Xpo7 is a broad-spectrum exportin and a nuclear import receptor

Abstract

Exportins bind cargo molecules in a RanGTP-dependent manner inside nuclei and transport them through nuclear pores to the cytoplasm. CRM1/Xpo1 is the best-characterized exportin because specific inhibitors such as leptomycin B allow straightforward cargo validations in vivo. The analysis of other exportins lagged far behind, foremost because no such inhibitors had been available for them. In this study, we explored the cargo spectrum of exportin 7/Xpo7 in depth and identified not only ∼200 potential export cargoes but also, surprisingly, ∼30 nuclear import substrates. Moreover, we developed anti-Xpo7 nanobodies that acutely block Xpo7 function when transfected into cultured cells. The inhibition is pathway specific, mislocalizes export cargoes of Xpo7 to the nucleus and import substrates to the cytoplasm, and allowed validation of numerous tested cargo candidates. This establishes Xpo7 as a broad-spectrum bidirectional transporter and paves the way for a much deeper analysis of exportin and importin function in the future.

Graphical abstract

Figure 1.

Identification of novel Xpo7 binders. (A) Xpo7 tagged with the ED domains of protein A was immobilized on anti–protein A beads and incubated with 2 ml hypotonic HeLa extract (Abmayr et al., 2006) in the absence or presence of 5 µM RanGTP (Q69L ΔC terminus mutant). 1/2,000 of the starting extracts and 1/10 of the bound fractions were analyzed by SDS-PAGE followed by Coomassie staining. Indicated bands were identified by MS. Tubulin dimer^# denotes tubulin α 4A and tubulin β 4B as the predominant forms. (B) Xpo7 affinity chromatography was performed as in A, but an extract from mouse spleen was used as starting material. Mw, molecular weight. (C) Starting extract (–Ran), Xpo7–RanGTP, and Xpo7 without Ran samples in B were analyzed by MS. The Venn diagram represents the number of identified unique proteins.

Figure 2.

Validation of cargo candidates as true Xpo7 binders. (A) H14-ZZ-bdNEDD8–tagged cargo candidates (human proteins) were immobilized on anti–protein A beads and incubated with a cytoplasmic HeLa extract in the absence or presence of 3 µM RanGTP. After washing off unbound material, the immobilized candidates and bound proteins were eluted by bdNEDP1 protease cleavage. 1/250 of the starting extracts and 1/10 of the eluates were analyzed by SDS-PAGE followed by immunoblotting with an anti-hsXpo7 antibody recognizing a C-terminal epitope (Mingot et al., 2004). (B) Binding assays were performed as in A, but candidates (from mouse) were incubated with purified recombinant Xpo7 in the absence or presence of RanGTP, and the eluates were analyzed by SDS-PAGE and Coomassie staining. Anp32e* corresponds with Anp32e isoform 3 (E9Q5H9). Mw, molecular weight.

Figure 3.

Xpo7 binds to and mediates nuclear export of tubulin. (A) A cytoplasmic HeLa extract was depleted of endogenous NTRs by the phenyl-Sepharose method (Ribbeck and Görlich, 2002). The sample was split, and single importins or exportins were added to the extract for formation of export complexes with ZZ-tagged RanGTP immobilized on IgG-Sepharose. Bound fractions were analyzed by immunoblotting with a monoclonal antibody recognizing α-tubulin. The panel also shows immunoblots against known cargoes of NTRs other than Xpo7. The bound fractions each correspond with 20× the amount of loaded starting extract. Note that tubulin was enriched only in the sample containing Xpo7. The signal observed with other NTRs resembles background binding to the matrix as seen for the sample without addition of any transport receptor. (B) [³⁵S]-labeled GST (nuclear injection marker) and human tubulin α6 were injected into the nuclei of Xenopus oocytes. Anti-Xpo7 antibodies (6 mg/ml) or 1× PBS were coinjected. Oocytes were dissected 1 or 3 h postinjection, and the nucleocytoplasmic distribution of the labeled proteins was analyzed by SDS-PAGE followed by autoradiography. The antibodies blocked tubulin export. (C) Radiolabeled tubulin and GST were coinjected with human Xpo7 or CRM1/Xpo1 into the nuclei of Xenopus oocytes. 4 h postinjection, oocytes were analyzed as in B. Note that export of tubulin is strongly stimulated by coinjection of Xpo7.

Figure 4.

Validation of export cargo candidates. GFP-fused candidate proteins from mouse were transiently expressed in HeLa cells, and their subcellular localization was recorded in live cells by confocal fluorescence microscopy. The effect of the nanobodies recognizing MBP, Xpo4, or Xpo7 (D18) was monitored by cotransfection with a vector encoding NES-RFP to stain the cytoplasmic compartment. In a separate experiment, anti-Xpo7 nanobody D11 was tested. Bar, 20 µm. The predominant cytoplasmic localization of Ttc39c (Q8VE09), Smyd3 (Q9CWR2), MetAP1 (Q8BP48), and Sestrin-2 (P58043) was disrupted only by anti-Xpo7 nanobodies, which led to nuclear accumulation. This suggests that these proteins can leak into nuclei and that they are kept cytoplasmic at steady state by Xpo7-dependent export. A milder disruption was observed for Sufu (Q9Z0P7), where block of Xpo7 increased the relative nuclear GFP signal. UniProt entry names are listed in parentheses. On the right, mean ratios of nuclear/cytoplasmic (N:C) GFP concentrations are plotted, with anti-MBP and anti-Xpo7 nanobody results averaged (n = 20–40). Error bars represent SD. Statistical significance (***, p < 10⁻⁶) was assessed using an unpaired two-sided t test. Means ± SD of each sample are listed in Table S1.

Figure 5.

Validation of import cargo candidates. The analysis of potential import cargo candidates was performed as described in Fig. 4 and included the human proteins Hat1 (O14929), Nampt (P43490), Smug1 (Q53HV7), Hmbs (P08397), and Hdac8 (Q9BY41). Note that the anti-Xpo7 nanobodies shift the here-validated import cargoes to a more cytoplasmic localization. Error bars represent SD. ***, p < 10⁻⁶. Bar, 20 µm.