The DEAD-Box Protein CYT-19 Uses Arginine Residues in Its C-Tail To Tether RNA Substrates

Abstract

DEAD-box proteins are nonprocessive RNA helicases that play diverse roles in cellular processes. The Neurospora crassa DEAD-box protein CYT-19 promotes mitochondrial group I intron splicing and functions as a general RNA chaperone. CYT-19 includes a disordered, arginine-rich “C-tail” that binds RNA, positioning the helicase core to capture and unwind nearby RNA helices. Here we probed the C-tail further by varying the number and positions of arginines within it. We found that removing sets of as few as four of the 11 arginines reduced RNA unwinding activity (k_cat/K_M) to a degree equivalent to that seen upon removal of the C-tail, suggesting that a minimum or “threshold” number of arginines is required. In addition, a mutant with 16 arginines displayed RNA unwinding activity greater than that of wild-type CYT-19. The C-tail modifications impacted unwinding only of RNA helices within constructs that included an adjacent helix or structured RNA element that would allow C-tail binding, indicating that the helicase core remained active in the mutants. In addition, changes in RNA unwinding efficiency of the mutants were mirrored by changes in functional RNA affinity, as determined from the RNA concentration dependence of ATPase activity, suggesting that the C-tail functions primarily to increase RNA affinity. Interestingly, the salt concentration dependence of RNA unwinding activity is unaffected by C-tail composition, suggesting that the C-tail uses primarily hydrogen bonding, not electrostatic interactions, to bind double-stranded RNA. Our results provide insights into how an unstructured C-tail contributes to DEAD-box protein activity and suggest parallels with other families of RNA- and DNA-binding proteins.

Graphical Abstract

DEAD-box helicase proteins participate in nearly all cellular processes involving RNA. They accomplish their diverse roles by using energy from ATP binding and hydrolysis to engage in a cycle of tight, yet regulated, binding to a segment of single-stranded RNA. For many DEAD-box proteins, this cycle produces local RNA unwinding that promotes rearrangements of structured RNAs or ribonucleoprotein complexes.^1–4

All DEAD-box proteins include a core of two RecA-like domains with a number of conserved motifs. The core domains bind ATP and RNA and unwind short RNA helices.^5–8 While the structures and activities of the core are conserved, many DEAD-box proteins also include amino- or carboxyl-terminal domains or extensions that vary widely in size and composition.⁹ These ancillary domains contribute to the diversity of cellular functions for individual DEAD-box proteins, in some cases by interacting with individual substrates^10–12 or groups of substrates^13–16 to direct DEAD-box proteins to their cellular targets.

CYT-19 is a DEAD-box protein that functions as a chaperone for several group I introns in the mitochondria of Neurospora crassa.¹⁷ CYT-19 can function in folding of group I and group II introns when expressed in a strain of Saccharomyces cerevisiae that lacks the functional ortholog, Mss116,¹⁸ indicating a role as a general RNA chaperone. To achieve this general chaperone activity, CYT-19 uses its ATP-dependent helicase activity to disrupt RNA structure, allowing misfolded RNAs additional opportunities to fold to the native state.^13,19,20 Single-molecule fluorescence and biochemical approaches showed that CYT-19 does not directly disrupt RNA tertiary structure but instead captures and unwinds transiently exposed helices, leading to an RNA chaperone activity that depends on the global stability of the RNA structure.^21,22 The dependence on global RNA stability is suggested to bias CYT-19 toward misfolded RNA structures, which are likely to be less stable on average than their native counterparts.^21,22

The conserved core of CYT-19 is flanked by a small region termed the C-tail. Eleven of 50 amino acids in the C-tail are arginine, making the region highly basic.²³ Biochemical experiments showed that removing the C-tail has a minimal effect on unwinding of short RNA helices in isolation but reduces the level of unwinding of helices that are appended to a group I intron RNA or even to just an additional helix.²³ In addition, small-angle X-ray scattering (SAXS) measurements showed that the C-tail is unstructured and lies adjacent to core domain 2, such that it is well positioned to interact with RNA structural elements.¹⁵ Together, these results led to a model in which the C-tail enhances RNA unwinding activity of CYT-19 by binding nonspecifically to structured RNA and positioning the helicase core for unwinding of nearby RNA helices. Binding of the C-tail to structured RNAs is also suggested to enable multiple rounds of local RNA unwinding in a single association event with structured RNA.^8,21 These roles are reminiscent of roles proposed for unstructured, basic tails of other classes of nucleic acid-binding proteins, including RNA processing proteins, DNA repair proteins, and transcription factors.^24–28

Here we further probed the mechanism of the C-tail in promoting RNA unwinding by CYT-19. To test the model in which the C-tail uses its arginine residues to interact with RNA, we used a series of CYT-19 C-tail mutants that varied in the number and distribution of arginines. Our results indicate that the C-tail requires positively charged residues, with maximal activity conferred by arginine, and that the C-tail enhances RNA affinity in parallel with enhanced RNA unwinding. In addition, a minimum or “threshold” number of arginines in the C-tail is required for detectable activation of RNA binding and unwinding. Interestingly, salt concentration dependences suggest that the C-tail does not depend primarily on electrostatic interactions but presumably relies instead upon hydrogen bonds with RNA. Our results raise the possibility that the RNA affinity of the C-tail is tuned through evolution for the optimal function of CYT-19 and suggests parallels with unstructured domains in diverse RNA- and DNA-binding proteins.

MATERIALS AND METHODS

Materials

RNA and DNA oligonucleotides were from Dharmacon and IDT, respectively, and purchased enzymes were from New England Biolabs. The Tetrahymena ribozyme (L-21/ScaI) was transcribed from a ScaI-digested plasmid using T7 RNA polymerase and purified using a Qiagen RNeasy column. Oligonucleotides were 5′-end-labeled with [γ-³²P]ATP by using T4 polynucleotide kinase and purified by non-denaturing polyacrylamide gel electrophoresis (PAGE).

Cloning of CYT-19 C-Tail Mutants

All clones were produced from a wild-type pMal-CYT-19 vector, a derivative of pMal-c2x from New England Biolabs, which encodes an N-terminal maltose-binding protein (MBP) tag. The Δ616–626 mutant was truncated by using site-directed mutagenesis primers to delete 33 nucleotides directly upstream of the stop codon. The sequences of the primers were 5′-CACAGGGCTCAGGATCTCGATTAACTGCAGGCAAG and 5′-GTGCCAAGCTTGCCTGCAGTTAATCGAGATCCTGAG. Sequence substitution mutants were produced by using six shorter overlapping oligos for polymerase chain reaction (PCR) synthesis to produce a fragment corresponding to the C-tail that included flanking NdeI and HindIII restriction sites. An NdeI site was incorporated upstream of the C-tail sequence as a silent mutation via QuikChange mutagenesis using the primer 5′-GAGCATCATGAAGCATATGGGCCGTG and the complementary sequence. The pMal-CYT-19 plasmid and the C-tail fragments were digested with NdeI and HindIII, and the digested plasmid was treated with phosphatase. Products were isolated with a Qiagen PCR cleanup kit and then ligated using T4 DNA ligase. The entire coding sequences of the clones were then verified via Sanger sequencing.

Expression and Purification of Wild-Type and Mutant CYT-19 Proteins

Proteins were expressed as fusions with an N-terminal MBP tag. Each protein was expressed and purified as described previously²³ except that the tag was retained (Figure S1). We found that the presence of this tag does not impact RNA unwinding activity significantly, as the second-order rate constants measured for MBP fusions of wild-type and Δ578–626 mutant proteins were the same within error as measured previously for the untagged proteins.²³

RNA Unwinding Assays

RNA unwinding activity was measured using gel mobility shift assays essentially as described previously.¹³ To monitor dissociation of the substrate from the Tetrahymena ribozyme, 25 nM ribozyme was incubated for 5 min at 25 °C in the presence of 10 mM Mg²⁺ and 50 mM Na⁺-MOPS at pH 7.0. This incubation gives predominantly misfolded ribozyme,^29,30 which provides a defined substrate for measuring unwinding of the P1 helix.^13,23 A trace amount (<1 nM) of ³²P-labeled oligonucleotide substrate (rSA₅, the sequence of which is CCCUCUA₅) was then incubated with the ribozyme for 15 min at 25 °C to allow binding. Immediately before the addition of CYT-19, excess unlabeled rSA₅ and saturating ATP-Mg²⁺ were added to the reaction mixture. Unless otherwise indicated, the reaction conditions were as follows: 10 nM ribozyme, 2.5 μM rSA₅, 50 mM KCl, 2 mM ATP, 10 mM MgCl₂, and 50 mM Na⁺-MOPS at pH 7.0. Reactions were quenched with a high concentration of MgCl₂ (80 mM), proteinase K (2 μM), and glycerol with xylene cyanol and mixtures subsequently loaded onto a 12% nondenaturing polyacrylamide gel run at 4 °C. Experiments monitoring the dissociation of substrate from model P1 helix complexes were performed under identical conditions except that the P1 helix was formed by incubating the unlabeled strand (600 nM) with trace rSA₅* for 1 h at 4 °C. The reaction mixture was then heated to 25 °C to measure RNA unwinding.

ATPase Activity Measurements

ATPase assays were performed as described previously²² at 25 °C in 10 mM Mg²⁺ and 50 mM Na⁺-MOPS at pH 7.0 with 50 μM ATP and trace [γ-³²P]ATP. Oligonucleotide substrates were added immediately before reactions were initiated by adding CYT-19 to a final concentration of 300 nM. We compared functional binding of CYT-19 to a ssRNA substrate with or without a helical DNA extension. We used DNA rather than RNA for the extension to avoid ATPase activity from binding of the helicase core to the helical extension rather than to the ssRNA. Reactions were quenched in 100 mM (final) ethylenediamine-tetraacetic acid (EDTA) and mixtures applied to a poly-ethylenimine cellulose thin-layer chromatography plate, which was developed with 0.5 M LiCl and 1 M formic acid.

Data Analysis

All data were plotted and fit using Kaleidagraph (Synergy Software). Second-order rate constants were calculated from the averages of at least two independent determinations. Reported uncertainty values reflect the standard error. The K_d values from ATPase curves were generated using Kaleidagraph based on pooled data from at least two different days.

RESULTS

To learn more about the functional roles and mechanisms of the CYT-19 C-tail, we asked whether its ability to promote RNA unwinding by CYT-19 depends on its 11 arginine residues. Therefore, we generated a mutant in which all 11 arginines of the C-tail were replaced with the polar but uncharged amino acid serine [Δ-11R (Table 1)].

Table 1.

CYT-19 C-Tail Construct Names and Corresponding Sequences

^*The C-terminal 54 amino acids of the wild-type CYT-19 protein sequence are shown. For mutants, all arginines are indicated and bolded residues in mutant sequences represent changes from the wild-type sequence.

We first tested the ability of the Δ-11R mutant CYT-19 to perform ATP-dependent unwinding of the 6 bp, “P1” RNA helix from within the misfolded Tetrahymena group I intron ribozyme, an activity that depends strongly on the C-tail (Figures 1A and S2).^13,23 In single-turnover measurements, the Δ-11R mutant gave a second-order rate constant of (1.2 ± 0.2) × 10⁶ M⁻¹ min⁻¹, more than 10-fold lower than that for wild-type CYT-19 [(1.9 ± 0.3) × 10⁷ M⁻¹ min⁻¹ (Figure 1B,C)]. Indeed, the value for the Δ-11R mutant was the same within error as that of a mutant that lacks the C-tail entirely, Δ578–626 [(1.5 ± 0.6) × 10⁶ M⁻¹ min⁻¹ (Figure S3; see also the summary of results in Figure 3)], suggesting that some or all of the arginine residues are required for detectable C-tail activity under these conditions.

Figure 1.

CYT-19 unwinding of the P1 helix from the Tetrahymena ribozyme. (A) Secondary structure of the ribozyme with the P1 helix highlighted in black. (B) Progress curves of P1 helix unwinding without CYT-19 (○) and with 25 nM (□) and 50 nM (◇) wild-type CYT-19. (C) Progress curves of P1 helix unwinding without CYT-19 (○) and with 20 nM (□), 40 nM (◇), and 60 nM (△) Δ-11R CYT-19. (D) Dependence of the P1 helix unwinding rate on protein concentration for wild-type CYT-19 (□), Δ-11R (○), and ΔRtoK (△).

Figure 3.

Second-order rate constants for P1 helix unwinding in the isolated construct (white), the P1–P2 construct (gray), and the ribozyme (black) for all mutants. Values and standard errors are listed in Table S1. Plots used to determine second-order rate constants are shown in Figures S3–S5.

To determine whether it was the loss of positive charge that led to the inactivation of the C-tail, we generated a mutant in which all 11 arginine residues of the C-tail were substituted with lysine [ΔRtoK (Table 1)]. This mutant gave a second-order rate constant of (8.0 ± 0.1) × 10⁶ M⁻¹ min⁻¹, 2–3-fold lower than that of wild-type CYT-19, indicating partial C-tail activity (Figure 1D). These results indicate that positively charged residues are critical and that arginines confer greater activity than lysines.

We next replaced subsets of the arginine residues within the C-tail with serine. In one mutant, denoted Δ-4R, the last four arginines clustered near the C-terminal end were replaced. In a second mutant, denoted Δ-7R, the remaining seven arginine residues were replaced while preserving the last four arginines (Table 1). As observed for the Δ-11R mutant above, both mutants displayed reduced levels of RNA unwinding that were comparable to that of the C-tail deletion mutant (Figure S3; see also the summary of results in Figure 3). We also created a Δ616–626 mutant to test for any effects of the nonarginines within the last 11 amino acids (Table 1). The Δ616–626 mutant displayed RNA unwinding activity comparable to that of the Δ-4R mutant (Figure S3E), indicating that the ability of the C-terminal 11 amino acids to promote RNA unwinding activity depends on the four arginines within this sequence. Together, these results indicate that neither the first seven nor the last four arginines are sufficient for detectable activity of the C-tail in RNA unwinding.

In both the Δ-4R and Δ-7R mutants, long uncharged stretches were introduced into the C-tail while preserving other R-rich sequences. It is notable that eight of the 11 arginines in the wild-type C-tail are found in pairs. To determine whether the overall distribution of the arginines rather than their number is critical for C-tail activity, we replaced the second arginine of each pair with serine. These substitutions removed four arginines without significantly changing the overall distribution of arginines along the C-tail [ΔRpair (Table 1)]. The ΔRpair mutant unwound the P1 helix with a low second-order rate constant of (4.1 ± 1.3) × 10⁵ M⁻¹ min⁻¹ (Figure S3G), indicating that even if the arginines are distributed along the C-tail, a decrease from 11 to seven arginines results in a complete loss of detectable C-tail function.

Together, the data indicate that the ability of the C-tail to increase the second-order rate constant for RNA unwinding depends on the presence of arginines within it. If the number of arginines is decreased from 11 to seven, the activity of the C-tail is lost regardless of whether the seven remaining arginines are clustered at the N-terminal end, clustered at the C-terminal end, or distributed throughout the C-tail. These results suggest that a critical number of arginines is necessary for the activity of the C-tail under these conditions, with this threshold number being between eight and 11 arginines (see Discussion).

To test directly whether the observed differences in unwinding activity were due to loss of RNA binding by the C-tail rather than changes in the properties of the helicase core of CYT-19, we measured unwinding of the isolated P1 RNA helix by wild-type CYT-19 and its mutants. The isolated P1 helix construct lacks adjacent structural elements (Figure 2A), and previous work showed that it is unwound by wild-type CYT-19 and the C-tail truncation with comparable efficiency.²³ We found that while there was some variability between the mutants, there was no systematic relationship between arginine content and RNA unwinding efficiency (Figures 2B, 3, and S4). We did not further examine whether the modest range of activity levels of the C-tail variants (5-fold) reflects subtle effects of C-tail sequences or the limits of experimental uncertainty. Overall, these results indicate that the basal RNA unwinding activity of the core is maintained in the mutants despite loss of the C-tail function. It is also notable that all of the CYT-19 constructs had lower activity for unwinding the isolated P1 helix than for the same helix attached to the ribozyme, mirroring a previous observation for the C-tail truncation mutant Δ578–626 and suggesting that another site or surface within CYT-19 may contact the intron RNA and make a modest contribution to RNA unwinding activity.²³

Figure 2.

Unwinding of minimal P1 helix constructs. (A) P1 helix construct. (B) Dependence of P1 helix unwinding rate of the construct shown in panel A on protein concentration for wild-type CYT-19 (□), Δ-11R (○), and ΔRpair (◇). (C) P1–P2 helix construct. (D) Dependence of P1 helix unwinding rate of the construct shown in panel C on protein concentration for wild-type CYT-19 (□), Δ-11R (○), and ΔRpair (◇). Data points on the y-intercept represent measured unwinding rates of the same helix in the absence of CYT-19.

We also tested unwinding of a construct in which P1 is attached only to the P2 RNA helix rather than to the entire intron (Figure 2C). This additional helix has been previously shown to restore C-tail-dependent activity to nearly the same level as the full intron.^13,23 As expected, the addition of P2 had little systematic effect on the unwinding activity of the compromised mutants relative to unwinding of the isolated P1 helix, while in side-by-side experiments the presence of P2 increased the unwinding efficiency of wild-type CYT-19 by nearly 100-fold to essentially the same level as that of the full intron (Figures 2D, 3, and S5).²³

Additional Arginines Enhance Activity of CYT-19 Relative to the Wild-Type Protein

Given that reducing the number of arginines in the C-tail compromises its activity in RNA unwinding, we tested whether adding five arginine residues to the C-tail would increase unwinding activity [Δ+5R (Table 1)]. Indeed, this “supertail” mutant gave a second-order rate constant of (7.2 ± 0.6) × 10⁷ M⁻¹ min⁻¹ for unwinding of the P1 helix from the Tetrahymena ribozyme, roughly 4-fold greater than that of wild-type CYT-19 (Figures S3H and 3). As expected, the Δ+5R mutant unwound the isolated P1 helix with an efficiency similar to those of the other CYT-19 constructs, and its activity was increased by attachment of the P2 hairpin (Figures S4H and S5H). Interestingly, the unwinding activity for this P2-P1 construct was no greater for the Δ+5R mutant than for wild-type CYT-19, suggesting that the ability of the five additional arginines to enhance C-tail function may depend on higher-order structural features in the group I intron. Together, these results indicate that the ability of the C-tail to enhance the RNA unwinding activity of CYT-19 for helices appended to structured RNA depends on arginine residues and can be “dialed” up or down by adding or removing arginines.

C-Tail Sequence Changes Modulate the RNA Affinity of CYT-19

The simplest expectation from the current and prior results²³ was that the changes in activity for the C-tail mutants reflect changes in binding affinity of CYT-19 for the RNA unwinding substrates. To test this model directly, we measured functional binding of wild-type CYT-19 and the C-tail mutants to an ssRNA sequence with or without a flanking hairpin using constructs analogous to the P1 helix constructs mentioned above (Figure 4A). To measure only functional binding, we determined affinities from the RNA concentration dependence of ATPase activity. We used a low, subsaturating concentration of ATP (50 μM) to ensure that RNA binding and release would be fast relative to ATP hydrolysis and therefore the K_1/2 value determined from the dependence of ATPase activity on RNA concentration would be equal to the K_d for RNA binding by CYT-19. For wild-type CYT-19, the maximal ATPase rate was roughly the same for the substrates with or without a helical extension (Figure 4B). However, the RNA concentration required for half-maximal stimulation (K_1/2 value) decreased substantially, from 53 ± 14 μM for the minimal substrate to 1.3 ± 0.3 μM for the construct that includes the P2 helical extension. This difference indicates that for wild-type CYT-19, the additional contacts made by the C-tail substantially increase the affinity for RNA substrates that include the helical extension.

Figure 4.

Flanking helix that provides a binding site for the C-tail. (A) ATPase substrates were an 11-nucleotide RNA (sequence highlighted in black) and the larger construct shown, which extends the same ssRNA sequence with the dsDNA helix shown. (B) Wild-type CYT-19 ATPase activity in the presence of the minimal RNA substrate (○; K_1/2 = 53 ± 14 μM) and the substrate that includes a helical extension (□; K_1/2 = 1.3 ± 0.3 μM). (C) Δ-11R ATPase activity in the presence of the minimal substrate (○; K_1/2 = 58 ± 29 μM) and the substrate that includes a helical extension (□; K_1/2 = 60 ± 3 μM). (D) Δ+5R ATPase activity in the presence of the minimal RNA substrate (○; K_1/2 = 37 ± 3 μM) and the substrate that includes a helical extension (□; K_1/2 = 0.28 ± 0.05 μM).

We then performed analogous measurements for representative C-tail mutants. The mutant lacking all 11 arginine residues (Δ-11R) gave K_1/2 values of 60 ± 3 and 58 ± 29 μM for the substrates with and without the helical extension, respectively (Figure 4C). Thus, for this mutant, the affinity for the minimal substrate is similar to that for wild-type CYT-19, but there is no significant enhancement of binding from the helical extension. In addition, the Δ+5R (supertail) mutant gave a K_1/2 value for the substrate with the extension of 0.28 ± 0.05 μM, 4-5-fold lower than that of wild-type CYT-19 (Figure 4D). As expected, the affinity of this mutant for the minimal substrate was similar to the values of the other versions of CYT-19. Together, these results indicate that the C-tail affects the affinity of CYT-19 for RNA substrates with helical extensions and that the differences in unwinding efficiency between different C-tail mutants of CYT-19 reflect differences in RNA binding affinity.

C-Tail Function Does Not Depend Primarily on Electrostatic Interactions

The C-tail has been suggested to bind RNA via electrostatic interactions of its arginine residues with the phosphate backbone of RNA.^15,23,31 To test this model, we measured the ability of the C-tail to enhance the second-order rate constant for unwinding of the P1 helix from the Tetrahymena ribozyme in the presence of various concentrations of KCl. The expectation was that electrostatic interactions would be weakened at higher KCl concentrations, reflecting less favorable counterion release upon complex formation.^32,33 Although the KCl concentration would also likely impact RNA binding and unwinding by the helicase core, these changes would impact the wild-type protein and Δ578–626 equally. Thus, differences in the KCl dependence between the wild-type and Δ578–626 proteins would provide a measure of the electrostatic contacts made by the C-tail.

Somewhat surprisingly, we found that the dependence of RNA unwinding efficiency on KCl concentration is the same within error for wild-type CYT-19 and Δ578–626 (Figure 5). For both proteins, the unwinding efficiency decreases at higher KCl concentrations, but the indistinguishable slopes indicate that this decrease arises from reduced RNA affinity and/or unwinding of the helicase core, not reduced RNA affinity of the C-tail. At lower KCl concentrations, there appears to be a break point, perhaps reflecting a change in the rate-limiting step. Analogous results were obtained for unwinding of the P1 helix appended to P2 (Figure S6). The lack of an effect of KCl concentration on RNA binding by the C-tail indicates that the C-tail does not rely primarily on electrostatic interactions for RNA binding, most likely relying instead on hydrogen bonds with the backbone and/or bases of the hairpin extensions (see Discussion).^32–35 The ΔRtoK mutant also displayed decreased RNA unwinding activity at higher KCl concentrations and gave an indistinguishable KCl concentration dependence (Figure 5), suggesting a reduced level of H-bonding but comparable electrostatic interactions when the arginine residues are replaced with lysine.

Figure 5.

Dependence of RNA unwinding rate on monovalent ion concentration. Second-order rate constants for unwinding of the P1 helix from within the Tetrahymena ribozyme decrease with an increasing monovalent ion concentration for Δ578–626 (□) with a dependence similar to that for wild-type CYT-19 (○) and the ΔRtoK mutant (△). Reaction mixtures included a variable concentration of KCl and a constant background of 20 mM Na⁺ to give the indicated monovalent cation concentration. There appears to be a break point at low ion concentrations, and the data at the lowest cation concentration (33 mM) were not included in the fits. Our standard reaction conditions, which included 70 mM monovalent cation (Figures 1–4), are well separated from this apparent break point.

DISCUSSION

Intrinsically disordered domains are critical to the function of diverse proteins such as transcription and translation factors, DNA repair and RNA processing proteins, and RNA chaperones.^24–26,28 Here we build on the understanding of how a C-tail acts in a DEAD-box protein that functions as an RNA chaperone: we show that the C-tail requires basic amino acids, with greater activity for arginine than for lysine, and that enhanced unwinding activity correlates directly with an increased affinity for RNA substrates.

To test the roles of basic amino acids in the C-tail, we systematically removed its arginine residues in sets. Although the degree of C-tail-dependent enhancement of RNA unwinding activity tracked with arginine content, the relationship was not a simple proportionality, as mutants that retained seven or fewer of the 11 arginine residues lacked detectable activity of the C-tail. We therefore suggest a “threhold” model for activation by the C-tail, such that a C-tail with seven or fewer arginines possesses insufficient affinity for RNA to contribute to CYT-19 binding under these conditions (Figure 6). For these mutants, even when the core is bound, enthalpic contributions from the remaining contacts are apparently insufficient to overcome the entropic penalty of localizing the C-tail (Figure 6B). While changes to the distribution of arginines, different RNA substrates beyond those tested here, or even different solution conditions may change the number of arginines representing this threshold level, we suspect that the presence of the threshold is a general property of interactions of CYT-19 with RNA.

Figure 6.

Threshold model of RNA binding by the C-tail. (A) Wild-type CYT-19 has a sufficient number of arginines to effectively bind nucleic acid extensions, contributing binding affinity and tethering the core to unwind adjacent RNA helices. (B) Mutants of the C-tail with an arginine content lower than a threshold level are unable to form stable interactions between the C-tail and RNA even when the helicase core of CYT-19 is bound to an adjacent helix.

For functional variants of the C-tail, we found that the increase in RNA unwinding efficiency (k_cat/K_M value) for RNA constructs that included a helical extension was mirrored by a similar increase in RNA binding affinity. These results suggest that the principal function of the C-tail is to act as a structural module that increases the RNA affinity of CYT-19 by interacting with the helical extension and tethering the helicase core. Supporting this model, the maximal ATPase value for wild-type CYT-19 is not increased by a helical extension, even when the extension is bound by the C-tail. Although our data do not provide evidence of allostery between the C-tail and the helicase core, it remains possible that such an interaction exists, with the C-tail communicating bound RNA to the helicase core in a way that increases the RNA binding affinity of the core but does not change the ATPase or RNA unwinding rates of the core once it is bound to RNA. Interestingly, the C-tail of the S. cerevisiae homologue Mss116 does increase the maximal rate for unwinding of longer helices,³¹ presumably either by activating the core or by contributing to RNA unwinding directly, analogous to a key role of unstructured domains in ATP-independent RNA chaperones.^24,28 Modest allostery has also been demonstrated between the helicase core and a structured ancillary domain in the bacterial DEAD-box protein YxiN.³⁶

To probe the physical origin of association of the C-tail with RNA, we measured the salt concentration dependence of its binding.^32–35 Our finding that C-tail binding, as reflected in enhanced RNA unwinding activity, does not result in counterion release suggests that it is not dominated by electrostatic interactions and presumably relies instead on hydrogen bonding between the arginines and flanking RNA. While hydrogen bonds occur between dipoles and are also expected to depend on ionic strength, the competing hydrogen bonds of each partner with water have very similar dependences, such that the observed affinity is not expected to depend significantly on ionic strength. Interestingly, analogous results have been obtained for transcription factor p53, in which an arginine-rich tail interacts with dsDNA in a salt-independent manner to increase the level of DNA binding of the structured core.³² The putative role of hydrogen bonding is further strengthened by our finding that lysines do not fully replace the functions of arginines in the C-tail, as lysine has the same charge but is less versatile for hydrogen bond formation.^37–39 Arginine-rich peptides have been shown to bind via hydrogen bond networks preferentially to discontinuous regions of RNA, such as the junction between A-form RNA and a bulge or loop.³⁷ Such regions are common in the physiological group I intron substrates for CYT-19, providing a possible evolutionary basis for the prevalence of arginines in the C-tail.

The increased RNA unwinding activity of the “supertail” mutant raises the question of why nature has not incorporated additional arginines into the wild-type CYT-19 protein. One possible explanation is that CYT-19 has attained a sufficient level of activity such that there is no longer evolutionary pressure for further improvement. Another possibility is that the level of activity of the C-tail is optimized for CYT-19 functions, analogous to evidence that the tails of some DNA-binding proteins are evolutionarily selected for optimal activity.^25,40 In vivo, an increase in C-tail binding affinity could result in excessively tight binding of CYT-19 to RNA, blocking RNA refolding and/or enzyme recycling. The foundation of knowledge and broad range of experimental approaches available for CYT-19 and its yeast homologue Mss116 are likely to make these proteins very useful for further dissection of the functions and potential evolutionary tuning of an unstructured tail within a DEAD-box protein.

ABBREVIATIONS

EDTA

ethylenediaminetetraacetic acid

MOPS

3-(N-morpholino)propanesulfonic acid

DTT

dithiothreitol

SDS

sodium dodecyl sulfate

PAGE

polyacrylamide gel electrophoresis

MBP

maltose-binding protein

dsDNA

double-stranded DNA

dsRNA

double-stranded RNA

ssRNA

single-stranded RNA

rSA₅

oligonucleotide substrate of the Tetrahymena ribozyme, CCCUCUA₅