Structural dissection of human translation elongation factor 1 B ( eEF 1 B ) : expression of full-length protein and its truncated forms

Aim. To gain more insights into properties of the human translation elongation factor eEF1B and its interaction with partners we intended to produce the full-length protein and its truncated forms. Methods. cDNAs encoding truncated forms of eEF1B were generated by PCR amplification with respective primers and cloned into vectors providing polyhistidine, glutathione S-transferase or maltose binding protein tags. The recombinant proteins were expressed in Escherichia coli and purified by affinity chromatography. An aggregation state of the proteins was analyzed by analytical gel filtration. Results. The expression, purification and storage conditions for the full-length recombinant His-eEF1B were optimized. Several truncated forms of eEF1B were also expressed and purified to homogeneity. Two short variants of C-terminal domain comprising amino acids 264–437 or 229–437 were obtained in monomeric state. Two short variants of N-terminal domain comprising amino acids 1–33 or 1–228, fused with glutathione S-transferase, were obtained and estimated to be dimers by gel filtration. The mutants of N-terminal domain comprising amino acids 1–92 or 1–163, fused with maltose binding protein, were obtained as soluble high molecular weight aggregates only. Conclusions. The purified recombinant HiseEF1B and several truncated forms of the protein were obtained and characterized. These protein variants will be used for further studies on the protein-protein interaction.


Introduction.
Elongation factor 1 (eEF1) is one of the major players during the elongation cycle of eukaryotic protein synthesis [1].eEF1 consists of two functionally distinct parts: eEF1A and eEF1B.eEF1A is a G-protein ensuring the delivery of aminoacylated tRNA to the ribosome.eEF1B acts as a guanine nucleotide exchange factor which catalyzes the conversion of inactive eEF1A* GDP into active eEF1A*GTP bound form.In higher eukaryotes, the eEF1B complex consists of three proteins eEF1Ba, eEF1Bb and eEF1Bg.
Both eEF1Ba and eEF1Bb are catalytic guanine nucleotide exchange subunits which interact with eEF1Bg, a structural component of eEF1B complex.Beside the presence in the complex a functional role of eEF1Bg is not clear.This subunit itself does not possess any exchange activity, but eEF1Bg isolated from Artemia salina enhanced the catalytic activity of eEF1Ba in vitro [2].eEF1Bg was found to interact with tubulin and based on this observation the role of this protein in anchoring eEF1B complex to the membranes and/or cytoskeleton was proposed [2].Moreover, eEF1Bg was shown to bind and bundle specifically the keratin intermediate filaments [3].The interaction between eEF1Bg and keratin regulates protein synthesis in the epithelial cells, but the physiological relevance of this bidirectional relationship remains to be defined [3].eEF1Bg was suggested to possess nonspecific RNA binding properties [4].eEF1Bg subunit might be also a regulatory element within the eEF1B complex.This subunit was found to be a substrate for the cell cycle protein kinase CDK1/cyclineB [5].The phosphorylation of eEF1Bg seems to be necessary for effective translation of valine-rich proteins [6].The overexpression of eEF1Bg subunit was observed in some gastric and esophageal carcinomas [7,8].Nuclear localization of eEF1Bg in both normal and cancer tissues suggests an unknown nucleus-specific role in human cells [9] eEF1Bg consists of two independently folding structural domains connected by a lysine-rich linker [6,10].The C-terminal domain is a highly conserved exceptionally protease resistant domain which consists of five stranded, anti-parallel b-sheets surrounded by five a-helices [10].A functional significance of this domain remains to be deciphered.By contrast, the N-terminal domain of eEF1Bg is homologous to Theta class glutathione S-transferases (GST) [11] and interacts with eEF1Ba [2].The crystallographic structure of yeast eEF1Bg N-terminal domain was solved and found to be very similar to those of GST enzymes [12].However, the activity of this domain toward the GST model substrate 1-chloro-2,4-dinitrobenzen (CDNB) was not detected [12].
To gain more insights into the properties of eEF1Bg and its interaction with partners we elaborated a purification procedure of the full-length protein, designed and expressed its truncated forms.We demonstrate that the recombinant eEF1Bg can be purified as a monomer and stored in the conditions preventing its multimerization.We show that individual N-and C-terminal domains of this protein could be also expressed in bacteria and purified under the conditions mostly preventing the formation of aggregates.By contrast, the attempt to express and obtain isolated subdomains from the eEF1Bg N-terminal domain was unsuccessful.
Materials and methods.pET16b/eEF1Bg plasmid containing ORF of human eEF1Bg protein was kindly provided by Dr. G. Janssen (Leiden University, The Netherlands ).
Supplementary information can be found on Web site http://www.biopolymers.org.ua.
All constructs described above were verified by DNA sequencing.
Test of protein overexpression in Escherichia coli.Respective producing strains containing plasmid DNA of interest were grown in 100 ml of LB medium supplemented with appropriate antibiotic at 37 °C to an A 600 = = 0.5.The expression was induced by addition of 1 mM isopropyl 1-thio-b-D-galactopyranoside (IPTG) for 3-4 h.Total bacterial extract (TE) was prepared as follows: 1 ml of culture was harvested by centrifugation and bacterial pellet was dissolved in 150 µl of sample buffer Laemmli with 6 M urea (SBLU1´) and boiled for 10 min.Extract of soluble proteins (SE) was prepared from the rest of culture.Bacteria were pelleted, washed twice with extraction buffer (30 mM Tris-HCl, pH 8.0, 30 mM KCl, 0.1 mM EDTA, 10 % glycerol, 2 mM DTT) followed by centrifugation.Pellet was resuspended in 1 ml of extraction buffer supplemented with 1 mM phenylmethyl sulfonyl fluoride (PMSF), sonicated and centrifuged at 16000 g for 20 min (4 °C). 10 µl of supernatant was added to 200 µl SBLU1´, boiled for 10 min.Equal volumes of total and soluble extracts were used for analysis on SDS-PAGE.
Expression of full-length His-eEF1Bg and its truncated forms in bacteria and their purification by affinity chromatography.The protein encoded by pET16b/ eEF1Bg plasmid was expressed in E. coli BL21(DE3) pLysS grown on LB medium supplemented with ampicillin (100 µg/ml).Culture (0.3 L) was grown at 37 °C to an A 600 = 0.5, transferred at 20 °C and grown till A 600 = = 0.8.The expression was induced by addition of 0.8 mM IPTG for 16 h.Cells were washed twice with icecold extraction buffer (30 mM Tris-HCl, pH 7.5, 30 mM KCl, 0.1 mM EDTA, 10 % glycerol, 5 mM 2mercaptoethanol), resuspended in 8 ml of the same buffer containing 1 mM PMSF, and sonicated.All subsequent steps were conducted at 4 °C.After centrifugation at 18000 ´g for 30 min, the clear supernatant was recovered and concentration of NaCl was adjusted to 500 mM, imidazole pH 8.0 to 20 mM and Tween-20 to 0.01 %.The obtained solution was mixed with Ni-NTA resin (2 ml of 50 % slurry, «Qiagen», USA), preequilibrated with the same buffer, and incubated on the orbital shaker for 1.5 h.The resin was washed with buffer A (25 mM Tris-HCl, pH 7.5, 500 mM KCl, 20 mM imidazole, pH 8.0, 10 % glycerol, 5 mM 2-mercaptoethanol) and then with buffer A containing 1 M NaCl and packed into column.eEF1Bg was eluted from the column by 220 mM imidazole, pH 8.0 on buffer A. Fractions were collected and analyzed by SDS-PAGE.The purest fractions were combined and dialyzed against storage buffer (30 mM Tris-HCl, pH 7.5, 150 mM KCl, 55 % glycerol, 2 mM DTT).Protein was stored at -20 °C.Protein concentrations were determined using a calculated absorption coefficient: 1.74 A 280 units × mg -1 × cm -1 .
Protein denaturation-renaturation procedure.The truncated forms of eEF1Bg obtained as soluble aggregates were subjected to denaturation-renaturation pro-cedure.Two different approaches were tested: matrix assisted (Ni-NTA) denaturation-renaturation and denaturation of the protein in solution with subsequent renaturation by stepwise dialysis.For matrix assisted denaturation the protein was first immobilized on Ni-NTA and then denatured by addition of 10 mM Tris, 100 mM NaH 2 PO 4 pH 8.0, 8 M urea (pH 8).Matrix was washed by the same buffer with pH 6.3.Denatured protein was subsequently eluted by the same buffer with pH 5.9 (monomers) and pH 4.5 (multimers).For renaturation the eluted protein was applied onto Superose 6 HR 10/30 column equilibrated with 20 mM imidazole HCl, pH 7.5, 150 mM NaCl, 10 % glycerol, 10 mM 2-mercaptoethanol.This allowed also estimating the molecular weight of renatured proteins.
The renaturation of immobilized on Ni-NTA protein was also carried out in another manner [14] using linear 6-1 M urea gradient on buffer 20 mM Tris HCl pH 7.4, 500 mM NaCl, 20 % glycerol at low flow rate (0.8 ml/min) during 1.5 h.After renaturation the protein was eluted by 250 mM imidazole pH 8.0.
The denaturation of protein in solution was performed in 20 mM Tris HCl, pH 7.5, buffer with 8 M urea.The renaturation was carried out by stepwise dialysis in the buffer solutions containing 20 mM imidazole HCl pH 7.5, 150 NaCl, 10 % glycerol, 10 mM 2-mercaptoethanol and 6, 4, 2 M urea during 6-8 h in each solution.The last dialysis was performed in the same buffer solution without urea.The molecular weight of renatured protein was estimated by analytical gel filtration.
Analytical gel filtration of proteins.To assess the aggregation state of the purified proteins size-exclusion chromatography on a Superose 6 HR 10/30 column (24 ml, «GE Healthcare») was performed.The column was equilibrated with 25 mM imidazole-HCl, pH 7.5, 150 mM NaCl, 10 % glycerol, 5 mM 2-mercaptoethanol.All samples were loaded in 0.1 ml and run at 0.4 ml/min; the elution was monitored at 280 nm.Elution of particular protein from column was described in term of corresponding K av value.K av = (V e -V 0 )/(V t -V 0 ), where V e is the elution volume of individual protein, V 0 is the void volume of the column, and V t is its total bed volume.V 0 and V t were determined with blue dextran (2 MDa) and 2-mercaptoethanol (78,1 Da), respectively.Following standard proteins were used for column calibration: thyroglobulin -669 kDa, ferritin -450 kDa, catalase -232 kDa, alcohol dehydrogenase -150 kDa, bovine serum albumin -67 kDa, ovalbumin -45 kDa, chymotrypsinogen A -25 kDa, cytochrome c -12,4 kDa.The K av value of each protein standard (on the liner scale) was plotted against the corresponding molecular weight (on the logarithmic scale).The straight line that best fits the points on the graph was drawn and used for molecular weight determination of the protein of interest (Supplement Fig. S2).
Antibodies and Western Blot Analysis.Mouse monoclonal antibodies for the detection of His-tagged recombinant proteins («Roche», Germany) were used at 0.5 µg/ ml working concentration.Mouse anti-GST monoclonal antibody («Pierce», USA) was used at 0.5 µg/ml working concentration.Anti-mouse secondary antibodies conjugated with peroxidase («Sigma», USA) were used at 1:10000 working dilution.Proteins were separated on 10 or 12 % SDS-PAGE and then transferred onto 0.45 µm nitrocellulose membrane («Bio-Rad», USA) using Trans-Blot Turbo transfer system («Bio-Rad») according to the manufacture recommendations.The membranes were treated subsequently with primary and secondary antibodies diluted in PBS containing 5 % non-fat dry milk and 0.1 % Tween-20 followed by extensive washing with PBS-0.1 % Tween-20 solution.The immune complexes were detected by Immobilon Western Chemiluminiscent HRP Substrate («Millipore», USA) on ChemiDoc system («Bio-Rad»).
Results and Discussion.Domain structure of the eEF1Bg protein and design of its truncated forms.The eukaryotic translation elongation factor eEF1Bg (Fig. 1) consist of two rather hydrophobic domains of about 200 amino acids each (N-terminal or Domain 1 and Cterminal or Domain 2), which are linked through a highly polar central lysine-rich stretch of about 30-60 residues [3,6,10,15].In order to understand the interaction between eEF1Bg and its partners we prepared a set of truncated forms of this protein fused to different tags.Based on the multiple sequence alignment we prepared four N-terminal deletion mutants without 33, 92, 228 (C-terminal domain with lysine-rich linker) and 263 (C-terminal domain alone) amino acids (Fig. 1, Supplement Fig. S1).A polyhistidine sequence was attached either to the N-end or to the C-end of the protein.
To overcome possible problems with the solubility of the N-terminal truncated forms we also prepared the C-terminal truncated forms: amino acid fragments 1-33, 1-92, 1-163 and 1-228 (N-terminal domain) were attached to GST or MBP tags (Fig. 1).All deletion mutants were expressed in E. coli and, if soluble, purified by affinity chromatography as described below.
Expression and purification of the full-length eEF1Bg in the conditions preventing its multimerization.The expression of eEF1Bg with N-terminal His-tag was successfully carried out in E. coli.After overnight IPTG induction at 20 °C His-eEF1Bg was recovered in the fraction of soluble proteins and purified by affinity chromatography to an apparent homogeneity (Fig. 2).
The fractions eluted from the Ni-NTA column were analyzed by gel filtration on Superose 6 HR.We observed that the aggregation state of His-eEF1Bg depended on its concentration.When His-eEF1Bg was loaded on the column at 0.3 mg/ml, it came out as a single sharp peak (Fig. 3, A), whereas at 1.6 mg/ml some high molecular weight aggregates appeared (Fig. 3, B).At higher concentration the ratio of aggregates increased significantly (Fig. 3, C).Thus, to avoid the aggregation the Ni-NTA column should not be overloaded by the His-eEF1Bg protein.We estimated that the extract of soluble proteins obtained form as much as 300 ml of His-eEF1Bg producing culture may by loaded on 1 ml of packed Ni-NTA matrix at the conditions described in Material and methods.
The molecular weight of human recombinant His-eEF1Bg was estimated by gel filtration to be about 200 kDa that is 4 fold higher than the theoretical value of the monomer (52 kDa).A sedimentation equilibrium analysis of purified His-eEF1Bg demonstrated that this protein behaves in solution as a monomer with a small portion of multimers (V.F. Shalak, unpublished results).Thus, recombinant His-eEF1Bg is not a globular protein and most probably has extended shape in solution.
The large soluble aggregates of recombinant rabbit [16], B. mori [17] and multimers of the human eEF1Bg [10] were previously observed.The aggregates of rabbit eEF1Bg could be dissociated by 500 mM NaCl and at this salt concentration the molecular weight of eEF1Bg was estimated by gel-filtration to be about 140 kDa [16].The preparation of the B. mori eEF1Bg was separated by gel filtration chromatography at 200 mM KCl into two fractions: large aggregates migrated close to the column void volume, whereas another fraction migrated as a single peak of about 150 kDa [17].In contrast, in our experiments with His-eEF1Bg the high salt con- centration (up to 1 M NaCl) did not reduce the amount of large aggregates.The addition of Tween-20 or Triton X-100 to the buffer solution also had no effect (data not shown).This disagreement with the literature data can be explained by the presence of N-terminal polyhistidine sequence absent in the rabbit eEF1Bg [10] that may influence the protein folding or its tendency for aggregation.It's worth mentioning that we did not succeed to express the full-length eEF1Bg with C-terminal His-tag in E. coli.Thus, depending on the concentration, the recombinant His-eEF1Bg can form soluble and irreversible under physiological conditions high molecular weight aggregates.To prevent the aggregation the purified His-eEF1Bg protein was kept at -20 °C in the storage buffer with 55 % glycerol at 2 mg/ml concentration.
Expression, purification and aggregation state of the N-terminally truncated forms of eEF1Bg fused with C-terminal or N-terminal His-tag.The expression of the N-terminally truncated forms fused with C-terminal or N-terminal His-tag (Fig. 1) is presented in Fig. 4. The expression of all variants with C-terminal His-tag was unsatisfactory: full-length eEF1Bg-His, truncated proteins eEF1Bg (93-437)-His and eEF1Bg (264-437)-His were not detected at all by western-blot analysis of corresponding bacterial extracts.The deletion mutants eEF1Bg (34-437)-His and eEF1Bg (229-437)-His were detected in total cell extracts only which suggested that these two proteins were in the insoluble aggregates (Fig. 4, left panel).
Further optimization of the expression conditions didn't improve the expression of these truncated forms (data not shown).
In contrast, the same truncated forms of eEF1Bg fused with the N-terminal His-tag (Fig. 1, A by affinity chromatography to homogeneity (Fig. 4, right panel) as described in Material and methods.
The aggregation state of all truncated forms obtained was estimated by gel filtration (Fig. 5).The His-eEF1Bg (34-437) and His-eEF1Bg (93-437) proteins formed large soluble aggregates eluted in the column void volume or close to it (> 1 MDa).The shorter forms His-eEF1Bg (229-437) and His-eEF1Bg (264-437) migrated as a sharp single peaks of 47 and 30 kDa, respectively (Fig. 5).The calculated molecular weight of His-eEF1Bg (264-437) is 21.4 kDa that is in agreement with the obtained experimental value, whereas the molecular weight of His-eEF1Bg (229-437) is 25.4 kDa that is almost two fold lower as compared to the experimental result.The structure of 19 kDa C-terminal domain of the human eEF1Bg (residues 276-437) was solved by NMR [10].This domain behaves mostly as a monomer in solution and has a contact lens shape.This should be true for the truncated mutant His-eEF1Bg (264-437), which is only 13 amino acids longer.By contrast, His-eEF1Bg (229-437) is most probably a non-globular protein: due to the presence of lysine rich linker (amino acids 228-263, Fig. 1) its shape might became considerably extended.Thus, both His-eEF1Bg (264-437) and His-eEF1Bg (229-437) variants are suitable for further studies with interacting partners.
We attempted also to perform a denaturation-renaturation procedure for the His-eEF1Bg (34-437) and His-eEF1Bg (93 -437) protein variants.The denaturation of both proteins was done in the buffer containing 8 M urea.For the renaturation two approaches were tested: renaturation of proteins immobilized on Ni-NTA matrix by reversed liner gradient of urea and in solution by a step-wise dialysis.Unfortunately, neither approach led to a sufficient recovery of both proteins in a monomeric form (data not shown).
As was mentioned above, we failed to express the truncated and full-length eEF1Bg with C-terminal Histag (Fig. 4, left panel), whereas the respective N-terminal His-tag proteins were purified as soluble proteins (Fig. 4 folding and/or solubility of a particular recombinant protein expressing in bacteria [18].Thus, for successful expression of the recombinant human eEF1Bg the Nterminal localization of an affinity tag appeared to be crucial. Expression, purification and aggregation state of the C-terminally truncated forms of eEF1Bg fused with Nterminal GST-or MBP-tag.The C-terminal eEF1Bg deletion mutants schematically shown in Fig. 1 were expressed in E. coli in the form of MBP and GST chimeric proteins (Fig. 6).Three truncated forms of eEF1Bg with His-MBP-tag demonstrated good expression, except the eEF1Bg (1-33) deletion mutant (Fig. 6, upper panel).In contrast, only GST-eEF1Bg (1-33) and GST-eEF1Bg (1-228) were expressed in bacteria as soluble proteins.Thus, for further purification we chose two truncated forms with different tags: eEF1Bg (1-92) and eEF1Bg (1-163) with His-MBP and eEF1Bg  and eEF1Bg (1-228) with GST-tag.GST-eEF1Bg (1-33) and GST-eEF1Bg (1-228) were successfully purified by affinity chromatography to an apparent homogeneity (not shown) and their aggregation state was checked by analytical gel filtration.Both proteins eluted from a column as sharp peaks of 60 and 103 kDa, respectively (Fig. 6).In GST-eEF1Bg (1-33) preparation a small contaminating protein (< 10 kDa) was detected.Taking into account that the theoretical molecular weight of GST-eEF1Bg (1-33) is 30.4 kDa and GST-eEF1Bg (1-228) is 52.5 kDa, both proteins most probably are dimers in solution.
Unlike GST fusions, the purification of His-MBP bound truncated forms was more problematic.We decided to purify the His-MBP-eEF1Bg (1-92) and His-MBP-eEF1Bg (1-163) mutants because the same truncated forms were not expressed as GST fusions (Fig. 6, upper panel).for purification, but this approach was not successful.The binding of His-MBP fusions to the Ni affinity matrix was very low.Then, we used the MBPTrap TM HP column that specifically interacts with maltose binding protein.Using this matrix both His-MBP-eEF1Bg (1-92) and His-MBP-eEF1Bg (1-163) proteins were purified to an apparent homogeneity (not shown) and their aggregation state was tested by gel filtration.Unfortunately, both proteins were obtained as soluble aggregates eluted in and close to the void volume of column (Fig. 6, lower panel).The denaturation procedure in 8 M urea with subsequent renaturation (described in Material and method) led to the partial dissociation of the aggregates but resulted in extremely low yield of the target proteins (data not shown).Thus, only two truncated forms GST-eEF1Bg (1-33) and GST-eEF1Bg (1-228) were obtained as soluble proteins suitable for further investigation.Despite all our efforts, the heavy aggregates of other forms were not possible to disrupt.In our opinion the reason for this failure lays in highly hydrophobic nature of the eEF1Bg N-terminal domain (41.9 % of hydrophobic residues).Most probably the removing of the relatively long amino acid fragments from this protein may lead to the incorrect folding followed by its irreversible aggregation via exposed hydrophobic regions.Presumably, the Nterminal domain of eEF1Bg could not be divided into separate subdomains.
Conclusions.The recombinant His-eEF1Bg and its truncated variants His-eEF1Bg (264-437), His-eEF1Bg (229-437), GST-eEF1Bg (1-33) and GST-eEF1Bg (1-228) were expressed in E. coli and purified under the conditions mostly preventing the formation of aggregates.For the successful expression of the recombinant full-length eEF1Bg and its truncated forms the affinity tag should be attached to the N-end of the protein.All the obtained protein variants will be used for further study on the protein-protein interaction.
Acknowledgments.Authors are grateful to Dr. G. Janssen (Leiden University, The Netherlands) for providing pET16b/eEF1Bg plasmid.
Funding.This work was supported by the Ukraine-France PICS (F2/2012-

Fig. 5 .
Fig. 5. Analysis of the eEF1Bg truncated forms by gel filtration on Superose 6 HR column.Each protein was applied on the column at 10 µM concentration