Interferon inducibility of mammalian tryptophanyl-tRNA synthetase: new perspectives.

Mammalian aminoacyl-tRNA synthetases are indispensible components of the cell's protein-synthesizing machinery. Surprisingly, recent experiments have demonstrated that synthesis of tryptophanyl-tRNA synthetase (WRS) is markedly enhanced after incubation of human cells with interferons. Why is this housekeeping enzyme interferon-inducible? Several hypotheses have been suggested. One hypothesis, that premature termination of protein synthesis was involved, was boosted by the discovery that the deduced amino acid sequence of the mammalian peptide chain release factor (RF) closely resembled that of WRS. Further investigation, however, suggests that the DNA encoding RF was wrongly identified and in fact encodes a rabbit WRS subunit. Other hypotheses on the interferon-inducibility of WRS, including the possibility that the protein performs other, regulatory functions in addition to its core enzymic activity, remain to be explored.

Translational termination comes of age.

Translational termination has been a largely ignored aspect of protein synthesis for many years. However, the recent identification of new release-factor genes, the mapping of release-factor functional sites and in vitro reconstitution experiments have provided a deeper understanding of the termination mechanism. In addition, protein-protein interactions among release factors and with other proteins have been revealed. The three-dimensional structures of a prokaryotic ribosome recycling factor and eukaryotic release factor 1 (eRF1) mimic the shape of transfer RNA, indicating that they bind to the same ribosomal site. Post-termination events in bacteria have been clarified, linking termination, ribosomal recycling and translation initiation.

Class-1 translation termination factors: invariant GGQ minidomain is essential for release activity and ribosome binding but not for stop codon recognition.

Previously, we have shown that all class-1 polypeptide release factors (RFs) share a common glycine-glycine-glutamine (GGQ) motif, which is critical for RF activity. Here, we subjected to site-directed mutagenesis two invariant amino acids, Gln185 and Arg189, situated in the GGQ minidomain of human eRF1, followed by determination of RF activity and the ribosome binding capacity for mutant eRF1. We show that replacement of Gln185 with polar amino acid residues causes partial inactivation of RF activity; Gln185Ile, Arg189Ala and Arg189Gln mutants are completely inactive; all mutants that retain partial RF activity respond similarly to three stop codons. We suggest that loss of RF activity for Gln185 and Arg189 mutants is caused by distortion of the conformation of the GGQ minidomain but not by damage of the stop codon recognition site of eRF1. Our data are inconsistent with the model postulating direct involvement of Gln185 side chain in orientation of water molecule toward peptidyl-tRNA ester bond at the ribosomal peptidyl transferase centre. Most of the Gln185 mutants exhibit reduced ability to bind to the ribosome, probably, to rRNA and/or (peptidyl)-tRNA(s). The data suggest that the GGQ motif is implicated both in promoting peptidyl-tRNA hydrolysis and binding to the ribosome.

Directed cleavage of ribopolynucleotides with nucleases restricted by multiple modification of substrate.

Simultaneous exhaustive modification of cytidine and uridine residues of rRNA with methoxyamine and sodium metabisulfite renders adjacent phosphodiester bonds resistant to pancreatic and T2 ribonucleases. Another method of T2 RNAase restriction is modification of cytidine with methoxyaminebisulfite followed by modification of guanosine residues with beta-ethoxy-alpha-ketobutyraldehyde. Mild alkaline treatment leads to demodification of uridine and guanosine residues leaving intact modified cytidine residues, thus providing a means of stepwise, directed cleavage of the polynucleotide. The series of combined cleavage procedures and methods of isolation of oligo(C), oligo(G) and oligopyrimidine tracts, as well as the procedure of selective cleavage at uridine residues elaborated in the course of the present studies may serve as a basis for more rational procedures of RNA sequencing.

Limited proteolysis of the tryptophanyl-tRNA synthetase.

Earlier studies have shown that native tryptophanyl-tRNA synthetase from beef pancreas is composed of two apparently identical subunits having a molecular weight of 60000 plus or minus 2000 each. Incubation of the pruified enzyme with trypsin under restrictive conditions results in splitting of each subunit to form an enzymatically inactive polypeptide chain of mol. wt 24500 plus or minus 1500. During proteolysis, two distinct intermediate forms of mol. wt 51000 plus or minus 2000 and 40000 plus or minus 2000 and fragments of mol. wt 14000 plus or minus 2500 are formed. The presence of substrates, viz. ATP, tryptophan or tryptophanyl adenylate, decreases the rate of proteolysis. However, a band pattern monitored by acrylamide gel electrophoresis is qualitatively indistinguishable from that obtained in the absence of substrates. Native and trypsin-modified subunits (the latter having a molecular weight of 24500) have been maleylated, reduced, carbosymethylated and subjected to exhaustive digestion by trypsin followed by peptide mapping. Comparison of the finger prints has shown that the trypsin-modified subunit represents a polypeptide with lowered content of dicarboxylic amino acids. That the number of peptides revealed after complete proteolysis of native and trypsin-modified subunits does not favour the presence of long repetitive sequences in each subunit, is at variance with some bacterial aminoacyl-tRNA synthetases. Study of the fluorescence polarisation of 1-anilino-8-napthalene sulphonate adsorbed on the dimeric tryptophanyl-tRNA synthetase, indicates that the molecule behaves as a complete entity in Brownian rotation. The trypsin-resistant end products, composed of two types of polypeptides (mol. wts 24500 and 14000), remain associated with each other. From the mol. wt of this associate it follows that each fragment is present in the associate in duplicate. When the purification procedure was carried out in the absence of a protease inhibitor, the active modified enzyme form was obtained. As judged from the molecular weight values, it is composed of two equal subunits corresponding to one of the products of limited proteolysis. The data presented are compatible with compact three-dimensional structure of tryptophanyl-tRNA synthetase having very limited regions exposed to exogenous or endogenous proteolysis.

Segment substitutions in alpha satellite DNA. Unusual structure of human chromosome 3-specific alpha satellite repeat unit.

We have sequenced the full-length copy of the alpha satellite higher-order repeated unit characteristic of human chromosome 3. Its internal structure, the regular alterations of J1 and J2 type monomers, is typical of the alphoid suprachromosomal family 1. This dimeric order is disrupted by the substitution of one J1 unit by an alien dimer which is not clearly related to any of the established monomeric types. We have also observed some other similar cases of segment substitutions in alpha satellite DNA. They probably represent a special type of molecular event which could be generated by gene conversion. Segment substitutions may be one of the important factors responsible for the extreme variability of localization patterns and actual sequences of alpha satellite DNA that should be taken into account in reconstructions of alpha satellite evolution.

Structural rearrangements and insertions of dispersed elements in pericentromeric alpha satellites occur preferably at kinkable DNA sites.

Centromeric region of human chromosome 21 comprises two long alphoid DNA arrays: the well homogenized and CENP-B box-rich alpha21-I and the alpha21-II, containing a set of less homogenized and CENP-B box-poor subfamilies located closer to the short arm of the chromosome. Continuous alphoid fragment of 100 monomers bordering the non-satellite sequences in human chromosome 21 was mapped to the pericentromeric short arm region by fluorescence in situ hybridization (alpha21-II locus). The alphoid sequence contained several rearrangements including five large deletions within monomers and insertions of three truncated L1 elements. No binding sites for centromeric protein CENP-B were found. We analyzed sequences with alphoid/non-alphoid junctions selectively screened from current databases and revealed various rearrangements disrupting the regular tandem alphoid structure, namely, deletions, duplications, inversions, expansions of short oligonucleotide motifs and insertions of different dispersed elements. The detailed analysis of more than 1100 alphoid monomers from junction regions showed that the vast majority of structural alterations and joinings with non-alphoid DNAs occur in alpha satellite families lacking CENP-B boxes. Most analyzed events were found in sequences located toward the edges of the centromeric alphoid arrays. Different dispersed elements were inserted into alphoid DNA at kinkable dinucleotides (TG, CA or TA) situated between pyrimidine/purine tracks. DNA rearrangements resulting from different processes such as recombination and replication occur at kinkable DNA sites alike insertions but irrespectively of the occurrence of pyrimidine/purine tracks. It seems that kinkable dinucleotides TG, CA and TA are part of recognition signals for many proteins involved in recombination, replication, and insertional events. Alphoid DNA is a good model for studying these processes.

Evidence for selection in evolution of alpha satellite DNA: the central role of CENP-B/pJ alpha binding region.

Conservation of DNA segments performing sequence-related functions is a landmark of selection and functional significance. Phylogenetic variability of alpha satellite and apparent absence of conserved regions calls its functional significance into question, even though sequence-specific alpha satellite-binding proteins pJ alpha and CENP-B have been discovered. Moreover, the function of pJ alpha is obscure and CENP-B binding satellite DNA, which is thought to participate in centromere formation, is found only in few species and not necessarily in all chromosomes. Analysis of alpha satellite evolution allows us to recognize the order in this variability. Here we report a new alpha satellite suprachromosomal family, which together with the four defined earlier, covers all known alpha satellite sequences. Although each family has its characteristic types of monomers, they all descend from two prototypes, A and B. We show that most differences between prototypes are concentrated in a short region (positions 35 to 51), which exists in two alternative states: it matches a binding site for pJ alpha in type A and the one for CENP-B in type B. Lower primates have only type A monomers whereas great apes have both A and B. The new family is formed by monomeric types almost identical to A and B prototypes, thus representing a living relic of alpha satellite. Analysis of these data shows that selection-driven evolution, rather than random fixation of mutations, formed the distinction between A and B types. To our knowledge, this is the first evidence for selection in any of the known satellite DNAs.

Compartmentalization of certain components of the protein synthesis apparatus in mammalian cells.

A comparative study on the localization of free cytosolic tryptophanyl-tRNA synthetase (TrpRS) and several components of the multi-aminoacyl-tRNA synthetase (ARS) complex (glutamyl-prolyl-tRNA synthetase (GluProRS), arginyl-tRNA synthetase (ArgRS)), and two non-synthetase polypeptides p38 and p43 has been carried out on ultrathin sections of cultured rabbit kidney cells by the immunogold technique using monoclonal antibodies raised against appropriate polypeptides. It has been shown that GluProRS, ArgRS and p38 polypeptide are distributed in the cells similarly to TrpRS and are located mainly in the vicinity of ribosomes. A smaller but significant portion of these proteins has been observed in the nuclei in the diffuse chromatin regions and in the vicinity of interchromatin granules. On the contrary, the main part of p43 protein was found in the cell nuclei; this indicates that this protein may exist in the cell separately from the cytoplasmic multi-ARS complex. Our results argue in favor of compartmentalization of both free ARS (such as TrpRS) and the multi-ARS complex in the vicinity of ribosomes. At the same time, the detection of some ARS in the diffuse chromatin regions in the nucleus implies that these enzymes may exhibit some non-canonical functions in addition to their role in protein synthesis.

Immunoelectron microscopic location of tryptophanyl-tRNA synthetase in mammalian, prokaryotic and archaebacterial cells.

Monoclonal antibody Am1 against conservative epitope of tryptophanyl-tRNA synthetase (WRS) was labeled with colloidal gold particles and used to localize the enzyme on ultrathin sections of eubacteria (Escherichia coli), archaebacteria (Methanococcus halophilus), rat pancreas tissue and rat fibroblasts (cell line RAT1). In all cell types immunoelectron microscopy revealed predominant cytoplasmic location of gold particles, as this could be expected from known biochemical data. In particular, in mammalian cells intensive labeling was observed in cytoplasmic regions rich in polysomes and free ribosomes. At the same time, the label was virtually absent in cytoplasmic regions where microfilament bundles were present. Significant concentrations of gold particles were found in mitochondria and nuclei. In the latter case, gold particles were located over diffuse chromatin regions and were virtually absent over compact chromatin. The density of diffuse chromatin in labeling may amount to about 50% of that found in the cytoplasm. Distribution of labeled antibodies over E. coli cells looks rather similar to that found for M. halophilus: gold particles are preferably concentrated over the cytoplasm and "boundary zone", i.e., a 30 nm wide cytoplasmic zone adjacent to the nucleoid border, while the label over nucleoid is virtually absent. Two main conclusions are drawn: (i) although in the animal cell homogenates WRS is recovered mainly as a soluble cytosolic enzyme, in intact cells it is associated with defined cellular organelles and compartments; this may be an evolutionarily acquired feature probably typical for multicellular organisms; (ii) the considerable density of labeling in diffuse (not compact) chromatin regions may be indicative of WRS involvement in the active chromatin functions (transcription, processing, transfer of gene products, etc.).

Tight linkage of retroviral-like sequences to a variant human c-mos gene in the human genome.

Chumakov et al. [Gene 17 (1982) 19-26] identified in the human gene library a number of recombinant phages that possess a homology to the v-mos gene. Here we report the unusual structure of one of these recombinants, lambda gp5. The 14.3-kb stretch of human DNA from this phage contains at least three regions of homology to the v-mos gene, together with multiple copies of Alu-family repeats. Moreover, we have shown the presence of retrovirus-related sequences in the close vicinity of the mos-homologous regions. These data point to the possibility of involvement of retrovirus in the process of c-mos gene amplification during the formation of a multigene family.

Cloning and nucleotide sequence of the structural gene encoding for human tryptophanyl-tRNA synthetase.

A structural gene encoding bovine (b) tryptophanyl-tRNA synthetase (WRS) has recently been cloned and sequenced [Garret et al., Biochemistry 30 (1991) 7809-7817]. Using part of this sequence as a hybridisation probe we have cloned and sequenced a structural gene encoding human polypeptide highly homologous with two mammalian proteins, bWRS [Garret et al., Biochemistry 30 (1991) 7809-7817; EMBL accession No. X52113] and rabbit peptide chain release factor [Lee et al., Proc. Natl. Acad. Sci. USA 87 (1990) 3508-3512]. Identification of the sequence encoding a human WRS is based on (i) the presence of 'HIGH' and 'KMSKS' structural motifs typical for class-I aminoacyl-tRNA synthetases [Eriani et al., Nature 347 (1990) 203-206]; (ii) coincidence of the number of SH groups per subunit estimated experimentally [Muench et al., Science 187 (1975) 1089-1091] and deduced from the cDNA sequence (six in both cases); (iii) close resemblance of two WRS polypeptides sequenced earlier [Muench et al., Science 187 (1975) 1089-1091] and the predicted structure in two different regions.

Nucleotide sequence of the anemic pigeon alpha-globin gene. Structural rearrangements in the cloned cDNA.

Pigeon alpha-globin cDNA, previously cloned into the PstI site of pBR322 to yield pGP22 (Scobeleva et al., 1981), was mapped and sequenced. The insert was found to have an unusual structural organization, with 33 bp corresponding to the 5'-untranslated region of mRNA located roughly at the middle and flanked downstream by the 81 triplets coding for the N-terminal half of the alpha-globin chain. Moreover, 25 triplets identical to the 3'-terminal region of the insert were located upstream from the 5'-untranslated region and were followed by the coding sequence of 82 to 134 triplets. The last seven codons from 135 to 142 triplets were missing. The structural rearrangements in the cloned cDNA are discussed. The nucleotide sequence of alpha-globin mRNA from anemic pigeon was determined; it differs remarkably from those for chicken and duck, although homologies can easily be found. The predicted amino acid sequence of pigeon anemic alpha-globin is given.

Nucleotide sequences in human chromosomal DNA from nonhepatic tissues homologous to the hepatitis B virus genome.

DNA extracted from human nonhepatic tissues (placenta and kidney) have been digested with restriction endonucleases and examined by the Southern procedure with cloned 32P-labelled DNA of hepatitis B virus (HBV). In placental DNAs of women with the history of a hepatitis B infection and in one out of four cases of patients with no known HBV exposure or manifestation, HBV-related chromosomal nucleotide sequences were detected. The integration of HBV-related sequences was observed also in human kidney DNA. Moreover, in the placenta of women who had hepatitis B infection prior to delivery, unusual unintegrated forms of HBV have been found. We conclude that HBV sequences can be found not only in hepatic tissue but also in placental and kidney DNA, both of HBV-exposed and in one case even of a nonexposed patient.

The human gene encoding tryptophanyl-tRNA synthetase: interferon-response elements and exon-intron organization.

Recently, we cloned and sequenced the cDNA encoding human tryptophanyl-tRNA synthetase (hWRS) [Frolova et al., Gene 109 (1991) 291-296]. Independently, it has been shown that this protein is induced by interferons (IFN) gamma and alpha [Fleckner et al., Proc. Natl. Acad. Sci. USA 88 (1991) 11520-11524; Rubin et al., J. Biol. Chem. 266 (1991) 24245-24248]. This unusual feature of a housekeeping enzyme raises the problem of how the gene is regulated. Since at present the genomic structure of hWRS is unknown, this issue remains unsolved. Here, the exon-intron organization of hWRS has been deciphered. This gene consists of at least 12 exons that span more than 35 kb of DNA. At least two alternative noncoding exons precede ten coding exons. Upstream from the first exon, two GGAAAN(N/-)GAAA sequences, which are considered to be IFN-stimulating response elements (ISRE), have been revealed. The same consensus was also found in the intron region in close vicinity to the 5' end of the second exon. Thus, the IFN-stimulated synthesis of hWRS is presumably due to gene activation at the transcriptional level. Alignment of hWRS amino acid sequences has shown that exons V to XI of hWRS encode regions of structural similarity with bacterial WRS, whereas the N-terminal portion of the protein encoded by exons II to IV exhibits no homology with bacterial WRS.(ABSTRACT TRUNCATED AT 250 WORDS)

Exons I and VII of the gene (Ker10) encoding human keratin 10 undergo structural rearrangements within repeats.

A genomic fragment containing the K51 gene previously isolated from a rat genomic library by hybridization with the v-mos probe in nonstringent conditions [Chumakov et al., Dokl. Akad. Nauk SSSR 290 (1986) 1252-1254], resembles a human keratin type-I-encoding gene [Shvets et al., Mol. Biol. 24 (1990) 663-677]. This genomic clone, K51, has been used as a probe to search for related human genes. A recombinant clone, HK51, with a 1.5-kb insert, was isolated from a human embryonic skin cDNA library, and its nucleotide (nt) sequence was determined. Analysis has shown that the cloned cDNA encodes human keratin 10 (Ker10). All presently known nt sequences of the human Ker10-encoding gene (Ker10) are not identical. Differences are concentrated in the 5'-end of the first exon and in the middle of the seventh exon within repeats. In spite of structural rearrangements in two of eight exons, the reading frame and position of the stop codon are preserved. The genetic rearrangements cause changes in hydrophobicity profiles of the N and C termini of Ker10. It was also noticed that insertion of one nt leads to the formation of an unusual 3'-end of the transcript.

A genome-integrated hepatitis B virus DNA in human neuroblastoma.

An integrated hepatitis B virus (HBV) DNA has been identified in a human neuroblastoma, cloned and sequenced. The integrated HBV DNA is not rearranged, although a 480-bp fragment is deleted. The integrated viral DNA covers the C gene fragment, the region of Pre-S and S, and the 3'-truncated X gene.

Human nucleotide sequences related to the transforming gene of a murine sarcoma virus: studies with cloned viral and cellular DNAs.

A recombinant plasmid, pI26, has been constructed by cloning into pBR322 a transforming gene of murine sarcoma virus (a Moloney strain, clone 124, MSV) synthesized by detergent-treated virions. From this plasmid a XbaI-HindIII fragment has been isolated which contains only mos-specific sequences. This mos-specific probe has been used for screening a human gene library cloned in bacteriophage lambda Charon 4A. Of these, 19 clones have been isolated containing mos-related sequences. By physical mapping and molecular hybridization it has been shown that these sequences are neighboured by DNA regions related to Moloney murine leukemia virus. Recombinant phages have also been found containing human inserts related to MLV, not to the mos gene. The possible existence of murine-like endogenous retroviruses in the normal human genome, including that of a sarcoma type, is discussed. By Northern blotting, expression of the cellular c-mos gene has been detected in mouse liver treated with a hepatocarcinogen. The general significance of the suggested model for evaluating the relationship between chemical carcinogenesis and oncogene expression is discussed.

Isolation of rat liver 5S RNA genes.

A method for the isolation of structural genes, whose transcripts do not contain terminal poly(A) sequences, is presented. Poly(A) tails of predicted length were synthesized at the 3'-OH ends of RNA molecules employing Escherichia coli ATP : RNA adenyltransferase (EC 2.7.7.19). The gene isolation was performed in two steps: (a) enrichment of DNA fragments carrying the genes of interest, (b) subsequent cloning and amplification of these fragments. The enrichment of given gene sequences requires separate purifications of each of two complementary DNA chains, followed by their annealing. Rat liver 5S RNA genes were isolated by this method. DNA (+)chains were obtained by hybridizing EcoRI-digested and denatured rat liver DNA fragments with the 5S RNA-poly(A). The complex was purified on an oligo(dT)-cellulose column. DNA (-)chains were purified by hybridizing DNA fragments with the 5S cDNA which was coupled to oligo(dT)-cellulose. In the isolated and annealed double-stranded DNA preparation the enrichment of 5S DNA was 350--400-fold, when compared with the total genomic DNA. These purified DNA sequences were inserted into the EcoRI site of pBR325. Out of 1200 recombinant-plasmid colonies, analyzed in the colony-hybridization test, 4 hybridized with [125I]5S RNA.

The coding region of the human c-mos pseudogene contains Alu repeat insertions.

We have determined the nucleotide sequence of an 841-bp fragment derived from a segment of the human genome previously cloned by Chumakov et al. [Gene 17 (1982) 19-26] and Zabarovsky et al. [Gene 23 (1983) 379-384] and containing regions homologous to the viral mos gene probe. This sequence displays homology with part of the coding region of the human and murine c-mos genes, contains several termination codons, and is interrupted by two Alu-family elements flanked by short direct repeats. Probably, the progenitor of the human c-mos gene was duplicated approximately at the time of mammalian divergence, was converted to a pseudogene, and acquired insertions of two Alu elements.