The primary structure of rabbit beta-globin mRNA as determined from cloned DNA

The structure of the human zeta-globin gene and a closely linked, nearly identical pseudogene

DNA sequencing studies indicate that only one of two closely linked human embryonic alpha-like globin genes, zeta (zeta), encodes a functional polypeptide. The other is a pseudogene (psi zeta) that differs by only 3 bp in the protein coding sequence, one of which converts the codon for amino acid 6 into a chain termination codon. Both zeta-globin genes differ from all other alpha-like genes thus far reported in that they contain large introns consisting, in part, of simple repeat sequences. Intron 1 of each gene contains a variation of the repeat sequence ACAGTGGGGAGGGG, while intron 2 contains the repeat sequence CGGGG. Comparison of the human zeta- and alpha-globin gene sequences reveals that the embryonic and adult alpha-like genes began to diverge from each other relatively early in vertebrate evolution (400 million years ago). In contrast, the beta-like embryonic globin gene, epsilon (epsilon), is the product of a much more recent evolutionary event (200 million years ago). Thus, even though the temporal and quantitative expression of zeta- and epsilon-globin genes must be coordinately controlled during development, their evolutionary histories are clearly distinct.

Purification of the mRNAs encoding the subunits of larval serum proteins 1 and 2 of Drosophila melanogaster

The messenger RNAs for the four subunits of larval serum proteins (LSPs) 1 and 2 were purified from third instar fat body total RNA by fractionating oligo(dT)-selected RNAs on the basis of size. Translations of RNA fractions in vitro showed that the LSP mRNAs were free of other fat body messenger activities, but they could not be completely resolved from each other. The molecular weights of the LSP mRNAs, estimated from their electrophoretic mobilities in 50% formamide/1.4% agarose gels, ranged from 1.0 X 10(6) to 1.35 X 10(6). These estimates are at least 20% larger than the size of RNA necessary to encode each LSP polypeptide.

High-efficiency cloning of full-length cDNA

A widely recognized difficulty of presently used methods for cDNA cloning is obtaining cDNA segments that contain the entire nucleotide sequence of the corresponding mRNA. The cloning procedure described here mitigates this shortcoming. Of the 10(5) plasmid-cDNA recombinants obtained per microgram of rabbit reticulocyte mRNA, about 10% contained a complete alpha- of beta-globin mRNA sequence, and at least 30 to 50%, but very likely more, contained the entire globin coding regions. We attribute the high efficiency of cloning full- or nearly full-length cDNA to (i) the fact that the plasmid DNA vector itself serves as the primer for first- and second-strand cDNA synthesis, (ii) the lack of any nuclease treatment of the products, and (iii) the fact that one of the steps in the procedure results in preferential cloning of recombinants with full-length cDNA's over those with truncated cDNA's.

Polyoma virus early and late mRNAs in productively infected mouse 3T6 cells

We mapped polyoma virus-specific mRNAs isolated from productively infected mouse 3T6 cells on the viral genome by analyzing nuclease S1-resistant RNA-DNA hybrids. The polyoma early mRNAs, which code for the three T antigens, have several 5' ends near 73 map units (m.u.). During the late phase of infection an additional 5' end is found near 71 m.u. All of the major early mRNAs have common 3' ends at 26.01 m.u. There is a minor species of early mRNA with a 3' end at 99.05 m.u. There are two proximal and two distal splice junctions in the early region which are used to generate three different spliced early mRNAs. There are three late mRNAs encoding the three virion proteins, VP1, VP2, and VP3. The late mRNAs have common 3' ends at 25.34 m.u. The late mRNAs have heterogeneous 5' leader sequences derived from the region between 65.53 and 68.42 m.u. The leader sequences are joined to the bodies of the messages coding for VP2, VP3, and VP1 at 66.59, 59.62, and 48.57 m.u., respectively. These results confirm and extend previous analyses of the fine structure of polyoma mRNAs.

RNA folding is unaffected by the nonrandom degenerate codon choice

The frequent suggestion that the nonrandom codon usage is explained by its forming more stable mRNAs is tested in 22 genes. Only the histones, globins, and the rat preproinsulin gene show a correlation between the preferred degenerate codons and the stability of the secondary structure of the their mRNAs. However, the examined members from the histone and globin gene families, both among the oldest, in evolutionary sense, eukaryotic genes, have a high GC content (approx. 56% compared to an average of 42% in all eukaryotes) which is reflected in their degenerate codon choice and thus in their more stable folding.

Cloning and sequence analysis of cDNA for porcine beta-neo-endorphin/dynorphin precursor

The primary structure of a precursor protein that contains beta-neo-endorphin, dynorphin and a third leucine-enkephalin sequence with a carboxyl extension has been deduced from the nucleotide sequence of cloned DNA complementary to the porcine hypothalamic mRNA encoding it. The three peptides are each bounded by Lys-Arg. This precursor protein, like adrenal preproenkephalin and the corticotropin/beta-lipotropin precursor, comprises multiple repetitive units and a cysteine-containing amino-terminal sequence preceded by a signal peptide.

Lysine tRNAs from Bacillus subtilis 168: function of the isoacceptors in a rabbit reticulocyte cell-free protein-synthesizing system

The two principal tRNA Lys isoaccepting species of Bacillus subtilis were compared in their functional activity in translating rabbit globin. Although neither species demonstrates any preference in reading either of the lysine codons, there is an overall preference for tRNa Lys 3 in lysine incorporation. The ratios of lysine incorporated by the two species into the different lysine-containing sites in the globin subunits vary over a more than two-fold range. As described in the accompanying paper, tRNA Lys 1 is a hypomodified form of tRNA Lys 3. Consistent with studies on other rRNA species, the fully modified isoacceptor functions preferentially. In contrast to these results, however, the fully modified isoacceptor (tRNA Lys 3) is found predominantly in rapidly dividing cells while the hypomodified isoacceptor (tRNA Lys 1) predominates in the stationary cells and spores of B.l subtilis.

Rat beta casein cDNA: sequence analysis and evolutionary comparisons

The complete sequence of a 1072 nucleotide rat beta-casein cDNA insertion in the hybrid plasmid pC beta 23 has been determined. Primer extension was employed to determine the sequence of an additional 82 5'-terminal nucleotides in beta-casein mRNA. Rat beta-casein mRNA consists of a 696 nucleotide coding region, flanked by 52 nucleotide 5' and 406 nucleotide 3' noncoding regions, including a 40 nucleotide poly(A) tail. The derived 216 amino acid sequence of rat beta-casein was compared to the previously determined sequences of beta-caseins from several other species. Approximately 38% of the amino acids have been conserved among the rat, ovine, bovine and human sequences and these conserved amino acids occurred in clusters throughout the protein. One such cluster containing the majority of the potential casein phosphorylation sites was located near the amino terminus. Contrary to the considerable divergence observed for the processed beta-casein, 14 of 15 amino acids in the signal peptide sequence of the precasein were identical between the rat and ovine caseins.

High-efficiency cloning of full-length cDNA

A widely recognized difficulty of presently used methods for cDNA cloning is obtaining cDNA segments that contain the entire nucleotide sequence of the corresponding mRNA. The cloning procedure described here mitigates this shortcoming. Of the 10(5) plasmid-cDNA recombinants obtained per microgram of rabbit reticulocyte mRNA, about 10% contained a complete alpha- of beta-globin mRNA sequence, and at least 30 to 50%, but very likely more, contained the entire globin coding regions. We attribute the high efficiency of cloning full- or nearly full-length cDNA to (i) the fact that the plasmid DNA vector itself serves as the primer for first- and second-strand cDNA synthesis, (ii) the lack of any nuclease treatment of the products, and (iii) the fact that one of the steps in the procedure results in preferential cloning of recombinants with full-length cDNA's over those with truncated cDNA's.

Nucleotide sequence of cloned cDNA encoding bovine arginine vasopressin-neurophysin II precursor

The sequence of a cDNA encoding the nonapeptide arginine vasopressin (AVP) and its carrier protein, neurophysin II (NpII) from bovine hypothalamus, proves that the 166-amino acid precursor molecule contains a signal peptide of 19 amino acids followed directly by AVP connected to NpII by a Gly-Lys-Arg sequence. The carboxy-terminal region of the precursor contains a naturally occurring glycopolypeptide of 39 amino acids which is separated from NpII by a single arginine residue.

The nucleotide sequence of the mouse embryonic beta-like y-globin messenger RNA as determined from cloned cDNA

We have determined the nucleotide sequence of two cloned cDNAs corresponding to the mRNA of mouse embryonic y2 globin. The combined overlapping sequences span a total of 480 bp, beginning at the codon corresponding to amino acido residue 21 and extending to the AATAAA sequence in the 3' untranslated region. Therefore, when the amino acid sequence encoded by the cDNA is combined with the available amino acid sequence, a complete y2 protein sequence can be obtained. Comparisons, at the nucleotide level, between the known beta- and beta-like globin sequences and the y2 sequence show that the embryonic, fetal-adult duplication occurred approx. 160 million years (MY) ago and that the embryonic-fetal duplication occurred approx. 100 MY ago.

Nucleotide-sequence heterogeneity and sequence rearrangements in influenza virus cDNA

Double-stranded cDNA has been synthesized from influenza virus RNA and cloned into derivatives of the bacteriophage M13 for sequence analysis. The characterization of over 200 clones has permitted an analysis both of nucleotide sequence heterogeneity and of clones containing unusual rearrangements of sequence. Heterogeneity, due to genetic variability in the RNA population and to in vitro synthetic errors, was detected at the low level of one nucleotide difference per 3 700 nucleotides. By contrast, gross sequence rearrangements were identified in eight clones. Inversions of sequence within the same cDNA molecule were the predominant type of rearrangement, and three mechanisms for producing such inversions are discussed. In addition, we observed rarer clones containing sequence from one RNA molecule joined to that from another molecule.

Nucleotide sequence of adenovirus 2 DNA fragment encoding for the carboxylic region of the fiber protein and the entire E4 region

The entire nucleotide sequence between coordinates 89.5 and 100% of the Ad 2 DNA genome has been determined using the Maxam and Gilbert method. This sequence of 3766 bp contains information relative to the carboxylic end of the fiber protein and to the entire E4 region. The position within the nucleotide sequence of various open reading frames and of several consensus splicing sequences was correlated with the location by EM and Sl digestion of the E4 mRNA. This correlation allows to suggest an additional splicing event in the maturation process of i or f mRNA and to deduce the structure of most E4 mRNA. The aminoacid sequences of the corresponding proteins are deduced allowing the location of several glycosylation sites. The presence of several open reading frames with a substantial coding capacity permits to postulate on the existence of additional genes located at the 3' end of the fiber gene and the 3' end of the E4 region. The existence of these putative additional genes might explain that termination of transcription is several hundred nucleotides beyond the main known poly A addition sites of the L5 and E4 regions.

Structure and properties of a bovine liver UGA suppressor serine tRNA with a tryptophan anticodon

A bovine liver serine tRNA with a variety of unusual features has been sequenced and characterized. This tRNA is aminoacylated with serine, although it has a tryptophan anticodon CmCA. In ribosome binding assays, this tRNA (tRNASERCmCA) binds to the termination codon UGA and shows little or no binding in response to a variety of other codons including those for tryptophan and serine. The unusual codon recognition properties of this molecule were confirmed in an in vitro assay where this tRNA suppressed UGA termination. This is the first naturally occurring eucaryotic suppressor tRNA to be so characterized. Other unusual features, possibly related to the ability of this tRNA to read UGA, are the presence of two extra nucleotides, compared to all other tRNAs, between the universal residues U at position 8 and A at position 14 and the presence of an extra unpaired nucleotide within the double-stranded loop IV stem. This tRNA is also the largest eucaryotic tRNA sequenced to date (90 nucleotides). Despite its size, however, it contains only six modified residues, tRNASerCmCA shows extremely low homology to other mammalian serine (47-52% homology) or tryptophan (49% homology) tRNAs.

Biology of hepatitis B virus

Immunochemical investigations of the viral antigens and molecular characterization of the viral DNA have elucidated the nature of the hepatitis B virus infection underlying acute, chronic, and oncogenic disorders of the liver in man. Cloning and sequencing of viral DNA have made possible studies on the structure of the genome and on certain aspects of the biology of the virus, hitherto constrained for a lack of tissue culture systems and laboratory animal models useful in its propagation.

Introduction by molecular cloning of artifactual inverted sequences at the 5' terminus of the sense strand of bovine parathyroid hormone cDNA

To study the structure and function of the gene for parathyroid hormone, we obtained recombinant plasmids containing bovine parathyroid hormone cDNA. The nucleotide sequence at the 5' terminus (relative to the sense strand) of the cDNA insert in a recombinant plasmid, pPTHi4, was different from that previously reported for the bovine parathyroid hormone cDNA insert of another recombinant plasmid, pPTHm1 [Kronenberg, H. M., McDevitt, B. F., Majzoub, J. A., Nathans, J., Sharp, P. A., Potts, J. T., Jr. & Rich, A. (1979) Proc. Natl. Acad. Sci. USA 76, 4981-4985]. The first 50 nucleotides of the pPTHm1 insert were an inverted complement of nucleotides 2-51 of the pPTHi4 insert. the cDNA insert of another plasmid, pPTHi8, contained a sequence identical to nucleotides 2-51 of the pPTHi4 insert but also contained an additional 42 bases at the 5' terminus. The first 41 bases of the pPTHi8 insert were an inverted repeat of an internal sequence of the pPTHi4 insert corresponding to nucleotides 184-224. Restriction endonuclease analysis of pPTHi8 indicated that the internal sequence corresponding to this region was retained. The nucleotide sequence of a restriction fragment hybridized to parathyroid hormone mRNA and extended toward the 5' terminus of the mRNA with reverse transcriptase confirmed that the sequence at the 5' terminus of the pPTHi4 insert was an accurate copy of the parathyroid hormone mRNA sequence. These data suggest that two types of sequence rearrangements may occur at the 5' terminus, as occurred in pPTHm1, and (ii) an inverted repeat of an internal sequence, as occurred in pPTHi8.

The universal dinucleotide asymmetry rules in DNA and the amino acid codon choice

Natural DNA sequences were recently found to contain distinct nearest neighbor patterns. Hetero-dinucleotides were demonstrated to appear consistently more (less) than their mirror-image counterparts. This paper shows that this asymmetric behavior does not stem from the coding requirements of the DNA. It also shows some codon patterns in prokaryotic and eukaryotic genomes which came up in the course of this work.

Nucleotide sequence analysis of a cloned duck beta-globin cDNA

The primary structure of the duck beta-globin mRNA was obtained from sequence analysis of the double-stranded in vitro-transcribed DNA cloned in plasma pBR322. The 646-bp long globin DNA insert comprises a coding sequence of 438 bp corresponding to 146 amino acids, a 5'-noncoding region 63 bp long, and a 3'-noncoding region of 113 bp prior to a stretch of adenosine residues. The salient features of each of these regions are discussed and compared with beta-globin mRNAs of other vertebrates.

Electron microscopic evidence of circular molecules in 9-S globin mRNA from rabbit reticulocytes

9S globin mRNA prepared by the proteinase K method from polysomes of rabbit reticulocytes consists of 40% circular molecules as revealed by electron microscopy, if spreading of the molecules is performed from a solution of 50% formamide, 0.5 M NaCl, 25 mM Tris, 10 mM EDTA, pH 8, after 16 h incubation at 42 degrees C. We assume a noncovalent nature of the circularization because of the fact that a total transformation into the well known linear form occurs if strong denaturing conditions for spreading were used. The biological significance of the circular globin mRNA molecules is unknown.

Preferential codon usage in genes

We present a method which permits comparison of the preferential use of degenerate codons within any gene. The method makes use of the triplet frequencies in the noncoding frames to assess whether a preference is specific to the reading frame. Preference is given a statistical meaning by use of the analysis of variance coupled to Duncan's multiple range test. Preferential use of degenerate codons is gene-specific and independent of gene size. The data suggest that any correlation between codon frequency distribution and tRNA levels is unreliable. In those animal genes examined, codons ending in C or G are preferred; in animal viruses tested, codons ending in U or A are preferred. Similarly, the bacterial genes and the genes of single-stranded DNA phages that we analyzed differed from each other as well as from eukaryotic genes in the third base of the codon.

Complete nucleotide sequence of a chicken alpha-globin cDNA

The entire sequence of a 541 bp insert in recombinant plasmid pHb1003 has been determined. This plasmid, which was shown to carry a cloned cDNA copy of the chicken alpha-globin mRNA, contains the complete structural gene as well as 19 bp of the 5'-untranslated region and 99 bp of the 3'-untranslated region. This sequence may encode a non-adult alpha-globin gene, especially since the cDNA clones were generated from phenylhydrazine-induced, globin-specific mRNA extracted from anemic white leghorns. The possibility that this alpha-globin might represent a stress globin is considered.

Nucleotide sequence of the EcoRI E fragment of adenovirus 2 genome

The entire nucleotide sequence of the Ad.2 EcoRI E fragment has been determined using the Maxam and Gilbert method. This sequence of 2222 bp, which maps between coordinate 83.4 and 89.7 contains information relative to the early 3 region and to the fiber gene. Altogether with fragment EcoRI D which has been recently sequenced, they cover the entire Early 3 region in which several mRNA were mapped. The aminoacid sequence of the 16K and 14K protein is deduced. The localization of the 14.5K mRNA directing the synthesis of the third E3 known protein is discussed, as well as the hypothetical existence of three other early 3 proteins, which would have a molecular weight of 11K. The initiator ATG triplet of the fiber protein has been found at coordinate 86.1, it is followed up to the end of the fragment by an open reading frame allowing deduction of 80% of the aminoacid sequence of this protein. Sequences known to be frequently present at the border of exon sequence were used to tentatively localize the additional "Z" late leader.

Method to determine the reading frame of a protein from the purine/pyrimidine genome sequence and its possible evolutionary justification

The periodic variations obtained by correlating the relative positions of purines and pyrimidines (and of the four bases thymine, cytosine, adenine, and guanine) in a wide variety of genomes of wholly or partly known sequence suggest that there may be enough of an earlier comma-free coding system (i.e., only readable in one frame) still present to permit determination of the reading frame and approximate extent of the present protein coding stretches. The characteristics of these variations support the hypothesis that these primitive messages were formed of coding triplets having the form RNY (R = purine; Y = pyrimidine; and N = purine or pyrimidine). The base sequences and reading frames that have a minimal deviation from such a message are still good predictors of actual coding regions and reading frames in spite of the many mutations that have occurred since such a genetic code was last in use. In fact, the right frame for almost all the proteins in a number of viruses and various prokaryotes and eukaryotes is deduced purely from purine/pyrimidine information and not by using the normal start and stop signals.

The use of synthetic oligonucleotides as hybridization probes. II. Hybridization of oligonucleotides of mixed sequence to rabbit beta-globin DNA

Two oligonucleotides 14-bases long were synthesized, one complementary to rabbit beta-globin DNA (R beta G14A) and the other with the same sequence except for a single base change (T for C) (R beta G14B). Hybridization conditions were established such that R beta G14A would hybridize to globin DNA while R beta G14B would not. We also synthesized a mixture of 13-base long oligonucleotides (R beta G13Mix), representing eight of the possible coding sequences for amino acids 15-19 of rabbit beta-globin. One of the eight is complementary to globin DNA. R beta G13Mix was found to hybridize specifically to globin DNA under conditions where oligonucleotides forming single base pair mismatches do not. Furthermore, R beta G13Mix was shown to hybridize specifically to colonies containing a plasmid with a globin DNA insert. These results are discussed with respect to a general procedure for screening recombinant clones for those containing DNA coding for a protein of known amino acid sequence.

Nonrandom patterns of codon usage and of nucleotide substitutions in human alpha- and beta-globin genes: an evolutionary strategy reducing the rate of mutations with drastic effects?

Nucleotide substitutions within a structural gene can cause two principal "drastic" phenotypic effects at the protein level: translatable leads to untranslatable and nonpolar hydrophobic in equilibrium hydrophilic amino acid substitutions. The sequence of nucleotides in the structural human alpha- and beta-globin genes and their variants were examined to determine whether codon usage, patterns of nucleotide substitutions, or both, reduced the relative and absolute rates of these unfavorable mutations. Based on translation of abnormal hemoglobins, it is likely that all 61 nontermination codons are potentially translatable, though only 47 are normally used. Moreover, codons that can mutate to a termination codon are never used whenever the corresponding amino acid is specified also by triplets that cannot mutate to termination by a single-step mutation. Thus, the number of opportunities to mutate to an untranslatable codon is reduced to the minimum compatible with the amino acid composition of these chains. The relative rates of U in equilibrium non-U substitutions were much lower than those of other substitutions. Because U residues must be involved in most termination mutations and in all nonpolar hydrophobic in equilibrium hydrophilic amino acid substitutions, there is a considerable reduction of mutational events, causing drastic phenotypic effects. These findings are likely to be the end result of evolutionary selection by yet unknown mechanisms.

Estimation of evolutionary distances between homologous nucleotide sequences

By using two models of evolutionary base substitutions--"three-substitution-type" and "two-frequency-class" models--some formulae are derived which permit a simple estimation of the evolutionary distances (and also the evolutionary rates when the divergence times are known) through comparative studies of DNA (and RNA) sequences. These formulae are applied to estimate the base substitution rates at the first, second, and third positions of codons in genes for presomatotropins, preproinsulins, and alpha- and beta-globins (using comparisons involving mammals). Also, formulae for estimating the synonymous component (at the third codon position) and the standard errors are obtained. It is pointed out that the rates of synonymous base substitutions not only are very high but also are roughly equal to each other between genes even when amino acid-altering substitution rates are quite different and that this is consistent with the neutral mutation-random drift hypothesis of molecular evolution.

The old REH theory remains unsatisfactory and the new REH theory is problematical - a reply to Holmquist and Jukes

In response to criticism of REH theory (Fitch 1980), Holmquist and Jukes (1981) have mostly avoided the criticism or misunderstood it. Since they themselves state in their response that "Amino acid sequence data alone cannot be used to estimate total nucleotide substitutions," they agree with the criticism. Most of their paper treats the newer theory (here designated as the REHN theory) which attempts to use the nucleotide sequences encoding proteins to better estimate total nucleotide substitutions (Holmquist and Pearl 1980). Since I made no criticism of REHN theory, their comments are frequently beside the point of my original criticism of REH theory. Nevertheless, it is shown here that REHN theory is also unsatisfactory in that: One, the varions are now more clearly defined but in such a way as to preclude the same codon from suffering a nucleotide substitution in more than one evolutionary interval. Two, the set of codons that accepts silent substitutions is identical to the set that accepts amino acid changing nucleotide substitutions. Three, the uncertainty in the REH estimate is considerable in that alternative excellent fits to the same observational data may give alternative REH values that differ significantly even before stochastic variation and selective bias are considered. Four, the fit of their model to data is an irrelevancy where there are zero degrees of freedom.

5'-terminal nucleotide sequence of the messenger RNA coding for bovine corticotropin/beta-lipotropin precursor

The complete 5'-terminal nucleotide sequence of the MRNA coding for the bovine common precursor of corticotropin and beta-lipotropin has been determined. The 5'-32P-labelled, 21-nucleotides-long, single-stranded DNA fragment complementary to a portion of the 5'-noncoding region of the mRNA was prepared from a cDNA clone and elongated by reverse transcriptase reaction with the mRNA as template. The DNA transcript formed was sequenced by the procedure of Maxam and Gilbert, and the resultant sequence was cross-checked by two-dimensional electrophoretic analysis of the partial alkaline digest of the 5'-32P-labelled mRNA. The 5'-terminal nucleotide residue was determined by two-dimensional thin-layer chromatography of the complete hydrolysis product of the 5'-32P-labelled mRNA. The nucleotide sequence determined, which partially overlaps the known sequence of the cloned cDNA, reveals the complete 5'-terminal sequence of the mRNA. This, in conjunction with our previous data, defines the complete primary structure of the mRNA. The mRNA is composed of 1098 nucleotides, including an unusually long 5'-noncoding sequence of 128 nucleotides. The presence of a 'cap' structure at the 5' terminus of the mRNA is suggested. The 5'-terminal 48 nucleotide residues of the mRNA are extremely purine-rich, having an A + G content of 83%, whereas all pyrimidine-rich segments are located downstream from there. Because the 5'-noncoding region of the mRNA contains three segments of potential secondary structure which partially overlap, it can exist in a number of alternative base-pairing configurations. However, its interaction with the 3'-terminal segment of 18-S rRNA at the site of maximal complementarity would fix the mRNA configuration in such a way as to bring the possible site of ribosome binding near the initiation codon.

The Role of the poly(A) sequence in mammalian messenger RNA

The poly(A) sequence is added to 3' termini of nuclear RNA segments destined to become part of the mRNA, and may play an essential role in the selection of these segments. It appears to be required for at least some of the splicing events involved in mRNA processing. In the cytoplasm, the poly(A) segment is the target of a degradation process which causes its gradual shortening, and leads to a heterogeneous steady-state poly(A)-size distribution. Complete loss of the poly(A) is probably followed by inactivation of the mRNA, since chains depleted of poly(A) do not accumulate in the cells. A role for this sequence in the promotion of mRNA stability is suggested by the behavior of globin mRNA depleted of poly(A) after injection into frog oocytes. The poly(A) shortening process may be part of the mRNA inactivation mechanism, as indicated by the greater sensitivity to degradation of the poly(A) of some short-lived mRNAs. However, the stochastic mRNA decay implies that new and old mRNA chains, with long and short poly(A) segments, respectively are equally susceptible to inactivation. The poly(A)-lacking histone mRNAs are stable only in cells engaged in DNA replication. Present knowledge favors a role for poly(A) in the control of mRNA stability. Loss of this sequence could be controlled through modulation of poly(A)-protein interactions or through masking of a sequence directly adjacent to the poly(A). In the nucleus, the poly(A) sequence could also serve as stabilizing agent, but, in addition, it might interact with the splicing machinery.