Primer extension analysis of tRNA gene transcripts synthesized in vitro and in vivo

A spider tRNA(Ala) requires a far upstream sequence element for expression

Within the series of timed differential accumulations of small RNAs we have shown to prelude the synthesis of fibroin in the large ampullate glands of Nephila clavipes (Nc), we are currently directing our attention to the alanine tRNAs. This work reports the subcloning of the members of a tRNAAla gene cluster and the optimization of their transcription in a heterologous cell-free system derived from Bombyx mori (Bm) silkglands. Our data show that the heterologous cell-free system supports the faithful and differential transcription of the individual spider alanine tRNA genes. We are thus making use of the extract to characterize the individual genes with respect to flank-contained regulatory elements through cell-free transcription of gene derivatives. The work has been initiated with pNTA3 because of its high transcriptional activity. Interestingly, the transcription of this gene requires a far upstream sequence, an uncommon modality in tRNA genes.

Differential expression of individual suppressor tRNA(Trp) gene gene family members in vitro and in vivo in the nematode Caenorhabditis elegans

Eight different amber suppressor tRNA (suptRNA) mutations in the nematode Caenorhabditis elegans have been isolated; all are derived from members of the tRNA(Trp) gene family (K. Kondo, B. Makovec, R. H. Waterston, and J. Hodgkin, J. Mol. Biol. 215:7-19, 1990). Genetic assays of suppressor activity suggested that individual tRNA genes were differentially expressed, probably in a tissue- or developmental stage-specific manner. We have now examined the expression of representative members of this gene family both in vitro, using transcription in embryonic cell extracts, and in vivo, by assaying suppression of an amber-mutated lacZ reporter gene in animals carrying different suptRNA mutations. Individual wild-type tRNA(Trp) genes and their amber-suppressing counterparts appear to be transcribed and processed identically in vitro, suggesting that the behavior of suptRNAs should reflect wild-type tRNA expression. The levels of transcription of different suptRNA genes closely parallel the extent of genetic suppression in vivo. The results suggest that differential expression of tRNA genes is most likely at the transcriptional rather than the posttranscriptional level and that 5' flanking sequences play a role in vitro, and probably in vivo as well. Using suppression of a lacZ(Am) reporter gene as a more direct assay of suptRNA activity in individual cell types, we have again observed differential expression which correlates with genetic and in vitro transcription results. This provides a model system to more extensively study the basis for differential expression of this tRNA gene family.

Identification of developmentally regulated sea urchin U5 snRNA genes

A PCR approach was used to isolate repeated U5 small nuclear RNA (snRNA) genes from the sea urchin Lytechinus variegatus. A 1.3 kb repeat, LvU5.0, and three other variants, LvU5.1-U5.3, that differ in the coding region and in the proximal sequence element (PSE) region were isolated. Southern Blot analysis indicate that the U5 snRNA genes, unlike other embryonically expressed snRNA genes (U1, U2 and U6), are not found in a simple tandem repeat, but instead, exist in several heterogeneous clusters each with a small number of genes. The U5 PSE has limited sequence similarity with the other sea urchin PSEs. However, when used in a mobility shift assay the U5 PSE forms a protein/DNA complex that is very similar to the complex formed with the U6 PSE. An RNase protection assay used to monitor the accumulation of U5 snRNA during development shows that at least two U5 variants are coordinately expressed during embryogenesis.

Toxicity of a heterologous leucyl-tRNA (anticodon CAG) in the pathogen Candida albicans: in vivo evidence for non-standard decoding of CUG codons

Plasmids containing derivatives of the Saccharomyces cerevisiae leucyl-tRNA (tRNA(3Leu)) gene that vary in anticodon sequence were constructed and transformed into the pathogen Candida albicans and S. cerevisiae. C. albicans could readily be transformed with plasmids encoding leucyl-tRNA genes with the anticodons CAA and UAA (recognizing the codons UUG and UUA) and expression of the heterologous tRNALeu could be demonstrated by Northern RNA blotting. In contrast, no transformants were obtained if the anticodons were UAG (codons recognized CUN, UUR) and CAG (codon CUG), indicating that the insertion of leucine at CUG codons is toxic for C. albicans. All tRNALeu-encoding plasmids transformed S. cerevisiae with equally high efficiencies. These results provide in vivo evidence that non-standard decoding of CUG codons is essential for the viability of C. albicans.

Establishment of a system for conditional gene expression using an inducible tRNA suppressor gene

We investigated the use of the prokaryotic tetracycline operator-repressor system as a regulatory device to control the expression of Dictyostelium discoideum tRNA genes. The tetO1 operator fragment was inserted at three different positions in front of a tRNA(Glu) (Am) suppressor gene from D. discoideum, and the tetracycline repressor gene was expressed under the control of a constitutive actin 6 promoter. The effectiveness of this approach was determined by monitoring the expression of a beta-galactosidase gene engineered to contain a stop codon that could be suppressed by the tRNA. When these constructs were introduced into Dictyostelium cells, the repressor bound to the operator in front of the tRNA gene and prevented expression of the suppressor tRNA. Addition of tetracycline (30 micrograms/ml) to the growth medium prevented repressor binding, allowed expression of the suppressor tRNA, and resulted in beta-galactosidase synthesis. The operator-repressor complex interfered with tRNA gene transcription when the operator was inserted immediately upstream (position +1 or -7) of the mature tRNA coding region. Expression of a tRNA gene carrying the operator at position -46 did not respond to repressor binding. This system could be used to control the synthesis of any protein, provided the gene contained a translational stop signal.

Establishment of a system for conditional gene expression using an inducible tRNA suppressor gene

We investigated the use of the prokaryotic tetracycline operator-repressor system as a regulatory device to control the expression of Dictyostelium discoideum tRNA genes. The tetO1 operator fragment was inserted at three different positions in front of a tRNA(Glu) (Am) suppressor gene from D. discoideum, and the tetracycline repressor gene was expressed under the control of a constitutive actin 6 promoter. The effectiveness of this approach was determined by monitoring the expression of a beta-galactosidase gene engineered to contain a stop codon that could be suppressed by the tRNA. When these constructs were introduced into Dictyostelium cells, the repressor bound to the operator in front of the tRNA gene and prevented expression of the suppressor tRNA. Addition of tetracycline (30 micrograms/ml) to the growth medium prevented repressor binding, allowed expression of the suppressor tRNA, and resulted in beta-galactosidase synthesis. The operator-repressor complex interfered with tRNA gene transcription when the operator was inserted immediately upstream (position +1 or -7) of the mature tRNA coding region. Expression of a tRNA gene carrying the operator at position -46 did not respond to repressor binding. This system could be used to control the synthesis of any protein, provided the gene contained a translational stop signal.

Expression of variant nuclear Arabidopsis tRNA(Ser) genes and pre-tRNA maturation differ in HeLa, yeast and wheat germ extracts

We have recently identified a tRNA gene cluster in the Arabidopsis nuclear genome. One tRNA(Ser) (AGA) gene and two tRNA(Tyr) (GTA) genes occur in tandem arrangement on a 1.5 kb unit that is amplified about 20-fold at a single chromosomal site. Here we have studied the in vitro expression of seven individually cloned tRNA(Ser) genes (pAtS1 to pAtS7) derived from this cluster. Five out of the seven tRNA(Ser) genes contain point mutations in the coding region which have in part adverse effects on the expression of these genes in different cell-free systems: (i) C10 and A62 in plant tRNA(Ser) genes, which correspond to G10 and C62, respectively, in all known vertebrate tRNA genes, result in a reduced transcription efficiency in HeLa but not in yeast extract. This indicates that yeast RNA polymerase III tolerates nucleotide substitutions at positions 10 [5' internal control region (ICR)] and 62 (3' ICR), whereas the vertebrate RNA polymerase III requires a more stringent consensus sequence. (ii) Processing of a pre-tRNA(Ser) with a mismatch in the aminoacyl stem is impaired in HeLa, yeast and wheat germ extracts, however, a mismatch in the anticodon stem is deleterious for HeLa and wheat germ but not for yeast processing enzymes. The unexpectedly high number of potential tRNA(Ser) pseudogenes in the cluster - quite in contrast to the tRNA(Ser) genes which mainly code for functional tRNAs - suggested that tRNA(Ser) (AGA) genes also occur elsewhere in the genome. We present evidence that single copies of tRNA(Ser) (AGA) genes do indeed exist outside the tRNA gene cluster.

Pleiotropic effect of a point mutation in the yeast SUP4-o tRNA gene: in vivo pre-tRNA processing in S. cerevisiae

The expression of mutant tyrosine-inserting ochre suppressor SUP4-o tRNA genes in vivo in S. cerevisiae was examined as a basis for further studies of tRNA transcription and processing. In vivo yeast precursor tRNAs have been identified by filter hybridization and primer extension analysis. We have previously shown that a mutant SUP4-o tRNA gene with a C52----A52 transversion at positive 52 (C52----A52(+IVS) allele) was transcribed but that the primary transcript was not processed correctly. We show here that 5' and 3' end processing as well as splicing are defective for this mutant but that the 5' end processing is restored when the intron is removed from the gene by oligonucleotide directed mutagenesis (C52----A52(-IVS) allele). Our results imply that the C52----A52 transversion by itself cannot account for the lack of susceptibility to RNase P cleavage but that the overall tertiary structure of the mutant tRNA precursor is destabilized by the intron/anticodon stem. A second consequence of the C52----A52 transversion is to prevent complete maturation of the tRNA precursor at its 3' end since intermediates containing incompletely processed 3' trailers accumulate in the yeast cells transformed with the C52----A52(-IVS) allele. A correct structure of the T stem might therefore define a structural feature required for the recognition of the 3' processing activity.

Analysis of mutant tRNA gene transcripts in vivo in Saccharomyces cerevisiae by abortive primer extension

When the primer extension of a synthetic oligonucleotide hybridized to a complementary region of RNA is made in the presence of only three deoxyribonucleosides triphosphates, elongation of the primer stops as soon as the missing nucleotide is needed. This abortive primer extension assay has been adapted to analyse tRNA gene transcripts and has two main advantages. First it is specific and allows the identification of particular tRNA gene products in an homologous system provided the gene bears a point mutation. Second, it is highly sensitive and can be used to complement and confirm results of Northern blot hybridization. This assay should be a useful tool in the further in vivo study of the transcription and processing of particular tRNA genes in the homologous system. In this report the expression of wild-type and mutant yeast Sup4- tyrosine inserting suppressor gene was studied.

Cloning and expression in vitro of a gene encoding tRNAArgACG from the nematode Caenorhabditis elegans

A gene (rtr-1) coding for the tRNAArgACG has been isolated and characterized from the nematode, Caenorhabditis elegans. The coding portion is not interrupted by an intron and is followed by a track of four thymidines associated with termination by RNA polymerase III. The predicted mature product is 76 nucleotides (nt) long including the CCA tail, and is specific for the most used Arg codon in C. elegans. The gene can be transcribed and processed in a homologous in vitro system. The 82-nt primary transcript begins at the first purine upstream from the mature tRNA 5' end and terminates after the first thymidine of the terminator signal.

Genes, variant genes, and pseudogenes of the human tRNA(Val) gene family are differentially expressed in HeLa cells and in human placenta

Pre-tRNAs(Val) were identified in unfractionated tRNA preparations from HeLa cells and human placenta and their 5' leader structures were deduced from the nucleotide sequences of the corresponding cDNAs. Several of these precursors can be assigned to nine out of the eleven members of the human tRNA(Val) gene family characterized so far, which demonstrates that these gene loci are actively transcribed in vivo. Among the expressed genes there are (a) genes for the two known tRNA(Val) isoacceptor species from human placenta, (b) gene variants that exhibit sequence alterations as compared to conventional genes, and (c) pseudogenes that produce processing-deficient precursors which are not matured to tRNAs. The transcription products of several yet unknown tRNA(Val) genes have also been detected. Furthermore, different expression patterns are observed in the two cell types studied. These data allow for the first time an insight into the in vivo expression of a human tRNA gene family.

An archaebacterial cell-free transcription system. The expression of tRNA genes from Methanococcus vannielii is mediated by a transcription factor

Our understanding of the mechanism of RNA biosynthesis in archaebacteria is limited, due in part to the inability of purified RNA polymerases to transcribe purified genes accurately in vitro. In the present study, we show that cell extracts of Methanococcus vannielii and Methanococcus thermolithotrophicus purified by gradient centrifugation synthesize a distinct transcript from templates harboring a cloned homologous tRNA(Val) and tRNA(Arg) gene. The in vitro transcripts initiate with GTP at the same sites as in Methanococcus cells. About 60% of the sequence of the in vitro RNA products was analyzed by dideoxyterminated primer extension and found to be identical with that of the precursors of tRNA(Val) and tRNA(Arg). This finding indicates that this RNA polymerase fraction both initiates and terminates transcription faithfully in vitro. After purification of a cell-free extract (S-100) of M. thermolithotrophicus by phosphocellulose chromatography, the endogenous RNA polymerase has lost its ability to transcribe the tRNA(Val) gene accurately. The activity directing specific expression of this template was reconstituted by the addition of a protein-fraction devoid of RNA polymerase activity. Thus, a transcription factor appears to be required for accurate cell-free expression of tRNA genes from M. vannielii.

Oocyte and somatic tyrosine tRNA genes in Xenopus laevis

Over a period of many months, Xenopus oocytes stockpile large quantities of tRNA for use during the first few hours of embryogenesis. To test the idea that these tRNAs are transcribed from one set of genes and that another set is used by somatic cells, we used synthetic oligonucleotides to analyze the sequence and steady-state levels of unspliced tyrosine tRNA precursors in Xenopus laevis oocytes, embryos, and cultured kidney cells. These analyses identify four kinds of tyrosine tRNA genes, two oocyte-type and two somatic-type, whose unspliced transcripts are distinguishable from one another by their different 5' leader and intervening sequences. The oocyte-type tyrosine tRNA precursors are present in oocytes, very abundant in gastrula embryos, but absent from postembryonic somatic cells. The somatic-type precursors are undetectable in oocytes but are found in gastrula and later stage embryos and in somatic cells. The major switch from oocyte-type to somatic-type transcripts occurs early during embryogenesis, between the midblastula transition and the onset of neurulation, but some oocyte-type precursors are also detectable in tadpoles.

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.

A family of non-allelic tRNA(ValGUU) genes from the cellular slime mold Dictyostelium discoideum

A haploid genome of the cellular slime mold Dictyostelium discoideum contains at least 14 non-allelic gene copies coding for a tRNA(ValGUU). The structure, genomic organization, and expression of these genes have been analyzed in relation to stages of the developmental cycle. So far, 13 tRNA(ValGUU) genes have been isolated and characterized. All genes contain identical mature tRNA-coding regions, and consequently identical gene internal promoter elements. However, different genes differ with respect to their 5'- and 3'-flanking regions, although a certain degree of sequence conservation seems apparent. Different members of this tRNA gene family appear to be randomly dispersed along the seven D. discoideum chromosomes, and not clustered at any one genomic location. In vivo expression of individual genes was studied in yeast. All but one tRNA(ValGUU) gene are actively transcribed, though with different efficiencies. There is also evidence that not all of these tRNA genes are constitutively transcribed in Dictyostelium throughout the developmental cycle. One characteristic primary transcript can only be detected in cells of the late preaggregation phase, whereas growing cells, cells in the stationary phase or cells harvested 4 h after the onset of development do not seem to carry this transcript. This product seems to be transcribed from a gene of an unusual structure. Although this particular gene has not yet been isolated, it can be predicted from the sequence of the cDNA synthesized from primary transcription products of this putative gene, that it is composed of nt 1-54 of a 3'-truncated tRNA(ValGUU) gene linked to a bona fide tRNA(ValGUU) gene.

Identification of a protein factor binding to the 5'-flanking region of a tRNA gene and being involved in modulation of tRNA gene transcription in vivo in Saccharomyces cerevisiae

Control mechanisms of tRNA gene transcription were studied in vivo in Saccharomyces cerevisiae. In order to be able to monitor in vivo transcription products of an individual tRNA gene, a 'tester gene' was used which is readily transcribed in vivo in yeast but does not cross-hybridize with any cellular yeast tRNA. A series of insertion mutants were constructed, modifying thereby the immediate and further distant 5'-flanking region of the 'tester tRNA gene'. Small linker molecules of different length and different sequence were inserted at positions -3 and -56 on the non-coding strand. Resulting tRNA gene variants were transformed into yeast cells and in vivo synthesized products were monitored by primer extension analysis. From the experimental data we suggest that a few essential nucleotides within the flanking region are able to determine the in vivo transcription activity of the 'tester tRNA gene'. Our results are rationalized on a biochemical level by protein binding assays: At least one protein binds to the 5'-flanking region of the 'tester tRNA gene' and different protein complexes are sequestered on active or less active tRNA gene variants.

Structural requirements for the synthesis of tRNATrp from Dictyostelium discoideum in yeast

Dictyostelium tRNA genes can generally be expressed in vivo in yeast. Among tested Dictyostelium tRNA genes a tRNATrp gene containing a 13 bp intron is transcribed with particularly poor apparent efficiency and the intron is not removed. Elimination of the intron from the gene increases the amount of transcription products significantly. Splicing can only occur if minimal base-pairing of the anticodon with intron sequences is possible. Accumulation of tRNA gene transcripts decreases with the inability of intron splicing. Products of neither amber (UAG) nor opal (UGA) suppressor variants of the tRNATrp gene from Dictyostelium are able to suppress corresponding non-sense mutations in defined structural yeast genes. This also holds true for suppressor tRNA gene variants with precisely deleted intron regions.

Influence of different 5'-flanking sequences of tRNA genes on their in vivo transcription efficiencies in Saccharomyces cerevisiae

We have investigated the influence of 5'-flanking sequences on the in vivo transcription activities in yeast. Since eukaryotic tRNA genes belong to multi-copy gene families monitoring of the activity of a particular tRNA gene is not possible. We therefore used two different tRNA genes from the cellular slime mould Dictyostelium discoideum which are efficiently transcribed and processed in vivo in yeast. The original 5'-flanking sequences of the two tRNA genes were replaced by random plasmid sequences. The modified tRNA genes were introduced into Saccharomyces cerevisiae and bulk tRNAs from the transformants were analyzed for the presence and the relative number of Dictyostelium tRNA gene transcripts. Substantial differences of steady-state levels of RNA transcribed were detected dependent on the 5'-flanking sequence of the tRNA gene. Minute structural changes, such as inserting two additional nucleotides in front of a tRNA gene, can lead to drastic activity changes. The efficiency of tRNA gene transcription can be conferred by sequences located more than 40 nucleotides upstream from the 5' end of the mature tRNA coding region.