Isolation of a Temperature-sensitive Mutant with an Altered tRNA Nucleotidyltransferase and Cloning of the Gene Encoding tRNA Nucleotidyltransferase in the Yeast Saccharomyces

We have isolated a yeast mutant, ts352, that is tem-perature-sensitive for growth. The mutation has a gen- eral effect on mRNA metabolism and a specific effect on tRNA biosynthesis.

We have isolated a yeast mutant, ts352, that is temperature-sensitive for growth. The mutation has a general effect on mRNA metabolism and a specific effect on tRNA biosynthesis. Cells shifted to the nonpermissive temperature accumulate tRNAs that are shorter than mature tRNAs. The increased ability of these tRNAs to accept ATP demonstrates that growth of the ts352 mutant at the nonpermissive temperature results in accumulation of tRNA with defective 3' ends. The activity of ATP (CTP):tRNA-specific tRNA nucleotidyltransferase can readily be measured in extracts from wild type but not mutant cells. We have cloned and sequenced the wild type allele of the ts352 gene and find significant similarity between the yeast protein sequence predicted from the DNA sequence and the protein predicted from the sequence of the Escherichi coli tRNA nucleotidyltransferase gene. Expression of the yeast gene on a multicopy plasmid increases the activity of the tRNA nucleotidyltransferase in extracts. We conclude that the defect in the ts352 mutant is in the gene coding for yeast tRNA nucleotidyltransferase and that we have isolated the yeast gene that codes for this enzyme. ATP (CTP):tRNA nucleotidyltransferase is the enzyme that catalyzes the incorporation of CMP and AMP residues into tRNAs that have an incomplete CCA sequence at their 3' end (1). The enzyme activity has been identified from both prokaryotic and eukaryotic sources but the role of the enzyme is best understood in Escherichia coli. The gene that encodes the E. coli enzyme has been isolated and sequenced (2). Nonsense mutations in the CCA gene of E. coli do not impair viability, so it is clear that all essential tRNA genes of E. coli encode the CCA sequence (3). A (4,5). The observation that mutants in the E. coli cca gene do display a slower growth rate implicates the enzyme in tRNA repair (4). In eukaryotes, however, tRNA genes do not have the CCA sequence found in mature tRNAs and it must be added post-transcriptionally (1). Although there is little direct evidence linking the enzyme to repair activity in eukaryotes, it presumably fulfills this role as well. Genetic and biochemical studies with the E. coli enzyme have enabled the isolation of the gene encoding the enzyme (2), the purification of the enzyme (6), and its continued analysis. We report here on genetic studies that have resulted in the isolation and characterization of the Saccharomyces cereuisiae tRNA nucleotidyltransferase gene and in the accompanying paper (7) on biochemical characterization of the enzyme from wild type and overproducing strains of yeast. Since yeast tRNA genes do not code the CCA sequence, the nucleotidyltransferase should be essential for growth. Consistent with that prediction is the isolation of a temperaturesensitive mutant strain defective in tRNA nucleotidyltransferase activity. This mutant accumulates tRNA molecules with incomplete 3' termini when shifted to the nonpermissive temperature.
By transforming the mutant strain with a yeast genomic library, we were able to isolate transformants without these defects, and if the complementing DNA is present in multiple copies, the nucleotidyltransferase activity is elevated.

RESULTS
Zdent@ation of the ts352 Mutant-The ts352 mutant was originally selected from a collection of temperature-sensitive mutants (8), because it showed increased steady-state levels of mRNAs produced by the poly(A)-binding protein gene (PAB) (ll), the ribosomal protein gene TCMl (lo), and the STEP gene (12) when shifted from the permissive (23 "C) to the nonpermissive (37 "C) temperature. To assure that the ts phenotype and the altered steady-state level of the mRNAs tested were caused by the same mutation, the original mutant (MATa ts352 ade2-101 his3A200 ura3-52 tyrl) was backcrossed to SS328 (MATa ade2-101 his3A200 ura3-52 lys2) and the resulting tetrads analyzed. All 49 tetrads examined showed a 2:2 segregation of the temperature-sensitive phenotype indicating that this phenotype was caused by a mutation in a single locus (data not shown). To determine if the temperature-sensitive phenotype and the mRNA accumulation are caused by the same mutation, RNA was analyzed from segregants of five tetrads grown at the permissive (23 "C) and nonpermissive (37 "C) temperatures. All five tetrads showed a cosegregation of the temperature-sensitive phenotype and elevated levels of PAB, TCMl, and STE2 mRNAs. The data from two tetrads is presented in Fig. 1. The STE2 gene, encoding the cY-pheromone receptor, is normally only expressed in cells of the a mating type and that pattern is maintained in ts352 if it is grown at the permissive temperature. A shift of ts352 cells to the restrictive temperature results in an increase of the STE2 mRNA in a cells and an inappropriate expression of the STE2 mRNA in (Y cells (Fig. 1). Although there were only a few MATa segregants tested in this analysis, we assume that the STE2 expression in LY cells under nonpermissive conditions is caused by the ts352 mutation. Isolation and Characterization of the Wild Type 352 Locus-Our strategy for the isolation of the wild type 352 locus was to transform the temperature-sensitive strain 352-1B. (MATo Yeast tRNA Nucleotidyltransferae Mutant ts352 ade2-I01 ura3-52 his3A200) with a yeast genomic library cloned into YCp50 (a gift from Scott Emr, California Institute of Technology). Transformants were selected for a Ura+ phenotype at 23 "C and subsequently tested for their ability to grow at 37 "C. Approximately 20,000 transformants were obtained and 15 were able to grow at 37 "C. All 15 transformants yielded the same plasmid which contained a 15kb fragment in YCp50. Subsequent subcloning experiments located the complementing activity to a 4.8-kb ClaI-BamHI fragment (Fig. 2) is restricted to the amino-terminal one-half of the protein (Fig. 3) but was sufficient to suggest that the ts352 mutant might be defective in the ATP(CTP):tRNA nucleotidyltransferase of yeast. To test this hypothesis we have characterized this enzyme activity and the tRNAs in segregants of tetrad 1 derived from the original backcross described above.
Nucleotidyltransferae Activity in Wild Type and ts352-Extracts were prepared from the segregants of tetrad 1 and tested for their ability to add radiolabeled CTP to tRNAs that have had their CCA ends removed by snake venom phosphodiesterase. As can be seen from the data in Table I, two of the four segregants have a reduced ability to carry out this reaction regardless of whether the assay was done at 23 or 37 "C. The two segregants with a deficiency in this enzyme activity are also the two that have the temperature-sensitive growth defect (data not shown).
We next used Northern analysis to examine the effect of the ts352 mutation on tRNA maturation. Cultures of each segregant were grown at 23 "C, harvested, and suspended in media prewarmed to 23 or 37 "C. tRNAs were isolated immediately following suspension in fresh media (zero time) and at I-h intervals and separated by gel electrophoresis. The RNAs were transferred to nylon membranes and probed with an oligonucleotide probe complementary to a tRNAAf& (13). As can be seen in Fig. 4, tRNAs from the segregants carrying the ts352 allele accumulate shorter tRNAs at once, and the accumulation continues throughout growth at the nonpermissive temperature.
Wild type cells do not show this accumulation of shorter tRNAs.
To determine whether the shorter tRNAs actually have an incomplete CCA sequence, we tested the ability of tRNAs from the 3-h 37 "C timepoints for their ability to accept ATP. CTP was included in all of the reactions so that tRNAs missing CTP could be repaired. The tRNAs from the segregants showed very different abilities to serve as substrates for ATP addition (Table II). The tRNAs from the temperaturesensitive segregants are the best substrates for the addition of ATP, indicating the presence of more 3' immature tRNAs in these strains than in the normal strains. We conclude from these experiments that the ts352 mutation results in a deficiency in tRNA nucleotidyltransferase activity.   The yeast sequence is shown in capital letters and the E. coli sequence in lower case letters. Identical amino acids and conservative amino acid changes are underlined. Cultures of each segregant (A-D) of tetrad 1 were grown at either 23 or 37 "C. tRNA samples taken at hourly intervals were separated by polyacrylamide gel electrophoresis and detected by hybridization with an oligonucleotide specific for tRNAA'".
possibility. Temperature sensitivity of the enzyme activity in mutant extracts would be a strong argument that the gene carrying the mutation codes for the enzyme in question but our assays of mutant extracts (Table I) failed to detect significant activity at any temperature. To obtain more definitive evidence that the gene we have isolated and sequenced is the gene coding for the nucleotidyltransferase, we inserted the 4%kb fragment into the multicopy plasmid pJDB207 (18) to create the plasmid pJDB207-352. We transformed pJDB207 and pJDB207-352 into the leu2 recipient strain W303-1B and prepared extracts from the transformants. As can be seen from Table III, the cells containing pJDB207-352 had elevated nucleotidyltransferase activity when compared with those transformed with pJDB207 alone. The simple control of mixing the extracts from the two cell types showed that there was no inhibitor present in the wild type extracts that could lead to an overestimate of the activity in the strain carrying the 352 gene on the multicopy plasmid (data not shown).

Strain and plasmid
Nanomoles CTP transferred/mm/mn W3031B-JDB207 1.2 W3031B-JDB207-352 190.0 DISCUSSION We have described the identification and characterization of a novel temperature-sensitive yeast mutant ts352. Based on the characteristics of this mutant, we propose that it is defective in the tRNA processing enzyme tRNA nucleotidyltransferase. tRNAs with incomplete 3' termini are more abundant in the mutant than in the wild type, and these shorter tRNAs accumulate upon incubation at 37 "C. The nucleotidyltransferase activity is strongly affected in the mutant as indicated by the observation that extracts prepared from ts352 mutants do not contain significant transferase activity whether assays are done at 23 or 37 "C. In vivo temperaturesensitive phenotypes are not always mimicked in vitro. This lack of correlation is, for example, observed with some temperature-sensitive alleles of prp(rna) genes (19,20).
The temperature-sensitive phenotype of the yeast mutant is in contrast to the phenotype of E. coli mutants with altered nucleotidyltransferase. In the latter, inactivation of tRNA nucleotidyltransferase results only in a decreased growth rate (3), presumably because all essential tRNAs in E. coli are coded by genes with a CGA sequence. In yeast, the CCA sequence must be added post-transcriptionally so the enzyme is essential for growth. The fact that a temperature-sensitive mutant was isolated indicates that there are not multiple nucleotidyltransferase genes in this organism.
As might be expected in a strain with a primary defect in tRNA biosynthesis, additional phenotypes are revealed in cells grown at the nonpermissive temperature. Originally ts352 was selected based on an increase in the steady-state levels of three different mRNAs. This suggests a link between an alteration in the translational machinery and the stability of mRNA. Degradation of mRNA may be influenced by several factors and is not understood completely (21). However, it is established that cycloheximide, an inhibitor of chain elongation in protein synthesis, leads to stabilization of many mRNAs (22-26). It seems reasonable that a depletion of functional tRNA molecules would lead to a decrease in the rate of chain elongation and thereby to a protection of mRNA from degradation.
Not only does the ts352 mutant display alterations in steady-state levels of mRNAs, the mutant has altered regulation of expression of the STE2 gene at the nonpermissive temperature. The STE2 gene, encoding the a-pheromone receptor, is expressed only in a cells and repressed in (Y cells by the action of the (~2 repressor (27). In ts352 mutant strains of the (Y mating type, STE2 is expressed at the nonpermissive temperature, suggesting that the 012 repressor level is depressed. Since this repressor is known to turn over rapidly in the nucleus, a decrease in its rate of synthesis due to decreases in the amount of functional tRNA could release the a-specific STE2 gene from repression. This hypothesis implies that the RNA polymerase II components and additional transcription factors essential for the STE2 transcription have a longer half-life than the (~2 repressor. The 352 locus was isolated by functional complementation of the ts352 mutation and the complementing activity localized by deletion mapping. Two lines of evidence in addition to the phenotypes of the ts352 mutant support the conclusion that the 352 locus codes for the yeast nucleotidyltransferase. First, if the sequences essential for complementation are present in multiple copies, the nucleotidyltransferase activity in extracts is elevated. Second, the protein predicted from the complementing DNA sequence shows significant homology to the sequence of the E. co/i enzyme. Interestingly, the similarity is restricted to the amino terminal end of the protein. Cudny et al. (2) noted the presence of the sequence Gly-X-Gly-X-X-Gly beginning at amino acids 66 and 272 in the E. coli protein. This sequence had been implicated previously in nucleotidebinding sites of some nucleotide-binding proteins (28), and in subsequent experiments Zhu et al. (29) demonstrated that changing glycine 70 to an aspartic acid resulted in the loss of AMP-incorporating activity. The yeast enzyme does not have a comparable sequence. The longest open reading frame in the yeast gene would code a protein of 62 kDa, somewhat smaller than the molecular mass reported in the early literature (30) and slightly larger than the size we determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis of enzyme purified as described in the accompanying manuscript (7). Further work with the mutant and the corresponding gene should allow further insight into the structure, biosynthesis, and biochemistry of nucleotidyltransferase in yeast and into the maturation of tRNA in eukaryotic cells.