Exportin-5 is a Kapβ family member. (A) Amino acid sequences of exportin-5, Crm1, and the S. cerevisiae Kapβ, Msn5p, were aligned using CLUSTAL. Identical residues are shown in black; closely related residues are shown in gray. (B) Unrooted phylogenetic tree of exportin-5 and the 14 S. cerevisiae Kapβs generated using the ProtPars program in the Phylip package. A total of 1,000 boot-strapped replica datasets were run from an alignment of the S. cerevisiae Kap and exportin-5 sequences that was generated using Clustal W. Each dataset was analyzed 13 times with the sequence input order randomized each time. The tree represents the consensus of all trees produced. Values on the branches refer to the proportion of times a given branch was recorded among all trees. Values >90% indicate strong support, values 90–60% indicate moderate to weak support, and values <60% indicate no support for a given branching pattern. (C) Northern blotting of mRNAs from human tissues was performed with a probe generated using the exportin-5 cDNA as a template. The same blot was hybridized with a β-actin probe as a loading control. The exportin-5 sequence data are available from GenBank/EMBL/DDBJ under accession no. AF298880.

Exportin-5 interacts with ILF3 in a RanGTP-dependent manner. (A) Schematic representation of human ILF3. Important elements found in this protein include the following: a putative bipartite NLS (striped box), two dsRBDs (shaded boxes), and an RGG RNA-binding motif (dotted box). The fragment that bound exportin-5 in the dihybrid screen is underlined. (B) Conjugation assays were performed as in the legend to Fig. 2 using VP16-ILF3(533–640) as the prey. Diploid yeast were replica plated onto selective medium (L⁻W⁻H⁻). (C) Equal amounts of GST or GST-ILF3(533–640) were immobilized on GSH beads and incubated with 200 nM recombinant exportin-5–His₆ ±3 μM RanQ69L. An aliquot of the supernatant (1/10 the total) was analyzed to ensure the beads were incubated with equal concentrations of exportin-5–His₆. Bound and unbound fractions were analyzed by SDS-PAGE and immunoblotting with anti-His₆, anti-Ran, and anti-GST antibodies. All lanes were exposed equally to film. (D) HEK293 cells were transfected with full-length human pKH3-ILF3. Cells were lysed after 48 h. HA-tagged proteins were immunoprecipitated with 12CA5, immobilized on protein A–Sepharose, and then incubated with 200 nM exportin-5–His₆ ±3 μM RanQ69L. HEK293 cells transfected with empty vector were used as a negative control. Bound proteins were analyzed by SDS-PAGE and immunoblotting with anti-His₆ and anti-HA antibodies.

Exportin-5 binds the dsRBDs of multiple proteins, but binding is blocked by dsRNA. (A) Equal amounts of GST-ILF3 dsRBD2, GST-Spnr dsRBD2, GST-staufen dsRBD3, or GST-PKR dsRBD1 were immobilized on GSH beads. Beads were incubated with 200 nM exportin-5–His₆ ±3 μM RanQ69L. Proteins were analyzed and detected as described in the legend to Fig. 3. (B) Amino acid sequences of ILF3 dsRBD2, Spnr dsRBD2, staufen dsRBD3, and PKR dsRBD1 were aligned using CLUSTAL. Residues conserved in all four dsRBDs are in black. Closely related residues found in at least three of the four sequences are in gray. Regions of secondary structure are denoted as boxes above the sequences. Asterisks denote conserved residues in the three regions that make essential contacts with the RNA. Residues essential for the correct packing of the α-helices against the β sheet are marked with diamonds. (C) GST-ILF3(404–592) and GST-ILF3(404–592)mut (1 μg each) were separated by SDS-PAGE and stained using Coomassie blue. GST-ILF3(404–592)mut is a construct that is deficient in dsRNA binding. A duplicate gel was transferred to nitrocellulose. The immobilized proteins were incubated with an α-³²P–labeled dsRNA probe. After washing to remove excess probe, the filter was exposed to film for ∼18 h before developing. (D) GST-ILF3(404–592) was immobilized on GSH-agarose beads and either ssRNA or dsRNA was added. After RNA binding, the beads were incubated with 100 nM exportin-5 ±3 μM RanQ69L. Proteins were analyzed and detected as described in the legend to Fig. 3.

Exportin-5 mediates export of ILF3 in digitonin- permeabilized cells. (A) HeLa cells were cotransfected with pEGFP-ILF3 and pKRed-RCC1. pKRed-RCC1 was used as a nuclear marker for transfected cells. The cells were permeabilized with digitonin, and export assays were performed in buffer alone or with buffer plus exportin-5–His₆ (100 μg/ml), exportin-5–His₆ (100 μg/ml) pretreated with LMB (1 μM), exportin-5–His₆ (100 μg/ml) plus WGA (200 μg/ml), or Crm1-His₆ (100 μg/ml). (B) Mean nuclear fluorescence values for GFP-ILF3 were obtained using Openlab (Improvision). All values were corrected for background fluorescence levels and normalized to the mean nuclear fluorescence levels of permeabilized buffer-treated cells. Each data point represents the mean nuclear fluorescence obtained from 30 randomly chosen cells. Error is expressed as ±1 SD from the mean. Bar, 10 μm.

Exportin-5 interacts with the RanGTPase and nucleoporins. (A) HF7c (MATa) yeast that express the indicated bait proteins as GAL4 DBD fusions were mated with the W303 (MATα) strain expressing the VP16 transactivation domain either alone (VP16) or as a fusion with Ran (VP16-Ran). Diploid yeast were selected on Leu⁻Trp⁻ plates and replica plated onto Leu⁻Trp⁻His⁻ ±3-aminotriazole (3AT) as indicated. (B) Exportin-5–His₆ was incubated with either GST-RanGDP or GST-RanGTP bound to GSH beads. Bound proteins were separated by SDS-PAGE and detected by Coomassie staining. (C) The apparent dissociation constant of the RanGTP–exportin-5 complex was estimated using a RanGAP protection assay. Ran[γ-³²P]GTP was preincubated for 15 min with the indicated concentrations of exportin-5–His₆. The GTPase reaction was started by addition of RanGAP. After 3 min, the proteins were filter bound and washed. Hydrolysis of GTP was determined by the amount of Ran[γ-³²P]GTP bound. Assays were in duplicate and are representative of two independent experiments. Error is expressed as the differences from the means. (D) HA–importin β and HA–exportin-5 were immunoprecipitated from HEK293 cell lysates to which RanQ69L was added as indicated. Untransfected HEK293 cells were used as a control for nonspecific nucleoporin binding (Mock IP). Lysate was prepared by solubilizing HEK293 cells with boiling sample buffer. Coimmunoprecipitated nucleoporins were detected with mAb RL1 (Snow et al., 1987). RL1 efficiently detected RanBP2, Can/Nup214, and Nup153 but not other nucleoporins in HEK293 lysate. HA–importin β and HA–exportin-5 were detected using mAb 12CA5. A protein not characterized previously as reacting with the RL1 mAb bound exportin-5 in the absence of RanQ69L and is denoted with an asterisk. (E) Heterotypic cell fusions were performed between BHK21 cells, transiently transfected with pKmyc–exportin-5, and a stably transfected HeLa cell line expressing GFP–streptavidin–NLS (GSN2). After fusion with polyethylene glycol, the cells were fixed and stained for myc-tagged proteins with 9E10 and a Texas red–conjugated secondary antibody. Nuclei were visualized using DAPI. The arrows denote the GSN2 cells in the fusion. Bar, 10 μm.

Exportin-5 binds ILF3 specifically and in an LMB- insensitive manner. (A) Equal amounts of GST-ILF3(533–640) were immobilized on GSH beads and incubated with 200 nM exportin-5–His₆, 500 nM Crm1-His₆, or 200 nM importin β–His₆ ±3 μM RanQ69L as indicated. Proteins were analyzed and detected as in the legend to Fig. 3. (B) Equal amounts of GST-PKI(36–50) or GST-ILF3(533–640) were immobilized on GSH beads. The beads were incubated with 200 nM Crm1-His₆ or exportin-5–His₆ ±3 μM RanQ69L. Where indicated, Crm1-His₆ and exportin-5–His₆ were incubated with 500 nM LMB at room temperature for 15 min before their addition to the binding assay. Proteins were analyzed and detected as described in the legend to Fig. 3. Exportin-5 is unable to bind GST-PKI(36–50) (unpublished data).

Nuclear export of ILF3 is dependent on exportin-5 in intact cells. (A) Results of coupled reverse transcription and PCRs on BHK21 (B) and HeLa (H) cell RNA. All reactions contained 1 μg of RNA and either exportin-5 or glyceraldehyde-3-phosphate dehydrogenase gene-specific primers. A fraction of each reaction (1/20) was visualized on an ethidium bromide-stained agarose gel. GGIBF3(533–640) (18 μM) was injected into the nuclei of (B) HeLa cells or (C) BHK21 cells using fluorescent dextran (TRITC-dextran, 1 mg/ml) as an injection marker. Where indicated, GGIBF3 was coinjected with either 2 μM exportin-5–His₆ or 2 μM Crm1-His₆. All injected cells were incubated 30 min before fixation and visualization of the injected import substrate. Bars, 10 μm.

Exportin-5 is a general export receptor for dsRBP. Exportin-5 and RanGTP associate with dsRBP in the nucleus. This export complex translocates through the NPC to the cytoplasm. RanGTP hydrolysis and binding to 3′ untranslated region elements on the target mRNA promotes release and retention of the dsRBP in the cytoplasm. RNA binding to the dsRBDs may inhibit import by preventing the interaction of an adjacent bipartite NLS with its nuclear import receptor (Imp). Release of the mRNA transcript due to RNA degradation would then permit import of the dsRBP. The shaded region of the dsRBP represents the dsRBD. A putative bipartite NLS is adjacent to the dsRBDs (see Fig. 3 A).