Report

Evidence for a Role of CRM1 in Signal-Mediated Nuclear Protein Export

See allHide authors and affiliations

Science  03 Oct 1997:
Vol. 278, Issue 5335, pp. 141-144
DOI: 10.1126/science.278.5335.141

Abstract

Chromosome maintenance region 1 (CRM1), a protein that shares sequence similarities with the karyopherin β family of proteins involved in nuclear import pathway, was shown to form a complex with the leucine-rich nuclear export signal (NES). This interaction was inhibited by leptomycin B, a drug that prevents the function of the CRM1 protein in yeast. To analyze the role of the CRM1-NES interaction in nuclear export, a transport assay based on semipermeabilized cells was developed. In this system, which reconstituted NES-, cytosol-, and energy-dependent nuclear export, leptomycin B specifically blocked export of NES-containing proteins. Thus, the CRM1 protein could act as a NES receptor involved in nuclear protein export.

Bidirectional transport across the nuclear envelope occurs through nuclear pore complexes. This process requires specific sequences found within transport substrates, soluble transport proteins, and nucleoporins. Thus, the import of nuclear proteins is governed by different nuclear localization sequences (NLS) that are presumably recognized by distinct receptors (karyopherins, importins, and transportins) that mediate the docking of the transport substrate at the cytoplasmic face of the nuclear pore (1) and by other soluble factors, including the small guanosine triphosphatase Ran and p10, that are responsible for the translocation step across the nuclear pore complex (2). Recent studies have shown that soluble factors involved in the docking step of nuclear import and in Ran-binding ability share amino acid sequences and structural homology domains. The CRM1 protein shares sequence homology in its NH2-terminal region with the karyopherin β family, as well as with the Ran–guanosine triphosphate (GTP)–binding domain of the Ran-GTP–binding protein family. This protein, which is encoded by an essential gene in yeast, is located at the nuclear pore complex as well as in the nucleoplasm (3,4). Thus, we examine whether CRM1 could be involved in nuclear export.

Although most nuclear proteins are potential shuttling proteins (5), amino acid sequences responsible for highly efficient nuclear export (NES) have recently been identified in an increasing number of proteins, in particular, the human immunodeficiency virus–type 1 (HIV-1) Rev protein, the protein kinase A inhibitor IkappaBα (IκBα), and the heterogeneous nuclear ribonucleoproteins (hnRNPs) A1 and K (6). With the exception of the hnRNPs, NES is a leucine-rich sequence in which leucine residues are critical for targeting proteins out of the nucleus. Molecular mechanisms governing NES-dependent nuclear protein export are less well documented than those of the nuclear protein import pathway. However, the existence of NES as well as its ability to saturate NES-dependent export strongly suggest the involvement of specific NES receptors in this process.

To analyze the role of CRM1 in nuclear protein export, we first tested the ability of human CRM1 protein to bind a leucine-rich sequence (NES) (7). For this purpose, interaction of human CRM1 with wild-type IκBα or IκBα-L234, a nuclear export mutant of IκBα in which leucine residues of NES have been replaced by alanine (6, 8), was analyzed. In vitro–translated human MYC-tagged CRM1 was mixed with in vitro–translated SV5-tagged wild-type IκBα or IκBα-L234 before being processed for immunoprecipitation with an antibody to the MYC tag (anti–MYC tag) or an anti–SV5 tag. CRM1 protein and wild-type IκBα coprecipitated with both antibodies. In contrast, no coprecipitation was observed between CRM1 and IκBα-L234 (Fig.1A), which suggests that CRM1 interacts with IκBα NES. To confirm this result, we conjugated biotinylated bovine serum albumin (BSA) to IκBα NES (CIQQQLGQLTLENL) or mutated NES (CIQQQAGQATAENA) (9) peptides and coupled it to streptavidin-agarose. In vitro–translated human CRM1 bound to a NES affinity column. However, no binding of CRM1 was observed when leucine residues of NES were substituted by alanines (Fig. 1B) (8) or when NES was replaced by NLS (10). Under the same experimental conditions, no specific binding was observed between NES and in vitro–translated human RIP (10). CRM1-NES interaction was prevented by the addition of an excess of NES peptide but was not affected by the mutated NES (Fig. 1B), which confirms that CRM1 interacts specifically with NES. However, the possibility that the specific interaction between CRM1 and NES could be mediated by a factor provided by reticulocyte lysate cannot be formally excluded.

Figure 1

CRM1 forms a leptomycin B–sensitive complex with NES. (A) Interaction between CRM1 and IκBα. 35S-Methionine– and35S-cysteine–labeled human MYC-tagged CRM1 translated in vitro was mixed with 35S-methionine– and 35S-cysteine–labeled SV5-tagged wild-type IκBα or IκBα-L234 translated in vitro before being processed for immunoprecipitation with an anti–MYC tag or an anti–SV5 tag (8). Immunoprecipitates were analyzed by 7% SDS-PAGE and autoradiography. (B) Interaction between CRM1 and NES.35S-Methionine– and 35S-cysteine–labeled CRM1 translated in vitro was incubated with streptavidin-agarose beads bound to biotinylated BSA-NES (biot-BSA-NES) or mutated NES (biot-BSA-NESmut) conjugates (8). The binding was performed with or without NES or mutated NES (NESmut) peptides (each 2 mg/ml). Bound (B) and unbound (U) fractions were collected and analyzed by 7% SDS-PAGE and autoradiography. The Bioprint acquisition system and Bioprofil program were used to quantify the autoradiograms. Values were obtained from five independent experiments. (C)35S-Methionine– and 35S-cysteine–labeled CRM1 translated in vitro was incubated with streptavidin-agarose beads bound to biotinylated BSA-NES conjugate (30 min at room temperature in PBS) with increasing concentrations of leptomycin B. Bound and unbound fractions were collected and analyzed by 7% SDS-PAGE and autoradiography.

The CRM1 homolog in Schizosaccharomyces pombe is the target of leptomycin B, an antifungal antibiotic that induces cell cycle arrest at the G1 and G2 phases in both mammalian and fission yeast cells (11). To test whether the leptomycin B–induced inhibition of CRM1 function was related to the ability of CRM1 to bind NES, we monitored the effect of the drug on CRM1-NES interaction. Addition of leptomycin B to 200 nM concentration completely blocked the formation of CRM1-NES complexes (Fig. 1C). We controlled the leptomycin B so that it did not bind directly to NES (10). Thus, CRM1 protein bound NES specifically in a leptomycin B–sensitive manner.

To analyze the role of the CRM1-NES interaction in the export of nuclear proteins, we developed an assay that reconstitutes nuclear export in vitro. HeLa cells were transiently transfected with cDNAs encoding fusion proteins consisting of MYC-tagged pyruvate kinase (PK), wild-type or mutated IκBα NES, and SV40 large T antigen NLS to direct the resulting proteins to the nucleus (NLS-PK-NES and NLS-PK-NESmut, respectively) (12). Eighteen hours after transfection, cells were treated with digitonin to permeabilize the plasma membrane and remove cytosolic components without affecting the integrity of the nuclear envelope (13). Exports of NLS-PK-NES and NLS-PK-NESmut from permeabilized cell nuclei were analyzed under different incubation conditions by indirect immunofluorescence with an anti–MYC tag (Fig.2A) (14). Incubation of permeabilized cells with buffer in the presence or absence of adenosine triphosphate (ATP) did not allow nuclear export of both proteins. In contrast, addition of Xenopus laevis egg extracts (cytosol) and ATP for 30 min at 23°C led to the disappearance of NLS-PK-NES from the nucleus, whereas the nuclear content of NLS-PK-NESmut was not affected. Treatment of the cytosol with apyrase abolished the disappearance of NLS-PK-NES from the nucleus. Because only a few amino acids within NES are different between both proteins, the disappearance of NLS-PK-NES from the nucleus treated with cytosol and ATP likely corresponded to an active nuclear export of this protein rather than an ATP-dependent hydrolysis that should also affect the mutated protein. To quantify results obtained by indirect immunofluorescence, we analyzed proteins from permeabilized cells treated in the different conditions by SDS–polyacrylamide gel electrophoresis (PAGE) and protein immunoblotting with an anti–MYC tag and an anti–hnRNP C (Fig. 2, B and C) (15). The hnRNP C protein was used as an internal control of a nonexported protein in both immunofluorescence and protein immunoblotting analyses (10,16). Neither buffer alone, buffer and ATP, or cytosol treated with apyrase affected the nuclear content of NLS-PK-NES or NLS-PK- NESmut. However, 75% of NLS-PK-NES was exported when both cytosol and ATP were added to permeabilized cells, whereas only 15% of NLS-PK-NESmut was transported under the same condition. In this in vitro assay, the replacement of total cytosol by the recombinant proteins required for import (karyopherins, Ran/TC4, and p10) promoted the nuclear import of a karyophilic substrate (BSA-NLS) but did not induce the nuclear export of NLS-PK-NES, which indicates that an essential component for nuclear export was provided by the total extracts (10). Thus, this in vitro assay allowed the reconstitution of a NES-, cytosol-, and energy-dependent nuclear export.

Figure 2

In vitro NES-dependent protein nuclear export assay. HeLa cells were transiently transfected with cDNAs encoding NLS-PK-NES or NLS-PK-NESmut (12). Eighteen hours after transfection, cells were permeabilized with digitonin (55 μg/ml) in transport buffer and incubated for 30 min at 23°C with BSA (20 mg/ml) in transport buffer in the absence (buffer) or presence of ATP (buffer + ATP) or with 45% X. laevis egg extracts in transport buffer in the presence of ATP (extracts + ATP) or in the absence of ATP [addition of apyrase (20 U/ml); extracts + apyrase] (13). After incubation under different conditions, cells were processed for immunofluorescence (A) or for protein immunoblotting (B) (14, 16). In both cases, NLS-PK-NES and NLS-PKNESmut were detected with a monoclonal anti–MYC tag. The nuclear DNA was visualized by costaining with 4′,6′-diamidino-2-phenylindole (DAPI). Photographs corresponding to the different conditions were taken with the same setting parameters. hnRNP C was used as an internal control of a nonexported protein in the same samples. (C) Quantitation of protein immunoblots from four independent experiments was performed with the Bioprint acquisition system and Bioprofil program. Hatched and white columns represent results obtained for NLS-PK-NES and NLS-PK-NESmut, respectively. Values correspond to the ratio between NLS-PK-NES and hnRNP C or NLS-PK-NESmut and hnRNP C contents measured on the same blot and normalized to the ratio measured after incubation in transport buffer without ATP (considered as 100%).

We next analyzed the role of CRM1-NES interaction in NES-dependent protein export by adding leptomycin B (Fig. 1). Cells producing NLS-PK-NES were permeabilized and treated with cytosol and ATP in the presence or absence of 200 nM leptomycin B (Fig.3, A and B). Nuclear export of NLS-PK-NES was analyzed by both indirect immunofluorescence and protein immunoblotting. Eighty percent of NLS-PK-NES was exported out of the nucleus with the addition of extracts and ATP, whereas, in the presence of leptomycin B, 90% of the protein stayed in the nucleus. No detectable effect of leptomycin B was observed when the drug was used at a 20 nM concentration (10). Thus, leptomycin B, which inhibits the interaction of CRM1 with NES, was able to block NES-dependent protein export in a similar concentration range. Leptomycin B at a concentration of 2 nM inhibits HIV-1 replication in primary human monocytes or in transient transfection in fibroblasts by preventing Rev function (17), which indicates that CRM1 also interacts with HIV-1 Rev protein. Because the drug could be accumulated by living cells, it may explain why lower concentrations are required for nuclear export inhibition in vivo than in vitro.

Figure 3

Leptomycin B inhibits NES-dependent protein nuclear export. In vitro nuclear export of NLS-PK-NES was performed as in Fig. 2 in BSA (20 mg/ml) (buffer), in 45% X. laevis egg extracts supplemented with ATP (extracts + ATP), or in 45% X. laevis egg extracts supplemented with ATP and 200 nM leptomycin B (extracts + ATP + leptomycin B). Nuclear export of NLS-PK-NES was analyzed by indirect immunofluorescence (A) or protein immunoblot (B) with an anti–MYC tag or an anti–hnRNP C.

CRM1 thus appears to form a specific complex with NES that is necessary for NES-mediated nuclear protein export. These data suggest that CRM1 could act as an NES receptor involved in nuclear export. Phe-Gly (FG) repeat–containing nucleoporins or related proteins such as RIP have been described as participating in NES-mediated Rev export in both yeast and higher eukaryotic cells. However, direct binding of recombinant Rev to recombinant FG repeats produced in Escherichia coli was not detectable in vitro (18). The interaction of CRM1 with NES may target the NES-containing substrates to the FG nucleoporins more efficiently. Moreover, CRM1 shares a sequence motif related to the Ran-GTP–binding site of Ran-GTP–binding proteins (4), and both p10 and Ran-GTP, but not Ran-dependent GTP hydrolysis, appear to be required in NES-mediated protein export (19). By analogy with the nuclear import process, Ran or a Ran-binding protein may also regulate the interaction of CRM1-NES–containing protein complexes with the nuclear pore complex before translocation out of the nucleus.

  • * To whom correspondence should be addressed. E-mail: dargemon{at}curie.fr

REFERENCES AND NOTES

View Abstract

Navigate This Article