Fig 1.
Flowchart illustrating the strategy for protein-ssRNA docking with structural fragments.
Fig 2.
Structures of ssRNA-protein test-cases complexes.
(A) Crystallographic structure of the human sex-lethal protein bound to a 5'-U8 ssRNA, PDB ID 1B7F. (B): Crystallographic structure of the polyA-binding protein bound to a 5'-A8 ssRNA, PDB ID 1CVJ. The protein is represented in gray cartoon, the RNA in stick colored as rainbow from blue (5', frag1) to red (3', frag6). The nucleotides discarded for docking experiments are colored in white.
Table 1.
Comparison to the bound form of the poses obtained by (I) bound docking, (II) position-specific and (III) non-position specific filtering of chain-forming poses.
Table 2.
Comparison to the bound form of the poses/chains obtained by (I) biased docking, then (II) chain-propensity filtering.
Fig 3.
Best approximation of the RNA bound form obtained by biased fragment docking and chain-propensity filtering.
The bound form of the RNA in 1B7F (A) or 1CVJ (B) is represented as white sticks, and the best approximation for each fragment as lines, colored from blue to orange by green for frag1 to frag6. In both cases, frag5 and frag6 are best approximated by the same docking pose.
Table 3.
Comparison to the bound form of the poses obtained by (I) unbound docking, then (II) chain-propensity filtering, then (III) 3Å-clustering and chain-propensity filtering.
Fig 4.
Best sampling obtained by unbound fragment docking and chain-propensity filtering.
1B7F (A) and 1CVJ (B), same legend as for Fig 3.
Fig 5.
Delimitation of the RNA-binding site by unbound docking.
The 7863 and 3268 chain-forming poses for 1B7F (A) or 1CVJ (B) are represented as green spheres, the bound RNA as red sticks and the protein as surface. All poses are visible on this view.
Fig 6.
Best solution obtained by unbound fragment docking, chain-propensity filtering, and poses clustering.
For 1B7F (A) or 1CVJ (B), the best prediction (in RMSD) for each fragment in frag1-5, out of a total of ~300 poses for the whole complex (small beads), is superimposed to the bound form of the RNA (sticks and large beads). The orientation of each nucleotide is distinguishable by colors: backbone in orange, sugar in green and base in blue. For clarity, the nucleotides are in an ultra-coarse grained representation, with three bead for the base, the sugar and the phosphate.
Fig 7.
Enrichment in correct solutions by poses and chains scoring.
Plotting of the RMSD to closest bound fragment of the poses (A/C) and chains (B/D) obtained by biased docking for 1B7F (A/B) and 1CVJ (C/D), according to their ranking by Sposes and Schains scoring functions respectively. The conformers were docked all together in a single docking run, and each pose was compared to each of the 6 bound fragments.