Molecular cloning, sequence analysis and in vitro expression of a rat tRNA gene cluster

BC1 RNA: transcriptional analysis of a neural cell-specific RNA polymerase III transcript

Rodent BC1 RNA represents the first example of a neural cell-specific RNA polymerase III (Pol III) transcription product. By developing a rat brain in vitro system capable of supporting Pol III-directed transcription, we showed that the rat BC1 RNA intragenic promoter elements, comprising an A box element and a variant B box element, as well as its upstream region, containing octamer-binding consensus sequences and functional TATA and proximal sequence element sites, are necessary for transcription. The BC1 B box, lacking the invariant A residue found in the consensus B boxes of tRNAs, represents a functionally related and possibly distinct promoter element. The transcriptional activity of the BC1 B box element is greatly increased, in both a BC1 RNA and a chimeric tRNA(Leu) gene construct, when the BC1 5' flanking region is present and is appropriately spaced. Moreover, a tRNA consensus B-box sequence can efficiently replace the BC1 B box only if the BC1 upstream region is removed. These interactions, identified only in a homologous in vitro system, between upstream Pol II and intragenic Pol III promoters suggest a mechanism by which the tissue-specific BC1 RNA gene and possibly other Pol III-transcribed genes can be regulated.

Isolation of a mouse genomic clone containing four tRNACys-encoding genes

A cluster of four tRNACys-encoding genes with the anticodon GCA was found on a murine genomic clone containing an 18-kb DNA insert. Three of the four genes encode the identical tRNA, whereas the fourth gene has an altered nucleotide (nt) sequence. Two of the genes within a 2-kb PvuII fragment have the same polarity and are separated by only 921 bp. These two tRNAs have a different primary sequence. The changes in the nt sequences occur within three stems of the tRNA cloverleaf structure and weaken the strength of the H-bonds within the stems. All four genes (designated i-iv) have the 3' structural element that has been proposed as the transcription termination signal [Bogenhagen and Brown, Cell 24 (1981) 261-270]. The remainder of the flanking regions of the three identical tRNAs are very similar to each other, whereas the flanking regions of the fourth tRNA are distinctly different.

The role of the 5'-flanking sequence of a human tRNA(Glu) gene in modulation of its transcriptional activity in vitro

The role of a tRNA-like structure within the 5'-flanking sequence of a human tRNA(Glu) gene in the modulation of its transcription in vitro by HeLa cell extracts has been investigated using several deletion mutants of a recombinant of the gene which lacked part or all of the tRNA-like structure. The transcriptional efficiency of four mutants was the same as that of the wild-type recombinant, two mutants had decreased transcriptional efficiency, one was more efficient, and one, lacking part of the 5' intragenic control region, was inactive. Correlation of the transcriptional efficiencies with the position and the size of the 5'-flanking sequence that was deleted indicated that the tRNA-like structure may be deleted without loss of transcriptional efficiency. Current models for the modulation of tRNA gene transcription by the 5'-flanking sequence are assessed in the light of the results obtained, and a potential model is presented.

Genes, variant genes and pseudogenes of the human tRNA(Val) gene family. Expression and pre-tRNA maturation in vitro

Nine different members of the human tRNA(Val) gene family have been cloned and characterized. Only four of the genes code for one of the known tRNA(Val) isoacceptors. The remaining five genes carry mutations, which in two cases even affect the normal three-dimensional tRNA structure. Each of the genes is transcribed by polymerase III in a HeLa cell nuclear extract, but their transcription efficiencies differ by up to an order of magnitude. Conserved sequences immediately flanking the structural genes that could serve as extragenic control elements were not detected. However, short sequences in the 5' flanking region of two genes show striking similarity with sequences upstream from two Drosophila melanogaster tRNA(Val) genes. Each of the human tRNA(Val) genes has multiple, i.e. two to four, transcription initiation sites. In most cases, transcription termination is caused by oligo(T) sequences downstream from the structural genes. However, the signal sequences ATCTT and CTTCTT also serve as effective polymerase III transcription terminators. The precursors derived from the four tRNA(Val) genes coding for known isoacceptors and those derived from two mutant genes are processed first at their 3' and subsequently at their 5' ends to yield mature tRNAs. The precursor derived from a third mutant gene is incompletely maturated at its 3' end, presumably as a consequence of base-pairing between 5' and 3' flanking sequences. Finally, precursors encoded by the genes that carry mutations affecting the tRNA tertiary structure are completely resistant to 5' and 3' processing.

Genomic organization of the human asparagine transfer RNA genes: localization to the U1 RNA gene and class I pseudogene repeat units

Previously isolated human DNA clones containing asparagine transfer RNA (tRNAAsn) genes have been used to determine the genomic organization of this multigene family in man. One clone also contained a gene for U1 RNA, and so the organization of the two multigene families could be directly compared. The majority, and perhaps all, of the human tRNAAsn genes map to the same chromosome bands as do the U1 RNA true genes and class I pseudogenes located on the short and long arms, respectively, of chromosome 1. These two gene clusters were independently isolated using a somatic-cell hybrid minipanel, and use of repeat-unit DNA polymorphisms showed that one tRNA gene clone maps to the short-arm gene cluster and the other to the long-arm gene cluster. Electron microscopy of heteroduplexes between these two clones showed duplex formation along the proposed region of overlap between them, indicating that the short- and long-arm gene clusters are structurally related. I suggest that the split into two distinct loci was facilitated by a pericentric chromosome inversion. This would have had the effect of positioning the genes currently on the long arm adjacent to the centromeric heterochromatin, perhaps resulting in a "position effect" on transcription of these genes. Restriction fragments of different sizes were found that were common to a majority of repeat units, depending on the restriction enzyme being used. Pulsed-field electrophoresis revealed that fragments of molecular weight of 180 kb were common to each unit (or multiples of units). These fragments also contained U1 RNA gene sequences. I therefore propose that these two gene families are closely linked on repeat units (or multiples of units) of 180 kb in size, which are probably organized in tandem arrays.

A human tRNA gene heterocluster encoding threonine, proline and valine tRNAs

A cluster of three tRNA genes encoding a tRNA(UGUThr), a tRNA(UGGPro), and a tRNA(AACVal), and two Alu-elements occur in a 6.0-kb human DNA fragment. The tRNA(Thr) gene is 2.7-kb upstream from the tRNA(Pro) gene, which is separated by 367 bp from the tRNA(Val) gene. One Alu-element actually overlaps the tRNA(Val) gene and is of opposite polarity to all three tRNA genes. All three tRNA genes are accurately transcribed in a homologous HeLa cell extract, since the ribonuclease T1 fingerprints of the tRNA transcripts are consistent with the nucleotide sequences of the tRNAs. The upstream region flanking the tRNA(Thr) gene has two tracts of alternating purine/pyrimidine residues potentially capable of adopting the Z-DNA conformation, and presumptive binding sites for two RNA polymerase II transcription factors. The tRNA(Thr) gene apparently has a substantially higher in vitro transcriptional efficiency than the other two tRNA genes in this cluster, and a tRNA(GCCGly) gene from another human DNA segment. Deletion constructs of the tRNA(Thr) gene retaining 272, 168, and 33 bp of original 5'-flanking DNA had about the same in vitro transcriptional efficiency, whereas that of the construct with only 2 bp of 5'-flanking human DNA was drastically reduced. The tRNA(Thr) gene constructs with 272 and 168 bp of original 5'-flanking DNA apparently reduce the transcriptional efficiencies of the proline and glycine tRNA genes, implicating the upstream region from the tRNA(Thr) gene as being crucial for its high transcriptional efficiency.

tRNAGlu(GAA) genes from the cellular slime mold Dictyostelium discoideum

The haploid genome of the cellular slime mold Dictyostelium discoideum contains at least 18 gene copies coding for a tRNAGlu(GAA). Using a combination of parasexual genetic analysis and molecular biology techniques, 14 of the 18 individual members of this gene family could be assigned to particular linkage groups. According ot this analysis four tRNAGlu genes are located on group I (C, H, I, K), two genes on group II (D,J), seven genes on either group III or VI (A, B, E, F, L, M, N), and one gene on group VII (G). Eight of the tRNAGlu(GAA) genes have been cloned and characterized. All genes are identical in that part of the gene which corresponds to the mature tRNA, thus representing true nonallelic members of this gene family. Different members of this gene family can be distinguished from each other because they reside on restriction fragments of different lengths and because each gene contains unique 5'- and 3'-flanking regions. Nevertheless, a certain degree of sequence conservation within these flanking regions is apparent for members of this gene family. According to in vivo expression analyses of individual genes in Saccharomyces cerevisiae, all isolated tRNAGlu(GAA) copies represent functional transcription units.

Metabolism of small RNAs in cultured human cells

There are gaps in what is known about the metabolism of some mammalian small RNA species. Our present observations can be summarized as follows. The level of radiolabeled mature U1 RNA doubled between 2 and 24 hr of label chase, while that of all other small RNA species tested did not change. These results are compatible with the possibility that the nucleotide precursor pool for U1 RNA transcription may be partly segregated, or that there may be a second pathway of U1 RNA formation. Precursors of nucleolar U3 RNA were detected whose electrophoretic mobilities are equivalent to those of transcripts approximately 14 and approximately 8 nucleotides longer than the mature species, and which are apparently cytoplasmic. The ladder of U2 RNA precursors showed a gap, suggesting that some of the cleavages during U2 RNA processing are endonucleolytic. We detected an apparent U5 RNA precursor whose electrophoretic mobility is equivalent to that of a species approximately 1 nucleotide longer than mature U5 RNA. There was a predominant band in the middle of the ladder of U4 RNA precursors (apparently approximately 3 nucleotides longer than mature U4 RNA) which suggests that U4 RNA maturation may pause briefly at that intermediate. There are apparently two additional species of mature hY3 RNA, which are less abundant and are about 1 and 2 bases longer than the major hY3 RNA species. An apparent hY3 RNA precursor was detected, which may be approximately 2-3 nucleotides longer than the main mature hY3 RNA species.

A human tRNA(iMet) gene produces multiple transcripts

A third nonallelic locus of the human methionyl-tRNA multigene family (tRNA(iMet-3) was isolated. This gene, unlike two other tRNA(iMet) loci, lacks a remarkable run of T and C residues which functions as a termination of transcription signal. Instead, three tandem termination signals, each containing no more than four thymidylate residues, function as relatively inefficient termination signals. As a result, polymerase readthrough generates at least three transcripts in vitro. The efficiency of apparent termination varies significantly at these sites. All resulting transcripts appear to be processed in vitro.

A human tRNA(iMet) gene produces multiple transcripts

A third nonallelic locus of the human methionyl-tRNA multigene family (tRNA(iMet-3) was isolated. This gene, unlike two other tRNA(iMet) loci, lacks a remarkable run of T and C residues which functions as a termination of transcription signal. Instead, three tandem termination signals, each containing no more than four thymidylate residues, function as relatively inefficient termination signals. As a result, polymerase readthrough generates at least three transcripts in vitro. The efficiency of apparent termination varies significantly at these sites. All resulting transcripts appear to be processed in vitro.

Two human tyrosine tRNA genes contain introns

A human lambda-phage recombinant which contains at least four tRNA genes, has been isolated. Two DNA fragments were subcloned to give the recombinant plasmids pM6 and pM6128. Nucleotide sequence analysis showed that each plasmid contained a different tyrosine acceptor tRNA (tRNATyr) gene. Both tRNATyr genes are interrupted by 21-bp introns. These recombinant plasmids have been shown to direct the in vitro synthesis of tRNA-sized products in a HeLa cell transcription system.

The nucleotide sequence, localization and transcriptional properties of a tRNALeuCUG gene from Drosophila melanogaster

The nucleotide sequence of a tRNALeuCUG gene from Drosophila melanogaster has been determined and compared with available tRNALeuCUG sequences from other eukaryotes, as well as with the tRNALeuUUG gene of D. melanogaster. The genomic location, determined by in situ hybridization, was found to be at site 66B on chromosome 3L. This localization probably places it within one of the known, but uncharacterized, clusters of tRNA genes in this organism. In addition, the transcriptional behaviour of this tRNALeuCUG gene in various in vitro systems is described and it seems that, although the gene is transcribed in all test systems, the very A + T-rich 5'-flanking sequence of this particular gene may be somewhat inhibitory to transcription in vitro.

The human tRNAVal gene family: organization, nucleotide sequences and homologous transcription of three single-copy genes

At least 13 independent tRNAVal gene loci were detected in the human genome. Three of these genes were isolated and shown to occur only once in the haploid genome. No further functional tRNA genes are located on the isolated clones. Two tRNAVal genes encode the known major and minor tRNAVal isoacceptors, the third may be a pseudogene because a corresponding tRNAVal is not yet known. Comparison of extragenic sequences did not reveal significant homologies, indicating the separation of these genes early in vertebrate evolution. An Alu-type repeat was found in two of the clones within several hundred bp distance from the tDNA. All three genes are transcriptionally active in a HeLa nuclear extract. We show here for the first time that homologous in vitro transcription of mammalian tRNA genes strongly depends on extragenic control regions: interestingly, as a consequence of different flanking regions, the transcription efficiencies vary by an order of magnitude among the genes for the major and the minor tRNAVal and thus reflect the concentrations of these tRNAs in vivo.

Nucleotide sequence and transcription of a human tRNA gene cluster with four genes

A bacteriophage lambda clone containing a 20-kb human DNA segment was isolated and found to harbor a cluster of four tRNA genes. An 8.2-kb HindIII subfragment encompassing the genes was cloned into pBR322 for restriction mapping and DNA sequence analysis. The genes were found to be arranged as two tandem pairs, separated by 3 kb. A proline tRNAAGG gene is separated from a leucine tRNAAAG gene by a 724-bp intergenic region in the first pair, and a second proline tRNAAGG gene is 316 bp from a threonine tRNAUGU gene in the second pair, with the leucine tRNA gene being of opposite polarity to the other three genes. A putative Alu-like element was found to occur within a 2.0-kb DNA fragment, at least 0.7 kb from the tRNA gene cluster. The coding sequences of the two proline tRNAAGG genes are identical. The coding regions of all four tRNA genes contain consensus internal split promoter sequences and do not have intervening sequences nor the CCA trinucleotide found in mature tRNAs. The 3'-flanking regions of these four tRNA genes have normal RNA polymerase III termination sites of at least four consecutive T nucleotides. No apparent homologies occur between the 5'-flanking regions of these genes. All four tRNA genes are accurately transcribed in an in vitro HeLa cell-free system, and the RNase T1 fingerprints of the mature-sized tRNA transcripts were found to be consistent with the DNA sequences of the genes.

Transcription factor binding is limited by the 5'-flanking regions of a Drosophila tRNAHis gene and a tRNAHis pseudogene

We determined the sequence of a Drosophila tRNA gene cluster containing a tRNAHis gene and a tRNAHis pseudogene in close proximity on the same DNA strand. The pseudogene contains eight consecutive base pairs different from the region of the bona fide gene which codes for the 3' portion of the anticodon stem of tRNAHis. The tRNAHis gene is transcribed efficiently in Drosophila Kc cell extract, whereas the pseudogene is not. The pseudogene is also a much poorer competitor than the real gene in a stable transcription complex formation assay, even though the sequence alteration in the pseudogene does not affect the sequence or spacing of the putative internal transcription control regions. Recombinant clones were constructed in which the 5'-flanking regions are exchanged. The transcription efficiencies and competitive abilities of the recombinant clones resemble those of the genes from which the 5' flank was derived; for example, the tRNAHis pseudogene with the 5'-flanking sequence of the tRNAHis gene is now efficiently transcribed. Deletion analysis of the pseudogene 5' flank failed to uncover an inhibitory element. Deletion analysis of the real gene showed very high dependence on the presence of the wild-type 5'-flanking sequence for factor binding to the internal control regions and stable complex formation. The 5'-flanking sequence of a Drosophila tRNAArg gene active in the Drosophila Kc cell extract does not restore transcriptional activity or stable complex formation. The tRNAHis gene and pseudogene behave atypically in HeLa cell extract. Both genes compete for HeLa transcription factors, but neither of them is efficiently transcribed. Removal of the 5'-flanking sequences of each gene and replacement with various sequences, including the tRNAArg gene 5' flank, does not allow increased transcription in HeLa cell extract.

Structure and transcription of eukaryotic tRNA genes

The availability of cloned tRNA genes and a variety of eukaryotic in vitro transcription systems allowed rapid progress during the past few years in the characterization of signals in the DNA-controlling gene transcription and in the processing of the precurser RNAs formed. This will be the subject matter discussed in this review.

Nucleotide sequence and transcription of a gene encoding human tRNAGlyCCC

A phage lambda clone containing a 13.1-kb human DNA fragment was isolated and found to contain a tRNA gene encoding a glycine tRNA. The nucleotide sequence of the gene and its flanking regions has been determined. The gene does not have an intervening sequence nor does it encode the 3'-terminal CCA sequence found in mature tRNAs. Although this tRNA gene has an anticodon sequence of CCC, it has a striking homology (96%) with a human glycine tRNA which has an anticodon of GCC. As in other eukaryotic tRNA genes, the coding region contains a characteristic internal split promoter sequence, and the 3'-flanking region has a typical RNA polymerase III termination site of five consecutive T residues. There is no apparent sequence in the 5'-flanking region which could serve as a regulatory element. This gene is accurately transcribed in vitro by RNA polymerase III using a HeLa cell-free system. During the course of in vitro transcription, larger precursor tRNAGlyCCC transcripts are processed to yield a mature-sized tRNA product. A precursor-product relationship was established by comparing the ribonuclease A fingerprints of the precursor and product tRNA transcripts.

Transcription factor binding is limited by the 5'-flanking regions of a Drosophila tRNAHis gene and a tRNAHis pseudogene

We determined the sequence of a Drosophila tRNA gene cluster containing a tRNAHis gene and a tRNAHis pseudogene in close proximity on the same DNA strand. The pseudogene contains eight consecutive base pairs different from the region of the bona fide gene which codes for the 3' portion of the anticodon stem of tRNAHis. The tRNAHis gene is transcribed efficiently in Drosophila Kc cell extract, whereas the pseudogene is not. The pseudogene is also a much poorer competitor than the real gene in a stable transcription complex formation assay, even though the sequence alteration in the pseudogene does not affect the sequence or spacing of the putative internal transcription control regions. Recombinant clones were constructed in which the 5'-flanking regions are exchanged. The transcription efficiencies and competitive abilities of the recombinant clones resemble those of the genes from which the 5' flank was derived; for example, the tRNAHis pseudogene with the 5'-flanking sequence of the tRNAHis gene is now efficiently transcribed. Deletion analysis of the pseudogene 5' flank failed to uncover an inhibitory element. Deletion analysis of the real gene showed very high dependence on the presence of the wild-type 5'-flanking sequence for factor binding to the internal control regions and stable complex formation. The 5'-flanking sequence of a Drosophila tRNAArg gene active in the Drosophila Kc cell extract does not restore transcriptional activity or stable complex formation. The tRNAHis gene and pseudogene behave atypically in HeLa cell extract. Both genes compete for HeLa transcription factors, but neither of them is efficiently transcribed. Removal of the 5'-flanking sequences of each gene and replacement with various sequences, including the tRNAArg gene 5' flank, does not allow increased transcription in HeLa cell extract.

Nucleotide sequences of two regions of the human genome containing tRNAAsn genes

The primary structures of two human tRNAAsn genes and 600-700 nucleotides of their flanking regions have been determined from two separate isolates of a fetal DNA library in phage lambda vector. The tRNA gene from one clone differs from the major mammalian tRNAAsn by a single base substitution at position 47, with an A replacing a G, while the tRNAAsn gene from the second clone has base substitutions at positions 17 and 65, with a G replacing a C and a T replacing a C, respectively. The sequences of the noncoding 5'- and 3'-flanking regions of both clones are over 90% homologous. As with other mammalian tRNA genes, these two human tRNAAsn genes contain CTTTTPu, which might act as a transcription termination signal, 11 bp 3' to the structural gene. In vitro transcription experiments in a HeLa cell extract demonstrate that both cloned tRNAAsn genes can be transcribed and processed to mature-sized tRNAs.