During interphase growth of eukaryotic cells, nuclear pore complexes (NPCs) are continuously incorporated into the intact nuclear envelope (NE) by mechanisms that are largely unknown. De novo formation of NPCs involves local fusion events between the inner and outer nuclear membrane, formation of a transcisternal membranous channel of defined diameter and the coordinated assembly of hundreds of nucleoporins into the characteristic NPC structure. Here we have used a cell-free system based on Xenopus egg extract, which allows the experimental separation of nuclear-membrane assembly and NPC formation. Nuclei surrounded by a closed double nuclear membrane, but devoid of NPCs, were first reconstituted from chromatin and a specific membrane fraction. Insertion of NPCs into the preformed pore-free nuclei required cytosol containing soluble nucleoporins or nucleoporin subcomplexes and, quite unexpectedly, major vault protein (MVP). MVP is the main component of vaults, which are ubiquitous barrel-shaped particles of enigmatic function. Our results implicate MVP, and thus also vaults, in NPC biogenesis and provide a functional explanation for the association of a fraction of vaults with the NE and specifically with NPCs in intact cells.

In order to maintain their size, proliferating cells must double their mass between two successive cell divisions. Interphase cell growth is accompanied by nuclear growth and expansion of the nuclear envelope (NE), concomitant with an increase in the number of NPCs. Thus, the total number of nuclear pore complexes (NPCs) doubles during the interphase of HeLa cells (Maul, 1977) and increases continuously throughout the cell cycle of yeast cells (Winey et al., 1997). An especially striking example of NE enlargement is provided by Xenopus oocytes during their diplotene growth phase, when the number of NPCs increases at a rate of ∼500 NPCs per minute (Scheer, 1973). Biogenesis of NPCs is considered to be a stepwise self-assembly process, beginning with a local fusion event between the inner and outer nuclear membrane, stabilization of the emerging membrane channel, and association of soluble nucleoporins or preformed nucleoporin complexes with the pore membrane in a specific order (Macaulay and Forbes, 1996; Goldberg et al., 1997; Burke and Ellenberg, 2002; Suntharalingam and Wente, 2003; Hetzer et al., 2005; Alber et al., 2007). However, the mechanisms involved in de novo formation of NPCs are as yet unknown.

Cell-free systems based on Xenopus egg extract are extremely valuable for studying and manipulating NE assembly (Gant and Wilson, 1997; Hetzer et al., 2005; Anderson and Hetzer, 2008). When sperm chromatin is incubated in extract derived from activated Xenopus eggs, a NE forms spontaneously at the surface of the added chromatin, resulting in the reconstitution of nuclei enclosed by a NE with functional NPCs (reviewed by Gant and Wilson, 1997; Hetzer et al., 2005). The in vitro formation of a NE around the added chromatin resembles the postmitotic reappearance of the NE in higher eukaryotic cells and involves closely linked and coordinated processes, such as recruitment of specific NPC subcomplexes and membrane vesicles to the chromatin surface, membrane fusion, membrane flattening, and assembly of functional NPCs (Gant and Wilson, 1997; Hetzer et al., 2005). However, the egg extract also supports subsequent growth of the reconstituted nuclei concomitant with the insertion of new NPCs into the expanding NE, similar to interphase nuclear growth (D'Angelo et al., 2006). Because different mechanisms may govern NPC assembly at telophase and during interphase growth (Hetzer et al., 2005; D'Angelo and Hetzer, 2008), we have developed a NPC insertion assay that allows the study of de novo formation of NPCs in a pore-free NE. Here we describe a novel NPC assembly factor, which turned out to be major vault protein (MVP).

The experimental system

Previously, we have noted that incubation of egg extract without addition of chromatin promoted the formation of flat membrane cisternae studded with pore complexes, termed annulate lamellae (AL) (Dabauvalle et al., 1991). In a subsequent study, we found that AL formation apparently depleted a specific type of membrane vesicles from the egg extract that were required for NPC assembly from the egg extract (Ewald et al., 1997). Once these pore-forming membranes were incorporated into AL, they were no longer available for subsequent nuclear reconstitution experiments. Hence, when we added chromatin to the pre-incubated extract, pore-free nuclei were reconstituted (Ewald et al., 1997). In an attempt to biochemically separate NPC-inducing and nuclear-membrane-forming vesicles, we fractionated membranes from `mitotic' Xenopus egg extract (see Newmeyer and Wilson, 1991) by centrifugation on a discontinuous sucrose gradient into light and heavy fractions (recovered from the 30% and 40% sucrose step, respectively) (Fig. 1A). Separation of membranes prepared from activated eggs [the `interphase extract' described by Newmeyer and Wilson (Newmeyer and Wilson, 1991)] was less efficient and yielded variable results. The two membrane fractions were then assayed for function either individually or in combination by addition to a nuclear reconstitution assay consisting of sperm chromatin and cytosol (S200). After a 1-hour incubation, the presence of NPCs was monitored by immunofluorescence microscopy using antibodies against the nucleoporin Nup62. Nup62 is a member of the FG-repeat containing nucleoporins; it localizes to the central NPC channel and marks the formation of completed NPCs (Dultz et al., 2008). In addition, the functional status of the newly formed NPCs was examined by scoring the nuclear uptake of the nucleolar protein fibrillarin. Only in the presence of functionally competent NPCs is fibrillarin imported into the nuclei where it forms numerous distinct dense bodies (Ewald et al., 1997). The distribution of the light and heavy membrane fractions was visualized by fluorescence microscopy after staining with green (DiOC18) or red (DilC18) fluorescent lipophilic dyes, respectively (Hetzer et al., 2000). Samples were also processed for transmission EM.

Fig. 1.

Experimental uncoupling of nuclear membrane formation and NPC assembly by using two different membrane fractions in the nuclear reconstitution assay. (A) Isolation of light and heavy membranes from mitotic egg extract. (B) Incubation of sperm chromatin in cytosol supplemented with ATP-regenerating system together with either the light (left panel) or heavy (middle panel) membrane fraction. The light membranes do not bind to chromatin in contrast to the heavy membranes, which form a closed double nuclear membrane around the chromatin without recognizable NPC, as shown by thin section EM (upper row). Consistent with the EM observations, fluorescence microscopy (bottom row) reveals the absence of DiOC18 pre-labeled light membranes (green) from chromatin, but shows association of the DilC18 pre-labeled heavy membranes (red) with the chromatin surface in a rim-like pattern, indicative of membrane fusion. DNA is counterstained with Hoechst 33258 (blue). When pore-free nuclei were first allowed to assemble by incubating chromatin with cytosol and heavy membranes for 1 hour, followed by addition of light membranes, NPC were recognized by EM (right panel, NPCs are indicated by arrows in the insert). Under these conditions, the differentially labeled membrane fractions both bind to the chromatin surface (bottom row). (C) Incubation of chromatin as outlined above (indicated on the left-hand side of the figure), followed by immunofluorescent detection of NPCs with antibodies against Nup62 and nuclear import with antibodies against fibrillarin. In the presence of functional NPCs, the nucleolar protein fibrillarin is imported into the reconstituted nuclei where it forms distinct nuclear bodies (Ewald et al., 1997). Corresponding phase contrast images and Hoechst fluorescence are also shown. Functional NPC are recognized only after sequential addition of the two membrane fractions (bottom row). Scale bars of the EM micrographs (upper part of B), 1 μm and 0.1 μm (inserts), all other bars 10 μm.

Fig. 1.

Experimental uncoupling of nuclear membrane formation and NPC assembly by using two different membrane fractions in the nuclear reconstitution assay. (A) Isolation of light and heavy membranes from mitotic egg extract. (B) Incubation of sperm chromatin in cytosol supplemented with ATP-regenerating system together with either the light (left panel) or heavy (middle panel) membrane fraction. The light membranes do not bind to chromatin in contrast to the heavy membranes, which form a closed double nuclear membrane around the chromatin without recognizable NPC, as shown by thin section EM (upper row). Consistent with the EM observations, fluorescence microscopy (bottom row) reveals the absence of DiOC18 pre-labeled light membranes (green) from chromatin, but shows association of the DilC18 pre-labeled heavy membranes (red) with the chromatin surface in a rim-like pattern, indicative of membrane fusion. DNA is counterstained with Hoechst 33258 (blue). When pore-free nuclei were first allowed to assemble by incubating chromatin with cytosol and heavy membranes for 1 hour, followed by addition of light membranes, NPC were recognized by EM (right panel, NPCs are indicated by arrows in the insert). Under these conditions, the differentially labeled membrane fractions both bind to the chromatin surface (bottom row). (C) Incubation of chromatin as outlined above (indicated on the left-hand side of the figure), followed by immunofluorescent detection of NPCs with antibodies against Nup62 and nuclear import with antibodies against fibrillarin. In the presence of functional NPCs, the nucleolar protein fibrillarin is imported into the reconstituted nuclei where it forms distinct nuclear bodies (Ewald et al., 1997). Corresponding phase contrast images and Hoechst fluorescence are also shown. Functional NPC are recognized only after sequential addition of the two membrane fractions (bottom row). Scale bars of the EM micrographs (upper part of B), 1 μm and 0.1 μm (inserts), all other bars 10 μm.

Experimental separation of nuclear membrane formation and NPC assembly

When both membrane fractions were pooled, a functional NE with NPCs was formed around chromatin (supplementary material Fig. S1), consistent with the well-known fact that NE formation in vitro requires both soluble and membrane components of the egg (Vigers and Lohka, 1991; Goldberg et al., 1997; Gant and Wilson, 1997). The light membranes alone (labeled with DiOC18, green emitting) did not bind to chromatin and a nuclear membrane did not form (Fig. 1B, left panel). This correlated with the absence of anti-Nup62 staining and the failure to import fibrillarin (Fig. 1C, upper row). By contrast, the heavy membranes labeled with DilC18 (red emitting) bound to chromatin, as judged by the fluorescent rim staining, and fused into a closed double membrane. However, the membrane lacked morphologically identifiable pores in EM ultrathin sections (Fig. 1B, middle panel). The absence of NPCs was confirmed by the absence of labeling with anti-Nup62 antibodies and of nuclear transport activity (Fig. 1C, middle row). However, when we added light membranes to the previously poreless nuclei, NPCs emerged in the pre-existing double membrane (Fig. 1B, right panel) and the resulting nuclei were stained with antibodies against Nup62 and imported fibrillarin (Fig. 1C, bottom row). Recruitment of the light membranes to the pore-free nuclei could be directly visualized by fluorescence microscopy and resulted in the superposition of red and green membrane fluorescence (Fig. 1B, right panel).

In conclusion, stepwise addition of two distinct membrane fractions allowed the experimental separation of nuclear membrane formation and NPC assembly. Pore-free nuclei enclosed by a double membrane were reconstituted by incubation of chromatin with cytosol and the heavy membrane fraction or, alternatively, by incubation of chromatin in preassembled extract, i.e. after sequestration of the pore-inducing membranes into AL (Ewald et al., 1997). Subsequent addition of the light membrane fraction to the pore-free nuclei triggered de novo insertion of functional NPCs into the pre-existing double nuclear membrane.

Functionally distinct membrane fractions with chromatin targeting, membrane fusion and NPC-inducing activities have also been described by others (Vigers and Lohka, 1991; Sasagawa et al., 1999; Oke and Inoue, 2003). However, in contrast to our present results, formation of intact NEs with NPCs required the simultaneous presence of these membrane fractions in the nuclear reconstitution assays. A direct comparison of the membrane fractions used by us and the fractions used by the authors mentioned above is difficult because we prepared `mitotic' membranes from eggs arrested at metaphase II, whereas in the other studies the membranes were isolated from the `interphase extract' (Newmeyer and Wilson, 1991). On the basis of functional characteristics, our heavy membrane fraction resembles the nuclear envelope precursor vesicle fraction 1 (PV1) described by Sasagawa et al. (Sasagawa et al., 1999) in that both membranes bind to chromatin and form a continuous NPC-free double nuclear membrane.

Previous studies on separating membrane-related events from NPC assembly relied on the use of chemical inhibitors such as the calcium chelator BAPTA or on the depletion or enrichment of proteins that are positively (RanGTP) or negatively (importin-β) involved in NPC assembly (Macaulay and Forbes, 1996; Harel et al., 2003; D'Angelo et al., 2006). Our in vitro system described here is different as it solely depends on the physical separation of two membrane populations. In any event, our results support the view that NPCs form de novo and not through the cleavage of existing NPCs (D'Angelo et al., 2006).

MVP is associated with the pore-inducing membrane fraction

Preliminary immunoblots revealed the presence of the transmembrane nucleoporins gp210 and POM121, the integral inner nuclear membrane protein LBR as well as several other proteins in both membrane fractions (supplementary material Fig. S2). The only notable difference was the apparent absence of the small GTPase Ran from the light membranes [for the role of Ran in NE assembly see Hetzer et al. (Hetzer et al., 2005)]. To identify potential NPC assembly factor(s), both membrane fractions were subjected to two-dimensional gel electrophoresis (16-BAC–SDS-PAGE) (Hartinger et al., 1996). A spot of ∼104 kDa was exclusively present in the light membranes and was identified by mass spectrometry as MVP. MVP is the major structural component of vaults, a large and abundant RNA-protein complex found in most eukaryotic cells. Vaults form barrel-shaped hollow structures (13 MDa, ∼55×30 nm) consisting of ∼96 molecules of MVP, two minor protein components (vault poly[ADP-ribose] polymerase and telomerase-associated protein-1) and about six copies of a small untranslated RNA (Suprenant, 2002; Mikyas et al., 2004; Anderson et al., 2007). MVP alone is sufficient to self-assemble into the characteristic vault structure (Stephen et al., 2001; Mikyas et al., 2004; Zheng et al., 2005). Remarkably, despite their ubiquitous expression, abundance and evolutionarily conservation, the function of vaults is still largely unknown (Suprenant, 2002; Steiner et al., 2006). Some authors found vaults to be enriched at the NPCs and implicated them in nucleocytoplasmic transport processes (Chugani et al., 1993; van Zon et al., 2006). Xenopus MVP (xMVP, GenBank accession number Q6PF69) is comprised of 849 amino acids, has a calculated Mr of 95.670, a pI of 5.16, and is 71% and 58% identical to human and Dicytostelium MVP, respectively.

The physical attachment of MVP and/or vaults to the light membranes was verified by flotation centrifugation experiments (supplementary material Fig. S3). In addition, we expressed full-length His-tagged xMVP in Escherichia coli and raised polyclonal antibodies against the recombinant protein. In western blots, the antibodies recognized recombinant xMVP (Fig. 2A, lane 2) (note that the apparent Mr of xMVP is ∼104.000 and is slightly greater than its calculated Mr). In addition, the antibodies identified endogenous MVP in a vault-enriched fraction prepared from Xenopus eggs (Fig. 2B, lane 2). Essentially identical results were obtained with commercial antibodies directed against human MVP (Fig. 2A,B, lanes 3). Next, various egg fractions were analyzed by immunoblotting experiments. As shown in Fig. 2C, MVP was recovered in the S100 egg extract (for definition see Fig. 2A and supplementary material Fig. S4), in the membrane-enriched fraction (P200) and in the light membrane fraction, but in neither the heavy membrane fraction nor the S200 cytosol. Interestingly, the majority of endogenous vaults present in the crude extract were associated with membranes and did not sediment at 100,000 g as expected for free vault particles. Only after membrane solubilization with Triton X-100, were the bulk of vault particles recovered in the pellet fraction (P100) (supplementary material Fig. S4). xMVP behaved like a peripheral protein as it was released from the light membranes with 2% Empigen (a zwitterionic detergent) and 750 mM NaCl. Consistent with this, a recent atomic-level model suggested the presence of membrane anchor residues at the surface of vaults (Anderson et al., 2007). When the extracted membranes were added to poreless nuclei, NPCs were no longer induced (not shown).

Fig. 2.

Characterization of antibodies against xMVP and association of MVP with the light membranes as revealed by immunoblotting experiments. (A) 1 μg of purified full-length His-tagged xMVP was either stained with Coomassie-blue (lane 1), probed with guinea pig antibodies raised against recombinant xMVP (lane 2) or with the monoclonal antibody to human MVP (lane 3). Note, that the apparent Mr of xMVP (∼104.000) is slightly higher than its calculated Mr (95.670). (B) Proteins of a vault-enriched fraction from Xenopus eggs were separated by SDS-PAGE and visualized by silver staining (lane 1). A sample run in parallel was blotted and incubated with anti-xMVP (lane 2) or anti-human MVP antibodies (lane 3). MVP is present as a distinct band in the vault fraction. (C) Proteins of the indicated egg fractions were separated by SDS-PAGE and stained with Coomassie blue (left). The MVP band is indicated by the arrowhead in lane 5. A gel run in parallel was probed with antibodies against xMVP (right). MVP is present in egg extract (lane 1), the washed membrane fraction (P200, lane 2) and the light membranes (lane 5), but is absent from the cytosol (S200, lane 3) and the heavy membranes (lane 4). Molecular mass standards are indicated in kDa.

Fig. 2.

Characterization of antibodies against xMVP and association of MVP with the light membranes as revealed by immunoblotting experiments. (A) 1 μg of purified full-length His-tagged xMVP was either stained with Coomassie-blue (lane 1), probed with guinea pig antibodies raised against recombinant xMVP (lane 2) or with the monoclonal antibody to human MVP (lane 3). Note, that the apparent Mr of xMVP (∼104.000) is slightly higher than its calculated Mr (95.670). (B) Proteins of a vault-enriched fraction from Xenopus eggs were separated by SDS-PAGE and visualized by silver staining (lane 1). A sample run in parallel was blotted and incubated with anti-xMVP (lane 2) or anti-human MVP antibodies (lane 3). MVP is present as a distinct band in the vault fraction. (C) Proteins of the indicated egg fractions were separated by SDS-PAGE and stained with Coomassie blue (left). The MVP band is indicated by the arrowhead in lane 5. A gel run in parallel was probed with antibodies against xMVP (right). MVP is present in egg extract (lane 1), the washed membrane fraction (P200, lane 2) and the light membranes (lane 5), but is absent from the cytosol (S200, lane 3) and the heavy membranes (lane 4). Molecular mass standards are indicated in kDa.

Recombinant MVP and purified vault complexes both promote NPC assembly

To assess the pore-inducing competence of MVP, full-length recombinant xMVP was purified, added to pre-assembled pore-free nuclei and the appearance of NPCs monitored by immunofluorescence microscopy with anti-Nup62 antibodies. It should be emphasized that for these experiments the light membranes were no longer used. MVP rapidly bound to the surface of the pore-free nuclei, and abundant NPCs stained with anti-Nup62 antibodies were now present (Fig. 3, middle row). The newly formed NPCs were functional and imported fibrillarin (Fig. 3, middle row). EM analyses confirmed the presence of nuclei enclosed by a NE as shown in the survey micrograph (Fig. 4A). At higher magnification, NPCs could be clearly identified both in transverse (Fig. 4B, arrows) and tangential sections (Fig. 4C). Essentially the same results were obtained with a vault-enriched fraction of Xenopus eggs instead of recombinant xMVP (Fig. 3, bottom row). Pore-free nuclei incubated with buffer alone as a control were negative with anti-Nup62 antibodies and did not import fibrillarin (Fig. 3, top row). Notably, the control nuclei were not stained with antibodies to MVP, consistent with the apparent absence of MVP from the heavy membranes as shown by western blot experiments (Fig. 2C).

We also assayed induction of NPC assembly by adding recombinant MVP to pore-free nuclei that were reconstituted in pre-incubated extract after formation of AL (see Ewald et al., 1997). The results were essentially identical to those described above with the notable exception that numerous cytoplasmic aggregates scored positive with both anti-Nup62 and anti-MVP antibodies (supplementary material Fig. S5). As previously shown, such Nup62-positive structures, which are readily visualized in the light microscope, represent stacks of AL (Dabauvalle et al., 1991). This observation is consistent with the notion that MVP-associated membranes are recruited into nascent AL and hence are no longer available for NE formation when chromatin is added to the pre-incubated extract. In fact, the pore-free nuclei assembled in pre-incubated egg extract were not stained with antibodies to MVP, in contrast to control nuclei assembled in fresh egg extract (supplementary material Fig. S5).

Taken together, our data demonstrate that in the presence of soluble nucleoporins or nucleoporin subcomplexes (cytosol fraction), both recombinant MVP and a fraction enriched in endogenous vaults trigger the de novo incorporation of NPCs into pre-existing NE double membranes. Consistent with earlier reports that expression of MVP alone is sufficient for vault formation (Stephen et al., 2001; Mikyas et al., 2004; Zheng et al., 2005), we found that purified His-tagged xMVP added to cytosol spontaneously assembled into larger structures with sedimentation characteristics similar to those of intact vaults (supplementary material Fig. S6). Hence, it is reasonable to assume that assembled vaults act as NPC-assembly factors. In fact, the topological arrangement of highly conserved tryptophan residues around the barrel region of vaults has been implicated in membrane binding and bending (Anderson et al., 2007). Interactions of the clustered tryptophan residues exposed at the vault surface with the cytoplasmic face of the outer nuclear membrane might locally deform the lipid bilayer, which in turn could facilitate the occurrence of fusion events between inner and outer nuclear membranes.

Fig. 3.

Recombinant MVP or purified vaults promote formation of NPCs. Pore-free nuclei were allowed to assemble (top diagram) and subsequently incubated as specified on the left-hand side. Nuclei were analyzed by immunofluorescence using antibodies to xMVP and to Nup62 to visualize NPCs, and fibrillarin to monitor nuclear import. Also shown are the corresponding phase contrast images and Hoechst staining of DNA. Both recombinant xMVP and purified vaults trigger assembly of functional NPCs into the pre-existing double membrane. Note the absence of MVP in control nuclei (+PBS) and the association of MVP with the restored nuclei (+His-MVP and +Vaults) in a distribution pattern resembling that of the newly formed NPCs as revealed by anti-Nup62 staining. Scale bars: 10 μm.

Fig. 3.

Recombinant MVP or purified vaults promote formation of NPCs. Pore-free nuclei were allowed to assemble (top diagram) and subsequently incubated as specified on the left-hand side. Nuclei were analyzed by immunofluorescence using antibodies to xMVP and to Nup62 to visualize NPCs, and fibrillarin to monitor nuclear import. Also shown are the corresponding phase contrast images and Hoechst staining of DNA. Both recombinant xMVP and purified vaults trigger assembly of functional NPCs into the pre-existing double membrane. Note the absence of MVP in control nuclei (+PBS) and the association of MVP with the restored nuclei (+His-MVP and +Vaults) in a distribution pattern resembling that of the newly formed NPCs as revealed by anti-Nup62 staining. Scale bars: 10 μm.

Although further studies are required to define the role of vaults in NPC assembly in more detail, we speculate that vaults might serve as structural elements that help to stabilize the newly formed membrane channels and prevent their uncontrolled expansion during the early phases of NPC assembly, when a rigid NPC scaffold has not yet been formed. Possible modes of NPC assembly have been described by several authors (Goldberg et al., 1997; Gant et al., 1998; Alber et al., 2007). Furthermore, by `sealing' the membrane channel during NPC formation, vaults might prevent the uncontrolled leakage of macromolecules across the NE. Given the functional importance of NPCs for nucleocytoplasmic communication it is not surprising that redundant mechanisms are involved in their assembly (Stavru et al., 2006). This might explain the finding that living cells tolerate downregulation of MVP by siRNA (Huffman and Corey, 2005) or can even bypass the complete absence of MVP in a knockout mouse model (Mossink et al., 2002). By contrast, the cell-free system used here is derived from purified subcellular fractions that might have lost compensatory mechanisms for NPC assembly. Participation of vaults in the fundamental process of NPC biogenesis, which in turn is tightly linked to cellular growth and proliferation processes, provides an explanation for the striking evolutionarily conservation of the MVP sequence and vault structure as well as the ubiquitous expression of vaults in normal cells and their upregulation in several tumor cells (Suprenant, 2002; Steiner et al., 2006).

Fig. 4.

Poreless nuclei were incubated with recombinant xMVP (corresponding to the experiment shown in Fig. 3, middle row) and analyzed by transmission EM. (A) Survey micrograph shows part of a nucleus. (B,C). Chromatin (CH) is enclosed by an NE, which is perforated by NPCs, as seen at higher magnification. Two NPCs are denoted by arrows in the transversely sectioned NE (B). The tangential section clearly reveals the transcisternal channel of the newly assembled NPCs (C). Scale bars: 0.5 μm (A) and 0.1 μm (B,C).

Fig. 4.

Poreless nuclei were incubated with recombinant xMVP (corresponding to the experiment shown in Fig. 3, middle row) and analyzed by transmission EM. (A) Survey micrograph shows part of a nucleus. (B,C). Chromatin (CH) is enclosed by an NE, which is perforated by NPCs, as seen at higher magnification. Two NPCs are denoted by arrows in the transversely sectioned NE (B). The tangential section clearly reveals the transcisternal channel of the newly assembled NPCs (C). Scale bars: 0.5 μm (A) and 0.1 μm (B,C).

Egg extract

Xenopus laevis were purchased from the Xenopus Express Farm (Le Bourg, France). Extract from activated eggs was prepared as described (Newport, 1987). After low speed centrifugation (10,000 g, 20 minutes at 4°C), the cleared supernatant was spun at 100,000 g for 1 hour at 4°C (see also supplementary material Fig. S4). The resulting supernatant was then centrifuged at 200,000 g for 2 hours at 4°C to obtain a membrane pellet (P200) and the cytosol (S200).

Membranes were prepared from nonactivated (mitotic) eggs as described by Newmeyer and Wilson (Newmeyer and Wilson, 1991) with the following modifications: The crude low speed extract was first centrifuged at 100,000 g for 1 hour at 4°C. The supernatant (S100) was diluted 1:1 with M-lysis buffer (80 mM β-glycerophosphate, 20 mM EGTA, 15 mM MgCl2, 1 mM DTT) to inhibit possible membrane fusion events, followed by centrifugation at 200,000 g for 1 hour at 4°C. The white portion of the membrane pellet was resuspended in 20 volumes of ice-cold buffer [hereafter termed `extract buffer', consisting of 250 mM sucrose, 50 mM KCl, 2.5 mM MgCl2, 1 mM DTT, 50 μg/ml cycloheximide and 5 μg/ml cytochalasin B (see Newport, 1987)] and pelleted onto a 60% sucrose cushion by centrifugation at 150,000 g for 20 minutes at 4°C. The resulting membrane layer was diluted with 200 μl of extract buffer and layered on top of a three-step sucrose gradient (200 μl of 30%, 150 μl of 40% and 100 μl of 50% sucrose made up in the same buffer). After centrifugation at 30,000 r.p.m. for 1 hour at 4°C in a Beckman SW50 Ti rotor, two distinct membrane fractions were recovered from the 30% (`light membranes') and 40% (`heavy membranes') sucrose steps by side puncture and used immediately or stored at –70°C until use (Fig. 1A).

Flotation centrifugation of membranes

An aliquot (80 μl) of the light membranes in 30% sucrose was mixed with the same volume of a 90% sucrose solution made up in the extract buffer and overlaid with 200 μl each of 40% and 30% sucrose solutions. After centrifugation at 35,000 r.p.m. for 4 hours at 4°C in a Beckman SW50 Ti rotor, the three fractions containing 60%, 40% and 30% sucrose were collected from the top. Proteins were precipitated with 20% trichloroacetic acid and assayed for MVP by immunoblotting.

Membrane extraction

Light membranes were incubated for 30 minutes at 4°C in an extract buffer containing 750 mM NaCl and 2% Empigen to release MVP and vaults. The MVP-depleted membranes were recovered by centrifugation at 200,000 g for 60 minutes at 4°C. In another set of experiments, membrane-free vaults were prepared by solubilizing the membranes with Triton X-100. To this end, 500 μl of the crude extract was incubated for 10 minutes at 22°C in the presence of either PBS or 2% Triton X-100 followed by centrifugation at 100,000 g for 1 hour at 4°C, as detailed in supplementary material Fig. S4. For immunoblotting analyses, 10 μl of the supernatant S100 (corresponding to one fiftieth of the supernatant volume) was diluted in 10 μl of double-concentrated SDS sample buffer. The pellet P100 was directly resupended in 150 μl of sample buffer, and 3 μl of this solution (corresponding to one fiftieth of the pellet fraction) was mixed with 17 μl of sample puffer.

Fluorescence labeling of membranes

Membranes were labeled with the fluorescent lipophilic dyes DiO-C18 or Dil-C18 (Invitrogen) essentially as described (Hetzer et al., 2000).

Nuclear reconstitution

De-membranated Xenopus sperm nuclei (Vigers and Lohka, 1991) were added to 50 μl cytosol (S200, see above) to give a final concentration of 106 nuclei/ml. All samples were supplemented with an ATP-regenerating system (Newport, 1987). A 20-μl aliquot of either the light or heavy membrane fraction or a 1:1 mixture of both was then added and the reaction incubated for 2 hours at 22°C. In order to separate membrane formation from NPC assembly, heavy membranes were added as described above. After 1 hour of incubation, pore-free nuclei surrounded by a closed double membrane were assembled. Subsequently, 20 μl of the light membranes, 2.5 μg recombinant His-MVP or 5 μl of a vault-enriched fraction were added and the sample incubated for another 1 hour.

Isolation of vault particles

Vaults were isolated from extract derived from 6 ml of packed activated Xenopus eggs following the protocol of Kedersha and Rome (Kedersha and Rome, 1986). After the final centrifugation step, vaults were diluted in 40 μl of extract buffer and stored in 5 μl aliquots at –70°C.

Recombinant Xenopus MVP (xMVP)

Full-length Xenopus laevis MVP cDNA was obtained as an EST clone from the German resource centre RZPD (clone IRBHp990E0413D). Sequencing confirmed the identity with the Xenopus laevis MVP sequence (Q6PF69). The xMVP open reading frame was inserted into the hexa-histidine expression vector pET-21a(+) and expressed in the E. coli strain Rosetta(DE3)pLys (Novagen, Darmstadt, Germany). His-tagged xMVP was extracted and purified by using a Ni-NTA column according to the manufacturer's protocol (Qiagen).

To examine whether recombinant xMVP assembles into larger structures upon incubation with cytosol, 7 μg of purified His-tagged xMVP was resuspended in 40 μl cytosol in the presence of an ATP-regenerating system. After 10 minutes at 22°C, the sample was diluted in 500 μl of the extract buffer and subjected to high speed centrifugation at 100,000 g at 4°C for 1 hour. The proteins of the supernatant were precipitated by overnight incubation with 20% trichloroacetic acid. Proteins of the supernatant and pellet were diluted in 15 μl of SDS sample buffer and subjected to immunoblotting analysis.

Antibodies

Antibodies were raised against recombinant His-xMVP in a guinea pig. Mouse monoclonal antibodies to human MVP were purchased from Biotrend (Cologne, Germany; order number BT02-0005-59). Antibodies to p97ATPase were from Dunn Labortechnik (Asbach, Germany; product number K44831M). The guinea pig antisera against Nup62 and mAb 72B9, recognizing the nucleolar protein fibrillarin, have been described (Ewald et al., 1997). LBR was identified with a guinea pig antiserum provided by Christine Dreger (German Cancer Research Center, Heidelberg, Germany) (Dreger et al., 2002). Rabbit antisera against gp 210 and POM 121 were a gift of Iain Mattaj (EMBL, Heidelberg, Germany) (Antonin et al., 2005) and rabbit sera against importin α, importin β and Ran were kindly provided by Mary S. Moore (Department of Anatomy, Ross University School of Medicine, Roseau, Dominica, West Indies).

Microscopy

Samples for immunofluorescence microscopy were processed as described (Dabauvalle et al., 1991). Images were recorded with a Zeiss Axiophot (Carl Zeiss, Oberkochen, Germany) equipped with epifluorescence optics and a CCD camera. EM was performed as described (Dabauvalle et al., 1991).

Gel electrophoresis of proteins and immunoblotting

SDS-PAGE and immunoblot analyses were performed as described previously (Ewald et al., 1997).

We thank Georg Krohne for preparing the guinea pig antibodies against xMVP and many helpful discussions. This work received financial support from the Deutsche Forschungsgemeinschaft (priority program 1175, grant DA 243/4-1).

Alber, F., Dokudovskaya, S., Veenhoff, L. M., Zhang, W., Kipper, J., Devos, D., Suprapto, A., Karni-Schmidt, O., Williams, R., Chait, B. T. et al. (
2007
). The molecular architecture of the nuclear pore complex.
Nature
450
,
695
-701.
Anderson, D. J. and Hetzer, M. W. (
2008
). Shaping the endoplasmic reticulum into the nuclear envelope.
J. Cell Sci.
121
,
137
-142.
Anderson, D. H., Kickhoefer, V. A., Sievers, S. A., Rome, L. H. and Eisenberg, D. (
2007
). Draft crystal structure of the vault shell at 9-A resolution.
PLoS Biol.
5
,
e318
.
Antonin, W., Franz, C., Haselmann, U., Antony, C. and Mattaj, I. W. (
2005
). The integral membrane nucleoporin pom121 functionally links nuclear pore complex assembly and nuclear envelope formation.
Mol. Cell
17
,
83
-92.
Burke, B. and Ellenberg, J. (
2002
). Remodelling the walls of the nucleus.
Nat. Rev. Mol. Cell. Biol.
3
,
487
-497.
Chugani, D. C., Rome, L. H. and Kedersha, N. L. (
1993
). Evidence that vault ribonucleoprotein particles localize to the nuclear pore complex.
J. Cell Sci.
106
,
23
-29.
Dabauvalle, M. C., Loos, K., Merkert, H. and Scheer, U. (
1991
). Spontaneous assembly of pore complex-containing membranes (“annulate lamellae”) in Xenopus egg extract in the absence of chromatin.
J. Cell Biol.
112
,
1073
-1082.
D'Angelo, M. A. and Hetzer, M. W. (
2008
). Structure, dynamics and function of nuclear pore complexes.
Trends Cell. Biol.
18
,
445
-466.
D'Angelo, M. A., Anderson, D. J., Richard, E. and Hetzer, M. W. (
2006
). Nuclear pores form de novo from both sides of the nuclear envelope.
Science
312
,
440
-443.
Dreger, C. K., König, A. R., Spring, H., Lichter, P. and Herrmann, H. (
2002
). Investigation of nuclear architecture with a domain-presenting expression system.
J. Struct. Biol.
140
,
100
-115.
Dultz, E., Zanin, E., Wurzenberger, C., Braun, M., Rabut, G., Sironi, L. and Ellenberg, J. (
2008
). Systematic kinetic analysis of mitotic dis- and reassembly of the nuclear pore in living cells.
J. Cell Biol.
180
,
857
-865.
Ewald, A., Hofbauer, S., Dabauvalle, M. C. and Lourim, D. (
1997
). Preassembly of annulate lamellae in egg extracts inhibits nuclear pore complex formation, but not nuclear membrane assembly.
Eur. J. Cell Biol.
73
,
259
-269.
Gant, T. M. and Wilson, K. L. (
1997
). Nuclear assembly.
Annu. Rev. Cell Dev. Biol.
13
,
669
-695.
Gant, T. M., Goldberg, M. W. and Allen, T. D. (
1998
). Nuclear envelope and nuclear pore assembly: analysis of assembly intermediates by electron microscopy.
Curr. Opin. Cell Biol.
10
,
409
-415.
Goldberg, M. W., Wiese, C., Allen, T. D. and Wilson, K. L. (
1997
). Dimples, pores, star-rings, and thin rings on growing nuclear envelopes: evidence for structural intermediates in nuclear pore complex assembly.
J. Cell Sci.
110
,
409
-420.
Harel, A., Chan, R. C., Lachish-Zalait, A., Zimmerman, E., Elbaum, M. and Forbes, D. J. (
2003
). Importin beta negatively regulates nuclear membrane fusion and nuclear pore complex assembly.
Mol. Biol. Cell
14
,
4387
-4396.
Hartinger, J., Stenius, K., Hogemann, D. and Jahn, R. (
1996
). 16-BAC/SDS-PAGE: a two-dimensional gel electrophoresis system suitable for the separation of integral membrane proteins.
Anal. Biochem.
240
,
126
-133.
Hetzer, M., Bilbao-Cortes, D., Walther, T. C., Gruss, O. J. and Mattaj, I. W. (
2000
). GTP hydrolysis by Ran is required for nuclear envelope assembly.
Mol. Cell
5
,
1013
-1024.
Hetzer, M. W., Walther, T. C. and Mattaj, I. W. (
2005
). Pushing the envelope: structure, function, and dynamics of the nuclear periphery.
Annu. Rev. Cell Dev. Biol.
21
,
347
-380.
Huffman, K. E. and Corey, D. R. (
2005
). Major vault protein does not play a role in chemoresistance or drug localization in a non-small cell lung cancer cell line.
Biochemistry
44
,
2253
-2261.
Kedersha, N. L. and Rome, L. H. (
1986
). Isolation and characterization of a novel ribonucleoprotein particle: large structures contain a single species of small RNA.
J. Cell Biol.
103
,
699
-709.
Macaulay, C. and Forbes, D. J. (
1996
). Assembly of the nuclear pore: biochemically distinct steps revealed with NEM, GTP gamma S, and BAPTA.
J. Cell Biol.
132
,
5
-20.
Maul, G. G. (
1977
). The nuclear and the cytoplasmic pore complex: structure, dynamics, distribution, and evolution.
Int. Rev. Cytol.
Suppl.
75
-186.
Mikyas, Y., Makabi, M., Raval-Fernandes, S., Harrington, L., Kickhoefer, V. A., Rome, L. H. and Stewart, P. L. (
2004
). Cryoelectron microscopy imaging of recombinant and tissue derived vaults: localization of the MVP N termini and VPARP.
J. Mol. Biol.
344
,
91
-105.
Mossink, M. H., van Zon, A., Franzel-Luiten, E., Schoester, M., Kickhoefer, V. A., Scheffer, G. L., Scheper, R. J., Sonneveld, P. and Wiemer, E. A. (
2002
). Disruption of the murine major vault protein (MVP/LRP) gene does not induce hypersensitivity to cytostatics.
Cancer Res.
62
,
7298
-7304.
Newmeyer, D. D. and Wilson, K. L. (
1991
). Egg extracts for nuclear import and nuclear assembly reactions.
Methods Cell Biol.
36
,
607
-634.
Newport, J. (
1987
). Nuclear reconstitution in vitro: stages of assembly around protein-free DNA.
Cell
48
,
205
-217.
Oke, Y. and Inoue, A. (
2003
). The interaction of two classes of nuclear vesicles is induced by the dissociation of soluble proteins from one class of vesicles.
Biol. Cell
95
,
87
-97.
Sasagawa, S., Yamamoto, A., Ichimura, T., Omata, S. and Horigome, T. (
1999
). In vitro nuclear assembly with affinity-purified nuclear envelope precursor vesicle fractions, PV1 and PV2.
Eur. J. Cell Biol.
78
,
593
-600.
Scheer, U. (
1973
). Nuclear pore flow rate of ribosomal RNA and chain growth rate of its precursor during oogenesis of Xenopus laevis.
Dev. Biol.
30
,
13
-28.
Stavru, F., Nautrup-Pedersen, G., Cordes, V. C. and Gorlich, D. (
2006
). Nuclear pore complex assembly and maintenance in POM121- and gp210-deficient cells.
J. Cell Biol.
173
,
477
-483.
Steiner, E., Holzmann, K., Elbling, L., Micksche, M. and Berger, W. (
2006
). Cellular functions of vaults and their involvement in multidrug resistance.
Curr. Drug Targets
7
,
923
-934.
Stephen, A. G., Raval-Fernandes, S., Huynh, T., Torres, M., Kickhoefer, V. A. and Rome, L. H. (
2001
). Assembly of vault-like particles in insect cells expressing only the major vault protein.
J. Biol. Chem.
276
,
23217
-23220.
Suntharalingam, M. and Wente, S. R. (
2003
). Peering through the pore: nuclear pore complex structure, assembly, and function.
Dev. Cell
4
,
775
-789.
Suprenant, K. A. (
2002
). Vault ribonucleoprotein particles: sarcophagi, gondolas, or safety deposit boxes?
Biochemistry
41
,
14447
-14454.
van Zon, A., Mossink, M. H., Houtsmuller, A. B., Schoester, M., Scheffer, G. L., Scheper, R. J., Sonneveld, P. and Wiemer, E. A. (
2006
). Vault mobility depends in part on microtubules and vaults can be recruited to the nuclear envelope.
Exp. Cell Res.
312
,
245
-255.
Vigers, G. P. and Lohka, M. J. (
1991
). A distinct vesicle population targets membranes and pore complexes to the nuclear envelope in Xenopus eggs.
J. Cell Biol.
112
,
545
-556.
Winey, M., Yarar, D., Giddings, T. H., Jr and Mastronarde, D. N. (
1997
). Nuclear pore complex number and distribution throughout the Saccharomyces cerevisiae cell cycle by three-dimensional reconstruction from electron micrographs of nuclear envelopes.
Mol. Biol. Cell
8
,
2119
-2132.
Zheng, C. L., Sumizawa, T., Che, X. F., Tsuyama, S., Furukawa, T., Haraguchi, M., Gao, H., Gotanda, T., Jueng, H. C., Murata, F. et al. (
2005
). Characterization of MVP and VPARP assembly into vault ribonucleoprotein complexes.
Biochem. Biophys. Res. Commun.
326
,
100
-107.

Supplementary information