Báo cáo khoa học: The inhibition of Ras farnesylation leads to an increase in p27Kip1 and G1 cell cycle arrest pdf

Analysis of cell cycle components showed that HR12 treatment of Rat1/ras cells led to elevated cellular levels of the cyclin-dependent kinase inhibitor p27Kip1and inhibition of the kinas

The inhibition of Ras farnesylation leads to an increase

Hadas Reuveni*, Shoshana Klein and Alexander Levitzki

Department of Biological Chemistry, Institute of Life Sciences, The Hebrew University of Jerusalem, Jerusalem, Israel

HR12 is a novel farnesyltransferase inhibitor (FTI) We have

shown previously that HR12 induces phenotypic reversion

of H-rasV12-transformed Rat1 (Rat1/ras) fibroblasts This

reversion was characterized by formation of cell–cell

con-tacts, focal adhesions and stress fibers Here we show that

HR12 inhibits anchorage independent and dependent

growth of Rat1/ras cells HR12 also suppresses motility and

proliferation of Rat1/ras cells, in a wound healing assay

Rat1 fibroblasts transformed with myristoylated H-rasV12

(Rat1/myr-ras) were resistant to HR12 Thus, the effects of

HR12 are due to the inhibition of farnesylation of Ras Cell

growth of Rat1/ras cells was arrested at the G1 phase of

the cell cycle Analysis of cell cycle components showed that

HR12 treatment of Rat1/ras cells led to elevated cellular levels of the cyclin-dependent kinase inhibitor p27Kip1and inhibition of the kinase activity of the cyclin E/Cdk2 complex This is the first time an FTI has been shown to lead

to a rise in p27Kip1levels in ras-transformed cells The data suggest a new mechanism for FTI action, whereby in ras-transformed cells, the FTI causes an increase in p27Kip1 levels, which in turn inhibit cyclin E/Cdk2 activity, leading

to G1 arrest

Keywords: farnesyl transferase inhibitor (FTI); p27Kip1; Ras; cell cycle

Localization of Ras proteins in the plasma membrane

follows a series of post-translational modifications [1] and is

crucial to the functioning of these proteins [2,3] The first

and essential step in this process is farnesylation, whereby a

farnesyl group (C15-isoprenoid) is covalently attached to the

cysteine residue of the C-terminal CAAX sequence of Ras

[4] Farnesylation is mediated by the enzyme

farnesyltrans-ferase (FT) The three C-terminal residues, AAX, are then

proteolytically cleaved and the new carboxy-terminus is

methylated H-Ras, N-Ras and K-Ras4A are also

palmi-toylated on one or more upstream cysteine residues

Mutationally activated ras genes are found in
30% of

all human cancers As farnesylation is required for the

oncogenic activity of activated Ras, there has been much

interest in the development of FT inhibitors (FTIs) for

anticancer treatment

We have developed an FTI, cysteine-N-methyl-valine-N-cyclohexyl-glycine-methionine-methyl ester, called HR12 [5] We have demonstrated recently [6] the compound’s ability to completely reverse the transformed phenotype of oncogenic H-Ras-transformed Rat1 (Rat1/ras) fibroblasts This reversion entailed the assembly of adheren junctions, concomitant with induction of cadherin and b-catenin Focal adhesions and actin stress fibers were formed, and the overall cell morphology was indistinguishable from that of nontransformed Rat1 cells

Cell adhesion affects cell growth and invasion Cadherin, the primary cell–cell adhesion molecule, acts as a suppressor

of cancer cell invasion [7,8], and the loss of cadherin function

is required for tumor progression in vivo [9,10] Moreover, the activation or overexpression of cadherin has been shown

to arrest cell growth at the G1 phase, following an increase in the p27Kip1 level and dephosphorylation of the retino-blastoma protein (pRb) [11,12] The present report shows that HR12 inhibits anchorage dependent and independent growth of Rat1/ras cells, and suppresses motility and proliferation in an in vitro ‘wound healing’ assay We further show that HR12 arrests Rat1/ras cells at the G1 phase of the cell cycle, following up-regulation of the cell cycle inhibitor p27Kip1and down-regulation of the kinase activity of the cyclin E/cyclin-dependent kinase-2 (Cdk2) complex Progression of mammalian cell division through the cell cycle is governed by the sequential formation, activation and subsequent inactivation of Cdk complexes [13] The activation of Cdks depends upon multiple levels of regula-tion: the synthesis of the cyclins and their assembly into cyclin/Cdk complexes [14], the phosphorylation of the Cdks, and the inhibitory action of the Cdk inhibitors (CKIs)

in these complexes [15,16] CKIs identified in mammalian cells are classified into two main categories: the INK4

Correspondence to A Levitzki, Department of Biological Chemistry,

Institute of Life Sciences, The Hebrew University of Jerusalem,

Jerusalem, Israel 91904.

Fax: + 972 2 6512958, Tel.: + 972 2 6585404,

E-mail: levitzki@vms.huji.ac.il.

Abbreviations: Cdk, cyclin dependent kinase; CKI, Cdk inhibitor;

Erk, extracellular-signal regulated kinase; FT, farnesyltransferase;

FTI, farnesyltransferase inhibitor; LLnL,

N-acetyl-leucyl-leucyl-norleucynal; mAb, monoclonal antibody; MAPK, mitogen activated

protein kinase; Mek, MAPK kinase; PI3K,

phosphatidylinositol-3¢OH-kinase; PKB, protein kinase B; pRb, retinoblastoma protein;

Rat1/ras, H-ras V12 -transformed Rat1 cell line; Rat1/myr-ras,

myristoylated H-ras V12 -transformed Rat1 cell line.

*Present address: Keryx Biopharmaceuticals, PO Box 23706,

Jerusalem, Israel.

(Received 11 March 2003, revised 29 April 2003, accepted 1 May 2003)

proteins, which bind to and specifically inhibit Cdk4 and

Cdk6 complexes [17], and the Kip/Cip inhibitors (p21Cip1,

p27Kip1 and p57Kip2) with broader specificity [15]

Over-expression of the CKIs causes G1 arrest [15,17–21]

Ras plays a central role in integrating mitogenic signals

and cell cycle progression Interference with normal Ras

function by injection of anti-Ras Igs or by the expression of

the dominant negative (DN) mutant, RasN17, blocks the

proliferation of NIH3T3 cells [22–24] In particular, Ras

was shown to control cell cycle progression at the early G1

stage by induction of cyclin D1, and to control the

progression and passage through the restriction point at

late G1, by down-regulation of the Cdk inhibitor p27Kip1

[25–27] Expression of DN-RasN17 in fibroblasts caused

p27Kip1 accumulation, resulting in suppression of Cdk

activities and G1 arrest [26,28] Oncogenic Ras-transformed

epithelial and fibroblast cells were shown to express reduced

levels of p27Kip1protein [29] p27Kip1is thus a key factor in

Ras regulation of progression through the late G1 phase

and through the restriction point, the latter being a

prerequisite for entry into the S phase

Reduced expression of the p27Kip1 protein has been

observed in a variety of human malignancies, and in

particular, the progressive loss of p27Kip1 is commonly

observed during the progression from normal cells to

benign and malignant tumors p27Kip1 appears to play a

role in the switch from cell proliferation to differentiation,

and loss of p27Kip1 is associated with a poorly

differen-tiated phenotype in several human malignancies,

suggest-ing that potentiation of p27Kip1might be a useful strategy

in cancer treatment (reviewed in [30,31]) FTIs, which

were designed as inhibitors of Ras localization in the

membrane, have been reported to elevate p21Cip1levels in

Rat1/ras cells [32,33] It has been claimed that the

elevation in p21Cip1levels was mediated by the inhibition

of non-Ras farnesylated proteins [32,34,35] For the first

time, we report here on an FTI that causes p27Kip1levels

in Rat1/ras cells to be elevated in a Ras-dependent

manner, resulting in inhibition of the kinase activity of the

cyclin E/Cdk2 complex We suggest that this is the

mechanism by which HR12 suppresses proliferation and

motility and arrests Rat1/ras cell growth at the G1 phase

of the cell cycle

Experimental procedures

Materials and cell cultures

All cell lines were maintained and treated in growth medium

[Dulbecco’s Modified Eagles Medium (DMEM) containing

10% fetal bovine serum (Biological Industries Bet-Haemek

Ltd, Israel)] Rat1/myr-ras cells were maintained under

G418 selection Rat1 and Rat1/ras cells were described

previously [6] Rat1/myr-ras cells [36] were kindly provided

by Yoel Kloog (Tel-Aviv University, Israel) HR12 was

synthesized and purified as described before [5,6]

Anchorage-dependent and independent cell growth

assays

Colony formation in soft agar was performed essentially

as described previously [37] A suspension of separated

Rat1/ras or Rat1/myr-ras cells was plated in agar at a density of 5 000 cells per well in a 96-well plate in growth medium containing 0.3% agar (50 lL per well), on top of

a layer of growth medium containing 1% agar (100 lL per well) Growth medium (50 lL) supplemented with HR12 at four times the indicated concentration was added on top Seven to nine days after plating, the cells were stained with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT; Sigma) and photographed The color developed by viable colonies was extracted by the addition of 100 lL per well of solubilization buffer containing 20% (w/v) SDS, 50% (v/v) dimethylforma-mide, 2% (v/v) acetic acid and 0.25M HCl Following incubation of the plate at 37C overnight, absorbance at

570 nm was read in an ELISA Reader The assays were performed in triplicates

For anchorage-dependent growth curves, Rat1 (1500 cells/well) and Rat1/ras (3300 cells/well) cells were plated on 96-well plates One day after seeding, cultures were treated with HR12 at various concentrations (triplicate samples were made for each concentration) in growth medium Medium with and without HR12 was replaced after two days Cells were counted 96 h after seeding

In vitro monolayer ‘wound healing’ assay Rat1/ras and Rat1/myr-ras were grown to confluence in

60 mm plates under growth medium, in the presence or absence of 20 lMHR12 Monolayers were wounded using a rubber policeman or a micropipette tip, and visualized using

a phase-contrast microscope Pictures were taken and the wound width was measured at various time points Cell cycle analysis

Rat1/ras and Rat1/myr-ras were grown to subconfluence in growth medium in the presence or absence of 20 lMHR12 for 48 h In the last 30 min of treatment, the cells were exposed to 10 lMbromodeoxyuridine (BrdU; Amersham), followed by harvesting and fixation in 70% ethanol The cells were stained with fluorescein isothiocyanate (FITC)-conjugated anti-BrdU Ig (Dako, Denmark) and propidium iodide (PI, Sigma) as described before [38] A total of 10 000 stained cells were analysed in a fluorescence-activated cell sorter

Immunostaining Immunostaining was conducted as described previously [6] Cells were plated on coverslips in DMEM containing 10% FBS and maintained at 37C with 5% CO2 After seeding (24 h), the medium was replaced with medium containing

20 lM HR12 Twenty four hours later, the medium was again replaced with fresh medium containing 20 lMHR12 Following 48 h of exposure to HR12, cells were fixed at

37C, as follows Cells were washed once with NaCl/Pi, fixed and permeabilized in a solution containing 3% paraformaldehyde, 50 mMMes buffer pH 6, 0.5% Triton X-100 and 5 mMCaCl2, for 30 s, followed by 1 h incuba-tion in the same soluincuba-tion without Triton The fixed cells were incubated with anti-(b-catenin) Ig (Transduction Laboratories, dilution 1 : 20 in NaCl/P) for 30 min at

room temperature, washed three times with NaCl/Piand

incubated with the secondary antibody, Cy3-conjugated

goat anti-(mouse IgG) (Jackson ImmunoResearch

Labor-atories USA, dilution 1 : 80 in NaCl/Pi) Following

stain-ing, coverslips were mounted in Elvanol Fluorescent images

were recorded with a Zeiss Axiophot microscope equipped

for fluorescence using· 66/1.4 or · 100/1.3 objectives

Immunoblotting

Cells were treated with HR12 at various concentrations

in growth medium for 48 h, and lysed in Laemmli sample

buffer (50 mM Tris/HCl, pH 6.8, 5% 2-mercaptoethanol,

3% SDS and 0.5 mgÆmL)1 bromophenol-blue) Aliquots

of cell extracts containing equal amounts of protein were

resolved by SDS/PAGE and electroblotted onto

nitrocellu-lose filters The membranes were blocked with lowfat milk

diluted 1 : 20 in NaCl/Tris containing 0.2% Tween-20

(blocking solution), incubated with primary Igs overnight at

4C, and then with horseradish peroxidase-conjugated

secondary antibodies for 75 min at room temperature

Immunoreactive bands were visualized using enhanced

chemiluminescence, and quantified using NIH-IMAGE1.61

program (http://rsb.info.nih.gov/nih-image/) Each

experi-ment was repeated at least two times Each figure shows a

representative blot, and its correspondingNIH-IMAGE

ana-lysis Arbitrary values are shown, except where otherwise

stated

Anti-p27Kip1monoclonal antibodies (mAb; cat#K25020)

and anti-(Rap1A/K-rev) mAb (cat#R22020) were provided

by Transduction Laboratories (KY) Polyclonal Igs against

cyclin D1 (M20), p21Cip1(C-19), Cdk2 (M-2) and Cdk6

(C-21) came from Santa-Cruz Biotechnologies Anti-Ras Ig

was produced from hybridoma Y13-259 Polyclonal

anti-phospho-pRb(Ser795) Ig came from New England BioLabs

(MA) Monoclonal anti-pRb (G3-245) came from

Pharmi-gen (San Diego, CA, USA) Monoclonal anti-cyclin E

(HE12) came from Upstate Biotechnology

Immunoprecipitation

Rat1/ras cells were treated with 20 lM HR12 in growth

medium for 24 and 48 h, and lysed at 4C in lysis buffer

containing 50 mMTris/HCl pH 7.5, 1 mM EGTA, 1 mM

EDTA, 1% Triton X-100, 0.27Msucrose, 1 mM sodium

orthovanadate, 20 mM2-glycerophosphate, 50 mMsodium

fluoride, 5 mM sodium pyrophosphate, 10 lgÆmL)1

soy-bean trypsin inhibitor, 10 lgÆmL)1 leupeptin, 1 lgÆmL)1

aprotonin, 313 lgÆmL)1benzamidine, 0.2 mM

4-(2-amino-ethyl)-benzenesulfonylfluoride (AEBSF) and 0.1%

2-merca-ptoethanol Lysates were centrifuged at 19 000 g for 10 min

and supernatants were subjected to immunoprecipitation

For each sample, 75 lL of 10% protein A–Sepharose were

incubated for 1 h at 4C with 2 lg of either anti-(cyclin E)

(M-20), Cdk2 (M-2), (cyclin D1) (72–13G) or

anti-Cdk6 (C-21) Igs Antibodies used for immunoprecipitation

were purchased from Santa Cruz Biotechnology

Super-natants (500 lg of each) were incubated with the Ig-coupled

protein A for 1 h at 4C As negative controls, Igs were

‘blocked’ by the inclusion of 2 lg of blocking peptide during

coupling The immunoprecipitates were washed twice with

lysis buffer and once with kinase buffer containing 50 m

Hepes pH 7.4, 10 mM magnesium acetate, 1 mM dithio-threitol and 1 lMATP

Cyclin-dependent kinase (Cdk) assays The assay for Cdk2 activity was performed by adding

40 lL of kinase buffer containing 10 lCi [c32P]ATP and 2.5 lg histone H1 (freshly prepared) to the anti-(cyclin E) immunoprecipitates The activities of Cdk4 and Cdk6 were measured as follows: 40 lL of kinase buffer containing 5 lCi [c32P]ATP and 1 lg of GST-pRb (C-terminal fragment of pRb; Santa Cruz Biotechnology) were added to the anti-(cyclin D1) and anti-Cdk6 immuno-precipitates The mixtures were agitated for 20 min (for Cdk2) or 30 min (for Cdk4/6) at 30C, and the reactions were halted by the addition of 15 lL per assay of 4· Laemmli sample buffer Samples were separated on 12% SDS/PAGE and electroblotted onto nitrocellulose filters The blots were exposed either to X-ray film or to a PhosphorImager screen to measure intensity of the 32 P-labelled substrates, and then blocked with blocking solution and immunoblotted with antibodies against the immunocomplex components (as described in Immuno-blotting)

Metabolic labeling Rat1/ras cells were cultured in 60 mm Petri dishes (120 000 cells per dish) The medium was replaced with fresh medium every 24 h HR12 (20 lM) was added to the relevant samples 24 h after the cells were plated Following 48 h exposure to HR12, the plates were washed three times with NaCl/Pi Starvation medium [dialysed FBS (10%) in medium lacking both methionine and cysteine; Biological Industries Beth HaEmek], with HR12 in the relevant samples, was added for 1 h

35S-Met/Cys Promix (200 lCiÆmL)1; Amersham-Pharma-cia) was then added N-acetyl-leucyl-leucyl-norleucynal (LLnL) (50 lM) was added to the appropriate samples After 3 h exposure to 35S-Met/Cys Promix, with or without LLnL, the plates were washed with NaCl/Piand the cells lysed Anti-p27Kip1 mAb (# K25020) was coupled to protein G-sepharose (Amersham-Pharmacia) and served for immunoprecipitation Following SDS/ PAGE and blotting, the membrane was exposed to X-ray film

Results

HR12 treatment of Rat1/ras cells inhibits anchorage-dependent and inanchorage-dependent cell-growth

We first characterized the effect of HR12 on anchorage-independent growth of Rat1/ras cells, using the assay for colony growth in soft agar HR12 treatment inhibited the growth of Rat1/ras cells in soft agar in a dose-dependent manner, with an IC50 value of 5 lM (Fig 1A) The inhibition led to a decrease in both colony size and the number of colonies The growth rate of Rat1/ras cells in a monolayer was also inhibited by HR12 in a dose dependent manner (IC50¼ 12 lM) This effect was selective, as the growth of the parental nontransformed Rat1 cells was not

affected at all by HR12 up to a concentration of 25 lM,

and only a minor effect of 50 lM HR12 was observed

(Fig 1B)

HR12 treatment of Rat1/ras cells suppressesin vitro monolayer ‘wound healing’

We then tested whether HR12 treatment of Rat1/ras cells suppresses the ability of cells present at the edges of a

‘wounded’ Rat1/ras monolayer to move out of the layer and

‘repair’ the wound This assay characterizes the proliferative and motility potentials of the cells, both of which are suppressed by cell–cell contacts Figure 2 shows that while Rat1/ras cells rapidly repaired the wound, HR12 treatment dramatically suppressed this process

HR12 induces arrest of Rat1/ras cells at the G1 phase

of the cell cycle

We next examined the effect of HR12 on the distribution

of Rat1/ras cells in the cell cycle To resolve the G1, S and G2/M phases, we double-labelled the cells with BrdU and propidium-iodide, as described in Experimental procedures Figure 3 shows that HR12 treatment of Rat1/ras cells induced G1 arrest, concomitant with a 25% reduction in the number of Rat1/ras cells in the S phase

The time course of HR12-induced inhibition of Ras processing correlates with the decrease in pRb phosphorylation

As phosphorylation of pRb is one of the key events required for G1/S transition, we examined whether HR12 affects pRb phosphorylation in Rat1/ras cells, and whether the timing of Ras-processing and pRb phosphorylation are correlated We treated Rat1/ras cells with 20 lMHR12 in growth medium and lysed the treated and untreated cells at the indicated times (Fig 4) Immunoblots of the lysates were probed with anti-Ras Ig and with anti-phospho-pRb(Ser795) Ig Unprocessed-Ras was separated from proc-essed Ras in 15% SDS/PAGE As we had shown previously, during the course of HR12 treatment, unprocessed Ras accumulated, whereas processed-Ras disappeared ([6] and Fig 4, upper panel) We found there to be a corresponding decrease in phosphorylation of pRb This dephosphoryla-tion followed the same kinetics as the inhibidephosphoryla-tion of Ras processing (Fig 4, lower panel) When HR12 was removed following 48 h treatment with 20 lMHR12, processed-Ras accumulated and pRb was phosphorylated simultaneously (Fig 4 ‘wash’) Thus, the inhibition of Ras processing caused

by HR12 was reversible, and relief of this inhibition correlated with the return of pRb phosphorylation

HR12 leads to an increase in p27Kip1 levels and to a decrease in pRb phosphorylation

in a dose-dependent manner

To analyse the cell cycle components affected by HR12 treatment, we prepared whole cell lysates of Rat1/ras cells that had been exposed to HR12 at various concentrations for 48 h Immunoblotting with Igs against cell cycle components led to several interesting findings First, the levels of the Cdk-inhibitor p27Kip1increased upon HR12 treatment in a dose-dependent manner (Fig 5) Second, the levels of the Cdk-inhibitor p21Cip1dropped (Fig 5) The levels of the Cdk-inhibitor p16INK4Awere also examined,

Fig 1 Inhibition of anchorage independent and dependent growth of

Rat1/ras cells by HR12 (A) Rat1/ras cells were grown in a layer of

0.3% soft agar in a 96-well plate, and exposed to HR12 at the indicated

concentrations, in triplicate After 7 days, the colonies were stained

with MTT and photographed Quantification was performed by

extraction of the color and measurement of the absorbance at 570 nm.

(B) HR12 selectively inhibited the growth of Rat1/ras cells, without

affecting the growth of nontransformed Rat1 cells Rat1 and Rat1/ras

cells were grown in monolayers in 96-well plates, and exposed to HR12

at the indicated concentrations Three days later the cells were

har-vested and counted.

but no increase was detected (data not shown) The levels of

cyclin D1 and cyclin E were not affected by HR12

treatment up to 40 lM (Fig 5); furthermore, their levels

remained unchanged over the course of HR12 treatment

(data not shown) Finally, a dose-dependent decrease in the hyper-phosphorylated form of pRb (Fig 5, pRb, upper band) was evident using an Ig against both the hyper- and the hypo-phosphorylated forms of pRb We note, how-ever, that in Rat1/ras cells the proportion of hyper-phosphorylated pRb was lower than in other cell lines (data not shown) Therefore, we also used an Ig specific for phospho-Ser795 of pRb: a dose-dependent reduction in phosphorylation was evident in both the hyper- and the hypo-phosphorylated bands (Fig 5, pS795-pRb) In sum-mary, we observed that treatment of Rat1/ras cells with increasing concentrations of HR12 led to a dose-dependent increase in p27Kip1levels accompanied by a corresponding, dose-dependent decrease in pRb-phosphorylation

Fig 2 HR12 suppresses in vitro monolayer ‘wound healing’of Rat1/ras

cells Rat1/ras cells were grown to confluence in the presence (right

column) or the absence (left column) of 20 l M HR12 At time 0, the

monolayer was wounded and phase-contrast photomicrographs were

taken at the indicated time points Medium with or without HR12 was

replaced every 24 h Quantification of the wound width vs time is

presented.

DNA content (PI)

no treatment

HR12

S

S

G1

G2/M

G1

G2/M

0 10 20 30 40 50 60

70

no treatment HR12

Fig 3 HR12 induces G1 arrest of Rat1/ras cells Rat1/ras cells were treated with 20 l M HR12 for 48 h, exposed to BrdU for 30 min, harvested and fixed in 70% ethanol The cells were double-stained with FITC-labelled anti-BrdU and propidium-iodide (PI), and analysed by flow cytometry.

HR12 inhibits the degradation of p27Kip1protein

in Rat1/ras cells

Three different mechanisms have recently been implicated in

the regulation of p27Kip1levels: (a) variations in the rate of

synthesis of the protein [25,29,39]; (b) variations in the rate

of degradation [40] and (c) transcriptional control [41] To

evaluate the contribution of HR12 to the stability of the

p27Kip1protein, we blocked the expression of new p27Kip1

protein by cycloheximide treatment of the cells, and observed

p27Kip1levels in whole cell lysates by immunoblotting with

anti-p27Kip1Igs Figure 6A shows that the t1/2of p27Kip1in

cells treated with HR12 is much longer (> 240 min) than the

t1/2of p27Kip1in untreated Rat1/ras cells (< 100 min) Thus,

HR12 leads to stabilization of the p27Kip1protein

To examine whether HR12 also affects the rate of

expression and/or synthesis of p27Kip1, we blocked

protea-some-mediated proteolysis by using LLnL, an inhibitor of the chymotryptic site on the proteasome [42,43] Rat1/ras cells were treated with 20 lM HR12 for 48 h and 50 lM LLnL was added to the cell medium for the last 3 h of treatment Figure 6B shows that p27Kip1levels in the cell lysate increased 2.5-fold as a result of LLnL treatment This result confirms the essential role of the proteasome in p27Kip1 down-regulation in Rat1/ras cells There was no significant difference between the amount of p27 synthesized after addition of LLnL when HR12 was absent (D1 in Fig 6B) and the amount of p27 synthesized after addition of LLnL when HR12 was present (D2 in Fig 6B), suggesting that HR12 does not influence the rate of synthesis of p27Kip1

To confirm this finding, we labelled newly synthesized proteins with35S-Met/Cys Promix during the last 3 h of HR12 treatment, as described in the Experimental proce-dures section The amount of label incorporated into immunoprecipitated p27Kip1in samples that were treated with HR12 was equivalent to or lower than the amount in untreated samples, whether or not LLnL was present during the metabolic labelling (Fig 6C) These findings confirm that HR12 does not enhance the rate of p27Kip1synthesis, indicating that the increase in amounts of p27Kip1in the presence of HR12 reflects a longer p27Kip1half-life

HR12 treatment of Rat1/ras cells leads to the accumulation of p27Kip1in the cyclin E/Cdk2 complex and to the inhibition of its kinase activity

We next examined whether G1 phase cyclin-dependent kinase activity is affected by the elevation in cellular p27Kip1 levels Rat1/ras cells were treated with 20 lMHR12 for 24 and 48 h, and lysates immunoprecipitated with anti-(cyclin E) (Fig 7A) or anti-Cdk2 (Fig 7B) Kinase activity

of the cyclin E/Cdk2 complex was measured using his-tone H1 and [c32P]ATP as substrates for the anti-(cyclin E) immunoprecipitates The mixtures were separated using SDS/PAGE, blotted onto a nitrocellulose filter and exposed

to a PhosphorImager screen to quantify the levels of phos-phorylated histone H1 The levels of the components of the immunocomplex were probed by immunoblotting the same blot with the relevant antibodies, as described in Experimen-tal procedures The kinase activity of the cyclin E/Cdk2 complex was significantly inhibited in Rat1/ras cells treated with HR12 Furthermore, the levels of Cdk-inhibitor p27Kip1 bound to the cyclin E/Cdk2 immunocomplexes in HR12-treated cells were at least three- to fourfold higher than those

of p27Kip1bound to the cyclin E/Cdk2 immunocomplexes in untreated cells (Fig 7A) Correspondingly, Fig 8B shows that the p27Kip1levels present in anti-Cdk2 immunoprecip-itates were significantly higher in the HR12-treated Rat1/ras cell complexes than in untreated cell complexes

HR12 treatment of Rat1/ras cells induces an increase in p27Kip1levels in the cyclin D1/Cdk6 and cyclin D1/Cdk4 complexes, with no inhibitory effect on their kinase activities

We evaluated the ability of HR12 to affect p27Kip1content

in the cyclin D1/Cdk6 and cyclin D1/Cdk4 G1 phase complexes and also evaluated its effect on the kinase activity

of these complexes The kinase activities of cyclin D1/Cdk6

p-pRb

0

40

80

120

time (hr) 1 3 5 15 24 48

Ras

Processed Ras (% of total Ras)

up

p

0

20

40

60

80

100

hours of exposure to HR12

Fig 4 The time course of the inhibition of Ras processing by HR12

correlates with the hypophosphorylation of pRb Rat1/ras cells grown in

medium containing 10% FBS were treated with 20 l M HR12 for the

indicated time periods, or exposed to 20 l M HR12 for 48 h, washed,

and incubated without the inhibitor for 24 h longer, before lysis

(wash) Lysates were immunoblotted with anti-Ras and

anti-phospho-Ser795-pRb (p-pRb) Igs (up) Unprocessed Ras, (p) processed Ras.

The upper graph shows the levels of processed Ras, as a percentage of

total Ras, over the course of HR12 treatment The lower graph shows

levels of pRb phosphorylation, compared to the untreated sample at

the same time point.

complexes, immunoprecipitated by anti-Cdk6 Ig, from

lysates of untreated and HR12-treated Rat1/ras cells, were

assayed using GST-pRb and [c32P]ATP as substrates The

immunocomplex components were visualized by

immuno-blotting as described above Figure 8A shows that cyclin

D1/Cdk6 complexes bound much higher levels of p27Kip1in

cells treated with HR12 than in untreated cells However, no

change in kinase activity was detected

Immunoprecipita-tion of cyclin D1/Cdk4 and cyclin D1/Cdk6 complexes

using anti-(cyclin D1) Igs revealed an HR12-induced

increase of p27Kip1 levels in the complexes This was

accompanied by increased kinase activity of these

immuno-precipitates (Fig 8B)

Fibroblasts transformed by farnesylation-independent

myristoylated-Ha-Ras are resistant to HR12-induced

G1 arrest, suppression ofin vitro monolayer ‘wound

healing’, cytoskeletal recovery and p27Kip1increase

To examine whether the effects of HR12 on the cell cycle,

cell motility and cell cycle components are mediated

exclusively by its effect on Ras, rather than on the

farnesylation of other protein(s), we examined the effects

of HR12 on Rat1 cells transformed by myr-Ras

(Rat1/myr-ras) Myr-Ras is an oncogenic Ha-Ras engineered to bind

the membrane constitutively through N-myristoylation with

no dependence on FT for its function Figure 9 shows that

treatment of Rat1/myr-ras with HR12 for 48 h had almost

no effect It did not change the cell cycle distribution

(Fig 9A) It had a minor effect on cell-growth in soft agar at

concentrations up to 25 lM(Fig 9B) The IC50of growth

inhibition in soft agar was about sevenfold higher for Rat1/

myr-ras cells than for Rat1/ras cells We have shown

previously that HR12 induces the assembly of adheren

junctions labelled with b-catenin and complete

morpho-logical reversion of Rat1/ras cells ([6] and Fig 9C) In the

Rat1/myr-ras cells, HR12 had no effect on b-catenin

distribution within the cells, as measured by

immunostain-ing (Fig 9C) Moreover, no morphological change of Rat1/

myr-ras cells was induced by HR12 treatment (Fig 9C)

HR12 did not affect the rate of ‘wound healing’ of Rat1/

myr-ras cells (Fig 9D), in contrast to its suppressive effect

on Rat1/ras cells (Fig 2) Finally, HR12 was not found to

affect p27Kip1 levels, pRb phosphorylation or cyclin D1

levels in Rat1/myr-ras cells (Fig 9E)

Discussion

HR12 effects are mediated by Ras inhibition

The inhibition of farnesyltransferase was developed

origin-ally as a strategy to block oncogenic Ras function

Nonetheless, the actual target of FTIs is a matter of controversy [44] We have reported recently on the devel-opment of a novel FTI, HR12 [5] We have shown that

Fig 5 HR12 treatment of Rat1/ras cells induces hypophosphorylation

of pRb and elevation of p27 Kip1 levels in a dose-dependent manner Rat1/

ras cells were exposed to HR12 at the indicated concentrations for

48 h, lysed and immunoblotted with Igs against phospho-Ser795-pRb

(p-pRb), pRb (pRb), p27Kip1, p21Cip1, cyclin D1 and cyclin E In the

case of pRb, the level of the phosphorylated protein was normalized to

the level of the total protein.

Cyclin D1

0 50 100 150

p21Cip1

0

60 80

40

20

Cyclin E

0 50 100 150

0 50 100 150

0 50 100

pS795-pRb pRb

M HR12

p27Kip1

Rat1/ras cells treated with HR12 undergo complete

mor-phological reversion and dramatic assembly of adheren

junctions, concomitant with an increase in cadherin and

b-catenin levels These effects are mediated via Ras [6]

In the current paper, we report upon the effects of HR12 on growth and on the cell cycle We find that HR12 suppresses anchorage-dependent and independent growth and motility

of Rat1/ras (Figs 1 and 2) Furthermore, treatment with HR12 leads to arrest of cell growth at the G1 phase of the cell cycle (Fig 3) It has been argued recently that FTIs inhibit the growth of Rat1/ras cells [32] and induce morphological reversion [45] through an inhibitory mech-anism that is Ras-independent and depends on the farnesy-lation of RhoB (the ‘FTI-RhoB hypothesis’, reviewed in [34,35,44]) In contrast, our results show clearly that the effects of HR12 are mediated via Ras In Rat1/myr-ras cells, Ras function is no longer dependent on farnesylation If the effects of HR12 were due to inhibition of a farnesylated protein other than Ras, the myristoylated-Ras transformed cells would have been affected by HR12 In Fig 9 we show that HR12 had no effect on cell cycle distribution (Fig 9A)

or the rate of ‘wound healing’ (Fig 9D) of Rat1/myr-ras cells Moreover, no cytoskeletal or morphological changes were observed in HR12-treated Rat1/myr-ras cells, while Rat1/ras cells were driven toward complete morphological and cytoskeletal reversion following HR12 treatment (Fig 9C) In accordance with the above data, HR12 had

no effect on p27Kip1levels or pRb phosphorylation in Rat1/ myr-ras cells (Fig 9E) Lastly, Rat1/myr-ras cells were much less sensitive to HR12 than Rat1/ras cells in a soft agar assay (Fig 9B) Resistance to HR12 was also seen with NIH3T3 fibroblasts transformed by myr-ras, unlike NIH3T3 cells transformed by farnesylation-dependent oncogenic ras (data not shown) Thus, the effects of HR12 on the proliferation, motility, cytoskeletal rearrange-ment and morphology of Rat1/ras cells are mediated through the inhibition of Ras farnesylation

P27Kip1inhibition of Cdk2 mediates HR12-induced G1 arrest

We show that HR12 treatment leads to accumulation of Rat1/ras cells in G1, with a corresponding reduction in the number of S phase cells (Fig 3) It has been shown that Ras controls progression through the late G1 phase of the cell cycle by controlling the levels of p27Kip1[25–27] Treating Rat1/ras cells with HR12, we saw a strong correlation

HR12-treated Rat1/ras cells

Untreated Rat1/ras cells

A

actin

0

50

100

actin

0

50

100

exposure to chx (min)

C

B

actin

LLnL: - - + +

HR12: - + - +

LLnL: - - + +

HR12: - + - +

0 2 4 6

8

∆2

∆1

Fig 6 HR12 enhances the half-life of p27Kip1protein, with no effect on its synthesis rate (A) HR12 leads to stabilization of the p27Kip1protein Rat1/ras cells were treated with 20 l M HR12 for 48 h, followed by the addition of 100 l M cycloheximide (chx) to the cell medium Lysates were prepared at the indicated time periods after chx addition, and immunoblotted with anti-p27 Kip1 Ig and with anti-actin Ig as a control The diagram shows quantification of the intensity of the p27Kip1 bands, calibrated to the intensity of the actin bands, where the zero time value was designated 100% (B, C) HR12 does not affect the synthesis rate of p27Kip1 (B) Rat1/ras cells were treated with 20 l M

HR12 for 48 h, and 50 l M LLnL was added to the medium 3 h before lysis Immunoblotting and quantification were performed as described above (C) Rat1/ras cells were treated with HR12 for 48 h, starved for

1 h, and labelled with35S-Met/Cys Promix in the presence or absence

of 50 l M LLnL for 3 h The lysates were immunoprecipitated with anti-p27 Kip1 , immunoblotted and exposed to X-ray film.

between the inhibition of Ras processing and the

accumu-lation of p27Kip1([6] and Fig 5) We observed an increase in

p27Kip1 levels in the cyclin E/Cdk2 complex, and a

corresponding reduction in the kinase activity of the

complex (Fig 7) Treatment of Rat1/ras cells with HR12

also led to an increase in the level of p27Kip1complexed with

Cdk4 and Cdk6, but their kinase activities were not

inhibited (Fig 8) This result is not surprising, for while

p27Kip1functions as an inhibitor of cyclin E/Cdk2, it also

plays a role in the assembly and activation of the cyclin D/

Cdk4 and cyclin D/Cdk6 complexes [46–48]

One of the best-characterized substrates of the Cdk

enzymes is the retinoblastoma protein (pRb)

Hypophos-phorylated pRb binds target proteins and arrests cells in the

G1 phase of the cell cycle This arrest is relieved by

Cdk-mediated hyperphosphorylation of pRb, which in turn

promotes the expression of factors that are essential for cell

cycle progression Treatment of Rat1/ras cells with HR12

led to a decrease in pRb phosphorylation (Fig 5) There

was a good correlation between the inhibition of

Ras-processing and of pRb dephosphorylation, in terms of both

kinetics and dose-responsiveness (Figs 4 and 5)

Our data contrast with those of Du et al [32,49], who

reported that their FTI led to an increase in p21CIP1levels, in

the same Rat1/ras model we used These authors, who did

not report any effect on p27Kip1, attribute the increase in

p21CIP1to the increase in geranylgeranylated RhoB caused

by inhibition of RhoB farnesylation (‘FTI-RhoB

Hypothe-sis’) We report here on a striking increase in p27Kip1levels

following HR12 treatment Moreover, we have shown that

this increase is correlated with increased amounts of p27Kip1

in complex with Cdk2 and with reduced Cdk2 kinase

activity Our data provide a plausible mechanism for the G1

arrest of Rat1/ras cells caused by HR12 We did not observe

an increase in p21CIP1levels, under the same conditions of

HR12 treatment (Fig 5)

HR12 leads to stabilization of p27Kip1

The amounts of p27Kip1 are regulated at the levels of

transcription [41], translation [39,50] and post-translational

degradation by the ubiquitin-proteasome pathway [40] Ras

has been reported to down-regulate p27Kip1 by all three

mechanisms: (a) control of p27Kip1degradation, by regula-tion of the RhoA pathway [27,29,51]; (b) repression of p27Kip1 synthesis, mediated either by the Raf/Mek/Erk pathway [29], the PI3K pathway [26] or the Rho pathway [52] and (c) repression of p27Kip1transcription through the activation of the PI3K/PKB pathway, which prevents the forkhead transcription factors from translocating to the nucleus [41]

The PI3K/PKB pathway is unlikely to be responsible for the observed increase in p27Kip1levels, as treatment of Rat1/ ras cells with HR12 for 48 h led to activation (rather than

Fig 7 HR12 treatment of Rat1/ras cells leads to an increase in the level

of p27Kip1in the Cyclin E/Cdk2 complex and inhibition of cyclin E/Cdk2

kinase activity (A) Rat1/ras cells were treated with 20 l M HR12 for 24

and 48 h Cell lysates were prepared and immunoprecipitated with

polyclonal anti-(cyclin E) Ig As a negative control, the anti-(cyclin E)

Ig was preincubated with a blocking peptide (BP) The

immunopre-cipitates were tested for kinase activity with histone-H1 as a substrate,

as described in Experimental procedures, followed by separation on

SDS/PAGE and blotting The blot was exposed to a PhosphorImager

screen or to X-ray film to quantify kinase activity ([32P]-H1) To

visualize the levels of the individual proteins in the immunoprecipitates

the same blot was immunoreacted with monoclonal anti-(cyclin E),

polyclonal anti-Cdk2 and monoclonal anti-p27Kip1Igs (B) Rat1/ras

cells were treated as in A, and immunoprecipitated with polyclonal

Cdk2 Ig Immunoprecipitates were immunoblotted with

anti-Cdk2 and anti-p27 Kip1 Ig.

repression) of PKB This activation of PKB was probably a

secondary event that arose as a consequence of the assembly

and activation of focal adhesions and cell–cell contacts [6]

The Raf/Mek/Erk pathway was strongly inhibited in Rat1/

ras cells treated with HR12 [6] However, we do not believe

this to be the regulatory pathway that leads to reduced levels

of p27Kip1because HR12 treatment had no effect on the rate

of p27Kip1synthesis and expression (Fig 6B) and inhibition

of Mek by PD98059 had no effect on p27Kip1levels in Rat1/ ras cells (data not shown) The half-life of the p27Kip1 protein was much longer in the presence of HR12 (Fig 6A), showing that HR12 stabilizes p27Kip1 The ubiquitin-proteasome pathway plays an essential role in p27Kip1 degradation, and indeed the specific proteasome inhibitor, LLnL, induced accumulation of p27Kip1protein in Rat1/ras cells (Fig 6B) Ras positively regulates RhoA [53], and RhoA leads to cyclin E/Cdk2 activation [54] The cyclin E/ Cdk2 complex phosphorylates p27Kip1at Thr187 and leads

it to degradation through the ubiquitin/proteasome path-way [27,55,56] There is a positive loop between p27Kip1 protein and cyclin E/Cdk2 in which p27Kip1serves both as a substrate and as an inhibitor of Cdk2 In summary, HR12 inhibits the degradation of the p27Kip1protein in Rat1/ras cells, possibly via the Ras-to-RhoA pathway

Is the increase in p27Kip1mediated by the induction

of cell–cell contacts?

p27Kip1levels are controlled by cadherin mediated cell–cell contacts that are themselves regulated by Ras [6,57] Levenberg et al and St Croix et al recently showed that overexpression or activation of cadherin leads to depho-sphorylation of pRb, increased levels of p27Kip1 and a reduction in cyclinE/Cdk2 levels, resulting in arrest of cell growth [11,12] Moreover, the levels of p27Kip1 mRNA remained constant in contact-inhibited cells [50] and the half-life of p27Kip1 protein was much longer in contact-inhibited cells than in cells growing exponentially [40] These phenomena are strikingly similar to the conse-quences of HR12 treatment: increased levels of cadherin, assembly of cell–cell contacts, stabilization of p27Kip1,

Fig 8 HR12 treatment of Rat1/ras cells does not induce inhibition of the kinase activity of cyclin D1/Cdk6 or cyclin D1/Cdk4 complexes Rat1/ras cells were treated as in Fig 7 and immunoprecipitated with anti-Cdk6 (A) or anti-(cyclin D1) (B) Igs The immunoprecipitates were tested for kinase activity with GST-pRb as a substrate, as des-cribed in Experimental procedures, followed by separation on SDS/ PAGE and blotting The blots were exposed to a PhosphorImager screen or to X-ray film to quantify kinase activity, [ 32

P]pRb To visualize the levels of the proteins in the immunoprecipitates the same blots were probed with polyclonal cyclin D1 or polyclonal anti-Cdk6 and with monoclonal anti-p27Kip1antibodies.

Fig 9 Resistance of Rat1/myr-ras cells to HR12 After 48 h treatment with 20 l M HR12 in medium containing 10% FBS, Rat1/myr-ras cells were (A) analysed for cell cycle distribution, (C) fixed and stained with anti-(b-catenin), (D) subjected to a wound healing assay, or (E) lysed and immunoblotted with antibodies against Ras, phospho-pRb (p-pRb), pRb, p27 Kip1 and cyclin D1 The growth of Rat1/myr-ras cells in soft agar was examined also (B) Rat1/myr-ras were resistant to HR12 effects, including suppression of ‘wound healing’, morphology rever-sion, assembly of adherens junctions, G1 arrest, up-regulation of p27Kip1and hypophosphorylation of pRb.