Review
Destination: inner nuclear membrane Santharam S. Katta1*, Christine J. Smoyer1*, and Sue L. Jaspersen1,2 1 2
Stowers Institute for Medical Research, Kansas City, MO 64110, USA Department of Molecular and Integrative Physiology, University of Kansas Medical Center, Kansas City, KS 66160, USA
The inner nuclear membrane (INM) of eukaryotic cells is enriched in proteins that are required for nuclear structure, chromosome organization, DNA repair, and transcriptional control. Mislocalization of INM proteins is observed in a wide spectrum of human diseases; however, the mechanism by which INM proteins reach their final destination is poorly understood. In this review we discuss how investigating INM composition, dissecting targeting pathways of conserved INM proteins in multiple systems and analyzing the nuclear transport of viruses and signaling complexes have broadened our knowledge of INM transport to include both nuclear pore complex-dependent and -independent pathways. The study of these INM targeting pathways is important to understanding nuclear organization and in both normal and diseased cells. INM proteins and nuclear organization The defining feature of eukaryotes is the nucleus. Unlike other organelles, the nucleus is formed by two lipid bilayers: an outer nuclear membrane (ONM) and an INM separated by a lumen. The ONM and INM merge at discrete regions of the nuclear envelope (NE) where nuclear pore complexes (NPCs) reside (Figure 1). Composed of 30 subunits present in multiple copies, NPCs are considered to be gatekeepers that restrict passage of macromolecules into and out of the nucleus in all eukaryotes [1,2]. Structural and molecular analysis has shown that the NPC contains a central channel lined with phenylalanine-glycine (FG)-rich repeat-containing nucleoporins (FG-Nups), which are thought to limit diffusion of molecules with a Stokes radius greater than 2.63 nm or a molecular weight of 40–60 kDa [3]. The NPC also contains a series of peripheral channels that may play a role in the transport of INM proteins (Figure 1; see below) [4–6]. The ONM is contiguous with the endoplasmic reticulum (ER) whereas the INM is distinct, containing several hundred to possibly a thousand proteins [7,8]. Studies of a handful of INM proteins, including the conserved SUN (for Sad1–UNC-84 homology) and LEM (for Lap1– emerin–MAN1) families, have revealed crucial roles for Corresponding author: Jaspersen, S.L. ([email protected]). Keywords: inner nuclear membrane; nuclear transport; NPC; SUN protein; LEM domain. * These authors contributed equally to this work. 0962-8924/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tcb.2013.10.006
INM proteins in nuclear structure, organization, and positioning (reviewed in [9–12]). One major function of SUN proteins is to connect the nucleus to the cytoplasmic cytoskeleton through their interaction with ONM proteins that bind to actin, dynein, microtubule-organizing centers, or intermediate filaments. Several SUN proteins as well as LEM domain-containing proteins also serve as scaffolds to cluster nuclear factors involved in transcriptional control, DNA repair, and meiotic recombination [11,13,14]. In metazoans, the function of SUN and LEM proteins in nuclear organization is partially dependent on lamins and other lamin-associated proteins, which form a NE-associated meshwork that is important for maintaining the structural integrity of the nucleus and the distribution of NPCs [15–18]. A myriad of human diseases, ranging from tissue-specific diseases of muscle, brain, bone, and fat to multisystem disorders such as the premature aging syndrome Progeria, are associated with mutations in the genes encoding lamins and INM components [16,19–21]. The etiology is unclear in many cases, but studies using tissue-culture cells, mouse, Caenorhabditis elegans, and Drosophila models have revealed interdependence between many INM proteins and lamins for their localization and/or function (e.g., [22–27]). Understanding how INM proteins are properly targeted and how their distribution is regulated in different cell types or under conditions such as stress or development is crucial to elucidate the mechanism of these disorders. Soluble cargo and the NPC The transport of soluble cargos across the NE has been extensively studied in many eukaryotic systems and a general set of principles for nucleocytoplasmic exchange has been established [1,28]. Trafficking of cargos into and out of the nucleus requires targeting information in the form of a nuclear localization sequence for entry (NLS; typically a short basic sequence or two basic sequences separated by a linker) or a nuclear export sequence for exit (NES; typically a short stretch of hydrophobic residues) [29]. These sequences are recognized by karyopherins (also known as importins and exportins) which facilitate movement through the central NPC channel. The ability of karyopherins to bind to their cargo depends on the small GTPase Ran (Ras-related nuclear protein). A gradient of Ran–GTP in the nucleus and Ran–GDP in the cytoplasm facilitates the binding and release of karyopherins and cargos, and it generates the directionality of transport (Figure 2). Although there is some diversity in import and export signals, and often redundancy between the karyopherins, these basic properties can account for transport of individual Trends in Cell Biology, April 2014, Vol. 24, No. 4
221
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Cytoplasm Central channel ∼50 nm
Peripheral channels ∼10 nm
Cytoplasmic FG Nups and filaments Yeast Metazoans Nup214 Nup159 – Nup358 hCG1 Nup42 Nup82 Nup88 Gle1 Gle1 Outer ring Nups Yeast Metazoans Nup120 Nup160 Nup133 Nup133 Nup96 Nup145C Nup107 Nup84 Nup85 Nup85 Seh1 Seh1 Sec13 Sec13 – Nup37 Nup43 – Cdc31 Centrin-2 – Aladin Nuclear FG Nups and basket Yeast Metazoans Nup60 – Nup153 Nup1 Nup2 Nup50 Mlp1/Mlp2 Tpr
Nucl Nu cleo eoplasm Transmembrane Nups Yeast Metazoans Ndc1 Ndc1 GP210 – Pom34 – Pom152 – Pom33 TMEM33 Pom121 –
Inner ring and linker Nups Yeast Metazoans Nup88 Nup82 Nic96 Nup93 Nup188/192 Nup188/205 Nup157/170 Nup155 Nup35 Nup53/59
Central FG Nups Yeast Metazoans Nup100/116/ Nup98/96 145N Nsp1 Nup62 Nup57 Nup54 Nup58/45 Nup49 TRENDS in Cell Biology
Figure 1. Schematic of the nuclear pore complex (NPC) showing subunits required for inner nuclear membrane (INM) targeting in yeast and metazoans. Biochemical and genetic analysis of the NPC shows that it is a modular structure. Its subunits, the nucleoporins (Nups), can be assigned to distinct functional subcomplexes based on their physical and genetic interactions with other Nups. Pore membrane proteins (Poms) connect to outer ring Nups to anchor the NPC in the nuclear envelope (NE). Linker Nups connect the outer ring to the inner ring Nups, and FG-Nups [phenylalanine-glycine (FG)-rich repeat-containing nucleoporins] form the central channel. Asymmetrically localized Nups form the nuclear basket and the cytoplasmic fibrils (based on [4–6]). Based on cryo-electron microscopy measurements of the human NPC, the diameters of the central and peripheral channels are approximately 50 and 10 nm, respectively [6]. Nups that have been tested for a role in INM trafficking are colored, and those that are required for transport of one or more INM proteins are in bold [37,43,48,51,52,84,85]. Nups in gray have not been assayed for a role in INM trafficking.
proteins and large protein complexes as well as RNAs, which are typically exported as part of a protein complex containing the targeting information [28–30]. It has been widely assumed that the localization of INM proteins would follow similar principles to those governing the localization of soluble cargo. The protein might diffuse or be trafficked to the INM by an INM protein-derived targeting sequence in an NPC-dependent pathway. This is particularly true in fungi that undergo a closed mitosis where NPC-mediated transport is the sole mechanism of macromolecular exchange between the nucleus and cytoplasm. Therefore, it was a surprise when studies of several conserved INM proteins failed to provide a simple paradigm for INM localization. Below we integrate our knowledge of INM trafficking and discuss growing evidence for at least four types of transport pathways. The ongoing study of these pathways will result in greater understanding of the nucleus, its evolution, and its function in genome organization and human diseases linked to NE dysfunction. Diffusion-retention The journey for a protein to the INM begins at the ER where the integral membrane protein is co-translationally 222
or post-translationally inserted into the ER membrane. Similarly to other integral membrane proteins found throughout the cell, insertion of INM proteins into the ER membrane likely involves the Sec61 translocon [31,32]. Because the ER is contiguous with the ONM, INM proteins are thought to freely diffuse through the ER to the ONM. Consistent with this hypothesis, photobleaching studies designed to assay the mobility of several INM proteins showed rapid diffusion from the ER to the NE [33–37]. The observation that their mobility decreased significantly once localized to the INM destination suggests that they are associated with relatively immobile components of the nucleus such as lamins or chromatin. Careful examination of NPC ultrastructure reveals that its peripheral channels might be sufficiently large (10 nm) to accommodate the diffusion or transport of an integral membrane protein assuming it has a small nucleoplasmic, or extralumenal, domain [6,38,39]. Studies of lamin B receptor (LBR), a multispanning INM component, showed a strong size-selection during INM transport: a version of LBR that had a 22 kDa extralumenal region localized to the INM whereas a 70 kDa version did not [40,41]. Similar studies on other INM proteins also
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Karyopherin α Karyopherin β NLS Cargo protein
Ran GDP
Ran GAP
Cargo protein NES
Ran GEF Ran GTP TRENDS in Cell Biology
Figure 2. Transport of soluble cargos. Targeting to the nucleus involves a nuclear localization sequence (NLS) that is recognized by an import-competent karyopherin (also called an importin) that shuttles the cargo through the nuclear pore complex (NPC) via its central channel. Binding of Ran (Ras-related nuclear protein)–GTP inside the nucleus causes the complex to disassemble. The karyopherin can be recycled to the cytoplasm while the cargo accumulates in the nucleus. During export, a nuclear export sequence (NES) is recognized by an export-competent karyopherin (also called an exportin) together with Ran–GTP. This ternary complex is transported through the NPC central channel to the cytoplasm, where nucleotide hydrolysis is stimulated, causing Ran–GTP to be converted to Ran–GDP, which releases the karyopherin and cargo. Abbreviations: GEF, guanine nucleotide exchange factor; GAP, GTPase activating protein.
indicate that there may be an upper size-limit on transport and have shown further roles for lamins as well as chromatin in the retention of proteins at the INM [35,37,42– 44]. This biphasic state of fast- and slow-moving pools of INM proteins associated with the ER and INM, respectively, combined with a strong size selection in the cargo, suggest that an INM protein is synthesized on the ER, diffuses freely between ONM and INM via the NPC, but is preferentially retained at the INM due to interactions with chromatin and/or lamins (Figure 3). The appealing simplicity of this model, referred to as diffusion-retention, may explain why it has predominated in the field until recently. LEM domain proteins: active transport? At first glance, nuclear trafficking of LEM domain-containing proteins such as Man1 (LEM domain-containing protein 3), Lap2b (lamina-associated polypeptide 2b), and emerin appears to follow the pattern of the diffusion-retention model (Table 1) [33,35,36,42,45]. For example, the localizations of human Man1 and Lap2b are largely dependent on the size and nuclear-binding function of their extralumenal domains (Table 1). A careful dissection of the N-terminus of Lap2b showed that it contains distinct chromatin- and lamin-binding domains; however, only the lamin-binding region is required for INM localization [42,45]. Similar results were also obtained for the Drosophila LEM proteins Otefin, Bocksbeutal, and Man1 [46,47]. It was therefore a surprise when analysis of the
budding-yeast LEM domain-containing proteins Heh1 (helix-extension-helix-1)/Src1 and Heh2 showed that their transport is not based on diffusion-retention but instead requires active transport similar to the pathway used by soluble cargos (Figure 4, Table 1) [48]. King and colleagues demonstrated that the accumulation of Heh1–YFP and Heh2–YFP at the INM is dependent on karyopherin-a (Kap60) and karyopherin-b (Kap95) together with the Ran GTPase cycle. They also identified a putative NLS in Heh2 and showed that this region (which includes adjacent residues) binds to karyopherins and is important for INM localization. Based on their work, the authors proposed a transport factor-based model for INM trafficking (Figure 3) [49]. Analysis of the primary sequence of other INM components revealed that many contain putative NLSs in their extralumenal domains, suggesting that this could be a widely used pathway for INM localization (Table 1) [49,50]. If Heh1 and Heh2 use active transport and require the same transport factors as soluble proteins, do they also use the central channel of the NPC for trafficking or are peripheral channels used? In yeast, as in higher eukaryotes, the size of the peripheral channels is small and would probably only accommodate a protein complex with a mass of 25–40 kDa [4,5,49,51]. It is therefore difficult to imagine how an integral membrane protein bound to a karyopherin-a/karyopherin-b complex would physically fit in the peripheral channel. However, it is also unclear how an INM protein might utilize the larger central channel, which is located in the center of the NPC approximately 50 nm away from the membrane region [6]. To test if Heh2 traverses through the central channel, Meinema et al. developed a clever strategy to trap translocation intermediates [52]. The authors constructed a synthetic INM protein by fusing the Heh2 NLS, its linker (L, the region between the NLS and transmembrane domain) and transmembrane domain to human FKBP12 and GFP (FKBP– GFP–NLS–L–TM). This was expressed in cells containing a version of the central channel NPC protein Nsp1 (nucleoskeletal-like protein/nucleoporin) fused to the FRB (FKBP12/rapamycin-binding) domain of human mTOR (mechanistic target of rapamycin). In the presence of rapamycin, FRB and FKBP12 will bind if they are in close proximity [53]. If the Heh2 NLS–linker construct uses the central NPC channel during transport, it should be possible to trap the synthetic INM protein in the NPC by addition of rapamycin. If it uses peripheral channels, rapamycin should have no effect because Nsp1–FRB and FKBP–GFP–NLS–L–TM will not come into physical proximity. NE aggregation of FKBP–GFP–NLS–L–TM in Nsp1–FRB cells in a rapamycin-dependent manner was observed, consistent with the idea that Heh2 uses the NPC central channel [52]. The fact that combinations of FG– Nups when deleted also affected transport further supported this possibility [52]. Curiously, trapping of the synthetic INM protein did not affect the transport of soluble cargos but did affect the NE accumulation of additional FKBP–GFP–NLS–L–TM, suggesting that trafficking of INM proteins is specifically blocked [52,54]. Thus, it appears that although INM and soluble transport pathways may at least partially overlap, there are differences in the actual transport mechanism. Dedicated NPCs and/or 223
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Diffusion retenon
Transport factor mediated Cyt y opl op asm
Cyt y opl op asm
Protein cargo
Protein cargo Karyopherin
Lamin or chroman binding domain
Central channel
Peripheral channel NLS
ER
Lamins
Ran GTP Chroman
Nucleoplasm
Sorng mof mediated
Nucleoplasm
Vesicle mediated
Ribosome Sec 61
Cytopl Cyt oplasm
Cytoplasm
Imporn α 16 Protein INM-SM
ER Protein
Nup 50/Nup 2
Nucleoplasm
Nucleoplasm TRENDS in Cell Biology
Figure 3. Four proposed INM targeting pathways. Integral membrane proteins are synthesized and inserted into the endoplasmic reticulum (ER) membrane either cotranslationally or post-translationally. (A) The diffusion-retention model suggests that the inner nuclear membrane (INM) protein is able to diffuse freely from the ER to the outer nuclear membrane (ONM), and to diffuse from the ONM to INM using peripheral nuclear pore complex (NPC) channels. Accumulation of the protein at the INM occurs due to tethering by chromatin and/or lamins. Because transport relies on peripheral channels, the extralumenal domain must be small (
Destination: inner nuclear membrane Santharam S. Katta1*, Christine J. Smoyer1*, and Sue L. Jaspersen1,2 1 2
Stowers Institute for Medical Research, Kansas City, MO 64110, USA Department of Molecular and Integrative Physiology, University of Kansas Medical Center, Kansas City, KS 66160, USA
The inner nuclear membrane (INM) of eukaryotic cells is enriched in proteins that are required for nuclear structure, chromosome organization, DNA repair, and transcriptional control. Mislocalization of INM proteins is observed in a wide spectrum of human diseases; however, the mechanism by which INM proteins reach their final destination is poorly understood. In this review we discuss how investigating INM composition, dissecting targeting pathways of conserved INM proteins in multiple systems and analyzing the nuclear transport of viruses and signaling complexes have broadened our knowledge of INM transport to include both nuclear pore complex-dependent and -independent pathways. The study of these INM targeting pathways is important to understanding nuclear organization and in both normal and diseased cells. INM proteins and nuclear organization The defining feature of eukaryotes is the nucleus. Unlike other organelles, the nucleus is formed by two lipid bilayers: an outer nuclear membrane (ONM) and an INM separated by a lumen. The ONM and INM merge at discrete regions of the nuclear envelope (NE) where nuclear pore complexes (NPCs) reside (Figure 1). Composed of 30 subunits present in multiple copies, NPCs are considered to be gatekeepers that restrict passage of macromolecules into and out of the nucleus in all eukaryotes [1,2]. Structural and molecular analysis has shown that the NPC contains a central channel lined with phenylalanine-glycine (FG)-rich repeat-containing nucleoporins (FG-Nups), which are thought to limit diffusion of molecules with a Stokes radius greater than 2.63 nm or a molecular weight of 40–60 kDa [3]. The NPC also contains a series of peripheral channels that may play a role in the transport of INM proteins (Figure 1; see below) [4–6]. The ONM is contiguous with the endoplasmic reticulum (ER) whereas the INM is distinct, containing several hundred to possibly a thousand proteins [7,8]. Studies of a handful of INM proteins, including the conserved SUN (for Sad1–UNC-84 homology) and LEM (for Lap1– emerin–MAN1) families, have revealed crucial roles for Corresponding author: Jaspersen, S.L. ([email protected]). Keywords: inner nuclear membrane; nuclear transport; NPC; SUN protein; LEM domain. * These authors contributed equally to this work. 0962-8924/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tcb.2013.10.006
INM proteins in nuclear structure, organization, and positioning (reviewed in [9–12]). One major function of SUN proteins is to connect the nucleus to the cytoplasmic cytoskeleton through their interaction with ONM proteins that bind to actin, dynein, microtubule-organizing centers, or intermediate filaments. Several SUN proteins as well as LEM domain-containing proteins also serve as scaffolds to cluster nuclear factors involved in transcriptional control, DNA repair, and meiotic recombination [11,13,14]. In metazoans, the function of SUN and LEM proteins in nuclear organization is partially dependent on lamins and other lamin-associated proteins, which form a NE-associated meshwork that is important for maintaining the structural integrity of the nucleus and the distribution of NPCs [15–18]. A myriad of human diseases, ranging from tissue-specific diseases of muscle, brain, bone, and fat to multisystem disorders such as the premature aging syndrome Progeria, are associated with mutations in the genes encoding lamins and INM components [16,19–21]. The etiology is unclear in many cases, but studies using tissue-culture cells, mouse, Caenorhabditis elegans, and Drosophila models have revealed interdependence between many INM proteins and lamins for their localization and/or function (e.g., [22–27]). Understanding how INM proteins are properly targeted and how their distribution is regulated in different cell types or under conditions such as stress or development is crucial to elucidate the mechanism of these disorders. Soluble cargo and the NPC The transport of soluble cargos across the NE has been extensively studied in many eukaryotic systems and a general set of principles for nucleocytoplasmic exchange has been established [1,28]. Trafficking of cargos into and out of the nucleus requires targeting information in the form of a nuclear localization sequence for entry (NLS; typically a short basic sequence or two basic sequences separated by a linker) or a nuclear export sequence for exit (NES; typically a short stretch of hydrophobic residues) [29]. These sequences are recognized by karyopherins (also known as importins and exportins) which facilitate movement through the central NPC channel. The ability of karyopherins to bind to their cargo depends on the small GTPase Ran (Ras-related nuclear protein). A gradient of Ran–GTP in the nucleus and Ran–GDP in the cytoplasm facilitates the binding and release of karyopherins and cargos, and it generates the directionality of transport (Figure 2). Although there is some diversity in import and export signals, and often redundancy between the karyopherins, these basic properties can account for transport of individual Trends in Cell Biology, April 2014, Vol. 24, No. 4
221
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Cytoplasm Central channel ∼50 nm
Peripheral channels ∼10 nm
Cytoplasmic FG Nups and filaments Yeast Metazoans Nup214 Nup159 – Nup358 hCG1 Nup42 Nup82 Nup88 Gle1 Gle1 Outer ring Nups Yeast Metazoans Nup120 Nup160 Nup133 Nup133 Nup96 Nup145C Nup107 Nup84 Nup85 Nup85 Seh1 Seh1 Sec13 Sec13 – Nup37 Nup43 – Cdc31 Centrin-2 – Aladin Nuclear FG Nups and basket Yeast Metazoans Nup60 – Nup153 Nup1 Nup2 Nup50 Mlp1/Mlp2 Tpr
Nucl Nu cleo eoplasm Transmembrane Nups Yeast Metazoans Ndc1 Ndc1 GP210 – Pom34 – Pom152 – Pom33 TMEM33 Pom121 –
Inner ring and linker Nups Yeast Metazoans Nup88 Nup82 Nic96 Nup93 Nup188/192 Nup188/205 Nup157/170 Nup155 Nup35 Nup53/59
Central FG Nups Yeast Metazoans Nup100/116/ Nup98/96 145N Nsp1 Nup62 Nup57 Nup54 Nup58/45 Nup49 TRENDS in Cell Biology
Figure 1. Schematic of the nuclear pore complex (NPC) showing subunits required for inner nuclear membrane (INM) targeting in yeast and metazoans. Biochemical and genetic analysis of the NPC shows that it is a modular structure. Its subunits, the nucleoporins (Nups), can be assigned to distinct functional subcomplexes based on their physical and genetic interactions with other Nups. Pore membrane proteins (Poms) connect to outer ring Nups to anchor the NPC in the nuclear envelope (NE). Linker Nups connect the outer ring to the inner ring Nups, and FG-Nups [phenylalanine-glycine (FG)-rich repeat-containing nucleoporins] form the central channel. Asymmetrically localized Nups form the nuclear basket and the cytoplasmic fibrils (based on [4–6]). Based on cryo-electron microscopy measurements of the human NPC, the diameters of the central and peripheral channels are approximately 50 and 10 nm, respectively [6]. Nups that have been tested for a role in INM trafficking are colored, and those that are required for transport of one or more INM proteins are in bold [37,43,48,51,52,84,85]. Nups in gray have not been assayed for a role in INM trafficking.
proteins and large protein complexes as well as RNAs, which are typically exported as part of a protein complex containing the targeting information [28–30]. It has been widely assumed that the localization of INM proteins would follow similar principles to those governing the localization of soluble cargo. The protein might diffuse or be trafficked to the INM by an INM protein-derived targeting sequence in an NPC-dependent pathway. This is particularly true in fungi that undergo a closed mitosis where NPC-mediated transport is the sole mechanism of macromolecular exchange between the nucleus and cytoplasm. Therefore, it was a surprise when studies of several conserved INM proteins failed to provide a simple paradigm for INM localization. Below we integrate our knowledge of INM trafficking and discuss growing evidence for at least four types of transport pathways. The ongoing study of these pathways will result in greater understanding of the nucleus, its evolution, and its function in genome organization and human diseases linked to NE dysfunction. Diffusion-retention The journey for a protein to the INM begins at the ER where the integral membrane protein is co-translationally 222
or post-translationally inserted into the ER membrane. Similarly to other integral membrane proteins found throughout the cell, insertion of INM proteins into the ER membrane likely involves the Sec61 translocon [31,32]. Because the ER is contiguous with the ONM, INM proteins are thought to freely diffuse through the ER to the ONM. Consistent with this hypothesis, photobleaching studies designed to assay the mobility of several INM proteins showed rapid diffusion from the ER to the NE [33–37]. The observation that their mobility decreased significantly once localized to the INM destination suggests that they are associated with relatively immobile components of the nucleus such as lamins or chromatin. Careful examination of NPC ultrastructure reveals that its peripheral channels might be sufficiently large (10 nm) to accommodate the diffusion or transport of an integral membrane protein assuming it has a small nucleoplasmic, or extralumenal, domain [6,38,39]. Studies of lamin B receptor (LBR), a multispanning INM component, showed a strong size-selection during INM transport: a version of LBR that had a 22 kDa extralumenal region localized to the INM whereas a 70 kDa version did not [40,41]. Similar studies on other INM proteins also
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Karyopherin α Karyopherin β NLS Cargo protein
Ran GDP
Ran GAP
Cargo protein NES
Ran GEF Ran GTP TRENDS in Cell Biology
Figure 2. Transport of soluble cargos. Targeting to the nucleus involves a nuclear localization sequence (NLS) that is recognized by an import-competent karyopherin (also called an importin) that shuttles the cargo through the nuclear pore complex (NPC) via its central channel. Binding of Ran (Ras-related nuclear protein)–GTP inside the nucleus causes the complex to disassemble. The karyopherin can be recycled to the cytoplasm while the cargo accumulates in the nucleus. During export, a nuclear export sequence (NES) is recognized by an export-competent karyopherin (also called an exportin) together with Ran–GTP. This ternary complex is transported through the NPC central channel to the cytoplasm, where nucleotide hydrolysis is stimulated, causing Ran–GTP to be converted to Ran–GDP, which releases the karyopherin and cargo. Abbreviations: GEF, guanine nucleotide exchange factor; GAP, GTPase activating protein.
indicate that there may be an upper size-limit on transport and have shown further roles for lamins as well as chromatin in the retention of proteins at the INM [35,37,42– 44]. This biphasic state of fast- and slow-moving pools of INM proteins associated with the ER and INM, respectively, combined with a strong size selection in the cargo, suggest that an INM protein is synthesized on the ER, diffuses freely between ONM and INM via the NPC, but is preferentially retained at the INM due to interactions with chromatin and/or lamins (Figure 3). The appealing simplicity of this model, referred to as diffusion-retention, may explain why it has predominated in the field until recently. LEM domain proteins: active transport? At first glance, nuclear trafficking of LEM domain-containing proteins such as Man1 (LEM domain-containing protein 3), Lap2b (lamina-associated polypeptide 2b), and emerin appears to follow the pattern of the diffusion-retention model (Table 1) [33,35,36,42,45]. For example, the localizations of human Man1 and Lap2b are largely dependent on the size and nuclear-binding function of their extralumenal domains (Table 1). A careful dissection of the N-terminus of Lap2b showed that it contains distinct chromatin- and lamin-binding domains; however, only the lamin-binding region is required for INM localization [42,45]. Similar results were also obtained for the Drosophila LEM proteins Otefin, Bocksbeutal, and Man1 [46,47]. It was therefore a surprise when analysis of the
budding-yeast LEM domain-containing proteins Heh1 (helix-extension-helix-1)/Src1 and Heh2 showed that their transport is not based on diffusion-retention but instead requires active transport similar to the pathway used by soluble cargos (Figure 4, Table 1) [48]. King and colleagues demonstrated that the accumulation of Heh1–YFP and Heh2–YFP at the INM is dependent on karyopherin-a (Kap60) and karyopherin-b (Kap95) together with the Ran GTPase cycle. They also identified a putative NLS in Heh2 and showed that this region (which includes adjacent residues) binds to karyopherins and is important for INM localization. Based on their work, the authors proposed a transport factor-based model for INM trafficking (Figure 3) [49]. Analysis of the primary sequence of other INM components revealed that many contain putative NLSs in their extralumenal domains, suggesting that this could be a widely used pathway for INM localization (Table 1) [49,50]. If Heh1 and Heh2 use active transport and require the same transport factors as soluble proteins, do they also use the central channel of the NPC for trafficking or are peripheral channels used? In yeast, as in higher eukaryotes, the size of the peripheral channels is small and would probably only accommodate a protein complex with a mass of 25–40 kDa [4,5,49,51]. It is therefore difficult to imagine how an integral membrane protein bound to a karyopherin-a/karyopherin-b complex would physically fit in the peripheral channel. However, it is also unclear how an INM protein might utilize the larger central channel, which is located in the center of the NPC approximately 50 nm away from the membrane region [6]. To test if Heh2 traverses through the central channel, Meinema et al. developed a clever strategy to trap translocation intermediates [52]. The authors constructed a synthetic INM protein by fusing the Heh2 NLS, its linker (L, the region between the NLS and transmembrane domain) and transmembrane domain to human FKBP12 and GFP (FKBP– GFP–NLS–L–TM). This was expressed in cells containing a version of the central channel NPC protein Nsp1 (nucleoskeletal-like protein/nucleoporin) fused to the FRB (FKBP12/rapamycin-binding) domain of human mTOR (mechanistic target of rapamycin). In the presence of rapamycin, FRB and FKBP12 will bind if they are in close proximity [53]. If the Heh2 NLS–linker construct uses the central NPC channel during transport, it should be possible to trap the synthetic INM protein in the NPC by addition of rapamycin. If it uses peripheral channels, rapamycin should have no effect because Nsp1–FRB and FKBP–GFP–NLS–L–TM will not come into physical proximity. NE aggregation of FKBP–GFP–NLS–L–TM in Nsp1–FRB cells in a rapamycin-dependent manner was observed, consistent with the idea that Heh2 uses the NPC central channel [52]. The fact that combinations of FG– Nups when deleted also affected transport further supported this possibility [52]. Curiously, trapping of the synthetic INM protein did not affect the transport of soluble cargos but did affect the NE accumulation of additional FKBP–GFP–NLS–L–TM, suggesting that trafficking of INM proteins is specifically blocked [52,54]. Thus, it appears that although INM and soluble transport pathways may at least partially overlap, there are differences in the actual transport mechanism. Dedicated NPCs and/or 223
Review
Trends in Cell Biology April 2014, Vol. 24, No. 4
Diffusion retenon
Transport factor mediated Cyt y opl op asm
Cyt y opl op asm
Protein cargo
Protein cargo Karyopherin
Lamin or chroman binding domain
Central channel
Peripheral channel NLS
ER
Lamins
Ran GTP Chroman
Nucleoplasm
Sorng mof mediated
Nucleoplasm
Vesicle mediated
Ribosome Sec 61
Cytopl Cyt oplasm
Cytoplasm
Imporn α 16 Protein INM-SM
ER Protein
Nup 50/Nup 2
Nucleoplasm
Nucleoplasm TRENDS in Cell Biology
Figure 3. Four proposed INM targeting pathways. Integral membrane proteins are synthesized and inserted into the endoplasmic reticulum (ER) membrane either cotranslationally or post-translationally. (A) The diffusion-retention model suggests that the inner nuclear membrane (INM) protein is able to diffuse freely from the ER to the outer nuclear membrane (ONM), and to diffuse from the ONM to INM using peripheral nuclear pore complex (NPC) channels. Accumulation of the protein at the INM occurs due to tethering by chromatin and/or lamins. Because transport relies on peripheral channels, the extralumenal domain must be small (
