Saccharomyces cerevisiae Possesses a Stress-Inducible Glycyl-tRNA Synthetase Gene

Aminoacyl-tRNA synthetases are a large family of housekeeping enzymes that are pivotal in protein translation and other vital cellular processes. Saccharomyces cerevisiae possesses two distinct nuclear glycyl-tRNA synthetase (GlyRS) genes, GRS1 and GRS2. GRS1 encodes both cytoplasmic and mitochondrial activities, while GRS2 is essentially silent and dispensable under normal conditions. We herein present evidence that expression of GRS2 was drastically induced upon heat shock, ethanol or hydrogen peroxide addition, and high pH, while expression of GRS1 was somewhat repressed under those conditions. In addition, GlyRS2 (the enzyme encoded by GRS2) had a higher protein stability and a lower KM value for yeast tRNAGly under heat shock conditions than under normal conditions. Moreover, GRS2 rescued the growth defect of a GRS1 knockout strain when highly expressed by a strong promoter at 37°C, but not at the optimal temperature of 30°C. These results suggest that GRS2 is actually an inducible gene that may function to rescue the activity of GRS1 under stress conditions.


Introduction
Aminoacyl-tRNA synthetases (aaRSs) are a structurally diverse group of enzymes, each of which couples a specific amino acid to its cognate tRNA. The resultant aminoacyl-tRNA is then carried by elongation factor (EF)-1 to ribosomes for protein synthesis. Typically, there are 20 aaRSs in prokaryotes, one for each amino acid [1][2][3][4]. In contrast, eukaryotes, such as yeast, commonly possess two distinct sets of aaRSs, one localized in the cytoplasm and the other in mitochondria. Each set aminoacylates the isoaccepting tRNAs within its respective cellular compartment, and is sequestered from isoacceptors confined to other compartments. Cytoplasmic and mitochondrial aaRSs with the same amino acid specificity are normally encoded by two distinct nuclear genes, regardless of the cellular compartment in which they are active. The gene that is regarded as being of eukaryotic origin encodes cytoplasmic aaRS, while its orthologue, which is thought to be of mitochondrial (or bacterial) origin, encodes the mitochondrial counterpart. Nevertheless, four Saccharomyces cerevisiae genes, ALA1 (encoding alanyl-tRNA synthetase) [5], GRS1 (encoding glycyl-tRNA synthetase [GlyRS]) [6], HTS1 (encoding histidyl-tRNA synthetase) [7], and VAS1 (encoding valyl-tRNA synthetase [ValRS]) [8], encode both the mitochondrial and cytosolic forms through alternative initiation of translation.
Based on conserved sequence motifs, the oligomeric structure, and aminoacylation function, aaRSs are segregated into two classes of ten enzymes each [9,10]. Class I enzymes first attach amino acids to the 29-OH of the terminal adenylate residue of tRNA before transferring it to 39-OH; class II enzymes couple them directly to the 39-OH. In addition, class I enzymes are mostly monomeric, while class II enzymes are often homodimers. Moreover, cognate enzymes, even from phylogenetically distant life forms, usually possess convincing sequence similarities in their catalytic domains and are in the same class, suggesting that they share a common lineage. The only exceptions to this rule are lysyl-tRNA synthetase (LysRS) and GlyRS. Both class I-and class IItype enzymes were recovered for LysRS [11], while two distinct oligomeric forms were identified for GlyRS: an a 2 b 2 heterotetramer and an a 2 homodimer [12]. While both forms of GlyRS share a common class II-defining architecture, they greatly differ in size and sequence [13]. As a result, these two forms of GlyRS are believed to have different origins [14]. To date, a 2 b 2 -type GlyRSs are found only in bacteria and plant chloroplasts, while a 2 -type GlyRSs have been recovered from all major domains of life.
In S. cerevisiae, two distinct nuclear GlyRS genes were identified. The first, GRS1, possesses both cytoplasmic and mitochondrial activities, while the second, GRS2, appears to be silent and dispensable for growth under normal conditions [15]. GRS1 encodes two distinct protein isoforms through alternative use of two in-frame initiator codons [6]. Such a dual-functional phenotype is conserved in the GRS1 homologue of Schizosaccharomyces pombe [15]. These findings prompted us to ask whether a similar feature is conserved in GlyRS genes of all other yeast species, and whether GRS2 is really a dysfunctional gene. We found that all yeasts studied possessed a dual-functional GRS1 homologue. Saccharomyces cerevisiae and Vanderwaltozyma polyspora were the only two yeasts known to contain a second GlyRS gene, a GRS2 homologue. As it turns out, GRS2 of S. cerevisiae is an inducible gene, the expression of which was activated by heat, H 2 O 2 , high pH, and ethanol. In contrast, expression of the housekeeping gene, GRS1, was repressed to some extent by these stimuli. Moreover, the enzyme specified by GRS2 was heatresistant. Hence, GRS2 may function to rescue the activity of GRS1 under certain stress conditions.

A Dual-Functional GRS1 Homologue is Present in All Yeasts
To further our understanding of the dual-functional feature of GRS1, we searched databases for all available yeast GlyRS genes and then analyzed their sequences. Sixteen GlyRS sequences were recovered from 14 yeast species. Except for S. cerevisiae and V. polyspora, which contained both GRS1 and GRS2 homologues, only one GlyRS gene, a GRS1 homologue, was present in those yeasts ( Figure 1A). As with GRS1 of S. cerevisiae, an alternative initiator candidate (AUG or non-AUG) was readily identified in the leading sequences of these GRS1 homologues, suggesting that they are dual functional ( Figure 1A). Note that all non-AUG initiator candidates identified here contained an ''A'' nucleotide at relative position 23 and encoded a canonical mitochondrial targeting signal. In contrast, no suitable alternative initiator candidate was found in the leading sequences of the GRS2 homologues of S. cerevisiae or V. polyspora. The GlyRS1 enzymes from S. cerevisiae and V. polyspora were highly similar in sequence (86% identity). In contrast, GlyRS2 enzymes were somewhat divergent, not only from their respective GlyRS1 counterparts, but also from one another (,60% identity) ( Figure 1B). Both GlyRS2 enzymes lacked an intact lysine-rich insertion domain of ,44 amino acid residues; this insertion domain is, thus far, restricted to yeast GlyRS1 sequences ( Figure 1C). In particular, a distinctive sequence motif embedded inside the insertion domain, KKKRKKK, was completely missing from GlyRS2 enzymes.

GRS2 Functions Better at a Higher Temperature
To gain further insights into the biological functions of GRS2 of S. cerevisiae, the open reading frame of this gene was cloned into pADH (a high-copy-number yeast shuttle vector with a constitutive ADH promoter and a short sequence coding for a His 6 tag), and the ability of the resultant construct to rescue the growth defects of a grs1 2 strain was tested. As shown in Figure 2, overexpression of GRS2 from a strong promoter failed to confer a positive phenotype to the gene under normal growth conditions; the GRS2 construct rescued neither the cytoplasmic nor mitochondrial defect of the knockout strain at 30uC, an observation also made by Turner et al. [15]. Unexpectedly, this construct successfully restored the growth phenotype of the null allele on FOA when tested at 37uC, suggesting that the enzyme encoded by GRS2, GlyRS2, prefers a higher temperature. On the other hand, fusion of a mitochondrial targeting sequence to GRS2 did not confer a mitochondrial phenotype to the gene at either temperature ( Figure 2B). A Western blot analysis using an anti-His 6 tag antibody further showed that GRS2 had an expression level ,50-fold lower than that of GRS1 under the control of an ADH promoter ( Figure 2C).

GRS2 is an Inducible Gene
To explore whether the expression of GRS2 is inducible, we compared relative levels of endogenous GRS1 and GRS2 messenger (m)RNAs by a semiquantitative reverse-transcription (RT)-polymerase chain reaction (PCR) (as illustrated in Figure 3A).
The expected sizes of the PCR-amplified complementary (c)DNA fragments were 755 and 629 bp for GRS1 and GRS2, respectively. Total RNA was isolated from cells grown under various culture conditions, such as normal conditions (defined here as a growth temperature at 30uC, a cell density of OD 600 ,0.6, and growth medium at pH 6.0), low temperature (16uC), high temperature (37uC), high external pH (pH 8.0), ethanol (3%), and H 2 O 2 (5 mM). Figure 3A shows that both the GRS1 and GRS2 DNA fragments were efficiently amplified by the designated primers using genomic DNA as the template (positive control). In contrast, no specific DNA fragments were amplified by the PCR using total RNA (from normal conditions) as the template (negative control). Consistent with a previous observation, only the GRS1 cDNA fragment was amplified by PCR using cDNA that had been reverse-transcribed from total RNA prepared from cells grown under normal conditions as the template ( Figure 3A) [15]. Surprisingly, expression of GRS2 was drastically induced by a high temperature, a high external pH, 3% ethanol, and 5 mM H 2 O 2 , but not by a low temperature. Just as surprising was the finding that expression of the housekeeping gene, GRS1, was somewhat repressed under these conditions.
To obtain more-accurate data on the relative levels of GRS1 and GRS2 mRNAs, a quantitative real-time RT-PCR was carried out using two distinct sets of primers. The data were first normalized to ACT1 (encoding actin) and then were compared to each other. As shown in Figure 3C, GRS1 had an mRNA level ,25-fold higher than that of GRS2 under normal conditions (left panel). Expression of GRS2 was enhanced 6-, 26-, and 5-fold by a high temperature, 3% ethanol, and a high external pH, respectively (right panel). In contrast, expression of GRS1 was reduced 2,4-fold by those stimuli (middle panel).

GlyRS2 is Much More Stable at 37uC
To compare the protein stability of GlyRS2 between 30uC (the optimal temperature for yeast) and 37uC (the induction temperature used here), a cycloheximide-chase assay was performed. GRS1 and GRS2 were first cloned into pGAL1, a high-copynumber yeast shuttle vector with an inducible GAL1 promoter and a short sequence coding for a His 6 tag. Constructs were transformed into INVSc1, and cultures of the transformants were then induced with galactose for 2 h, followed by the addition of cycloheximide to terminate protein synthesis. Cycloheximidetreated cells were grown at 30uC or switched to 37uC, and harvested at various intervals following induction. Protein extracts were prepared for Western blot analyses using an anti-His 6 tag antibody. As shown in Figure 4, GlyRS2 was much more stable at the higher temperature, 37uC. Up to 60% of GlyRS2 was degraded after 4 h of cycloheximide treatment at 30uC, but only 15% of the protein was degraded within the same time period at 37uC. In contrast, there was no significant difference in protein stability of GlyRS1 at these two temperatures throughout the time period tested.

GlyRS1 is ,5-Fold More Active Than GlyRS2 In Vitro
To compare the aminoacylation activity of GlyRS2 between 30 and 37uC, recombinant GlyRS2-His 6 was purified from a transformant of S. cerevisiae containing a plasmid-borne GRS2 gene by Ni-NTA column chromatography to homogeneity, and unfractionated S. cerevisiae tRNA was used as the substrate for the aminoacylation reactions. As a control, recombinant GlyRS1-His 6 was also purified and assayed. Figure 5 shows that the aminoacylation activity of GlyRS2 was ,5-fold lower than that of GlyRS1 at 30uC.
Kinetic parameters for these enzymes were subsequently determined using unfractionated yeast tRNA as the substrate. As shown in Table 1, the K M and k cat values for GlyRS1 were respectively 0.53 mM and 0.36 s 21 at 30uC, while the K M and k cat values for GlyRS2 were respectively 0.61 mM and 0.06 s 21 under the same condition. Increasing the reaction temperature from 30 to 37uC enhanced the K M value of GlyRS1 for tRNA Gly ,2.8-fold, while reducing the K M value of GlyRS2 ,2.3-fold. In contrast, the k cat values for these two enzymes were not affected by the temperature changes. Paradoxically, these two enzymes had almost identical catalytic efficiencies (k cat /K M ) at 37uC. Taken together, these results suggested that GlyRS2 had a lower K M value for tRNA Gly at the induction temperature than at the optimal growth temperature. Similar observations were made with an in vitro-transcribed tRNA Gly (see Appendix S1).

Discussion
Many yeast cytoplasmic tRNA synthetases contain an N-or Cterminal polypeptide extension that is absent from their bacterial relatives [16]. These domains are generally rich in lysine residues and involved in non-specific tRNA binding. Examples are found in yeast glutaminyl-and valyl-tRNA synthetases [17][18][19] and tRNA synthetases of higher eukaryotes [20][21][22][23]. These domains act in cis as an auxiliary tRNA-binding domain and enhance the tRNAbinding affinity of the enzymes. In contrast, the appended domains of some yeast tRNA synthetases were found to participate in protein-protein interactions, such as those of yeast glutamyl-, methionyl- [24], and seryl-tRNA synthetases [25]. Such interactions were also shown to enhance the tRNA binding and aminoacylation of the enzymes. Although the yeast GlyRS1 enzymes lack an appended domain, they possess a lysine-rich insertion domain, which distinguishes yeast GlyRS1 from other eukaryotic GlyRS sequences (Figure 1) [15]. It is believed that this insert confers certain selective advantages (such as enhanced tRNA binding) upon the dual-functional yeast enzyme. Contrary to this hypothesis, our results showed that GlyRS2, which lacks such an insert, has a K M value for yeast tRNA Gly similar to that of GlyRS1 under normal conditions (at 30uC) ( Table 1).
Evidence presented herein shows that expression of GRS2 was drastically induced by high temperature, high external pH, ethanol, and H 2 O 2 , while expression of GRS1 was somewhat repressed by these stimuli (Figure 3). Moreover, GlyRS2 had a much higher protein stability at a high temperature, 37uC ( Figure 4). These findings underscore the possibility that GlyRS2 functions under certain stress conditions where GlyRS1 is insufficient, unavailable, or rendered inactive. One such scenario was found in Escherichia coli, where two distinct LysRS genes were  recovered. The lysS gene encodes a housekeeping synthetase, while the lysU gene is heat-inducible. Individually, each of the two LysRS genes is dispensable for growth, but deletion of lysS leads to a cold-sensitive phenotype [26,27].
The identity elements for tRNA Gly comprise the discriminator base N73, the first three base pairs of the acceptor stem (1:72, 2:71, and 3:70), and C35 and C36 in the anticodon [2]. The most conspicuous difference between bacterial and eukaryotic tRNA Gly isoacceptors is the discriminator base, N73, which is nearly always a U in bacteria and an A in eukaryotes. For example, in E. coli, N73 is a U; in A. thaliana, N73 is a U in the mitochondrial-encoded tRNA Gly and an A in the three nuclearencoded cytosolic tRNAs Gly [28]. For that reason, plant cytosolic tRNAs Gly is charged by an a 2 -dimeric GlyRS, while plant mitochondrial-encoded tRNA Gly is aminoacylated by an (ab) 2dimeric GlyRS [29]. This raised the question of how a single GlyRS can charge both mitochondrial-and nuclear-encoded tRNAs Gly in yeast. As it turns out, tRNAs Gly in both compartments possess the eukaryote-type discriminator base, A73. Perhaps, the ancient mitochondrial tRNA Gly once possessed the bacterium-type discriminator base U73, but the discriminator base was later mutated to A73 as an outcome of tRNA Gly /GlyRS coevolution. However, the discovery that the a 2 -dimeric GlyRS from Thermus thermophilus efficiently charges E. coli and yeast tRNAs Gly (with U73 and A73, respectively) breaks down the relationship between the oligomeric structure of the enzyme (a 2 or a 2 b 2 ) and the nature of its tRNA substrate (A73 or U73) [30].
We recently showed that Sch. pombe possesses two distinct nuclear ValRS genes, one encoding the cytoplasmic form and the other its mitochondrial counterpart. Unexpectedly, both of these genes are of mitochondrial origin [31]. A phylogenetic analysis further indicated that the ValRS gene in all eukaryotes, including those in the amitochondriate protists Giardia lamblia and Trichomonas vaginalis, is of mitochondrial origin [32]. Partly for this reason, the E. coli homologue fused to Arc1p or other tRNA-binding domains can easily substitute for VAS1 of S. cerevisiae [33]. It is conceivable that the ValRS gene of eukaryotic origin was lost in all eukaryotes during evolution [32]. In contrast, the evolution of eukaryotic GlyRS appears to have taken a different route. The GlyRS gene in all eukaryotes is of eukaryotic origin [15]. Conceivably, the GlyRS gene of mitochondrial origin was lost in all yeasts and possibly all eukaryotes. Therefore, the evolutionary footprints, including duplication, reduction, acquisition, and loss, of a particular type of gene (eukaryotic or mitochondrial) appear to be independent for each aminoacylation activity, even in the same organism. But regardless of the detailed interpretation, the most striking finding reported here is the discovery of an inducible GlyRS gene, which may function to rescue the activity of the housekeeping homologue under stress conditions.

Construction of plasmids
Cloning of the GRS1 and GRS2 genes from S. cerevisiae and V. polyspora into pADH (a high-copy-number yeast shuttle vector with a constitutive ADH promoter and a short sequence coding for a His 6 tag) and pGAL1 (a high-copy-number yeast shuttle vector with an inducible GAL1 promoter and a short sequence coding for a His 6 tag) followed a protocol described earlier [6]. Briefly, the open reading frames of these genes (extending from 212 bp to the full-length coding sequence) were amplified by a polymerase chain reaction (PCR) as an EagI-XhoI fragment (,2000 bp) using a pair of gene-specific primers. After enzyme digestion, the PCRamplified fragment was cloned into the EagI/XhoI sites of pADH (for complementation and Western blotting) or pGAL1 (for protein purification and cycloheximide-chase assays).
Complementation assays for the cytoplasmic and mitochondrial activities of GRS1 and GRS2 followed protocols described earlier [6]. The yeast grs12 strain, RJT3/II-1, used for complementation assays was described elsewhere [15]. Western blot analyses [34] and cycloheximide-chase assays [35] followed protocols described earlier. The membrane was probed with a horseradish peroxidase (HRP)-conjugated anti-His 6 antibody and then exposed to x-ray film following the addition of the appropriate substrates.

Reverse-Transcription (RT)-PCR and Quantitative Real-Time RT-PCR
To determine the relative levels of specific GRS1 and GRS2 mRNAs in S. cerevisiae, a semiquantitative RT-PCR experiment was carried out following protocols provided by the manufacturer (Invitrogen, Carlsbad, CA). Total RNA was first isolated from the transformant, and then treated with DNase to remove contaminating DNA. Aliquots (,3 mg) of the RNA were then reversetranscribed into single-stranded complementary (c)DNA using an oligo-dT primer. After RNase H treatment, the single-stranded cDNA products were amplified by PCR using three gene-specific primers: GSP1, GSP2, and GSP3. GSP1 is complementary to nucleotides +424 to +447 of GRS1, GSP2 is complementary to nucleotides +430 to +455 of GRS2, and GSP3 is complementary to a consensus sequence of GRS1 and GRS2 (nucleotides +1153 to +1178 of GRS1 and nucleotides +1033 to +1058 of GRS2). As a positive control, the genomic DNA of S. cerevisiae was also isolated and used as a template for the PCR amplification of GRS1 and GRS2.
A quantitative real-time RT-PCR experiment was subsequently used to obtain more-accurate data [36]. Two sets of primers, GSP4/GSP5 and GSP6/GSP7, were respectively used to amplify the cDNA fragments of GRS1 and GRS2. GSP4 and GSP5 are respectively complementary to nucleotides +473 to +492 and +628 to +651 of GRS1, while GSP6 and GSP7 are respectively complementary to nucleotides +1302 to +1322 and +1477 to +1498 of GRS2. The relative mRNA levels of GRS1 and GRS2 were first normalized to those of ACT1 and then compared to each other. The forward and reverse primers for ACT1 are respectively complementary to nucleotides +490 to +509 and +668 to +686 of this gene. Quantitative data were obtained from three independent experiments and averaged.

Aminoacylation Assay
Aminoacylation reactions were carried out at 30uC (or 37uC) in a buffer containing 50 mM HEPES (pH 7.5), 50 mM KCl, 15 mM MgCl 2 , 5 mM dithiothreitol, 10 mM ATP, 0.1 mg/ml bovine serum albumin (BSA), 100 mM S. cerevisiae tRNA, and 20 mM glycine (4 mM 3 H-glycine; Moravek Biochemicals, Brea, CA). The specific activity of 3 H-glycine used was 35.0 Ci/mmol. Purification of His 6 -tagged GlyRS1 and GlyRS2 proteins was as previously described [35]. The final concentration of the enzymes used in the reactions was 20 nM. Determination of active protein concentrations by active site titration was as previously described [37]. Reactions were quenched by spotting 10-ml aliquots of the reaction mixture onto Whatman filters (Maidstone, UK) soaked in 5% trichloroacetic acid and 1 mM glycine. The filters were washed three times for 15 min each in ice-cold 5% trichloroacetic acid before liquid scintillation counting. Data were obtained from three independent experiments and averaged.

Supporting Information
Appendix S1 Kinetic Parameters for Aminoacylation of an in vitro-transcribed yeast tRNA Gly by GlyRS1 and GlyRS2. (TIF)