Solution Structure of the QUA1 Dimerization Domain of pXqua, the Xenopus Ortholog of Quaking

The STAR protein family member Quaking is essential for early development in vertebrates. For example, in oligodendrocyte cells it regulates the splicing, localization, translation and lifetime of a set of mRNAs that code for crucial components of myelin. The Quaking protein contains three contiguous conserved regions: a QUA1 oligomerization element, followed by a single-stranded RNA binding motif comprising the KH and QUA2 domains. An embryonic lethal point mutation in the QUA1 domain, E48G, is known to affect both the aggregation state and RNA-binding properties of the murine Quaking ortholog (QKI). Here we report the NMR solution structure of the QUA1 domain from the Xenopus laevis Quaking ortholog (pXqua), which forms a dimer composed of two perpendicularly docked α-helical hairpin motifs. Size exclusion chromatography studies of a range of mutants demonstrate that the dimeric state of the pXqua QUA1 domain is stabilized by a network of interactions between side-chains, with significant roles played by an intra-molecular hydrogen bond between Y41 and E72 (the counterpart to QKI E48) and an inter-protomer salt bridge between E72 and R67. These results are compared with recent structural and mutagenesis studies of QUA1 domains from the STAR family members QKI, GLD-1 and Sam68.


.2) Site directed mutagenesis
The overall protocol followed was adapted from the manufacturer's instructions for the QuikChange Site-Directed Mutagenesis kit (Strategene). Forward and reverse primers (Table S3) were prepared to a final concentration of 125 ng μL -1 . The reaction mixture was prepared according to Table S4 and PCR was performed using the conditions reported in Table S3. 1 μL of DpnI restriction enzyme was then added to the PCR reaction mixture and incubated at 37 °C for 1 h. 10 μL of the DpnI digestion products were mixed with E. coli DH5α competent cells, which were then transformed by the heat shock method: incubation at 42 °C for 90 s, immediately followed by incubation on ice for 2 min. Cells were then plated onto LB-amp agar plates and incubated overnight at 37 °C.  GGA GAA GCC GAA GCC GAC TCC AGA  RP: TCA TGA ATT CTC AGT ACA TAT CTT TCC GTA CTC TGC T   QUA1-C59S  FP: GCC TGC CCA ACT TCT CCG GGA TAT TTA CCC  RP: GGG TAA ATA TCC

.3) Analytical ultracentrifugation
Sedimentation velocity (SV) and sedimentation equilibrium (SE) experiments were performed with an An60-Ti rotor in an Optima XL-1 (Beckman-Coulter) analytical centrifuge, using an Epon double-sector centrepiece and interference optics, at 20 °C. 160 µL of buffer was added into the reference section of the center-piece, while 140 µL of protein solution was added into the sample compartment at 0.5 mg mL -1 , 1.0 mg mL -1 and 2.0 mg mL -1 concentrations in three different cells. SV experiments were performed at 60,000 rpm. SE experiments were performed at 25,000 rpm, 36,000 rpm and 42,000 rpm with an initial equilibration period of 8 h and further equilibration period of 4 h at higher speeds. Values for the buffer density (1.00499 g mL -1 ), viscosity (0.010214 g cm -1 s -1 ) and partial specific volume of the protein sample (0.7411 mL g -1 ) were determined using SEDNTERP [Lebowitz . All results were analyzed using SEDPHAT: SV data were fitted to a continuous sedimentation coefficient distribution model with 150 increments in the range 1.0 S to 2.0 S; and SE data were fitted globally to a single species model [Schuck, 2003].

(1.4) NMR experiments and analysis
Protein NMR spectra were collected using standard pulse sequences [Cavanagh et al., 2006]  ; 32* × 90* × 512* points and τ max of 5 ms, 12 ms and 51 ms in the 13 C, 1 H and 1 H dimensions ( 13 C/ 15 N X-filtered NOESY; at 800 MHz). All NOESY spectra were recorded with 100 ms mixing times. Water suppression was achieved by 'flipback' methods, using shaped selective pulses to return water magnetization to the z-axis prior to acquisition. Pulsed field gradients were used to suppress undesired coherence pathways and the residual water signal. All spectra were processed and interpreted using the CcpNmr Azara and Analysis packages [Vranken et al., 2005].
The 15 N longitudinal relaxation rate (R 1 ) experiment was recorded with ten delay times (10, 50, 100, 150, 250, 400, 550, 700, 850, and 1000 ms) and the transverse relaxation rate (R 2 ) experiment with eight delay times (14.4 (twice), 28.8, 43.2, 57.6, 86.4, 72.0, 100.8 and 115.2 ms (twice)). A heteronuclear steady state { 1 H}-15 N NOE experiment was also recorded. The decay data were fit to the equation I(t) = A.exp (-Bt). For each residue the R 2 /R 1 ratio was used to determine an effective rotational correlation time according to equation 8 of [Kay et al., 1989]; the mean and standard deviation of these values are reported in the main text as the overall rotational correlation time τ C .

(1.5) Circular dichroism
Circular dichroism (CD) spectra were recorded with 0.1 mg mL -1 protein in 10 mM potassium phosphate buffer at pH 8.0 on an AVIV 315 spectrometer using quartz cuvettes. Spectra for secondary structure determination were collected between 190 nm to 250 nm. Secondary structure content was estimated using the SELCON3 algorithm on the DichroWEB server [Whitmore et al., 2004]. Thermal denaturation experiments were monitored at 209 nm, raising the temperature in 1 °C min -1 steps from 20 °C to 90 °C and then back to 20 °C at the same rate. Spectra were averaged over 10 scans and background spectra were subtracted as appropriate. The thermal denaturation profile was fitted to the two-state equation [Fersht, 1999]: where α N , β N , α D and β D describe the partial differential ellipticities of the native and denatured states, T is the absolute temperature, ΔS is the entropy of unfolding and ΔH is the enthalpy of unfolding. The melting temperature T m was estimated from ΔH/ΔS.   show typical results from a sedimentation velocity experiment at a protein concentration of 2 mg mL -1 : (A) concentration profiles; (B) residuals plots; and (C) a plot indicating the distribution of sedimentation coefficients necessary to fit the data; these experiments demonstrate that QUA1-C59S is present in solution as a single species. Parts (D) and (E) show typical results from a sedimentation equilibrium experiment at 25,000 rpm: (D) concentration profiles for sample concentrations of 0.5 mg mL -1 (blue), 1.0 mg mL -1 (red) and 2.0 mg mL -1 (black), and (E) a residuals plot; a global single species fit yielded a mean molecular weight of 12.2 ± 0.3 kDa. Figure S4: Summary of NMR data used to define elements of secondary structure in pXqua QUA1-C59S. The amino acid sequence is indicated, along with sequential and medium range NOE connections, secondary chemical shift values for 1 H α and 13 C α nuclei, and a schematic showing the boundaries of α-helices detected in the final solution structure. Figure S5: Per residue predictions of α-helical content made using the AGADIR server [Lacroix et al., 1998] for: (A) the wild type pXqua QUA1 sequence; (B) QUA1-C59S; (C) QUA1-C59S/E72G. Figure S6: CD studies of the secondary structure content and thermal stability of pXqua QUA1-C59S, showing: (A) far ultraviolet CD spectrum of a 0.1 mg mL -1 protein sample; and (B) thermal unfolding monitored by following the CD signal at 209 nm. Using the spectrum shown in (A), the DichroWEB server [Whitmore et al, 2004] reported an α-helical content of 60 %. Fitting of the thermal denaturation profile yielded a melting temperature T m of 65 °C.