Dual Binding Mode of the Nascent Polypeptide-associated Complex Reveals a Novel Universal Adapter Site on the Ribosome

Abstract

Nascent polypeptide-associated complex (NAC) was identified in eukaryotes as the first cytosolic factor that contacts the nascent polypeptide chain emerging from the ribosome. NAC is present as a homodimer in archaea and as a highly conserved heterodimer in eukaryotes. Mutations in NAC cause severe embryonically lethal phenotypes in mice, Drosophila melanogaster, and Caenorhabditis elegans. In the yeast Saccharomyces cerevisiae NAC is quantitatively associated with ribosomes. Here we show that NAC contacts several ribosomal proteins. The N terminus of βNAC, however, specifically contacts near the tunnel exit ribosomal protein Rpl31, which is unique to eukaryotes and archaea. Moreover, the first 23 amino acids of βNAC are sufficient to direct an otherwise non-associated protein to the ribosome. In contrast, αNAC (Egd2p) contacts Rpl17, the direct neighbor of Rpl31 at the ribosomal tunnel exit site. Rpl31 was also recently identified as a contact site for the SRP receptor and the ribosome-associated complex. Furthermore, in Escherichia coli peptide deformylase (PDF) interacts with the corresponding surface area on the eubacterial ribosome. In addition to the previously identified universal adapter site represented by Rpl25/Rpl35, we therefore refer to Rpl31/Rpl17 as a novel universal docking site for ribosome-associated factors on the eukaryotic ribosome.

Introduction

The biosynthesis of proteins by ribosomes is an essential process in all living cells. As soon as a newly synthesized polypeptide emerges from the ribosomal exit tunnel several ribosome-associated factors, which are involved in maturation, folding, and/or sorting of the protein, contact the nascent chain (1–3). In eubacteria, these associated factors include trigger factor, the modifying enzymes peptide deformylase and methionine aminopeptidase, signal recognition particle (SRP)⁴ and its receptor (SR), and the translocation pore SecY. In contrast, in eukaryotes a larger variety of such factors must gain access to the nascent chain (the chaperones RAC and SSB), the nascent polypeptide-associated complex (NAC), various methionine aminopeptidases and N-acetyltransferases, the ribosome-associated membrane protein ERj1p as well as SRP, SR, and the translocon Sec61. Because the available interaction space around the ribosomal tunnel exit is rather limited, these factors must have developed different strategies to encounter the nascent chain. The ribosomal exit tunnel, which itself is mainly formed by ribosomal RNA, is surrounded at its exit site by a small set of ribosomal proteins. Whereas some of these are universally conserved through all domains of life (Rpl17, Rpl25, Rpl26, and Rpl35 in yeast corresponding to L22, L23, L24, and L29 in eubacteria), others are either restricted to eubacteria (L17 and L32) or only found in archaea and eukaryotes (Rpl19, Rpl31, and Rpl39) (4–7). Based on cross-linking experiments and structural studies, Rpl25/Rpl35 (L23/L29 in eubacteria) have been identified so far as a general docking site for the ribosome-associated factors trigger factor, SRP, the translocon Sec61 or SecY, the ribosome-associated membrane protein ERj1p, and the “insertases” YidC or Oxa1, respectively (8–13).

In eukaryotes, NAC was identified as the first ribosome-associated factor to contact the emerging polypeptide chain (14). NAC is a highly abundant heterodimeric cytosolic protein complex composed of αNAC and βNAC, which show substantial homology to each other (15). Although not essential in Saccharomyces cerevisiae, the importance of NACs in vivo function is emphasized by early embryonically lethal phenotypes of NAC mutants in higher eukaryotes, such as Mus musculus, Drosophila melanogaster, and Caenorhabditis elegans (16–18). Intracellular levels of the individual NAC subunits change in the context of several human diseases, such as Alzheimer disease, Down syndrome, AIDS, malignant brain tumors, and a role for NAC in apoptosis was proposed (16, 19–23). NAC does not show similarity to other proteins and its cellular function is still poorly understood.

Early protease protection assays suggested that NAC might function as a shield for newly synthesized polypeptides on the ribosome against inappropriate interaction with cytosolic factors (24). Cycles of binding and releasing NAC were proposed to expose the polypeptide to the cytosol in “quantal units,” rather than amino acid by amino acid. Thereby NAC would contribute to fidelity in co-translational processes such as targeting and folding (25). Additional studies suggested that NAC is involved in regulation of ribosome access to the translocation pore in the ER membrane in co-translational protein translocation (26–28). Yeast NAC was shown to play a role in the attachment of cytosolic ribosomes to mitochondria (29, 30) and in translation-coupled import of proteins into mitochondria (31, 32). For the human αNAC also transcription-related functions have been described (33, 34). Both NAC subunits seem to differ with respect to their function. Although both subunits contact the nascent chain on the ribosome, as was shown by cross-linking, βNAC alone was sufficient for binding to the ribosome and prevention of ribosome interaction with the translocon (15, 35). In contrast, only the heterodimeric complex prevents inappropriate interaction with the nascent chain (15).

The availability of numerous completed archaebacterial genomes revealed that all archaeal taxa contain one gene with apparent homology to αNAC (36). The crystal structure of archaeal NAC uncovered that it forms a homodimer exhibiting two folded domains (37). A ubiquitin-associated domain (UBA domain) at the C terminus of archaeal NAC, which is also found in all eukaryotic αNAC proteins, and the central NAC domain. The latter provides the dimerization interface. The central NAC domain exhibits a unique novel protein fold. It resembles a flattened β barrel that exposes several hydrophobic residues on one of its concave surfaces. This domain is structurally conserved from archaea to humans emphasizing its importance for NACs function.

The observation of cross-links between NAC and very short nascent chains (24) imply that the NAC binding site on the ribosome must be in close proximity to the tunnel exit on the large ribosomal subunit. Recently, Rpl25 was suggested as a NAC binding site on the ribosome based on a heterologous cross-linking assay using yeast NAC and Escherichia coli ribosomes (38).

To further characterize the function of NAC we wanted to investigate its interaction with the ribosome in more detail. Therefore we followed a cross-linking approach to identify the major binding site of NAC on the ribosome. Based on these results we constructed various NAC mutants and studied the sedimentation behavior of these NAC variants. In this report we show that NAC interacts with several ribosomal proteins. The pivotal contact is made to Rpl31 via the N terminus of βNAC. We show that the ability to interact with ribosomes can be conferred by fusing this N terminus to a protein that is otherwise not associated with ribosomes. In addition, we show that αNAC contacts Rpl17, the direct neighbor of Rpl31 on the ribosomal surface and propose a model for NAC interaction with this novel universal docking site at the ribosomal tunnel exit.

EXPERIMENTAL PROCEDURES

Materials

Chemicals and secondary antibodies were purchased from Merck and Sigma, restriction enzymes and Vent polymerase from New England Biolabs, protease inhibitor mixture was from Roche Applied Science, and RNase inhibitor from Promega. Cross-linking reagents were purchased from Pierce and Molecular Biosciences. Samples were analyzed on precast BisTris gels from Invitrogen.

Strains

S. cerevisiae strain W303-1A and the NAC knock-out strain YJF24 (Δegd2; Δegd1; Δbtt1) were a gift of B. Wiedmann (35). Strain MH272-3fα and the ΔRpl31A/B strain were provided by S. Rospert (39). Cloning was performed in E. coli XL1 Blue (Stratagene), expression in E. coli ER2566 (New England Biolabs).

Cloning, Expression, and Purification

Cloning experiments were performed following standard protocols (40). The egd2 and egd1 genes were amplified from yeast genomic DNA and cloned together into pET28a (Novagen) leading to a His tag fusion followed by a thrombin cleavage site at the N terminus of αNAC. The genes were expressed in E. coli ER2566 and the protein complex purified via nickel-nitrilotriacetic acid-agarose (Qiagen, according to the manufacturer's manual) followed by heparin affinity chromatography (POROS 20 HE, Applied Biosystems) and anion exchange chromatography (POROS 20 HQ, Applied Biosystems). The His tag was removed from αNAC by thrombin cleavage during the purification procedure. In addition, egd1 and its alanine substitution mutants were cloned into pRS425-GAL and egd2 into pYES2.1 (Invitrogen) for expression in YJF24.

Antibodies and Immunoblotting Procedures

The genes coding for yeast Rpl4, Rpl35, and Rpl25 were cloned into pET28a, expressed in E. coli, purified via a His tag, and used for immunization of chicken (David Biotechnologies, Regensburg). Preimmune sera did not react with any protein in the Western blot with total cell extracts from S. cerevisiae prepared according to Yaffe and Schatz (41). Proteins were separated on 10% NuPAGE BisTris gels using MES buffer (Invitrogen) followed by transfer onto nitrocellulose or polyvinylidene difluoride membrane using Bis-Bicine transfer buffer (25 mm Bicine, 25 mm BisTris, 1 mm EDTA free acid, 20% methanol). Secondary antibodies were conjugated with horseradish peroxidase and detected using the ECL kit (Amersham Biosciences).

Site-specific Mutagenesis

Site-specific introduction of unique cysteine residues in the N terminus of βNAC at positions 3, 5, 7, 9, or 11 as well as replacement of ⁶EKL⁸ and ¹⁰KLQ¹² against AAA in βNAC were performed following the “QuikChange Site-directed Mutagenesis” protocol (Stratagene). Fusion genes coding for the first 14, 23, or 39 N-terminal amino acids of βNAC fused via a (Gly₄,Ser)₃ linker to the full-length gene of the maltose-binding protein from E. coli were constructed using standard PCR techniques, cloned into pYES2.1, and expressed in YJF24.

Isolation of Ribosomes from Yeast

Yeast cultures were grown to an A₆₀₀ of 1.5 on YPD medium, cells were collected, washed with ice-cold ddH₂O and 1% KCl, followed by incubation with 100 mm Tris-HCl, pH 8.0, 10 mm DTT for 15 min at 30 °C. Cells were collected, resuspended in lysis buffer (20 mm Hepes-KOH, pH 7.5, 100 mm KOAc, 5 mm Mg(OAc)₂, 4 mm DTT, 150 mm sucrose, 500 μm phenylmethylsulfonyl fluoride, protease inhibitor) and disrupted by three passages through an EmulsiFlex-C5 High Pressure homogenizer at 1,400 bar. Cell debris was separated by spinning for 15 min in a SS34 rotor at 15,500 rpm. The supernatant was spun again in a Ti-60 rotor for 30 min at 35,700 rpm. The resulting supernatant was termed S100. Ribosomes were pelleted from this S100 extract through a high salt (500 mm KOAc) or low salt (100 mm KOAc) sucrose cushion (20 mm Hepes-KOH, pH 7.5, 100–500 mm KOAc, 5 mm Mg(OAc)₂, 2 mm DTT, protease inhibitor, 1 m sucrose) in a TLA 100.3 rotor for 1 h at 100,000 rpm.

Chemical Cross-linking of NAC and Ribosomal Proteins Using AMAS

A 2-fold molar excess of yeast NAC was incubated with high salt-stripped ribosomes isolated from the NAC knock-out yeast strain YJF24 as follows: 80 S ribosomes (4 μm) were incubated with 8 μm recombinant NAC (in 20 mm Hepes, pH 7.5, 150 mm KOAc, 20 mm Mg(OAc)₂, protease inhibitor mixture complete) for 2 min at 26 °C followed by 5 min on ice. AMAS (N-(α-maleimidoacetoxy)succinimide ester) in dimethyl sulfoxide was added to a final concentration of 1.6 mm. After 2 h at 4 °C the reaction was stopped by a 50-fold excess of glycine and incubation at 25 °C for 30 min.

Reactions were brought to high salt by addition of KOAc to 0.5 m, and incubated on ice for 30 min before the ribosome-bound material was pelleted through a 0.5 m sucrose cushion (20 mm Hepes, pH 7.5, 500 mm KOAc, 20 mm Mg(OAc)₂, protease inhibitor mixture complete) at 355,000 × g in a TLA-100.2 rotor. Pellets were resuspended in RNase buffer (50 mm Tris-HCl, pH 8.0, 0.5 m NaCl, 2 mm EDTA, 0.5 mg/ml RNase A, protease inhibitor mixture complete), incubated for 1 h at 37 °C before addition of SDS sample buffer and separation on 10% BisTris gels (Invitrogen).

Site-specific Cross-linking of NAC and Ribosomal Proteins Using BPIA

Site-specific cross-linking with recombinant NAC carrying unique cysteins at the N terminus of βNAC was performed in principle according to Ref. 42. 6 μm NACs (in 20 mm Hepes-KOH, pH 7.5, 150 mm KOAc, 20 mm Mg(OAc)₂, 5 mm TCEP (Tris-(2-carboxyethyl)-phosphine), protease inhibitor complete) were incubated for 30 min at 30 °C to reduce possibly formed disulfide bridges. BPIA (benzophenone-4-iodoacetamide) in dimethyl formamide was added to a final concentration of 12 μm followed by a 1-h incubation at room temperature in the dark. DTT was added in a 5-fold excess and incubation was continued for 15 min. High salt purified 80 S ribosomes from YJF24 were added to a final concentration of 3 μm, resulting in 2 molar excess on NACs over ribosomes. Samples were incubated for 2 min at 26 °C and 5 min on ice in the dark. The cross-linker was activated by irradiation with UV light (Blak Ray lamp model B100 AP 100 W 365 nm for 5 min at 5 cm distance). The samples were then high salt-treated and ribosome-bound material was treated as described above.

Analyses of NACs Association with Ribosomes

Ribosomes were isolated under low salt (100 mm KOAc) conditions. Yeast cultures (W-303-1A or YJF24 expressing egd1 or egd2 from plasmids) were grown on YPD medium or the appropriate selection medium to an A₆₀₀ of 1.0 to 1.5. Cells were harvested (7.000 × g, 6 min, 4 °C), washed with ice-cold ddH₂O followed by 1% KCl. Cells were then incubated in 100 mm Tris-HCl, pH 8.0, 10 mm DTT for 15 min at 30 °C, pelleted again, resuspended in lysis buffer (20 mm Hepes-KOH, pH 7.5, 100 mm KOAc, 20 mm Mg(OAc)₂, 4 mm DTT, 150 mm sucrose, 500 μm phenylmethylsulfonyl fluoride, protease inhibitor mixture complete) and disrupted by 3 passages through an EmulsiFlex-C5 High Pressure homogenizer at 1,400 bar. Cell debris was separated at 16,000 × g for 15 min. The supernatant was spun again at 100,000 × g for 30 min to yield the S100 extract that was quick frozen in liquid nitrogen or directly used for sedimentation of ribosomes. A 1-ml sucrose cushion (20 mm Hepes-KOH, pH 7.5, 100 mm KOAc, 20 mm Mg(OAc)₂, 4 mm DTT, 1 m sucrose, 500 μm phenylmethylsulfonyl fluoride, protease inhibitor mixture complete) was overlaid with 2 ml of S100 and ribosomes were pelleted for 1 h at 540,000 × g and 4 °C. The supernatant was carefully removed and ribosomes were resuspended in ribosome buffer (20 mm Hepes-KOH, pH 7.5, 100 mm KOAc, 20 mm Mg(OAc)₂, 1 mm DTT, 500 μm phenylmethylsulfonyl fluoride, protease inhibitor mixture complete). Equivalent amounts of ribosomes and supernatant were analyzed by immunoblotting.

In Vitro Binding of NAC Variants to Ribosomes

High salt-stripped ribosomes (4 μm) from yeast (YJF24) or E. coli (W3110) were incubated with a 2 molar excess of recombinant NAC (8 μm in 20 mm Hepes-KOH, pH 7.5, 150 mm KOAc, 20 mm Mg(OAc)₂, 1 mm DTT, protease inhibitor) for 2 min at 26 °C followed by 5 min on ice. Ribosomes were sedimented through a 1 m sucrose cushion (20 mm Hepes-KOH, pH 7.5, 100 mm KOAc, 20 mm Mg(OAc)₂, 1 mm DTT, 1 m sucrose, protease inhibitor) for 1 h at 355,000 × g at 4 °C. Equivalent amounts of the ribosomes and the supernatant were analyzed by immunoblotting.

RESULTS

NAC Is Associated with Ribosomes via the N Terminus of βNAC

To characterize the binding partner(s) of NAC on the ribosome we first verified that the amino-terminal 11 amino acids of βNAC are critical for ribosome association in yeast as described before (43). We set up a system that would allow us to study different variants of NAC in the same cellular background. For this purpose we expressed egd1 (coding for yeast βNAC) and egd2 (coding for yeast αNAC) from two separate plasmids under control of the GAL1 promoter in a yeast NAC knock-out strain (YJF24: Δegd1, Δegd2, Δbtt1) (35). Growth of W303-1A (the corresponding wild type strain) and YJF24+egd2+egd1 on d-galactose as carbon source resulted in comparable amounts of αNAC and βNAC in both strains (see supplemental Fig. S1a). Furthermore, when we sedimented ribosomes from S100 extracts prepared from two different strains, comparable amounts of NAC remained associated with the ribosome (see supplemental Fig. S1a). For subsequent assays we sedimented the ribosomes through a sucrose cushion at a physiological salt concentration and again comparable amounts of NAC sedimented with the yeast ribosomes in the WT situation as well as when NAC was expressed from a plasmid in a NAC knock-out strain (see supplemental Fig. S1b). When we replaced the plasmid that contained the egd1 gene against a plasmid coding for a truncated βNAC, which is missing the amino-terminal 11 amino acids (egd1Δ1–11), expression levels of the NAC subunits in strain YJF24 did not change in comparison with the ones observed in the WT strain (Fig. 1a, compare lanes 1 and 10). However, sedimentation of the truncated NAC with ribosomes was completely abolished (Fig. 1a, lanes 11 and 12 as opposed to lanes 2 and 3).

FIGURE 1.

Association of different NAC mutants with 80 S ribosomes. a, ribosomes were sedimented through low salt sucrose cushions. 0.5 A₂₆₀ units of the pellet (P) together with $\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$10$}\right.$ of the equivalent volume of total (T = S100) and supernatant (S) were analyzed via Western blot. The upper half of the membrane was analyzed with antibodies directed against αNAC and βNAC. The lower half with an antibody directed against Rpl4 to verify that comparable amounts of ribosomes were sedimented in each assay. b, association of maltose-binding protein fused to different parts of the βNAC-N terminus with 80 S ribosomes. Maltose-binding protein fused to different parts of the N terminus of βNAC was expressed in YJF24. Ribosomes were sedimented from S100 extracts. Supernatants and pellets were checked for co-sedimentation of the respective fusion protein in a Western blot using antibodies directed against Rpl35 or MBP. Whereas the first 14 amino acids of βNAC were not sufficient for co-sedimentation of MBP with ribosomes the first N-terminal 23 amino acids were sufficient. Schematic representations of the fusion constructs are shown at the bottom. The start of the NAC domain is marked by a red box.

The N terminus of βNAC is predicted to form an α-helical secondary structure. Comparison of the N termini of βNAC from diverse eukaryotes, including plants, animals, and fungi revealed that there are two absolutely conserved (Lys⁷, Leu⁸) and several highly conserved (Glu⁶, Lys¹⁰, Leu¹¹, Gln¹²) amino acids within this N-terminal sequence of βNAC (see supplemental Fig. S2). We replaced two motifs independently in βNAC (⁶EKL⁸ and ¹⁰KLQ¹²) against alanines (AAA) and verified that the predicted secondary structure for the respective peptide sequence was not changed by the mutations using the program PHD (see supplemental Fig. S3)(44). We expressed these altered versions of βNAC together with αNAC in YJF24. Expression levels of the mutant βNAC (E6A,K7A,L8A or K10A,L11A,Q12A) was comparable with WT βNAC (compare lanes 1, 4, and 7 in Fig. 1a). When we sedimented ribosomes from S100 extracts derived from these strains, the replacement of ⁶EKL⁸ by AAA in βNAC lead to a severe reduction of ribosome association (see Fig. 1a compare lanes 3 and 6). The effect was even more profound when we changed ¹⁰KLQ¹² to AAA. In this case ribosome association of NAC was almost abolished (see Fig. 1a compare lanes 3 and 9). Together these results emphasize the importance of the conserved ⁶EKLXKLQ¹² motif at the N terminus of βNAC for NACs association with the ribosome.

The First 23 Amino Acids of βNAC Are Sufficient for Ribosome Association

Next we asked whether the N terminus of βNAC is not only necessary but sufficient to direct a protein to the ribosome that is usually not associated with ribosomes. We fused sequences coding for peptides with increasing length of the N terminus of βNAC via a linker sequence to the malE gene from E. coli coding for the maltose-binding protein (MBP). The fusion proteins were expressed in YJF24 and tested for sedimentation with ribosomes from S100 extracts prepared from the respective yeast strains. Fig. 1b shows that full-length MBP, as expected, does not associate with ribosomes on its own (lanes 1 and 2). Adding the first 14 amino acids of βNAC did not result in cosedimentation with the ribosome (lanes 3 and 4), indicating at the same time that the chosen linker sequence (Gly₄,Ser)₃ does not artificially lead to ribosome association of the fusion protein. However, the first 23 amino acids can mediate ribosome association of βNAC-(1–23)-MBP (lanes 5 and 6). Extension of the βNAC part in the fusion protein to amino acid 39 did not significantly increase the amount of associated protein (lanes 7 and 8). Interestingly, these first 23 amino acids do not contain the motif (Arg²⁴–Lys³⁰) that was recently identified by Wegrzyn et al. (38) as being responsible for NACs association with the ribosome via Rpl25 (see “Discussion”).

NAC Interacts with the Ribosome via Multiple Contact Sites

Eukaryotic NAC contains several areas that are predicted to be intrinsically disordered and thus the flexible ends of the two polypeptide chains could contact multiple distinct areas on the ribosomal surface. Because WT NAC contains many lysines spread throughout both polypeptide chains (αNAC 16 Lys and βNAC 18 Lys) we performed cross-linking assays to probe for the contact sites of NAC on the ribosome using a bifunctional cross-linker (AMAS), which reacts with lysines under the chosen conditions but only spans a distance of 4.4 Å. NAC was incubated with high salt-purified ribosomes from strain YJF24 in the presence of AMAS. Subsequently, ribosomes were pelleted through sucrose cushions to remove non-cross-linked NAC and assayed for cross-linking products on a Western blot decorated with antibodies against βNAC. Fig. 2a shows that βNAC forms two strong cross-links to ribosomal proteins with an apparent molecular mass of 15 and 25 kDa (lane 5). Particularly, the cross-link to the 15-kDa protein depended on the presence of the N-terminal 11 amino acids of βNAC, because it was not detected when the cross-linking assay was performed using NAC-(βNACΔ1–11) (see Fig. 2a, lane 4). In a first attempt to identify the ribosomal protein interacting with NAC we used antisera against Rpl25, Rpl35, and Rpl4 but none of these cross-links reacted with the antibodies (see supplemental Fig. S4 for the Western blot using anti-Rpl25 antibodies).

FIGURE 2.

Cross-linking of different NAC variants to 80 S ribosomes. a, NAC or NAC missing the first 11 N-terminal amino acids of βNAC was incubated with 80 S ribosomes from a yeast NAC knock-out strain in the presence of AMAS. Ribosomes were pelleted and cross-linking products were detected by Western blotting with an antibody directed against βNAC. b, NAC carrying a unique cysteine at position 7 of βNAC was incubated with 80 S ribosomes from a yeast NAC knock-out strain in the presence of the cross-linker BPIA. Cross-linking products were detected by Western blotting with an antibody directed against βNAC. In the presence of ribosomes βNAC cross-linked to only one ribosomal protein of an apparent molecular mass of 15 kDa (see lane 3). Lanes 1 and 2 represent NAC or ribosomes alone with BPIA showing no cross-linking product.

The N Terminus of βNAC Cross-links to Ribosomal Protein Rpl31

To test whether the N terminus of βNAC was directly responsible for the cross-link to the 15-kDa protein or whether the cross-link was rather influenced by the presence of the N terminus, we made use of the fact that neither αNAC nor βNAC contain any cysteine residues. We introduced unique cysteines at positions 3, 5, 7, 9, or 11 of βNAC, expressed the respective protein together with αNAC in E. coli, and purified the resulting complexes to homogeneity. We labeled the single cysteine via the iodoacetamide group of a photoactivatable cross-linker (BPIA), and added high salt-purified 80 S ribosomes from YJF24. Upon activation by UV light this cross-linker reacts with C-H bonds (45). To remove non-cross-linked NAC, ribosomes were pelleted through a sucrose cushion before analysis by Western blotting and detection of cross-linking products by an antibody specific for βNAC. All five different NAC complexes resulted in a single cross-link to a ribosomal protein with an apparent molecular mass of 15 kDa. Fig. 2b shows the representative result for NAC (βNAC^K7C). The cross-link shown in lane 3 did not react with a polyclonal antibody directed against Rpl25.

To identify the 15-kDa cross-linking partner of βNAC we repeated the assay on a preparative scale incubating 80 S ribosomes from YJF24 with NAC carrying a C-terminal His tag on βNAC in the presence of AMAS. The overall cross-linking pattern was not changed by the presence of the His tag (data not shown). After purification of the cross-linking products the N terminus of the cross-linked 15-kDa protein was determined by Edman degradation. Apart from the N terminus of βNAC, cycles 2 to 8 resulted in the sequence GLKDVVT, which represents the N terminus of Rpl31. Unfortunately, the cross-linking product to the 25-kDa protein was not susceptible to Edman degradation, most likely due to N-terminal blockage. Rpl31 is only present in eukaryotic and archaeal ribosomes and has no homolog in eubacteria. It is located between Rpl17 and Rpl19 and is in close proximity to the tunnel exit site (see Fig. 6).

FIGURE 6.

Cryo-EM-based molecular model of the exit site of the large ribosomal subunit of S. cerevisiae showing the two universal adapter sites. a, the illustration was made using PyMOL (DeLano Scientific LLC) based on PDB code 1S1I. Ribosomal RNA is shown in gray. Ribosomal proteins surrounding the exit site are highlighted in different colors. Yellow circles highlight the two universal adapter sites (UAS) on the opposite sites of the exit tunnel. b, the ribosome is rotated by 90° toward the viewer and a detailed view of the tunnel exit is shown. Colors are as described in a. c, current working model for NACs interaction with the ribosome. Rpls are colored as described in a. For a clearer view Rpl26 and RpL19 were omitted.

Because no antibody directed against yeast Rpl31 was available we choose a different approach to confirm the identity of Rpl31 as the 15-kDa cross-linking partner of βNAC. We fused an HA tag to the N or C terminus of Rpl31B and expressed the respective gene under control of a GAL1 promoter in YJF24. Incorporation of Rpl31-HA into ribosomes was confirmed in a Western blot with an antibody against HA (data not shown). We repeated the cross-linking experiments with ribosomes from theses strains together with purified NAC and AMAS. Samples were treated as described above and after Western blotting analyzed with an antibody against βNAC or HA, respectively. Independent of whether the HA tag was fused to the N or C terminus of Rpl31, in both cases the cross-link to Rpl31 now appeared as a double band when detected with an antibody against βNAC (Fig. 3, lanes 2 and 4). The observed double band reflects the fact that the ribosomes now represented a mixed population either carrying the WT copy of Rpl31 or the HA-tagged version. When an identical sample was analyzed with an antibody against HA only the upper band resulted in a signal (see Fig. 3, lane 5) indicating that the cross-link between βNAC and the 15-kDa protein was indeed caused by a contact between βNAC and Rpl31.

FIGURE 3.

Cross-linking of NAC to 80 S ribosomes carrying an HA-tagged version of Rpl31. NAC was incubated with AMAS and 80 S ribosomes purified from yeast NAC knock-out strains expressing additional Rpl31 carrying an HA tag either on the N terminus (HA-Rpl31) or C terminus (Rpl31-HA). Ribosomes were pelleted and analyzed by Western blotting using an antibody directed against βNAC. In both cases the cross-link to the ribosomal protein with the apparent molecular mass of 15 kDa now appeared as a double band (lanes 2 and 4). Only the upper of the two bands also reacted with an antibody against the HA tag (lane 5).

Surface-exposed Peptides in Rpl31 Are Potential Candidates for NAC Binding

A major problem in identifying binding regions in ribosomal proteins is the fact that most ribosomal proteins are not soluble when purified out of the context of the ribosome. Therefore we screened 50 cellulose-bound peptides representing the amino acid sequence of Rpl31 for binding of full-length NAC versus NAC-(βNACΔ1–11) to identify possible interaction partners of the N-terminal sequence of βNAC in Rpl31. The array contained pentadecamer peptides with a 13-residue overlap. Full-length NAC discriminated between distinct amino acid side chains, because it bound only to a subset of peptides. In particular it bound to peptides representing Val²³ to Lys⁴¹ in Rpl31 (see supplemental Fig. S5, a and b) where no binding was detected with NAC-(βNACΔ1–11). A second interaction site at Arg⁷⁸ to Ser⁹¹ turned out to be unspecific because it was not only also observed with NAC-(βNACΔ1–11) but also with the antibody control, suggesting that the positive signal at these spots was due to an epitope-independent unspecific binding of the antibody used for detection. When we searched for the localization of the peptide Val²³ to Lys⁴¹ in Rpl31 in the context of the yeast ribosome we found that it is surface accessible and therefore a likely candidate for interaction with other proteins (see supplemental Fig. S5c).

Rpl31 Is Critical for NACs Association with the Ribosome

After having identified the contact between βNAC and Rpl31 we wondered whether this contact was critical for NACs association with the ribosome. To test this hypothesis we prepared S100 extracts from a ΔRpl31A/B strain and its maternal WT strain MH272-3fα (39). Both extracts were adjusted according to their general protein content judged by their absorption at 280 nm. Based on these criteria comparable amounts of S100 extracts were sedimented through 1 m sucrose cushions at different salt concentrations. Corresponding samples of the total, the supernatant, and the ribosomal pellet were separated on NuPAGE gels and blotted onto nitrocellulose. The latter was decorated with antibodies directed against βNAC, Rps8, Rpl35, and Rpl25. As described before by the Rospert group (39), we find a clearly reduced level of large ribosomal subunits in the ΔRpl31A/B strain (compare the signals for Rpl25 and Rpl35 in lanes 1 and 4 or lanes 3 and 6 in Fig. 4). The level of small ribosomal subunits in the WT and ΔRpl31A/B strain are rather similar resulting in an accumulation of 40 S subunits, as also observed by Peisker et al. (39). More importantly, in the WT strain roughly 50% of the cellular NAC cosediments with the ribosomes through the sucrose cushion at low salt conditions (100 mm KOAc), whereas in the ΔRpl31A/B strain only a very minor fraction of cellular NAC is detected in the pellet fraction. When pelleting ribosomes through 200 mm KOAc, a considerable fraction of the cellular NAC still sedimented with the ribosomes in the WT extract, whereas no NAC was detected in the pellet fraction of the ΔRpl31A/B strain. At 300 mm KOAc, some NAC still sedimented with ribosomes in the WT situation. This result indicates that in the ΔRpl31A/B strain where Rpl25 is still present in the ribosome and does not seem to be underrepresented (compare signals for Rpl25 and Rpl35), the NAC interaction with the ribosome is severely affected. This result demonstrates indeed that the most important contact for NAC association with the ribosome is established at least in part by a direct interaction with Rpl31 and that the observed cross-link is not just caused by close proximity.

FIGURE 4.

Association of NAC with 80 S ribosomes from a ΔRpl31A/B strain. Ribosomes were sedimented from S100 extracts through sucrose cushions at different salt concentrations. Equivalent samples of the total (T), the supernatant (S), and the pellet (P) were analyzed by Western blotting using anti-βNAC antibodies. Sedimentation of equal amounts of ribosomes was monitored by antibodies against Rpl25 and Rpl35 for the large ribosomal subunit and antibodies against Rps8 for the small ribosomal subunit.

αNAC Contacts Several Ribosomal Proteins Including Rpl17

Because the cross-linking approach with AMAS resulted in at least two contact sites for βNAC and in the ΔRpl31A/B extract a minor fraction of NAC still sedimented with ribosomes, we wondered whether αNAC also contributes to the affinity of NAC for the ribosome by contacting ribosomal proteins. We therefore probed aliquots from the cross-linking assays with AMAS for interaction partners of αNAC. Fig. 5 shows that upon addition of ribosomes αNAC apparently contacts several ribosomal proteins (lane 3). A comparison of the molecular masses of the proteins surrounding the tunnel exit site revealed that Rpl17 and Rpl19, which flank Rpl31 on either side, are likely candidates for these cross-linking partners. Decoration of the Western blot with a polyclonal antiserum directed against Rpl17 showed an additional cross-link (see Fig. 5, lane 6, asterisk) in the presence of NAC in contrast to incubation of ribosomes and cross-linker alone (see Fig. 5, lane 5). A cross-link of the same size was detected with an antiserum directed against αNAC, suggesting that αNAC contacts Rpl17.

FIGURE 5.

Immunodetection of cross-linking products of αNAC. NAC was incubated with 80 S ribosomes from a yeast NAC knock-out strain in the presence of AMAS. Ribosomes were pelleted and analyzed by Western blotting. Cross-linking products were detected with antibodies directed against αNAC (lanes 1–3) or Rpl17 (lanes 4–6). The two cross-links of the same apparent molecular mass are labeled with an asterisk.

DISCUSSION

In the yeast S. cerevisiae, NAC is quantitatively associated with the ribosome (35, 46). Here we show that the pivotal contact for NACs robust interaction with the ribosome is established via Rpl31, a ribosomal protein in close proximity to the tunnel exit site. Rpl31 is exclusively present in eukaryotes and archaea. The contact to Rpl31 is mediated by the extreme N terminus of βNAC, which most likely exhibits an α-helical secondary structure. This N terminus contains several absolutely and some highly conserved amino acid residues and is enriched in positively charged residues that face one side of the α-helix. Exchange of these conserved residues (⁶EKL⁸ or ¹⁰KLQ¹²) against alanines severely reduced NACs association with ribosomes, whereas the predicted secondary structure of the respective polypeptide region was unaltered. Deletion of the first 11 amino acids of βNAC even abolished ribosome association. The ability to robustly interact with the ribosome could be transferred to the MBP of E. coli (which normally does not interact with ribosomes) by fusing the N-terminal 23 amino acids of yeast βNAC to MBP.

Notably, the N-terminal 23 amino acids of βNAC, which are sufficient to mediate ribosome binding, do not contain the recently identified “NAC signature” (²⁴RRK(X)_nKK³⁰), which was suggested to be responsible for the NAC interaction with the ribosome via Rpl25 (L23 in eubacteria) (38). In contrast to the binding motif described here, which is part of the first α-helix of yeast βNAC, the previously observed “NAC signature” is located in an intrinsically disordered region between two α-helices at the N terminus of βNAC. The previously observed loss of ribosome interaction upon replacement of ²⁴RRK²⁶ or ²⁹KK³⁰ by alanines could also have been caused by overall changes in the local secondary structure as predicted by the program PHD (see supplemental Fig. S3) leading to the changed behavior. In addition, Wegrzyn et al. (38) used a fusion protein composed of yeast α and β NAC in most of their experiments and their cross-linker spanned a distance (10 Å) more than twice as long as the one used in this report (4.4 Å), possibly explaining the differing results.

During our studies we realized that yeast NAC artificially interacts with E. coli ribosomes when still carrying a His tag on the N terminus of αNAC, whereas this interaction is lost when the His tag was removed (see supplemental Fig. S6). Furthermore, our antibody against Rpl25 did not detect any additional cross-link when incubating ribosomes with AMAS in the presence or absence of NAC. Finally, whereas in a WT extract about 50% of NAC sedimented with the ribosomes through a sucrose cushion at low salt conditions, sedimentation of NAC with ribosomes in a ΔRpl31A/B extract was strongly reduced. Raising the salt concentration to 200 mm completely abolished NACs cosedimentation with ΔRpl31A/B ribosomes, whereas a considerable amount of NAC still sedimented with WT ribosomes. Furthermore, this result indicates that the interaction with Rpl31 is not the only contact NAC establishes upon ribosome binding. In addition to αNAC contacting Rpl17, the tendency of NAC to bind nucleic acids (15) is likely to contribute to its robust interaction with the ribosome.

The finding that Rpl31 is the critical anchor point for NAC on the ribosome offers a possible explanation for the observation that NAC prevents mistargeting of translating ribosomes to the ER-based translocon in vitro (26–28). Halic identified Rpl31 as the contact site of the SR in the so-called “docking complex” composed of the translating ribosome, SRP and SR (47). In particular, the structure revealed that SRβ contacts the 7 S RNA of SRP, helix 99 of 25 S rRNA, and concomitantly the ribosomal protein Rpl31. At the same time the NG domain of SRP54 changes its location and thereby frees the main contact site for interaction with the translocon. A competition of SR and NAC for the same binding site could be an explanation for NACs influence on ribosome nascent chain complex binding to the ER-based translocon in this context.

Initially, NAC was identified as a very early interaction partner of nascent chains on eukaryotic ribosomes (14). Surprisingly, strong cross-links to α and β NAC were already obtained when the first cross-linker-carrying lysine in the nascent chain was positioned only 27 amino acids away from the peptidyl transferase center (24). An amino acid at this position would be expected to still be enclosed by the ribosomal tunnel. On the other hand, Bhushan et al. (48) just recently presented a model based on single-particle cryoelectron microscopy reconstructions of eukaryotic 80 S ribosomes containing different nascent chains. According to this model the lower part of the ribosomal tunnel opens up to a compartment that they name “vestibule” where tertiary protein folding might be already occurring. Conversely, one could also speculate that such a vestibule would allow an early contact of ribosome-associated factors to the nascent chain.

The Rospert group recently showed that the ribosome-associated chaperone RAC also binds in the vicinity of Rpl31. However, RACs ribosome binding is most likely qualitatively different and may also include participation of the ribosomal RNA (39).

Taking into account the two identified binding sites for the different NAC subunits, our current working model for NACs interaction with the ribosome is illustrated in Fig. 6c. Although the N terminus of βNAC contacts Rpl31, αNAC contacts Rpl17. It is quite possible that this is a contact to the central NAC domain, which is structurally conserved from archaea to humans. By anchoring itself mainly to a ribosomal protein that is not directly localized at the rim of the exit tunnel, NAC could still remain associated with the ribosome even when other factors engage the exit site nearby. Moreover, the identification of a contact between αNAC and Rpl17 offers the first evidence of how archael NAC, which does not contain a β-subunit, could interact with the ribosome.

Apparently, the additional ribosome-associated factors in eukaryotes have developed different strategies to interact with the ribosomal surface surrounding the exit site. Whereas Rpl25/Rpl35 (L23/L29 in eubacteria) was identified as the first universal binding site on the ribosome providing contact sites for trigger factor, SRP, and the translocon, as well as YidC/Oxa1, we now propose the existence of a second universal adapter site on the eukaryotic ribosome established by ribosomal proteins Rpl31/Rpl17, located on the opposite site of the tunnel exit and serving as an interaction site for NAC, RAC, and the SRP receptor (see Fig. 6). The second docking site seems not to be restricted to eukaryotes. In E. coli the same ribosomal surface area that is formed by Rpl17 and Rpl31 in yeast is built up by L22 (the homolog of Rpl17) and L32/L17 (unique to eubacteria). Indeed, the co-translationally acting enzyme peptide deformylase, which removes the formyl group from the starter methionine of the nascent chain is anchored via a basic C-terminal α-helix between L22 and L32 (12). Future studies will have to assess what molecular details determine which factor binds to this second universal docking site or whether they can simultaneously engage the ribosome.