The nuclear pore complex: structure and function

Vincent Duheron, Birthe Fahrenkrog

Institute for Molecular Biology and Medicine, Université Libre de Bruxelles, 6041 Charleroi, Belgium

Corresponding author: Birthe Fahrenkrog
bfahrenk@ulb.ac.be; phone: +32 2 650 9951; fax: +32 2 650 9950.

October 2014

Abstract

Nuclear pore complexes (NPCs) are large multi-protein complexes, which are embedded in the nuclear envelope and which are regulating the molecular exchange between the nucleus and the cytoplasm. While electron microscopy and cryo-electron tomography studies have provided high-resolution pictures of the NPC structure as entity, the challenge nowadays is to elucidate the organization and the functions of nuclear pore proteins (nucleoporins or Nups) inside and outside the NPC. Nucleoporins are not only involved in nucleocytoplasmic transport, but in an increasing number of other cellular processes, such as kinetochore organization, cell cycle regulation, DNA repair, and gene expression. The implication of nucleoporins in these diverse processes links them also to a wide variety of human diseases, such as cancer and autoimmune diseases. Here we review the progress made in defining the molecular arrangement of nucleoporins within the NPC and use the example of Nup153 to illustrate the versatility of individual nucleoporins and their implication in various human diseases.

Keywords: nucleus, nuclear pore complex; nucleoporins, FG repeats, nucleocytoplasmic transport, Nup153, disease, HIV-1

1. Introduction

Nuclear and cytoplasmic compartments of interphase eukaryotic cell are separated by the nuclear envelope (NE), which is formed from two closely juxtaposed membranes, the outer nuclear membrane (ONM) and the inner nuclear membrane (INM), respectively. The ONM is continuous with the rough endoplasmic reticulum, whereas the INM contains unique transmembrane proteins, which establish contacts with chromatin and the nuclear lamina. Large multi-protein complexes known as nuclear pore complexes (NPCs) perforated the NE to allow molecular trafficking between the cytoplasm and the nucleus. Nucleocytoplasmic transport comprises passive diffusion of small molecules and ions, as well as signal- and receptor-mediated translocation of proteins and ribonucleoprotein complexes that are larger than ~40 kDa (Görlich et al., 1995; Melchior et al., 1993; Moroianu and Blobel, 1995; Moroianu et al., 1995a; Moroianu et al., 1995b; Radu et al., 1995a; Radu et al., 1995b; Rexach and Blobel, 1995; Sweet and Gerace 1995; Kehlenbach et al., 1998; Chook et al., 1999; Keminer and Peters, 1999; for review see: Görlich and Kutay, 1999). Nuclear transport receptors are collectively known as karyopherins and they comprise importins and exportins. Karyopherins recognize signal sequences within their cognate cargoes: nuclear localization signals (NLSs) for cargo destined for the nucleus and nuclear export signals (NESs) for cargo that is exported from the nucleus (Moroianu and Blobel, 1995; Moroianu et al., 1995a; Moroianu et al., 1995b; Moroianu et al., 1996; Wen et al., 1995; Conti et al., 1998; Feng et al., 1999). The directionality of nucleocytoplasmic transport is determined by the small GTPase Ran, which is switching between its GDP- and GTP-bound states (Melchior et al., 1993; Moore and Blobel, 1993; Weis et al., 1996; Izaurralde et al., 1997; Görlich and Kutay, 1999). RanGTP is predominantly found in the nucleus and its binding to importins displaces import cargo from the receptor (Görlich et al., 1996; Izaurralde et al., 1997). The RanGTP-importin complex is directed to the cytoplasm, where the cytoplasmic GTPase-activating protein RanGAP1 catalyzes GTP hydrolysis, together with the Ran-binding proteins RanBP1 and RanBP2. GTP hydrolysis results in the disassembly of the RanGTP-importin complex and the subsequent recycling of the importin for a new round of nuclear import (Kutay et al., 1997a; for review see: Görlich and Kutay, 1999). RanGTP furthermore forms trimeric complexes with exportins and export cargoes to stabilize their interaction during nuclear export (Arts et al., 1998; Fornerod et al., 1997; Kaffman et al., 1998; Kutay et al., 1997a; Kutay et al., 1998). After translocation through the NPC, cytoplasmic GTP hydrolysis allows the dissociation of the trimeric complex and the redirection of the exportin to the nucleus. RanGDP is reimported by its own nuclear import receptor, NTF2 (Ribbeck et al., 1998; Bayliss et al., 1999; Chaillan-Huntington et al., 2000), and the GDP on Ran becomes exchanged to GTP by the action of the chromatin-bound Ran guanine nucleotide exchange RCC1, which is critical to maintain the intracellular Ran gradient (Görlich et al., 1996; Boche and Fanning, 1997; Izaurralde et al., 1997). Cargo translocation through NPCs is a fast process with hundreds of proteins, RNA particles, and metabolites passing through each NPC every second (Peters et al., 1986; Schulz and Peters, 1987; Peters et al., 1990; Middeler et al., 1997; Ribbeck and Görlich, 2001; Yang et al., 2004; Kubitscheck et al., 2005; Dange et al., 2008; Siebrasse et al., 2012).

2. Nucleoporins

The NPC is composed of approximately 30 different proteins, termed nucleoporins (Nups), which are broadly conserved between yeast, vertebrates and plants (Rout et al., 2000; Cronshaw et al., 2002; Neumann et al., 2010). Secondary structure predictions of nucleoporins allowed their classification into three groups (Schwartz, 2005; Devos et al., 2006): the first group of nucleoporins is characterized by the presence of transmembrane α-helices and cadherin-like domains, which mediate the anchoring of NPCs to the NE and which stabilize the interaction between the INM and the ONM. The second group is composed of nucleoporins containing α-solenoid and β-propeller folds, which may optimize interactions between nucleoporins as these particular structures are mainly found in architectural, scaffold nucleoporins. This scaffold nucleoporins connect the transmembrane nucleoporins to nucleoporins containing phenylalanine-glycine (FG)-repeats, i.e. the third group of nucleoporins (Schwartz, 2005; Devos et al., 2006).

FG nucleoporins represent about one third of the nucleoporins and they are characterized by the presence of repeated clusters of FG motifs that fall into three predominant classes: FxFG (x refers to any residue), GLFG (L refers to leucine) and FG. Other less frequent motifs are PAFG, PSFG and SAFG (Devos et al., 2006; Denning and Rexach, 2007). FG motifs are separated by linker sequences, which are typically enriched in charged and polar amino acids and are "disorder"-associated (Dunker et al., 2001; Denning and Rexach, 2007). The linker sequences lack hydrophobic, "order"-associated amino acids, so that this particular amino acid distribution renders FG domains natively unfolded with an absence of secondary structure (Denning et al., 2002; Denning et al., 2003; Fahrenkrog et al., 2002; Lim et al., 2006a).

The extend of charged amino acids varies between individual nucleoporins and two subgroups appear to exist: FG domains with low content of charged amino acids, which are capable of low affinity interactions among each other, whereas FG domains with a high content of charged amino acids abolish interactions with each other (Patel et al., 2007; Yamada et al., 2010; Xu and Powers, 2013). In this context, GLFG motifs are more susceptible to interact with each other than other motifs (Patel et al., 2007). The cohesiveness of FG domains and the spacing between the FG motifs is further dependent on the amino acid composition of the linker sequences (Patel et al., 2007; Dolker et al., 2010). An interaction between FG domains may contribute to a cohesive barrier that hinders diffusion of small proteins through the NPC (Shulga et al., 2000; Ribeck and Görlich, 2001; Hülsmann et al., 2012). FG nucleoporins are located throughout the NPC (Rout et al., 2000; Fahrenkrog et al., 2002; Walther et al., 2002; Paulillo et al., 2005; Schwarz-Herion et al., 2007; Chatel and Fahrenkrog, 2012) and they are mediating the interaction to nuclear transport receptors. Crystal structures have shown that the interaction between FG repeats and transport receptors mainly involves the Phe residues of the FG repeats, together with the flanking Gly residues that provide conformational flexibility, and hydrophobic residues of the receptor (Quimby et al., 2001; Fribourg et al., 2001; Bayliss et al., 2002a; Bayliss et al., 2002b; Liu and Stewart, 2005; Vognsen et al., 2013). Karyopherins possess several FG-binding sites and their translocation through the NPC is accomplished due to multiple and rapid binding events to the FG-nucleoporins (Rexach and Blobel, 1995; Kutay et al., 1997; Bayliss et al., 2000; Allen et al., 2001; Gilchrist et al., 2002; Tetenbaum-Novatt et al., 2012). While originally it was assumed that an "affinity gradient" between karyopherins and nucleoporins determines translocation from the cytoplasmic filaments to the nuclear basket (Ben-Efraim and Gerace, 2001; Pyhtila and Rexach, 2003), it has now been shown that only FG-nucleoporins that are symmetrically localized to both sides of the NPC are essential for translocation and cell viability, whereas the asymmetric FG-nucleoporins are dispensable for transport (Strawn et al., 2004). Yeast NPCs can accommodate the loss of 50% of their FG-repeats with only little effect on nucleocytoplasmic transport or NPC permeability. Moreover, Zeitler et al. were able to alter or invert the asymmetric distribution of FG-nucleoporins without altering nucleocytoplasmic transport (Zeitler et al., 2004).

The disordered configuration of FG-nucleoporins and the transient, dynamic interactions within the NPC make the elucidation of the nucleocytoplasmic transport mechanisms difficult. Numerous translocation models have been elaborated, such as the Brownian affinity/virtual gating/polymer brush model (Rout et al., 2000; Rout et al., 2003; Lim et al., 2006; Lim et al., 2007), the selective phase/hydrogel model (Ribbeck and Görlich, 2002; Frey et al., 2006; Frey et al., 2007), the reduction of dimensionality model (Peters, 2005; Moussavi-Baygi et al., 2011), and the forest model (Yamada et al., 2010). Details of the models are discussed in depth in several excellent recent review articles (Tetenbaum-Novatt and Rout, 2010; Wente and Rout, 2010; Lim and Deng, 2009).

3. NPC architecture

Nucleoporins assemble into distinct subcomplexes (Fig. 1A), which can be isolated as such from interphase NPCs and frequently from disassembled mitotic NPCs as well. These nucleoporin subcomplexes serve as building blocks for the NPC (Matsuoka et al., 1999; Allen et al., 2002; Suntharalingam and Wente, 2003; Schwartz, 2005). Electron microscopic and tomographic analyses of NPCs from different species, such as yeast, Xenopus, and human, have demonstrated that, despite the large evolutionary distance between these species, the basic structural organization of the NPC is evolutionary conserved (Xenopus: Hinshaw et al., 1992; Akey and Radermacher, 1993; Stoffler et al., 2003; Frenkiel-Krispin et al., 2010; human: Maimon et al., 2012; Bui et al., 2013; yeast: Fahrenkrog et al., 1998; Rout et al., 2000; Kiseleva et al., 2004; Dictyoselium: Beck et al., 2004; Beck et al., 2007; Grossman et al., 2012). Accordingly, the NPC consists of an eight-fold symmetric central framework, eight cytoplasmic filaments and a nuclear basket (Fig. 1B). The central framework has an hourglass-like shape and is composed of three connected rings: a central spoke ring sandwiched between a cytoplasmic ring and a nuclear ring. The spoke ring resides within the NE and is anchored to the region where the inner and outer nuclear membranes fuse. Eight cytoplasmic filaments are attached to the cytoplasmic ring, while the nuclear ring is also decorated with eight filaments that join into a distal ring, thereby forming the NPC's nuclear basket (Fahrenkrog and Aebi, 2003; Beck and Medalia, 2008; Grossman et al., 2012; Maimon et al., 2012).

Among the NPC subcomplexes, the Nup107-160 complex is the most studied and best characterized one due to its pivotal role for NPC assembly. The Nup107-160 complex is composed of nine nucleoporins, i.e. Nup160, Nup133, Nup107, Nup96, Nup85, Nup43, Nup37, Sec13, and Seh1 (Belgareh et al., 2001; Vasu et al., 2001) and it is usually associated with the putative transcription factor Elys/Mel-28 (Rasala et al., 2006; Szymborska et al., 2013). Depletion of any member of the Nup107-160 complex leads to defects in NPC assembly (Boehmer et al., 2003; Harel et al., 2003; Walther et al., 2003). The Nup107-160 complex is symmetrically located to both sides of the NPC (Belgareh et al., 2001; Vasu et al., 2001; Bui et al., 2013; Szymborska et al., 2013) and EM analysis of the isolated Nup107-160 complex and its yeast homologue, the Nup84p complex, has revealed that the complex adopts a Y-shaped architecture (Fig. 2A) (Siniossoglou et al., 1996; Siniossoglou et al., 2000; Lutzmann et al., 2002; Kampmann and Blobel, 2009; Bui et al., 2013). The relative positions of each nucleoporin within the Y complex are known (Kampmann and Blobel, 2009; Bui et al., 2013), whereas its orientation within the NPC is not ultimately clear. Three main models are discussed: (1) the fence model predicts that 32 copies of the Nup84p complex are organized in four octameric rings, which are packed in antiparallel orientation, and which are linked by vertical hetero-octamers of either Nup85p-Seh1p or Nup145p-Sec13p (Fig. 2B) (Debler et al., 2008); (2) the lattice model predicts that the Nup84p complex resides on the cytoplasmic and nuclear side of the NPC, sandwiching the Nic96p complex (see below), with the long axis of Nup84p complex positioned almost parallel to the nucleocytoplasmic transport axis (Fig. 2C) (Brohawn and Schwartz, 2009b); (3) the head-to-tail model suggests that Nup84p complexes are arranged head-to-tail into two antiparallel rings on the cytoplasmic and nuclear face of the NPC (Fig. 2D) (Alber et al., 2007a; Seo et al., 2009). Recent advances in light microscopy confirmed the head-to-tail octameric ring arrangement of the Nup107-160 complex directly in cells (Szymborska et al., 2013) and electron tomography data indicate that Nup107-160 complexes assemble at both faces of the NPC as paired octameric rings with a shift of about 11 nm towards each other, which in turn appears to be important for the structural plasticity of the NPC (Bui et al., 2013).

Sandwiched between the Nup107-160 complex and the cytoplasmic and nuclear ring is the Nup93 complex, which is composed of five nucleoporins, i.e. Nup93, Nup205, Nup188, Nup155 and Nup35/53, and it is evolutionary conserved (Grandi et al., 1995; Schlaich et al., 1997; Grandi et al., 1997; Miller et al., 2000; Galy et al., 2003; Hawryluk-Gara et al., 2005; Hawryluk-Gara et al., 2008; Theerthagiri et al., 2010; Amlacher et al., 2011; Sachdev et al., 2012; Vollmer and Antonin, 2014). It occupies a central position in the NPC architecture, serving as a link between the different nucleoporin subcomplexes within the NPC scaffold. Hence the Nup93 complex can potentially link the NPC to the NE via Nup35/53 and Nup155: both interact with the transmembrane nucleoporin Ndc1 and Nup155 additionally with Pom121 (Mansfeld et al., 2006; Onischenko et al., 2009; Mitchell et al., 2010; Eisenhardt et al., 2014). No direct interaction between the Nup107-160 and the Nup93 complex has been demonstrated in vertebrates, but for the yeast complexes (Lutzmann et al., 2005), suggesting that their vertebrate homologues bind each other as well.

The Nup93 complex is also interacting with nucleoporins of the Nup62 complex via the N-terminal region of Nup93 and Nup62 (Grandi et al., 1997; Sachdev et al., 2012), which links the Nup93 complex to the channel nucleoporins Nup62, Nup58, Nup54 and Nup45 (Finlay et al., 1991; Guan et al., 1995; Vasu and Forbes, 2001; Xenopus homologues: Macaulay et al., 1995; yeast homologues: Grandi et al., 1993). The FG-nucleoporins of the Nup62 complex are believed to project their FG-domains towards the center of the NPC channel thereby contributing to nucleocytoplasmic transport and the NPC permeability barrier (Finlay et al., 1991; Paschal and Gerace, 1995; Clarkson et al., 1996; Hu et al., 1996; Yoshimura et al., 2013). How the peripheral NPC structures, the cytoplasmic filaments and the nuclear basket, are exactly linked to the NPC scaffold remains to be seen.

While the comparison of NPCs from different organisms revealed a shared global architecture, variations especially in the height of the central framework were observed (Fig. 1C). For instance, the central framework of NPCs from Xenopus laevis are ~ 95 nm high with an outer diameter of ~ 125 nm (Frenkiel-Krispin et al., 2010; Stoffler et al., 2003), while human NPCs have a central framework with a height of ~ 85 nm and an outer diameter of ~ 120 nm (Maimon et al., 2012; Bui et al., 2013). NPCs from yeast and Dictyoselium discoideum are more compact with a central framework of ~ 60 nm in height and an outer diameter of ~ 120 nm (Fahrenkrog et al., 1998; Yang et al., 1998; Beck et al., 2004; Beck et al., 2007; Kiseleva et al., 2004; Frenkiel-Krispin et al., 2010). Closer analyses of the central spoke ring revealed a similar thickness (~ 35 nm) of this ring in different species, indicating that the size differences within the central framework arise from species-specific variations in the arrangements of the cytoplasmic and nuclear rings (Frenkiel-Krispin et al., 2010). Despite these differences in the dimensions of the central framework, the diameter of the central pore in the mid plane of the NE is about 40 and 50 nm (Fahrenkrog et al., 1998; Yang et al., 1998; Stoffler et al., 2003; Beck et al., 2007, Fiserova et al., 2009; Frenkiel-Krispin et al., 2010; Bui et al., 2013), which is close to the size limit of 39 nm in diameter for cargo translocation (Panté and Kann, 2002). Besides the central pore, metazoan NPCs exhibit peripheral channels of ~ 5-10 nm in diameter (Hinshaw et al., 1992; Akey and Radermacher, 1993; Yang et al., 1998; Stoffler et al., 2003; Beck et al., 2007; Maimon et al., 2013). These peripheral channels seem to be absent in yeast NPCs likely due to the smaller size of the central spoke complex (Yang et al., 1998). As of today, the function of these peripheral channels is not ultimately clear and controversially discussed (Hinshaw et al., 1992; Akey, 1995; Soullam and Worman, 1995; Kiseleva et al., 1998; Danker et al., 1999; Shahin et al., 2001; Stoffler et al., 2003; Ohba et al., 2004; Frenkiel-Krispin et al., 2010).

4. The nucleoporin Nup153: domain organization and localization

The nucleoporin Nup153 is a constituent of the NPC's nuclear basket (Pante et al., 1994; Pante et al., 2000; Walther et al., 2001; Fahrenkrog et al., 2002; Hase and Cordes, 2003), and it is, based on its amino acid sequence, comprised of roughly three major domains: (i) a N-terminal domain, which further includes a nuclear envelope targeting cassette (NETC; amino acids 1-144), a nuclear pore association region (NPAR; amino acids 39-339) and a RNA binding domain (RBD; amino acids 250-400) (Bastos et al., 1996; Enarson et al., 1998; Ball et al., 2004), (ii) a central zinc finger domain and (iii) a FG-rich C-terminal domain (for review see: Ball and Ullman, 2005) (Fig. 3A). The zinc-finger domain of Nup153 binds DNA in vitro (Sukegawa and Blobel, 1993) as well as RanGDP and RanGTP, albeit with a preference for the GDP-bound form (Higa et al., 2007, Schrader et al., 2008; Partridge et al., 2009). The cellular functions of these interactions have thus far remained elusive. Important for NE breakdown in reconstituted nuclei from Xenopus egg extracts, however, is the recruitment of the coatomer COPI complex, a major participant of membrane remodeling, to NPCs via Nup153's zinc-finger domain (Liu et al., 2003; Prunuske et al., 2006). The FG-domain of Nup153 is known to bind several nuclear transport factors, such as importin β, transportin and NXF1, which is important for Nup153's role in nucleocytoplasmic transport (Shah et al., 1998; Moroianu et al., 1995; Moroianu et al., 1997; Nakielny et al., 1999; Bachi et al., 2000; Brownawell and Macara, 2002; Kuersten et al., 2002; Walther et al., 2003). The FG domain of Nup153 is natively unfolded and has a length of about 800 nm, when fully extended (Lim et al., 2006). The actual conformation of the FG domain within the NPC is unknown: it might adapt the form of a polymer brush that acts as entropic barrier, which transiently collapses to a more compact conformation upon interaction with transport receptors, such as importin β (Lim et al., 2006; Lim et al., 2007). Alternatively, Nup153's FG-domain might form a hydrogel as seen at high concentrations in vitro (Milles et al., 2011; Milles et al., 2013), in which only transport receptors and transport complexes remain soluble (Ribbeck and Görlich, 2002; Frey et al., 2006; Frey and Görlich, 2007). Or a combination of both (Milles and Lemke, 2011), as it has been shown for the yeast nucleoporin Nsp1p (Ader et al., 2010).

The association of Nup153 with the nuclear basket of the NPC (Cordes et al., 1993; Pante et al., 1994) is rather complex, which became evident from the application of domain-specific antibodies against Nup153 in combination with immuno-EM analyses. Accordingly, the N-terminal domain locates to the nuclear ring moiety of the nuclear basket (Pante et al., 2000; Walther et al., 2001; Fahrenkrog et al., 2002), whereas the zinc-finger domain is found at the distal ring (Fahrenkrog et al., 2002). Nup153's FG-domain was detected all over the nuclear basket and occasionally on the cytoplasmic side of the NPC, suggesting a high mobility of this domain (Nakielny et al., 1999; Fahrenkrog et al., 2002; Paulillo et al., 2005; Paulillo et al., 2006; Lim et al., 2007) (Fig. 3B). Not only the FG-domain of Nup153 appears to be mobile, but the entire protein: fluorescent recovery after photobleaching (FRAP) experiments have revealed that GFP-tagged Nup153 is highly dynamic and shuffles between a NPC-associated and a nucleoplasmic population (Daigle et al., 2001; Rabut et al., 2004; Griffis et al., 2004). Nup153's mobility is dependent on active transcription, which suggests that Nup153 may associate with mRNA cargoes to facilitate their recruitment to the NPC (Griffis et al., 2004). Binding to mRNA might occur either directly via the N-terminal RNA binding domain (Ball et al., 2004) or via the mRNA export factor NXF1, which binds the C-terminal FG-domain of Nup153 (Bachi et al., 2000).

5. Nup153 assembly into nuclear pores

The NE of eukaryotic cells disassembles at the end of mitotic prophase, which is accompanied with NPC disassembly. NPC reassembly starts in late anaphase/early telophase in a highly ordered process during which Nup153 belongs to the early recruited nucleoporins (Bodoor et al., 1999; Haraguchi et al., 2000; Daigle et al., 2001; Dultz et al., 2008). Nup153 incorporation in the reforming NPC requires the Nup107-160 complex and depletion of this complex impairs Nup153 recruitment to NPCs (Boehmer et al., 2003; Walther et al., 2003; Harel et al., 2003; Krull et al., 2004). Nup153 recruitment to the newly reforming NPC might occur in two phases: a first pool of Nup153 accumulates at the periphery of chromosomes at the end of anaphase (Bodoor et al., 1999; Haraguchi et al., 2000; Dultz et al., 2008), shortly after Nup107-160 recruitment (Dultz et al., 2008), while a second and major pool of Nup153 associates with the NPC later in telophase (Dultz et al., 2008), at the same time as the nuclear lamina (Smythe et al., 2000). The assembly and stable association of Nup153 with the NE and NPCs is in fact dependent on the lamina and mutations in lamin A have been shown to displace Nup153 from NPCs (Smythe et al., 2000, Hübner et al., 2006). NPCs lacking Nup153 are more mobile within the NE (Walther et al., 2001), suggesting that a Nup153-lamina interaction is important for NPC anchoring in the NE. Consistently, Nup153 was found to bind lamin B3 (also known as lamin LIII) from Xenopus egg extracts (Smythe et al., 2000) and recombinant human lamin A and lamin B1 (Al Haboubi et al., 2011). Moreover, Nup153 depletion from HeLa cells altered the localization of lamin A/C and Sun1, an INM protein that binds to lamin A (Zhou and Pante, 2010), which coincided with a rearrangement of the cytoskeleton and a modified cell morphology (Zhou and Pante, 2010).

The early recruited pool of Nup153 might be important for the reformation of a functional NPC and the correct recruitment of other nucleoporins to the NPC. Along this line depletion of Nup153 from Xenopus egg extracts resulted in the loss of several nucleoporins from the nuclear basket, amongst others Tpr (Walther et al., 2001). Tpr is thought to be the central architectural element of the nuclear basket (Krull et al., 2004) and is lately recruited during NPC reassembly (Bodoor et al., 1999; Haraguchi et al., 2000). Consistently, Nup153 localization to the NPC is not Tpr dependent (Frosst et al., 2002; Hase and Cordes, 2003), whereas the importance of Nup153 for Tpr recruitment is controversial: Nup153 depletion lead to a complete absence of Tpr from the NE on the one hand (Hase and Cordes, 2003), but not in other cases (Lussi et al., 2010; Mackay et al., 2010; Umlauf et al., 2013). In contrast to that, it is without any doubt that the association of Nup50 with NPCs necessitates the presence of Nup153 (Hase and Cordes, 2003; Mackay et al., 2010). Nup50 is a mobile nucleoporin associated with the nuclear basket (Guan et al., 2000; Smitherman et al., 2000) and its recruitment to reforming NPCs is closely associated with Nup153's with a similar, biphasic recruitment in ana- and telophase (Dultz et al., 2008).

6. Versatile functions of Nup153

Nup153 participates in nucleocytoplasmic transport via its FG domain, which interacts with numerous transport receptors such as importin β (Moroianu et al., 1995; Shah et al., 1998), transportin (Shah and Forbes, 1998), importin-5 (Yaseen and Blobel, 1997) and importin-7 (Walther et al., 2003), as well as with the nuclear export receptors CRM1 (Nakielny et al., 1999), exportin-5 (Brownawell and Macara, 2002), exportin-t (Kuersten et al., 2002) and NXF1 (Bachi et al., 2000). Moreover, Nup153 binds import adapter proteins, such as importin α (Moroianu et al., 1997) and the specialized nuclear import receptor for Ran, NTF2 (Cushman et al., 2004), suggesting that Nup153 is a decisive player in many transport pathways. The high mobility of Nup153's FG domain and its spread distribution throughout the NPC (Fahrenkrog et al., 2002; Paulillo et al., 2005; Lim et al., 2006; Lim et al., 2007) may promote the translocation of nuclear import complexes to the nuclear side of the NPC (Ogawa et al., 2012) and/or promote the entry of nuclear export complexes to the NPC (Soop et al., 2005).

Because of the key position of the NPC at the interface between the cytoplasm and the nucleus, it became evident that nucleoporins either directly or indirectly participate in numerous cellular processes, both in interphase and mitosis. Not surprisingly, Nup153 is therefore essential for cell viability (Harborth et al., 2001; Galy et al., 2003) and alterations in Nup153 expression impair numerous cellular pathways by altering the localization of a large number of factors. For example, Nup153 contributes to gene expression via micro RNAs and small interfering RNAs, as it is important for the nuclear localization of Dicer1, the protein that cuts double stranded RNAs (Ando et al., 2011). Moreover, in Drosophila melanogaster and in human cells Nup153 is required for correct dosage compensation of the X-chromosome (Mendjan et al., 2006; Vaquerizas et al., 2010). Nup153 directly binds chromatin on large nucleoporin-associated regions, which probably promotes the formation of an open chromatin environment (Vaquerizas et al., 2010).

Nup153 is critical for the nuclear localization of 53BP1, a mediator protein in DNA damage response (DDR), to DNA double strand breaks (DSBs) (Wang et al., 2002). Cells depleted for Nup153 have an increased sensitivity to DSB inducing drugs, which is in part due to a cytoplasmic mislocalization of 53BP1 (Lemaitre et al., 2012; Moudry et al., 2012). Furthermore, Nup153 depletion compromised the phosphorylation of the two cell cycle checkpoint kinases CHK1 and CHK2, suggesting an improper activation of the G2/M checkpoint, and it promotes the use of the homologous recombination pathway over non-homologous end joining to repair DSBs (Lemaitre et al., 2012). Nup153 becomes phosphorylated by the Ataxia telangiectasia mutated (ATM) kinase, a key kinase during DSB repair (Wan at al., 2013). ATM-mediated phosphorylation of Nup153 provokes the interaction between Nup153 and exportin-5, which in turn promotes the nuclear export of pre-miRNAs in response to DNA damage (Zhang et al., 2011; Wan at al., 2013). The implication of miRNAs in DDR is still poorly understood on a mechanistic level, but it is known that a loss of Dicer and miRNAs resulted in increased levels of DNA damage (Wan et al., 2011).

Nup153 is not only impacting cellular processes due to its central role in nucleocytoplasmic transport, but also due to direct interactions with numerous partners, such as transcription factors (Zhong et al., 2005; Xu et al., 2002), signaling molecules (Nybakken et al., 2005), membrane remodeling proteins (Liu et al., 2003; Prunuske et al., 2006), and SUMO specific proteases (Chow et al., 2012). SUMOylation is a post-translational modification characterized by the addition of a small ubiquitin-like peptide on target proteins to modulate their activity (for review see: Wang and Dasso, 2009). Like ubiquitination, SUMOylation occurs through a three-step enzymatic cascade, which requires the action of an ATP-dependent E1 activating enzyme, an E2 conjugating enzyme and a SUMO-specific E3 ligase. SUMOylation is a reversible process and SUMO removal is catalyzed by SUMO-specific isopeptidases SENPs, two of which, SENP1 and SENP2, are located at the NE and NPCs (Hang and Dasso, 2002; Zhang et al., 2002; Bailey and O'Hare, 2004; Chow et al., 2012; Cubenas-Potts et al., 2013). SENP1 and SENP2 directly interact with Nup153, suggesting that Nup153 is responsible for their NE localization and that it might be implicated in the deSUMOylation of proteins (Hang and Dasso, 2002; Chow et al., 2012; Cubenas-Potts and Matunis, 2013).

Nup153 not only has functions in interphase, but also in mitosis. Its role in mitosis might be at least in part due to its direct interaction with the spindle assembly checkpoint (SAC) protein Mad1 (Lussi et al., 2010). The SAC assures correct chromosome segregation at the metaphase-anaphase transition (see review Lara-Gonzalez et al., 2012) and overexpression of Nup153 leads to a SAC override and partial mislocalization of Mad1, which coincides with chromosome mis-segregation, multipolar spindle formation, multinucleation and cytokinesis failure (Lussi et al., 2010). Cytokinesis, the final step of mitosis, is also hampered by siRNA-mediated depletion of Nup153 (Mackay et al., 2009; Lussi et al., 2010). Nup153 depletion causes mislocalization of Aurora B kinase, which is regulating many mitotic processes (see review van der Waal et al., 2012), during cytokinesis. This mislocalization of Aurora B leads to the activation of the so-called abscission checkpoint and consequently abscission delay and an increased number of cells in cytokinesis (Mackay et al., 2009; Mackay et al., 2010; Lussi et al., 2010). The underlying molecular mechanism that leads to the activation of the abscission checkpoint remains to be elucidated.

7. Nup153-related disorders

Because of the pivotal role of NPCs for nucleocytoplasmic exchange, alterations in NPC components and/or nucleocytoplasmic transport have a strong impact on cell growth and survival. Therefore, it is not surprising that nucleoporins and likewise Nup153 are implicated in a large number of disorders, such as cancer and autoimmune disease (Table 1). Increased expression of Nup153 due to a 6p22 genomic translocation was detected in urothelial carcinoma and retinoblastoma (Orlic et al., 2006; Heidenblad et al., 2008). Moreover, in a screen for pancreatic cancer genes, Nup153 was found amplified in the pancreatic cell line PL5 (Shain et al., 2013). This study suggested an oncogenic function for Nup153 by modulating the TGF-β signaling pathway. Nup153 is known to regulate the intracellular distribution of the TGF-β signal transducer SMAD2 (Xu et al., 2002). It stoichiometrically competes with FAST1, a nuclear retention factor for SMAD2, and Nup153 up-regulation leads to an increased cytoplasmic localization of SMAD2, which disrupts TGF-β signaling and may enforce proliferation (Shain et al., 2013). Nup153 is furthermore important for tumor cell migration (Zhou and Pante, 2010). Depletion of Nup153 from HeLa cells and the breast cancer cell line MDA-231, respectively, induced rearrangements of the actin cytoskeleton and microtubules. The alterations of the cytoskeleton led to impaired cell migration and polarization of MDA-231 cells and fibrosarcoma HT1080 cells (Zhou and Pante, 2010). Further work is necessary to address the molecular details of Nup153's role in cancer biology.

Small nucleotide polymorphisms (SNPs) in NUP153 were found associated with disorders due to hyperbilirubinemia (Datta et al., 2012) and schizophrenia (Lin et al., 2009). The transport of biliverdin reductase, an important enzyme for bilirubin conjugation, is affected by a SNP near the NUP153 locus on chromosome 6. This leads to increased levels of unconjugated bilirubin, which in turn is associated with severe disorders, such as bilirubin encephalopathy and Gilbert's syndrome (Datta et al., 2012).

Autoantibodies against Nup153 were found in sera from patients suffering from various autoimmune diseases, such as systemic lupus erythematosis, autoimmune thyroiditis, hepatitis B and hepatitis C virus (HBV and HCV, respectively) infections (Nesher et al., 2001; Enarson et al., 2004). To explain the presence of these Nup153 autoantibodies some investigators suggested a molecular mimicry mechanism. Sera from patients with various autoimmune diseases possess antibodies against HBV DNA polymerase subtype adr (Gregorio et al., 1999). Surprisingly, sera from patients with chronic HBV infection react with the peptide 827-846 of Nup153. This peptide sequence of Nup153 presents similarities to residues 57-76 of HBV DNA polymerase, which may lead to cross-reactions between the Nup153 peptide 827-846 and antibodies directed against the HBV DNA polymerase peptide 57-76 (Gregorio et al., 1999).

As mentioned earlier, Nup153 can directly interact with the Ig-fold domain of nuclear lamins (Smythe et al., 2000; Al Haboubi et al., 2011). Laminopathies are disorders caused by mutations in the LMNA gene, often localized in the Ig-fold domain of the lamin A protein (Goldman et al., 2004; Worman and Bonne, 2007). These mutations reduce the population of Nup153 present at the NE maybe by impairing Nup153-lamin interaction or by promoting accumulation of Nup153 at mutated lamin A aggregates (Hübner et al., 2006). This lack of Nup153 at the nuclear periphery might contribute to a compromised nuclear protein import (Busch et al., 2009).

8. Nup153 and HIV

Nup153 is furthermore a target for viruses to enable their replication. Thereby viruses utilize different strategies to take advantage of Nup153's central function in nucleocytoplasmic transport. Polioviruses or rhinoviruses, for example, express a proteinase enzyme that is causing the cleavage and degradation of Nup153, Nup98 and Nup62, which consequently inhibits nucleocytoplasmic transport (Belov et al., 2000) and increases the permeability of the NPC. This increased permeability promotes passive diffusion and the re-localization of host cell nuclear proteins that take part in viral replication (Meerovitch et al., 1993; McBride et al., 1996; Gustin and Sarnow, 2001; Gustin and Sarnow, 2002; Park at al., 2008; Fitzgerald et al., 2013). Inhibition of nucleocytoplasmic transport might reduce the export of host mRNAs to the cytoplasm, which in turn is reducing the competition for the translation machinery and which is promoting viral protein synthesis.

Other viruses, such as human immunodeficiency virus 1 (HIV-1), need to enter the nucleus for their replication, and therefore to pass through the NPC. It is accepted that the HIV-1 nuclear import is an active and energy-dependent process (Suzuki and Craigie, 2007), which requires the help of import receptors (Konig et al., 2008). Numerous nuclear transport factors were identified that are implicated in HIV-1 replication (Brass et al., 2008; Konig et al., 2008) and among the best characterized are Nup153 and Nup358. Both nucleoporins are required for HIV-1 nuclear entry as their depletion decreases HIV-1 integration into the host genome (Konig et al., 2008; Di Nunzio et al., 2012). Nup153 and Nup358 can interact with the HIV-1 capsid (Schaller et al., 2011; Di Nunzio et al., 2013b; Matreyek et al., 2013) and Nup358 may serve as a docking site for the HIV-1 pre-integration complex (PIC) at the nuclear pore, while Nup153 might play a role in the nuclear entry of the PIC (Lee et al., 2010; Matreyek and Engelman, 2011). Nup153 may interact with PICs docked on Nup358 and facilitate their transport through the NPC. Nup153 depletion does not only decrease HIV-1 nuclear entry, but also reduces the number of HIV-1 integration sites into transcriptionally active regions (Koh et al., 2013; Di Nunzio et al., 2013). It has been proposed that Nup153 is directly guiding the PIC to the genomic integration regions. Besides the interaction with the PIC, Nup153 also interacts with HIV-1 integrase in vitro when both proteins are produced in recombinantly in bacteria (Woodward et al., 2009), but this interaction was not detected when both proteins were co-expressed in mammalian cells (Di Nunzio et al., 2013), suggesting that the HIV-1 capsid is the viral determinant for the requirement of Nup153 and that this interaction is critical for HIV-1 nuclear import (Matreyek and Engelman, 2011; Di Nunzio et al., 2013). Future investigations are required to further support this notion.

9. Conclusion

Progress in imaging techniques in recent years, both on the light and electron microscopic level, in combination with X-ray crystallography has provided novel insights into the overall structure of the NPC as well as into the organization and localization of individual nucleoporins within the NPC architecture. While the structural roles of nucleoporins and their functions in nucleocytoplasmic transport have been appreciated for a long time, new functions for nucleoporins - direct and indirect - are regularly discovered. The more we learn about this multifunctional usage of nucleoporins it becomes apparent that they are central for many cellular functions and processes, both in interphase and mitosis. Many of these processes, such as the regulation of gene expression, have direct relevance for human health and dysfunction of nucleoporins is not surprisingly frequently associated with human malignancies and viral infections. How defects in nucleoporins alter cellular pathways precisely and how this leads to a particular disease, however, has remained largely elusive. Future investigations in this context are urgently needed.

Acknowledgements

This work was supported by grants from the Fonds de la Recherche Scientifique-FNRS Belgium (grants T.0237.13, 1.5019.12, and F.6006.10), the Fonds Brachet and the Fonds Van Buuren.