Further characterization of the helicase activity of eIF4A. Substrate specificity.

Eukaryotic initiation factor (eIF) 4A is the archetypal member of the DEAD box family of RNA helicases and is proposed to unwind structures in the 5'-untranslated region of mRNA to facilitate binding of the 40 S ribosomal subunit. The helicase activity of eIF4A has been further characterized with respect to substrate specificity and directionality. Results confirm that the initial rate and amplitude of duplex unwinding by eIF4A is dependent on the overall stability, rather than the length or sequence, of the duplex substrate. eIF4A helicase activity is minimally dependent on the length of the single-stranded region adjacent to the double-stranded region of the substrate. Interestingly, eIF4A is able to unwind blunt-ended duplexes. eIF4A helicase activity is also affected by substitution of 2'-OH (RNA) groups with 2'-H (DNA) or 2'-methoxyethyl groups. These observations, taken together with results from competitive inhibition experiments, suggest that eIF4A may interact directly with double-stranded RNA, and recognition of helicase substrates occurs via chemical and/or structural features of the duplex. These results allow for refinement of a previously proposed model for the mechanism of action of eIF4A helicase activity.

Eukaryotic initiation factor (eIF) 4A is the archetypal member of the DEAD box family of RNA helicases and is proposed to unwind structures in the 5-untranslated region of mRNA to facilitate binding of the 40 S ribosomal subunit. The helicase activity of eIF4A has been further characterized with respect to substrate specificity and directionality. Results confirm that the initial rate and amplitude of duplex unwinding by eIF4A is dependent on the overall stability, rather than the length or sequence, of the duplex substrate. eIF4A helicase activity is minimally dependent on the length of the single-stranded region adjacent to the double-stranded region of the substrate. Interestingly, eIF4A is able to unwind blunt-ended duplexes. eIF4A helicase activity is also affected by substitution of 2-OH (RNA) groups with 2-H (DNA) or 2-methoxyethyl groups. These observations, taken together with results from competitive inhibition experiments, suggest that eIF4A may interact directly with double-stranded RNA, and recognition of helicase substrates occurs via chemical and/or structural features of the duplex. These results allow for refinement of a previously proposed model for the mechanism of action of eIF4A helicase activity.
The initiation of protein synthesis in eukaryotic systems is an intricate process involving at least 12 protein factors that work in concert to bring the mRNA, the initiating methionyl-tRNA, and the 40 S and 60 S ribosomal subunits together to form an 80 S complex capable of peptide chain elongation (see Refs. 1-3 for reviews on translation initiation). In this process, eukaryotic initiation factor (eIF) 1 4A works in conjunction with eIF4B, eIF4F, and eIF4H to promote the binding of mRNA to the 40 S ribosomal subunit (1,3,4). eIF4A is proposed to function in this important regulatory step by unwinding secondary structure in the 5Ј-untranslated region of the mRNA, thus facilitating the binding of the 40 S ribosomal subunit to the mRNA and allowing for subsequent scanning to the initiator AUG codon.
eIF4A is the archetype of the DEAD box family of proteins (5). DEAD box (and related DEXH box) proteins contain eight highly conserved amino acid sequence motifs and have been implicated in virtually every cellular process involving RNA unwinding and/or rearrangement. These include transcription, ribosomal biogenesis, pre-mRNA splicing, RNA export, translation, and RNA degradation (6,7). In vitro analysis of this family of proteins has demonstrated that all members possess RNA-dependent ATPase activity, and many are capable of melting short RNA duplex structures in an ATP-dependent manner (8 -13). In addition to providing the energy for unwinding RNA duplexes, it has been postulated that the ATPase activities of DEAD/DEXH proteins may function in part as kinetic proofreading devices and maintain fidelity in complex biological processes (14,15). Furthermore, many viruses also encode DEAD/DEXH proteins that exhibit RNAor DNA-dependent ATPase activities and ATP-dependent RNA or DNA unwinding activities, which are presumably needed for viral replication (16 -22). Recently, Jankowsky et al. (23) have elegantly demonstrated that the NPH-II protein of vaccinia virus is able to function as a processive and directional RNA helicase.
The eight conserved motifs of eIF4A span 400 amino acid residues and represent the core sequence of DEAD/DEXH proteins in the superfamily II group of helicases (5,6,24). Extensive mutational analyses of the conserved regions of eIF4A and other DEAD/DEXH proteins have demonstrated that they are important for ATP binding, ATP hydrolysis, RNA binding, RNA unwinding, and coupling of these different activities (17,(25)(26)(27)(28)(29)(30). Two groups have presented partial crystal structures of yeast eIF4A (31,32). Both show that the three-dimensional fold, strand topology, and the orientation of conserved motifs in the N-terminal portion of eIF4A are strikingly similar to domain I of the HCV NS3, PcrA, and Rep helicases (31,32). Recently, the crystal structure of the C-terminal portion eIF4A has been solved and again shows that tertiary structure and positions of the conserved motifs in eIF4A have the same topology as the equivalent domains of the Rep, PcrA, and HCV NS3 helicases (33). Such similarities strongly suggest a common structural and mechanistic theme among these helicases (34).
eIF4A has been the subject of intense biochemical study. A detailed kinetic and thermodynamic analysis of the RNA-activated ATPase activity of eIF4A has been performed, which showed that the binding of RNA and ATP to eIF4A is coupled (35). These data, along with limited proteolysis experiments, suggest that eIF4A undergoes a series of ligand-dependent conformational changes as it binds its substrates (RNA and ATP), hydrolyzes ATP, and releases products (36). The effect of changing oligonucleotide length, backbone composition, and identity of the 2Ј-substituent of the ribose moiety has been investigated and shown to alter the nucleic acid binding and ATPase activities of eIF4A (37).
Previously, we have demonstrated that eIF4A functions alone as an ATP-dependent RNA helicase (8). Results indicated that the initial rate of eIF4A-dependent unwinding decreased as the stability of the duplex is increased and that eIF4A helicase activity is nonprocessive (8). A simple model of unwinding was proposed suggesting that eIF4A affects strand separation by unwinding a small number of base pairs, consequently making the duplex behave as a shorter, less stable duplex in a thermodynamic sense. If there is sufficient destabilization of the duplex by eIF4A, complete stand separation occurs (less stable duplex substrates). Conversely, if the duplex is not sufficiently destabilized by eIF4A, the partially unwound strands reanneal (i.e. snap back), and complete strand separation does not occur (more stable duplex substrates) (8). In addition, the helicase activity of eIF4A may be stimulated by either eIF4B or eIF4H or if eIF4A exists as part of the eIF4F complex (4,8,38). However, it has not been determined if there is a minimal size of the single-stranded region adjacent to the duplex region of the substrate needed to promote helicase activity or if eIF4A alone may function bidirectionally. Furthermore, the effect of changing the 2Ј-substituents of the ribose sugars or the phosphate backbone of the duplex on the helicase activity of eIF4A is unknown. In a continuing effort to understand the mechanism by which eIF4A and other DEAD/DEXH proteins unwind nucleic acid duplexes, an investigation of the helicase activity of eIF4A with respect to duplex substrate specificity was performed.

Methods
Purification of eIF4A from Rabbit Reticulocyte Lysate-Purification of eIF4A follows the standard procedure used to purify protein translation initiation factors that have been previously published by this laboratory (39,40).
DNA and RNA Oligonucleotides-DNA oligonucleotides synthesized by the Molecular Biology Core Laboratory, Case Western Reserve University, were oligonucleotide purification cartridge-purified and stored in distilled H 2 O. DNA oligonucleotides synthesized by IDT, Inc., were polyacrylamide gel electrophoresis-purified, lyophilized, and resuspended in distilled H 2 O. RNA oligonucleotides synthesized by Cybersyn were polyacrylamide gel electrophoresis-purified, lyophilized, and resuspended in distilled H 2 O. RNA oligonucleotides synthesized by Dharmacon were deprotected per the manufacturer's instructions, lyophilized, and resuspended in distilled H 2 O. Quantitation of each oligonucleotide was performed by UV spectroscopy, and a value of 33 g per 1 A 260 was used in determining concentration. Integrity and proper size of each oligonucleotide was assessed by 32 P-end-labeling each oligonucleotide (described below) and analysis on a denaturing (7 M urea) 20% polyacrylamide gel with known size standards.
Transcription of RNA-R-44, -41SD, -41S-, -41Sϩ, and -44Lϩ (see Table I) were synthesized by in vitro transcription using T7 RNA polymerase. The templates for transcription were composed of PTT-23 annealed to PTB-44, -41SD, -41S-, -41Sϩ, or -44Lϩ (see Table I). Transcription reactions were performed using Ambion's Megashortscript TM transcription kit per the manufacturer's instructions and included 1 l of [␣-32 P]CTP (specific activity 3000 Ci/mmol, 10 mCi/ml) as a tracer label. Products of the reactions were purified on a denaturing (7 M urea) 14% polyacrylamide (19:1 acrylamide:bisacrylamide) gel. The proper bands were located by autoradiography, excised from the gel, and the RNA eluted from the gel slices in 400 l of 0.5 M sodium acetate, pH 5.2, 1 mM EDTA, and 3% phenol:chloroform (v/v) for 3-4 h at 4°C. The gel slices were removed from the solution by centrifugation through quick-sep columns, and the recovered RNAs were phenol:chloroformextracted once and ether-extracted twice, followed by ethanol (1 ml) precipitation for 12-24 h at Ϫ20°C. The precipitated RNAs were recovered by centrifugation at 14,000 rpm at 4°C for 60 min. The supernatant was removed, and the RNA pellets washed in ice-cold 70% ethanol, dried, and resuspended in distilled H 2 O. Quantitation of each purified transcript was performed by UV spectroscopy as described above.  Table I) and 10 pmol of [␥-32 P]ATP (specific activity 3000 Ci/mmol, 10 mCi/ml) were combined with 10 units of T4 polynucleotide kinase (10 l final volume), and reactions were performed as described (41). Percent incorporation of the label into these oligonucleotides is 90 -95%, and specific activities of the oligonucleotides are on the order of 1 ϫ 10 6 cpm/pmol. Forty picomoles of DNA-PS, 2Ј-MOE, or 2Ј-MOE/PS oligonucleotides (see Table I) and 100 pmol of [␥-32 P]ATP (specific activity 6000 Ci/mmol, 150 mCi/ml) were combined with 10 units of T4 polynucleotide kinase (10 l final volume) and reactions were performed as described above. Percent incorporation of the label into these oligonucleotides is 1-5%, and specific activities of the oligonucleotides are on the order of 5ϫ10 5 cpm/pmol.
Helicase Substrates-Duplexes used in the helicase reactions are made by combining long oligonucleotides with appropriate complementary 32 P-labeled short oligonucleotides (see Table I for individual oligonucleotide lengths and sequences and figures for duplexes used in specific experiments) in a 1.25:1 ratio, respectively. This excess of long strand is to ensure that essentially all of the labeled short oligonucleotide is in the duplex species. The complementary strands are combined in 1ϫ hybridization buffer (1ϫ Tris-EDTA (TE) plus 100 mM KCl), and the concentration of duplex was 0.5 pmol/l. Samples were heated to 95°C for 5 min and then slow-cooled to 4°C over 90 min (0.1°C/5 s) using a programmable thermocycling instrument. Under these conditions, ϳ90 -95% of the labeled bottom (short) strand hybridizes to the top (long) strand. The duplex was stored at Ϫ20°C and diluted in 1ϫ hybridization buffer for use in the helicase assay.
dsRNA and dsDNA Competitor Duplexes-The R-17-3Ј/R-17C and D-20/D-20C duplexes used in the competition assay ( Fig. 4) were prepared as above except the R-17-3Ј and R-17C or D-20 and D-20C oligonucleotides were combined in a 1:1 ratio at a concentration of 105 or 89 pmol/l, respectively. Control experiments showed that no significant strand separation occurred at the lowest final concentration of duplex used in the competition experiment (10 nM) over the time course of the reaction (data not shown).
Helicase Assay-The helicase assay was performed as described in detail previously (8). Unwinding of duplex substrates was monitored by following displacement of the radiolabeled short strand from the duplex. In general, 20-l reactions contained 20 mM HEPES-KOH, pH 7.5 (final pH 7.2), 70 mM KCl, 2 mM dithiothreitol, 1 mg/ml BSA, and 1 mM magnesium acetate (Buffer A). Buffer BЈ contained 20 mM MES-KOH, pH 6.0, 15 mM KCl, 15 mM potassium acetate, 2 mM dithiothreitol, 1 mg/ml BSA, and 1 mM magnesium acetate. The concentration of ATP was 1 mM; duplex concentration was 1.8 -1.9 nM (36 -38 fmol), and eIF4A concentration was 0.8 M (16 pmol), unless stated otherwise. All reactions were mixed in siliconized tubes to minimize material sticking to tube walls and kept on ice throughout the assembly process. Reactions were incubated at 35°C (unless otherwise stated) for the duration of the time course. In most circumstances, eIF4A was the last component added to the reaction, just before incubation at 35°C; however, equivalent results were obtained when duplex or competitor was added as the last component. Reactions were terminated by adding 5 l of a solution containing 50% glycerol, 2% SDS, 20 mM EDTA, and 0.05% bromphenol blue and xylene cyanol dyes. The products of unwinding reactions were analyzed by separation of displaced labeled short strand from duplex species by electrophoresis on 15 or 18% native polyacrylamide gels (19:1 acrylamide:bisacrylamide) at 4°C for 2-3 h at 200 V in 1ϫ TBE (TBE, Tris borate/EDTA) buffer. Gels were pre-electrophoresed at 4°C for 30 min. After electrophoresis, gels were scanned directly using an Ambis radioanalytical scanner. Gels were then exposed to Kodak X-Omat AR film at Ϫ80°C for using intensifying screens.
Quantitation of the Helicase Assay-Quantitation of the helicase assay was as described previously (8). In brief, quantitation of counts/ min in duplex and displaced/single-stranded RNA bands was performed using the Ambis software. The degree of unwinding for each reaction was determined by measuring the percent of counts/min in duplex and displaced/single-stranded RNA bands using the following formulas: % duplex ϭ cpm duplex/(cpm duplex ϩ cpm displaced strand) and % displaced strand ϭ cpm displaced strand/(cpm duplex ϩ cpm displaced strand). This method of quantitation was used to account for slightly different yields in total counts/min due to variations in pipetting and gel loading steps. The total yield in counts/min among different reactions did not vary by more than Ϯ10% of the average counts/min value of all the reactions in the given experiment. Furthermore, all reagents and protein preparations used in the assay were judged to be free of RNase and phosphatase contamination. A single reaction without (Ϫ) eIF4A was incubated at 4°C during the course of each experiment, and the amount of duplex in this control reaction (typically 36 -38 fmol) was normalized to 100%, and the amount of (background) single-stranded RNA (typically 2-4 fmol) was normalized to 0%. All other reactions were scaled to these values. For the competition experiments (Fig. 4), the percentage of inhibition is given by 100 Data Treatment-Treatment of data has been described previously (8). All data reported are the average of 3-5 separate experiments. Standard errors were less than Ϯ10% in all experiments. Plots and graphs were constructed using Prism software by GraphPad or Slide-Write 4.0. Duplex stability (⌬G) values for RNA/RNA duplex substrates were calculated for 35°C using the nearest neighbor method of analysis as described by Turner and Sugimoto (42). Duplex stability (⌬G) values for RNA/DNA and DNA/RNA duplex substrates were calculated for 35°C using the nearest neighbor method of analysis as described by Sugimoto et al. (43). Duplex stability (⌬G) values for DNA/DNA duplex substrates were calculated for 35°C using the nearest neighbor method of analysis as described by SantaLucia et al. (44). Changing the nonbridging oxygen atom in the oligonucleotide phosphodiester (PO) backbone to sulfur (phosphorothioate (PS) backbone) decreases the stability of the duplex by 0.3 kcal/mol per modification (45). Changing the 2Јgroup of the ribose sugar to a methoxyethyl (MOE) group increases the stability of the duplex by 0.2 kcal/mol per modification (46 -48).
Initial rates of unwinding duplex substrates by eIF4A were deter- mined by measuring the amount of unwinding in the linear (0 -2 min) portion of the reaction. Linear fits were applied, and the initial rate (in fmol/min) was taken from the slope of this line (data not shown). The amplitude (maximum % duplex unwound) of duplex unwinding was determined by allowing the helicase reaction to proceed to completion (time course experiments indicated that the reaction was complete in 15 min, data not shown and Ref. 8). This value is then used to make comparisons with respect to the ability of eIF4A to unwind various duplex substrates. This method of comparison is valid since a direct linear correlation exists between the ln(amplitude of unwinding by eIF4A) and duplex stability (⌬G value) for duplexes with stabilities between Ϫ17.9 kcal/mol and Ϫ24.7 kcal/mol. (Fig. 1C).

RESULTS
eIF4A Helicase Activity and Duplex Stability-Previous studies from this laboratory demonstrated the initial rate of duplex unwinding by eIF4A decreased as the length and stability of the duplex (⌬G value) increased (8). Furthermore, it was observed that a linear correlation exists between the ln(initial rate of unwinding) and ⌬G value of the duplex. It was demonstrated that eIF4A was able to efficiently unwind an 11-base pair (bp) duplex having a stability of Ϫ17.9 kcal/mol (8); however, former studies showed that eIF4A was unable to unwind a 10-bp duplex with a stability of approximately Ϫ25.0 kcal/mol (38). Taken together, this suggests that the amount of unwinding by eIF4A is dependent only on the stability, rather than the length, of the duplex substrate. Thus, we wished to test more rigorously the effects of duplex length and stability on eIF4A helicase activity. In addition, it was previously observed that the maximum percentage of duplex unwound by eIF4A also decreased as the stability of the duplex increased (8). Therefore, we also further investigated the relationship between the maximum percentage of duplex unwinding (amplitude) by eIF4A and duplex stability.
To examine the effect of duplex stability independent of duplex length on the unwinding activity of eIF4A, three 13-bp duplexes of different stabilities (⌬G ϭ Ϫ17.9, Ϫ20.6, and Ϫ23.3 kcal/mol) were designed and tested in the helicase assay. Fig.  1A shows typical experimental data for the unwinding of RNA duplexes by eIF4A. It is apparent from Fig. 1A that the degree of unwinding by eIF4A decreases as the stability of the duplex FIG. 1. The degree of unwinding duplexes by eIF4A is dependent on the stability (rather than the length) of the duplex for short RNA duplex substrates. A, autoradiograms illustrating the degree of unwinding three 13-bp duplexes of differing stabilities and one 16-bp duplex by eIF4A (R-41S-/R-13S-, ⌬G ϭ Ϫ17.9 kcal/mol; R-41SD/R-13SD, ⌬G ϭ Ϫ20.6 kcal/mol; R-41Sϩ/R-13Sϩ, ⌬G ϭ Ϫ23.3 kcal/mol; R-44Lϩ/R-16Lϩ, ⌬G ϭ Ϫ20.5 kcal/mol). Lane designated as Ϫprotein was incubated at 4°C. ⌬ denotes that the duplex was heated to 95°C for 5 min. Lane designated as Ϫprotein, 35°C demonstrates that minimal duplex unwinding (thermal melting) occurs in the absence of eIF4A (15 min). Note that the 1st three lanes (Ϫprotein; ⌬; Ϫprotein, 35°C) show representative controls for the 13-bp, Ϫ17.9 kcal/mol duplex (R-41S-/R-13S-); controls for the other duplex substrates used showed similar results. Lanes designated as ϩeIF4A indicate the duplex was incubated with 0.8 M eIF4A for 15 min (representing maximal unwinding) with the indicated duplex substrate. B, plot illustrating the linear relationship between ln(initial rate of unwinding) (q) or ln(initial rate of thermal melting) (E) and duplex stability for the R-44/R-10, -11, -12, -13, -14, and -15 duplexes. C, plot illustrating the linear relationship between the ln(amplitude of unwinding) versus duplex stability for the R-44/R-11, -12, -13, and -14 duplexes (q) and the R-41S-/R-13S-, R-41SD/R-13SD, R-41Sϩ/R-13Sϩ, and R-44Lϩ/R-16Lϩ duplexes (Ⅺ). The ln(amplitude of thermal melting) versus duplex stability for the R-44/R-10, -11, -12, and -13 duplexes is designated by (E). Note that the values of the slopes in B and C are equivalent, suggesting that the relationship between the ln(amplitude of unwinding) and duplex stability is similar to the relationship between the ln(initial rate of unwinding) and duplex stability. Error bars are omitted for clarity; the standard error is less than Ϯ5% for all data points in this series of experiments. is increased. In addition, a 16-bp duplex (⌬G ϭ Ϫ20.5 kcal/mol) was unwound to approximately the same degree as the 13-bp duplex of the same stability (⌬G ϭ Ϫ20.6 kcal/mol). These results show that the amplitude of duplex unwinding decreases as the stability, but not the length, of the duplex is increased. Furthermore, it shows that two duplexes of different length but of equal stability are unwound by eIF4A with similar efficiencies. These results indicate that duplex stability, rather than length, is the primary determinant for how efficiently eIF4A is able to unwind an RNA duplex.
Previous studies demonstrated that the ln(initial rate of unwinding) by eIF4A or the ln(initial rate of thermal melting) 2 shows a linear correlation with duplex stability and suggests a model for how eIF4A unwinds RNA duplexes (see Introduction and Ref. 8). Due to slight differences in the length of the long strand of the duplex substrates used in this study when compared with our previous work (R-44, 44 nt versus R-1, 50 nt, respectively), the initial rates of eIF4A-dependent unwinding and thermal melting were measured for the R-44/R-10, -11, -12, -13, -14, and -15 duplexes (data not shown). The ln(initial rate of unwinding or thermal melting) was plotted versus duplex stability (Fig. 1B), and the results are equivalent to those obtained previously (8).
To investigate a potential relationship between the maximum percentage (amplitude) of eIF4A-dependent unwinding and duplex stability, the amplitudes of eIF4A-dependent unwinding and thermal melting were measured for the R-44/R-10, -11, -12, -13, -14, and -15 duplexes (Table II and data not shown). The ln(amplitude of unwinding) and the ln(amplitude of thermal melting) for these substrates were plotted versus duplex stability (Fig. 1C, closed circles and open circles, respectively). In addition, the ln(amplitude of unwinding) for the duplexes substrates shown in Fig. 1A was also plotted versus duplex stability (Fig. 1C, open squares). It can be seen from Fig.  1C that a linear correlation exists between the ln(amplitude of unwinding or thermal melting) and duplex stability. This relationship is very similar to that shown in Fig. 1B, and it is stressed that the values of the slopes are equivalent in both Fig. 1, B and C (Ϸ0.24). It should be noted that although the length and nucleotide sequences of the duplexes tested in Fig.  1A are different from the R-44 series of duplexes tested (Table  I), the ln(amplitude of eIF4A-dependent unwinding) for the two sets of points may be fitted to a single line.
Several conclusions may be drawn from the data presented in Fig. 1. First, when taken together, the data presented in Fig.  1, A and C, demonstrate that the overall stability of the duplex, rather than the length or nucleotide sequence, determines how efficiently eIF4A is able to unwind that duplex. Second, comparison of Fig. 1, B with C, indicates that the relationship between the amplitude of unwinding and duplex stability is similar to that observed between the initial rate of unwinding and duplex stability. Since the amplitude of unwinding may be linearly correlated with duplex stability, measuring the amplitude provides an efficient means for comparing how effectively eIF4A is able to unwind duplexes with different stabilities and/or physical and chemical modifications. This is desirable since the amplitude of unwinding provides a more accurate measure of eIF4A unwinding activity due to a higher signal to noise ratio when compared with measurement of initial rates. It is stressed that the amplitude measurements presented here are used only to compare the overall macroscopic degree of unwinding various duplex substrates by eIF4A and are not used to provide information concerning the microscopic kinetic steps of the unwinding reaction.
Effect of the Length and Directionality of the Single-stranded Region on eIF4A Helicase Activity-Previously, it was demonstrated (37) that the RNA-stimulated ATPase activity of eIF4A was dependent on the length the ssRNA activator. These results were consistent with former studies (49, 50) and together suggest a binding site size of ϳ15 nucleotides (nt) for eIF4A. Therefore, it was 2 Thermal melting is defined as any strand separation of duplex nucleic acid that occurs as a passive thermal process; the term unwinding is reserved for strand separation of duplex nucleic acid that occurs in an eIF4A-dependent manner.  of interest to investigate if the helicase activity of eIF4A is also affected by length of the single-stranded region adjacent to the double-stranded region in a duplex substrate. In addition, previous results demonstrated that a combination of either eIF4A ϩ eIF4B or eIF4F (eIF4A⅐eIF4E⅐eIF4G) ϩ eIF4B was able to unwind RNA duplexes in both 5Ј 3 3Ј and 3Ј 3 5Ј directions (38). Thus, it is unknown if this bidirectional activity is a function inherent to eIF4A or if eIF4B is necessary for bidirectional unwinding.
To address how the helicase activity of eIF4A is affected by the length and directionality of the single-stranded region adjacent to the duplex region, a series of 13-bp duplex substrates with either 5Ј-or 3Ј-single-stranded regions ranging from 0 to 25 nt was designed ( Fig. 2A). The sequences of both the singlestranded and duplex regions are exactly the same for the 5Ј 3 3Ј and 3Ј 3 5Ј substrates (Table II). Presented in Fig. 2B are experimental data of eIF4A unwinding duplexes with 0-(bluntended) and 25-nt 5Ј-single-stranded regions, and the complete set of data is summarized graphically in Fig. 2C. These data indicate that eIF4A alone is capable of bidirectional unwinding activity, and the degree of unwinding is approximately equal in both directions (i.e. the lines in Fig. 2C are similar), although eIF4A appears to be slightly more active in unwinding in a 5Ј 3 3Ј direction at shorter (0 -8 nt) single-stranded region lengths. It can also be seen from Fig. 2C that there is little, if any, change in the degree of unwinding as the length of the singlestranded region is increased beyond 8 nt for either 5Ј 3 3Ј or 3Ј 3 5Ј unwinding. In addition, eIF4A unwinding activity decreases only slightly as the length of single-stranded region is shortened below 8 nt. Most surprising was the observation that eIF4A is capable of unwinding a duplex with neither 5Ј-nor 3Ј-single-stranded regions (blunt-ended) with ϳ70% of the activity observed for duplexes with at least 8-nt single-stranded regions.
eIF4A Unwinding of Blunt-ended Duplexes and Interaction with dsRNA-Since the above result demonstrating that eIF4A could unwind a blunt-ended RNA duplex was unexpected, additional experiments designed to explore this observation further were performed. To determine whether significantly different amounts of eIF4A were required for the unwinding of blunt-ended duplexes as compared with duplexes containing a single-stranded region, the effect of eIF4A concentration on unwinding activity was investigated. Fig. 3 shows the effect of increasing eIF4A concentration using the R-38-5Ј/R-13 (25-nt single-stranded region) and R-13/R-13C (blunt-ended) duplexes. The plot shows that the amplitude of unwinding for both duplexes is linear with respect to the concentration of eIF4A (0 -1.6 M) and, further, that the amplitude of unwinding the blunt-ended duplex is consistently 70% of the amplitude of unwinding a duplex with a 5Ј single-stranded region of 25 nt. Furthermore, the initial rate of unwinding the blunt-ended duplex was also ϳ70% of the initial rate of unwinding the duplex containing a 5Ј single-stranded region of 25 nt (0.8 M eIF4A, data not shown). Thus, the efficiency of unwinding a blunt-ended duplex by eIF4A is only minimally decreased (Ϸ30%) when compared with the efficiency of unwinding a duplex substrate containing a single-stranded region adjacent to the duplex region.
The results demonstrating that eIF4A can unwind a bluntended duplex indicates that eIF4A may interact directly with dsRNA. To investigate this possibility further, competitive inhibition experiments using blunt-ended dsRNA were performed. If eIF4A interacts with a nonlabeled dsRNA, then increasing concentrations of this species should abrogate the unwinding of a labeled duplex substrate due to competitive inhibition. The unwinding assay was performed using the R-41S-/R-13S-duplex as a substrate, and a 17-bp blunt-ended dsRNA competitor (R-17-3Ј/R-17C) was titrated into the reaction. It should be noted that the dsRNA competitor's sequence is noncomplementary to the duplex substrate, and control experiments showed the dsRNA competitor was too stable to be unwound efficiently by eIF4A under the reaction conditions used (data not shown). Since eIF4A binds to ssRNA, the competition experiment was also performed with a 17-nt ssRNA (RNA-17C) to serve as a control (data not shown and Fig. 3B). Fig. 3A presents the experimental data for adding increasing concentrations of the dsRNA competitor into the unwinding reaction. Fig. 3B summarizes the results obtained from this series of experiments using the ssRNA and dsRNA competitors. As expected, ssRNA served as a competitive inhibitor to the helicase activity of eIF4A, with half-maximal inhibition of helicase activity observed at 0.05 M ssRNA. The blunt-ended dsRNA duplex also acted as an inhibitor, although it was less effective when compared with the ssRNA in competing for eIF4A, with half-maximal inhibition of helicase activity observed at 0.11 M dsRNA (Fig. 3B). This result implies that eIF4A binds directly to dsRNA. The ability of ssDNA and dsDNA to inhibit the unwinding reaction was also investigated. This will be addressed in more detail below.
Effect 2Ј-Ribose Group and Backbone Modifications on the Duplex Unwinding Activity of eIF4A-Although DEAD/DEXH proteins are classified as RNA helicases, several members of this family of proteins are also able to unwind RNA/DNA, DNA/RNA, and/or DNA/DNA duplexes 3 (16,18,20,51,52). Also, Peck and Herschlag (37) demonstrated the following. (i) eIF4A is able to bind DNA with a similar affinity as RNA (although DNA was unable to stimulate the ATPase activity of eIF4A). (ii) Replacing the nonbridging oxygen atoms in phosphodiester (PO) linkages with sulfur atoms to make phosphorothioate (PS) linkages had no significant effect on the maximal stimulation of the ATPase activity of eIF4A; however, there was a substantial increase in the affinity of eIF4A for the modified oligonucleotide. (iii) Changing the 2Ј-OH substituent of the ribose sugar to 2Ј-OCH 3 reduced the ATPase stimulation at least 100-fold without affecting the oligonucleotide affinity. Thus, to investigate further the substrate specificity and mechanism of the helicase activity of eIF4A, experiments were performed in which the 2Ј-group of the ribose sugar and/or the phosphate backbone of the duplex substrate was chemically modified.
First, the ability of eIF4A to unwind RNA/DNA, DNA/RNA, and DNA/DNA duplex substrates was investigated. Several RNA/RNA duplexes were compared with RNA/DNA (the single-stranded region is RNA), DNA/RNA (the single-stranded region is DNA), and DNA/DNA duplexes of similar sequence, length, and stability in the eIF4A-dependent unwinding assay. Presented in Fig. 5A is selected data of eIF4A unwinding activity using RNA/RNA, RNA/DNA, DNA/RNA, and DNA/DNA duplex substrates. Table II lists the entire data set, including the type of duplex, number of base pairs, calculated stability, and the amplitude of unwinding by eIF4A. All substrates had a 5Ј-single-stranded region of at least 26 nt. As expected, eIF4A unwound the RNA/RNA duplexes with the amplitude of unwinding correlating to the stability of the duplex, and these results were equivalent to those obtained previously (8). As evident from Fig. 5A and Table II, eIF4A was also able to unwind RNA/DNA and DNA/RNA duplexes, and the magnitude of unwinding was also dependent on the stability of the duplex (i.e. as the stability of the duplex increased, the amplitude of unwinding by eIF4A decreased). Furthermore, different types of duplexes with similar stabilities were unwound with similar efficiencies (e.g. compare R-44/R-13 with R-44/D-16 and compare R-44/R-12 with D-44/R-17). In contrast, eIF4A was unable to unwind either DNA/DNA duplex tested. These observations suggest eIF4A may interact with both RNA and DNA, since both RNA/DNA and DNA/RNA duplexes are unwound. However, since DNA/DNA duplexes are not unwound, this indicates that eIF4A requires an RNA strand for helicase activity.
Next, we investigated the ability of eIF4A to unwind a duplex substrate in which the long strand was RNA, but the short strand had either a phosphorothioate (PS) backbone and/or a 2Ј-methoxyethyl (MOE) modification. Fig. 5B presents a diagram of the backbone (PO 3 PS) and 2Ј-substituent (2Ј-H 3 2Ј-MOE) modifications used. Note that changing PO 3 PS decreases the stability of the duplex by 0.3 kcal/mol per modification, and substituting MOE for the 2Ј-OH increases the stability of the duplex by 0.2 kcal/mol per modification (45)(46)(47)(48). The unwinding assay was performed with the modified substrates, and the RNA/RNA and RNA/DNA duplexes were used in parallel reactions for comparison. Fig. 5A presents selected raw data for unwinding RNA/DNA-PS, RNA/2Ј-MOE, and RNA/2Ј-MOE/PS duplexes, and the amplitudes of unwinding for these modified duplexes by eIF4A are listed in Table II. The PS modification appeared to have no effect on the helicase activity of eIF4A, as the magnitude of unwinding the R-44/D-PS-16 duplex was in the expected range for its stability (compare with the stabilities and amplitudes of unwinding the R-44/R-11 and R-44/R-12 duplexes, Table II). In contrast, the amplitudes of unwinding the duplex substrates containing either 2Ј-MOE or 2Ј-MOE/PS modifications were much less than predicted by their relative stabilities (compare R-44/2Ј-MOE-11, ⌬G Ϸ Ϫ16.7 kcal/mol, amplitude Ϸ10% with R-44/R-11, ⌬G Ϸ Ϫ17.9 kcal/mol, amplitude Ϸ70%, etc.) In fact, all of the duplexes with 2Ј-MOE or 2Ј-MOE/PS modifications were unwound to Ϸ10% or less ( Fig. 5A and Table II). Taken together, these results demonstrate that PO 3 PS modifications of the DNA strand in an RNA/DNA duplex have little or no effect on unwinding activity, whereas a 2Ј-H 3 2Ј-MOE modification of one strand in a duplex abrogates unwinding by eIF4A.
To investigate further potential interactions of eIF4A with ssDNA, dsDNA, ssDNA-PS, or 2Ј-MOE, competition experiments similar to those presented in Fig. 4 were performed. The ability of ssDNA and dsDNA to inhibit the unwinding reaction was tested, and in contrast to ssRNA and dsRNA, neither ssDNA nor dsDNA caused inhibition of eIF4A unwinding activity when titrated into the reaction (data not shown and Fig.  4B). Similarly, an ssDNA-PS oligonucleotide was not an effec-tive inhibitor of unwinding (Ͻ15%, 1.0 M ssDNA-PS) when tested in the competitive inhibition assay (data not shown). To investigate the possibility that eIF4A was able to bind, but not unwind an RNA/2Ј-MOE duplex, both a single-stranded 2Ј-MOE and a blunt-ended RNA/2Ј-MOE duplex (R-13C/2Ј-MOE-13) were tested in the competition assay. Results indicated that neither was effective at inhibiting the helicase reaction (up to 1.0 M was tested, data not shown). Taken together, these results demonstrate that under the assay conditions used, the relative affinity of eIF4A for ssDNA, dsDNA, single-stranded 2Ј-MOE, or RNA/2Ј-MOE duplexes is much weaker than for ssRNA or dsRNA.
That single-stranded DNA did not act as a competitive inhibitor was surprising, since data had shown that eIF4A does bind to DNA when measured via the RNA-dependent ATPase activity of eIF4A (37). Therefore, this inconsistency was investigated further. Comparing reaction conditions between those used here for the unwinding activity and those used by Peck and Herschlag (37) for the RNA-dependent ATPase activity showed major differences in the pH, ion type, and ionic strength. For the helicase reaction, the typical buffer contains 20 mM HEPES-KOH, pH 7.2, and 70 mM potassium chloride (Buffer A), whereas for the ATPase reaction, the typical buffer contains 25 mM MES-KOH, pH 6.0, and 10 -30 mM potassium acetate (Buffer B). Therefore, similar competitive inhibition studies to those above were performed under Buffer B conditions. It should be noted that due to the buffer in which reticulocyte eIF4A is purified and stored, it was possible to only decrease the total potassium chloride concentration in the reaction to 15 mM, and thus a final concentration of 15 mM potassium acetate was used to give a combined monovalent ionic strength of 30 mM (Buffer BЈ). Under these Buffer BЈ conditions, the unwinding activity of eIF4A in the absence of any potential inhibitor was ϳ80% of the unwinding activity observed under Buffer A conditions (data not shown). When the ssRNA or ssDNA competitor was added to a final concentration of 1.0 M, the unwinding activity of eIF4A was reduced to ϳ25 and 70%, respectively, of the total unwinding activity observed under Buffer BЈ conditions (data not shown). Thus, limited interaction of eIF4A with ssDNA is observed when the pH and ionic strength of the buffer are altered from conditions A to conditions BЈ. These results indicate that altering the buffer conditions can result in differences in the ability of eIF4A to interact with single-stranded and double-stranded nucleic acids, as monitored by inhibition of the unwinding reaction.

DISCUSSION
Substrate Specificity of eIF4A-Previous work (8) demonstrated that the initial rate and amplitude of duplex unwinding by eIF4A was a function of duplex stability. This relationship has been more rigorously investigated by using a series of RNA duplex substrates in which the effects of duplex length and duplex stability on the degree of unwinding have been tested independently of each other. This study confirms that the degree of duplex unwinding by eIF4A is dependent only on the overall stability (⌬G value) of the duplex and is independent of duplex length or nucleotide sequence (Fig. 1, A and C). In addition, current results demonstrate that like the initial rate, the ln(amplitude of eIF4A-dependent unwinding or thermal melting) is also linearly correlated with the stability of the duplex (Fig. 1, B and C), and the values of these slopes are equivalent to those of the ln(initial rate of unwinding or thermal melting) versus duplex stability. This indicates that the amplitude of unwinding may also be used to compare the efficiency of eIF4A-dependent unwinding among different types of duplex substrates. It is stressed that the unwinding assay used here allows the measurement of only fully unwound duplexes. Thus, only the overall macroscopic rate (or amplitude) of the unwinding process may be observed, and the many potential microscopic kinetic steps of the unwinding reaction (including but not limited to: ATP and duplex binding, ATP hydrolysis, duplex unwinding, release of products, duplex reannealing, and "snap-back") are unable to be characterized in detail. Although limited, this assay does allow for the comparison of eIF4A helicase activity with respect to differences in duplex stability, chemical composition, or physical characteristics and, combined with previous RNA binding and ATP hydrolysis data, allows simple models of the duplex unwinding mechanism to be proposed (see below).
Investigation of the dependence of unwinding on the direction and length of the single-stranded region of duplex substrates indicated that the efficiency of eIF4A helicase activity is approximately equal for the 5Ј 3 3Ј and 3Ј 3 5Јdirections and that bidirectional unwinding is a property of eIF4A and is not dependent on the presence of other initiation factors (Fig. 2). This bidirectional helicase activity supports the idea that eIF4A is involved in melting short RNA duplex structures in mRNA, rather than acting as a processive, directional helicase. Furthermore, the unwinding activity of eIF4A is minimally dependent on the length of the single-stranded region adjacent to the duplex region of the substrate. Most interesting is the fact that eIF4A is able to unwind a blunt-ended duplex with only a modest (30%) decrease in activity when compared with duplexes with 4 -25-nt single-stranded regions (Figs. 2 and 3). This suggests that a single-stranded region preceding the duplex is not a requirement for eIF4A-dependent duplex unwinding and that eIF4A may interact directly with dsRNA. This is supported by the competition experiments showing that like ssRNA, dsRNA acts as an inhibitor of eIF4A helicase activity (Fig. 4). Taken together, this suggests that eIF4A also binds to dsRNA but with a decrease in relative affinity when compared with ssRNA. The ability of eIF4A to unwind blunt-ended duplexes in vitro may indicate its being able to disentangle highly structured and/or sequestered duplex region in the 5Ј-untranslated region of mRNAs.
Although the data support the idea that eIF4A unwinds a blunt-ended duplex by directly interacting with the doublestranded region, it cannot be explicitly ruled out that eIF4A is recognizing short single-stranded regions at the ends of the duplex due to "breathing" or "fraying" of the terminal base pair(s). However, if breathing was a major feature of the ability of eIF4A to unwind blunt-ended duplexes by presenting short single-stranded regions, then one would expect to observe a much greater increase in the degree of unwinding as the singlestranded region adjacent to the duplex was increased from 0 to 4 nts. Furthermore, the potential for breathing in the 13-bp blunt-ended duplex is decreased as 2 of the 3 bp at both ends of the duplex are G:C. In addition, it has been demonstrated that the site size for optimal binding and ATPase activity by eIF4A is 13-15 nt (37, 49).
Experiments designed to test the effect of chemical modifications of the duplex on helicase activity show that eIF4A is able to unwind RNA/RNA, RNA/DNA, and DNA/RNA but not DNA/DNA duplexes ( Fig. 5 and Table II). In addition, the stability of the duplex, regardless of composition (RNA/RNA, RNA/DNA, or DNA/RNA) or length, dictates how efficiently eIF4A will unwind a particular duplex (Table II). Results indicated that although a change from a PO to a PS backbone had no effect on eIF4A helicase activity, substrates containing a 2Ј-MOE modification abrogated unwinding activity by eIF4A ( Fig. 5 and Table II).
Taken together, the results of this study provide additional detail with respect to how eIF4A interacts with and unwinds duplex substrates. First, eIF4A is not limited to interacting with only ssRNA but also may interact directly with dsRNA. Second, it is evident that an RNA strand (2Ј-OH) is required for eIF4A helicase activity. This agrees with the finding that 2Ј-OH groups are necessary for stimulating eIF4A ATPase activity (37,49). Third, eIF4A may bind the RNA strand in either orientation (5Ј 3 3Ј or 3Ј 3 5Ј) since duplex substrates with 5Ј and 3Ј single-stranded regions are unwound with similar efficiencies by eIF4A. This is further supported by the fact that both RNA/DNA and DNA/RNA duplex are unwound, since the orientation of the RNA strand relative to the DNA strand is opposite for each duplex. Finally, the molecular basis for the inability of eIF4A to unwind an RNA/2Ј-MOE or RNA/2Ј-MOE/PS duplex is likely due to the large 2Ј-MOE group obstructing the binding interface between eIF4A and the duplex.
In support of this hypothesis, it has been shown via NMR structural studies that the MOE group occupies the minor groove of an RNA/2Ј-MOE duplex (53,54) and suggests that the contacts made between eIF4A and double-stranded nucleic acids may involve the minor groove of the duplex. The data presented here suggest that eIF4A may recognize and bind to only one strand (which must be RNA) of doublestranded nucleic acids and that any interaction with the complementary strand is minimal. This is supported by the results that RNA/RNA, RNA/DNA, RNA/DNA-PS, and DNA/ RNA but not DNA/DNA duplexes are unwound by eIF4A and that ssRNA and dsRNA serve as competitive inhibitors of the helicase reaction, whereas ssDNA, ssDNA-PS, and dsDNA do not serve as competitive inhibitors. Thus, the recognition of single-stranded and double-stranded nucleic acids (as helicase substrates) by eIF4A is based on the presence of a 2Ј-OH ribose moiety. The possibility that eIF4A interacts with only one (RNA) strand of a double-stranded nucleic acid is similar to recent observations (55) made with the HCV NS3 helicase using a series of RNA/RNA, RNA-Me/RNA-Me and RNA/ RNA-Me hybrid duplexes. This is also similar to the RecBC helicase, which interacts and translocates along only one strand (the 3Ј 3 5Ј strand) of DNA duplex substrates (56).
Whereas our results suggest that eIF4A may interact with only one strand of a nucleic acid duplex and that this strand must be RNA (2Ј-OH), it cannot be explicitly ruled out that the differences with which eIF4A interacts with dsRNA and dsDNA are based on structural, rather than chemical, features of double-stranded nucleic acids. That is, eIF4A is able to recognize A-form helices (dsRNA), but not B-form helices (dsDNA), and implies that eIF4A would interact with both strands of the duplex substrate. Structural studies of RNA/ DNA (or DNA/RNA) duplexes indicate that while these molecules have an A-form character, the RNA and DNA chains in these hybrid duplexes adopt different helical conformations (57). The RNA strand retains the A-form helical conformation, whereas the DNA strand adopts a conformation that is neither A-form nor B-form but is intermediate in character between these two forms (57). This may explain how eIF4A recognizes and unwinds RNA/DNA and DNA/RNA duplexes. However, whereas the structures of short ssRNA and ssDNA oligonucleotides differ in their degree of rigidity, neither adopts a "standard" structural form like dsRNA and dsDNA. Thus, the ability of eIF4A to interact with (or unwind) ssRNA, dsRNA, RNA/ DNA, RNA/DNA-PS, and DNA/RNA but not ssDNA, ssDNA-PS, or dsDNA may be due to differences in either the chemical composition of the nucleic acids (2Ј-OH versus 2Ј-H) or helical structure (A-versus B-form helices). It is also quite possible that both the chemical composition and physical characteristics of single-stranded and double-stranded nucleic acids play a role in eIF4A substrate recognition.
A Revised Model for Duplex Unwinding by eIF4A-Previous results obtained in this laboratory led to a simple model by which eIF4A unwinds duplexes in an ATP-dependent manner (see Introduction and Ref. 8). The results obtained in this report support the basic tenets of this model and offer more details concerning the mechanism of unwinding by eIF4A. Our previous model suggested that eIF4A must bind to a ssRNA region before translocation and duplex unwinding occurs. However, the results presented in this report suggest that eIF4A is capable of binding to the double-stranded region of a partial or fully duplexed nucleic acid substrate. Therefore, a revised model of how eIF4A initiates unwinding is presented in Fig. 6. This model shows that unwinding may initiate from two possible pathways. In the first pathway (1), eIF4A⅐ATP binds to the single-stranded region of the substrate and translocates to the double-stranded region, where it initiates unwinding and displaces the release strand. Note that this pathway is identical to that proposed previously; FIG. 6. Revised model of eIF4A-dependent RNA duplex unwinding. Schematic diagram illustrating possible pathways by which eIF4A unwinds RNA duplex substrates. In this model, eIF4A is depicted as a two domains (N-terminal, trapezoid; C-terminal, oval) connected by a distended polypeptide linker (curved line) in accordance with the structural results obtained by Caruthers and McKay (33). Both pathways begin with eIF4A binding to ATP (q). Pathway 1 shows eIF4A binding the ssRNA region of a duplex substrate. In this pathway, eIF4A must translocate to the dsRNA region, which may require ATP hydrolysis and nucleotide exchange. eIF4A, now positioned at the duplex region (equal to n base pairs), hydrolyzes ATP to ADP (E) and unwinding initiates at the ss/ds junction. A small number of base pairs (e.g. 4) are unwound, and if enough destabilization occurs, the strands separate from each other. This is followed by complex (eIF4A⅐ADP⅐RNA) dissociation. Pathway 2 shows eIF4A binding to a blunt-ended duplex. Note that eIF4A is positioned directly at the double-stranded region, and no translocation occurs. eIF4A hydrolyzes ATP, and unwinding initiates at either the end of the duplex (pathway 2A) or at an internal region of the duplex (pathway 2B). A small number of base pairs (e.g. 4) are unwound, and if enough destabilization occurs, the strands separate from each other. This is followed by complex (eIF4A⅐ADP⅐RNA) dissociation. (i) The degree of strand separation (as opposed to reannealing or snap-back) is dependent on duplex stability. (ii) It is possible that eIF4A may bind directly to the double-stranded region of a substrate containing both and single-stranded and doublestranded regions and proceed via pathway 2A or 2B. See text for additional details. however, it is unclear if more than one ATP hydrolysis event is needed for translocation and unwinding. Since the binding of eIF4A to ssRNA is weak (K d Ϸ100 M, (35,49,58)), it is possible that eIF4A undergoes several rounds of ATP hydrolysis, dissociation, and reassociation to the substrate before interacting with the double-stranded region of the molecule. Thus, the presence of the single-stranded region may help to build a higher local concentration of eIF4A⅐ATP near the double-stranded region of the substrate, and thus "translocation" is actually a dissociation/reassociation event as opposed to continuous movement along the ssRNA. This may explain why the degree of eIF4A-dependent unwinding is slightly higher with substrates containing an ssRNA region when compared with blunt-ended substrates.
In the second pathway (2), eIF4A⅐ATP binds directly to and initiates unwinding at the duplex region. Note that this pathway eliminates the necessity for a translocation event and that a single ATPase event may be sufficient for strand separation. It is unclear, however, whether eIF4A initiates unwinding at the end of the duplex (Pathway 2A) or at an internal region of the duplex (Pathway 2B). Pathway 2 is likely the only means by which eIF4A could unwind bluntended and DNA/RNA duplexes. Note that in both pathways, the ability of eIF4A to unwind the duplex completely is still related to the overall stability of the substrate, and recognition of the duplex may be due to chemical and/or structural features of the substrate.
eIF4A: Nucleic Acid Binding, ATPase Activity, and Helicase Activity-Comparison of the results presented in this report with previous studies (35)(36)(37)49) show a number of inconsistencies among the nucleic acid binding, ATP hydrolysis, and helicase activities of eIF4A. These differences focus on the chemical and physical properties of various nucleic acids that will (or will not) serve as substrates for binding, stimulating ATP hydrolysis and helicase activity. Many of these inconsistencies may be explained by the fact that the binding ATPase and helicase experiments were performed under markedly different reaction conditions, including buffer composition, protein and substrate concentration, and source of eIF4A (8,(35)(36)(37)49). This is strongly supported by a series of experiments demonstrating that coupling between the binding of ligands (ssRNA and ATP) and the number of conformational states available to eIF4A are markedly different under various salt and pH conditions (36).
These apparent differences between the ATPase and helicase activities of eIF4A imply that the relationship between these two enzymatic activities may be much more complex than originally predicted. In general, it appears that the large stimulation of eIF4A ATPase activity observed in the presence of ssRNA may not be absolutely required for efficient strand separation. This may reflect differences in the microscopic rate constants and/or rate-limiting steps between the ATPase and helicase reactions. A similar disparity between the ATPase and helicase activities of the HCV NS3 helicase has recently been investigated in detail, and it was concluded that the binding of ssRNA by HCV NS3 necessary for maximal stimulation of ATPase activity is not directly related to nucleic acid duplex unwinding (55).
In conclusion, this report has further characterized the helicase activity of eIF4A with respect to substrate specificity and provides additional insight into the mechanism of duplex unwinding by DEAD/DEXH box proteins in general. Further investigation will be required to fully understand the complex relationship between the nucleic acid binding, ATP hydrolysis, and helicase activities of this important class of proteins.