Neurobiology of Aging 22 (2001) 517–528

www.elsevier.com/locate/neuaging

Cdc2 phosphorylation of nucleolin demarcates mitotic stages and
Alzheimer’s disease pathology૾
Alex Dranovskya,b, Inez Vincentc, Luisa Gregorib, Alexander Schwarzmanb, David Colfleshd,
Jan Enghilde, Warren Strittmattere, Peter Daviesc, Dmitry Goldgaberb,*
a

Medical Scientist Training Program, State University of New York, Stony Brook, NY 11794, USA
Department of Psychiatry and Behavioral Science, State University of New York, Stony Brook, NY 11794, USA
c
Department of Pathology, Albert Einstein College of Medicine, Bronx, NY 10461, USA
d
University Microscopy Imaging Center, State University of New York, Stony Brook, NY 11794, USA
e
Department of Neurology, Duke University, Durham, NC 27710, USA

b

Received 28 August 2000; received in revised form 2 November 2000; accepted 3 November 2000

Abstract
Nucleolin is a major multifunctional nuclear phosphoprotein that is phosphorylated by Cdc2 kinase in mitosis and that participates in
a number of cellular processes. The monoclonal antibody TG-3 generated against neurofibrillary tangles (NFT) found in Alzheimer’s disease
(AD) is highly specific for mitotic cells in culture. We here demonstrate that phosphorylation of nucleolin by Cdc2 kinase generates the TG-3
epitope. The unique pool of TG-3 immunoreactive nucleolin appears abruptly during the prophase. It is associated with chromosomes
through the metaphase and it gradually disappears during separation of chromosomes and exit from mitosis. In the brain, nucleolin was
localized not only to nuclei but also to neuronal cytoplasm, and it is a marker for early NFT. In patients with AD, Cdc2 phosphorylated
nucleolin was present in NFT. These findings suggest that phosphorylation of nucleolin by Cdc2 kinase is a critical event and the point of
convergence of two distinct pathways, mitosis and neurodegeneration. © 2001 Elsevier Science Inc. All rights reserved.
Keywords: Nucleolin; Mitosis; Neurodegeneration; Alzheimer’s disease

1. Introduction
Several recent reports suggest that reactivation of certain
cell cycle events in postmitotic neurons may underlie neurofibrillary degeneration in AD [6,25,28,29,43,44]. Evidence for involvement of cell cycle specific proteins in
neurodegeneration stems from studies on the role of phosphorylation in the formation of neurofibrillary tangles
(NFT), one of the neuropathological hallmarks of AD. NFT
consist of paired helical filaments (PHF) and straight filaments. Phosphorylation by proline directed serine-threonine
kinases results in PHF properties in tau (a structural component of PHF) [31]. Experiments in vitro have shown that
many proline directed kinases phosphorylate recombinant
tau [1,27,31,46]. However, the biologic relevance of such
૾ This paper is part of the special issue ‘Cell Cycle’ (Guest Editor Inez
Vincent).
* Corresponding author. Tel.: ϩ1-631-444-1369; fax: ϩ1-631-4447534.
E-mail address: dgoldgaber@mail.psychiatry.sunysb.edu (D. Goldgaber).

experiments is not clear due to the substrate promiscuity of
these kinases [30] and the ability of tau to form PHF in the
absence of phosphorylation in vitro [48].
In vivo support for tau phosphorylation by mitotic proline directed kinases was provided when a kinase related to
the major mitotic kinase Cdc2 was immunolocalized to NFT
in the AD brain [26,50]. Moreover, Vincent and colleagues
recently demonstrated that Cdc2 kinase was activated in AD
neurons [43]. Furthermore, immunoreactivity of other cell
cycle markers was detected in the brains of patients affected
by AD [6,28 –29]. Aberrant activation of mitotic mechanisms in postmitotic neurons may facilitate cellular events
leading to NFT formation and neurodegeneration. Evidence
was obtained for the involvement of mitotic mechanisms in
NFT formation when monoclonal antibodies (TG/MC series) raised against immunopurified NFT showed specificity
for both the AD brain in immunohistochemical studies, and
for mitotic cells in culture [44]. The same study provided
provocative evidence that monoclonal antibody MPM-2,
which was raised against mitotic cells [10], was highly
immunoreactive with AD, but not with control brains.

0197-4580/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S 0 1 9 7 - 4 5 8 0 ( 0 0 ) 0 0 2 4 8 - 7

518

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

Moreover, MPM-2 recognized NFT, the neuritic components of plaques, and morphologically healthy neurons in
the AD brain, thus suggesting that it is a marker for the early
stages of neurodegeneration. Similar to the TG/MC series,
MPM-2 recognized mitotic phosphoepitopes [10,16,39].
The presence of mitotic phosphoepitopes in degenerating
neurons suggests that mitotic and neurodegenerative pathways converge in AD. Therefore, antibodies that demonstrate high specificity for both AD pathology and mitotic
events can serve as ideal tools for identifying common
molecules within the two pathways.
Of all the antibodies reported by Vincent and colleagues,
TG-3 was the most specific for AD neurofibrillary pathology demonstrating no detectable immunoreactivity in control brains. Therefore, identifying and characterizing the
TG-3 antigen or antigens are essential to our understanding
of the convergence of neurofibrillary degeneration and mitosis. The first TG-3 immunoreactive protein in the brains of
patients with AD was recently identified as an abnormally
phosphorylated microtubule associated protein tau [20].
Other potentially biologically relevant TG-3 antigens in the
brain remain unknown. However, in cultured cells synchronized in mitosis, TG-3 recognizes not tau but a 105 kDa
protein [44]. We here report that the cellular TG-3 antigen
is the major nucleolar phosphoprotein nucleolin after phosphorylation by Cdc2 kinase. We also report that in the
human brain, nucleolin is present in neuronal cytoplasm and
colocalizes with NFT in AD.

2. Materials and methods
2.1. Antibodies
Monoclonal antibody TG-3 (hybridoma medium) was
described in a previous publication [44] and used at a 1:10
dilution for Western blotting. Anti-nucleolin monoclonal
antibody MS3 (hybridoma medium) was kindly provided by
Drs. H. Busch and L. Perlaky from Baylor College of
Medicine, and used at a dilution of 1:10 for immunocytochemistry and at a 1:1000 dilution for Western blotting [42].
MS3 ascites fluid was used for immunoprecipitations at a
1:10 dilution. Monoclonal antibody D3 (hybridoma medium), kindly provided by Dr. J.-S. Deng from the University
of Pittsburgh, was used at a 1:10 dilution for immunocytochemistry [11]. Anti-nucleolin monoclonal antibody CC98
(hybridoma medium), kindly provided by Dr. N.H. Yeh,
National Yang Ming College, Taiwan, was used at a 1:5
dilution for immunocytochemistry [8]. Anti-nucleolin
monoclonal antibody 4E2 (hybridoma medium) from Research Diagnostics (Flanders, NJ) was used at a 1:5 dilution
for immunocytochemistry. Rabbit polyclonal antisera W15
generated against human nucleolin and nucleolin-maltosebinding protein fusion protein were generously provided by
Dr. N. Maizels, Yale University [17]. Each polyclonal antiserum was used at a 1:1000 dilution. Non-immune rabbit

serum was used as a control. Monoclonal antibody AT180
(Innogenetics, Belgium), which recognizes phospho-threonine 231 (Thr231) of tau [15], was used at a 1:10 dilution
for the blocking experiments. Monoclonal antibodies
against ␤-tubulin (Sigma, St. Louis, MO) and histone protein mAb052 (Chemicon, Temecula, CA) were used at a
1:200 final dilution. The specificity of each primary monoclonal and polyclonal antibody was characterized by the
authors who generated the antibody and was described in
the papers cited above. Secondary goat anti-mouse IgG1a
and goat anti-mouse IgM antibodies conjugated to fluorescein and rhodamine (Jackson Immunochemicals, West
Grove, PA) were used at a 1:200 final dilution for immunocytochemistry. Secondary antibodies conjugated to biotin
along with streptavidin conjugated to CY3 and CY5 fluorochromes (Southern Biotech, Birmingham, AL) were all
used at a 1:500 dilution for the immunohistochemistry of
brain sections. Secondary antibodies conjugated to horseradish peroxidase (Amersham, Oakville, Ontario) were used
at a 1:3000 final dilution for Western blotting.
2.2. Cell culture and immunocytochemistry
HEp-2 human epithelial cells (ATCC, Rockville, MD)
were cultured in DME containing 10% bovine calf serum,
and penicillin-streptomycin (Gibco, Burlington, Ontario)
was cultured at 37°C in 5% CO2. Cells were washed with
PBS, harvested by trypsinization, counted and plated at
1.25 ϫ 104 - 2.5 ϫ 104 cells/cm2 on 8 well chambered
coverslips (VWR, Piscataway, NJ). On the following day,
cells were fixed in 2% paraformaldehyde/0.1% glutaraldehyde for 40 min, permeabilized with 0.5% Triton X-100 for
10 min and stored in PBS at 4°C for immunocytochemistry.
Prior to immunocytochemistry, subconfluent monolayers
were blocked with 5% bovine serum albumin (BSA) in PBS
for 30 min and rinsed in 3 volumes of PBS 3 times followed
by three 10 min washes in PBS. Immunostaining of monolayers was performed in the chambered coverslips. Primary
antibodies were applied for 2 h at room temperature or
overnight at 4°C in a high humidity incubation chamber.
The monolayers were washed as described above and incubated with secondary antibodies in the presence of DAPI
(Sigma, St. Louis, MO) for 30 min. The monolayers were
washed as described above and stored in the ANTI-FADE
reagent (Molecular Probes, Eugene, OR) at 4°C. Rodamine.
Fluorescein was visualized by the Noran Confocal Odyssey
system through a Nikon inverted Diaphot microscope, and
images were captured using Image One software.
2.3. Protein purification
HEp-2 cells were grown in suspension according to the
conditions described above. Five liters of logarithmically
growing cells were synchronized in mitosis by treatment
with 2 mg/ml of nocodazole (Sigma, St. Louis, MO) over-

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

night. Nocodazole was diluted from a DMSO solution so
that the final DMSO concentration was 0.05%. Cells were
harvested by centrifugation at 250 ϫ g for 2 min, rinsed
with PBS, centrifuged, and cell pellets were stored at
Ϫ80°C. All extraction and purification procedures were
performed at 4°C. Cells were extracted in 5 volumes of
buffer A for 15 min, gently vortexed, and centrifuged at 400
ϫ g for 10 min. Buffer A contained 0.015 M Tris-Cl pH 7.4,
0.08 M KCl, 2 mM EDTA-KOH, 1% Nonidet P-40, 0.2 mM
spermine, 0.5 mM spermidine, 1 mM phenylmethane-sulphonyl fluoride, 10 nM Microsystin-LR, 1 mM Na-o-vanadate (all from Sigma), and protease inhibitor cocktail
(Boehringer Mannheim, Indianapolis, IN). The supernatant
was chromatographed on DEAE-Sepharose (Pharmacia,
Piscataway, NJ) preequilibrated with buffer A, washed, and
eluted with a 0.1– 0.5 M NaCl step gradient in buffer A
using 0.1 M increments. TG-3 reactive fractions were identified by Western blotting using the ECL detection system
(Amersham, Oakville Ontario). The TG-3 reactive fraction
was diluted in 15 mM Tris, 2 mM EDTA, pH 7.4 (buffer B),
chromatographed on HiTrap-Heparin mini column (Pharmacia, Piscataway, NJ) preequilibrated in buffer B, washed,
and eluted with a 0.1– 0.5 M NaCl step gradient using 0.1 M
increments. TG-3 immunoreactivity was determined by
Western blotting. The TG-3 reactive fraction was separated
by SDS-PAGE and stained with Coomassie Brilliant Blue.
The purified TG-3 reactive 105 kDa protein was subjected
to N-terminal sequencing and the obtained sequence was
compared by BLAST search against the Genbank database.

519

was lysed and analyzed by Western blotting with TG-3 as
described above.
2.6. Chemical cleavage
TG-3 positive nucleolin was purified from nocodazole
treated HEp-2 cells and subjected to SDS-PAGE. The TG-3
positive band was cut out of the gel, and the gel slice was
subjected to chemical digestion with N-chlorosuccinimide,
which specifically cleaves at tryptophane (positions 481 and
641 in nucleolin), or with hydroxylamine, which specifically cleaves between asparagine and glycine (positions
136, 137 and 519, 520 in nucleolin). N-chlorosuccinimide
was used at 10 ml of 0.1 M in 2 M urea, and 50% acetic acid
for 30 min at room temperature with rotation.
Hydroxylamine was used at 2 M in 6 M guanidine-HCl
at pH 9.0 for 12 h at room temperature with rotation. Gel
slices were washed with 15 mM Tris, pH 6.8, 0.1% SDS,
and 2-mercaptoethanol for 20 min. After chemical cleavage,
the gel slice was placed on top of a second SS-gel and
electrophoresed in a direction perpendicular with respect to
the first gel. The separated cleavage products were transferred onto PVDF membrane and analyzed by Western
blotting. The membranes were first immunostained with
TG-3, then stripped with SDS and 2␤- mercaptoethanol [18]
and reprobed with MS3 to detect nucleolin. In the control
experiments the cleaving reagent was omitted.
2.7. Immunohistochemistry

2.4. Immunoprecipitation
Nocodazole-treated HEp-2 cells from 50 ml of culture
media were collected by centrifugation and washed twice
with PBS. The following procedures were performed at
4°C. Cells were lysed with 1 ml of RIPA buffer [18] and
centrifuged at 8000 ϫ g for 30 s. 30 ␮l of supernatant were
mixed with 3 ␮l of MS3 ascites fluid. The reaction mixtures
were incubated with rotation for 1 h, and 70 ␮l of antimouse IgG agarose (Sigma, St. Louis, MO) were added for
an additional 30 min. The resin was washed three times with
1 ml of RIPA buffer, and the immunoprecipitated proteins
were eluted by the addition of SDS-PAGE sample buffer
containing 2-mercaptoethanol. The sample was divided into
two equal parts and subjected to SDS-PAGE, transferred
onto PVDF and probed using either MS3 antibodies or
TG-3 antibodies.
2.5. “Mitotic shake”
HEp-2 cells were grown in 175 mm2 flasks as described
above. Mitotic cells were mechanically dislodged from the
rest of the monolayer by dropping the flasks from three feet.
The dislodged mitotic cells were harvested by centrifugation of the supernatant at 1000 ϫ g for 5 min. The cell pellet

All human subjects used in this study were evaluated
both clinically and neuropathologically at the Albert Einstein College of Medicine as described elsewhere [44]. For
neuropathological evaluation, tissue was immersion-fixed in
4% paraformaldehyde and independently evaluated with
antibody Alz-50 and Thioflavin S. Brains from patients with
clinical symptoms of AD showed marked neuropathological
lesions. No clinical diagnosis of AD was noted in the control cases. Immunostaining was performed essentially as
described elsewhere by free floating 50 ␮m vibratome sections in 24 well culture dishes [44]. For the blocking experiment the tissue was incubated overnight at 4°C with
monoclonal antibody AT180 which recognized phosphorylated threonine 231 (Thr231) on tau and does not recognize
nucleolin. The tissue was then washed as described elsewhere [44], incubated with unlabeled anti-mouse IgG1
overnight at 4°C, washed again, incubated with TG-3 and
nucleolin rabbit antisera W15 overnight at 4°C, and washed.
TG-3 was detected with TRITC conjugated anti-mouse
IgM, while W15 was detected with biotinylated anti-rabbit
antibodies followed by streptavidin conjugated CY5. All
secondary incubations were performed as described above.
Fluorochromes were visualized by the Bio-Rad Confocal
system through a Nikon inverted Diaphot microscope.

520

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

Fig. 1. Purification and identification of the TG-3 antigen. (A) Purification and Western blotting of the TG-3 reactive 105 kDa protein. HEp-2 cells were
synchronized with nocodazole, detergent extracted, and fractionated by sequential DEAE and Heparin chromatography. TG-3 immunoreactivity eluted in a
single fraction from both columns with 0.4 M and 0.5 M NaCl respectively. (B) Western blot analysis (lanes 1–5) and immunoprecipitations with
anti-nucleolin MS3 antibody (lanes 6, 7). Lysates of synchronized (lanes 4 –7) and asynchronous (lanes 1–3) HEp-2 cells were analyzed with MS3 (lanes
1, 4) or TG-3 (lanes 2, 3, 5). Lane 3, cells collected by the mitotic shake technique (see methods).

3. Results
3.1. The TG-3 antigen in mitotic cells is nucleolin
TG-3 consistently recognized a major 105 kDa band in
extracts from HEp-2 cells that were synchronized in mitosis
by treatment with nocodazole (Fig. 1B, lane 5). This band
was also observed in extracts from synchronized MSN neuroblastoma [44] and HeLa cells (data not shown). The
appearance of this band was not an artifact of nocodazole
treatment because an identical band was observed in extracts from mitotic HEp-2 cells that were enriched in mitotic
cells by mechanical dislodging from the flask (Fig. 1B, lane
3). The TG-3 antigen was thus purified from nocodazole
synchronized HEp-2 cells by sequential column chromatography and subjected to N-terminal sequencing (Fig. 1A).
The sequence of the first 18 amino acids of the TG-3 antigen
was identical to the N-terminal sequence of human nucleolin, a protein with a molecular mass of 105 kDa [37].
To exclude the possibility that a minor protein of a
similar size, copurifying with nucleolin, could account for
the TG-3 immunoreactivity, we immunoprecipitated cell
lysates with anti-nucleolin monoclonal antibody MS3 and
analyzed the immunoprecipitate with TG-3. A band with the
molecular weight of 105 kDa was immunoprecipitated by
the MS3 antibody and detected by the TG-3 antibody (Fig.

1B, lane 7). These results support our initial finding that
both MS3 and TG-3 antibodies recognize nucleolin in mitotic cells.
As a further confirmation, the protein purified from nocodazole synchronized cells was subjected to chemical
cleavage with either N-chlorosuccinimide or hydroxylamine
(Fig. 2). Human nucleolin has two recognition sites for each
reagent, distributed in such a way that cleavage generates
three fragments, two of which were predicted to be detectable with the antibodies TG-3 and MS3 (see methods).
Major bands migrating at 92 kDa and 74 kDa were observed
after cleavage with N-chlorosuccinimide (Fig. 2A) and hydroxylamine (Fig. 2B), respectively. It should be noted that
although nucleolin contains two cleavage sites for each
reagent, under our experimental conditions we detected the
product of only one major cleavage event. After each cleavage reaction, both MS3 and TG-3 stained identical bands,
suggesting that they recognized the same protein. Thus, the
results from the chemical cleavage experiments, in conjunction with the sequencing and immunoprecipitation experiments, proved that the TG-3 antibody recognized nucleolin
in mitotic cells.
3.2. TG-3 reactive nucleolin is phosphorylated by Cdc2
Nucleolin is a major cellular phosphoprotein, which has
nine putative phosphorylation sites for the major mitotic

Fig. 2. Identification of the TG-3 antigen as mitotic nucleolin by chemical cleavage. The TG-3 antigen was partially purified from synchronized cells as in
Fig. 1, separated by SDS-PAGE, and subjected to partial chemical digestion with N-chlorosuccinimide (A) and hydroxylamine (B). The cleavage products
of each digest were separated by SDS-PAGE and transferred onto two membranes. Each membrane was probed with the TG-3 antibody (lanes 1,2), stripped,
tested by ECL, and reprobed with the MS3 antibody (lanes 3,4). Controls contained no cleaving reagent (lanes 1, 3).

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

Fig. 3. Phosphorylation of nucleolin by cdc2 confers TG-3 immunoreactivity. Nucleolin partially purified from asynchronous cells was phosphorylated by Cdc2/cyclinB in the presence of 32P labeled ATP. The phosphorylated nucleolin was further immunoprecipitated with anti-nucleolin
monoclonal antibody MS3 and Western-blotted with either MS3 (panel B)
or TG-3 (panel C) antibodies. (A) Autoradiograph of samples separated by
SDS-PAGE (Lanes 1 and 2). Partially purified preparations were incubated
with [32P]ATP in the presence (lane 1) and the absence (lane 2) of the
cdc2/cyclinB complex. The 105 kDa TG-3 immunoreactive band appears
only in the sample where Cdc2 was present (lane 1, panel C). (Lanes 3– 6)
Immunoprecipitations of nucleolin with MS3 from samples prepared as in
lanes 1 and 2. 32P is incorporated into nucleolin both in the presence (lane
3, panel A) and absence (lane 4, panel A) of Cdc2. The 105 kDa, TG-3
immunoreactive protein is immunoprecipitated only in the sample where
cdc2 was present (lane 3, panel C). No staining was detected when no
nucleolin (lane 5) or no primary antibody (lane 6) was used in the experiment. (ϩ) and (-) above lanes indicate presence or absence of the reagent.

kinase Cdc2 and has been shown to be phosphorylated by
Cdc2 both in vitro and in vivo [2]. Previous studies demonstrated that TG-3 recognizes phosphoepitopes [44,20].
Since TG-3 recognizes mitotic and not interphase cells,
the following experiments were performed in order to test if
TG-3 immunoreactivity is the result of phosphorylation of
nucleolin by Cdc2. Interphase nucleolin, which is not reactive with TG-3, was partially purified from asynchronous
HEp-2 cells. Nucleolin enriched samples were phosphorylated with [32P]ATP in the presence or absence of Cdc2
kinase. After phosphorylation, nucleolin was immunoprecipitated with the MS3 antibody, separated by SDS-PAGE,
and analyzed by autoradiography and immunoblotting with
either MS3 or TG-3 antibodies (Fig. 3). Autoradiography
revealed 32P incorporation into numerous bands suggesting
that the nucleolin enriched fraction contained many proteins
that could be phosphorylated by Cdc2 kinase (Fig. 3A, lane
1). A significant incorporation of 32P into the protein im-

521

munoprecipitable with MS3 in the absence of Cdc2 suggested that nucleolin was also phosphorylated by copurified
kinases (Fig. 3A, lane 4).
While many bands showed Cdc2 dependent incorporation of 32P, only a single band migrating at 105 kDa was
immunoreactive with TG-3 (Fig. 3C, lane 1). The appearance of the single TG-3 immunoreactive band at 105 kDa
after phosphorylation with Cdc2 kinase demonstrated the
antigenic specificity of the TG-3 phosphoepitope-specific
antibody for the nucleolin protein in the relatively crude
preparation. Furthermore, immunoprecipitation of nucleolin
from the enriched fraction with non-phosphospecific MS3
(see Valdez et al., 1995 for epitope mapping) [42] revealed
that nucleolin became immunoreactive with TG-3 only after
phosphorylation by Cdc2 (Fig. 3C, lane 3). Phosphorylation
reactions carried out in the absence of Cdc2 did not produce
epitopes recognizable by TG-3 (Fig. 3C, lanes 2 and 4).
Therefore, TG-3 is specific for phosphoepitopes produced
by Cdc2 and not by copurifying endogenous kinases. Finally, phosphorylation was time and kinase concentration
dependent, and TG-3 immunoreactivity was proportional to
the amount of 32P incorporation (data not shown). Together,
these data demonstrate the specificity of the TG-3 antibody
for nucleolin phosphorylated by Cdc2.
3.3. TG-3 reactive nucleolin undergoes temporal and
spatial changes during mitosis in culture
TG-3 immunoreactivity was examined in an asynchronous population of HEp-2 (Fig. 4) and HeLa cells (data not
shown) by laser confocal microscopy. Dual labeling with
monoclonal antibodies to either ᮀ-tubulin or to histone
proteins was used to identify cells in different stages of
mitosis. TG-3 reactive nucleolin (TG-3 nucleolin) appeared
abruptly in cells entering mitosis. Interphase cells were not
reactive with TG-3 (Fig. 4). TG-3 nucleolin was localized to
the nuclei in cells in late prophase and early prometaphase
at the beginning of centrosomal segregation (Figs. 4A-C).
TG-3 nucleolin was also localized to the mitotic centrosomes starting from the earliest stages of polar migration
(Figs. 4C and 4F). TG-3 nucleolin became dispersed
throughout the mitotic cytosol as the prometaphase progressed (Figs. 4E and 4F). Optical sectioning revealed that
the condensed chromosomes were immunoreactive with
TG-3 as judged by the overlap with immunostaining that
was detected by anti-histone antibody mAb053 (Fig. 4L).
The chromosomal localization of TG-3 nucleolin became
most apparent in metaphase when the majority of the immunoreactivity was localized to the aligned chromosomes
(Figs. 4H and 4K). While chromosomal TG-3 immunoreactivity was most intense, some also localized to the mitotic
cytosol. The intensity of cytosolic TG-3 immunoreactivity
varied between the metaphase cells of the different cell lines
but was always lower than immunoreactivity associated
with chromosomes (data not shown). TG-3 immunoreactivity began to decrease at the earliest stages of chromosomal

522

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

Fig. 4. Distribution of TG-3 immunoreactive nucleolin throughout mitosis. Laser confocal micrograph of asynchronous HEp-2 cells dual labeled with
neurofibrillary tangle monoclonal antibody, which recognizes nucleolin phosphorylated by Cdc2: TG-3 in red, monoclonal antibody to ␤-tubulin in green (A,
D, G), and monoclonal antibodies to histone proteins in green (J). Colocalization is shown in yellow (C, F, I, L). Note the appearance of TG-3
immunoreactivity (B). Only the cells in mitosis are immunoreactive with TG-3 (B, H, K) and the immunoreactivity is localized to the nucleus and to the
segregating centrosomes (A-arrow, B, C). During late prometaphase, TG-3 immunoreactivity becomes dispersed throughout the mitotic cytosol (E, F). The
centrosomes retain TG-3 immunoreactivity (F). In metaphase, the aligned chromosomes are immunoreactive with TG-3 (H) and chromosomal TG-3 staining
colocalizes with histone staining (K). The centrosomes lose their TG-3 immunoreactivity. In anaphase TG-3 immunoreactivity begins to decrease (H-cell 3).
A marked difference in intensity of TG-3 immunoreactivity is observed between metaphase (K-cell 1) and telophase (K-cell 2).

Fig. 5. Distribution of TG-3 immunoreactive and non-reactive nucleolin in the early stages of mitosis. Laser confocal micrographs of asynchronous HEp-2
cells dual labeled with monoclonal antibody to nucleolin: MS3 (green) and TG-3 (red). A nuclear distribution of nucleolin is observed in interphase cells.
In the early stages of mitosis, nucleoli are intensely immunoreactive with both TG-3 and MS3. Note that the overlap in the mitotic cells is partial (A, B),
suggesting that not all nucleolin is phosphorylated by Cdc2 in these cells. DAPI insert indicates cell in prometaphase (D-arrow)

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

segregation, which was determined by histone specific
mAb053 (Figs. 4J and 4L) and DAPI staining (data not
shown). Whereas the intensity of TG-3 nucleolin immunoreactivity proceeded to decrease from metaphase to telophase, its spatial distribution did not change and the chromosomes persisted to be the predominant TG-3-immunoreactive
structures (Figs. 4H, cell 3; Fig. 4K, cell 2). Cells in cytokinesis
and interphase showed only background levels of immunostaining.
3.4. TG-3 reactive and non-reactive nucleolin are
differentially distributed during mitosis
Asynchronous HEp-2 cells were dual-labeled with phosphoepitope specific TG-3 and with MS3 which is specific
for the region of nucleolin that is devoid of phosphorylation
sites [42]. Whereas TG-3 was highly specific for nucleolin
in cells in mitosis, MS3 was immunoreactive with nucleolin
in all phases of the cell cycle (Fig. 5). As expected, the most
prominent MS3 immunoreactive structures in interphase
cells were nuclei and nucleoli. MS3 immunoreactivity was
more dispersed throughout the nucleus in the early stages of
mitosis. In prophase, nuclear TG-3 immunoreactivity only
partially colocalized with the MS3 immunoreactivity (Fig.
5A). In metaphase, MS3 immunoreactivity was distributed
throughout the mitotic cell. MS3 and TG-3 immunoreactivity colocalized primarily around the chromosomes since
chromosomal TG-3 immunoreactivity was most pronounced. MS3 immunoreactivity of cytosol persisted in
anaphase, telophase, and cytokinesis, while TG-3 immunoreactivity gradually became undetectable (Fig. 5C). The
partial overlap of MS3 and TG-3 immunoreactivities in all
stages of mitosis suggests that TG3 reactive nucleolin represents a unique subpopulation of mitotic nucleolin.
3.5. Differential distribution of nucleolin in the normal
human brain and in AD
The functions of nucleolin have been extensively studied
in cultured cells and in vitro [14,41]. Nucleolin has been
immunocytochemically localized primarily to the nucleolus
and the nucleus. However, the distribution and the role of
nucleolin in human brain cells remain unknown. Therefore,
the distribution of nucleolin in the brain was examined in
four individuals affected by AD and in eight age-matched
controls. Identical immunolocalization was obtained with
all examined anti-nucleolin monoclonal and polyclonal antibodies, strongly suggesting that the distribution of immunoreactivity reflected the true distribution of nucleolin and
that it was not the result of cross reactivity with other
proteins. The images on panels of Fig. 6 and 7 were obtained with anti-nucleolin monoclonal antibody MS3 and
illustrate a typical distribution of nucleolin in the hippocampus. The specificity of TG-3 for AD brains has been previously described [44], and it is here presented for illustrative
value (Fig. 6A and 6F).

523

Cells corresponding to all major areas of the hippocampal formation, including the dentate gyrus, were positive for
nucleolin. Nucleolin was detected in both large neuronal
nuclei and smaller glial nuclei (Figs. 6C and 6D). The
distribution of nucleolin in glia appeared primarily nuclear,
while neuronal nucleolin was readily detected in both cytoplasm and nuclei. Moreover, some neurons displayed
mostly cytoplasmic distribution of nucleolin, while in other
neurons, nucleolin was primarily nuclear (Fig. 6B). In addition to staining the neuronal cell bodies, nucleolin was
detected in the apical dendrites of the pyramidal cells (Fig.
6D). Neurons with predominantly nuclear, and neurons with
predominantly cytoplasmic nucleolin were observed in CA3
and CA4, whereas more cytoplasmic and robust dendritic
staining was observed in CA2 and CA1 (Figs. 6C and 6D).
The variability in the distribution of nucleolin between cells
is indicative of a dynamic function for the protein in neurons. Occasionally a small number of NFT were observed in
the hippocampus from individuals who showed no cognitive
impairments prior to their death. The appearance of these
early pathological changes in the brain prior to development
of clinical symptoms is believed to characterize AD progression. CA1 is the earliest region in the hippocampus to
be affected by neurodegenerative changes [5]. Interestingly,
an intense nucleolin immunoreactivity in densely packed
material typical of NFT within some pyramidal cells in the
CA 1 region was observed in four of our control cases (Fig.
6E, representative example).
To further characterize the association of nucleolin with
AD pathology, brains from 3 individuals affected by AD
were immunostained with mAb MS3 (Figs. 6G,H,I, and J).
The distribution of nucleolin in the AD brains paralleled the
distribution of nucleolin in the control brains. However, the
neurons in the AD brains exhibited a marked decrease in
nucleolin immunoreactivity throughout the hippocampal
formation (Fig. 6G and 6H). Nucleolin immunoreactivity,
shown in red, labeled small granular structures in many of
the neurons in the CA3 and CA4 regions (Fig. 6H). Nucleolin localized to NFT in the CA1 and CA2 regions of the
hippocampal formation (Figs. 6I and 6J). Using the classification of NFT proposed by Wischik (1989) [49], we found
that the NFT immunolabeled with MS3 had the appearance
of type 1 (early intracellular), type 2 (late intracellular) and
type 3 (extracellular) NFT. Type 1 NFT were observed as
intensely fluorescent, densely packed material traversing the
pyramidal neuron and surrounding the nucleus. Large swollen pyramidal neurons with intensely labeled NFT and a
nucleus that was positioned near the plasma membrane
resembled type 2 NFT. Finally, intense filament-like staining of cell remnants with no noticeable nucleus was consistent with type 3 NFT.
3.6. TG-3 reactive nucleolin is present in NFT
A recent study demonstrated that in AD brains, TG-3
antibody recognized microtubule associated protein tau

524

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

Fig. 6. Distribution of nucleolin in control and AD brains. Light and immunofluorescent micrographs of control (B-E) and AD (G-J) hippocampal sections
immunostained with monoclonal antibody MS3. Control (A) and AD (F) brain sections immunostained with TG-3 were used as a reference for illustrative
value. CA3-CA4 (B, C): Nucleolin is present in large neuronal nuclei and neuronal cytoplasm, and in small glial nuclei (B, bar ϭ 0.05 mm; C, bar ϭ 0.1
mm). Nucleolin is primarily nuclear in some neurons, primarily cytoplasmic in others and was detected in both compartments in still other neurons (B). In
CA1 (D, F), note the intense immunoreactivity in the apical dendrites (D, bar ϭ 0.05 mm). Note the cytoplasmic distribution of nucleolin and presence of
nucleolin in early NFT in the control brain (E, bar ϭ 0.05 mm). A decrease in nucleolin immunoreactivity in all regions of the AD hippocampus was
consistently observed (G-J). In CA3-CA4 note infrequently observed immunoreactive plaque-like structures (G, arrows, bar ϭ 0.2 mm), and the granular
appearance of nucleolin staining (H, bar ϭ 0.2 mm). CA1 (I, J) - note the abundance of nucleolin positive NFT. Type 1, 2, and 3 NFT were observed (bar ϭ
0.05 mm).

phosphorylated on Thr231 [20]. It has not been established,
however, if TG-3 recognizes other proteins in AD brains.
After we established that TG-3 is highly specific for mitotic
nucleolin in cell cultures, it became important to determine
if TG-3 immunreactive nucleolin was present in the brains
of patients with AD and, if yes, to differentiate it from TG-3
immunoreactive tau.
In order to identify a population of NFTs with TG-3

reactive nucleolin, dual-label confocal microscopy was first
performed with TG-3 and MS3 antibodies on AD cases that
showed abundant MS3 immunoreactive NFT (Fig. 7A).
There were many TG-3 immunoreactive NFT and an intense neuritic staining by the TG-3 antibody. MS3 stained
NFT and nuclei, but there was no staining of neurites. Thus,
there were many NFT that were positive with both antibodies (Fig. 7A).

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

525

Fig. 7. TG-3 reactive nucleolin in the AD brain. (A) Laser confocal micrograph of AD affected hippocampus dual labeled with antibody TG-3 (red) and
anti-nucleolin mAb MS3 (green) (bar ϭ 2.5 ᮀ). Green nuclei indicate TG-3 nonreactive nucleolin in neurons. Yellow NFT indicate overlap. TG-3 recognizes
microtubule associated protein tau in neurites (red). (B, C) TG-3 recognition of phosphorylated Thr 231 on tau was blocked by preincubation with anti-tau
monoclonal antibody AT-180 which recognizes phosphorylated Thr231 on tau, but not mitotic nucleolin. Blocked and control sections from the same region
of the same brain were dual labeled with TG-3 and anti-nucleolin antiserum. Note the complete disappearance of TG-3 staining from compartments where
no nucleolin was present (B). The amount of total TG-3 immunoreactivity and the ratio of TG-3/nucleolin overlap were quantitated and compared in blocked
(ϩAT180) and control (-AT180) sections (C). Blocking tau from recognition by TG-3 decreased the amount of total TG-3 staining, but did not decrease the
amount of TG-3/nucleolin overlap, suggesting that TG-3 recognizes nucleolin in tangles. The data are presented as the ratio of the amount of TG-3/nucleolin
overlap to total TG-3.

We then took advantage of another monoclonal antibody,
AT180, which is also specific for tau phosphorylated on
Thr231 [15]. However, AT180 did not recognize nucleolin
in mitotic cells in cell cultures by immunocytochemistry or
immunoblotting (data not shown). In AD brains all cellular
structures recognized by TG-3 were also recognized by
AT180, indicating that the phosphorylated Thr231 on tau
was accessible to both antibodies in the same cellular compartments (data not shown). Since AT180 recognizes phosphorylated tau, but not mitotic nucleolin, AD brain tissue
sections were preincubated with AT180 to block phosphorylated Thr231 on tau from being recognized by TG-3. The
tissue sections were subsequently dual labeled with TG-3
(to detect the TG-3 immunoreactive protein which was not
tau) and anti-nucleolin polyclonal antisera (to detect how
much of the residual TG-3 immunoreactivity overlapped
nucleolin) (Figs. 7B and 7C). After preincubation with
AT180, TG-3 immunoreactivity disappeared preferentially
from areas where no nucleolin immunoreactivity was
present. TG-3 immunoreactivity was completely blocked in
dystrophic neurites and neuritic components of plaques
(Fig. 7B). The laser confocal fluorescence was quantified,
and the proportion of total TG-3 immunoreactivity represented by the overlap of TG-3 and nucleolin was computed
(Fig. 7C). A six-fold increase in the overlap ratio was seen
in the sample blocked by AT180 compared to the control
sample, demonstrating that AT180 blocked TG-3 reactive
tau, but not TG-3 reactivity colocalized with nucleolin. The
overlapping of residual TG-3 with nucleolin indicated that
TG-3 reactive nucleolin was present in the NFT.

4. Discussion
4.1. Nucleolin phosphorylated by Cdc2 is the TG-3 antigen
Nucleolin is a major multifunctional nuclear phosphoprotein [14,41] and was one of the first identified nuclearcytoplasmic shuttling proteins [3]. It is thought to be intricately involved in ribosomal biogenesis and it is a marker
for cellular proliferation [13,22]. Nucleolin is a physiological substrate for Cdc2 kinase [2,32], and its biologic functions are highly regulated by phosphorylation in a cell cycle
specific manner [21,35]. Here we report the identification of
the cellular antigen for the AD specific antibody TG-3 as a
form of nucleolin that is phosphorylated by Cdc2 in mitosis.
Nucleolin has 9 proline directed consensus sequences for
Cdc2 recognition. Phosphorylation of nucleolin by Cdc2
and casein kinase II (CKII) has been reported in vitro and in
vivo [2,32,24]. The incorporation of 32P in the absence of
Cdc2 shown in the present study, complements the previous
observation that CKII copurifies with and phosphorylates
nucleolin [24]. However, our data indicate that in mitotic
cells, TG-3 recognizes nucleolin phosphorylated by Cdc2
and not other kinases, and TG-3 does not recognize other
phosphoproteins in cultured cells. These results firmly establish the TG-3 antibody as a novel and useful tool for
studying nucleolin in mitosis.
4.2. Cdc2 phosphorylated nucleolin and mitosis
The abrupt appearance of TG-3 reactive nucleolin in
early mitosis corresponds to activation of Cdc2, thus sup-

526

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

porting the in vitro phosphorylation results discussed above.
A recent report indicates that nucleolin is also phosphorylated by cell cycle kinases that are active prior to mitosis
[33]. The specificity of TG-3 for cells in mitosis presents in
situ evidence that TG-3 is specific for a phosphoepitope
generated by Cdc2 and not for phosphoepitopes generated
by other kinases. TG-3 immunoreactivity appears in the
nucleus at the onset of mitosis. Here we report the translocation of nucleolin phosphorylated by Cdc2 to the cytoplasm during late prophase and early prometaphase.
Schwab and Dreyer (1997) [35] recently demonstrated that
Cdc2 phosphorylation localizes microinjected X.laevis
nucleolin to the cytoplasm. The redistribution of Cdc2 phosphorylated nucleolin to the cytoplasm in late prophase provides functional in situ support for the above result and
raises the possibility that phosphorylation may play a role in
targeting nucleolin outside the nucleus at this stage of the
cell cycle. Interestingly, the in situ detection of nucleolin in
the cytoplasm has been elusive, suggesting that under normal conditions the amount of nucleolin in the cytoplasm of
cultured cells is below the level detectable by immunocytochemical methods. However, in polio-virus infected cells
[47] and in rat intestinal epithelial cells grown on a glass
surface [51], nucleolin was found in the cytoplasm.
The localization of Cdc2 phosphorylated nucleolin to the
chromosomes is consistent with previously reported observations [12]. Moreover, colocalization of MS3 and TG-3
immunoreactivity in the perichromosomal region suggested
that Cdc2 phosphorylation of nucleolin could be important
for its chromosomal interaction. It has been hypothesized
that interaction of Cdc2 phosphorylated nucleolin with histone H1 plays a role in chromatin condensation in mitosis
[21].
While most of the nucleolin is located in the mitotic
cytosol, most of the TG-3 reactive nucleolin is localized to
mitotic chromosomes. In every stage of mitosis, MS3 immunoreactivity only partially overlapped the TG-3 immunoreactivity, suggesting heterogeneity of mitotic nucleolin.
This partial overlap also establishes that TG-3 reactive
nucleolin represents a unique subpopulation of mitotic
nucleolin. Functional studies with nucleolin mutated in its
Cdc2 phosphorylation domains will help to elucidate the
biologic role of phosphorylation in the interaction of
nucleolin with chromosomes and in the cytoplasmic targeting of nucleolin.
We observed a decrease in TG-3 immunoreactivity starting at anaphase, and it became undetectable in cytokinesis.
The decrease in TG-3 immunoreactivity during mitotic exit
may be due to dephosphorylation or degradation of nucleolin. Recent reports suggest that nucleolin is more susceptible to degradation in nonproliferating cells [our unpublished
observations; 8]. Our observation that TG-3 immunoreactivity decreases in late mitosis while MS3 immunoreactivity
remains unchanged suggests the possibility that TG-3 reactive nucleolin is first dephosphorylated and then degraded.
Interestingly, the Saccharomyces pombe nucleolin homolog

gar2 is essential for cytokinesis and phosphorylated by
Cdc2 [23].
Perhaps dephosphorylation and/or degradation of chromosome associated nucleolin has functional significance for
chromosomal segregation or cell division. Activation of
phosphatase Cdc14 triggers mitotic exit in budding yeast
[36,45]. Similar events in mammalian cells may explain the
disappearance of TG-3 phosphoepitope, which parallels the
separation of chromosomes and mitotic exit. The recently
demonstrated participation of nucleolin in special cases of
recombination and replication may provide clues to the role
of Cdc2 phosphorylated TG-3 immunoreactive nucleolin in
mitosis [17,40,4,9]. Thus, our finding of the dynamic association of TG-3 reactive nucleolin with mitotic chromosomes may represent the long sought but elusive direct link
between nucleolus and cell cycle regulation in higher organisms [7].
4.3. Nucleolin in the brain
This study represents the first characterization of the
distribution of nucleolin in the human brain. The presented
data were highly reproducible when several polyclonal antisera and monoclonal antibodies to nucleolin were used on
sections of human brain tissues obtained from both archival
biopsy and autopsy material.
Three novel observations were made. First, nucleolin
was detected in neuronal cytoplasm. Second, nucleolin was
present in NFTs, including early tangles, and NFT-associated nucleolin was TG3-positive, suggesting that it was
phosphorylated by Cdc2 kinase. Third, in AD, the overall
level of nucleolin was dramatically decreased in the temporal cortex as well as in the CA1 and CA2 regions of the
hippocampus. As reported by others in cultured cells [41,
14,38] and as described here, nucleolin was found in nuclei
and nucleoli of neurons and glia cells in the brain. Surprisingly, nucleolin was also present in the cytoplasm of granular neurons of the dentate gyrus and pyramidal neurons of
the hippocampus and of the lower neocortical layers. Moreover, nucleolin was identified in the apical dendrites.
The finding of nucleolin in the cytoplasm was surprising
because it is generally believed that nucleolin, as the name
indicates, is present in the nucleolus and that it can also be
detected in the nucleus. Nucleolin has been found in the
cytoplasm and plasma membrane by biochemical methods
[41,14,38]. However, immunocytochemical detection visualized nucleolin normally in the nucleus and nucleolus of
cultured cells suggesting that the quantity of nucleolin outside the nucleus is insufficient for its detection under normal
conditions. Nucleolin can be detected in cytoplasm of cultured cells by immunocytochemical methods only under
certain conditions [51,47]. Yu et al. [51] detected nucleolin
in the cytoplasm of rat intestinal epithelial IEC-6 cells
grown on glass slides. However, in cells grown on laminin,
nucleolin was found in the usual location of the nucleus and
nucleolus. Waggoner and Sarnow (1998) [47] reported a

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

massive nucleolar-cytoplasmic relocalization of nucleolin in
poliovirus-infected cells. Recently a nucleolin-related protein, probably nucleolin, was detected in the cytoplasm of
bovine photoreceptor cells. These studies suggest that in
response to certain as yet unidentified signals, nucleolin
may relocalize to the cytoplasm. Thus, it is possible to
speculate that in the brain, nucleolin is localized to the
neuronal cytoplam in response to signals that are absent in
cultured cells.
The association of nucleolin with neuronal degeneration
stems from the localization of nucleolin to the NFT. It is
important to stress that brain tissue from some of the so
called control age-matched cases, where no clinical symptoms of AD were apparent, had a small number of nucleolin
positive, early (type 1) NFT in the CA1. Since CA1 is
thought to be the earliest hippocampal region affected by
NFT during the course of AD progression [5], nucleolin was
inferred to be a morphological marker for early neurofibrillary changes.
Dual labeling experiments with antibodies to nucleolin
and TG-3 yielded partial colocalization. Recognition of
hyperphosphorylated tau by TG-3 serves to explain the
TG-3 immunoreactivity that does not overlap with nucleolin
immunoreactivity. Blocking of tau from being recognized
by TG-3 by preincubation with an excess of monoclonal
antibody AT180 led to the disappearance of TG-3 immunoreactivity in dystrophic neurites and neuritic components
of plaques, thus confirming tau presence in these structures.
Persistence of residual TG-3 immunoreactivity in NFT suggests that in NFT, TG-3 primarily recognized proteins other
than tau. Inefficient blocking of Thr231 on tau in NFT is not
likely since AT180 is intensely immunoreactive with NFT.
The colocalization of nucleolin immunoreactivity with the
residual TG-3 immunoreactivity strongly suggests that
Cdc2 phosphorylated nucleolin is one of the major TG-3
immunoreactive proteins in NFT.
Detection of TG-3 immunoreactive nucleolin in NFT
marks the first identification of a normal cellular substrate
for mitotic kinases, which appears to undergo mitotic modification in AD. Several recent reports indicating that mitotic kinases are active in the AD brain contribute important
information to our understanding of mitotic mechanisms
that are reactivated in AD pathogenesis. However, in order
to understand the mechanisms of the resulting neuronal
degeneration, substrates for mitotic kinases in AD must be
identified. It is important to reemphasize that nucleolin is a
marker for early (pre-clinical) neurofibrillary changes and
thus its mitotic modification may prove important in the
early development of neuronal degeneration associated with
AD.
The possibility that nucleolin is involved in both chromosomal segregation during mitosis as well as AD may
help to explain the observed co-occurrence of Down syndrome (DS) and AD within certain families [19,34]. The
majority of DS cases are the result of non-dysjunction of
maternal chromosome 21 in miosis, or a defect in chromo-

527

somal segregation. Therefore, the co-occurrence of DS and
AD within specific families suggests a converging cellular
mechanism between chromosomal segregation and neuronal
degeneration. Specifically, proteins and pathways involved
in chromosomal segregation and cell division in general
may serve as signals for, or intermediates in, the degeneration of post-mitotic neurons in AD. The possible involvement of nucleolin in mitosis and AD makes it a candidate
for such a protein.

Acknowledgments
This work was supported by NIA, Alzheimer’s Association, and Long Island Alzheimer Foundation (LIAF) to
DG, AG/OD 12721 to IV and MSTP GM08444, AFAR, and
LIAF fellowships to AD. The authors thank M. Cammer for
assistance with confocal images; G. Jicha for assistance
with a phosphorylation experiment; Drs. C. Dingwall, and
S. Strickland for critical reading of the manuscript; Drs. E.
Bromet, P. Fisher, W. Quitschke, S. Simon, and W. Van
Nostrand for helpful comments on the manuscript.

References
[1] Baumann, K., E.M. Mandelkow, J. Biernat, H. Piwnica-Worms, and
E. Mandelkow. Abnormal Alzheimer’-like phosphorylation of tauprotein by cyclin-dependent kinases cdk2 and cdk5. FEBS Lett.
1993;336:417– 424.
[2] Belenguer, P., M. Caizergues-Ferrer, J.C. Labbe, M. Doree, and F.
Amalric. Mitosis-specific phosphorylation of nucleolin by p34Cdc2
protein kinase. Mol Cell Biol. 1990;10:3607–3618.
[3] Borer, R.A., C.F. Lehner, H.M. Eppenberger, and E.A. Nigg. Major
nucleolar proteins shuttle between nucleus and cytoplasm. Cell. 1989;
56:379 –390.
[4] Borggrefe, T., M. Wabl, A.T. Akhmedov, and R. Jessberger. A
B-cell-specific DNA recombination complex. J Biol Chem. 1998;273:
17025–17035.
[5] Braak H, Braak E. Staging of Alzheimer’s disease-related neurofibrillary changes. Neurobiol Aging 1995;16:271– 8.
[6] Busser, J., D.S. Geldmacher, and K. Herrup. Ectopic cell cycle
proteins predict the sites of neuronal cell death in Alzheimers disease
brain. J Neurosc. 1998;18:2801–2807.
[7] Carmo-Fonseca, M., L. Mendes-Soares, and I. Campos. To be or not
to be in the nucleolus. Nature Cell Biol. 2:E107–E112, 2000.
[8] Chen, C.M., S.Y. Chiang, and N.H. Yeh. Increased stability of
nucleolin in proliferating cells by inhibition of its self-cleaving activity. J Biol Chem. 1991;266:7754 –7758.
[9] Daniely, Y, J.A. Borowiec. Formation of a complex between nucleolin and replication protein after cell stress prevents initiation of DNA
replication. J. Cell Biol. 149:799 – 809, 2000.
[10] Davis, F.M., T.Y. Tsao, S.K. Fowler, and P.N. Rao. Monoclonal
antibodies to mitotic cells. Proc Nat Acad Sci. 1983;80:2926 –2930.
[11] Deng, J.S., B. Ballou, and J.K. Hofmeister. Internalization of antinucleolin antibody into viable HEp-2 cells. Mol Biol Rep. 1996;23:
191–195.
[12] Dundr, M., U.T. Meier, N. Lewis, D. Rekosh, M.L. Hammarskjold,
and M.O.J. Olson. A class of nonribosomal nucleolar components is
located in chromosome periphery and in nucleolus-derived foci during anaphase and telophase. Chromos. 1997;105:407– 417.

528

D. Dranovsky et al. / Neurobiology of Aging 22 (2001) 517–528

[13] Ghisolfi, L., G. Joseph, M. Erard, J.M. Escoubas, C. Mathieu, and F.
Amalric. Nucleolin–pre-rRNA interactions and preribosome assembly. Mol Biol Rep. 1990;14:113–114.
[14] Ginisty, H., H. Sicard, B. Roger, and P. Bouvet. Structure and functions of nucleolin. J Cell Sci. 1999;112:761–772.
[15] Goedert, M., R. Jakes, R.A, Crowther, P. Cohen, E. Vanmechelen, M.
Vandermeeren, and P. Cras. Epitope mapping of monoclonal antibodies to the paired helical filaments of Alzheimer’s disease: identification of phosphorylation sites in tau protein. Biochem J. 1994;301:
871– 877.
[16] Gorbsky GJ. Cell cycle progression and chromosome segregation in
mammalian cells cultured in the presence of the topoisomerase II
inhibitors ICRF-187 [(ϩ)-1,2-bis(3,5-dioxopiperazinyl-1-yl)propane;
ADR-529] and ICRF-159 (Razoxane). Cancer Res 1994;54:1042– 8.
[17] Hanakahi, L.A., L.A. Dempsey, M.J. Li, and N. Maizels. Nucleolin is
one component of the B cell-specific transcription factor and switch
region binding protein, Lr1. Proc Natl Acad Sci. 1997;94:3605–3610.
[18] Harlow, E, D. Lane. Antibodies (Cold Spring Harbor: Cold Spring
Harbor Laboratory), 1988.
[19] Heston LL, Mastri AR. The genetics of Alzheimer’s disease: associations with hematologic malignancy and Down’s syndrome. Arch
Gen Psychiatry 1977;34:976 – 81.
[20] Jicha, G.A., E. Lane, I. Vincent, L. Otvos, Jr., R. Hoffmann, and P.
Davies. A conformation- and phosphorylation-dependent antibody
recognizing the paired helical filaments of Alzheimer’s disease.
J Neurochem. 1997;69:2087–2095.
[21] Kharrat, A., J. Derancourt, M. Doree, F. Amalric, and M. Erard, M.
Synergistic effect of histone H1 and nucleolin on chromatin condensation in mitosis: role of a phosphorylated heteromer. Biochem.
1991;30:10329 –10336.
[22] Lapeyre, B., H. Bourbon, and F. Amalric. Nucleolin, the major
nucleolar protein of growing eukaryotic cells: an unusual protein
structure revealed by the nucleotide sequence. Proc Natl Acad Sci
USA. 1987;84:1472–1476.
[23] Leger-Silvestre, I., M.P. Gulli, J. Noaillac-Depeyre, M. Faubladier,
H. Sicard, M. Caizergues-Ferrer, and N. Gas. Ultrastructural changes
in the Schizosaccharomyces pombe nucleolus following the disruption of the gar2ϩ gene, which encodes a nucleolar protein structurally
related to nucleolin. Chromosoma. 1997;105:542–552.
[24] Li, D., G. Dobrowolska, and E.G. Krebs. The physical association of
casein kinase 2 with nucleolin. J Biol Chem. 1996;271:15662–15668.
[25] Li, J.H., M. Xu, H. Zhou, J.Y. Ma, and H Potter. Alzheimer presenilins in the nuclear membrane, interphase kinetochores, and centrosomes suggest a role iIn chromosome segregation. Cell. 1997;90:
917–927.
[26] Liu, W.K., R.T. Williams, F.L. Hall, D.W. Dickson, and S.H. Yen.
Detection of a cdc2-related kinase associated with Alzheimer paired
helical filaments. Am J Path. 1995;146:228 –238.
[27] Mawal-Dewan, M., C.S. Parimal, M. Abdel-Ghani, D. Shalloway,
and E. Racker. Phosphorylation of tau protein by purified p34 cdc28
and a related protein kinase. J Biol Chem. 1992;267:19705–19709.
[28] McShea, A., P.L.R. Harris, K.R. Webster, A.F. Wahl, and M.A.
Smith. Abnormal expression of the cell cycle regulators P16 and
Cdk4 in Alzheimerıs disease. Am J Path. 1997;150:1933–1939.
´
[29] Nagy, Z., M.M. Esiri, and A.D. Smith. Expression of cell division
markers in the hippocampus in Alzheimers disease and other neurodegenerative conditions. Acta Neuropath. 1997;93:294 –300.
[30] Nigg EA. Cellular substrates of p34cdc2 and its companion cyclindependent kinases. Trends Cell Biol 1993;3:296 –301.
[31] Paudel, H., J. Lew, Z. Ali, and J.H. Wang. Brain proline-directed
kinase phosphorylates tau on sites that are abnormally phosphorylated
in tau associated with Alzheimer’s disease helical filaments. J Biol
Chem. 1993;268:23512–23518.

[32] Peter, M., J. Nakagawa, M. Doree, J.C. Labbe, and E.A. Nigg, E.A.
Identification of major nucleolar proteins as candidate mitotic substrates of cdc2 kinase. Cell. 1990;60:791– 801.
[33] Sarcevic, B., R. Lilischkis, and R.L. Sutherland. Differential phosphorylation of T-47d human breast cancer cell substrates by D1-, D3-,
E-, and a-type cyclin-Cdk complexes. J Biol Chem. 1997;272:33327–
33337.
[34] Schupf, N., D. Kapell, J.H. Lee, R. Ottman, and R. Mayeux. Increased risk of Alzheimer’s disease in mothers of adults with Down’s
syndrome. Lancet. 1994;344:353–356.
[35] Schwab MS, Dreyer C. Protein phosphorylation sites regulate the
function of the bipartite NLS of nucleolin. Eur J Cell Biol 1997;73:
287–97.
[36] Shou, W., J.H. Seol, A. Shevchenko, C. Baskerville, D. Moazed,
Z.W.S. Chen, J. Jang, A. Shevchenko, H. Charbonneau, and R.J.
Deshaies. Exit from mitosis is triggered by Tem1-dependent release
of the protein phosphatase Cdc14 from nucleolar RENT complex.
Cell. 1999;97:233–244.
[37] Srivastava, M., P.J. Fleming, H.B. Pollard, and A.L. Burns. Cloning
and sequencing of the human nucleolin cDNA. FEBS Lett. 1989;250:
99 –105.
[38] Srivastava, M, H.B. Pollard. Molecular dissection of nucleolin’s role
in growth and cell proliferation: new insights. FASEB J. 1999;13:
1911–1922.
[39] Taagepera, S., M.S. Campbell, and G.J. Gorbsky. Cell-cycle-regulated localization of tyrosine and threonine phosphoepitopes at the
kinetochores of mitotic chromosomes. Exp Cell Res. 1995;221:249 –
260.
[40] Thyagarajan, B., R. Lundberg, M.C. Rafferty, and C. Campbell.
Nucleolin promotes homologous DNA pairing in vitro. Somatic Cell
Mol Genet. 1998;24:263–272.
[41] Tuteja, R, N. Tuteja. Nucleolin: a multifunctional major nucleolar
phosphoprotein. Crit Rev Biochem Mol Biol. 1999;33:407– 436.
[42] Valdez, B. C., D. Henning, R.K. Busch, M. Srivastava, and H. Busch.
Immunodominant RNA recognition motifs of human nucleolin/C23.
Mol Immunol. 1995;32:1207–1213.
[43] Vincent, I., G. Jicha, M. Rosado, and D.W. Dickson. Aberrant expression of mitotic cdc2/cyclin B1 kinase in degenerating neurons of
Alzheimers disease brain. J Neurosci. 1997;17:3588 –3598.
[44] Vincent, I., M. Rosado, and P. Davies. Mitotic mechanisms in Alzheimer’s disease? J Cell Biol. 1996;132:4134 – 4125.
[45] Visinti, R., E.S. Hwang, and A. Amon. Cfi1 prevents premature exit
from mitosis by anchoring Cdc14 phosphatase in the nucleolus. Nature. 1999;398:818 – 823.
[46] Vulliet, R., S.M. Halloran, R.B., Braun, A.J. Smith, and G. Lee.
Proline-directed phosphorylation of human tau protein. J Biol Chem.
1992;267:22570 –22574.
[47] Waggoner, S, P. Sarnow, P. Viral ribonucleoprotein complex formation and nucleolar cytoplasmic relocalization of nucleolin in poliovirus-infected cells. J Virol. 1998;72:6699 – 6709.
[48] Wille, H., G. Drewes, J. Biernat, E.M. Mandelkow, and E. Mandelkow. Alzheimer-like paired helical filaments and antiparallel
dimers formed from microtubule-associated protein tau in vitro. J Cell
Biol. 1992;118:573–584.
[49] Wischik CM. Cell biology of the Alzheimer tangle. Cur Op Cell Biol
1989;1:115–22.
[50] Yen, S.H., A. Kenessey, S.C. Lee, and D.W. Dickson. The distribution and biochemical properties of a Cdc2-related kinase, KKIALRE,
in normal and Alzheimer brains. J Neurochem. 1995;65:2577–2584.
[51] Yu, D.H., M.Z. Schwartz, and R. Petryshyn. Effect of laminin on the
nuclear localization of nucleolin in rat intestinal epithelial Iec-6 cells.
Biochem. Biophys. Res. Comm. 247:186 –192;1998.