OUP user menu

Identifying Pex21p as a protein that specifically interacts with yeast seryl-tRNA synthetase

Sanda Rocak, Irena Landeka, Ivana Weygand-Durasevic
DOI: http://dx.doi.org/10.1111/j.1574-6968.2002.tb11331.x 101-106 First published online: 1 August 2002


The interaction of Saccharomyces cerevisiae seryl-tRNA synthetase (SerRS) with peroxin Pex21p was identified in a two-hybrid screen with SerRS as bait. This was confirmed by an in vitro binding assay with truncated Pex21p fused to glutathione S-transferase. Furthermore, purified Pex21p acts as an activator of yeast seryl-tRNA synthetase in aminoacylation in vitro, revealing the functional significance of the Pex21p–SerRS interaction. Pex21p is a protein involved in the peroxisome biogenesis [Purdue, P.E., Yang, X. and Lazarow, P.B., J. Cell Biol. 143 (1998) 1859–1869]. Since eukaryotic aminoacyl-tRNA synthetases are known to participate in assembles with other synthetases and non-synthetase proteins, we propose that this unusual interaction reflects another function of the peroxin.

  • Aminoacyl-tRNA synthetase
  • Protein–protein interactions
  • Protein synthesis
  • two-hybrid system

1 Introduction

Aminoacyl-tRNA synthetases (aaRSs) catalyze the aminoacylation of tRNAs by their cognate amino acid. The accuracy of this process is a prerequisite for proper translation of the genetic code [1,2]. aaRSs in general do not require protein cofactors for activity and specificity of substrate recognition. Instead, idiosyncratic domains are attached or inserted in the conserved class defining catalytic core [35], and are responsible for binding and recognition of cognate tRNAs [1,2]. However, in higher eukaryotes, these enzymes are known to participate in complex formation with other synthetases and non-synthetase proteins [612]. Extra peptides, acquired during evolution, are often added as extensions to the N- or C-termini of the enzymes, [6,13,14], and link together different synthetases into a complex[11]. In humans such a complex is composed of class I and class II synthetases that are monomers (IleRS, LeuRS, MetRS, GlnRS, ArgRS) or dimers (LysRS, AspRS), of a bifunctional enzyme (GluProRS) and the three non-synthetase proteins of molecular mass (MW) of 18, 38 and 43 kDa [6,15]. Interactive domains of these proteins and the topology of the complex have been studied by two-hybrid analysis [9,12]. p18 and p43 have been characterized as proteins that optimize normal function of synthetases [16,17] or help the association of the multisynthetase complex with EF-1[18], while p38 was defined as a ubiquitous component of the complex, serving as a core protein for the anchorage of the other proteins[12]. Although in yeast such large complexes have not been found, a smaller complex containing two class I synthetases, MetRS and GluRS, and the protein Arc1p, was identified [1921], confirming the existence of an evolutionary intermediate toward higher-order organization of aaRSs. Arc1p is the first example of a protein that is not itself an enzyme but is required for the efficient aminoacylation in eukaryotic cells. It enhances the binding affinity of cognate tRNAs and consequently increases aminoacylation specificity. Furthermore, it seems that the interaction with Arc1p provides a means of regulating GluRS and MetRS subcellular distribution[22].

Seryl-tRNA synthetase is a class II enzyme that is not known to interact in vivo with other cellular components. It has been reported, however, that yeast SerRS copurifies with IleRS and TyrRS on Bio-Gel A-5M chromatography columns[23], giving an indication of the involvement of SerRS in forming larger synthetase complexes. Furthermore, these particular data revealed the existence of these synthetases both as monomeric enzymes and as parts of the high-molecular mass complex as well. These findings might be biologically relevant, since yeast SerRS, like all eukaryotic cytosolic seryl-tRNA synthetases, contains basic C-terminal extension which is dispensable for cell viability, but influences the enzyme's conformation and its substrate recognition properties [14,24]. In order to find out whether SerRS from Saccharomyces cerevisiae specifically interacts with other cell components in vivo, we screened a yeast cDNA library using the two-hybrid system. Five positives were selected and sequenced. We identified yeast peroxin Pex21p as SerRS interacting protein, which was confirmed by an in vitro binding assay using truncated Pex21p fused to glutathione S-transferase (GST). Here we discuss the potential relevance of this unusual interaction.

2 Materials and methods

2.1 Yeast strains

The S. cerevisiae two-hybrid reporter strain L40 (MAT a trp1 leu2 his3 LYS2::lexA-HIS3 URA3:lexA-lacZ)[25] was a gift from Prof. Igor Stagljar, Institute of Veterinary Biochemistry, University Zurich-Irchel, Switzerland. The S. cerevisiae strain S2088α (MATαura3-52 trp1 lys2-801 leu2Δ1 his3-Δ200 pep4::HIS3 prb-Δ1.6R can1 GAL) was used for preparation of protein extracts as described[14].

2.2 Plasmid constructions

Yeast full-length SES1 gene was cloned in the yeast bait plasmid pAB151. Resulting construct pAB151SES1, encoding LexA–SerRS fusion protein, did not activate reporter genes expression by itself. The recombinant plasmid expressing GST–Pex21Δ63 fusion protein was generated using pGEX-6P-3 system (Amersham, Pharmacia). Plasmid pBTM116DNA2, which was used as a non-specific bait, was a gift from Prof. Igor Stagljar[26]. For purification of full-length Pex21p, GST–Pex21p fusion was expressed from pGEX-6P-3pex21.

2.3 Yeast two-hybrid screening

The transformant cells carrying recombinant bait plasmid were subsequently transformed with yeast cDNA library cloned in the yeast expression vector pACT (Clontech). Of approximately 1.8×106 colonies screened, those growing on selective medium lacking leucine, triptophane and histidine were isolated and tested for β-galactosidase activity as described in protocols from Clontech.

2.4 GST-binding assay

GST and GST–Pex21pΔ63 (Pex21p lacking 63 N-terminal amino acids) were produced in Escherichia coli BL21(DE3). Aliquots of total protein extracts were incubated for 15 min at room temperature (RT) with 25 μl of GST–Sepharose 4B (50% slurry, Amersham, Pharmacia) suspended in phosphate buffer saline (PBS), pH 7.4. After extensive washing in PBS, beads were washed in the binding buffer containing 50 mM HEPES, pH 7.9, 100 mM KCl, 5 mM MgCl2, 1 mg ml−1 bovine serum albumin (BSA), 0.2 mM phenylmethylsulfonyl fluoride and 2 mM dithiothreitol (DTT). Pure SerRS or crude yeast extract was then applied to the precharged beads and incubated for 30 min at 4°C in the binding buffer. The beads were then washed four times with the same buffer and pellets were analyzed by SDS–PAGE. Recovered SerRS was visualized by immunoblotting using polyclonal antibodies against yeast SerRS[27].

2.5 Purification of Pex21p

Fresh overnight culture of E. coli BL21 transformed with pGEX-6P-3pex21 was diluted 1:100 in 1 l Luria–Bertani containing ampicilline (100 μg ml−1). Cultures were grown at 30°C to OD600∼1.3. IPTG was added to 1 mM final concentration and cultures were incubated for additional 4 h at 30°C. Bacteria were pelleted and resuspended in 20 ml PBS pH 7.4. The cells were then lysed on ice by mild sonication and centrifuged at 10 000×g for 15 min at 4°C. Supernatant was added to 1 ml of GST–Sepharose beads suspended in PBS pH 7.4, and incubated for 30 min at RT. Beads were then washed three times with 10 bed volumes of PBS and once with 10 bed volumes of protease cleavage buffer (50 mM Tris–HCl, pH 7.0, 150 mM NaCl, 1 mM EDTA, 1 mM DTT). PreScission Protease (30 U, Pharmacia) was added to bound GST–Pex21p and incubated for 4 h at 4°C. Eluted proteins were collected in volume of 1 ml. For analyses of purified Pex21p, aliquots were boiled in 5× sample buffer and loaded on SDS–polyacrylamide gels. Proteins were visualized by Coomassie blue staining.

2.6 Aminoacylation assay

Aminoacylation was done at 30°C as previously described[14] in reaction mixtures containing 50 mM Tris–HCl, pH 7.5, 15 mM MgCl2, 4 mM DTT, 5 mM ATP, 8 mg ml−1 unfractionated brewer's yeast tRNA (Roche), which corresponds to approximately 14 mM tRNASer, and 125 mM 14C-labeled serine. SerRS concentration was 54 nM. SerRS was preincubated for 30 min at 30°C with or without Pex21p and then added to respective reaction mixtures to initiate reaction. Preincubation with BSA was done as a control. All values represent the average of three independent determinations, which varied by less than 10%.

3 Results

3.1 Construction of LexA–SerRS hybrid protein

The bait plasmid was constructed by fusing full-length gene encoding yeast SerRS to the 3′-end of the LexA in the yeast expression vector pAB151. The hybrid protein was expressed in yeast reporter strain L40, which contains yeast HIS3 and bacterial lacZ genes under the control of synthetic promoters bearing LexA-binding sites. Stable expression was determined by immunoblotting using anti-LexA antibody (Clontech).

3.2 Screening yeast cDNA library

The yeast cDNA library fused to the gene for the transcription activator GAL4 activation domain (GAL4ad) was introduced into yeast cells expressing bait protein LexA–SerRS. If a GAL4ad-fused protein interacts with the bait protein, it will activate expression of the HIS3 reporter gene and result in colony growth on media lacking histidine. From 1.75×106 independent clones screened, 720 grew on the selective media (Trp, Leu, His). The colonies larger than 2 mm in diameter were selected, supposing that they express proteins that form interactions of different strengths. Of these, 35 colonies expressed β-galactosidase activity as well. The library plasmids carrying respective clones were named pACTpoz1–pACTpoz35. The plasmids were isolated from selected yeast cells and introduced into E. coli HB101, the host that selects for library plasmids due to deficiency in leucine synthesis that can be complemented by expression of yeast LEU2 gene. Isolated plasmids bearing potentially interactive clones were subjected to extensive restriction analyses. Representative clones were retransformed in yeast strain L40 carrying three different constructs (pAB151 expressing LexA, pAB151SES1 expressing LexA–SerRS, pAB151DNA2 expressing LexA–DNA2) and screened for interactions. False positive results were not dependent on the presence of LexA–SerRS, even in the absence of SerRS or in the presence of another bait they activated transcription. Plasmids activating both reporter genes only in the presence of LexA–SerRS were selected for sequencing (Fig. 1).

Figure 1

SerRS specifically interacts with Pex21p in yeast two-hybrid system. A: The L40 reporter strain, expressing LexA and GAL4ad or fusion proteins, was assayed for histidine prototrophy. Filter-lift assay was used to detect β-galactosidase activity. LexA-fusions contain full-length SerRS (bait) and full-length DNA2 (false bait). The results are shown with three independent transformants for each transformation. True positive results activate both HIS3 and LacZ expression, in contrast to false positives that are not dependent on presence of LexA–SerRS. B: Quantification of β-galactosidase activity for the SerRS–Pex21p interaction using liquid ONPG-assay. LexA and LexA–SerRS are used as a control. The bars indicating SerRS–Pex21p interaction are marked as Pex21p. Numbers in brackets represent starting amino acid of Pex21p.

3.3 Identification of the interactive protein

All five positive clones represented yeast ORF YGR239C, which is unambiguously assigned to pex21, the gene on the chromosome VII encoding Pex21, a cytoplasmic protein involved in peroxisome biogenesis (peroxin). Sequence analyses revealed that five identified clones differ in their 5′-ends, giving rise to expression of fusion polypeptides of different lengths. The longest clone (poz23) encodes fusion proteins starting with amino acid 2 of Pex21p. The others are shorter, starting with amino acid 3 (poz13), 8 (poz12), 26 (poz20) and 63 (poz11). The interaction was confirmed in vitro using GST pull-down. The shortest isolated clone, named poz11, was fused to GST and expressed in E. coli BL21, same as GST alone. Proteins were bound to GST-beads and incubated with pure SerRS or total protein extract from S. cerevisiae S2088α. Bound proteins were analyzed on SDS–PAGE followed by Western blotting using primary anti-SerRS antibody. Pex21 specifically binds a 53-kDa protein, which corresponds to the size of SerRS subunit. The same kind of protein complex was detected both with pure SerRS and with crude yeast protein extract (Fig. 2).

Figure 2

Pex21p interacts with SerRS in vitro. Bacterial extracts from E. coli BL21 (DE3) cells overexpressing GST or GST-tagged Pex21pΔ63, were bound to GST-beads and incubated with 1.5 μg of pure SerRS or 500 μg of yeast crude extract. After extensive washing, bound fractions were resolved by SDS–PAGE, and subjected to: (A) Coomassie blue staining, showing expression (lanes 2 and 4), binding and purification (lanes 3 and 5) of GST and GST–Pex21pΔ63 fusions on GST-beads, and (B) immunoblotting using polyclonal anti-SerRS antibody to detect SerRS in bound fractions after incubation with pure SerRS (lane 3) or crude yeast extract (lane 4).

Interestingly, although SerRS is a homodimer, we were not able to detect expected interaction between SerRS subunits. Furthermore, our trials to use a different bait plasmid, pGBT9 (Clontech), for the expression of the hybrid protein GAL4–SerRS, were also unsuccessful. It confirms a previous observation that the efficiency of the two-hybrid approach is greatly influenced by the polarity of the protein fusions as much as the type of the fusion[12]. It is obvious that the geometry of the fusion and conformation of the hybrid protein differ in GAL4 and LexA-system, which directly reflects the success of two-hybrid approach.

3.4 Pex21p activates yeast seryl-tRNA synthetase

To test the functional significance of the Pex21p–SerRS interaction, aminoacylation assay was done with and without preincubation of SerRS with Pex21p. In both cases, charging of tRNASer by SerRS was increased. Activation was more pronounced when SerRS was preincubated with Pex21p. As a control, SerRS was preincubated with BSA, to exclude the effect of stabilization with inert protein. BSA did not show any significant effect on activity of serylation (Fig. 3).

Figure 3

Pex21p activates yeast seryl-tRNA synthetase. A: Kinetic analyses of SerRS aminoacylation reaction at different concentrations of Pex21p. Amounts of seryl-tRNASer formed were plotted against the molar ratio of Pex21p and SerRS. B: Time-course of the activity of SerRS with and without Pex21p. Molar ratio of Pex21p and SerRS was 4:1. Preincubation with Pex21p leads to higher activation of SerRS.

4 Discussion

In order to elucidate the specificity of serylation in yeast, we have used the benefits of the yeast two-hybrid system[28] to search for the proteins that interact with SerRS in vivo. Screening yeast cDNA library with SerRS fusion protein as bait resulted in identification of only one interacting protein, represented as five independent positive clones of different lengths. Isolated cDNAs were completely homologous to the sequence encoding yeast peroxin Pex21p.

Pex21p is a protein of 288 amino acid residues with a calculated MW of 33 045 Da. It is a member of yeast peroxin family and is engaged in peroxisomal biogenesis. Together with Pex18p (which shares 23% overall identity), Pex21p is required for peroxisomal localization of Pex7 and therefore is important for protein targeting via the peroxisomal targeting signal 2 (PTS2) pathway. The sequence alignment of Pex21p with yeast Pex18p, Yarovia lypolitica Pex20p and mammalian Pex5pL revealed a conserved motif that represents a common Pex7-binding site[29]. In Pex18p and Pex21p, this conserved region is in the C-terminal part of the protein. Interestingly, Pex18p was not identified as a SerRS-interacting protein in the two-hybrid screen we performed. This may indirectly lead to the assumption that the C-terminal domain of Pex21p is not responsible for interaction with SerRS, suggesting that the interactive domain may include the middle part of Pex21p. In vivo, the lack of Pex21p does not significantly perturb peroxisomal biogenesis since it is redundant to Pex18p[30]. Different contributions of two peroxins to PTS2 targeting, Pex18p being more involved, leaves a possibility that Pex21p exhibits additional function in the cell. In support of such an idea is a recent finding that yeast TyrRS interacts with Knr4, a protein involved in cell wall synthesis[31]. Similarly, p43 – which is shown to be identical to proEMAPII[32]– is a multifunctional protein that assists aminoacylation in normal cells and releases the functional cytokine upon apoptosis[16]. Likewise, heat shock protein 90, which is a molecular chaperone responsible for protein folding and maturatin in vivo, interacts with human glutamyl-prolyl-tRNA synthetase and mediates protein–protein interactions during the association of several human synthetases[33].

In the context of interaction with SerRS, it can be proposed that Pex21p acts as a cofactor that enhances the efficiency of serylation performed by SerRS. A tempting hypothesis is that Pex21p acts as a functional analogue of Arc1p in stabilizing the interaction between SerRS and its cognate tRNA. Since Arc1p binds preferentially to a subset of tRNA species, forming only a weak interaction with tRNASer[21], one can speculate that a different auxiliary protein is required for increasing the specificity and stability of SerRS:tRNASer complex formation. In this respect, such a protein–protein interaction can be also seen as one of numerous quality control mechanisms that ensure fidelity of translation [2,34]. Although many peroxins are intraperoxisomal or the parts of peroxisomal membrane, the experiments of Purdue et al.[30] revealed a primarily cytosolic distribution of Pex21p. However, the analysis of its amino acid sequence by PSORT II (http://www.psort.nibb.ac.jp) predicts a potential nuclear localization signal, also found in Arc1p. This opens the possibility that Pex21p functions in the nucleus as well, maybe as a part of an aminoacylation-dependent nuclear tRNA export pathway in yeast [3537]. The other scenario may include auxiliary function of seryl-tRNA synthetase. That would not be the first example in which a member of the synthetase family has been shown to function in an unexpected manner.

We have recently started mapping the SerRS/Pex21p interactive domains. Until this is done, we can only speculate about the regions involved in the protein association. It is likely that the N-terminal region of Pex21p is not involved in the interaction with SerRS, since the shortest clone isolated encodes truncated peroxin beginning with amino acid 63. The same clone was satisfactory used for in vitro binding assay (Fig. 2), which confirms the interaction detected in vivo. Purification of Pex21p and studying its effect on SerRS activity revealed that this unusual interaction reflects another function of the peroxin, related to the increased efficiency of serylation.


We are indebted to Igor Stagljar (Institute of Veterinary Biochemistry, University Zurich-Irchel, Switzerland) for his generous help with the two-hybrid system and for providing the strains and plasmids. This work was supported by grants from International Center for Genetic Engineering and Biotechnology, Trieste, the Ministry of Science and Technology of the Republic of Croatia, and National Institutes of Health (NIH/FIRCA).


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
View Abstract