Design and structural characterization of autoinhibition-compromised full-length Ran

Abstract

Dear Editor, The Ras-related nuclear protein Ran is a small GTPase that functions in nuclear transport, mitotic spindle formation, nuclearenvelope/nuclear-pore complex assembly, and other diverse cytoplasmic activities. Unlike other Ras superfamily proteins, Ran contains a unique autoinhibitory C-terminal tail (C-tail) that accounts for an estimated tenfold lower affinity for GTP as compared with GDP. Multiple missense cancer mutations at the C-tail have been observed, but the biological significance is unknown. The study of recombinant Ran proteins often helps provide insights into cellular and disease mechanisms. To purify active Ran (GTP or GTP analoguebound), one could either mutate the catalytic residue Q69 to slow down the intrinsic GTP hydrolysis or charge the protein with an excess of GTP analogue. These methods are complicated, inefficient, and prone to protein denaturation. Another strategy is to delete the C-terminal 37 residues or only the DEDDDL terminus. However, these mutants are of limited usage due to their inability to bind to key regulators, such as RanBP1 (Ran-binding protein 1). To generate more broadly applicable GTP-bound proteins, and to set a foundation for investigating cancer mutations, four mutations were designed to disrupt the interaction between the C-tail and the G-domain (Fig. 1a). In the RanGDP crystal structure, C-terminal residues L182, M189, and Y197 are each inserted into a surface cavity on the G-domain, and the C-terminal helix is intimately packed against G-domain residue A133. A133D, L182A, M189D, and Y197A were designed to disrupt these interactions. Among these residues, A133 and L182 are strictly conserved from fungi to humans (Fig. 1b). Together with Ran, Ran (unable to hydrolyse GTP), Ran (without the C-terminal 37 residues), and Ran (without the DEDDDL C-terminus), these proteins were purified from Escherichia coli. The designed mutants were similar to Ran in size exclusion chromatography (single peak) and purification yields, and more stable than Ran (Fig. 1c). The C-tail shields a large hydrophobic area in RanGDP. Unlike 1-179, the dislodged and flexible C-tail of GTP-bound WT or mutant Ran remains able to inhibit hydrophobic contacts and prevent protein aggregation. The purified proteins were over 90% nucleotide-bound, though no nucleotides were added in the purification stages. To determine the status of Ran-bound nucleotides, the proteins were denatured using 100mM NaOH, and the samples were analysed using a Q-sepharose anion exchange column (Fig. 1d). As expected, Ran, Ran, and Ran were 6, 87, and 35% GTP-bound, respectively. Though Ran does not hydrolyse GTP, only 12% was GTP-bound. Notably, the C-dis (C-tail disrupting) mutants Ran, Ran, and Ran were substantially GTP-bound (79-85% GTP). Only 23% of Ran was loaded with GTP, suggesting that Y197 is less important for autoinhibition. Although the intrinsic GTP hydrolysis of Ran is slow, a reduced level of bound GTP was observed for Ran after ten days of storage at 4 °C (Fig. 1e). To prevent intrinsic GTP hydrolysis, a Q69L/L182A double mutant was generated, which exhibited 100% GTP binding after purification. In the absence of free nucleotides at 4 °C, this double-mutant remained 100% GTP-bound for at least ten days without obvious precipitation. After one month of storage at −80 °C, Ran was similarly GTP-bound as freshly purified Ran (Fig. 1d). RanBP1, which relieves the blockage of RanGTP hydrolysis by karyopherins, displays a high affinity for RanGTP but not RanGDP. RanBP1 pull-down was used to check the binding activity of freshly purified mutants. Unlike Ran and Ran, which exhibited dosedependent binding, Ran bound to RanBP1 at all concentrations (Fig. 1f). When all Ran mutants were tested at a low concentration, the C-dis mutants were more bound to RanBP1 than the rest of the proteins (Fig. 1g), in agreement with the greater fraction of bound GTP. The binding to RanBP1 by different mutants was specific (Fig. 1h). RanGTP forms a nuclear export complex with CRM1 (Exportin-1) and NES cargoes (nuclear export signal-containing proteins) in the nucleus. In pull-downs with CRM1, L182A dose-dependently enhanced CRM1 binding to GST-tagged NES (Fig. 1i). Further, no CRM1-binding difference was detected for parallelly purified Ran in the presence or absence of free GTP, suggesting that supplementing GTP in the purification buffers may not be necessary. Because the C-tail of Ran is involved in RanBP1 binding, three Cterminal mutants (Ran, Ran, and Ran) were crystallized in complex with RanBP1 and CRM1 to examine whether these mutations alter RanBP1 binding (Fig. 1j, Supplementary Table 1). Since the L182 side chain does not directly contact RanBP1 (Fig. 1k), the differences between the L182A and WT structures (4HAT) are likely caused by differences in the NES groove, but not the mutation. One evidence is that the C-tail of Ran is also similar to that of Ran (Fig. 1l). However, the C-tail of Ran exhibits substantial local changes (Fig. 1m). In the Ran structure, M189 is loosely packed on the edge of a hydrophobic pocket in RanBP1 (Fig. 1n). In Ran, D189 (being more hydrophilic) is flipped out towards the solvent, and a previously solvent-exposed proline (P191) is inserted into that hydrophobic pocket. The movement of P191 drags towards the pocket a one-turn helix (A192–A195), which was originally part of a longer helix (A192–T206). The adjacent end of the now-muchshorter helix is shifted about 2 Å from its original position (Fig. 1m). It should be noted that this change is not due to crystal packing. Both the C-tail electron density and the RanBP1 contact area for Ran are increased as compared to the other mutants (Fig. 1o), suggesting possibly tighter RanBP1 binding. The observed binding changes reflect higher protein dynamics at the binding interface, as seen in other small GTPase interactions. The designed C-dis mutants simultaneously possess two advantages: (1) C-dis mutations enable the effortless and cost-effective purification of highly GTP-bound Ran. Via simple mutation to disrupt the autoinhibitory C-tail, RanGTP could be easily purified without adding any GTP or GTP analogues, nor performing the previously necessary GTP-charging steps. (2) The C-dis mutants are more stable than 1–179, and are able to bind to RanBP1 and RanBP2. The double-mutant Ran was suitable for conducting isothermal