Báo cáo khoa học: The regulation of the endosomal compartment by p53 the tumor suppressor gene docx

Here, we demonstrate that p53 regulates transcription of the genes TSAP6 and CHMP4C, which enhance exosome production, and CAV1 and CHMP4C, which produce a more rapid endosomal clearance

tumor suppressor gene

Xin Yu1, Todd Riley2,3and Arnold J Levine1,3

1 The Cancer Institute of New Jersey, University of Medicine and Dentistry of New Jersey, New Brunswick, NJ, USA

2 The BioMaPS Institute at Rutgers University, Piscataway, NJ, USA

3 School of Natural Sciences, The Institute for Advanced Study, Princeton, NJ, USA

The p53 protein is a transcription factor that is

acti-vated by a wide variety of stress signals [1] This

results in a transcriptional program that responds to

the stress and returns the cell to homeostasis,

permit-ting it to function normally without errors The most

commonly studied stress is DNA damage, and the p53

program responds with cell-cycle arrest and the

synthe-sis of repair functions, cellular senescence or apoptosynthe-sis

However, the p53 pathway also responds to stress

sig-nals, resulting in a transcriptional program impacting

upon a large number of other cellular processes For

example, the p53 pathway shuts down the insulin-like

growth factor-1 (IGF-1)⁄ AKT-1 ⁄ mammalian target of

rapamycin (mTor) pathways in response to nutrient

starvation [2–4] and activates autophagy, monitors ribosomal biogenesis and regulates the metabolic path-ways [5] The p53 pathway also regulates the synthesis

of cytokines that can attract cells to a senescent signal-ing cell [6–11] Recently, it has become clear that p53-activating stress signals can have an impact upon the endosomal compartment in a cell and alter membrane and vesicle trafficking [3,12], leading to autophagy and exosome production Exosomes are 50–150 nm vesicles generated from the late endosome⁄ multivesicular bodies (MVBs) in a cell by invagination into the MVB, trapping cytoplasmic components and mem-brane proteins Exosomes exit into the extracellular space after MVBs fuse with the plasma membrane

Keywords

EGF receptor; endosomal compartment;

exosome production; internalization; p53

regulation

Correspondence

A J Levine, School of Natural Sciences,

The Institute for Advanced Study, Princeton,

NJ 08540, USA

Fax: +1 609 951 4438

Tel: +1 609 734 8118

E-mail: alevine@ias.edu

(Received 23 September 2008, revised 29

January 2009, accepted 4 February 2009)

doi:10.1111/j.1742-4658.2009.06949.x

The endosomal compartment of the cell is involved in a number of func-tions including: (a) internalizing membrane proteins to multivesicular bodies and lysosomes; (b) producing vesicles that are secreted from the cell (exosomes); and (c) generating autophagic vesicles that, especially in times

of nutrient deprivation, supply cytoplasmic components to the lysosome for degradation and recycling of nutrients The p53 protein responds to various stress signals by initiating a transcriptional program that restores cellular homeostasis and prevents the accumulation of errors in a cell As part of this process, p53 regulates the transcription of a set of genes encod-ing proteins that populate the endosomal compartment and impact upon each of these endosomal functions Here, we demonstrate that p53 regulates transcription of the genes TSAP6 and CHMP4C, which enhance exosome production, and CAV1 and CHMP4C, which produce a more rapid endosomal clearance of the epidermal growth factor receptor from the plasma membrane Each of these p53-regulated endosomal functions results in the slowing of cell growth and division, the utilization of cata-bolic resources and cell-to-cell communication by exosomes after a stress signal is detected by the p53 protein These processes avoid errors during stress and restore homeostasis once the stress is resolved

Abbreviations

ChIP, chromatin immunoprecipitation; Chmp, charged multivesicular body protein; EGFR, epidermal growth factor receptor; IGF-1, insulin-like growth factor; mTOR, mammalian target of rapamycin; MVB, multivesicular body.

These vesicles can communicate with the immune

sys-tem (dendritic cells) immunizing the host, fuse with

adjacent cells, presumably communicating

physiologi-cal signals, or contribute to the extracellular matrix

[13,14] The increased rate of exosome production in

cells with an activated p53 response is caused, in part,

by the p53-regulated gene TSAP6, a member of the

Steap family of proteins (Steap3), which functions in

an unknown way to enhance exosome production

[12,15–19]

Endosomes have a number of functions in a cell

Endosomal vesicles sample the environment and bring

components to the lysosome for degradation They

transport membrane proteins, including receptors for

growth or cell maintenance, to the intracellular

com-partments (MVBs and lysosomes) They internalize

receptors that have engaged their ligands and are

sig-naling, either reducing the signals or setting up new

locations for signaling Some of these receptors are

degraded in lysosomes, whereas others are trafficked

back onto the cell surface in a regulated process

Reg-ulation of these processes permits cells to be responsive

to outside signaling or to ignore such messages

The MVB contains a set of  30 different sorting

proteins that are quite conserved from yeast to humans

[20–23] The mammalian MVB is composed of several

sets of protein complexes termed Stam⁄ Hrs, ESCRT-I,

ESCRT-II, ESCRT-III and Vps4 [20,21] These

com-plexes are sequentially recruited to the site of MVB

formation and result in the progressive trafficking of

vesicles (cargo) through this organelle It is during this

process that decisions are made to traffic cargo outside

the cell (exosomes), into the plasma membrane for

degradation (lysosomes) or into an autophagic vesicle

The ESCRT-III protein complex on the MVB is

com-posed of a series of charged MVB proteins (Chmp)

1A, 1B, 2A, 2B, 3, 4A, 4B, 4C, 5 and 6 [20,21] The

experiments presented here demonstrate that

CHMP4Cis a p53-regulated gene whose transcription

and protein increase after a p53 stress response This is

correlated with higher rates of exosome production

and faster rates of clearing the epidermal growth factor

receptor (EGFR) from the plasma membrane

There are at least two routes via which to clear

pro-teins such as EGFR from the plasma membrane into

the cell: clathrin-coated pits or a caveolae-mediated

pathway [24,25] In this study, the caveolin-1 (CAV1)

gene, encoding one of three caveolin proteins, is shown

to be a p53-regulated gene The EGFR and caveolin-1

proteins colocalize in the plasma membrane and the

EGFR is then internalized at a faster rate after a p53

stress response, demonstrating for the first time that

the p53 response down-modulates the availability of

growth receptors at the cell surface, making the cell less sensitive to growth and division signals Interest-ingly, the CAV1 gene has been called by some a tumor suppressor gene that is absent in some breast cancer cells [26] In some animal models that deleted the CAV1 gene, animals were more susceptible to oncogene- or carcinogen-induced tumorigenesis [27] However, caveolin-1 protein levels were found to be very high in some multidrug resistant cells [28], aggres-sive prostate cancers [29] and malignant breast lesions [30,31] Clearly, this is not a consistent pattern of observations from which to draw any firm conclusions Thus, the functions of the endosome compartment, exosome production, endosome production, and the regulation and recycling of cell-surface receptors all increase after a p53 response to stress The net result is

to shut down growth and division, utilize the cell’s reserves and communicate stress signals to other cells In this fashion, the p53 protein helps to down-modulate cell growth and division after stress, and utilizes cellular reserves to maintain cells during periods of stress

Results

CHMP4C is a p53-regulated gene Previous experimental results [2,3,12] have identified the endosome compartment of the cell as a place where several types of cellular stress are responded to by a p53-mediated transcription of genes that enhances endosomal functions such as autophagy and exosome production For this reason, the DNA sequences in the promoter–enhancer regions of all known genes for endosome compartment components in the human cell were screened for potential p53 DNA-binding sites To carry out this screening, we developed an algorithm designed to detect p53 regulatory DNA sequences (p53MHH) The position of )512 to )450 nucleotides 5¢ to the transcriptional start site of the first exon of the CHMP4C gene was a perfect match to a p53 DNA-binding site, with the two sites separated by an

18 bp spacer (Fig 1A) CHMP4C is part of the ESC-RT-III protein complex that is essential for endosome function in a cell Regulation of CHMP4C expression

by p53 was tested in the human cell lines H460 (wild-type p53) and H1299 (null-p53) For this we used quantitative real-time PCR to follow the steady-state levels of CHMP4C mRNA in the cell at 24 h after irradiation Activation of p53 by irradiation increased the levels of CHMP4C RNA by 4.5-fold in H460 cells, although H1299 cells did not show increased CHMP4C RNA (Fig 1B) A different stress agent,

0 1 2 3 4 5

H460

0 1 2 3 4 5

0 0.5 1 1.5 2 2.5 3 3.5 4

No treatment IR H1299

0 0.5 1 1.5 2 2.5 3 3.5 4

No treatment IR

M

300 bp

200 bp

Input -A

H1299

- +

H460

0 1 2 3 4 5 6 7 8 9

pGL3-Vector pGL3-promoter

pGL3-CHMP4Cseq

wt-p53 mt-p53 (22/23)

mt-p53 (273)

-No treatment Etop

No treatment Etop 0

50 100 150 200 250

siRNA-NS siRNA-p53

p21 in H460

0 0.5 1 1.5 2 2.5 3

siRNA-NS siRNA-p53 p53 in H460

-0

1

2

3

4

5

6

7

siRNA-NS siRNA-p53

CHMP4C in H460

A

B

Fig 1 Expression of CHMP4C was regulated by p53 (A) A p53 recognition sequence (p53RE) was identified by the algorithm p53MHH Uppercase letters represent the two repeats of PuPuPuC(A ⁄ T)(T ⁄ A)GPyPyPy in the p53RE The lowercase letters within the repeats repre-sent the spacer The lowercase letters on the flank of the repeats reprerepre-sent the flanking sequences around the p53RE (B) Regulation of gene expression was measured by real-time PCR after the cells (H460 and H1299) had been treated with c-radiation and the cells (H460) had undergone transfection of siRNA against nonspecific sequence (siRNA-NS) or siRNA against p53 (siRNA-p53), followed by etoposide treatment (C) Putative p53RE was able to be bound by the p53 protein shown in the ChIP assay The samples were the input, the reco-vered DNA from incubation with no antibody, with IgG or with antibody against p53 (DO-1) Both H460 and H1299 cells were treated with irradiation (D) DNA sequence, including the p53RE, induced luciferase activity with co-transfection of wild-type p53, but not with p53 mutants (22 ⁄ 23) and (273), in the luciferase activity assay The pGL3–vector and pGL3–promotor plasmids were tested as parallel controls to the constructs of pGL3–p53RE in the CHMP4C sequence.

etoposide, also increased CHMP4C mRNA levels

five-fold in H460 cells, and this induction was decreased

nine-fold by an siRNA directed against the p53 mRNA

(Fig 1B) When V138 cells, a cell line with a

tempera-ture-sensitive p53 protein, were shifted from a

nonper-missive temperature to a pernonper-missive one, the levels of

CHMP4CRNA increased by 2.5-fold (data not shown)

These data make it clear that the p53 protein can

regu-late the levels of this endosomal protein after activation

of p53 by a series of diverse stress exposures in several

different cell lines Cell lines with a mutant p53 gene

(e.g H1299⁄ V138 with a ts mutation in Val138 of p53

protein, or H1299 with a deletion in p53 gene) failed to

regulate CHMP4C We used chromatin

immunoprecipi-tation (ChIP) to show that, after a stress response in

H460 cells, the p53 protein could be shown to bind to

chromatin in the )512 to )450 nucleotide region, the

predicted DNA site that regulates the CHMP4C gene

(Fig 1C) In order to test the p53-dependent

transcrip-tional activity through this putative p53 responsive

element, a construct containing this sequence cloned in

front of a luciferase expression vector was introduced

into a p53-null H1299 cell, with co-transfection of a

wild-type p53 expression vector There was a seven-fold

increase in luciferase activity By contrast, there was no

increase in luciferase activity when it was co-transfected

with a vector of no p53 expression, or two different

p53-mutant vectors (codon 22⁄ 23 mutant or codon

273 mutant) of the p53 protein (Fig 1D) Clearly,

CHMP4Cis a p53-regulated gene

Chmp4C plays an important role in exosome

production

Previously, it has been shown that exosome production

is regulated by p53 [12] In p53-null cells, no exosomes

were detected in the cell media, even after c-radiation

[12] Exosome production is conveniently isolated by

differential centrifugation to pellet the exosomes from

the culture medium The many cellular proteins in an

exosome preparation can then be visualized either by

staining the proteins (Fig 2A) or by using an antibody

to Hsp90b or PGK1 (Fig 2B) as a marker to show

exosome production after separating by SDS⁄ PAGE

H1299 cells with no p53 fail to produce exosomes with

or without c-radiation, whereas the introduction of a

wild-type p53 expression vector into these same cells

produces high levels of exosomes, as shown in a

SDS⁄ PAGE by staining [12] The addition of the

CHMP4CcDNA (added as a YFP–CHMP4C so as to

visualize Chmp4C protein) to H1299 cells restored

exo-some production in the medium, as measured by either

stained proteins in exosomes (Fig 2A) or western blots

for Hsp90b and PGK1 (Fig 2B) In addition, the Chmp4C protein was detected in both the cell extract and in exosomes (Fig 2B)

In H460 cells, in which exosomes were produced only after activation of p53 (by etoposide) (Fig 2C), pretreatment of the cells with siRNA directly against CHMP4C, followed by exposure with etoposide, led to

a failure to produce exosomes (Fig 2C) Thus, it is clear that the p53-mediated increase in the transcrip-tion rate of CHMP4C is required to increase the levels

of exosome production Two different p53-regulated genes, TSAP6 and CHMP4C, can each increase the rate of exosome production when introduced sepa-rately into cells with no p53 [12,15,16] Although there was an occasional overlap in the localizations of the two proteins, Tsap6 and Chmp4C, in some cells by fluorescent staining, no consistent evidence was found that these two proteins act together in a complex in a cell (data not shown) This is interpreted most simply

as either CHMP4C and TSAP6 are on different, but parallel, pathways for exosome production, or they are

in the same pathway, but each raises the rate of pro-duction of exosomes (two distinct rate-limiting steps)

p53 regulates the internalization of EGFR from the plasma membrane into the endosome Because the p53 response regulates the activity of the endosome compartment, and endosomal processes reg-ulate the levels and activity of growth factor receptors

at the cell surface, we next explored whether a p53 response could accelerate removal of the EGFR from the plasma membrane In H460 cells, the localization

of the EGFR was determined by fluorescent immuno-staining In the absence of etoposide, EGFR was all at the plasma membrane (Fig 3A) After treatment with etoposide, EGFR molecules progressively moved into the internal compartments of the cells so that by 6–8 h most of the EGFR was internalized (Fig 3A) In order

to determine if EGFR molecules were moving into the endosome, the molecules were stained with red fluores-cence and several endosomal proteins were counter-stained with green fluorescence Figure 3B shows the staining of the TfR protein, an early endosome compo-nent, whereas Fig 3C shows staining of the LAMP1 protein, a late endosome–lysosome protein, and Fig 3D shows staining of Chmp4C in the MVB late endosome Clearly, EGFR proteins become progres-sively associated with the different endosomal com-partments, with a rather clear colocalization with Chmp4C in the MVB

To confirm that p53 activity was responsible for the rapid clearance of EGFR from the plasma membrane,

we tested for EGFR internalization in H1299 cells,

which do not express p53 protein Even though the

H1299 cells were treated with etoposide, the location

of the EGFR molecules did not change from

predomi-nantly the plasma membrane to the cytoplasm

(Fig 4A) We also tested EGFR internalization in H24

cells in which p53 expression is controlled by the

pres-ence of tetracycline With tetracycline withdrawal, the

p53 expression is increased (Fig 4B), resulting in

EGFR internalization in these cells (Fig 4C)

Colocal-ization of EGFR with the endosomal compartment

proteins TfR and LAMP1 was also observed in these

experiments (Fig 4C) Clearly, EGFR internalization

from the plasma membrane into the endosomal

com-partment can be regulated by a p53 response Removal

of the EGFR from the cell surface by a p53-responsive

mechanism was shown to occur in several very

differ-ent cell lines: H460 treated with a DNA damaging

agent, H24 with a Tet-off controlled p53 expression,

and V138 with a temperature-sensitive p53 protein

(data not shown); this failed to occur in cells without a

p53 gene (H1299 cells) The EGFR protein level

decreased upon p53 activation, as determined by

western blot (Fig S1), and this is consistent with

the observation of EGFR clearance from the plasma

membrane to the endosome compartment and then to the lysosomes for degradation

Caveolin-1 expression is regulated by p53 Receptors such as EGFR may utilize a caveolae-medi-ated pathway for internalization [32] Caveolin-1 is the major component of caveolae [33] Based solely upon increased levels of CAV1 mRNA [34–36], or, separately reported, p53-binding assays (EMSA and luciferase assay) [34–36], it had previously been reported that CAV1 is a p53-regulated gene However, the CAV1 gene has a rather poor consensus p53 DNA-binding site in the 5¢ location (Fig 5A), which is nonetheless predicted to be a p53-responsive element by the p53MH algorithm [37] When H460 cells were treated with etoposide or radiation the steady-state levels of CAV1 mRNA increased, by seven- and two-fold respectively, at 24 h after treatment (Fig 5B) Simi-larly, when V138 cells were shifted to the permissive temperature for activation of p53 protein, there was a five-fold increase in the levels of CAV1 mRNA at 24 h after the temperature shift (Fig 5B) Irradiation of H1299 cells with no p53 protein failed to increase the levels of CAV1 mRNA in those cells (Fig 5B)

YFP- Chmp4C

210

125

80

49

35

29

21 7.1 kDa

Exosomes

YFP- Chmp4C

GAPDH

Cell lysates

YFP- Chmp4C

PGK1

Exosomes

YFP- Chmp4C

YFP- Chmp4C

210

125

80

49

35

29

21 7.1 kDa

Exosomes

H460

siRNA- Chmp4C

Fig 2 Exosome production was regulated by expression of Chmp4C (A,B) Overexpression of Chmp4C restored exosome production (A) Silver staining of a typical SDS ⁄ PAGE of the exosome production isolated by ultracentrifugation from H1299 cells (p53 null), with or without overexpression of Chmp4C (B) Western blot of the isolated exosomes and the cell lysates Hsp90b and PGK1 were used as markers for exosome production (C) Suppression of exosome production by knockdown of Chmp4C expression The silver staining of an SDS ⁄ PAGE is shown ), no etoposide (Etop) addition; +, the addition of etoposide M, molecular markers.

Irradiation of HCT116 cells with a wild-type p53 gene

increased gene expression by four-fold, but the isogenic

cell line without a p53 gene failed to produce more

CAV1mRNA after this treatment (Fig 5B)

Caveolin-1 protein levels in H460 and VCaveolin-138 cells were induced

after p53 activation (Fig S1) In H460 cells, ChIP with

a p53-specific antibody (DO-1) immunoselected the

same region of DNA predicted in Fig 5A to regulate

this gene, whereas a no-antibody control failed to

detect this DNA, and H1299 cells with no p53 protein

also failed to bind to this DNA (Fig 5C) Clearly, this

DNA sequence can bind the p53 protein after a stress

signal When this sequence ()202 to )185 bp in

Fig 5A) was cloned and placed into a luciferase expres-sion vector it stimulated luciferase activity more than 80-fold compared with wild-type p53 protein, but not compared with the two different p53 mutant proteins that fail to stimulate p53-regulated transcription (Fig 5C) Clearly, these additional criteria demonstrate that the p53 protein regulates the CAV1 gene at the promoter region ()202 to )185 bp), rather than at the reported sequence ()297 to )259 bp) [34–36], increas-ing its rate of transcription

It has previously been reported that caveolin-1 interacted with EGFR under various conditions [38– 40] Using coimmunoprecipitation, we provided

0 h

6 h

4 h

2 h

C

0 h

6 h

4 h

2 h

EGFR TfR EGFR / TfR

B

D

A

Fig 3 p53 activation promoted EGFR internalization through the endosome compartment (A) H460 cells were treated with etoposide and

at 0, 2, 4, 6 and 8 h cells were washed and stained for EGFR (red) (B) H460 cells treated with etoposide were stained for EGFR (red) and TfR (green) (C) H460 cells treated with etoposide were stained for EGFR (red) and LAMP1 (green) (D) H460 cells transfected with YFP– CHMP4C, followed by etoposide treatment for 8 h, were stained for EGFR (red) The location of Chmp4C protein was visualized in green Bars, 10 lm.

evidence that, in the H460 cell line, caveolin-1

inter-acted with EGFR (Fig S1) To test the role

caveolin-1 may have in EGFR clearance from the plasma

membrane after p53 activation, we treated H460 cells

with etoposide and costained the cells with antibodies

against EGFR and caveolin-1 The green fluorescent

signals represent EGFR and the red signals represent

caveolin-1 (Fig 5D) Without etoposide treatment,

EGFR and caveolin-1 are localized mainly on the

plasma membrane, and both molecules show

signifi-cant overlap in their locations at the membrane

(Fig 5D, 0 h, no etoposide treatment) With time

after etoposide treatment and p53 activation (4, 6

and 24 h), caveolin-1 has a stronger signal, forms

patched structures and moves into the cell (Fig 5D,

b, e, h, k) At the same time, EGFR also changes to

a more granulated appearance and moves into the cell (Fig 5D, a, d, g, j) Merger of these two mole-cules shows a progressive colocalization inside the cell

in the endosomal compartment with increasing time (Fig 5D, c, f, i, l) A similar experiment was carried out with V138 cells (ts p53) employing a temperature shift from a nonpermissive to a permissive tempera-ture for p53 activity The results confirmed the con-clusions presented in Fig 5D, activation of p53 increased the removal of both EGFR and caveolin-1 proteins from the surface and they colocalized within the cell These data provide clear evidence, in several independent cell lines with several diverse ways to activate p53, that the internalization of EGFR and caveolin-1 from the cell surface can be mediated by gene products produced after the activation of p53

Day 0

Day 2

Day 1

d

e

f

Relative expression level 0 2 4 6 8 10 12

LAMP1

24 h

4 h

0 h

A

B

C

Fig 4 EGFR internalization was mediated

by p53 activation (A) H1299 cells (without

p53 expression) were treated with

etopo-side at 0, 4 and 24 h, followed by staining

for EGFR (red) (B) In H24 cells in which

p53 expression was under the control of

tetracycline, the expression of p53 was

determined by quantitative PCR after

tetra-cycline withdrawal (C) H24 cells were

incu-bated with and without tetracycline and

stained for EGFR (red) and TfR (green) or

LAMP1 (green) Bars, 10 lm.

300 bp

200 bp

CAV1 +DO-1

H460

H1299

IR + + +

No Ab

100 bp

M

300 bp

200 bp

CAV1 +DO-1

H460

H1299

IR – + + +

No Ab

100 bp

0

20

40

60

80

100

120

–202 to –185 bp (A) –297 to –259 bp (B)

No p53 wt-p53 mt-p53 (22/23) mt-p53 (273)

No p53 wt-p53 mt-p53 (22/23) mt-p53 (273)

V138

0 1 2 3 4 5 6

39 °C 32 °C

H1299

0 1 2 3

H460

0 1 2 3

0 1 2 3 4 5 6 7 8 9 10

treatment Etoposide

0

1

2

3

4

5

HCT116wt HCT116-p53null HCT116wt HCT116-p53null

No treatment IR

No treatment IR

Intron 1 +1

297 to –259bp (B)

Consensus p53 responsive element (p53RE):

PuPuPuC (A/T)(T/A)GPyPyPy N(spacer) PuPuPuC(A/T)(T/A)GPyPyPy

Intron 1 +1

Exon Exon Intron 1

+1

– 297 to –259bp (B)

––202 to -185bp (A)

CAV1

0 h

4 h

24 h

6 h

A

B

C

D

Fig 5 Caveolin-1 expression was regulated by p53 (A) A putative p53RE was identified The uppercase letters at )202 to )185 bp repre-sent the p53RE predicted by p53MH The uppercase letters at )297 to )259 bp represent the p53RE as reported previously [34,35] The lowercase letters represent the genomic sequence (B) Caveolin-1 expression was measured by quantitative PCR in H460, H1299, HCT116 and V138 cells upon treatment with etoposide, irradiation or temperature shift (C) The putative p53RE was able to be bound by the p53 protein shown in the ChIP assay, and the DNA sequence including the p53RE induced luciferase activity with co-transfection of wild-type p53, but not with p53 mutants (22 ⁄ 23) and (273) in the luciferase activity assay (D) H460 cells were treated with etoposide followed by staining for EGFR (green) and caveolin-1 (red) Bar, 10 lm.

The p53 protein responds to stress signals in a cell by

initiating cell-cycle arrest, senescence or apoptosis At

the same time, the p53 transcriptional program shuts

down the IGF-1⁄ AKT-1 and mTOR pathways [2,3] and

activates several of the endosomal compartment

activi-ties including autophagy and exosome production [12]

As part of this p53 program to down-modulate cellular

growth and division, the levels of several proteins in the

endosomal compartment are increased (caveolin-1,

Chmp4C), resulting in lower amounts of cell surface

growth receptors (EGFR) which are internalized and

sent to the MVB At the same time, there is an increased

rate of exosome production, which results from the

stimulation of the transcription of TSAP6 and

CHMP4Cby p53 These events communicate a cellular

stress event to the immune system, adjacent cells and the

extracellular matrix Thus, the endosomal compartment

participates in a coordinated response to both shut

down cellular processes after stress and to alert adjacent

cells and the immune system of these events (Fig 6)

The p53 network also alters several other

physiologi-cal processes in the cell that are undoubtedly related

to the functions of the endosome compartment p53

regulates genes encoding proteins that alter glycolytic

activities and oxidative phosphorylation [5] Thus, a

p53-responsive stress signal can result in: (a) the cell

shutting down its commitment to cell growth and

divi-sion; (b) the removal of growth receptors from the cell

surface; (c) an increase in the rate of exosome

produc-tion signaling to surrounding cells and the immune

system (along with secreting cytokines); (d) alteration

of its metabolic and energy production sources; and (e)

depending upon the cell type and whether it is a

nor-mal cell or cancer cell, the cell undergoing cell-cycle

arrest, senescence or apoptosis and autophagy

Throughout this process, there is an elaborate set of

negative and positive feedback loops to regulate and

increase or decrease p53 levels and its activities [41]

We are beginning to appreciate the coordinated nature

of these networks and how each p53-regulated gene fits

into this picture Clearly, an integral part of this

coordi-nated system is the p53-regulated control of the

endo-somal compartment of the cell (Fig 6) The results

presented here begin to outline the way in which the p53

protein regulates the transcription of selected genes to

accomplish this integrated response There is an

interest-ing level of redundancy in these endosomal-regulated

processes: (a) four p53-regulated genes turn off the

IGF–mTOR pathways; (b) autophagy is activated by the

negative control of mTOR and the positive control of

an autophagy gene MAP1LC3A (unpublished data); (c)

exosome production is stimulated by both Chmp4C and Tsap6; and (d) both Chmp4C and caveolin-1 enhance removal of the EGFR from the cell surface These activi-ties all function to slow cell growth and division, con-serve and reutilize cellular resources, and notify other cells and organ systems (the immune response) about the stresses These functions are also an important part of cell and tissue repair after cell damage (DNA or chemical damage), virus infection or hypoxia This coordinated effort by the p53 pathway integrates the molecular, cellu-lar and systemic levels of activities and demonstrates how a stress response is independent of scale The endo-somal compartment of a cell, regulated by its protein constituents, can coordinate interactions at each of these scales and respond to stress in a p53-regulated fashion

Experimental procedures

Cell culture, DNA damage treatment and transfection

H460 and H1299 cells were cultured as described previously [12] H1299⁄ V138 cells (from J Chen, H Lee Moffitt Cancer

Stress signals (such as DNA damage)

Upstream mediators (such as ATM/ATR, Chk1/Chk2)

p53 Mdm2

Core regulation

Cell cycle arrest Apoptosis Senescence

Chmp4C Caveolin-1 TSAP6

Autophagy

Lysosome function

Exosome secretion

Communication between cells

Receptor endocytosis, Protein trafficking Endosome functions

Downstream effectors and pathways

Fig 6 p53 regulation of cellular pathways upon stress responses p53 activation by stress signals regulates not only cell-cycle arrest and apoptosis, but also endosome functions and autophagy which are involved in protein trafficking and signal transferring inside the cell and between the cells See text for details.

Center, FL, USA), with a stably transfected

temperature-sensitive mutant form of p53 (Ala138 to valine) into H1299

cells [42], were cultured in Dulbecco’s modified Eagle’s

med-ium, supplemented with 10% fetal bovine serum and

500 gÆmL)1 G418 The H24 cell line was from C Prives

(Columbia University, NY, USA), established to express

tet-racycline-regulated p53 [43] HCT116 p53+⁄ +and HCT116

p53) ⁄ )cells (from B Vogelstein at John Hopkins University,

USA) were cultured in McCoy’s 5A with 10% fetal bovine

serum All cells were grown at 37C with 5% CO2 The cells

were treated with DNA damage reagent, 20 lm etoposide or

irradiated with 5 Gy as described previously [12]

pRC⁄ CMV-wt p53 and mutant p53 (mt22 ⁄ 23 and

mt273H) expression plasmids were generated as described

previously [44] The plasmids of pcDNA-3.1–HA–TSAP6

and the vector were from A Telerman (Molecular

Engines Laboratories, France) The plasmid of YFP–

CHMP4C was from P Bieniasz (Rockefeller University,

USA) The pGL3 luciferase reporter vectors (pGL3-Basic

and pGL3-promoter vectors) were purchased from

Pro-mega (Madison, WI, USA) siRNA against CHMP4C

and p53 was purchased from Dharmacon (Chicago, IL,

USA) The siRNA for a nonspecific siRNA (NS) was

purchased from Qiagen (Valencia, CA, USA) Plasmids

were transfected with Lipofectamin 2000 (Invitrogen,

Carlsbad, CA, USA) or jetPEI (ISC Bioexpress,

Kays-ville, UT, USA), and the siRNA was transfected with

oligofectamin (Invitrogen)

Exosome isolation

Exosome isolation was carried out as described previously

[12]

Western blot

Cell lysates were made as described previously [12] The cell

lysates or isolated exosomes were run on SDS⁄ PAGE

(4–20%) (Invitrogen) and transferred to Immobilon-P

mem-branes (Millipore) Memmem-branes were blotted with the

following antibodies (Santa Cruz Biotechnology, Santa

Cruz, CA, USA): Hsp90b (D-19, sc-1057); PGK1 (Y-12,

sc-17943); GFP (FL, sc-8334); and GAPDH (FL-335,

sc-25778)

Real-time PCR

Total RNA was isolated at different time points using a

RNeasy Kit (Qiagen) followed by one-step reverse

tran-scription (TaqMan Reverse Transcription Reagents;

Applied Biosystems, Foster City, CA, USA) and real-time

PCR (ABI Prism 7000 sequence detection system; Applied

Biosystems) b-Actin was used as an internal control

Probes and primer sets were purchased as predeveloped

assays from Applied Biosystems Triplicate samples were taken and each experiment was repeated The relative induction⁄ repression level was calculated by the ratio of the value of the gene to that of b-actin and then to the controls

ChIP assay ChIP assays were performed using a Upstate ChIP Assay Kit (Lake Placid, NY, USA) according to the manufac-turer’s instructions The antibody against p53 (DO-1, sc-126) was from Santa Cruz Biotechnology The primer sets were designed to encompass the potential p53-binding elements in the human CHMP4C and CAV1 genes The sequences for the promoter region of CHMP4C gene are

as follows: 5¢-CCTGACATTAGGAAAAGAGATGGCC-3¢ and 5¢-ATGAGTGTGTGGACACAAAGGCTTCC-3¢ The sequences for the CAV1 gene are as follows: 5¢-CGGGG TACCGGGAAAATTGTTGCCTCAGG-3¢, 5¢-CCGCTCG AGGGTTTGTTCTGCTCGCGG-3¢ (A) and 5¢-CCGCTC GAGCCCCAAGGTTCTGGCAGCAG-3¢ (B)

Luciferase activity reporter assay H1299 cells (p53 null) were plated in a 12-well plate, one day before transfection Cells were transiently co-trans-fected with the constructed luciferase reporter plasmids pGL3-putative p53RE sequences and either wild-type or

a mutant p53 plasmid, and pRL-TK plasmid (Promega) was used as an internal control Forty-eight hours after transfection, whole-cell extracts were prepared and a luciferase assay was performed according to the manufac-turer’s instructions (Promega) Each transfection was performed with repeats and standard deviations were calculated

Immunofluorescent staining and confocal microscopy

Cells were cultured on glass coverslips, treated with etoposide (20 lm), temperature shifting for various periods

or cultured in the media without tetracycline (see above for details), and rinsed with phosphate buffered saline (NaCl/

Pi) Cells were fixed in 4% paraformaldehyde in NaCl⁄ Pi for 10 min, followed by permeabilization with 0.5% Triton X-100 in NaCl⁄ Pi for 10 min Cells were then incubated with primary antibody for 1 h followed by washing with NaCl⁄ Pi and incubation with Alexa Fluor )555 or )488 conjugated secondary antibody The cells were visualized with a Zeiss Axiovert 200M fluorescence microscope under confocal settings The primary antibodies used included EGFR (1005, sc-03), CD71 (i.e TfR, 3B8 2A1, sc-32272), LAMP1 (H5G11, sc-18821), caveolin-1 (N-20, sc-894) and

HA (Y-11, sc-805)