Chikungunya virus capsid protein contains nuclear import and export signals

Semliki Forest virus (SFV) is a well-studied Old World alphavirus and its capsid protein is reported to have an N-terminal NLS [4, 15]. To validate the role of the corresponding region in CHIKV CP, the putative NLS containing sequence was cloned in fusion with EGFP and HEK293 were transfected with pNLS-EGFP. As shown in Figure 1A and B, the NLS-EGFP fusion protein was quantitatively imported to the nucleus with 5.48 times (± 1.41) higher nuclear accumulation when compared to the control EGFP expression (1.59 ± 0.42 times nuclear). As in other alphaviruses, a substantial amount of CP was seen in nucleoli [4]. Karyopherins are known to directly interact in the cytoplasm with proteins bearing a NLS and to carry these through nuclear pores into the nucleus. To investigate possible mechanisms of CHIKV CP nuclear translocation, we analyzed the interactions of this protein with various karyopherins. Recombinant CHIKV CP was expressed in E. coli as a His-tagged protein and purified using the Ni-NTA column system. HEK293 cells transiently transfected with various FLAG-tagged karyopherins (Karα1-4) were lyzed and incubated with the purified CHIKV CP. Pull-down assays revealed that the viral CP interacts strongly and specifically with Karα4 (Figure 2A, lane 4). To confirm the observed interaction between CHIKV CP and Karα4, HEK293 cells were cotransfected with FLAG-Karα4 and CHIKV-CP-EGFP. Cells were lysed and the protein complex was co-immunoprecipitated with an anti-FLAG antibody. Bound proteins were subjected to SDS-PAGE followed by Western blot with an anti-GFP antibody. As shown in Figure 2B, the band detected in lane 4 indicated that CHIKV CP co-immunoprecipitates with Karα4. The experiment was performed five times and similar results were obtained each time. No binding was observed in the control experiments using only beads or a non-specific antibody.

Karα molecules contain two NLS binding sites that directly recognize NLS sequences of cargo proteins. The primary NLS binding site is located at the N-terminal arm (alpha helices that form a hairpin structure) repeats 2 to 4, while the secondary NLS binding site is located at the C-terminal arm repeats 7 to 9 (Figure 3A). The NLS major binding site is involved in the interaction with monopartite NLS, while the minor binding site associates with bipartite NLS. The acidic residues of the binding site undergo electrostatic interactions with positively charged functional groups of basic amino acids in NLS peptide regions. The hydrophobic chains of basic residues, like lysine and arginine in the NLS peptide, also form hydrophobic interactions with conserved residues of the binding site [16‐19]. To further characterize the specific mechanism of CHIKV CP binding to Karα4 molecule, we created the deletion mutant Karα4Δ259, which lacks the minor binding site, and analyzed whether this region is required for the interaction with CHIKV CP by pull-down assay. Indeed, Karα4Δ259 was able to interact with CHIKV CP both alone (Figure 3B, lane 5), and in the presence of excess full-length Karα4 (Karα4-FL) protein (Figure 3B, lane 6). This demonstrates that CP binds to the Karα4 NLS major binding site. One can conclude that CHICKV CP and Karα4 binding occurs as a monopartite NLS-driven interaction.

To verify whether CHIKV CP is a nuclear-cytoplasmic shuttling protein, we used the specific CRM1-mediated nuclear export inhibitor leptomycin B (LMB) [20]. HEK293 cells were transfected with pCHIKV-CP-EGFP and maintained in the presence of 15 ng/ml LMB. As shown in Figure 4A and B, treatment with LMB led to an almost exclusive nuclear accumulation of CHIKV-CP-EGFP (6.68 ± 1.41 times nuclear), whereas in mock treated cells the protein remained distributed in both compartments, the nucleus and the cytoplasm (1.42 ± 0.34 times nuclear). The experiment was repeated using HeLa cells and similar results were observed (data not shown). The presence of this NES was further confirmed through GST pull-down assays. Purified CHIKV CP directly interacted with purified GST-tagged CRM1 whereas no interaction was observed in control experiments (Figure 4C).

To identify a region of CHIKV CP that may contain a nuclear export signal, several deletion mutants were constructed (Figure 5A). Deletion of 46 aa from the C-terminus did not have any effect on the nuclear-cytoplasmic translocation process, as the truncated protein was distributed throughout the cell (Figure 5B and C). Next, the region between aa 143 and 155 that is rich in leucine residues was examined as a potential NES site. This stretch of amino acids is highly conserved among Old World alphavirus isolates (Figure 6A). Structural analysis revealed that this region can form a protruding ridge (Figure 6B) to make it accessible for binding to the CRM1 protein. The domain that consists of a pi-helix and a loop can be flexible in solution and thus, may fold to dock with the CRM1 binding site. A larger C-terminal deletion of 118 aa removed this potential NES and rendered the mutant EGFP fusion protein highly nuclear (6.43 ± 1.91 times more nuclear; Figure 5B and C). To further characterize the NES, two leucine residues at aa positions 149 and 152 were exchanged against alanine by means of an overlap PCR. Expression of the altered protein designated NES mutant resulted in a significant decrease of cytoplasmic localization (5.57 ± 1.67 times more nuclear; Figure 5B and C). These results provide evidence that the region between aa 149 and 152 acts as a NES that enables the CHIKV CP to be exported from the nucleus to the cytoplasm via CRM1.

A structural model of docking between CHIKV CP NES and CRM1 shows the mode of binding between the molecules (Figure 6C). The molecular interaction is similar to that observed between snurportin-1 NES and CRM1 (PDB 3NBY) despite minor differences in the spacing of hydrophobic residues. The latter can also be superpositioned in the same orientation (Figure 6D) and form a consensus NES which is structurally equivalent. The modeled structure of CHIKV CP NES and snurportin-1 possesses all the hydrophobic residue side chains aligned to one face of helical structure. The comparison of binding energy between NES and CRM1 shows the global binding energy for CHIKV CP NES and CRM1 to be −71.55 joules, whereas the value for the snurportin-1 NES is −96.78. NES conserved hydrophobic residues binding to CRM1 pockets contribute most to the global binding energy (Table 1). This shows further insight to the mode of interaction between NES and CRM1 for nuclear export, and on the role of NES hydrophobic residues to dock with CRM1.

The Chik-Wü virus isolate (Genbank: EU037962) or subclones thereof (GenBank: HM369441) served as templates for PCR amplification of the CHIKV capsid and all subclones [3, 48]. To clone pCHIKV-CP-EGFP, full-length CHIKV CP was amplified from our previously described expression plasmid in which the protease cleavage site had been inactivated by exchanging serine 213 to alanine [3]. The amplimer created with the oligonucleotide primers HindIIIChikcap_fwd 5′-aagcttatggagttcatcccaacccaaac-3′ and ChikcapXmaI_rev 5′-atcccgggtccactcttcggctc-3′ was inserted into pEGFP-N1 (Clontech) into the restriction sites HindIII and XmaI in frame with EGFP. The NLS region in CP located between aa positions 60 and 100 was inserted into pEGFP-N1 after amplification with primers HindIIINLScap_fwd 5′-aagcttatgaagccacgcaggaatcgg-3′ and XmaINLScap_rev 5′-atcccgggtcggcttctttttcttttg-3′ to create pNLS-EGFP. Construction of Chikcap Δ46 in which 46 aa were deleted from the C-terminus has been described previously [3]. The mutant with the larger C-terminal deletion Δ118 was constructed with HindIIIChikcap_fwd and Chikcap118_rev 5′-atcccgggtcccctttacgtgtgctgg-3′. To mutate leucines of CP at position 149 and 152 to alanine, an overlap PCR was performed with the above and overlapping primers 5′-cttaaaggccgctttggccgcgtccgc-3′ and 5′-gcggacgcggccaaagcggcctttaag-3′. Expression in E. coli of the complete CHIKV CP was carried out by inserting a PCR-derived product amplified with chk_petcapsid_fwd 5′-ggatccatggagttcatcccaacc-3′ and chk_petcapsid_rev 5′-ctcgagtccactcttcggctc-3′ from pChikcap [3] into pET28a (Invitrogen) between BamHI and XhoI sites. pGEX-CRM1 containing GST-tagged CRM1 was a kind gift of Nabeel R. Yaseen [49]. The mammalian expression vectors pCAGGS containing either Karα1, 2, 3, or 4 with N-terminal FLAG-tags were a generous gift of P. Palese and have been described previously [50, 51]. The deletion mutant pCAGGS Δ259 Karα4 was constructed by digesting the full-length vector with Tth111I followed by Klenow filling and religation.

HEK293 cells untransfected or transfected with either of the FLAG-tagged constructs pCAGGSKarα 1, 2, 3, 4, or Δ259Karα4 were lyzed with RIPA buffer (Thermo Scientific) and incubated after adding purified His-tagged CHIKV CP. Ten μl of resuspended Ni-NTA beads were added to bacterially expressed CHIKV CP, mixed, and incubated further at 4°C. Samples were centrifuged and the supernatant was discarded. The beads were washed with RIPA buffer and boiled for 10 min after adding 40 μl Laemmli buffer. The samples were subjected to SDS-PAGE and transferred to nitrocellulose membranes. The membranes were probed with monoclonal anti-FLAG antibodies (Sigma) followed by anti-mouse secondary antibodies conjugated with horseradish peroxidase (Santa Cruz Biotechnology). Bands were visualized with enhanced chemiluminescence detection reagents (Amersham). GST pull-down assay was performed using glutathione-Sepharose resin (GE Healthcare) according to manufacturer’s instructions. Briefly, GST or GST-CRM1 protein immobilized on glutathione-Sepharose was incubated with 2 μg of purified CHIKV CP at 4°C overnight. Bound proteins were eluted by boiling the Sepharose in Laemmli buffer, resolved by a 12% SDS-PAGE, transferred to nitrocellulose, and analyzed by Western blot.

For Co-IPs, HEK293 grown on Petri dishes (10 cm) were transfected with FLAG-tagged pCAGGSKarα4 and pCHIKV-CP-EGFP using the Turbofect (Fermentas) transfection reagent according to manufacturer’s instructions. Cell lysates were prepared 48 h after transfection in 1.5 ml of RIPA buffer. Aliquots of 500 μg total protein were incubated with 1 μg of mouse anti-FLAG antibodies or control IgG for 1 h at 4°C. A 50% suspension of protein AG agarose (Santa Cruz Biotechnology) was added to the lysate and incubated for 2 h at 4°C while shaking. The agarose beads were collected by centrifugation and washed by three with RIPA buffer. After the final wash, the supernatant was discarded and the beads were resuspended in Laemmli buffer and boiled for 10 min. The immunoprecipitated material was subjected to 12% SDS-PAGE and analyzed by Western blot using an anti-GFP antibody.

HEK293 cells were maintained in DMEM supplemented with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin. HEK293 seeded in 12-well plates at a density of 2 × 10⁵ per well were transfected with Turbofect (Fermantas). For the experiments with LMB, cells were treated with LMB at a concentration of 15 ng/ml for 6 h. To detect EGFP-tagged proteins by fluorescence microscopy, the cells were fixed for 10 min in PBS with 4% formaldehyde, washed twice in PBS, and mounted on slides with an anti-fade mounting media (Dako). The slides were analyzed on a Leica confocal microscope and cells with moderate GFP expression were selected for analysis.

NIH Image J v1.30 software was used to quantify the nuclear localization of GFP fusion proteins as described before [52]. All confocal images were quantified at settings where the intensity of GFP fluorescence ranged up to 250 pixel values. To examine the nuclear localization of the GFP fusion proteins, a square with an area of 225 pixels was used to measure the mean intensities of six different areas in the nucleus and the cytoplasm of a representative cell from each of three transfections. Thus, eight representative cells analyzed yielded a total of 48 areas from each, the nucleus and the cytoplasm. Relative nuclear localization is reported as the ratio of the mean intensity of GFP fluorescence in the nucleus and cytoplasm (nuclear/cytoplasmic ratio, N/C ratio).

Various Old World alphavirus CP amino acid sequences were aligned with the ClustalW2 - Multiple Sequence Alignment tool http://​www.​ebi.​ac.​uk/​Tools/​msa/​clustalw2/​. The structures of CHIKV CP and Karα4 were modeled using the automated homology modelling server 3D-JIGSAW [53] which used PDB ID 1VCP and 1BK6, respectively, as template. Figures were generated using PyMOL: http://​www.​pymol.​org. To perform in silico protein-protein docking, the NES region of CHIKV CP was modeled using the loop building tool of the SPDBV [54]. The hydrophobic conserved CHIKV CP NES residues were kept positioned in the NES binding site and superimposed to the respective NES residues as in chain A of PDB ID 3NBY followed by energy minimization. The docking was further optimized by a flexible docking method using flex dock server [55]. The energetically most favorable structure was used for the CHIKV CP NES and CRM1 interaction analysis and binding energy was calculated using firedock server [56, 57].