© 2012. Published by The Company of Biologists Ltd.
Karyopherin-independent spontaneous transport of amphiphilic proteins through the nuclear pore
Masahiro Kumeta, Hideki Yamaguchi, Shige H. Yoshimura, and Kunio Takeyasu
Graduate School of Biostudies, Kyoto University, Kyoto 606-8501, Japan
Email: [email protected]
Phone/Fax: +81-75-753-7905
Running title: Spontaneous nuclear transport
Keywords: nucleocytoplasmic transport / nuclear pore complex / surface hydrophobicity
Journal of Cell Science
Accepted manuscript
Corresponding author: Masahiro Kumeta
1
JCS online publication date 3 September 2012
Summary
Highly selective nucleocytoplasmic molecular transport is critical to eukaryotic cells, which is
illustrated by the size-filtering diffusion and karyopherin-mediated passage mechanisms. However, a
considerable number of large proteins without nuclear localization signals are localized to the
nucleus. Here, we provide evidence for spontaneous migration of large proteins in a
Accepted manuscript
karyopherin-independent manner. Time-lapse observation of nuclear transport assay revealed that
several large molecules spontaneously and independently pass through the NPC. The amphiphilic
motifs were shown to be sufficient to overcome the selectivity barrier of the NPC. Furthermore, we
report that the characteristic amphiphilic property of these proteins enables altered local
conformation in hydrophobic solutions, so that elevated surface hydrophobicity facilitates passage
through the nuclear pore. The molecular dynamics simulation revealed the conformational change of
Journal of Cell Science
the amphiphilic structure that exposes the hydrophobic amino acid residues to the outer surface in
hydrophobic solution. These results contribute to the understanding of both nucleocytoplasmic
molecular sorting and the nature of the permeability barrier.
2
Introduction
Nucleocytoplasmic communication is critical to both basal and adaptive activities of eukaryotic cells,
which is accomplished by molecular transport through the nuclear pore complex (NPC). The NPC is
a symmetric, octameric structure that is embedded in the nuclear envelope with an outer diameter of
~120 nm and a height of ~200 nm. The inside of the pore is composed of characteristic subunits that
Accepted manuscript
contain hydrophobic phenylalanine-glycine motifs (FG-Nups) and serves as a selective barrier for
molecular trafficking (Denning et al., 2003). One of the well-known selective properties of the NPC
is size dependent filtration. Water, ions and proteins smaller than ~40 kDa can pass through the NPC
via passive diffusion, whereas proteins larger than ~40 kDa rarely diffuse through the NPC (Mohr et
al., 2009). On the other hand, karyopherin-mediated nuclear transport pathways are known to
strongly influence the subcellular distribution of proteins. Karyopherins recognize specific cargo,
Journal of Cell Science
such as proteins containing nuclear localization signals (NLS, recognized by importin β) and nuclear
export signals (NES, recognized by CRM1), and mediate the passage through the NPC via
hydrophobic interactions with FG-Nups (Mosammaparast and Pemberton, 2004).
However, these canonical nuclear transport pathways are not sufficient to explain the
general nuclear localization of proteins. Bioinformatic analysis revealed that among ≈ 1500 proteins
with localization restricted to the nucleus in Saccharomyces cerevisiae, 57% contain a classical NLS
with the remainder expected to use another mechanism to enter the nucleus (Lange et al., 2007). In
addition to proteins constrained to the nucleus, there are also relevant examples of nuclear transport
of cytoplasmic/cytoskeletal proteins. Interestingly, evidence for the association of large cytoskeletal
proteins with nuclear functions has been accumulating (Kumeta et al., 2011). Actinin-4 interacts with
INO80 chromatin remodeling complex and regulates gene expressions, whereas several different
forms of spectrin are involved in DNA repair and the formation of nuclear bodies. Nuclear
translocation of β-catenin, primarily a focal adhesion component, is a constituent of the canonical
3
Wnt signaling pathway, which is an indispensable step for asymmetric cell division during
embryonic development. Such protein classes often contain amphiphilic helices and include: i)
cytoskeletal proteins containing spectrin repeats (Kumeta et al., 2010); ii) nuclear shuttling/signaling
molecules containing armadillo repeats (Yokoya et al., 1999); and iii) a group of proteins containing
HEAT repeats, to which, interestingly, the karyopherins themselves belong. Recently, it was
revealed through the use of chemically modified BSA that molecular surface hydrophobicity is
focused on the amphiphilic nature of certain proteins and found that these proteins undergo a
conformational change to adapt to a hydrophobic environment, with a resulting increase in surface
hydrophobicity facilitating spontaneous passage through the NPC.
Journal of Cell Science
Accepted manuscript
sufficient to overcome the selectivity barrier of the NPC (Naim et al., 2009). In this study, we
4
Results and Discussion
Karyopherin-independent nuclear transport of large proteins
A nuclear transport assay has been developed for analyzing nuclear transport of proteins using
digitonin-treated semi-permeabilized cells (Adam et al., 1990). We utilized this assay in time-lapse
observations to quantitatively analyze the nuclear transport of large molecules and transport kinetics.
Accepted manuscript
Full-length cDNAs for actinin-4, βI-spectrin, and β-catenin were cloned and GFP-fused recombinant
proteins were purified using the Sf9 insect cell expression system (Fig. 1A). Time-lapse observation
of the nuclear import of purified proteins clearly showed that these proteins continuously migrate
into the nucleus in a karyopherin-independent manner (Fig. 1B,C). GFP-importin-β rapidly
accumulated in the nucleus in a few minutes. It showed slightly stronger signals at the nuclear
envelope, due to its docking activity to the pore (Nachury and Weis, 1999). Compared to
Journal of Cell Science
GFP-importin-β, actinin-4, βI-spectrin, and β-catenin showed slower nuclear transport. It should be
noted that these proteins exceed the size limitation for passive diffusion through the NPC, and the
assay solution does not contain any transport mediators. Fibrous signals of GFP-actinin-4 outside the
nucleus at 300 and 700 seconds are derived from its binding to actin cytoskeleton. β-Catenin showed
aggregated fluorescent signals during the incubation, probably due to its self-association property in
this experimental condition, or some remaining binding partners in the semi-permeabilized cells.
GFP was used as a passively diffusing control. Fluorescein-conjugated 70kDa dextran, a hydrophilic
polysaccharide, serves as a negative control that is excluded from the nucleus. Analyses of nuclear
efflux, observed after removing the extracellular fluorescent molecules, showed that the signals from
actinin-4, βI-spectrin, and β-catenin were decreased (Fig. 1D). These results demonstrated that
nuclear shuttling of these molecules, both import and export, can be achieved in a
karyopherin-independent manner. Importin-β was not actively exported from the nucleus under this
experimental condition, which may be due to strong binding to nuclear factors or the depletion of
5
RanGTP in the nucleus. The signal intensities were quantified and rate constants were calculated by
curve fitting. Rate constants for import (kin) for actinin-4, βI-spectrin, and β-catenin were 5.27, 5.06
and 4.69 (10-3s-1), approximately six times less than that of importin-β (31.86, Fig. 1E). Rate
constants for export (kout) were estimated by analyzing signal influx and efflux, and were similar for
all proteins with the exception of importin-β. Namely, several large molecules spontaneously and
independently passed through the NPC. Compared to the GFP control that showed similar rates of
Accepted manuscript
import and export, the values for other proteins used in this experiment were significantly different,
suggesting that their passages are not the simple diffusion, and their imports and exports are
regulated in different manners.
The amphiphilic spectrin repeat is sufficient for nuclear targeting
A feature that is common to these proteins is their amphiphilic property, which is due to the presence
Journal of Cell Science
of amphiphilic α-helices. β-Catenin contains armadillo repeats and does not require karyopherins for
the nuclear targeting (Malik et al., 1997; Yokoya et al., 1999). Actinins and spectrins possess
multiple spectrin repeats (SRs), which consist of three amphiphilic α-helices bundled in a rod-shaped
structure through the interaction of each hydrophobic surface of the helices (Yan et al., 1993).
To test if the amphiphilic motif itself is generally permeable to the NPC, ten different SR
regions from four typical SR proteins (actinin-1, actinin-4, βI-spectrin and dystrophin) were chosen
to assess the role of amphiphilic SR regions in nuclear targeting (Fig. 2A). Expression of GFP-fused
fragments containing three to six SRs in HeLa cells resulted in consistent nuclear localization (Fig.
2B). In comparison to results obtained using a protein containing three tandem GFP repeats that
lacked an NLS and was not found in the nucleus, it appears that SRs are capable of mediating
nuclear targeting.
The GFP-fused SR fragments were also assessed with a semi-intact nuclear transport assay.
After incubation with GFP-fragments, semi-intact cells were washed and fixed before observation.
6
All fragments accumulated in the nucleus, demonstrating the capability for spontaneous nuclear
transport despite exceeding the theoretical size limitation of the NPC (Fig. 2C). A dominant-negative
importin-β mutant (IBM), which lacks cargo-binding and Ran-binding domains, is known to inhibit
karyopherin-dependent nuclear transport by masking hydrophobic subunits of the NPC (Kutay et al.,
1997). When IBM was added to the assay, nuclear GFP signal was significantly decreased (≈ 50% on
average, Fig. 2D,E). IBM-mediated suppression was found to be concentration dependent (Fig. 2F),
Accepted manuscript
with the kin values for GFP-βI-spectrin SR1-6 decreasing with increasing IBM concentrations (Fig.
2G).
The semi-intact nuclear transport assay clearly demonstrated the spontaneous nuclear
targeting of amphiphilic SR fragments (Fig. 2C). Some of these fragments tend to accumulate in the
nucleus when they are expressed in HeLa cells (Fig. 2B), while the full length SR proteins are
equilibrated in the transport assay (Fig. 1B). In live cells, unidentified transport mechanisms may
Journal of Cell Science
contribute to such accumulation, in addition to their spontaneous passage. Another possibility is that
such fragments are retained in the nucleus by interacting with binding partners, which are easily
removed from the nucleus during digitonin-permeabilization in the semi-intact assay.
Amphiphilic proteins increase their surface hydrophobicity in hydrophobic solutions
A hydrophobic fluorescent probe, 4,4’-bis-1-anilinonaphthalene 8-sulfonate (bis-ANS), is known to
emit fluorescence when it binds to a hydrophobic surface on a molecule. When equal amounts of
purified proteins were subjected to bis-ANS measurement, GFP-fused actinin-4, βI-spectrin, and
β-catenin showed higher fluorescent signal intensities compared to GFP alone or GFP-GST,
consistent with higher surface hydrophobicity (Fig. 3A). When the bis-ANS measurements were
performed in a hydrophobic environment in the presence of increasing amounts of trifluoroethanol
(TFE), the fluorescence from GFP and GFP-GST were decreased due to attenuation of hydrophobic
interactions between the probe and proteins. Interestingly, importin-β and β-catenin exhibited
7
increased bis-ANS binding in the presence of 5% to 10% TFE, and actinin-4 exhibited slightly
enhanced fluorescent signal in 5% to 15% TFE compared to controls. Relative signal intensities
compared with GFP alone clearly demonstrated increased surface hydrophobicity of amphiphilic
proteins in a hydrophobic environment (Fig. 3B).
A helical wheel model of actinin-4 SR3, based on a known NMR structure (PDB: 1WLX),
showed significantly higher hydrophobic amino acid content in the inner face of the helical bundle
Accepted manuscript
compared to the outer surface (Fig. 3C). Starting with this NMR structure, molecular dynamics
simulations were performed to survey conformational changes in SR. In the simulations, the
permittivity constants (represented as “e”) of the solution were set to 70 and 20. In a hydrophilic
solution (e = 70), the three helices are closely aligned and hydrophobic amino acid residues are
folded inside the bundle (Fig. 3D, hydrophobic amino acid residues are shown in red). In contrast, in
hydrophobic environments (e = 20), the helices are bent and exhibit an open structure that exposes
Journal of Cell Science
the hydrophobic residues to the outer surface (Fig. 3E). This simulation analysis supports our
experimental results concerning the surface properties of amphiphilic proteins.
Ultraviolet circular dichroism (CD) spectroscopy was performed to analyze the helical
contents of proteins in hydrophobic environment. The negative peak of CD angle at 222nm is known
to represent the helical contents (Greenfield, 2006). Both actinin-4 and βI-spectrin did not show
significant differences of this value in 0% and 50% TFE-containing solutions (-38662 to -40776, and
-36147 to -36175, respectively) (Fig. 3F), suggesting that each helix was not disrupted in this
experimental condition. The helical contents in each solution were estimated to be 44.2% and 48.1%,
and 45.4% and 47.1%, for actinin-4 and βI-spectrin, respectively, based on the spectrum analyses
using K2D3 prediction program (Louis-Jeune et al., 2011). Combined with surface hydrophobicity
measurement and molecular dynamics simulation, it is reasonable to conclude that configurational
change in amphiphilic helices is sufficient to increase surface hydrophobicity of the protein to
facilitate spontaneous transport through the NPC. We propose that such dynamic conformational
8
change in the molecules is an important factor for NPC permeability, including in the transport of
karyopherins themselves.
Biological implication of spontaneous molecular migration into the nucleus
There is increasing evidence showing the relationships between molecular surface hydrophobicity
and NPC permeability. Karyopherins are known to have a similar molecular conformation and a
Accepted manuscript
greater surface hydrophobicity than other cytoplasmic proteins (Ribbeck and Gorlich, 2002).
Because importin-β interacts with hydrophobic nucleoporins inside the NPC, hydrophobic
interactions are expected to be a driving force for its passage (Bayliss et al., 2000; Bayliss et al.,
2002; Bednenko et al., 2003; Otsuka et al., 2008). These findings suggest the possibility that any
protein that interacts hydrophobically with the NPC can pass through the pore unaided. Indeed, BSA
chemically modified to increase its surface hydrophobicity was shown to overcome the selectivity
Journal of Cell Science
barrier of the NPC (Naim et al., 2009). Considering that hydrophobic interaction is one of the major
driving forces for protein folding/self assembly in general (Gerstman and Chapagain, 2005), it is
likely that a variety of proteins capable of undergoing conformational changes to adapt to the
hydrophobic environment of the NPC can spontaneously migrate in and out of the nucleus.
In contrast to the Brownian ratchet mechanism of molecular transport into the mitochondria,
it is believed that molecules maintain their conformation during passage through the NPC. However,
we propose that certain molecules alter their surface hydrophobicity by undergoing a local
conformational change, enabling spontaneous migration into the nucleus (Fig. 4A,B). This mode of
passage through the NPC is likely to be involved in many important cellular events. Nuclear
export-based mechanisms seem to play a dominant role in subcellular localization of several nuclear
shuttling proteins (Nix and Beckerle, 1997; Petit et al., 2005; Stuven et al., 2003; Wada et al., 1998).
We postulate that the nuclear border is more permissive than previously thought, and that many
proteins, including cytoplasmic proteins, are actually capable of migrating into the nucleus. This
9
spontaneous migration model will provide important clues for unveiling the nature of the NPC, as
well as understanding complicated molecular sorting mechanisms in cells, which are carefully
Journal of Cell Science
Accepted manuscript
orchestrated to achieve basal and adaptive cellular functions.
10
Materials and Methods
cDNA constructs and recombinant proteins
Complementary DNA for GFP was digested from pEGFP-C1 vector at NcoI/BglII sites and cloned
into NcoI/BamHI sites of pFastBac-HT-B vector. cDNAs for actinin-4, βI-spectrin, and β-catenin
were amplified by PCR from HeLa total RNA reverse transcribed with the SuperScript First-Strand
Accepted manuscript
Synthesis System (Invitrogen). cDNA for dystrophin was purchased from Kazusa DNA Research
Institute. The cDNAs were cloned into the pEGFP-C1 vector for expression in HeLa cells. For
protein purification, cDNAs were cloned into the pFastBac-HT-B-GFP vector and expressed in Sf9
insect cells using the Bac-to-Bac baculovirus expression system (Invitrogen).
Nuclear transport assay
Semi-permeabilized cells were prepared by digitonin treatment, as previously described (Kumeta et
Journal of Cell Science
al., 2010). 70kDa fluorescein- and rhodamine B-conjugated dextran were purchased from Molecular
Probes. Microscopic observations were performed at room temperature using a confocal
laser-scanning microscope (LSM 5 PASCAL, Carl Zeiss) with a 63× Plan-Apo objective. Time-lapse
images were taken after incubation with transport substrates. Signal intensities of ten different nuclei
were quantified and statistically analyzed. Curve fitting was performed to obtain rate constants for
import (kin) and export (kout) by applying the following formula: Y = exp(-tkout)Y0 + (kin / kout) X {1 exp(-tkout)} where, X: cytoplasmic (background) signal intensity; Y: nucleoplasmic signal intensity;
t: time; and Y0: Y at time = 0.
Surface hydrophobicity measurement
The hydrophobic fluorescent probe bis-ANS was purchased from Sigma. Proteins (2 µg) were
dissolved in 50 µl PBS containing 0% to 50% v/v TFE and incubated for 5 minutes. Ten micromolar
bis-ANS was added and incubated for 5 minutes. Fluorescence was quantified using a plate reader
11
(Wallac 1420 ARVO SX, Perkin Elmer) with excitation and emission wavelengths of 405 and 460
nm, respectively.
Molecular dynamics simulation
The NMR structure of SR was obtained from a protein structure database (PDB: 1WLX). Model 1 of
20 was selected for analysis. The simulation was performed using the Amber11 molecular dynamics
package and AmberTools (version 1.5) (D.A. Case and Gohlke, 2010). Two Na+ ions were placed
Accepted manuscript
around SR as counter ions. After short energy minimizations, three production runs were conducted
using the generalized Born model (Hawkins et al., 1995; Hawkins et al., 1996; Tsui and Case, 2000)
with dielectric constants of 70 and 20. The time increment was 1 fs and trajectories were calculated
up to 10 ns. No cut-off for long-range interactions was set. The temperature was kept at 300K using
the Langevin thermostat (collision frequency = 1.0 ps-1).
CD spectroscopy
Journal of Cell Science
cDNA for actinin-4 was cloned into pGEX-5X vector (GE) and expressed in bacteria. After
purification by glutathione sepharose beads (GE), GST-tag was excised with Factor Xa protease
(GE). βI-Spectrin was cloned into pFastBac-HT-B vector, expressed in Sf9 cells, and purified with
Ni-NTA agarose (Quiagen). CD spectra were obtained with a J-805 Spectropolarimeter (JASCO)
with temperature controller using a 1mm optical cuvette. Spectra were recorded from 240 nm to 200
nm at 25Co, using 0.05 mg/ml purified proteins in PBS or PBS with 50% v/v TFE. An average of
five runs was obtained by sampling every 0.2 nm with 10 nm/min scanning speed.
12
Acknowledgements
This work was supported by Grand-in-Aid for Research Activity Start-up (to M.K.) from JSPS,
Grant-in-Aid for Scientific Research on Innovative Areas “Spying minority in biological phenomena
(No.3306: 24115512)” (to M.K.) from MEXT, Japan, and Funding Program for Next Generation
World-Leading Researchers (to S.H.Y.) of MEXT, Japan. We thank Dr. Shotaro Otsuka and Mr.
Journal of Cell Science
Accepted manuscript
Yutaka Takashima for helpful discussions and technical supports.
13
References
Ad am, S. A ., Mar r, R. S. a nd Ger ace, L. (1990). Nuclear protein import in permeabilized
mammalian cells requires soluble cytoplasmic factors. J Cell Biol 111 , 807-16.
Ba yliss, R ., Little woo d, T. a nd S te war t, M. (2000). Structural basis for the interaction
between FxFG nucleoporin repeats and importin-beta in nuclear trafficking. Cell 1 02 , 99-108.
Ba yliss, R ., Little woo d, T. , S tra wn, L. A., We nte, S. R. a nd S te w art, M. (2002).
GLFG and FxFG nucleoporins bind to overlapping sites on importin-beta. J Biol Chem 2 7 7, 50597-606.
Bed nenko, J., C ingo la ni, G. a nd Ger ace, L. (2003). Importin beta contains a
Journal of Cell Science
Accepted manuscript
COOH-terminal nucleoporin binding region important for nuclear transport. J Cell Biol 16 2 , 391-401.
D. A. C a se, T. A . D., T. E . C he atham, II I, C. L. Simm erling, J. Wa ng, R. E. D uke, R.
Luo, R. C. Wa lke r, W. Zha ng, K. M. Merz, B. P. Ro ber ts, B. Wa ng, S. H a yik, A . Ro itber g, G.
Se abr a, I . K olossvár y, K. F. Wo ng, F. P ae sani, J. Va nicek, J. L iu, X. Wu, S.R. Bro ze ll, T.
Steinbrec he r, H. a nd Go hlke, Q. C., X. Ye , J. Wa ng, M.- J. H sie h, G. C ui, D.R . Ro e, D.H.
Ma the w s, M. G. See tin, C. S a gui, V. B a bin, T. Luc hk o, S . G usaro v, A . Ko va le nk o, a nd P. A .
Ko llm an. (2010). AMBER 11. University of California, San Francisco.
De nning, D. P. , P a te l, S. S., U ver sk y, V. , Fink, A. L. a nd Re xac h, M. (2003). Disorder
in the nuclear pore complex: the FG repeat regions of nucleoporins are natively unfolded. Proc Natl Acad
Sci U S A 1 00 , 2450-5.
Ger stm a n, B. S. a nd C ha p aga in, P. P. (2005). Self-organization in protein folding and the
hydrophobic interaction. J Chem Phys 1 23 , 054901.
Gree nfield, N. J. (2006). Using circular dichroism spectra to estimate protein secondary
structure. Nat Protoc 1, 2876-90.
Ha wkins, G. D., Cra me r, C. J. a nd Tr uhla r, D. G. (1995). Pairwise solute descreening of
solute charges from a dielectric medium. Chem. Phys. Lett. 2 4 6, 122-129.
Ha wkins, G. D., Cra me r, C. J. a nd Tr uhla r, D. G. (1996). Parametrized Models of
Aqueous Free Energies of Solvation Based on Pairwise Descreening of Solute Atomic Charges from a
Dielectric Medium. J. Phys. Chem. 1 00 , 19824-39.
Kum eta, M., Yo shim ur a, S. H., H ara ta, M . a nd Ta k e ya su, K. (2010). Molecular
mechanisms underlying nucleocytoplasmic shuttling of actinin-4. J Cell Sci 1 23 , 1020-30.
Kum eta, M., Yo shim ur a, S. H., Hej na, J. a nd Ta ke ya su, K. (2011). Nucleocytoplasmic
shuttling of cytoskeletal proteins: molecular mechanism and biological significance. Int J Cell Biol 2 01 2,
494902.
Kuta y, U., I za urr alde, E., Bisc hof f, F. R., Ma ttaj, I. W. a nd Gor lic h, D. (1997).
Dominant-negative mutants of importin-beta block multiple pathways of import and export through the
nuclear pore complex. Embo J 1 6, 1153-63.
La nge, A ., M ills, R. E., L a nge, C. J., S te w art, M., De vine, S. E. a nd C orb ett, A . H.
14
(2007). Classical nuclear localization signals: definition, function, and interaction with importin alpha. J
Biol Chem 2 8 2, 5101-5.
Lo uis- Je une, C., A ndr ade -N a varr o, M . A. a nd Pere z- Ira txe ta, C. (2011). Prediction of
protein secondary structure from circular dichroism using theoretically derived spectra. Proteins.
Ma lik, H. S., Eick b ush, T. H. a nd G oldf arb, D. S. (1997). Evolutionary specialization of
the nuclear targeting apparatus. Proc Natl Acad Sci U S A 94 , 13738-42.
Mo h r, D., Fr e y, S., Fisc he r, T. , G uttle r, T. a nd Gorlic h, D. (2009). Characterisation of
the passive permeability barrier of nuclear pore complexes. Embo J 28 , 2541-53.
Mo sam ma par a st, N. a nd P em ber to n, L. F. (2004). Karyopherins: from nuclear-transport
mediators to nuclear-function regulators. Trends Cell Biol 1 4, 547-56.
Accepted manuscript
be inverted. Proc Natl Acad Sci U S A 9 6 , 9622-7.
Journal of Cell Science
Nac hur y, M. V. a nd We is, K. (1999). The direction of transport through the nuclear pore can
Individual binding pockets of importin-beta for FG-nucleoporins have different binding properties and
Na im, B., Zba id a, D., D aga n, S., K apo n, R. a nd Re ic h, Z. (2009). Cargo surface
hydrophobicity is sufficient to overcome the nuclear pore complex selectivity barrier. Embo J 2 8 ,
2697-705.
Nix, D. A . a nd B ecker le, M. C . (1997). Nuclear-cytoplasmic shuttling of the focal contact
protein, zyxin: a potential mechanism for communication between sites of cell adhesion and the nucleus.
J Cell Biol 1 38 , 1139-47.
Otsuk a, S., Iw a sa ka, S., Yo ne d a, Y. , Ta ke yasu, K. a nd Yo shim ur a, S. H. (2008).
different sensitivities to RanGTP. Proc Natl Acad Sci U S A 1 0 5, 16101-6.
Petit, M. M., Me ulem a ns, S. M., A le n, P. , Ayo ubi, T. A ., J a nse n, E. a nd Va n de Ve n,
W. J . (2005). The tumor suppressor Scrib interacts with the zyxin-related protein LPP, which shuttles
between cell adhesion sites and the nucleus. BMC Cell Biol 6 , 1.
Ribb eck, K. a nd Gor lic h, D. (2002). The permeability barrier of nuclear pore complexes
appears to operate via hydrophobic exclusion. Embo J 2 1 , 2664-71.
Stuve n, T. , H artm a nn, E. a nd Gor lic h, D. (2003). Exportin 6: a novel nuclear export
receptor that is specific for profilin.actin complexes. Embo J 2 2, 5928-40.
Tsui, V. a nd C ase, D. A. (2000). Theory and applications of the generalized Born solvation
model in macromolecular simulations. Biopolymers 56 , 275-91.
Wa d a, A ., F uk uda, M., M ishim a, M. a nd Nishid a, E. (1998). Nuclear export of actin: a
novel mechanism regulating the subcellular localization of a major cytoskeletal protein. Embo J 1 7,
1635-41.
Ya n, Y. , Winogr a d, E., Viel, A., Cro nin, T. , H arriso n, S. C. a nd Br a nto n, D. (1993).
Crystal structure of the repetitive segments of spectrin. Science 2 6 2, 2027-30.
Yo k o ya, F. , I ma mo to, N., Ta c hiba na, T. a nd Yo ne da, Y. (1999). beta-catenin can be
transported into the nucleus in a Ran-unassisted manner. Mol Biol Cell 1 0 , 1119-31.
15
Figure legends
Fig. 1. Nuclear entry of large non-NLS proteins.
(A) GFP and GFP-fused actinin-4, βI-spectrin and β-catenin were expressed in the Sf9 insect cell
expression system and purified. Proteins were loaded onto 5% to 20% acrylamide gels and visualized
with coomassie brilliant blue. (B) Time-lapse observations of nuclear transport of GFP-fused
Journal of Cell Science
Accepted manuscript
proteins in semi-permeabilized HeLa cells were recorded. The images were acquired over an
approximate 12-minute period. Images taken before and after adding the proteins are shown (scale
bar = 10 µm). Fluorescein-conjugated 70kDa dextran was used as a negative control. (C) Averaged
nuclear fluorescent signal. Signal intensities in ten different cells per experiment were measured in
three independent experiments. Data are presented as mean ± SD. (D) Efflux of nuclear GFP
fluorescence was analyzed by removing the fluorescent molecules after the import assay. Data were
processed as described for (C). (E) Rate constants for import (kin) and export (kout) were calculated by
curve fitting of the influx and efflux data, as described in the Methods.
Fig. 2. Nuclear transport of amphiphilic SR fragments.
(A) Ten multiple SR-containing regions from actinin-1, actinin-4, βI-spectrin and dystrophin
(indicated as i-x) were subcloned into GFP-fusion expression vectors. The fragments contain the
following amino acid regions from their respective proteins: i) 286-751; ii) 267-732; iii) 302-953; iv)
955-1590; v) 1592-2129; vi) 342-666; vii) 722-1262; viii) 1269-1779; ix) 1784-2428; and x)
2473-3039. SRs are indicated by boxes and are numbered. (B) The multiple SR-containing fragments
indicated in (A) were expressed in HeLa cells. GFP and a protein containing three tandem GFP
repeats were expressed in HeLa cells, with and without fusion of a NLS, as controls (scale bars = 10
µm). (C) Nuclear transport of multiple SR-containing fragments (as indicated in A) purified from a
16
bacterial expression system. The semi-permeabilized cells were washed and fixed after incubating
with the fragments. (scale bars = 10 µm). (D) Nuclear transport assay in the presence of IBM. IBM
was pre-incubated prior to the incubation with the fragments (scale bars = 10 µm). (E) Quantitation
of the effect of IBM on nuclear transport of SR fragments. Values are presented as mean ± SD. (F) A
concentration-dependent effect of IBM on nuclear transport of GFP-βI spectrin SR1-6. (G) Kin
Accepted manuscript
values in the presence of IBM were calculated by curve fitting as described in Methods.
Fig. 3. Conformational changes of amphiphilic proteins in a hydrophobic solution.
(A) Surface hydrophobicity of amphiphilic proteins was evaluated by measuring the fluorescence
signals of bis-ANS incubated with 2 µg of each protein in PBS supplemented with TFE (0% to 50%,
v/v). Measurements were repeated three times and values are represented as mean ± SD. (B) Relative
values of the bis-ANS signal intensities against GFP, in the presence of 0%, 5% and 10% TFE.
Journal of Cell Science
Values of three independent experiments are represented as mean ± SD. (C) Helical wheel model of
the third SR of actinin-4 based on a NMR model (PDB: 1WLX). Hydrophobic amino acids are
indicated in gray and hydrophilic amino acids in blue. (D) Molecular structure of the third SR of
actinin-4 in a high permittivity condition (e = 70) revealed by a molecular dynamics simulation. The
entire structure is represented in a ribbon model (left), and the zoomed image is shown with
hydrophobic amino acid residues indicated in red (right). The three helices are indicated as H1, H2,
and H3. Indicated amino acids are: A21, F24, W27, M28, A31, M32, L35, M38-V41, I44, I47, L50,
I51, A53, F57, A64, W107, V110, and V114. (E) Molecular structure of the third SR of actinin-4 in a
low permittivity condition (e = 20) is shown in the same manner as in (D). (F) CD spectra of
actinin-4 and βI-spectrin in PBS (lined in blue) or PBS supplemented with 50% v/v TFE (red). The
CD angles (deg cm2 dmol-1) at 222 nm were -38662 and -40776 (actinin-4, in PBS and PBS+50%
TFE), and -36147 and -36175 (βI-spectrin).
17
Fig. 4. The spontaneous migration model.
(A) Spontaneous migration is a carrier-free transport system for proteins with amphiphilic properties.
These molecules increase their surface hydrophobicity in hydrophobic environments, which provide
enhanced interaction with FG-Nups. (B) Properties of transport pathways through the NPC. While
karyopherin itself spontaneously passes through the pore by increasing its surface hydrophobicity, it
is uncertain whether its cargo also adapts to the hydrophobic environment. Spontaneous migration is
determined by release/retention in the nucleoplasm or cytoplasm, in addition of its spontaneous
migrating property.
Journal of Cell Science
Accepted manuscript
a bidirectional passage mechanism. Actual localization of the protein in intact cells may be
18
Fig 1. Nuclear entry of large non-NLS proteins
B
actinin-4
195-
Journal of Cell Science
I-spectrin
-catenin
influx
efflux
kin
kout
kout
(10-3s-1) (10-3s-1) (10-3s-1)
5.27
10.71
10.86
βI-spectrin 5.06
β-catenin 4.69
importin-β 31.86
9.25
7.26
3.66
2.89
11.64
0.08
GFP
8.58
8.31
6.32
dextran
0.94
19.77
-
β
β
importin-
actinin-4
β
dextran
Accepted manuscript
1158962473828208-
20
60
300
dextran
3.5
3
2.5
2
1.5
1
0.5
0
0
D
βI-spectrin
importin-β
actinin-4
β-catenin
GFP
700
relative signal intensity
time (s)
0
m
E
C
relative signal intensity
A
250
500
750
time (s)
1.2
1
0.8
0.6
0.4
0.2
0
0
250
500
750
time (s)
Fig 2. Nuclear transport of amphiphilic SR fragments
A
B
1
Actinin-1
1 2 3 4
892
1
Actinin-4
1 2 3 4
1
1 2 3 4 5 6 7 8 9
iii
Dystrophin
1
1 2 3
1 1 1 1 1 1 1 1
0 1 2 3 4 5 6 7
iv
4 5 6 7 8 9
vii
2364
iii
iv
v
vi
vii
viii
ix
x
GFP
NLS-GFP
GFPx3
v
1 1 1 1 1 1 1 1 1 1
0 1 2 3 4 5 6 7 8 9
viii
2 2 2 2 2
0 1 2 3 4
ix
3685
NLS-GFPx3
x
C
D
i
ii
iii
iv
v
i
ii
iii
iv
v
vi
vii
viii
ix
x
vi
vii
viii
ix
x
E
IBM (-)
IBM (+)
1.4
1.2
1
0.8
0.6
0.4
0.2
0
F
relative signal intensity
relative signal intensity
Journal of Cell Science
Accepted manuscript
vi
ii
ii
i
βI-spectrin
i
911
GST
0.5uM
0.02uM
2uM
0.1uM
1.4
competitor
1.2
1
0.8
0.6
0.4
0.2
0
βI-spectrin
0
dystrophin
G
300
600
900
1200
time (s)
kin
(10-3s-1)
2μM GST
3.44
0.02μM IBM
3.16
0.1μM IBM
3.01
0.5μM IBM
2.58
2μM IBM
2.03
Fig 3. Conformational changes of amphiphilic proteins in a hydrophobic solution
B
signal intensity
160000
140000
GFP
GFP-actinin-4
GFP-β-catenin
GFP-importin-β
GFP-GST
120000
100000
80000
60000
40000
20000
0
0
20
30
40
TFE (%)
50
helix 2
15
10
5
0
actinin-4 β-catenin importin-β
H1
H1
H2
H2
H3
H3
βI-spectrin
0
-10,000
210
220
230
240
(nm)
-20,000
-30,000
θ -40,000
θ
-50,000
-60,000
CD [ ] (deg cm2 dmol-1)
actinin-4
10,000
200
0% TFE
TFE 10%
E
helix 3
F
TFE 5%
20
D
helix 1
CD [ ] (deg cm2 dmol-1)
Journal of Cell Science
Accepted manuscript
C
10
TFE 0%
relative signal intensity
A
50% TFE
10,000
0
-10,000
200
210
220
230
240
(nm)
-20,000
-30,000
-40,000
-50,000
-60,000
0% TFE
50% TFE
GST
Fig 4. The spontaneous migration model
A
B
Size-filtering Diffusion
Spontaneous Migration
protein with
amphiphilic property
small molecule
Size limitation
cytoplasm
Change of surface
hydrophobicity
NPC
Transporter
Journal of Cell Science
Accepted manuscript
nucleoplasm
conformational
change
examples:
water, ions,
small molecules
Transport against
concentration
gradient
Examples
karyopherins,
β-catenin,
SR-proteins
Size-filtering
diffusion
Spontaneous
migration
Karyopherinmediated
transport
＋
－
－
－
＋
－
－
(－)
＋
－
－
＋
water, ions,
small
molecules
karyopherins,
β-catenin,
SR-proteins
NL(E)Sproteins