Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Stopped-flow Kinetic Analysis of eIF4E and Phosphorylated eIF4E Binding to Cap Analogs and Capped Oligoribonucleotides: EVIDENCE FOR A ONE-STEP BINDING MECHANISM *

Stopped-flow Kinetic Analysis of eIF4E and Phosphorylated eIF4E Binding to Cap Analogs and Capped... THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 21, pp. 14927–14938, May 26, 2006 © 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. Stopped-flow Kinetic Analysis of eIF4E and Phosphorylated eIF4E Binding to Cap Analogs and Capped Oligoribonucleotides EVIDENCE FOR A ONE-STEP BINDING MECHANISM Received for publication, February 22, 2006 Published, JBC Papers in Press, March 15, 2006, DOI 10.1074/jbc.M601653200 ‡ § ‡1 Sergey V. Slepenkov , Edward Darzynkiewicz , and Robert E. Rhoads From the Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center, Shreveport, Louisiana 71130-3932 and the Department of Biophysics, Warsaw University, Warsaw 02-089, Poland Recruitment of eukaryotic mRNA to the 48 S initiation complex to eIF4G, a protein that also interacts with the RNA helicase eIF4A to is rate-limiting for protein synthesis under normal conditions. promote unwinding of mRNA secondary structure, with the multisub- Binding of the 5-terminal cap structure of mRNA to eIF4E is a unit factor eIF3 to recruit the 43 S initiation complex, and with the critical event during this process. Mammalian eIF4E is phosphoryl- cytoplasmic poly(A)-binding protein to enhance initiation of poly(A)- ated at Ser-209 by Mnk1 and Mnk2 kinases. We investigated the containing mRNAs. Initiation codon recognition is followed by dissoci- interaction of both eIF4E and phosphorylated eIF4E (eIF4E(P)) with ation of eIFs and joining of the 60 S ribosomal subunit to form the cap analogs and capped oligoribonucleotides by stopped-flow elongation-competent 80 S initiation complex. kinetics. For m GpppG, the rate constant of association, k , was eIF4E has been extensively investigated in organisms that range from on dependent on ionic strength, decreasing progressively up to 350 mM yeast to mammals (2–7). Besides translation, eIF4E also functions in nucle- KCl, but the rate constant of dissociation, k , was independent of ocytoplasmic transport of mRNA, sequestration of mRNA in a nontrans- off ionic strength. Phosphorylation of eIF4E decreased k by 2.1–2.3- latable state, and stabilization of mRNA against decay in the cytosol (8–10). on fold at 50–100 mM KCl but had progressively less effect at higher The three-dimensional structures of human, mouse, and Saccharomyces ionic strengths, being negligible at 350 mM. Contrary to published cerevisiae eIF4E have been solved (11–13). The complex of full-length evidence, eIF4E phosphorylation had no effect on k . Several human eIF4E with m GpppA is bell-shaped, with the cap analog situated in off observations supported a simple one-step binding mechanism, in a deep slot in the concave surface of the protein. The cap-binding pocket contrast to published reports of a two-step mechanism. The kinetic consists of separate recognition components for the m G base, the triphos- function that best fit the data changed from single- to double-expo- phate moiety, and the second nucleoside residue. The alkylated base is nential as the eIF4E concentration was increased. However, meas- stacked between Trp-56 and Trp-102. (Amino acid numbers refer to the uring k for dissociation of a pre-formed eIF4Em GpppG complex human sequence (14).) Glu-103 and Trp-102 form H-bonds with the off suggested that the double-exponential kinetics were caused by dis- N-1, N-2, and O-6 protons of m G. Trp-56 also interacts directly with sociation of eIF4E dimers, not a two-step mechanism. Addition of a the ribose moiety, and Arg-157 and Lys-162 interact directly with the - 12-nucleotide chain to the cap structure increased affinity at high and -phosphate oxygen atoms. The second nucleoside moiety of ionic strength for both eIF4E (24-fold) and eIF4E(P) (7-fold), pri- m GpppA is fixed by the flexible C-terminal loop. To date, there have marily due to a decrease in k . This suggests that additional stabi- been no structures reported for eIF4E in complex with capped mRNA off lizing interactions between capped oligoribonucleotides and eIF4E, or even short oligoribonucleotides. which do not occur with cap analogs alone, act to slow dissociation. Cap analogs bind to eIF4E in a tight complex, a step that has been studied primarily by equilibrium techniques (15–33). Intrinsic Trp fluorescence quenching of N-terminal truncated mouse eIF4E (residues 28–217) by The efficiency of mRNA translational initiation is strongly enhanced titration with cap analogs indicates that the free energy of m G stacking and hydrogen bonding is separate from the free energy of triphosphate chain by the 5-terminal cap, m GpppN (1). The cap specifically binds to eIF4E, which may be the first canonical initiation factors to interact interactions (26). Electrostatically steered eIF4E-cap analog association is with mRNA during its recruitment to the ribosome. eIF4E in turn binds accompanied by hydration of the complex and a shift in ionic equilibria. The kinetics of cap analog binding to wheat eIF(iso)4F (34) and mouse eIF4E-(28–217) (35, 36) have also been studied with rapid mixing tech- * This work was supported by National Institutes of Health Grants 2R01GM20818 (to niques. The authors of these studies have interpreted their data as indicat- R. E. R.) and 1R03TW006446 (to R. E. R. and E. D.) and Howard Hughes Medical Institute ing a two-step binding reaction. Another group measured the equilibrium Grant 55005604 (to E. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked dissociation constant, K , and the rate constant of dissociation, k , for d off “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 human eIF4E by surface plasmon resonance (SPR) (37). Surprisingly, the To whom correspondence should be addressed: 1501 Kings Highway, Shreveport, LA 71130-3932. Tel.: 318-675-5161; Fax: 318-675-5180; E-mail: [email protected]. experimentally determined k values and the calculated k values (K off on d The abbreviations used are: eIF4E, eukaryotic initiation factor 4E; ARCA, anti-reverse cap k /k ) differed from those reported for the stopped-flow studies (35, 36) off on analog; DTT, dithiothreitol; eIF4E(P), human eIF4E phosphorylated at the physiological by 2–3 orders of magnitude. site, which is Ser-209 for the human protein; IEF, isoelectric focusing; k , the observed obs rate constant; k , the rate constant for dissociation; k , the rate constant for association; off on Mammalian eIF4E is phosphorylated at Ser-209 (38, 39). Although 7 1 3 7 m GpppG, P -7-methylguanosine-5 P -guanosine-5 triphosphate; m GpppN, same as 7 7,3-O 1 several eIF4E kinases have been reported, the strongest evidence points m GpppG except with any nucleotide base in place of G; m GpppG, P -3-O,7- dimethylguanosine-5 P -guanosine-5 triphosphate; Mnk, mitogen-activated pro- to Mnk1 and Mnk2 as the physiological kinases (40, 41). Mnk is acti- tein kinase-interacting kinase; RU, response units; SPR, surface plasmon resonance; vated via the extracellular signal-regulated kinase (ERK) and p38 path- GST, glutathione S-transferase; MOPS, 4-morpholinepropanesulfonic acid; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; V, volt(s). ways in response to mitogens, cytokines, or cellular stress (40–42). MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14927 This is an Open Access article under the CC BY license. Kinetic Analysis of Cap Binding to eIF4E Several types of observations link increased eIF4E phosphorylation with pET11d-eIF4E (64). Protein expression was induced with 0.4 mM iso- increased rates of protein synthesis, including stimulation of cultured propyl--D-thiogalactopyranoside overnight at 15 °C. We found that cells with mitogens (43–45), use of the Mnk inhibitor CGP57380 (46, these conditions (lowering the isopropyl--D-thiogalactopyranoside 47), sequence variants of eIF4E that cannot bind eIF4G (48), shut-off of concentration and temperature) increased the yield of soluble eIF4E cellular protein synthesis by adenovirus infection (49), and induction of compared with the published procedure (64) and produced 7–10 mg/ apoptosis (50). The opposite conclusion, that eIF4E phosphorylation liter of cell suspension. Cells (3 g) were lysed with a French press in 25 does not affect the rate of protein synthesis, or even decreases it, has ml of buffer A (50 mM HEPES/KOH, pH 7.6, 0.3 M KCl, 10 mM EDTA, 1 been drawn from studies with Mnk overexpression (51), cell-free trans- mM DTT, 1% streptomycin sulfate, and protease inhibitor mixture). lation (52), recovery of cultured cells from hypertonic stress (53, 54), and After centrifugation at 32,000 g for 40 min at 4 °C, the soluble fraction Indinavir, a human immunodeficiency virus protease inhibitor (55). was diluted 1:1 with 100 M GTP and loaded onto a 3-ml m GTP- Whole animal studies have also generated conflicting results. Trans- Sepharose column. This and other chromatographic steps were moni- genic Drosophila that express a non-phophorylatable form of eIF4E are tored spectrophotometrically at  . The column was washed with 280 nm small, have morphological defects, and are less viable (56). The Drosoph- 40 column volumes of buffer B (50 mM HEPES/KOH, pH 7.6, 0.2 M KCl, ila homolog of Mnk, Lk6, is dispensable under a high protein diet, but its 1mM EDTA, and 1 mM DTT). eIF4E was eluted with buffer B except loss causes growth reduction when amino acids in the diet are restricted that the KCl concentration was 0.5 M KCl. Use of KCl for elution rather (57, 58). Yet when Mnk1 and Mnk2 are knocked out in mice, no phe- than the more traditional m GTP (65) prevents formation of tight eIF4E notype is observed (59). Thus, at the levels of protein synthesis, cell complexes with the cap analog (66), which can affect the measurement growth, and intact animal physiology, the roles of Mnk and eIF4E phos- of kinetic parameters for the binding reaction. eIF4E was at least 95% phorylation are controversial. pure as judged by SDS-PAGE (see “Results”). For eIF4E purification The current study was motivated by several gaps in our understanding of from E. coli cells and Mnk purification from HEK 293 cells, protein eIF4E-cap interactions as well as discrepancies in the published literature. concentrations were determined with the protein assay reagent from First, protein synthesis is a dynamic process and, during each round of Bio-Rad, in which bovine serum albumin was used as standard. For initiation, the mRNA cap is presumably both bound and released by eIF4E. kinetic experiments, protein concentrations were determined spectro- 1 1 A full understanding of the biochemical consequences of eIF4E phospho- photometrically, assuming   53,400 M cm at pH 7.2. 280 nm rylation therefore requires a knowledge of k and k values for both phos- on off eIF4E Phosphorylation—Recombinant mouse Mnk kinase was pro- phorylated and unphosphorylated eIF4E (eIF4E and eIF4E(P)). Second, it is duced as described previously (42) but with modifications. Plasmids pEBG- not known whether the discrepancies in kinetic values noted above (35–37) Mnk2 and pEBG-Mnk1T2A2 were generously provided by Christopher result from the methodologies employed (SPR versus stopped-flow), the Proud (University of British Columbia). These encode, respectively, wild- forms of eIF4E studied (full-length versus N-terminal truncated), or the type Mnk2 fused to glutathione S-transferase (GST::Mnk2) and an inac- methods of producing eIF4E(P) (enzymatic phosphorylation versus intein- tive variant, GST::Mnk1T2A2 (41). HEK 293T cells, grown in 10-cm mediated ligation). Third, the k for eIF4E(P)-cap analog association has on dishes in Dulbecco’s modified Eagle’s medium supplemented with 10% not yet been measured directly. Fourth, kinetic parameters measured by the fetal bovine serum, were transfected with plasmids in complex with stopped-flow technique have not been reported for the binding of capped Lipofectamine 2000 (Invitrogen). After 44 h, cells were harvested and oligoribonucleotides to eIF4E. Fifth, all previous studies of eIF4E interac- lysed at 4 °C by incubation with buffer C (10 mM HEPES/KOH, pH 7.4, tions with capped oligonucleotides (20, 22, 37, 60) have utilized a mixture of 50 mM NaF,2mM EDTA, 2 mM sodium orthovanadate, 0.1% 2-mercap- normally capped and reverse-capped oligoribonucleotides, the latter of toethanol, 1% Triton X-100, and protease inhibitor mixture) at a ratio of which do not bind eIF4E. Sixth and finally, the proposed two-step mecha- cell pellet to buffer C of 1:8 (w/v) over 30 min with rotation. The sus- nism for cap binding to eIF4E (26, 34, 35) is at variance with observations for pension was centrifuged for 30 min at 20,000  g, and GST::Mnk2 was a viral cap-binding protein (61, 62). purified from the supernatant on glutathione-Sepharose 4B equili- We report here stopped-flow kinetic results for full-length human eIF4E brated with buffer D (20 mM MOPS/KOH, pH 7.4, 20 mM KCl, 15 mM and eIF4E(P) binding to cap analogs and capped oligoribonucleotides, the MgCl , 0.5 mM EDTA, 1 mM DTT, 25 mM -glycerolphosphate, 1 mM latter being capped entirely in the correct orientation. Our findings differ sodium orthovanadate, and 5% glycerol) at a supernatant:Sepharose from those of previous reports with regard to the magnitude of kinetic ratio of 10:3 (v/v). Glutathione-Sepharose beads with bound GST::Mnk2 constants, the effects of eIF4E phosphorylation on k and k , and the on off and GST::Mnk1T2A2 were stored at 80 °C. mechanism of cap binding. With regard to the latter, our data support a Recombinant eIF4E was phosphorylated in vitro essentially as described one-step rather than two-step binding model. Furthermore, the stopped- (37) but with some modifications. Recombinant eIF4E (225 g) was incu- flow results provide a kinetic basis for enhanced binding of eIF4E to capped bated at room temperature (23 °C) in a 0.5-ml reaction mixture containing oligoribonucleotides compared with cap analogs. 150 l (packed volume) of glutathione-Sepharose-bound GST::Mnk2 and 500 M ATP in buffer D for 6 h with rotation. The kinase was removed by EXPERIMENTAL PROCEDURES centrifugation, and then eIF4E(P) was passed over a PD-10 desalting col- umn (Amersham Biosciences) to separate it from ATP and to transfer it Materials—All common reagents were of analytical grade and were 7 7 purchased from Sigma unless otherwise stated. m GTP, m GTP-Sepha- into an appropriate buffer for kinetic experiments. In parallel experi- ments, eIF4E was mock-phosphorylated with GST::Mnk1T2A2. The rose 4B, and glutathione-Sepharose 4B were purchased from Amer- efficiency of eIF4E phosphorylation was determined by isoelectric sham Biosciences. Protease inhibitor mixture was from Roche Diagnos- 7 7,3-O tics. The syntheses of m GpppG and m GpppG were performed as focusing on CleanGel IEF polyacrylamide gels (Amersham Biosciences) described previously (63). The concentrations of dinucleotide cap ana- in the presence of 8 M urea, 30 mM CHAPS, and 2% Pharmalyte 3–10 3 1 log solutions were determined by absorbance (  22.6 10 M according to the manufacturer’s instructions. IEF was conducted at 255 nm cm at pH 7.0). 15 °C on a Multiphor II apparatus for horizontal separation equipped eIF4E Expression and Purification—Full-length wild-type human with a cooling platform and MultiDrive XL power supply (Amersham eIF4E (14) was expressed in Escherichia coli BL21 (COE) pLys(S) from Biosciences). Protein bands were stained with Coomassie Blue. In pre- 14928 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E liminary experiments to determine the optimal time for enzymatic 1000 data points throughout the reaction (50–200 ms). Fluorescence phosphorylation, we also used [- P]ATP. changes were examined up to 200 s. The binding reaction was investigated under pseudo first-order condi- Synthesis of Oligoribonucleotides—T7 RNA polymerase was prepared as previously described (67). The template consisted of two comple- tions ([protein] [ligand]) unless otherwise noted. For most experiments, mentary oligodeoxyribonucleotides (Integrated DNA Technology Inc., concentrations of protein and ligand in the syringes of the stopped-flow instrument were 0.1–0.2 and 1–20 M, respectively. Samples were Coralville, IA), 5-TAATACGACTCACTATAG-3 and 5-TTTTT- degassed prior to loading into the syringes. Reactions were initiated by ATGCGCCTATAGTGAGTCGTATTA-3, which contain the T7 pro- moter followed by a single-stranded 12-nucleotide 5-overhang. This mixing equal volumes of protein and ligand at 25 °C in buffer E except containing 50, 100, 150, or 350 mM KCl as noted. The temperature was sequence was chosen because it gave the highest yield of several tested controlled with a circulating external water bath (T  0.2 °C). The (68). The oligodeoxynucleotides, each at 183 M in 20 mM Tris-HCl, pH stopped-flow traces shown under “Results” are the average of 5–15 individ- 7.5, were annealed by heating at 90 °C and slow cooling to room tem- ual traces. Control experiments on specificity of binding were conducted perature. Transcription reactions were performed as described previ- with GpppG and with the uncapped form of the oligoribonucleotide. For ously (68) with some modifications. Reaction mixtures (1 ml) con- stopped-flow experiments where the m GpppG concentration was limit- tained 40 mM HEPES/KOH, pH 7.9, 10 mM MgCl ,2mM spermidine, ing (0.2–0.4M), the protein concentration was varied in the range of 1–10 10 mM NaCl, 10 mM DTT, 0.1 mg/ml bovine serum albumin, 1000 M. Dissociation of the pre-formed eIF4Em GpppG complex was fol- units/ml RNA Guard (Amersham Biosciences), 2.5 mM each of ATP, lowed by measuring the increase in intrinsic Trp fluorescence when equal CTP, UTP, and GTP, 40 g/ml DNA template, and 100 gof volumes of the complex and buffer were mixed in the stopped-flow cell at T7 RNA polymerase. Capped oligoribonucleotides were produced 150 mM KCl. In this case, eIF4E and m GpppG concentrations in the by lowering the GTP concentration to 0.5 mM and including 7,3-O syringe were equal (2 M). m GpppG, an “anti-reverse cap analog” (ARCA), at 3 mM, Curve Fitting and Error Analysis—Stopped-flow traces representing which prevents heterogeneity in synthesized oligoribonucleotides binding of cap analogs or a capped oligoribonucleotide to eIF4E or due to reversed incorporation (63). The binding affinities of 7 7,3-O eIF4E(P) were analyzed using Curfit, a curve-fitting program that uses a m GpppG and m GpppG for eIF4E are the same within 0.6% Marquardt algorithm (71). Data were fit to both single- and double- (69). Reaction mixtures were incubated at 35 °C for 2 h, after which exponential functions. For single-exponential fits, the DNA template was digested with DNase RQ1 (Promega). Oli- goribonucleotides were purified by size-exclusion chromatography F(t)Fexp(k t) F (Eq. 1) obs on Bio-Gel P6 columns with diethylpyrocarbonate-treated water as eluent. The purity of oligoribonucleotides was determined by PAGE where k is the observed first-order rate constant,F is the amplitude, obs on 20% gels containing 8 M urea (70). The A /A ratio for oligori- and F is the final value of fluorescence. For double-exponential fits, 260 280 ∞ bonucleotides was 1.85. Concentrations (in g/ml) were determined F(t)F exp(k t) F exp(k t) F (Eq. 2) 1 obs1 2 obs2 as 33  A . 260 nm Steady-state Fluorescence Measurements—A StrobeMaster lifetime where k and k are the observed rate constants for the first and obs1 obs2 spectrometer (Photon Technology Inc., South Brunswick, NJ) with the second components of a double-exponential reaction, respectively, and SE-900 steady-state fluorescence option was used to obtain fluorescence F and F , are the amplitudes for the first and second components of 1 2 emission spectra of eIF4E and eIF4E(P) in the presence or absence of a double-exponential reaction, respectively. An assessment of each fit m GTP. Fluorescence was monitored at an excitation wavelength of 295 was made from the residuals, which measure the differences between nm (2-nm bandwidth) and emission bandwidth of 5 nm. Spectra were actual data and the calculated fit. These are designated  in the case of recorded with a photon counting detector (Photon Technology Inc., model single-exponential fits and  in the case of double-exponential fits. The 710) from 305 to 400 nm at 25 °C. Samples were maintained in a quartz program KaleidaGraph (Synergy Software, Reading, PA; version 3.06) cuvette (1-cm path length) with constant stirring at concentrations of both was used for least-squares fitting of data with linear equations and deter- protein and m GTP of 1M in buffer E (50 mM HEPES/KOH, pH 7.2, 1 mM mination of standard errors for parameters obtained from the fits. Equi- DTT, 0.5 mM EDTA, and 50 mM KCl). Under these conditions, the K for librium constants are reported either as dissociation constants (K )or the m GTP-eIF4E interaction is 9nM (26). Thus, essentially all of the association constants (K  1/K ). a d m GTP is bound to eIF4E. The degree of quenching of intrinsic Trp fluo- RESULTS rescence (Q) is an indicator of both purity and native conformation of eIF4E. For mammalian eIF4E, Q is 65% (26). The preparations used in Preparation of eIF4E and eIF4E(P)—The objective of the current this study had Q  64  4% for eIF4E and Q  59  5% for eIF4E(P), study was to apply fast kinetic techniques to elucidate the mecha- indicating that both proteins were essentially native. nism of eIF4E-cap interactions, with particular emphasis on the Kinetic Methods—Rapid kinetic measurements of eIF4E and effect of three parameters: phosphorylation of eIF4E, ionic strength, eIF4E(P) interaction with free dinucleotide cap analogs (m GpppG and linkage of the cap to an oligoribonucleotide chain. This required 7,3-O and m GpppG) and an ARCA-capped oligoribonucleotide were preparation of milligram quantities of highly purified, native eIF4E made with a stopped-flow spectrometer with fluorescence detection and eIF4E(P). In previous reports in which recombinant eIF4E was (Applied Photophysics Ltd., Leatherhead, UK, model SX.18MV). The used for kinetic studies, the protein was purified from inclusion bod- dead time of the instrument was 1.2 ms. The excitation wavelength was ies solubilized with either urea (37) or guanidine hydrochloride (35, 295 nm with 2.3-nm bandwidth, and stray excitation radiation was elim- 36) and then renatured by dialysis. It is possible to calculate what inated with a 320-nm Oriel long-pass filter. Excitation occurred through fraction of a given eIF4E preparation is in the native state by deter- a 2-mm path in the stopped-flow optical cell, and emission was meas- mining the quenching of intrinsic Trp fluorescence (Q) by steady- ured through a 10-mm path. The instrument time constant was 0.5% of state fluorescence spectroscopy (see “Experimental Procedures”). In the reaction half-time. For very fast reactions, no in-line filtering was the only case where this was reported, 57  16% of the renatured used. An oversampling option of the instrument permitted us to collect protein preparation was found to be active, presumably due to MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14929 Kinetic Analysis of Cap Binding to eIF4E imperfect refolding in vitro (35). To avoid this problem, we purified cedures”). Another possible source of inactive protein is the pres- human eIF4E exclusively from the soluble fraction of E. coli lysates. ence of contaminating cap analog in complex with eIF4E (66), due to The protein appeared as a homogeneous band by SDS-PAGE (Fig. the conventional method of eluting eIF4E from m GTP-Sepharose 1A). eIF4E prepared in this way had a value of Q indicating that it with m GTP (65). Such eIF4E preparations contain up to 60% bound existed entirely in the native conformation (see “Experimental Pro- m GTP, which cannot be removed by dialysis or ion exchange chro- matography (26). To avoid this problem, we eluted the protein from 7 7 m GTP-Sepharose with 0.5 M KCl rather than m GTP. eIF4E(P) was prepared by in vitro phosphorylation with activated mouse Mnk2. It has previously been shown that this kinase phospho- rylates human eIF4E exclusively at Ser-209 in vitro (41, 42), which is the same as the in vivo site (38, 39). The degree of phosphorylation was measured by two methods: IEF and calculating the moles of phosphate incorporated per mole of eIF4E from the specific activity of [- P]ATP in the kinase reaction. Both methods revealed that phosphorylation was essentially stoichiometric (Fig. 1B and data not shown). Both untreated and kinase-treated proteins appeared on the IEF gels as doublets, the minor form constituting 11% of the total. The minor form has been observed previously in preparations of both phosphorylated and unphosphorylated eIF4E (72, 73). The structural basis for the heteroge- neity is unknown. It may represent post-translational modification, e.g. incomplete N-terminal acetylation (74), or it may be an artifact of IEF. For instance, the concentration of CN in the 8 M urea used in the IEF gel can reach 20 mM and can carbamylate NH groups in proteins (75), removing one positive charge. Regardless of the sources of microhet- FIGURE 1. Characterization of purified recombinant human eIF4E and eIF4E(P). A, erogeneity, both forms had nearly the same molecular mass (Fig. 1A), SDS-PAGE (10% gel) of eIF4E after column chromatography on m GTP-Sepharose. The were able to bind m GTP-Sepharose, and were stoichiometrically phos- mobilities of molecular mass markers are indicated on the left. B, phosphorylation of eIF4E as monitored by IEF. Recombinant human eIF4E was treated with in vivo activated phorylated (Fig. 1B). eIF4E(P) prepared in this way had a value of Q GST::Mnk2 and analyzed by IEF in the range of pH 3–10, as described under “Experimen- indicating that it existed entirely in the native conformation (see “Exper- tal Procedures.” The polarity during IEF is indicated by and. In both A and B, protein is visualized by Coomassie Blue staining. imental Procedures”). FIGURE 2. Kinetics of the m GpppG-induced decrease in intrinsic Trp fluorescence of eIF4E and eIF4E(P) as monitored by stopped-flow fluorescence. A, reaction of 0.1 M eIF4E with m GpppG at 350 mM KCl and 25 °C (upper panel). Solid lines represent fitting of the data with Equation 1. Fluorescence is measured in volts (V). Traces are offset vertically 7 1 to allow each to be seen individually. For m GpppG at 1, 3, 5, and 9 M, k was 118, 176, 229, and 362 s , respectively. Residuals for fits to Equation 1 are shown in the lower panels. obs 7 1 B, same as A except that eIF4E(P) was used. For m GpppG at 1, 3, 5, and 9 M, k was 112, 164, 236, and 376 s , respectively. A control reaction for both proteins was performed with obs 3 7 9 M GpppG (top trace), which binds eIF4E with an affinity that is 10 -fold lower than that of m GpppG (26). 14930 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E Association of eIF4E with m GpppG—We initially attempted to one-step process, k is predicted to vary linearly with cap analog con- obs measure kinetic parameters for the interaction of m GTP, the most centrations (76) (Equation 3). commonly used cap analog, with eIF4E and eIF4E(P) by stopped-flow k  k [m GpppG] k (Eq. 3) experiments but found that rate of fluorescence change was too fast to obs on off be followed. We therefore turned to m GpppG, which binds eIF4E with 7 7 Also, under pseudo first-order conditions, the binding reaction is not only 7% the affinity of m GTP (26). However, m GpppG is more similar sensitive to small amounts of inactive protein in the preparation. to the natural mRNA cap than m GTP because of the second nucleoside The rapid mixing of eIF4E with m GpppG resulted in a decrease in moiety. We measured the reaction over a range of m GpppG concen- intrinsic Trp fluorescence that was dependent on m GpppG concen- trations under pseudo first-order reaction conditions, where [eIF4E] 7 tration (Fig. 2A, upper panel), the apparent rate constant of the [cap analog]. The first stopped-flow study of eIF4E-m GpppG interac- decrease being k . By contrast, there was no decrease in fluores- obs tions (34), which utilized wheat eIF(iso)4F, was also conducted under cence with GpppG (top trace), which binds eIF4E with an affinity pseudo first-order conditions. A later study of mouse eIF4E-(28–217) 3 7 that is 10 -fold lower than that of m GpppG (26). The data were fit (35) used second-order reaction conditions, i.e. the concentrations of with a single-exponential function (Equation 1) (solid lines). The eIF4E and cap analog were approximately equal, and both were varied. residuals, representing the deviation between the calculated and The latter experimental approach makes treatment of the data more actual data, indicate that the single-exponential function fit the complicated and can potentially cause artifacts (see below). Under points over the entire time of the measurement (lower panels). pseudo first-order conditions, if the association reaction is a simple Treatment of the data using a double-exponential function (Equa- tion 2) did not improve the fit (not shown). We followed the reaction over 200 s to test whether there was a second slow phase but saw no evidence for one (not shown). The binding reaction of eIF4E(P) also was probed over the same range of m GpppG concentrations (Fig. 2B, upper panel). The traces again followed a single-exponential function over all cap analog concentra- tions. The residuals did not vary over the time course (lower panels) nor were they diminished by a double-exponential fit (not shown). To ensure that any differences between eIF4E and eIF4E(P) were not an artifact due to nonspecific changes occurring during the 6-h kinase reaction, we also conducted a mock kinase reaction in which the active GST::Mnk2 was replaced with GST::Mnk1T2A2 (77) (see “Experimen- tal Procedures”). The k values for binding of m GpppG were indis- obs tinguishable between unphosphorylated eIF4E and mock-phosphoryl- ated eIF4E (data not shown). Plots of k versus m GpppG concentration for eIF4E and obs eIF4E(P) were linear (Fig. 3). The slope and y intercept are k and on k , respectively (Equation 3). The values obtained for k for both off on 6 1 1 eIF4E and eIF4E(P) were in the range of 33–292  10 M s , depending on the ionic strength, and those obtained for k were off 70–87 s (Table 1). Such high values indicate that equilibrium is reached rapidly and that slower methods of obtaining kinetic con- stants may not yield accurate results (see “Discussion”). K values calculated from these kinetic constants ranged from 0.24 to 2.48 M, again depending on the ionic strength (Table 1). These are somewhat higher than K values obtained by equilibrium methods (26, 32). For instance, at 100 mM KCl, the kinetic approach yielded K  0.45 M for eIF4E (Table 1), whereas the equilibrium approach yielded 0.135 M in one study (26) and 0.08 M in another (32). For eIF4E(P) at 100 mM KCl, kinetic measurements gave K  1.06 M, whereas equilib- FIGURE 3. Dependence of k on m GpppG concentration as a function of KCl con- obs rium measurements gave K  0.172 M (32). The reason for these centration. k obtained in experiments similar to those shown in Fig. 2 at the indicated obs 7 discrepancies is not known, but it should be noted that, besides the KCl concentrations is plotted versus [m GpppG] for eIF4E (panel A) and eIF4E(P) (panel B). Data are fit with Equation 3. Numerical values for k , k , and K are given in Table 1. method of K determination, the protein preparations were differ- on off d TABLE 1 Kinetic constant for association and dissociation of eIF4E and eIF4E(P) with m GpppG obtained from pre-steady-state experiments Data are obtained from experiments similar to Fig. 3, with application of Equation 3. k  10 k K on off d KCl concentration eIF4E eIF4E(P) eIF4E eIF4E(P) eIF4E eIF4E(P) 1 1 1 mMM s s M 50 292  16 138 870  18 73  12 0.24  0.06 0.53  0.09 100 184  10 80 483  23 85  12 0.45  0.13 1.06  0.16 150 102 756 287  15 79  6 0.85  0.16 1.41  0.12 350 33 234 282  12 72  8 2.48  0.39 2.12  0.27 MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14931 Kinetic Analysis of Cap Binding to eIF4E rate of caps with eIF4E but not the dissociation rate. This contradicts results obtained by SPR (37) (see “Discussion”). The KCl concentration dependence of K , as calculated from these kinetic parameters, closely resembled that of k (Fig. 4B). K values on a decreased 10-fold for eIF4E and 4-fold for eIF4E(P) from 50 to 350 mM KCl. The difference in K values between eIF4E and eIF4E(P) was most pronounced (2.2–2.4-fold) at 50 and 100 mM KCl. This was progressively eliminated at higher KCl concentrations, and by 350 mM, there was no difference in K values between eIF4E and eIF4E(P). The Apparent Kinetic Mechanism of m GpppG Binding Depends on eIF4E Concentration—The linearity of the k versus m GpppG plots obs (Fig. 3) is consistent with a one-step, second-order reaction mechanism, on 7 7 eIF4E m GpppG| -0 eIF4Em GpppG* off SCHEME 1 where the asterisk (*) denotes a reduction in Trp fluorescence of eIF4E in complex with m GpppG. This conflicts with published studies indi- cating that cap binding to eIF4E occurs in a two-step process (26, 34, 35). The authors of one of these reports (35) found that the experimental data could be fit better with a double- than a single-exponential func- tion. They attributed the slow phase to a second step in the binding reaction, in which eIF4E fluorescence is further reduced, k k 1 2 7 7 7 eIF4E m GpppG| -0 eIF4Em GpppG*| -0 eIF4Em GpppG** k k 1 2 SCHEME 2 where the asterisk (*) denotes a complex with reduced fluorescence relative to free eIF4E, and the double asterisk (**) denotes a different complex with even lower fluorescence. The mechanisms represented by Schemes 1 and 2 can be distinguished experimentally (see below). Under pseudo first-order conditions, the kinetic mechanism and parameters for a simple association between eIF4E and m GpppG FIGURE 4. Dependence of k and K for binding of m GpppG to eIF4E and eIF4E(P) on a (Scheme 1) theoretically do not depend on the choice of which com- on KCl concentration. Data for eIF4E (open circles) and eIF4E(P) (filled circles) are derived from Table 1. ponent is limiting. For the experiments shown in Figs. 2–4, the eIF4E concentration was limiting and the m GpppG concentration was varied. We measured the same association reaction but at a limit- ent: full-length human eIF4E phosphorylated enzymatically (Table ing m GpppG concentration (0.1 M) and variable eIF4E concen- 1) versus truncated mouse eIF4E phosphorylated by intein technol- trations, so that [m GpppG]  [eIF4E]. To our surprise, the kinetics of ogy (26, 32). Despite these differences in K values, the change in K d d the reaction under these conditions were different. As the eIF4E for eIF4E versus eIF4E(P) was similar for kinetic and equilibrium concentration was increased from 0.5 to 5 M, the stopped-flow determinations: a 2.35-fold increase in K by kinetic methods (Table traces became increasingly biphasic. Representative results for 0.5 1) and a 2.15-fold increase by equilibrium methods (32). and 2 M eIF4E are shown in Fig. 5, A–D. The same data for [eIF4E] The Effect of eIF4E Phosphorylation on k Is Diminished at High Ionic on 0.5 M are presented in Fig. 5, A and C, but with either single- or Strength—It is known that electrostatic forces play an important role double-exponential fits, respectively. The points were fit well by a sin- in determining K for eIF4E-cap analog interactions. Specifically, the 1 gle-exponential function (Equation 1), with k  71  4s and F obs affinity for cap analogs of natural rabbit reticulocyte eIF4E (16) and 0.0155 V (Fig. 5A). Application of a double-exponential function (Equa- recombinant mouse eIF4E-(28 –217) (26, 32) decreases with increas- tion 2) did not improve the fit (Fig. 5C). By contrast, at [eIF4E] 2.0 M, ing ionic strength. To determine the basis for the effect of KCl on K , d the data were fit poorly by a single-exponential function (Fig. 5B), we plotted k and k from Table 1 against KCl concentrations (Fig. on off whereas they were fit well by a double-exponential function (Fig. 1 1 4A). Over a range of KCl concentrations from 50 to 350 mM, k on 5D), with k  205  20 s and k  8.1  4s . The ampli- obs1 obs2 decreased 9-fold for eIF4E and 4-fold for eIF4E(P), whereas k tudes for the fast and slow phases were similar, 0.080 and 0.078 V, off did not change significantly (inset). Furthermore, there was no sta- respectively. The relative improvement of the fit with a double- tistical difference in k between eIF4E (open symbols) and eIF4E(P) 2versus single-exponential function, (   )/ , was 4. As the off s d d (filled symbols). Thus, phosphorylation diminishes the association eIF4E concentration was raised, the amplitude of the slow phase 14932 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E 7 7 7 FIGURE 5. Effect of eIF4E concentration on the kinetics of m GpppG binding to eIF4E, and the kinetics of m GpppG dissociation from a pre-formed eIF4Em GpppG complex. A, kinetic traces were obtained as described in the legend to Fig. 2 after mixing 0.1 M m GpppG with 1 M eIF4E at 350 mM KCl. (The final concentration of eIF4E was 0.5 1 7 M.). Data are fit with a single-exponential function, F(t)  0.0155 exp(71 s  t)  0.0135. B, same as A except that 0.2 M m GpppG and 4 M eIF4E were used. Data are fit with 1 1 a single-exponential function, F(t) 0.0871 exp(57.5 s  t) 0.0227. C, the same data points in A are shown but fit with a double-exponential function, F(t) 0.0092 exp(142 s 1 1 1 t) 0.00797 exp(39 s  t) 0.0143. D, the same data points in B are shown but fit with a double-exponential function, F(t) 0.08 exp(205 s  t) 0.078 exp(8.1 s  t) 0.012. 7 7 E, dissociation of the eIF4Em GpppG complex. eIF4E and m GpppG (2 M each) were mixed in one syringe of the stopped-flow apparatus in buffer E containing 150 mM KCl. After incubation for 5 min at 25 °C, the eIF4Em GpppG complex was diluted in the stopped-flow cell with an equal volume of the same buffer. The trace follows Equation 1 (solid line) with k being equal to 178  9s . obs progressively increased in relation to that of the fast phase (Table 2). eIF4E ^ (eIF4E) Our observation that the mechanism appears to change from a one- SCHEME 3 step to a two-step process as the eIF4E concentration is increased might be explained by formation of inactive eIF4E dimers (or multimers). We When the concentration of active eIF4E is lowered due to interaction suggest that there is a pre-existing equilibrium that becomes important with m GpppG during the stopped-flow experiment, the equilibrium at high protein concentrations. MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14933 Kinetic Analysis of Cap Binding to eIF4E shifts and the dimers slowly dissociate, yielding active monomers that- Fig. 5E shows a representative dissociation trace obtained at 150 mM can then bind ligand. KCl. The trace followed single-exponential kinetics during 20 s of meas- urement (only the first 0.1 s are shown) and yielded k  178  9s . obs on We used K determined in Table 1 to calculate the free concentrations 7 7 (eIF4E) ^ eIF4E m GpppG| -0 eIF4Em GpppG* 7 of eIF4E and m GpppG after dilution with buffer. Using these data and off k determined in Table 1, we calculated k for the reaction in the on off SCHEME 4 reverse direction according to the following equation. This dissociation reaction produces a second, slow phase of eIF4E- k  k  k ([cap]  [eIF4E] ) (Eq. 4) off obs on free free m GpppG interaction. Scheme 4 could possibly explain our results (Fig. 5D and Table 2) as well as those of previous studies, which were con- This yielded a value of k  84  18 s . This is the same, within off ducted at molar concentrations of either wheat eIF(iso)4F (34) or mouse experimental error, as k determined from the reaction conducted in off eIF4E-(28–2-17) (35) that were at least 5-fold higher than in Figs. 2–4. the forward direction, 87  15 s (Table 1). These results are consist- Thus, there are two alternative hypotheses to explain the double-expo- ent with a one-step but not two-step mechanism for cap binding to nential kinetics observed at high eIF4E concentrations: dissociation of eIF4E. inactive dimeric eIF4E (Scheme 4) and a two-step binding reaction Reaction of eIF4E and eIF4E(P) with a Capped Oligoribonucleo- (Scheme 2). tide—Although most studies of cap binding to eIF4E have been per- At 50 mM KCl, the authors of one study (35) calculated k  30 1 formed with mono- or dinucleotide cap analogs, the more physiologi- 1 1 1 s and k  1s .At150mM KCl, they obtained k  50 s and 2 1 cally relevant ligand is a capped RNA molecule. Unfortunately, use of k  6s . Thus, the step defined by k in Scheme 2 is rate- full-length mRNA would preclude steady-state or stopped-flow meas- 2 2 limiting for the reverse direction (dissociation of the urements based on fluorescence quenching because the optical density eIF4Em GpppG complex) in the two-step model. This provides a would be too high at the concentrations required for experiments means to distinguish experimentally between the one- and two-step such as shown in Fig. 2. We therefore chose to study a capped models. The dissociation kinetics of the eIF4Em GpppG complex oligoribonucleotide. can be followed by diluting the complex with buffer in the stopped- As noted under “Experimental Procedures,” use of conventional cap flow cell. A first-order rate constant is obtained from the increase in analogs in the in vitro synthesis of capped RNAs results in approxi- fluorescence upon dissociation of the complex. For the one-step mately half of the caps being incorporated in the incorrect orientation model with pre-equilibrium between monomers and dimers (63). Instead, we synthesized a capped 12-mer oligoribonucleotide con- 7,3-O (Scheme 4), k determined in the forward reaction is expected to be off taining the anti-reverse cap analog (ARCA) m GpppG (63) to equal to k determined in the reverse reaction. For the two-step off ensure that all caps were in the correct orientation. Previous equilib- 7,3-O 7 model (Scheme 2), the apparent k determined in the reverse reac- off rium measurements have shown that m GpppG and m GpppG do tion is the limiting k and is expected to be considerably smaller not differ significantly in their affinity for eIF4E (69). However, we than the apparent k determined in the forward reaction. off needed to determine whether the kinetic parameters were also the same 7,3-O 7 We therefore measured the rate of dissociation of the pre-formed for m GpppG and m GpppG. The results indicated that the kinetic eIF4Em GpppG complex by diluting the complex 2-fold with buffer. constants were statistically the same for binding of the conventional cap and the ARCA to eIF4E. For instance, at 100 mM KCl, k was 184 on TABLE 2 6 1 1 7 6 1 1 10  10 M s for m GpppG (Table 1) and 205  24  10 M s Effect of eIF4E concentration on the amplitudes of the first and the for the ARCA (Table 3). Similarly, k was statistically the same for off second phases for reaction of eIF4E with m GpppG at limiting 7 1 1 m GpppG (83 23 s ) and the ARCA (76 27 s ). This held true for m GpppG concentrations eIF4E(P) as well (Tables 1 and 3). Data are derived from experiments similar to Fig. 5D fit with Equation 2. Stopped-flow traces for the binding of the capped oligoribonucle- eIF4E concentration F F 1 2 otide to eIF4E and eIF4E(P) at 350 mM KCl are shown in Fig. 6. Reactions M % followed single-exponential kinetics, because application of a double- 1.0 69 631  30 1.5 61 839  5 exponential function did not improve the fit (data not shown). Plots of 2.0 52 748  7 k versus ligand concentration were linear (Fig. 7), indicating that the 3.0 30 770  2 obs binding of capped oligonucleotides also follows second-order kinetics in As noted under “Experimental Procedures,” the amplitudes of fluorescence changes F were measured in V. However, different stopped-flow instrument agreement with a one-step binding mechanism. Values for k and k on off settings were used for different eIF4E concentrations, so that a direct comparison were derived from the slope and y intercept, respectively, and are given of absolute values of F is not meaningful. Instead, the relative amplitudes at each eIF4E concentration are presented. in Table 3. TABLE 3 Kinetic constant for association and dissociation of eIF4E and eIF4E(P) with an ARCA and an ARCA-capped oligoribonucleotide Data are obtained from experiments similar to Fig. 7, with application of Equation 3. KCl concentration Protein Ligand 100 mM 350 mM 6 6 k  10 k K k  10 k K on off d on off d 1 1 1 1 1 1 M s s MM s s M eIF4E ARCA 205  24 76  27 0.37  0.14 41  4 107  11 2.61  0.37 eIF4E Capped oligo ND ND ND 101 411  10 0.11  0.10 eIF4E(P) ARCA 84 774  20 0.88  0.25 32  3 106  13 3.31  0.51 eIF4E(P) Capped oligo 394  54 103  50 0.26  0.13 37 318  6 0.49  0.17 ND, not determined because the rate was too fast to be measured. 14934 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E 1 1 M s . An even greater difference, 9.7-fold, was observed for k . The off combination of a faster association rate and a slower dissociation rate resulted in a K that was 24-fold lower for capped oligoribonucleotide compared with cap analog. Phosphorylation of eIF4E decreased k for the capped oligoribo- on nucleotide sufficiently that it could be measured at both 100 and 350 mM KCl (Table 3). The capped oligoribonucleotide associated with eIF4E(P) 4.7-fold faster than the cap analog at 100 mM KCl (k  394 54 10 on 6 1 1 versus 84  7  10 M s ). At 350 mM KCl, the effect of the oligori- bonucleotide chain was eliminated, because k values were statistically on the same for capped oligoribonucleotide and cap analog. Because this was not the case for unphosphorylated eIF4E, it appears that Ser-209 phosphorylation negates the increase in association rate contributed by the oligoribonucleotide chain, perhaps because of charge repulsion. No statistical difference in k could be detected between capped oligoribo- off nucleotide and cap analog at 100 mM KCl, but at 350 mM, the capped oligoribonucleotide dissociated 5.9-fold more slowly than the cap ana- log. For both eIF4E and eIF4E(P) at 350 mM KCl, the decrease in k is off the major determinant for the higher affinity of capped oligoribonucle- otide compared with cap analog. DISCUSSION FIGURE 6. Kinetics of capped oligoribonucleotide binding to eIF4E and eIF4E(P). Protein synthesis is a dynamic process in which cap binding and Reactions were conducted as described in the legend to Fig. 2 except with a 12-mer 7,3-O release presumably occur at least once for each round of initiation. We oligoribonucleotide capped with the anti-reverse cap analog (ARCA) m GpppG. A, reaction of eIF4E with the indicated concentrations of capped oligoribonucleotide at therefore sought to compare kinetic rather than equilibrium parameters 350 mM KCl. Traces follow Equation 1 (solid line) with k being 109, 211, 351, and 444 s obs for cap interaction with eIF4E versus eIF4E(P). Because the binding and for 1, 2, 3, and 4 M oligoribonucleotide, respectively. B, same as A except with eIF4E(P). 1 7 k was 55, 78, 119, and 205 s for 1, 2, 3, and 5 M m GpppG, respectively. dissociation reactions are fast, an appropriate method is to measure obs pre-steady-state kinetics in rapid mixing experiments. Whereas eIF4E and eIF4E(P) have been compared by equilibrium techniques, and eIF4E interactions with cap analogs have been studied by stopped-flow tech- niques, eIF4E and eIF4E(P) have not previously been compared by stopped-flow techniques. Furthermore, the binding of eIF4E or eIF4E(P) to capped oligoribonucleotides, which are structurally more similar to mRNA than cap analogs, has not previously been studied by stopped-flow techniques. Magnitude of Kinetic Constants—Our stopped-flow data for cap ana- logs are in reasonable agreement with those of Blachut-Okrasinska et al. 1 7 (35), who found k  24–35 s for the interaction between m GpppG off and mouse eIF4E-(28–217) in stopped-flow experiments conducted from 50 to 350 mM KCl, compared with our values of 72–85 s (Table 1). Our 2-fold higher k values may be because of human versus mouse eIF4E off (although they are 99% identical), the presence of the N-terminal 27 amino acid residues in our eIF4E preparations, or the fact that the nat- ural Thr-205 is replaced with Cys-205 in eIF4E-(28–217) phosphoryl- ated by intein technology. Dlugosz et al. (36) did not study m GpppG, 7 7 but they did obtain k values for m GTP and m GDP by stopped-flow off measurements of mouse eIF4E-(28–217) that were in the range of 40–75 s . Similarly, k for vaccinia virus VP39 was reported to be off 41–60 s (61). Our k values for eIF4E and eIF4E(P) interaction with a capped oli- off goribonucleotide are at variance with the findings of Scheper et al. (37), 7,3-O FIGURE 7. Dependence of k for binding of eIF4E to either the ARCA m Gp- obs 2 despite the fact that human eIF4E was used in both cases and the ppG or to an ARCA-capped 12-mer oligoribonucleotide. Data are derived from exper- iments similar to those shown in Fig. 6. Experiments were conducted at either 350 mM KCl method of phosphorylation was the same. These authors determined (panel A) or 100 mM KCl (panel B). Data are fit with Equation 3. Numerical values for k , on k for a capped 18-mer oligoribonucleotide at 100 mM KCl to be 0.06 off k , and K are given in Table 3. off d 1 1 s for mock-phosphorylated eIF4E, 0.078 s for untreated eIF4E, and 0.58 s for eIF4E(P), whereas our k value for eIF4E(P) under similar For unphosphorylated eIF4E, the binding of capped oligoribonucle- off otide was too fast to be measured by the stopped-flow technique at 100 conditions was 103 s , or 186-fold greater (Table 3). This discrepancy mM KCl. We therefore estimate that it is close to the diffusion limit (1 may be because of the inability of SPR to measure kinetic parameters for 9 1 1 10 M s ). At 350 mM KCl, k was 2.5-fold greater for capped oli- reactions as fast as the binding of eIF4E to cap analogs or capped oli- on 6 6 goribonucleotide than cap analog, 101  4  10 versus 41  4  10 goribonucleotides. The upper limits for k and k measured by SPR are on off MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14935 Kinetic Analysis of Cap Binding to eIF4E 7 1 1 1 10 M s and 0.1 s , respectively (78), which are several orders of not observe biphasic kinetics until eIF4E concentrations exceeded 0.5 magnitude lower than with stopped-flow techniques. This conclusion is M (Table 2). Third, the value of k determined from the reverse reac- off further supported by the low analyte response observed by Scheper et al. tion is the same as k determined from the forward reaction. off (37). From the 1:1 stoichiometry of the binding reaction, the molecular Effect of eIF4E Phosphorylation on Cap Binding—Two similar models masses of a capped 18-mer oligoribonucleotide and eIF4E, and the have been proposed for the role of eIF4E phosphorylation in translation ligand response (79), one can calculate a theoretical analyte response of initiation. Marcotrigiano et al. (11) suggested a “clamping” mechanism 586 RU. The analyte responses reported by Scheper et al. (37) do not in which the generation of a salt bridge of Ser-209 with Lys-159 acts as a exceed 200 RU for eIF4E or 50 RU for eIF4E(P). One of the causes for a clamp that stabilizes the cap in the binding slot. Tomoo et al. (13) used low maximum analyte response is when reaction rates exceed the range molecular dynamics simulations to suggest that a hydrogen-bonded measurable by the SPR technique (78, 80) cluster of water molecules or polar amino acid residues could form Mechanism of eIF4E Association with Cap Analogs as Inferred from around the dianionic form of Ser(P)-209. This could potentially block Kinetics Measurements—Cap binding to eIF4E has been studied exten- the release of the cap from the binding slot. In both cases, stabilization is sively by both equilibrium and kinetic methods (see Introduction). Most envisioned to facilitate the assembly of initiation complexes loaded with of these studies have utilized a decrease in intrinsic Trp fluorescence mRNA. The principal prediction from both hypotheses is a significant upon cap binding, although some have utilized SPR (37), isothermal decrease in k upon phosphorylation of eIF4E because of closure of the off titration calorimetry (87), or NMR (12). Results from some of these cap-binding pocket. Also, in such a mechanism, k is expected to be on studies have led the authors to propose a two-step binding mechanism significantly impaired if mRNA were to interact with eIF4E in which a (26, 34, 35). The first step is envisioned as being the ligand entering the salt bridge had already been formed between Ser-209 and Lys-159. A cap-binding slot and anchoring, via the triphosphate moiety, to basic different type of model was proposed by Scheper et al. (37) on the basis amino acid residues. The second step is a change within the m G-bind- of their results from SPR experiments showing that phosphorylation of ing slot that leads to a further fluorescence decrease. Molecular dynam- eIF4E accelerates the rate of cap dissociation. They speculated that ics simulations have also supported a model in which there is a confor- phosphorylation facilitates the release of eIF4E and other initiation fac- mational change in eIF4E upon cap binding (13) (although this does not tors from the 5-end of the mRNA. per se constitute proof for two kinetically distinct steps). By contrast, a Our determination of k and k for eIF4E and eIF4E(P) allows us to on off one-step binding model has been proposed for another cap-binding test these predictions directly. We did not find any significant difference protein, the vaccinia virus VP39 (61). Cap binding by VP39 involves a 7 in k values between eIF4E and eIF4E(P) at any salt concentrations off cation- sandwich of m G between two aromatic amino acid residues, investigated, either for cap analogs or capped oligoribonucleotide similar to the cap-eIF4E interaction. However, in this case, phosphate (Tables 1 and 3). By contrast, Scheper et al. (37) found a 10-fold groups do not contribute to the binding process. increase in k for eIF4E(P) interaction with capped oligoribonucleotide Our data also show that association between eIF4E and cap analogs as off compared with eIF4E. Furthermore, k from that study (calculated well as oligoribonucleotides behave kinetically as a simple one-step on from K and k ) is slightly higher for eIF4E(P) binding to capped oli- process at all salt concentrations investigated and regardless of whether d off goribonucleotide compared with eIF4E, although we find k to be 2–3- eIF4E is phosphorylated. The discrepancy between our results and those on fold lower for eIF4E(P) compared with eIF4E, as measured with both cap of others (26, 34, 35) could be explained by the existence of a rate- limiting, parallel reaction resulting from dissociation of pre-formed analogs (Tables 1 and 3) and the capped oligoribonucleotide (at 350 mM KCl) (Table 3). Thus, our results do not support any of the previously eIF4E dimers or oligomers. We postulate an equilibrium between the reactive monomer and unreactive dimers or higher order oligomers proposed hypotheses. Instead, we propose that phosphorylation of Ser- (Scheme 3). The fast phase of fluorescence change is because of 209, which is located at the entrance to the cap-binding slot (Fig. 8A), m GpppG binding to monomeric eIF4E according to Scheme 1. The diminishes the rate of association by charge repulsion but has no effect slow phase is a result of the rate-limiting dissociation of unreactive on the rate of dissociation. This model is consistent with the observation oligomers to yield reactive monomers, which in turn react with the cap that the effect of eIF4E phosphorylation is progressively eliminated as analog according to Scheme 4. However, the cap-binding reaction per se the KCl concentration is increased (Refs. 32 and 35; Fig. 4A); at high salt is still a one-step process. concentrations, the charge on Ser(P)-209 is shielded and its inhibitory The evidence supporting this alternative mechanism is as follows. effect on the association rate is masked (Fig. 4A). First, under pseudo first-order reaction conditions (with [eIF4E] limit- Reaction of eIF4E and eIF4E(P) with Capped Oligoribonucleotides: an ing and0.5 M), the experimental data were fit by a single-exponential Additional Binding Site for mRNA?—We found that a capped 12-mer function and dependence of k on m GpppG concentrations was lin- obs oligoribonucleotide interacts with both eIF4E and eIF4E(P) in a very fast ear. Second, the slower second phase was observed only at elevated reaction. The addition of an oligoribonucleotide chain to the cap struc- eIF4E concentrations, which is consistent with concentration-depen- ture does not change the kinetic mechanism of binding. It is still a dant protein self-association. Whereas most of our experiments were one-step reaction, but both k and k are different for binding of the on off conducted at eIF4E concentrations of 0.1–0.2 M, those of Sha et al. capped oligoribonucleotide compared with the cap analog. For eIF4E(P) (34) were conducted at 0.5 M and those of Blachut-Okrasinska et al. at 100 mM KCl, k is 5-fold greater for capped oligoribonucleotide on (35) were conducted at 0.1–4.1 M. Examination of the latter data (35) than cap analog (Table 3). The reaction for unphosphorylated eIF4E is reveals that the double-exponential function improves the fit only when too fast to be measured by stopped-flow kinetics at 100 mM KCl, but at the eIF4E concentration is increased (Tables A1, A2, and A3). We did 350 mM, the reaction is slowed enough to reveal that k is 2.5-fold on greater for capped oligoribonucleotide than cap analog. Interestingly, The analyte response can be calculated from the expression (80), S  (AR/LR)  (LM/ the combination of two inhibitory effects on k , introduction of a phos- on AM), where S is the stoichiometry, AR is the analyte response, LR is the ligand response, LM is the ligand molecular mass, and AM is the analyte molecular mass. LR phate group on Ser-209, and shielding of positively charged amino acid is not given in Ref. 37 but is given in a publication cited in that work (85) as 150 RU for residues by high ionic strength, causes k to be the same for capped on SPR of eIF4E on immobilized capped oligoribonucleotides. LM and AM are 6,400 and 25,000 Da, respectively, and S is 1.0. AR is therefore calculated to be 586 RU. oligoribonucleotide and cap analog. 14936 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E is k , which decreases 2–3-fold. Is such a difference biologically signif- on icant? Even though 48 S complex formation is thought to be rate-limit- ing for initiation of protein synthesis (81, 82), the association of mRNA with eIF4E and eIF4E(P) is so fast (as inferred from measurements with a capped oligoribonucleotide) that it may not be rate-limiting for 48 S initiation complex formation, especially because k is near the diffusion on limit for both eIF4E and eIF4E(P). Perhaps some other process in 48 S complex formation is rate-limiting. Arguing against this suggestion is the fact that the binding affinity of various cap analogs to eIF4E is pos- itively correlated with the translational efficiency of mRNA capped with those analogs, as measured both in vitro (63, 69, 83) and in vivo (84). This correlation would not hold if cap-eIF4E interactions did not make a significant contribution to the rate of initiation. Given that eIF4E is involved in recruitment of the unwinding machinery (1), it is possible that the effect of eIF4E phosphorylation on the rate of initiation is man- ifested in subsequent steps rather than cap binding per se. Acknowledgments—We are grateful to Christopher Proud for plasmids pEBG- Mnk2 and pEBG-Mnk1T2A2, Ewa Grudzien for advice on capped oligoribo- nucleotide synthesis, Srikripa Chandrasekaran for preparation of T7 polym- erase, and Stephan Witt for the use of the stopped-flow instrument. REFERENCES 1. Kapp, L. D., and Lorsch, J. R. (2004) Annu. Rev. Biochem. 73, 657–704 2. Rhoads, R. E., Joshi-Barve, S., and Rinker-Schaeffer, C. (1993) Prog. Nucleic Acids Res. Mol. Biol. 46, 183–219 3. Jankowska-Anyszka, M., Lamphear, B. J., Aamodt, E. J., Harrington, T., Darzyn- kiewicz, E., Stolarski, R., and Rhoads, R. E. (1998) J. Biol. Chem. 273, 10538–10542 4. Dyer, J. R., Pepio, A. M., Yanow, S. K., and Sossin, W. S. (1998) J. Biol. Chem. 273, 7 29469–29474 FIGURE 8. Three-dimensional models of human eIF4E in complex with m GpppA. 5. Rodriguez, C., Freire, M., Camilleri, C., and Robaglia, C. (1998) Plant J. 13, 465–473 Data are taken from Ref. 86. A, ribbon drawing showing phosphorylation at Ser-209. B, 6. Robalino, J., Joshi, B., Fahrenkrug, S. C., and Jagus, R. (2004) J. Biol. Chem. 279, space-filling model showing surface charge distribution. The color code for amino acid residues are: basic, green; acidic, magenta; hydrophobic, blue; and uncharged polar, yel- 10532–10541 low. The images were created using the Brookhaven Protein Data Bank data file 1WKW 7. Joshi, B., Cameron, A., and Jagus, R. (2004) Eur. J. Biochem. 271, 2189–2203 and RasMol. 8. Gorlich, D., and Mattaj, I. W. (1996) Science 271, 1513–1518 9. Strudwick, S., and Borden, K. L. (2002) Differentiation 70, 10–22 At high salt concentrations, a difference in k between capped oli- off 10. Richter, J., and Sonenberg, N. (2005) Nature 433, 477–480 goribonucleotide and cap analog also became apparent. For eIF4E, it 11. Marcotrigiano, J., Gingras, A.-C., Sonenberg, N., and Burley, S. K. (1997) Cell 89, was 10-fold lower for capped oligoribonucleotide, and for eIF4E(P), it 951–961 12. Matsuo, H., Li, H., McGuire, A. M., Fletcher, C. M., Gingras, A.-C., Sonenberg, N., was7-fold lower. Thus, both k and k contribute to a decrease in K on off d and Wagner, G. (1997) Nat. Struct. Biol. 4, 717–724 when comparing capped oligoribonucleotide to the cap analog: 24- 13. Tomoo, K., Shen, X., Okabe, K., Nozoe, Y., Fukuhara, S., Morino, S., Sasaki, M., fold for eIF4E and7-fold for eIF4E(P) at 350 mM KCl. These data are in Taniguchi, T., Miyagawa, H., Kitamura, K., Miura, K., and Ishida, T. (2003) J. Mol. Biol. agreement with previous equilibrium experiments where the binding of 328, 365–383 14. Rychlik, W., Domier, L. L., Gardner, P. R., Hellmann, G. M., and Rhoads, R. E. (1987) natural rabbit eIF4E to globin mRNA was found to be 5.3–5.5-fold 7 Proc. Natl. Acad. Sci. U. S. A. 84, 945–949 stronger than its binding to m GpppG (18, 22). Binding to capped oli- 15. McCubbin, W. D., Edery, I., Altmann, M., Sonenberg, N., and Kay, C. M. (1988) J. Biol. goribonucleotides was also stronger than to m GpppG but varied with Chem. 263, 17663–17671 the degree of oligoribonucleotide secondary structure (22). 16. Carberry, S. E., Rhoads, R. E., and Goss, D. J. (1989) Biochemistry 28, 8078–8083 17. Carberry, S. E., Darzynkiewicz, E., Stepinski, J., Tahara, S. M., Rhoads, R. E., and Goss, The pronounced increase in affinity for capped oligoribonucleotide D. J. (1990) Biochemistry 29, 3337–3341 suggests that some additional stabilizing interactions may be formed 18. Goss, D. J., Carberry, S. E., Dever, T. E., Merrick, W. C., and Rhoads, R. E. (1990) that slow dissociation of the larger ligand. TheG caused by addition Biochemistry 29, 5008–5012 of the oligoribonucleotide chain to the cap structure is 1.9 kcal/mol 19. Carberry, S. E., Darzynkiewicz, E., and Goss, D. J. (1991) Biochemistry 30, 1624–1627 20. Carberry, S. E., and Goss, D. J. (1991) Biochemistry 30, 4542–4545 for eIF4E and 1 kcal/mol for eIF4E(P). This is sufficient for the forma- 21. Carberry, S. E., and Goss, D. J. (1991) Biochemistry 30, 6977–6982 tion of an additional salt bridge or hydrogen bond between the body of 22. Carberry, S. E., Friedland, D. E., Rhoads, R. E., and Goss, D. J. (1992) Biochemistry 31, the mRNA and amino acid residues near the entrance of the cap-bind- 1427–1432 ing slot. There are several basic amino acid residues located on the 23. Minich, W. B., Balasta, M. L., Goss, D. J., and Rhoads, R. E. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 7668–7672 surface of eIF4E near the cap-binding slot (Arg-112, Arg-157, Lys-159, 24. Wang, Y. H., Sha, M., Ren, W. Y., van Heerden, A., Browning, K. S., and Goss, D. J. and Lys-162) (Fig. 8B). Additional basic residues are found on the lateral (1996) Biochim. Biophys. Acta 1297, 207–213 surface of a groove where the polynucleotide chain of mRNA conceiv- 25. Ren, J., and Goss, D. (1996) Nucleic Acids Res. 24, 3629–3634 ably could be bound (Lys-206, Lys-212, Lys-192, Lys-119, and Lys-52). It 26. Niedzwiecka, A., Marcotrigiano, J., Stepinski, J., Jankowska-Anyszka, M., Wyslouch- Cieszynska, A., Dadlez, M., Gingras, A.-C., Mak, P., Darzynkiewicz, E., Sonenberg, N., is possible that additional interactions occur between the polynucle- Burley, S. K., and Stolarski, R. (2002) J. Mol. Biol. 319, 615–635 otide chain and one or more of these basic side chains. 27. Wieczorek, Z., Zdanowski, K., Chlebicka, L., Stepinski, J., Jankowska, M., Kierdaszuk, Role of eIF4E Phosphorylation in Protein Synthesis—Our results indi- B., Temeriusz, A., Darzynkiewicz, E., and Stolarski, R. (1997) Biochim. Biophys. Acta cate that the only kinetic parameter affected by eIF4E phosphorylation 1354, 145–152 MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14937 Kinetic Analysis of Cap Binding to eIF4E 28. Wieczorek, Z., Darzynkiewicz, E., and Lonnberg, H. (1998) J. Photochem. Photobiol. B 56. Lachance, P. E. D., Miron, M., Raught, B., Sonenberg, N., and Lasko, P. (2002) Mol. Biol. 43, 158–163 Cell. Biol. 22, 1656–1663 29. Wieczorek, Z., Niedzwiecka-Kornas, A., Chlebicka, L., Jankowska, M., Kiraga, K., 57. Arquier, N., Bourouis, M., Colombani, J., and Leopold, P. (2005) Curr. Biol. 15, 19–23 Stepinski, J., Dadlez, M., Drabent, R., Darzynkiewicz, E., and Stolarski, R. (1999) Z. 58. Reiling, J., Doepfner, K., Hafen, E., and Stocker, H. (2005) Curr. Biol. 15, 24–30 Naturforsch. 54, 278–284 59. Ueda, T., Watanabe-Fukunaga, R., Fukuyama, H., Nagata, S., and Fukunaga, R. (2004) 30. Zuberek, J., Jemielity, J., Stepinski, J., Lewdorowicz, M., Niedzwiecka, A., Haber, D., Mol. Cell. Biol. 24, 6539–6549 Stolarski, R., Rhoads, R., and Darzynkiewicz, E. (2003) Nucleosides Nucleotides Nucl. 60. Goss, D. J., Carberry, S. E., Dever, T. E., Merrick, W. C., and Rhoads, R. E. (1990) Acids 22, 1703–1706 Biochim. Biophys. Acta 1050, 163–166 31. Zuberek, J., Wyslouch-Cieszynska, A., Niedzwiecka, A., Dadlez, M., Stepinski, J., 61. Hu, G., Tsai, A.-L., and Quiocho, F. A. (2003) J. Biol. Chem. 278, 51515–51520 Augustyniak, W., Gingras, A.-C., Zhang, Z., Burley, S. K., Sonenberg, N., Stolarski, R., 62. Fechter, P., and Brownlee, G. G. (2005) J. Gen. Virol. 86, 1239–1249 and Darzynkiewicz, E. (2003) RNA (N.Y.) 9, 52–61 63. Stepinski, J., Waddell, C., Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (2001) 32. Zuberek, J., Jemielity, J., Jablonowska, A., Stepinski, J., Dadlez, M., Stolarski, R., and RNA (N.Y.) 7, 1486–1495 Darzynkiewicz, E. (2004) Biochemistry 43, 5370–5379 64. Stern, B. D., Wilson, M., and Jagus, R. (1993) Protein Expression Purif. 4, 320–327 33. Khan, M., and Goss, D. (2004) Biochemistry 43, 9092–9097 65. Webb, N. R., Chari, R. V. J., DePillis, G., Kozarich, J. W., and Rhoads, R. E. (1984) 34. Sha, M., Wang, Y. H., Xiang, T., van Heerden, A., Browning, K. S., and Goss, D. J. Biochemistry 23, 177–181 (1995) J. Biol. Chem. 270, 29904–29909 66. von der Haar, T., Ball, P. D., and McCarthy, E. G. (2000) J. Biol. Chem. 275, 35. Blachut-Okrasinska, E., Bojarska, E., Niedzwiecka, A., Chlebicka, L., Darzynkiewicz, 30551–30555 E., Stolarski, R., Stepinski, J., and Antosiewicz, J. M. (2000) Eur. Biophys. J. 29, 67. Zawadzki, V., and Gross, H. (1991) Nucleic Acids Res. 19, 1948 487–498 68. Milligan, J., Groebe, D., Witherell, G., and Uhlenbeck, O. (1987) Nucleic Acids Res. 15, 36. Dlugosz, M., Blachut-Okrasinska, E., Bojarska, E., Darzynkiewicz, E., and 8783–8798 Antosiewicz, J. (2003) Eur. Biophys. J. 31, 608–616 69. Jemielity, J., Fowler, T., Zuberek, J., Stepinski, J., Lewdorowicz, M., Niedzwiecka, A., 37. Scheper, G. C., van Kollenburg, B., Hu, J., Luo, I., Goss, D. J., and Proud, C. G. (2002) Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (2003) RNA (N.Y.) 9, 1108–1122 J. Biol. Chem. 277, 3303–3309 70. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1989) Molecular Cloning: A Laboratory 38. Joshi, B., Cai, A.-L., Keiper, B. D., Minich, W. B., Mendez, R., Beach, C. M., Stepinski, Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY J., Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (1995) J. Biol. Chem. 270, 71. Bevington, P. R. (1969) Data Reduction and Error Analysis for the Physical Sciences, 14597–14603 McGraw-Hill, New York 39. Flynn, A., and Proud, C. G. (1995) J. Biol. Chem. 270, 21684–21688 72. Rychlik, W., Gardner, P. R., Vanaman, T. C., and Rhoads, R. E. (1986) J. Biol. Chem. 40. Fukunaga, R., and Hunter, T. (1997) EMBO J. 16, 1921–1933 261, 71–75 41. Waskiewicz, A. J., Flynn, A., Proud, C. G., and Cooper, J. A. (1997) EMBO J. 16, 73. Hiremath, L., Hiremath, S., Rychlik, W., Joshi, S., Domier, L., and Rhoads, R. (1989) 1909–1920 J. Biol. Chem. 264, 1132–1138 42. Scheper, G. C., Morrice, N. A., Kleijn, M., and Proud, C. G. (2001) Mol. Cell. Biol. 21, 74. Bradshaw, R., Brickey, W., and Walker, K. (1998) Trends Biochem. Sci. 23, 263–267 743–754 75. Righetti, P. G. (1983) Isoelectric Focusing: Theory, Methodology and Applications, 43. Kaspar, R. L., Rychlik, W., White, M. W., Rhoads, R. E., and Morris, D. R. (1990) J. Biol. Elsevier Science, Amsterdam Chem. 265, 3619–3622 76. Fersht, A. (1999) Structure and Mechanism in Protein Science, W. H. Freeman and 44. Rychlik, W., Rush, J. S., Rhoads, R. E., and Waechter, C. J. (1990) J. Biol. Chem. 265, Company, New York 19467–19471 77. Waskiewicz, A. J., Johnson, J. C., Penn, B., Mahalingam, M., Kimball, S. R., and 45. Manzella, J. M., Rychlik, W., Rhoads, R. E., Hershey, J. W. B., and Blackshear, P. J. Cooper, J. A. (1999) Mol. Cell. Biol. 19, 1871–1880 (1991) J. Biol. Chem. 266, 2383–2389 78. Myszka, D. (1997) Curr. Opin. Biotechnol. 8, 50–57 46. Worch, J., Tickenbrock, L., Schwable, J., Steffen, B., Cauvet, T., Mlody, B., Buerger, H., 79. Ptushkina, M., von der Haar, T., Karimm, M. M., Hughers, J. M., and McCarthy, Koeffler, H., Berdel, W., Serve, H., and Muller-Tidow, C. (2004) Oncogene 23, J. E. G. (1999) EMBO J. 18, 4068–4075 9162–9172 80. Pharmacia Biosensor AB (1994) BIApplications Handbook, Pharmacia Biosensor AB, 47. Walsh, D., and Mohr, I. (2004) Genes Dev. 18, 660–672 Uppsala, Sweden 48. Pyronnet, S., Imataka, H., Gingras, A.-C., Fukunaga, R., Hunter, T., and Sonenberg, N. 81. Darnbrough, C., Legon, S., Hunt, T., and Jackson, R. J. (1973) J. Mol. Biol. 76, 379–403 (1999) EMBO J. 18, 270–279 82. Safer, B., Kemper, W., and Jagus, R. (1978) J. Biol. Chem. 253, 3384–3386 49. Cuesta, R., Xi, Q., and Schneider, R. J. (2000) EMBO J. 19, 3465–3474 83. Grudzien, E., Stepinski, J., Jankowska-Anyszka, M., Stolarski, R., Darzynkiewicz, E., 50. Orton, K. C., Ling, J., Waskiewicz, A. J., Cooper, J. A., Merrick, W. C., Korneeva, N. L., and Rhoads, R. E. (2004) RNA (N.Y.) 10, 1479–1487 Rhoads, R. E., Sonenberg, N., and Traugh, J. A. (2004) J. Biol. Chem. 279, 84. Grudzien, E., Kalek, M., Jemielity, J., Darzynkiewicz, E., and Rhoads, R. E. (2006) 38649–38657 J. Biol. Chem. 281, 1857–1867 51. Knauf, U., Tschopp, C., and Gram, H. (2001) Mol. Cell. Biol. 21, 5500–5511 85. Ptushkina, M., von der Haal, T., Vasilescu, S., Frank, R., Birkenhager, R., and 52. McKendrick, L., Morley, S. J., Pain, V. M., Jagus, R., and Joshi, B. (2001) Eur. J. Bio- McCarthy, J. E. G. (1998) EMBO J. 17, 4798–4808 chem. 268, 5375–5385 86. Tomoo, K., Matsushita, Y., Fujisaki, H., Abiko, F., Shen, X., Taniguchi, T., Miyagawa, 53. Morley, S. J., and Naegele, S. (2002) J. Biol. Chem. 277, 32855–32859 H., Kitamura, K., Miura, K., and Ishida, T. (2005) Biochim. Biophys. Acta 1753, 54. Naegele, S., and Morley, S. J. (2004) J. Biol. Chem. 279, 46023–46034 191–208 55. Hong-Brown, L. Q., Brown, C. R., and Lang, C. H. (2004) Am. J. Physiol. 287, 87. Niedzwiecka, A., Stepinski, J., Darzynkiewicz, E., Sonenberg, N., and Stolarski, R. C1482–C1492 (2002) Biochemistry 41, 12140–12148 14938 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry American Society for Biochemistry and Molecular Biology

Stopped-flow Kinetic Analysis of eIF4E and Phosphorylated eIF4E Binding to Cap Analogs and Capped Oligoribonucleotides: EVIDENCE FOR A ONE-STEP BINDING MECHANISM *

Download PDF
12 pages

Loading next page...
 
/lp/american-society-for-biochemistry-and-molecular-biology/stopped-flow-kinetic-analysis-of-eif4e-and-phosphorylated-eif4e-e9jNdSjJX6

References (85)

Publisher
American Society for Biochemistry and Molecular Biology
Copyright
Copyright © 2006 Elsevier Inc.
ISSN
0021-9258
eISSN
1083-351X
DOI
10.1074/jbc.m601653200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 21, pp. 14927–14938, May 26, 2006 © 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. Stopped-flow Kinetic Analysis of eIF4E and Phosphorylated eIF4E Binding to Cap Analogs and Capped Oligoribonucleotides EVIDENCE FOR A ONE-STEP BINDING MECHANISM Received for publication, February 22, 2006 Published, JBC Papers in Press, March 15, 2006, DOI 10.1074/jbc.M601653200 ‡ § ‡1 Sergey V. Slepenkov , Edward Darzynkiewicz , and Robert E. Rhoads From the Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center, Shreveport, Louisiana 71130-3932 and the Department of Biophysics, Warsaw University, Warsaw 02-089, Poland Recruitment of eukaryotic mRNA to the 48 S initiation complex to eIF4G, a protein that also interacts with the RNA helicase eIF4A to is rate-limiting for protein synthesis under normal conditions. promote unwinding of mRNA secondary structure, with the multisub- Binding of the 5-terminal cap structure of mRNA to eIF4E is a unit factor eIF3 to recruit the 43 S initiation complex, and with the critical event during this process. Mammalian eIF4E is phosphoryl- cytoplasmic poly(A)-binding protein to enhance initiation of poly(A)- ated at Ser-209 by Mnk1 and Mnk2 kinases. We investigated the containing mRNAs. Initiation codon recognition is followed by dissoci- interaction of both eIF4E and phosphorylated eIF4E (eIF4E(P)) with ation of eIFs and joining of the 60 S ribosomal subunit to form the cap analogs and capped oligoribonucleotides by stopped-flow elongation-competent 80 S initiation complex. kinetics. For m GpppG, the rate constant of association, k , was eIF4E has been extensively investigated in organisms that range from on dependent on ionic strength, decreasing progressively up to 350 mM yeast to mammals (2–7). Besides translation, eIF4E also functions in nucle- KCl, but the rate constant of dissociation, k , was independent of ocytoplasmic transport of mRNA, sequestration of mRNA in a nontrans- off ionic strength. Phosphorylation of eIF4E decreased k by 2.1–2.3- latable state, and stabilization of mRNA against decay in the cytosol (8–10). on fold at 50–100 mM KCl but had progressively less effect at higher The three-dimensional structures of human, mouse, and Saccharomyces ionic strengths, being negligible at 350 mM. Contrary to published cerevisiae eIF4E have been solved (11–13). The complex of full-length evidence, eIF4E phosphorylation had no effect on k . Several human eIF4E with m GpppA is bell-shaped, with the cap analog situated in off observations supported a simple one-step binding mechanism, in a deep slot in the concave surface of the protein. The cap-binding pocket contrast to published reports of a two-step mechanism. The kinetic consists of separate recognition components for the m G base, the triphos- function that best fit the data changed from single- to double-expo- phate moiety, and the second nucleoside residue. The alkylated base is nential as the eIF4E concentration was increased. However, meas- stacked between Trp-56 and Trp-102. (Amino acid numbers refer to the uring k for dissociation of a pre-formed eIF4Em GpppG complex human sequence (14).) Glu-103 and Trp-102 form H-bonds with the off suggested that the double-exponential kinetics were caused by dis- N-1, N-2, and O-6 protons of m G. Trp-56 also interacts directly with sociation of eIF4E dimers, not a two-step mechanism. Addition of a the ribose moiety, and Arg-157 and Lys-162 interact directly with the - 12-nucleotide chain to the cap structure increased affinity at high and -phosphate oxygen atoms. The second nucleoside moiety of ionic strength for both eIF4E (24-fold) and eIF4E(P) (7-fold), pri- m GpppA is fixed by the flexible C-terminal loop. To date, there have marily due to a decrease in k . This suggests that additional stabi- been no structures reported for eIF4E in complex with capped mRNA off lizing interactions between capped oligoribonucleotides and eIF4E, or even short oligoribonucleotides. which do not occur with cap analogs alone, act to slow dissociation. Cap analogs bind to eIF4E in a tight complex, a step that has been studied primarily by equilibrium techniques (15–33). Intrinsic Trp fluorescence quenching of N-terminal truncated mouse eIF4E (residues 28–217) by The efficiency of mRNA translational initiation is strongly enhanced titration with cap analogs indicates that the free energy of m G stacking and hydrogen bonding is separate from the free energy of triphosphate chain by the 5-terminal cap, m GpppN (1). The cap specifically binds to eIF4E, which may be the first canonical initiation factors to interact interactions (26). Electrostatically steered eIF4E-cap analog association is with mRNA during its recruitment to the ribosome. eIF4E in turn binds accompanied by hydration of the complex and a shift in ionic equilibria. The kinetics of cap analog binding to wheat eIF(iso)4F (34) and mouse eIF4E-(28–217) (35, 36) have also been studied with rapid mixing tech- * This work was supported by National Institutes of Health Grants 2R01GM20818 (to niques. The authors of these studies have interpreted their data as indicat- R. E. R.) and 1R03TW006446 (to R. E. R. and E. D.) and Howard Hughes Medical Institute ing a two-step binding reaction. Another group measured the equilibrium Grant 55005604 (to E. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked dissociation constant, K , and the rate constant of dissociation, k , for d off “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 human eIF4E by surface plasmon resonance (SPR) (37). Surprisingly, the To whom correspondence should be addressed: 1501 Kings Highway, Shreveport, LA 71130-3932. Tel.: 318-675-5161; Fax: 318-675-5180; E-mail: [email protected]. experimentally determined k values and the calculated k values (K off on d The abbreviations used are: eIF4E, eukaryotic initiation factor 4E; ARCA, anti-reverse cap k /k ) differed from those reported for the stopped-flow studies (35, 36) off on analog; DTT, dithiothreitol; eIF4E(P), human eIF4E phosphorylated at the physiological by 2–3 orders of magnitude. site, which is Ser-209 for the human protein; IEF, isoelectric focusing; k , the observed obs rate constant; k , the rate constant for dissociation; k , the rate constant for association; off on Mammalian eIF4E is phosphorylated at Ser-209 (38, 39). Although 7 1 3 7 m GpppG, P -7-methylguanosine-5 P -guanosine-5 triphosphate; m GpppN, same as 7 7,3-O 1 several eIF4E kinases have been reported, the strongest evidence points m GpppG except with any nucleotide base in place of G; m GpppG, P -3-O,7- dimethylguanosine-5 P -guanosine-5 triphosphate; Mnk, mitogen-activated pro- to Mnk1 and Mnk2 as the physiological kinases (40, 41). Mnk is acti- tein kinase-interacting kinase; RU, response units; SPR, surface plasmon resonance; vated via the extracellular signal-regulated kinase (ERK) and p38 path- GST, glutathione S-transferase; MOPS, 4-morpholinepropanesulfonic acid; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; V, volt(s). ways in response to mitogens, cytokines, or cellular stress (40–42). MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14927 This is an Open Access article under the CC BY license. Kinetic Analysis of Cap Binding to eIF4E Several types of observations link increased eIF4E phosphorylation with pET11d-eIF4E (64). Protein expression was induced with 0.4 mM iso- increased rates of protein synthesis, including stimulation of cultured propyl--D-thiogalactopyranoside overnight at 15 °C. We found that cells with mitogens (43–45), use of the Mnk inhibitor CGP57380 (46, these conditions (lowering the isopropyl--D-thiogalactopyranoside 47), sequence variants of eIF4E that cannot bind eIF4G (48), shut-off of concentration and temperature) increased the yield of soluble eIF4E cellular protein synthesis by adenovirus infection (49), and induction of compared with the published procedure (64) and produced 7–10 mg/ apoptosis (50). The opposite conclusion, that eIF4E phosphorylation liter of cell suspension. Cells (3 g) were lysed with a French press in 25 does not affect the rate of protein synthesis, or even decreases it, has ml of buffer A (50 mM HEPES/KOH, pH 7.6, 0.3 M KCl, 10 mM EDTA, 1 been drawn from studies with Mnk overexpression (51), cell-free trans- mM DTT, 1% streptomycin sulfate, and protease inhibitor mixture). lation (52), recovery of cultured cells from hypertonic stress (53, 54), and After centrifugation at 32,000 g for 40 min at 4 °C, the soluble fraction Indinavir, a human immunodeficiency virus protease inhibitor (55). was diluted 1:1 with 100 M GTP and loaded onto a 3-ml m GTP- Whole animal studies have also generated conflicting results. Trans- Sepharose column. This and other chromatographic steps were moni- genic Drosophila that express a non-phophorylatable form of eIF4E are tored spectrophotometrically at  . The column was washed with 280 nm small, have morphological defects, and are less viable (56). The Drosoph- 40 column volumes of buffer B (50 mM HEPES/KOH, pH 7.6, 0.2 M KCl, ila homolog of Mnk, Lk6, is dispensable under a high protein diet, but its 1mM EDTA, and 1 mM DTT). eIF4E was eluted with buffer B except loss causes growth reduction when amino acids in the diet are restricted that the KCl concentration was 0.5 M KCl. Use of KCl for elution rather (57, 58). Yet when Mnk1 and Mnk2 are knocked out in mice, no phe- than the more traditional m GTP (65) prevents formation of tight eIF4E notype is observed (59). Thus, at the levels of protein synthesis, cell complexes with the cap analog (66), which can affect the measurement growth, and intact animal physiology, the roles of Mnk and eIF4E phos- of kinetic parameters for the binding reaction. eIF4E was at least 95% phorylation are controversial. pure as judged by SDS-PAGE (see “Results”). For eIF4E purification The current study was motivated by several gaps in our understanding of from E. coli cells and Mnk purification from HEK 293 cells, protein eIF4E-cap interactions as well as discrepancies in the published literature. concentrations were determined with the protein assay reagent from First, protein synthesis is a dynamic process and, during each round of Bio-Rad, in which bovine serum albumin was used as standard. For initiation, the mRNA cap is presumably both bound and released by eIF4E. kinetic experiments, protein concentrations were determined spectro- 1 1 A full understanding of the biochemical consequences of eIF4E phospho- photometrically, assuming   53,400 M cm at pH 7.2. 280 nm rylation therefore requires a knowledge of k and k values for both phos- on off eIF4E Phosphorylation—Recombinant mouse Mnk kinase was pro- phorylated and unphosphorylated eIF4E (eIF4E and eIF4E(P)). Second, it is duced as described previously (42) but with modifications. Plasmids pEBG- not known whether the discrepancies in kinetic values noted above (35–37) Mnk2 and pEBG-Mnk1T2A2 were generously provided by Christopher result from the methodologies employed (SPR versus stopped-flow), the Proud (University of British Columbia). These encode, respectively, wild- forms of eIF4E studied (full-length versus N-terminal truncated), or the type Mnk2 fused to glutathione S-transferase (GST::Mnk2) and an inac- methods of producing eIF4E(P) (enzymatic phosphorylation versus intein- tive variant, GST::Mnk1T2A2 (41). HEK 293T cells, grown in 10-cm mediated ligation). Third, the k for eIF4E(P)-cap analog association has on dishes in Dulbecco’s modified Eagle’s medium supplemented with 10% not yet been measured directly. Fourth, kinetic parameters measured by the fetal bovine serum, were transfected with plasmids in complex with stopped-flow technique have not been reported for the binding of capped Lipofectamine 2000 (Invitrogen). After 44 h, cells were harvested and oligoribonucleotides to eIF4E. Fifth, all previous studies of eIF4E interac- lysed at 4 °C by incubation with buffer C (10 mM HEPES/KOH, pH 7.4, tions with capped oligonucleotides (20, 22, 37, 60) have utilized a mixture of 50 mM NaF,2mM EDTA, 2 mM sodium orthovanadate, 0.1% 2-mercap- normally capped and reverse-capped oligoribonucleotides, the latter of toethanol, 1% Triton X-100, and protease inhibitor mixture) at a ratio of which do not bind eIF4E. Sixth and finally, the proposed two-step mecha- cell pellet to buffer C of 1:8 (w/v) over 30 min with rotation. The sus- nism for cap binding to eIF4E (26, 34, 35) is at variance with observations for pension was centrifuged for 30 min at 20,000  g, and GST::Mnk2 was a viral cap-binding protein (61, 62). purified from the supernatant on glutathione-Sepharose 4B equili- We report here stopped-flow kinetic results for full-length human eIF4E brated with buffer D (20 mM MOPS/KOH, pH 7.4, 20 mM KCl, 15 mM and eIF4E(P) binding to cap analogs and capped oligoribonucleotides, the MgCl , 0.5 mM EDTA, 1 mM DTT, 25 mM -glycerolphosphate, 1 mM latter being capped entirely in the correct orientation. Our findings differ sodium orthovanadate, and 5% glycerol) at a supernatant:Sepharose from those of previous reports with regard to the magnitude of kinetic ratio of 10:3 (v/v). Glutathione-Sepharose beads with bound GST::Mnk2 constants, the effects of eIF4E phosphorylation on k and k , and the on off and GST::Mnk1T2A2 were stored at 80 °C. mechanism of cap binding. With regard to the latter, our data support a Recombinant eIF4E was phosphorylated in vitro essentially as described one-step rather than two-step binding model. Furthermore, the stopped- (37) but with some modifications. Recombinant eIF4E (225 g) was incu- flow results provide a kinetic basis for enhanced binding of eIF4E to capped bated at room temperature (23 °C) in a 0.5-ml reaction mixture containing oligoribonucleotides compared with cap analogs. 150 l (packed volume) of glutathione-Sepharose-bound GST::Mnk2 and 500 M ATP in buffer D for 6 h with rotation. The kinase was removed by EXPERIMENTAL PROCEDURES centrifugation, and then eIF4E(P) was passed over a PD-10 desalting col- umn (Amersham Biosciences) to separate it from ATP and to transfer it Materials—All common reagents were of analytical grade and were 7 7 purchased from Sigma unless otherwise stated. m GTP, m GTP-Sepha- into an appropriate buffer for kinetic experiments. In parallel experi- ments, eIF4E was mock-phosphorylated with GST::Mnk1T2A2. The rose 4B, and glutathione-Sepharose 4B were purchased from Amer- efficiency of eIF4E phosphorylation was determined by isoelectric sham Biosciences. Protease inhibitor mixture was from Roche Diagnos- 7 7,3-O tics. The syntheses of m GpppG and m GpppG were performed as focusing on CleanGel IEF polyacrylamide gels (Amersham Biosciences) described previously (63). The concentrations of dinucleotide cap ana- in the presence of 8 M urea, 30 mM CHAPS, and 2% Pharmalyte 3–10 3 1 log solutions were determined by absorbance (  22.6 10 M according to the manufacturer’s instructions. IEF was conducted at 255 nm cm at pH 7.0). 15 °C on a Multiphor II apparatus for horizontal separation equipped eIF4E Expression and Purification—Full-length wild-type human with a cooling platform and MultiDrive XL power supply (Amersham eIF4E (14) was expressed in Escherichia coli BL21 (COE) pLys(S) from Biosciences). Protein bands were stained with Coomassie Blue. In pre- 14928 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E liminary experiments to determine the optimal time for enzymatic 1000 data points throughout the reaction (50–200 ms). Fluorescence phosphorylation, we also used [- P]ATP. changes were examined up to 200 s. The binding reaction was investigated under pseudo first-order condi- Synthesis of Oligoribonucleotides—T7 RNA polymerase was prepared as previously described (67). The template consisted of two comple- tions ([protein] [ligand]) unless otherwise noted. For most experiments, mentary oligodeoxyribonucleotides (Integrated DNA Technology Inc., concentrations of protein and ligand in the syringes of the stopped-flow instrument were 0.1–0.2 and 1–20 M, respectively. Samples were Coralville, IA), 5-TAATACGACTCACTATAG-3 and 5-TTTTT- degassed prior to loading into the syringes. Reactions were initiated by ATGCGCCTATAGTGAGTCGTATTA-3, which contain the T7 pro- moter followed by a single-stranded 12-nucleotide 5-overhang. This mixing equal volumes of protein and ligand at 25 °C in buffer E except containing 50, 100, 150, or 350 mM KCl as noted. The temperature was sequence was chosen because it gave the highest yield of several tested controlled with a circulating external water bath (T  0.2 °C). The (68). The oligodeoxynucleotides, each at 183 M in 20 mM Tris-HCl, pH stopped-flow traces shown under “Results” are the average of 5–15 individ- 7.5, were annealed by heating at 90 °C and slow cooling to room tem- ual traces. Control experiments on specificity of binding were conducted perature. Transcription reactions were performed as described previ- with GpppG and with the uncapped form of the oligoribonucleotide. For ously (68) with some modifications. Reaction mixtures (1 ml) con- stopped-flow experiments where the m GpppG concentration was limit- tained 40 mM HEPES/KOH, pH 7.9, 10 mM MgCl ,2mM spermidine, ing (0.2–0.4M), the protein concentration was varied in the range of 1–10 10 mM NaCl, 10 mM DTT, 0.1 mg/ml bovine serum albumin, 1000 M. Dissociation of the pre-formed eIF4Em GpppG complex was fol- units/ml RNA Guard (Amersham Biosciences), 2.5 mM each of ATP, lowed by measuring the increase in intrinsic Trp fluorescence when equal CTP, UTP, and GTP, 40 g/ml DNA template, and 100 gof volumes of the complex and buffer were mixed in the stopped-flow cell at T7 RNA polymerase. Capped oligoribonucleotides were produced 150 mM KCl. In this case, eIF4E and m GpppG concentrations in the by lowering the GTP concentration to 0.5 mM and including 7,3-O syringe were equal (2 M). m GpppG, an “anti-reverse cap analog” (ARCA), at 3 mM, Curve Fitting and Error Analysis—Stopped-flow traces representing which prevents heterogeneity in synthesized oligoribonucleotides binding of cap analogs or a capped oligoribonucleotide to eIF4E or due to reversed incorporation (63). The binding affinities of 7 7,3-O eIF4E(P) were analyzed using Curfit, a curve-fitting program that uses a m GpppG and m GpppG for eIF4E are the same within 0.6% Marquardt algorithm (71). Data were fit to both single- and double- (69). Reaction mixtures were incubated at 35 °C for 2 h, after which exponential functions. For single-exponential fits, the DNA template was digested with DNase RQ1 (Promega). Oli- goribonucleotides were purified by size-exclusion chromatography F(t)Fexp(k t) F (Eq. 1) obs on Bio-Gel P6 columns with diethylpyrocarbonate-treated water as eluent. The purity of oligoribonucleotides was determined by PAGE where k is the observed first-order rate constant,F is the amplitude, obs on 20% gels containing 8 M urea (70). The A /A ratio for oligori- and F is the final value of fluorescence. For double-exponential fits, 260 280 ∞ bonucleotides was 1.85. Concentrations (in g/ml) were determined F(t)F exp(k t) F exp(k t) F (Eq. 2) 1 obs1 2 obs2 as 33  A . 260 nm Steady-state Fluorescence Measurements—A StrobeMaster lifetime where k and k are the observed rate constants for the first and obs1 obs2 spectrometer (Photon Technology Inc., South Brunswick, NJ) with the second components of a double-exponential reaction, respectively, and SE-900 steady-state fluorescence option was used to obtain fluorescence F and F , are the amplitudes for the first and second components of 1 2 emission spectra of eIF4E and eIF4E(P) in the presence or absence of a double-exponential reaction, respectively. An assessment of each fit m GTP. Fluorescence was monitored at an excitation wavelength of 295 was made from the residuals, which measure the differences between nm (2-nm bandwidth) and emission bandwidth of 5 nm. Spectra were actual data and the calculated fit. These are designated  in the case of recorded with a photon counting detector (Photon Technology Inc., model single-exponential fits and  in the case of double-exponential fits. The 710) from 305 to 400 nm at 25 °C. Samples were maintained in a quartz program KaleidaGraph (Synergy Software, Reading, PA; version 3.06) cuvette (1-cm path length) with constant stirring at concentrations of both was used for least-squares fitting of data with linear equations and deter- protein and m GTP of 1M in buffer E (50 mM HEPES/KOH, pH 7.2, 1 mM mination of standard errors for parameters obtained from the fits. Equi- DTT, 0.5 mM EDTA, and 50 mM KCl). Under these conditions, the K for librium constants are reported either as dissociation constants (K )or the m GTP-eIF4E interaction is 9nM (26). Thus, essentially all of the association constants (K  1/K ). a d m GTP is bound to eIF4E. The degree of quenching of intrinsic Trp fluo- RESULTS rescence (Q) is an indicator of both purity and native conformation of eIF4E. For mammalian eIF4E, Q is 65% (26). The preparations used in Preparation of eIF4E and eIF4E(P)—The objective of the current this study had Q  64  4% for eIF4E and Q  59  5% for eIF4E(P), study was to apply fast kinetic techniques to elucidate the mecha- indicating that both proteins were essentially native. nism of eIF4E-cap interactions, with particular emphasis on the Kinetic Methods—Rapid kinetic measurements of eIF4E and effect of three parameters: phosphorylation of eIF4E, ionic strength, eIF4E(P) interaction with free dinucleotide cap analogs (m GpppG and linkage of the cap to an oligoribonucleotide chain. This required 7,3-O and m GpppG) and an ARCA-capped oligoribonucleotide were preparation of milligram quantities of highly purified, native eIF4E made with a stopped-flow spectrometer with fluorescence detection and eIF4E(P). In previous reports in which recombinant eIF4E was (Applied Photophysics Ltd., Leatherhead, UK, model SX.18MV). The used for kinetic studies, the protein was purified from inclusion bod- dead time of the instrument was 1.2 ms. The excitation wavelength was ies solubilized with either urea (37) or guanidine hydrochloride (35, 295 nm with 2.3-nm bandwidth, and stray excitation radiation was elim- 36) and then renatured by dialysis. It is possible to calculate what inated with a 320-nm Oriel long-pass filter. Excitation occurred through fraction of a given eIF4E preparation is in the native state by deter- a 2-mm path in the stopped-flow optical cell, and emission was meas- mining the quenching of intrinsic Trp fluorescence (Q) by steady- ured through a 10-mm path. The instrument time constant was 0.5% of state fluorescence spectroscopy (see “Experimental Procedures”). In the reaction half-time. For very fast reactions, no in-line filtering was the only case where this was reported, 57  16% of the renatured used. An oversampling option of the instrument permitted us to collect protein preparation was found to be active, presumably due to MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14929 Kinetic Analysis of Cap Binding to eIF4E imperfect refolding in vitro (35). To avoid this problem, we purified cedures”). Another possible source of inactive protein is the pres- human eIF4E exclusively from the soluble fraction of E. coli lysates. ence of contaminating cap analog in complex with eIF4E (66), due to The protein appeared as a homogeneous band by SDS-PAGE (Fig. the conventional method of eluting eIF4E from m GTP-Sepharose 1A). eIF4E prepared in this way had a value of Q indicating that it with m GTP (65). Such eIF4E preparations contain up to 60% bound existed entirely in the native conformation (see “Experimental Pro- m GTP, which cannot be removed by dialysis or ion exchange chro- matography (26). To avoid this problem, we eluted the protein from 7 7 m GTP-Sepharose with 0.5 M KCl rather than m GTP. eIF4E(P) was prepared by in vitro phosphorylation with activated mouse Mnk2. It has previously been shown that this kinase phospho- rylates human eIF4E exclusively at Ser-209 in vitro (41, 42), which is the same as the in vivo site (38, 39). The degree of phosphorylation was measured by two methods: IEF and calculating the moles of phosphate incorporated per mole of eIF4E from the specific activity of [- P]ATP in the kinase reaction. Both methods revealed that phosphorylation was essentially stoichiometric (Fig. 1B and data not shown). Both untreated and kinase-treated proteins appeared on the IEF gels as doublets, the minor form constituting 11% of the total. The minor form has been observed previously in preparations of both phosphorylated and unphosphorylated eIF4E (72, 73). The structural basis for the heteroge- neity is unknown. It may represent post-translational modification, e.g. incomplete N-terminal acetylation (74), or it may be an artifact of IEF. For instance, the concentration of CN in the 8 M urea used in the IEF gel can reach 20 mM and can carbamylate NH groups in proteins (75), removing one positive charge. Regardless of the sources of microhet- FIGURE 1. Characterization of purified recombinant human eIF4E and eIF4E(P). A, erogeneity, both forms had nearly the same molecular mass (Fig. 1A), SDS-PAGE (10% gel) of eIF4E after column chromatography on m GTP-Sepharose. The were able to bind m GTP-Sepharose, and were stoichiometrically phos- mobilities of molecular mass markers are indicated on the left. B, phosphorylation of eIF4E as monitored by IEF. Recombinant human eIF4E was treated with in vivo activated phorylated (Fig. 1B). eIF4E(P) prepared in this way had a value of Q GST::Mnk2 and analyzed by IEF in the range of pH 3–10, as described under “Experimen- indicating that it existed entirely in the native conformation (see “Exper- tal Procedures.” The polarity during IEF is indicated by and. In both A and B, protein is visualized by Coomassie Blue staining. imental Procedures”). FIGURE 2. Kinetics of the m GpppG-induced decrease in intrinsic Trp fluorescence of eIF4E and eIF4E(P) as monitored by stopped-flow fluorescence. A, reaction of 0.1 M eIF4E with m GpppG at 350 mM KCl and 25 °C (upper panel). Solid lines represent fitting of the data with Equation 1. Fluorescence is measured in volts (V). Traces are offset vertically 7 1 to allow each to be seen individually. For m GpppG at 1, 3, 5, and 9 M, k was 118, 176, 229, and 362 s , respectively. Residuals for fits to Equation 1 are shown in the lower panels. obs 7 1 B, same as A except that eIF4E(P) was used. For m GpppG at 1, 3, 5, and 9 M, k was 112, 164, 236, and 376 s , respectively. A control reaction for both proteins was performed with obs 3 7 9 M GpppG (top trace), which binds eIF4E with an affinity that is 10 -fold lower than that of m GpppG (26). 14930 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E Association of eIF4E with m GpppG—We initially attempted to one-step process, k is predicted to vary linearly with cap analog con- obs measure kinetic parameters for the interaction of m GTP, the most centrations (76) (Equation 3). commonly used cap analog, with eIF4E and eIF4E(P) by stopped-flow k  k [m GpppG] k (Eq. 3) experiments but found that rate of fluorescence change was too fast to obs on off be followed. We therefore turned to m GpppG, which binds eIF4E with 7 7 Also, under pseudo first-order conditions, the binding reaction is not only 7% the affinity of m GTP (26). However, m GpppG is more similar sensitive to small amounts of inactive protein in the preparation. to the natural mRNA cap than m GTP because of the second nucleoside The rapid mixing of eIF4E with m GpppG resulted in a decrease in moiety. We measured the reaction over a range of m GpppG concen- intrinsic Trp fluorescence that was dependent on m GpppG concen- trations under pseudo first-order reaction conditions, where [eIF4E] 7 tration (Fig. 2A, upper panel), the apparent rate constant of the [cap analog]. The first stopped-flow study of eIF4E-m GpppG interac- decrease being k . By contrast, there was no decrease in fluores- obs tions (34), which utilized wheat eIF(iso)4F, was also conducted under cence with GpppG (top trace), which binds eIF4E with an affinity pseudo first-order conditions. A later study of mouse eIF4E-(28–217) 3 7 that is 10 -fold lower than that of m GpppG (26). The data were fit (35) used second-order reaction conditions, i.e. the concentrations of with a single-exponential function (Equation 1) (solid lines). The eIF4E and cap analog were approximately equal, and both were varied. residuals, representing the deviation between the calculated and The latter experimental approach makes treatment of the data more actual data, indicate that the single-exponential function fit the complicated and can potentially cause artifacts (see below). Under points over the entire time of the measurement (lower panels). pseudo first-order conditions, if the association reaction is a simple Treatment of the data using a double-exponential function (Equa- tion 2) did not improve the fit (not shown). We followed the reaction over 200 s to test whether there was a second slow phase but saw no evidence for one (not shown). The binding reaction of eIF4E(P) also was probed over the same range of m GpppG concentrations (Fig. 2B, upper panel). The traces again followed a single-exponential function over all cap analog concentra- tions. The residuals did not vary over the time course (lower panels) nor were they diminished by a double-exponential fit (not shown). To ensure that any differences between eIF4E and eIF4E(P) were not an artifact due to nonspecific changes occurring during the 6-h kinase reaction, we also conducted a mock kinase reaction in which the active GST::Mnk2 was replaced with GST::Mnk1T2A2 (77) (see “Experimen- tal Procedures”). The k values for binding of m GpppG were indis- obs tinguishable between unphosphorylated eIF4E and mock-phosphoryl- ated eIF4E (data not shown). Plots of k versus m GpppG concentration for eIF4E and obs eIF4E(P) were linear (Fig. 3). The slope and y intercept are k and on k , respectively (Equation 3). The values obtained for k for both off on 6 1 1 eIF4E and eIF4E(P) were in the range of 33–292  10 M s , depending on the ionic strength, and those obtained for k were off 70–87 s (Table 1). Such high values indicate that equilibrium is reached rapidly and that slower methods of obtaining kinetic con- stants may not yield accurate results (see “Discussion”). K values calculated from these kinetic constants ranged from 0.24 to 2.48 M, again depending on the ionic strength (Table 1). These are somewhat higher than K values obtained by equilibrium methods (26, 32). For instance, at 100 mM KCl, the kinetic approach yielded K  0.45 M for eIF4E (Table 1), whereas the equilibrium approach yielded 0.135 M in one study (26) and 0.08 M in another (32). For eIF4E(P) at 100 mM KCl, kinetic measurements gave K  1.06 M, whereas equilib- FIGURE 3. Dependence of k on m GpppG concentration as a function of KCl con- obs rium measurements gave K  0.172 M (32). The reason for these centration. k obtained in experiments similar to those shown in Fig. 2 at the indicated obs 7 discrepancies is not known, but it should be noted that, besides the KCl concentrations is plotted versus [m GpppG] for eIF4E (panel A) and eIF4E(P) (panel B). Data are fit with Equation 3. Numerical values for k , k , and K are given in Table 1. method of K determination, the protein preparations were differ- on off d TABLE 1 Kinetic constant for association and dissociation of eIF4E and eIF4E(P) with m GpppG obtained from pre-steady-state experiments Data are obtained from experiments similar to Fig. 3, with application of Equation 3. k  10 k K on off d KCl concentration eIF4E eIF4E(P) eIF4E eIF4E(P) eIF4E eIF4E(P) 1 1 1 mMM s s M 50 292  16 138 870  18 73  12 0.24  0.06 0.53  0.09 100 184  10 80 483  23 85  12 0.45  0.13 1.06  0.16 150 102 756 287  15 79  6 0.85  0.16 1.41  0.12 350 33 234 282  12 72  8 2.48  0.39 2.12  0.27 MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14931 Kinetic Analysis of Cap Binding to eIF4E rate of caps with eIF4E but not the dissociation rate. This contradicts results obtained by SPR (37) (see “Discussion”). The KCl concentration dependence of K , as calculated from these kinetic parameters, closely resembled that of k (Fig. 4B). K values on a decreased 10-fold for eIF4E and 4-fold for eIF4E(P) from 50 to 350 mM KCl. The difference in K values between eIF4E and eIF4E(P) was most pronounced (2.2–2.4-fold) at 50 and 100 mM KCl. This was progressively eliminated at higher KCl concentrations, and by 350 mM, there was no difference in K values between eIF4E and eIF4E(P). The Apparent Kinetic Mechanism of m GpppG Binding Depends on eIF4E Concentration—The linearity of the k versus m GpppG plots obs (Fig. 3) is consistent with a one-step, second-order reaction mechanism, on 7 7 eIF4E m GpppG| -0 eIF4Em GpppG* off SCHEME 1 where the asterisk (*) denotes a reduction in Trp fluorescence of eIF4E in complex with m GpppG. This conflicts with published studies indi- cating that cap binding to eIF4E occurs in a two-step process (26, 34, 35). The authors of one of these reports (35) found that the experimental data could be fit better with a double- than a single-exponential func- tion. They attributed the slow phase to a second step in the binding reaction, in which eIF4E fluorescence is further reduced, k k 1 2 7 7 7 eIF4E m GpppG| -0 eIF4Em GpppG*| -0 eIF4Em GpppG** k k 1 2 SCHEME 2 where the asterisk (*) denotes a complex with reduced fluorescence relative to free eIF4E, and the double asterisk (**) denotes a different complex with even lower fluorescence. The mechanisms represented by Schemes 1 and 2 can be distinguished experimentally (see below). Under pseudo first-order conditions, the kinetic mechanism and parameters for a simple association between eIF4E and m GpppG FIGURE 4. Dependence of k and K for binding of m GpppG to eIF4E and eIF4E(P) on a (Scheme 1) theoretically do not depend on the choice of which com- on KCl concentration. Data for eIF4E (open circles) and eIF4E(P) (filled circles) are derived from Table 1. ponent is limiting. For the experiments shown in Figs. 2–4, the eIF4E concentration was limiting and the m GpppG concentration was varied. We measured the same association reaction but at a limit- ent: full-length human eIF4E phosphorylated enzymatically (Table ing m GpppG concentration (0.1 M) and variable eIF4E concen- 1) versus truncated mouse eIF4E phosphorylated by intein technol- trations, so that [m GpppG]  [eIF4E]. To our surprise, the kinetics of ogy (26, 32). Despite these differences in K values, the change in K d d the reaction under these conditions were different. As the eIF4E for eIF4E versus eIF4E(P) was similar for kinetic and equilibrium concentration was increased from 0.5 to 5 M, the stopped-flow determinations: a 2.35-fold increase in K by kinetic methods (Table traces became increasingly biphasic. Representative results for 0.5 1) and a 2.15-fold increase by equilibrium methods (32). and 2 M eIF4E are shown in Fig. 5, A–D. The same data for [eIF4E] The Effect of eIF4E Phosphorylation on k Is Diminished at High Ionic on 0.5 M are presented in Fig. 5, A and C, but with either single- or Strength—It is known that electrostatic forces play an important role double-exponential fits, respectively. The points were fit well by a sin- in determining K for eIF4E-cap analog interactions. Specifically, the 1 gle-exponential function (Equation 1), with k  71  4s and F obs affinity for cap analogs of natural rabbit reticulocyte eIF4E (16) and 0.0155 V (Fig. 5A). Application of a double-exponential function (Equa- recombinant mouse eIF4E-(28 –217) (26, 32) decreases with increas- tion 2) did not improve the fit (Fig. 5C). By contrast, at [eIF4E] 2.0 M, ing ionic strength. To determine the basis for the effect of KCl on K , d the data were fit poorly by a single-exponential function (Fig. 5B), we plotted k and k from Table 1 against KCl concentrations (Fig. on off whereas they were fit well by a double-exponential function (Fig. 1 1 4A). Over a range of KCl concentrations from 50 to 350 mM, k on 5D), with k  205  20 s and k  8.1  4s . The ampli- obs1 obs2 decreased 9-fold for eIF4E and 4-fold for eIF4E(P), whereas k tudes for the fast and slow phases were similar, 0.080 and 0.078 V, off did not change significantly (inset). Furthermore, there was no sta- respectively. The relative improvement of the fit with a double- tistical difference in k between eIF4E (open symbols) and eIF4E(P) 2versus single-exponential function, (   )/ , was 4. As the off s d d (filled symbols). Thus, phosphorylation diminishes the association eIF4E concentration was raised, the amplitude of the slow phase 14932 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E 7 7 7 FIGURE 5. Effect of eIF4E concentration on the kinetics of m GpppG binding to eIF4E, and the kinetics of m GpppG dissociation from a pre-formed eIF4Em GpppG complex. A, kinetic traces were obtained as described in the legend to Fig. 2 after mixing 0.1 M m GpppG with 1 M eIF4E at 350 mM KCl. (The final concentration of eIF4E was 0.5 1 7 M.). Data are fit with a single-exponential function, F(t)  0.0155 exp(71 s  t)  0.0135. B, same as A except that 0.2 M m GpppG and 4 M eIF4E were used. Data are fit with 1 1 a single-exponential function, F(t) 0.0871 exp(57.5 s  t) 0.0227. C, the same data points in A are shown but fit with a double-exponential function, F(t) 0.0092 exp(142 s 1 1 1 t) 0.00797 exp(39 s  t) 0.0143. D, the same data points in B are shown but fit with a double-exponential function, F(t) 0.08 exp(205 s  t) 0.078 exp(8.1 s  t) 0.012. 7 7 E, dissociation of the eIF4Em GpppG complex. eIF4E and m GpppG (2 M each) were mixed in one syringe of the stopped-flow apparatus in buffer E containing 150 mM KCl. After incubation for 5 min at 25 °C, the eIF4Em GpppG complex was diluted in the stopped-flow cell with an equal volume of the same buffer. The trace follows Equation 1 (solid line) with k being equal to 178  9s . obs progressively increased in relation to that of the fast phase (Table 2). eIF4E ^ (eIF4E) Our observation that the mechanism appears to change from a one- SCHEME 3 step to a two-step process as the eIF4E concentration is increased might be explained by formation of inactive eIF4E dimers (or multimers). We When the concentration of active eIF4E is lowered due to interaction suggest that there is a pre-existing equilibrium that becomes important with m GpppG during the stopped-flow experiment, the equilibrium at high protein concentrations. MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14933 Kinetic Analysis of Cap Binding to eIF4E shifts and the dimers slowly dissociate, yielding active monomers that- Fig. 5E shows a representative dissociation trace obtained at 150 mM can then bind ligand. KCl. The trace followed single-exponential kinetics during 20 s of meas- urement (only the first 0.1 s are shown) and yielded k  178  9s . obs on We used K determined in Table 1 to calculate the free concentrations 7 7 (eIF4E) ^ eIF4E m GpppG| -0 eIF4Em GpppG* 7 of eIF4E and m GpppG after dilution with buffer. Using these data and off k determined in Table 1, we calculated k for the reaction in the on off SCHEME 4 reverse direction according to the following equation. This dissociation reaction produces a second, slow phase of eIF4E- k  k  k ([cap]  [eIF4E] ) (Eq. 4) off obs on free free m GpppG interaction. Scheme 4 could possibly explain our results (Fig. 5D and Table 2) as well as those of previous studies, which were con- This yielded a value of k  84  18 s . This is the same, within off ducted at molar concentrations of either wheat eIF(iso)4F (34) or mouse experimental error, as k determined from the reaction conducted in off eIF4E-(28–2-17) (35) that were at least 5-fold higher than in Figs. 2–4. the forward direction, 87  15 s (Table 1). These results are consist- Thus, there are two alternative hypotheses to explain the double-expo- ent with a one-step but not two-step mechanism for cap binding to nential kinetics observed at high eIF4E concentrations: dissociation of eIF4E. inactive dimeric eIF4E (Scheme 4) and a two-step binding reaction Reaction of eIF4E and eIF4E(P) with a Capped Oligoribonucleo- (Scheme 2). tide—Although most studies of cap binding to eIF4E have been per- At 50 mM KCl, the authors of one study (35) calculated k  30 1 formed with mono- or dinucleotide cap analogs, the more physiologi- 1 1 1 s and k  1s .At150mM KCl, they obtained k  50 s and 2 1 cally relevant ligand is a capped RNA molecule. Unfortunately, use of k  6s . Thus, the step defined by k in Scheme 2 is rate- full-length mRNA would preclude steady-state or stopped-flow meas- 2 2 limiting for the reverse direction (dissociation of the urements based on fluorescence quenching because the optical density eIF4Em GpppG complex) in the two-step model. This provides a would be too high at the concentrations required for experiments means to distinguish experimentally between the one- and two-step such as shown in Fig. 2. We therefore chose to study a capped models. The dissociation kinetics of the eIF4Em GpppG complex oligoribonucleotide. can be followed by diluting the complex with buffer in the stopped- As noted under “Experimental Procedures,” use of conventional cap flow cell. A first-order rate constant is obtained from the increase in analogs in the in vitro synthesis of capped RNAs results in approxi- fluorescence upon dissociation of the complex. For the one-step mately half of the caps being incorporated in the incorrect orientation model with pre-equilibrium between monomers and dimers (63). Instead, we synthesized a capped 12-mer oligoribonucleotide con- 7,3-O (Scheme 4), k determined in the forward reaction is expected to be off taining the anti-reverse cap analog (ARCA) m GpppG (63) to equal to k determined in the reverse reaction. For the two-step off ensure that all caps were in the correct orientation. Previous equilib- 7,3-O 7 model (Scheme 2), the apparent k determined in the reverse reac- off rium measurements have shown that m GpppG and m GpppG do tion is the limiting k and is expected to be considerably smaller not differ significantly in their affinity for eIF4E (69). However, we than the apparent k determined in the forward reaction. off needed to determine whether the kinetic parameters were also the same 7,3-O 7 We therefore measured the rate of dissociation of the pre-formed for m GpppG and m GpppG. The results indicated that the kinetic eIF4Em GpppG complex by diluting the complex 2-fold with buffer. constants were statistically the same for binding of the conventional cap and the ARCA to eIF4E. For instance, at 100 mM KCl, k was 184 on TABLE 2 6 1 1 7 6 1 1 10  10 M s for m GpppG (Table 1) and 205  24  10 M s Effect of eIF4E concentration on the amplitudes of the first and the for the ARCA (Table 3). Similarly, k was statistically the same for off second phases for reaction of eIF4E with m GpppG at limiting 7 1 1 m GpppG (83 23 s ) and the ARCA (76 27 s ). This held true for m GpppG concentrations eIF4E(P) as well (Tables 1 and 3). Data are derived from experiments similar to Fig. 5D fit with Equation 2. Stopped-flow traces for the binding of the capped oligoribonucle- eIF4E concentration F F 1 2 otide to eIF4E and eIF4E(P) at 350 mM KCl are shown in Fig. 6. Reactions M % followed single-exponential kinetics, because application of a double- 1.0 69 631  30 1.5 61 839  5 exponential function did not improve the fit (data not shown). Plots of 2.0 52 748  7 k versus ligand concentration were linear (Fig. 7), indicating that the 3.0 30 770  2 obs binding of capped oligonucleotides also follows second-order kinetics in As noted under “Experimental Procedures,” the amplitudes of fluorescence changes F were measured in V. However, different stopped-flow instrument agreement with a one-step binding mechanism. Values for k and k on off settings were used for different eIF4E concentrations, so that a direct comparison were derived from the slope and y intercept, respectively, and are given of absolute values of F is not meaningful. Instead, the relative amplitudes at each eIF4E concentration are presented. in Table 3. TABLE 3 Kinetic constant for association and dissociation of eIF4E and eIF4E(P) with an ARCA and an ARCA-capped oligoribonucleotide Data are obtained from experiments similar to Fig. 7, with application of Equation 3. KCl concentration Protein Ligand 100 mM 350 mM 6 6 k  10 k K k  10 k K on off d on off d 1 1 1 1 1 1 M s s MM s s M eIF4E ARCA 205  24 76  27 0.37  0.14 41  4 107  11 2.61  0.37 eIF4E Capped oligo ND ND ND 101 411  10 0.11  0.10 eIF4E(P) ARCA 84 774  20 0.88  0.25 32  3 106  13 3.31  0.51 eIF4E(P) Capped oligo 394  54 103  50 0.26  0.13 37 318  6 0.49  0.17 ND, not determined because the rate was too fast to be measured. 14934 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E 1 1 M s . An even greater difference, 9.7-fold, was observed for k . The off combination of a faster association rate and a slower dissociation rate resulted in a K that was 24-fold lower for capped oligoribonucleotide compared with cap analog. Phosphorylation of eIF4E decreased k for the capped oligoribo- on nucleotide sufficiently that it could be measured at both 100 and 350 mM KCl (Table 3). The capped oligoribonucleotide associated with eIF4E(P) 4.7-fold faster than the cap analog at 100 mM KCl (k  394 54 10 on 6 1 1 versus 84  7  10 M s ). At 350 mM KCl, the effect of the oligori- bonucleotide chain was eliminated, because k values were statistically on the same for capped oligoribonucleotide and cap analog. Because this was not the case for unphosphorylated eIF4E, it appears that Ser-209 phosphorylation negates the increase in association rate contributed by the oligoribonucleotide chain, perhaps because of charge repulsion. No statistical difference in k could be detected between capped oligoribo- off nucleotide and cap analog at 100 mM KCl, but at 350 mM, the capped oligoribonucleotide dissociated 5.9-fold more slowly than the cap ana- log. For both eIF4E and eIF4E(P) at 350 mM KCl, the decrease in k is off the major determinant for the higher affinity of capped oligoribonucle- otide compared with cap analog. DISCUSSION FIGURE 6. Kinetics of capped oligoribonucleotide binding to eIF4E and eIF4E(P). Protein synthesis is a dynamic process in which cap binding and Reactions were conducted as described in the legend to Fig. 2 except with a 12-mer 7,3-O release presumably occur at least once for each round of initiation. We oligoribonucleotide capped with the anti-reverse cap analog (ARCA) m GpppG. A, reaction of eIF4E with the indicated concentrations of capped oligoribonucleotide at therefore sought to compare kinetic rather than equilibrium parameters 350 mM KCl. Traces follow Equation 1 (solid line) with k being 109, 211, 351, and 444 s obs for cap interaction with eIF4E versus eIF4E(P). Because the binding and for 1, 2, 3, and 4 M oligoribonucleotide, respectively. B, same as A except with eIF4E(P). 1 7 k was 55, 78, 119, and 205 s for 1, 2, 3, and 5 M m GpppG, respectively. dissociation reactions are fast, an appropriate method is to measure obs pre-steady-state kinetics in rapid mixing experiments. Whereas eIF4E and eIF4E(P) have been compared by equilibrium techniques, and eIF4E interactions with cap analogs have been studied by stopped-flow tech- niques, eIF4E and eIF4E(P) have not previously been compared by stopped-flow techniques. Furthermore, the binding of eIF4E or eIF4E(P) to capped oligoribonucleotides, which are structurally more similar to mRNA than cap analogs, has not previously been studied by stopped-flow techniques. Magnitude of Kinetic Constants—Our stopped-flow data for cap ana- logs are in reasonable agreement with those of Blachut-Okrasinska et al. 1 7 (35), who found k  24–35 s for the interaction between m GpppG off and mouse eIF4E-(28–217) in stopped-flow experiments conducted from 50 to 350 mM KCl, compared with our values of 72–85 s (Table 1). Our 2-fold higher k values may be because of human versus mouse eIF4E off (although they are 99% identical), the presence of the N-terminal 27 amino acid residues in our eIF4E preparations, or the fact that the nat- ural Thr-205 is replaced with Cys-205 in eIF4E-(28–217) phosphoryl- ated by intein technology. Dlugosz et al. (36) did not study m GpppG, 7 7 but they did obtain k values for m GTP and m GDP by stopped-flow off measurements of mouse eIF4E-(28–217) that were in the range of 40–75 s . Similarly, k for vaccinia virus VP39 was reported to be off 41–60 s (61). Our k values for eIF4E and eIF4E(P) interaction with a capped oli- off goribonucleotide are at variance with the findings of Scheper et al. (37), 7,3-O FIGURE 7. Dependence of k for binding of eIF4E to either the ARCA m Gp- obs 2 despite the fact that human eIF4E was used in both cases and the ppG or to an ARCA-capped 12-mer oligoribonucleotide. Data are derived from exper- iments similar to those shown in Fig. 6. Experiments were conducted at either 350 mM KCl method of phosphorylation was the same. These authors determined (panel A) or 100 mM KCl (panel B). Data are fit with Equation 3. Numerical values for k , on k for a capped 18-mer oligoribonucleotide at 100 mM KCl to be 0.06 off k , and K are given in Table 3. off d 1 1 s for mock-phosphorylated eIF4E, 0.078 s for untreated eIF4E, and 0.58 s for eIF4E(P), whereas our k value for eIF4E(P) under similar For unphosphorylated eIF4E, the binding of capped oligoribonucle- off otide was too fast to be measured by the stopped-flow technique at 100 conditions was 103 s , or 186-fold greater (Table 3). This discrepancy mM KCl. We therefore estimate that it is close to the diffusion limit (1 may be because of the inability of SPR to measure kinetic parameters for 9 1 1 10 M s ). At 350 mM KCl, k was 2.5-fold greater for capped oli- reactions as fast as the binding of eIF4E to cap analogs or capped oli- on 6 6 goribonucleotide than cap analog, 101  4  10 versus 41  4  10 goribonucleotides. The upper limits for k and k measured by SPR are on off MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14935 Kinetic Analysis of Cap Binding to eIF4E 7 1 1 1 10 M s and 0.1 s , respectively (78), which are several orders of not observe biphasic kinetics until eIF4E concentrations exceeded 0.5 magnitude lower than with stopped-flow techniques. This conclusion is M (Table 2). Third, the value of k determined from the reverse reac- off further supported by the low analyte response observed by Scheper et al. tion is the same as k determined from the forward reaction. off (37). From the 1:1 stoichiometry of the binding reaction, the molecular Effect of eIF4E Phosphorylation on Cap Binding—Two similar models masses of a capped 18-mer oligoribonucleotide and eIF4E, and the have been proposed for the role of eIF4E phosphorylation in translation ligand response (79), one can calculate a theoretical analyte response of initiation. Marcotrigiano et al. (11) suggested a “clamping” mechanism 586 RU. The analyte responses reported by Scheper et al. (37) do not in which the generation of a salt bridge of Ser-209 with Lys-159 acts as a exceed 200 RU for eIF4E or 50 RU for eIF4E(P). One of the causes for a clamp that stabilizes the cap in the binding slot. Tomoo et al. (13) used low maximum analyte response is when reaction rates exceed the range molecular dynamics simulations to suggest that a hydrogen-bonded measurable by the SPR technique (78, 80) cluster of water molecules or polar amino acid residues could form Mechanism of eIF4E Association with Cap Analogs as Inferred from around the dianionic form of Ser(P)-209. This could potentially block Kinetics Measurements—Cap binding to eIF4E has been studied exten- the release of the cap from the binding slot. In both cases, stabilization is sively by both equilibrium and kinetic methods (see Introduction). Most envisioned to facilitate the assembly of initiation complexes loaded with of these studies have utilized a decrease in intrinsic Trp fluorescence mRNA. The principal prediction from both hypotheses is a significant upon cap binding, although some have utilized SPR (37), isothermal decrease in k upon phosphorylation of eIF4E because of closure of the off titration calorimetry (87), or NMR (12). Results from some of these cap-binding pocket. Also, in such a mechanism, k is expected to be on studies have led the authors to propose a two-step binding mechanism significantly impaired if mRNA were to interact with eIF4E in which a (26, 34, 35). The first step is envisioned as being the ligand entering the salt bridge had already been formed between Ser-209 and Lys-159. A cap-binding slot and anchoring, via the triphosphate moiety, to basic different type of model was proposed by Scheper et al. (37) on the basis amino acid residues. The second step is a change within the m G-bind- of their results from SPR experiments showing that phosphorylation of ing slot that leads to a further fluorescence decrease. Molecular dynam- eIF4E accelerates the rate of cap dissociation. They speculated that ics simulations have also supported a model in which there is a confor- phosphorylation facilitates the release of eIF4E and other initiation fac- mational change in eIF4E upon cap binding (13) (although this does not tors from the 5-end of the mRNA. per se constitute proof for two kinetically distinct steps). By contrast, a Our determination of k and k for eIF4E and eIF4E(P) allows us to on off one-step binding model has been proposed for another cap-binding test these predictions directly. We did not find any significant difference protein, the vaccinia virus VP39 (61). Cap binding by VP39 involves a 7 in k values between eIF4E and eIF4E(P) at any salt concentrations off cation- sandwich of m G between two aromatic amino acid residues, investigated, either for cap analogs or capped oligoribonucleotide similar to the cap-eIF4E interaction. However, in this case, phosphate (Tables 1 and 3). By contrast, Scheper et al. (37) found a 10-fold groups do not contribute to the binding process. increase in k for eIF4E(P) interaction with capped oligoribonucleotide Our data also show that association between eIF4E and cap analogs as off compared with eIF4E. Furthermore, k from that study (calculated well as oligoribonucleotides behave kinetically as a simple one-step on from K and k ) is slightly higher for eIF4E(P) binding to capped oli- process at all salt concentrations investigated and regardless of whether d off goribonucleotide compared with eIF4E, although we find k to be 2–3- eIF4E is phosphorylated. The discrepancy between our results and those on fold lower for eIF4E(P) compared with eIF4E, as measured with both cap of others (26, 34, 35) could be explained by the existence of a rate- limiting, parallel reaction resulting from dissociation of pre-formed analogs (Tables 1 and 3) and the capped oligoribonucleotide (at 350 mM KCl) (Table 3). Thus, our results do not support any of the previously eIF4E dimers or oligomers. We postulate an equilibrium between the reactive monomer and unreactive dimers or higher order oligomers proposed hypotheses. Instead, we propose that phosphorylation of Ser- (Scheme 3). The fast phase of fluorescence change is because of 209, which is located at the entrance to the cap-binding slot (Fig. 8A), m GpppG binding to monomeric eIF4E according to Scheme 1. The diminishes the rate of association by charge repulsion but has no effect slow phase is a result of the rate-limiting dissociation of unreactive on the rate of dissociation. This model is consistent with the observation oligomers to yield reactive monomers, which in turn react with the cap that the effect of eIF4E phosphorylation is progressively eliminated as analog according to Scheme 4. However, the cap-binding reaction per se the KCl concentration is increased (Refs. 32 and 35; Fig. 4A); at high salt is still a one-step process. concentrations, the charge on Ser(P)-209 is shielded and its inhibitory The evidence supporting this alternative mechanism is as follows. effect on the association rate is masked (Fig. 4A). First, under pseudo first-order reaction conditions (with [eIF4E] limit- Reaction of eIF4E and eIF4E(P) with Capped Oligoribonucleotides: an ing and0.5 M), the experimental data were fit by a single-exponential Additional Binding Site for mRNA?—We found that a capped 12-mer function and dependence of k on m GpppG concentrations was lin- obs oligoribonucleotide interacts with both eIF4E and eIF4E(P) in a very fast ear. Second, the slower second phase was observed only at elevated reaction. The addition of an oligoribonucleotide chain to the cap struc- eIF4E concentrations, which is consistent with concentration-depen- ture does not change the kinetic mechanism of binding. It is still a dant protein self-association. Whereas most of our experiments were one-step reaction, but both k and k are different for binding of the on off conducted at eIF4E concentrations of 0.1–0.2 M, those of Sha et al. capped oligoribonucleotide compared with the cap analog. For eIF4E(P) (34) were conducted at 0.5 M and those of Blachut-Okrasinska et al. at 100 mM KCl, k is 5-fold greater for capped oligoribonucleotide on (35) were conducted at 0.1–4.1 M. Examination of the latter data (35) than cap analog (Table 3). The reaction for unphosphorylated eIF4E is reveals that the double-exponential function improves the fit only when too fast to be measured by stopped-flow kinetics at 100 mM KCl, but at the eIF4E concentration is increased (Tables A1, A2, and A3). We did 350 mM, the reaction is slowed enough to reveal that k is 2.5-fold on greater for capped oligoribonucleotide than cap analog. Interestingly, The analyte response can be calculated from the expression (80), S  (AR/LR)  (LM/ the combination of two inhibitory effects on k , introduction of a phos- on AM), where S is the stoichiometry, AR is the analyte response, LR is the ligand response, LM is the ligand molecular mass, and AM is the analyte molecular mass. LR phate group on Ser-209, and shielding of positively charged amino acid is not given in Ref. 37 but is given in a publication cited in that work (85) as 150 RU for residues by high ionic strength, causes k to be the same for capped on SPR of eIF4E on immobilized capped oligoribonucleotides. LM and AM are 6,400 and 25,000 Da, respectively, and S is 1.0. AR is therefore calculated to be 586 RU. oligoribonucleotide and cap analog. 14936 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006 Kinetic Analysis of Cap Binding to eIF4E is k , which decreases 2–3-fold. Is such a difference biologically signif- on icant? Even though 48 S complex formation is thought to be rate-limit- ing for initiation of protein synthesis (81, 82), the association of mRNA with eIF4E and eIF4E(P) is so fast (as inferred from measurements with a capped oligoribonucleotide) that it may not be rate-limiting for 48 S initiation complex formation, especially because k is near the diffusion on limit for both eIF4E and eIF4E(P). Perhaps some other process in 48 S complex formation is rate-limiting. Arguing against this suggestion is the fact that the binding affinity of various cap analogs to eIF4E is pos- itively correlated with the translational efficiency of mRNA capped with those analogs, as measured both in vitro (63, 69, 83) and in vivo (84). This correlation would not hold if cap-eIF4E interactions did not make a significant contribution to the rate of initiation. Given that eIF4E is involved in recruitment of the unwinding machinery (1), it is possible that the effect of eIF4E phosphorylation on the rate of initiation is man- ifested in subsequent steps rather than cap binding per se. Acknowledgments—We are grateful to Christopher Proud for plasmids pEBG- Mnk2 and pEBG-Mnk1T2A2, Ewa Grudzien for advice on capped oligoribo- nucleotide synthesis, Srikripa Chandrasekaran for preparation of T7 polym- erase, and Stephan Witt for the use of the stopped-flow instrument. REFERENCES 1. Kapp, L. D., and Lorsch, J. R. (2004) Annu. Rev. Biochem. 73, 657–704 2. Rhoads, R. E., Joshi-Barve, S., and Rinker-Schaeffer, C. (1993) Prog. Nucleic Acids Res. Mol. Biol. 46, 183–219 3. Jankowska-Anyszka, M., Lamphear, B. J., Aamodt, E. J., Harrington, T., Darzyn- kiewicz, E., Stolarski, R., and Rhoads, R. E. (1998) J. Biol. Chem. 273, 10538–10542 4. Dyer, J. R., Pepio, A. M., Yanow, S. K., and Sossin, W. S. (1998) J. Biol. Chem. 273, 7 29469–29474 FIGURE 8. Three-dimensional models of human eIF4E in complex with m GpppA. 5. Rodriguez, C., Freire, M., Camilleri, C., and Robaglia, C. (1998) Plant J. 13, 465–473 Data are taken from Ref. 86. A, ribbon drawing showing phosphorylation at Ser-209. B, 6. Robalino, J., Joshi, B., Fahrenkrug, S. C., and Jagus, R. (2004) J. Biol. Chem. 279, space-filling model showing surface charge distribution. The color code for amino acid residues are: basic, green; acidic, magenta; hydrophobic, blue; and uncharged polar, yel- 10532–10541 low. The images were created using the Brookhaven Protein Data Bank data file 1WKW 7. Joshi, B., Cameron, A., and Jagus, R. (2004) Eur. J. Biochem. 271, 2189–2203 and RasMol. 8. Gorlich, D., and Mattaj, I. W. (1996) Science 271, 1513–1518 9. Strudwick, S., and Borden, K. L. (2002) Differentiation 70, 10–22 At high salt concentrations, a difference in k between capped oli- off 10. Richter, J., and Sonenberg, N. (2005) Nature 433, 477–480 goribonucleotide and cap analog also became apparent. For eIF4E, it 11. Marcotrigiano, J., Gingras, A.-C., Sonenberg, N., and Burley, S. K. (1997) Cell 89, was 10-fold lower for capped oligoribonucleotide, and for eIF4E(P), it 951–961 12. Matsuo, H., Li, H., McGuire, A. M., Fletcher, C. M., Gingras, A.-C., Sonenberg, N., was7-fold lower. Thus, both k and k contribute to a decrease in K on off d and Wagner, G. (1997) Nat. Struct. Biol. 4, 717–724 when comparing capped oligoribonucleotide to the cap analog: 24- 13. Tomoo, K., Shen, X., Okabe, K., Nozoe, Y., Fukuhara, S., Morino, S., Sasaki, M., fold for eIF4E and7-fold for eIF4E(P) at 350 mM KCl. These data are in Taniguchi, T., Miyagawa, H., Kitamura, K., Miura, K., and Ishida, T. (2003) J. Mol. Biol. agreement with previous equilibrium experiments where the binding of 328, 365–383 14. Rychlik, W., Domier, L. L., Gardner, P. R., Hellmann, G. M., and Rhoads, R. E. (1987) natural rabbit eIF4E to globin mRNA was found to be 5.3–5.5-fold 7 Proc. Natl. Acad. Sci. U. S. A. 84, 945–949 stronger than its binding to m GpppG (18, 22). Binding to capped oli- 15. McCubbin, W. D., Edery, I., Altmann, M., Sonenberg, N., and Kay, C. M. (1988) J. Biol. goribonucleotides was also stronger than to m GpppG but varied with Chem. 263, 17663–17671 the degree of oligoribonucleotide secondary structure (22). 16. Carberry, S. E., Rhoads, R. E., and Goss, D. J. (1989) Biochemistry 28, 8078–8083 17. Carberry, S. E., Darzynkiewicz, E., Stepinski, J., Tahara, S. M., Rhoads, R. E., and Goss, The pronounced increase in affinity for capped oligoribonucleotide D. J. (1990) Biochemistry 29, 3337–3341 suggests that some additional stabilizing interactions may be formed 18. Goss, D. J., Carberry, S. E., Dever, T. E., Merrick, W. C., and Rhoads, R. E. (1990) that slow dissociation of the larger ligand. TheG caused by addition Biochemistry 29, 5008–5012 of the oligoribonucleotide chain to the cap structure is 1.9 kcal/mol 19. Carberry, S. E., Darzynkiewicz, E., and Goss, D. J. (1991) Biochemistry 30, 1624–1627 20. Carberry, S. E., and Goss, D. J. (1991) Biochemistry 30, 4542–4545 for eIF4E and 1 kcal/mol for eIF4E(P). This is sufficient for the forma- 21. Carberry, S. E., and Goss, D. J. (1991) Biochemistry 30, 6977–6982 tion of an additional salt bridge or hydrogen bond between the body of 22. Carberry, S. E., Friedland, D. E., Rhoads, R. E., and Goss, D. J. (1992) Biochemistry 31, the mRNA and amino acid residues near the entrance of the cap-bind- 1427–1432 ing slot. There are several basic amino acid residues located on the 23. Minich, W. B., Balasta, M. L., Goss, D. J., and Rhoads, R. E. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 7668–7672 surface of eIF4E near the cap-binding slot (Arg-112, Arg-157, Lys-159, 24. Wang, Y. H., Sha, M., Ren, W. Y., van Heerden, A., Browning, K. S., and Goss, D. J. and Lys-162) (Fig. 8B). Additional basic residues are found on the lateral (1996) Biochim. Biophys. Acta 1297, 207–213 surface of a groove where the polynucleotide chain of mRNA conceiv- 25. Ren, J., and Goss, D. (1996) Nucleic Acids Res. 24, 3629–3634 ably could be bound (Lys-206, Lys-212, Lys-192, Lys-119, and Lys-52). It 26. Niedzwiecka, A., Marcotrigiano, J., Stepinski, J., Jankowska-Anyszka, M., Wyslouch- Cieszynska, A., Dadlez, M., Gingras, A.-C., Mak, P., Darzynkiewicz, E., Sonenberg, N., is possible that additional interactions occur between the polynucle- Burley, S. K., and Stolarski, R. (2002) J. Mol. Biol. 319, 615–635 otide chain and one or more of these basic side chains. 27. Wieczorek, Z., Zdanowski, K., Chlebicka, L., Stepinski, J., Jankowska, M., Kierdaszuk, Role of eIF4E Phosphorylation in Protein Synthesis—Our results indi- B., Temeriusz, A., Darzynkiewicz, E., and Stolarski, R. (1997) Biochim. Biophys. Acta cate that the only kinetic parameter affected by eIF4E phosphorylation 1354, 145–152 MAY 26, 2006• VOLUME 281 • NUMBER 21 JOURNAL OF BIOLOGICAL CHEMISTRY 14937 Kinetic Analysis of Cap Binding to eIF4E 28. Wieczorek, Z., Darzynkiewicz, E., and Lonnberg, H. (1998) J. Photochem. Photobiol. B 56. Lachance, P. E. D., Miron, M., Raught, B., Sonenberg, N., and Lasko, P. (2002) Mol. Biol. 43, 158–163 Cell. Biol. 22, 1656–1663 29. Wieczorek, Z., Niedzwiecka-Kornas, A., Chlebicka, L., Jankowska, M., Kiraga, K., 57. Arquier, N., Bourouis, M., Colombani, J., and Leopold, P. (2005) Curr. Biol. 15, 19–23 Stepinski, J., Dadlez, M., Drabent, R., Darzynkiewicz, E., and Stolarski, R. (1999) Z. 58. Reiling, J., Doepfner, K., Hafen, E., and Stocker, H. (2005) Curr. Biol. 15, 24–30 Naturforsch. 54, 278–284 59. Ueda, T., Watanabe-Fukunaga, R., Fukuyama, H., Nagata, S., and Fukunaga, R. (2004) 30. Zuberek, J., Jemielity, J., Stepinski, J., Lewdorowicz, M., Niedzwiecka, A., Haber, D., Mol. Cell. Biol. 24, 6539–6549 Stolarski, R., Rhoads, R., and Darzynkiewicz, E. (2003) Nucleosides Nucleotides Nucl. 60. Goss, D. J., Carberry, S. E., Dever, T. E., Merrick, W. C., and Rhoads, R. E. (1990) Acids 22, 1703–1706 Biochim. Biophys. Acta 1050, 163–166 31. Zuberek, J., Wyslouch-Cieszynska, A., Niedzwiecka, A., Dadlez, M., Stepinski, J., 61. Hu, G., Tsai, A.-L., and Quiocho, F. A. (2003) J. Biol. Chem. 278, 51515–51520 Augustyniak, W., Gingras, A.-C., Zhang, Z., Burley, S. K., Sonenberg, N., Stolarski, R., 62. Fechter, P., and Brownlee, G. G. (2005) J. Gen. Virol. 86, 1239–1249 and Darzynkiewicz, E. (2003) RNA (N.Y.) 9, 52–61 63. Stepinski, J., Waddell, C., Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (2001) 32. Zuberek, J., Jemielity, J., Jablonowska, A., Stepinski, J., Dadlez, M., Stolarski, R., and RNA (N.Y.) 7, 1486–1495 Darzynkiewicz, E. (2004) Biochemistry 43, 5370–5379 64. Stern, B. D., Wilson, M., and Jagus, R. (1993) Protein Expression Purif. 4, 320–327 33. Khan, M., and Goss, D. (2004) Biochemistry 43, 9092–9097 65. Webb, N. R., Chari, R. V. J., DePillis, G., Kozarich, J. W., and Rhoads, R. E. (1984) 34. Sha, M., Wang, Y. H., Xiang, T., van Heerden, A., Browning, K. S., and Goss, D. J. Biochemistry 23, 177–181 (1995) J. Biol. Chem. 270, 29904–29909 66. von der Haar, T., Ball, P. D., and McCarthy, E. G. (2000) J. Biol. Chem. 275, 35. Blachut-Okrasinska, E., Bojarska, E., Niedzwiecka, A., Chlebicka, L., Darzynkiewicz, 30551–30555 E., Stolarski, R., Stepinski, J., and Antosiewicz, J. M. (2000) Eur. Biophys. J. 29, 67. Zawadzki, V., and Gross, H. (1991) Nucleic Acids Res. 19, 1948 487–498 68. Milligan, J., Groebe, D., Witherell, G., and Uhlenbeck, O. (1987) Nucleic Acids Res. 15, 36. Dlugosz, M., Blachut-Okrasinska, E., Bojarska, E., Darzynkiewicz, E., and 8783–8798 Antosiewicz, J. (2003) Eur. Biophys. J. 31, 608–616 69. Jemielity, J., Fowler, T., Zuberek, J., Stepinski, J., Lewdorowicz, M., Niedzwiecka, A., 37. Scheper, G. C., van Kollenburg, B., Hu, J., Luo, I., Goss, D. J., and Proud, C. G. (2002) Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (2003) RNA (N.Y.) 9, 1108–1122 J. Biol. Chem. 277, 3303–3309 70. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1989) Molecular Cloning: A Laboratory 38. Joshi, B., Cai, A.-L., Keiper, B. D., Minich, W. B., Mendez, R., Beach, C. M., Stepinski, Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY J., Stolarski, R., Darzynkiewicz, E., and Rhoads, R. E. (1995) J. Biol. Chem. 270, 71. Bevington, P. R. (1969) Data Reduction and Error Analysis for the Physical Sciences, 14597–14603 McGraw-Hill, New York 39. Flynn, A., and Proud, C. G. (1995) J. Biol. Chem. 270, 21684–21688 72. Rychlik, W., Gardner, P. R., Vanaman, T. C., and Rhoads, R. E. (1986) J. Biol. Chem. 40. Fukunaga, R., and Hunter, T. (1997) EMBO J. 16, 1921–1933 261, 71–75 41. Waskiewicz, A. J., Flynn, A., Proud, C. G., and Cooper, J. A. (1997) EMBO J. 16, 73. Hiremath, L., Hiremath, S., Rychlik, W., Joshi, S., Domier, L., and Rhoads, R. (1989) 1909–1920 J. Biol. Chem. 264, 1132–1138 42. Scheper, G. C., Morrice, N. A., Kleijn, M., and Proud, C. G. (2001) Mol. Cell. Biol. 21, 74. Bradshaw, R., Brickey, W., and Walker, K. (1998) Trends Biochem. Sci. 23, 263–267 743–754 75. Righetti, P. G. (1983) Isoelectric Focusing: Theory, Methodology and Applications, 43. Kaspar, R. L., Rychlik, W., White, M. W., Rhoads, R. E., and Morris, D. R. (1990) J. Biol. Elsevier Science, Amsterdam Chem. 265, 3619–3622 76. Fersht, A. (1999) Structure and Mechanism in Protein Science, W. H. Freeman and 44. Rychlik, W., Rush, J. S., Rhoads, R. E., and Waechter, C. J. (1990) J. Biol. Chem. 265, Company, New York 19467–19471 77. Waskiewicz, A. J., Johnson, J. C., Penn, B., Mahalingam, M., Kimball, S. R., and 45. Manzella, J. M., Rychlik, W., Rhoads, R. E., Hershey, J. W. B., and Blackshear, P. J. Cooper, J. A. (1999) Mol. Cell. Biol. 19, 1871–1880 (1991) J. Biol. Chem. 266, 2383–2389 78. Myszka, D. (1997) Curr. Opin. Biotechnol. 8, 50–57 46. Worch, J., Tickenbrock, L., Schwable, J., Steffen, B., Cauvet, T., Mlody, B., Buerger, H., 79. Ptushkina, M., von der Haar, T., Karimm, M. M., Hughers, J. M., and McCarthy, Koeffler, H., Berdel, W., Serve, H., and Muller-Tidow, C. (2004) Oncogene 23, J. E. G. (1999) EMBO J. 18, 4068–4075 9162–9172 80. Pharmacia Biosensor AB (1994) BIApplications Handbook, Pharmacia Biosensor AB, 47. Walsh, D., and Mohr, I. (2004) Genes Dev. 18, 660–672 Uppsala, Sweden 48. Pyronnet, S., Imataka, H., Gingras, A.-C., Fukunaga, R., Hunter, T., and Sonenberg, N. 81. Darnbrough, C., Legon, S., Hunt, T., and Jackson, R. J. (1973) J. Mol. Biol. 76, 379–403 (1999) EMBO J. 18, 270–279 82. Safer, B., Kemper, W., and Jagus, R. (1978) J. Biol. Chem. 253, 3384–3386 49. Cuesta, R., Xi, Q., and Schneider, R. J. (2000) EMBO J. 19, 3465–3474 83. Grudzien, E., Stepinski, J., Jankowska-Anyszka, M., Stolarski, R., Darzynkiewicz, E., 50. Orton, K. C., Ling, J., Waskiewicz, A. J., Cooper, J. A., Merrick, W. C., Korneeva, N. L., and Rhoads, R. E. (2004) RNA (N.Y.) 10, 1479–1487 Rhoads, R. E., Sonenberg, N., and Traugh, J. A. (2004) J. Biol. Chem. 279, 84. Grudzien, E., Kalek, M., Jemielity, J., Darzynkiewicz, E., and Rhoads, R. E. (2006) 38649–38657 J. Biol. Chem. 281, 1857–1867 51. Knauf, U., Tschopp, C., and Gram, H. (2001) Mol. Cell. Biol. 21, 5500–5511 85. Ptushkina, M., von der Haal, T., Vasilescu, S., Frank, R., Birkenhager, R., and 52. McKendrick, L., Morley, S. J., Pain, V. M., Jagus, R., and Joshi, B. (2001) Eur. J. Bio- McCarthy, J. E. G. (1998) EMBO J. 17, 4798–4808 chem. 268, 5375–5385 86. Tomoo, K., Matsushita, Y., Fujisaki, H., Abiko, F., Shen, X., Taniguchi, T., Miyagawa, 53. Morley, S. J., and Naegele, S. (2002) J. Biol. Chem. 277, 32855–32859 H., Kitamura, K., Miura, K., and Ishida, T. (2005) Biochim. Biophys. Acta 1753, 54. Naegele, S., and Morley, S. J. (2004) J. Biol. Chem. 279, 46023–46034 191–208 55. Hong-Brown, L. Q., Brown, C. R., and Lang, C. H. (2004) Am. J. Physiol. 287, 87. Niedzwiecka, A., Stepinski, J., Darzynkiewicz, E., Sonenberg, N., and Stolarski, R. C1482–C1492 (2002) Biochemistry 41, 12140–12148 14938 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 21 •MAY 26, 2006

Journal

Journal of Biological Chemistry – American Society for Biochemistry and Molecular Biology

Published: May 26, 2006

There are no references for this article.