American Journal of Pathology, Vol. 168, No.

5, May 2006
Copyright American Society for Investigative Pathology
DOI: 10.2353/ajpath.2006.050603

Growth Factors, Cytokines, Cell Cycle Molecules

Mortalin-Based Cytoplasmic Sequestration of

p53 in a Nonmammalian Cancer Model

Charles Walker, Stefanie Bottger, and Ben Low

From the Department of Zoology, Center for Marine Biology and
Marine Biomedical Research Group, University of New
Hampshire, Durham, New Hampshire

In nature the soft shell clam Mya arenaria develops a

fatal neoplasm that shares molecular similarity with
an unrelated group of human cancers. In leukemic
clam hemocytes, wild-type p53 and mortalin proteins
co-localize in the cytoplasm. A similar phenotype,
characterized by cytoplasmic sequestration of wildtype p53 protein, has been observed in several
human cancers (undifferentiated neuroblastoma,
retinoblastoma, colorectal and hepatocellular
carcinomas, and glioblastoma). In some of these cancers p53 is tethered in the cytoplasm by mortalin
when the latter protein is overexpressed. Using coimmunoprecipitation we have demonstrated that
mortalin and p53 proteins are complexed in the cytoplasm of leukemic clam hemocytes (and not in normal hemocytes). In addition, treatment of leukemic
clam hemocytes with MKT-077, a cationic inhibitor of
mortalin, disrupts the interaction of mortalin and
p53 proteins, resulting in translocation of some p53
to the nucleus. Based on these data, we introduce
leukemic clam hemocytes as novel and easily accessible, in vivo and in vitro models for human cancers
displaying a similar mortalin-based phenotype. Treatment of these models with novel chemotherapeutics
may help reveal the molecular mechanism(s) involved in inactivating p53 by this form of cytoplasmic
sequestration. (Am J Pathol 2006, 168:1526 1530; DOI:
10.2353/ajpath.2006.050603)

Analyzing the effects of cytotoxic compounds on malignancy is currently limited to vertebrate models, which are
expensive to maintain and highly regulated. Alternatively,
studies of naturally occurring cancers in nonmammalian
model organisms that are inexpensive to maintain can
yield important data about molecular mechanisms that
are held in common with human diseases.
A number of explanations have been proposed to account for the cytoplasmic sequestration of p53 in mam-

1526

malian cancer cells. Among these are the overexpression

of mortalin (a Hsp70 family member) and Parc (a parkinlike ubiquitin ligase) proteins that can independently
serve as cytoplasmic tethers for p53.13 The soft shell
clam Mya arenaria displays a fatal leukemia at multiple
sites along the coasts of New England and south to the
Chesapeake Bay4 6 (individuals from virtually any New
England commercial source have 1% or greater incidence of this disease). The underlying molecular mechanisms that govern the development and progression of
clam leukemia may reflect those seen in human diseases
with similar mortalin-generated phenotypes.
Here we present the soft-shell clam as a naturally
occurring, preclinical model for studies of mortalin-based
cytoplasmic sequestration of wild-type p53 protein. In
this study, we characterize normal (NCH) and leukemic
(LCH) clam hemocytes in vivo and LCH in vitro using
immunocytochemistry, co-immunoprecipitation, and subcellular localization of clam p53 (Map53) and clam mortalin (Mamot) proteins before and after treatment with the
cationic inhibitor MKT-077.

Materials and Methods

Animals
Soft shell clams (n 70 to 150) were collected at the
lowest tides of each month from Marsh Island in New
Bedford Harbor at Fairhaven, MA (41o 38.0 N 70o 55.0
W) and were maintained at the University of New Hampshire Coastal Marine Laboratory in New Castle, NH. To
differentiate normal and leukemic clams, a small aliquot
(10 l) of hemolymph was aspirated from the pericardial
sinus using a 26-gauge needle and incubated in 96-well
microtiter plates for 2 hours at 8C. Clams were classified
using a Zeiss (Thornwood, NY) IM inverted microscope
Supported by the National Cancer Institute (grants CA71008-01 and
CA104112-01 to C.W.W.) and a United States Department of Agriculture
Hatch (grant 353 to C.W.W.).
Accepted for publication January 30, 2006.
Address reprint requests to Charles W. Walker, Professor of Zoology,
Department of Zoology, Center for Marine Biology and Marine Biomedical
Research Group, Rudman Hall, The University of New Hampshire,
Durham, NH 03824. E-mail: cwwalker@christa.unh.edu.

Cytoplasmic Sequestration of p53

1527
AJP May 2006, Vol. 168, No. 5

as normal (0% round, nonmotile LCH; 100% attached

NCH), early incipient leukemic (1 to 50% LCH), late incipient leukemic (50 to 99% LCH), or fully leukemic
(100% LCH). The average number of clams that was
100% leukemic in 51 collections was 5.64%; the range
was 0 to 11% throughout a 5-year period.

LCH in Vitro
To prepare LCH for in vitro experiments, hemolymph was
obtained from the pericardial sinus of freshly collected
fully leukemic clams (100% leukemic) and then centrifuged. Pelleted LCH were resuspended in 10 ml of chemically defined culture medium7 to which 10% heat-inactivated fetal bovine serum was added. LCH were seeded
(4 to 7 104 ml1) in 100- to 150-ml spinner flasks at 32
rpm in an environmental chamber maintained at 8 to
10C. Penicillin, ampicillin, and streptomycin were also
added to all media at 200 U ml1, 25 g ml1, and 200
g L1, respectively.

Mamot Antibody Synthesis

A polyclonal antibody to Mamot was raised against a
21-residue peptide (NH2-CRLREAAEKAKIELSSSLQTDCOOH) synthesized by New England Peptide, Inc.
(Gardner, MA) based on authentic Mamot sequence from
a region containing residues within the MKT-077 binding
domain (GenBank Accession number AY326398).8 Bovine serum albumin was conjugated to the N-terminal
cysteine using Imject maleimide-activated bovine serum
albumin (Pierce Biotechnology Inc., Rockford, IL). The
hapten-peptide conjugate was used to immunize two
New Zealand White rabbits (Millbrook Farm, Amherst,
MA). The antibody was affinity-purified with Affi-Gel (BioRad, Hercules, CA) and screened for its ability to recognize Mamot protein using Western blot analysis.

Immunocytochemistry and Protein Localization

in Clam Hemocytes
Cytospins of 100 l of freshly collected hemocytes
from normal (0% leukemic) and fully leukemic (100%
leukemic) clams or of LCHs in vitro were fixed and
permeabilized by immersion in acetone. Primary antibodies for immunocytochemistry (1:50 l of primary
antibody:buffer) included -tubulin (clone DM-1A,
mouse anti-chicken monoclonal from ICN, Irvin, CA; to
demonstrate the microtubular array involved in cytoplasmic transport of Map53 protein),9 clam mortalin
(Mamot, rabbit anti-clam polyclonal), and clam p53
(Map53, rabbit anti-clam polyclonal).10 Resulting
preparations for -tubulin were developed with the
appropriate peroxidase-labeled secondary antibody
(Vectastain ABC Elite IgG kit; Vector Laboratories, Burlingame, CA) followed by treatment with 3,3-diaminobenzidine. To demonstrate co-localization of Mamot
and Map53, polyclonal antibodies were directly conjugated to Quantum Dots 525 (green fluorescence, mor-

talin) and 655 (red fluorescence, p53) (Invitrogen,

Carlsbad, CA). Control cytospins received identical
treatment minus the conjugated antibody. LCHs were
fixed in equal amounts of methanol and acetone, permeabilized in phosphate-buffered saline (PBS) containing 0.5% Triton X-100, blocked with PBS containing
0.05% Triton X-100 and 2% bovine serum albumin, and
incubated for 12 hours in 0.5 l of conjugated Mamot
and 1 to 3 l of conjugated Map53 antibodies (differences in amounts of antibodies relating to differences
in preconjugation antibody concentrations). After rinsing the slides thoroughly with PBS containing 0.2%
Triton X-100 and mounting the cells with polyvinyl alcohol-based mounting medium (Sigma-Aldrich, St.
Louis, MO), fields of 200 cells were counted and
scored as positive or negative on a Zeiss Axioplan II
microscope equipped with epifluorescence, an AxioCam MR camera, and AxioVision 4.4 software (Carl
Zeiss, Inc.).
To determine the distribution of Map53 protein between the nucleus and cytoplasm, 10 ml of normal and
1 ml of leukemic clam hemolymph (containing 1.5
108 LCH/ml1) were extracted from the pericardial sinus. Normal and leukemic hemolymph was centrifuged
at 1200 rpm for 10 minutes. Hemocytes were isolated
after centrifugation and nuclear and cytoplasmic proteins were extracted using an NE-PER kit (Pierce).
Total protein was determined using a modified Lowry
procedure. Distribution of nuclear and cytoplasmic
Map53 protein in NCH and LCH cytoplasm was assayed by Western blot using our anti-Map53 polyclonal
antibody; all lanes were loaded with 24 g of total
protein.10

Co-Immunoprecipitation of Mortalin and p53

Hemolymph was extracted from a leukemic clam (1 ml;
3.45 108 cells/ml) and a normal clam (5 ml; 1.55
106 cells/ml). Total cytoplasmic protein was extracted
using the NEPER extraction kit (Pierce) and quantified
using a modified Bradford procedure (Bio-Rad Laboratories, Hercules, CA). Co-immunoprecipitation of
Map53 and Mamot was accomplished using an antibody-coupling gel to precipitate the bait protein
(Map53) and co-immunoprecipitate the interacting
prey protein (Mamot). Anti-p53 (Map53, rabbit anticlam polyclonal) was coupled to an amine-reactive gel
(ProFound co-immunoprecipitation kit, Pierce) using
slow agitation at 22C overnight. Cytoplasmic lysates
for both NCH and LCH (100 l) were incubated with the
antibody-coupled gel in separate spin-columns. Two
negative controls were also prepared, a control gel that
used an inactivated form of the gel and a quenched
gel, in which a quenching buffer was applied in place
of the Map53 antibody. Both negative controls were
incubated with a 50/50 mixture of NCH and LCH lysates and processed alongside the treatments in an
otherwise identical manner. The first three elutions from
each co-immunoprecipitation were stored for Western
blot analysis. Elutions were separated by electrophore-

1528
Walker et al
AJP May 2006, Vol. 168, No. 5

Figure 2. Western blot of NCH and LCH showing nuclear and cytoplasmic
distribution of Map53 in the former and cytoplasmic sequestration of p53 in
the latter. A larger p53 family member protein is also identified in LCH.

Co-Immunoprecipitation of Mortalin by p53

Figure 1. Nomarski image of living, freshly collected normal clam agranulocytic (AG) and granulocytic (G) NCH. Scale bar, 20 m.

sis on a 4 to 15% Tris-HCl gel and then transferred to

a polyvinylidene difluoride membrane. Concurrently,
nuclear and cytoplasmic lysates from LCH were electrophoresed and analyzed as positive controls. The
membrane was probed with Map53 antibody first to
ensure presence of the bait protein followed by probing with Mamot antibody. Bands for Mamot were visualized using anti-rabbit horseradish peroxidase antibody and chemiluminescent detection reagents
(Amersham, Arlington Heights, IL).

Treatment with MKT-077 and Disruption of

Mortalin/p53 Co-Localization
To determine the appropriate concentration of MKT-077
to yield disruption of mortalin/p53 binding in vitro, MKT077 at final concentrations of 1 to 7 mol/L was added to
a suspension of LCH. Cells were rotated at 1 rpm for 8
hours, prepared as cytospins, and assayed using direct
observation of Romanovski-stained cells, the terminal
dUTP nick-end labeling (TUNEL) assay (In Situ cell death
detection kit, AP; Roche Laboratories, Indianapolis, IN)
and simultaneous Quantum Dot immunocytochemistry for
Mamot and Map53. Preparations were observed on a
Zeiss Axioplan II microscope to determine the LC50 (concentration at which 50% of the cells died) and to detect
translocation of Map53 into the nucleus.

Results

Presence of the bait Map53 protein was detected in

cytoplasmic lysates from both NCH and LCH (data not
shown). Co-immunoprecipitation using Map53 as bait
demonstrated that complexes between Map53 and
Mamot were found in cytoplasmic lysates from LCH but
not in those from NCH (Figure 3). Both negative controls
confirmed the validity of these results because neither
produced a positive signal. Furthermore, detection using
the Mamot antibody showed that the interaction occurred
in LCH cytoplasm but not in the nucleus.

Distribution of p53 and Mortalin in LCH in Vivo

and in Vitro
LCH in vivo (100% LCH at 5 108 cells ml1) had an
average diameter of 7 to 10 m (Figure 4a). These nearly
round, mitotic hemocytes were monotonous in appearance and attached only loosely to plastic or glass. Results of immunocytochemistry of freshly collected LCH
indicated positive reactions for -tubulin (Figure 4b and
inset), Mamot (Figure 4c), and Map53 (Figure 4d) proteins in the cytoplasm; none of these proteins was detected in the nucleus. Both Map53 and Mamot proteins
co-localized near the centriole (perinuclear) or in relationship to the well developed microtubular array (pancytoplasmic); such overlap can be detected by the yellow
color in Figure 4e. Western blot analysis of the distribution
of Map53 confirms its sequestration in the cytoplasm in
LCH (Figure 2). LCH in vitro were nearly round, averaged
5 to 7 m in diameter, attached weakly to plastic and
glass by short pseudopodia, and had obvious mitotic
figures (Figure 4a). Distributions of -tubulin, Mamot, and
Map53 in the cytoplasm and nucleus of LCH in vitro were
identical to those seen in LCH in vivo.

Distribution of p53 and Mortalin in NCH in Vivo

Normal clam hemolymph contained amitotic, terminally
differentiated, agranular, and granular NCH (100% NCH
at 1 to 6 106 cells ml1) with an average diameter of
7 to 9 m (Figure 1). Immunocytochemistry10 and Western blotting (Figure 2) indicated positive reactions for
Map53 protein in the cytoplasm and the nucleus of
freshly collected NCH, whereas Mamot was only identified in the cytoplasm of NCH.

Figure 3. Co-immunoprecipitation of Mamot and Map53 in NCHs and LCHs.

Std, protein standard; C1&2, first and second elute from LCH lysates; N1&2,
first and second elute from NCH lysates; NA, first elute from inactive gel
loaded with NCH and LCH lysate mix (negative control); Q, first elute from
quenched gel loaded with NCH and LCH lysate mix (negative control); Ccyt,
cytoplasmic protein extract from LCHs; Cnuc, nuclear protein extract from
LCH.

Cytoplasmic Sequestration of p53

1529
AJP May 2006, Vol. 168, No. 5

Figure 4. a: Romanovski-stained cytospin of freshly collected LCHs; inset shows similarly stained cytospin of LCHs maintained in vitro (arrows point out mitotic
divisions). b: Cytospin of several LCHs showing localization of -tubulin using indirect immunofluorescence (dark complex to one side of the nucleus). c e:
Quantum-dot immunocytochemical localization in LCH of Mamot (c, notice concentration near the centriole, light green dot), of Map53 (d) and merge to show
cytoplasmic co-localization of Mamot and Map53 (e, yellow indicates co-localization of the two proteins). f: Disruption of Mamot and Map53 binding and
translocation of some Map53 to the nuclei of cells treated with 3.5 mol/L MKT-077. N, nucleus in all figures. Scale bars, 20 m.

1530
Walker et al
AJP May 2006, Vol. 168, No. 5

Treatment of LCH with MKT-077 Disrupts

Mortalin/p53 Co-Localization
Treatment of LCH with 1 to 7 mol/L MKT-077 displayed
a dose-dependent increase in the translocation of Map53
protein into the nucleus at 6 to 8 hours (Figure 4f). A dose
of 3.5 mol/L MKT-077 was designated as the LC50 (50%
of cells undergoing apoptosis based on observations
from the TUNEL assay) (data not shown).

Discussion
In this study, immunocytochemistry, co-immunoprecipitation, and subcellular localization of clam p53 (Map53)
and clam mortalin (Mamot) proteins before and after
treatment with MKT-077 were determined for normal
(NCH) and leukemic (LCH) clam hemocytes in vivo and in
vitro. We show that Map53 and Mamot proteins formed
complexes in the cytoplasm of LCH and not in NCH and
that mortalin-based cytoplasmic sequestration of Map53
could be disrupted by MKT-077.
Cytoplasmic sequestration of Map53 protein is a characteristic feature of the phenotype of LCH,10 and Map53
and Mamot are often concentrated near the centriole or
mirror that of the well developed microtubular array. This
distribution of Map53 protein reflects the pattern established by mortalin protein tethers in mice and humans.
Cytoplasmic sequestration leads to inactivation of p53
family proteins in all of these organisms and is especially
common in human lymphomas and neuroblastomas (up
to 96% in undifferentiated neuroblastomas).11,12 Cytoplasmic tethering of p53 is also mediated by mortalin
and/or Parc protein anchors in undifferentiated neuroblastoma cells in mice and humans.13
In mammals, reversal of cytoplasmic sequestration of
wild-type p53 protein can be achieved by a variety of
means all of which result in translocation of p53 to the
nucleus. Because these are wild-type proteins, their renewed presence in the nucleus reactivates them to direct
p53-related transcriptional activities, usually resulting in
apoptosis of the cancer cells. In all cases so far observed, reversal of cytoplasmic sequestration and resultant nuclear function of p53 protein requires either a reduction in the amount or availability of the cytoplasmic
anchoring protein or in DNA damage-dependent overproduction and activation of p53 family proteins.
In leukemic clams, treatment with the rhodocyanin dye
MKT-077 results in translocation of some Map53 protein
from the cytoplasm to the nucleus and ultimately in apoptosis of LCH. Competition between Map53 and MKT077 for the mortalin p53 binding site explains this response after treatment with MKT-077. This treatment
results in a reduction of the effective concentration of
Mamot in LCH. Because Map53 is sequestered in the
cytoplasm, co-localizes with Mamot (which has a binding

site for p53 that is 97% conserved with that of human

mortalin), co-imunoprecipitates with Mamot, and is translocated to the nucleus after treatment with MKT-077, we
suggest that Map53 is tethered in the cytoplasm of LCH
by Mamot in a manner similar to that seen for p53 and
mortalin proteins in a number of unrelated human
cancers.

Acknowledgments
We thank Drs. John Hranitz and Kris Brubaker from
Bloomsburg University for their help with the detection of
co-immunoprecipitated Mamot using chemiluminescent
Western blots; Brian Ricketts for accomplishing the early
sequencing of the Mamot cDNA; and Dan Aird and Kristin Rier for performing the affinity purification of the
Mamot antibody.

References
1. Nikolaev AY, Li M, Puskas N, Qin J, Gu W: Parc: a cytoplasmic anchor
for p53. Cell 2003, 112:29 40
2. Dundas SR, Lawrie LC, Rooney PH, Murray GI: Mortalin is overexpressed by colorectal adenocarcinomas and correlates with poor
survival. J Pathol 2005, 205:74 81
3. Wadhwa R, Taira K, Kaul SC: Mortalin: a potential candidate for
biotechnology and biomedicine. Histol Histopathol 2002,
17:11731177
4. Walker HA, Lorda E, Saila SB: A comparison of the incidence of five
pathological conditions in soft-shell clams, Mya arenaria, from environments with various pollution histories. Mar Environ Res 1981,
5:109 123
5. Reinisch CL, Charles AM, Stone AM: Epizootic neoplasia in soft shell
clams collected from New Bedford Harbor. Hazard Waste 1984,
1:73 81
6. Laevitt DF, Capuzzo JM, Smolowitz RM, Miosky DL, Lancaster BA,
Reinisch CL: Hematopoietic neoplasia in Mya arenaria: prevalence
and indices of physiological condition. Mar Biol 1990, 105:313323
7. Sible J, Marsh AG, Walker CW: Effect of extrinsic polyamines on
post-spawning testes of the sea star (Asterias vulgaris)
(Echinodermata): implications for seasonal regulation of spermatogenesis. Invertebrate Reprod Dev 1992, 19:257265
8. Wadhwa R, Sugihara T, Yoshida A, Nomura H, Reddel RR, Simpson
R, Maruta H, Kaul SC: Selective toxicity of MKT-077 to cancer cells is
mediated by its binding to the hsp70 family protein mot-2 and reactivation of p53 function. Cancer Res 2000, 60:6818 6821
9. Giannakakou P, Nakano M, Nicolaou KC, OBrate A, Yu J, Blagosklonny MV, Greber UF, Fojo T: Enhanced microtubule-dependent trafficking and p53 nuclear accumulation by suppression of microtubule
dynamics. Proc Natl Acad Sci USA 2002, 99:1085510860
10. Kelley ML, Winge P, Heaney JD, Stephens RE, Farell JH, Van
Beneden RJ, Reinisch CL, Lesser MP, Walker CW: Expression of
homologues for p53 and p73 in the softshell clam (Mya arenaria), a
naturally-occurring model for human cancer. Oncogene 2001,
20:748 758
11. Ostermeyer AG, Runko E, Winkfield B, Ahn B, Moll UM: Cytoplasmically sequestered wild-type p53 protein in neuroblastoma is relocated to the nucleus by a C-terminal peptide. Proc Natl Acad Sci USA
1996, 93:15190 15194
12. Moll UM, LaQuaglia M, Benard J, Riou G: Wild-type p53 protein
undergoes cytoplasmic sequestration in undifferentiated neuroblastomas but not in differentiated tumors. Proc Natl Acad Sci USA 1995,
92:4407 4411