Figure 1

Coarse-grained modeling shows the importance of electrostatic interactions for the binding between poly-PR and Kapβs. (a) (Left panel) 1-bead-per-amino-acid (1BPA) representation of poly-PR with 7 and 25 repeat units. Both Proline and Arginine residues are shown in red. (Right panel) Our 1BPA approach to create coarse-grained (CG) models for Kapβs. The CG beads are placed at the location of α-carbon atoms, here shown for two α-helices of impβ1 on top of the crystal structure. Red, blue, and green beads correspond to negatively-charged residues (D and E), positively-charged residues (R and K), and residues with aromatic rings (F, Y, and W), respectively. Other residues are shown with grey beads. The all-atom crystal structure is depicted in yellow using the New Cartoon representation in VMD. (b) 1BPA representation of the various Kapβs modeled in the current study that serve either as importin or exportin. Importins have superhelical structures whereas exportins have a more ring-like structure in case of KAP124 and CRM1, and U-shaped structure in case of XPO5. The crystal structures used to make the CG models are listed in Table S3. (c) (Top panel) The time-averaged number of contacts \({C}_{t}\) between PR25 and Kapβs plotted against the net charge per residue (NCPR) of Kapβs at a monovalent salt concentration \({C}_{\mathrm{salt}}\) of 200 mM. The values on the vertical axis are normalized by the sequence length of the importins/exportins (\({N}_{\mathrm{Kap\beta }}\)) and the sequence length of poly-PR (\({N}_{\mathrm{PR}}\)). The NCPR values on the horizontal axis is calculated based on the sequence lengths and amino acid compositions of the Kapβ models listed in Table S3 (column 6) and Table S5. Among the selected Kapβs, the importins Imp5, Impβ1, TNPO1, TNPO3 (highlighted with a horizontal black line) have been shown to directly bind to PR25 in in vitro experiments¹⁹. However, no binding has been observed for CRM1 (highlighted with a horizontal red line) in vitro experiments performed at the same monovalent salt concentration¹⁹. A linear correlation can be seen between the normalized \({C}_{t}\) and the negative NCPR. The dashed line shows a linear fit. The error bars are half of the standard deviation (see the SI for details). (Bottom panel) The PR25-Kapβs binding probability \({P}_{\mathrm{b}}\) plotted against NCPR of Kapβs at \({C}_{\mathrm{salt}}=\) 200 mM.