Báo cáo khoa học: Two separate regions essential for nuclear import of the hnRNP D nucleocytoplasmic shuttling sequence ppt

Two separate regions essential for nuclear import of the
hnRNP D nucleocytoplasmic shuttling sequence
Maiko Suzuki*, Megumi Iijima*, Akira Nishimura*, Yusuke Tomozoe, Daisuke Kamei
and Michiyuki Yamada
Graduate School of Integrated Science, Yokohama City University, Yokohama, Japan
In eukaryotic cells, molecules all move into and out of
the nucleus through nuclear pore complexes (NPCs)
which span the nuclear envelope. Small molecules
diffuse passively through the NPCs, while molecules of
more than about 60 kDa are transported by an energy-
dependent process. Most proteins are transported into
and out of the nucleus by nuclear transport receptors
and directionality is determined by high and low Ran-
GTP concentrations in the nucleus and cytoplasm,
respectively, generated by a RanGTPase system [1–5].
Many of nuclear proteins contain classical nuclear
localization sequences (NLS) consisting of one or two
clusters of basic amino acids termed basic type mono-
partite or bipartite NLS, respectively. They are impor-
ted into the nucleus by the nuclear import receptor
importin b with or without an adaptor importin a.
Other groups of nuclear RNA binding proteins, such as
hnRNP A1, SR proteins and HuR, are imported into
the nucleus by an mRNA synthesis-dependent process
and they shuttle continuously between the nucleus and
the cytoplasm [6–10]. Their NLS is bound by the nuc-
lear transport receptor transportin (Trn), but not recog-
nized by importin a ⁄ b [11–13]. These NLSs also serve
as a nuclear export sequence (NES) [2]. hnRNP A1 has
the best characterized nucleocytoplasmic shuttling
sequence M9, which is a 38 amino acid sequence in the

C-terminal domain [11,14,15]. M9 mutational analysis
has provided information on a consensus Trn)1 inter-
action motif [16]. There are various Trn-1 binding pro-
teins, such as TAP, poly(A)-binding protein II, and
Keywords
AUF1; hnRNP D; nucleocytoplasimic
shuttling sequence; nuclear transport;
transportin
Correspondence
M. Yamada, Graduate School of Integrated
Science, Yokohama City University, 22–2
Seto, Kanazawa-ku, Yokohama 236–0027,
Japan
Fax: +81 45 787 2413
Tel: +81 45 787 2214
E-mail:
*These authors contributed equally to this
work.
(Received 16 April 2005, revised 6 June
2005, accepted 14 June 2005)
doi:10.1111/j.1742-4658.2005.04820.x
Heterogeneous nuclear ribonucleoprotein (hnRNP) D ⁄ AUF1 functions in
mRNA genesis in the nucleus and modulates mRNA decay in the cyto-
plasm. Although it is primarily nuclear, it shuttles between the nucleus and
cytoplasm. We studied the nuclear import and export of the last exon-enco-
ding sequence common to all its isoforms by its expression as a green fluor-
escent protein-fusion protein in HeLa cells and by heterokaryon assay.
The C-terminal 19-residue sequence (SGYGKVSRRGGHQNSYKPY) was
identified as an hnRNP D nucleocytoplasmic shuttling sequence (DNS).
In vitro nuclear transport using permeabilized cells indicated that nuclear

import of DNS is mediated by transportin-1 (Trn-1). DNS accumulation in
the nucleus was dependent on Trn-1, Ran, and energy in multiple rounds
of nuclear transport. Use of DNS with deletions, alanine scanning muta-
genesis and point mutations revealed that two separate regions (the N-ter-
minal seven residues and the C-terminal two residues) are crucial for
in vivo and in vitro transport as well as for interaction with Trn-1. The
N- and C-terminal motifs are conserved in the shuttling sequences of
hnRNP A1 and JKTBP.
Abbreviations
DAPI, 4¢,6-diamino-2-phenylindole; DNS, hnRNP D ⁄ AUF1 nucleocytoplasmic shuttling sequence; EGFP, enhanced green fluorescent protein;
GST, glutathione S-transferase; mt, mutant type; NES, nuclear export sequence; NLS, nuclear localization sequence; NPC, nuclear pore
complexes; PAD, peptidylarginine deiminase; RU, resonance unit; SPR, surface plasmon resonance; Trn-1, transportin 1; wt, wild type.
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3975
HuR, but no obvious consensus Trn-1 binding sequence
has been found in these naturally occurring proteins
[12,17–19].
hnRNP D ⁄ AUF1 consists of two RNA binding
domains (RBDs) and has a high content of glycine in
the C-domain, like hnRNP A1 [20,21]. Four isoforms,
D01 ⁄ p37, D02 ⁄ p40, D1 ⁄ p42 and D2 ⁄ p45, are formed
by alternative splicing and are found in many tissues
and various types of cultured cells [22–24]. They func-
tion in trans-acting transcriptional factors, alternative
splicing factors in the nucleus and in modulation of
AU-rich element-directed mRNA decay in the cyto-
plasm [21,25–28]. They are found mainly in the nuc-
leus, but rapidly shuttle between the nucleus and the
cytoplasm [22,29–31]. Besides, their subcellular distri-
bution in cells changes in response to environmental
stimuli such as temperature shift, cell differentiation

and mRNA synthesis inhibition [26,30–32]. It has been
shown by protein–protein blotting that hnRNP D2
binds to Trn-1 at the C-terminal 112 amino acid
sequence [13]. Recently, the C-terminal 50 amino acid
and 35 amino acid sequences of D01 and D02 were
found as an NLS [29,30]. However, there is no clear
evidence for the nucleocytoplasmic shuttling activity
of NLS and the involvement of Trn-1 in the nuclear
import. The C-terminal 21 amino acid sequence enco-
ded by the hnRNP D last exon 8 is noted to be
homologous with the 25 amino acid shuttling sequence
in the JKTBP C-terminal tail [33]. In this study, we
attempted to determine whether the nuclear import
and export sequences are located in the same region.
We identified the hnRNP D nucleocytoplasmic shut-
tling sequence (DNS) as a 19 amino acid sequence and
found that the N- and C-terminal portions of DNS are
important for the nuclear import mediated by Trn-1.
Results
Determination of hnRNP D NLS
The exon 8 of hnRNP D encodes the 21 amino acids
sequence common to all the isoforms D01, D02, D1,
and D2 (Fig. 1A) [23]. To examine whether the exon-8
encoding sequence has NLS activity, we prepared plas-
mid constructs encoding the D02 C-terminal 25 amino
acids (282–306) and mutants of this sequence with
increasing N- and C-terminal deletions as fusion pro-
teins with the C-end of a composite EGFP-GST-PAD
protein ( 69 kDa) (Fig. 1B). These plasmids were
used to transfect HeLa cells and after their expression,

their subcellular localization in the cells were examined
by fluorescence microscopy (Fig. 1C). Fluorescence
micrographs of the cells revealed that the empty
vector-encoding composite GFP-GST-PAD protein,
used as a control, was exclusively present in the cyto-
plasm (panel a), indicating that it is larger than the
passive diffusion protein. The N-terminal deletion D02
mutants 282–306, 288–306, and 292–306 were present
only in the nucleus (Fig. 1C, panels b, c and d). The
shorter mutants, 293–306, 294–306, and 295–306, were
found mainly in the nucleus but also slightly in the
cytoplasm (Fig. 1C, panels e, f and g). However, a one
residue shorter 11-residue mutant 296–306 and an
eight-residue mutant 299–306 were found exclusively in
the cytoplasm like the control with an empty vector
(Fig. 1C, panels h and i and a). The nuclear localiza-
tions of C-terminal deletion mutants were also studied
in the same way (Fig. 1D). The C-3 and -6 amino
acids deletion mutants 288–303 and 288–300 were
found only in the cytoplasm like the control (Fig. 1D,
panels c, d and a), while the 19-residue mutant 288–
306 was found in the nucleus (panel b). Immunoblot-
ting of the cell lysates using anti-GFP confirmed the
expression of mutant fusion proteins of the expected
size of  70 kDa (data not shown), indicating that the
cytoplasmic fluorescent signal was not that of a degra-
ded protein. These results indicated that D02 NLS is
mapped to the C-terminal 19 amino acids (288–306)
encoded by exon 8.
Role of amino acid residues of a D02 NLS

in nuclear import
To determine the role of amino acids of the C-terminal
19 residue NLS in nuclear import, we performed alan-
ine scanning mutagenesis experiments. Mutants mt1-
mt5 were prepared using the construct encoding an
EGFP-GST-PAD-D02 NLS (288–306) fusion gene (wt)
as a template by sequential consecutive three amino
acid replacements by a cluster of three alanines
(Fig. 2A). These constructs were examined for nuclear
import in the same way as described above. As shown
in Fig. 2B, mt3 was located in the nucleus as the wt
(Fig. 2B, panels e and b), whereas mt1, mt2 and mt4
were mostly located in the nucleus, but also signifi-
cantly in the cytoplasm (Fig. 2B, panels c, d and f). In
contrast, the C-terminal three amino acid substitution
mutant mt5 was located exclusively in the cytoplasm,
likely the control with an empty vector (Fig. 2B,
panels a and g). This prompted us to test the two last
amino acid substitution mutants mt6–9 for nuclear
import (Fig. 2C). The mutants mt6 (P305A ⁄ Y306A),
mt7 (P305A), mt8 (Y306A) and mt9 (Y306D) were
located in the cytoplasm (Fig. 2C, panels b–e), while
the wt was imported into the nucleus (panel a). These
results indicated that both C-terminal residues PY are
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
3976 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
A
B
C
D

Fig. 1. The C-terminal location of an NLS in hnRNP D. (A) hnRNP D ⁄ AUF1 isoforms. Boxes 2 and 7 show alternative splicing exons 2 and 7 enco-
ding 19 and 49 amino acid residues, respectively. Box 8 denotes the exon 8-encoded sequence. The first and last amino acid residue numbers
are shown under the corners of boxes. (B) Plasmid constructs for nuclear transport of the D02 C-terminal sequence (282–306) and its N-terminal
and C-terminal deletion mutant sequences. 1, a control empty vector pEGFP-GST-PAD encoding a GFP-labeled composite protein (604 amino
acids); 2–11, plasmid constructs encoding D02 (282–306) and its N-terminal and C-terminal deletion mutant sequences, respectively, represen-
ted as a fusion protein linked to the C-end of the composite protein. (C) Subcellular localizations of D02 (282–306) and N-terminal deletion
mutants expressed as GFP-fusion proteins in HeLa cells. HeLa cells were transfected with the plasmids described above and incubated on
coverslips for their expression for 24 h and then were studied by fluorescence microscopy. Panels a–i, fluorescent signals of the cells expressing
GFP-labeled proteins shown on the top of each panels; panels j–r, nuclear DNA stained with DAPI of the cells in the same views as in panels a–i,
respectively. (D) Subcellular localizations of C-terminal deletion mutants described on the top of each panel were studied as described in (C). Pan-
els a–d, fluorescent signals of cells; panels e–h, nuclear DNA stained with DAPI of the cells in the same view as in panels a–d, respectively. Plus
and minus signs on the right of (B) show, respectively, positive and negative signals for nuclear import and nuclear export described below.
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3977
essential for the nuclear import. Results on mt9
suggested that phosphorylation of the tyrosine residue
is not related to the nuclear import. Then, the role
of these two residues PY in nuclear import of a
full-length D02 (1–306) was examined by mutation.
Full-length D02 and the mutants were expressed as
EGFP fusion proteins but not as EGFP-GST-PAD
fusion proteins, and were studied in the same way as
above (Fig. 3). Mutants D02 (1–306) P305A ⁄ Y306A
and Y306A showed cytoplasmic localization and their
signals appeared as numerous speckles around the
nuclear periphery (Fig. 3, panels c and d), while a
control of EGFP was seen throughout the cells and
wild type D02 was seen exclusively in the nucleus
(Fig. 3, panels a and b). Mutant D02 (D288-291), with
a four amino acid SGYG (288–291) deletion, was also

A
B
C
Fig. 2. Effects of amino acid-substitutions of an hnRNP D NLS on nuclear import. (A) A wild type (wt) NLS (D02 288–306) and various
mutant types (mt1–9) with replacements by Ala and Asp in the indicated sites were represented as a fusion protein linked to the C-terminal
end of a composite protein GFP-GST-PAD in the pEGFP-GST-PAD vector described in Fig. 1B. Dashed lines indicate the same amino acids
in the sequence as shown at the top. Plus and minus signs on the right show, respectively, positive and negative signals for nuclear import
described below. (B) Subcellular localizations of NLS Ala scan mutant fusion proteins (wt and mt1–5) in cells. HeLa cells were transfected
with the above plasmid constructs and grown for 24 h for expression. The cells were studied by fluorescence microscopy. Panels a–g, fluor-
escent signals of the cells expressing GFP-labeled proteins shown on the top of each panel; panels h–n, nuclear DNA stained with DAPI of
the cells in the same views as in panels a–g. (C) Subcellular localization of NLS C-terminal end mutants (mt6–9). Panels a–e, fluorescent
signals of the cells expressing GFP-labeled proteins shown on the top of each panel; panels f–j, nuclear DNA stained with DAPI of the cells
in the same views as in the panels a–e, respectively.
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
3978 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
located in the nucleus (panel e). Immunoblotting of
cell lysates confirmed the expression of intact mole-
cules (data not shown). These results taken together
indicated that the N-terminal seven residues and the
last two C-terminal residues PY are essential for nuc-
lear import of the D02 C-terminal segment (288–306).
Identification of an hnRNP D NLS as a
nucleocytoplasmic shuttling sequence
To investigate whether the above described NLS has
nucleocytoplasmic shuttling activity, we used hetero-
karyon assay. The constructs encoding NLS in pEG-
FP-GST were expressed as GFP-tagged proteins in
HeLa cells for 22 h to label the nucleus, and the cells
were then fused to nontransfected murine 3T3 cells in
the presence of a protein synthesis inhibitor, cyclohexi-

mide, and further incubated for 1 h to see whether
GFP- labeled protein migrated from the HeLa nucleus
to the 3T3 nucleus in the heterokaryons. JKTBP2
served as a control for a nuclear retention protein
remaining in the original HeLa nucleus (Fig. 4, panel
a) and a full length D02 was used as a control positive
for shuttling and was found in the nucleus of both
HeLa and 3T3 cells (Fig. 4, panel b; arrow shows the
position of the mouse nucleus) as expected [30,33]. Of
the three NLS segments, D02 (282–306) and D02
(288–306) became located in the nucleus of both HeLa
and 3T3 cells (Fig. 4, panels c and d). However, the
four-residue shorter D02 (292–306) was found in the
HeLa but not the 3T3 nucleus (Fig. 4, panel e), indica-
ting that deletion of the four N-terminal residues
SGYG (288–291) of D02 (288–306) has a more deleteri-
ous effect on nuclear export than on import. It is note-
worthy that the substitution of GHQ (298–300) by
these alanines of mt3 did not affect shuttling activity so
much as nuclear import activity (Fig. 4, panel f). These
results indicated that the D02 C-terminal 19 residue
sequence (288–306) constitutes the hnRNP D nucleo-
cytoplasmic shuttling sequence. This was termed DNS.
Trn-1-dependent import of hnRNP D
We examined whether the nuclear import of DNS ⁄ D02
(288–306) was mediated by Trn-1. Nuclear import sub-
strates were prepared as GST-GFP fusion proteins and
tested for in vitro nuclear import activity using digito-
nin-permeabilized HeLa cells supplemented with either
reticulocyte lysates or a reconstituted mixture of Trn-1,

Ran mix (RanGDP, NTF2 and RanGAP) and an
energy-regenerating system. DNS 288–306 was effect-
ively imported into the nucleus at 30 °C but not at
4 °C in the presence of reticulocyte lysates during a 30-
min incubation period (Fig. 5A, left, panels a and b).
This import into the nucleus was inhibited almost com-
pletely by the addition of a 40-molar excess of hnRNP
A1 (1–320), but not significantly by a shortened form
(1–196) of hnRNP A1 ⁄ UP1 lacking the M9 domain
(right panels a–c). This M9-mediated inhibition sugges-
ted that DNS nuclear import is mediated by a nuclear
transport receptor of Trn, but not other importins. Use
of the reconstituted transport mixture instead of reticu-
locyte lysates indicated that DNS accumulation in the
nucleus was dependent on Ran mix, an energy-regener-
ating system and Trn-1 (Fig. 5B). Ran and energy were
required only when high substrate and low Trn-1 con-
centrations were used in the assay.
Next, to compare the NLS activity in vivo and
in vitro, the DNS N- and C-terminal deletion mutants
described above were tested for nuclear import activity
in the presence or absence of Trn-1 (Fig. 5C).
DNS ⁄ 288–306 was imported into the nucleus in a Trn-
Fig. 3. Importance of C-terminal residues of a full-length hnRNP D02 for subcellular distribution. Plasmid constructs carrying a full length
D02 wild type, D02 mutant (P305A ⁄ Y306A), D02 mutant (Y306A) and D02 deletion mutants (D288-291) gene linked in frame to the 3¢ end of
the EGFP gene in pEGFP vector were used to transfect HeLa cells. After 24 h expressions the cells were studied for subcellular localizations
by fluorescence microscopy. Panels a–e, fluorescent signals of cells expressing the GFP-labeled mutants shown on the top of each panel;
panels f–j, nuclear DNA stained with DAPI of the cells in the same views as in panels a–e.
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3979

1-dependent manner as efficiently as full D02 (1–306)
(Fig. 5C, panels a–d). Nuclear imports of the DNS N-
deletion mutants 292–306 and 293–306 were decreased
to a low level but significantly higher levels than the
levels in the absence of Trn-1 (Fig. 5C, panels e–h).
However, the shorter N-deletion mutant 296–306
showed no nuclear import activity (Fig. 5C, panels i
and j). The C-deletion mutant (288–303) lacking the
last three residues of DNS showed no nuclear import
activity (Fig. 5C, panels k and l). These results indica-
ted that the C-terminal 19-residue sequence of D is
necessary and sufficient for the in vitro nuclear import
which is mediated through a Trn-1 system. DNS NLS
(292–306) and (293–306) revealed that the nuclear
import was much lower in vitro than in vivo (compare
Fig. 5C, panels e–h, with Fig. 1C, panels d and e).
Direct interaction of DNS NLS with Trn-1
We analyzed the interaction of DNS N-and C-terminal
deletion NLS mutants with Trn-1 by GST pull-down
assay and surface plasmon resonance (SPR) (Fig. 6).
GST-tagged NLS mutant proteins immobilized to
glutathione-beads were incubated with HeLa cell
extracts at 4 °C for 4 h. NLS interacting proteins iso-
lated by the beads were probed for Trn-1 by immuno-
blotting using anti-Trn-1 (Fig. 6A, upper panel).
Protein blots stained with Amido Black 10B indicated
excess amounts of GST-NLS mutant proteins (Fig. 6A,
lower panel). DNS (288–306) bound considerable Trn-1
(lane 3), while the N-deletion mutants 292–306, 293–
306, 294–306, and 295–306 bound Trn-1 slightly (lanes

4–7). However, the even shorter mutants 296–306 and
299–306 revealed no Trn-1 binding, like GST as a con-
trol (lane 8 and 9 and 2). This is consistent with the
finding that the minimum 12-residue C-terminal
sequence (295–306) can be transported into the nucleus
in vivo (Fig. 1C, panel g). Then, the DNS Ala scan
mutants mts1–5 and DNS C-two residue single or dou-
ble substitution mutants mts 6–8 described in Fig. 2A
were studied for interaction with cellular Trn-1 in the
same way (Fig. 6B). Ala scan mutants mt1 and mt3
showed weaker interaction with Trn-1 than wtDNS
(lanes 3, 4 and 6), but mt2, mt4 and mt5 showed no
A
B
Fig. 4. Identification of hnRNP D nucleocytoplasmic shuttling sequence. (A) Nucleocytoplasmic shuttling of hnRNP D NLS mutants in hetero-
karyons. HeLa cells transfected with pEGFP-C constructs encoding JKTBP2 and D02 as a GFP fusion protein and with pEGFP-GST con-
structs encoding D02 (282–306), D02 (288–306), D02 (292–306) and mt3 as a GFP-GST fusion protein were expressed for 22 h. The cells
were then fused with 3T3 cells and cultured for another 1 h in the presence of cycloheximide. Arrows point out the murine nucleus. Panels
a–f, fluorescent signals of the GFP- or GFP-GST-tagged proteins indicated at the top of each panel; g–l, nuclear morphology upon staining
with Hoechst 33342; m–r, images of heterokaryons merged upon nuclear staining (dashed lines show approximate outlines). A heterokaryon
in the same column is in the same view. (B) Alignment of the hnRNP D shuttling sequence with those of JKTBP1, hnRNP A1 and consen-
sus Trn)1 interaction motif. Identical amino acid and similar amino acid residues in a column were colored pink. J, Hydrophilic amino acid; Z,
hydrophobic amino acid; X, any residue.
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
3980 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
interaction with Trn-1 (lanes 5, 7 and 8). Both the last
two amino acid C-terminal mutants mts 6, 7, and 8
showed almost no interaction with Trn-1 (lanes 9–11).
Next, the direct interactions between the two purified
recombinant DNS NLS mutants and Trn-1 at 25 °C

were studied by SPR. The response signal was monit-
h
r
GFP
DAPI
288-306
292-306
296-306
++
+
+
ab c d
efg kl
mno p
q
1-306
288-303
+
ij
st
GFP
DAPI
Trn-1
DAPI
GST-GFP-DNS
ab
cd ef
Competitors
hnRNP A1 UP-1
A

B
30
o
C4
o
C
C
+ Trn-1
Ran
+ energy
+ Ran
energy
+ Apyrase
+ Ran
+ energy
+ Ran
+ energy
Tr n- 1
293-306
+
abc
d
GFP
DAPI
Trn-1
uvwx
a
e
bcd
hfg

Fig. 5. Nuclear imports of full-length D02 and DNS deletion mutants in permeabilized HeLa cells. (A) The right panels show inhibition of the
nuclear import of DNS ⁄ (288–306) by hnRNP A1. Permeabilized HeLa cells were incubated with 0.1 l
M GST-EGFP-DNS fusion proteins in the
presence or absence of a 40-fold molar excess of the competitors indicated on the top of each panels in transport buffer (10 lL) supplemen-
ted with 4 lL of reticulocyte lysates at 30 °C for 30 min. The cells were stained with anti-GST IgG followed with goat anti rabbit IgG (H + L)-
biotin conjugate and streptoavidin. Panels a–c, localization of DNS; panels d–f, nuclear DNA stained with DAPI in the same views as in panels
a–c. The left panels show temperature dependent nuclear import of DNS. Permeabilized cells were incubated with 2 l
M GST-EGFP-DNS
fusion protein as described above at the indicated temperature of 30 °Cor4°C and studied for GFP fluorescent signal and nuclear DNA. Pan-
els a and b, localization of DNS; panels c and d, nuclear DNA in the same views as in panels a and b, respectively. (B) Dependency of DNS
nuclear import on Trn-1, RanGTP generating system and energy-regenerating system. Permeabilized cells were incubated with 4 l
M GST-
EGFP-DNS, 0.2 l
M Trn-1, Ran mixture and an energy-regenerating system as described in the Experimental procedures, except that either
Trn-1, Ran mixture or the energy-regenerating system was omitted and on omission of the latter apyrase (1 unit) was added to deplete resid-
ual ATP and GTP as stated at the top of the panels. Note high substrate and low Trn-1. The cells were studied for GFP-fluorescent signals and
nuclear DNA stain. Upper and lower panels in a column show the same view. (C) Permeabilized cells were incubated with 2 l
M GST-EGFP-
D02 (1–306) and -DNS N-terminal and C-terminal deletion mutants as transport substrates stated at the tops of the panels in the presence (+)
or absence (–) of 2 l
M Trn-1 in a nuclear import mixture containing Ran mixture and energy mixture and the cells were studied by fluores-
cence microscopy. a and b, full-length D02; c and d, DNS ⁄ 288–306, e and f, g and h, and i and j, N-deletion DNS mutants 292–306, 293–306,
and 296–306, respectively; k and l, C-deletion DNS mutant 288–303. Upper and lower panels in columns show the same view.
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3981
ored for 2 min in flow of 100 nm Trn-1 over the various
DNS NLS mutant-GST fusion proteins which had been
immobilized on sensor chips. Figure 6C–E shows their
sensorgrams. Figure 6D indicates that full-length D02
interacted with Trn-1 as well as DNS (curves 1 and 2).

The DNS N-deletion mutants D02 (292–306) and D02
(293–306) interacted with Trn-1 much less well than
DNS, but significantly more than the control GST-GFP
(Fig. 6C, curves 1–3 and 5). However, the DNS C-dele-
tion mutants D02 (288–303), D02 (288–300) and D02
(288–297) did not interact with Trn-1 at all, like the con-
trol GST-GFP (Fig. 6D, curves 3–6 and Fig. 6C, curve
4). The Ala scan mutants mt3 and mt1 interacted with
Trn-1 much less well than DNS (Fig. 6E, curves 1–3).
Other Ala scan mutants, mts4, 2 and 5, and the C-two
amino acid substitution mutants mts6, 7, 8 and 9 did
not interact with Trn-1 significantly like GST (Fig. 6E,
curves 4–11). These results substantiate the importance
of the N-seven amino acids and the last two C-terminal
amino acids PY in DNS for interaction with Trn-1.
Affinities of D02, DNS and DNS mutants for Trn-1
Kinetic parameters of the association rate constants
(k
a
), dissociation rate constants (k
d
) and dissociation
constants (K
D
) of D02, DNS, and DNS mutants for
Trn-1 were determined at various concentrations of
Trn-1 (2.5–40 nm)at25°C by SPR (Table 1). The K
D
A
B

C
D
E
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
3982 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
of D02 was 4.0 nm and the K
D
of DNS 9.2 nm, that is
double that of D02. This larger K
D
was accounted for
by the twofold larger k
d
of DNS than that of D02 and
their nearly similar k
a
values. The N-four residue dele-
tion mutant 292–306 of DNS and a one residue shorter
mutant 293–306 had, respectively, about six and 14
times larger K
D
values than that of DNS. These larger
K
D
values were largely accounted for by the decreased
k
a
values. These results indicate that the DNS N-seven
amino acids sequence contributes greatly to the associ-
ation with Trn-1. The DNS Ala scan mutants mt1,

2 and 3 had about seven times larger K
D
values than
DNS, largely accounted for by their lower k
a
values.
Other mutants, including the C-terminal mutants,
showed too weak bindings to determine as described
above. For comparison, the K
D
, k
a
, and k
d
values of
hnRNP A1 for Trn-1 were estimated to be 6.3 nm,
2.2 · 10
5
m
)1
Æs
)1
, and 1.4 · 10
)3
s
)1
, respectively. The
K
D
values of A1 M3 NLS (238–320) and TAP-NLS

(61–102) for Trn-1 have been determined as 2.8 and
18.7 nm, respectively, by fluorescence titration and
SPR [34]. This indicates that affinity of D for Trn-1 is
in a similar order to those of A1 and TAP.
Discussion
In this study we identified the NLS and NES of D02
as an hnRNP D nucleocytoplasmic shuttling sequence
(DNS), which is located at the C-terminal tail. The
sequence identified was 19 amino acids long,
SGYGKVSRRGGHQNSYKPY (residues 288–306),
which is encoded by exon 8 common to all D isoforms.
Mutational analysis of DNS indicated that two separ-
ate regions in DNS, the N-terminal seven amino acids
and the two C-terminal amino acids, are essential for
nuclear import mediated by Trn-1.
Heterokaryon assay indicated that DNS as well as
hnRNP D rapidly shuttles between the nucleus and
the cytoplasm. A DNS mutant lacking an N-terminal
SGYG sequence was imported into the nucleus, but
could not be exported from the nucleus. It would
appear that nuclear export of DNS occurs in a facilita-
ted manner but not diffusion. Nuclear export of
hnRNP D has been shown to be insensitive to lepto-
mycin B and therefore to be independent of a nuclear
export receptor of CRM1 ⁄ exportin-1 [30]. Consistent
with this finding, the DNS sequence has no similarity
Table 1. Kinetic parameters of D02, DNS, and DNS N-deletion and
Ala scan mutants interacting with Trn-1. GST-fusion forms of D02,
DNS and DNS mutants (Figs 1 and 2) were bound as a ligand to an
anti-GST Ig-immobilized sensor chip and sensorgrams were

obtained by injecting various concentrations (2.5, 5, 10, 20 and
40 n
M) of Trn-1 as an analyte at 25 °C and then were analyzed
using
BIACORE kinetic software.
Ligand
Trn-1 as an analyte
k
a
(1 ⁄ M
)1Æ
s
)1
) k
d
(1 ⁄ s) K
D
(nM)
D02 7.5 · 10
5
3.0 · 10
)3
4.0
DNS ⁄ 288–306 7.2 · 10
5
6.6 · 10
)3
9.2
292–306 1.4 · 10
5

3.5 · 10
)3
24.7
293–306 0.7 · 10
5
4.0 · 10
)3
56.2
mt1 1.2 · 10
5
3.6 · 10
)3
29.8
mt2 1.1 · 10
5
3.1 · 10
)3
29.2
mt3 1.4 · 10
5
4.3 · 10
)3
30.3
Fig. 6. Interaction between D02 NLS mutants and Trn-1. (A) Interaction of GST-tagged DNS N-deletion NLS mutants with cellular Trn-1. Var-
ious D02 NLS mutants fused with the C-terminal end of GST in place of GFP-GST-PAD described in Fig. 1B were produced in Escherichia
coli, and purified as glutathione-Sepharose bead-bound forms. The bead-bound GST-NLS fusion proteins were incubated with HeLa cell
extracts for 4 h at 4 °C and then washed. Bead-bound proteins were eluted and analyzed by immunoblotting using anti-Trn-1. The upper
panel shows the immunoblots; lane 1, 16 lg of cell extracts used as a source for pull-down assay; lanes 2–9, pull-down assays from 160 lg
of cell extracts; 2, GST; 3, DNS ⁄ 288–306; 4, D02 (292–306); 5, D02 (293–306); 6, D02 (294–306); 7, D02 (295–306); 8, D02 (296–306); 9,
D02 (299–306); lower panel: the same blots stained with Amide Black 10B. Only positively stained sections of the immuno and protein blots

are shown. Arrows on the right show Trn-1 and GST-NLS. (B) Interaction of GST-tagged DNS Ala scan mutants and point mutation mutants
with cellular Trn-1. DNS Ala scan mutants and C-terminal point mutation mutants described in Fig. 2 were allowed to express GST fusion
protein and purified. They were subjected to GST pull-down with cell extracts as described above. The upper panel shows the immunoblots
probed with anti-Trn-1. The lower panel shows the same blots stained for protein. Lanes; 1, one-tenth of cell extracts; 2–7, pull-down assays
of cell extracts with GST-fusion proteins; 2, GST; 3, DNS (wt); 4, mt1; 5, mt2; 6, mt3; 7, mt4; 8, mt5; 9, mt6; 10, mt7; 11, mt8. (C) SPR ana-
lysis of interaction of Trn-1 with DNS N-deletion NLS mutants. The purified Trn-1 and NLS mutant proteins were used as analyte and ligands,
respectively. The response signal was monitored for 2 min at 25 °C on injection of Trn-1 (100 n
M in HBS-EP) over the DNS various N-dele-
tion mutant GST-GFP fusion proteins which had been immobilized on an anti-GST antibody-bound sensor chip and then HBS-EP. 1,
DNS ⁄ 288–306; 2 (292–306); 3 (293–306); 4, C-deletion mutant (288–303); 5, GST-GFP. (D) SPR analysis of interaction of Trn-1 with DNS
C-deletion NLS mutants. The response signal was monitored on a flow of Trn-1 over the DNS various C-deletion NLS mutants GST-GFP fusion
protein-immobilized sensor chips as described above. 1, full length D02; 2, DNS; 3–6 * merging lines: 3 (288–303); 4 (288–300); 5 (288–297);
6, GST-GFP. Note no measurable difference in 3–6. (E) SPR analysis of interaction of Trn-1 with DNS Ala scan mutants and C-terminal
mutants. Response signals were monitored over the various DNS Ala scan mutants and single point mutants described in (B) as described
above. 1, DNS; 2, mt3; 3, mt1; 4–11 *, from upper line to lower line: 4, mt4; 5, mt2; 6, mt7; 7, mt5; 8, mt6; 9, mt9; 10, mt8; 11, GST.
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3983
to a leucine-rich NES which is contained in many
CRM1-mediated nuclear export proteins [35]. In addi-
tion, DNS N-deletion mutants suggested that N-ter-
minal four residues are important for nuclear export,
but not necessarily essential for nuclear import. To
understand whether export sequence is only limited to
an N-terminal region of DNS or contains a whole
DNS sequence, further investigation is needed.
Two smaller isoforms (D01 and D02) were found to
contain an exon 6 and 8 encoded 35 amino acid NLS,
but not NES [30]. It has been suggested that larger iso-
forms (D1 and D2) contain inactive NLS with an exon
7 insertion containing an NES and their associations

with a smaller isoform are involved in nuclear import
and consequently D shuttling occurs [30]. However, in
this work, mutational analysis of DNS in the D02
molecule suggested that D02 is able to shuttle alone
between the nucleus and the cytoplasm. This discrep-
ancy remains to be understood.
The DNS is rich in hydrophilic amino acids and
glycine, and differs from the classical basic type NLS.
Successive deletions of up to seven residues from the
N-terminal portion of DNS gradually reduced the
nuclear import activity in vivo and in vitro and also
similarly decreased its binding to Trn-1. In contrast,
deletion of the last three C-terminal amino acids
KPY, alanine substitution and even point mutation
of either of the last two C-terminal amino acids PY
completely abolished the in vivo and in vitro nuclear
import activity as well as the binding to Trn-1. Ala-
nine scanning mutagenesis of the sequence linking the
two motifs had moderate effects on nuclear import
and binding to Trn-1. Thus, the N- SGYGKVS (288–
294) and C-PY (305–306) motifs in DNS are more
crucial for nuclear import than the internal 10-residue
sequence separated by the two motifs. Interestingly,
these two motifs are conserved in JKTBP1 and
ABBP1 C-terminal tail sequences and an N-terminal
portion (271–289) of M9 when they are aligned
(Fig. 4B and [33]) The N-SGYGKVS motif is also
conserved in the C-portion of the consensus Trn)1
interaction motif (12 residues), which is derived from
randomized M9s and is necessary for both the import

and export activity of M9 [16]. However, the PY
motif is located 10 residues C-terminal of the consen-
sus Trn interaction motif (Fig. 4B). Some differences
between the nuclear imports in vivo and in vitro of
DNS N-deletion and substitution mutants were
observed (compare Fig. 1C with Fig. 5C). Nuclear
import in vivo appeared to plateau within 1 day.
However, it is not known at what time point plateau
was reached. Therefore, the results of in vivo and
in vitro could not be directly compared.
The D1 C-terminal 112 amino acid sequence on a
blot has been shown to bind to Trn-1 [13]. In vitro
transport assay provided convincing evidence that
DNS nuclear import is mediated through NPC by
Trn-1. Ran and energy were required for nuclear
import at a low concentration of 0.1–0.2 lm Trn-1 with
a high concentration of 4 lm DNS, but not at a high
concentration of 1 lm Trn-1. This indicates that the
translocation of the substrate–Trn-1 complex through
NPCs to the nucleoplasmic side is independent of both
Ran and energy and that RanGTP and GTP energy
are required only for the release of substrate from the
substrate Trn-1 complex and for multiple rounds of
Trn-1-mediated nuclear import, as was found in M9
Trn-1-mediated nuclear transport [36–38]. SPR analysis
provided clear evidence that DNS interacts directly
with Trn-1. In DNS, both the N-terminal seven amino
acid sequence SGYGKVS (288–294) and the last two
C-terminal residues PY are essential for binding to
Trn-1. The shorter, import-deficient N- and C-terminal

deletion mutants, which also show no ability for in vitro
nuclear import, do not bind Trn-1. DNS N-deletion
and Ala scan mutants, which have reduced ability for
nuclear import, revealed 6–14 times larger K
D
values
with a decreased k
a
and fairly invaried k
d
for Trn-1 as
compared with DNS. Whether these DNS mutations
affect release of substrate from the substrate–Trn com-
plex on binding to RanGTP remains to be studied.
Structural studies on the Trn-1 ⁄ karyopherin b2A-
RanGppNp complex indicated that the structural
change of Trn-1 upon binding RanGTP in its N-ter-
minal arch is transmitted through a long internal
acidic loop to the substrate–Trn complex in its
C-terminal arch concomitantly with release of the sub-
strate [34,39]. As the C-terminal tail of D is predicted
to have no secondary structure, the extended DNS
conformation might be stabilized at the N- and C-ter-
minal residues by binding to the C-terminal arch of
Trn-1, as found in basic type bipartite NLS importin a
complexes [40,41]. Trn-1 recognizes various kinds of
nuclear proteins and the nuclear pore complex proteins
nup98 and nup153 as substrates [3,11,12,17–19,33].
DNS and its mutants will aid in understanding such a
broad recognition of Trn-1 on the basis of the crystal

structure of the Trn-1-substrate complex.
Experimental procedures
Clonings of cDNAs for hnRNP D02, Trn-1,
RanGAP and Ran
A human full length hnRNP D02 cDNA was isolated from
a human monocytic SKM-1 cDNA library using a RBD1
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
3984 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
of JKTBP1 cDNA as a probe in the course of JKTBP clo-
ning [42]. Its entire coding sequence was amplified by PCR
with a set of 5¢ and 3¢ EcoRI-attached primers (0.5 lm) and
6% (v ⁄ v) dimethylsulfoxide, and subcloned into an EcoRI
site of pGEX-6P-1 (Amersham BioScience, Inc., Piscata-
way, NJ, USA) and pEGFP-C vectors (Clontech, Mountain
View, CA, USA). The entire coding regions of human
Trn-1, RanGAP and Ran were amplified by RT-PCR using
HeLa cell poly(A)
+
RNA as a template using the Titan
one-tube reverse transcription-PCR system (Roche Diag-
nostics, Indianapolis, IN, USA) and then cloned at a Bam-
HI ⁄ SalI site of pGEX-6P-3.
Plasmid constructs for hnRNP D NLS mutants
N-Terminal deletion NLS mutants were constructed by
amplification by PCR using hnRNP D02 cDNA as a tem-
plate and a primer set of a 5¢ EcoRI-attached primer
designed for the defined N-terminal amino acid sequence of
NLS and a 3¢ SalI-attached primer designed for 3¢-UTR.
The amplified cDNAs were subcloned at an EcoRI ⁄ SalI site
in appropriate expression vectors: the bacterial expression

vectors used were pGEX-6P-1 and pGEX-6P-3-EGFP-C
composite vector encoding a GST-EGFP fusion protein;
the mammalian expression vectors were pEGFP-C, pEG-
FP-GST encoding an EGFP-GST fusion protein, and
pEGFP-GST-PAD encoding a tripartite protein EGFP-
GST-PAD (PAD is an N-terminal 104 residues sequence of
a NLS deficient full-length PAD IV ⁄ V(1–663) [43]).
For the preparation of C-terminal deletion NLS mutants,
alanine scanning mutants and substitution mutants, syn-
thetic oligomers (30–35-mer) were used as PCR primers,
containing a stop codon or codons for defined single or tri-
ple consecutive amino acid substitutions, and a pEGFP-
GST-PAD-D02 NLS was used as a template, with 18 cycles
of a 12 min period using the Quick Change XL site-directed
mutagenesis kit (Stratagene, La Jolla, CA, USA). Integrity
was verified by sequencing.
Preparation of GST-fusion proteins
All recombinant proteins were expressed as fusion proteins
with GST in Escherichia coli BL21 cells with 0.1 mm isopro-
pyl thio-b-d-galactoside for 6–12 h at 25 °C. The cells were
collected and suspended in buffer [150 mm NaCl, 10 mm
Tris ⁄ HCl (pH 8.0)], 5 mm EDTA, and 1 mm phenyl-
methansulfonyl fluoride containing 5 mm dithiothreitol, 1%
(v ⁄ v) Triton X-100 and 200 lgÆmL
)1
lysozyme. The cell sus-
pension was sonicated 20 times for 10 s periods and then cen-
trifuged at 12 000 g for 30 min. The resultant supernatant
was mixed with glutathione-Sepharose 4B beads (Amersham
Biosciences) for 1 h and the beads were washed four times

with buffer [10 mm Tris ⁄ HCl (pH 8.0), 5 mm EDTA, and
5mm dithiothreitol]. Bound GST-fusion proteins were eluted
into elution buffer [100 mm Tris ⁄ HCl (pH 8.0)] and 20 mm
glutathione. GST-free recombinant proteins were prepared
by the digestion with Precission protease (Amersham Bio-
sciences) at 5 °C for 4 h.
In vivo nuclear import assay and heterokaryon
assay
Assays were performed as described previously [33]. HeLa
cells (4 · 10
6
cells) suspended in 0.5 mL of RPMI-1640 med-
ium were transfected by electroporation with pEGFP-GST-
PAD (30 l g) carrying various NLS. About 1 · 10
5
cells were
grown on a 15 mm glass coverslip and the rest were grown
for analysis of expression of EGFP-fusion proteins in a 9-cm
dish. After 24 h incubation, the cells grown on the coverslip
were fixed with 4% (v ⁄ v) paraformaldehyde (Sigma, St
Louis, MO, USA) in NaCl ⁄ P
i
(–) and stained with 1 lgÆmL
)1
DAPI in NaCl ⁄ P
i
(–). The cells in mounting solution were
studied for subcellular distributions by fluorescence micros-
copy. Fluorescent signals were recorded with a cooled CCD
camera.

For heterokaryon assay, 3.8 · 10
4
HeLa cells transfected
with various plasmids were grown on a 0.1% (w ⁄ v) gelatin-
coated glass coverslip for 24 h. First the slides were rinsed
and overlaid with 5 · 10
4
Balb ⁄ C NIH3T3 cells and cultured
for 3 h, and then for 15 min with 20 lgÆmL
)1
cycloheximide.
The cells were then rinsed and fused by exposure to 100 lL
of 50% (w ⁄ v) polyethylene glycol 3400 (Polysciences,
Warrington, PA) in RPMI 1640 medium for 2 min at 37 °C,
and after washing twice with NaCl ⁄ P
i
(–), were incubated in
Dulbecco’s modified Eagle’s medium containing 20 lgÆmL
)1
cycloheximide at 37 °C for 1 h. The cells were then fixed
and stained with 25 lgÆmL
)1
Hoechst 33342 (Sigma) in
NaCl ⁄ P
i
(–) and observed by fluorescence microscopy.
In vitro nuclear import assay
Exponentially growing HeLa cells on an eight-well multitest
slide (MP) which had been coated with 50 l g ÆmL
)1

poly(l-lysine) (30 000–70 000, Sigma) in NaCl ⁄ P
i
(–) for a
few minutes were permeabilized by treatment with
40 lgÆmL
)1
digitonin (Wako, Osaka, Japan) in transport
buffer (20 mm Hepes pH 7.3, 2 mm magnesium acetate,
5mm sodium acetate, 110 mm potassium acetate, 0.1 mm
EGTA, 1 mm dithiothreitol and 0.25 m sucrose) for 5 min
on ice and were rinsed twice with transport buffer as
described previously [33,44,45]. The cells were overlaid with
import reaction mixture (10 lL) containing 2 lm GST-
EGFP-hnRNP D as a cargo, 2 lm Trn-1 as a receptor,
Ran mixture (3 lm RanGDP, 0.3 lm NTF2 (Sigma) and
0.2 lm RanGAP) and an energy-regenerating system
[0.2 mm ATP, 0.2 mm GTP, 4 mm creatine phosphate and
10 lgÆmL
)1
creatine kinase (Roche Diagnostics)] in the
transport buffer and incubated at 30 °C for 30 min 1 U
apyrase (grade VI, Sigma) was added to the reaction mix
without an energy-regenerating system to deplete residual
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3985
ATP. Without transport factors, reticulocyte lysates
(Promega, Madison, WI, USA) was used. After the reac-
tion, the cells were fixed with 4% (v ⁄ v) paraformaldehyde
for 30 min at 4 °C and were studied for the nuclear local-
ization by fluorescence microscopy.

Interaction between hnRNP D NLS and Trn-1
A GST-pull down assay was performed as described previ-
ously [33]. GST-hnRNP D02 and GST-NLS mutant fusion
protein were purified as a glutathione bead-bound form.
About 20 lL of a 50% slurry of the packed beads was
mixed with HeLa cell extract (5–8 mg) in a volume of
1.5 mL binding buffer at 4 °C for 4 h. Bound proteins were
eluted from the beads with SDS sample buffer and analyzed
by immunoblotting using 1000-fold diluted Trn-1 monoclo-
nal antibody D45. Protein concentrations were determined
by the Bradford method, using bovine serum albumin as a
standard [46].
SPR analysis was performed at 25 °C with a BIA-
CORE type 3000 instrument. Anti-GST IgG was fixed to
a sensor chip CM5 at a concentration of 16820 resonance
units (RU) by amine coupling according to instructions
of the supplier (BIAcore BR-10023 and BR-100–50) and
then GST-GFP-NLS proteins were immobilized to the
chip at a concentration of about 100 RU by injecting
0.074–0.1 pmolÆlL
)1
solution in HBS-EP. The SPR signal
was monitored by injection of Trn-1 (0.075–0.1 pmo-
lÆlL
)1
) over the chip at a flow rate of 20 lLÆmin
)1
for
2 min and then HBS-EP. HBS-EP contained 0.01 m
Hepes, pH 7.4, 0.15 m NaCl, 3 mm EDTA, and

0.005% (v ⁄ v) Tween 20. Surface regeneration was accom-
plished with 10 mm glycine ⁄ HCl, pH 2.2. k
a
, k
d
and
K
D
were determined at various concentrations of Trn-1
with biaevaluation version 3.0 (with a global fitting
program).
Acknowledgements
We are grateful for Dr Gideon Dreyfuss (University
of Pennsylvania School of Medicine, PA) for a gift of
antibody D45 to Trn-1 and to Dr Ichiro Tanaka for
the use of a fluorescence microscope. We thank
Manabu Takahashi and Tsuyoshi Imasaki for their
help in this work and Hidenobu Kawamura and Dr
Mamoru Sato for discussion and advice. This work
was partly supported by grants-in-aids for Promotion
of Research at Yokohama City University.
References
1Go
¨
rlich D & Kutay U (1999) Transport between the cell
nucleus and the cytoplasm. Annu Rev Cell Dev Biol 15,
607–660.
2 Nakielny S & Dreyfuss G (1999) Transport of proteins
and RNAs in and out of the nucleus. Cell 99, 677–690.
3 Fried H & Kutay U (2003) Nucleocytoplasmic trans-

port: taking an inventory. Cell Mol Life Sci 60, 1659–
1688.
4 Macara IG (2001) Transport into and out of the nuc-
leus. Microbiol Mol Biol Rev 65, 570–594. table of con-
tents.
5 Bednenko J, Cingolani G & Gerace L (2003) Nucleo-
cytoplasmic transport: navigating the channel. Traffic 4,
127–135.
6 Vautier D, Chesne P, Cunha C, Calado A, Renard JP
& Carmo-Fonseca M (2001) Transcription-dependent
nucleocytoplasmic distribution of hnRNP A1 protein in
early mouse embryos. J Cell Sci 114, 1521–1531.
7 Siomi MC, Eder PS, Kataoka N, Wan L, Liu Q &
Dreyfuss G (1997) Transportin-mediated nuclear import
of heterogeneous nuclear RNP proteins. J Cell Biol 138,
1181–1192.
8 Caceres JF, Screaton GR & Krainer AR (1998) A speci-
fic subset of SR proteins shuttles continuously between
the nucleus and the cytoplasm. Genes Dev 12, 55–66.
9 Fan XC & Steitz JA (1998) HNS, a nuclear-cytoplasmic
shuttling sequence in HuR. Proc Natl Acad Sci USA 95,
15293–15298.
10 Pinol-Roma S & Dreyfuss G (1992) Shuttling of
pre-mRNA binding proteins between nucleus and cyto-
plasm. Nature 355, 730–732.
11 Pollard VW, Michael WM, Nakielny S, Siomi MC,
Wang F & Dreyfuss G (1996) A novel receptor-mediated
nuclear protein import pathway. Cell 86, 985–994.
12 Guttinger S, Muhlhausser P, Koller-Eichhorn R, Bren-
necke J & Kutay U (2004) Transportin2 functions as

importin and mediates nuclear import of HuR. Proc
Natl Acad Sci USA 101, 2918–2923.
13 Siomi MC, Fromont M, Rain JC, Wan L, Wang F,
Legrain P & Dreyfuss G (1998) Functional conservation
of the transportin nuclear import pathway in divergent
organisms. Mol Cell Biol 18, 4141–4148.
14 Siomi H & Dreyfuss G (1995) A nuclear localization
domain in the hnRNP A1 protein. J Cell Biol 129,
551–560.
15 Michael WM, Choi M & Dreyfuss G (1995) A nuclear
export signal in hnRNP A1: a signal-mediated, tempera-
ture-dependent nuclear protein export pathway. Cell 83,
415–422.
16 Bogerd HP, Benson RE, Truant R, Herold A, Phing-
bodhipakkiya M & Cullen BR (1999) Definition of a
consensus transportin-specific nucleocytoplasmic trans-
port signal. J Biol Chem 274, 9771–9777.
17 Nakielny S, Shaikh S, Burke B & Dreyfuss G (1999)
Nup153 is an M9-containing mobile nucleoporin with a
novel Ran-binding domain. EMBO J 18, 1982–1995.
18 Fontoura BM, Blobel G & Yaseen NR (2000) The
nucleoporin Nup98 is a site for GDP ⁄ GTP exchange on
3986 FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS
hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence M. Suzuki et al.
ran and termination of karyopherin beta 2-mediated
nuclear import. J Biol Chem 275, 31289–31296.
19 Truant R, Kang Y & Cullen BR (1999) The human tap
nuclear RNA export factor contains a novel transpor-
tin-dependent nuclear localization signal that lacks
nuclear export signal function. J Biol Chem 274, 32167–

32171.
20 Krecic AM & Swanson MS (1999) hnRNP complexes:
composition, structure, and function. Curr Opin Cell
Biol 11, 363–371.
21 Guhaniyogi J & Brewer G (2001) Regulation of mRNA
stability in mammalian cells. Gene 265, 11–23.
22 Inoue A, Arao Y, Omori A, Ichinose S, Nishio K,
Yamamoto N, Kinoshita Y & Mita S (2003) Identifica-
tion of S1 proteins B2, C1 and D1 as AUF1 isoforms
and their major role as heterogeneous nuclear ribonu-
cleoprotein proteins. Biochem J 372, 775–785.
23 Wagner BJ, DeMaria CT, Sun Y, Wilson GM & Brewer
G (1998) Structure and genomic organization of the
human AUF1 gene: alternative pre-mRNA splicing gen-
erates four protein isoforms. Genomics 48, 195–202.
24 Lu JY & Schneider RJ (2004) Tissue distribution of
AU-rich mRNA binding proteins involved in regulation
of mRNA decay. J Biol Chem 279, 12974–12979.
25 Wilson GM, Lu J, Sutphen K, Sun Y, Huynh Y &
Brewer G (2003) Regulation of A + U-rich element-
directed mRNA turnover involving reversible phos-
phorylation of AUF1. J Biol Chem 278, 33029–33038.
26 Laroia G, Cuesta R, Brewer G & Schneider RJ (1999)
Control of mRNA decay by heat shock-ubiquitin-pro-
teasome pathway. Science 284, 499–502.
27 Tolnay M, Baranyi L & Tsokos GC (2000) Heteroge-
neous nuclear ribonucleoprotein D0 contains transacti-
vator and DNA-binding domains. Biochem J 348,
151–158.
28 Hanakahi LA & Maizels N (2000) Transcriptional acti-

vation by LR1 at the Emu enhancer and switch region
sites. Nucleic Acids Res 28, 2651–2657.
29 Chen CY, Xu N, Zhu W & Shyu AB (2004) Functional
dissection of hnRNP D suggests that nuclear import is
required before hnRNP D can modulate mRNA turn-
over in the cytoplasm. RNA 10, 669–680.
30 Sarkar B, Lu JY & Schneider RJ (2003) Nuclear import
and export functions in the different isoforms of the
AUF1 ⁄ heterogeneous nuclear ribonucleoprotein protein
family. J Biol Chem 278, 20700–20707.
31 Arao Y, Kuriyama R, Kayama F & Kato S (2000) A
nuclear matrix-associated factor, SAF-B, interacts with
specific isoforms of AUF1 ⁄ hnRNP D. Arch Biochem
Biophys 380, 228–236.
32 Loflin P, Chen CY & Shyu AB (1999) Unraveling a
cytoplasmic role for hnRNP D in the in vivo mRNA
destabilization directed by the AU-rich element. Genes
Dev 13, 1884–1897.
33 Kawamura H, Tomozoe Y, Akagi T, Kamei D,
Ochiai M & Yamada M (2002) Identification of the
nucleocytoplasmic shuttling sequence of heterogeneous
nuclear ribonucleoprotein d-like protein JKTBP and
its interaction with mRNA. J Biol Chem 277, 2732–
2739.
34 Chook YM, Jung A, Rosen MK & Blobel G (2002)
Uncoupling Kapbeta2 substrate dissociation and ran
binding. Biochemistry 41, 6955–6966.
35 Fornerod M & Ohno M (2002) Exportin-mediated
nuclear export of proteins and ribonucleoprotein. In
Nuclear Transport (Weiss K, ed), pp. 67–91. Springer-

Verlag, Heidelberg.
36 Ribbeck K, Kutay U, Paraskeva E & Gorlich D (1999)
The translocation of transportin-cargo complexes
through nuclear pores is independent of both Ran and
energy. Curr Biol 9, 47–50.
37 Ribbeck K & Gorlich D (2001) Kinetic analysis of
translocation through nuclear pore complexes. EMBO J
20, 1320–1330.
38 Nakielny S & Dreyfuss G (1998) Import and export of
the nuclear protein import receptor transportin by a
mechanism independent of GTP hydrolysis. Curr Biol 8,
89–95.
39 Chook YM & Blobel G (1999) Structure of the nuclear
transport complex karyopherin-beta2-Ran x GppNHp.
Nature 399, 230–237.
40 Hodel MR, Corbett AH & Hodel AE (2001) Dissection
of a nuclear localization signal. J Biol Chem 276, 1317–
1325.
41 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.
42 Kamei D, Tsuchiya N, Yamazaki M, Meguro H &
Yamada M (1999) Two forms of expression and geno-
mic structure of the human heterogeneous nuclear ribo-
nucleoprotein d-like JKTBP gene (HNRPDL). Gene
228, 13–22.
43 Nakashima K, Hagiwara T & Yamada M (2002)
Nuclear localization of peptidylarginine deiminase V
and histone deimination in granulocytes. J Biol Chem

277, 49562–49568.
44 Adam SA, Marr RS & Gerace L (1990) Nuclear protein
import in permeabilized mammalian cells requires solu-
ble cytoplasmic factors. J Cell Biol 111, 807–816.
45 Go
¨
rlich D, Seewald MJ & Ribbeck K (2003) Characteri-
zation of Ran-driven cargo transport and the RanGT-
Pase system by kinetic measurements and computer
simulation. EMBO J 22, 1088–1100.
46 Bradford MM (1976) A rapid and sensitive method for
the quantitation of microgram quantities of protein util-
izing the principle of protein-dye binding. Anal Biochem
72, 248–254.
FEBS Journal 272 (2005) 3975–3987 ª 2005 FEBS 3987
M. Suzuki et al. hnRNP D ⁄ AUF1 nucleo-cytoplasmic shuttling sequence