Crystal structure of importin-a bound to a peptide bearing
the nuclear localisation signal from chloride intracellular
channel protein 4
Andrew V. Mynott
1
, Stephen J. Harrop
1
, Louise J. Brown
2
, Samuel N. Breit
3
, Bostjan Kobe
4,5
and
Paul M. G. Curmi
1,3
1 School of Physics, University of New South Wales, Sydney, NSW, Australia
2 Department of Chemistry and Biomolecular Sciences, Macquarie University, Sydney, NSW, Australia
3 St Vincent’s Centre for Applied Medical Research, St Vincent’s Hospital and University of New South Wales, Sydney, NSW, Australia
4 School of Chemistry and Molecular Biosciences and Centre for Infectious Disease Research, University of Queensland, Brisbane, Qld,
Australia
5 Institute for Molecular Bioscience, University of Queensland, Brisbane, Qld, Australia
Introduction
The importin-a:b nuclear import pathway is one of the
best understood nuclear trafficking systems in the cell
[1]. The pathway operates via the importin-a receptor,
an armadillo (ARM) repeat protein, that recognizes and
binds directly to cargo protein in the cytoplasm. The im-
portin-a:importin-b:cargo complex travels through the
nuclear pore, with importin-b primarily responsible for
negotiating passage through the nuclear pore complex.
This transport process is dependent on the ability of im-
portin-a to recognize specific nuclear localization signals
(NLSs) presented by the cargo protein. The acidic envi-
ronment of the importin-a binding sites confers a high
Keywords
chloride intracellular channel protein; CLIC4;
importin-a; nuclear localization signal (NLS);
nucleocytoplasmic transport
Correspondence
P. Curmi, School of Physics, University of
New South Wales, Sydney, NSW 2052,
Australia
Fax: +61 2 9385 6060
Tel: +61 2 9385 4552
E-mail:
(Received 17 November 2010, revised 31
January 2011, accepted 23 February 2011)
doi:10.1111/j.1742-4658.2011.08086.x
It has been reported that a human chloride intracellular channel (CLIC)
protein, CLIC4, translocates to the nucleus in response to cellular stress,
facilitated by a putative CLIC4 nuclear localization signal (NLS). The
CLIC4 NLS adopts an a-helical structure in the native CLIC4 fold. It is
proposed that CLIC4 is transported to the nucleus via the classical nuclear
import pathway after binding the import receptor, importin-a. In this
study, we have determined the X-ray crystal structure of a truncated form
of importin-a lacking the importin-b binding domain, bound to a CLIC4
NLS peptide. The NLS peptide binds to the major binding site in an
extended conformation similar to that observed for the classical simian
virus 40 large T-antigen NLS. A Tyr residue within the CLIC4 NLS makes
surprisingly favourable interactions by forming side-chain hydrogen bonds
to the importin-a backbone. This structural evidence supports the hypothe-
sis that CLIC4 translocation to the nucleus is governed by the importin-a
nuclear import pathway, provided that CLIC4 can undergo a conforma-
tional rearrangement that exposes the NLS in an extended conformation.
Database
Structural data are available in the protein Data Bank under the accession number
3OQS.
Structured digital abstract
l
CLIC4 and importin alpha bind by x-ray crystallography (View interaction)
Abbreviations
ARM, armadillo; CLIC, chloride intracellular channel; NLS, nuclear localization signal; RSCC, real space correlation coefficient; TAg, simian
virus 40 (SV40) large T-antigen.
1662 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
affinity to clusters of basic residues in the NLS. Mono-
partite NLSs consist of a single cluster of basic amino
acids, approximately six residues long, which generally
interact with the major binding site in importin-a. Struc-
tural studies have shown that an NLS binds importin-a
in an extended conformation, suggesting that functional
NLSs need to be unfolded and flexible within the cargo
protein. Recent studies have demonstrated an interaction
between importin-a and the chloride intracellular chan-
nel (CLIC) protein, CLIC4 [2,3].
The structure of a soluble form of CLIC4 shows
that it adopts the canonical glutathione S-transferase
fold with an N-terminal thioredoxin-like domain and
an a-helical C-terminal domain [4]. CLIC4 can form
poorly selective anion channels that are sensitive to
redox conditions [5] and, like other CLIC family mem-
bers, it is hypothesized that CLIC4 undergoes a struc-
tural transition from the soluble form to an integral
membrane form. CLIC4 is functionally important in
the cell and has recently been implicated in angiogene-
sis, with the observation that suppressed CLIC4
expression leads to the disruption of tubular morpho-
genesis [6]. CLIC4 is upregulated in human and mouse
differentiating cells [7] and has also been implicated in
the regulation of tumour growth [8,9].
In the event of cellular stress, CLIC4 translocates to
the nucleus in human osteosarcoma cells as well as
mouse S1 keratinocytes, where it is involved in an
apoptosis pathway independent of the apoptotic prote-
ase activating factor [2]. After translocation, CLIC4
localizes near the nuclear envelope and in the nucleo-
plasm. Immunoprecipitation experiments have shown
that tumour necrosis factor-a or etoposide treatment
of keratinocytes increases the constitutive interaction
between CLIC4 and various members of the nuclear
import machinery, including Ran, nuclear transport
factor-2 and importin- a [2]. Mutagenesis of a cluster
of basic residues in the putative CLIC4 NLS site
(199KVVAKKYR206 to 199TVVAITYG206) is suffi-
cient to prevent nuclear translocation, suggesting that
this monopartite NLS-like sequence has an active role
in the nuclear import process [2]. This indicates that
the binding of CLIC4 to importin-a via this putative
NLS is responsible for nuclear translocation; however,
in the crystal structure of soluble CLIC4, the putative
NLS adopts a helical conformation that would pre-
clude binding to importin-a (refer to Fig . 2D).
More recently, it has been shown that CLIC4
nuclear translocation is induced in mouse S1 keratino-
cytes by treatment with nitric oxide [3]. The nuclear
translocation is accompanied by S-nitrosylation of a
Cys residue in CLIC4, Cys234. The S-nitrosylation of
CLIC4 has been found to induce a conformational
change which destabilizes the native conformation.
Such a destabilization may facilitate the interaction
between the otherwise helical CLIC4 NLS and impor-
tin-a. It has been shown that S-nitrosylation of CLIC4
enhances the interaction with importin-a, as deter-
mined by immunoprecipitation [3].
In this article, we present the X-ray crystal structure
of mouse importin-a (70–529) bound to a peptide cor-
responding to the CLIC4 NLS. The importin-a (70–
529) construct used to obtain the importin-a:CLIC4
NLS complex lacks the first 69 residues that corre-
spond to the flexible importin-b binding domain. The
importin-b binding domain is known to have an
autoinhibitory function, whereby an internal NLS-like
sequence competes for the importin-a binding site,
reducing binding affinity for cargo proteins and help-
ing to facilitate the release of the cargo within the
nucleus [10,11]. The removal of the autoinhibitory
domain to create a truncated importin-a avoids possi-
ble competition for the binding site between this inter-
nal NLS and the CLIC4 NLS peptide.
The importin-a C-terminal domain (residues 70–529)
consists of 10 ARM structural repeats that form two
well-characterized cargo binding sites, referred to as
the major and minor binding sites [1]. These sites are
located in the concave face of the protein near regions
of invariant Trp and Asn arrays. The major binding
site spans ARM repeats 2, 3 and 4 (Fig. 1A). In the
major binding site, the positions of six NLS residues
are labelled P1–P6, following the directionality of a
bound NLS from the N-terminus to the C-terminus,
which runs antiparallel to importin-a.
Our structure shows that the CLIC4 NLS peptide
binds to the importin-a major binding site in an
extended conformation consistent with a classical im-
portin-a
:NLS complex. There is no clear interaction
between the CLIC4 NLS peptide and the minor bind-
ing site of importin-a. In the major binding site, elec-
tron density clearly defines the peptide residues
201VAKKYRN207, which have been included in the
final model with Lys203 occupying the critical P2 bind-
ing position. Our results reveal that Lys199 at the
putative CLIC4 NLS N-terminus is disordered in the
crystal and is therefore not necessary for peptide bind-
ing. The core binding pockets P2–P5 are occupied by
residues KKYR, a rather atypical NLS because of the
presence of a bulky aromatic Tyr residue in the P4
binding position. Surprisingly, the Tyr205 side-chain is
favourably placed at P4, forming hydrogen bonds with
the importin-a main chain. An analysis of normalized
B factors demonstrates a localized reduction in atomic
flexibility experienced by importin-a residues as a
result of the binding of the CLIC4 NLS peptide.
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1663
The importin-a:CLIC4 NLS structure presented in
this article adds to a growing body of knowledge on
the structural mechanisms that govern the classical
nuclear import model. It also clarifies that the CLIC4
NLS can indeed bind directly to importin-a on condi-
tion that it can unfold into an extended conformation.
Results
Structure of the importin-a:CLIC4 NLS peptide
complex
The structure of importin-a (70–529) bound to the
CLIC4 NLS peptide (198VKVVAKKYRN207) was
solved at 2.0 A
˚
resolution using synchrotron radiation
(Table 1). The model of importin-a in the CLIC4 NLS
peptide complex includes residues 72–496 and closely
resembles the full-length importin-a structure that
incorporates the N-terminal autoinhibitory domain
(PDB:1IAL, rmsd of 0.20 A
˚
across 425 C
a
atoms in
residues 72–496). The major binding site spanning
ARM repeats 2–4 has a similar conformation to the
equivalent region in apo importin-a (70–529), with an
rmsd of 0.16 A
˚
(46 C
a
atoms) across the inner H3 heli-
ces, suggesting that there are minimal backbone con-
formational changes as a result of peptide binding.
Electron density corresponding to residues 201–207
(VAKKYRN) of the CLIC4 NLS peptide was unam-
biguously identified in the importin-a major binding
site between ARM repeats 2–4. The F
o
) F
c
map con-
structed by omitting the peptide from model phases is
shown in Fig. 1A. The importin-a minor binding site
contains no electron density that unambiguously corre-
sponds to the CLIC4 NLS, and therefore no peptide
was modelled at this site. The CLIC4 NLS binds in an
extended conformation that runs antiparallel to impor-
tin-a, analogous with other NLS cargo. The average
atomic B factors for importin-a in the structure are
32.1 A
˚
2
for main-chain atoms, 35.7 A
˚
2
for side-chain
atoms and 33.8 A
˚
2
overall (3244 atoms). For the pep-
tide, B factors are slightly higher than those for
importin-a: 36.9 A
˚
2
for main-chain atoms, 40.2 A
˚
2
for
side-chain atoms and 38.7 A
˚
2
overall (62 total atoms).
Electron density analysis
Both the main-chain and side-chain atoms of the mod-
elled CLIC4 peptide show a good fit to the electron
density (Fig. 1B). The peptide residues 202–207 corre-
spond to the key binding positions P1–P6, with the
critical P2 position occupied by Lys203. This means
that the core basic motif, 203KKYR206, fills the
central binding pockets P2–P5 in which the majority
of peptide side-chain interactions take place with
importin-a. Therefore, the CLIC4 NLS satisfies the
accepted consensus sequence for an optimal NLS,
P2
K(K ⁄ R)·(K ⁄ R)
P5
[12]. The residue at P4, which has
been shown to contribute the least, energetically, to
peptide binding of the four main binding pockets [12],
is unambiguously occupied by Tyr205 as defined by
2F
o
) F
c
electron density.
The N-terminal peptide residues, 198VKV, are disor-
dered in the crystal and are thus likely to be highly
flexible. It is particularly noteworthy that the basic res-
idue, Lys199, defined as part of the putative CLIC4
NLS (KVVAKKYR) [2], does not contribute to pep-
tide binding. If the N-terminal flanking region
increases CLIC4 NLS affinity for importin-a,itis
unclear how it does so from our crystal structure. The
terminating carboxyl group of the peptide at Asn207 is
well defined in the 2F
o
) F
c
electron density map,
despite making no interactions with importin-a.
The presence of a Tyr residue at P4 is a strong indi-
cation that the bound peptide corresponds to the
CLIC4 NLS. There is also weak and unaccounted for
density approximately 4 A
˚
from the aromatic plane of
Tyr205 in a position that may correspond to a cation–p
interaction. The cation in this case is likely to be an
Na
+
ion from the crystallization buffer, with a low
occupancy (< 50%). As a result of the weak nature of
the Tyr205 cation–p bond, it seems unlikely that it will
have a significant effect on the CLIC4 NLS peptide
binding to importin-a.
We have also analysed the veracity of the CLIC4
NLS model built in the major binding site by inspect-
ing the F
clic4nls
o
À F
apo
o
data–data difference Fourier (see
Materials and methods). This Fourier analysis reduces
bias when interpreting the density of a peptide bound
to importin-a and thus provides additional support for
our structure. The results are shown in Fig. 1D. As
expected, the electron density is strong along the pep-
tide main chain with well-defined carbonyl and amide
backbone groups. The one exception to this is the
location of the amide group of Lys204 at P3, where
there is a break in the main-chain density at the 2.8r
map level. The corresponding position in the apo struc-
ture has particularly strong density at this point, which
may correspond to a water molecule. Peptide side-
chain density is also well defined in the F
clic4nls
o
ÀF
apo
o
map. At P1, the density is weak, but resembles Ala202.
The Lys residue at P2 (Lys203) is well defined despite
the presence of a partially occupied water at this
location in the apo structure. Lys204 at P3 and Arg206
at P5 are also well defined. Perhaps the most definitive
characteristic of the F
clic4nls
o
ÀF
apo
o
difference Fourier
is the strong and unambiguous electron density
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1664 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
corresponding to Tyr205 at P4. There is a strong posi-
tive difference density peak near the Tyr205 O–C
f
bond at the 9.1r map level, the strongest density peak
in the F
clic4nls
o
ÀF
apo
o
difference map. The presence of
Tyr205 is definitive evidence that the CLIC4 NLS
peptide binds importin-a (70–529).
CLIC4 NLS interactions with importin-a
The CLIC4 NLS forms an extensive network of inter-
actions with importin-a through both main-chain and
side-chain atoms, similar to other importin-a structures
with a bound monopartite NLS [13–15]. Hydrogen
Fig. 1. The importin-a:CLIC4 NLS peptide complex. (A) The F
o
) F
c
‘omit’ electron density map over all atoms in importin-a. Positive con-
tours are shown at 2.8r in grey. Density corresponding to the bound CLIC4 NLS peptide is clearly visible in the major binding site. (B) Ste-
reoimage of the CLIC4 NLS peptide and 2F
o
) F
c
map. The CLIC4 NLS peptide bound to the importin-a major binding site is shown as a
stick representation. Colour code for atoms: carbon, cyan; nitrogen, blue; oxygen, red. Electron density is contoured at 1.5r in grey. Binding
positions P1–P6 and the N- and C-termini are labelled. (C) Schematic representation of hydrogen bonds (broken lines, < 3.5 A
˚
) between
importin-a and the CLIC4 NLS peptide,
P1
AKKYRN
P6
. Backbone carbonyl oxygens and amide nitrogens are shown as red and blue spheres,
respectively. Nitrogen and oxygen side-chain atoms are shown as blue and red squares, respectively. (D) Stereoimage of the CLIC4 NLS
bound to importin-a, showing the F
clic4nls
o
ÀF
apo
o
data–data difference Fourier map. Grey contours represent positive difference density at
2.8r. (E) Stereoimage of hydrogen-bond interactions (broken lines, < 3.5 A
˚
). The CLIC4 NLS peptide is shown as a ball and stick represen-
tation, where carbons are black, nitrogens are blue and oxygens are red. Importin-a is shown in cartoon representation (cyan) with bonded
residues shown as sticks.
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1665
bonding by the main chain of the CLIC4 NLS peptide
involves importin-a side chains in the conserved
WxxxN motifs of ARM repeats 2-4, which include res-
idues Trp142, Trp184, Asn146, Asn188 and Asn235
(Fig. 1C). In the CLIC4 NLS peptide, this corresponds
to hydrogen-bonded carbonyl and amide groups from
every second residue in the major binding site: P1
(Ala202), P3 (Lys204) and P5 (Arg206). The peptide
side chains in binding positions P2 (Lys203), P4
(Tyr205) and P5 (Arg206) form hydrogen bonds to the
importin-a main chain and side chains. In addition,
Lys203 forms a critical salt bridge with
impa
Asp192
(‘impa’ denotes importin-a; bond length, 2.80 A
˚
;
N
f
) O
d
) at P2, the most energetically significant inter-
action involved in importin-a recognition of NLSs
[16,17].
Other basic residues in the peptide, Lys204 and
Arg206, fill negatively charged pockets at P3 and P5
Fig. 2. Analysis of the importin-a:CLIC4 NLS peptide complex. (A–C) Importin-a is coloured by the normalized B factor score, B
Àapo
z
, over a
blue–magenta colour spectrum ()3r to +3r). (A) The bound CLIC4 NLS is shown on the molecular surface of importin-a in the major binding
site. (B) The importin-a C
a
backbone is shown as a cartoon tube representation in the same orientation as in (A). Important residues are
shown as a stick representation. (C) Full-length images of importin-a coloured by the B
Àapo
z
score. Residues in grey have not been included
in the calculation of B
Àapo
z
. (D) The CLIC4 crystal structure (PDB:2AHE) is shown as a cartoon representation. The N-terminal thioredoxin
domain (blue) and C-terminal a-helical domain (green) are coloured separately. The CLIC4 NLS residues are highlighted in cyan. Inset: The
NLS is shown as a stick representation (carbons, cyan; oxygens, red; nitrogens, blue). Hydrogen bonds are represented by broken lines. (E)
A multiple sequence alignment of the CLIC4 NLS motif in human CLICs. Conserved residues are red, nonconserved residues are black and
perfect conservation is highlighted with red fill. The sequence of CLIC3 is added for comparison. Binding positions P1–P6 are shown in
an alignment corresponding to our importin-a:CLIC4 NLS peptide complex. Sequence alignment was performed using
CLUSTALW [44] and
ESpript [45].
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1666 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
without forming salt bridges. At P3, there is a pocket
formed between
impa
Trp184 and
impa
Trp231 which
favourably accepts the extended and positively charged
Lys204 side chain, with
impa
Glu266 (5.8 A
˚
,N
f
) O
e
distance) and
impa
Asp270 (4.5 A
˚
,N
f
) O
d
distance)
positioned at the end of the binding pocket. Similarly,
the P5 pocket formed between
impa
Trp142 and
impa
Trp184 favourably accepts Arg206, with
impa
Glu180
(4.9 A
˚
,N
f1
) O
e1
distance) positioned at the end of the
binding pocket.
In total, the CLIC4 NLS main chain and side chains
make 174 atom-to-atom van der Waals’ contacts with
importin-a and 13 hydrogen bonds, including one salt
bridge at P2. In addition, there are 10 hydrogen bonds
formed between the peptide and water molecules, three
of which involve the terminating carboxyl group. The
van der Waals’ contact area between the CLIC4 NLS
peptide and importin-a has been calculated for each
peptide residue by integrating over contact areas using
MolProbity [18,19]. Main-chain contributions to the
contact area were found to be approximately equal
(2–5 A
˚
2
). Side-chain contributions vary to a greater
extent, reflecting differences in the binding pockets.
The Lys and Arg residues in binding positions P2 and
P5 have the largest side-chain van der Waals’ contact
areas, both corresponding to 15.9 A
˚
2
. This is closely
followed by the Tyr residue at P4, with a contact area
of 13.5 A
˚
2
and the Lys residue at P3 with a contact
area of 10.2 A
˚
2
. Residues outside the central binding
area have significantly lower values. These results are
summarized in Table 2.
The solvent-accessible surface area on importin-a
buried by the peptide is 513.1 A
˚
2
, with the largest con-
tribution of buried surface area from the Trp array
that includes Trp142 (51.3 A
˚
2
), Trp184 (70.2 A
˚
2
) and
Trp231 (63.7 A
˚
2
). The surface area buried on the pep-
tide is 744.7 A
˚
2
, which corresponds to 59.2% of the
total peptide surface area. The real space correlation
coefficient (RSCC) for each residue has also been cal-
culated by comparing the importin-a:CLIC4 NLS
experimental electron density with density calculated
from the model. The CLIC4 NLS main chain fits the
density well, with an average RSCC of 0.96. Side
chains have greater RSCC variability, with an average
of 0.89 over all residues and 0.94 for those in P2-P5
(Table 2).
The CLIC4 NLS Tyr residue
By solving the structure of the importin-a:CLIC4 NLS
complex, we have shown that the major binding
pocket P4 is unambiguously occupied by a Tyr residue:
the first importin-a:NLS structure that has an aromatic
residue present in the core binding region. Tyr205
adopts a common rotamer with a score of 82.9%
(v
1
$ 180, v
2
$ 80), calculated by comparing the side
chain with a high-quality reference dataset using Mol-
Probity [18]. Interestingly, it appears that this unique
NLS residue is not only accommodated in the P4 site,
but makes significant interactions with importin-a.
This primarily occurs by the formation of a hydrogen
bond between the side-chain hydroxyl group of Tyr205
and the importin-a main chain at the C-terminus of
the ARM1 H3 helix. The hydrogen bond is possibly
shared between the carbonyl oxygen of
impa
Leu104
(bonding distance from oxygen to oxygen, 2.83 A
˚
) and
the carbonyl oxygen of
impa
Arg106 (oxygen to oxygen,
2.64 A
˚
). Although the amide nitrogen of
impa
Arg106
(N–O, 3.16 A
˚
) is of hydrogen-bonding distance, the
geometry for this interaction is not favourable. Analy-
sis of importin-a:CLIC4 NLS contacts using MolPro-
bity suggests that the Tyr205–Arg106 (O–O) hydrogen
bond is the most favourable with optimal geometry
(refer to Fig. S1).
The bulky side chain of Tyr205 also makes extensive
hydrophobic interactions with surrounding importin-a
Table 1. Data collection and refinement statistics.
Data collection
Source (k) Australian Synchrotron
(0.95 A
˚
)
Detector ADSC Q210
Space group P2
1
2
1
2
1
Unit cell dimensions (A
˚
): a, b, c 78.6, 89.6, 100.1
Resolution (A
˚
)
a
2.0 (2.11–2.00)
Observations 318 168 (27 080)
Unique reflections 46 758 (5520)
Completeness (%)
a
96.7 (80.3)
Mean I ⁄ r
a
12.2 (1.8)
R
merge
(%)
a,b
9.6 (77.6)
Wilson B value (A
˚
2
) 30.4
Refinement and structure
R factor (%)
c
19.6
R
free
(%) 23.6
Scaling R factor (R
iso
)
d
16.3
Number of nonhydrogen atoms (mean B value, A
˚
2
)
Importin-a 3238 (35.2)
Peptide 62 (40.4)
Water 356 (45.5)
Ramachandran plot (%)
e
Favoured region 98.6
Allowed region 1.4
Disallowed 0
a
Outer shell statistics are shown in parentheses.
b
R
merge
= R
hkl-
R
i
|I
i
) <I>| ⁄ R
hkl
R
i
I.
c
R factor = R
hkl
||F
obs
| ) |F
calc
|| ⁄ R
hkl
|F
obs
|.
d
R
iso
¼
P
hkl
jF
clic4nls
o
ÀF
apo
o
j=
P
hkl
jF
apo
o
j.
e
Calculated by MolProbity [18].
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1667
residues, whereby the phenol ring fits into a hydropho-
bic pocket formed by the loop connecting ARM
repeats 1 and 2 (Leu104–Pro111). Indicative of the
tightness of the fit, there are a large number (30) of
atom-to-atom van der Waals’ contacts between the
Tyr205 side chain and importin-a (Table 2). The num-
bers of contacts are comparable with those of the basic
residues at P2 (Lys203: 31 contacts), P3 (Lys204: 20
contacts) and P5 (Arg206: 45 contacts).
Normalized B factor analysis
In order to analyse changes in the conformational
dynamics of importin-a binding site residues, as a
result of the presence of a bound CLIC4 NLS peptide,
the relative B factor score, B
Àapo
z
, was calculated. This
score represents a change in flexibility of each residue
in apo importin-a and the corresponding residue in the
importin-a:CLIC4 NLS complex. The B
Àapo
z
score was
determined by comparing normalized B factors from
the importin-a:CLIC4 NLS structure with those from
apo importin-a (see Materials and methods). A nega-
tive score represents a decrease in flexibility, a positive
score represents an increase in flexibility and a score of
zero represents no change. Residues near the importin-
a C-terminus (430–496) were considered to be outliers
(Z > 4) and are thus excluded from the analysis. The
B
Àapo
z
scores have a zero mean and standard deviation
of unity (Fig. S1). The B
Àapo
z
score is colour mapped
onto a molecular surface representation of the impor-
tin-a molecule in Fig. 2. Here, the scores were aver-
aged over main-chain (B
Àapo mc
z
) atoms and side-chain
(B
Àapo sc
z
) atoms for each residue.
We note that the major binding site corresponds to
a cluster of negative B
Àapo
z
scores, demonstrating a
reduced flexibility on binding the CLIC4 NLS peptide,
whereas the minor binding site is unchanged. This is
particularly significant over the Trp and Asn arrays,
which includes residues that interact directly with the
CLIC4 NLS peptide backbone: Asn146 ()4.8
sc
);
Asn188 ()2.4
sc
); Asn235 ()2.0
sc
); Trp142 ()2.2
sc
);
Trp184 ()3.9
sc
). In addition, Ser149 ()3.3
sc
) is notably
constrained by the CLIC4 NLS in a single conforma-
tion compared with dual rotamer conformations in the
apo structure.
The presence of the Lys203 side chain at P2 has
marginal effects on the hydrogen-bonding partners
Thr155 ()0.6
res
) and Asp192 ()0.8
res
), but a significant
effect on the third hydrogen-bonding partner, the car-
bonyl group of Gly150 ()3.1
res
). This can be explained
by the presence of a water molecule in the apo struc-
ture which is predicted to make hydrogen bonds with
Thr155 and Asp192, but not Gly150. We also note
that the Lys N
f
–Gly150 bond has the added effect of
reducing flexibility in the connecting loop between
ARM repeats 1 and 2 (Ser149–Ser152).
At the P4 binding site, we see that Tyr205 has a sta-
bilizing effect on residues in the ARM1–ARM2 con-
necting loop. The hydrogen bond formed between
Tyr205 and the carbonyl oxygen of
impa
Arg106
()2.8
res
) decreases significantly the B
Àapo
z
score for the
residue. This is also seen for the other Tyr205-shared
hydrogen-bonding partner,
impa
Leu104 ()2.3
res
). Van
der Waals’ contacts involving Tyr205 appear to stabi-
lize both
impa
Pro110 ()1.2
res
) and
impa
Ser105 ()3.3
res
).
The CLIC4 NLS tightens residue mobility across the
ARM2–ARM3 connecting loop from Leu103 to Ile112.
Arg206 at P5 forms a hydrogen bond with
impa
Gln181
which has a corresponding decrease in its B
Àapo
z
score of
)1.9
sc
. This B factor analysis does not reflect significant
changes in residue flexibility as a result of long-range
electrostatic interactions between CLIC4 NLS side
chains at P3 (Lys204) or P5 (Arg206) and acidic resi-
dues in importin-a (Glu180, Glu266, Asp270).
Table 2. Characteristics of the importin-a:CLIC4 NLS peptide interaction.
NLS residue
Interactions
a
Van der Waals’ contact area (A
˚
2
)
b
Buried surface
area (A
˚
2
) RSCC
c
Main chain Side chain Main chain Side chain Total
Val201 0 ⁄ 1 ⁄ 00⁄ 0 ⁄ 0 0.4 0 0.4 26.7 0.89 ⁄ 0.74
P1
Ala202 1 ⁄ 8 ⁄ 00⁄ 6 ⁄ 0 3.5 5.3 8.8 89.4 0.91 ⁄ 0.81
P2
Lys203 0 ⁄ 6 ⁄ 13⁄ 31 ⁄ 1 3 15.9 18.9 145.3 0.96 ⁄ 0.97
P3
Lys204 3 ⁄ 5 ⁄ 01⁄ 20 ⁄ 0 2.2 8.0 10.2 128.4 0.98 ⁄ 0.92
P4
Tyr205 0 ⁄ 8 ⁄ 01⁄ 30 ⁄ 1 3.3 13.5 16.8 133.8 0.96 ⁄ 0.93
P5
Arg206 3 ⁄ 9 ⁄ 01⁄ 45 ⁄ 4 5.6 15.9 21.5 174.8 0.98 ⁄ 0.94
P6
Asn207 0 ⁄ 6 ⁄ 30⁄ 0 ⁄ 0 2.8 0 2.8 46.3 0.96 ⁄ 0.79
Totals 7 ⁄ 42 ⁄ 46⁄ 132 ⁄ 6 20.8 58.6 79.4 744.7 0.96 ⁄ 0.89
a
Shown as importin-a hydrogen bonds ⁄ van der Waals’ contacts ⁄ ordered solvent hydrogen bonds.
b
Calculated using MolProbity by integrating
contact dots with 16 dots ⁄ A
˚
2
[18].
c
The real space correlation coefficient (RSCC) is listed as main chain ⁄ side chain.
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1668 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
Discussion
The importin-a:CLIC4 NLS peptide complex
In this article, we have presented the 2.0 A
˚
resolution
X-ray crystal structure of the CLIC4 NLS peptide
(198VKVVAKKYRN207) bound to importin-a (70–
529). The monopartite NLS peptide binds in the major
binding groove of importin-a in an extended confor-
mation consistent with previously solved structures [1].
In the case of CLIC4, this extended NLS conformation
differs greatly from its helical conformation as seen in
the soluble CLIC4 structure (Fig. 2D). This important
feature means that the putative CLIC4 NLS needs to
undergo a structural transition if it is to be a biologi-
cally active NLS. It also suggests that there is tighter
control over CLIC4 translocation to the nucleus in
comparison with other nuclear-destined proteins that
do not require their NLS to undertake a structural
rearrangement.
The F
clic4nls
o
ÀF
apo
o
data–data difference Fourier was
inspected to ensure that electron density in the impor-
tin-a major binding site is not confused with residual
density from the apo importin-a structure. The results
of this difference Fourier show that all peptide side
chains in P1–P6 positions have well-defined
F
clic4nls
o
ÀF
apo
o
density, supporting the assignment of
CLIC4 NLS residues in these binding positions. This
analysis is therefore a recommended method of struc-
tural examination in other importin-a:NLS complexes
to ensure the correct identification of the bound
peptide.
Side-chain interactions involving CLIC4 NLS Lys
residues at P2 and P3 are equivalent to those observed
in the importin-a:simian virus 40 (SV40) large T-antigen
(TAg) NLS complex [13]. This includes the key electro-
static interaction at P2 involving Lys203 and
impa
Asp192. In classical NLSs, the Lys at P2 is known to
provide the most substantial energetic contribution to
peptide binding [12] and is critical for nuclear import,
which can be eliminated through mutagenesis of this
single amino acid [16]. Although the work presented by
Suh et al. [2] obliterated the putative CLIC4 NLS motif
(199KVVAKKYR206 to 199TVVAITYG206), rendering
CLIC4 unable to translocate to the nucleus, a single
point mutation at Lys203 should be sufficient to prevent
nuclear translocation.
A prominent difference between the CLIC4 NLS
and classical NLSs is the presence of a bulky aromatic
residue within the core NLS motif. Our structure
shows how a Tyr residue can be accommodated in the
major binding site at P4, and its presence is confirmed
by strong, unambiguous density in the F
clic4nls
o
ÀF
apo
o
difference map. We note that a structure of the phos-
pholipid scramblase 1 NLS bound to importin-a has
been reported, in which a Trp residue is located three
residues downstream from P4 [20].
The Tyr residue is found to interact with importin-a
through both hydrophobic and hydrogen-bonding
interactions. Although it is an unusual NLS residue
containing a bulky phenol side group, it neatly fits at
its location with the hydroxyl group reaching the im-
portin-a main chain near the C-terminus of the ARM1
H3 helix. Most of the hydrophobic contacts with im-
portin-a residues lining the P4 pocket are equivalent to
those made by the aliphatic portion of an Arg residue
in the TAg NLS, whereas the hydrogen bonds shared
between the Tyr hydroxyl group and carbonyl groups
from
impa
Leu104 and
impa
Arg106 are reproduced in the
importin-a:TAg NLS P4 binding position involving the
N
f
atom of
TAg
Arg130. Because of the proximity of
the Tyr205 hydroxyl group to importin-a, CLIC4 NLS
peptide binding should be prevented by Tyr phosphor-
ylation. Therefore, the phosphorylation state of CLIC4
could foreseeably act as a switch between an ‘active’
and ‘inactive’ CLIC4 NLS.
Structural studies have shown that the importin-a
P4 binding position tolerates a wide range of residues,
including Arg from the TAg NLS (PDB:1EJL [13]),
Lys in the nucleoplasmin NLS (PDB:1EE5 [21]) and
androgen receptor NLS (PDB:3BTR [14]), Val in the
c-myc NLS (PDB:1EE4 [21]), Leu in the retinoblas-
toma NLS or Ser in the N1N2 NLS (1PJM and 1PJN
[22]) and Ile in the influenza A PB2 subunit NLS
(PDB:2JDQ [23]). However, an oriented peptide
library screening study demonstrated that Tyr has a
higher specificity for the P4 binding position compared
with these hydrophobic residues (Val, Leu and Ile)
[24]. Interestingly, the contribution to NLS binding
energy of the residue at P4 is similar whether it is
occupied by a positively charged Arg side chain or a
hydrophobic Val side chain [12]. This is despite the
extra hydrogen bonding formed by Arg at P4 and a
possible helix dipole interaction with H3 helices of
ARM repeats 1 and 2. Although the hydrogen bonds
formed by Tyr205 may suggest that a Tyr at P4 is
more favourable than a Phe or other bulky residue, it
is likely to have a similar energetic contribution to
peptide binding as Val or Arg.
The CLIC4 NLS contains an Arg residue (Arg206)
in binding position P5 which adopts a common and
extended rotamer conformation. The P5 position is
defined by importin-a residues Trp142 and Trp184,
which both adopt a rare rotamer conformation, facili-
tating the formation of this hydrophobic pocket. The
site is normally occupied by a NLS Lys residue in
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1669
importin-a:NLS structures, where the aliphatic portion
of the side chain fits into the hydrophobic alcove and
allows the charged head group access to the solvent on
the other side. The extended side chain of Arg206
allows for the same favourable interactions as a Lys
residue, including a hydrogen bond to
impa
Gln181 and
the formation of a network of solvent interactions.
The positively charged guanidinium group of Arg206
is also compensated by the nearby acidic residue
impa
Glu180, with an interaction distance (N
f1
) O
e1
)of
4.9 A
˚
.
Conservation of the CLIC NLS
The CLIC4 NLS 199KVVAKKYR206 is highly con-
served in vertebrate species and across CLICs 1–6,
apart from CLIC3 (Fig. 2E). The N-terminal KVV
motif in the CLIC NLS is mostly conserved; however,
its importance to aiding the recognition by importin-a
is likely to be small, as we observed no interaction
between these residues and importin-a in our structure.
This is particularly noteworthy considering the pres-
ence of Lys199 in this flanking region. We therefore
suggest that the CLIC4 NLS motif, currently reported
as 199KVVAKKYR206, could be abbreviated to
include only those residues that bind directly to impor-
tin-a, 202AKKYRN207.
The KKYR motif, which occupies the core binding
region P2–P5, controls importin-a recognition of the
NLS and is conserved in all human CLICs, except
CLIC3. The significance of the motif is clear, with
92% of the hydrogen bonds formed between importin-
a and the CLIC4 NLS occurring in this region. It is
interesting to note that the KKYR motif is ignored by
NLS predicting tools, such as PredictNLS [25] and
NLStradamus [26]. This suggests that the criteria for
predicting an NLS need to be expanded to account for
residue variability, particularly in the P4 binding posi-
tion. Future studies may seek to elucidate the impor-
tance of the KKYR motif in governing the nuclear
import for other CLICs, and we note that CLIC1 has
been shown to localize to the nucleus with an unex-
plained import pathway [27].
Import pathways
The importin-a:CLIC4 NLS complex presented here
supports the hypothesis that CLIC4 can enter the
nucleus via an importin-a-mediated nuclear import
pathway. However, the recognition of full-length
CLIC4 by importin-a is strictly dependent on its abil-
ity to undergo a change in conformation that exposes
a linear NLS ready for binding.
Experiments showing that CLIC4 translocates to the
nucleus are not clear on the exact mechanism by which
translocation occurs. Although mutagenesis of the
putative CLIC4 NLS implies an interaction with im-
portin-a, immunoprecipitation experiments have shown
that CLIC4 also associates with nuclear transport fac-
tor-2 and Ran [2]. The nuclear transport factor-2 path-
way is normally used for the nuclear import of
RanGDP [28], but nuclear transport factor-2 can also
interact with cargo complexes [29]. The nuclear trans-
port of CLIC4 may therefore reveal new general para-
digms for nucleocytoplasmic transport processes.
Triggers for structural change
Typically, NLSs are located in solvent-exposed regions
of unstructured domains, such as flexible ends or loop
regions. The CLIC4 NLS motif, as seen in the soluble
CLIC4 crystal structure, is unusual for an NLS as it is
neither unfolded nor situated at domain termini
(Fig. 2D). As we have shown that the CLIC4 NLS
binds to importin-a in an extended conformation, this
confirms a previous hypothesis [4] that structural
changes in CLIC4 are required to expose the NLS
region. The NLS in CLIC4 forms a structurally stable
helical conformation at the C-terminus of helix 6. The
key KKYR binding motif in the NLS is positioned so
that Arg206 terminates the helix and the two Lys resi-
dues are solvent exposed. A number of side-chain
hydrogen-bonding interactions involve helix 9 and the
C-terminal tail. Given the metamorphic nature of the
CLIC module, as demonstrated by the structural tran-
sition of monomeric CLIC1 to a dimeric form [30], we
anticipate that there is a structural rearrangement that
can unfold the CLIC4 NLS.
A recent study has observed a post-translational
modification to CLIC4 which could represent the
required trigger for structural change [3]. This modifi-
cation is the S-nitrosylation of a Cys residue located
between helix 8 and helix 9, near the CLIC4 NLS site.
It has been shown that Cys234 is S-nitrosylated when
exposed to a nitric oxide agent, S-nitrosocysteine [31],
despite the thiol side chain being buried in an interdo-
main interface. Consistent with this hypothesis is a
correlation between nitric oxide donors and the rate of
nuclear translocation of CLIC4 [32], as well as an
increased association between CLIC4 and importin-a
after nitrosylation [3].
Using the high-resolution crystal structure of human
thioredoxin (PDB:2HXK), which contains a buried
S-nitrosocysteine (Cys62), the coordinates of the SNO
group were transferred to Cys234 in CLIC4 (refer to
Fig. S1). Both Cys residues adopt the most common
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1670 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
rotamer assisting nitric oxide transfer. In this model, a
strong steric clash is expected between the S-nitrosylat-
ed Cys234 and His196 in helix 6, which can be allevi-
ated by a shift of at least 1 A
˚
in the interdomain
interface. Accordingly, experiments have shown that
the S-nitrosylation of CLIC4 induces a structural
change [3]. We note that an S-nitrosylation trigger
leading to conformational change has been observed
previously in blackfin tuna myoglobin, where the
S-nitrosylation of a Cys residue causes a structural
domain shift [33]. Whether the S-nitrosylation of
Cys234 results in a similar structural rearrangement of
CLIC4, and provides a trigger for NLS conforma-
tional change, requires further investigation.
The possibility of trigger events to expose NLSs
offers an interesting criterion for cargo selection by
nuclear import receptors. Typically, NLSs are identifi-
able as a result of their high content of basic residues,
which allows them to bind tightly to an acidic binding
site in importin-a. However, it is clear that the prerequi-
site for NLS binding, based on sequence alone, is very
limited because a core region of just four residues is nec-
essary for importin-a peptide recognition. Furthermore,
there is a certain degree of tolerance to nonbasic resi-
dues within the importin-a binding region, particularly
at P4, where we find the Tyr residue in CLIC4 NLS. In
order to select specific cargo proteins destined for the
nucleus, it is likely that importin-a screens additional
criteria, such as the flexibility of the NLS and its solvent
exposure within a folded protein. This form of regula-
tion is governed more by the cargo protein itself, rather
than importin-a, and, as such, is a specific control
mechanism that has not been thoroughly considered
among the many levels of import regulation.
For a protein such as CLIC4, in which nuclear trans-
location is linked to cellular events, it is understandable
that a higher level of regulation is needed relative to
proteins that regularly shuttle across the nuclear mem-
brane. The folded state of the CLIC4 NLS may provide
the necessary barrier to binding importin-a, so that an
environmental trigger is required before the NLS
unfolds into an extended and exposed sequence that is
recognized by the import receptor.
Materials and methods
Protein expression and purification
The generation of the mouse importin-a (70–529) expression
construct has been described previously [13]. The construct
consists of residues 70–529 from Mus musculus importin-a
isoform a2, inserted into the pET-30a N-terminal His-tag
expression vector (Novagen, Madison, WI, USA). The plas-
mid was transformed into BL21 (DE3) cells and grown at
37 °C overnight in Luria–Bertani medium containing
30 lgÆmL
)1
of kanamycin. Flasks of 400 mL of Luria–
Bertani medium (30 lgÆmL
)1
kanamycin) were then inocu-
lated with 10 mL of the overnight culture. The cells were
grown at 37 °C until an absorbance at 600 nm of 1.0 cm
)1
was reached, at which point the culture was induced with
1mm isopropyl thio-b-d-galactoside. The temperature was
then lowered to 30 °C and expression continued for approxi-
mately 5 h. Cells were harvested by centrifugation at
11 000 g for 8 min and resuspended in 35 mL NaCl ⁄ P
i
with
2mm dithiothreitol. At this stage, protease inhibitors were
also added (Complete Protease Inhibitor Cocktail; Roche
Applied Science, Penzberg, Germany). The cells were lysed
using a French pressure cell and the resulting cell debris was
centrifuged at 39 000 g for 45 min at 4 °C.
The supernatant was added to a 3-mL solution of either
nickel nitrilotriacetic acid agarose resin (Qiagen, Valencia,
CA, USA) or Profinity immobilized metal affinity chro-
matography (IMAC) resin (Bio-Rad, Hercules, CA, USA),
and was mixed under gentle rotation for 1 h. The resin
was then loaded into an Econopak gravity flow column
(Bio-Rad) and unbound protein was eluted and discarded.
The resin was washed with 200 mL of 20 mm Hepes, pH 7.0,
500 mm NaCl, 1 mm MgCl
2
, 0.1 mm dithiothreitol and up
to 20 mm imidazole. Bound protein was then step eluted
with 10 mL of a buffer containing 20 mm Hepes,
pH 7.0, 500 mm NaCl, 0.1 mm dithiothreitol and 150 mm
imidazole.
In addition to step elution of importin-a (70–529) after
binding to IMAC resin, an imidazole gradient was
employed in some instances for greater control over protein
elution and higher purity. A HisTrapÔ FF column (GE
Healthcare, Chalfont St. Giles, Buckinghamshire, UK) with
a volume of 1 mL and prepacked with Ni-SepharoseÔ 6
Fast Flow resin was first pre-equilibrated with a buffer con-
sisting of 20 mm Tris, pH 8.0, and 500 mm NaCl before
binding, and then washed with 20 mm Tris, pH 8.0,
500 mm NaCl, 1 mm MgCl
2
and 20 mm imidazole. The
protein was eluted over an imidazole gradient from 20 to
250 mm at a flow rate of 0.5 mLÆmin
)1
over a volume of
40 mL.
Fractions with the majority of protein, as assessed by
UV absorption at 280 nm (A
280
), were then pooled and
dialysed into 20 mm Tris, pH 8.0, 100 mm NaCl and 2 mm
dithiothreitol. The purification protocol described here
follows the methods reported in the literature [13,22].
For ion exchange chromatography, protein was loaded
onto a HiloadÔ 26 ⁄ 10 Q-SepharoseÔ anion exchange col-
umn (GE Healthcare) immediately after imidazole gradient
IMAC purification. Importin-a was allowed to bind to the
N
+
(CH
3
)
3
charged groups of the Q-SepharoseÔ HP resin
in Buffer A (50 mm NaCl, 20 mm Tris, pH 8.0, 2 mm dith-
iothreitol), before being eluted over a salt gradient from
50 mm to 1 m at a flow rate of 4 mLÆmin
)1
(total volume
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1671
of 400 mL). Eluted fractions with the highest concentration
of protein were pooled and concentrated using an Amicon
YM-10 Centriprep centrifugal concentrator (Amicon,
Billerica, MA, USA) in a Sorvall SH-3000 swinging bucket
rotor (Sorvall, Waltham, MA, USA) at 1800 g. The con-
centrated protein from ion exchange chromatography was
loaded on a Superdex 200 10 ⁄ 300 GL (GE Healthcare)
analytical size exclusion column and eluted with 20 mm
Tris, pH 8.0, 100 mm NaCl and 2 mm dithiothreitol.
Fractions corresponding to importin-a (70–529) were pooled
and concentrated to approximately 18 mgÆmL
)1
, flash
frozen in liquid nitrogen and stored at )80 °C until use.
NLS peptide synthesis
The CLIC4 NLS peptide was synthesized (Sigma-Genosys,
Sydney, Australia) and used in the crystallization of impor-
tin-a:CLIC4 peptide complexes. Although the proposed
CLIC4 NLS motif includes residues 199–206 (199KVVAK-
KYR206) [2], the CLIC4 NLS synthetic peptide used for
crystallization includes one-residue extensions at both ter-
mini (NH
2
-198VKVVAKKYRN207-COOH). The purity
was assessed by analytical RP-HPLC to be better than
95% and the peptide has a molecular mass of 1204.5 Da
and pI of 10.9. Lyophilized peptides were first dissolved in
water or a suitable assay buffer before use. The stock con-
centration of each peptide sample varied between 1 and
3mgÆmL
)1
for different crystallization trials.
Crystallization and data collection
The importin-a (70–529):CLIC4 NLS crystals were obtained
by co-crystallization using hanging drop vapour diffusion.
The reservoir consisted of 500 lL of 0.7 m sodium citrate,
10 mm dithiothreitol and 70 mm Hepes at pH 7.4. Drops
consisted of 0.75 l L of importin-a (70-529) at 18 mgÆmL
)1
,
0.75 lL of CLIC4 NLS peptide at 3 mgÆmL
)1
and 1 lLof
reservoir. These conditions resulted in a 7.6 times molar
excess of the CLIC4 NLS peptide over importin-a (70–529).
There was no incubation of peptide and protein other than
the immediate mixing in the crystallization drop. Crystals
grew to maturity in approximately 2 weeks at 20 °C. Crys-
tals were gradually transferred into a cryoprotectant solu-
tion consisting of the reservoir solution supplemented with
20% glycerol. The importin-a:CLIC4 NLS crystal was flash
cooled in liquid nitrogen and stored in a cryogenic dry ship-
ping dewar. It was then mounted at 100 K in a nitrogen
cryostream on beamline PX-1 (3BM-1B) at the Australian
Synchrotron [34], and data collection was performed using
an ADSC Quantum 210 (Q210) detector.
A number of initial images were screened at orthogonal
u angles to gauge crystal quality before autoindexing and
determining the optimum rotation range using strategy
from within mosflm [35]. Crystals were found to have the
symmetry of the orthorhombic space group P2
1
2
1
2
1
with
unit cell dimensions of approximately a =79A
˚
, b =90A
˚
and c = 100 A
˚
(see Table 1). The final diffraction dataset
was obtained over 180 images with a 1° oscillation angle
(u) and 5-s exposure time per image. The crystal to detector
distance was set at 220 mm and the beam width was set to
200 lm · 200 lm.
Structure determination and refinement
After datasets had been initially processed and integrated in
mosflm, the reflection data file was then passed through a
suite of programs in the ccp4 crystallography package [36].
Scaling of intensities and inspection of data quality were
performed in scala. In particular, the resolution cut-off
was determined by analysing the signal to noise parameter,
<I ⁄ r(I)>, as it varies against resolution. The resolution
corresponding to an <I ⁄ r(I)> value of 2.0 was used as an
approximate limit to the high-resolution bin, as described
in the literature [37]. The fraction of data used for the R
free
calculation was set to 5%.
Phases were determined using molecular replacement with
the maximum likelihood search function in phaser [38]. A
model of importin-a was built using the full-length importin-
a structure (PDB:1IAL [11]) as a search probe, with the sol-
vent and NLS peptide model removed, and atomic B factors
reset to an appropriate value as suggested from the Wilson
plot. An initial solvent model was built using arp ⁄ warp [39],
which was then manually checked and validated using the
visualization program coot [40]. Water molecules that were
not well ordered, as well as water molecules placed in or near
the importin-a binding site, were removed.
Iterative refinement of the model was performed with
multiple passes of refmac5 maximum likelihood refinement
[41] and coot manual refinement. Electron density for a
bound peptide was generally much clearer in the major bin-
ding site compared with the minor binding site. Data reduc-
tion and refinement statistics are summarized in Table 1.
Electron density analysis
Fourier analysis was used to compare the importin-
a:CLIC4 NLS peptide complex (observed structure factors
F
pep
o
) with an apo structure of importin-a (70–529)
(observed structure factors F
apo
o
). This structure is described
in ref. [42]. The structure factors from each dataset were
scaled using the ccp4 program scaleit [36]. This calculates
a scaling R factor of 16.3%, supporting isomorphism
between the two crystals. We refer to the Fourier difference
of these structure factors as a data–data difference Fourier,
F
pep
o
ÀF
apo
o
, which is calculated as:
F ¼ F
pep
o
À F
apo
o
ÀÁ
Á e
ia
apo
c
where the phases (a
apo
c
) are obtained from the apo impor-
tin-a structure.
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1672 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
B factor analysis
In order to analyse changes in the conformational dynamics
of importin-a residues, a comparison was made between
the atomic B factors of apo importin-a and the importin-
a:CLIC4 NLS complex. These two crystals are isomor-
phous: they were both solved in the space group P2
1
2
1
2
1
,
with similar lattice dimensions (CLIC4 NLS crystal:
a = 78.6 A
˚
, b = 89.6 A
˚
, c = 100.1 A
˚
; apo crystal:
a = 79.0 A
˚
, b = 89.8 A
˚
, c = 100.3 A
˚
), and the scaling R
factor is relatively low (16.3%). We have calculated a nor-
malized B factor score using a method described previously
[43]. The B factors of atoms in structure j are first normal-
ized according to the following equation:
B
z
¼
B
i
À l
j
r
j
where B
z
is the normalized B factor, B
i
is the B factor of
atom i, l
j
is the mean B factor (excluding water molecules
and peptide atoms) and r
j
is the standard deviation of B
factors. The normalized B factors have a zero mean and unit
variance. All atoms that satisfy B
z
‡ 4 are treated as outliers
and discarded. After discounting outliers, B
z
values were
recalculated. The normalized B factors from apo importin-a
were then subtracted from those in the importin-a:CLIC4
NLS structure, and the difference was again normalized
using the equation above. The final values are referred to as
the B
Àapo
z
score, and were calculated for all nonhydrogen
atoms in the importin-a:CLIC4 NLS structure.
Acknowledgements
Data contributing to this research were obtained on
the PX-1 beamline at the Australian Synchrotron, Vic.,
Australia.
References
1 Marfori M, Mynott A, Ellis JJ, Mehdi AM, Saunders
NFW, Curmi PM, Forwood JK, Bode
´
n M & Kobe
B(2010) Molecular basis for specificity of nuclear
import and prediction of nuclear localization. Biochim
Biophys Acta – Mol Cell Res doi:10.1016/j.bbamcr.
2010.10.013.
2 Suh KS, Mutoh M, Nagashima K, Fernandez-Salas E,
Edwards LE, Hayes DD, Crutchley JM, Marin KG,
Dumont RA, Levy JM et al. (2004) The organellular
chloride channel protein CLIC4 ⁄ mtCLIC translocates
to the nucleus in response to cellular stress and acceler-
ates apoptosis. J Biol Chem 279, 4632–4641.
3 Malik M, Shukla A, Amin P, Niedelman W, Lee J, Jiv-
iden K, Phang JM, Ding J, Suh KS, Curmi PM et al.
(2010) S-nitrosylation regulates nuclear translocation of
chloride intracellular channel protein CLIC4. J Biol
Chem 285, 23818–23828.
4 Littler DR, Assaad NN, Harrop SJ, Brown LJ, Pank-
hurst GJ, Luciani P, Aguilar MI, Mazzanti M, Berry-
man MA, Breit SN et al. (2005) Crystal structure of the
soluble form of the redox-regulated chloride ion channel
protein CLIC4. FEBS J 272, 4996–5007.
5 Singh H & Ashley RH (2007) CLIC4 (p64H1) and its
putative transmembrane domain form poorly selective,
redox-regulated ion channels. Mol Membr Biol 24, 41–52.
6 Tung JJ, Hobert O, Berryman M & Kitajewski J (2009)
Chloride intracellular channel 4 is involved in endothe-
lial proliferation and morphogenesis in vitro. Angiogen-
esis 27, 27.
7 Suh KS, Mutoh M, Mutoh T, Li L, Ryscavage A,
Crutchley JM, Dumont RA, Cheng C & Yuspa SH
(2007) CLIC4 mediates and is required for Ca
2+
-
induced keratinocyte differentiation. J Cell Sci 120,
2631–2640.
8 Suh KS, Crutchley JM, Koochek A, Ryscavage A,
Bhat K, Tanaka T, Oshima A, Fitzgerald P & Yuspa
SH (2007) Reciprocal modifications of CLIC4 in tumor
epithelium and stroma mark malignant progression of
multiple human cancers. Clin Cancer Res 13, 121–131.
9 Suh KS, Mutoh M, Gerdes M & Yuspa SH (2005)
CLIC4, an intracellular chloride channel protein, is a
novel molecular target for cancer therapy. J Invest
Dermatol Symp Proc 10, 105–109.
10 Fanara P, Hodel MR, Corbett AH & Hodel AE (2000)
Quantitative analysis of nuclear localization signal
(NLS)–importin alpha interaction through fluorescence
depolarization. Evidence for auto-inhibitory regulation
of NLS binding. J Biol Chem 275, 21218–21223.
11 Kobe B (1999) Autoinhibition by an internal nuclear
localization signal revealed by the crystal structure of
mammalian importin alpha. Nat Struct Biol 6, 388–397.
12 Hodel MR, Corbett AH & Hodel AE (2001) Dissection
of a nuclear localization signal. J Biol Chem 276, 1317–
1325.
13 Fontes MR, Teh T & Kobe B (2000) Structural basis of
recognition of monopartite and bipartite nuclear locali-
zation sequences by mammalian importin-alpha. J Mol
Biol 297, 1183–1194.
14 Cutress ML, Whitaker HC, Mills IG, Stewart M &
Neal DE (2008) Structural basis for the nuclear import
of the human androgen receptor. J Cell Sci 121,
957–968.
15 Conti E, Uy M, Leighton L, Blobel G & Kuriyan J
(1998) Crystallographic analysis of the recognition of a
nuclear localization signal by the nuclear import factor
karyopherin alpha. Cell 94, 193–204.
16 Kalderon D, Roberts BL, Richardson WD & Smith AE
(1984) A short amino acid sequence able to specify
nuclear location. Cell 39, 499–509.
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1673
17 Colledge WH, Richardson WD, Edge MD & Smith AE
(1986) Extensive mutagenesis of the nuclear location
signal of simian virus 40 large-T antigen. Mol Cell Biol
6, 4136–4139.
18 Davis IW, Leaver-Fay A, Chen VB, Block JN,
Kapral GJ, Wang X, Murray LW, Arendall WB III,
Snoeyink J, Richardson JS et al. (2007) MolProbity:
all-atom contacts and structure validation for
proteins and nucleic acids. Nucleic Acids Res 35,
W375–W383.
19 Word JM, Lovell SC, LaBean TH, Taylor HC, Zalis
ME, Presley BK, Richardson JS & Richardson DC
(1999) Visualizing and quantifying molecular goodness-
of-fit: small-probe contact dots with explicit hydrogen
atoms. J Mol Biol 285, 1711–1733.
20 Chen MH, Ben-Efraim I, Mitrousis G, Walker-Kopp
N, Sims PJ & Cingolani G (2005) Phospholipid scramb-
lase 1 contains a nonclassical nuclear localization signal
with unique binding site in importin alpha. J Biol Chem
280, 10599–10606.
21 Conti E & Kuriyan J (2000) Crystallographic analysis
of the specific yet versatile recognition of distinct
nuclear localization signals by karyopherin alpha.
Structure 8, 329–338.
22 Fontes MR, Teh T, Jans D, Brinkworth RI &
Kobe B (2003) Structural basis for the specificity
of bipartite nuclear localization sequence binding
by importin-alpha. J Biol Chem 278, 27981–
27987.
23 Tarendeau F, Boudet J, Guilligay D, Mas PJ, Bougault
CM, Boulo S, Baudin F, Ruigrok RW, Daigle N, Ellen-
berg J et al. (2007) Structure and nuclear import func-
tion of the C-terminal domain of influenza virus
polymerase PB2 subunit. Nat Struct Mol Biol 14,
229–233.
24 Yang SN, Takeda AA, Fontes MR, Harris JM, Jans
DA & Kobe B (2010) Probing the specificity of binding
to the major nuclear localization sequence-binding site
of importin-alpha using oriented peptide library screen-
ing. J Biol Chem 285, 19935–19946.
25 Cokol M, Nair R & Rost B (2000) Finding nuclear
localization signals. EMBO Rep 1, 411–415.
26 Nguyen Ba AN, Pogoutse A, Provart N & Moses AM
(2009) NLStradamus: a simple Hidden Markov Model
for nuclear localization signal prediction. BMC Bioin-
formatics 10, 202.
27 Valenzuela SM, Martin DK, Por SB, Robbins JM,
Warton K, Bootcov MR, Schofield PR, Campbell TJ &
Breit SN (1997) Molecular cloning and expression of a
chloride ion channel of cell nuclei. J Biol Chem 272,
12575–12582.
28 Hu W & Jans DA (1999) Efficiency of importin alpha ⁄ -
beta-mediated nuclear localization sequence recognition
and nuclear import. Differential role of NTF2. J Biol
Chem 274, 15820–15827.
29 Van Impe K, Hubert T, De Corte V, Vanloo B, Bou-
cherie C, Vandekerckhove J & Gettemans J (2008) A
new role for nuclear transport factor 2 and Ran:
nuclear import of CapG. Traffic 9, 695–707.
30 Littler DR, Harrop SJ, Fairlie WD, Brown LJ,
Pankhurst GJ, Pankhurst S, DeMaere MZ, Campbell
TJ, Bauskin AR, Tonini R et al. (2004) The intracellu-
lar chloride ion channel protein CLIC1 undergoes a
redox-controlled structural transition. J Biol Chem 279,
9298–9305.
31 Greco TM, Hodara R, Parastatidis I, Heijnen HF, Den-
nehy MK, Liebler DC & Ischiropoulos H (2006) Identi-
fication of S-nitrosylation motifs by site-specific
mapping of the S-nitrosocysteine proteome in human
vascular smooth muscle cells. Proc Natl Acad Sci USA
103, 7420–7425.
32 Suh KS, Malik M, Shukla A & Yuspa SH (2007)
CLIC4, skin homeostasis and cutaneous cancer: surpris-
ing connections. Mol Carcinog 46, 599–604.
33 Schreiter ER, Rodriguez MM, Weichsel A, Montfort
WR & Bonaventura J (2007) S-nitrosylation-induced
conformational change in blackfin tuna myoglobin.
J Biol Chem 282, 19773–19780.
34 McPhillips TM, McPhillips SE, Chiu HJ, Cohen AE,
Deacon AM, Ellis PJ, Garman E, Gonzalez A, Sauter
NK, Phizackerley RP et al. (2002) Blu-Ice and the Dis-
tributed Control System: software for data acquisition
and instrument control at macromolecular crystallogra-
phy beamlines. J Synchrotron Radiat 9, 401–406.
35 Leslie AG (2006) The integration of macromolecular
diffraction data. Acta Crystallogr D: Biol Crystallogr
62, 48–57.
36 Collaborative Computational Project Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr D: Biol Crystallogr 50, 760–763.
37 Evans P (2006) Scaling and assessment of data quality.
Acta Crystallogr D: Biol Crystallogr 62, 72–82.
38 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn
MD, Storoni LC & Read RJ (2007) Phaser crystallo-
graphic software. J Appl Crystallogr 40, 658–674.
39 Langer G, Cohen SX, Lamzin VS & Perrakis A (2008)
Automated macromolecular model building for X-ray
crystallography using ARP ⁄ wARP version 7. Nat
Protoc 3 , 1171–1179.
40 Emsley P & Cowtan K (2004) Coot: model-building
tools for molecular graphics. Acta Crystallogr D: Biol
Crystallogr 60, 2126–2132.
41 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by the maxi-
mum-likelihood method. Acta Crystallogr D: Biol Crys-
tallogr 53, 240–255.
42 Mynott AV (2009) Structural investigations of CLIC
proteins and importin-alpha recognition of nuclear locali-
sation signals. Thesis, University of New South Wales,
Sydney, Australia.
Crystal structure of importin-a:CLIC4 NLS peptide A. V. Mynott et al.
1674 FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS
43 Smith DK, Radivojac P, Obradovic Z, Dunker AK &
Zhu G (2003) Improved amino acid flexibility parame-
ters. Protein Sci 12, 1060–1072.
44 Thompson JD, Higgins DG & Gibson TJ (1994)
CLUSTAL W: improving the sensitivity of progressive
multiple sequence alignment through sequence weight-
ing, position-specific gap penalties and weight matrix
choice. Nucleic Acids Res 22 , 4673–4680.
45 Gouet P, Courcelle E, Stuart DI & Metoz F (1999) ES-
Pript: analysis of multiple sequence alignments in Post-
Script. Bioinformatics 15, 305–308.
Supporting Information
The following supplementary material is available:
Fig. S1. Analysis of the importin-a:CLIC4 NLS com-
plex.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
A. V. Mynott et al. Crystal structure of importin-a:CLIC4 NLS peptide
FEBS Journal 278 (2011) 1662–1675 ª 2011 The Authors Journal compilation ª 2011 FEBS 1675