DEAD-box RNA helicase domains exhibit a continuum between complete functional independence and high thermodynamic coupling in nucleotide and RNA duplex recognition

Abstract

DEAD-box helicases catalyze the non-processive unwinding of double-stranded RNA (dsRNA) at the expense of adenosine triphosphate (ATP) hydrolysis. Nucleotide and RNA binding and unwinding are mediated by the RecA domains of the helicase core, but their cooperation in these processes remains poorly understood. We therefore investigated dsRNA and nucleotide binding by the helicase cores and the isolated N- and C-terminal RecA domains (RecA_N, RecA_C) of the DEAD-box proteins Hera and YxiN by steady-state and time-resolved fluorescence methods. Both helicases bind nucleotides predominantly via RecA_N, in agreement with previous studies on Mss116, and with a universal, modular function of RecA_N in nucleotide recognition. In contrast, dsRNA recognition is different: Hera interacts with dsRNA in the absence of nucleotide, involving both RecA domains, whereas for YxiN neither RecA_N nor RecA_C binds dsRNA, and the complete core only interacts with dsRNA after nucleotide has been bound. DEAD-box proteins thus cover a continuum from complete functional independence of their domains, exemplified by Mss116, to various degrees of inter-domain cooperation in dsRNA binding. The different degrees of domain communication and of thermodynamic linkage between dsRNA and nucleotide binding have important implications on the mechanism of dsRNA unwinding, and may help direct RNA helicases to their respective cellular processes.

Figure 1.

Illustration of motifs and domain structure of Hera and YxiN. (A) The N- and C-terminal RecA-like domains are shown in black (RecA_N) and gray (RecA_C), respectively. DD is the dimerization domain and RRM is an ancillary RNA recognition motif. The dotted line indicates the boundary of the helicase cores. (B) Homology model of free YxiN_core in the open state (left). Motifs involved in RNA binding are colored blue, and motifs participating in nucleotide binding and hydrolysis are colored red. Green motifs coordinate communication between nucleotide and RNA binding sites. RNA and ATP binding (black arrows) lead to the closed state of the YxiN_core/RNA/nucleotide complex (right). dsRNA unwinding (red arrows) is coupled to ATP hydrolysis causing the release of ADP, ssRNA and inorganic phosphate P_i. The structures of free YxiN_core and YxiN_core/U6 RNA/mAMPPNP complex were modeled by SWISSMODEL using crystal structures of free mjDEAD (PDB: 1hv8) and bound eIF4A-III (PDB: 2hyi) as templates, respectively (12). Structures in (B) were created by PYMOL.

Figure 2.

Binding of helicase cores and individual RecA domains to nucleotide in the absence and presence of excess RNA. (A) For RecA_N and helicase cores, the relative fluorescence emission at 444 nm after excitation at 280 nm (FRET from Trp to mAMPPNP) was used as a probe to monitor binding. Constant ∼0.8–1.0 μM protein was titrated with mAMPPNP. K_d values (± standard error of the mean from two independent experiments) according to analysis with Equation (1) are 733 ± 101 μM (Hera RecA_N), 514 ± 62 μM (Hera_core), 571 ± 7 μM (YxiN RecA_N) and 324 ± 87 μM (YxiN_core). For RecA_C (lacking tryptophans), binding was followed via the fluorescence emission at 444 nm of 1.0 μM mAMPPNP (excitation at 365 nm) with increasing concentration protein. No changes in fluorescence emission are observed. (B) Binding of YxiN and Hera cores to nucleotide in the presence of excess RNA, monitored by fluorescence emission at 520 nm upon excitation at 365 nm (FRET from mAMPPNP to the fluorescein coupled to the dsRNA). 400 nM YxiN_core (red) or Hera_core (black) was titrated with mAMPPNP in the presence of saturating 3.0 μM 14mer dsRNA. Data are corrected for background fluorescence of mAMPPNP at 520 nm. The K_d values are 14.8 μM and 2.1 μM for YxiN and Hera_core, respectively. The incubation time was 2–3 min.

Figure 3.

Binding of helicase cores and individual RecA domains to 14mer dsRNA. (A) Changes in relative fluorescence emission of 50 nM fluorescein-labeled 14mer dsRNA upon incubation with different concentrations of Hera (black) or YxiN (red). There is no detectable fluorescence change by YxiN_core, RecA_N or RecA_C, indicative of no binding to the 14mer dsRNA. The dissociation constants (K_d) for Hera are 889 ± 236 nM (Hera RecA_N), 327 ± 161 nM (Hera RecA_C) and 7 ± 3 nM (Hera_core). The errors are standard deviations of three independent measurements. (B) Fluorescence anisotropy of 50 nM 14mer dsRNA upon incubation with Hera (black) and YxiN (red). Hera binding to dsRNA reduces the anisotropy. Addition of YxiN to the RNA has no effect on the anisotropy. The K_d values for Hera/RNA complexes (three measurements) were similar to those determined in (A). (C) Representative time-domain fluorescence intensity decays (I, in photon counts) of 500 nM 14mer dsRNA measured at room temperature with the emission polarizer oriented at 54.7° (magic angle) in the absence (green) and presence of 10 μM Hera_core (black) or YxiN_core (red). Fluorescence decays of the RecA_N and RecA_C subunits were similar to the respective cores (Supplementary Figure S2). Three fluorescence lifetimes (Table 1) were determined from the decays, deconvoluted by the IRF (yellow). (D) Intensity-weighted average lifetimes of 14mer dsRNA in the absence (green) and presence of 10 μM Hera (black) or YxiN (red) subunits. All measurements were performed in 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM MgCl₂, 2 mM 2-mercaptoethanol.

Figure 4.

Time-resolved anisotropy decays of fluorescein-labeled 14mer dsRNA in the absence (A) and presence of 10 μM Hera_core (B). The parallel (VV, black) and perpendicular (VH, red) anisotropy decays are described by three lifetimes and two rotational correlation times (Table 1). The IRF is shown in green. (C) Support plane error analyses of the long (Φ₁ [ns]) and short (Φ₂ [ns]) rotational correlation components of free RNA (black) and Hera_core-dsRNA complex (red). The dotted lines indicate the 75% confidence interval and are colored accordingly.

Figure 5.

Binding of helicase cores to RNA in the presence of nucleotide. Relative changes in fluorescence emission at 520 nm of 5–50 nM 14mer dsRNA after direct excitation at 496 nm. 14mer dsRNA was titrated with Hera_core (A) or YxiN_core (B) in the presence of 1.0 mM mAMPPNP (∼70% protein saturation). Data from two independent measurements (open and crossed squares) were analyzed in a global fit according to Equation (1). The K_d values are 15.1 nM and 2 nM for Hera and YxiN_core, respectively. (C) Reduction in steady-state fluorescence anisotropy of 14mer dsRNA upon binding of Hera_core (filled squares) or YxiN_core (open squares) to the RNA. The conditions were the same as in (A). The equilibration time was at least 15 min.

Figure 6.

Continuum from complete functional independence of RecA-like domains to high thermodynamic coupling. Summary of nucleotide and dsRNA binding mechanisms for different DEAD-box helicases. The black star denotes an intermediate. Dissociations constants are included. The large arrow head (left, gray) symbolizes increasing thermodynamic coupling from Mss116 through Hera to YxiN. *The model for functional independence of the Mss116p RecA-like domains in nucleotide and RNA binding was derived from (10). The K_d values are from (18). The K_d values for dissociation of RNA from the nucleotide/RNA/complex of Mss116 (18) and Hera_core were calculated from the energetic balance of the thermodynamic cycle.