Karyopherin α7 (KPNA7), a divergent member of the importin α family of nuclear import receptors
© Kelley et al; licensee BioMed Central Ltd. 2010
Received: 15 March 2010
Accepted: 11 August 2010
Published: 11 August 2010
Skip to main content
© Kelley et al; licensee BioMed Central Ltd. 2010
Received: 15 March 2010
Accepted: 11 August 2010
Published: 11 August 2010
Classical nuclear localization signal (NLS) dependent nuclear import is carried out by a heterodimer of importin α and importin β. NLS cargo is recognized by importin α, which is bound by importin β. Importin β mediates translocation of the complex through the central channel of the nuclear pore, and upon reaching the nucleus, RanGTP binding to importin β triggers disassembly of the complex. To date, six importin α family members, encoded by separate genes, have been described in humans.
We sequenced and characterized a seventh member of the importin α family of transport factors, karyopherin α 7 (KPNA7), which is most closely related to KPNA2. The domain of KPNA7 that binds Importin β (IBB) is divergent, and shows stronger binding to importin β than the IBB domains from of other importin α family members. With regard to NLS recognition, KPNA7 binds to the retinoblastoma (RB) NLS to a similar degree as KPNA2, but it fails to bind the SV40-NLS and the human nucleoplasmin (NPM) NLS. KPNA7 shows a predominantly nuclear distribution under steady state conditions, which contrasts with KPNA2 which is primarily cytoplasmic.
KPNA7 is a novel importin α family member in humans that belongs to the importin α2 subfamily. KPNA7 shows different subcellular localization and NLS binding characteristics compared to other members of the importin α family. These properties suggest that KPNA7 could be specialized for interactions with select NLS-containing proteins, potentially impacting developmental regulation.
Eukaryotic cells are defined by the separation of DNA from the rest of the cell by the nuclear envelope, a double bilayer made selectively permeable by Nuclear Pore Complexes (NPC) . Transport of proteins between the nucleus and the cytoplasm is carried out by karyopherins, a family of proteins made up of importins and exportins [2, 3]. Classical nuclear localization signal (NLS) dependent nuclear import is carried out by importin α and importin β Importin α family members bind NLS cargo, and bind to importin β through an N-terminal importin β binding domain (IBB). Importin β mediates translocation of the NLS-Importin α-Importin β import complex into the nucleus through direct interactions with the NPC. Once in the nucleus, RanGTP binds to importin β and induces dissociation of the import complex . Exportin mediated nuclear export is regulated by RanGTP through a related mechanism. Whereas RanGTP dissociates import complexes by binding importins, exportins must bind to RanGTP in order to bind nuclear export signal (NES) containing cargoes . The heterotrimeric export complex then translocates through the NPC and is dissociated in the cytoplasm by RanGAP stimulated conversion of RanGTP to RanGDP.
While there are at least 10 importin β family members which can bind directly to cargo and mediate import , importin β is unique in its ability to bind the importin α family of nuclear transport receptors (also called karyopherin α) [2, 3]. Importin α binds to two major classes of NLS, both characterized by basic amino acids; a monopartite NLS, such as the SV40 NLS, which consists of a single cluster of basic amino acids; and a bipartite NLS, such as the retinoblastoma (RB) NLS, which consists of two clusters of basic amino acids, separated by a ~10 residue spacer . The architecture of importin α proteins is composed of Armadillo (ARM) repeats, a three a-helix motif named for the D. melanogaster homologue of β catenin . The binding site for a monopartite NLS is located between the 2nd and 4th ARM repeats and is called the major site . Importin α binds to the C-terminus of bipartite NLS sequences with the major site and to the N-terminal element of the bipartite NLS using a smaller site created by the 7th and 8th ARM repeats called the minor site [8–10]. The accessibility of these NLS binding sites is regulated by an autoinhibitory mechanism. The IBB of importin α contains basic amino acids that bind to the NLS binding surface when the receptor is in an autoinhibited state [11–14]. Importin α binding to NLS cargo and to importin β is, therefore, a cooperative process because importin β binding to the IBB relieves the autoinhibition of importin α. Relief of autoinhibition facilitates Importin α binding to NLS cargo. After nuclear import, the complex is dissociated by the cooperative effects of RanGTP binding to importin β, and binding of importin α to CAS . CAS is an exportin which forms a trimeric complex consisting of CAS, RanGTP and importin α, and is responsible for recycling importin α to the cytoplasm .
Yeasts encode a single importin α, but higher eukaryotes encode three importin α subfamilies, designated importin α1, α2, and α3. There are six previously described human importin α forms, each encoded by different gene. Importin α family members show preferences for specific types of NLS cargo [17–20], although there is also some functional redundancy. Most NLSs are very similar in sequence leading to the suggestion that importin α specificity is conferred by the protein context of the NLS imparted by the cargo . It appears that additional importin α proteins are required for a multicellular organism because different importin α's have specific cargoes to import during differentiation . Studies in D. melanogaster have shown that importin α2 is not required in adult flies, but is required for oogenesis  and that importin α3 is favored for a step in larval development . Similarly, it has been shown in C. elegans that while importin α3 is ubiquitously expressed, α1 and α2 are expressed only in the germline, and are not sufficient to replace α3 in somatic cells . In mouse embryonic stem cells, switching from importin α1 (KPNA2) expression to importin α5 (KPNA1) expression has been shown to induce neural differentiation . These data suggest that cargo interactions with specific importin α isotypes are important for controlling differentiation pathways.
Here we describe a novel importin α in humans, the seventh member of the importin α family, which we term karyopherin alpha 7 (KPNA7). This new α family member is most similar to KPNA2 and a member of the α2 subfamily. We show that KPNA7 binds to importin β through an N-terminal IBB, but that its importin β binding affinity appears to be much higher than that of the other KPNAs. KPNA7 localization is predominantly nuclear in HeLa cells, highlighting another difference between 7 and the other importin α family members.
Importin α family identity matrix.
KPNA and importin α nomenclature.
The alignment of KPNA7 with KPNA's 1-6 (Fig 1) revealed that the IBB of KPNA7 has a region of low identity to the consensus. Separate phylogenetic analysis of the ARM repeats and IBB of the KPNAs show that the KPNA7 IBB is more divergent than the body of KPNA7 (Fig. 2B,C), as the branch length separating the KPNA7 IBB from S. cerevisiae importin α is longer than the branch length to the KPNA7 body. The α1 subfamily has the least divergent body, while the IBBs of this subfamily are almost as divergent as the KPNA7 IBB. KPNA2, KPNA3, and KPNA4 show the same branch lengths for both their bodies and IBB's, suggesting they evolved at the same rate. KPNA7 has an N-terminal domain which is 48% identical to the KPNA2 IBB, and only 24% identical to the KPNA5 IBB. The strong difference in the IBB of KPNA7 from all other KPNAs suggests that the KPNA7 IBB is has diverged in response to different selective pressures.
Because reticulocyte lystate contains proteins (including transport factors) that might influence the outcome of these binding assays, we used recombinant proteins to formally test whether the N-terminus of KPNA7 functions as an IBB. GST fusions of the IBB domains from KPNA2 and KPNA7 were immobilized on glutathione beads, recombinant importin β was added, and bound fractions analyzed by immunoblotting. Importin β bound to the IBB domains of both KPNA2 and KPNA7; the interactions were Ran-sensitive because they were reduced significantly by including RanQ69L (preloaded with GTP) in the binding reaction (Fig. 3C, first two lanes). We also found that preformed complexes of importin β with IBB domains of KPNA2 and KPNA7 were dissociated by addition of RanGTP (Fig. 3C), a reaction that mimics Ran-dependent disassembly of import complexes. Based on semi-quantitative immunoblotting, RanGTP addition resulted in dissociation of 60% of the preformed importin β-KPNA2 complexes but only 3% of the preformed importin β- KPNA7 complexes. This result, together with data showing that full-length KPNA7 binds efficiently to importin β (Fig. 2A), suggests that KPNA7 import complexes could be more stable than complexes formed between importin β and other KPNA proteins.
The competitive autoinhibition of NLS binding by the IBB may be altered by changes in the IBB affinity for importin α. Thus we decided to test whether the KPNA7ΔIBB would show enhanced binding to the SV40-NLS compared to full length KPNA7. Unexpectedly, neither full length nor IBB deletion forms of KPNA7 display significant levels of binding to the SV40 NLS (Fig 3D). The difference in SV40 NLS binding affinities between KPNA1 or KPNA 2 and KPNA7 suggests that there may be a difference in their NLS binding surfaces.
After we deposited the human KPNA7 sequence to GenBank (EU126604), Tejomurtula et al. reported the sequence of bovine KPNA7 and showed that it binds bovine nucleoplasmin . We therefore tested the ability of human KPNA7 to bind the human NPM NLS (NPM 150-175) (Fig 5B). The NPM NLS bound to KPNA1 very strongly, and showed some binding to KPNA3, KPNA4, and KPNA6 (although most of the bound KPNA6 appears to be a degradation product). Under our reaction conditions, neither KPNA7 nor KNPA7ΔIBB bound to the human NPM NLS.
Here we describe the characterization of a new member of the human importin α family, KPNA7. Sequence analysis revealed that KPNA7 is most similar to KPNA2 (54%), and that it is the most divergent member of the importin α family. Sequence divergence in KPNA7 occurs throughout the protein, including the IBB domain. Our binding data indictates that the IBB of KPNA7 binds importin β stronger than other KPNA proteins. The NLS binding surface is highly conserved in all of the importin α family members, but we observed clear differences among the different family members for binding the RB NLS, SV40 NLS, and the NPM NLS. KPNA7 bound to the RB NLS with equivalent affinity to KNPA2, but failed to bind both the SV40 NLS and the NPM NLS. In addition, by immunofluorescence microscopy KPNA7 localizes to the nucleus, a distribution that contrasts with the predominantly cytoplasmic localization of KPNA2. Our data emphasizes that despite a high degree of structural conservation, there are important differences in the NLS binding specificity and nucleocytoplasmic trafficking of importin -a family members.
Based on the conservation of the residues which make up the RB NLS and SV40 NLS binding surfaces of KPNA2, we reached the conclusion that the NLS binding surface is almost perfectly conserved in all seven members of the importin α family. It was surprising, therefore to observe KPNA-specific binding to both the RB NLS and SV40 NLS. Differences in the IBBs of importin α family members could have potentially altered the affinity for NLS binding through autoinhibition , but deletion of the IBB from KPNA7 did not alter its affinity for either the SV40 NLS or the NPM NLS. Assignment of the conserved residues is based upon the ClustalW alignment of the alpha family members, which was then superimposed upon the crystal structure of mouse KPNA2. The key residues are perfectly conserved, but this method of examining the conservation does not take into account gaps, deletions, or the effects of substitutions outside the binding surface on the ultimate structure of the binding surface. There is variability between α family members in the rest of their structure that may result in slight changes in the shape of the binding surface or relative location of the conserved residues which may then alter the NLS specificity of a given importin α. Evidence that the protein context of an NLS plays a role in the specificity for a particular importin α  suggests that an otherwise generic NLS may acquire specificity for a specific importin α due to structural constraints imposed by its protein context. Subtle differences in the conformation of the NLS binding surface may therefore lead to specificity for a contextualized NLS, and promiscuous binding to an unstructured NLS encountered out of its proper protein context. These results emphasize the need for experimental validation of predictions made for paralogues. Understanding how highly similar members of the same protein family achieve NLS selective binding will require structural analysis.
KPNA7 is the most divergent of the seven proteins in the human importin α family, and the IBB domain is more divergent than the ARM repeats, which maintain identity in the amino acids necessary for interaction with both the RB NLS and the SV40 NLS. The lack of conservation in the KPNA7 IBB suggests that it does not have the same selective pressures as KPNAs 1 through 6. There are two known functions for the IBB, binding to importin β, and autoinhibition through binding to importin α. Importin β binding could be affected by the amino acid changes in the KPNA7 N-terminus, but our data show very strong binding to importin β. This increased binding to importin β may represent an enhanced affinity of the IBB for importin β, but it could also be due to decreased autoinhibitory binding of the IBB to importin α.
A striking feature of KPNA7 is its localization to the nucleus under steady state conditions, while KPNA2 is predominantly cytoplasmic. This nuclear localization suggests that nucleocytoplasmic shuttling of KPNA7 differs from that of the other members of the importin α family. There are three possibilities for the altered subcellular distribution; 1) enhanced nuclear import by importin β; 2) decreased nuclear export by CAS; and 3) inefficient release of KPNA7 from target NLS, resulting in nuclear retention. The most likely of these situations, given our evidence that KPNA7 binds importin β with much higher affinity than other a family members is that this enhanced binding results in increased nuclear import of KPNA7.
The recent study on bovine KPNA7 showed that it is expressed preferentially in the embryo, suggesting a possible role in early in development . Given that KPNA1 plays a critical role in neuronal differentiation , restricted expression of KPNA isoforms could provide a mechanism for selective import of factors that drive key events during development.
We identified and characterized KPNA7, a highly divergent member of the importin a family of nuclear import receptors. The importin α family in vertebrates comprises seven proteins, KPNA1 through KPNA7. Sequence analysis and molecular modeling revealed the residues that form the NLS binding surface are nearly perfectly conserved throughout the importin a family. Nonetheless, there are striking differences in NLS recognition amongst the different KPNA proteins. Tertiary structure in KPNA proteins could play a major role in creating binding surfaces that can discriminate between different NLS's.
RNA from the prostate cancer cell line LnCaP was used as the template to amplify the open reading frame of KPNA7. Multiple alignments were made using the ClustalW alignment tool in Bioedit http://www.mbio.ncsu.edu/BioEdit/BioEdit.html. The Protdist neighbor function was used to create phylogenetic trees, and Dendroscope http://www-ab.informatik.uni-tuebingen.de/software/dendroscope was used to visualize the phylogenetic trees.
The parent vector used for generating the KPNA proteins was pCMVTNT (Promega). T7 or HA tags were cloned as annealed oligonucleotides with a 5' Kozak sequence into pCMVTNT using Xho1 and EcoR1 sites. KPNAs 1 through 6 were purchased from Origene, PCR amplified and cloned into pCMVTNT-T7 using EcoR1 and Sma1 sites (KPNA1, KPNA2, KPNA3, KPNA5, and KPNA6) or Kpn1 and Sma1 sites (KPNA4). Importin β was cloned using Mlu1 and Xba1 sites.
GST fusions of the indicated NLS constructs were expressed and purified by standard methods. His-tagged RanQ69L protein was expressed in the E. coli strain BL21 by induction for 16 hrs at 18 C, and purified on Talon beads. Recombinant Ran protein was loaded with GTP (100 uM) in the presence of EDTA (10 mM) for 20 min at 30 C. Bound nucleotide was stabilized by the addition of excess MgCl2 (20 mM), and unincorporated nucleotide was removed by chromatography on a PD-10 gel filtration column. The pCMVTNT-T7 KPNA constructs were translated in vitro with the TNT Coupled Reticulocyte Lysate System (cat# L4610, Promega) using T7 polymerase and 35S-methionine (cat# 51006, MP Biomedicals). 10 μg of each GST fusion protein was bound to glutathione beads, and incubated with 250,000 CPM of translated protein for 4 hours at 4°C. The beads were washed with buffer (25 mM Tris pH 7.5, 50 mM NaCl, 5 mM MgCl2, 0.1 mM EDTA, 0.5 mg/ml BSA, 0.1% NP-40, 1 mM DTT and protease inhibitors) and eluted in 2x SDS sample buffer. Samples were resolved on 10% acrylamide gels, fixed in methanol/acetic acid, incubated in Autofluor (National Diagnostics), dried and exposed to film at -80°C. Importin β binding to GST-IBB domains was performed in a similar manner except that His-tagged, recombinant importin β (1 μM final concentration) and Ran (15 μM final concentration, loaded with GTP) were used. Binding was assessed by semi-quantitative immunoblotting using IRDye conjugated secondary antibody and the Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE).
A peptide corresponding to a unique sequence (DQDYEFIDYEC) in the C-terminus of KPNA7 was synthesized, coupled to Keyhole Limpet hemocyanin, and injected into rabbits (Cocalico Biologicals, Inc.). Antibody was purified from serum using immobilized peptide.
HeLa cells were grown in DMEM containing 5% Newborn Calf Serum (NCS), 5% Fetal Bovine Serum, 1 mM Sodium Pyruvate at 37°C in 5% CO2. For transfection experiments, cells were plated on glass coverslips for 24 hours and transfected with KNPA plasmid using Transfectin (Biorad) according to the manufacturer's protocol. Cells were grown for an additional 24 hours and processed for immunofluorescence microscopy. Coverslips were washed with PBS, fixed with 4% formaldehyde for 20 minutes, permeabilized with 0.5% TritonX-100 for 5 minutes, blocked (2% BSA and 2% NCS in PBS), and incubated with primary antibody. Primary Antibodies used were 16B12 anti-HA antibody (HA.11 16B12, Covance), Rch1 antibody (cat# 610485, BD Biosciences), and anti-KPNA7. Secondary antibodies used were Donkey anti-Mouse FITC (cat# 715-095-150, Jackson ImmunoResearch) and Donkey anti-Rabbit Cy3 (cat# 711-165-152, Jackson ImmunoResearch).
We thank Dirk Gorlich and Steve Adam for their kind gifts of plasmids, and the Paschal laboratory for helpful discussions. These studies were supported by the National Cancer Institute (PO1 CA104016).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.