiments, showing the average and standard deviation. Because termination is exponential, we retain the convention of plotting the results semi-logarithmically (Orozco et al. 2002). (B) G/U-less poly(A) signals. The average and standard deviation of three (element iii) or the average and range of two (element iv) entirely separate experiments are shown. (C) ΔCTD polymerase. For methods, see text and Park et al. (2004). Closed triangles show the average and range for two different experiments with element iii (normalized to v).

FIGURE 4.

The G/U-rich region is not required for poly(A)-dependent pausing/slowing in vitro. (A) The AAUAAA hexamer directs pausing in vitro. The gel lanes show representative results of an in vitro elongation assay with reporter plasmids containing either a single hexamer (insert iv) or its mutant (insert v) centered at the position of the arrow in the map beneath the gel lanes. The reporter map shown is the same as that for pAP〈cat〉 in Figure 6A of Tran et al. (2001) except for the hexamer insert (there are additional irrelevant differences outside the region mapped). The percentage readthrough was calculated (see figure) by normalizing the post/pre ratio for element iv to that for its mutant (i.e., element v). The result is shown in the middle bar of the Figure 4B histogram. (B) Hexamer-mediated pausing is comparable to pausing directed by a full poly(A) signal. The percentage readthroughs are shown for a full poly(A) signal as well as for two different G/U-less constructs. The post/pre ratio for each element was normalized to that for its respective mutant (i normalized to ii; iii and iv normalized to v). (C) Size exclusion of transcripts depends on the integrity of the DNA. Two plasmids, one expressing a 117-nt G-less cassette and the other a 137-nt cassette, were transcribed in parallel, and then either active (117 construct) or heat denatured (137 construct) DNase I was added to each mixture for 3 min. The reactions were stopped with salt, Sarkosyl, and EDTA; mixed together; and then loaded on a size-exclusion column. The RNA from each fraction was isolated, digested with RNase T1, and displayed on an 8% polyacrylamide gel. The experiment was repeated, but with the DNase I treatments reversed, with similar results. (D) Size exclusion of transcripts depends on the tether to the polymerase. Transcription was carried out in the presence of a DNA oligonucleotide that directs RNase H to cut the transcript about 800 nt downstream of the promoter (Rigo et al. 2005). Size-exclusion chromatography was then carried out to separate the ternary complexes from the released RNA. The column fractions were electrophoresed in two groups at different times, so the intensities of the two gels relative to each other may not exactly correspond. (E) Direct evidence that ternary complexes slow down after crossing an AAUAAA hexamer. Elongation assays were carried out on reporters containing either element iii or element v of Figure 1B, and the transcription mixtures were then fractionated by size-exclusion chromatography. An aliquot of each fraction was analyzed for plasmid DNA by agarose gel electrophoresis to identify the fractions containing ternary complexes. These fractions (7 and 8) were then analyzed by using the methods of A to determine what proportion of ternary complexes had reached the end of the post-cassette (and, conversely, what proportion still remained within the “window”). The error bars represent the range of values from two separate experiments. The reporter used is the same as in Figure 5, A and B, but with a G/U-less element in place of the poly(A) signal.

FIGURE 5.

CstF is not required for poly(A)-dependent pausing/slowing in vitro. (A) CstF is not required for poly(A) signal–mediated pausing in vitro. The Western blot documents physical depletion of CstF and shows equal amounts of untreated extract, mock (anti-E1B)-depleted extract, and anti-CstF64 depleted extract run on an 8% SDS–polyacrylamide gel and transferred to PVDF membrane (Roche) followed by overnight incubation with α-CstF64 antibody 6A9. To visualize the blot, we used the BM Chemiluminescence Western Blotting Kit Mouse/Rabbit (Roche), and the blot was exposed to Hyper film (Amersham). We note that two CstF bands appear in this blot, previously attributed to different post-translationally modified forms (Takagaki et al. 1990). However, two bands are not always apparent as seen in parts B and C and as reported by Wallace et al. (1999). The cleavage assay documents functional depletion of CstF. The RNA substrate was incubated with either mock or CstF depleted extract for 2 h followed by RNA isolation. The in vitro pausing assay was carried out by using a plasmid reporter closely related (and identical in the region mapped) to the pAP〈C₉〉 construct in Figure 5 of Tran et al. (2001). (B) Similar to A except mock depletion was with normal goat antibody and probing was with 3A7. (C) Similar to A but using the reporter of Figure 4A and a G/U-less insert (insert iii, and also v for normalization). Probing was with 3A7.