The Human Homologue of the Yeast Prt1 Protein Is an Integral Part of the Eukaryotic Initiation Factor 3 Complex and Interacts with p170*

(Received for publication, September 16, 1996)

Nathalie Méthot Dagger §, Eran Rom Dagger , Henrik Olsen par and Nahum Sonenberg Dagger **

From the Dagger  Department of Biochemistry and McGill Cancer Centre, McGill University, Montréal, Québec, Canada H3G 1Y6 and par  Human Genome Science Inc., Rockville, Maryland 20850

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

Eukaryotic initiation factor 3 (eIF3) is a large multisubunit complex that stabilizes the ternary complex, eIF2·GTP·tRNAiMet, and promotes mRNA binding to the 40 S ribosomal subunit. eIF3 also functions as a ribosome subunit anti-association factor. The molecular mechanisms by which eIF3 exerts these functions are poorly understood. We describe here the cloning of the cDNA encoding the human homologue of the yeast eIF3 subunit Prt1. The human PRT1 cDNA encodes a protein of predicted molecular mass of 98.9 kDa that migrates at 116 kDa on SDS-polyacrylamide gels. Human and yeast Prt1 share 31% identity and 50% similarity at the amino acid level. The homology is distributed throughout the entire protein, except for the amino terminus, and is particularly high in the central portion of the protein, which contains a putative RNA recognition motif. hPrt1 is recognized by an antibody raised against eIF3, and an affinity-purified antibody to recombinant hPrt1 recognizes a protein migrating at 116 kDa in a purified eIF3 preparation. Far Western analysis shows that hPrt1 interacts directly with the p170 subunit of eIF3. Mapping studies identify the RNA recognition motif as the region required for association with p170. Taken together, these experiments demonstrate that hPrt1 is a component of eIF3. Our data, combined with those of Hershey and co-workers, suggest that mammalian eIF3 is composed of at least 10 subunits: p170, p116 (hPrt1), p110, p66, p48, p47, p44, p40, p36, and p35.


INTRODUCTION

Eukaryotic protein synthesis requires the participation of translation initiation factors, which assist in the binding of the mRNA to the 40 S ribosomal subunit (reviewed in Refs. 1 and 2). Ribosome binding is facilitated by the cap structure (m7GpppN, where N is any nucleotide) that is present at the 5' end of all cellular mRNAs (except organellar). Biochemical fractionation studies elucidated the general pathway for translation initiation and led to the characterization of several translation initiation factors (Refs. 3, 4, 5; reviewed in Ref. 1). It is believed that the mRNA cap structure is initially bound by eukaryotic initiation factor (eIF)1 4F, which, in conjunction with eIF4B, melts RNA secondary structure in the 5'-untranslated region of the mRNA to promote ribosome binding. The 40 S ribosomal subunit, in a complex with eIF3, eIF1A, and eIF2·GTP·tRNAiMet, binds at or near the cap structure and scans vectorially the 5'-untranslated region in search of the initiator AUG codon (reviewed in Refs. 1 and 2).

Eukaryotic initiation factor 3 (eIF3) is the largest translation initiation factor, with at least 8 different polypeptide subunits and a total mass of approximately 550-700 kDa (6, 7, 8). In mammals, the apparent molecular masses of the eIF3 subunits are 35, 36, 40, 44, 47, 66, 115, and 170 kDa (Refs. 8, 9, 10; see also "Discussion" for new additional subunits). eIF3 is a moderately abundant translation initiation factor, with 0.5-1 molecules/ribosome in HeLa cells and rabbit reticulocyte lysates (9, 11). eIF3 serves several functions during translation initiation (reviewed in Ref. 12). eIF3 binds to the 40 S ribosomal subunit and prevents joining of the 60 S subunit. It also interacts with the ternary complex and stabilizes the binding of the latter to the 40 S ribosomal subunit (5, 13, 14, 15). eIF3 cross-links to mRNA and 18 S rRNA (16, 17), an activity mainly attributed to the 66-kDa subunit (or 62-kDa GCD10 in yeast; Refs. 18 and 19). eIF3 co-purifies with eIF4F and eIF4B, two initiation factors involved in the mRNA binding step (6). A direct interaction between the eIF4G subunit of eIF4F and eIF3 has been demonstrated (20), and a role for eIF3 serving as a bridge between the 40 S ribosomal subunit and eIF4F-bound mRNA has been postulated (20).

The complex structure of eIF3 and its pleiotropic roles in translation initiation have rendered the study of this factor difficult. The protein sequence for only three of the yeast subunits (SUI1/p16, GCD10/p62, and PRT1/p90) has been published (18, 21, 22). However, several additional mammalian and yeast subunits have been cloned recently.23 The yeast protein p90, also known as Prt1, is the best characterized. Prt1 is an integral subunit of eIF3 (19, 23). A conditional lethal mutation in the PRT1 gene reduces the binding of the ternary complex to the 40 S ribosomal subunit (24). Other mutations, which confer temperature sensitivity, are located in the central and carboxyl-terminal portion of Prt1. An NH2-terminal deletion, which removes the Prt1 putative RNA recognition motif (RRM; for reviews, see Refs. 25, 26, 27), acts as a trans-dominant negative inhibitor (28).

To further understand the structure and function of mammalian eIF3, we have cloned the putative human homologue of the yeast eIF3 subunit Prt1. The human Prt1 cDNA (hPrt1) is predicted to encode a protein of 873 amino acids, and shares 31% identity and 50% similarity with the yeast Prt1. Immunological characterization show that hPrt1 is part of the eIF3 complex and co-migrates with the 115-kDa polypeptide of eIF3. Far Western analysis demonstrates that hPrt1 is incorporated into the eIF3 complex by virtue of a direct interaction with the largest subunit of eIF3, p170. The p170 interaction site is located in a conserved region of hPrt1, which contains part of the RRM.


EXPERIMENTAL PROCEDURES

Materials

Materials were obtained from the following sources. Restriction enzymes were from New England Biolabs. T7 RNA polymerase, RNasin, and rabbit reticulocyte lysate were from Promega. T7 DNA polymerase sequencing kit was from Pharmacia Biotech Inc. Protein A-Sepharose was from Repligen. Heart muscle kinase was from Sigma. Hybond-N+ nylon membrane, chemiluminescence system was from Amersham Corp. Nitrocellulose membrane was from Schleicher & Schuell. Polyvinylidene difluoride membrane was from Millipore. [gamma -32P]ATP (6000 Ci/mmol), [alpha -32P]dCTP (3000 Ci/mmol), and [35S]methionine (1000 Ci/mmol) were from DuPont NEN. Oligonucleotides were prepared at the Sheldon Biotechnology Center, McGill University. All other reagents were at least reagent grade.

Isolation of hPrt1 cDNA Clones

The expressed sequence tag used in this study was identified using established expressed sequence tag (EST) methods described previously (29), and this partial cDNA clone encoding the human homologue of the yeast Prt1 protein was used to obtain the full-length cDNA clone. Full-length cDNA clones for hPrt1 were isolated from a lambda gt11 human placenta cDNA library (generous gift from Morag Park, McGill University). A 250-base pair DNA was generated by the polymerase chain reaction (PCR) using the hPrt1 EST clone as template and the following primers: 5'-TTTACGGTTTTGACAAGTA-3' and 5'-TTAGGAGACCAACG-3'. The amplified DNA was 32P-labeled by random priming using [alpha -32P]dCTP, random hexamers, and the Klenow fragment of DNA polymerase (30), and used as a probe in cDNA screening and Northern blot analysis. For cDNA screening, 5 × 105 phages displayed on duplicate sets of filters (Hybond-N+, Amersham) were prehybridized in 5 × SSPE (20 × SSPE is 3.6 M NaCl, 0.2 M Na3PO4, 0.02 M EDTA, pH 7.7), 5 × Denhardt's solution (1 × Denhardt's is 0.1% bovine serum albumin, 0.1% Ficoll, 0.1% polyvinylpyrrolidone), 0.5% SDS, and 40 µg/ml heat-denatured salmon sperm DNA, for 4 h at 65 °C. Hybridization was performed in the same buffer containing the hPrt1 probe at 1 × 106 cpm/ml for 16 h at 65 °C. Filters were washed to a final stringency of 0.1 × SSPE, 0.1% SDS at 65 °C, and exposed to Kodak XAR films for 72 h with intensifying screens. Phages from positive clones were used to prepare plate lysates, and DNA was purified, digested with SalI, and ligated into pBluescript that had been digested with SalI. Oligonucleotides used for sequencing were derived from either pBluescript or from the hPrt1 EST DNA sequence. The nucleotide sequence for full-length hPrt1-1 was obtained from both strands of independent overlapping clones using the dideoxy chain termination method (31) and the T7 polymerase sequencing kit (Pharmacia). Regions of compression were re-sequenced using 7-deaza-dGTP.

Vectors, Proteins

The full-length cDNA (clone 3-6) was excised from the lambda gt11 phage by SalI digestion and inserted into pBluescript KS in the T7 promoter orientation. The resulting vector is designated as KST7hPrt1-6. Constructs for truncated hPrt1 proteins were generated by PCR using primers in which an EcoRI site had been engineered. Cleavage of the PCR product with EcoRI and ligation into pARDelta [59/69] (32) or pGEX2T[128/129] (32) that had been digested with EcoRI preserves the hPrt1 open reading frame and creates a GST-FLAG-HMK or FLAG-HMK fusion protein. For pARDelta NDelta 146, the forward primer was 5'-ACCGGAATTCAAAATGGACGCGGACGAGCCCTC-3' and the reverse primer 5'-AGCGGAATTCTTAAATCCCCCACTGCAG-3'. For pGEX N255, the hPrt1 open reading frame was first amplified by PCR and inserted into pGEX2T[128/129]. The resulting vector was linearized with HindIII, blunt-ended with the Klenow fragment of Escherichia coli DNA polymerase, and religated. Religation creates a stop codon 3 amino acids downstream of hPrt1 residue 255. pGEX 146-255 was obtained by linearizing pGEX NDelta 146 with HindIII, blunt-ending with Klenow, and religating. Vectors were transformed in either E. coli BL21 or BL21 pLysS. Bacteria were grown in LB broth to an optical density of 0.5, and protein expression was induced with 1 mM isopropyl-1-thio-beta -D-galactopyranoside for 1 h at 37 °C. Cells were pelleted and lysed in lysis buffer (phosphate-buffered saline, 1 mM EDTA, 1 mM DTT, 0.1 mM phenylmethylsulfonyl fluoride) by 6 sonication cycles. Debris was removed by centrifugation. GST fusion proteins were purified on glutathione-Sepharose (Pharmacia) as described previously (33). FLAG-HMK fusion proteins were affinity-purified over an anti-FLAG column (Kodak) according to the manufacturer's specifications. pACTAG-hPrt1 was made by linearizing pACTAG-2 (34) with NotI, and inserting the hPrt1 open reading frame that had been excised from KST7hPrt1 cut with NotI. hPrt1 is expressed from this vector as a fusion protein bearing three hemagglutinin (HA) tags at its amino terminus.

In Vitro Transcription and Translation

KST7hPrt1-6 was digested with DraI, and the linearized plasmid was used as template for in vitro transcription using T7 RNA polymerase (Promega) under conditions recommended by the supplier. Translation reactions were performed in nuclease-treated rabbit reticulocyte lysate (Promega) in a final volume of 15 µl. Reaction mixtures contained 10 µl of lysate, 10 µCi of [35S]methionine (1000 Ci/mmol), 15 units of RNasin (Promega), 20 µM amino acid mixture (minus methionine), and 100 ng of RNA. The reactions were incubated 60 min at 30 °C and stopped by the addition of three volumes of Laemmli buffer. Translation products were analyzed by 9% SDS-polyacrylamide gel electrophoresis. Gels were fixed, treated in 16% salicylic acid (Sigma), dried, and processed for autoradiography.

Immunoprecipitations, Western Blots

For HA-hPrt1 protein expression, HeLa cells that had been cultured in DMEM supplemented with 10% fetal bovine serum (FBS) were infected for 1 h with recombinant vaccinia virus vTF7-3 with the T7 RNA polymerase cDNA inserted into its genome (35), and transfected with 5 µg of plasmid DNA using Lipofectin (Life Technologies, Inc.) in DMEM without FBS. Cells were incubated 2 h with the DNA-Lipofectin mixture, and returned to DMEM-10% FBS for 12 h before harvesting. The cells were lysed in 20 mM Tris-HCl, pH 7.4, 75 mM KCl, 5 mM MgCl2, 1 mM DTT, 10% glycerol, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, aprotinin (25 ng/ml), and pepstatin (1 ng/ml). Cellular debris and nuclei were removed by centrifugation, and protein content was assayed by the Bradford method. Immunoprecipitations were performed on 200 µg of extract using alpha -HA antibody. Briefly, extracts were diluted to 500 µl in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate) and incubated on ice for 30 min with 1 µg of antibody. Protein-A-Sepharose (Repligen) was added and allowed to mix at 4 °C for 60 min. The beads were washed five times in RIPA buffer before addition Laemmli buffer and boiling for 5 min. Immunoprecipitates were then loaded on a 10% SDS-polyacrylamide gel, blotted onto nitrocellulose and probed with a goat anti-rabbit eIF3 antibody. Immunoreactive species were visualized using the Renaissance chemiluminescence system (ECL; Amersham). Affinity-purified antibodies against recombinant hPrt1 were obtained using a protocol described by Asano et al. (43). E. coli extracts expressing hPrt1 NDelta 146 were fractionated by SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose. The bands containing NDelta 146 were excised, blocked in Blotto (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.075% Tween 20, 0.5% milk powder), incubated with crude alpha -eIF3, and washed. Antibodies bound to the membrane were eluted with 2 M glycine, 1 mM EGTA, pH 2.5, and neutralized by the addition of 1 M Tris-HCl, pH 8.8. To eliminate contamination with p110 (hNip1), 50 µg of GST-p110 immobilizes on nitrocellulose were present during the incubation with crude alpha -eIF3 antibodies. Western blotting was performed with antibodies at the following dilutions: alpha -eIF3, 1:3000; alpha -p170, 1:10. For Western blots performed with the monoclonal alpha -p170 antibody, horseradish-peroxidase alpha -mouse IgM (Pierce) secondary antibodies were used; for alpha -eIF3, alpha -goat IgG-horseradish peroxidase; for alpha -hPrt1, alpha -goat IgG-alkaline phosphatase.

Northern Blot

Total RNA from HeLa cells was isolated using Trizol (Life Technologies, Inc.) and fractionated by electrophoresis on a 1% agarose/formaldehyde gel overnight at 40 V. RNA was blotted to Hybond N filters overnight and UV cross-linked to the membrane using UV light (Stratalinker). The membrane was prehybridized and hybridized under conditions identical to those used for the cDNA library screening, and exposed for 24 h to a Kodak BioMax film with intensifying screen.

Far Western Blots

Partially purified FLAG-HMK hPrt1 fusion proteins (1-3 µg) were 32P-labeled using heart muscle kinase (Sigma) as described (32). Proteins were resolved by SDS-polyacrylamide gel electrophoresis and blotted on polyvinylidene difluoride (Millipore) or nitrocellulose membranes. The membranes were blocked overnight with 5% milk in HBB buffer (25 mM HEPES-KOH, pH 7.5, 25 mM NaCl, 5 mM MgCl2, 1 mM DTT), and incubated 4 h in hybridization buffer (20 mM HEPES-KOH, pH 7.5, 75 mM KCl, 2.5 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, 0.1% Nonidet P-40, 1% milk) containing the 32P-labeled FLAG-HMK- or GST-FLAG-HMK-hPrt1 at 250,000 cpm/ml and unlabeled purified GST at 1 µg/ml. The membranes were washed three times with hybridization buffer and processed for autoradiography.


RESULTS

Cloning and Features of hPrt1

EST 112738 from Human Genome Science Inc. encodes a predicted protein with homology to the yeast p90 eIF3 subunit, Prt1. The cDNA sequence, 2 kilobase pairs in length, contains a polyadenylation signal and a short polyadenylate tail. An ATG codon is present at the 5' end of the clone. However, this ATG is not preceded by stop codons (Fig. 1A). It was therefore possible that EST 112738 contains an incomplete cDNA. A 32P-labeled probe derived from the 5' end of the EST sequence was generated and used in a Northern analysis on HeLa cell RNA. A single RNA species, migrating at 3.1 kilobases, hybridized with the probe (Fig. 1B). We concluded that 1 kilobase pair was missing from EST 112738. To obtain the full-length cDNA sequence, a human placenta lambda gt11 cDNA library (generous gift from M. Park) was screened with the same probe used in the Northern analysis. Many (40) positive plaques were obtained, and some of the cDNAs extended further upstream of the 5'-most sequence of EST 112738. One of these cDNAs, 3-6, contained a 3-kilobase pair insert with a predicted open reading frame of 873 amino acids encoding a protein with a molecular mass of 98.9 kDa. An ATG codon is located 53 nucleotides downstream of the 5' end but is not preceded by an in-frame stop codon (Fig. 1A). Thus, it is formally possible that clone 3-6 encodes a partial cDNA. However, since the length of the cDNA closely resembles that of the mRNA, and because the translation product from clone 3-6 migrates at the expected molecular weight (see below), the first ATG most probably constitutes the authentic initiation codon. An in-frame CTG codon 24 nucleotides upstream of the first AUG is present and could potentially serve as the initiation codon. We have named the protein encoded by this cDNA hPrt1, for human Prt1.


Fig. 1. The hPrt1 cDNA. A, schematic representation of the hPrt1 cDNA. A partial hPrt1 cDNA was obtained from Human Genome Science. A 200-base pair probe derived from the 5' end of this cDNA was used to screen a human placenta cDNA library. The coding region of clone 3-6 is indicated as an open box, while the untranslated regions are shown as solid lines. The two cDNAs are identical in the overlapping region. Some of the unique restriction sites are indicated. B, Northern blot analysis on total HeLa cell RNA using the probe shown in A.
[View Larger Version of this Image (13K GIF file)]


The cDNA sequence of hPrt1 is predicted to encode a protein containing a canonical RRM located between amino acids 185 and 270 (Fig. 2). The identification of the hPrt1 RRM is based on the consensus structural core sequence of RRMs (25), which include the presence of RNP-1 and RNP-2 sequence, and hydrophobic amino acids found at defined positions within the RRM. The presence of an RRM was also previously predicted for the yeast Prt1 protein (22, 28). A BLAST search (36) with the hPrt1 RRM sequence revealed that the RRM is most highly related to the fourth RRM of the poly(A)-binding protein, PABP. No other common protein motifs are evident. hPrt1 is acidic, with a predicted PI of 4.8. The middle portion of the protein is unusually rich in tryptophan residues (close to 5% tryptophan content over 400 amino acids). Amino acid sequence comparison between human and Saccharomyces cerevisiae Prt1 reveal extensive sequence identity (31% identity, 50% homology) across the entire protein except for the first 140 amino acids (Fig. 3). The similarity between yeast and human Prt1 is more striking in the middle portion of the protein, which encompasses the RRM. Several but not all of the tryptophan are conserved, suggesting that they are functionally important. A recent GenBankTM entry for the putative Prt1 from Schizosaccharomyces pombe is also highly conserved with hPrt1 (~35% identity at the protein level; data not shown). The hPrt1 protein contains 2 protein kinase A, 6 protein kinase C, and 17 casein kinase II consensus phosphorylation sites.


Fig. 2. Nucleotide and amino acid sequence of hPrt1. 5'- and 3'-untranslated regions are in lowercase letters. The RNP-1 consensus sequence (thick lined box) and the RNP-2 consensus sequence (thin lined box) of the RRM are indicated. The polyadenylation signal is underlined. Numbers on the right refer to the amino acid position. The hPrt1 sequence has been deposited in GenBankTM (accession number U62583[GenBank]).
[View Larger Version of this Image (57K GIF file)]



Fig. 3. Amino acid alignment of Prt1 sequences from human and S. cerevisiae. Identical residues are boxed. The alignment was performed using the Megalign program from the DNA*Star software, J. Hein method. Numbers refer to the amino acid position in each respective Prt1 species.
[View Larger Version of this Image (52K GIF file)]


hPrt1 Is an Integral Component of eIF3

It is conceivable that hPrt1 is a subunit of eIF3. To prove this, an immunological characterization of hPrt1 and eIF3 was performed. First, we translated in vitro an RNA derived from the hPrt1 cDNA. A single polypeptide, migrating at 116 kDa on a 9% SDS-polyacrylamide gel, was obtained (Fig. 4A, lane 2). The translation product co-migrated with a 115-kDa protein in a HeLa extract that immunoreacted with alpha -eIF3 antibodies (lane 3). Thus, the size of hPrt1 is similar to one of the eIF3 subunits. Next, we tested the ability of an alpha -eIF3 polyclonal antibody to recognize hPrt1. To this end, we expressed hPrt1 fused to the HA epitope tag in HeLa cells using a recombinant vaccinia virus expression system (35). Extracts from infected cells were blotted onto nitrocellulose and probed with a polyclonal alpha -eIF3 antibody (Fig. 4B). eIF3 subunits (p170 and p115) in extracts from cells transfected with the parental vector (pACTAG-2; Ref. 34) or pACTAG-hPrt1 were readily identifiable (lanes 1 and 2). A 125-kDa protein that cross-reacts with alpha -eIF3 was present in extracts from cells transfected with pACTAG-hPrt, but was absent from cells transfected with the vector alone (compare lanes 1 and 2). To confirm the identity of this protein as HA-hPrt1, immunoprecipitations using an alpha -HA antibody were performed and the products were probed with an alpha -eIF3 antibody. The immunoprecipitated HA-hPrt1 co-migrated with the 125-kDa polypeptide, and cross-reacted with the alpha -eIF3 antibody (lane 4). The slower mobility of HA-hPrt1 relative to hPrt1 is probably due to the three HA epitopes present in the fusion protein. Immunoprecipitates from cells transfected with the parental vector (lane 3) or with a vector encoding HA-La autoantigen (lane 5) failed to cross-react with alpha -eIF3. We conclude that hPrt1 is recognized by an antibody directed against eIF3. Finally, we wished to determine whether antibodies directed against hPrt1 could recognize a 115-kDa polypeptide in purified human eIF3. Attempts to generate antibodies against hPrt1 in rabbits failed. To circumvent this problem, affinity-purified hPrt1-specific antibodies from alpha -eIF3 antisera were prepared from crude eIF3 antibodies, using a bacterially expressed hPrt1 fragment. These antibodies recognized a protein migrating at approximately 115 kDa in a highly purified human eIF3 preparation and in HeLa extracts (Fig. 4C), and did not cross-react with hNip1, a 110-kDa protein also recently shown to be an eIF3 component (see "Discussion" and Ref. 43). Together with the previous data, these experiments strongly suggest that hPrt1 is one of the polypeptides found in the 115-kDa band of eIF3, the other being hNip1 (43).


Fig. 4. hPrt1 is a subunit of eIF3. A, the in vitro translated product of the hPrt1 cDNA co-migrates with one of the rabbit eIF3 subunits. In vitro translation (IVT) in rabbit reticulocyte lysate: lane 1, unprogrammed lysate; lane 2, programmed with hPrt1 RNA. HeLa extract was resolved on a 9% SDS-polyacrylamide gel, blotted onto nitrocellulose, and probed with alpha -eIF3 antibodies (lane 3). B, an alpha -eIF3 antibody cross-reacts with hPrt1. Western blotting using alpha -eIF3 antibodies is shown. Lanes 1 and 2, total cell extract (20 µg). Lanes 3-5, immunoprecipitation with anti-HA antibody. C, Western blotting with alpha -hPrt1 antibody.
[View Larger Version of this Image (24K GIF file)]


hPrt1 Interacts Directly with the p170 Subunit of eIF3

To further substantiate the finding that hPrt1 is a subunit of human eIF3, we examined the possibility that hPrt1 interacts directly with one or more eIF3 subunits. To this end, hPrt1 was tagged with a FLAG peptide linked to a heart muscle kinase phosphorylation site peptide (FLAG-HMK) or fused to a glutathioneS-transferase-FLAG-HMK sequence (GST-FLAG-HMK). We opted to use fragments of hPrt1 rather than the full-length protein due to the low yield and extensive degradation of full-length hPrt1 in E. coli. Fig. 5 illustrates the various fragments that were used. The proteins were purified using a FLAG antibody or glutathione-Sepharose resin, and were 32P-labeled with heart muscle kinase. The labeled proteins were then used to detect interacting proteins by the Far Western assay with HeLa cytoplasmic extracts, rabbit reticulocyte lysate, and different preparations of eIF3. Two of the probes, GST N255 (amino acids 1-255 of hPrt1) and NDelta 146 (amino acids 147-873), interacted with a 170-kDa protein in HeLa and rabbit reticulocyte lysates (Fig. 6, panels A and B, lanes 1 and 2). The hPrt1 probes reacted with a 140-kDa polypeptide in eIF3 preparation 1 (panels A and B, lane 3), and 140- and 170-kDa polypeptides in eIF3 preparation 2 (panels A and B, lane 4). A 32P-labeled probe consisting of a GST-HMK fusion fragment only failed to recognize any proteins (data not shown).The 170-kDa polypeptide in HeLa, rabbit reticulocyte lysate, and eIF3 preparation 2 could be the largest subunit of eIF3, p170. This polypeptide is sensitive to degradation, and the two eIF3 preparations used here differ by the extent of p170 proteolysis. An immunoblot using a monoclonal antibody directed against p170 (11) reveals the extent of p170 degradation in the eIF3 preparations (Fig. 6C), and clearly shows that a 140-kDa degradation product of p170 is present in preparation 1 (lane 3), and to a lesser extent, in preparation 2 (lane 4). This experiment demonstrates that hPrt1 interacts directly with the p170 subunit of eIF3. No other eIF3 subunits were recognized by the hPrt1 probes in this assay (data not shown).


Fig. 5. Schematic representation of hPrt1 deletions used in Far Western analysis. The amino acid sequence contained within each hPrt1 fragment is as follows: N255, hPrt1 residues 1-255; NDelta 146, 147-873. Also shown are two hPrt1 fragments: amino acids 147-255 and 147-209. The position of the RRM is shown by the gray box. The FLAG-heart muscle kinase tag is depicted by a black box. All hPrt1 fragments except NDelta 146 were purified as GST-HMK fusion proteins as described under "Experimental Procedures." NDelta 146 was purified on an alpha -FLAG column, as described.
[View Larger Version of this Image (15K GIF file)]



Fig. 6. hPrt1 interacts with p170 through a region that encompasses the RRM. A, B, D, and E, Far Western blots on HeLa cell extracts (70 µg), rabbit reticulocyte lysate (70 µg), and purified eIF3 preparations (1 µg) using various 32P-labeled hPrt1 fragments. The identity of the fragment is indicated at the bottom of each panel. C, immunoblot using a monoclonal antibody directed against the p170 subunit of eIF3 on membranes identical to those shown in panels A and B.
[View Larger Version of this Image (24K GIF file)]


The finding that both the NDelta 146 and N255 fragments of hPrt1 react with p170 suggest that the site of protein-protein interaction is located between amino acids 147 and 255. This segment of hPrt1, which encompasses most of the RRM, was assessed for its ability to interact with p170 independently of other sequences. A fragment containing amino acids 147-255 of hPrt1, used as a probe in a Far Western assay, indeed interacted with p170 (Fig. 6D). To further delimit the interaction site, fragment 147-209 was tested, and failed to interact with p170 (data not shown). Taken together, these data indicate that a portion of the RRM is crucial for the association between hPrt1 and p170.


DISCUSSION

We have cloned a human cDNA that is homologous to the yeast eIF3 subunit, Prt1. In vitro translation of hPrt1 RNA yielded a polypeptide of 116 kDa that co-migrated with one of the eIF3 subunits. Immunological characterization revealed that hPrt1 immunoreacts with alpha -eIF3 and that affinity-purified alpha -hPrt1 antibodies recognize a polypeptide of approximately 115 kDa in a highly purified eIF3 preparation. A direct interaction between hPrt1 and the p170 subunit of mammalian eIF3 has been demonstrated. Based on these data, we conclude that hPrt1 corresponds to one of the two 115-kDa subunits of eIF3. The immunoprecipitates of HA-hPrt1 did not contain other eIF3 subunits. It is likely that HA-hPrt1 does not incorporate well into the endogenous eIF3 because of the stability of the complex. Alternatively the HA tag may hinder association of HA-hPrt1 with eIF3.

Recently, Hershey and co-workers have isolated a human cDNA predicted to encode a 110-kDa protein that showed homology to the yeast Nip1 protein. Although Nip1 is not present in yeast eIF3 preparations, the human homologue is a subunit eIF3. Using affinity-purified antibodies against hPrt1 and hNip1, Asano et al. (43) demonstrated that these polypeptides co-localize with the 40 S ribosomal subunit and other eIF3 subunits. Our data combined with Hershey's group work suggest that mammalian eIF3 contains two subunits that migrate at approximately 115 kDa. Indeed, examination of various rat or rabbit eIF3 preparations resolved on SDS-polyacrylamide gel electrophoresis clearly shows two polypeptides migrating at this position (8, 9). In human eIF3, these two polypeptides were shown to correspond to immunologically distinct proteins, the largest (116 kDa) being hPrt1 and the smallest (110 kDa) corresponding to hNip1 (43). Thus, the mammalian eIF3 complex consists of at least nine polypeptides: p170, p116 (hPrt1), p110 (hNip1), p66, p47, p44, p40, p36, and p35. In addition, a new subunit, p48, has also been identified.4

Mutations in the PRT1 gene of S. cerevisiae impair translation initiation in vivo at 37 °C (37, 38). One of the mutants, prt1-1, failed to promote binding of the ternary complex eIF2·GTP·tRNAiMet to the 40 S ribosomal subunit (24). Evans et al. (28) identified six mutations in PRT1 that impair translation initiation. Two of these mutations alter amino acids that are conserved between yeast and human Prt1. Human Prt1, when expressed in the prt1-1 yeast strain, was unable to rescue the temperature sensitive phenotype.5 This was somewhat surprising since yeast eIF3 functions in a mammalian methionyl-puromycin assay system (19). Methionyl-puromycin synthesis is dependent on the binding of the ternary complex to the 40 S ribosome, requires only washed ribosomes, tRNAiMet, eIF1A, eIF2, eIF3, eIF5, and eIF5A (39), but does not measure mRNA binding to the ribosome. It is clear that yeast eIF3 can replace mammalian eIF3 for some, but not all normal eIF3 functions, and that hPrt1 is unable to fulfill all the roles of yeast Prt1. One reason for the inability of hPrt1 to replace Prt1 in vivo is that it may not incorporate into the yeast eIF3 complex. A Far Western analysis on yeast extracts using the hPrt1 N255 fragment did not reveal any interacting proteins.6

Both yeast and human Prt1 contain an amino-terminal RRM (residues 185-270 in hPrt1). The RRM contains the sequence elements that are responsible for specific protein-protein interactions with the p170 subunit of eIF3 (Fig. 6). It is unlikely that the interaction between hPrt1 and p170 is mediated through RNA, since hPrt1 was unable to bind a radiolabeled RNA probe as measured by UV photocross-linking and Northwestern assays.5 Furthermore, treatment of the FLAG-HMK NDelta 146 probe and the nitrocellulose membrane with RNase A did not reduce the intensity of the interaction with p170.5 It is possible that the RRM is functional as an RNA binding module only within the eIF3 complex, and that its RNA binding activity and specificity are modulated by p170. Precedents for protein-protein interactions altering the RNA binding activity of an RRM-containing protein exist. The spliceosomal protein U2B" is unable on its own to distinguish between the U1 and U2 small nuclear RNAs, but will bind specifically to U2 small nuclear RNA in the presence of the U2A' protein (40, 41). U2A' and U2B" associate in the absence of RNA, an interaction that is mediated by the RRM (41). The major RNA-binding protein of yeast eIF3 is the p62 subunit (19). It would be interesting to compare the RNA binding specificities of p62 and the whole eIF3. We have shown previously that the p170 subunit of eIF3 interacts directly with eIF4B (42). The sites of interaction of eIF4B and hPrt1 on p170 do not appear to be the same, since hPrt1 reacts very strongly with the 140-kDa degradation product of p170, while eIF4B does not. The multiple protein-protein interactions involving p170 suggest that this subunit may play a central role in eIF3 function.


FOOTNOTES

*   This work was supported in part by a grant from the Medical Research Council of Canada (to N. S.). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U62583[GenBank].


§   Recipient of a studentship from the Medical Research Council of Canada. Present address: Zoologisches Institut der Universitat Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland.
   Supported by a long term fellowship from the Human Frontier Science Program.
**   To whom correspondence should be addressed: McGill University, Department of Biochemistry and McGill Cancer Centre, 3655 Drummond St., Rm. 807, Montréal, PQ, Canada H3G 1Y6. Tel.: 514-398-7274; Fax: 514-398-1287; E-mail: sonenberg{at}medcor.mcgill.ca.
1    The abbreviations used are: eIF, eukaryotic initiation factor; DTT, 1,4-dithiothreitol; EST, expressed sequence tag; GST, glutathione S-transferase; HA, hemagglutinin; HMK, heart muscle kinase; PCR, polymerase chain reaction; RRM, RNA recognition motif; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum.
2    K. Asano and J. W. B. Hershey, submitted for publication.
3    K. Johnson and W. C. Merrick, submitted for publication.
4    K. Asano and J. W. B. Hershey, personal communication.
5    N. Méthot, unpublished data.
6    N. Méthot, E. Rom, H. Olsen, and N. Sonenberg, unpublished observations.

Acknowledgments

We thank Hans Trachsel and William C. Merrick for the generous gifts of eIF3 preparations, and H. Trachsel for the p170 monoclonal antibody. We are grateful to K. Asano and J. W. B. Hershey for communicating their results prior to publication and for the experiment shown in Fig. 4C. We are indebted to Morag Park for the placenta lambda gt11 cDNA library.


REFERENCES

  1. Merrick, W. C., and Hershey, J. W. B. (1996) in Translational Control (Hershey, J. W. B., Mathews, M. B., and Sonenberg, N., eds), pp. 31-69, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
  2. Pain, V. M. (1996) Eur. J. Biochem. 236, 747-771 [Abstract]
  3. Benne, R., and Hershey, J. W. B. (1978) J. Biol. Chem. 253, 3078-3087 [Medline] [Order article via Infotrieve]
  4. Safer, B., Adams, S. L., Kemper, W. M., Berry, K. W., Floyd, M., and Merrick, W. C. (1976) Proc. Natl. Acad. Sci. U. S. A. 73, 2584-2588 [Abstract]
  5. Trachsel, H., Erni, B., Schreier, M. H., and Staehelin, T. (1977) J. Mol. Biol. 116, 755-767 [Medline] [Order article via Infotrieve]
  6. Schreier, M. H., Erni, B., and Staehelin, T. (1977) J. Mol. Biol. 116, 727-753 [Medline] [Order article via Infotrieve]
  7. Benne, R., and Hershey, J. W. B. (1976) Proc. Natl. Acad. Sci. U. S. A. 73, 3005-3009 [Abstract]
  8. Behlke, J., Bommer, U.-A., Lutsch, G., Henske, A., and Bielka, H. (1986) Eur. J. Biochem. 157, 523-530 [Abstract]
  9. Meyer, L. J., Milburn, S. C., and Hershey, J. W. B. (1982) Biochemistry 21, 4206-4212 [Medline] [Order article via Infotrieve]
  10. Milburn, S. C., Duncan, R. F., and Hershey, J. W. B. (1990) Arch. Biochem. Biophys. 276, 6-11 [Medline] [Order article via Infotrieve]
  11. Mengod, G., and Trachsel, H. (1985) Biochim. Biophys. Acta 825, 169-174 [Medline] [Order article via Infotrieve]
  12. Hannig, E. M. (1995) BioEssays 17, 915-919 [Medline] [Order article via Infotrieve]
  13. Gupta, N. K., Roy, A. L., Nag, M. K., Kinzy, T. G., MacMillan, S., Hileman, R. E., Dever, T. E., Wu, W., Merrick, W. C., and Hershey, J. W. B. (1990) in Post-transcriptional Control of Gene Expression (McCarthy, J. E. G., and Tuite, M. F., eds), pp. 521-526, Springer-Verlag, Berlin
  14. Goumans, H., Thomas, A., Verhoeven, A., Voorma, H. O., and Benne, R. (1980) Biochim. Biophys. Acta 608, 39-46 [Medline] [Order article via Infotrieve]
  15. Peterson, D. T., Merrick, W. C., and Safer, B. (1979) J. Biol. Chem. 254, 2509-2516 [Abstract]
  16. Nygard, O., and Westermann, P. (1982) Nucleic Acids Res. 10, 1327-1334 [Abstract]
  17. Westermann, P., and Nygard, O. (1984) Nucleic Acids Res. 12, 8887-8897 [Abstract]
  18. Garcia-Barrio, M. T., Naranda, T., Vazquez de Aldana, C. R., Cuesta, R., Hinnebusch, A. G., Hershey, J. W. B., and Tamame, M. (1995) Genes Dev. 9, 1781-1796 [Abstract]
  19. Naranda, T., MacMillan, S. E., and Hershey, J. W. B. (1994) J. Biol. Chem. 269, 32286-32292 [Abstract/Free Full Text]
  20. Lamphear, B. J., Kirchweger, R., Skern, T., and Rhoads, R. E. (1995) J. Biol. Chem. 270, 21975-21983 [Abstract/Free Full Text]
  21. Naranda, T., MacMillan, S. E., Donahue, T. F., and Hershey, J. W. B. (1996) Mol. Cell. Biol. 16, 2307-2313 [Abstract]
  22. Hanic-Joyce, P. J., Singer, R. A., and Johnston, G. C. (1987) J. Biol. Chem. 262, 2845-2851 [Abstract/Free Full Text]
  23. Danaie, P., Wittmer, B., Altmann, M., and Trachsel, H. (1995) J. Biol. Chem. 270, 4288-4292 [Abstract/Free Full Text]
  24. Feinberg, B., McLaughlin, C. S., and Moldave, K. (1982) J. Biol. Chem. 257, 10846-10851 [Abstract]
  25. Birney, E., Kumar, S., and Krainer, A. R. (1993) Nucleic Acids Res. 21, 5803-5816 [Abstract]
  26. Burd, C. G., and Dreyfuss, G. (1994) Science 265, 615-621 [Medline] [Order article via Infotrieve]
  27. Nagai, K., Oubridge, C., Ito, N., Avis, J., and Evans, P. (1995) Trends Biochem. Sci. 20, 235-240 [CrossRef][Medline] [Order article via Infotrieve]
  28. Evans, D. R. H., Rasmussen, C., Hanic-Joyce, P. J., Johnston, G. C., Singer, R. A., and Barnes, C. A. (1995) Mol. Cell. Biol. 15, 4525-4535 [Abstract]
  29. 3174Adams, M. D., Fleischmann, R. D., Fuldner, R. A., Bult, C. J., Lee, N. H., Kirkness, E. F., Weinstock, K. G., Gocayne, J. D., White, O., et al. (1995) Nature 377, (Suppl.), 3-174
  30. Feinberg, A. P., and Vogelstein, B. (1984) Anal. Biochem. 137, 266-267 [Medline] [Order article via Infotrieve]
  31. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 5463-5467 [Abstract]
  32. Blanar, M. A., and Rutter, W. J. (1992) Science 256, 1014-1018 [Medline] [Order article via Infotrieve]
  33. Méthot, N., Pause, A., Hershey, J. W. B., and Sonenberg, N. (1994) Mol. Cell Biol. 14, 2307-2316 [Abstract]
  34. Charest, A., Wagner, J., Shen, S.-H., and Trembley, M. L. (1995) Biochem. J. 308, 425-432 [Medline] [Order article via Infotrieve]
  35. Fuerst, T. R., Niles, E. G., Studier, F. W., and Moss, B. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 8122-8126 [Abstract]
  36. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. P. (1990) J. Mol. Biol. 215, 403-410 [CrossRef][Medline] [Order article via Infotrieve]
  37. Hartwell, L. H., and McLaughlin, C. S. (1968) J. Bacteriol. 96, 1664-1671 [Medline] [Order article via Infotrieve]
  38. Hartwell, L. H., and McLaughlin, C. S. (1969) Proc. Natl. Acad. Sci. U. S. A. 62, 468-474 [Abstract]
  39. Benne, R., Brown-Luedi, M. L., and Hershey, J. W. B. (1979) Methods Enzymol. 60, 15-35 [Medline] [Order article via Infotrieve]
  40. Scherly, D., Boelens, W., Dathan, N. A., van Venrooij, W. J., and Mattaj, I. W. (1990) Nature 345, 502-506 [CrossRef][Medline] [Order article via Infotrieve]
  41. Scherly, D., Dathan, N. A., Boelens, W., van Venrooij, W. J., and Mattaj, I. W. (1990) EMBO J. 9, 3675-3681 [Abstract]
  42. Méthot, N., Song, M. S., and Sonenberg, N. (1996) Mol. Cell. Biol. 16, 5328-5334 [Abstract]
  43. Asano, K., Kinzy, T. G., Merrick, W. C., and Hershey, J. W. B. (1997) J. Biol. Chem. 272, 1101-1109 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.