eEF2 (eukaryotic elongation factor 2) contains a post-translationally modified histidine residue, known as diphthamide, which is the specific ADP-ribosylation target of diphtheria toxin, cholix toxin and Pseudomonas aeruginosa exotoxin A. Site-directed mutagenesis was conducted on residues within the diphthamide-containing loop (Leu693–Gly703) of eEF2 by replacement with alanine. The purified yeast eEF2 mutant proteins were then investigated to determine the role of this loop region in ADP-ribose acceptor activity of elongation factor 2 as catalysed by exotoxin A. A number of single alanine substitutions in the diphthamide-containing loop caused a significant reduction in the eEF2 ADP-ribose acceptor activities, including two strictly conserved residues, His694 and Asp696. Analysis by MS revealed that all of these mutant proteins lacked the 2′-modification on the His699 residue and that eEF2 is acetylated at Lys509. Furthermore, it was revealed that the imidazole ring of Diph699 (diphthamide at position 699) still functions as an ADP-ribose acceptor (albeit poorly), even without the diphthamide modification on the His699. Therefore, this diphthamide-containing loop plays an important role in the ADP-ribosylation of eEF2 catalysed by toxin and also for modification of His699 by the endogenous diphthamide modification machinery.
- enzyme structure and function
- eukaryotic elongation factor 2 (eEF2)
- Pseudomonas aeruginosa exotoxin A (ExoA)
eEF2 (eukaryotic elongation factor 2), a member of the GTPase superfamily, is a single polypeptide chain with a molecular mass varying from 90 to 110 kDa . The eEF2 protein functions at the elongation step of protein synthesis, where it catalyses the GTP hydrolysis-dependent translocation of peptidyl-tRNA from the A site to the P site within the ribosome, and the translocation of the deacylated tRNA from the P site to the E site (reviewed in ). The eEF2 protein is highly conserved among eukaryotes, showing 58–100% sequence identity and 71–100% sequence homology (Y. Zhang, unpublished work). It is less well conserved among the Archaea, but, nonetheless, the conservation is still noteworthy (29–36% identity and 47–56% homology). The eEF2 protein contains a post-translationally modified histidine residue known as diphthamide, which, from previous NMR spectral analysis, was suggested to be 2-[3-carboxyamido-3-(trimethylammonio)propyl]histidine  and was conclusively identified by X-ray crystallography [4,5]. This residue is located at position 699 within domain IV of Saccharomyces cerevisiae eEF2, which is the specific ribosylation target of the diphthamide-specific bacterial toxins [3,4,6]. Although the diphthamide is the site of toxin attack, it does not appear to be essential for normal eEF2 function [7–9]; however, it may interact with the backbone phosphate groups of two conserved bases of the small ribosomal subunit in eukaryotes in a proof-reading function . The diphthamide of eEF2 is the exclusive cellular substrate for inactivation by DT (diphtheria toxin), Pseudomonas aeruginosa ExoA (exotoxin A) and cholix toxin [11–13]. Diphthamide has only been found in eEF2 and is completely conserved in the ribosome translocase throughout all eukaryotic and Archaea evolution [14,15]. Five proteins (dph1, dph2, dph3, dph4 and dph5) have been identified as participants in the biosynthesis of diphthamide [9,16–18]. Yeast mutants with deletions of DPH1–DPH4 genes show complete resistance to DT [7–9]. Moreover, the eEF2 from the yeast mutant lacking DPH5 can be ADP-ribosylated very weakly by DT, although it is undetectable when using the standard enzyme assay . The dph2 protein, together with others (dph1, dph3 and dph4), function at the first step in diphthamide biosynthesis, in which they transfer the 3-amino-3-carboxypropyl group to the C-2 of the imidazole ring of the precursor histidine of eEF2 . A cellular ADPRT (ADP-ribosyltransferase) that is believed to possess a similar catalytic mechanism to the diphthamide-specific bacterial toxins has been identified [14,19], a finding indicating that the highly conserved diphthamide residue in eukaryotic cells may not be simply a target of DT and ExoA, but may serve as a site for the regulation of eEF2 activity . Unlike ADP-ribosylation by DT, ExoA and cholix toxin, the eukaryotic cellular ADPRT only covalently modifies a fraction of eEF2 within the eukaryotic cell, and the ADP-ribose can subsequently be removed through the action of ADP-ribosyl hydrolases, which results in the re-activation of eEF2 .
The location of the site(s) for ADP-ribosylation within the eEF2 protein is not without controversy, however, since there have been a number of reports suggesting that covalent binding of ADP-ribose to the eEF2 occurs, which is different from its NAD+ and endogeneous transferase-dependent ADP-ribosylation [21,22]. Several high-resolution X-ray structures consisting of the Ps. aeruginosa toxin in complex with yeast eEF2 were recently reported , and it was suggested that residues in eEF2 are important for the toxin-catalysed ADPRT reaction mechanism (substrate-assisted catalysis). The quaternary ammonium group of diphthamide is located close to the β-phosphate group of the NAD+ substrate and both His694 and Asp696 of eEF2 form H-bonds with the N-ribose-2′OH in ADPr-eEF2 (ADP-ribosylated eEF2) . Recently, three domain-IV mutants of eEF2 (D696A, I698A and H699N) were characterized for their ability to maintain translation fidelity and for their resistance to DT . In the present study we sought to elucidate further the role of the diphthamide loop region within eEF2 that contains both His694 and Asp696 in the ADP-ribosylation reaction catalysed by Pseudomonas ExoA. ASM (alanine-scanning mutagenesis) was used to survey the diphthamide loop within yeast eEF2 (Leu693 to Gly703) for key residues that may be involved in the ADP-ribose acceptor activity of eEF2 or that may serve as recognition sites for the diphthamide biosynthetic enzymes in yeast. In addition, a knockout yeast strain with the deletion of the DPH2 gene was used to determine the role of the 2′-substituent on the imidazole ring of the diphthamide residue in the toxin-catalysed ADP-ribosylation of eEF2. To determine the extent of diphthamide modification at His699, LC–MS/MS (liquid chromatography–tandem MS) was used to analyse tryptic digests of those samples.
The S. cerevisiae strain YEFD12h/pURA3-EFT1 (MATa ade2 lys2 ura3 his3 leu2 trp1 eft1 Δ:HIS3 eft2 Δ:TRP1 pURA3-EFT1) was used for the electroporation experiments; the S. cerevisiae strains with the DPH2 and DPH5 knockout were purchased from Open Biosystems (Huntsville, AL, U.S.A.). The MKK-A yeast strain, which expresses the H699A mutant of the eEF2 protein, was a gift from Dr Kenji Kohno (Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Nara, Japan) . The yeast mutant strain harbouring the H694A eEF2 mutant caused a significantly slower growth phenotype in yeast, indicating that the mutant eEF2 protein is defective when the same strain and plasmid system as described previously  was used. However, it did sustain growth of the yeast culture after long incubations. We used electroporation, which is a more efficient method for obtaining yeast transformants especially when there is a defect associated with the protein product encoded on the plasmid. Using our usual protocol, we re-sequenced the mutant eEF2 gene from the working yeast strain to verify that the H694A mutation was still present in our cultures, since yeasts are well-known to produce revertants at a high rate.
Site-directed mutagenesis and plasmid shuffling
Single alanine mutations were made in the pTKB612 plasmid using the QuikChange® XL Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA, U.S.A.), which encodes the EFT2 (elongation factor 2) gene with a His6 (hexahistidine) tag at the C-terminus and contains the LEU2 (β-isopropylmalate dehydrogenase) selectable marker . The appropriate codon changes in the mutated plasmids were verified by cycle sequencing using an Applied Biosystems (Foster City, CA, U.S.A.) 3730 DNA analyser. The mutant plasmids were used to transform the yeast strain YEFD12h/pURA3-EFT1 by electroporation and the transformation mixture was streaked on to SD–Leu [SD (synthetic dropout) medium lacking leucine] plates. The yeast colonies that grew on SD–Leu plates were streaked on SD+0.1% 5-FOA (5-fluoro-orotic acid) plates in order to eliminate the pURA3-EFT1 (wild-type) plasmid. After 5-FOA selection, the viable yeast colonies should only contain the mutated pTKB612 plasmid and should grow on SD–Leu plates, not on SD–uracil medium. In order to ensure that no reversion of the target mutation site had occurred, the pTKB612 plasmids in the mutant strains were then extracted from the host yeast strain and were re-sequenced as described above. Starter cultures were always grown from original mutant yeast-cell stocks in order to ensure that reversion to the wild-type genotype did not occur.
Expression and purification of His6-tagged eEF2 proteins
A single yeast colony was inoculated in 50 ml of YPD (yeast/peptone/dextrose) medium containing ampicillin (100 μg/ml). The cells were grown at 30 °C with shaking until the culture attenuance (D600) was greater than 1.5, and the starter culture was transferred into 8 litres of YPD medium also containing ampicillin (100 μg/ml) with 10 ml of starter culture/2 litres of YPD medium. The cells were harvested when the D600 was 1.5, care being taken to ensure that the culture was not allowed to grow for longer than 30 h. The yeast cell pellets were resuspended in lysis buffer [50 mM KH2PO4, 1 M KCl, 1% Tween 20, 10% (v/v) glycerol, 10 mM imidazole and 0.2 mM PMSF, pH 8.0]. Cells were lysed using a bead beater (Biospec Products, Bartlesville, OK, U.S.A.) using 425–625 μm-diameter glass beads (Sigma, St Louis, MO, U.S.A.). The pH of the lysate was adjusted to 7.5 with 1 M K2HPO4 and the lysate was centrifuged at 30000 g for 20 min at 4 °C, filtered through two sheets of Whatman no. 1 filter paper, one sheet of Whatman no. 5 filter paper and then through low-protein-binding 0.45-μm-pore-size membrane (Millipore, Billerica, MA, U.S.A.). The filtrate was then loaded on to 2 ml of Ni2+-charged chelating Sepharose Fast Flow resin (Amersham Biosciences, Uppsala, Sweden), repeating the process once. The column was then washed with Wash buffer (50 mM KH2PO4, 1 M KCl, 1% Tween 20, 10% glycerol, 20 mM imidazole and 0.2 mM PMSF, pH 8.0). The His6-tagged eEF2 protein was eluted with Elution buffer (50 mM KH2PO4, 1 M KCl, 1% Tween 20, 10% glycerol, 250 mM imidazole and 0.2 mM PMSF, pH 8.0). Each fraction was analysed by SDS/PAGE, and the eEF2-containing fractions were pooled then dialysed overnight at 4 °C in 2 litres of dialysis buffer (20 mM Tris/HCl, 5 mM MgCl2, 0.5 mM EDTA, 10% glycerol, 30 mM KCl, 1 mM DTT (dithiothreitol) and 0.2 mM PMSF, pH 7.6] with one buffer change after 6 h. Filtered dialysed protein was loaded to a 1 ml Mono Q column (Amersham Biosciences) previously equilibrated with dialysis buffer. The protein was eluted with Q-500 buffer (20 mM Tris/HCl, 5 mM MgCl2, 0.5 mM EDTA, 10% glycerol, 500 mM KCl, 1 mM DTT and 0.2 mM PMSF, pH 7.6) with a linear gradient from 6% Q-500 to 25% Q-500 for 3 ml and 25% to 60% Q-500 over 25.5 ml at 0.5 ml/min. His6-tagged recombinant eEF2 proteins were pooled and concentrated, filtered through 0.2-μm-pore-size HT Tuffryn® low-protein-binding polysulfone membrane (Pall Corporation, Ann Arbor, MI, U.S.A.) and were stored at −80 °C until further use.
Expression and purification of Δdph2-eEF2 and H699A proteins
The expression of Δdph2 and H699A mutant eEF2 proteins was similar to the procedure described above for the various alanine mutants of eEF2. After lysis of the yeast cells and clarification, the supernatant was dialysed into 9 vol. of S-0 buffer (20 mM Mes, 5 mM MgCl2, 0.5 mM EDTA, 10% glycerol, 1 mM DTT and 0.1 mM PMSF, pH 6.7) overnight at 4 °C. The dialysed fraction was then clarified by centrifugation at 35000 g for 20 min at 4 °C and 45000 g for 25 min and filtered as described above. The filtrate was then loaded on to a 15 ml SP (sulfopropyl)-Sepharose Fast Flow column (Amersham Biosciences) previously equilibrated with 10% (v/v) S-300 (100% S-300 is 20 mM Mes, 5 mM MgCl2, 0.5 mM EDTA, 10% glycerol, 300 mM KCl, 1 mM DTT and 0.1 mM PMSF, pH 6.7). After loading, the column was washed with 60 ml of 10% S-300 buffer. The protein was then eluted with 60 ml of 18.3% S-300 buffer and was directly loaded to a 10 ml Source Q15 column (Amersham Biosciences), previously equilibrated with 10% Q-500 buffer. After loading, the column was washed with 50 ml of 10% Q-500 buffer at 1.5 ml/min. The protein was eluted with a 100 ml gradient, from 10 to 50%, of Q-500 solution. The fractions were analysed by SDS/PAGE, and the eEF2-containing fractions were pooled and diluted to 10% Q-500 with Q-0 (Q-500 without KCl) buffer. The diluted solution was loaded on to a 1 ml Mono Q column that had previously been equilibrated with 10% Q-500 buffer. A 3 ml gradient was run from 10 to 25% Q-500 buffer and a 25.5 ml gradient was developed from 25% to 60% Q-500 at 0.5 ml/min. eEF2 proteins were pooled, analysed by SDS/PAGE and concentrated in Amicon Ultra concentrators (Amicon Inc, Danvers, MA, U.S.A.). The protein obtained was filtered and stored at −80 °C.
Overexpression and purification of ExoAc (the catalytic fragment of ExoA)
ExoAc was expressed and purified as described previously . All protein concentrations were determined by absorption spectroscopy using the calculated molar absorption coefficients (ϵ) for each protein at 280 nm .
Tryptophan fluorescence emission spectra
The tryptophan fluorescence emission spectra of all the eEF2 proteins (1.0 μM) were obtained at 25 °C in 20 mM Tris/HCl/150 mM KCl buffer, pH 7.9. The excitation wavelength was 295 nm and the emission was monitored from 310 nm to 450 nm with all slits set at 4 nm.
Chemical denaturation of eEF2 proteins with GdmCl (guanidinium chloride)
eEF2 proteins were titrated with GdmCl in 20 mM Tris/HCl and 150 mM KCl solution at pH 7.9. The concentration range of GdmCl was from 0 to 7 M and the starting concentration of eEF2 was 2.0 μM. The unfolding of eEF2 proteins was monitored by the relative shifts in the emission wavelength of the tryptophan fluorescence spectra. The excitation was 295 nm and the emission was from 310 nm to 450 nm with all slits held at 4 nm. The data were analysed as described previously . The fluorescence measurements for Trp λem,max (tryptophan fluorescence wavelength emission maximum) and chemical denaturation were conducted with a PTI Alphascan spectrophotometer (Photon Technology International, South Brunswick, NJ, U.S.A.) interfaced with a computer using Felix™ version 1.41 software and equipped with a water-jacketted sample chamber set to 25 °C.
Gel labelling of eEF2 proteins
The purified eEF2 proteins (24 μM) were mixed with 15 μM ϵ-NAD+ (nicotinamide–1,N6-ethenoadenine dinucleotide, a fluorescent analogue of NAD+; Sigma) in the presence or absence of 12 μM ExoAc and incubated at room temperature (25 °C) for 30 min. The reaction was stopped by adding Laemmli loading buffer and the mixtures were then loaded on SDS/12.5%-(w/v)-PAGE gels and the bands were visualized under UV reflective light with a Hoechst Blue (460 nm) filter in a FluorChem™ 8900 Imaging System (Alpha Innotech, San Leandro, CA, U.S.A.).
MS analysis of eEF2 proteins
Protein samples were buffer-exchanged into 100 mM ammonium carbonate buffer, pH 8.0. A 2 μg sample of each protein in 100 μl of the ammonium carbonate buffer were reduced using DTT, alkylated with idoacetamide, and then digested with modified trypsin (Promega, Madison, WI, U.S.A.) at 37 °C for 4 h. A 5 μl portion of the digest solution was then injected for LC–MS/MS analysis. LC–MS/MS analysis of the protein digest was conducted with a Waters Chromatogrpahy Division (Millipore Corp., Milford, MA, U.S.A.) CapLC system coupled to a Q-TOF (quadrupole time-of-flight) mass spectrometer (Global Ultima; Micromass, Manchester, U.K.). Briefly, the digested peptides of the protein were separated on a capillary LC column and detected by MS. MS/MS was performed on peptide peaks detected with collision-induced dissociation to obtain sequence information. MassLynx and PEAKS software  were used for assignment of peptide sequences from the MS/MS data and the sequences of peptides were confirmed with manual verification. In addition, LC–MS analysis of the intact protein was carried out on a Waters CapLC system coupled to a Q-TOF2 spectrometer (Micromass). Proteins samples (10 μl) were injected into, and separated on, a reversed-phase column (Jupiter; C4; 5 μm particle size; 300 Å pore size; 150 mm long×1 mm internal diameter; Phenomenex, Torrance, CA, U.S.A.). The separation was achieved with gradient elution at a flow rate of 40 μl/min, beginning at 5% (v/v) acetonitrile and reaching 80% acetonitrile in 30 min. Mass spectra were recorded from 400 to 1800 m/z in the positive electrospray-ionization mode.
Fluorescence-based ADPRT assay
The ADPRT activity of the various enzyme samples was tested as described previously . Briefly, the excitation and emission monochromators were set to 305 nm and zero order (no diffraction) respectively, and a 309-nm-cut-off filter (Oriel Corporation, Stratford, CT, U.S.A.) was included on the sample-chamber side of the emission monochromator to eliminate scattered excitation light and to maximize the signal. Buffer (20 mM Tris/HCl, pH 7.9), ϵ-NAD+ (various concentrations were used up to 500 μM from an ϵ-NAD+ stock solution prepared in distilled water; ϵ265 6000 M−1·cm−1) and 14 μM eEF2 (at saturating levels) were combined in an Ultramicro (3 mm pathlength) cuvette (Helma Inc., Concord, ON, Canada). The cuvette was equilibrated at 25 °C for 10 min. Toxin dilutions were performed in silicone-treated tubes with the buffer described above. The reaction was started by rapidly adding ExoAc (final concentrations of the toxin ranged from 5 nM to 19 μM, depending upon the mutant protein) and the progress of the reaction was monitored by an increase in fluorescence during the production of ϵ-ADP-ribose.
Diphthamide-containing loop in eEF2
Figure 1(A) shows a multiple sequence alignment for part of domain IV of the eEF2 protein from a selected list of some higher eukaryotes and Archaea, including baker's yeast (S. cerevisiae), a thermophile (Thermoplasma volcanium), chicken (Gallus gallus), mouse (Mus musculus) and human (Homo sapiens). eEF2 is a highly conserved protein among eukaryotes and Archaea , and this level of sequence conservation is reflected in the domain IV segment shown in Figure 1(A). In the diphthamide-containing loop under study (Leu693-Gly703), among eukaryotes (17 species) and Archaea (20 species), His694, Asp696, His699, Arg700 are absolutely conserved, with Gly701 showing 97% conservation. This region possesses a loop conformation, except for a very interesting segment in the centre that includes Ala697-Arg700, which is a short 310 helix. The His699 residue is diphthamide, a post-translationally modified histidine  that is the site of ADP-ribosylation by both DT and ExoA  and is also believed to be a site for ribosylation by endogenous cellular enzymes [14,19]. In the present study, the role of this loop region in the ADP-ribosylation reaction catalysed by ExoA, and also its role in diphthamide modification, was investigated by ASM (Leu693-Gly703). The recently obtained X-ray structure of Pseudomonas exotoxin A in complex with eEF2  is shown in Figures 1B (full complex) and 1C (active site and interfacial regions only). It is clear that parts of the diphthamide loop within eEF2 make contact with the active-site elements of ExoA during the ADP-ribosylation reaction (Figure 1C). Loop 4 (Glu486-Ile493) within ExoAc makes intimate van der Waals contacts with residues Gly702 and Gly703 in the eEF2 diphthamide-containing loop (Figure 1C; ). Additionally, Gln483 in the toxin forms an H-bond with the backbone of Gly703 in eEF2, and also Trp669 in eEF2 forms several contacts with the residues Arg490-R492 within Loop 4 of ExoAc (Figure 1C; ). These interactions are at the heart of the proposed mechanism for toxin-catalysed ADP-ribosylation, where it was previously suggested that eEF2 facilitates its own covalent modification in a substrate-assisted mechanism .
Global structural integrity of the eEF2 mutants
The intrinsic Trp λem,max is a sensitive measure of the folded status of a protein in solution [27,30,31] and can be used to report changes in the stability of a mutant compared with its corresponding wild-type protein. Yeast eEF2 contains eight tryptophan residues, including three in the G′ domain and five in domain IV. The effect on the folded structure of the eEF2 protein caused by single amino acid substitution should be observable by a shift in Trp λem,max . The Trp λem,max values for all of the eEF2 mutant proteins ranged between 333 and 338 nm, and none of the mutant-protein values deviated significantly from the wild-type Trp λem,max value (335 nm). These data indicate that the global structure of the mutant proteins was unaffected by the alanine substitutions and as well as the diphthamide→histidine (Δdph2 -eEF2 knockout) mutation. In addition, chemical denaturation is also a powerful method for studying the folded stability of a protein [27,32]. On the basis of the denaturation profiles obtained for the wild-type and mutant eEF2 proteins, the D1/2 (midpoint of the transition between folded and unfolded state) is a reliable measure of protein stability, since it is not dependent upon subtle differences between the slopes of the baseline for the denaturation profiles . The D1/2 values ranged between 2.0 and 2.1 M, again indicating that all the mutant proteins tested were structurally intact and possessed stabilities similar to that of the wild-type protein. A further measure of the folded integrity and stability of the eEF2 mutant proteins comes from the viability of the respective host yeast cells, since the only copy of the eEF2 protein is provided by the plasmid harbouring the mutated EFT2 gene. Since eEF2 is an essential protein in eukaryotic cells , cell viability can be compromised if the mutant eEF2 is less stable or possesses reduced activity compared with the wild-type protein. Therefore, on this basis, the viability of the yeast eEF2 mutant strains provides further evidence that the alanine substitutions within the Leu693-Gly703 loop region of eEF2 did not compromise the folded stability or the function of the mutant eukaryotic ribosome translocase proteins.
MS of eEF2 mutant proteins
MS analysis was performed on the intact eEF2 proteins, and their masses are summarized in Table 1. The mass data confirmed that the wild-type protein was modified with a mass shift of 143 Da, which corresponds to the diphthamide modification of His699. The Δdph2 -eEF2 protein was also modified with a mass shift of 102 Da, which corresponds to a diphthamide intermediate. However, the observed protein masses do not match the calculated ones, all showing a similar mass discrepancy (~60 Da) suggestive of the presence of unknown modifications. On reduction of these proteins, a mass shift of 7±1 Da was observed for all proteins, suggesting that there are four disulfide intrachain bonds formed among the eight cysteine residues in eEF2. One of these bonds was Cys366–Cys372, as revealed from LC–MS/MS data of peptide ions at 826.03 m/z obtained from an unreduced tryptic digest sample (results not shown). Extensive analysis by LC–MS/MS after both tryptic and endoproteinase Asp-N digestion of the eEF2 proteins gave an amino acid sequence with 95% coverage. The first methionine residue at the N-terminus was not present in any of the sequences, as confirmed by LC–MS/MS analysis of the trypsin-digested samples. It was also found that Lys509 was acetylated (AcK in L-V-E-G-L-AcK-R) and that this was the only acetylation detected in the eEF2 protein. This acetylation is not likely to have occurred during sample preparation, since none of the other lysine residues were acetylated. Taking the acetylation into account, the observed masses were 35±2 Da higher than the calculated masses for the intact proteins. We do not know the reason for this mass discrepancy, but it is likely to be due to a salt adduct, since it cannot be traced at the peptide level. Since this mass discrepancy is observed for all proteins, it is not an important factor for the purpose of this comparative study.
To determine whether the alanine-substitution mutants of eEF2 and Δdph2- and Δdph5-eEF2 possessed the diphthamide modification at the 2-position of the His699 residue, LC–MS/MS analysis was conducted on these proteins. On tryptic digestion, eEF2 proteins generated a peptide (T85 in Figure 2, Val686–Arg700; VNILDVTLHADAIHR; molecular mass 1685.92 Da) containing the His699 (diphthamide) residue. Results of LC–MS/MS analysis of wild-type eEF2, mutant eEF2 and Δdph2- and Δdph5-eEF2 proteins are shown in Table 1. The wild-type eEF2 gave a peptide with a mass of 1829.04 Da, detected as an [M+H]2+ ion at 915.02 m/z, or as an [M+2H]3+ ion at 610.35 m/z, corresponding to a diphthamide modification at His699 of 143.12 Da. The LC–MS/MS spectrum of ions at 610.35 m/z shown in Figure 2(A) confirmed the amino acid sequence of the wild-type peptide and that the diphthamide modification was on His699. Owing to the presence of the quaternary amine group, all the y ions, except y1, were observed with a loss of 60 Da, which corresponds to [HN(CH3)]+ from the diphthamide on the His699 side chain. The detected m/z value, the charge state and the fragment pattern of the peptide ions agree well with the fact that one of the charges is due to the presence of the quaternary amine group, which is permanently charged. On the other hand, as expected, the T85 peptides were detected exclusively as non-modified from all of the toxin-resistant mutant eEF2 proteins: H694A, H699A, I698A, D696A, G701A and G703A (Table 1). These results indicate that the residues within the loop region surrounding His699 are of critical importance for recognition by the diphthamide biosynthetic enzymes. Δdph2-eEF2 contained no modification at the His699 residue in T85, since only the unmodified T85 peptide was detected from the Δdph2-eEF2 sample (Table 1). In contrast, Δdph5-eEF2 possessed an intermediate modification on His699 in T85 corresponding to 101.06 Da. The LC–MS/MS spectrum of the T85 peptide showed that all y ions (except for y1, which did not contain the modified histidine residue) were 101 Da heavier than their unmodified histidine counterparts (Figure 2B).
Detection of ADP-ribose modification at His(Diph)699
LC–MS/MS analysis of a trypsin digest of wild-type eEF2 protein labelled with ADP-ribose in the presence of toxin revealed the modified peptide T85. This peptide was detected at 790.7 m/z as [M+2H]3+, corresponding to ADP-ribose attachment at His699 [Diph699 (diphthamide at position 699)] of the diphthamide-containing peptide T85 (Figure 3). Its MS/MS spectrum is shown in Figures 3(B) and 3(C). Intense b and y* ion series were observed, and these allowed the determination of the protein sequence and the specific residue location of the modification at His699. The y* ions are not classical y fragment ions, but indicate further losses of 59 Da of N(CH3) as a neutral ion and of 542 Da of ADP-ribose unit as a positively charged ion respectively. Other ions generated by fragmentation through the backbone of ADP-ribose were also observed and are shown in Figure 3(C). From the LC–MS/MS analysis of a trypsin digest of Δdph2-eEF2 protein treated with toxin, no evidence was found for the existence of any modification at the imidazole ring of His699. This is possibly due to the weak ADP-ribose acceptor activity and, hence, very low abundance of this species.
Visualization of ADP-ribosylated eEF2 proteins
The eEF2 protein can be labelled with ϵ-ADP (ethenoadenosine diphosphate)-ribose by toxin in the presence of ϵ-NAD+ , and the fluorescent product, ϵ-ADP-ribose-eEF2, was visualized by UV excitation on SDS/PAGE gels (Figures 4A–4D). This gel-labelling procedure is a quick method that facilitates screening for toxin-resistant mutant eEF2 proteins. In order to verify that the eEF2 proteins were ADP-ribosylated by the enzymatic activity of the toxin and not by an endogenous transferase activity, intrinsic to and/or associated with eEF2, a control was included for each mutant protein tested in which eEF2 proteins were only incubated with ϵ-NAD+ (Figures 4A–4D). In no instance was there any sign of ADP-ribose labelling of eEF2 in the absence of toxin. The gel results indicated that, except for the W669A and L693A mutant proteins (Figure 4A, lanes 4 and 5, and lanes 6 and 7, respectively), the mutant eEF2 proteins could not be ADP-ribosylated by toxin under conditions where wild-type eEF2 is an excellent acceptor of ADP-ribose. This suggests that either the diphthamide-containing loop within eEF2 is an important determinant of the ADP-ribose acceptor activity of the elongation factor in the toxin-catalysed reaction or that this sequence is required by the diphthamide biosynthetic machinery in yeast.
Activity of the diphthamide-deficient mutant (Δdph2-eEF2)
To determine the molecular details of the ADP-ribose acceptor activity, the Δdph2-eEF2 protein was included as a substrate in the ADPRT assay for an extended time (6 h) and the concentration of toxin/enzyme was also increased from 10 nM to 32.1 μM (3200-fold increase). At such a high concentration of toxin, the toxin's intrinsic GH (NAD+ glycohydrolase) activity had to be considered, because, as a competing reaction also catalysed by the toxin , it may be expected to contribute significantly to the overall reaction progress. The grey curve in Figure 5(A) is the kinetic time-course trace for the GH reaction (toxin and ϵ-NAD+ only; no eEF2 acceptor present). During the first 6 min of the reaction, the Δdph2-eEF2 was incubated with ϵ-NAD+ in buffer only and the black trace shows that the Δdph2-eEF2 mutant protein possessed no autoribosylation activity. We have previously characterized the water-catalysed hydrolysis reaction and showed that it is nearly zero and is only slightly above the background noise, compared with the toxin-catalysed GH activity . Following the 6 min incubation, the addition of ExoAc (indicated by an arrow) resulted in a gradual increase in the rate of the ADPRT reaction, with Δdph2-eEF2 acting as the ADP-ribose acceptor (as judged by an increase in fluorescence intensity). The inset to Figure 5(A) illustrates the whole time course for both the GH and ADPRT reactions, and it is clear that the ADP-ribose acceptor activity of Δdph2-eEF2 is significantly greater than the endogenous GH activity catalysed by the toxin. Figure 5(B) depicts a gel showing the extent of labelling of the Δdph2-eEF2 protein in the presence and absence of toxin. These labelling data provide further evidence that Δdph2-eEF2 can serve as an ADP-ribose acceptor, with the appearance of a faint ϵ-ADP-ribose fluorescent band after reaction with a high toxin concentration for an extended time period (4.5 h) (Figure 5B1, lane 2). By contrast, Figure 5(C) indicates that the H699A mutant of eEF2 could not be labelled under these same conditions (Figure 5C1, lane 2) and hence cannot function as an acceptor of ADP-ribose in the toxin-catalysed reaction. The Δdph2-eEF2 (i.e. Diph699H) protein, which lacks the modified side-chain group (2-position of the imidazole ring of His699), was still able to function, albeit poorly, as an acceptor for ADP-ribose at high toxin concentration. However, the LC–MS/MS analysis did not confirm the existence of any ADP-ribosylation at the imidazole ring of H-699 from the Δdph2-eEF2 samples, probably because of the low abundance of this label on the Δdph-eEF2 protein.
Relative ADP-ribose acceptor activities of the eEF2 mutant proteins
Unfortunately, gel-labelling experiments did not allow quantification of the eEF2 mutant protein ADP-ribose acceptor activity. Table 2 shows the relative ADP-ribose acceptor activities for wild-type and several (nine in total) mutant eEF2 proteins. W669A and L693A proteins exhibited similar acceptor activity to the wild-type protein. Although Trp669 is strictly conserved in eukaryotes (but not in Archaea; Figure 1A) and, indeed, has, on the basis of the X-ray structure of the toxin–eEF2 complex, some interaction with ExoAc , the replacement of Trp669 by alanine did not affect the ADP-ribosylation of eEF2. Likewise, Leu693 is also conserved in eukaryotes; however, it does not seem to participate in the ADP-ribose acceptor activity of eEF2 (Table 1; L693A ADP-ribose acceptor activity is 0.89 relative to the wild-type). In contrast, all of the conserved residues located within the diphthamide-containing loop of eEF2, except Leu693, were sensitive to replacement by alanine, resulting in a marked decrease in the ADP-ribose acceptor activity of the proteins. The single alanine substitutions within the diphthamide loop, except for mutant L693A, resulted in decreases in the relative ADP-ribose acceptor activity that ranged from 47000 to 114000-fold, with the H699A mutant being completely devoid of acceptor activity (Table 2). In an experiment designed to determine the effect of eEF2 binding to toxin–NAD+ upon the extent of strain imposed within the C–N bond of bound NAD+, the GH activity was measured for ExoA (toxin) in the presence of NAD+ and the H699A eEF2 mutant. It was shown that H699A mutant binding with eEF2 increases the GH activity by nearly 3-fold (results not shown), suggesting that additional strain is imposed upon the critical glycosidic bond of the NAD+ substrate when eEF2 docks with the toxin/enzyme. The diphthamide loop within eEF2 contains two alanine residues (Ala695 and Ala697), which could not be studied by ASM, and methyl side chains, which are not likely to be important for interaction with eEF2 or with the enzymes in the diphthamide-specific machinery of the yeast cell (Figure 1A). The R700A mutation within eEF2 was lethal to the yeast cells, and it was concluded that this mutant eEF2 protein must have lost its ability to participate in normal ribosomal function (results not shown).
Conformation of the diphthamide-specific loop in eEF2 structures
To date, seven X-ray structures of eEF2 have been determined, either as the free protein in solution or in a complex with ExoAc [4,5,16], and these vary in resolution from 2.1 to 3.1 Å (1 Å=0.1 nm). An important parameter associated with crystallographic structure determination is a number called the B-factor, or temperature factor, for each atom. In a general sense, B-factors indicate the precision of the atom positions within the X-ray model. Atom positions can be uncertain because of disorder in the crystal from which the structure was determined or due to the inherent mobility of a particular region of the protein (often loops). The B-factors reflect the mobility or flexibility of various parts of the molecule. Values of 60 or greater may imply disorder (for example, free movement of a side chain or alternative side-chain conformations). Values of 20 and 5 correspond to uncertainties of 0.5 and 0.25 Å respectively. The average B-factor values for the eEF2 protein in solution vary from 38 for the eEF2–sordarin structure to nearly 50 and 60 for the ADP-ribose–eEF2 and apo-eEF2 structures [no sordarin (an antifungal agent) or GTP bound] respectively. The structures of the eEF2 protein in complex with Ps. aeruginosa exotoxin A showed B-factors from 46 to 55, indicating that the extent of disorder was relatively consistent throughout the seven X-ray structures of the eEF2 protein. However, the specific normalized B-factor profile for the diphthamide-specific loop within eEF2 revealed some interesting trends (Figure 6). First, the degree of mobility or flexibility of this loop was greatest at its centre, where the diphthamide residue (Diph699) is located. Secondly, the three X-ray structures for uncomplexed eEF2 (apo-eEF2, eEF2–sordarin and ADP-ribose–eEF2) showed the most flexible diphthamide loop, and the four eEF2 structures in complex with toxin has relatively rigid diphthamide loop regions (Figure 6). It is also noteworthy that the overall shapes of the B-factor profiles for all seven eEF2 X-ray structures was generally similar; however, there was a pronounced trough in three of the four eEF2–toxin complexes (ADP-ribose–eEF2–ExoAc, eEF2–βTAD-ExoAc, apo-eEF2–ExoAc) in the loop sequence from Arg685 to Val691 (βTAD is β-methylenethiazole-4-carboxamide–adenine dinucleotide, a non-hydrolysable analogue of NAD+). However, this region of the diphthamide loop does not directly contact the toxin in these complexes , so the origin of this apparent hindered mobility is not known. Although the overall magnitude of the B-factors within the diphthamide loop of the eEF2–toxin complexes is smaller than for the free eEF2 structures, the relative change in the B-factors for the centre of the loop was much greater for the three of the four eEF2–toxin structures (results not shown).
Function of the conserved residues within the diphthamide loop of eEF2
The conserved residues within the eEF2 diphthamide-containing loop may be contributing important and specific interactions during the ADP-ribosylation reaction catalysed by toxin. However, in agreement with a recent report , the conserved residues with the diphthamide loop of eEF2 are also important for dipthamidylation of His699 in yeast eEF2 (Table 1). Our previously described structure for ADP-ribosylated eEF2 indicated that residues within the diphthamide loop of eEF2 make contact with the NAD+ substrate during its covalent modification by toxin . In fact, the conserved loop residues His694 and Asp696 form a triangular network of H-bonds with the ribose of the ADP-ribose moiety attached to the diphthamide residue. Circumspection of the X-ray structure of the diphthamide loop region within eEF2 inspires the notion that this region shows considerable flexibility and that it may function in a similar manner to intrinsically disordered proteins that are design to adapt to a variety of binding partners . It is highly conceivable that, during the various stages of the ADP-ribosylation reaction catalysed by toxin, the diphthamide loop undergoes a partial helix-to-coil transition , moulding itself according to the complementary surface of the toxin/enzyme. Additionally, in a similar manner, this loop may also accommodate the diphthamide biosynthetic enzyme complex during its multi-step synthesis of the C-2 substituent on His699 of yeast eEF2 . This is supported by the relative change in the normalized B-factors for this loop region where the segment from Thr692 to Ile705 shows enhanced relative mobility compared with the baseline regions (residues) of this loop (results not shown). Notably, the X-ray structure of this loop region indicates that the segment from Asp696 to Arg700 possesses partial helical secondary structure that may modulate according to its contact surface in a binding partner, perhaps both during diphthamidylation of His699 and ADP-ribosylation by diphthamide-specific toxins in the cytoplasm. This ‘chameleon’-like behaviour may be an important feature of this loop in facilitating the covalent modification of its resident diphthamide residue during acceptor activity from both diphthamide biosynthetic machinery and exogenous toxins. Thus, these secondary-structure transitions may be important to facilitate the correct positioning of diphthamide for both the diphthamidylation and the ADP-ribosylation reactions. Additionally, one or more of the conserved residues within the diphthamide loop may actively participate in these two reaction mechanisms. However, because of the strict requirement of the C-2 substituent of His699 (diphthamide) for ADP-ribosylation (Tables 1 and 2) , the role of these residues in the toxin-catalysed reaction could not be dissected. Notwithstanding, the conservation of His694 and Asp696 elicits the intriguing possibility that both of these conserved residues participate in the transfer component of the reaction mechanism in a substrate-assisted catalytic mechanism . Single replacements of these residues by alanine would then be expected to impair the reaction in a manner akin to replacement of the toxin/enzyme's active-site residues, namely Glu553 and His440 [37,38]. However, this effect is masked by the loss of the C-2 substituent on His699 (Tables 1 and 2). As secondary-structure breakers, the three consecutive glycine residues (Gly701-Gly702-Gly703 in eukaryotic eEF2) undoubtedly endow the diphthamide-containing loop with a certain degree of conformational flexibility, which may allow this loop to sample multiple conformations, some of which may be prerequisite to the transfer of the ADP-ribose moiety from NAD+ to N-3 of Diph699 (His699). However, because these residues are also important for the diphthamidylation of His699, it was not possible to test this hypothesis (Table 1).
Role of the diphthamide residue
Five proteins have been identified that are responsible for the biosynthesis of diphthamide , implying the importance of this residue in cell function. However, results of diphthamide knockout mutants in yeast have been less than convincing concerning the role of the diphthamide in eukaryotic cell function, including protein synthesis [8,9]. A systematic mutagenesis study of eEF2 Diph699 was conducted by Kimata and Kohno, in which the histidine residue was replaced by (in turn) 19 other amino acids, resulting in eEF2 mutant proteins that could not be ADP-ribosylated by diphtheria toxin . These eEF2 variants were either inactive or caused temperature-sensitive growth in the host yeast cells. Ortiz et al. found that yeast strains harbouring the eEF2 mutants D696A, I698A and H699N, and those lacking the diphthamide modification enzymes, showed increased −1 frameshifting . Our present results strongly suggest that the imidazole ring of Diph699 is essential for the acceptance of the ADP-ribose moiety during the toxin-catalysed reaction. However, the C-2 substituent of the imidazole ring is also an important and major participant in the transferase reaction. In fact, in the absence of this positively charged amidated ammonium propyl 2′-substituent, the naked imidazole ring at position 699 is a very poor acceptor of the ADP-ribose group (Table 2; 114000-fold decrease in ADP-ribose acceptor activity). It is noteworthy that the extent of ribosylation of Δdph2-eEF2 by toxin was very small (1–2% of the wild-type value) and could only be detected by fluorescence methods, but not by MS (Figure 5 and Table 1). Therefore this observation is not likely to represent a significant in vivo activity for mutant eEF2 ADP-ribose acceptor activity.
One possibility for the role of the C-2 substituent on the diphthamide may be to trigger the transferase reaction by interacting with the phosphate(s) of the NAD+, thus alleviating the toxin-induced strain on the scissile glycosidic bond within NAD+, which may serve to direct the N-3 of the diphthamide imidazole towards the oxacarbenium ion in a process that may also involve migration of the electrophile during the nucleophilic substitution reaction instigated by the toxins [5,39]. The replacement of Diph699 by alanine further proved that Diph699 is the only specific ADP-ribosylation site recognized by Ps. aeruginosa ExoA (Table 2; no ADP-ribose acceptor activity for H699A). Since the H699A mutant of eEF2 was completely devoid of ADP-ribose acceptor activity, it must be concluded that the minimal requirement for ribosylation of eEF2 by toxin is an imidazole ring at residue 699 within the diphthamide loop.
Enigmatic function of the diphthamide residue in eukaryotic cells
The biosynthesis of diphthamide requires at least six proteins, with only five having so far been identified . In our experience, knockout of the genes of these biosynthetic proteins did not significantly change the growth phenotype of the yeast cells. Diphthamide is completely conserved throughout eukaryotes and Archaea, but is not found in eubacteria . The physiological function of the diphthamide residue is still unknown; however, it is positioned at the apex of eEF2 domain IV and, when eEF2 is bound to the ribosome in the post-translocational state, this part of eEF2 is close enough to interact with the A-site codon–anticodon area of the small ribosomal subunit . Notably, the diphthamide residue seems to be able to interact with the backbone phosphate groups of two universally conserved adenine residues in helix 44 of the small ribosomal subunit, which is known to have a central role in discriminating against non-cognate tRNA binding in the A site . Recently, Ortiz et al. demonstrated that diphthamide plays a role in maintaining translational fidelity and in preventing a −1 frameshift . Furthermore, the DPH2 gene involved in biosynthesis of diphthamide has been previously identified as the tumour suppressor gene OVCA1, implying that diphthamide might have a fundamental role in regulating progression of the cell cycle .
This work was funded by the Canadian Institutes of Health Research (grant to A. R. M.) and by ORDCF (grant to G. L). We are indebted to Dr Kenji Kohno for the gift of the H699A mutant of eEF2. We thank Dr René Jørgensen (Department of Molecular and Cellular Biology, University of Guelph, Guelph, ON, Canada) for critically reading this manuscript before its submission and for constructive comments. We also thank Ms Dawn White and Mr Gerry Prentice (both at the Department of Molecular and Cellular Biology, University of Guelph) for excellent technical support during the course of this study.
Abbreviations: ϵ-ADP, ethenoadenosine diphosphate; ADPr-eEF2, ADP-ribosylated eukaryotic elongation factor 2; ADPRT, ADP-ribosyltransferase; ASM, alanine-scanning mutagenesis; Diph699, diphthamide residue at position 699; DT, diphtheria toxin; DTT, dithiothreitol; eEF2, eukaryotic elongation factor 2; ExoA, exotoxin A; ExoAc, the catalytic fragment of exotoxin A; 5-FOA, 5-fluoro-orotic acid; GH, NAD+ glycohydrolase; GdmCl, guanidinium chloride; His6, hexahistidine; LC–MS/MS, liquid chromatography–tandem MS; SD, synthetic dropout medium; Trp, λem,max, tryptophan fluorescence wavelength emission maximum; YPD, yeast/peptone/dextrose
- © The Authors Journal compilation © 2008 Biochemical Society