Computational protocol: Crystal Structure of the S. solfataricus Archaeal Exosome Reveals Conformational Flexibility in the RNA-Binding Ring

Similar protocols

Protocol publication

[…] Our Rrp4-exosome isoform was concentrated to ∼10 mg/mL in a buffer containing 100 mM Tris-HCl pH 8.6, 200 mM Mg(Ac)2, and 25 mM (NH4)2SO4), and crystallized by hanging vapor diffusion against a well solution containing 30% PEG4000 with 100 mM Tris-HCl pH 8.6, 200 mM Mg(Ac)2. Crystals appeared after two months incubation at room temperature among heavy precipitations. The crystals were soaked in the well solution plus 10 mM NaSO4 for 10 minutes before flash-frozen in liquid nitrogen. A complete data set to 2.9 Å was collected at 90K from CHESS beamline A1. The crystal belongs to C2 space group with unit cell dimensions of 151 Å×145 Å×97 Å, and γ = 93.8°, as indexed and scaled by HKL2000 . The apo Rrp4-exosome structure was solved by molecular replacement using PHASER from the structure of the catalytic core of the S. solfataricus exosome (PDB code: 2BR2) as the search model. Initial search using the nine-subunit complex resulted in very poor and untraceable maps in seven out of the nine subunits, while searching attempt with three copies each of individual Rrp41or Rrp42 subunits resulted in more than 100 steric clashes. Finally, a search using three pairs of Rrp41–42 heterodimer produced a reasonable solution and electrondensity map revealing traceable difference between the search model and our structure. Rigid body refinement by Refmac further improved the overall electron density and revealed extra electron densities corresponding to Rrp4 in the Fo-Fc difference map. Initially, strict three-fold non-crystallographic symmetry (NCS) was applied to refine the catalytic ring structure (Rrp41/42 heterotrimer) but was dropped when refinement can no longer improve the model and Rfree. Only two of the three Rrp4 subunits can be correctly placed by molecular replacement in PHASER using the published Rrp4 structure (PDB code 2JE6) , while attempting to position the third subunit resulted in broken densities. Therefore the third Rrp4 subunit was manually traced using COOT and built from three- to ten-amino acid-long peptide segments of Rrp4. At this stage the strict NCS matrix was removed from subsequent refinements. Successive steps of manual refinement of all nine subunits using COOT allowed tracing of more flexible regions as well as locating three sulfate ions at catalytic core domains., Multiple rounds of manual fitting followed by restrained and rigid body refinement in Refmac5 reduced Rwork/Rfree to 0.31/0.34 and positioned all three Rrp4 proteins into density map. At this stage the, Rwork/Rfree stopped dropping due to over refinement in some of the subunits and under refinement in others. Therefore, we treated each of the nine subunits separately and independently subject them to energy minimization, simulated annealing, grouped B-factor refinement in CNS while holding remaining eight subunits in position, which reduced Rwork/Rfree to 0.29/0.31. We then defined a total of 15 translation-liberation-screw (TLS) groups in the entire structure: three of each Rrp41, Rrp42, N-ter, S1 and KH domain of Rrp4 subunit, to be used in TLS refinement in Refmac5 which ultimately reduced refinement statistics to the final Rwork/Rfree = 0.268/0.289 (, ). The TLS refinement result is visualized by Raster3D . The structure is verified by simulated annealing omit maps by excluding 7.5 percent of structure in each calculation. […]

Pipeline specifications

Software tools Coot, REFMAC5, CNS, Raster3D
Applications Drug design, Protein structure analysis
Organisms Sulfolobus solfataricus, Bacteria