Reflections on the history of pre-mRNA processing and highlights of current knowledge: a unified picture

Abstract

Several strong conclusions emerge concerning pre-mRNA processing from both old and newer experiments. The RNAPII complex is involved with pre-mRNA processing through binding of processing proteins to the CTD (carboxyl terminal domain) of the largest RNAPII subunit. These interactions are necessary for efficient processing, but whether factor binding to the CTD and delivery to splicing sites is obligatory or facilitatory is unsettled. Capping, addition of an m(7)Gppp residue (cap) to the initial transcribed residue of a pre-mRNA, occurs within seconds. Splicing of pre-mRNA by spliceosomes at particular sites is most likely committed during transcription by the binding of initiating processing factors and ∼50% of the time is completed in mammalian cells before completion of the primary transcript. This fact has led to an outpouring in the literature about "cotranscriptional splicing." However splicing requires several minutes for completion and can take longer. The RNAPII complex moves through very long introns and also through regions dense with alternating exons and introns at an average rate of ∼3 kb per min and is, therefore, not likely detained at each splice site for more than a few seconds, if at all. Cleavage of the primary transcript at the 3' end and polyadenylation occurs within 30 sec or less at recognized polyA sites, and the majority of newly polyadenylated pre-mRNA molecules are much larger than the average mRNA. Finally, it seems quite likely that the nascent RNA most often remains associated with the chromosomal locus being transcribed until processing is complete, possibly acquiring factors related to the transport of the new mRNA to the cytoplasm.

FIGURE 1.

Nascent RNA in HeLa cells. (A) Growing cells were labeled with ³H uridine for 10 (•-•), 20 (○-○), or 45 sec (▴-▴). (The cells were pretreated with 0.04 μg/mL of actinomycin D for 25 min [Perry and Kelley 1970] so that no pre-rRNA, only hnRNA, was labeled.) Nuclear RNA was extracted, treated with DMSO to assure no aggregation, and sedimented with markers (Δ ³²P hnRNA, steady state; 3-h label), and, in another tube, 45S (14 kb), 32S (6.7 kb), and 18S (2 kb) preribosomal and ribosomal RNAs (from Fig. 3 in Derman et al. 1976). (B) PolyA⁺ nuclear RNA, pulse-labeled. Cells were labeled for 30 sec with ³H uridine or ³H adenosine as in A, hnRNA was extracted and molecules larger than 5 kb (28S) selected and submitted to polyU sepharose binding, elution with 30% formamide (denaturing concentration), and sedimented through a second sucrose gradient to determine their size. The ³H adenosine-labeled RNA was assayed both for total CPM (○-○) and for labeled polyA (data not shown in this figure); ∼60% of the polyA was in molecules 6.7 kb and larger (32S marker) and ∼40% was in smaller polyA⁺ molecules (from Fig. 4 in Derman and Darnell 1974). (C) ³H uridine flow into polyA⁺ and polyA⁻ hnRNA fragments. Cells were briefly labeled as indicated; total hnRNA was extracted, broken with brief alkali treatment to ∼500-nt fragments and polyA⁺ and polyA⁻ fragments counted (from Fig. 2 in Salditt-Georgieff et al. 1980).

FIGURE 2.

Addition of polyA during transcription of late adenovirus 2 pre-mRNA. Diagram at top shows adenovirus genome (0–100, 1 unit = 360 bases). mRNA products have three leader sequences found in all late mRNAs. Primary transcript begins at ∼16.4 and map positions for five sites of polyA addition are indicated (L1, 2, 3, 4, and 5). Spliced example shown includes a coding sequence (body) with polyA at L4 site. Graphs below show two experiments (A, B, C on left; A, B on right with cells 18 h after adenovirus infection). ³H uridine-labeled nuclear RNA (45 sec) was separated by sucrose gradient sedimentation. Fractions were assayed for RNAse-resistant hybridization to indicated adenovirus DNA fragments (•-•) from selected regions of the genome (on right of each graph). Size markers of rRNA (45, 32, or 18S markers located by OD₂₆₀) were added as size markers (horizontal dashed lines). In experiment on the right, adenovirus cytoplasmic mRNA labeled for 15–18 h with ³²P served as an additional size maker (○-○) (Fig. 2 in Nevins and Darnell 1978b).

FIGURE 3.

Time course of splicing adenovirus mRNA leader sequences. Adenovirus 2-infected HeLa cells at 16 h after infection were labeled with ³H uridine for indicated times (10 to 75 min), and total cell RNA was isolated. DNA containing all three spliced leader sequences was cloned from cDNA copies of late Ad-2 mRNA. Hybridization of labeled mRNA from each time sample to the DNA segment containing the three leader sequences was followed by RNAse treatment and proteolytic enzyme digestion to purify the DNA-RNA hybrid. RNA-DNA hybrids were recovered and immediately separated by electrophoresis. (Left) One example of gel electrophoresis and autoradiogram of hybridized Ad-2 RNAs from cells labeled for indicated times. (Right) Graph is plot of labeled, hybridized RNA as a function of time. Leader 1-2 linkage was rapid (T_1/2 < 1–2 min), and L1-L2-L3 linkage lagged by ∼15 min (in Figs. 3, 4 in Keohavong et al. 1982).

FIGURE 4.

Electron micrographs of Drosophila pre-mRNA synthesis and partial processing during transcription. Detergent-spread Drosophila embryonic cells were examined; two different pre-mRNAs are shown (A,D). Tracings in B and E show diagram of RNP fibrils deriving from transcription of each gene and the position of small and large presumed RNAP particles bound to the nascent RNA which survive the detergent treatment required to prepare the spreads. Graphs at right (C,F) represent summaries of the position of large and small particles on pre-mRNA fibrils that provide an estimate of intron length and elapsed time for processing (see text; Fig. 2 in Beyer and Osheim 1988).

FIGURE 5.

Using nascent RNA to study pre-mRNA synthesis and processing. (A) Short pulse-label times with ³H uridine (∼1 min or less) revealed nascent RNA profile (see Fig. 1A). (TSS) transcription short site, (pA) polyA site. Hybridization of such labeled RNA to segments of a known DNA (a, b, c, d) template were used to map individual transcription units that prove fast polyA addition time (see Figs. 1C, 2) and demonstrate the first recognized cotranscriptional splicing (see Fig. 3). (B) Experimental discharge of nascent RNA with DRB treatment followed by removal allows nascent RNA synthesis to resume from at (or very near) the transcription start site. Given total genome sequencing, progress of transcription across any selected gene can be monitored by choosing primers at various sites and RT-PCR amplifying nascent, newly started RNA. This allows a measurement of elapsed time to transcribe across a gene and with selected primers in both exons and introns to determine extent of cotranscriptional splicing (see Pandya-Jones and Black 2009; Singh and Padgett 2009; Wada et al. 2009). (C) Studying nascent RNA from individual inducible genes. Nascent RNA has been studied on induced genes recently hybridized to tiling arrays of DNA (see Wada et al. 2011) and by RNA_seq in several labs after preparing chromatin-associated RNA (see Bhatt et al. 2012; A Pandya-Jones, DM Bhatt, C-H Lin, A-J Tong, ST Smale, and DL Black, in prep.).

FIGURE 6.

Choosing polyA sites. (A) The DSX pre-mRNA transcript is processed differently in females and males. In females, TRA proteins (SR-like proteins, yellow ellipses) are present and required to splice exon 3 to exon 4 which has a translation STOP codon and polyA site at the end of exon 4. The identified binding sites for the TRA protein (green) and other recruited factors are a series of repeated 13-mers which are actually in exon 4. In males, the absence of TRA proteins is associated with skipping exon 4 and splicing of exon 3 to exon 5 and then to exon 6 which has a polyA site (in Fig. 2 of Maniatis and Tasic 2002). (B) (Top) Processing pre-mRNA from the gene encoding calcitonin (CT) and calcitonin gene-related protein (CGRP). The fourth exon (black) is included in the CT mRNA ending the translation of the preprotein that encodes the short CT hormone in exon 4. In neuronal cells, exon 4 is skipped and exon 5 and exon 6, which encodes the CGRP sequence in the pre-protein and has a functional polyA site, are included (Amara et al. 1984). (Bottom) The intron between exons 4 and 5 contains a processing enhancer element (black oval) that recruits a protein complex initiated by the binding of the SR-like protein, SRp20. This “splicing factor” has a strong affinity for CF1, a known polyA site factor. Two proposed modes of action (direct and indirect) of SRp20 for favoring exon 4 inclusion in thyroid cells are shown (in Figs. 1 and 8 in Lou et al. 1998). (C) Immunoglobulin heavy chains (µ locus) in B cells and plasma cells. Selection of polyA sites for either the secreted form of the protein (first or µs site) or the membrane-bound form (µm polyA site) lead to two proteins that differ in a –COOH terminal peptide. While the pre-mRNA could be differentially spliced to include either of the –COOH terminal domains, the selection of polyA sites of the μS encoding exon is greatly favored by a large increase in CstF 64 (cross-hatched circles) in plasma cells which drives choice of the “weaker” first polyA site. The open oval (CPSF) is another required polyA factor (in Fig. 8B in Takagaki et al. 1996).