DEAD-Box Proteins from Escherichia coli Exhibit Multiple ATP-Independent Activities

Abstract

DEAD-box proteins (DBPs) are a widespread class of ATP-dependent RNA helicases that play a key role in unwinding RNA duplexes. In recent years, certain DBPs have also been found to exhibit activities that do not require ATP. To gain a better understanding of prokaryotic RNA metabolism, we investigated whether Escherichia coli DBPs harbor any ATP-independent activities. We show that each of the four E. coli DBPs tested in this study can accelerate the association of cRNA molecules, can stimulate strand displacement, and can function as an RNA chaperone without utilizing ATP. To the best of our knowledge, these prokaryotic DBPs constitute the first examples of proteins that harbor each of these three activities. The identification of these auxiliary functions indicates that the E. coli DBPs are versatile factors that possess significant RNA remodeling activity in addition to their canonical RNA helicase activity and might therefore participate in a greater variety of cellular processes than has been previously appreciated.

INTRODUCTION

DEAD-box proteins (DBPs) are an important and ubiquitous class of proteins that are characterized as RNA helicases due to their ability to unwind RNA duplexes (8, 21, 33). These factors belong to the SF2 superfamily of RNA helicases and contain nine characteristic motifs distributed over 350 to 400 amino acids. These amino acid sequences contribute to RNA and ATP binding and to the coupling of ATP hydrolysis with RNA unwinding (12). By virtue of their helicase activity, DBPs help to resolve incorrectly folded RNAs as well as allow RNA sequences to engage in multiple rounds of interactions with complementary sequences. Such molecular reorganization has been found to be important for the regulated folding and biogenesis of many different RNA molecules, including ribosomal and spliceosomal RNAs (14, 26). An absence of DBPs is frequently associated with a variety of phenotypic defects, including cell lethality, which attests to a key role for these proteins in different aspects of cellular RNA metabolism.

Of the several dozen DBPs whose ATP-dependent helicase activity has been biochemically verified, a selected number have also been found to harbor activities that do not require ATP (15, 34, 36, 37). The ATP-independent activities that have been identified for different DBPs include an ability to accelerate the annealing of complementary strands, to catalyze strand displacement, and/or to promote the attainment of functional RNA conformation, referred to as RNA chaperone activity (31). While such functions can render these proteins even more versatile for RNA remodeling, it is unclear whether their occurrence reflects a common characteristic of DBPs or not.

We have been studying DBPs from Escherichia coli with a view to understanding their function in prokaryotic organisms. E. coli possesses five DBPs: DbpA, DeaD (also referred to as CsdA), RhlB, RhlE, and SrmB (17, 24). DbpA, which is considered a classic example of a sequence-specific helicase, binds tightly to stem-loop 92 of 23S rRNA, an interaction that greatly stimulates its ATPase and helicase activity (10). DeaD and SrmB are each implicated as ribosome processing factors. The absence of either protein results in the formation of aberrant ribosomal particles and the accumulation of unprocessed rRNA, particularly at low temperatures (6, 7). RhlB is best known for its interaction with two ribonucleases, RNase E and polynucleotide phosphorylase, which together with the glycolytic enzyme, enolase, constitute an mRNA-degrading complex called the “degradosome” (4, 30). Very little is known about RhlE, although recently it was found to modulate the function of DeaD and SrmB inside the cell (19).

Although all five E. coli DBPs possess ATPase and ATP-dependent RNA helicase activity, there is very little detailed information about their biochemical properties, with the exception of DbpA, which has been characterized in considerable detail (2, 5, 10, 13, 16). In particular, it has been unclear whether the E. coli DBPs, most of which are relatively small, harbor any ATP-independent activities that have been associated with the larger eukaryotic DBPs. With a view to bridging this knowledge gap, we investigated the potential ATP-independent functions of DeaD, RhlB, RhlE, and SrmB. Here, we show that each of these four proteins has a potent RNA-annealing activity, can stimulate the exchange of strands between single-stranded RNA and duplex RNA, and possesses RNA chaperone activity. These observations provide evidence for multifunctionality by these prokaryotic proteins and strongly suggest that their ATP-independent activities contribute to the function of these proteins inside the cell.

MATERIALS AND METHODS

Proteins.

Expression clones for DeaD, RhlB, RhlE, SrmB, and Hfq were derived from the ASKA set of cloned E. coli genes (25). The clones were transformed into a derivative of BL21(DE3) that has a deletion in the gene for RNase I (Δrna::kan) and contains an F′::tet lacI^q episome. The transformed strains were grown to mid-log phase at 37°C, induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and harvested after 2 to 3 h. Proteins were purified using Ni-nitrilotriacetic acid (NTA) beads, as described previously (20). The purified proteins were fractionated by SDS-PAGE alongside known amounts of purified bovine serum albumin (BSA) and quantified by staining and densitometry of the protein bands. Purified Mss116p was generously provided by Alan Lambowitz (University of Texas, Austin, TX).

Oligonucleotides.

The sequences of the RNA oligonucleotides used are as follows: R17, 5′ AAGUGAUGGUGGUGGGG 3′; R17*, 5′ CCCCACCACCAUCACUU 3′; R34*, 5′ CCCCACCACCAUCACUUAAAAAAAAAAAAAAAAA 3′; R12, 5′ AAGUGAUGGUGG 3′; R14, 5′ AAGUGAUGGUGGUG 3′; R19, 5′ GCUUAACCUUACAACGCCG 3′; and R16*, 5′ GGUUGUGAGGUUAAGC 3′. D17 and D34* are DNA oligonucleotides that correspond in sequence to the RNA oligonucleotides R17 and R34*, respectively. Oligonucleotides were synthesized by Sigma-Genosys or obtained from Murray Deutscher's laboratory (University of Miami Miller School of Medicine).

RNA-annealing assays.

Oligonucleotides were 5′-end labeled with ³²P and gel purified. Annealing assay mixtures contained <0.03 nM labeled oligonucleotides (∼500 cpm), various amounts of unlabeled complementary oligonucleotides, and proteins, as indicated. Annealing assays were performed at room temperature in buffer C (50 mM MOPS [morpholinepropanesulfonic acid] [pH 7.0], 0.01% Triton X-100) by the separate addition of each of the annealing oligonucleotides and protein in a total volume of 5 μl. The reactions were allowed to proceed for 1 min and terminated by the addition of 1 μl of unlabeled trapping oligonucleotide that corresponds to the sequence of the labeled oligonucleotide but is present in significant excess. The reaction products were extracted with 6 μl of phenol and centrifuged, and 5 μl of the recovered supernatant was added to 1.2 μl of loading dye (0.3% bromophenol blue, 0.3% xylene cyanol, 50% glycerol). The reaction products were fractionated on a 15% polyacrylamide gel, and the extent of annealed product formation was visualized by phosphorimaging. Control reactions were performed to ensure that the addition of the trapping oligonucleotide effectively prevents further annealing of the labeled oligonucleotide with a complementary oligonucleotide. All assays with the E. coli DBPs were performed at least twice using independent protein preparations.

Strand displacement assays.

Approximately 500 cpm of a complex comprised of gel-purified radiolabeled R12 annealed to unlabeled R34* was incubated at room temperature in a 5-μl volume of buffer C for 1 h with unlabeled R14 and DBPs. The reaction products were extracted with phenol and fractionated on a 15% polyacrylamide gel followed by phosphorimaging. Strand displacement was indicated by the release of R12 from the duplex into free form.

RNA chaperone assays.

H1 and H2 transcripts containing separate halves of the td intron were synthesized by in vitro transcription, as described previously (35). The transcripts were incubated with 1 μCi of [γ-³²P]GTP in 10 μl of splicing buffer (4 mM Tris [pH 7.6], 3 mM MgCl₂, 0.4 mM spermidine, and 4 mM dithiothreitol [DTT]) for 15 min at 37°C or 55°C without any added proteins or at 37°C with proteins present at 150 nM. Reactions were terminated by the addition of 10 μl of stop solution (40 mM EDTA, 300 mg/ml tRNA), and products were electrophoresed on a 5% polyacrylamide 8 M urea gel and analyzed by phosphorimaging.

RESULTS

DeaD, RhlB, RhlE, and SrmB stimulate strand annealing.

Previous studies have indicated that prokaryotic DBPs possess both ATPase and ATP-dependent RNA helicase functions, but whether these proteins also possess any ATP-independent functions has not been determined. To investigate such possibilities, four E. coli DBPs, DeaD, RhlB, RhlE, and SrmB, were purified as hexahistidine-tagged proteins (Fig. 1A). Preliminary assays indicated that these proteins stimulate an ATP-independent annealing of complementary oligonucleotides in a manner that is dependent upon protein concentration and reaction time. To characterize these effects in more detail, a labeled 17-nucleotide (nt) oligonucleotide (R17) and different concentrations of a 34-nt RNA oligonucleotide (R34*) that has 17 nt complementary to R17 were coincubated with either no protein added, with control proteins added, or with each of the four E. coli DBPs, DeaD, RhlB, RhlE, and SrmB. The concentration of the unlabeled oligonucleotide was varied over a range of concentrations that resulted in the formation of different levels of an annealed complex, which was detected by gel electrophoresis. For these assays, a fixed annealing time of 1 min was used, and the annealing reactions were terminated by the addition of a large excess of unlabeled R17. In the absence of any protein, or with a nonspecific protein BSA added, very small amounts of complex formation were observed when subnanomolar concentrations of R34* were used (Fig. 1B). In the presence of each of the four E. coli DBPs, however, very significant amounts of complex formation could be detected even with 0.04 nM R34*, the lowest concentration used (Fig. 1C). In each case, the extent of annealed product formation was similar to that observed using 5 nM R34* when DBPs were not present. When higher concentrations of R34* were used with DBPs present, virtually all of R17 was found complexed with R34*. These results indicate that each of the four E. coli DBPs possesses a significant RNA-RNA-annealing activity. Further examination of the annealing reaction indicated that the extent of annealing was not affected by the addition of 1 mM ATP or ADP (data not shown).

Fig. 1.

E. coli DBPs stimulate the annealing of complementary RNAs. (A) Gel electrophoresis of purified DBPs. Purified His-tagged DeaD, RhlB, RhlE, and SrmB (0.3 μg each) were fractionated by SDS-PAGE and visualized by dye staining. A molecular weight standard (M) was run in parallel. The calculated molecular masses of tagged DeaD, RhlB, RhlE, and SrmB are 73, 49, 52, and 52 kDa, respectively. The molecular masses of the protein bands present in the standard, in kDa, are indicated on the left. (B to D) Annealing reactions were performed with radiolabeled R17, various concentrations of R34*, as indicated, and 50 nM the indicated proteins. The products were separated by gel electrophoresis and visualized by phosphorimaging. F, free R17; A, R17 annealed to R34*. (B) Annealing reactions without any added protein or with a nonspecific protein (BSA) added. (C) Annealing reactions with DeaD, RhlB, RhlE, or SrmB added. (D) Annealing reactions with Mss116p or Hfq added.

To compare the annealing activity of the E. coli DBPs with those of previously characterized proteins, we investigated, in parallel, the annealing of R17 and R34* in the presence of Mss116p, a DBP from Saccharomyces cerevisiae, whose ATP-independent activities have been characterized in considerable detail (15). A second reference protein was Hfq, a small E. coli protein that promotes the annealing of a large number of small regulatory RNAs to their mRNA targets in vivo (27, 28). When Mss116p was used, a high degree of annealing stimulation was observed, similar to that observed with the E. coli DBPs (Fig. 1D). Stimulated annealing was also observed with Hfq, but a high degree of annealing required increased concentrations of R34* compared to the E. coli DBPs. Based on these results, we conclude that each of the four E. coli DBPs is a potent stimulator of RNA-RNA annealing, with a stimulatory activity comparable to Mss116p and exceeding that of Hfq, an E. coli protein that has a well-defined role in promoting the interaction of regulatory RNAs with their targets.

The strand-annealing activities of DeaD, RhlB, RhlE, and SrmB are relatively nonspecific.

To characterize the oligonucleotide sequence and length requirements for annealing stimulation by E. coli DBPs, annealing assays were performed using different oligonucleotide combinations.

First, we investigated whether the high degree of DBP-mediated stimulation observed for the 17- and 34-nt combinations of the oligonucleotides used in Fig. 1 was dependent upon the length of the longer oligonucleotide. For these purposes, a shorter derivative of R34* was used in annealing assays. When a 17-nt oligonucleotide (R17*) fully complementary to R17 was used in place of R34*, substantial amounts of complex formation were observed with each of the four E. coli DBPs and 0.2 nM this oligonucleotide (Fig. 2A). Without any added protein, similar levels of complex formation were observed only when 100-fold-higher levels of R17* were used. Therefore, reducing the length of one of the annealing oligonucleotides from 34 to 17 nt, by removing the noncomplementary portion of R34*, had little effect on the ability of the E. coli DBPs to promote strand annealing.

Fig. 2.

The annealing activity of E. coli DBPs is relatively nonspecific. Annealing reactions were performed using different sets of complementary oligonucleotides with no protein or with 50 nM DeaD, RhlB, RhlE, or SrmB added. The concentrations of the unlabeled oligonucleotides are indicated in each case. F, free (uncomplexed)-labeled RNA; A, annealed RNA-RNA duplex. (A) Annealing reactions between labeled R17 and unlabeled R17*. (B) Annealing reactions between labeled R14 and unlabeled R34*. (C) Annealing reactions between labeled R19 and unlabeled R16*.

Next, we investigated whether a derivative of R17 that is 3 nt shorter (R14) could be efficiently annealed to R34* by the E. coli DBPs. Using radiolabeled R14 and unlabeled R34*, the addition of 0.2 nM R34* was found to be sufficient to promote complex formation when any of the four DBPs was present (Fig. 2B). In contrast, over 100-fold-higher levels of R34* were needed to promote equivalent levels of complex formation without protein. Therefore, oligonucleotides as short as 14 nt can be efficiently annealed by E. coli DBPs.

Finally, to test whether oligonucleotides corresponding to a different sequence can be annealed efficiently by DBPs, a set of oligonucleotides (R19 and R16*) that correspond to the two complementary ends of immature E. coli 23S rRNA were used (9). Annealing assays performed with these oligonucleotides indicated that each of the DBPs promoted annealing at low oligonucleotide concentrations, whereas two-magnitude-higher concentrations of R16* were required to promote similar levels of annealing in their absence (Fig. 2C). Therefore, each of the four E. coli DBPs has a marked effect on annealing that appears to be relatively independent of oligonucleotide length and sequence.

Nucleotide specificity of the strand-annealing activity of E. coli DBPs.

To further characterize the requirements for strand annealing, we investigated whether the E. coli DBPs can stimulate RNA-DNA and DNA-DNA annealing in addition to RNA-RNA annealing. To analyze their effect on RNA-DNA interactions, annealing reactions between radiolabeled R17 and an unlabeled DNA oligonucleotide (D34*), which corresponds to the sequence of R34*, were set up. As a control, annealing reactions using R17 and R34* were performed in parallel. In the absence of any added protein, comparably high levels of either R34* or D34* (10 to 50 nM) were required for any significant degree of annealing to take place (Fig. 3A). However, when any of the four E. coli DBPs was added to reaction mixtures containing 2 nM D34*, a significant amount of complex formation could be observed. Therefore, each of the four E. coli DBPs stimulates annealing between RNA and DNA. However, a comparison with the RNA-RNA-annealing reaction indicated that the extent of R17 annealing with D34* was lower than that observed between R17 and R34*. Therefore, the ability of the E. coli DBPs to promote RNA-DNA annealing, while significant, is less than their ability to promote RNA-RNA annealing.

Fig. 3.

Nucleotide specificity of annealing stimulation by DBPs. (A) Annealing reactions were performed between labeled R17 and unlabeled RNA or DNA oligonucleotide R34* or D34*, respectively. Annealing reaction mixtures without any protein contained the indicated concentrations of R34* or D34*, in nM. Annealing reaction mixtures with the E. coli DBPs (at 50 nM) contained 2 nM R34* or D34*. F, free (uncomplexed) R17; A, R17 annealed to R34* or D34*. (B) Annealing reactions were performed with labeled D17 and unlabeled D34*. The concentration of D34* used, in nM, is indicated. DBPs were used at 50 nM. F, free D17; A, D17 annealed to D34*.

Despite the lower degree of annealing stimulation observed, annealing reactions between RNA and DNA oligonucleotides were sufficiently stimulated to suggest that the E. coli DBPs might also stimulate DNA-DNA annealing. This possibility was tested by incubating cDNA oligonucleotides, D17 and D34*, either in the absence or the presence of the E. coli DBPs (Fig. 3B). Without any protein, high levels of annealing were observed only with 100 nM D34*. In contrast, the addition of each of the four E. coli DBPs allowed annealing to proceed to significant levels at 5 nM D34*. Therefore, each of the E. coli DBPs stimulates the annealing of cDNA molecules as well.

E. coli DBPs possess strand displacement activity.

Another biochemical function that has been associated with certain DBPs is ATP-independent strand displacement or exchange activity, an activity that promotes the release of an RNA strand from a base-paired duplex (32). This activity is distinct from canonical helicase activity, because it requires neither ATP nor magnesium ions, each of which is necessary for helicase activity. To determine whether any of the E. coli DBPs promotes strand displacement, a suitable assay was developed (Fig. 4A). The components of the assay include an RNA duplex composed of a labeled 12-nt oligonucleotide (R12) annealed to R34* and an unlabeled 14-nt single-stranded exchange oligonucleotide (R14) whose sequence overlaps with that of R12. In the event that R12 can be displaced by the more stably binding R14, the release of R12 from the complex can be detected by gel electrophoresis. When these components were incubated without any added protein, no significant release of R12 was observed unless very high levels of R14 (200 nM) were present (Fig. 4B). Similar results were observed when BSA was included in the strand displacement reaction (data not shown). In the presence of each of the DEAD-box proteins, however, a considerable fraction of R12 was released from the duplex with just 5 nM R14 present. Hence, each of the E. coli DBPs also promotes strand displacement reactions.

Fig. 4.

Strand displacement activity of E. coli DBPs. (A) Schematic description of strand displacement assay. An RNA duplex (thick lines) composed of an unlabeled oligonucleotide and a radiolabeled complementary oligonucleotide (indicated by an asterisk) is incubated with a single-stranded oligonucleotide (thin line) similar in sequence to the radiolabeled strand. Displacement of the radiolabeled oligonucleotide results in its release as a single strand. (B) Limiting concentrations of a labeled oligonucleotide duplex (R12/R34*) and an exchange oligonucleotide (R14) were incubated in the absence or presence of E. coli DBPs. Strand displacement reactions were performed using the indicated concentration of R14 and with either no added protein or with 50 nM DeaD, RhlB, RhlE, or SrmB added. Strand displacement was visualized by the release of labeled R12 into free form following gel electrophoresis. A, radiolabeled R12 annealed to R34*; F, free R12.

E. coli DBPs possess RNA chaperone activity.

Having shown that the E. coli DBPs possess ATP-independent strand-annealing and strand displacement activities, we investigated whether these proteins can also function as ATP-independent RNA chaperones. A commonly utilized assay for RNA chaperone activity measures trans-splicing of a split bacteriophage T4 td intron (Fig. 5A) (11). The transcribed RNAs normally adopt a nonproductive conformation in vitro, and efficient splicing, which initiates with the ligation of guanosine to intron RNA, requires either the presence of an RNA chaperone or elevated temperatures to promote the attainment of productive RNA conformations. To determine whether the E. coli DBPs can act as RNA chaperones, the two halves of the td intron and radiolabeled GTP were incubated with no proteins, with BSA, or with each of the four E. coli DBPs. One other protein that was used was Mss116p, which has been shown to harbor RNA chaperone activity using different assays (15). Incubation of the RNAs without any proteins or with BSA resulted in undetectable incorporation of radiolabeled GTP into RNA at 37°C, while incubating the reaction mixtures without protein at 55°C or with Mss116p addition at 37°C resulted in the generation of splicing products, with the major radiolabeled product corresponding to guanosine ligated to the 5′ intron (Fig. 5B). Significantly, the addition of each of the four E. coli DBPs also yielded splicing products at 37°C. Therefore, each of the E. coli DBPs can function as an RNA chaperone, in line with the expectations derived from their other ATP-independent activities.

Fig. 5.

RNA chaperone activity of E. coli DBPs. (A) Schematic description of the RNA chaperone activity assay. In this assay, the pre-mRNA of the td intron is split into two halves (H1 and H2), with each half containing exon (thick lines) and intron sequences (thin lines). The addition of radiolabeled GTP (*G) initiates the splicing reaction, resulting in cleavage at the exon-intron boundary of H1 and ligation of the labeled nucleotide to intron sequences. The in vitro trans-splicing reaction is inefficient at 37°C due to the inability of H1 and H2 to adopt a productive conformation. The addition of an RNA chaperone protein can restore productive conformation to the RNAs, resulting in the generation of splicing products. (B) Testing DBPs for RNA chaperone activity. Radiolabeled GTP and in vitro-transcribed RNAs containing the halves of the td intron were incubated at 37°C or 55°C without protein or at 37°C with BSA, with each of the four E. coli DBPs or with Mss116p added. The reactions were analyzed by gel electrophoresis and phosphorimaging. The primary GTP-ligated radiolabeled product is indicated by an arrow.

DISCUSSION

DBPs have long been considered to be an important class of RNA remodeling factors based on their ATP-dependent RNA helicase activity. In recent years, several of these proteins have also been found to be capable of remodeling RNA structure without an ATP requirement. The ATP-independent activities that have been identified for different DBPs include the stimulation of strand-annealing, strand displacement, and RNA chaperone activity. Since these activities render DBPs even more versatile for performing molecular rearrangements, an assessment of these activities is crucial to gain an understanding of DBP function in the cell.

In this work, we show that each of the four E. coli DBPs analyzed here has a potent RNA-annealing activity. This activity appears to be relatively nonspecific, requires only a short stretch of single-stranded pairing region, and promotes annealing to DNA in addition, albeit with lower efficiency. Thus, apart from the molecular rearrangements promoted due to their canonical helicase activity, these prokaryotic proteins are potentially capable of promoting the annealing of complementary RNAs inside the cell, as well as RNA-DNA and DNA-DNA interactions. In addition, each of the four DBPs was found to display strand displacement and RNA chaperone activity. To the best of our knowledge, the four E. coli DBPs characterized here represent the first examples of proteins that have been found to possess all three ATP-independent functions: strand exchange, strand displacement, and RNA chaperone activities.

In biochemical assays, each E. coli DBP stimulated RNA-RNA annealing as efficiently as the S. cerevisiae DBP, Mss116p, and more efficiently than Hfq, a reference protein that has an established role as an RNA annealer. Nonetheless, it is interesting that the in vivo defects caused by a lack of these DBPs are generally less severe than those due to the absence of Hfq (3, 18). Under standard laboratory growth conditions, an absence of Hfq causes a variety of pleiotropic defects, whereas an absence of DBPs causes generally mild growth and ribosome assembly defects. While the cellular phenotypes might seem at variance with the expectations derived from biochemical data, it is worth pointing out that Hfq is expressed at a notably high level of 30,000 to 60,000 molecules per cell, one that is likely to be far higher than the currently known expression levels of the E. coli DBPs (1, 23). However, under cold-shock conditions, DeaD becomes highly overexpressed as well, and interestingly its absence then has a profound effect on growth, with growth cessation occurring at temperatures below 20°C (17, 22). Moreover, DeaD also becomes a significant component of the RNA degradosome at low temperatures (29). Therefore, the physiological effects of these RNA regulators appear to be primarily dictated by their expression levels rather than the efficacy with which they promote RNA remodeling reactions in vitro.

Another question of relevance is which DBP amino acids are responsible for their ATP-independent functions. All DBPs contain a core domain of 350 to 400 amino acids encompassing nine conserved motifs that are associated with RNA and ATP binding, with the amino acids responsible for helicase activity included within this core domain. DBPs from eukaryotic organisms usually contain significant extensions outside the core domain, and it has been suggested that ATP-independent activities of such DBPs could be influenced by their flanking domains. For example, removal of the 86-amino-acid arginine- and glycine-rich C-terminal domain of Ded1, a yeast DBP, reduced the efficiency of annealing stimulation significantly (37). However, three of the four E. coli DBPs studied here—RhlB, RhlE, and SrmB—are 420 to 460 amino acids in size, significantly smaller than most of the eukaryotic DBPs, yet these proteins also possess ATP-independent functions. This suggests that the ATP-independent functions of DBPs could be inherently associated with the core domain rather than with the flanking sequences. Extrapolating to other DBPs, we predict that the majority of these factors will also be found to harbor ATP-independent functions that derive from their core sequences. If that should prove to be the case, instead of referring to DBPs primarily as RNA helicases, a revision of their current status to include their ATP-independent functions will be warranted. In particular, for DBPs that possess both ATP-dependent and ATP-independent activities, it will be of interest to ultimately determine how significantly each set of functions contribute to their role as RNA remodeling factors inside the cell.