The Saccharomyces cerevisiae HCR1 Gene Encoding a Homologue of the p35 Subunit of Human Translation Initiation Factor 3 (eIF3) Is a High Copy Suppressor of a Temperature-sensitive Mutation in the Rpg1p Subunit of Yeast eIF3*

The complex eukaryotic initiation factor 3 (eIF3) was shown to promote the formation of the 43 S preinitiation complex by dissociating 40 S and 60 S ribosomal subunits, stabilizing the ternary complex, and aiding mRNA binding to 40 S ribosomal subunits. Recently, we described the identification of RPG1(TIF32), the p110 subunit of the eIF3 core complex in yeast. In a screen for Saccharomyces cerevisiae multicopy suppressors of the rpg1-1 temperature-sensitive mutant, an unknown gene corresponding to the open reading frameYLR192C was identified. When overexpressed, the 30-kDa gene product, named Hcr1p, was able to support, under restrictive conditions, growth of the rpg1-1temperature-sensitive mutant, but not of a Rpg1p-depleted mutant. Anhcr1 null mutant was viable, but showed slight reduction of growth when compared with the wild-type strain. Physical interaction between the Hcr1 and Rpg1 proteins was shown by co-immunoprecipitation analysis. The combination of Δhcr1 andrpg1-1 mutations resulted in a synthetic enhancement of the slow growth phenotype at a semipermissive temperature. In a computer search, a significant homology to the human p35 subunit of the eIF3 complex was found. We assume that the yeast Hcr1 protein participates in translation initiation likely as a protein associated with the eIF3 complex.

Eukaryotes initiate protein synthesis by dissociation of an 80 S ribosome into 40 S and 60 S subunits, formation of the Met-tRNA i ⅐eIF2 1 ⅐GTP ternary complex and binding of this complex to the 40 S subunit, joining the resulting 43 S preinitiation complex with mRNA, and finally, formation of the 48 S initiation complex at the initiation codon, followed by addition of the 60 S ribosomal subunit. Most of these processes are aided by eIF3, the most complex of all the initiation factors described so far (reviewed in Ref. 1). This factor is believed to prevent premature association of 40 S and 60 S ribosomal subunits, to stabilize the ternary complex bound to the 40 S ribosomal subunit, and to promote the binding of the 43 S preinitiation complex to mRNA (2)(3)(4)(5)(6)(7). In the yeast Saccharomyces cerevisiae, eIF3 is thought to form a core complex composed of five subunits, Rpg1p (Tif32p), Nip1p, Prt1p, Tif34p, and Tif35p (8 -15). All of them have homologues in the mammalian eIF3 complex. This core complex was found to be associated with additional proteins, among them Sui1p and Gcd10p (16,17). However, the latter was shown to be involved in specific initiator methionyl-tRNA modification during its maturation, and its role as an eIF3-associated protein (if any) is unclear (18). Interestingly, Hinnebusch and co-workers (9,19,20) recently reported that yeast eIF3 binds the joining factor eIF5; this suggests a novel function for eIF3 in start codon recognition, GTP hydrolysis by eIF2, and 60 S ribosomal subunit joining. A similar model has recently also been postulated for mammals (21).
Genetic approaches in yeast resulted in cloning of several genes coding for translation initiation factors or their subunits (eIF4B, eIF2, eIF2B, and eIF3) (for reviews, see Refs. 1, 22, and 23). For instance, a multicopy suppressor screen done with a temperature-sensitive allele of the known eIF3 subunit Tif34p resulted in identification of another eIF3 subunit, eIF3-p33 (Tif35p) (13,14). Another example is the STM1/TIF3 gene, the protein product of which was identified as a multicopy suppressor of a temperature-sensitive mutation in eIF4A and shown to represent the yeast homologue of mammalian eIF4B (24). We recently reported the characterization of the temperature-sensitive (ts) mutant of the large subunit of yeast eIF3, Rpg1p (Tif32p) (10). When rpg1-1 ts mutant cells are incubated at a restrictive temperature (37°C) or the conditional mutant cells are depleted of Rpg1p, translation ceases, and cells accumulate in the G 1 phase of the cell cycle and do not respond to ␣-pheromone (10,11). We have shown that the essential S. cerevisiae Rpg1 protein is the homologue of the mammalian eIF3-p170 protein (5) required for translation initiation in vivo and in vitro, interacts with Prt1p, and represents the largest component of the eIF3 core complex. Involvement of Rpg1p in the process of translation initiation was clearly demonstrated in a cell-free system dependent on exogenous eIF3 (25): we showed that the sucrose gradient fraction containing Rpg1p possesses the biochemical activity ascribed to eIF3, e.g. the restoration of translation in an extract in which an endogenous eIF3 subunit had been inactivated. The mammalian homologue of Rpg1p, p170, was proved to interact directly with the human homologue of Prt1p (26), which is in a good agreement with our result. In addition, it has been also proposed to bind eIF4B (7) and to be necessary for binding and stabilization of the ternary complex (27). We have demonstrated that this protein is not able to replace yeast Rpg1p in the functional rescue experiment. Up to now, none of the mammalian eIF3 subunits has worked in a similar type of experiment (26,28), even though the yeast eIF3 factor functions in a mammalian methionylpuromycin assay system (15). In addition to Prt1p, yeast Rpg1p has recently been shown to interact with Nip1p (9). Interestingly, Nip1p also interacts with Sui1p and eIF5 (19), both of which are involved in the start codon recognition process previously not considered to be influenced by eIF3. It is clear that further experiments are needed to define the complex spectrum of eIF3 functions. Thus, biochemical strategies and genetic studies carried out with the eIF3 core complex subunits can help to reveal important interactions between either eIF3 subunits themselves and/or with other components of the translational apparatus and subsequently to better understand the function of eIF3 in translation initiation.
To learn more about the large subunit of yeast eIF3 and its physiological interactions, a multicopy suppressor study was carried out to identify proteins that can functionally complement the temperature-sensitive phenotype of the rpg1-1 ts allele. Here, we present biochemical and genetic data demonstrating that the nonessential S. cerevisiae HCR1 gene (high copy suppressor of RPG1) codes for a protein that appears to be the yeast homologue of the p35 subunit of mammalian eIF3 and that, in high dosage, partially suppresses the growth defect of the rpg1-1 ts mutant caused by incubation at the restrictive temperature. Moreover, both Hcr1 and Rpg1 proteins were shown to interact with each other, which in light of the previously described interaction between their mammalian homologues, eIF3-p35 and eIF3-p170 (29), suggests a possible involvement of Hcr1p in translation initiation.
YCpLVHCR1-cMyc is a CEN4,LEU2 plasmid that contains the Myctagged HCR1 gene under its endogenous promoter. This plasmid was constructed in two steps. First, an HCR1 fragment generated by polymerase chain reaction using the original library clone YEpLVHCR2.18 as a template was digested using SalI-KpnI polymerase chain reactionincorporated artificial sites (primers HCR1-SalI (5Ј-CTGCAGGTC-GACTCTAGTGCA-3Ј) and HCR1-KpnI (5Ј-CTATAGGTACCTTAGCG-GCCGCCCATAAAGTCGTCATCACCAGG-3Ј)) and ligated into SalI-KpnI of YEpLVHCR1, producing YEpLVHCR1-N, which also contained a NotI restriction site inserted by polymerase chain reaction immediately in front of the stop codon of HCR1. The second step of plasmid construction consisted of an in-frame insertion of a NotI-digested Myc tag into the NotI site of YEpLVHCR1-N. The resulting YEpLVHCR1-cMyc was used for transformation of the haploid yeast strain W303 to generate strain YLVH10-Me expressing the Myc-tagged Hcr1 protein in high copy number in order to test the function of the Myc-tagged HCR1 gene. Strain YLVH10-Mc was produced by introduction of YC-pLVHCR1-cMyc, a construct generated by insertion of the 1.5-kb SalI-KpnI fragment carrying the Myc-tagged HCR1 gene into SalI-KpnI-cut YCplac111 (34).
Disruption of the HCR1 Gene-A gene disruption was carried out in a diploid S. cerevisiae strain using the one-step transplacement method described previously (35). For the disruption of HCR1, the plasmid YEpHCR1-D was constructed as follows. A 1.2-kb NdeI-NheI fragment representing the entire coding sequence and part of the promoter region of HCR1 was replaced by a 2.2-kb NdeI-XbaI fragment of pJJ250 (36) containing LEU2 in the original library clone YEpLVHCR2.18. The SpeI-SpeI fragment of the former plasmid containing the LEU2 gene was isolated and transformed into the diploid strain W303 to produce strain YLVH13 d (see Fig. 1A). The resulting strain was then sporulated, and tetrads were dissected. Integration into the HCR1 locus and disruption of only one HCR1 copy were confirmed by Southern analysis.
Western Blot Analysis and Co-immunoprecipitations-Crude cell extracts or immunoprecipitates were fractionated by SDS-polyacrylamide gel electrophoresis (37) using the Mini-Protean system (Bio-Rad) and immunoblotted onto nitrocellulose (38). Immunodetection of proteins was carried out using monoclonal antibody directed against Rpg1p (10) and anti-Myc hybridoma supernatant antibody. As a secondary antibody, anti-mouse IgG antibodies conjugated with alkaline phosphatase or with horseradish peroxidase, respectively, were used. Proteins were visualized either using alkaline phosphatase substrates (nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate) in 100 mM Tris-HCl, 100 mM NaCl, and 5 mM MgCl 2 (pH 9.5) or as described in the ECL manual (Amersham Pharmacia Biotech). Preparation of crude cell extracts as well as immunoprecipitates used was described previously (10).

RESULTS
Isolation of Multicopy Suppressors of the rpg1-1 Temperature-sensitive Mutant-To isolate proteins that interact with the eIF3 complex, we set out to identify multicopy suppressors of the rpg1-1 ts allele. A yeast genomic library (kindly provided by Alexander Schleiffer) constructed in the high copy number vector YEplac181 (34) was introduced into the recipient strain YLV314U (11), carrying the chromosomally integrated rpg1-1 ts allele. After overnight incubation of the plates at a permissive temperature (25°C), the transformants were incubated for up to 7 days at a restrictive temperature (37°C). A total number of 34 transformants growing at this temperature were identified, the rescued plasmids of which were able to suppress the growth at 37°C of the rpg1-1 ts mutant after back-transformation into the YLV314U strain. Out of these plasmids, 19 contained the wild-type RPG1 gene, indicating that our screen was saturated. Restriction analysis of the remaining 15 suppressor plasmids revealed five different clones, all containing a 1.52-kb SpeI fragment. The insert of one of the plasmids from each clone was sequenced, and the sequence was compared with sequences in the yeast genome data base. All suppressor clones contained DNA from chromosome XII corresponding to the hypothetical open reading frame region YLR192C, which we then named HCR1 (high copy suppressor of RPG1) (Fig.  1A). Further subcloning into the plasmid YEplac181 confirmed the minimal region required for suppression to be a 1.2-kb SpeI-StuI DNA fragment exclusively containing the open reading frame of YLR192C/HCR1 (strain YLVH11) (Figs. 1A and 2). To show that suppression was due to the presence of multiple copies of the HCR1 gene, the 1.2-kb SpeI-StuI DNA fragment was inserted into the centromeric plasmid YCplac111 (34). The presence of this plasmid in the mutant strain YLV314U did not suppress the mutant phenotype (data not shown). Moreover, expression of the rpg1-1 ts allele on a multicopy plasmid did not lead to suppression of growth arrest of the rpg1-1 ts mutant strain at the restrictive temperature. 2 Thus, it is highly unlikely that overexpression of Hcr1p suppressed the rpg1-1 ts mutation by an effect on the expression of the mutant rpg1-1 ts gene. This was documented by Northern blot analysis carried out with the YLVH11, YLV314U, and W303 strains grown overnight at 25°C and subsequently shifted to 37°C for 12 h, where no significant alteration in RPG1 transcript level was observed (data not shown). These findings strongly suggest that Hcr1p is indeed a specific multicopy suppressor of the rpg1-1 ts mutant.
Hcr1p May Be a Homologue of the Human p35 Subunit of eIF3-As mentioned above, the DNA fragment that suppresses the rpg1-1 ts mutation is 1239 nucleotides long and contains the HCR1 open reading frame. The predicted ATG codon of HCR1 is preceded by an in-frame stop codon at position Ϫ48. The open reading frame encodes a protein containing 265 amino acids with a molecular mass of 29.56 kDa and a pI of 5.03. This ATG codon is preceded by four adenines and thus should represent an efficient translation start site (39). The predicted stop codon is followed by additional in-frame stop codons at positions 847 and 913. The accession number for the DNA sequence in the NCBI Database is U14913. When the amino acid sequence deduced for Hcr1p was run against the PROSITE data base (40), sites for protein phosphorylation by protein kinase C and casein kinase II as well as N-glycosylation and N-myristoylation modification sites were identified. When the amino acid sequence was compared with sequences in the protein data bases, two interesting matches showed up. The first is the p35 subunit of human translation initiation factor 3 recently characterized by Block et al. (29), displaying ϳ26% identity and 42% similarity. These authors already noted this amino acid sequence homology. The other is a protein of unknown function from Arabidopsis thaliana. Based on this finding, we postulate that Hcr1p represents the S. cerevisiae homologue of the p35 subunit of human eIF3 (Fig. 1B).
Suppression of the rpg1-1 ts Mutation by HCR1-Overexpression of Hcr1p leads to suppression of the temperaturesensitive phenotype of the rpg1-1 ts mutant (Fig. 2). However, the suppression seems not to be complete since the doubling time of the suppressed rpg1-1 ts cells was estimated to range from 8 to 10 h, compared with 1.5 h for the wild-type cells (Fig.  3). We have previously reported that the majority of the rpg1-1 ts mutant cells accumulate in the G 1 phase of the cell cycle as unbudded cell doublets upon shift to the restrictive temperature, displaying a marked cell separation defect (11). As judged from morphological and flow cytometric analyses, the rpg1-1 ts mutant cells carrying the high copy gene are delayed in G 1 , having similar morphological characteristics as the rpg1-1 ts cells themselves (data not shown). Whereas the rpg1-1 ts mutant cells contain ϳ15% budded cells (11), we have observed Ͼ60% budded cells in the HCR1-suppressed population. This suggests that overexpression of Hcr1p somehow helps the mutant cells to overcome the G 1 arrest and to reinitiate budding. No effect was observed when HCR1 was overexpressed in the wild-type W303 strain (strain YLVH10). YLVH10 cells proliferated with the same generation time as W303 cells, and no morphological abnormalities were seen.
To test the allele specificity of the suppressor effect, we subcloned the HCR1 gene into the plasmid YEplac192 (34) and transformed the resulting construct into the Rpg1p depletion strain YLV041 containing a methionine-suppressible copy of RPG1 (10), producing strain YLVH12. The cells of this strain were then tested for growth on medium containing methionine, which shuts off transcription of the RPG1 gene from the MET3 promoter. Since no suppression was observed (data not shown), we conclude that high dosage HCR1-driven complementation is rpg1-1 ts allele-specific. This also demonstrates that Hcr1p is not able to functionally replace Rpg1p. In addition, overexpressed Hcr1p is unable to suppress the ts growth defect of prt1-1, another translation initiation mutant tested (data not 2 L.Valášek, unpublished observation.  FIG. 2. Suppression of the rpg1-1 ts phenotype by overexpression of Hcr1p. The YLV314U strain carrying the rpg1-1 ts mutation was transformed with HCR1 on a multicopy plasmid (strain YLVH11) and tested for growth suppression at 37°C. As controls, the original temperature-sensitive YLV314U and wild-type W303 strains are shown. Plates were incubated for 3 days.
shown). The rpg1-1 ts allele specificity observed may indicate that Hcr1p directly interacts with Rpg1p.
Hcr1p Interacts with Rpg1p in Vitro-The physical interaction between Tif34p and its multicopy suppressor eIF3-p33 was suggested to play an important role in the functional rescue observed in this case (13). In addition, the recently identified homologue of Hcr1p, the p35 subunit of human eIF3, has been shown to interact with another eIF3 subunit, p170, actually representing the mammalian homologue of yeast Rpg1p (29). To test whether the rpg1-1 ts multicopy suppressor Hcr1p could interact with Rpg1p, we constructed a single copy expression plasmid for the c-Myc-tagged version of Hcr1p and introduced it into the wild-type W303 strain, generating strain YLVH10-Mc. The fusion protein exhibiting an apparent molecular mass of ϳ55 kDa was expressed in the YLVH10-Mc strain (Fig. 4A, lane 1). Three additional bands migrating with a smaller size likely represent proteolytic breakdown products. The tagged protein was shown to complement the ts defect of the rpg1-1 ts mutant, demonstrating that it is functional (data not shown). Crude extracts from strains expressing a tagged or an untagged version of Hcr1p were prepared and used for immunoprecipitation experiments with antibodies directed against Rpg1p or Myc-tagged Hcr1p. The resulting immune complexes were separated by SDS-polyacrylamide gel electrophoresis and probed by immunoblot analysis using either anti-Rpg1p or anti-Myc antibodies. As demonstrated in Fig. 4 (B,  lane 5), immunoprecipitation of Myc-tagged Hcr1p resulted in coprecipitation of Rpg1p. Conversely, immunoprecipitation of Rpg1p using anti-Rpg1p monoclonal antibody coprecipitated Hcr1p (Fig. 4A, lane 3). The 50-kDa bands seen in Fig. 4A correspond to the IgG heavy chain. In control experiments, anti-Myc antibody immunoprecipitated neither HCR1 nor RPG1 gene products of the untagged Hcr1p strain (Fig. 4, A  and B, lanes 6). Similarly, in an immunoprecipitate obtained with anti-Rpg1p antibody, untagged Hcr1p was not detectable by anti-Myc antiserum (Fig. 4A, lane 4). Thus, these data indicate that both proteins interact directly and/or reside on the same multiprotein complex.
HCR1 Is Not Essential for Cell Proliferation-To determine whether the HCR1 gene is essential for cell viability, one copy of the entire coding sequence in a diploid wild-type strain (W303) was replaced by the LEU2 gene ( Fig. 1A; see "Experimental Procedures"). Correct integration was verified by Southern blot analysis (data not shown). After sporulation and tetrad dissection, all four spores in each tetrad were able to form a colony, and the LEU2 gene segregated 2 ϩ :2 Ϫ , indicating that the HCR1 gene is not essential for cell proliferation (data not shown). The hcr1::LEU2 null mutant strain YLVH13 showed neither heat nor cold sensitivity when compared with an isogenic wild-type strain. The generation time of the null mutant grown in YP and 2% glucose calculated from absorbance measurements of exponentially growing cultures was found to be somewhat longer (ϳ2.0 h) at both 30 and 37°C compared with the wild-type strain generation time of ϳ1.5 h (data not shown). The mating efficiency was estimated to be ϳ10-fold lower than that of the wild-type strain (data not shown), indicating that mating ability is reduced in the mutant strain. These results suggest a defect, perhaps on the transla-  A and B,  lanes 3 and 4) or anti-Myc antibody (A and B, lanes 5 and 6). Immune complexes were then separated by 12.5% SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and probed with monoclonal antibody against the Myc tag (A) or against Rpg1p (B). The crude extracts containing tagged or untagged Hcr1p are shown in A and B (lanes 1 and 2, respectively). Note that Hcr1p is predicted to be a protein with a molecular mass of ϳ30 kDa, but that the fusion with the Myc tag and perhaps some post-translational modifications resulted in a larger size. tional level, in the hcr1-deleted cells; however, we were not able to prove this by analyzing polysome profiles of the hcr1 null mutant strain (data not shown).
To further investigate the functional relation between both Hcr1 and Rpg1 proteins, we investigated a possible synthetic interaction of the hcr1 null mutant with the rpg1-1 ts mutation. The idea of synthetic lethality is based on the observation that certain double mutants are inviable under conditions in which the parental single mutants are viable (41), which suggests a biochemical function in the same pathway. Several different combinations of translation factor mutations were found to be lethal or to show a synthetic enhancement (42). Cells of YLVH13 (hcr1) and YLV314U (rpg1-1) strains were crossed, and the resulting diploid strain, YLVH14 d , was sporulated. After dissection of tetrads and outgrowth of the spores, the resulting strains were evaluated for amino acid auxotrophy and for temperature sensitivity. This revealed that the combination of the hcr1 null allele with the rpg1-1 ts mutant allele did not result in an inviable double mutant (data not shown). However, comparison of the growth characteristics at the semipermissive temperature (33°C) of the double mutant YLVH14 and the rpg1-1 ts mutant itself (strain YLV314U) showed a significant decrease in the growth rate of the double mutant YLVH14 (Fig. 5). This could be caused by a synthetic enhancement of the rpg1-1 ts phenotype defective in translation initi-ation at 33°C or, alternatively, may be caused by an additive effect of the hcr1 deletion on protein synthesis already affected by the rpg1-1 ts mutation. Since a polysome profile of the rpg1-1 ts mutant grown at 33°C itself shows a severe reduction in the polysome/monosome ratio and remarkable accumulation of the 80 S ribosomes, we were not able to detect further changes when looking at the polysome profile of the double mutant YLVH14. Nevertheless, the data presented here indicate the existence of a genetic interaction between these two proteins.

DISCUSSION
In a high dosage suppressor screen of the rpg1-1 ts mutation, a novel nonessential S. cerevisiae gene named HCR1 was isolated. The screen appeared to be saturated and highly specific as illustrated by the fact that, besides the RPG1 gene (19 clones), 15 additional clones were isolated, all of which contained the HCR1 gene. In high dosage, the Hcr1 protein suppresses the rpg1-1 ts mutant, but does not rescue the rpg1 null mutant (a Rpg1p-depleted strain); hence, it does not bypass the requirement of Rpg1p for growth. The deletion of the HCR1 gene resulted in a slight reduction of the growth rate and a significant reduction of the mating ability of the corresponding strain. In an immunoprecipitation experiment, Hcr1p was shown to associate with Rpg1p. Furthermore, we demonstrate that the deletion of HCR1 in combination with the rpg1-1 ts mutation leads to a synthetic enhancement of the slow growth phenotype of the rpg1-1 ts mutant at the semipermissive temperature. These data indicate that the two proteins interact both physically and genetically. Taken together with the facts that Hcr1p shares ϳ27% identity and 42% similarity with the p35 subunit of human eIF3 and that the latter protein was shown to associate with eIF3-p170 (29), i.e. the mammalian homologue of yeast Rpg1p (5), we suggest that Hcr1p represents the yeast homologue of eIF3-p35.
At present, we can only speculate about the mechanism by which Hcr1p suppresses the rpg1-1 ts mutation. Enhanced levels of the Hcr1 protein may stabilize just Rpg1p or the entire multisubunit eIF3 complex at the restrictive temperature by specific protein-protein interaction or may stimulate one or several of the eIF3 activities in the cell, i.e. binding of the ternary complex Met-tRNA i ⅐eIF2⅐GTP or mRNA to the 40 S ribosome. This would be somewhat reminiscent of Tif35p, another subunit of yeast eIF3. This protein was identified as a multicopy suppressor of a tif34 ts mutation and shown to directly interact with Tif34p (13). Tif34p plays a central role in the assembly of the subunits of eIF3 into the multiprotein complex (28). To address, at least partly, the question of whether Hcr1p works as a factor involved in assembly and/or maintenance of eIF3, we also tested whether it acts as a multicopy suppressor of the prt1-1 ts mutation. This is not the case. This might mean that the suppressor effect is Rpg1p-specific, i.e. suppression occurs at the level of the translation initiation block, if applicable, caused by the rpg1-1 ts mutation.
An alternative possibility is that a high dosage of Hcr1p stabilizes RPG1 mRNA. In this case, the elevated level of the rpg1-1 ts protein would result in sufficient Rpg1p activity to allow slow growth of cells. However, Northern blot analysis and the fact that overexpression of the rpg1-1 ts protein in the rpg1-1 ts mutant does not rescue growth clearly contradict this possibility.
When we analyzed the suppression of the rpg1-1 ts mutation by overexpression of Hcr1p by measuring the growth rates of the corresponding strains, we noticed that it is rather weak (Fig. 3). We do not know yet the precise function of Rpg1p and hence the biochemical defect conferred by the rpg1-1 ts mutation. However, since Rpg1p is involved in translation initiation, we assume that Hcr1p overexpression partially corrects the translation initiation defect. Rpg1p, the largest subunit of the eIF3 core complex, may have several functions, and the rpg1-1 ts mutation may impair more than one of them. The weak suppression of rpg1-1 by HCR1 might then be explained by suppression of only some of the affected functions of the rpg1-1 protein by Hcr1p. This would be sufficient to overcome the G 1 arrest and to reinitiate budding (likely by enhancing the rate of the otherwise severely reduced protein synthesis).
If Hcr1p is involved in translation initiation as suggested by the suppression of the ts mutation in the RPG1 (TIF32) gene and from its homology to the human eIF3-p35 subunit, one might have expected that the deletion of the HCR1 gene would be lethal. However, this is not the case. Other nonessential genes encoding translation initiation factors are STM1, encoding the yeast homologue of mammalian eIF4B, which was identified as a multicopy suppressor of the eIF4A ts mutation (24); and GCN3, encoding a subunit of the nucleotide exchange factor eIF2B (43). Gcn3p is a regulatory subunit of eIF2B and plays an important role in up-regulation of translation of GCN4 mRNA under conditions of amino acid starvation (44). In analogy to Gcn3p, it is tempting to speculate that Hcr1p might be a regulator of eIF3 activity, acting by binding to Rpg1p under special physiological conditions, but being dispensable for eIF3 activities under normal growth conditions. In accordance with this possibility, Hcr1p was found not to belong to the firmly associated eIF3 complex referred to as the eIF3 core complex (8,9). Together with other proteins, among them Sui1p and maybe Gcd10p, Hcr1p may belong to a group of eIF3-associated proteins whose functions await elucidation. Moreover, when Hershey and co-workers (29) analyzed human eIF3-p35, they found it to be a very "sticky" protein, interacting with nearly all subunits of human eIF3, especially with p66 and the p170 homologue of yeast Rpg1p. Further experiments are necessary to find the function for Hcr1p in the translation initiation machinery and to explain the functional connection with Rpg1p (Tif32p). This may lead to a better understanding of the function of the p35 subunit of the human eIF3 complex.