Mapping the dynamic organization of the nuclear pore complex inside single living cells


Most cellular activities are executed by multi-protein complexes that form the basic functional modules of their molecular machinery1. Proteomic approaches can provide an evermore detailed picture of their composition, but do not reveal how these machines are organized dynamically to accomplish their biological function. Here, we present a method to determine the dissociation rates of protein subunits from complexes that have a traceable localization inside single living cells. As a case study, we systematically analysed the dynamic organization of vertebrate nuclear pore complexes (NPCs), large supramolecular complexes of about 30 different polypeptides2. NPC components exhibited a wide range of residence times covering five orders of magnitude from seconds to days. We found the central parts of the NPC to be very stable, consistent with a function as a structural scaffold, whereas more peripheral components exhibited more dynamic behaviour, suggesting adaptor as well as regulatory functions. The presented strategy can be applied to many multi-protein complexes and will help to characterize the dynamic behaviour of complex networks of proteins in live cells.


NPCs are very large protein complexes, embedded in the two membranes of the nuclear envelope, that mediate all traffic between the cytoplasm and the nucleoplasm. In mammalian cells, they are composed of multiple copies of about 30 proteins termed 'nucleoporins'2. Only two nucleoporins, Pom121 and gp210, are integral membrane proteins that are consequently believed to anchor NPCs in the nuclear membranes3, whereas all others are synthesized as soluble proteins in the cytoplasm. Several nucleoporins can be biochemically isolated as subcomplexes that are believed to be the building blocks for the assembly of the whole NPC3. In addition, the majority of nucleoporins have numerous repeats containing FG dipeptides that are thought to be essential for nucleocytoplasmic traffic through their low affinity for transport receptors such as importin β3.

How NPCs mediate rapid and selective nucleocytoplasmic translocation of macromolecules is still controversial. They have often been described as a 'stationary phase' between dynamic cytoplasm and nucleoplasm, merely providing static binding sites for rapidly trafficking transport receptors. In interphase mammalian cells, NPCs are indeed stably anchored in the nuclear envelope4 (see Supplementary Information, Fig. S2 and Movie S1) and some nucleoporins have been shown to stay associated with NPCs over hours4,5,6. In contrast to this static view of the NPC, several nucleoporins have been suggested to be dynamically associated with NPCs4,7,8,9 or even to shuttle between the cytoplasm and the nucleoplasm7,10,11. Such dynamic properties have been proposed to provide several functions in nucleocytoplasmic transport including targeting, translocation, recycling and release of transport complexes from NPCs7,8,9,10,11,12 or regulation of the small GTPase Ran10. How many nucleoporins are dynamically or stably associated with NPCs and what their rate of dissociation from the NPC are remains an open question. To map the functional dynamics of the entire pore complex, we systematically and quantitatively analysed the dissociation rates of green fluorescent protein (GFP)-tagged nucleoporins from the NPC in single living mammalian cells.

For this, we produced a set of stable cell lines expressing GFP-tagged nucleoporins at low levels to overcome the toxicity of overexpression and to allow efficient incorporation of the tagged proteins into the long-lived complex. For 19 out of the roughly 30 nucleoporins, monoclonal cell lines were successfully generated that expressed full-length proteins, proliferated normally and had no gross morphological defects of the nucleus (Fig. 1). All GFP-tagged nucleoporins localized to NPCs, as shown by the punctuate pattern on the nuclear envelope (Fig. 1). Many nucleoporins also exhibited a pool in the cytoplasm (or the endoplasmic reticulum for gp210), the nucleoplasm or both, consistent with the reported immunolocalization of the endogenous proteins8,13,14 (Fig. 1). The amount of GFP-tagged nucleoporins at NPCs relative to the endogenous protein was assessed by quantitative western blot analysis, using the two least variegated cell lines. The Pom121–EGFP3 line showed a twofold overexpression of the tagged Pom121 and a twofold downregulation of the endogenous protein, that is, a 2.5-fold higher combined expression (Fig. 2c). Similarly, the EGFP3–Nup107 line had a total twofold higher expression compared with wild-type cells (Fig. 2c). The EGFP3–Nup107 molecules were nevertheless functional as they could complement the essential endogenous Nup107 in RNA interference (RNAi) experiments, where they showed dynamics identical to those in the untreated cells (see Supplementary Information, Fig. S1). In addition, EGFP3–Nup107 showed identical dynamics in HeLa cells, ruling out an influence of c.ellular background from different mammalian species on the dissociation rate (see Supplementary Information, Fig. S1). To make our data comparable to future measurements of nucleoporin dynamics, it was essential to provide a quantitative description of the steady-state level of NPC-bound GFP-tagged nucleoporin for each nucleoporin cell line. Therefore, we measured the amount of GFP-tagged nucleoporins localizing to single NPCs by comparison with virus-like particles containing 120 GFP molecules using quantitative confocal microscopy (Fig. 2a,b)15. NPCs contained between three and 50 GFP-tagged nucleoporins (Table 1), which was similar to the endogenous amounts estimated from proteomics studies (8–48)2. Thus, all tagged nucleoporins were correctly targeted and incorporated into NPCs.

Figure 1: Stable cell lines expressing GFP-tagged nucleoporins.

Whole-cell and NPC localization of a representative cell of each stable line expressing GFP-tagged nucleoporins, illustrated by confocal sections in the cell centre and on the lower surface of the nucleus. Scale bars represent 5 μm. Western blots with anti-GFP antibody show expression of full-length tagged nucleoporin. In some lines no signal could be detected, which was probably because of their high variegation and/or low expression level and/or low number of GFP tag copies associated with each nucleoporin. Markers, in descending order, indicate relative molecular weights of 200,000, 150,000, 100,000, 75,000, 50,000 and 37,000.

Figure 2: Expression levels of the GFP-tagged nucleoporins.

(a) Average fluorescence intensity distribution of single virus-like-particles (VLPs) containing 120 GFP molecules. (b) Distribution of the number of GFP-tagged nucleoporins in single NPCs measured using the calibration shown in a for the Pom121–EGFP3- and EGFP3–Nup107-expressing cell lines as an example. The other 18 lines were analysed identically (Table 1). (c) Quantitative western blot of serial dilutions of the Pom121–EGFP3- and EGFP3–Nup107-expressing cell lines probed with anti-Pom121 or anti-Nup107-C antibodies to compare expression levels of tagged and endogenous protein. Nup153, probed with monoclonal antibody (mAb) 414, was used as a loading control.

Table 1 Dissociation rates and number of GFP-tagged nucleoporins per NPC

To assay the time of association of the GFP-tagged nucleoporins with the NPCs, we performed inverse fluorescence recovery after photobleaching experiments (iFRAPs)16. In an iFRAP, the entire cellular fluorescence is photobleached except for a small region containing the tagged protein bound in the target complex (Fig. 3a and 4a). Its dissociation is then reflected by the loss of fluorescence from this region over time monitored by confocal microscopy (see Methods section). Figure 3a shows single iFRAP experiments for four nucleoporins with different dynamic behaviour. At least seven similar experiments were performed for all 19 GFP-tagged proteins, and were averaged to generate the mean fluorescence decays plotted in Fig. 3b–d. iFRAPs are more suitable to analyse binding interactions than classical FRAPs because the free pool of the tagged protein is bleached away, allowing precise measurement of the fluorescence changes of only the bound molecules. We determined dissociation rates (or their inverse, the mean residence time) of nucleoporins from iFRAP experiments using kinetic modelling, assuming that the interaction of a free nucleoporin (or subcomplex) with its receptor at the NPC is described by a reversible bimolecular chemical reaction in steady state (Fig. 4b, c).

Figure 3: Nucleoporin dynamics at nuclear pore complexes.

(a) Selected images and corresponding plots of fluorescence decay kinetics from the unbleached region of representative iFRAP experiments showing four nucleoporins (Nup93, Nup62, Nup98 and gp210) that illustrate different dissociation rates. (bd) Dissociation kinetics obtained for all 19 investigated nucleoporins and importin β. The three different plots show: (b) 'dynamic' nucleoporins and importin β, (c) 'structural adaptor' nucleoporins and (d) 'structural scaffold' nucleoporins. Each curve represents the mean of at least seven independent experiments; the standard deviations, not shown for clarity, were less than 10% of the initial fluorescence intensity for the initial time points and less than 20% for the later time points.

Figure 4: Kinetic modelling.

(a) The interaction of a free nucleoporin with its receptor at the NPC was modelled by a reversible chemical reaction. The dissociation rate of the reaction was determined by fitting the dissociation kinetics obtained from iFRAP experiments (see Methods section) with either (b) a single exponential decay (for example, Pom121) or (c) a bi-exponential decay (for example, Nup153) when a single exponential model clearly did not fit the experimental data. (d) The dissociation kinetics of Nup62 was modelled equally well with a single exponential decay or a bi-exponential decay with the dissociation rates determined for Nup214 and Nup58 (also see Supplementary Information, Fig. S4c).

Based on clustering of their residence times, we could classify the 19 nucleoporins into three distinct classes (Table 1; Figs 3b–d and 5a). Ten nucleoporins were barely exchangeable with dissociation rates below 7 × 10−6 s−1 or a residence time of more than 35 h (Table 1 and Fig. 3d). All the tested members of the Nup107–160 subcomplex (Nup133, Nup107, Nup85, Nup43, Nup37, Sec13 and Seh1; ref. 17) were found in this category, consistent with studies on individual proteins5,6. This indicates that this subcomplex binds as a unit to the NPC and supports the findings that it has an essential function for proper NPC assembly18,19. Similarly, stable interactions with the NPC were also observed for Nup214, Nup93 and Aladin. We thus propose that those nucleoporins form part of the 'structural scaffold' of the NPC. Scaffold nucleoporins may not all be essential for NPC integrity, as nuclei lacking Nup214 or Aladin have no gross NPC ultrastructural defects20,21.

Figure 5: Map of the dynamic organization of the nuclear pore complex.

(a) Distribution of the average residence times of the 19 different nucleoporins displayed on a logarithmic pseudo-colour scale, illustrating the three dynamic classes. (b) Spatial map of the nucleoporins based on published localizations superimposed onto a drawn-to-scale scheme representing nuclear pore complex structure. The asterisk represents the other members of the Nup107–160 complex measured in this study (Seh1, Sec13, Nup37, Nup43 and Nup85). The average residence time for each nucleoporin is represented quantitatively by the pseudo-colour scale in a. Scale bar represents 20 nm.

The second class consisted of six nucleoporins that had dissociation rates between 10−4 s−1 and 10−5 s−1 or residence times between 2.5 and 30 h (Table 1 and Fig. 3c). This class of nucleoporins, which included Pom121, Nup62, Nup58, CG1, Nup35 and Nup98, might therefore function as 'structural adaptors' of the NPC, which would assemble on top of the very stable 'structural scaffold'. Pom121 was the most stable in this group, with an average residence time of about 20 h, consistent with a role in anchoring the NPC in the nuclear membrane. Nup62 had an overall dissociation rate intermediate between Nup214 and Nup58 (Table 1), consistent with its participation in two separate NPC subcomplexes with either protein3. Our kinetic model would predict that on average about 31% of Nup62 is present in the Nup214–Nup88–Nup62 subcomplex and about 69% is present in the Nup62–Nup58–Nup45–Nup45 subcomplex if the dissociation rates of Nup214 and Nup58 are assumed to be representative of the two subcomplexes (Fig. 4d). Nup98 had an average residence time at NPCs of about 3 h, which was significantly more stable than reported in classical FRAP experiments on transiently overexpressing HeLa cells8. This difference could be explained by much lower expression levels and/or better incorporation of Nup98–EGFP into NPCs in our stable cell lines, or by the difficulty to exclude the mobile nucleoplasmic fraction of Nup98 in FRAP, but not in iFRAP, experiments.

Only three nucleoporins had dissociation rates larger than 10−4 or residence times at NPCs below 2.5 h (Table 1; and Fig. 3b). An average residence time of about 5 min was observed for gp210, which was surprising as it is considered to be a core component of NPC structure both for membrane anchorage and NPC assembly22. This result is consistent with the observation that endogenous gp210 can exchange between nuclei in heterokaryon assays (B. Burke, personal communication) as well as its late recruitment during post-mitotic nuclear envelope formation14. Our data argues for a non-structural function of gp210, as already suggested by its cell-type and tissue-specific expression23, the lack of a clear penetrant phenotype of the Caenorhabditis elegans homologue by RNAi24, and observations that it is not essential for nuclear pore formation in vitro (W. Antonin, personal communication). The two other dynamic nucleoporins were Nup50 and Nup153. Both had iFRAP kinetics that could not be modelled accurately with a single type of binding interaction with the NPC. A model assuming two separate binding sites at the NPC fitted the data and allowed us to determine two dissociation rates (Fig. 4c; also see Supplementary Information, Fig. S4b). About 90% of Nup50 had a dissociation rate of roughly 0.05 s−1, or a residence time of 20 s, whereas 10% was associated with NPCs for more than 3 min. Nup153 also had a fast and a slower exchanging pool (0.038 s−1 (60%) and 0.001 s−1 (40%); about 1 min and 13 min residence times, respectively; Table 1). Those results are consistent with previous characterization of Nup153 dynamics by FRAP4,25, and with the finding that Nup2p — the yeast analogue of Nup50 (ref. 11) — was also dynamically associated with NPCs10. The bi-exponential dissociation kinetics of those nucleoporins could reflect their interaction with at least two distinct receptors at NPCs. In particular, the slower fraction (about eight fluorescent molecules per NPC; Table 1) of Nup153 may function in anchoring the nucleoporin Tpr at the nuclear basket. Alternatively, it is also possible that those nucleoporins can change their dissociation rate from a single receptor at NPCs, for instance after a conformational change induced by their interaction with a transport complex. The dynamic properties of Nup153 and Nup50 could impose termination of nuclear import on the nucleoplasmic side of NPCs11,26; however, transient association with the NPC is not a feature of Nup214, which, similarly to Nup50 and Nup153, has been suggested to be involved in the termination of export on the cytoplasmic side of the NPC26. Furthermore, even the highest dissociation rates we measured (less than two Nup50 molecules were released from a single NPC each second; Table 1) are not sufficient to sustain import rates of more than 100 transport events per second, which have been measured27,28. This indicates that complete dissociation of Nup50 and Nup153 from the NPC is not necessary for every import cycle. To assess this directly, we performed iFRAP of cells transiently expressing the GFP-tagged import receptor importin β (Fig. 3b). In contrast to the most dynamic nucleoporin Nup50, importin β release from the NPC was limited by diffusion (see Supplementary Information, Fig. S3) and we can thus assign only a lower limit of about 0.5 s−1 on its dissociation rate, that is, at least more than one order of magnitude faster than the most dynamic nucleoporin we have measured (Table 1; also see note added in proof).

Although photobleaching techniques are popular for studying protein mobility, they have still rarely been used to measure the biochemical properties of protein interactions in living cells16,29. Dissociation rates of protein–protein interactions span ten orders of magnitude30, ranging from 10−7 s−1 to 103 s−1. We show here that, combined with quantitative data analysis by kinetic modelling, iFRAPs can determine a large subset of those dissociation rates in living cells, from 10−6 s−1 to 10−1 s−1. Lower dissociation rates would need several days to monitor fluorescence redistribution, whereas higher dissociation rates would become diffusion-limited and would have to be analysed by other methods. The NPC turned out to be a good illustration of the power of this method, as we measured dissociation rates covering the entire theoretical range (Fig. 5a). Combined with the known ultrastructural organization of the NPC, we now have a systematic map of the majority of the components that dynamically organize of a single protein complex in hand (Fig. 5b). The central part, with the notable exception of gp210, is a very stable structural scaffold that is likely to form the first seed to assemble the complex18,19. All studied WD-repeat-containing nucleoporins belonged to this class (Table 1), probably reflecting the role of this domain in establishing the multiple protein–protein interactions required for assembling the NPC scaffold. In contrast, most FG-repeat-containing nucleoporins were more dynamically associated with NPCs, consistent with the FG regions being basically unstructured12 and having no specific function in subcomplex stability or inter-subcomplex interactions. Those peripheral exchangeable nucleoporins may thus perform adaptor, regulatory and/or transport functions and most probably have only a secondary role in assembly of the NPC. In general, iFRAPs are straightforward to implement on laser scanning confocal microscopes and the data analysis by kinetic modelling as presented here is generally applicable as long as the studied protein complexes can be traced during the time-course of the experiment. Applied systematically to many supramolecular assemblies, this approach has the power to provide important parameters to model how the cellular machinery works in the context of the complex system of the intact cell.

Note added in proof: Using single-molecule fluorescence microscopy, Yang et. al (Proc. Natl Acad. Sci. USA 101, 12887–12892; 2004) have recently shown that the import substrate NLS–2×GFP interacts with NPCs for an average of 10 ± 1 ms during transport.


Plasmids and cell culture.

All plasmids used in this study are based on pEGFP-Nx or pEGFP-Cx vectors (BD Biosciences, Palo Alto, CA) and were made by placing the indicated number of GFPs in frame at the carboxyl or amino terminus of the full-length cDNAs of the nucleoporins of interest (see Table 1), except for gp210–EGFP3–CT where the GFPs were placed between the transmembrane and lumenal domains of the protein. Stable lines using normal rat kidney cells (NRK) were produced and maintained as described4 with the constructs that showed correct localization when transiently expressed.

Western blots and quantification of GFP-tagged nucleoporins localizing to NPCs.

To verify expression of full-length GFP-tagged nucleoporins (Fig. 1), cells were extracted essentially as described6 (see Supplementary Information, Methods). To compare stoichiometry of endogenous and GFP-tagged nucleoporins at NPCs (Fig. 2), we prepared a crude nuclear envelope extract (see Supplementary Information, Methods). The blots were probed with an anti-GFP monoclonal antibody mix (1:1,000; Roche, Basel, Switzerland), monoclonal antibody 414 (1:2,000; Covance Research Products, Berkeley, CA), anti-Pom121 (1:2,000) or anti-hNup107-C6 (1:5,000).

To measure the number of GFP-tagged nucleoporins localizing to single NPCs, we compared fluorescence intensity of individual NPCs with the fluorescence intensity of rotavirus-like particles known to contain 120 GFP molecules, essentially as described15 (see Supplementary Information, Methods for details).

iFRAP experiments.

For live-cell imaging, the cells were prepared as described4. iFRAPs were performed on a customized LSM510 confocal microscope4 using a 40× 1.3 NA oil Fluar objective and a completely open pinhole. For experiments lasting more than 1 h, 10 μg ml−1 of cycloheximide (Calbiochem, Merck Biosciences, Darmstadt, Germany) was added to the medium to avoid new synthesis of GFP-tagged nucleoporins. Most cellular fluorescence was bleached using full laser power and leaving only a fraction (typically 10–20%) of NPCs unbleached. Post-bleach images were then acquired for up to 10 h at regular time intervals, depending on the dynamics of the nucleoporin. Image acquisition was semi-automated, using a visual basic macro developed for the LSM software that automatically corrected focus fluctuations, tracked moving cells and parallelized image acquisition at multiple locations31.

Kinetic modelling.

To model the interaction of free ligands (nucleoporins) with their receptor (NPCs), we considered the reversible bimolecular reaction:

where kon and koff are the association and dissociation rates of the ligand L with its receptor R. We assume L to be fluorescently tagged, R to be anchored in a traceable multi-protein complex, diffusion not to be limiting and the system to be at steady state at the beginning of the experiment. The steady state is then perturbed by bleaching a fraction of L and LR. After the bleach, we consider the interactions of the fluorescent ligand (Lf) with its receptors in the bleached (Rb) and unbleached regions (Ru):

It can then be demonstrated (see Supplementary Information, Methods) that:


Thus, under the above assumptions, the fluorescence difference between the unbleached and the bleached regions follows an exponential decay whose rate is the dissociation rate of the considered bimolecular reaction. Similarly, if the ligand interacts with distinct receptors, or can change its dissociation rate from a single receptor, the fluorescence variation follows a multi-exponential decay.

Image processing and analysis.

Images of iFRAP time series were first registered to correct for cellular movements (translations and rotations) using either the TurboReg plugin ( for ImageJ ( or the Autoaligner software (Bitplane, Zürich, Switzerland). Then, for each time point the average fluorescence intensities of unbleached (Fu(t)) and bleached (Fb(t)) NPCs was measured in the respective regions. The background (BG(t))-subtracted intensities were normalized by the corresponding background-subtracted pre-bleach values (Fu(tpre)-BG(tpre) and Fb(tpre)-BG(tpre) respectively), and if necessary corrected for acquisition photobleaching. To obtain the dissociation kinetics (D(t)), we subtracted the fluorescence of the bleached NPCs from that of the unbleached NPCs. Its initial time was set to the time of acquisition of the first post-bleach image (tpost) and its initial intensity to 1. The dissociation kinetic is thus given by the formula:

To obtain the dissociation rate of the NPC-bound nucleoporins, the dissociation kinetics of single iFRAP experiments were fitted with a single or a bi-exponential decay using the Berkeley Madonna software ( To evaluate the quality of the fits we checked that they did not systematically deviate from the data by plotting the average residuals (see Supplementary Information, Fig. S4). Single exponential decay fitted most observed dissociation kinetics (for example, Pom121; Fig. 4b; also see Supplementary Information, Fig. S4a). Only Nup50 and Nup153 exhibited obvious bi-exponential decays (Fig. 4c; also see Supplementary Information, Fig. S4b) indicating that those nucleoporins were found in at least two states within the NPCs. We could not confidently distinguish multiple slow dissociation rates for all other nucleoporins but, owing to the noise of individual datasets, we cannot exclude their presence. For instance, the dissociation kinetics of Nup62 can also be modelled as the sum of the dissociations of Nup214 and Nup58 (Fig. 4d; also see Supplementary Information, Fig. S4c).

Note: Supplementary Information is available on the Nature Cell Biology website.


  1. 1

    Hartwell, L. H., Hopfield, J. J., Leibler, S. & Murray, A. W. From molecular to modular cell biology. Nature 402, C47–C52 (1999).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  2. 2

    Cronshaw, J. M., Krutchinsky, A. N., Zhang, W., Chait, B. T. & Matunis, M. J. Proteomic analysis of the mammalian nuclear pore complex. J. Cell Biol. 158, 915–927 (2002).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  3. 3

    Suntharalingam, M. & Wente, S. R. Peering through the pore: nuclear pore complex structure, assembly, and function. Dev. Cell 4, 775–789 (2003).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  4. 4

    Daigle, N. et al. Nuclear pore complexes form immobile networks and have a very low turnover in live mammalian cells. J. Cell Biol. 154, 71–84 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  5. 5

    Enninga, J., Levay, A. & Fontoura, B. M. Sec13 shuttles between the nucleus and the cytoplasm and stably interacts with Nup96 at the nuclear pore complex. Mol. Cell. Biol. 23, 7271–7284 (2003).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  6. 6

    Belgareh, N. et al. An evolutionarily conserved NPC subcomplex, which redistributes in part to kinetochores in mammalian cells. J. Cell Biol. 154, 1147–1160 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  7. 7

    Nakielny, S., Shaikh, S., Burke, B. & Dreyfuss, G. Nup153 is an M9-containing mobile nucleoporin with a novel Ran-binding domain. EMBO J. 18, 1982–1995 (1999).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  8. 8

    Griffis, E. R., Altan, N., Lippincott-Schwartz, J. & Powers, M. A. Nup98 is a mobile nucleoporin with transcription-dependent dynamics. Mol. Biol. Cell 13, 1282–1297 (2002).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  9. 9

    Zolotukhin, A. S. & Felber, B. K. Nucleoporins nup98 and nup214 participate in nuclear export of human immunodeficiency virus type 1 Rev. J. Virol. 73, 120–127 (1999).

    CAS  PubMed  PubMed Central  Google Scholar 

  10. 10

    Dilworth, D. J. et al. Nup2p dynamically associates with the distal regions of the yeast nuclear pore complex. J. Cell Biol. 153, 1465–1478 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  11. 11

    Lindsay, M. E., Plafker, K., Smith, A. E., Clurman, B. E. & Macara, I. G. Npap60/Nup50 is a tri-stable switch that stimulates importin-alpha:beta-mediated nuclear protein import. Cell 110, 349–360 (2002).

    CAS  Article  Google Scholar 

  12. 12

    Denning, D. et al. The nucleoporin Nup60p functions as a Gsp1p–GTP-sensitive tether for Nup2p at the nuclear pore complex. J. Cell Biol. 154, 937–950 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  13. 13

    Guan, T. et al. Nup50, a nucleoplasmically oriented nucleoporin with a role in nuclear protein export. Mol. Cell. Biol. 20, 5619–5630 (2000).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  14. 14

    Bodoor, K. et al. Sequential recruitment of NPC proteins to the nuclear periphery at the end of mitosis. J. Cell Sci. 112, 2253–2264 (1999).

    CAS  Google Scholar 

  15. 15

    Dundr, M., McNally, J. G., Cohen, J. & Misteli, T. Quantitation of GFP-fusion proteins in single living cells. J. Struct. Biol. 140, 92–99 (2002).

    CAS  Article  Google Scholar 

  16. 16

    Dundr, M. et al. A kinetic framework for a mammalian RNA polymerase in vivo. Science 298, 1623–1626 (2002).

    CAS  Article  Google Scholar 

  17. 17

    Loiodice, I. et al. The entire nup107–160 complex, including three new members, is targeted as one entity to kinetochores in mitosis. Mol. Biol. Cell 15, 3333–3344 (2004).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  18. 18

    Harel, A. et al. Removal of a single pore subcomplex results in vertebrate nuclei devoid of nuclear pores. Mol. Cell 11, 853–864 (2003).

    CAS  Article  Google Scholar 

  19. 19

    Walther, T. C. et al. The conserved Nup107–160 complex is critical for nuclear pore complex assembly. Cell 113, 195–206 (2003).

    CAS  Article  Google Scholar 

  20. 20

    Walther, T. C. et al. The cytoplasmic filaments of the nuclear pore complex are dispensable for selective nuclear protein import. J. Cell Biol. 158, 63–77 (2002).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  21. 21

    Cronshaw, J. M. & Matunis, M. J. The nuclear pore complex protein ALADIN is mislocalized in triple A syndrome. Proc. Natl Acad. Sci. USA 100, 5823–5827 (2003).

    CAS  Article  PubMed  Google Scholar 

  22. 22

    Drummond, S. P. & Wilson, K. L. Interference with the cytoplasmic tail of gp210 disrupts “close apposition” of nuclear membranes and blocks nuclear pore dilation. J. Cell Biol. 158, 53–62 (2002).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  23. 23

    Olsson, M., Scheele, S. & Ekblom, P. Limited expression of nuclear pore membrane glycoprotein 210 in cell lines and tissues suggests cell-type specific nuclear pores in metazoans. Exp. Cell Res. 292, 359–370 (2004).

    CAS  Article  PubMed  Google Scholar 

  24. 24

    Galy, V., Mattaj, I. W. & Askjaer, P. Caenorhabditis elegans nucleoporins Nup93 and Nup205 determine the limit of nuclear pore complex size exclusion in vivo. Mol. Biol. Cell 14, 5104–5115 (2003).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  25. 25

    Griffis, E. R., Craige, B., Dimaano, C., Ullman, K. S. & Powers, M. A. Distinct functional domains within nucleoporins Nup153 and Nup98 mediate transcription-dependent mobility. Mol. Biol. Cell 15, 1991–2002 (2004).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  26. 26

    Macara, I. G. Transport into and out of the nucleus. Microbiol. Mol. Biol. Rev. 65, 570–594 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  27. 27

    Smith, A. E., Slepchenko, B. M., Schaff, J. C., Loew, L. M. & Macara, I. G. Systems analysis of Ran transport. Science 295, 488–491 (2002).

    CAS  Article  PubMed  Google Scholar 

  28. 28

    Ribbeck, K. & Gorlich, D. Kinetic analysis of translocation through nuclear pore complexes. EMBO J. 20, 1320–1330 (2001).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  29. 29

    Carrero, G., McDonald, D., Crawford, E., de Vries, G. & Hendzel, M. J. Using FRAP and mathematical modeling to determine the in vivo kinetics of nuclear proteins. Methods 29, 14–28 (2003).

    CAS  Article  PubMed  Google Scholar 

  30. 30

    Janin, J. in Protein–Protein Recognition (ed. Kleanthous, C.) 1–32 (Oxford University Press, New York, 2000).

    Google Scholar 

  31. 31

    Rabut, G. & Ellenberg, J. J. Microsc. Automatic real-time 3D cell tracking by fluorescence microscopy. 216, 131–137 (2004).

    CAS  Article  Google Scholar 

Download references


We would like to thank E. Izaurralde for CG1, Nup93, Nup98 and importin β cDNAs; B. Glick for pBK-SEC13–GFP; D. Görlich for Nup50 and Nup58 cDNAs; M. Fornerod for pBS-KS(-)CAN; R. Wozniak for the gp210 cDNA; J. Hanover for the Nup62 cDNA; J. Cohen for GFP–VLPs; I. Loïodice for EGFP–Seh1; M. Van Overbeek for EGFP–Nup85; J. Cronshaw and M. Matunis for EGFP–Nup37, EGFP–Nup43, EGFP–Nup35 and EGFP–Aladin; E. Hallberg for the anti-Pom121 antibody; I. Mattaj, E. Izaurralde and the members of the Ellenberg lab for critical reading of the manuscript; and J.C. Courvalin for stimulating discussions. J.E. acknowledges support from the German Research Council (DFG, EL 246/1-1) and the Human Frontiers Science Program (RGP0031/2001-M). V.D. is funded by the Institut Curie, the Centre National de la Recherche Scientifique, the Association pour la Recherche contre le Cancer, and la Ligue Contre le Cancer (Comité de Paris).

Author information



Corresponding author

Correspondence to Jan Ellenberg.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Supplementary Information, Movie S1 (MOV 2677 kb)

Supplementary Information, Figures

Supplementary Information, Fig. S1, Fig. S2, Fig. S3, Fig. S4 and Supplementary Methods (PDF 818 kb)

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Rabut, G., Doye, V. & Ellenberg, J. Mapping the dynamic organization of the nuclear pore complex inside single living cells. Nat Cell Biol 6, 1114–1121 (2004).

Download citation

Further reading


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing