nucleoporin domain topology is linked to the transport status of the nuclear pore complex

doi:10.1016/j.jmb.2005.06.034 J. Mol. Biol. (2005) 351, 784–798

Nucleoporin Domain Topology is Linked to theTransport Status of the Nuclear Pore Complex

Sara M. Paulillo1, Erica M. Phillips2, Joachim Koser1, Ursula Sauder1

Katharine S. Ullman3, Maureen A. Powers2 and Birthe Fahrenkrog1*

1M.E. Muller Institute forStructural Biology, BiozentrumUniversity of BaselKlingelbergstr. 70, 4056 BaselSwitzerland

2Department of Cell BiologyEmory University School ofMedicine, 615 Michael StreetAtlanta, GA 30322, USA

3Department of OncologicalSciences, Huntsman CancerInstitute, University of UtahSalt Lake City, UT 84112USA

0022-2836/$ - see front matter q 2005 E

Abbreviations used: NPC, nucleaphenylalanine-glycine; EM, electronnucleoporin; AL, annulate lumellaeescent protein.E-mail address of the correspond

[email protected]

Nuclear pore complexes (NPCs) facilitate macromolecular exchangebetween the nucleus and cytoplasm of eukaryotic cells. The vertebrateNPC is composed of w30 different proteins (nucleoporins), of whicharound one third contain phenylalanine-glycine (FG)-repeat domains thatare thought to mediate the main interaction between the NPC and solubletransport receptors. We have recently shown that the FG-repeat domain ofNup153 is flexible within the NPC, although this nucleoporin is anchoredto the nuclear side of the NPC. By using domain-specific antibodies, wehave now mapped the domain topology of Nup214 in Xenopus oocytes andin human somatic cells by immuno-EM. We have found that whereasNup214 is anchored to the cytoplasmic side of the NPC via its N-terminaland central domain, its FG-repeat domain appears flexible, residing onboth sides of the NPC. Moreover, the spatial distribution of the FG-repeatdomains of both Nup153 and Nup214 shifts in a transport-dependentmanner, suggesting that the location of FG-repeat domains within the NPCcorrelates with cargo/receptor interactions and that they concomitantlymove with cargo through the central pore of the NPC.

q 2005 Elsevier Ltd. All rights reserved.

Keywords: nuclear pore complex; nuclear transport; FG-repeats; Nup214;Nup153
*Corresponding author
Introduction

Nuclear pore complexes (NPCs) facilitate theexchange of proteins and RNPs between the nucleusand cytoplasm of eukaryotic cells.1,2 The vertebratew125 MDa NPC3 is composed of w30 differentproteins4 collectively termed nucleoporins (Nups).Extensive electron microscopy (EM) studies mainlyin Xenopus oocytes have led to a draft of the 3-DNPC architecture.5–8 Accordingly, the NPC iscomposed of an 8-fold symmetric central frame-work that is continuous with a cytoplasmic and anuclear ring moiety. Eight cytoplasmic filamentsdecorate the cytoplasmic ring, whereas the nuclearring is topped by the nuclear basket. Enclosed bythe central framework is the central pore, w90 nmlong and 40–50 nm in diameter in its narrowest

lsevier Ltd. All rights reserve

r pore complex; FG,microscopy; Nup,; GFP, green fluor-

ing author:

part,5,9 which allows diffusion of small ions andmolecules and facilitated, signal-dependent nucleartransport in and out of the nucleus.10–12

For typical facilitated nuclear transport, receptorsof the karyopherin family recognize nuclear importsignals (NLSs) or nuclear export signals (NESs) ontheir cargo.1,2 The transport receptor, in turn, is ableto mediate the interaction of the cargo–receptorcomplex with the NPC. It is thought that the main,albeit not the exclusive,13 interaction between thetransport receptors and the NPC is mediated by asubset of nucleoporins that harbor FG (phenyl-alanine-glycine)-repeat domains. Based on X-raycrystallography and biophysical analysis,FG-repeat domains exhibit little secondarystructure.14–20

Immuno-EM studies of the vertebrate nucleo-porin Nup153 demonstrated a high degree ofmobility and structural flexibility of its FG-repeatdomain.21 Nup153 is a constituent of the nuclearbasket of the NPC22 and immuno-EM analysisindicates that its N-terminal domain is anchoredto the nuclear ring,21,23 whereas its central zinc-finger domain resides at the distal ring.21 Thew700

d.

Transport-dependent Nucleoporin Domain Topology 785

amino acid residues C-terminal domain of Nup153harbors w40 FG-repeats, and seems to be flexiblewithin the NPC, since it can be detected at thenuclear basket and even at the cytoplasmicperiphery of the central pore.21 Nup153 is knownto be critical for both nuclear import and nuclearexport.23–27

The vertebrate nucleoporin Nup214, also calledCAN, is an FG-repeat nucleoporin previouslymapped to the cytoplasmic face of the NPC.22,23,28

Like Nup153, Nup214 participates in both nuclearimport and export.29 Nup214 depletion, however,does not lead to a complete block of nuclearimport.29,30 Based on its amino acid sequence,Nup214 can be subdivided into three distinctdomains: an N-terminal domain, a central domainwith a region of predicted coiled-coil at its Nterminus, and a C-terminal FG-repeat domain thatspans w800 amino acid residues31 and mediatesinteractions with various transport factors.31–34

Based on the structural flexibility and mobility ofthe FG-repeat domain of Nup153, we askedwhether this might be a general attribute ofFG-repeat domains, in particular those ofperipheral nucleoporins such as Nup214. We there-fore aimed to map the domain topology of Nup214within the 3-D architecture of the NPC. To achievethis, we raised three antibodies against the distinctdomains of human Nup214 and determined theprecise location of these domains in intact Xenopusnuclei as well as in somatic human cells byimmuno-EM. Our data clearly show that the threedomains of Nup214 are located at distinct siteswithin the NPC and that the FG-repeat domain ofNup214 is flexible within the NPC. Moreover,our data demonstrate that the localization ofthe FG-repeat domains of both Nup153 andNup214 is responsive to transport activity at thenuclear pore.

Results

Production and characterization of domain-specific antibodies to Nup214

In order to gain a better understanding of thetopology and domain accessibility of Nup214within the NPC, a number of domain-specificantibodies were generated (see Materials andMethods and Figure 1(a)). To characterize thespecificity of these antibodies, their reactivity wastested in an immunoblot (Figure 1(b)). Eachantibody selectively recognized Nup214 in HeLacell extracts, Xenopus egg extracts and amongisolatedXenopusWGA-binding proteins (Figure 1(b),lanes 1–3 each panel). The Nup214-A antibodyrecognizes an additional protein of w50 kDa inXenopus egg extracts (Figure 1(b), arrow head), butnot in HeLa cell extracts. This w50 kDa protein isunlikely to be NPC-associated, however, as it is notdetected in an NPC-rich annulate lamellae (AL)fraction (see 214A, Figure 1(b)). All antibodiesrecognize to some extent a protein of w40 kDa(Figure 1(b), asterisk). This band is still present inAL prepared in the presence of BAPTA to preventNPC assembly (most noticeable in the 214D blot,Figure 1(b)), suggesting that this band is a cross-reacting membrane protein. Together, these testsindicate that, while some cross-reactivity cannot beruled out, reactivity at the NPC itself should clearlyreflect the presence and domain exposure ofNup214.

The N-terminal domain of Nup214 localizes tothe cytoplasmic side of the NPC

To localize the distinct domains of Nup214 withinthe NPC, we used domain-specific antibodies for

Figure 1. Domain-specificNup214 antibodies. (a) Schematicrepresentation of Nup214. Bluelines indicate the positions ofindividual FG-repeats. Recombi-nant domains used as immuno-gens are indicated (see Materialsand Methods). (b) Nup214 anti-bodies were used to probeimmunoblots. HeLa cell extract(H), Xenopus egg extracts (X),Xenopus WGA-binding proteins(W) and Xenopus annulatelamellae (AL) assembled in thepresence or absence of the inhibi-tor BAPTA (G) were separatedby SDS-PAGE and processed forimmunoblotting (see Materialsand Methods).

Table 1. Comparison of the location of different Nup214domains in Xenopus nuclei versus HL-60 cells

Antibody Xenopus laevis HL-60 cells

Anti-Nup214-A C31 nmG6 nm C20 nmG13 nmAnti-Nup214-B C21 nmG12 nm C9 nmG10 nmAnti-Nup214-D C11 nmG10 nm C14 nmG12 nmAnti-Nup214-D K29nmG12 nm K30 nmG12 nmAnti-Nup214-D K86 nmG13 nm K75 nmG18 nm

Numbers represent the distance from the central plane of theNPC. C, cytoplasmic; K, nuclear.

786 Transport-dependent Nucleoporin Domain Topology

immuno-EM. Although the antibodies were raisedagainst human Nup214, Xenopus oocyte nuclei wereused due to the better structural resolution andhigher density of NPCs in comparison to somaticcells. Intact nuclei were isolated and incubated withanti-Nup214-A antibody, a secondary anti-rabbitIgG antibody conjugated to 10 nm colloidal gold,and then analyzed by thin section immuno-EM. Wefound the N-terminal domain of Nup214 to resideclose to the nuclear membrane at a mean distancefrom the central plane of the NE of 31 (G6) nm(Table 1), corresponding to an epitope at thecytoplasmic ring of the NPC. The anti-Nup214-Aantibody showed some degree of cross-reactivity inXenopus nuclei with an epitope at the nuclear basket(data not shown). Based on Western blot analysis(Figure 1(b)) we suppose this to be the protein ofw50 kDa, which is recognized by the antibody inXenopus but not in human.

To confirm the exclusive cytoplasmic location ofthe N-terminal domain of Nup214 we determinedits location in HL-60 cells, human promyelocyticleukocytes. As shown in Figure 2(a), in these humansomatic cells the anti-Nup214-A recognized anepitope exclusively on the cytoplasmic side of theNE. Quantification of the gold particle distribution(Figure 2(b) and Table 1) revealed a meandistance of 20(G13) nm from the central plane ofthe NE and a corresponding average radial distanceof 15(G13) nm, representing an epitope near or atthe cytoplasmic ring of the NPC.

Location of the central domain of Nup214

We next examined the location of the centraldomain of Nup214 within the NPC. Intact Xenopusoocyte nuclei were isolated and incubated withanti-Nup214-B (see Figure 1), a secondary anti-rabbit IgG antibody conjugated to 10 nm colloidalgold and analyzed by thin section immuno-EM.

As shown in Figure 2(c), the anti-Nup214-Bantibody exclusively labeled the cytoplasmic faceof the NPC. Quantification of the gold particledistribution (Figure 2(d) and Table 1) revealed anaverage distance of 21(G12) nm from the centralplane of the NE. A corresponding mean radialdistance of 17(G15) nm indicated that this locationresides near or at the cytoplasmic ring of the NPCnear the entry to the central pore. These results wereconfirmed in HL-60 cells (see Table 1).

The C-terminal domain of Nup214 is flexible

Next, we wanted to resolve the location of theC-terminal FG-repeat domain of Nup214. As before,isolated intact Xenopus nuclei were incubated withthe repeat domain antibody (anti-Nup214-D) and asecondary anti-rabbit IgG antibody conjugated to10 nm colloidal gold. The epitope recognized by theantibody was analyzed by thin section immuno-EM.

As shown in Figure 2(e), the anti-Nup214-Drecognized epitopes on both the cytoplasmic andthe nuclear face of the NPC. Quantification of thelabeling pattern (Figure 2(f) and Table 1) relative tothe central plane of the NPC revealed that 87% ofthe particles were detected at a mean distance of11(G10) nm. With corresponding mean radialdistances of 12(G11) nm, this labeling pattern isconsistent with a location near or at the cytoplasmicring. 13% of the gold particles were found on thenuclear side of the NPC, of which 2% were detectedat a mean distance of K29(G12) nm and 11% atK86(G12) nm, reflecting epitopes within thenuclear basket of the NPC. Again, these resultswere confirmed in HL-60 cells and a similardistribution was observed (see Table 1). Moreover,similar results were obtained in Xenopus nucleiwith a distinct anti-peptide antibody raised againstthe FG-repeat domain of human Nup214(R. Kehlenbach and B.F., unpublished results).Taken together, the multiple locations of theC-terminal, FG-repeat domain of Nup214 indicatea flexible topology of this domain.

Recombinant expression of epitope-taggedhNup214 into Xenopus NPCs

Previously we have shown that tagged hNup153incorporates into the NPC after microinjection of aplasmid encoding epitope-tagged hNup153 into thenuclei of Xenopus oocytes.21 We, therefore, aimed toconfirm the location of the N terminus and themultiple locations of the C terminus of Nup214 bysimilarly employing epitope-tagged versions ofNup214. To achieve this, we generated a plasmidencoding hNup214 with either a N-terminal greenfluorescent protein (GFP)-tag (GFP-Nup214) or aC-terminal myc-tag (Nup214-myc). By transienttransfection of different cultured cell lines we firstverified that the fusion proteins were expressed andincorporated into the NPC (Supplementary FigureS1). Next, we microinjected the plasmids into thenuclei of Xenopus oocytes. To determine the locationof the incorporated proteins we used either apolyclonal anti-GFP antibody or a monoclonalanti-myc antibody, both directly conjugated to8 nm colloidal gold.

As shown in Figure 3(a), the N-terminal GFP-taglocalizes predominantly to the cytoplasmic face ofthe NPC (mean distance from the central planeof the NPC 25(G9) nm with radial distances of12(G10) nm). However, about 23% of the goldparticles associated with the nuclear side of the

Figure 2. Domain-topology of Nup214 within the NPC. (a) HL-60 cells were pre-embedding labeled with anti-Nup214-A antibody and a secondary anti-rabbit IgG antibody conjugated to 10 nm gold and prepared for thin sectionEM. Shown are selected examples of gold-labeled NPCs in cross-section. c, cytoplasm; n, nucleus. (b) Quantitativeanalysis of the gold particle distribution associated with the NPC. 36 gold particles were scored. (c) Intact isolatedXenopus nuclei were incubated with anti-Nup214-B antibody and (e) anti-Nup214-D antibody and a secondary anti-rabbit IgG antibody conjugated to 10 nm colloidal gold. Shown are a stretch along the NE in cross-sections (top) and agallery of selected examples of gold-labeled NPCs (bottom). c, cytoplasm; n, nucleus. (d) Quantification of the goldparticles associated with the NPC after labeling with the anti-Nup214-B antibody and (f) with the anti-Nup214-Dantibody. A total of 166 and 259, respectively, gold particles were scored. Scale bars represent 100 nm.


Figure 3. Incorporation of tagged hNup214 into Xenopus NPCs. (a) Immunolocalization of N-terminally GFP-taggedNup214 (GFP-Nup214) expressed in Xenopus oocytes with a polyclonal anti-GFP antibody directly conjugated to 8 nmcolloidal gold. Selected examples of labeled NPCs in cross-sections are shown. c, cytoplasm; n, nucleus.(b) Quantification of the gold particle distribution associated with the NPC after labeling with an anti-GFP antibody.Ninety-seven gold particles were scored. (c) Immunolocalization of C-terminally myc-tagged Nup214 (Nup214-myc)expressed in Xenopus oocytes with a monoclonal anti-myc antibody directly conjugated to 8 nm colloidal gold. Selectedexamples of gold-labeled NPCs are shown. c, cytoplasm; n, nucleus. (d) Quantitative analysis of the distribution of goldparticles associated with the NPC after labeling with an anti-myc antibody. Ninety-two gold particles were scored.(e) Schematic representation of the domain-topology of Nup214 in the 3-D architecture of the NPC. Nup214 is anchoredto the cytoplasmic ring of the NPC by its N-terminal and central-domain (red), whereas the C-terminal FG-repeatdomain (yellow) is flexible and can be mapped at the cytoplasmic face of the NPC (left), the nuclear periphery of thecentral pore (middle), and even at the distal ring of the nuclear basket (right). c, cytoplasm; n, nucleus. Scale barsrepresent 100 nm.


NPC (Figure 3(b)). This nuclear signal may be duein part to some mislocalization previously shown tobe caused by overexpression.35

As shown in Figure 3(c), the C-terminal myc-taglocalizes to both faces of the NPC. Quantification ofthe labeling distribution revealed that 63% of the

gold particles were associated with the cytoplasmicside of the NPC (mean distance from the centralplane of the NPC 31(G11) nmwith a correspondingradial distance of 19(G13) nm). 37% of the goldparticles labeled the nuclear face of the NPC. Ofthese, 13% of the gold particles were found at a


mean distance of K40(G7) nm, and 24% at a meandistance of K76(G14) nm from the central planerepresenting locations within the nuclear basket ofthe NPC. Taking into account that some goldparticles might be associated with the nuclear faceof the NPC due to overexpression (w23%; seeabove), we estimate thatw14% of the gold particlesshould reflect specific location of this region to thenuclear side of the NPC. This extent of C-terminalmyc-tag detection is similar to the labeling seenwith the domain-specific antibody against theFG-repeat domain of Nup214 (anti-Nup214-D; seeabove and Figure 2(e) and (f)).

Taken together, our immuno-EM data usingdomain-specific antibodies and incorporation ofepitope-tagged hNup214 into Xenopus oocyte NPCssuggest that the N-terminal and central region ofNup214 are anchored near or at the cytoplasmicring of the NPC, whereas the C-terminal FG-repeatdomain is flexible and present on both thecytoplasmic and nuclear periphery of the NPC(Figure 3(e)).

Attenuation of nuclear transport constrainsFG-repeat domains to their anchoring site

The pronounced heterogeneous distribution ofFG-repeat domains could be due to mobility that islinked to nuclear transport activity, especially incargo translocation through the central pore of theNPC. This hypothesis predicts that the location ofthe FG-repeat domains within the NPC would beinfluenced by the transport activity of the NPC. Wetherefore analyzed whether the location of theFG-repeat domains of the nucleoporins Nup153and Nup214 is dependent on the transport state ofthe cell.

Nuclear transport is readily attenuated at 4 8C.36

To study the influence of attenuated nucleartransport on the location of the FG-repeat domainsof Nup153 and Nup214, nuclei were isolated fromXenopus oocytes and incubated at 4 8C for one hour.After fixation in 4% formaldehyde, the nuclei werelabeled with anti-Nup153-C2 antibody21 or anti-Nup214-D antibody to determine the location of theFG-repeat domains of these nucleoporins underthese conditions.

At steady-state, the FG-repeat domain of Nup153localizes to the nuclear basket as well as to someextent (w12%) to the cytoplasmic periphery of thecentral pore.21 In contrast, when nuclear transport isarrested, the anti-Nup153-C2 antibody exclusivelydecorates the nuclear face of the NPC (Figure 4(a)).76% of the gold particles were found in the area ofthe distal ring of the NPC (mean distance from thecentral plane of the NPCK77(G12) nm) and 24% ofthe gold particles were closer to the nuclear ring(mean distance of K36(G9) nm from the centralplane) (Figure 4(b)). Under the same conditions, theexpression level of Nup153 does not change withinthe nuclei and the location of the N-terminal andthe central Zn-finger domain of Nup153 are not

altered in comparison to steady-state conditions(Supplementary Figure S2A).The FG-repeat domain of Nup214 localizes to the

cytoplasmic periphery of the central pore and,additionally, to the nuclear side of the NPC atsteady-state (Figures 2 and 3). At 4 8C, when nucleartransport is attenuated, the FG-repeat domain ofNup214 localizes exclusively to the cytoplasmicperiphery of the central pore at a mean distance of25(G11) nm (Figure 4(d)). Attenuation of nucleartransport at 4 8C has no effect on the expressionlevel of Nup214 in the oocyte nuclei or significantlyinfluenced the location of the central domain ofNup214 (Supplementary Figure S2B).Similar results for Nup153 were obtained when

nuclear transport was attenuated by either WGA ora dominant-negative mutant of importin b,importinb45–462, known to inhibit distinct nucleartransport pathways37 (Supplementary Figure S3A).In the case of Nup214, however, the distribution ofthe FG-repeat domain after attenuating nucleartransport by either WGA or importinb45–462 wassimilar to steady-state conditions (SupplementaryFigure S3B). Therefore, attenuating nuclear trans-port at 4 8C constrains the distribution of the FG-repeat domains of Nup153 as well as Nup214 to thesite at which the corresponding nucleoporin isanchored within the NPC, whereas WGA andimportinb45–462 have different influence on thedistribution of the FG-repeat domains of Nup153 ascompared to Nup214.

Nuclear import cargo influences the localizationof FG-repeat domains

Next we attempted to interfere with steady-statenuclear transport by challenging the systemwith anexcess of import cargo and then analyzing how thisinfluences the location of the FG-repeat domains ofNup153 and Nup214 within the NPC. As importcargo we used nucleoplasmin, a nuclear proteinthat readily becomes imported in Xenopusoocytes.38,39 To study the effect of import cargo onthe location of the FG-repeat domains of Nup153and Nup214, we isolated nuclei from Xenopusoocytes and incubated them in an import competentmixture containing excess nucleoplasmin. Thenuclei were incubated for various lengths of timein this import mixture and then labeled with anti-Nup153-C2 or anti-Nup214-D antibody to deter-mine the location of these FG-repeat domains.As shown in Figure 5(a) (and Supplementary

Figure S4A), after five minutes of incubation, theFG-repeat domain of Nup153 localizes to both sidesof the NPC, as observed previously.21 In compari-son to steady-state conditions, however, the numberof NPCs that are labeled on their cytoplasmic sidehas significantly increased to 39% in the presence ofexcessive nucleoplasmin versus 12% at steady-state.21 The percentage of cytoplasmic labelingfurther increases to 50% after 15 minutes ofincubation (Figure 5(a)). Finally, after 30 minutesincubation the FG-repeat domain of Nup153

Figure 4. Attenuated nuclear transport constrains the FG-repeat domains of Nup153 and Nup214 to their anchoringsite in the NPC. (a) Isolated Xenopus nuclei were incubated at 4 8C prior to labeling with an antibody against theFG-repeat domain of Nup153 (anti-C2-Nup153) directly conjugated to 8 nm colloidal gold. c, cytoplasm; n, nucleus.(b) Quantitative analysis of the gold particles associated with the NPC after labeling with the anti-C2-Nup153 antibodyat 4 8C. Eighty-eight gold particles were scored. (c) Immunolabeling of isolated Xenopus nuclei after incubation at 4 8Cwith the anti-Nup214-D antibody and a secondary anti-rabbit IgG antibody conjugated to 10 nm colloidal gold. c,cytoplasm; n, nucleus. (d) Quantification of the gold particle distribution associated with the NPC after labeling with theanti-Nup214-D antibody at 4 8C. A total of 134 gold particles were scored. Scale bars represent 100 nm.


localizes exclusively to the nuclear face of the NPC(Figure 5(a)) with approximately 90% of the goldbeing located at the distal ring of the NPC. Importmixture alone without addition of nucleoplasminhas no influence on the location of the FG-repeatdomain of Nup153 (Supplementary Figure S5A).The location of the N-terminal and the central Zn-finger domain of Nup153 remained unchanged inthe presence of import mixture containing excessnucleoplasmin (Supplementary Figure S5B and C).

Under the same incubation conditions, the FG-repeat domain of Nup214 also shows a higherproportion of nucleoplasmic location in comparisonto steady-state conditions (Figure 5(b)). Whereas atsteady-state conditions about 13% of the NPCs were

labeled on the nuclear side (see above), afterincubation for five minutes in excess nucleoplasminthis fraction increased to 43%, and further to about60% after 15 minutes and 30 minutes (Figure 5(b);see also Supplementary Figure S4B). As the numberof gold particles associated with the nuclear face ofthe NPC increased, the cytoplasmic value decreasedcorrespondingly from w57% after five minutes tow33% after 30 minutes incubation with nucleo-plasmin.

Taken together, our data suggest that the localiz-ations of the FG-repeat domains of Nup153 andNup214 shift, corresponding to a wave of transportmoving through the NPC and possibly reflectingsteps that promote a certain arrangement of the

Figure 5. The location of the FG-repeat domains of Nup153 and Nup214 in the presence of excess import cargo.Quantitative analysis of the gold particles associated with the NPC after labeling with (a) the anti-C2-Nup153 antibodyand (b) the anti-Nup214-D antibody, respectively, in the presence of excess nucleoplasmin. Import of nucleoplasmin intoisolated intact Xenopus oocytes nuclei was allowed for the indicated time-points before the nuclei were shortly pre-fixedin formaldehyde and pre-embedding labeled with the corresponding antibody. The following number of gold particleswere scored for the individual experiments: 55 (anti-C2-Nup153, five minutes), 60 (anti-C2-Nup153, 15 minutes), 106(anti-C2-Nup153, 30 minutes), 132 (anti-Nup214-D, five minutes), 178 (anti-Nup214-D, 15 minutes), and 163 (anti-Nup214-D, 30 minutes).


FG-repeat domain in the bulk of transport events ateach particular time point.

The influence of RNA export on the localizationof FG-repeat domains

Next we wanted to examine how excess exportcargo would influence the location of the FG-repeatdomains of Nup153 and Nup214. We thereforeisolated poly(AC) RNA from HeLa cells andmicroinjected this RNA into the nuclei of Xenopusoocytes. Nuclei were isolated following variousincubation times and labeled with either the anti-Nup153-C2 antibody or the anti-Nup214-D anti-body. Since RNA export is slower than proteinimport,40 longer incubation times were performed.

As illustrated in Figure 6(a) (and SupplementaryFigure S6A), the FG-repeat domain of Nup153

localizes predominantly to the nuclear face of theNPC 30 minutes after microinjection of poly(AC)RNA into the nuclei of Xenopus oocytes. Comparedto steady-state conditions21 (Figure 6(a), first panel),however, the number of the gold particles associ-ated with the cytoplasmic side of the NPC increasedtow26% (Figure 6(a), secondpanel). The cytoplasmiclocation of the FG-repeat domain of Nup153 furtherincreased by 60 minutes of incubation with excesspoly(AC) RNA (Figure 6(a), third panel), when 37%of the gold particles were detected on this side of theNPC. By 90 minutes, the cytoplasmic proportion ofthe FG-repeat domain of Nup153 reached its maxi-mum to w63% (Figure 6(a), fourth panel).Microinjection of poly(AC) RNA into the nucleus

of Xenopus oocytes results in a predominant nuclearlocalization of the FG-repeats of Nup214 after30 minutes of incubation (Figure 6(b); see also

Figure 6. The influence of export cargo on the location of the FG-repeat domains of Nup153 and Nup214.Quantification of the gold particle distribution associated with the NPC after microinjection of poly(AC) RNA into thenuclei of Xenopus oocytes and pre-embedding labeling with the (a) anti-C2-Nup153 and the (b) anti-Nup214-D antibody.The following numbers of gold particles were scored for the individual experiments: 130 (anti-C2-Nup153, 30 minutes),104 (anti-C2-Nup153, 60 minutes), 49 (anti-C2-Nup153, 90 minutes), 80 (anti-Nup214-D, 30 minutes), 212 (anti-Nup214-D, 60 minutes), and 80 (anti-Nup214-D, 90 minutes).


Supplementary Figure S6B). 89% of the goldparticles were on the nuclear side of the NPC andpredominantly at or near the distal ring of the NPC,whereas only w11% of the gold particles werefound to be associated with the cytoplasmic face ofthe NPC. At 60 minutes after microinjection of RNAinto Xenopus nuclei (Figure 6(b), third panel) w80%of the gold particles were associated with thenuclear side of the NPC (67% of these in the areaof the central pore) and w20% associated with thecytoplasmic face.After 90minutes (Figure6(b), fourthpanel), w92% of the gold particles were detected onthe cytoplasmic face of the NPC and onlyw8% on itsnuclear side, similar to steady-state conditions.Largely analogous results were obtained after micro-injection of total RNA isolated from HeLa cells intothe nuclei of Xenopus oocytes, except for the 90 -minutes time point in the case of Nup153 (Sup-plementary Figures S7 and S8). This might reflect adifference inNup153 conformation in the presence ofexcess mRNA versus, for example, ribosomal RNA,which is present in the total RNA preparation.

In conclusion, the location of Nup153 and Nup214FG-repeat domains shifts concomitant with a trans-port wave of export cargo through the NPC.

Localization of FG-repeat domains in nuclei thatlack RNA export cargo

Next, we assessed whether depletion of cargowould also bias the location of the FG-repeatdomains of Nup153 and Nup214 within the NPC.To do so, we injected actinomycin D into the nucleiof Xenopus oocytes to inhibit endogenous RNApolymerases. After three hours incubation ofactinomycin D treatment, the location of the FG-repeat domains of Nup153 andNup214 in manuallyisolated Xenopus nuclei was determined byimmuno-EM.

As shown in Figure 7(a) (see also SupplementaryFigure S9A), depletion of RNA polymerase I and IItranscripts has only a slight effect on the location ofthe FG-repeat domain of Nup153. A total of 8% ofthe gold particles were detected on the cytoplasmic

Figure 7. Actinomycin D and its influence on the location of the FG-repeat domains of Nup153 and Nup214.Actinomycin D was microinjected into the nuclei of Xenopus oocytes three hours prior to nuclei isolation and pre-embedding labeling with the anti-C2-Nup153 and the anti-Nup214-D antibody, respectively. Quantification of the goldparticle distribution associated with the NPC after labeling with the (a) anti-C2-Nup153 antibody and (b) the anti-Nup214-D antibody, respectively. A total of 95 gold particles were scored for the anti-C2-Nup153 antibody and 140 forthe anti-Nup214-D antibody.


face of the NPC. On the nuclear side, 22% of thegold particles were found at or near the nuclear ringof the NPC and 70% at or near the distal ring. Incomparison to steady-state conditions, this rep-resents a movement of the FG-repeats towards thedistal ring on the nuclear side of the NPC (70%versus 34% at steady-state), whereas the cytoplasmicvalue is rather unaffected (8% versus 12% at steady-state21).

In contrast, the location of the FG-repeat domainof Nup214 is more strongly influenced by thetreatment of Xenopus nuclei with actinomycin D.Compared to steady-state, actinomycin D treatmentleads to increased nuclear localization of Nup214FG-repeat domain (Figure 7(b) and see Supplemen-tary Figure S9B). In the absence of transcription,76% of the gold particles were associated with thenuclear side of the NPC, whereas only 25% of thegold particles were found on the cytoplasmic face ofthe NPC.

Taken together, preventing new transcription inXenopus nuclei appears to cause an arrest of the FG-repeat domains of Nup153 and Nup214 near or atthe distal ring of the NPC.

Discussion

Along with the identification of presumably allvertebrate nucleoporins ,4 many of these nucleo-porins have been mapped within the NPC bydifferent immuno-EM techniques.21,23,30,41–44

Nucleoporins are large proteins with, at least inthe case of the FG-repeat nucleoporins, flexibility insecondary structure that implies a complex top-ology within the NPC. Individual antibodiesagainst a particular nucleoporin therefore do notyield a holistic view of the organization of this

nucleoporin within the 3-D architecture of the NPC.In contrast, use of domain-specific antibodies againstnucleoporins, such as Nup15321 and Tpr44,45 hashelped to resolve discrepancies regarding theirlocalization and organization within the vertebrateNPC. Moreover, in the case of the FG-repeatnucleoporin Nup153, this approach revealed thatwhile its N-terminal domain and central zinc-fingerdomain are stationary within the NPC, theC-terminal FG-repeat domain exhibits a widerdistribution. With the present study we have nowresolved the domain organization and topology ofanother large FG-repeat nucleoporin, Nup214,within the NPC. Moreover, by immuno-EM weshow for the first time that the location of the FG-repeat domains of Nup153 and Nup214 isinfluenced by the transport state of the cell.

Multiple-site topology of Nup214 within the NPC

To determine the spatial organization of a givennucleoporin within the 3-D NPC architecture,domain-specific antibodies are required.21,44 Bythis approach, we show here that the N-terminaldomain of Nup214 resides near or at the cyto-plasmic ring of the NPC close to the entry to thecentral pore (see Figures 2(a) and 3(a) and (e)),consistent with a previous finding30 in which apolyclonal antibody against an N-terminal frag-ment of Xenopus Nup214 was used. We furtherfound that the central domain of Nup214 is alsostationary and localizes near or at the cytoplasmicring of the NPC (see Figures 2(c) and 3(e)). This is inagreement with the location reported by Kraemer etal.28 using a polyclonal antibody specific for thecentral region of human Nup214.Here, we have for the first time mapped the FG-

repeat domain of Nup214 within the NPC.


Interestingly, this domain is not confined to aspecific subdomain of the NPC. Instead, by usingdomain-specific antibodies and by expressingC-terminal myc-tagged Nup214 in Xenopus oocytes,we could show that this domain is flexible in itsarrangement, appearing at both the cytoplasmicand nuclear face of the NPC (see Figures 2(e) and3(c) and (e)).

Taken together, our immuno-EM studies on thedomain topology of Nup214 suggest that it isanchored to the cytoplasmic face of the NPC by itsN-terminal and/or central domain, whereas the FG-repeat domain of Nup214 is flexible within the NPCand might exist in a rather unstructured, more orless extended conformation.

Natively unfolded, extended character of FG-repeat domains

The C-terminal domain of Nup214 spans about800 amino acid residues (residues 1225–2091 of thehuman protein). Biophysical measurements havesuggested that the FG-repeat domains of yeastnucleoporins are natively unfolded, i.e. they havelittle secondary structure.19,20 This leads to thepossibility that the FG-repeat domains of vertebratenucleoporins, too, have little secondary structure,and might have a more or less extended confor-mation. If extended, the FG-repeat domain ofNup214 could span a distance of up tow275 nm.46 This length would easily allow therepeat domain to span the distance from thecytoplasmic periphery of the NPC to the distalring of the nuclear basket. Indeed, secondarystructure prediction reveals the FG-repeat domainof Nup214 to be predominantly unstructured(S.M.P., O. Mayans and B.F., unpublished data).Moreover, w67% of the amino acids of the FG-repeat domain of human Nup214/CAN havedisorder promoting character (data not shown47,48).These disorder promoting residues are scatteredacross the FG-repeat domain, further supportingthe notion that the FG-repeat domain has anextended conformation. Based on immuno-EMdata an extended conformation has also beensuggested for the FG-repeat domain of Nup153,which spans about 700 amino acid residues.21

Again, about 62% of the amino acids of the FG-repeat domain of Xenopus Nup153 have disorderpromoting character (data not shown). Further-more, atomic force microscopy measurementsindeed revealed that the FG-repeat domain ofNup153 has an extended conformation (R. Lim,J. Koeser, and U. Aebi, personal communication).

Natively unfolded, extended proteins usuallyencompass multiple binding domains and arethus capable of simultaneous interactions withmultiple binding partners.47,48 Both Nup153 andNup214 are able to interact with a number ofdifferent nuclear transport receptors.25,31,33,35 Theflexible nature of their FG-repeat domains wouldallow Nup153 and Nup214 to act as scaffolds for theassembly of transport complexes. The export

complex that mediates the recycling of Kap60p,the yeast importin a, is formed directly at the yeastFG-nucleoporin Nup2p.49 Furthermore, it has beenshown that the U1 snRNP import complex isformed while associated with Nup214.34

Flexible FG-repeat domains: implications fornuclear transport

Our localization studies have revealed the FG-repeat domain of Nup214 to be flexible within theNPC. Similarly, we have documented the FG-repeatdomain of Nup153 to also be flexible within theNPC.21 Here, we have now shown that the locationof the FG-repeat domains of Nup153 and Nup214 iscorrelated to the transport state of the cell,suggesting that they may associate with cargo–receptor complexes and accompany them throughthe central pore.

Arresting nuclear transport constrains the spatialdistribution of flexible FG-repeat domains of bothnucleoporins to their respective anchoring siteswithin the NPC, i.e. the distal ring of the NPC in thecase of Nup153 and near the cytoplasmic ring in thecase of Nup214 (see Figure 4). A surge of nuclearimport promotes cytoplasmic exposure of the FG-repeat domain of Nup153 followed by a distri-bution predominantly on the nuclear face of theNPC (see Figure 5). The spatial distribution of theFG-repeat domain of Nup214 appears responsive toimport cargo aswell. In the presence of excess importcargo, Nup214’s FG-repeat domain transientlyaccumulates at the nuclear NPC periphery. Similarly,the FG-repeat domain of Nup153 predominates onthe nuclear side under these conditions. Theseresults indicate that the spatial distribution of theFG-repeat domains within the NPC correlates withcargo/receptor interactions.

Microinjecting excess poly(AC) RNA into thenucleus leads first to a higher frequency of the FG-repeat domains of Nup214 on the nuclear face of theNPC (see Figure 6). Next, the nuclear portion of theFG-repeat domains of Nup153 and Nup214decreases as the cytoplasmic portion increases. Atlate time points, when the burst of poly(AC) RNAexport is presumably complete, the FG-repeatdistribution within the NPC returns to steady-state conditions. In the case of RNA export thespatial location of the FG-repeat domains ofNup153 and Nup214 is biased toward the point ofcargo contact, and seems to shift concomitantlywith the cargo through the central pore of the NPC.

Surprisingly, recent studies in yeast have shownthat FG-repeat domains of asymmetric nucleo-porins are not essential for receptor-mediatednuclear transport and that the importin-b pathwayis ratherunaffectedbydepleting themaximalnumberof FG-repeats.13,50 Asymmetric nucleoporins, such asNup153, may play a role in receptor-independentnuclear transport, as shown for the nuclear importof the transcription factor PU.1.51

Taken together, our findings suggest that FG-repeat domains have a more active role in cargo


translocation through the NPC than simply actingas a scaffold for the formation of cargo–receptorcomplexes or providing docking sites for cargo–receptor complexes at the NPC. Indeed, our resultsindicate that FG repeat domains accompany andguide cargo through the NPC’s central pore to itsfinal destination. In this context, transport sub-strates might be passed from an asymmetric FG-repeat nucleoporin residing on the cytoplasmic faceof the NPC, such as Nup214, to a symmetric FG-repeat nucleoporin, for example p62, to anasymmetric FG-repeat nucleoporin residing on thenuclear side of the NPC, such as Nup153, or viceversa. It is also conceivable that transport substratesare directly passed from one extended asymmetricFG-repeat nucleoporin to the next, thereby optimiz-ing cargo translocation through the central pore ofthe NPC. It will be interesting to analyze if the FG-repeat domains of symmetric nucleoporins, whichare, as in the case of p62, relatively short ascompared to the FG-repeat domains of Nup214and Nup153, have an extended conformation aswell. Such a scenario, in fact, might explain why thedeletion of individual or a minimal number of FG-repeat domains have only a slight effect on nucleartransport kinetics.

Material and Methods

All experimental procedures were performed at roomtemperature (rt) unless otherwise stated.

Antibody production and purification

Full length human Nup214 in pBluescript was a giftfrom Dr Gerard Grosveld (St. Jude Children’s ResearchHospital). Four non-overlapping fragments were createdby PCR for expression and antibody production. 214Acorresponds to the N-terminal domain (amino acidresidues 1–469), 214B corresponds to the coiled-coilsection of the central domain (amino acid residues702–931), 214C corresponds to the non-coiled-coil sectionof the central domain (amino acid residues 1026–1364),and 214D corresponds to the FG repeat domain (aminoacid residues 1684–2091). Following PCR, each DNAfragment was inserted into the pTrcHis-TOPO vector(Invitrogen, Carlsbad, CA) for expression of N-terminalhexahistidine-tagged protein. Recombinant fusionproteins were expressed in BL21(DE3) cells and purifiedon His-Bind resin (Novagen/EMD Biosciences, Madison,WI). Expression of the 214C fragment was insufficient forantibody production. The 214B and 214D proteins weresolubilized and purified in the presence of 6 M urea.Purified protein was used to inoculate rabbits (SpringValley Laboratories; Sykesville, MD) and the rabbit serawere affinity purified using the corresponding recombi-nant protein bound to a CNBr-activated Sepharose 4 FastFlow support (Amersham Pharmacia Biosciences,Buckinghamshire, England).For immunoblots, proteins were separated by 8% (w/v)

acrylamide SDS-PAGE and transferred to PVDF mem-brane. The membrane was blocked with either 2% (w/v)BSA (anti-214A) or 5% (w/v) non-fat milk (anti-214B andanti-214D) in PBS with 0.2% Tween-20. Affinity purifiedantibodies were used at 1/1000 dilution in immunoblots.

Antibody signals were detected by chemiluminescencewith ECL substrate (Amersham Pharmacia Biosciences).

Immuno-EM of isolated nuclei from Xenopus oocytes

Mature (stage 6) oocytes were surgically removed fromfemale Xenopus laevis, and their nuclei were isolated asdescribed.22 Colloidal gold particles, w8 nm in diameter,were prepared by reduction of tetrachloroauric acid withsodium citrate in the presence of tannic acid andantibodies were conjugated to colloidal gold particles asdescribed.52 Isolated nuclei were labeled as described21

with the following modifications. Nuclei were incubatedin a solution of anti-Nup214 antibodies diluted 1:1000 inPBS for two hours and washed twice in PBS. Afterwashing, nuclei were incubated for two hours in an anti-rabbit IgG antibody conjugated to 10 nm colloidal gold(BBI International, Cardiff, UK). Labeled nuclei werefixed and processed for EM as described.21

Immuno-EM of human cultured cells

HL-60 cells were cultivated in DulbeccoModified EagleMedium (DMEM; Vitromex, Geilenkirchen, Germany)containing 10% (v/v) fetal calf serum plus penicillin andstreptomycin (Gibco BRL, Grand Island, NY). Aftertrypsinisation, cells were washed twice in PBS and fixedfor 15 minutes in freshly prepared PBS containing 4%paraformaldehyde. After washing twice in PBS, cellswere permeabilized with PBS containing 0.1% TritonX-100 for two minutes, and washed again twice in PBS.For immunolabeling, cells were resuspended in the

corresponding anti-Nup214 antibody diluted 1:100 in PBSand incubated for two hours. Next cells were washedtwice in PBS containing 0.1% BSA, and resuspended insecondary anti-rabbit-IgG antibody conjugated to 10 nmcolloidal gold (see above) diluted 1:2 in PBS containing0.1% BSA and incubated for two hours. After two washesin PBS containing 0.1% BSA, the cells were fixed in 2%glutaraldehyde for one hour, washed twice in PBS andpost-fixed in 1% OsO4 for one hour. Fixed samples weredehydrated, embedded in Epon 812 resin (Fluka, Buchs,Switzerland), and processed for EM as described.21

Microinjection and immuno-EM of tagged humanNup214 in Xenopus nuclei

Full-length human Nup214 was subcloned frompET21b (a kind gift from Dr Doris Kraemer, Universityof Wurzburg) into EcoRI and NotI cut pcDNA3.1/myc-His (Invitrogen Corporation, Carlsbad, CA) to produceC-terminally tagged hNup214-myc. To produceN-terminally tagged GFP-Nup214, full-length humanNup214 was subcloned from pET21b into EcoRI andSacII cut pEGFP-C1 (Clontech, Palo Alto, CA). Formicroinjection into nuclei, freshly isolated oocytes fromX. laevis were prepared and processed for microinjectionas described.21 The localization of the fusion proteinswithin the NPC were determined by using a polyclonalanti-GFP and a monoclonal anti-myc antibody, respect-ively, directly conjugated to 8 nm colloidal gold.

Immunolocalization of FG-repeat domains at 4 8C

Freshly isolated nuclei from Xenopus oocytes werecollected in low salt buffer (LSB) containing 1 mM KCl,0.5 mM MgCl2, 10 mM Hepes (pH 7.5), and equilibratedto 4 8C for one hour. Next the nuclei were fixed for


15 minutes in LSB containing 4% formaldehyde. Thenuclei were washed twice in LSB for five minutes each,and incubated in anti-C2-Nup153 antibody directlyconjugated to 8 nm gold21 for two hours to determinethe localization of the FG-repeat domains of Nup153. Forlocalization of the FG-repeat domain of Nup214, nucleiwere incubated in anti-Nup214-D antibody and second-ary anti-rabbit-IgG antibody conjugated to 10 nmcolloidal gold as described above. After labeling thenuclei were prepared for EM as described.21

Nuclear import of nucleoplasmin

Xenopus nucleoplasmin was expressed from a pQE70vector and purified (Qiagen GmbH, Hilden, Germany) asdescribed.53 Nuclei were isolated manually from Xenopusoocytes, collected in LSB and incubated in an importmixture (50% HeLa cytosol, 20 mM Hepes (pH 7.3),110 mM potassium acetate, 5 mM sodium acetate, 1 mMEGTA, 2 mM DTT, 1 mM ATP, 5 mM creatine phosphate,20 units/ml creatine phosphokinase, and 1 mg/ml of eachof aprotinin, leupeptin, and pepstatin)38 containing 100 ngrecombinantly expressed nucleoplasmin as indicated. Nextthe nuclei were fixed in 4% formaldehyde and labeledwithanti-C2-Nup153 and anti-Nup214-D antibody as describedabove.

Export of poly(AC) RNA and inhibition oftranscription

Total RNA was purified from HeLa cells using theRNAeasy kit (Qiagen GmbH, Hilden, Germany)following the instructions of the manufacturer. Fromthis total RNA, poly(AC) RNA was purified using theGenElute mRNA miniprep kit (Sigma, St. Louis, MO)following the instructions of the manufacturer. Oocyteswere microinjected into their nuclei with 10–20 nl ofpurified poly(AC) RNA (0.1 mg/ml) and incubated fordifferent time points as indicated. Nuclei were isolated,collected in LSB, fixed in 4% formaldehyde and labeledwith anti-C2-Nup153 and anti-Nup214-D antibody,respectively, as described above. Actinomycin D (Sigma,St. Louis, MO) was dissolved in ethanol to 5 mg/ml.10–20 nl of a 5 mg/ml diluted solution were microinjectedinto nuclei of Xenopus oocytes and incubated for threehours. Next the nuclei were isolated, fixed and labeledwith anti-C2-Nup153 and anti-Nup214-D antibody,respectively, as described.

Acknowledgements

The authors thank Doris Kraemer (University ofWurzburg, Germany) and Gerard Grosveld (St.Jude Children’s Research Hospital) for providing uswith clones of human Nup214, Dirk Gorlich(University of Heidelberg, Germany) for theplasmid to express nucleoplasmin, and UlrikeKutay (ETH Zurich, Switzerland) for the importin-b45–462 fragment. Ueli Aebi is greatly acknowledgedfor critically reading of the manuscript and LaurentKreplak for helpful discussions. This work wassupported by research grants from the SwissNational Science Foundation (to B.F.) and theNational Institutes of Health (grants GM59975 to

M.A.P. and GM61275 to K.S.U.), by the Kanton BaselStadt, and the M.E. Muller Foundation.

Supplementary Data

Supplementary data associated with thisarticle can be found, in the online version, atdoi:10.1016/j.jmb.2005.06.034

The Supplementary Material comprises nineFigures.

References

1. Fried, H. & Kutay, U. (2003). Nucleocytoplasmictransport: taking an inventory. Cell. Mol. Life Sci. 60,1659–1688.

2. Mosammaparast, N. & Pemberton, L. F. (2004).Karyopherins: from nuclear-transport mediators tonuclear-function regulators. Trends Cell Biol. 14,547–556.

3. Reichelt, R., Holzenburg, A., Buhle, E. L., Jr, Jarnik,M., Engel, A. & Aebi, U. (1990). Correlation betweenstructure and mass distribution of the nuclear porecomplex and of distinct pore complex components.J. Cell Biol. 110, 883–894.

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

5. Stoffler, D., Feja, B., Fahrenkrog, B., Walz, J., Typke, D.& Aebi, U. (2003). Cryo-electron tomography pro-vides novel insights into nuclear pore architecture:implications for nucleocytoplasmic transport. J. Mol.Biol. 328, 119–130.

6. Beck, M., Forster, F., Ecke, M., Plitzko, J. M., Melchior,F., Gerisch, G. et al. (2004). Nuclear pore complexstructure and dynamics revealed by cryoelectrontomography. Science, 306, 1387–1390.

7. Akey, C. W. & Radermacher, M. (1993). Architectureof the Xenopus nuclear pore complex revealed bythree-dimensional cryo-electron microscopy. J. CellBiol. 122, 1–19.

8. Hinshaw, J. E., Carragher, B. O. & Milligan, R. A.(1992). Architecture and design of the nuclear porecomplex. Cell, 69, 1133–1141.

9. Pante, N. & Kann, M. (2002). Nuclear pore complex isable to transport macromolecules with diameters ofabout 39 nm. Mol. Biol. Cell, 13, 425–434.

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

11. Fahrenkrog, B. & Aebi, U. (2003). The nuclear porecomplex: nucleocytoplasmic transport and beyond.Nature Rev. Mol. Cell. Biol. 4, 757–766.

12. Fahrenkrog, B., Koser, J. & Aebi, U. (2004). Thenuclear pore complex: a jack of all trades? TrendsBiochem. Sci. 29, 175–182.

13. Strawn, L. A., Shen, T., Shulga, N., Goldfarb, D. S. &Wente, S. R. (2004). Minimal nuclear pore complexesdefine FG repeat domains essential for transport.Nature Cell Biol. 6, 197–206.

14. Bayliss, R., Littlewood, T. & Stewart, M. (2000).Structural basis for the interaction between FxFGnucleoporin repeats and importin-beta in nucleartrafficking. Cell, 102, 99–108.

15. Bayliss, R., Kent, H. M., Corbett, A. H. & Stewart, M.

http://dx.doi.org/doi:10.1016/j.jmb.2005.06.034


(2000). Crystallization and initial X-ray diffractioncharacterization of complexes of FxFG nucleoporinrepeats with nuclear transport factors. J. Struct. Biol.131, 240–247.

16. Bayliss, R., Leung, S. W., Baker, R. P., Quimby, B. B.,Corbett, A. H. & Stewart, M. (2002). Structural basisfor the interaction between NTF2 and nucleoporinFxFG repeats. EMBO J. 21, 2843–2853.

17. Fribourg, S., Braun, I. C., Izaurralde, E. & Conti, E.(2001). Structural basis for the recognition of anucleoporin FG repeat by the NTF2-like domain ofthe TAP/p15 mRNA nuclear export factor. Mol. Cell.8, 645–656.

18. Grant, R. P., Hurt, E., Neuhaus, D. & Stewart, M.(2002). Structure of the C-terminal FG-nucleoporinbinding domain of TAP/NXF1. Nature Struct. Biol. 9,247–251.

19. Denning, D. P., Patel, S. S., Uversky, V., Fink, A. L. &Rexach, M. (2003). Disorder in the nuclear porecomplex: the FG repeat regions of nucleoporins arenatively unfolded. Proc. Natl Acad. Sci. USA, 100,2450–2455.

20. Denning, D. P., Uversky, V., Patel, S. S., Fink, A. L. &Rexach, M. (2002). The Saccharomyces cerevisiaenucleoporin Nup2p is a natively unfolded protein.J. Biol. Chem. 277, 33447–33455.

21. Fahrenkrog, B., Maco, B., Fager, A. M., Koser, J.,Sauder, U., Ullman, K. S. & Aebi, U. (2002). Domain-specific antibodies reveal multiple-site topology ofNup153 within the nuclear pore complex. J. Struct.Biol. 140, 254–267.

22. Pante, N., Bastos, R., McMorrow, I., Burke, B. & Aebi,U. (1994). Interactions and three-dimensional localiz-ation of a group of nuclear pore complex proteins. J.Cell Biol. 126, 603–617.

23. Walther, T. C., Fornerod, M., Pickersgill, H., Goldberg,M., Allen, T. D. &Mattaj, I. W. (2001). The nucleoporinNup153 is required for nuclear pore basket formation,nuclear pore complex anchoring and import of asubset of nuclear proteins. EMBO J. 20, 5703–5714.

24. Bastos, R., Lin, A., Enarson, M. & Burke, B. (1996).Targeting and function in mRNA export of nuclearpore complex protein Nup153. J. Cell Biol. 134,1141–1156.

25. Shah, S., Tugendreich, S. & Forbes, D. (1998). Majorbinding sites for the nuclear import receptor are theinternal nucleoporin Nup153 and the adjacent nuclearfilament protein Tpr. J. Cell Biol. 141, 31–49.

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

27. Ullman, K. S., Shah, S., Powers, M. A. & Forbes, D. J.(1999). The nucleoporin nup153 plays a critical role inmultiple types of nuclear export. Mol. Biol. Cell, 10,649–664.

28. Kraemer, D., Wozniak, R. W., Blobel, G. & Radu, A.(1994). The human CAN protein, a putative oncogeneproduct associated with myeloid leukemogenesis, is anuclear pore complex protein that faces the cyto-plasm. Proc. Natl Acad. Sci. USA, 91, 1519–1523.

29. van Deursen, J., Boer, J., Kasper, L. & Grosveld, G.(1996). G2 arrest and impaired nucleocytoplasmictransport in mouse embryos lacking the proto-oncogene CAN/Nup214. EMBO J. 15, 5574–5583.

30. Walther, T. C., Pickersgill, H. S., Cordes, V. C.,Goldberg, M. W., Allen, T. D., Mattaj, I. W. &

Fornerod, M. (2002). The cytoplasmic filaments ofthe nuclear pore complex are dispensable for selectivenuclear protein import. J. Cell Biol. 158, 63–77.

31. Fornerod, M., van Deursen, J., van Baal, S., Reynolds,A., Davis, D., Murti, K. G. et al. (1997). The humanhomologue of yeast CRM1 is in a dynamic sub-complex with CAN/Nup214 and a novel nuclear porecomponent Nup88. EMBO J. 16, 807–816.

32. Kuersten, S., Arts, G. J., Walther, T. C., Englmeier, L. &Mattaj, I. W. (2002). Steady-state nuclear localizationof exportin-t involves RanGTP binding and twodistinct nuclear pore complex interaction domains.Mol. Cell. Biol. 22, 5708–5720.

33. Katahira, J., Strasser, K., Podtelejnikov, A., Mann, M.,Jung, J. U. & Hurt, E. (1999). The Mex67p-mediatednuclear mRNA export pathway is conserved fromyeast to human. EMBO J. 18, 2593–2609.

34. Rollenhagen, C., Muhlhausser, P., Kutay, U. & Pante,N. (2003). Importin beta-depending nuclear importpathways: role of the adapter proteins in the dockingand releasing steps. Mol. Biol. Cell, 14, 2104–2115.

35. Boer, J. M., van Deursen, J. M., Croes, H. J., Fransen, J.A. & Grosveld, G. C. (1997). The nucleoporin CAN/Nup214 binds to both the cytoplasmic and thenucleoplasmic sides of the nuclear pore complex inoverexpressing cells. Expt. Cell Res. 232, 182–185.

36. Pante, N. & Aebi, U. (1996). Sequential binding ofimport ligands to distinct nucleopore regions duringtheir nuclear import. Science, 273, 1729–1732.

37. Kutay, U., Izaurralde, E., Bischoff, F. R., Mattaj, I. W. &Gorlich, D. (1997). Dominant-negative mutants ofimportin-beta block multiple pathways of import andexport through the nuclear pore complex. EMBO J. 16,1153–1163.

38. Adam, S. A., Sterne-Marr, R. & Gerace, L. (1992).Nuclear protein import using digitonin-perme-abilized cells. Methods Enzymol. 219, 97–110.

39. Dingwall, C., Sharnick, S. V. & Laskey, R. A. (1982).A polypeptide domain that specifies migration ofnucleoplasmin into the nucleus. Cell, 30, 449–458.

40. Pante, N., Jarmolowski, A., Izaurralde, E., Sauder, U.,Baschong, W. & Mattaj, I. W. (1997). Visualizingnuclear export of different classes of RNA by electronmicroscopy. RNA, 3, 498–513.

41. Enninga, J., Levay, A. & Fontoura, B. M. (2003). Sec13shuttles between the nucleus and the cytoplasm andstably interacts with Nup96 at the nuclear porecomplex. Mol. Cell. Biol. 23, 7271–7284.

42. Belgareh, N., Rabut, G., Bai, S. W., van Overbeek, M.,Beaudouin, J., Daigle, N. et al. (2001). An evolution-arily conserved NPC subcomplex, which redistri-butes in part to kinetochores in mammalian cells.J. Cell Biol. 154, 1147–1160.

43. Griffis, E. R., Xu, S. & Powers, M. A. (2003). Nup98localizes to both nuclear and cytoplasmic sides of thenuclear pore and binds to two distinct nucleoporinsubcomplexes. Mol. Biol. Cell, 14, 600–610.

44. Frosst, P., Guan, T., Subauste, C., Hahn, K. & Gerace,L. (2002). Tpr is localized within the nuclear basket ofthe pore complex and has a role in nuclear proteinexport. J. Cell Biol. 156, 617–630.

45. Krull, S., Thyberg, J., Bjorkroth, B., Rackwitz, H. R. &Cordes, V. C. (2004). Nucleoporins as components ofthe nuclear pore complex core structure and Tpr asthe architectural element of the nuclear basket. Mol.Biol. Cell, 15, 4261–4277.

46. Strelkov, S. V., Herrmann, H., Geisler, N., Wedig, T.,Zimbelmann, R., Aebi, U. & Burkhard, P. (2002).


Conserved segments 1A and 2B of the intermediatefilament dimer: their atomic structures and role infilament assembly. EMBO J. 21, 1255–1266.

47. Dunker, A. K., Lawson, J. D., Brown, C. J., Williams,R. M., Romero, P., Oh, J. S. et al. (2001). Intrinsicallydisordered protein. J. Mol. Graph. Model. 19, 26–59.

48. Uversky, V. N. (2002). What does it mean to benatively unfolded? Eur. J. Biochem. 269, 2–12.

49. Matsuura, Y., Lange, A., Harreman, M. T., Corbett, A.H. & Stewart, M. (2003). Structural basis for Nup2pfunction in cargo release and karyopherin recycling innuclear import. EMBO J. 22, 5358–5369.

50. Zeitler, B. & Weis, K. (2004). The FG-repeat asym-metry of the nuclear pore complex is dispensable forbulk nucleocytoplasmic transport in vivo. J. Cell Biol.167, 583–590.

51. Zhong, H., Takeda, A., Nazari, R., Shio, H., Blobel, G.& Yaseen, N. R. (2005). Carrier-independent nuclearimport of the transcription factor PU.1 via RANGTP-stimulated binding to NUP153. J. Biol. Chem. 280,10675–10682.

52. Baschong, W. & Wrigley, N. G. (1990). Small colloidalgold conjugated to Fab fragments or to immuno-globulin G as high-resolution labels for electronmicroscopy: a technical overview. J. Electron Microsc.Technol. 14, 313–323.

53. Gorlich, D., Prehn, S., Laskey, R. A. & Hartmann, E.(1994). Isolation of a protein that is essential for thefirst step of nuclear protein import. Cell, 79, 767–778.

Edited by W. Baumeister

(Received 11 April 2005; received in revised form 8 June 2005; accepted 13 June 2005)Available online 5 July 2005

nucleoporin domain topology is linked to the transport status of the nuclear pore complex

Documents