Position-dependent sequence elements downstream of AAUAAA are required for efficient rabbit beta-globin mRNA 3' end formation

We previously demonstrated that a critical 35 bp region 3' of the AAUAAA is required for rabbit beta-globin mRNA 3' end formation. Recently, we synthesized and tested sequence elements derived from this region. Here, we report that a GU-rich and a U-rich sequence element are both required for efficient rabbit beta-globin mRNA 3' end formation. The efficiency of processing is restored to the wild-type level when the two elements are placed together and is greatly diminished when only one element is present. The level of 3' end formation is also decreased when the distance between the two elements is expanded. These results demonstrate that the GU-rich and U-rich elements function synergistically to restore efficient mRNA 3' end formation and that they most likely form a single requisite sequence 3' of the AAUAAA. Furthermore, we show that the effect of the GU-rich and U-rich sequence elements is position-dependent.     

The consensus sequence YGTGTTYY located downstream from the AATAAA signal is required for efficient formation of mRNA 3'' termini.

Our previous DNA sequence comparisons of 3' terminal portions from equivalent herpes simplex virus type 1 (HSV-1) and HSV-2 genes identified a conserved sequence (consensus YGTGTTYY; Y = pyrimidine) located approximately 30bp downstream from the AATAAA signal. We report here that this signal is located downstream from 67% of the mammalian mRNA 3' termini examined. Using constructions with the bacterial chloramphenicol acetyl transferase (CAT) gene linked to an HSV 'terminator' fragment, we show that deletions in the 'terminator' reduce CAT activities and the levels of CAT mRNA 3' termini. Specifically: (1) deletions of downstream sequences which extend up to the consensus YGTGTTYY signal reduce CAT levels to values 35% of those obtained with undeleted plasmids, (2) a deletion of a further 14bp, which removes the YGTGTTYY consensus but not the poly A site, reduces CAT activities to 1%-4%. The levels of CAT mRNA 3' termini reflect the reductions in CAT activities however, levels of mRNA 5' termini are unaffected by these deletions. The RNA produced in the absence of the YGTGTTYY signal is present in the cytoplasm although no CAT activity is detectable.     

Signal sequence for generation of mRNA 3'' end in the Saccharomyces cerevisiae GAL7 gene.

We have identified a signal sequence (designated core signal) necessary to specify formation of mRNA 3' end of the GAL7 gene in Saccharomyces cerevisiae within a DNA segment 26 bp long. The sequence was located 4-5 nucleotides upstream from the 3' end, i.e. the polyadenylation site, of the GAL7 mRNA. Replacement of a DNA segment encompassing the polyadenylation site with a pBR322 DNA, leaving the core signal intact, resulted in alteration of the mRNA 3' end by several nucleotides, suggesting the existence of an additional signal (designated end signal) at or near the polyadenylation site. The normal end formation was abolished when the core signal was placed in the reverse orientation. A considerable fraction of pre-mRNA synthesized in vitro with SP6 RNA polymerase on the template of a DNA fragment containing these signals was cleaved and polyadenylated presumably at the in vitro 3' end during incubation in a cell-free system of yeast. By contrast pre-mRNA synthesized on the template with the core signal alone was processed but much less efficiently. No such processing was seen when the pre-mRNA either lacked the core signal or contained it in the reverse orientation.     

Three distinct RNA sequence elements are required for efficient apolipoprotein B (apoB) RNA editing in vitro.

Apolipoprotein B (apoB) mRNA is edited in rat liver and intestine to convert a CAA glutamine codon to a UAA translational stop codon by the direct conversion of cytidine to uridine at nucleotide 6666. We have proposed the 'mooring sequence' model for apoB RNA editing, in which editing complexes (editosomes) assemble on specific apoB mRNA flanking sequences to direct this site-specific editing event. One sequence element (approx. nts 6671-81, the presumed 'mooring sequence') has been previously identified as necessary for editing. We have identified two additional sequence elements which are necessary for efficient editing: (1) a 5' 'Regulator' region which modulates editing efficiency and (2) a 'Spacer' region between the editing site and the 3' mooring sequence, whose distance is critical for efficient editing. Utilizing this data, we have induced editing at a cryptic site and have defined a 22 nucleotide 'cassette' of specific apoB sequence which is sufficient to support wild-type levels of editing in vitro in a background of distal apoB RNA sequence.     

Homologous mRNA 3'' end formation in fission and budding yeast.

Sequences resembling polyadenylation signals of higher eukaryotes are present downstream of the Schizosaccharomyces pombe ura4+ and cdc10+ coding regions and function in HeLa cells. However, these and other mammalian polyadenylation signals are inactive in S. pombe. Instead, we find that polyadenylation signals of the CYC1 gene of budding yeast Saccharomyces cerevisiae function accurately and efficiently in fission yeast. Furthermore, a 38 bp deletion which renders this RNA processing signal non-functional in S. cerevisiae has the equivalent effect in S. pombe. We demonstrate that synthetic pre-mRNAs encoding polyadenylation sites of S. pombe genes are accurately cleaved and polyadenylated in whole cell extracts of S. cerevisiae. Finally, as is the case in S. cerevisiae, DNA sequences encoding regions proximal to the S. pombe mRNA 3' ends are found to be extremely AT rich; however, no general sequence motif can be found. We conclude that although fission yeast has many genetic features in common with higher eukaryotes, mRNA 3' end formation is significantly different and appears to be formed by an RNA processing mechanism homologous to that of budding yeast. Since fission and budding yeast are evolutionarily divergent, this lower eukaryotic mechanism of mRNA 3' end formation may be generally conserved.     

Control of cleavage site selection during mRNA 3'' end formation by a yeast hnRNP.

Endonucleolytic cleavage of pre-mRNAs is the first step during eukaryotic mRNA 3' end formation. It has been proposed that cleavage factors CF IA, CF IB and CF II are required for pre-mRNA 3' end cleavage in yeast. CF IB is composed of a single polypeptide, Nab4p/Hrp1p, which is related to the A/B group of metazoan heterogeneous nuclear ribonucleoproteins (hnRNPs) that function as antagonistic regulators of 5' splice site selection. Here, we provide evidence that Nab4p/Hrp1p is not required for pre-mRNA 3' end endonucleolytic cleavage. We show that CF IA and CF II devoid of Nab4p/Hrp1p are sufficient to cleave a variety of RNA substrates but that cleavage occurs at multiple sites. Addition of Nab4p/Hrp1p prevents these alternative cleavages in a concentration-dependent manner, suggesting an essential and conserved role for some hnRNPs in pre-mRNA cleavage site selection.     

Point mutations in the stem-loop at the 3'' end of mouse histone mRNA reduce expression by reducing the efficiency of 3'' end formation.

Mammalian histone mRNAs end in a highly conserved stem-loop structure, with a six-base stem and a four-base loop. We have examined the effect of mutating the stem-loop on the expression of the histone mRNA in vivo by introducing the mutated histone genes into CHO cells by stable transfection. Point mutations have been introduced into the loop sequence and into the UA base pair at the top of the stem. Changing either the first or the third base of the conserved UYUN sequence in the loop to a purine greatly reduced expression, while changing both U's to purines abolished expression. A number of alterations in the stem sequence, including reversing the stem sequence, reversing the two base pairs at the base of the stem, or destroying the UA base pair at the top of the stem, also abolished expression. Changing the UA base pair to a CG or a UG base pair also reduced expression. The loss of expression is due to inefficient processing of the pre-mRNA, as judged by the efficiency of processing in vitro. Addition of a polyadenylation site or the wild-type histone processing signal downstream of a mutant stem-loop resulted in rescuing the processing of the mutant pre-histone mRNA. These results suggest that if the histone pre-mRNA is not rapidly processed, then it is degraded.     

RNA-RNA interaction is required for the formation of specific bicoid mRNA 3'' UTR-STAUFEN ribonucleoprotein particles.

The formation of the anterior pattern of the Drosophila embryo is dependent on the localization of the mRNA of the morphogen Bicoid (bcd) to the anterior pole of the egg cell. Staufen protein (STAU) is required in a late step of the localization to anchor the bcd mRNA in the anterior cytoplasm. We have shown previously that endogenous STAU associates specifically with injected bcd mRNA 3'-untranslated region (UTR), resulting in the formation of characteristic RNA-protein particles that are transported along microtubules of the mitotic spindles in a directed manner. The regions recognized by STAU in this in vivo assay are predicted to form three stem-loop structures involving large double-stranded stretches. Here, we show that the STAU interaction requires a double-stranded conformation of the stems within the RNA localization signal. In addition, base pairing between two single-stranded loops plays a major role in particle formation. This loop-loop interaction is intermolecular, not intramolecular; thus dimers or multimers of the RNA localization signal must be associated with STAU in these particles. The bcd mRNA 3' UTR can also dimerize in vitro in the absence of STAU. Thus, in addition to RNA-protein interactions, RNA-RNA interaction might be involved in the formation of ribonucleoprotein particles for transport and localization.     

Chloroplast mRNA 3'' end processing requires a nuclear-encoded RNA-binding protein.

The protein coding regions of plastid mRNAs in higher plants are generally flanked by 3' inverted repeat sequences. In spinach chloroplast mRNAs, these inverted repeat sequences can fold into stem-loop structures and serve as signals for the correct processing of the mature mRNA 3' ends. The inverted repeat sequences are also required to stabilize 5' upstream mRNA segments, and interact with chloroplast protein in vitro. To dissect the molecular components involved in chloroplast mRNA 3' end processing and stability, a spinach chloroplast protein extract containing mRNA 3' end processing activity was fractionated by FPLC and RNA affinity chromatography. The purified fraction consisted of several proteins and was capable of processing the 3' ends of the psbA, rbcL, petD and rps14 mRNAs. This protein fraction was enriched for a 28 kd RNA-binding protein (28RNP) which interacts with both the precursor and mature 3' ends of the four mRNAs. Using specific antibodies to this protein, the poly(A) RNA-derived cDNA for the 28RNP was cloned and sequenced. The predicted amino acid sequence for the 28RNP reveals two conserved RNA-binding domains, including the consensus sequences RNP-CS1 and CS2, and a novel acidic and glycine-rich N-terminal domain. The accumulation of the nuclear-encoded 28RNP mRNA and protein are developmentally regulated in spinach cotyledons, leaves, root and stem, and are enhanced during light-dependent chloroplast development. The general correlation between accumulation of the 28RNP and plastid mRNA during development, together with the result that depletion of the 28RNP from the chloroplast protein extract interferes with the correct 3' end processing of several chloroplast mRNAs, suggests that the 28RNP is required for plastid mRNA 3' end processing and/or stability.     

Requirements for accurate and efficient mRNA 3'' end cleavage and polyadenylation of a simian virus 40 early pre-RNA in vitro.

Using a pre-RNA containing the simian virus 40 early introns and poly(A) addition site, we investigated several possible requirements for accurate and efficient mRNA 3' end cleavage and polyadenylation in a HeLa cell nuclear extract. Splicing and 3' end formation occurred under the same conditions but did not appear to be coupled in any way in vitro. Like splicing, 3' end cleavage and polyadenylation each required Mg2+, although spermidine could substitute in the cleavage reaction. Additionally, cleavage of this pre-RNA, but not others, was totally blocked by EDTA, indicating that structural features of pre-RNA may affect the ionic requirements of 3' end formation. The ATP analog 3' dATP inhibited both cleavage and polyadenylation even in the presence of ATP, possibly reflecting the coupled nature of these activities. A 5' cap structure appears not to be required for mRNA 3' end processing in vitro because neither the presence or absence of a 5' cap on the pre-RNA nor the addition of cap analogs to reaction mixtures had any effect on the efficiency of 3' end processing. Micrococcal nuclease pretreatment of the nuclear extract inhibited cleavage and polyadenylation. However, restoration of activity was achieved by addition of purified Escherichia coli RNA, suggesting that the inhibition caused by such a nuclease treatment was due to a general requirement for mass of RNA rather than to destruction of a particular nucleic acid-containing component such as a small nuclear ribonucleoprotein.