If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
An Investigator of the Howard Hughes Medical Institute. To whom correspondence may be addressed: Laboratory of Cell Biology, Howard Hughes Medical Institute, The Rockefeller University, 1230 York Avenue, New York, NY 10065. Tel.: 212-327-8096; Fax: 212-327-7880;
To whom correspondence may be addressed: Laboratory of Cell Biology, Howard Hughes Medical Institute, The Rockefeller University, 1230 York Ave., New York, NY 10065. Tel.: 212-327-8181; Fax: 212-327-7880;
* The authors declare that they have no conflicts of interest with the contents of this article. 1 Both authors contributed equally to this work. 2 Present address: Dept. of Chemistry, State University of New York at Binghamton, P. O. Box 6000, Binghamton, NY, 13902.
Three out of ∼30 nucleoporins, Nup62, Nup54, and Nup58, line the nuclear pore channel. These “channel” nucleoporins each contain an ordered region of ∼150–200 residues, which is predicted to be segmented into 3–4 α-helical regions of ∼40–80 residues. Notably, these segmentations are evolutionarily conserved between uni- and multicellular eukaryotes. Strikingly, the boundaries of these segments match our previously reported mapping and crystal data, which collectively identified two “cognate” segments of Nup54, each interacting with cognate segments, one in Nup58 and the other one in Nup62. Because Nup54 and Nup58 cognate segments form crystallographic hetero- or homo-oligomers, we proposed that these oligomers associate into inter-convertible “mid-plane” rings: a single large ring (40–50 nm diameter, consisting of eight hetero-dodecamers) or three small rings (10–20 nm diameter, each comprising eight homo-tetramers). Each “ring cycle” would recapitulate “dilation” and “constriction” of the nuclear pore complex's central transport channel. As for the Nup54·Nup62 interactome, it forms a 1:2 triple helix (“finger”), multiples of which project alternately up and down from mid-plane ring(s). Collectively, our previous crystal data suggested a copy number of 128, 64, and 32 for Nup62, Nup54, and Nup58, respectively, that is, a 4:2:1 stoichiometry. Here, we carried out solution analysis utilizing the entire ordered regions of Nup62, Nup54, and Nup58, and demonstrate that they form a dynamic “triple complex” that is heterogeneously formed from our previously characterized Nup54·Nup58 and Nup54·Nup62 interactomes. These data are consistent both with our crystal structure-deduced copy numbers and stoichiometries and also with our ring cycle model for structure and dynamics of the nuclear pore channel.
) has been difficult to determine experimentally as most procedures employ cell fractionation, where differential losses of nups, especially weakly attached ones, are likely to occur at the numerous theoretical plates of cell fractionation. Nevertheless, symmetry considerations provide important clues. Exhibiting an 8-fold rotational symmetry, each NPC contains at least eight (or multiples thereof) copies of a given nucleoporin. Moreover, as the central core of the NPC displays an additional 2-fold axis of symmetry (coincident with the NPC's mid-plane), the copy number for nucleoporins of the NPC core is at least 16 (or multiples thereof).
A case in point for discrepancies in reported copy numbers and stoichiometries are the three “channel” nucleoporins, Nup62, Nup54, and Nup58 (Refs.
). However, this region is not contiguous, but is subdivided into 3–4 segments of ∼40–80 residues in length, each predicted to exhibit α-helical secondary structure. Remarkably, the topology of segmentation is evolutionarily highly conserved between channel nups of uni- and multicellular eukaryotes indicating functional relevance for the segmentation of each ordered region. Notably, our mapping and crystal studies on the three channel nups matched the boundaries of these segments (
). Collectively, these studies led us to propose a model for the dynamics and structure of the central channel of the NPC. The salient features of this model are mid-plane rings (consisting of “dilated” and “constricted” conformers of Nup54 and Nup58) with attached triple helices of Nup62 and Nup54 (“fingers”), projecting alternately to either the nucleoplasmic or the cytoplasmic side of the mid-plane rings (
) arrived at this conclusion based on analytical ultracentrifugation studies of recombinant fragments comprising the entire ordered region of each of the three channel nucleoporins. However, as these authors utilized only a single loading concentration instead of the multiple loading concentrations that are required for global modeling of multi-component systems to accurately determine equilibrium constants (
) are “non-canonical” in nature and therefore of doubtful physiological relevance.
In this study, we report data on assembly and disassembly of complexes formed by the ordered region of each of the three channel nups. By employing several biochemical and biophysical approaches, we found that these fragments form dynamic assemblies of heterogeneous stoichiometries rather than forming complex(es) of a uniform 1:1:1 stoichiometry as proposed by Ulrich et al. (
), which we refer to as interactomes, to be the principal contributors, even in the context of the heterogeneous and dynamic assemblies formed in solution by the entire ordered regions of the three channel nups. Hence, the previously detected interactomes (
) between segments of the ordered regions of the three channel nucleoporins are indeed “canonical.” The data reported here also provide further support for our “ring cycle” model for the molecular structure and dynamics of the central channel of the nuclear pore complex (
). Briefly, coding sequences of Nup58, Nup54, and Nup62 were cloned in a modular tricistronic vector with a thrombin-cleavable His6 tag at the N terminus of Nup58. The coding sequences of the three genes were separated by an ∼60-bp ribosome-binding site-encoding linker. Protein complex purification was performed as reported earlier (
), using His6 affinity chromatography (through His-tagged Nup58), followed by thrombin cleavage, anion exchange chromatography, and size-exclusion chromatography. After thrombin cleavage, a peptide sequence (GSHM) from the vector precedes the actual Nup58 sequence. Nup54 mutant H469F/Q473L was also expressed as His6-tagged protein from the pET28a vector and purified as described above. All the experiments were performed at 4 °C unless otherwise specified. Protein concentrations were determined from absorbance at 280 nm using extinction coefficients calculated from the protein sequence with the ProtParam tool from the ExPASy webserver (
For size-exclusion chromatography analysis, 200 μl of protein sample at the given protein concentration was loaded on a Superdex200 10/300 GL column (GE Healthcare) that was pre-equilibrated with 10 mm Tris-Cl (pH 8.0), 150 mm NaCl, 1 mm DTT (buffer A), unless otherwise indicated. The column was calibrated in the same buffer with a gel filtration calibration kit (GE Healthcare) using standards of known molecular weights.
Circular Dichroism Spectroscopy Analysis
Secondary structure content of the complex was estimated by far UV wavelength CD scans. CD data are reported as molar ellipticities [Θ]. α-Helical content of the protein complex was estimated from the CD wavelength scans using the webserver K2D3 (
Purified proteins were subjected to analytical size-exclusion chromatography at 4 °C. The size-exclusion chromatography column (Superdex200 10/300 GL) was connected to multiangle light scattering (MALS) and refractive index detectors (DAWN HELEOS and Optilab rEX, Wyatt Technology). The temperature of the Wyatt detectors was controlled at 4 °C (the entire system was placed in a refrigeration unit kept at 4 °C). Weight-averaged molar masses were determined by MALS as described by Solmaz et al. (
) and averaged from at least two experiments (standard deviations were calculated; representative experiments are shown), using the ASTRA 184.108.40.206 software (Wyatt Technology).
Discontinuous Native Gel Electrophoresis
The analysis of proteins using native-PAGE was performed in Tris-glycine buffer (pH 8.0) according to the method of Laemmli with minor modifications. Polyacrylamide gels with 10% acrylamide in the separating gel and 4% acrylamide in the stacking gel (acrylamide:bis-acrylamide, 29:1, respectively) were used. The protein samples were mixed with equal volumes of 2× native loading buffer (250 mm Tris-Cl, pH 6.8, 0.1% bromphenol blue, and 35% glycerol). To prevent denaturation of the protein complex, the samples were run at a constant voltage of 20 V at 4 °C for 14 h. The native gel was negatively stained using the GelCode E-zinc reversible staining kit (Pierce). The identified protein bands were cut, washed with deionized water, and dried. Dry gel was crushed by a pipette tip, and the proteins were eluted with 300 μl of buffer containing 250 mm Tris-Cl (pH 6.8) and 0.1% SDS overnight at room temperature. Eluted proteins were concentrated with a 3-kDa cutoff filtration unit (Millipore). Concentrated proteins were analyzed by SDS-PAGE and silver staining (SilverXpress, Life Technologies) of the SDS-PAGE gel.
Matrix-assisted Depletion Assay
The MBP-Nup62(322–525)·Nup54(346–510)·Nup58(239–415) complex from R. norvegicus was generated by co-expression of the proteins using a polycistronic vector where N termini of Nup58 and Nup62 were tagged to His6 and maltose-binding protein (MBP) tags, respectively. The MBP-Nup62·Nup54·Nup58 complex was purified through His6 affinity chromatography. The His6 tag of Nup58 was cleaved with thrombin. The sample was dialyzed in buffer A. The purified protein complex was bound to amylose resin (New England Biolabs Inc.) equilibrated with buffer A. The MBP-Nup62·Nup54·Nup58-amylose resin was washed with buffer A and used for subsequent studies. 75 μl of amylose-MBP-Nup62·Nup54·Nup58 resin was incubated in a 0.8-ml centrifuge column (Thermo Scientific), at 4 °C with increasing amounts of wild-type Nup54 (453–494) in a final volume of 375 μl. The mixture was then spun at 1000 × g for 2 min at 4 °C. The amount of Nup58 released from the amylose matrix into the supernatant, as a function of the concentration of wild-type Nup54 (453–494), was calculated through absorbance of the supernatant at 280 nm, after the blank subtraction.
Limited Proteolysis and Mass Spectrometry
1 mg/ml of trypsin, chymotrypsin, and elastase stock solutions were separately added to 3 mg/ml of the Nup62·Nup54·Nup58 complex in 1:100 (v/v) ratio. Reaction mixtures were incubated at room temperature, and samples were taken out at different time points. Reactions were quenched by adding guanidine hydrochloride to a final concentration of 5 m. The protein fragments generated in the limited proteolysis experiments were identified as intact peptides measured to within 1 atomic mass unit of mass accuracy, by electrospray ionization-ion trap mass spectrometry (Bruker Amazon), after separation by microbore reversed-phase chromatography on a 300 Å pore polystyrene-divinylbenzene column (PLRP-S, Agilent).
Modular Segmentation of Channel Nucleoporins
The channel nups Nup62, Nup54, and Nup58 contain ordered regions of ∼150–200 residues and natively unfolded regions of up to ∼300 residues that are marked by repeating phenylalanine-glycine (FG) motifs. Fig. 1 (A–C) illustrates the boundaries of the predicted ordered region in each of the three channel nups and indicates segmentation of each region into 3–4 predicted α-helical segments of ∼40–80 amino acid residues in length. As indicated (Fig. 1D), the boundaries of the predicted helical segments of each ordered region closely matched the experimental data (
) we obtained from mapping and crystal structure studies (Fig. 1, A–D). This segmentation pattern is highly conserved between uni- and multicellular eukaryotes (Fig. 2). Each of these segments has been traced and assigned to modular structures that build mid-plane rings and their projecting fingers in a model of a reconstructed central channel, in both its dilated and its constricted forms (
Biophysical and Biochemical Analyses of Triple Complex
We purified a “triple complex” comprising the entire ordered region of each of the three channel nups (Fig. 3A). Size-exclusion chromatography of the purified triple complex at increasing concentrations showed a corresponding shift of its broadly eluting peak to earlier elution positions (Fig. 3B) (
), consistent with dynamic and concentration-dependent mass heterogeneity of the triple complex. We analyzed the complex by circular dichroism spectroscopy to determine its secondary structure content. Wavelength spectra of the triple complex showed no significant difference at various concentrations (0.075–0.6 mg/ml) on the α-helical content of the structure, which was calculated to be about 68% (Fig. 3, C and D). This value closely matched the predicted α-helical content for the complex (Fig. 2). Hence, the observed concentration effects on the elution profile (Fig. 3B) (
) are not attributable to improper folding of α-helical regions.
To assess the range of mass heterogeneity of the broadly eluting peak of the triple complex, we coupled size-exclusion chromatography to MALS analyses to determine molar masses (Fig. 4, A and B). The large range in molar masses exhibited by the entirety of the eluted peak is listed in Fig. 4C. Notably, the molar mass distribution across the peak ranges from 40 to 92 kDa at 2 mg/ml and from 46 to 192 kDa at 5 mg/ml. Within the considerable mass heterogeneity across the elution profile of the triple complex, masses as low as 40–46 kDa (Fig. 4C) were detected in the descending part of the peak. As the theoretical mass of a 1:1:1 triple complex would be ∼64 kDa, the descending part of the peak is made up of two rather than three protein molecules. It should also be noted that in a mixture of heterogeneous assemblies that are in continuous dynamic equilibrium, the individual oligomeric states are not fully separated, and therefore their masses are averaged. Thus, individual assemblies with lower or higher mass may exist. The mass distribution of the ternary complex and its concentration-dependent variability support the observation that the triple complex is heterogeneous and forms multiple oligomeric states. Furthermore, the mass distribution accommodates various structural states expected from our model of the NPC transport channel, and it can be excluded that the complex forms a stable and homogenous 1:1:1 assembly, as suggested previously (
To further assess the degree of heterogeneity of the triple complex, we used multidimensional analysis. Aliquots from fractions across the peak eluted from size-exclusion chromatography (Fig. 5A) were first subjected to electrophoresis under native conditions (Fig. 5B), followed by electrophoresis of the resulting resolved fractions under denaturing conditions (SDS) (Fig. 5C). We observed that each of the four analyzed fractions was composed of distinct mixtures of varying composition of the three channel nups. Notably, the amount of Nup58 is variable. For example, one of these four fractions (Fig. 5B) contained primarily Nup62 and Nup54 when analyzed in a second dimension under denaturing conditions (Fig. 5C). When only one-dimensional SDS-PAGE analysis was carried out on fractions across the peak eluted by size-exclusion chromatography (Fig. 5A), the staining intensity for the three components of the triple complex appeared to be consistent with 1:1:1 stoichiometry of Nup62, Nup54, and Nup58 (Fig. 5D), thus camouflaging the dynamic heterogeneity of the triple complex as illustrated in Fig. 5, B and C.
Disrupting Interactions in the Triple Complex
To probe the stability of the triple complex, we subjected it to size-exclusion chromatography in the presence of the non-ionic detergent n-octyl-β-d-glucopyranoside. Strikingly, the Nup58 component of the triple complex was dissociated during size-exclusion chromatography in 65 mmn-octyl-β-d-glucopyranoside and eluted in later fractions, separated from the peak of complexes containing Nup54 and Nup62 (Fig. 6). These data indicate that Nup58 engages in triple complex formation primarily through hydrophobic interactions, in agreement with our crystal data showing the Nup54·Nup58 interface to be largely hydrophobic in nature (
In a second approach, we probed the preassembled triple complex in a competition assay with a cognate segment of Nup54 (residues 453–494) as used for determination of the crystal structure of the Nup54·Nup58 interactome (
). These components were incubated, and the mixture was subjected to size-exclusion chromatography followed by SDS-PAGE analysis (Fig. 7). Wild-type Nup54 segment pulled Nup58 (comprising the entire ordered region of Nup58) out of the triple complex (Fig. 7B). Competition was quantitated using a matrix-assisted depletion assay (
) in which amylose resin-bound MBP triple complex was titrated with varying amounts of wild-type Nup54 (453–494). The amount of Nup58 depleted from the matrix was monitored through increase in the absorbance in the solution. Indeed, incubation of the triple complex with increasing concentrations of Nup54 segment pulled out increasing amounts of Nup58 (Fig. 7D). We conclude that Nup58 is anchored to the triple complex primarily via its interaction with a cognate segment of Nup54 (
) and that interactions of other segments of Nup58, if they would contribute to triple complex formation, would be too weak to detect in this assay. Strikingly, a thermostable mutant of the Nup54 segment (
) where two crucially localized hydrophilic residues were changed to hydrophobic ones (H469F and Q473L) was unable to accomplish an equivalent pullout of Nup58 from the triple complexes (Fig. 7C). Together these data are consistent with the previously identified Nup54·Nup58 interactome (
) being an indispensable contributor for triple complex formation.
In a third approach, we exposed the triple complex to various proteases and identified the resulting fragments by mass spectrometry, with the expectation that interacting regions should be more protected from proteolysis. Strikingly, the location of protease-protected regions of the triple complex roughly coincided with the boundaries of our previously identified and characterized interactomes (
), either of Nup54 and Nup58 or of Nup54 and Nup62, are indeed principal contributors to the triple complex assembly, we co-expressed various deletion constructs of the three channel nups, purified the resulting complexes by affinity pulldown via the His6 tag of Nup58, and analyzed the eluates by SDS-PAGE (Fig. 9). We found an N-terminally truncated region of Nup58 (Fig. 9, B and E), but not its C-terminally truncated counterpart (Fig. 9, C and F), to be sufficient to efficiently co-purify Nup62 and Nup54 (compare with Fig. 9, A and D). Moreover, a truncated Nup54 cognate segment of Nup62, but not an N-terminally truncated counterpart of Nup62, could be co-purified with Nup54 and Nup58 (Fig. 9, G–J). These data confirm that our crystallographically defined Nup54·Nup58 and Nup54·Nup62 hetero-interactomes are indeed crucial for forming the heterogeneous and dynamic assemblies of the triple complex.
Here, we report solution studies on a triple complex formed by the ∼150–200-residue-long ordered region of each of the three channel nucleoporins, Nup62, Nup54, and Nup58. We show that recombinant proteins (each representing the entire ordered region of the three channel nups) form a dynamic mixture of heterogeneous complexes. We biophysically and biochemically characterized this triple complex and employed truncated forms or cognate segments of the ordered region for probing assembly and disassembly of the triple complex. We find that our previously identified hetero-interactomes (
) formed by cognate segments of the ordered region of the three channel nups are indeed also the principal drivers for triple complex assembly in solution.
Cognate segments of the ordered region of each of the channel nups have previously been identified and characterized by mapping and atomic resolution crystal structures, which led us to propose a model for the structure and dynamics of the central channel of the nuclear pore (
). Two salient structural features of this model are mid-plane rings and attached triple helices (fingers) that form cytoplasmic and nucleoplasmic entries to the mid-plane rings (Fig. 10). With regard to dynamics, we have proposed a ring cycle model, whereby conformers of a single large (dilated) ring convert to conformers of three small (constricted) rings, reflecting opened and closed states of the central channel of the nuclear pore (Fig. 10), with transport factors suggested to stabilize the open form of the central channel (
). Indeed, a recent study yielded data in support of the proposed regulation of the central transport channel. Transport factor binding to the Nup58 segment (extended C-terminally to contain the requisite phenylalanine-glycine repeat binding site for the transport factor) was shown to allosterically stabilize dilated Nup54·Nup58 conformers, at the expense of their constricted counterparts (
Collectively, our studies led to crystal structure-informed copy numbers of 128, 64, and 32 molecules of Nup62, Nup54, and Nup58, respectively, amounting to a mass of 12.3 MDa for the central channel, roughly 10% of the estimated mass of the entire NPC (
). Our model also suggests that the hitherto unassigned segments of the ordered regions of the three channel nups provide binding sites for neighboring nups. Specifically, a segment located in the N-terminal part of the ordered region of Nup58 was proposed to anchor mid-plane rings to neighboring nucleoporins, and a C-terminal segment of the ordered region of Nup62 was proposed to link the base of the fingers to a neighboring nucleoporin. Although structural information for interactions of channel nups with neighboring nups is still missing, there is biochemical evidence for this arrangement (
) and assigned to both the principal elements of the central channel (mid-plane ring and fingers) and putative binding sites in neighboring nucleoporins. Of the three channel nups, Nup54 stands out as it provides two cognate segments, one for forming the mid-plane ring and the other for forming the fingers, respectively, whereas only one segment each of Nup62 and two adjacent segments of Nup58 (Fig. 1D) contribute to the finger and mid-plane elements of the central channel. Notably, the structure-informed stoichiometry of 4:2:1 for Nup62, Nup54, and Nup58 has been supported by in vivo quantitative fluorescence intensity data from the fission yeast where endogenous nups were genomically replaced with their GFP-tagged versions (
). We show that the principal interactions of the triple complex are established by the Nup54·Nup58 interactome (responsible for mid-plane ring) and the Nup54·Nup62 interactome (building the fingers) (FIGURE 9, FIGURE 10). Accordingly, the hitherto “unsaturated” segments of Nup58 and Nup62 that potentially link the structural elements of the central channel to neighboring nucleoporins do not participate in assembly or in disassembly of the triple complex.
Collectively, these studies on the triple complex are entirely consistent with our previous mapping and crystal studies (
) and validate our approach not only to gain insight into the dynamics and structure of the central channel of the NPC, but also to probe the structure and dynamics of surrounding nucleoporins, which are likely to undergo structural and dynamic changes accompanying and accommodating the huge diameter changes proposed for the central channel of the NPC (
A. S, S. R. S., G. B., and I. M. designed experiments, analyzed data, and wrote the paper. A. S., S. R. S., and I. M. performed experiments.
Technical assistance from Amrita Venkateswaran, Ndeye Fatou Gueye, and Hyung Bum Kim is acknowledged. We thank Elias Coutavas, Richard Wing, Thalia Farazi, Erik Debler, Nimisha Singh, and Richa Jaiswal for critical reading of the manuscript and David King for mass spectrometry analysis.