GCD14p, a Repressor of GCN4 Translation, Cooperates with Gcd10p and Lhp1p in the Maturation of Initiator Methionyl-tRNA in Saccharomyces cerevisiae

Abstract

Gcd10p and Gcd14p were first identified genetically as repressors of GCN4 mRNA translation in Saccharomyces cerevisiae. Recent findings indicate that Gcd10p and Gcd14p reside in a nuclear complex required for the presence of 1-methyladenosine in tRNAs. Here we show that Gcd14p is an essential protein with predicted binding motifs for S-adenosylmethionine, consistent with a direct function in tRNA methylation. Two different gcd14 mutants exhibit defects in cell growth and accumulate high levels of initiator methionyl-tRNA (tRNA_i^Met) precursors containing 5′ and 3′ extensions, suggesting a defect in processing of the primary transcript. Dosage suppressors of gcd10 mutations, encoding tRNA_i^Met (hcIMT1 to hcIMT4; hc indicates that the gene is carried on a high-copy-number plasmid) or a homologue of human La protein implicated in tRNA 3′-end formation (hcLHP1), also suppressed gcd14 mutations. In fact, the lethality of a GCD14 deletion was suppressed by hcIMT4, indicating that the essential function of Gcd14p is required for biogenesis of tRNA_i^Met. A mutation in GCD10 or deletion of LHP1 exacerbated the defects in cell growth and expression of mature tRNA_i^Met in gcd14 mutants, consistent with functional interactions between Gcd14p, Gcd10p, and Lhp1p in vivo. Surprisingly, the amounts of NME1 and RPR1, the RNA components of RNases P and MRP, were substantially lower in gcd14 lhp1::LEU2 double mutants than in the corresponding single mutants, whereas 5S rRNA was present at wild-type levels. Our findings suggest that Gcd14p and Lhp1p cooperate in the maturation of a subset of RNA polymerase III transcripts.

Eukaryotic cells respond to different conditions of starvation and stress by down-regulating the rates of protein biosynthesis. One aspect of this response in mammalian cells involves phosphorylation on serine 51 of the alpha subunit of translation initiation factor 2 (eIF2α) (53). This phosphorylation event reduces the activity of eIF2B, the guanine nucleotide exchange factor for eIF2 that recycles eIF2-GDP to the active form eIF2-GTP (20, 40, 46, 59). Only eIF2-GTP can bind initiator methionyl-tRNA (tRNA_i^Met) to produce the eIF2-GTP-tRNA_i^Met ternary complex that delivers initiator tRNA to the 43S preinitiation complex; thus, phosphorylation of eIF2 inhibits general protein synthesis initiation. In the budding yeast Saccharomyces cerevisiae, eIF2α becomes phosphorylated when cells are deprived of an amino acid or purine, leading to depletion of ternary complex levels, just as occurs in mammalian cells (19, 31). Interestingly, it also specifically stimulates translation of the distinctive mRNA encoding Gcn4p, a transcriptional activator of more than 40 genes involved in amino acid biosynthesis (29), thereby alleviating the nutrient limitation conditions that trigger phosphorylation of eIF2α in yeast.

The molecular mechanism that couples GCN4 translation to amino acid availability has been explored in great detail over past several years (reviewed in references 31 and 32). Four short upstream open reading frames (uORFs) present in the GCN4 mRNA leader sequence prevent ribosomes from initiating translation at the GCN4 start codon under conditions of amino acid sufficiency. Amino acid starvation triggers activation of the protein kinase Gcn2p, which phosphorylates eIF2α and thereby inhibits eIF2-GTP-tRNA_i^Met ternary complex formation. This situation specifically favors GCN4 translation because, after translating uORF1, many ribosomes cannot acquire the ternary complex in time to recognize the start codons at uORF2, -3, and -4 and thus continue scanning downstream. The majority of these ribosomes rebind the ternary complex by the time they reach the GCN4 AUG codon and reinitiate there instead (19, 20).

Recessive mutations in any of the genes encoding the four essential subunits of eIF2B, GCD1, GCD2, GCD6, and GCD7, mimic the effects of eIF2 phosphorylation and derepress GCN4 mRNA translation in the absence of Gcn2p and amino acid starvation (29, 30). These gcd mutations also lead to unconditional slow growth or temperature sensitivity on nutrient-rich medium, phenotypes that result from defects in general translation initiation (9, 16, 22, 27). It is thought that gcd1, gcd2, gcd6, and gcd7 mutations impair the ability of eIF2B to recycle eIF2-GDP to eIF2-GTP and thereby decrease the rate of ternary complex formation in vivo (31, 46). This causes lower rates of translation initiation on most mRNAs but leads to increased translation of GCN4 mRNA, which is inversely coupled to the availability of ternary complexes in the cell (20).

Recessive mutations in GCD10 also lead to temperature-sensitive growth and constitutive derepression of GCN4 translation in the absence of Gcn2p (26, 42). GCD10 is an essential gene and was shown to encode a 62-kDa protein that copurified and coimmunoprecipitated with subunits of eIF3 (23). eIF3 stimulates several steps of the initiation pathway in mammalian cells (among others, binding of ternary complexes to 40S ribosomes [47]). Accordingly, we proposed that gcd10 mutations might derepress GCN4 translation by decreasing the ability of eIF3 to promote rebinding of ternary complexes to 40S subunits that have translated uORF1 and continued scanning downstream on the GCN4 mRNA leader. This would allow these subunits to skip uORF4 and reinitiate translation at the GCN4 AUG without any reduction in the level of ternary complex formation (23). More recently, we found that gcd10 mutations reduce the expression of mature tRNA_i^Met at a posttranscriptional step and that the phenotypes of gcd10 mutants are suppressed by multiple copies of the genes IMT1 to IMT4, encoding tRNA_i^Met (3). Interestingly, gcd10 mutants lack 1-methyladenosine (m¹A) at position 58 in tRNA_i^Met and in all other tRNAs (17 or more) containing this modified base. The absence of m¹A seems to impair specifically the expression of mature tRNA_i^Met through degradation of the unprocessed precursors containing 5′ and 3′ extensions (3). The reduction in mature tRNA_i^Met levels in gcd10 mutants should diminish ternary complex formation, decreasing the rate of general translation initiation and producing constitutive derepression of GCN4 translation (20). Thus, the impaired maturation of tRNA_i^Met can also explain the known phenotypes of gcd10 mutants (3).

We previously identified GCD14 and GCD15 as novel genes in S. cerevisiae that are required for translational repression of GCN4 mRNA under amino acid-replete conditions (17). Here we report the molecular cloning of GCD14 and show that it encodes an essential protein containing binding motifs for S-adenosylmethionine (S-AdoMet). We found that gcd14 mutants have reduced levels of mature tRNA_i^Met and accumulate tRNA_i^Met precursors containing 5′ and 3′ extensions. Similar to the case for gcd10 mutants, the presence of IMT1 to IMT4 in high copy number suppresses the phenotypes of gcd14-1 and gcd14-2 mutants, and overexpression of IMT4 overcomes the lethality of a gcd14::URA3 deletion. Gcd14p is physically associated with Gcd10p in cell extracts (3), and we observed additive effects of gcd14 and gcd10 mutations on cell growth and expression of mature tRNA_i^Met. Interestingly, we found that gcd14 mutations also lead to reduced expression of other tRNAs and the RNA components of RNases P (38) and MRP (54) when combined with a deletion of LHP1, encoding a yeast homologue of the human autoantigen La. These findings and others (3) suggest that Gcd10p and Gcd14p functionally cooperate with Lhp1p to promote the maturation of a subset of RNA polymerase III transcripts.

MATERIALS AND METHODS

Plasmids.

Plasmids constructed in this work are represented in Fig. 1. Plasmid pRC5 was isolated from a yeast genomic library in the low-copy-number vector YCp50 (51) by selecting for complementation of the 3-amino-1,2,4-triazol resistance (3AT^r) and Slg⁻ phenotypes of strain Hm316 (gcd14-2). Plasmids pRC55 and pRC56 contain, respectively, the 2.3-kb XbaI and the 2.8-kb SpeI-ClaI fragments from pRC5 subcloned into the corresponding sites of pRS316 (56).

FIG. 1.

Analysis of the GCD14 coding region. A partial restriction map of the GCD14 region is shown in the center, indicating the positions of the two ORFs, GCD14 and SPT10 (44), identified in the plasmid isolated from the genomic library bearing GCD14 (pRC5). The direction of transcription is indicated by the arrows. The top line depicts the genomic DNA insert in pRC5, and below that are depicted the fragments present in subclones constructed from pRC5 in low-copy-number vector pRS316 (pRC55 and pRC56). The ability of each subclone to complement the phenotypes of gcd14-1 and gcd14-2 mutants is shown on the right. Below the map of the GCD14 region are shown enlargements of the 2.6-kb SpeI-ClaI region in plasmid pRC56 and of the 3.2-kb SpeI-ClaI region in pRC560 bearing the gcd14::URA3 allele, in which a 1.1-kb URA3 fragment replaces ∼730 bp of the GCD14 coding sequence. Restriction enzyme sites: B, BamHI; Bg, BglII; C, ClaI; H, HindIII; Hp, HpaI; S, Sau3A; X, XbaI; Sp, SpeI.

Plasmid pRC56 was digested with HpaI and BglII to delete an internal 730-bp fragment of the GCD14 coding sequence, and the BglII end was filled in with Klenow fragment. The resulting linearized plasmid was ligated to a blunt-ended 1.1-kb HindIII-HindIII fragment containing the URA3 gene, generating pRC560.

Plasmid pRC57 was generated by subcloning into pRS306 a ∼2.8-kb HpaI-BamHI fragment from pRC5, extending from 1 kb upstream of the GCD14 ATG to ∼30 nucleotides 5′ of the stop codon. Plasmids pRC64 and pRC65 were constructed as follows. (i) The 2.8-kb SpeI-ClaI fragment from pRC56 was cloned into the corresponding sites of the SK+ polylinker of Bluescript M13+ (Stratagene), generating plasmid pRC6. (ii) A NotI site was introduced upstream of the GCD14 stop codon by in vitro mutagenesis with a MUTA-GENE in vitro mutagenesis kit (Bio-Rad), using plasmid pRC6 and an oligonucleotide (RC 22) containing a NotI site (5′CGATCCACGGAAAAACGCGGCCGCTAATTAAATGATTAAC3′; the NotI site is underlined, and the GCD14 stop codon is in boldface). The resulting plasmid was named pRC61. (To avoid methylation of the ClaI site located downstream of the GCD14 gene in plasmid pRC61, it was always amplified in the dam mutant Escherichia coli strain GM119.) (iii) A 2.4-kb ClaI fragment containing all of the GCD14 coding sequences was subcloned in YCp50 or in the high-copy-number YEp24 vector (50), generating plasmids pRC62 and pRC63, respectively. (iv) A 115-bp NotI fragment, encoding three in-tandem copies of the hemagglutinin (HA) epitope, was inserted at the NotI site in pRC62 and pRC63, to create plasmids pRC64 and pRC65, respectively. Sequence analysis confirmed the in-frame insertion of the HA-coding sequence just upstream of the GCD14 stop codon.

A 670-bp HpaI-EcoRI fragment from the 3′ coding region of the GCD14 gene was fused in frame to the trpE coding sequence in pATH2 (21) to generate pTRPC14. Construction of the correct in-frame fusion was confirmed by DNA sequencing.

Plasmid pJA128 was constructed by inserting the ca. 3.5-kb XbaI-XhoI fragment bearing GCD10 from pMG107 (23) into YCplac111 (24). Plasmid pJA103, containing the lhp1::hisG::URA3::hisG allele, was constructed by PCR amplification of 200- and 350-bp fragments containing, respectively, the 5′ and 3′ flanking sequences at LHP1 by using the appropriate primers. The 5′ PCR product was digested with EcoRI and BglII and cloned into EcoRI- and BglII-digested pNKY51 (1) to create pJA103′. The 3′ PCR product was digested with BamHI and SphI and cloned into BamHI- and SphI-digested pJA103′ to create pJA103.

All other plasmids used in this work have been previously described: pE107 (GCD10) (23); p1775 (IMT4) (20); p2632 (IMT1), p2633 (IMT2), p2634 (IMT3), p2635 (IMT4), and p2626(LHP1) (3); pNK985 (1); pRS315 (56); pRS426 (13); and pKOIII (60).

Yeast strain constructions.

The genotypes and origins of all yeast strains used in this study are summarized in Table 1. Strains Hm295 (gcd14-1), Hm296 (gcd14-2), and Hm316 (gcd14-2) were selected as 3AT^r and Slg⁻ ascospores from genetic crosses previously described (17). Isogenic Hm296G and Hm296g strains were obtained by transforming Hm296 to Ura⁺ with the integrating plasmid pRC57, containing an incomplete copy of GCD14. Depending on the location of the crossover, integration of pRC57 yielded a nontandem duplication of wild-type GCD14 and the truncated gcd14 allele, or the gcd14-2 and truncated gcd14 alleles, separated by plasmid sequences. The former (in Hm296G) led to a 3AT-sensitive phenotype, whereas the latter (in Hm296g) conferred 3AT resistance.

TABLE 1.

Yeast strains

Strain	Genotype	Source or reference
H117	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 (HIS4::lacZ URA3)	26
H160	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-1 (HIS4::lacZ URA3)	17
H168	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-2 (HIS4::lacZ URA3)	17
H752	MATa gcn2::LEU2 leu2-3 leu2-112 ura3-52	26
H753	MATα gcn2::LEU2 leu2-3 leu2-112 ura3-52	26
H1515	MATa leu2,3-112 trp1Δ63 ura3-52	A. G. Hinnebusch
Hm295	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-1 (HIS4::lacZ ura3-52)	17
Hm296	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-2 (HIS4::lacZ ura3-52)	17
Hm296g	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-2 (gcd14 URA3)	This study
Hm296G	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-2 (GCD14 URA3)	This study
Hm298	MATa gcn2-101 gcn3-101 his1-29 ura3-52 gcd10-505 (HIS4::lacZ ura3-52)	23
Hm316	MATα gcd2-101 his1-29 ino1 leu2-3 leu2-112 ura3-52 gcd14-2	17
Hm397	MATa gcn2-101 gcn3-101 his1-29 ino1 leu2::hisG ura3-52 gcd10-505 gcd14-2	This study
Hm406	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 (HIS4::lacZ ura3-52) leu2::hisG::URA3::hisG lhp1::LEU2	This study
Hm407	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-1 (HIS4::lacZ ura3-52) leu2::hisG::URA3::hisG lhp1::LEU2	This study
Hm408	MATα gcn2-101 gcn3-101 his1-29 ino1 ura3-52 gcd14-2 (HIS4::lacZ ura3-52) leu2::hisG::URA3::hisG lhp1::LEU2	This study
F104	MATα ura3-52	G. Fink
F105	MATa ura3-52	G. Fink
F294	MATa prt1-1 ura3-52 ade2-1 leu2-3 leu2-112	G. Johnston
JAy113	MATa leu2,3-112 trp1Δ63 ura3-52 lhp1::hisG::URA3::hisG (p2705::GCD10-HA)	This study
JAy206	MATα gcd10Δ1 ura3-52 trp1 leu2Δ 1 his3Δ200 pep4::HIS4 prb1Δ6 can1 (pJA128::GCD10)
JAy142	MATα gcd10Δ1 ura3-52 trp1 leu2Δ1 his3Δ200 pep4::HIS4 prb1Δ6 can1(GCD10-HA URA3)	3
JAy143	MATα gcd10Δ1 ura3-52 trp1 leu2Δ1 his3Δ200 pep4::HIS4 prb1Δ6 can1 (GCD10 URA3)	3
W303 1A	MATa ura3-52 leu2-3 leu2-112 trp1-289 his3-11 ade2-1 can1-100	F. del Rey
W303 1B	MATα ura3-52 leu2-3 leu2-112 trp1-289 his3-11 ade2-1 can1-100	F. del Rey
YNG1	MATa/α ura3-52 /ura3-52 leu2-3 leu2-112/leu2-3 leu2-112 trp1-289 /trp1-289 his3-11/his3-11 ade2-1/ade2-1 can1-100/can1-100	This study
YNG174	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 (HIS4::lacZ ura3-52)	This study
YNG175	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 (HIS4::lacZ ura3-52) gcd14-1	This study
YNG176	MATa gcn2-101 gcn3-101 his1-29 ino1 ura3-52 (HIS4::lacZ ura3-52) gcd14-2	This study
YRC1	MATa/α gcn2-101/gcn2::LEU2 ura3-52/ura3-52 leu2-3 leu2-112/leu2-3 leu2-112 his1-29/HIS1 ino1/INO1 gcd14-2/GCD14	This study
YRC2	MATa/α ura3-52/ura3-52	This study

A 3.2-kb SpeI-ClaI fragment from pRC560 containing gcd14::URA3 was used for one-step gene disruptions (52) of one of the two GCD14 alleles in diploid strain YRC1 or YRC2, constructed by crossing Hm316 with H753 and F104 with F105, respectively (Table 1). Total DNA was isolated from Ura⁺ transformants and analyzed by Southern blotting to verify that replacement by the null allele was successful. Heterozygous GCD14/gcd14::URA3 or gcd14/gcd14::URA3 diploids were sporulated, and tetrad analysis was conducted as described in Results.

The diploid strain YNG1 was constructed by crossing W3031A (ura3-52 leu2-3,112) with W3031B (ura3-52 leu2-3,112) (Table 1). One allele of GCD14 in YNG1 was replaced by gcd14::URA3 (as described above), and a Ura⁺ Leu⁻ heterozygous GCD14/ gcd14::URA3 diploid was transformed with the high-copy-number LEU2 plasmid p1775 containing IMT4 (20). A Leu⁺ diploid transformant was sporulated, and tetrad analysis was conducted with the following results: a 4:0 segregation for viability was observed in ascospores of 7 tetrads of 12 analyzed, showing that p1775 (IMT4) rescued the lethality of a gcd14::URA3 deletion.

To construct strain Hm397, a gcd10-505 gcd14-2 ascospore clone was isolated from a tetratype tetrad obtained from a cross between Hm296 and Hm298 (23). The genotypes of all four ascopore clones from this tetrad were transformed with empty vectors or low-copy-number plasmids containing GCD10 or GCD14, followed by phenotype testing of the transformants. The LEU2 gene was disrupted in the gcd10-504 gcd14-2 strain by transforming it with a 6.5-kb BglII fragment from pNK985 (1), containing the 3.8-kb hisG::URA3::hisG cassette inserted at the EcoRI site of the LEU2 gene, and Leu⁻ Ura⁺ transformants were selected. The transformants were plated on SD medium containing 1 mg of 5-fluoro-orotic acid (5-FOA) per ml (6) to select the Ura⁻ derivative Hm397 (see Table 1 for genotypes). Hm397 was transformed with low-copy-number plasmids containing GCD10 (pJA128), GCD14 (pRC56), or empty vectors (pRS315 or pRS316) to generate the isogenic transformants with the genotypes GCD10 gcd14-2 (Hm420), gcd10-505 GCD14 (Hm421), GCD10 GCD14 (Hm422), and gcd10-505 gcd14-2 (Hm423).

LHP1 was replaced in GCD14 and gcd14 strains with a null allele in which the coding sequence was completely replaced with LEU2 (60) as follows. A Ura⁻ derivative of strain H117 was selected in 5-FOA medium, and this strain, named YNG174 (GCD14), along with Hm295 (gcd14-1) and Hm296 (gcd14-2), was transformed with the 6.5-kb BglII fragment from pNK985 (1) described above to disrupt the LEU2 gene in each strain with hisG::URA3::hisG sequences. Leu⁻ Ura⁺ transformants were selected and transformed with a 4.6-kb SalI-XbaI fragment from pKOIII (60) containing the null allele lhp1::LEU2. Total RNA was isolated from Leu⁺ transformants and analyzed by Northern blotting to verify the absence of LHP1 mRNA in isogenic strains Hm406 (GCD14 lhp1::LEU2), Hm407 (gcd14-1 lhp1::LEU2), and Hm408 (gcd14-2 lhp1::LEU2).

Strain JAy113 was constructed by transforming H1515 (MATa ura3-52 leu2-3,112 trp1Δ63) to Ura⁺ with a 4.3-kb SphI/EcoRI fragment isolated from pJA103 containing lhp1::hisG::URA3::hisG. Disruption of LHP1 was confirmed by PCR amplification of fragments diagnostic of lhp1::hisG::URA3::hisG from JAy113 genomic DNA. Strain JAy206 was constructed by transforming JAy143 (3) to Leu⁺ with pJA128, bearing single-copy GCD10 and LEU2, and a stable Ura⁻ derivative exhibiting wild-type growth was isolated on medium containing 1 mg of 5-FOA per ml (6).

Media and genetic techniques.

Sensitivity to 3AT (Sigma; catalog no. A8056) was tested as previously described (28). Transformation of yeast strains was carried out as described by Ito et al. (34). Standard genetic techniques and media were as previously described (55).

Chromosomal mapping of GCD14.

GCD14 was physically mapped to chromosome X by using a Chromo-Blot (Clontech), in which yeast chromosomes were fractionated by clamped homogeneous electrical field electrophoresis and blotted to a nylon membrane. Two individual blot lanes were probed, hybridizing a radiolabeled 2.8-kb HpaI-BamHI fragment from pRC5 that contains most of the GCD14 coding sequence. The same probe and a radiolabeled 2.4-kb XbaI fragment from pRC55 were used to hybridize a set of filters containing an ordered lambda library of yeast genomic DNA (49). The first probe hybridized to clone 70692, and the second hybridized with the same clone and also with clone 70091, both bearing inserts corresponding to containing a region encompassing TIF2 and URA2 sequences on in the left arm of chromosome X. To map GCD14 genetically, we crossed H168 (gcn2-101 gcn3-101 gcd14-2 ino1) with H753 (gcn2::LEU2 GCD14 INO1) and analyzed cosegregation of the 3AT^r, Slg⁻, and Ino⁻ phenotypes in 45 tetrads. We observed 32 parental ditype, 0 nonparental ditype, and 13 tetratype asci, thus locating GCD14 at 14.4 centimorgans from INO1.

DNA sequence analysis and cloning of gcd14 alleles.

The nucleotide sequence of a 2.8-kb SpeI-ClaI fragment in pRC56 was obtained for both strands by using a Sequenase version 2.0 kit (U.S. Biochemical Corp). Reverse primer, M13-40 primer, and several oligonucleotides derived from the previously determined sequences were used as primers, and pRC56 was used as double-stranded DNA template. The DNA sequence was analyzed by using the DNAsis sequence analysis program (Hitachi software), and comparisons of the GCD14 sequence were made with BLAST programs (2).

The gcd14-1 and gcd14-2 alleles were cloned by PCR with genomic DNAs from strains H160 and H168, respectively, as templates and the Expand High Fidelity PCR System (Boehringer Mannheim). The primers consisted of oligonucleotides having 20 bases corresponding to sequences at the beginning (200 bp upstream of the ATG) or the end of the GCD14 coding sequence. Amplification products corresponding to gcd14-1 or gcd14-2 were isolated from two independent PCRs in each case, subcloned into the pGEM vector (Promega), and sequenced on both strands by using the appropriate oligodeoxynucleotides derived from the previously determined GCD14 sequence as primers.

Production of Gcd14p-specific antiserum.

The TrpE-Gcd14p fusion protein encoded in plasmid pTRPC14 was overexpressed in E. coli HB101 and partially purified from the insoluble fraction of whole-cell extracts by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis as already described (21). Gel slices containing ∼200 μg of the fusion protein were used to inoculate rabbits. Immunization of the rabbits and collection of the immune antisera were conducted by Hazelton Laboratories (Vienna, Va.).

RNA isolation and Northern blot analysis.

Total RNA was isolated as previously described (36). Samples were dried, resuspended in loading buffer (90% formamide, 1 mM EDTA [pH 8.0], 0.1% bromophenol blue, 0.1% xylene cyanol), and heated to >90°C for 10 min before separation on a 6 or 8% polyacrylamide-bisacrylamide (19:1)–8.3 M urea gels by electrophoresis at 250 V for approximately 3.5 h. Gels were run at 250 V for 20 min prior to separation of RNAs. Following electrophoresis, gels were soaked in 0.5× Tris-borate-EDTA for 20 min, and RNA was transferred to positively charged nylon membranes (Boehringer Mannheim) in the same buffer by electrotransfer (Trans-Blot Cell; Bio-Rad) for 3.5 h at 19 V. RNA was immobilized on membranes by UV cross-linking with a UV-Stratalinker 2400 (Stratagene) according to the manufacturer’s instructions. tRNAs on the membranes were detected by hybridization with radiolabeled oligonucleotides in hybridization buffer (0.25 M Na₂HPO₄ [pH 7.5], 7% SDS, 1% bovine serum albumin, 1 mM EDTA) at 50 to 52°C for 12 to 20 h. Oligonucleotides were radiolabeled at the 5′ ends with [γ-³²P]ATP (6,000 Ci/mmol) and T4 polynucleotide kinase (Pharmacia). Following hybridization, membranes were washed once with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS for 30 min at room temperature and once in 1× SSC–0.1% SDS at 60°C for 20 min prior to detection by autoradiography. Direct quantitation of all hybridization signals was conducted by phosphorimager analysis with a BAS-1500 PhosphorImager and MacBAS Ver.2.x software (Fuji Film). For reprobing, membranes were stripped of radioactive probes by incubation at 100°C in 1.0% SDS and cooled to room temperature.

The oligonucleotides used as probes in Northern blots and the corresponding transcript species to which they specifically hybridize (RNA) were as follows: (i) 5′-TCGTTTCGATCCCGAGGACATCAGGGTTATGA-3′ (tRNA_i^Met), (ii) 5′-TCGGTTTCGATCCCGAGGACATCAGGGTTATGA-3′ (tRNA_e^Met), (iii) 5′-TGCTCGAGGTGGGGWTTGAACCCACGACGG-3′ (tRNA_UAU^Ile), (iv) 5′-CAGTTGATCGGACGGGAAAC-3′ (5S rRNA), (v) 5′-TGCGTTCTTCATCGATGCGAGAACC-3′ (5.8S rRNA), (vi) 5′-GACGTCCTACGATTGCACTC-3′ (RPR1 RNA), (vii) 5′-TCCCCCGGGTGAATCCATGGACCAAGA-3′ (NME1 RNA), (viii) 5′-GGTTCATCCTTATGCAGGG-3′ (U6 RNA), (ix) 5′-AGCCGAACTTTTTATTCCATTCG-3′ (pre-tRNA_CGA^Ser) (61), and (x) 5′-AGCCCAAGAGATTTCGAGTCTCTCG-3′ (tRNA_CGA^Ser) (61).

Nucleotide sequence accession number.

The EMBL accession number for the GCD14 nucleotide sequence is Z54149.

RESULTS

Isolation and characterization of the wild-type GCD14 gene and of the gcd14-1 and gcd14-2 mutant alleles.

Mutations in GCD14 were isolated as suppressors of the 3AT phenotype of a gcn2-101 gcn3-101 double mutant (17). All gcn mutations confer sensitivity to 3AT because they impair derepression of GCN4, and of histidine biosynthetic genes regulated by Gcn4p, in response to histidine starvation imposed by 3AT (28). gcd14 mutations restore derepression of GCN4 translation and thus confer 3AT resistance in a gcn2 gcn3 background. In addition, recessive gcd14 mutations lead to a slow-growth phenotype (Slg⁻) on rich media at 28°C which is exacerbated at both lower and higher temperatures (18 and 37°C) (Ts⁻ phenotype) (data not shown). We cloned the wild-type allele of GCD14 from a yeast genomic library on plasmid pRC5 (Fig. 1) by complementing the 3AT^r and Slg⁻ phenotypes conferred by gcd14-2 in the gcn2-101 gcd14-2 strain Hm316. To prove that we had cloned authentic GCD14, we mapped the chromosomal location of the genomic insert in pRC5 to chromosome X of S. cerevisiae. A radiolabeled 2.8-kb HpaI-BamHI fragment obtained from pRC5 hybridized exclusively to chromosome X in a yeast Chromo-Blot (see Materials and Methods). The cloned sequence was mapped to the left arm of chromosome X, in a region close to SPT10/PBS2/TRK1, by hybridizing the same HpaI-BamHI fragment to filters containing an ordered lambda library of yeast genomic clones (see Materials and Methods). From tetrad analysis we estimated that gcd14-2 is about 14.4 centimorgans from INO1 (see Materials and Methods). The fact that INO1 is located 50 kb from the cloned sequence on the left arm of chromosome X verifies that the insert in pRC5 contains the wild-type allele of GCD14.

To define the boundaries of GCD14, subclones of the genomic insert in pRC5 were constructed in low-copy-number plasmids and tested for complementation of the gcd14-1 and gcd14-2 mutations (Fig. 1). The results of this analysis localized GCD14 to the 2.8-kb SpeI-ClaI fragment in pRC56 (Fig. 1 and data not shown), since deletions that remove sequences from the 5′ or the 3′ end (pRC55) of that fragment completely abolished complementation activity (Fig. 1). The SpeI-ClaI fragment was sequenced on both strands and found to contain a single ORF of 1,149 bp that is predicted to encode a protein of 383 amino acids with a molecular mass of 43,917 Da. (Fig. 2). Comparison of the deduced amino acid sequence of Gcd14p with protein sequences in the GenBank, EMBL, and SWIS-PROT databases revealed that the gene had 60% similarity and 47.5% identity with Schizosaccharomyces pombe CPD1, a gene proposed to encode the 42-kDa subunit of DNA polymerase δ (62).

FIG. 2.

Deduced amino acid sequence of Gcd14p. The predicted amino acid sequence of S. cerevisiae Gcd14p is given in single-letter code. Two regions of sequence similarity observed in several S-AdoMet-dependent methyltransferases are boxed and indicated as motif I and motif II (35). The conserved G residue at position 5 of motif I is G118 in Gcd14p. A glutamate residue commonly located 17 to 19 residues C terminal to motif I is found 17 residues C terminal to motif I in Gcd14p, at position 139 (*), and a cluster of hydrophobic residues, hhXh (D/E), at positions 135 to 138 (underlined, LFSF) precedes the E element. The central invariant aspartate in motif II is conserved at position 209 in Gcd14p. Motifs I and II are separated by 83 residues in Gcd14p. A putative nuclear localization signal (7) (shaded box) is located between residues K282 R294 in the Gcd14p sequence.

The predicted polypeptide sequence of Gcd14p contains two regions of sequence similarity described for several S-AdoMet-dependent methyltransferases (35). These motifs always occur in the same order separated by intervals of similar lengths, and it was suggested that they contribute to binding of the substrate S-AdoMet or the product S-adenosylhomocysteine of S-AdoMet-dependent methyltransferases. Motif I (VIEAGTGSG), located between residues 114 and 122 in Gcd14p (Fig. 2), is similar to conserved regions described for DNA adenine and cytosine methyltransferases (33) and was also found in 69 of 84 non-DNA methyltransferases (35). Gcd14p also contains a conserved glutamate (E) residue located 17 residues C terminal to motif I and a cluster of hydrophobic residues, LFSF, at positions 135 to 138 preceding the E element, thus matching the established consensus hhXh (D/E). Motif II (APWDAIPH), located between amino acid residues 206 and 213 in Gcd14p (Fig. 2), is present in 46 of the 84 enzyme sequences analyzed (35), including one bacterial tRNA methyltransferase (33). Generally, motifs I and II are separated by 57 ± 13 amino acids, whereas in Gcd14p the separation is 82 residues; however, the RNA methyltransferases and porphyrin precursor methyltransferases are known exceptions to this rule.

We identified the mutations in gcd14-1 and gcd14-2 by PCR amplification of both alleles from genomic DNAs of H160 and H168, respectively, followed by automatic DNA sequencing (see Materials and Methods). The gcd14-1 allele contains a single point mutation, consisting of a transversion from G to C at nucleotide 265, replacing a histidine codon (GAC) with an aspartate codon (CAC) at amino acid residue 89. This mutation falls in a region of ca. 23 amino acid residues that is very well conserved between CPD1 of S. pombe and GCD14; however, it does not alter the predicted S-AdoMet binding motifs in Gcd14p. The gcd14-2 allele contains a single G-to-T transversion in nucleotide 3 of the GCD14 coding sequence, replacing the GCD14 methionine start codon ATG (Met) with ATT (Ile).

Immunoblot analysis of whole-cell extracts with antibodies against Gcd14p (see Materials and Methods) showed that the gcd14-1 product was expressed at a level greater than or equal to that of the wild type, whereas the gcd14-2 product was substantially diminished in abundance (data not shown). Presumably, a low level of gcd14-2 protein can be produced by inefficient translation initiation at the ATT start codon.

GCD14 is an essential gene.

The slow-growth phenotype of the gcd14-1 and gcd14-2 mutations suggests that Gcd14p has an essential function beyond its role in GCN4-specific translational control. To test this possibility, we constructed a deletion-insertion null allele, gcd14::URA3, in plasmid pRC560 (Fig. 1) and used it to replace one allele of GCD14 in ura3-52/ura3-52 diploid strains YRC1 (gcn2/gcn2 GCD14/gcd14-2 ura3-52/ura3-52) and YRC2 (GCN2/GCN2 GCD14/GCD14 ura3-52/ura3-52) (see Materials and Methods). We verified by Southern blot analysis that the gcd14::URA3 allele replaced one of the two copies of GCD14 in the Ura⁺ transformants of YRC1 and YRC2 that were obtained (data not shown). Tetrad analysis of these transformants revealed that in 24 asci from each diploid, only two of the four spores formed colonies on rich medium after 4 days at 28°C (data not shown). All viable spores were Ura⁻ (ura3-52) (and thus contain GCD14), and those from the YRC1 transformant were also 3AT^s (ura3-52 gcn2 GCD14), indicating that the gcd14-2 allele had been replaced by gcd14::URA3. These results demonstrate that GCD14 is essential for growth.

To confirm that GCD14 was the only essential gene whose function had been disrupted, a Ura⁻ derivative of the heterozygous gcd14::URA3/GCD14 transformant of diploid strain YRC2 was isolated by growth on 5-FOA medium (6), transformed with a URA3 low-copy-number plasmid containing only the GCD14 coding region (pRC56 [Fig. 1]), and induced to sporulate. Tetrads that contained four viable ascospores were dissected, showing that GCD14 complements the lethal phenotype of haploid spores bearing the gcd14::URA3 mutation.

gcd14-2 cells exhibit a modest defect in general translation initiation.

The essential nature of the GCD14 gene together with its role as a repressor of GCN4 mRNA translation (17) suggested to us that Gcd14p could be required for general initiation of translation. To investigate that possibility, we analyzed total polysome profiles from wild-type and mutant gcd14-2 strains by fractionating whole-cell extracts on sucrose gradients by velocity sedimentation. The gcd14-2 mutants do not form colonies at 37°C; however, they exhibited little or no growth defect in liquid medium after several hours at 37°C (data not shown). Quantitation of total polysome profiles for gcd14-2 and GCD14 strains grown at 23°C to mid-logarithmic phase and then incubated for 4 h at 37°C revealed no significant differences in the polysome/monosome (P/M) ratio. Accordingly, we next examined polysome profiles for the isogenic strains Hm296g (gcd14-2) and Hm296G (GCD14) grown at 39°C, where the gcd14-2 mutant exhibits a more detectable growth defect in liquid medium. As shown in Fig. 3, the P/M ratio was considerably lower in the gcd14-2 mutant (P/M ratio, 2.4) than in its isogenic GCD14 parent strain (P/M ratio, 5.7) after 1.5 h at 39°C. There was also a small reduction in the average size of the polysomes in the gcd14-2 mutant at 39°C (see Fig. 3 legend). Although these differences in polysome size and content are not dramatic, they were consistently observed in three independent experiments (see Fig. 3 legend). These data suggest that the gcd14-2 mutants exhibit a modest defect in general translation initiation. The magnitude of this defect is less than that observed in gcd1 or gcd2 mutants, consistent with the lesser derepression of GCN4 translation associated with gcd14 versus gcd1 or gcd2 mutations (17, 26).

FIG. 3.

Polysome profiles of isogenic gcd14-2 and GCD14 strains. Isogenic strains Hm296g and Hm296G were cultured overnight in yeast extract-peptone-dextrose at 23°C and used to inoculate two flasks containing 300 ml of yeast extract-peptone-dextrose to an optical density at 600 nm of ∼0.1. Cultures were incubated at room temperature in a rotary shaker at 250 rpm. At an optical density at 600 nm of ∼4.0, cells were collected and resuspended in the appropriate volume of yeast extract-peptone-dextrose prewarmed to 39°C and grown to an optical density at 600 nm of ∼2.0. From both flasks, 150 ml was collected immediately for the 39°C, t = 0 samples, and 150 ml of fresh prewarmed yeast extract-peptone-dextrose was added back to each flask. Samples of 150 ml were collected from both flasks after 1.5 h and processed for polysome analysis by velocity sedimentation of whole-cell extracts on sucrose gradients (22). Gradients were fractionated while being scanned at 254 nm, and the resulting absorbance profiles are shown, with the tops of the gradients on the left. The positions of 40S, 60S, and 80S ribosomal species are indicated by arrows. For each strain, the top two gradients correspond to samples incubated at 39°C for 0 h (t = 0) and the two bottom gradients contain samples obtained after 1.5 h of incubation at 39°C. Areas delimited inside the profiles correspond to the fractions of the total absorbance profiles represented by 2-mer and 3-mer polysomes. In the gcd14-2 strain, the 2-mers and 3-mers represented 26% of the polysome mass at the permissive temperature but comprised 32% of the polysomes after 1.5 h at 39°C. In contrast, 2-mers and 3-mers represented 24 and 25% of the polysome mass at 23°C and after 1.5 h at 39°C, respectively, in the wild-type strain Hm296G. The P/M ratios were calculated at t = 0 and t = 1.5 h for three independent experiments, and the following mean values and standard errors (in parentheses) were obtained: at t = 0, P/M = 4.2 (0.83) for GCD14 and 2.8 (0.44) for gcd14-2; at t = 1.5 h, P/M = 5.2 (0.35) for GCD14 and 3.5 (0.72) for gcd14-2.

Genes encoding initiator tRNA^Met or Lhp1p are dosage suppressors of the Gcd⁻ and Slg⁻ phenotypes of gcd14 mutants.

gcd10 and gcd14 mutations lead to constitutive derepression of GCN4 translation in the absence of Gcn2p (Gcd⁻ phenotype), conferring resistance to 3AT in a gcn2 strain background (17, 26). We found recently (3) that the 3AT^r and Ts⁻ phenotypes conferred by gcd10 mutations in a gcn2 background can be fully suppressed by the presence in high copy number of any of the four genes, IMT1 to IMT4, encoding tRNA_i^Met (11, 14) and can be partially suppressed by high-copy-number LHP1 (39, 60). In view of the physical interaction found between Gcd10p and Gcd14p (3), we asked whether the dosage suppressors of gcd10 mutations could also suppress the 3AT^r and Slg⁻ phenotypes conferred by gcd14 mutations. Each of the four IMT genes on high-copy-number plasmids (hcIMT) suppressed the 3AT^r (data not shown) and Slg⁻ (Fig. 4) phenotypes in gcd14-1 gcn2-101 and gcd14-2 gcn2-101 mutants to about the same extent as did a low-copy-number plasmid containing GCD14, whereas high-copy-number LHP1 (hcLHP1) only partially suppressed the mutant Slg⁻ phenotype (Fig. 4 and data not shown). hcGCD10 had no suppressor activity in the gcd14 mutants (Fig. 4), and neither single-copy GCD14 nor hcGCD14 suppressed the phenotypes of mutants containing different gcd10 alleles (data not shown). Furthermore, suppression by hcIMT and hcLHP1 seems to be specific for gcd10 and gcd14, as temperature-sensitive mutations in genes encoding subunits of translation initiation factors eIF2 and eIF3 (sui2-1 and prt1-1, respectively) were not suppressed by these dosage suppressors (3).

FIG. 4.

Phenotypes of gcd14 dosage suppressors. Eight independent transformants of strains Hm295 (gcd14-1) and Hm296 (gcd14-2) carrying empty vector pRS426 (gcd14), a low-copy-number plasmid bearing GCD14 (pRC56), or high-copy-number plasmids bearing IMT1 (p2632), IMT2 (p2633), IMT3 (p2634), IMT4 (p2635), GCD10 (pE107), or LHP1 (p2636) were streaked for single colonies on minimally supplemented SD plates and incubated at 28°C (top plates) or 37°C (bottom plates) for 2 days. The locations of the transformants on plates are indicated inside the central schematic.

gcd14 mutants are defective in processing the primary precursors of the initiator tRNA^Met.

High-copy-number IMT1, -2, -3, and -4 and LHP1 suppress the phenotypes of gcd10 mutants because these mutants contain reduced amounts of mature tRNA_i^Met, and the dosage suppressors compensate for this defect by increasing mature tRNA_i^Met to nearly wild-type (hcLHP1) or higher-than-wild-type (hcIMT) levels (3). Accordingly, we used Northern analysis to investigate whether steady-state levels of mature tRNA_i^Met were also reduced in gcd14 mutants H160 (gcd14-1) and H168 (gcd14-2) compared to that in their parental wild-type strain (H117). The probe for tRNA_i^Met, containing sequences complementary to nucleotides 33 to 64 in tRNA_i^Met, hybridizes to both mature tRNA_i^Met and different precursor tRNA_i^Met species bearing unique 5′-3′ extensions encoded by the different IMT genes (3) (see Materials and Methods). The Northern data in Fig. 5 were quantified with a phosphorimager, and the results are given in Table 2. Relative to the level of elongator tRNA^Met (tRNA_e^Met), we calculated that the amounts of mature tRNA_i^Met in the mutant strains were ca. half of that in the wild-type strain both at 28°C and after 1.5 h at 37°C (Fig. 5A; Table 2). In addition, the levels of tRNA_i^Met precursors were 2.3- and 3.2-fold higher at 28 and 37°C, respectively, leading to a fivefold increase in the ratio of precursor to mature tRNA_i^Met in the gcd14 mutants versus the wild type (Fig. 5A; Table 2). These data strongly suggest that Gcd14p is required for efficient processing of the nascent tRNA_i^Met transcripts. It is noteworthy that tRNA_i^Met species migrating faster than the wild-type mature form were detected in both gcd14 mutants under all conditions (species f in Fig. 5A), suggesting that fully 5′-3′-processed tRNA_i^Met is either unstable or incorrectly processed in these mutants. Similar results were obtained for a gcd14-2 mutant by Anderson et al. (3), although the accumulation of pre-tRNA_i^Met at 37°C was less extensive than that observed here.

FIG. 5.

gcd14 mutants are defective in processing the primary precursors of initiator tRNA^Met and tRNA_UAU^Ile. Northern blot analysis of total RNA (10 μg) isolated from strains H117 (GCD14), H160 (gcd14-1), and H168 (gcd14-2) grown in yeast extract-peptone-dextrose medium at 28°C to mid-exponential phase (0 h at 37°C) and shifted to 37°C for 1.5 h is shown. The blot was probed with a radiolabeled oligonucleotide that specifically hybridized to tRNA_i^Met (A) and then was stripped and reprobed with radiolabeled oligonucleotides specific for tRNA_e^Met (B) or tRNA_UAU^Ile (C) (see Materials and Methods). The positions of pre-tRNA_i^Met species, mature tRNA_i^Met, tRNA_e^Met, and primary (upper band) and 5′- and 3′-end-processed intron-containing (lower band) pre-tRNA_UAU^Ile and mature tRNA_UAU^Ile are indicated on the left. Indicated on the right are the positions of pre-tRNA_i^Met containing 5′ and 3′ extensions encoded by IMT2 and IMT3 (b), pre-tRNA_i^Met containing 5′ and 3′ extensions encoded by IMT1 and IMT4 (c), mature tRNA_i^Met (e), and aberrantly processed tRNA_i^Met species (f) (see text for details).

TABLE 2.

Comparison of tRNA_i^Met and tRNA_UAU^Ile levels in GCD14 and gcd14 isogenic strains^a

Parameter and tRNA	Value for:
	GCD14 strain		gcd14-1 strain		gcd14-2 strain
	28°C	37°C, 1.5 h	28°C	37°C, 1.5 h	28°C	37°C, 1.5 h
Relative level (% of wild type at 28°C)
pre-tRNAiMet	100	126	237	320	234	324
tRNAiMet	100	95	48	46	51	46
tRNAeMet	100	102	100	82	101	165
Primary pre-tRNAUAUIle	100	172	172	192	167	172
Intron-containing pre-tRNAUAUIle	100	80	113	76	139	94
tRNAUAUIle	100	115	75	79	101	87
Ratio of primary precur-sors to mature tRNA
tRNAiMet	0.03	0.04	0.16	0.23	0.15	0.23
tRNAUAUIle	0.05	0.08	0.12	0.13	0.10	0.11

^aThe relative intensities of the hybridization signals in the autoradiograms shown in Fig. 5 were quantified by phosphorimaging analysis, and the data were normalized to the tRNA_e^Met signals. The amounts of each tRNA species (precursor or mature forms) measured in each strain at 28°C or after 1.5 h at 37°C were expressed relative to the corresponding values in the wild-type strain at 28°C, which were set to 100%. The ratios of pre-tRNA_i^Met to mature tRNA_i^Met, or of primary pre-tRNA_UAU^Ile to mature tRNA_UAU^Ile are listed in the lower half of the table. 

As described for gcd10-504 (3), the gcd14-1 and gcd14-2 mutations had little or no effect on the size or steady-state amount of elongator tRNA^Met (Fig. 5B; Table 2). In the case of tRNA_UAU^Ile, the gcd14-1 and gcd14-2 mutations led to small increases in the amounts of the primary transcript containing both 5′ and 3′ extensions plus the intron (slower-migrating species of pretRNA_UAU^Ile) and possibly a small reduction in the levels of mature tRNA_UAU^Ile in the mutant strains (Fig. 5C; Table 2). Consequently, the precursor tRNA/mature tRNA ratios for this tRNA were slightly elevated in the mutant versus the wild-type strain (Table 2), albeit not to the extent observed for tRNA_i^Met. A similar result was obtained for tRNA_GTA^Tyr (data not shown). These results raise the possibility that Gcd14p is required primarily for efficient removal of 5′ and 3′ extensions from initiator tRNA^Met but may make a small contribution to the processing of certain other tRNAs besides the initiator.

We confirmed by Northern analysis that the presence of hcIMT or hcLHP1 plasmids in the gcd14 mutants led to increased levels of mature tRNA_i^Met, accounting for their suppressor phenotypes. Transformants of the gcd14-1 strain containing hcIMT1, hcIMT2, hcIMT3, or hcIMT4 had levels of mature tRNA_i^Met that were 3- to 7-fold higher than that observed in a vector transformant and 1.5- to 3.5-fold higher than that seen in the wild-type GCD14 transformant (data not shown). High-copy LHP1 led to only a 1.5-fold increase in the level of mature tRNA_i^Met in the gcd14-1 mutant Hm295 at 37°C, reaching a level slightly below that observed in the GCD14 strain (data not shown). This last finding is in accordance with the fact that hcLHP1 is a relatively weak suppressor of the Gcd⁻ (data not shown) and Slg⁻ (Fig. 4) phenotypes of gcd14-1.

Additive effects of gcd10-504 and gcd14-2 mutations on processing and accumulation of tRNA_i^Met.

In view of the similar phenotypes of gcd14 and gcd10 mutations, we investigated possible genetic interactions between them. The gcd10-505 gcd14-2 double mutant Hm397 was transformed with low-copy-number plasmids containing GCD10, GCD14, or empty vectors to generate four isogenic transformants with the following genotypes: (i) gcd10-505 gcd14-2, (ii) GCD10 gcd14-2, (iii) gcd10-505 GCD14, and (iv) GCD10 GCD14. As shown in Fig. 6A, the gcd10-505 gcd14-2 double mutant grew more slowly at 28 and 34°C than did either single mutant, indicating additivity of the growth defects conferred by these mutations. Northern analysis of the transformants grown at 37°C showed that the gcd14-2 single mutant contained high levels of tRNA_i^Met precursors and a modest reduction in mature tRNA_i^Met levels relative to the wild-type strain (Fig. 6B; Table 3), similar to the results described above. Compared to the gcd14-2 mutant, the gcd10-505 single mutant at 37°C showed a similar accumulation of the precursors but a somewhat greater reduction in mature tRNA_i^Met expression at 37°C (Fig. 6B, lanes 6 to 8). Both mutants showed 7- to 10-fold increases in the precursor tRNA/mature tRNA ratios for tRNA_i^Met (Table 3). These last findings are interesting because they suggest that Gcd10p is also required for efficient end processing of tRNA_i^Met. The gcd10-505 gcd14-2 double mutant at 37°C showed a greater reduction in mature tRNA_i^Met levels than did either single mutant, exhibiting levels only 30% of that seen in the wild-type transformant (Fig. 6B, lanes 5 to 8; Table 3). The additivity of these defects in mature tRNA_i^Met expression suggests that Gcd10p and Gcd14p participate in the same process involved in pre-tRNA_i^Met processing.

FIG. 6.

Exacerbation of gcd14-2 phenotypes by a gcd10-505 mutation. (A) The double mutant Hm423 (gcd10-505 gcd14-2) and isogenic GCD10 gcd14-2 (Hm420), gcd10-505 GCD14 (Hm421), and GCD10 GCD14 (Hm422) strains were streaked for single colonies on minimally supplemented SD plates and incubated at 28, 34, or 37°C for 2 days. The genotypes of the transformants are indicated adjacent to the appropriate sectors of the 28°C plate. (B) Northern blot analysis of total RNA (10 μg) isolated from the same strains analyzed in panel A, conducted as described for Fig. 5 except that cells were grown in minimally supplemented SD at 28°C (t = 0) (lanes 1 to 4) and then shifted to 37°C for 4 h (lanes 5 to 8). The blot was probed for 5S rRNA, tRNA_i^Met, and tRNA_UAU^Ile as described for Fig. 5.

TABLE 3.

Effects of combining gcd10-505 and gcd14-2 mutations on expression of tRNA_i^Met and tRNA_UAU^Ilea

Parameter and tRNA	Value for:
	gcd10 gcd14 strain		gcd14 strain		gcd10 strain		GCD10 GCD14 strain
	28°C	37°C, 4 h	28°C	37°C, 4 h	28°C	37°C, 4 h	28°C	37°C, 4 h
Relative level (% of wild type at 28°C)
Pre-tRNAiMet	1,031	535	661	1,236	639	930	100	198
tRNAiMet	145	31	101	60	95	45	100	92
Primary pre-tRNAUAUIle	182	82	164	260	195	216	100	160
Intron-containing pre-tRNAUAUIle	104	62	103	98	100	71	100	69
tRNAUAUIle	76	85	80	77	82	65	100	95
Ratio of primary precursor to mature tRNA
tRNAiMet	0.37	0.88	0.34	1.06	0.35	1.06	0.05	0.11
tRNAUAUIle	0.35	0.14	0.30	0.50	0.35	0.50	0.15	0.25

^aThe relative intensities of the hybridization signals in the autoradiograms shown in Fig. 6B were quantified by phosphorimaging analysis, and the data were normalized to the 5S rRNA signals. The amounts of each tRNA species (precursor or mature forms) measured in each strain at 28°C or after 4 h at 37°C were expressed relative to the corresponding values in the wild-type strain at 28°C, which were set to 100%. The ratios of pre-tRNA_i^Met to mature tRNA_i^Met or of primary pre-tRNA_UAU^Ile to mature tRNA_UAU^Ile are listed in the lower half of the table. 

Following the temperature shift, the double-mutant culture increased in mass by only about 30% (data not shown), whereas the level of mature tRNA_i^Met was reduced by ca. 69% (Table 3). Therefore, it appears that the reduction in mature tRNA_i^Met levels cannot be accounted for by dilution of the preexisting mature molecules during cell growth following the temperature shift. This implies that a substantial fraction of the mature tRNA_i^Met was degraded at 37°C in the double mutant. It is noteworthy that the level of tRNA_i^Met precursors in the double mutant at 37°C, while higher than that in the wild-type strain, was lower than that seen in the single mutants. This may indicate that the unprocessed tRNA_i^Met precursors which accumulate in the double mutant also are degraded more rapidly at 37°C than in the single mutants. Finally, it is interesting that the single and double mutants showed ca. 10-fold-higher levels of the tRNA_i^Met precursors at 28°C compared to the wild-type strain, even though there was little or no reduction in mature tRNA_i^Met expression under these conditions (Table 3). Perhaps processing of tRNA_i^Met is slower in the mutants at all temperatures but this leads to a deficit in mature tRNA_i^Met levels relative to other stable RNAs only at the high growth rates occurring at elevated temperatures.

Similar to the results shown above for the gcd14 single mutants (Fig. 5), we observed modest (15 to 35%) reductions in mature tRNA_UAU^Ile expression in the single and double mutants analyzed in Fig. 6B; however, we did not observe additive reductions in the levels of mature tRNA_UAU^Ile upon combining gcd14-2 and gcd10-505 mutations (Table 3). There were also small increases in the amounts of the tRNA_UAU^Ile primary transcript in the single and double mutants at 28°C and in the single mutants at 37°C (Fig. 6B), suggesting that processing of this precursor occurred more slowly in the mutants. Interestingly, the primary tRNA_UAU^Ile precursor did not accumulate in the double mutant at 37°C, perhaps indicating that it is more susceptible to degradation, as suggested above for pre-tRNA_i^Met. Thus, as concluded above, the gcd10 and gcd14 mutations have considerably smaller effects on the maturation of pre-tRNA_UAU^Ile versus pre-tRNA_i^Met, particularly in the gcd10-505 gcd14-2 double mutant (Table 3). In fact, as shown next, maturation of tRNA_i^Met appears to be the only essential function of Gcd14p.

GCD14 is dispensable in yeast strains overexpressing initiator tRNA^Met.

Having shown that GCD14 is essential and that gcd14 mutations impair the maturation of pre-tRNA_i^Met, we asked whether overexpression of an IMT gene would allow cells to survive in the absence of Gcd14p. To test this possibility, we replaced one of the two GCD14 alleles in the wild-type diploid strain YNG1 with the gcd14::URA3 deletion-insertion allele and verified the gene replacement by Southern blot analysis (data not shown). After sporulation of the resulting GCD14/gcd14::URA3 diploid, tetrad analysis revealed the expected 2⁺:2⁻ segregation for cell viability, and all viable spores were Ura⁻ (bearing GCD14). When the GCD14/gcd14::URA3 diploid, which is homozygous for leu2, was first transformed with a high-copy-number LEU2 plasmid p1775 bearing IMT4 (20) and then subjected to tetrad analysis, 20 of 32 tetrads showed a 4⁺:0⁻ segregation for cell viability and 2⁺:2⁻ segregation for the Ura phenotype. Importantly, all of the Ura⁺ spores (bearing gcd14::URA3) were Leu⁺ (bearing hcIMT4 on p1775) (data not shown). These results showed that p1775 (hcIMT4) overcame the lethality of the gcd14::URA3 deletion. The gcd14::URA3 hcIMT4 and GCD14 hcIMT4 strains derived from one tetrad grew similarly at 28 and 37°C (Fig. 7A, spores 7A to 7D). In a second tetrad, however, the two gcd14::URA3 hcIMT4 strains grew more slowly than did the GCD14 hcIMT4 strains, particularly at 37°C (Fig. 7A, spores 5A to 5D). Thus, at least in certain genetic backgrounds (e.g., spores 7C and 7D), overexpression of tRNA_i^Met from IMT4 makes Gcd14p dispensable for growth at 28 and 37°C.

FIG. 7.

GCD14 is not essential in the presence of hcIMT4. (A) Ascospore clones from two four-spored tetrads (designated 5 and 7) were obtained from a heterozygous GCD14/gcd14::URA3 diploid (YNG1) containing high-copy-number plasmid p1775 (hcIMT4) (20). The four strains from each tetrad (5A to 5D and 7A to 7D), all containing hcIMT4, were streaked for single colonies on minimally supplemented SD medium and incubated at 28 or 37°C for 3 days. The position of each spore clone and its GCD14 genotype is indicated adjacent to the appropriate plate sector. +, wild-type GCD14; Δ, gcd14::URA3. (B) Northern blot analysis of total RNA (10 μg) isolated from the spore clones of tetrad 5 (strains 5A to 5D) bearing hcIMT4 (described for panel A). The strains were grown to mid-logarithmic phase in minimally supplemented SD medium at 28°C (t = 0 at 37°C) and shifted to 37°C for 4 h. The blot was probed for tRNA_i^Met, tRNA_e^Met, or tRNA_UAU^Ile as described for Fig. 5.

We conducted Northern blot analysis of the eight strains from the two tetrads just described, all containing hcIMT4. As shown in Fig. 7B, the gcd14::URA3 hcIMT4 strains accumulated 3- to 10-fold-higher levels of the pre-tRNA_i^Met species, most of which derive from hcIMT4 (3), than did the congenic GCD14 hcIMT4 strains. Thus, it appears that processing of this overproduced precursor is impaired in strains lacking GCD14. The mature tRNA_i^Met was reduced by ca. half at 37°C in only two of the four gcd14::URA3 hcIMT4 strains (spore clones 5D and 7D) versus the four GCD14 hcIMT4 strains (Fig. 7B and data not shown). Thus, in certain genetic backgrounds (spore clones 5B and 7C), overproduction of pre-tRNA_i^Met from hcIMT4 completely overcame the requirement for Gcd14p to express high levels of mature tRNA_i^Met. However, species migrating faster than mature tRNA_i^Met were visible in the two deletion strains in tetrad 5 (Fig. 7B) and also in tetrad 7 (data not shown), suggesting that a fraction of the mature fully processed tRNA_i^Met present in the gcd14::URA3 hcIMT4 strains may be improperly processed or partially degraded and thus be nonfunctional in translation. The gcd14::URA3 hcIMT4 strains showed little or no defect in expression of elongator tRNA^Met and the mature form of tRNA_UAU^Ile (Fig. 7B); however, there was a modest accumulation of the primary (slower-migrating) precursor for tRNA_UAU^Ile, particularly at 37°C, reaching at levels ca. 1.5- to 4-fold higher than those seen in GCD14 hcIMT4 strains (Fig. 7B).

Deletion of LHP1 exacerbates the phenotypes of gcd14 mutations.

To investigate the role of Lhp1p in processing of tRNA_i^Met precursors in gcd14 mutants, we disrupted LHP1 with the LEU2 gene in wild-type GCD14 and gcd14 mutant strains (see Materials and Methods). The lhp1::LEU2 mutation did not affect the growth rate of the wild-type strain H117, in agreement with previous results (60); however, it greatly exacerbated the Slg⁻ phenotypes of the gcd14-1 and gcd14-2 mutants (Fig. 8A). In addition, we observed additive effects of the lhp1::LEU2 and gcd14 mutations on the levels of mature tRNA_i^Met. Whereas the lhp1::LEU2 allele reduced mature tRNA_i^Met by only ca. 27% in GCD14 cells, it caused reductions of ca. 60% in gcd14-1 and gcd14-2 cells (Fig. 8B; Table 4). The levels of tRNA_i^Met precursors also were greatly diminished by the lhp1::LEU2 mutation in all three backgrounds, with reductions of 77, 86, and 68% in the GCD14, gcd14-1, and gcd14-2 strains, respectively (Fig. 8B; Table 4). Lhp1p has been implicated in stabilizing precursors and promoting correct 3′ end processing of certain tRNAs (61). Thus, to account for the reduced amounts of mature tRNA_i^Met seen in the double mutants, we suggest that deletion of LHP1 leads to degradation of a substantial fraction of the tRNA_i^Met precursors and that this degradation is more extensive in the gcd14 mutants than in the GCD14 strain.

FIG. 8.

Synthetic interactions of gcd14 and lhp1::LEU2 mutations. (A) Strains YNG174 (GCD14), YNG175 (gcd14-1), and YNG176 (gcd14-2) and the isogenic lhp1::LEU2-containing derivatives Hm406 (GCD14 lhp1::LEU2), Hm407 (gcd14-1 lhp1::LEU2), and Hm408 (gcd14-2 lhp1::LEU2) were streaked for single colonies on minimally supplemented SD plates and incubated at 28 or 37°C for 2 days. The locations of the strains are indicated by their relevant genotypes in the schematic at the top. (B) Northern blot analysis of total RNA (20 μg) isolated from the strains described for panel A and separated by electrophoresis in a 6% polyacrylamide–8.3 M urea gel. Strains were grown in minimally supplemented SD at 28°C. The blot was probed for tRNA_i^Met, tRNA_UAU^Ile, tRNA_e^Met, and tRNA_CGA^Ser by using the appropriate oligonucleotides (see Materials and Methods). Indicated on the right of the top panel are the positions of various precursor and processed forms of tRNA_i^Met, as described in the legend to Fig. 5. The positions of precursor and mature tRNA_CGA^Ser species are indicated on the right of the bottom panel: a primary precursor containing 5′ and 3′ extensions (g), a processing intermediate containing only the 3′ extension (h), a 5′- and 3′-end-processed intron-containing precursor (i), and mature tRNA_CGA^Ser (j). (C) The Northern blot in panel B was probed for RPR1 RNA, NME1 RNA, 5S rRNA, and U6 RNA by using the appropriate oligonucleotides (see Materials and Methods). The positions of precursor and mature RPR1 species are indicated on the left.

TABLE 4.

Effects of combining gcd14 and lhp1::LEU2 mutations on expression of tRNA_i^Met, tRNA_e^Met, tRNA_UAU^Ile, tRNA_CGA^Ser, and RPR1, NME1, and U6 RNAs^a

Parameter and tRNA	Value for:
	GCD14 strain	GCD14 lhp1::LEU2 strain	gcd14-1 strain	gcd14-1 lhp1::LEU2 strain	gcd14-2 strain	gcd14-2 lhp1::LEU2 strain
Relative level (% of wild type at 28°C)
Pre-tRNAiMet	100	23	126	18	236	74
tRNAiMet	100	73	37	15	52	20
tRNAeMet	100	96	112	114	120	92
Primary pre-tRNAUAUIle	100	62	118	79	116	79
Intron-containing pre-tRNAUAUIle	100	96	58	61	91	71
tRNAUAUIle	100	104	65	66	97	59
Primary pre-tRNACGASer	100	38	57	39	60	26
Intermediate precursor	100	—b	59	—	56	—
Intron-containing pre-tRNACGASer	100	61	49	24	30	33
tRNACGASer	100	94	94	32	75	50
Pre-RPR1	100	51	49	15	82	46
RPR1	100	73	68	24	60	52
NME1	100	61	100	28	98	28
U6	100	96	85	73	80	55
Ratio of primary precursor to mature RNA
tRNAiMet	0.08	0.025	0.30	0.08	0.35	0.31
tRNAUAUIle	0.04	0.03	0.08	0.05	0.05	0.06
tRNACGASer	0.28	0.11	0.17	0.34	0.22	0.15
RPR1	0.51	0.36	0.36	0.33	0.70	0.45

^aThe relative intensities of the hybridization signals in the autoradiograms shown in Fig. 8B and C were quantified by phosphorimaging analysis, and the data were normalized to the 5S rRNA signals. The amounts of each tRNA or RNA species (precursors or mature forms) measured in each strain at 28°C were expressed relative to the corresponding values in the wild-type strain at 28°C, which were set to 100%. The ratios of pre-tRNA_i^Met to mature tRNA_i^Met, of primary pre-tRNA_UAU^Ile to mature tRNA_UAU^Ile, of primary pre-tRNA_CGA^Ser to mature tRNA_CGA^Ser, and of pre-RPR1 to mature RPR1 are listed in the lower half of the table. 

^b—, undetectable. 

In contrast to the findings for tRNA_i^Met, there were relatively small effects of deletion of LHP1 on the levels of the precursor and mature forms of tRNA_UAU^Ile in all three strains, and there was little indication that the lhp1::LEU2 and gcd14 mutations had additive effects on expression of the fully processed form of this tRNA (Fig. 8B; Table 4). Thus, similar to our findings for gcd10 gcd14 double mutants (Fig. 6), expression of tRNA_UAU^Ile was relatively insensitive to mutations which showed strong additive effects on the production of mature tRNA_i^Met. No defects whatsoever in expression of elongator tRNA^Met in the single or double mutants relative to that in the wild type were observed (Fig. 8B; Table 4). However, we observed a substantial reduction (70%) in the level of mature tRNA_CGA^Ser in the gcd14-1 lhp1::LEU2 double mutant compared to the corresponding gcd14-1 single mutant and an additive effect of a lesser degree for this tRNA in the gcd14-2 lhp1::LEU2 double mutant (Fig. 8B; Table 4). Unlike the situation for tRNA_i^Met, the amounts of tRNA_CGA^Ser precursors were decreased, rather than increased, by the gcd14 single mutations; however, for both tRNAs, deletion of LHP1 greatly reduced the amounts of precursors in both GCD14 and gcd14 backgrounds. Thus, Gcd14p and Lhp1p may both function in stabilizing the tRNA_CGA^Ser precursors. It is unknown whether tRNA_CGA^Ser contains m¹A at position 58 (57).

We next considered the possibility that expression of other RNA polymerase III transcripts was also reduced in the gcd14 lhp1::LEU2 double mutants. As shown in Fig. 8C, the levels of 5S rRNA were nearly indistinguishable among the gcd14 and lhp1::LEU2 single and double mutants. In contrast, we observed strong additive effects of gcd14-1 and lhp1::LEU2 on expression of both the RPR1 and NME1 transcripts. These are the RNA components of RNase P and RNase MRP, which are involved in the processing of tRNAs and rRNA, respectively (38, 54). We observed a similar interaction between gcd14-2 and lhp1::LEU2 for NME1 RNA, but a lesser effect for RPR1 RNA, in this double mutant versus the gcd14-1 lhp1::LEU2 strain (Table 4). It is interesting that deletion of LHP1 reduced the amount of RPR1 precursor in all three strains, and this effect was especially marked in the gcd14-1 background. As suggested above for tRNA_i^Met, Lhp1p may increase the stability of pre-RPR1 RNA, and this function may be critically required for production of mature RPR1 in gcd14-1 mutants. Finally, U6 RNA showed modest reductions in both gcd14 lhp1::LEU2 double mutants compared to that in the corresponding single mutants, similar in magnitude to those described above for tRNA_UAU^Ile (Fig. 8B; Table 4). Taken together, the results in Fig. 8B and C strongly suggest that Gcd14p and Lhp1p cooperate in maturation and accumulation of a subset of RNA polymerase III transcripts. In the case of tRNA_i^Met, a requirement for Gcd14p can be readily observed in mutants defective for only this protein, whereas for tRNA_CGA^Ser, NME1, and RPR1, a strong dependence on Gcd14p is revealed only in the absence of Lhp1p.

We showed previously that Gcd14p copurified with a polyhistidine-tagged form of Gcd10p on nickel affinity resin and that neither protein was stably associated with the PRT1-encoded subunit of eIF3 (3). Moreover, Gcd10p and Gcd14p were not detected in a highly purified eIF3 complex (48), and both proteins showed prominent nuclear localization (3). Thus, it appears that Gcd10p and Gcd14p reside in a nuclear protein complex distinct from eIF3. In view of the genetic interactions between gcd14 mutations and the lhp1::LEU2 allele, it was of interest to determine whether Lhp1p is tightly associated with the Gcd10p-Gcd14p complex. To answer this question, we investigated whether Lhp1p could be coimmunoprecipitated from cell extracts with an HA-tagged form of Gcd10p. It was shown previously that the GCD10-HA allele employed for this study and wild-type GCD10 were indistinguishable in the ability to complement the lethal phenotype of a gcd10::URA3 mutation (3). Whole-cell extracts were prepared from isogenic GCD10-HA or GCD10 strains containing either a high-copy-number plasmid bearing LHP1 or an empty vector and were immunoprecipitated with monoclonal anti-HA antibodies. As expected, comparable proportions of Gcd14p and HA-Gcd10p were immunoprecipitated with anti-HA antibodies from the extracts prepared from strains containing GCD10-HA, but neither protein was immunoprecipitated from the GCD10 extract (Fig. 9A, lanes 4, 6, and 8). No detectable Lhp1p was coimmunoprecipitated with HA-Gcd10p and Gcd14p, even when Lhp1p was greatly overexpressed (Fig. 9A, lanes 4 and 6, and B, lane 6). In accordance with this finding, no Lhp1p was detectable in a preparation of the Gcd10p-Gcd14p complex purified to homogeneity (48a). We conclude that Lhp1p is not tightly associated with the Gcd10p-Gcd14p complex in cell extracts.

FIG. 9.

Gcd14p, but not Lhp1p, is tightly associated with Gcd10p in cell extracts. (A) Whole-cell extracts were prepared as described previously (47) from strains H1515 (LHP1) and YJA113 (lhp1) and from transformants of YJA142 (GCD10HA) or YJA143 (GCD10) bearing high-copy-number plasmid p2626 containing LHP1 or empty vector YEp24. Aliquots containing 200 μg of total cell protein were mixed with 2 to 4 μg of anti-HA monoclonal antibody HA.11 (Babco), which was prebound to protein A-Sepharose for 2 h on ice, and the mixture was incubated for 1 h at 4°C. After three washes with lysis buffer (20 mM Tris-HCl [pH 7.4], 100 mM KCl, 1.0 mM Mg acetate, 0.1% [vol/vol] Triton X-100) containing one complete protease inhibitor tablet (Boehringer Mannheim) per 25 ml, immunoprecipitates were collected by centrifugation and proteins were eluted by boiling in Laemmli buffer (37). The total proteins immunoprecipitated from 200 μg (pellet [P]; lanes 4, 6, and 8) or 50 μg of the starting whole-cell extracts (input [I]; lanes 3, 5, and 7) were separated by SDS-polyacrylamide gel electrophoresis (36) and transferred to a nitrocellulose membrane (Millipore) in 25 mM Tris–192 mM glycine–0.1% SDS containing 20% (vol/vol) methanol. Lanes 1 and 2 contain 50 μg of the starting whole-cell extracts from strains H1515 and YJA113, respectively, which were included to establish the identity of Lhp1p. The membrane was blocked overnight at 4°C in BLOTTO (5% [wt/vol] nonfat dry milk, 10 mM Tris-HCl [pH 7.4], 150 mM NaCl, 0.05% [vol/vol] Tween 20). A single immunoblot was probed with 3 μg of rabbit anti-HA antibody HA.11 (Babco) per ml, stripped, reprobed with a 1:1,000 dilution of anti-Gcd14p antibodies (see Materials and Methods), and reprobed with a 1:1,000 dilution of anti-Lhp1p antibodies (60). Immune complexes were detected with horseradish peroxidase-conjugated sheep antimouse (Amersham) or donkey antirabbit (Amersham) secondary antibodies and an enhanced chemiluminescence kit (ECL; Amersham). (B) A longer exposure of lanes 5 and 6 of the blot in panel A probed with anti-Lhp1p antibodies, included to detect small amounts of Lhp1p in the immunoprecipitates.

DISCUSSION

Evidence that Gcd14p is required for efficient processing of tRNA_i^Met.

Recessive gcd14 mutations confer constitutive derepression of GCN4 mRNA translation in the absence of eIF2α phosphorylation (17), a phenotype generally indicating reduced formation of the eIF2-GTP-tRNA_i^Met ternary complex (20). gcd14 mutants also have a Slg⁻ phenotype, and we found that general translation initiation is slightly impaired in these mutants, at least at elevated temperatures (Fig. 3). In addition, deletion of GCD14 is lethal. These findings suggest that Gcd14p has an essential function required for ternary complex formation in vivo. In agreement with this conclusion, we found that gcd14 mutants contain approximately 50% lower steady-state levels of mature tRNA_i^Met than are present in isogenic wild-type strains (Tables 2 to 4). They also contain processed tRNA_i^Met molecules that appear to be smaller than wild-type mature tRNA_i^Met (Fig. 5 to 7, species f) and thus may be defective in aminoacylation or ternary complex formation. The Gcd⁻ and Slg⁻ phenotypes of the gcd14 mutants were fully suppressed by introducing multiple copies of the IMT genes encoding tRNA_i^Met, which restored expression of mature tRNA_i^Met to greater-than-wild-type levels. Moreover, the presence of hcIMT4 overcame the lethality of a gcd14 deletion. The phenotypes of nonlethal gcd14 mutations also were partially suppressed by hcLHP1, which increased the level of mature tRNA_i^Met to just below wild-type levels. Together, these findings indicate that the essential function of Gcd14p is required for efficient and accurate production of mature tRNA_i^Met.

The fact that gcd14 mutants exhibit tRNA_i^Met precursor tRNA/mature tRNA ratios 5- to 10-fold higher than wild type (Fig. 5; Table 2) is most easily explained by a defect in pre-tRNA_i^Met processing. The precursor molecules that accumulate in gcd14 mutants contain both 5′ and 3′ extensions, indicating that both cleavage of the 5′ extension by RNase P (18) and removal of the 3′ extension by endo- or exonucleases occur less efficiently in gcd14 mutants. The accumulation of molecules shorter than wild-type mature tRNA_i^Met in gcd14 mutants could indicate that a portion of the 5′- and 3′-processed tRNA_i^Met lacks the CCA triplet added posttranscriptionally to the 3′ ends of all tRNAs (18). Alternatively, these shorter molecules could be degradation intermediates.

Evidence that Gcd14p and Gcd10p function together to promote maturation of tRNA_i^Met in vivo.

Nonlethal gcd10 mutations lead to reduced expression of mature tRNA_i^Met and can be suppressed by hcIMT and hcLHP1 (3), just as shown here for gcd14 mutations (Fig. 4). Moreover, the lethal effect of deleting GCD10 also could be suppressed by overexpression of tRNA_i^Met from high-copy-number IMT4 (3). Thus, the essential functions of both proteins, at least at low growth temperatures, are required only for expression of mature tRNA_i^Met. In addition to these genetic similarities, we found that Gcd10p and Gcd14p can be coimmunoprecipitated (Fig. 9) and copurified (3) from whole-cell extracts. The physical association between these two proteins has been confirmed by using a polyhistidine-tagged form of Gcd10p to isolate a Gcd10p-Gcd14p complex from whole-cell extracts by affinity chromatography (3) and by yeast two-hybrid analysis (4). In addition, epitope-tagged forms of both proteins Gcd10p and Gcd14p were shown by immunofluorescence to exhibit prominent nuclear localization in yeast cells (3). These findings strongly suggest that Gcd10p and Gcd14p reside in a nuclear complex that functions directly in the maturation of tRNA_i^Met.

Here we provided additional in vivo evidence for this conclusion by showing that gcd14-2 and gcd10-505 mutations had additive effects on cell growth and expression of mature tRNA_i^Met. The double mutant at 28°C accumulated higher levels of pre-tRNA_i^Met than did either single mutant, suggesting compounded defects in processing. Moreover, the gcd10-505 single mutant accumulated tRNA_i^Met precursors concomitant with diminished amounts of mature tRNA_i^Met (Fig. 6B), implicating Gcd10p in pre-tRNA_i^Met processing. In a previous study we showed that expression of mature tRNA_i^Met was reduced by ca. fivefold after several hours at 37°C in a gcd10-504 mutant; however, we observed no accumulation of pre-tRNA_i^Met at the restrictive temperature in that strain (3). This could be explained by proposing that the gcd10-504 and gcd10-505 (studied here) mutations both lead to defects in processing of pre-tRNA_i^Met but that the unprocessed molecules are more susceptible to degradation in the gcd10-504 mutant than in the gcd10-505 or gcd14 strains analyzed here. In accordance with this explanation, we obtained evidence that both precursor and mature tRNA_i^Met were highly susceptible to degradation in the gcd10-505 gcd14-1 double mutant at 37°C (Fig. 6B), similar to our previous findings for the gcd10-504 single mutant (3).

Thus far, we have observed no defect in pre-tRNA_i^Met synthesis or processing in cell extracts prepared from gcd10 or gcd14 mutants (3a). Therefore, Gcd10p and Gcd14p do not appear to be essential components of the processing machinery. Our discovery that Gcd14p contains two sequence motifs conserved among S-AdoMet-dependent methyltransferases, including at least one tRNA methyltransferase (35), raised the possibility that one or more nucleotides in pre-tRNA_i^Met is undermethylated in gcd14 mutants. This could alter the conformation of pre-tRNA_i^Met in a way that impairs removal of the 5′ or 3′ extensions or addition of the CCA triplet to the 3′ end. Consistent with this possibility, we recently found that m¹A is lacking in total tRNA isolated from gcd10Δ cells (3). This base modification occurs at position 58 in the initiator and elongator forms of tRNA^Met, in tRNA_GTA^Tyr (also examined here), and in 15 other tRNAs but is not found at any other positions in yeast tRNAs (57). In the initiator, the m¹A58 residue is involved in a unique tertiary substructure not observed in any elongator tRNAs, involving three residues unique to eukaryotic initiators, A60, A54, and A20 (5). Thus, the lack of methylation at position 58 could perturb the structure of initiator tRNA^Met, and impair its processing and stability, without similarly affecting elongator tRNA^Met or other tRNAs bearing m¹A58. It is unknown whether tRNA_UAU^Ile and tRNA_CGA^Ser, whose processing and accumulation were slightly impaired in gcd14 mutants, contain m¹A58.

It remains to be determined whether the Gcd14p-Gcd10p complex is directly responsible for m¹A58 formation in yeast tRNA and whether Gcd14p has methylase activity, as predicted from the presence of S-AdoMet binding motifs. Considering that Gcd10p has RNA binding activity (23), it might have a chaperone function that would facilitate methylation of pre-tRNA_i^Met by Gcd14p in wild-type cells. The results in Fig. 6B seem to indicate that the gcd10-505 mutation leads to degradation of the unprocessed precursors that accumulate in GCD10 gcd14-2 cells. Thus, Gcd10p may also be required to protect hypomethylated pre-tRNA_i^Met from exonucleolytic degradation. If Gcd10p performs a chaperone function, it could conceivably interact with mature tRNA_i^Met in the cytoplasm and promote formation of the ternary complex with eIF2 and GTP. This hypothetical function might explain why Gcd10p copurified with eIF3 activity (23), as stabilizing the ternary complex is a function ascribed to eIF3 (47).

Evidence that Lhp1p is required to protect tRNA_i^Met from degradation in gcd14 mutants.

In otherwise wild-type cells, deletion of LHP1 led to substantially decreased levels of precursor tRNA_i^Met and to only a small reduction in mature tRNA_i^Met levels (Fig. 8B; Table 4). This could be explained by postulating a role for Lhp1p in transcription of the IMT genes. However, in view of the allele-specific interactions between gcd14 mutations and the lhp1::LEU2 deletion (Fig. 8B), plus the evidence that Gcd14p functions in methylation and processing of pre-tRNA_i^Met, we favor the idea that Lhp1p affects tRNA_i^Met expression posttranscriptionally. As suggested for other yeast tRNAs (61), binding of Lhp1p to the poly(U) stretch at the 3′ end of pre-tRNA_i^Met could stabilize the conformation needed for precise endonucleolytic 3′ cleavage and block access by 3′-5′ exonucleases. In the absence of Lhp1p, a significant fraction of the pre-tRNA_i^Met would be incorrectly trimmed at the 3′ end and subsequently degraded, reducing the yield of mature tRNA_i^Met.

Deletion of LHP1 in the gcd14 mutants led to larger reductions in both precursor and mature tRNA_i^Met than in GCD14 cells (Fig. 8B), suggesting a greater need for the putative chaperone function of Lhp1p in gcd14 versus GCD14 cells. It was shown previously that a mutation in yeast tRNA_CGA^Ser makes its processing completely dependent on Lhp1p, and the processing intermediates appear to be rapidly degraded when Lhp1p is depleted from cells (61). It was proposed that Lhp1p was essential to stabilize the conformation of the mutant tRNA_CGA^Ser to permit correct endonucleolytic cleavage removal of the 3′ trailer. Similarly, the absence of m¹A at position 58 could perturb the structure of the initiator tRNA^Met in gcd14 mutants and create a stronger requirement for binding of Lhp1p to the 3′ trailer to achieve precise endonucleolytic 3′ end processing. When Lhp1p was overexpressed in gcd14-1 cells, it led to increased steady-state levels of both precursor and mature full-length tRNA_i^Met and of an intermediate precursor transcribed from IMT2 and IMT3 at 37°C containing the 3′ extension (data not shown). Through increased binding to the 3′ trailers of these precursors, the overproduced Lhp1p could promote correct endonucleolytic cleavage, as suggested above, accounting for the ability of hcLHP1 to increase expression of mature tRNA_i^Met in gcd14 mutants at elevated temperatures. This type of functional interaction between Gcd14p and Lhp1p in the maturation of tRNA_i^Met is consistent with the lack of strong physical interaction between these two proteins (Fig. 9).

Combining the deletion of LHP1 with the gcd14-1 mutation led unexpectedly to large reductions in the expression of NME1 RNA, RPR1 RNA, and tRNA_CGA^Ser, in addition to tRNA_i^Met. The gcd14-2 lhp1::LEU2 mutant also contained substantially reduced amounts of NME1 RNA. The double mutants exhibited small reductions in levels of tRNA_UAU^Ile and U6 RNA compared to the corresponding single mutants and had wild-type levels of elongator tRNA^Met and 5S rRNA (Fig. 8B). Thus, it appears that Gcd14p and Lhp1p cooperate in the maturation of a subset of RNA polymerase III transcripts. Of the transcripts analyzed thus far, only tRNA_i^Met shows a strong requirement for Gcd14p in cells containing Lhp1p. The removal of Lhp1p reveals additional strong requirements for Gcd14p in the production of mature RPR1 RNA, NME1 RNA, and tRNA_CGA^Ser and a lesser requirement in U6 expression. Similarly to the tRNAs, RPR1 RNA is transcribed as a 369-nucleotide precursor RNA with 84 nucleotides of 5′ leader and up to 30 nucleotides of 3′ trailing sequences (38). An intriguing possibility is that Gcd14p is required for methylation of pre-RPR1 RNA and that the hypomethylated RPR1 RNA present in gcd14-2 mutants would be highly susceptible to degradation in cells lacking Lhp1p, as postulated above for tRNA_i^Met. It remains to be determined whether RPR1 RNA is methylated. Obviously, we cannot rule out the possibility that combining the gcd14 and lhp1::LEU2 mutations in the same strains leads to an unforeseen indirect effect on transcription or maturation of NME1 RNA, RPR1 RNA, or tRNA_CGA^Ser.

The fact that gcd14 mutations alone lead to 30 to 40% reductions in the levels of mature RPR1 RNA (Fig. 8B; Table 4) raises the possibility that a reduction in RNase P activity contributes to the defect in processing of tRNA_i^Met in gcd14 single mutants. Perhaps the absence of m¹A58, with the attendant defects in tertiary structure, renders tRNA_i^Met more susceptible than other tRNAs to reductions in RNase P activity. By the same token, it could be proposed that the dramatic reductions in RPR1 RNA levels observed in the gcd14-1 lhp1::LEU2 double mutant could largely account for the additive effects of these mutations in reducing mature tRNA_i^Met levels. However, it was shown previously that depletion of Rpp1p, a protein subunit of RNase P (12, 58), or a mutation in the POP1-encoded subunit of RNase P (41) leads to accumulation of unprocessed tRNA precursors concomitant with reduced amounts of the mature tRNAs. The fact that both precursors and mature tRNA_i^Met were greatly diminished in the gcd14-1 lhp1::LEU2 double mutant (Fig. 8B) suggests that the unprocessed tRNA precursors, which should accumulate in response to reduced RNase P levels, are rapidly degraded in the absence of both Lhp1p and wild-type Gcd14p.

It will be interesting to identify the complete role of the Gcd10p-Gcd14p complex in modification and processing of different RNA polymerase III transcripts and to investigate further functional molecular interactions between this complex and Lhp1p. We recently found that deletion of CPD1, the GCD14 homologue in fission yeast, is not lethal but leads to accumulation of tRNA_i^Met precursors (11a). Thus, it appears that the function of Gcd14p in tRNA_i^Met maturation is evolutionarily conserved.