The methylated constituents of L cell messenger RNA: evidence for an unusual cluster at the 5' terminus.

An analysis of the methylated constituents of L cell mRNA by a combination of chromatographic methods and enzymatic treatments indicates that they comprise both 2'-O-methyl nucleosides and N6-methyl adenine, and/or 1-methyl adenine, and suggests that the 2'-O-methyl nucleotides, Ym, are part of an unusual class of sequences forming the 5' terminus of mRNA. These sequences seem to contain two 2'-O-methyl residues and a terminal residue that is not phosphorylated but, nevertheless, is blocked with respect to polynucleotid kinase reactivity. A strong candidate is a sequence of the type XppY1mpY2mpZp..., where X represents a blocking group which is itself occasionally methylated. The sequences isolated from total poly(A)+ mRNA contain all four species of 2'-O-methylated nucleoside, indicating some variability among different mRNA species. The methylated sequences do not appear to be enriched in the mRNA which hybridizes with repetitive DNA. The average L cell mRNA molecule also contains three residues of N6-methyl adenine. These residues are not part of the poly(A) segment, but appear to be located internal to the poly(A) near the 3' end of the mRNA molecules.     

Methylated messenger RNA in mouse kidney.

Polyadenylated messenger RNA from mouse kidney labeled in vivo exhibited a pattern of methylation distinct from that of rRNA and tRNA. After mice were given L-[methyl-3H]methionine, 4% of the polyribosomal RNA label was bound to oligo (dT)-cellulose; 20-24% of orotate- or adenine-labeled polyribosomal RNA eluted in the poly(A)+ RNA fraction under similar conditions. [3H]Methyl radioactivity was not incorporated into low molecular weight (5-5.8 S) rRNA, indicating the extent of nonmethylpurine ring labeling was negligible. [3H]Methyl-labeled poly(A)+ RNA sedimented heterogeneously in sodium dodecyl sulfate containing gradients similarly to poly(A)+ mRNA labeled with [3H]orotic acid. Based on an average molecular length of 2970 nucleotides, renal mRNA was estimated to contain 8.6 methyl moieties per molecule. Analysis of alkaline-hydrolyzed RNA sampled by DEAE-Sephadex-urea chromatography provided estimates of the relative amounts of base and ribose methylation. Although 83% of the [3H]methyl radioactivity in rRNA was in the 2'-0-methylnucleotide fraction, no methylated dinucleotides were found in mRNA. In poly(A)+ mRNA 60% of the [3H]methyl label was in the mononucleotide fraction; the remainder eluted between the trinucleotide and tetranucleotide markers and had a net negative charge between -4 and -5. The larger structure, not yet charcterized, could result from two or three consecutive 2'-0-ribose methylations and is estimated to contain 2.6 methyl residues. Alternatively, the oligonucleotide could be a 5'-terminal methylated nucleotide species containing 5'-phosphate(s) in addition to the 3'-phosphate moiety resulting from alkaline hydrolysis. Either structure could have a role in the processing or translation of mRNA in mammalian cells.     

Heterogeneity of the 5' terminus of hen ovalbumin messenger ribonucleic acid.

The 5'-terminal sequence of hen ovalbumin mRNA was investigated using a novel labeling method. Ovalbumin mRNA was purified by hybridization to complementary DNA coupled to cellulose. The mRNA thus purified was shown to be 97.9% pure by hybridization with plasmid DNA containing sequences to the messengers coding for conalbumin and ovomucoid, the next two most abundant messengers of oviduct. After digestion with RNase T1 and alkaline phosphatase, 5'-terminal capped oligonucleotides were selected by binding to anti-m7G-Sepharose. These were then labeled using RNA ligase and [5'-32P]pCp, separated by two-dimensional gel electrophoresis, and sequenced by partial digestion with base-specific ribonucleases. A nested set of three capped oligonucleotides was identified. Their structures and relative abundances were m7GpppAUACAG, 3% m7GpppACAUACAG, 61+; and m7GpppGUACAUACAG, 36%.     

H4 histone messenger RNA decay in cell-free extracts initiates at or near the 3' terminus and proceeds 3' to 5'

The relative decay of four human messenger RNAs, gamma globin, delta globin, c-myc and H4 histone, were compared in a cell-free system. Under appropriate conditions, they are degraded in vitro in approximately the same relative order as in vivo: histone faster than c-myc and delta globin faster than gamma globin. Degradation of polysome-associated H4 histone mRNA and of deproteinized histone mRNA begins at or near the 3' terminus. At least a portion of the mRNA then continues to be degraded in a 3' to 5' direction. Discrete 3'-terminal degradation hold-up points are observed, suggesting that 3' to 5' degradation occurs non-uniformly. Cycloheximide and puromycin inhibit protein synthesis but do not affect the rate or directionality of histone mRNA decay in vitro. We conclude that the rate-limiting step in H4 histone mRNA decay occurs at or near the 3' terminus and that at least a portion of the mRNA molecule is subsequently degraded 3' to 5', probably via a processive exonuclease.     

m7G5''ppp5''GmptcpUp at the 5'' terminus of reovirus messenger RNA.

In the presence of S-adenosyl methionine the 5' terminal guanosine residue of in vitro synthesized reovirus mRNA becomes methylated at the 2'-OH position. In addition, 7-methyl guanylic acid is condensed covalently at the 5' terminus resulting in the formation of a 5' to 5' triphosphate bridge. Analysis of the 5' terminal sequence of methylated reovirus mRNA revealed that it has the structure m7G5'ppp5'GmpCpUp.     

Structure of the 5' terminus of hen oviduct lysozyme messenger ribonucleic acid

Lysozyme mRNA (mRNAlys) was purified from hen oviduct poly(A)-containing RNA by hybridization, labeled with NaB[3H]4 and digested with RNase T1. This revealed the presence of equal amounts of two major oligonucleotides having structures of m7Gppp(Np)7 and m7Gppp(Np)4 plus minor amounts of m7Gppp(Np)2 and m7GpppNp. The total mRNAlys contained the cap structures m7Gpppm6Am, m7GpppGm, m7GpppAm, m7GpppCm, m7GpppA, and m7GpppG, in decreasing order of abundance. The m7Gppp(Np)7 oligonucleotide contained only A-caps and the m7Gppp(Np)4, only G-caps. 32P-labeled 5'-terminal T1-oligonucleotides were prepared, and at least 12 different types were observed, the most abundant being m7Gppp(Np)7 and m7Gppp(Np)4. Their sequences were determined to be m7Gppp(m6)AmNmUCCCG and m7GpppGmNmAG. Taken together with the findings of Grez et al. (Grez, M., Land, H., Giesecke, K., Schutz, G., Jung, A., and Sippel, A. E. (1981) Cell 25, 743-752), these results indicate that in the genomic sequence AGCTTGCAGTCCCGT, 52% of the mRNAlys molecules begin at the underlined A residue and 38% at the underlined G residue.     

Methylated constituents of heterogeneous nuclear RNA: presence in blocked 5' terminal structures.

A substantial portion of the hnRNA of mouse L cells contains internal residues of N6-methyl adenylate and blocked 5' terminal sequences which are apparently of the type m7G5' ppp5' YmpZp..., in which 7-methyl guanosine is joined by a 5'-5' pyrophosphate linkage to a 2'-0-methylated residue, Ym. These sequences are indistinguishable from those comprising one of the two classes of blocked 5' sequences found in mRNA, and are quite distinct from those comprising the other class. The remarkable similarity in 5' terminal methylated sequences of hnRNA and a major fraction of mRNA appears to extend even to the relative occurrence of each of the four 2'-0-methylated species in position Ym.     

Methylated constituents of Aedes albopictus poly (A)-containing messenger RNA.

Poly (A)-containing mRNA prepared from cultured mosquito (Aedes albopictus) cells was found to contain methylated 5'-terminal "caps" as well as internal m6A residues. Both type I [m7G(5')ppp(5')Xmp] and type II [m7G(5')ppp(5')XmpYmp] caps were present, at molar ratio of ca five to one. All four common RNA bases were represented in the second position (Xm) of the caps, adenine being the most abundant and N6-methyladenine being absent. The four bases were also represented in the third position (Ym), but here uracil was the predominant base. There was approximately one internal m6A residue for every three caps. These studies demonstrate that mRNA from an invertebrate source can have a methylation pattern comparable with that of mammalian cells in it complexity.     

At least 104 nucleotides are transposed from the 5' terminus of the avian sarcoma virus genome to the 5' termini of smaller viral mRNAs.

Cells producing avian sarcoma virus (ASV) contain at least three virus-specific mRNAs, two of which are encoded within the 3' half of the viral genome. Each of these viral RNAs can hybridize with single-stranded DNA(cDNA5') that is complementary to a sequence of 101 nucleotides found at the 5' terminus of the ASV genome, but not within the 3' half of the genome. We proposed previously (Weiss, Varmus and Bishop, 1977) that this nucleotide sequence may be transposed to the 5' termini of viral mRNAs during the genesis of these RNAs. We now substantiate this proposal by reporting the isolation and chemical characterization of the nucleotide sequences complementary to cDNA5' in the genome and mRNAs of the Prague B strain of ASV. We isolated the three identified classes of ASVmRNA (38, 28 and 21S) by molecular hybridization; each class of RNA contained a "capped" oligonucleotide identical to that found at the 5' terminus of the ASV genome. When hybridized with cDNA5', each class of RNA gave rise to RNAase-resistant duplex hybrids that probably encompassed the full extent of cDNA5'. The molar yields of duplex conformed approximately to the number of virus-specific RNA molecules in the initial samples; hence most if not all of the molecules of virus-specific RNA could give rise to the duplexes. The duplexes prepared from the various RNAs all contained the capped oligonucleotide found at the 5' terminus of the viral genome and had identical "fingerprints" when analyzed by two-dimensional fractionation following hydrolysis with RNAase T1. In contrast, RNA representing the 3' half of the ASV genome did not form hybrids with cDNA5'. We conclude that a sequence of more than 100 nucleotides is transposed from the 5' end of the ASV genome to the 5' termini of smaller viral RNAs during the genesis of these RNAs. Transposition of nucleotide sequences during the production of mRNA has now been described for three families of animal viruses and may be a common feature of mRNA biogenesis in eucaryotic cells. The mechanism of transposition, however, and the function of the transposed sequences are not known.     

Comparison of proteins bound to heterogeneous nuclear RNA and messenger RNA in HeLa cells.

Ribonucleoprotein particles containing either heterogeneous nuclear RNA or polyribosomal messenger RNA were isolated from growing HeLa cells in order to compare their respective protein components. The major obstacle to analysing the proteins bound to HeLa cell mRNA proved to be the cosedimentation of a large fraction of the mRNP² particles with ribosomal subunits following puromycin or EDTA disassembly of polyribosomes. This was circumvented by oligo(dT)-cellulose chromatography, in which essentially all of the ribosomal subunits passed through the column without retention, while approximately 80% of the pulse-labeled, poly(A)-containing mRNP became bound and could be eluted with formamide. Polyacrylamide gel electrophoresis of the non-bound fraction (ribosomal subunits) revealed polypeptides between 15,000 and 55,000 molecular weight, with no detectable components greater than 55,000. The oligo-(dT)-bound mRNP contained a much simpler protein complement, consisting of three major components having molecular weights of 120,000, 76,000 and 52,000.In the case of the nuclear ribonucleoprotein particles that contain heterogeneous nuclear RNA, oligo(dT)-cellulose chromatography revealed two classes of particles. The first contained 10 to 20% of the hnRNA, did not bind to oligo(dT)-cellulose in 0.25 m-NaCl, 10 mm-sodium phosphate buffer, pH 7.0 (4 °C), and contained primarily a single polypeptide component having an estimated molecular weight of 40,000 (“informofers”). A second population of hnRNP particles comprised approximately 80% of the hnRNA, displayed strong binding to oligo(dT)-cellulose at 0.25 m-NaCl, and contained a very complex population of proteins, having molecular weights between 40,000 and 180,000, the same as unfractionated hnRNP. The results indicate that, at the resolution of gel electrophoresis and at the sensitivity of Coomassie blue dye, the proteins bound to HeLa cell hnRNA are qualitatively distinct from those bound to polyribosomal mRNA and, in addition, that the hnRNP proteins are the more complex of the two. These results are discussed in relation to the possible nucleotide sequence elements in hnRNA and mRNA to which these specific proteins are bound.     

A new RNA helicase isolated from HeLa cells that catalytically translocates in the 3' to 5' direction.

We have purified an RNA helicase to near homogeneity from nuclear extracts of HeLa cells. The enzyme migrated as a 130-kDa protein upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis and exhibited a sedimentation coefficient of 6.4 on glycerol gradient centrifugation. The enzyme translocated in a 3' to 5' direction and acted catalytically, displacing at least a 4-fold molar excess of duplex RNA compared with the enzyme added. All eight common nucleoside triphosphates supported RNA helicase activity at relatively low concentrations (Km in values in the 15-20 microM level). In the presence of RNA and some single-stranded DNAs, the RNA helicase hydrolyzed all nucleoside triphosphates to nucleoside diphosphates and inorganic phosphate. The enzyme displaced deoxyribooligonucleotides provided they were hydrogen-bonded to RNA possessing 3' single-stranded regions, but it did not displace ribooligonucleotides hydrogen-bonded to DNA containing 3' single-stranded regions. The enzyme, in the absence of ATP, binds to both single-stranded RNA and DNA, but the amount of complex formed with RNA was 20-fold greater than the complex formed with DNA. In both cases, the complex formed in the absence of ATP was rapidly reversed by the addition of ATP and not by adenyl-5'-yl (beta,gamma-methylene)-diphosphate. We propose that the enzyme can bind to both single-stranded RNA and DNA and hydrolyze ATP, but by virtue of its greater stability on RNA, the enzyme can only translocate on RNA possessing 3' single-stranded regions.     

Differential enzymatic accessibilities of the 5' and 3' splice sites of beta-globin pre-messenger RNA in splicing competent HeLa cell nuclear extract

Inhibition of oligonucleotide-directed cleavage of pre-mRNA using exogenously added E. coli RNase H has been utilized as a probe for mRNA-protein interaction. We now show that such an RNase H-like activity is present in splicing competent Hela cell nuclear extract. Using this extract and in vitro transcribed beta-globin pre-mRNA, we have demonstrated that synthetic oligonucleotides, complementary to the splice site sequences, direct preferential cleavage of the 5' splice site. Thus, these experiments using complementary oligonucleotide-directed, endogenous RNase H-like cleavage of pre-mRNA, suggest a useful probe for studying the mRNA-protein complex in vitro.     

Do eukaryotic mRNA 5' noncoding sequences base-pair with the 18 S ribosomal RNA 3' terminus?

Protein synthesis initiation on prokaryotic mRNAs involves base-pairing of a site preceding the initiation codon with the 3' terminal sequence of 16 S rRNA. It has been suggested that a similar situation may prevail in eukaryotic mRNAs. This suggestion is not based on experiments, but on observation of complementarities between mRNA 5' noncoding sequences and a conserved sequence near the 18 S rRNA 3' terminus. The hypothesis can be evaluated by comparing the number of potential binding sites found in the 5' noncoding sequences with the number of such sites expected to occur by chance. A method for computing this number is presented. The 5' noncoding sequences contain more binding sites than expected for a random RNA chain, but the same is true for 3' noncoding sequences. The effect can be traced to a clustering of purines and pyrimidines, common to noncoding sequences. In conclusion, a close inspection of the available mRNA sequences does not reveal any indication of a specific base-pairing ability between their 5' noncoding segments and the 18 S rRNA 3' terminus.