Protein glutaminylation is a yeast-specific posttranslational modification of elongation factor 1A

Ribosomal translation factors are fundamental for protein synthesis and highly conserved in all kingdoms of life. The essential eukaryotic elongation factor 1A (eEF1A) delivers aminoacyl tRNAs to the A-site of the translating 80S ribosome. Several studies have revealed that eEF1A is posttranslationally modified. Using MS analysis, site-directed mutagenesis, and X-ray structural data analysis of Saccharomyces cerevisiae eEF1A, we identified a posttranslational modification in which the α amino group of mono-l-glutamine is covalently linked to the side chain of glutamate 45 in eEF1A. The MS analysis suggested that all eEF1A molecules are modified by this glutaminylation and that this posttranslational modification occurs at all stages of yeast growth. The mutational studies revealed that this glutaminylation is not essential for the normal functions of eEF1A in S. cerevisiae. However, eEF1A glutaminylation slightly reduced growth under antibiotic-induced translational stress conditions. Moreover, we identified the same posttranslational modification in eEF1A from Schizosaccharomyces pombe but not in various other eukaryotic organisms tested despite strict conservation of the Glu45 residue among these organisms. We therefore conclude that eEF1A glutaminylation is a yeast-specific posttranslational modification that appears to influence protein translation.

Ribosomal translation factors are fundamental for protein synthesis and highly conserved in all kingdoms of life. The essential eukaryotic elongation factor 1A (eEF1A) delivers aminoacyl tRNAs to the A-site of the translating 80S ribosome. Several studies have revealed that eEF1A is posttranslationally modified. Using MS analysis, site-directed mutagenesis, and X-ray structural data analysis of Saccharomyces cerevisiae eEF1A, we identified a posttranslational modification in which the ␣ amino group of mono-L-glutamine is covalently linked to the side chain of glutamate 45 in eEF1A. The MS analysis suggested that all eEF1A molecules are modified by this glutaminylation and that this posttranslational modification occurs at all stages of yeast growth. The mutational studies revealed that this glutaminylation is not essential for the normal functions of eEF1A in S. cerevisiae. However, eEF1A glutaminylation slightly reduced growth under antibiotic-induced translational stress conditions. Moreover, we identified the same posttranslational modification in eEF1A from Schizosaccharomyces pombe but not in various other eukaryotic organisms tested despite strict conservation of the Glu 45 residue among these organisms. We therefore conclude that eEF1A glutaminylation is a yeast-specific posttranslational modification that appears to influence protein translation.
One of the most conserved biological processes is ribosomal protein synthesis, comprising initiation, elongation, termination, and recycling steps. Each step is dependent on specific translation factors. Eukaryotic elongation factor 1A (eEF1A), 4 a large GTPase that is one of the most abundant cytosolic proteins, is important for the binding, stabilization, and delivery of aminoacylated tRNA to the translating ribosome. Correct codon-anticodon pairing of the aminoacyl-tRNA with the mRNA at the ribosomal A-site triggers ribosome-dependent hydrolysis of GTP and leads to dissociation of eEF1A from the ribosome. eEF1A is reactivated by nucleotide exchange factor eEF1B (1) and then able to reassociate with charged tRNA to start a new translation cycle. Besides its essential role in mRNA translation, eEF1A participates in many other cellular functions and is reportedly involved in actin bundling, nuclear export (2), signal transduction (3), apoptosis, proteasomal degradation, and tumorigenesis (4,5).
Crystal structures of archaeal, mammalian, and yeast eEF1A (partially in complex with subunits of eEF1B) showed that the elongation factor consists of three domains (I-III) (6 -8). Domain I is the GTP-binding domain and resembles GTPases of the Ras family. Domains II and III are likely to act as a rigid functional unit and are involved in aminoacyl-tRNA binding. These domains were also reported to be implicated in the interaction with cytoskeletal proteins (9,10).
eEF1A is extensively posttranslationally modified by lysyl acetylation, methylation (11), ubiquitination, nitrosylation, glutathionylation, phosphorylation (12), C-terminal methyl esterification (13), and the attachment of ethanolamine phosphoglycerol (EPG) (14). Most posttranslational modifications of eEF1A occur in domains II and III, whereas fewer modifications are found in the enzymatic GTPase domain. However, domain I of eEF1A is posttranslationally modified by Legionella pneumophila glucosyltransferases Lgt1-3, which attach glu- cose onto Ser 53 of yeast and mammalian eEF1A (15,16). Although toxin-induced modification of eEF1A results in inhibition of protein synthesis in yeast and mammalian cells, the roles and functional consequences of most endogenous posttranslational modification of eEF1A are still not clear (17). eEF1A belongs to the most conserved proteins across all kingdoms of life (18). However, in contrast to the prokaryotic orthologue, eukaryotic and archaeal EF1A contain an additional subdomain, the helix A*-loop-helix AЈ region (amino acids 36 -69). This double helix insert is part of the switch I region within the GTPase domain. Extensive structural alterations of this region during GDP/GTP exchange and during interaction with the ribosome suggest a pivotal role of this additional helix-loop-helix region in eukaryotic organisms (7).
Here we describe, for the first time, a novel type of posttranslational modification, the glutaminylation of a glutamate residue, that occurs within the helix A*-loop-helix AЈ region of yeast eEF1A. We show that the glutamine residue is attached via its ␣ amino group to the side chain carboxyl group of Glu 45 within eEF1A. Glutaminylation of eEF1A was detected in Saccharomyces cerevisiae and Schizosaccharomyces pombe. Experiments with S. cerevisiae expressing Glu 45 eEF1A mutants instead of the wildtype elongation factor indicate that glutaminylation slightly enhances growth defects under translational stress conditions.

Posttranslational modification of the switch I region in eEF1A
We set out to specifically analyze the helix A*-loop-helix AЈ region in EF1A for unconventional posttranslational modifications. eEF1A was purified from S. cerevisiae using affinity chromatography with His 6 -tagged guanine nucleotide exchange factor eEF1B (19). MS analysis revealed that the tryptic peptide 42-FEKEAAELGK-51 [Mϩ2H] 2ϩ ϭ 625.3251 was shifted to a higher mass by 128.05858 Da Ϯ 2 ppm (Fig. 1A). The additional mass suggested a modification with an organic molecule with the sum formula C 5 H 8 N 2 O 2 , which is exactly the mass of the amino acid glutamine reduced by an H 2 O molecule. Further CID MS-MS fragmentation analysis showed that the additional mass was clearly attached to the side chain of Glu 45 (Fig. 1B). The attachment of a single glutamine to proteins has so far not been reported. We termed this novel type of posttranslational modification "glutaminylation."

Confirmation of eEF1A glutaminylation at Glu 45 by site-directed mutagenesis
To confirm the glutaminylation of Glu 45 in eEF1A, we performed MS-MS analysis with the site-directed mutants eEF1A E45A and eEF1A E45D expressed in and isolated from S. cerevisiae (Fig. 2). Mutation of Glu 45 to alanine prevented modification by glutamine. Similarly, the mutant eEF1A E45D with shortened side-chain carbonyl was not modified. Thus, glutaminylation seems to be highly specific for glutamate at position 45. Notably, using wild-type eEF1A, we were not able to identify unmodified eEF1A in yeast cells. Also, at different cell stages of yeast growth, eEF1A was always completely glutaminylated (supplemental Fig. 1).

Glutaminylation of yeast eEF1A L-Glutamine is linked to Glu 45 via the ␣ amino group
To analyze glutaminylation in molecular detail and clarify how the glutaminyl moiety is attached, we used the 1.67-Åresolution crystal structure of S. cerevisiae eEF1A in complex with the C-terminal catalytic domain of the exchange factor eEF1B (PDB code 1F60 (8)), which is the highest-resolution structure available for this protein in the Protein Data Bank. An unassigned electron density close to residue Glu 45 permitted addition of a glutaminyl moiety in a defined orientation (Fig. 3, A and B). The ␣ amino group of the glutamine is covalently linked to the ␦ C atom of the carboxylic group of the Glu 45 side chain. This is a peptide bond-like connection with a characteristic planar geometry and a short C-N bond. The bond formation results in the loss of one oxygen atom of the carboxylic group from Glu 45 and equates to the release of a water molecule (Fig. 3C). The modification extends the glutamate side chain as a branched moiety, with the side chain and the carboxylic group of glutamine both exposed to the surface of the GTPase domain of eEF1A. Structure refinement comparing the fit of L-and D-glutamine enantiomers clearly indicated the attachment of L-glutamine. Multiple posttranslational modifications were described previously for eEF1A (11)(12)(13)(14). Of those, monomethylation of Lys 30 and trimethylation of Lys 79 (11) could also be assigned to electron density features, and these modifications were included in the final refined structure (PDB code 5O8W) (supplemental Fig. 2, A and B). The attachment of a glutamatederived glutamyl is a mechanism known for several proteins, e.g. in polyglutamylation of ␣ and ␤ tubulin (20), ␥-glutamylation in glutathione metabolism, or xenobiotics detoxification (21). In contrast, the attachment of a single glutamine via its ␣ amino group has not been reported previously.

Mouse, bovine, fish, insect, and archaeal eEF1A are not glutaminylated at Glu 45
Glu 45 is highly conserved within the helix A*-loop-helix AЈ region of eukaryotic eEF1A and also present in several archaeal EF1A molecules (Fig. 4). To determine whether glutaminylation of Glu 45 is also conserved, we isolated elongation factors from fission yeast (S. pombe), archaea (Haloferax volcanii), zebrafish (Danio rerio) ZF4 fibroblasts, bovine (Bos taurus) liver, mouse (Mus musculus) brain and liver, and wax moth larvae (Galleria mellonella) homogenate and analyzed the proteins by LC/MS-MS. Identified peptidic fragments were wellassigned and clearly indicated that glutaminylation at Glu 45 was present in eEF1A from fission yeast (S. pombe) but not present in elongation factors isolated from any of the other organisms ( Fig. 5 and data not shown). In addition, we analyzed the crystal structures of eEF1A from rabbit (PDB code 4C0S (7), Aeropyrum pernix (PDB codes 3WXM (22) and 3VMF (23)), and Sulfolobus acidocaldarius (PDB codes 1SKQ (24) and 1JNY (6)). In

Glutaminylation of yeast eEF1A
agreement with our MS data, unassigned electron density around Glu 45 or equivalent residues was not detectable in any of the three structures containing mammalian or archaeal eEF1A. Thus, glutaminylation of eEF1A seems to be restricted to yeast.

Glutaminylation of eEF1A increases yeast growth defects under translational stress conditions
To analyze the biological effects of Glu 45 glutaminylation of eEF1A, we performed growth assays under various cell stress conditions with S. cerevisiae strains in which wild-type eEF1A was substituted by eEF1A-E45A, -E45D, or -E45K (supplemental Fig. 3A). Yeast cells expressing the mutant versions of eEF1A were viable and did not display growth defects under a variety of stress conditions, including temperature stress, osmotic stress, or endoplasmic reticulum stress in the presence of benomyl, caffeine, or polymyxin (supplemental Fig. 3B).
Interestingly, the response of the tested eEF1A mutants toward translation-specific stress conditions induced by the antibiotics Geneticin, paromomycin, and, to a lesser extent, anisomycin showed that modification at Glu 45 by glutaminylation (eEF1A Glu 45 -glut) slightly enhanced growth defects compared with mutant eEF1A E45D that was not modified by attachment of glutamine (Fig. 6). Moreover, introduction of a local positive charge and bulky side chain, as in the E45K mutant, increased the susceptibility of the mutant yeast strain toward translation-specific antibiotics. These data suggest that glutaminylation of eEF1A at the critical position Glu 45 might regulate translation under specific growth conditions.

Glycosylation of Ser 53 of eEF1A by the Legionella effector Lgt3 is not influenced by glutaminylation
The results of the previous experiments with translational inhibitors suggested that site-directed mutation (eEF1A E45K) or glutaminylation of Glu 45 might alter the conformation of the helix A* loop region of the switch I region of eEF1A. Glu 45 is structurally located close to Ser 53 , the specific modification site of the L. pneumophila toxin effector glucosyltransferases Lgt1, 2, and 3 (15,16). To analyze whether glutaminylation of Glu 45 results in structural alterations of the helix A*-loop-helix AЈ region unfavorable for Lgt3-catalyzed glucosylation, we compared the initial velocities of glycosylation of non-glutaminylated (eEF1A purified from mouse liver) versus glutaminylated eEF1A (purified from S. cerevisiae) (supplemental Fig. 4). Additionally, we compared the glucosylation of the ternary complex (eEF1A, GTP, and Phe-tRNA Phe ) constituting the bona fide substrate of Lgt3. We found that the initial glucosylation rate of glutaminylated and non-glutaminylated eEF1A within the ternary complex was similar, showing that Glu 45 glutaminylation does not influence modification of the elongation factor at Ser 53 by Lgt3.

Discussion
Here we show, by using tandem mass spectrometric analysis, site-directed mutagenesis, and structural data that eEF1A from yeast is modified by the attachment of a single glutamine moiety to amino acid Glu 45 within the GTPase domain. Structural data reveal covalent linkage of L-glutamine via the ␣ amino group.
Although glutamylation and polyglutamylation (the attachment of glutamic acids) are well-known posttranslational modifications (e.g. modification of ␣and ␤-tubulin (20)) catalyzed by ␥-glutamyltransferase (21), the posttranslational modification of proteins by attachment of glutamine (glutaminylation) via the ␣ amino group has not been reported previously.
The chemical principle of the novel posttranslational modification reaction reported here appears to be similar to the reaction catalyzed by glutaminyl-tRNA synthetase (Gln-RS) during the charging of tRNA at the 3Ј acceptor stem region, which results in the same covalent linkage of a glutaminyl moiety (25). Several aminoacyl-tRNA synthetase (aa-RS) paralogs with unknown functions exist in various species, including yeast. The class II lysyl-tRNA synthetase paralog GenX/PoxA/YjeA was shown to attach a lysine residue to another lysine of Escherichia coli elongation factor P (26,27). This lysine modification resembles the glutaminylation reaction of eEF1A. Thus, one may speculate that glutaminylation of eEF1A might also be catalyzed by a glutaminyl-tRNA synthetase paralog. Notably, the site of glutaminylation in eEF1A is located close to the 3Ј aminoacyl acceptor stem of the eEF1A-bound aa-tRNA (supplemental Fig. 5).
So far, it is unclear whether glutaminylation in yeast is a transient and reversible modification. Despite extensive efforts, we were not able to identify unmodified eEF1A in yeast cells. Anal-

Glutaminylation of yeast eEF1A
yses performed at different cell stages of yeast growth exclusively resulted in detection of completely glutaminylated eEF1A (supplemental Fig. 1). By contrast, we were not able to detect glutaminylation of mammalian, fish, or insect eEF1A. The site of glutaminylation and the nature of the acceptor amino acid is highly specific. We observed that amino acid point mutations of the glutaminyl attachment site (Glu 45 ) of yeast eEF1A prevented glutaminylation, even when glutamate was replaced by negatively charged aspartic acid with a C1 atom-shortened side chain. This demonstrates that the attachment of glutamine is strictly specific for Glu 45 .
Besides glutaminylation of eEF1A, there are three other "unusual" posttranslational modifications also found in eukaryotic ribosome-associated factors (28): EPG modification of Glu 301 and Glu 374 in eEF1A, hypusine modification of eukaryotic initiation factor 5A (eIF5A), and diphthamide modification of eukaryotic elongation factor 2 (eEF2) (29). The precise functional roles of these modifications are not well-understood. Although the diphthamide modification might play a role in translation fidelity (30), the function of hypusine and EPG modifications are still enigmatic (28). Expression of a glutaminyldeficient mutant of eEF1A (E45A or E45D) in S. cerevisiae, after deletion of endogenous eEF1A, demonstrated that attachment of glutamine to eEF1A is not essential for growth. Surprisingly, translation stress conditions induced by Geneticin, paromomycin, or anisomycin showed growth retardation of native glutaminylated eEF1A (local neutral charge) or eEF1A with an introduced bulkier side chain of lysine at position Glu 45 (local positive charge) in comparison with the E45D mutant (local negative charge). Thus, in the eEF1A E45K mutant, as in the Glu 45 -glutaminylated eEF1A, the effect of the bulky and positively charged side chain might influence the interaction with the ribosomal factor-binding site and, therefore, translation efficiency. These data suggest that, in yeast, glutaminylation might have a regulatory function in protein synthesis. Previously, we reported that Ser 53 of eEF1A is modified by glucosylation catalyzed by the L. pneumophila effectors Lgt1-3 (15,16). We found that Lgt-induced glucosylation did not differ with glutaminylated and non-glutaminylated eEF1A, which is a further indication that modification of Glu 45 did not cause drastic structural changes of the molecule.
Recently, the mammalian elongation complex structure with eEF1A bound to the 80S ribosome has been reported (31). In this complex, the conserved Glu 45 is oriented close to the sarcin-ricin loop. The sarcin-ricin loop is suggested to stimulate the GTPase activity of eEF1A after codon-anticodon recognition at the A-site of the ribosome (31, 32). Analysis of the described mammalian 80S ribosome complex reveals that glutaminylation of amino acid Glu 45 of eEF1A would be accommodated in this complex. On the other hand, it is conceivable that the glutaminylation is involved in binding of the eEF1A ternary complex to the factor-binding site of the ribosome and, therefore, has the potential to influence translation (supplemental Fig. 5).
Taken together, here we describe a novel posttranslational modification of yeast eEF1A by glutaminylation. In all organisms, the genomic content is restricted to a specific protein repertoire. Fine-tuning of protein functions (e.g. in signaling events) and/or functional extensions are achieved by a large array of posttranslational modifications, resulting in additional diversification of the proteome. Translation elongation factors especially appear to be substrates of an extended spectrum of posttranslational modifications not observed in other proteins. Our findings add another type of posttranslational modification to this spectrum. It remains to be clarified whether glutaminylation is restricted to eEF1A or also observed in other proteins. Identification of the responsible enzyme will be an essential requirement for future detailed characterization of this novel posttranslational modification.

Glutaminylation of yeast eEF1A Protein purification
The cloning, recombinant expression and purification of Lgt3 (gene lgt3/lpg1488) from L. pneumophila strain Philadelphia-1 was described previously (36). For recombinant eEF1B (37) and Lgt3 protein production, the corresponding genes were induced with 0.2 mM isopropyl-␤-D-thiogalactopyranoside at 22°C overnight on a shaker. Bacterial cells were collected by centrifugation and lysed by French press on ice. The proteins were purified by nickel affinity chromatography using HisTrap columns (GE Healthcare) connected to an ÄKTA purifier (GE Healthcare) and stored in 10% glycerol/TBS solution at Ϫ80°C. Purification of native EF1A from S. cerevisiae, S. pombe, H. volcanii, zebrafish ZF4 cells, bovine liver tissue, mouse brain or liver, or total G. mellonella homogenate was performed by a method based on eEF1A interaction with His-tagged eEF1B␣ as described previously (37).

Engineering of S. cerevisiae strains with eEF1A mutations
Construction of S. cerevisiae variant with inactivated chromosomal copies of eEF1A and containing wild-type eEF1A on an Ura3 marker-containing plasmid was described previously (38). This strain was transformed individually with the pRS313based plasmids coding for eEF1A wild-type, E45D, E45K, and E45A and passed over 5-fluoroorotic acid-containing agar plates (39) to remove the initial Ura3 marker-carrying plasmid coding for wild-type eEF1A.

Yeast growth assay
To estimate yeast growth phenotypes, engineered S. cerevisiae cells were titrated 10-fold from the starting value of A 600 ϭ 1.0. From each dilution, an aliquot of 5 l of suspension was dropped onto SD agar supplemented with the corresponding marker substances. Where indicated, additional stress agents were included (20 mM DTT, 2% galactose, 50 ng/ml Geneticin, 350 g/ml paromomycin, 5 g/ml anisomycin, 1 M sorbitol, 1 M NaCl, 1 M KCl, 10 g/ml benomyl, 0.15% caffeine, or 200 g/ml polymyxin), or the pH level was adjusted correspondingly. Petri plates were incubated for 3-5 days at 30°C (or at the temperature mentioned in the figure legends) before photography. In comparison with liquid culture experiments, assays on solid medium led to results that were more conclusive.

Glucosyltransferase assay
Glucosylation was performed with 140 nM recombinant Histagged Lgt3 and eEF1A or the eEF1A-GTP-Phe-aatRNA Phe ternary complex (3 M) in a total volume of 20 l. Production of Phe-tRNA Phe and ternary eEF1A complex formation was performed as described previously (36). The standard glycosylation reaction was performed at 30°C for the indicated times in 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM MnCl 2 , and 10 M UDP-[ 14 C]glucose. The reaction was stopped by addition of SDS sample buffer and heating at 95°C for 5 min. Subsequently, the samples were subjected to polyacrylamide gel electrophoresis (40). Proteins were stained with Coomassie Brilliant Blue R250, and radiolabeled bands were analyzed by phosphorimaging and quantified using ImageQuant 5.2 (GE Healthcare) and Sigma Plot.

LC/MS-MS analysis
For in-gel digestion, the excised gel bands were destained with 30% acetonitrile, shrunk with 100% acetonitrile, and dried in a vacuum concentrator (Concentrator 5301, Eppendorf, Hamburg, Germany). Digests with trypsin were performed at 37°C and digests with chymotrypsin at 25°C overnight in 0.1 M NH 4 HCO 3 (pH 8). About 0.1 g of protease was used for one gel band. Peptides were extracted from the gel slices with 5% formic acid. LC/MS-MS analyses were performed on a Q-TOF (quadrupole time-of-flight) mass spectrometer (Agilent 6520, Agilent Technologies) and on an ion trap mass spectrometer (Agilent 6340, Agilent Technologies) equipped with an ETD source. Both instruments were coupled to a 1200 Agilent nanoflow system via a HPLC chip cube electrospray ionization interface. Peptides were separated on an HPLC chip with an analytical column of 75-m inner diameter and 150-mm length and a 40-nl trap column, both packed with Zorbax 300SB C-18 (5-m particle size). Peptides were eluted with a linear acetonitrile gradient with 1%/min at a flow rate of 300 nl/min (starting with 3% acetonitrile). The Q-TOF was operated in the 2-GHz extended dynamic range mode. MS-MS analyses were performed using data-dependent acquisition mode. After an MS scan (2 spectra/s), a maximum of three peptides were selected for MS-MS (2 spectra/s). Singly charged precursor ions were excluded from selection. Internal calibration was applied. The ETD ion trap was operated in data-dependent acquisition mode. After an MS scan (standard enhanced mode), a maximum of three peptides were selected for MS-MS (standard enhanced mode). The ICC control for the survey scan was set to 350,000, and the maximum accumulation time was set to 300 ms. The accumulation time for fluoranthene was set to 4 ms (according to an ICC of 500,000 -600,000), and the ETD reaction was set to 100 ms. Resonance excitation (smart decomposition) was used for doubly charged peptides. Mascot Distiller 2.3 was used for raw data processing and for generating peak lists, essentially with standard settings for the Agilent Q-TOF or Agilent ion trap, respectively. Mascot Server 2.3 was used for database searching with the following parameters: peptide Glutaminylation of yeast eEF1A mass tolerance, 20 ppm (Q-TOF)/1.1 Da (ion trap); MS-MS mass tolerance, 0.05 Da (Q-TOF)/0.3 Da (ion trap); enzyme, "trypsin" or "chymotrypsin" with two uncleaved sites allowed; variable modifications, carbamidomethyl (C), Gln-Ͼ pyroGlu (N-term. Q), oxidation (M), and hexose (ST). For protein identification, a custom database was used.

X-ray structure analysis and refinement
The vicinity of Glu 45 in the 1.67-Å-resolution X-ray structure of eEF1A from S. cerevisiae (PDB code 1F60 (8)) was inspected for unassigned electron density, which was calculated using TLS (translation, libration, screw motion) refinement in phenix.refine (41) with coordinates and structure factors retrieved from the Protein Data Bank. The mF o -dF c difference electron density map clearly indicated the position and orientation of the posttranslational modification. Water molecules in the proximity of Glu 45 were removed, and the glutaminylation was manually inserted as covalent attachment to Glu 45 in a peptide bond-like manner between the ␣ amino group of glutamine and the ␦ C atom of the Glu 45 side chain. The structure was refined employing restraints for a peptide-like bond according to Ref. 42. Specifically, refinement targets for atomic distances were set to 1.336 Å (N Gln -C␦ E45 ), 1.229 Å (C␦ E45 -O⑀ E45 ), 1.459 Å (C␣ Gln -N Gln ), and 1.525 Å (C␦ E45 -C␥ E45 ). Refinement targets for the angles were set as follows: 121.7°( C␣ Gln -N Gln -C␦ E45 ), 122.7°(N Gln -C␦ E45 -O⑀ E45 ), 120.1°(O⑀ E45 -C␦␥ E45 ), and 117.2°(N Gln -C␦ E45 -C␥ E45 ). In addition, planarity restraints were used for atoms C␣ Gln , N Gln , C␦ E45 , C␥ E45 , and O⑀ E45 . After refinement, these distances and the plane were identical to the ideal values. The quality of the model and fit to the electron density were compared for L-and D-glutamine, and the ideally fitting L-enantiomer was included in the final structure. Water molecules were inserted, and model building was completed using COOT (43). This included insertion of three polyethylene glycol moieties and two additional posttranslational modifications an N⑀-monomethylation of Lys 30 and N⑀,N⑀,N⑀-trimethylation of Lys 79 . Electron density was visible for an additional short ␣-helical peptide that was likely to be part of the C terminus of eEF1A, as fragmentary electron density at 1 and continuous electron density at 0.5 connected that peptide to the C terminus of eEF1A. However, the connecting loop was not ordered enough to allow refinement. The structure was refined using phenix.refine with TLS to a crystallographic R and R free of 15.6% and 18.4%, respectively. The quality of the structure was checked using MolProbity (44). There was no Ramachandran outlier, and 97.44% of the residues were in the favored regions. Data refinement statistics are summarized in supplemental Table 1. The figures were prepared using the PyMOL molecular graphics system (version 1.5.0.4, Schrödinger, LLC).
Author contributions-T. J. designed and conducted experiments, analyzed the data, and wrote the paper. A. S., Z. H., and J. D. performed mass spectrometric analyses. C. W., C. H., and G. R. A. conducted structural analyses. Y. B. and S. R. performed yeast experiments. T. T. performed glycosylation experiments. K. A. designed the study, analyzed the data, and wrote the paper. All authors discussed the results and commented on the manuscript.