Nucleocytoplasmic transfer of the NF2 tumor suppressor protein merlin is regulated by exon 2 and a CRM1-dependent nuclear export signal in exon 15

Michael Kressel 1 Beatrice Schmucker 0 0 Department of Otorhinolaryngology/Head and Neck Surgery, Friedrich-Alexander University Medical Center , 91054 Erlangen , Germany 1 Institute of Anatomy I, Friedrich-Alexander University of Erlangen , 91054 Erlangen , Germany The neurofibromatosis 2 protein merlin is a classical tumor suppressor protein. Germline mutations predispose to the development of schwannomas, meningiomas and ependymomas. Merlin has been implicated in cellular migration and adhesion. This function is reflected in its subcellular localization at the plasma membrane and known interacting partners. Merlin has been regarded as an exception in not exerting a functional role within the nucleus as other tumor suppressors do. Here, we show that detection of wild-type protein in the nucleus is a rare event. However, splicing out of exon 2 leads to unrestricted entry into the nucleus. Skipping of adjacent exon 3 has no comparable effect ruling out an unspecific effect due to misfolding of the 4.1/JEF domain. Exon 2 functions as a cytoplasmic retention factor as it is able to confer sole cytoplasmic localization to a GFP fusion protein. Nuclear entry of merlin is thus regulated by alternative splicing within the 4.1/JEF domain and analogous to band 4.1 protein. Merlin's ability to enter the nucleus is complemented by a full nuclear-cytoplasmic shuttle protein with a functional Rev-type nuclear export sequence (NES) within exon 15 that facilitates export via the CRM1/exportin pathway. Deletion of this NES or treatment with the CRM1-specific inhibitor leptomycin Bleads to overall nuclear accumulation of merlin isoforms missing exon 2. A cellular function different to the wild-type protein is implied for naturally occurring splice variants lacking exon 2. A putative effect of merlin as a transcriptional regulator and identification of nuclear binding partners remains to be elucidated. - Due to its high sequence conservation during evolution and developmental arrest at an early stage in homozygous mutant embryos, the neurofibromatosis type 2 (NF2) tumor suppressor protein named merlin or schwannomin can be regarded as a key regulatory protein for development of the organism. Insights into its biological function have been gained by experimental studies and in the model organisms mouse and Drosophila. In humans, germline mutations of the NF2 gene lead to development of a tumor condition characterized by schwannomas, meningiomas and ependymomas (1). The N-terminal half of the NF2 protein comprises a conserved 38 kDa domain, which was originally identified in the red blood cell protein band 4.1 (2). Subsequently this domain was found to be present in several different protein families and thus has been dubbed the 4.1/JEF domain (for band 4.1, Janus kinase family, ERM protein family, Focal adhesion kinase family) (3). Based on amino acid sequence homology of its 4.1/JEF domain the NF2 protein is most closely related to members of the ERM family of proteins consisting of ezrin, radixin and moesin. In addition to the N-terminal 4.1/JEF domain, the NF2 protein shares the same secondary structure with the ERM proteins (4). Furthermore, activity of the NF2 protein is conformationally regulated in the same way as the ERM proteins (5). In its dormant form the molecule undergoes an intramolecular selfassociation between the N-terminal 4.1/JEF domain and the C-terminus (6,7). Binding affinity of the C-terminus to the 4.1/JEF domain fosters the formation of intermolecular interactions leading to both homodimerization between NF2 proteins and heterodimerization between NF2 protein and ERM family members (8). Activation of the small GTPase Rac by growth promoting stimuli has recently been shown to lead to NF2 phosphorylation at serine 518 by the Rac effector kinase PAK2, which is associated with unfolding of the protein and loss of intramolecular self-association (9,10). The growth suppressive potential of the NF2 tumor suppressor protein has been correlated with the hypophosphorylated wild-type protein undergoing intramolecular head to tail binding (11). This is the predominant form of NF2 found in culture under conditions of growth arrest and contact inhibition, which is localized at the interface between the plasma membrane and the cytoskeleton (12,13,14). In this realm the NF2 protein interacts with several binding partners comprising transmembraneous membrane proteins, cytoskeletal elements and endosomal proteins: CD44, b1-integrin, bIIspectrin, actin, syntenin, SCHIP-1, NHE-RF, Rho-GDI and hepatocyte growth factor-regulated tyrosine kinase substrate (HRS) (2,13,1523). Overexpression of merlin is able to inhibit Rac mediated signaling, which is believed to be essential for the metastasis and tumor suppressive potential of NF2 (9). By yeast two-hybrid analysis HRS has been identified as an NF2 binding protein (23). HRS stimulates formation of early endosomes and sorting of activated tyrosine kinase receptors into multivesicular bodies for subsequent degradation in lysosomes. In Drosophila mutant HRS is associated with inability to limit signaling induced by the activated tyrosine kinase receptors, i.e. epidermal growth factor receptor and torso (24). HRS furthermore binds to STAM, a protein implicated in the suppression of cytokine induced cell growth (25,26). Therefore, in addition to its inhibitory effect on activated Rac NF2 protein binding to HRS provides an alternative pathway explaining its tumor suppressive effect. The NF2 gene is subject to alternative splicing (2732). The two most commonly occurring isoforms differ in the C-terminus of the protein. Isoform I (wild-type protein) correlates with the longest and predominant splice form and consists of exons 117 with splicing out of exon 16. Isoform II retains exon 16 resulting in a frameshift and premature stop and consequently in a shortened protein with an altered C-terminus. Recently novel full-length NF2 isoforms have been described with splicing out of more exons, which are targeted subcellularly to cytoplasmic granules or the cell nucleus (32). Nuclear localization of NF2 as indicated by the novel isoforms is, however, interesting, as it might implicate an additional functional role in this compartment. Matching the exon composition of the isoforms with their subcellular localization we hypothesized that splicing out of exons 2 and/or 3 is a decisive step for entry into the nuclear compartment. This is reminiscent of the prototypical member of the 4.1 superfamily the erythrocyte band 4.1 protein. For the band 4.1 it has been reported that alternative splicing events regulate nuclear localization (33,34). Nuclear occurring band 4.1 protein is known to be a component of nuclear speckled domains and to associate with pre-mRNA splicing factors (35,36). One of the splicing events contributing to nuclear localization of band 4.1 is skipping of exon 5 (34). Stimulated by potential splicing events in the 4.1/JEF domain (...truncated)