You are on page 1of 8

LETTERS

Temporal and spatial control of nucleophosmin by the


Ran–Crm1 complex in centrosome duplication
Wei Wang1, Anuradha Budhu1, Marshonna Forgues1 and Xin Wei Wang1,2

Centrosome duplication is tightly controlled during faithful Recent studies indicate that cellular components that regulate nucleo-
cell division, and unnecessary reduplication can lead to cytoplasmic transport are independently involved in spindle assembly9,10.
supernumerary centrosomes and multipolar spindles that are In principle, these processes are accomplished by specific receptors of
associated with most human cancer cells 1–5. In addition to the β-importin family: for example, the importin receptors (importins
nucleocytoplasmic transport, the Ran–Crm1 network is involved α and β) that bind to nuclear localization signals (NLS) and the export
in regulating centrosome duplication to ensure the formation receptor Crm1 that binds to nuclear export signals (NES). These proc-
of a bipolar spindle6–8. Here, we discover that nucleophosmin esses require a small GTPase, Ran, which controls the interaction of
(NPM) may be a Ran–Crm1 substrate that controls centrosome these receptors with their substrates. The guanine nucleotide-exchange
duplication. NPM contains a functional nuclear export signal factor RCC1 facilitates Ran binding to Crm1, whereas RanBP1, a major
(NES) that is responsible for both its nucleocytoplasmic regulator of Ran, promotes Crm1 dissociation from Ran. Importins α
shuttling and its association with centrosomes, which are and β may negatively regulate NLS-containing proteins to modulate
Ran–Crm1-dependent as they are sensitive to Crm1-specific microtubule assembly. We have recently shown that Crm1 may regu-
nuclear export inhibition, either by leptomycin B (LMB) or by late the fidelity of centrosome duplication by acting as a licencing fac-
the expression of a Ran-binding protein, RanBP1. Notably, tor to prevent unscheduled duplication6. A fraction of Ran, Crm1 and
LMB treatment induces premature centrosome duplication RanBP1 is found on centrosomes6,7,11. Inactivation of Crm1, either by
in quiescent cells, which coincides with NPM dissociation a Crm1-specific inhibitor, LMB12, or through hepatitis B virus (HBV)
from centrosomes. Moreover, deficiency of NPM by RNA HBx protein interaction with its NES motif13, results in supernumerary
interference results in supernumerary centrosomes, which can centrosomes and multipolar spindles6,13. Similar multipolar spindles
be reversed by reintroducing wild-type but not NES-mutated and mitotic abnormalities are also a consequence of Ran mutations or
NPM. Mutation of a potential proline-dependent kinase overexpression of RanBP1 (ref. 8). We hypothesize that the Ran–Crm1
phosphorylation site at residue 95, from threonine to aspartic complex may negatively regulate the initiation of centrosome duplica-
acid (T95D) within the NES motif, abolishes NPM association tion, possibly through its association with NES-containing proteins6.
and inhibition of centrosome duplication. Our results are NPM may negatively regulate centrosome duplication14,15. NPM is a
consistent with the hypothesis that the Ran–Crm1 complex ubiquitously expressed phosphoprotein that is mainly localized in the
may promote a local enrichment of NPM on centrosomes, nucleolus, and shuttles between the nucleus and the cytoplasm during
thereby preventing centrosome reduplication. the cell cycle16. NPM associates with unduplicated centrosomes and dis-
sociates from centrosomes upon phosphorylation by CDK2–cyclin E,
The centrosome is the principal microtubule-organizing centre of ani- which coincides with the initiation of centrosome duplication and DNA
mal cells, and functions in concert with the spindle poles to direct the replication15,17. NPM reassociates with centrosomes during mitosis15,18.
assembly of a bipolar spindle during mitosis1,2. Centrosome duplication Sequence analysis of NPM orthologues indicated the presence of a puta-
is initiated at the G1/S boundary and is completed at the S phase of the tive hydrophobic leucine-rich NES motif (IxxPxxLxL) within residues 94–
cell cycle, which coincides with DNA replication. To ensure the forma- 102 that is conserved from Xenopus laevis to human (Fig. 1a). A putative
tion of a bipolar mother cell, duplication takes place only once per cell bipartite NLS motif and nucleolus localization signal (NoLS) are also evi-
cycle, leading to an equal division into two daughter cells1,2. This feature dent (Fig. 1a). The presence of both NES and NLS motifs in NPM suggests
indicates that there are negative regulatory proteins that act as licencing that a Ran GTPase-dependent pathway regulates NPM nucleocytoplasmic
factors to prevent unscheduled duplication. However, the precise mecha- transport. Moreover, it is possible that the Ran–Crm1 complex modulates
nism of how centrosome duplication is controlled is largely unknown. centrosome duplication by regulating NPM transport and localization.

1
Liver Carcinogenesis Section, Laboratory of Human Carcinogenesis, Center for Cancer Research, NCI, NIH, Bethesda, MD 20892, USA.
2
Correspondence should be addressed to X.W.W. (e-mail: xw3u@nih.gov)

Published online: 24 July 2005; DOI: 10.1038/ncb1282

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005 823

©2005 Nature Publishing Group


LETTERS
a b Phase GFP−NPM hC23 Merge
h-NPM MEDS-MDMD-MS PLRPQNYLFG CELKADK-DYH FKVDNDENEH QLSLRTVSLG 50
m-NPM MEDS-MDMD-MS PLRPQNYLFG CELKADK-DYH FKVDNDENEH QLSLRTVSLG 50
r-NPM MEDS-MDMD-MS PLRPQNYLFG CELKADK-DYH FKVDNDENEH QLSLRTVSLG 50 − LMB
ch-NPM MEDSAMDMESMG PLRPQTFLFG CELKAEK-EYQ FKVDDEENEH QLSLRTVTLG 52
xl-NPM MEDS-MDMDNIA PLRPQNFLFG CELKADKKEYS FKVEDDENEH QLSLRTVSLG 52
Consensus MEDS.MDM.... PLRPQ..LFG CELKA.K..Y. FKV...ENEH QLSLRTV.LG
PDK NES
h-NPM AGAKDELHIV EAEAMNYEGS PIKVTLATLK MSVQPTVSLG GFE ITPPVVL RLK 103
m-NPM AGAKDELHIV EAEAMNYEGS PIKVTLATLK MSVQPTVSLG G FEITPPVVL RLK 103
r-NPM AGAKDELHIV EAEAMNYEGS PIKVTLATLK MSVQPTVSLG GFE ITPPVVL RLK 103
ch-NPM AGAKDELHVV EAEALDYEGN PTKVVLASLK MSVQPTVSLG GFE ITPPFVL RLK 105 + LMB
xl-NPM ASAKDELHVV EAEGINYEGK TIKIALASLK PSVQPTVSLG GFE ITPPVIL RLK 105
Consensus A.AKDELH.V EAE...YEG. ..K..LA.LK .SVQPTVSLG GFE ITPP..L RLK
NPM-ALK NPM-RAR NLS
h-NPM CGSGPVHISG QHLVAVEEDA ESEDEEEEDV KLLSISGKRS APGGGSKVPQ KKV 156
m-NPM CGSGPVHISG QHLVAVEEDA ESEDEDEEDV KLLGMSGKRS APGGGNKVPQ KKV 156
r-NPM CGSGPVHISG QHLVAVEEDA ESEDEDEEDV KLLGMSGKRS APGGGNKVPQ KKV 156
ch-NPM CGSGPVYVSG QHLVALEEEP ESEDE-EEDT KIGNASTKRP ASGGGAKTPQ KKP 157
xl-NPM SGSGPVYVSG QHLVALEDLE SSDDE-DEEH EPSPKNAKRI APDSASKVPR KKT 157
Consensus .GSGPV..SG QHLVA.E... .S.DE..E.. .......KR. A.....K.P. KK. c 100

(percentage of fusion cells)


NPM-MLF1 CDC2
h-NPM KLAADEDDDD DDEEDDDEDD DDDDFDDEEA EEKAPVKKSI RDTPAKNAQK SNQ 209

GFP−NPM-positive
80
m-NPM KLDEDDEDDD EDDEDD-EDD DDDDFDEEET EEKVPVKKSV RDTPAKNAQK SNQ 208

human nuclei
r-NPM KLDEDDDEDD EDDEDD-EDD DDDDFDEEET EEKVPVKKSV RDTPAKNAQK SNQ 208
ch-NPM KLSEDDEDDD EDEDDD--ED DEDDLDDDEE EIKTPMKKPA REPAGKNMQK AKQ 208 60
xl-NPM RLEEEEEDSD EDDDDD--DE DDDDEDDDEE EEETPVKK-T DSTKSKAAQK LNH 207
Consensus .L.......D .D..DD.... D.DD.D..E. E...P.KK.. .....K..QK ...
40
CDC2 CDC2
h-NPM NGKDSKPSST PRS-KGQESF KKQE-K--TP KTPKGPSSVE DIKAKMQASI EKG 258
m-NPM NGKDLKP-ST PRS-KGQESF KKQE-K--TP KTPKGPSSVE DIKAKMQASI EKG 256 20
r-NPM NGKDLKP-ST PRS-KGQESF KKQE-K--TP KTPKGPSSVE DIKAKMQASI EKG 256
ch-NPM NGKDSKP-ST PAS-KTKTPD SKKD-KSLTP KTPKVPLSLE EIKAKMQASV DKG 258
xl-NPM NGKASALSTT QKTPKTPEQK GKQDTKPQTP KTPKTPLSSE EIKAKMQTYL EKG 260
0
Consensus NGK......T ....K..... .K...K..TP KTPK.P.S.E .IKAKMQ... .KG Control LMB
NoLS
h-NPM GSLPKVEAKF INYVKNCFRM TDQEAIQDLW QWRKSL 294
m-NPM GSLPKVEAKF INYVKNCFRM TDQEAIQDLW QWRKSL 292
r-NPM GSLPKVEAKF INYVKNCFRM TDQEAIQDLW QWRKSL 292
ch-NPM CSLPKLEPKF ANYVKNCFRT EDQKVIQALW QWRQTL 294
xl-NPM NVLPKVEVKF ANYVKNCFRT ENQKVIEDLW KWRQSLKDGK 300
Consensus ..LPK.E.KF .NYVKNCFR. ..Q..I..LW .WR..L....

d e
NLS (141−157) WT NESM NESD NLS1D
NES (94−102) NoLS (288−290)
WT GFP - 294
A-A
NESM GFP -
NESD GFP -
NLS1D GFP -
NLS2D GFP -
G-G
NoLSM GFP -
A-A NLS2D NoLSM NESM-NLS1D NLS1/2D
NESM-NLS1D GFP -
NLS1/2D GFP -

Figure 1 Nucleocytoplasmic shuttling of NPM depends on the presence of cells, identified under phase contrast (phase), GFP–NPM signal, human
functional NES and NLS motifs mediated by the Ran network. C23 signal, as well as their merged images, are shown. Magnification:
(a) Sequence comparison of NPMs from five different species was performed ×400. (c) Percentage of heterokaryon fusion cells with the GFP–NPM signal
by the ClustalW program. The NES (yellow), NLS (sky blue) and nucleolus present in human nuclei was quantified by counting more than 100 fusion
localization signal (NoLS; green) motifs are highlighted. Potential CDC2 cells. Values were obtained from three independent experiments, shown
phosphorylation sites (red) and a proline-dependent kinase (PDK) site as mean ± s.d. (d) Schematic depiction of GFP–NPM and its NES, NLS
(red) were determined by Motif Scan. NPM fusion protein sites (grey) that and NoLS mutants. WT, wild type; NESM, mutated at residues L100A and
result from translocation of the N-terminal NPM to ALK, RAR, or MLF1 L102A; NESD, deleted at residues 94–102; NLS1D, deleted at residues
in lymphoma and leukaemia are also indicated. Prefixes: h, human; m, 134–142; NLS2D, deleted at residues 150–155; NoLSM, mutated at
mouse; r, rat; ch, Chinese hamster; xl, Xenopus laevis. (b) The ability of residues W288G and W290G; NESM-NLS1D, with both a L100A/L102A
NPM to shuttle from the nucleus to the cytoplasm was determined by the mutation and a 134–142 deletion; NLS1/2D, deleted at residues 134–142
mouse-human heterokaryon assay in the absence (upper panels) or presence and 150–155. (e) Intracellular localization of the NPM mutants. HCA2-
(lower panels) of the nuclear export inhibitor leptomycin B (LMB). Human hTERT cells were transfected with various mutants, incubated for 24 h, fixed
nucleolin-specific antibody C23 (red) was used to identify human nuclei with paraformaldehyde, and examined under a fluorescence microscope.
(arrows) from mouse nuclei in the heterokaryon. Representative fusion Representative images are shown. Magnification: ×400.

We tested this hypothesis by using a heterokaryon assay in which a antibody was used to distinguish human cells from murine cells (Fig. 1b;
mouse–human cell fusion was induced by polyethylene glycol19,20 to evalu- arrows). The green fluorescent protein (GFP)–NPM signal, predomi-
ate whether NPM in mouse nuclei was exported to the cytoplasm, and then nantly localized in the nucleolus, was frequently found not only in murine
transported into human nuclei. A human-specific anti-nucleolin (C23) nuclei, but also in human nuclei where they were a part of a heterokaryon

824 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005

©2005 Nature Publishing Group


LETTERS

a b Endogenous NPM c
NPM γ-tubulin Merge GFP−NPM
GFP−NPM-NESM

NPM centrosomes (%)


100

NPM centrosomes (%)


100
80

Cells with
75

Cells with
G0/G1
60
50
40
25
20
0
0

MB

P1
l
tro
G0/G1 S/G2 M

HB

nB
n

+L
Co

Ra
S/G2

d e

n ≥ 2 centrosomes (%)
100 n=1

Cells with n = 1 or
n≥2 G1: 91% G1: 91%
75 S: 4% S: 3%
G2/M: 5% G2/M: 6%
M 50

25

0
− LMB + LMB
− LMB + LMB

f
Control 50 nM LMB
NPM

abnormal centrosomes (n>2) (%)


g h
NPM centrosomes (%)

100
40
Mitotic cells with
Centrosomes

Mitotic cells with


80
30
60
40 20

20 10
0
0
B
l
tro

LM

B
l
tro
n

LM
Merge

Co

n
nM

Co

nM
50

50
Figure 2 The centrosome association of NPM and initiation of centrosome of 40 nM LMB for 2 h, or with expression of HBx or GFP–RanBP1. More
duplication are sensitive to Ran–Crm1-specific nuclear export inhibition. than 100 cells were analysed. Data are an average of three independent
(a) Localization of NPM on G0/G1 phase, S/G2 phase, or mitotic phase experiments and are shown as mean ± s.d. (d) The percentage of cells
centrosomes. HCA2-hTERT cells were fixed and stained with anti-NPM that contain either one (n = 1) or two or more (n ≥ 2) centrosomes was
(green panels) and anti-γ-tubulin for centrosomes (red panels). The following calculated in serum-starved HCA2-hTERT cells treated with or without LMB.
representative cells with different stages of centrosomes (arrows) and their At least 100 cells were counted for each experiment, and data are shown as
merged images from NPM and γ-tubulin staining are shown: normal cells mean ± s.d. of three separate experiments. (e) Cell-cycle profiles of serum-
with a single centrosome representing the G0/G1 phase; an interphase starved cells in d were determined by FACS analysis. (f) Effect of LMB on
cell with two adjacent centrosomes representing the S/G2 phase; and binding of NPM to mitotic centrosomes. Mitotic HCA2-hTERT cells were
a mitotic cell with two distanced centrosomes representing M phase. collected by shaking and were reseeded in chamber slides in the absence or
Magnification: ×630. (b) The percentage of interphase cells from serum- presence of 50 nM LMB for 2 h. Representative images from control or LMB-
starved HCA2-hTERT cells with NPM-bound centrosomes or mitotic cells treated mitotic cells, as analysed by anti-NPM (green) or anti-γ-tubulin (red)
from unsynchronized cells with or without the expression of GFP–NPM or antibodies are shown. Arrows indicate centrosomes. Magnification: ×630.
GFP–NPM-NESM was quantified. At least 100 cells were counted for each (g, h) The percentage of mitotic cells that contain detectable NPM signal on
cell-cycle phase. Data are averages of three independent experiments, shown centrosomes (g) or contain more than two centrosomes for each mitotic cell
as mean ± s.d. (c) NPM association with centrosomes in serum-starved (h) was quantified. At least 100 cells were analysed. Data are an average of
HCA2-hTERT cells (G0/G1 phase) was quantified in the presence or absence three separate experiments, shown as mean ± s.d.

containing both murine and human nuclei (Fig. 1b; upper panels). No GFP-positive human nuclei when cells were treated with LMB (Fig. 1b,
human nuclei were positive for GFP–NPM signals when unfused or fused c). These results indicate that NPM can shuttle between the nucleus and
to each other (data not shown). As a control, human nucleolin was not cytoplasm, and the process may depend on a Ran–Crm1 pathway that is
shuttled from human nuclei to murine nuclei for up to 16 h after cell fusion mediated through its NES motif.
(see Supplementary Information, Fig. S1). About 60% of GFP–NPM-posi- To determine whether NPM nucleocytoplasmic shuttling is directed
tive heterokaryons contained GFP-positive human nuclei (Fig. 1c). In by its NES, NLS and NoLS motifs, GFP–NPM mutants were constructed
contrast, less than 10% of GFP–NPM-positive heterokaryons contained by site-directed mutagenesis (Fig. 1d). Wild-type GFP–NPM was

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005 825

©2005 Nature Publishing Group


LETTERS
a b
predominantly localized in the nucleolus (Fig. 1e). The NES missense
Sucrose gradient fraction
mutant at residues L100A and L102A (NESM) or the deletion mutant 200

concentration (µg ml−1)


3 4 5 6 7 8 9 10 11
at residues 94–102 (NESD) was largely distributed in the nucleoplasm
γ-tubulin 150

Total protein
(Fig. 1e). These mutant proteins were also found in the nucleolus of cells
pretreated with Triton X-100 to remove soluble NPM signals (data not 100
NPM
shown). These results imply that the NESM and NESD mutant proteins 50
lack the ability to export NPM to the cytoplasm and thus accumulate Crm1 0
in large excess in the nucleus. Mutations at the NLS motif abolished 0 5 10 15 20 25 30
Bottom Top Fraction
localization in the nucleoplasm, as positive staining was mainly found
in the cytoplasm and nucleolus (Fig. 1e). Consistent with a previous LMB (nM)
c
report21, a NoLS mutant at W288G and W290G (NoLSM) was mainly 0 0.2 2 20
found in the nucleus (Fig. 1e), but no longer localized to the nucleolus γ-tubulin
upon Triton pretreatment (data not shown). Consistently, both NESM 100 87 97 74
and NESD mutants lost their ability to shuttle from murine to human NPM
nuclei, whereas both wild type and NoLSM still retained these activities
100 47 23 21
(data not shown). In contrast, NESM-NLS1D was diffusely distributed in
Crm1
both the cytoplasm and the nucleus (Fig. 1e). It appeared that this mutant
100 140 149 79
was not preferentially localized in a specific subcellular compartment
when both import and export motifs were mutated. Taken together, Figure 3 Centrosome-enriched fractions contain NPM, which is sensitive
these results indicate that the NES, NLS and NoLS motifs are functional to the nuclear export inhibitor, LMB. (a) Various fractions obtained from
in directing NPM subcellular localization. discontinuous sucrose gradient centrifugation were analysed by western
blotting with anti-γ-tubulin (top panel), anti-NPM (middle panel) or anti-
Because Crm1 and NPM were recently found to be associated with
Crm1 (bottom panel). Corresponding fraction numbers are indicated.
centrosomes6,15,18, we determined whether their localization required (b) Total protein concentrations of each fraction were determined. (c)
a functional NES motif. Consistent with previous reports15,18, syn- Association of NPM with centrosomes was inhibited by LMB. An aliquot of
chronized G0/G1 cells contained a single centrosome, and over 80% 20 µl of fraction 8 was incubated with various doses of LMB for 30 min,
followed by centrifugation to obtain a centrosome pellet. The pellets were
of these cells also showed a detectable NPM signal on centrosomes
analysed by western blotting with anti-γ-tubulin (top panel), anti-NPM
(Fig. 2a, b). A centrosome-associated NPM signal was absent in inter- (middle panel) and anti-Crm1 (bottom panel). Percentages of the NPM,
phase cells that contained two centrosomes (representing cells in S/G2 γ-tubulin and Crm1 signal intensities, normalized to their untreated
phase) from a non-synchronized population (Fig. 2a, middle panels; controls, are indicated below each blot.
Fig. 2b). In contrast, 100% of mitotic centrosomes showed a detect-
able NPM signal (Fig. 2a, lower panels; Fig. 2b). Similar results were To further examine whether Crm1 and NPM are associated with cen-
obtained in cells expressing GFP–NPM, but not GFP–NPM-NESM trosomes, we enriched centrosomes from quiescent cells by discontinu-
(Fig. 2b). However, greater than 98% of the cells at either interphase ous sucrose gradient fractionation22. Consistent with previous reports15,22,
(containing one or two centrosomes) or mitosis (two centrosomes) fractions 7–9 from HeLa cells, corresponding to 60% (w/w) sucrose, con-
showed detectable GFP–Crm1 on centrosomes (see Supplementary tained the peak signal of γ-tubulin, NPM and Crm1 (Fig. 3a), whereas
Information, Fig. S2A, B). the level of soluble cellular proteins was very low in these fractions
To determine whether regulation of centrosome duplication is (Fig. 3b). On the basis of immunofluorescence staining, only fractions
dependent on NES-mediated NPM transport, serum-starved cells were 7–9 contained centrosomes (see Supplementary Information, Fig. S3B).
either treated with LMB or nucleofected with RanBP1 or HBx. As shown The isolated centrosomes in these fractions retained NPM and Crm1
in Fig. 2c, RanBP1, HBx and LMB completely blocked NPM transport (see Supplementary Information, Fig. S3C); however, no other nuclear
to centrosomes. LMB treatment resulted in a significant decrease in proteins such as nucleolin and a subunit of Ku antigen with relative
the number of cells with a single centrosome (from 90% to 30%), and molecular mass 86,000 (Mr 86K) were detected (see Supplementary
an increase in the number of cells (from 5% to 70%) with more than Information, Fig. S3A). Similar results were obtained from HCA2-
one centrosome (Fig. 2d). Similar results were obtained when another hTERT cells (see Supplementary Information, Fig. S3). Whereas these
bona fide centrosome marker, anti-centrin antibody, was used (see studies do not prove that isolated centrosomes are absolutely free from
Supplementary Information, Fig. S2C, D). In contrast, most of the cells contamination by other cellular proteins, they indicate that fraction 8
treated with LMB still remained in G0/G1, as analysed by fluorescence- is highly centrosome-enriched and that Crm1 and NPM co-separate
activated cell sorting (FACS) (Fig. 2e). Next, we determined whether in this fraction. Following treatment of fraction 8 with LMB, a dose-
LMB acts directly to regulate NPM localization to the centrosome or dependent reduction of the NPM signal was observed (from 100%
indirectly through its effect on nucleocytoplasmic transport. In mitotic to 21%), whereas the levels of Crm1 and γ-tubulin remained similar.
cells, LMB consistently and effectively blocked NPM localization to Taken together, these results suggest that NPM binding to centrosomes
centrosomes, with an increased frequency of mitotic cells containing depends on its NES and is mediated by Crm1. Consistently, an anti-NPM
more than two centrosomes (Fig. 2f–h). Taken together, these results antibody, but not a control mouse IgG, co-immunoprecipitated with
suggest that disruption of Crm1 function by LMB, RanBP1 or HBx leads Crm1 (see Supplementary Information, Fig. S3F).
to NPM dissociation from centrosomes and a subsequent initiation of The results above indicate that LMB-mediated NPM dissociation
premature centrosome duplication. from centrosomes in quiescent cells (G0/G1) may result in unscheduled

826 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005

©2005 Nature Publishing Group


LETTERS
A

a b c

1 1 1

2 2 2

d e f g

B C
NA

40
NA

NA
siR

siR

siR
ol

ol
ntr

ntr

centrosomes (n>2) (%)


NP

NP

30
Co

Co

Cells with abnormal

NPM 20
100 107 68 18

10
β-actin
0
A2
l

A
tro

RN
iR
on

ls
C

si
tro

PM
on

N
C

Figure 4 Deficiency of NPM by RNA interference is associated with centrosomes. (B) Western blot analysis was used to determine the level of
unscheduled centrosome duplication and abnormal spindles. (A) HCA2- NPM in HCA2-hTERT cells transiently transfected with either a control siRNA
hTERT cells transiently transfected with the NPM siRNA2 oligonucleotides oligonucleotide or two NPM siRNA oligonucleotides for 120 h. The same blot
were co-immunostained either with anti-γ-tubulin (red) and anti-NPM was probed with anti-β-actin antibody as a loading control. Percentages of the
(green) (a–c), or with anti-α-tubulin (red) and anti-γ-tubulin (green) (d–g). NPM signal intensities, normalized to their corresponding β-actin signals and
Nuclei were stained with DAPI (blue) (c, f). Representative interphase cells untreated control, are indicated. (C) The percentage of cells with more than
(a–c) or a mitotic cell (d–g) are shown. Magnification: ×1,000. Panel g is a two centrosomes was determined from HCA2-hTERT cells transfected with
merged image of panels d–f. The inserts 1 and 2 in a are ×5 magnifications or without either control oligonucleotide or NPM siRNA2 oligonucleotides for
of the two areas with centrosomes, as indicated by arrows 1 and 2. One cell 120 h. In the NPM siRNA2 group, only the cells with a diminished expression
(arrow 1) that has a diminished NPM signal contains multiple centrosomes of NPM were evaluated. At least 300 cells were evaluated for each group.
whereas the other cell (arrow 2) with normal NPM expression contains two Data are averages of three independent experiments, shown as mean ± s.d.

centrosome duplication. To further examine the physiological effects of supernumerary centrosomes (more than two centrosomes in each cell)
NPM on centrosome duplication, we used NPM siRNA to knock down were often observed in NPM-siRNA2-transfected cells that were nega-
the expression of endogenous NPM. NPM siRNA2 and to a lesser degree tive for NPM staining (Fig. 4A, a–c; and see Supplementary Information,
NPM siRNA1, but not control siRNA, inhibited NPM expression (Fig. 4A, Fig. S4A). Approximately 30% of NPM-knocked-down cells contained
B). Whereas control cells always contained one or two centrosomes, supernumerary centrosomes (Fig. 4C). Moreover, amplified centrosomes

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005 827

©2005 Nature Publishing Group


LETTERS
a b Fig. S4C). These results indicate that the NPM NES motif is required
for NPM-mediated suppression of premature centrosome duplication,

or >2 centrosomes (%)


100 n=1 n=2 n >2

Cells with n = 1, 2
GFP−NPM which is independent of the essential function of NPM in cell growth.
Endogenous 75
A motif scanning analysis indicates that the threonine residue (T95)
NPM 50
β-actin
within the NES motif of NPM is a potential proline-dependent kinase
25 phosphorylation site. It is possible that phosphorylation at T95 may
FP

Co P−N M
ol S D
NP siR siRN iRNA
NP NPM PM trol

NA
0 alter the NES properties, thereby preventing its binding to Crm1 and
P
siR A + G + G

E
siR
n

+ G P−N

D
NA
NA

FP PM
ol
Co

P
s

ES
docking on centrosomes. To test this hypothesis, we made two addi-

GF
ntr
A

siR
siR

+ G P−N
F

−N
ntr
F

Co

A+
tional mutants of GFP–NPM in which the T95 residue was either
N

ol
M

F
iRN

ntr
NP

siR A + G
NA
N

changed to alanine (mimicking non-phosphorylation) or aspartic

Co
s

NA
M

N
acid (mimicking phosphorylation at T95). Whereas the GFP-T95A
M

NP

NP siR
M

M
mutant was still able to suppress premature centrosome duplication

M
NP
induced by NPM siRNA, the GFP-T95D mutant partially lost this
ability (Fig. 5c). Consistently, GFP-T95A, but not GFP-T95D, was
efficiently associated with centrosomes (Fig. 5d).
GFP−NPM centrosomes (%)
centrosomes (n>2) (%)

c d
Cells with abnormal

20
100 We recently suggested that the Ran–Crm1 complex may have a role in
15
regulating centrosome duplication because inactivation of Crm1 either
Cells with

75
10 by expression of RanBP1, HBx or by LMB leads to centrosome amplifica-
50
5 tion and multipolar spindles6,8. The effect of HBx on Crm1 is dependent
25
0 on the presence of a NES motif6,13. Consistently, Ran, Crm1 and RanBP1
NA FP FP
NA FP PM

5D
NP siR A + G A + G A

FP 5A
NP NP NP l siRN l
NP siR siR siR A

0
ro

Interphase Mitosis are localized on centrosomes6,7,11. Moreover, alteration of Ran network


ntr ont

+ G -T9
-T9
siR + G −N

(n = 1) (n = 2)
C

components, including Ran and RanBP1, results in multipolar spindles


M
o

GFP−NPM
or mitotic centriole splitting, an event leading to unscheduled initiation
Co

GFP-T95A
M
M N

GFP-T95D of centrosome duplication8. These studies led us to hypothesize that the


M
M

Ran–Crm1 network may function by regulating the local concentrations


of NES-containing cellular factors near centrosomes to ensure the forma-
tion of a bipolar spindle. This is analogous to the Ran–importins–NLS
Figure 5 The NPM NES motif is required for NPM-mediated suppression
of centrosome duplication. (a) HCA2-hTERT cells were transfected with or protein network for stimulating microtubule nucleation during spindle
without NPM siRNA2 together with various NPM expression vectors, and were assembly9. In this study, we have used both genetic and biochemical
analysed by western blot with anti-NPM and anti-β-actin antibodies. (b) In approaches to demonstrate that NPM contains a functional NES, acts as
parallel, the above cells were examined by indirect immunofluorescence with
a Crm1substrate, and that its local transport is mediated by Ran–Crm1.
anti-γ-tubulin antibody. The percentage of cells that contain one (n = 1), two
(n = 2) or more than two (n > 2) centrosomes was determined. Data are an Furthermore, Crm1 and NPM colocalize on centrosomes, and NPM
average of three independent experiments, shown as mean ± s.d. (c, d) HCA2- co-immunoprecipitates with Crm1. Thus, the Ran–Crm1 complex may
hTERT cells were cotransfected with NPM siRNA and GFP-T95A or GFP-T95D be associated with NPM at centrosomes through its NES motif, thereby
mutants. The percentage of cells with abnormal centrosomes (greater than
providing a novel mechanistic insight regarding how NPM is regulated
two centrosomes per cell) was quantified in 5 days after transfection (c) and
the percentage of interphase (single centrosome cells) and mitotic cells with to control centrosome duplication.
centrosomes that had a detectable GFP–NPM signal was quantified in 14 days NPM contains multiple phosphorylation sites and its function may
after transfection (d). Only the GFP-positive cells were analysed. Data are an be regulated by phosphorylation. NPM may be a CDK2–cyclin E target
average of three independent experiments, shown as mean ± s.d.
through phosphorylation sites at residues 186–239, triggering initiation
of both centrosome duplication and DNA replication15. In the present
could be nucleated in mitotic cells expressing NPM siRNA2 (Fig. 4A, d–g), work, we have identified another potential phosphorylation site within
suggesting that supernumerary centrosomes could function in forming the NES motif upstream of the proposed CDK2–cyclin E domain, which
spindle poles. Similar results were obtained with cells stained for a centriole- also appears to be involved in regulating the function of NPM on centro-
specific anti-centrin antibody (see Supplementary Information, Fig. S4B). some duplication. Consistently, the T95D mutant had a decreased abil-
Thus, loss of NPM is associated with centrosome reduplication. ity to bind to centrosomes and was inactive in suppressing centrosome
To further determine the requirement of the NPM NES motif in the duplication. Thus, it is possible that phosphorylation at T95 may regulate
suppression of centrosome duplication, we performed a rescue experi- NPM binding to Crm1.
ment in HCA2-hTERT cells in which the expression of endogenous On the basis of our results, we hypothesize that the Ran–Crm1 complex
NPM was blocked by NPM siRNA2 and replaced with a recombinant may regulate NES-containing substrates at centrosomes, and this proc-
GFP-NPM gene (Fig. 5). Consistently, NPM siRNA resulted in supernu- ess may be regulated by phosphorylation. Such a mechanism to shuttle
merary centrosomes (Fig. 5b). This activity was effectively inhibited by proteins to the sites of action is particularly efficient for processes such
coexpressing GFP–NPM, but not GFP–NESD (Fig. 5b). Interestingly, as the initiation of centrosome duplication and entry to the cell cycle,
silencing endogenous NPM by NPM siRNA resulted in decreased cell and probably also acts as a switch to synchronize these two processes.
growth (see Supplementary Information, Fig. S4C), suggesting that NPM Although our results have not ruled out the possibility that other NES-
may be an essential gene. However, both GFP–NPM and GFP–NESD containing proteins participate in this process, we suggest that NPM is an
can rescue the slow growth phenotype (see Supplementary Information, important substrate. Thus, the following model can be envisaged: during

828 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005

©2005 Nature Publishing Group


LETTERS

cell progression from S to G2 phase, a major portion of NPM is located in Cells and transfection. NIH 3T3, HCA2-hTERT and HeLa cells were main-
the nucleolus, mainly functioning as a regulator of ribosome biogenesis tained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10%
and cell proliferation23 due to its strong NLS activity. After the nuclear fetal bovine serum, penicillin/streptomycin (1×; 100 U ml−1, 100 µg ml−1), and
200 mM glutamine. GFP–NPM and GFP–NPM mutant plasmids, GFP–Crm1,
membrane breaks down at the beginning of mitosis, NPM reassociates GFP–RanBP1 and pcDNA3-Hbxadr-Hatag13 were transfected into NIH 3T3 or
with mitotic centrosomes through its interaction with the Ran–Crm1 HCA2-hTERT cells using the Nucleofection protocol according to the manu-
complex, preventing centrosome reduplication. Thus, a newly divided G1 facturer’s instructions (Amaxa Biosystems, Gaithersburg, MD). NPM siRNA1
cell contains a NPM-bound centrosome. NPM is then phosphorylated oligonucleotides (target sequence 5′-ccgtcttatttcatttctgta-3′) (Qiagen, Valencia,
and dissociates from the G1 centrosome to initiate centrosome duplica- CA), NPM siRNA2 oligonucleotides (target sequence 5′-atggaatgttatgataggaca-3′)
(Qiagen), and luciferase siRNA oligonucleotides (target sequence 5′-cgtacgcg-
tion and DNA replication. Thus, the Ran network may regulate cell-cycle gaatacttcga-3′) (Dharmacon, Lafayette, CO) were transfected into HCA2-hTERT
progression to ensure faithful cell division24. It is possible that Ran–Crm1 cells with TransIT-TKO transfection reagent (Mirus Corporation, Madison,
may be utilized by other NES-containing proteins such as p53, whose WI) according to the manufacturer’s instructions. The NPM siRNA oligonucle-
stability can be regulated by NPM25,26, or HDM2, a protein containing a otides correspond to the sequences localized in the 3′ UTR of the NPM gene.
functional NES motif20. The oncogenic viral protein HBx binds and inac- Cotransfection of GFP–NPM or GFP–NPM mutants with NPM siRNA was done
using the Nucleofection protocol.
tivates the Ran–Crm1-dependent network6,13, thus disrupting the function
of this complex. This may illustrate a mechanism by which HBx induces Heterokaryon assays. Nucleocytoplasmic shuttling was detected using a heter-
abnormal centrosome duplication and abnormal mitotic spindles, thereby okaryon assay as previously described19,20. Immediately after NIH 3T3 cells in
contributing to HBV-mediated hepatocarcinogenesis27. suspension were nucleofected with GFP–NPM and its mutants, the cells were
seeded on 2-well chambers together with HCA2-hTERT cells. After 20 h, the cells
were treated with 50 µg ml−1 cycloheximide (Sigma, St Louis, MO) in cell culture
METHODS medium for 20 min at 37 °C, and then cell fusion was induced with 50% (w/v)
Plasmid construction. For expression of GFP–NPM, the plasmid pEGFP- polyethylene glycol 8000 (Sigma) in serum-free DMEM for 3 min at 37 °C. The
C2FlagB23 (ref. 28) was kindly provided by K. Nagata (University of Tsukuba, cells were cultured again in cell culture medium containing 50 µg ml−1 cyclohex-
Japan). To generate a plasmid expressing GFP–NPM NESM, pEGFP-C2FlagB23 imide with or without 40 nM of LMB for 2 h, followed by fixing and immunos-
was used as template for two-step overlapping PCR. The following primer pairs taining. Similarly, NIH 3T3 cells were nucleofected with GFP–PML, fused to
were used for the first rounds of PCR: (a) 5′-acacttcgcccttgcgaccactggtggtg-3′ HCA2-hTERT cells, and cultured for up to 16 h in culture medium containing
and (b) 5′-tcgaattctgcagtcgac-3′; and (c) 5′-gtcgcaagggcgaagtgtggttcagggccag- 50 µg ml−1 of cycloheximide, followed by fixing and immunostaining.
3′ and (d) 5′-tgatcagttatctagatcc-3′. The products of these reactions were then
purified, annealed and used as a template for a final PCR using primers (b) Cell-cycle analysis. Serum-starved HCA2-hTERT cells used in NPM centro-
and (d). The final PCR product was digested with BamHI and subcloned into somal localization assays were also evaluated by cell-cycle analysis. Cells were
the pEGFP-C2FlagB23 cut with BamHI. Similarly, GFP-T95A was generated trypsinized, washed twice with phosphate-buffered saline, pelleted, and fixed
using two-step overlapping PCR with primer pairs (b) and 5′-taagaccactggt- with 70% ethanol overnight at 4 °C. Samples were pelleted, washed once with
ggtgctatttcaaag-3′, and (d) and 5′-tagcaccaccagtggtcttaag-3′ for first rounds PBS, treated with 10 µg ml−1 RNaseA and stained with 50 µg ml−1 of propidium
of PCR, and primers (b) and (d) for final PCR. The final PCR product was iodide for 30 min at 37 °C, and then subjected to FACScan analysis.
digested with BamHI and subcloned into pEGFP-C2FlagB23 at the BamHI site.
GFP–NPMT95D was generated with primer pairs (b) and 5′-taagaccactggtgggtc- Cell proliferation assay. GFP–NPM plasmid and NPM siRNA co-transfected
tatttcaaag-3′, and (d) and 5′-tagacccaccagtggtcttaag-3′ for first rounds of PCR, HCA2-hTERT cells were frozen at −70 °C at different time points. The frozen cells
and primers (b) and (d) for final PCR. The final PCR product was digested with were thawed and subjected to a cell proliferation assay using the CyQUANT assay
BamHI and subcloned into the pEGFP-C2FlagB23 cut with BamHI. To gener- according to the manufacturer’s instructions (Molecular Probes, Eugene, OR).
ate the GFP–NPM NESD expression plasmid, PCR fragments from pEGFP-
C2FlagB23 with primers (b) and 5′-tggtgttaagcttcaaagcccccaag-3′ cut with Centrosome isolation. Centrosomes were prepared from quiescent HeLa and
KpnI and HindIII, and PCR fragments from pEGFP-C2FlagB23 with primers HCA2-hTERT cells as described22. In brief, HeLa or HCA2-hTERT cells were
(d) and 5′-ttaaggtaagcttgtggttcagggccag-3′ cut with BamHI and HindIII were cultured for 2 days, followed by serum starvation for 24 h. Cells were pelleted,
ligated into the pEGFP-C2FlagB23 digested with BamHI and KpnI. To gener- washed with TBS and TBS 1/10–8% sucrose buffer, lysed and subjected to cen-
ate the GFP–NPM NLS1D expression plasmid, PCR fragments from pEGFP- trifugation at 2,500g for 10 min to remove cell debris, nuclei and chromatin. The
C2FlagB23 with primers (b) and 5′-ggcagaaagcttcacatcctcctcctcttc-3′ cut with supernatant was filtered with a nylon mesh and treated with 2 U ml−1 DNase I,
PstI and HindIII, and PCR fragments from pEGFP-C2FlagB23 with primers followed by centrifugation with a 60% sucrose cushion at the bottom of centrifu-
(d) and 5′-gatgtgaagctttctgcccctggaggtggt-3′ digested with BamHI and HindIII gation tube for 30 min at 11,000g. The centrosomes in the sucrose cushion were
were ligated into the pEGFP-C2FlagB23 cut with BamHI and PstI. To gener- diluted with lysis buffer and subjected to a discontinuous gradient centrifugation
ate the GFP–NPM NLS2D expression plasmid, PCR fragments from pEGFP- consisting of 2.5 ml of 70%, 1.5 ml of 50% and 1.5 ml of 40% sucrose solution at
C2FlagB23 with primers (b) and 5′-agcagcaagctttacgctaccacctccaggggcagc-3′ 85,000g for 1.5 h with a SW40Ti rotor. Centrosome fractions (300 µl per fraction)
cut with PstI and HindIII, and PCR fragments from pEGFP-C2FlagB23 with were collected from the bottom of a centrifugation tube. Protein concentrations
primers (d) and 5′-aaagtaaagcttgctgctgatg-3′ cut with BamHI and HindIII of each fraction were determined by the Bio-Rad protein assay dye reagent (Bio-
were ligated into the pEGFP-C2FlagB23 cut with BamHI and PstI. To gener- Rad, Hercules, CA) according to the manufacturer’s instructions. The fractions
ate the GFP–NPM NoLSM expression plasmid, PCR fragments from pEGFP- containing centrosomes were determined by immunoblot analysis for γ-tubulin,
C2FlagB23 with primers (b) and 5′-tccggtggatccttaaagagacttcctcccctgcccgag-3′ and immunofluoescence for γ-tubulin and α-tubulin as described22. The aliquots
cut with BamHI were subcloned into the pEGFP-C2FlagB23 cut with BamHI. from the centrosome fractions were diluted with 1 ml of 10 mM PIPES buffer
To generate the GFP–NPM NESM-NLS1D expression plasmid, PCR fragments (pH 7.2) and subjected to LMB (0.2, 2 or 20 nM) treatment for 30 min at room
from pGFP-NPM NLS1D plasmid with primer pairs (a) and (b), and (c) and (d) temperature, followed by centrifugation at 20,000g for 10 min. Pellets were washed
were purified, annealed and used as a template for a final PCR using primers three times in 10 mM PIPES. Proteins associated with LMB-treated centrosomes
(b) and (d). The purified product was cut with BamHI and subcloned into the were then detected by immunoblot analysis.
pEGFP-C2FlagB23 digested with BamHI. To generate the GFP–NPM NLS1/2D
expression plasmid, a fragment of GFP–NPM NLS1D digested by KpnI and Immunoblot analysis. HCA2-hTERT cells were homogenized in NP40 lysis
HindIII and a fragment of GFP–NPM NLS2D cut by BamHI and HindIII were buffer (1% NP40, 25 mM Tris pH 7.5, 150 mM NaCl, 2 mM EDTA, 1 mM
ligated in the pEGFP-C2FlagB23 plasmid digested with BamHI and KpnI. sodium orthoVanadate, 100 µM GTPγS and protease inhibitors), and subjected to

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005 829

©2005 Nature Publishing Group


LETTERS
immunoprecipitation with mouse monoclonal anti-NPM (FC-61991; Zymed, San Received 17 February 2005; accepted 22 June 2005
Francisco, CA). After washing protein A/G beads, immunoprecipitation products Published online at http://www.nature.com/naturecellbiology.
were detected by immunoblot analysis. Immunoblot analysis was performed as 1. Sluder, G. & Hinchcliffe, E. H. Control of centrosome reproduction: the right number
previously described6. Samples were resolved on SDS–PAGE and transferred to at the right time. Biol. Cell 91, 413–427 (1999).
nitrocellulose membranes (Invitrogen, Carlsbad, CA). The blots were blocked in 2. Brinkley, B. R. Managing the centrosome numbers game: from chaos to stability in
cancer cell division. Trends Cell Biol. 11, 18–21 (2001).
5% (w/v) dry skimmed milk in TBST buffer for 1 h at room temperature, probed
3. Nigg, E. A. Centrosome aberrations: cause or consequence of cancer progression?
with appropriate antibodies; that is, mouse monoclonal anti-nucleolin antibody Nature Rev. Cancer 2, 815–825 (2002).
(4E2; Medical & Biological Laboratories Co., Nagoya, Japan), mouse monoclonal 4. Doxsey, S. Duplicating dangerously: linking centrosome duplication and aneuploidy.
anti-NPM (FC-61991, Zymed), mouse monoclonal anti-γ-tubulin (GTU-88, Mol. Cell 10, 439–440 (2002).
Sigma), mouse monoclonal anti-Crm1 (17; BD Transduction Laboratories, San 5. Lingle, W. L. & Salisbury, J. L. The role of the centrosome in the development of
malignant tumors. Curr. Top. Dev. Biol. 49, 313–329 (2000).
Diego, CA), mouse monoclonal anti-β-actin (AC-15, Sigma) and mouse mono- 6. Forgues, M. et al. Involvement of Crm1 in hepatitis B virus X protein-induced aberrant
clonal anti-Ku antigen (Ku15, Sigma), and then detected by chemiluminescence centriole replication and abnormal mitotic spindles. Mol. Cell Biol. 23, 5282–5292
ECL (Amersham, Piscataway, NJ) or supersignal west femto chemiluminescence (2003).
ECL (Pierce, Rockford, IL). 7. Keryer, G. et al. Part of Ran is associated with AKAP450 at the centrosome: involvement
in microtubule-organizing activity. Mol. Biol Cell 14, 4260–4271 (2003).
8. Di Fiore, B., Ciciarello, M. & Lavia, P. Mitotic functions of the Ran GTPase network: the
Indirect immunofluorescence assay and confocal analysis. Cell samples on importance of being in the right place at the right time. Cell Cycle 3, 305–313 (2004).
4-well or 2-well chamber slides were fixed and stained as previously described6. 9. Dasso, M. Running on Ran: nuclear transport and the mitotic spindle. Cell 104, 321–
Samples were blocked with 10% normal donkey serum for 1 h at room tempera- 324 (2001).
10. Weis, K. Regulating access to the genome: nucleocytoplasmic transport throughout the
ture and stained with various primary antibodies for 1 h at 37 °C, followed by
cell cycle. Cell 112, 441–451 (2003).
either Alexa 488 FITC-conjugated anti-mouse/rabbit IgG, or Alexa 568 Texas- 11. Di Fiore, B. et al. Mammalian RanBP1 regulates centrosome cohesion during mitosis.
Red-conjugated anti-mouse/rabbit IgG (Molecular Probes), or AMCA-conjugated J. Cell Sci. 116, 3399–3411 (2003).
anti-mouse/rabbit antibodies (Jackson Labs, West Grove, PA). Mouse monoclonal 12. Kudo, N. et al. Leptomycin B inactivates CRM1/exportin 1 by covalent modification
at a cysteine residue in the central conserved region. Proc. Natl Acad. Sci. USA 96,
anti-nucleolin antibody (4E2, Medical & Biological Laboratories Co.) was used
9112–9117 (1999).
to identify human nuclei. Mouse monoclonal anti-NPM (FC-61991, Zymed) or 13. Forgues, M. et al. Interaction of the hepatitis b virus x protein with the Crm1-dependent
rabbit polyclonal anti-NPM antibodies (Santa Cruz Biotechnology, Santa Cruz, nuclear export pathway. J. Biol. Chem. 276, 22797–22803 (2001).
CA) were used to detect NPM expression and localization. Mouse monoclonal 14. Tarapore, P., Okuda, M. & Fukasawa, K. A mammalian in vitro centriole duplication
system: evidence for involvement of CDK2/cyclin E and nucleophosmin/B23 in centro-
Crm1 antibody (17, BD Transduction Laboratories) was used to determine its
some duplication. Cell Cycle 1, 75–81 (2002).
subcellular localization. Mouse monoclonal anti-γ-tubulin (GTU-88, Sigma), rab- 15. Okuda, M. et al. Nucleophosmin/B23 is a target of CDK2/cyclin E in centrosome
bit polyclonal anti-γ-tubulin (Sigma), Rhodamine-conjugated goat polyclonal duplication. Cell 103, 127–140 (2000).
anti-γ-tubulin antibodies (Santa Cruz Biotechnology), or mouse monoclonal 16. Borer, R. A., Lehner, C. F., Eppenberger, H. M. & Nigg, E. A. Major nucleolar proteins
shuttle between nucleus and cytoplasm. Cell 56, 379–390 (1989).
anti-centrin antibody (kindly provided by J. Salisbury, Mayo Clinic), were used
17. Tokuyama, Y., Horn, H. F., Kawamura, K., Tarapore, P. & Fukasawa, K. Specific phos-
to detect centrosomes. Mouse monoclonal anti-α-tubulin antibody (clone B-5- phorylation of nucleophosmin on Thr(199) by cyclin-dependent kinase 2-cyclin E and
1-2, Sigma) was used to detect spindles. To determine localization of NPM at its role in centrosome duplication. J. Biol. Chem. 276, 21529–21537 (2001).
centrosomes, cells were pre-extracted with 0.1% Triton X-100 and immediately 18. Zatsepina, O. V. et al. The nucleolar phosphoprotein B23 redistributes in part to the
spindle poles during mitosis. J. Cell Sci. 112, 455–466 (1999).
fixed as described above. Nuclei were stained with 4,6-diamidino-2-phenylindole
19. Pinol-Roma, S. & Dreyfuss, G. Shuttling of pre-mRNA binding proteins between nucleus
(DAPI). Determination of cellular proteins associated with isolated centrosomes and cytoplasm. Nature 355, 730–732 (1992).
was performed as previously described22. Conventional or confocal fluorescence 20. Roth, J., Dobbelstein, M., Freedman, D. A., Shenk, T. & Levine, A. J. Nucleo-cytoplas-
microscopic analysis was done essentially as previously described13. For deter- mic shuttling of the hdm2 oncoprotein regulates the levels of the p53 protein via a
pathway used by the human immunodeficiency virus rev protein. EMBO J. 17, 554–564
mining localization of NPM on mitotic centrosomes, HCA2-hTERT cells and
(1998).
HCA2-hTERT cells transfected with GFP–NPM plasmid were cultured for 2 days. 21. Nishimura, Y., Ohkubo, T., Furuichi, Y. & Umekawa, H. Tryptophans 286 and 288 in the
Mitotic cells were collected by shaking, which were cultured for up to 2 h in cell C-terminal region of protein B23.1 are important for its nucleolar localization. Biosci.
culture medium with or without 50 nM of LMB. Biotechnol. Biochem. 66, 2239–2242 (2002).
22. Moudjou, M. & Bornens, M. Cell Biology: A Laboratory Handbook (ed. Celis, J. E.)
111–119 (Academic Press, San Diego, 1998).
BIND identifier. One BIND identifier (www.bind.ca) is associated with this
23. Itahana, K. et al. Tumor suppressor ARF degrades B23, a nucleolar protein involved in
manuscript: 312465. ribosome biogenesis and cell proliferation. Mol. Cell 12, 1151–1164 (2003).
24. Clarke, P. R. & Zhang, C. Ran GTPase: a master regulator of nuclear structure and func-
Note: Supplementary Information is available on the Nature Cell Biology website. tion during the eukaryotic cell division cycle? Trends Cell Biol. 11, 366–371 (2001).
25. Stommel, J. M. et al. A leucine-rich nuclear export signal in the p53 tetramerization
ACKNOWLEDGEMENTS domain: regulation of subcellular localization and p53 activity by NES masking. EMBO
We thank K. Nagata for the generous gift of GFP-tagged NPM, P. Lavia for the J. 18, 1660–1672 (1999).
RanBP1 expression vector and J. Salisbury for anti-centrin antibody; C. Harris 26. Colombo, E., Marine, J. C., Danovi, D., Falini, B. & Pelicci, P. G. Nucleophosmin
regulates the stability and transcriptional activity of p53. Nature Cell Biol. 4, 529–533
for invaluable comments; S. Garfield, S. Wincovitch and B. J. Taylor for superior
(2002).
technical support; K. MacPherson for bibliographical help; and the National Cancer 27. Butel, J. S. Viral carcinogenesis: revelation of molecular mechanisms and etiology of
Institute-Center for Cancer Research Fellows Editorial Board for editorial assistance. human disease. Carcinogenesis 21, 405–426 (2000).
28. Okuwaki, M., Tsujimoto, M. & Nagata, K. The RNA binding activity of a ribosome biogen-
COMPETING FINANCIAL INTERESTS esis factor, nucleophosmin/B23, is modulated by phosphorylation with a cell cycle-depend-
The authors declare that they have no competing financial interests. ent kinase and by association with its subtype. Mol. Biol. Cell 13, 2016–2030 (2002).

830 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 8 | AUGUST 2005

©2005 Nature Publishing Group

You might also like