JOURNAL OF CELLULAR PHYSIOLOGY 155:340-348 (19931

Mevalonate Controls Cytoskeleton Organization and Cell Morphology in Thyroid

Epithelial Cells M A U R l Z l O BIFULCO,* CHIARA LAEZZA, SALVATORE M. ALOJ, AND CORRADO CARBl CEOSKNR and Dipartimento di Biologia e Patologia Cdlulare e Molecolare, Universita di

Napoli, 80131 Napoli (M.B., C.L., S.M.A., C.C.), and Dipartimento di Medicina Sperimentale e Clinica, Universita di Reggio Calabria, 88 100 Catanzaro (M.B.), ltaly

Blockade of mevalonate synthesis by the 3-hydroxy-3-methylglutaryl Coenzyme A reductase inhibitor mevinolin (lovastatin) causes FRTL-5 thyroid cells to un- dergo significant morphological changes; these include a transition from a flat, polygonal to a round shape, the development of cytoplasmic arborizations, and the loss of contact between neighboring cells. lmmunofluorescence studies of cytoskeletal structures show that, at early times after administering the drug, and before the round phenotype develops, stress fibers disassemble while the periph- eral actin filaments, which are adjacent to the cytoplasmic face of the plasma membrane, appear largely unaffected. Subsequently, when this cortical actin network becomes fragmented, cells start to round up and become separated from neighbors. Microtubules become disconnected from the plasma membrane and retract toward the cell center, although they do not appear depolymerized; in- deed, at this stage, cytoplasmic elongations contain mostly intact microtubules. After exposure to mevinolin FRTL-5 cells also lose vinculin-related substrate con- tacts. Treatment of cells with either cycloheximide or colchicine abolishes mor- phological changes induced by mevinolin, suggesting that ongoing protein syn- thesis and microtubule integrity are prerequisites for the drug to be effective. Both cytoskeletal and morphological perturbations can be reversed by mevalonate, but not by cholesterol or the non-sterol derivatives of mevalonate such as dolichol, ubiquinone, and isopentenyladenine, individually or in Combination. It is sug- gested that mevalonate deficiency may impair formation of isoprenylated proteins important for cytoskeletal organization and stability. o 1993 WiIey-Liss, Inc.

Eucaryotic cells synthesize cholesterol and several non-sterol isoprenoid compounds, such as ubiquinone, dolichol, isopentenyladenine, and isoprenylated pro- teins, via a pathway in which mevalonate, the product of the reaction catalyzed by 3-hydroxy-3-methylglu- tarylCoA (HMG-CoA) reductase, is the rate limiting precursor for the formation of all these products (for review, see Goldstein and Brown, 1990).

Blocking mevalonate synthesis causes arrest of cell proliferation (Quesney-Huneeus et al., 1979; Haben- icht et al., 1980; Fairbanks et al., 1984) and changes in cell morphology, such as cell rounding and neurite out- growth (Schmidt et al., 1982; Maltese and Sheridan, 1985). Many attempts aimed a t the identification of the metabolite of mevalonate and the molecular mecha- nisms responsible for the effects described above have failed to provide convincing results. Thus, while it is well established that mevalonate controls cell morphol- ogy and cell duplication, its mode of action is, as yet, not understood, nor is there any clear insight into the cellu- lar components, whose structural organization is per- turbed by mevalonate deficiency and responsible for morphological alterations. It has been suggested that mevalonate, andlor a derivative thereof, may influence cytoskeletal organization and/or cell attachment to the 0 1993 WILEY-LISS, INC.

substrate (Schmidt et al., 1982; Maltese, 1984), but these possibilities have not been investigated further. In this study we investigated in detail the effects of inhibition of mevalonate synthesis on cell morphology and cytoskeletal organization in FRTL-5 cells. These cells provide a suitable model for such study since their progression from quiescence into the cell cycle is associ- ated with a burst of mevalonate synthesis caused by a large transcriptional activation of the HMG-CoA re- ductase gene (Bifulco et al., 1990; Grieco et al., 1990) and alterations of cell morphology and microfilament organization (Tramontano et al., 1982).

In this study we show that in FRTL-5 cells the mor- phological changes caused by inhibition of mevalonate synthesis are preceded by a major reorganization of the cytoskeleton and involve the cell substrate adhesion system, the stress fibers, the cortical actin, and the microtubular network. All these changes can be pre- vented or reversed by mevalonate but not by choles-

Received August 4,1992; accepted December 4,1992.

*To whom reprint requestskorrespondence should be addressed.

34 1 MEVALONATE METABOLISM AND CYTOSKELETON ORGANIZATION

terol, ubiquinone, dolichol, and isopentenyladenine. We discuss the possible mechanisms of mevalonate ac- tions and speculate on the relevance of protein isopre- nylation for the stability of cytoskeletal structures and their interaction.

MATERIALS AND METHODS Cells

FRTL-5 cells (ATCC CRL 8305) are a strain of rat thyroid cells whose characteristics and culture condi- tions have been extensively described (Ambesi-Impi- ombato et al., 1980, 1982). They were grown in Coon’s modified Ham’s F-12 medium supplemented with 5% calf serum and a six-hormone mixture of TSH, insulin, hydrocortisone, transferrin, glycyl-L-histidyl-L-lysine acetate, and somatostatin. This mixture will be re- ferred as 6H.

Materials Mevinolin (lovastatin) was a gift from Dr. A.W. Al-

berts of the Merck, Sharp and Dohme Institute for Therapeutic Research (Alberts et al., 1980; Alberts 1988). Prior to adding mevinolin to the culture me- dium, the lactone was converted to the sodium salt (Kita et al., 1980). Mevalonic acid lactone, dolichol, ubiquinone, isopentenyladenine, cycloheximide, colchi- cine, and cytochalasin D were obtained from Sigma. Rhodamine isothiocyanate (R1TC)-conjugated phalloi- din, monoclonal antibodies to a-tubulin and vinculin were purchased from Sigma. Rabbit antiserum to mouse laminin was from BRL (Bethesda, MD).

Immunofluorescence Cells were seeded onto 1.2 cm diameter glass cover-

slips and cultured for 4 days to subconfluence in a me- dium containing 5% calf serum and the 6H mixture. Subsequently the cells were shifted for 72 h in a me- dium from which the hormonal supplement was re- moved (quiescent control cells). Mevinolin (10 pM) was then added for various periods of time. Cells were fixed with 3.7% formaldehyde and 0.1% glutaraldehyde in phosphate buffered saline (PBS), incubated for 10 min with 1% sodium borohydride in PBS, and, if needed, permeabilized with 0.2% Triton X-100 in PBS. The cells were then exposed to the specific primary antibodies for 40 min, washed in PBS, and reacted with fluorescein- or rhodamine-tagged goat anti-rabbit (Miles Scientific, Najerville, IL) or goat anti-mouse (Jackson, Immuno- Research, West Grove, PA) immunoglobulins. After fi- nal washes, the coverslips were mounted on a micro- scope slide using a 50% solution of glycerol in PBS. To visualize actin filaments, permeabilized cells were in- cubated with rhodamine isothiocyanate (RITC) or fluo- rescein isothiocyanate (F1TC)-conjugated phalloidin (Sigma), washed, and then mounted as described above.

RESULTS Morphology and cytoskeletal organization

Exposure of quiescent FRTL-5 cells (i.e., cells main- tained in the absence of TSH) to 10 pM mevinolin (lov- astatin) resulted in cell retraction from the substratum, rounding up and loss of contacts between neighboring cells, and development of cytoplasmic arborizations. Non-specific toxic effect of mevinolin could be ruled out

since, at 10 pM, the drug had no measurable effect on thyroglobulin synthesis, CAMP production, and 1- up- take (Bifulco et al., manuscript in preparation). There was no significant decrease of cellular cholesterol level, although a modest increase of LDL-receptor binding activity could be measured (data not shown).

Staining of quiescent FRTL-5 cells with rhodamine- isothiocyanate (RITC) labeled phalloidin showed that actin is polymerized into microfilament bundles (Fig. la). An actin rich subplasmalemmal region (cortical actin) was most prominent in well-spread cells at the edges of cell clusters (Fig. la) . A well-developed array of microtubules could be visualized by a monoclonal antibody to a-tubulin. Microtubules extended towards the cell periphery where the extremities were bent run- ning parallel to the cortical actin (Fig. lc). Since vimen- tin staining showed a diffuse pattern, no intermediate filaments could be resolved (data not shown). Exposure of quiescent FRTL-5 cells to 10 pM mevinolin for 24 h resulted in a major pertubation of the cytoskeletal structures described above. Thus the highly structured actin microfilaments had completely disappeared (Fig. lb ) along with the collapse of the microtubular network (Fig. Id). Morphological and cytoskeletal perturbations caused by MVA synthesis inhibition could be prevented by incubating FRTL-5 cells with the protein synthesis inhibitor cycloheximide (5 ng/ml) during mevinolin treatment (data not shown).

Temporal relation of cytoskeletal and morphological alterations

In order to investigate the chronological relationship of the morphological changes and cytoskeletal pertuba- tion, we examined the time course of cytoskeletal mod- ifications by fixing cells at various time intervals after mevinolin treatment. As shown in Figure 2, not all cells were affected synchronously; however, at each time se- lected for the observation there were a sufficient num- ber of cells affected, showing distinct changes which were typical of the time of exposure to the drug. Two hours after mevinolin addition, cell morphology was essentially unmodified (i,e., most cells still maintained a polygonal shape). At this stage, however, microfila- ments were disorganized and fragmented, with the ex- ception only of the cortical actin, which was mostly intact (Fig. 2a; compare to Fig. l a , control). Microtu- bules did not appear damaged yet (Fig. 2b); as in control cells (Fig. lc), they extended from the centrosome to- ward the cell periphery ending in the subplasmalem- ma1 region where the cortical actin was present. Four hours after mevinolin addition, the filamentous actin network at the cortical surface became progressively fragmented (Fig. 2c) and the microtubules appeared to retract from the submembrane region toward the cyto- sol. It is interesting to note that intact microtubules were still in place where intact domains of the cortical actin were present (the peripheral end of microtubules still co-localized with actin when both structures were visualized in double fluorescence experiments) (Fig. 2c,d). Changes in morphology occurred only after the disappearance of the cortical actin (5 h or longer after mevinolin challenge) (Fig. 2e,f). This observation sug- gests that in FRTL-5 cells, morphology is determined, to a large extent, by the state of polymerization of corti-

342

CONTROL

BIFULCO ET AL.

M EV I NOLl N

Fig. 1. Effect of mevalonate deprivation on cytoskeleton organiza- tion. Cells were cultured on glass coverslips (see Materials and Meth- ods), fixed, and stained with RITC-phalloidin (a,b), and anti a-tubulin (c,d). Cells were grown to semiconfluence in the 6H medium and 5 4 calf serum and then maintained for 72 h in a medium containing no hormones supplement (quiescent control cells). Subsequently cells

were incubated in the presence (b,d) 3r absence (a,c) of 10 p M mevino- lin for 24 h. Simultaneous double fluorescence staining (a,c) and (b,d). Mevinolin treated cells develop cytoplasmic processes with loss of con- tacts among neighboring cells. Stres.; fibers were lost (b) and microtu- bules collapsed (d). Bar, 20 pm.

cal actin rather than cytoplasmic stress fibers. The ob- servation that microtubules become "unhooked" when cortical actin disassembles supports the existence of areas of interaction between the two types of cytoskele- tal elements in the proximity of the plasma membrane (Niggli and Burger, 1987). By 10 h after mevinolin addition microfilaments have disappeared and a large fraction of microtubules collapsed in the cell cytoplasm (Fig. 2g,h). However, they were not completely depoly- merized and, in some cases, could still be seen in the cytoplasmic projections (Fig. 2h).

Cell-substrate adhesion system Quiescent FRTL-5 cells, in the absence of TSH, de-

velop scattered vinculin streaks alongside and near the end of stress fibers (Fig. 3a,b) (Garbi e t al., 1990). These sites of actin-membrane association correspond to focal adhesion areas. Since the most prominent morphologi- cal changes caused by mevinolin involved retraction from the substratum and rounding up of the cells, we

were prompted to investigate the perturbation of the cell-substrate adhesion systcm. FRTL-5 cells, treated with mevinolin for 24 h, were exposed to a monoclonal antibody to vinculin (see Materials and Methods). Im- munofluorescence staining revealed the disappearance of vinculin a t the basal surface and a concomitant dis- ruption of actin stress fibers 1 Fig. 3c,d).

FRTL-5 cells produce components of the extracellular matrix and develop laminin-mediated close contacts (Garbi et al., 1988). Analysis of mevinolin treated cells, following reaction with anti-laminin antibodies in the immunofluorescence staining, showed that arrange- ment of the extracellular matrix was different from that of control cells (data not shown). While the stress- fibers-dependent adhesion system was greatly and pre- cociously affected by mevinolin, perturbation of the ex- tracellular matrix organization, as revealed by laminin staining, became apparent or ly at later stages. Thus it could not be unequivocally established whether the changes of laminin staining patterns after mevinolin

MEVALONATE METABOLISM AND CYTOSKELETON ORGANIZATlON

ACTIN TUBULIN

Fig. 2. Time dependence of the effect of mevinolin on cytoskeleton organization. Actin (a,c,e,g) and tubulin (b,d,f,h) double fluorescence staining. Cells were cultured on glass coverslips as described (Materi- als and Methods). At time 0 (quiescent control cells; Fig. 3a,c) 10 FM mevinolin was added and incubation continued for the following inter- vals of time: 2 h (a,b); 4 h (c,d); 5 h (e,D; 10 h (g,h). Actin staining showed a progressive loss of microfilaments (a), fragmentation of the

were caused by a direct effect of the drug or were the consequence of cell rounding.

Reversal of mevinolin-induced changes by mevalonate

The rounded up appearance of a FRTL-5 cell exposed to mevinolin for 24 h (Fig. 4a, arrow) reverts to a mor-

343

cortical actin (c), and loss of contacts among neighboring cells (e ,g) . Peripheral actin and microtubules’ extremities have in some places a superimposable distribution (arrows and triangles in a,b,c,dl. In e and f the arrow indicates a cell where microfilaments have disassembled but cortical actin and the microtubules network is mostly intact. In h, arrowhead shows collapsed microtubules. In g and h arrows point to a cytoplasmatic projection containing intact microtubules. Bars, 10 pm.

phology nearly indistinguishable from that of control cells within 2 h following the addition of 700 pM MVA (Fig. 4b, arrow). Neither cholesterol nor non-sterol compounds such as ubiquinone, dolichol, or isopentenyl- adenine, individually or in combination, were able to revert the rounded phenotype. Reversal of cell morphol- ogy concurred with the reorganization of the cytoskele-

344 BIFULCO ET AL

Fig, 3. Mevinolin effect on cell-substrate adhesion. Double fluores- cence staining with RITC-phalloidin (a,c,e) and anti-vinculin (b,d,f). Control cells (a,b), 24 h mevinolin treated cells (c,d), cells exposed for 24 h to mevinolin and for 4 additional hours to mevalonate (e,D. Vin-

culin containing focal contacts disappeared in mevinolin treated cells (d) and reappeared (D along with reorganized actin stress fibers (e 1 after exposure to mevalonate. Bar, 20 pn. Arrowheads in a and 1) indicate stress fibers endings (a) corresponding to vinculin streaks (b).

ton: the time course of the process was remarkably sim- ilar to that of cell rounding and cytoskeletal disruption.

The reappearance of cytoplasmic microfilament bun- dles was the first visible event taking place within 1 h after MVA addition (Fig. 4d). Reorganization of the cortical actin could be observed within 2-3 h. It was

strate adhesion was reverted by MVA since with the reappearance of the normal flat morphology also the actin stress fibers and the vinculin containing struc- tures regained their normal appearance (Fig. 4e,f).

preceded by the appearance of discrete segments of po- lymerized actin assembled a t the cell periphery, and by remodeling of the edges of the cells which were smead-

perturbation Of microfilaments requires the integrity of microtubules

ing on the" substrat& (Fig. 4e). These regions may represent anchoring sites for the bent microtubules' extremities (Fig. 40. Also the pertubation of cell sub-

In order to identify the primary target for mevinolin effects on cytoskeleton, we performed studies in which either microtubules or microfilaments were disrupted

MEVALONATE METABOLISM AND CYTOSKELETON ORGANIZATION 345

Fig. 4. Reversal of mevinolin induced cytoskeleton disorganization by mevalonate. Cells were treated with 10 FM mevinolin for 24 h (a,c); subsequently 700 FM mevalonate was added and pictures were taken 1 h (d ) and 2 h (b,e,f) afterwards. Phase contrast microscopy in a and b shows the same cell cluster of mevinolin treated cells before and 2 h after mevalonate addition (the arrow points to identical cells). Fluo- rescence staining with RITC-phalloidin (c,d,e) and anti a-tubulin (fi.

Simultaneous double fluorescence (e,D. Note in d the reappearance of cytoplasmic microfilament bundles while morphology is not yet re- turned similar to control cells. In e and f cells recovered a quite normal morphology. Cortical actin appears in the process of being reassem- bled (el and a distinct set of microtubules appear to extend toward the cell periphery in close contact to a segment of polymerized actin ( e , f, arrows and arrowheads). a,b: Bar, 40 pm. c,d,e,f: Bar, 20 pm.

by specific agents before the administration of the drug. FRTL-5 cells, maintained for 1 h in the presence of 1 pM colchicine, displayed a complete disappearance of polymerized microtubules (Fig. 5c; compare to Fig. lc); however, their morphology, as assessed by phase con-

trast microscopy, and the presence of microfilaments, as probed by fluorescent staining with RITC-phalloi- din, was not significantly changed (Fig. 5a,b; compare to Fig. la). When the effect of mevinolin was investi- gated on colchicine pretreated cells, no significant

346 BIFULCO ET AL.

Fig. 5. Effect of colchicine pretreatment on mevinolin induced per- turbations. Colchicine (1.0 pM) was added to control cells for 1 h (a,b,c). Then cells were washed and incubated with 10 pM mevinolin for 18 h (d,e,D. Phase contrast microscopy (a,d). Double fluorescence staining with RITC-phalloidin (b,e) and anti a-tubulin (c,D. After

changes in morphology and microfilament assembly could be seen (Fig. 5d,e). Thus, it would appear that the integrity of microtubular structure is a prerequisite for mevinolin to affect FRTL-5 cell morphology and cy- toskeletal organization. These studies also suggest that for FRTL-5 cells to maintain their flat, poligonal mor- phology, the integrity of actin containing structures is a mandatory requirement. We investigated this further in studies in which FRTL-5 cells were incubated for 2 h in the presence of 1 pM cytochalasin D. Under these conditions microfilaments completely disappeared and cell morphology was altered to some extent, although microtubules did not appear to be affected. When cy- tochalasin D treated cells were exposed to mevinolin for 24 h, changes in morphology became more dramatic (data not shown). Further indirect evidence of the im- portance of the integrity of microtubules, and not of microfilaments, for the mevinolin effect comes from studies of FRTL-5 cells treated with TSH. It has been previously shown that TSH, through CAMP, causes cy- toskeleton perturbations characterized by the disap- pearance of microfilament structures without any ap- parent effect on microtubules (Tramontano et al., 1982). Preincubation with TSH did not interfere with mevinolin effect on microtubules and morphology (data not shown).

DISCUSSION Morphological changes caused by inhibition of me-

valonate synthesis have been described in very differ-

colchicine pretreatment mevinolin failed to induce either morphologi- cal changes (d) or microfilament disassembly (e) and did not interfere with microtubule growth from the microtubules organizing center after colchicine removal (0. a,d: Bar, 80 pm. b,c,e,f: Bar, 20 pm.

ent cell strains (Schmidt et al., 1982; Maltese, 1984; Maltese and Sheridan, 1985; Meng-Sheng et al., 1991). However, no detailed data have been provided, as yet, on the involvement of cytoskeletal structures as a caus- ative element for such changes. Rather, they have been related to inhibition of proliferation and induction of differentiation (Maltese, 1984; Maltese and Sheridan, 1985). The FRTL-5 cells, under the conditions of this study, allow the investigation of morphological changes induced by mevinolin in the absence of any change of differentiation and/or inhibition of DNA synthesis. In- deed, when maintained in the absence of TSH, FRTLd are quiescent and DNA synthesis is essentially ar- rested (Ambesi-Impiombato et al., 1982; Valente et al., 1983) and mevinolin has no measurable effect on differ- entiated functions such as 1- uptake and thyroglobulin synthesis. The data presented in this study provide am- ple evidence that mevalonte deficiency is, by far, the predominant effect of mevinolin and no other, non-spe- cific, cytotoxic effects are responsible for the morpho- logical changes observed. They also show that cytoskel- eta1 components are the early target of mevalonate deficiency; these are primarily the stress fibers and the region involved in cell-substratum contact, whose con- stitutive elements provide anchorage for the actin fila- ments bundled into stress fibers (Burridge et al., 1988).

Focal adhesions are characterized by a multiplicity of proteins functionally important since they mediate in- teractions of the extracellular matrix with the cytoskel- eton. Studies of maturation of these structures have

347 MEVALONATE METABOLISM AND CYTOSKELETON ORGANIZATION

also noted that more stable focal adhesions are often associated with the ends of microtubules, and it has been suggested that association with microtubules not only increases stability of focal adhesions but also pro- motes nucleation of stress fibers (Rinnerthaler et al., 1988). The time course of mevinolin effects shows that perturbations of cytoskeletal components precede the appearance of morphological alterations and, in all likelihood, are the causative element for such alter- ations. It also shows that actin containing structures are affected by mevinolin a t different times; thus, the cytoplasm spanning microfilaments appear disorga- nized and fragmented when cortical actin is still unaf- fected, suggesting the existence of a different regula- tion for actin containing structures according to their location within the cell. Cortical actin seems to control cell shape, more so than microfilaments, since no rounding up is observed 4 h after exposing FRTL-5 cells to mevinolin, when cytoplasmic microfilaments have disappeared. Cell shape changes coincide with frag- mentation of cortical actin 5 h (or longer) after drug treatment. Perturbation of microtubules, albeit exten- sive, is not as dramatic as those of microfilaments; thus, areas of nearly normal microtubular appearance are still visible when the cells have lost their polygonal shape and the actin containing structures are dis- solved. However, in spite of the relatively lesser in- volvement of microtubules, their integrity appears to be a prerequisite for mevinolin to affect both actin mi- crofilaments and cell morphology. The finding that massive dissolution of microtubules does not affect cell shape to any significant extent could be explained since colchicine does not perturb stress fibers and/or cortical actin which, as discussed above, are primary determi- nants of cell shape. Why then is microtubular integrity required for mevinolin to affect polymerized actin? One plausible explanation is suggested by the observation that inhibition of protein synthesis also prevents cy- toskeletal and morphological effects of mevinolin. This would argue in favor of one or several proteins, of rela- tively short half-life, involved in microtubular mech- ano-chemical function, which are critical for the inter- action with focal adhesions (see above) and other cytoskeletal elements. Lack of such proteins caused by cycloheximide, or structural alteration caused by col- chicine, could disrupt the interaction of microtubules and polymerized actin. It could be speculated that, un- der inhibition of mevalonate synthesis, defective preny- lation of regulatory proteins would somehow affect the state of actin containing structures, provided that their interaction with microtubules is preserved. This specu- lation would also localize prenylated regulatory pro- teins a t the cytoplasmic side of the plasma membrane, in a strategic position to control the dynamic interac- tion of microtubules and cortical actin.

The mechanism whereby inhibition of mevalonate synthesis causes perturbation of cytoskeletal struc- tures and modifications of cell shape remains elusive. Since acute cAMP elevation also causes cytoskeletal perturbation in FRTL-5 (Tramontano et al., 1982), i t had to be ruled out that mevinolin could affect cAMP production in these cells. Thus, mevinolin addition, for 24 h, did not affect cAMP levels, both in the presence or absence of TSH (data not shown).

The data reported in this study are consistent with the contention that mevalonate deprivation may affect the structure of one or more proteins physiologically involved in cytoskeletal organization. Vinculin has been reported to be the substrate of purified src kinase (Richert et al., 1982) and contains high levels of phos- photyrosine in Rous sarcoma virus transformed cells. These observations have led to the proposal that struc- tural modifications of vinculin may be responsible for cytoskeletal alterations observed in transformed cells (Sefton and Hunter, 1981). Subsequent studies of phor- bol ester and calcium induced phosphorylation of vincu- lin in normal cells have failed to show any correlation with morphological changes (Werth and Pastan, 1984). Therefore, phosphorylation of vinculin alone may not be sufficient, in and of itself, to induce cytoskeletal alterations and morphological changes. We do not know whether mevinolin affects phosphorylation of cy- toskeleton proteins in FRTL-5 cells; however, we do know that inhibition of mevalonate synthesis impairs protein prenylation. Thus the data presented in this study are consistent with the proposal that cytoskeletal andlor cytoskeleton associated proteins are substrate for isoprenylation and that this post-translational mod- ification may have an important bearing on cytoskele- tal organization and stability (Schmidt et al., 1984; Sin- ensky and Logel, 1985; Maltese, 1990). There have been several recent reports which lend support to the hy- pothesis that isoprenylation of cellular proteins affects cell morphology as well as cell differentiation and pro- liferation (Maltese and Sheridan, 1987). These include studies of regulation of oocyte maturation and microtu- bule organization by low molecular weight GTP bind- ing protein of the ras family (Birchmeier et al., 1985; Schmidt et al., 1986), whose membrane interaction and functional properties depend on isoprenylation (Han- cock et al., 1989). Isoprenylation of valine-14 p21rho (a member of the ras family) causes morphological alter- ations in Xenopus oocytes (Mohr et al., 1990) and block- ade of p21'"" isoprenylation, by mevalonate inhibition, causes rounding up and spike formation in PC12 cells (Meng-Sheng et al., 1991). We believe that mevinolin induced inhibition of mevalonate biosynthesis causes an impairment of isoprene moieties formation; this would alter post-translational modification of cytoskel- eta1 andlor cytoskeleton associated proteins and result in altered organization and stability of such proteins.

Concurrent with our study, and totally independent of it, Fenton et al. (1992), working on an entirely differ- ent cell system, the K1 cell line, derived from human renal carcinoma, have investigated the effect of mevin- olin on cell morphology and cytoskeletal organization. Their attention has been focused on the structural changes of actin cables caused by the drug, without providing data on the fate of cortical actin and the role of microtubules. They also propose that intracellular actin polymerization is regulated by prenylation, based on the observation that microinjection of a competitive inhibitor of prenyl transferase causes loss of actin ca- bles in 25-35% of the treated cells. This finding, al- though indirectly, supports our contention that impair- ment of prenylation of cytoskeletal proteins is the most likely molecular mechanism whereby mevinolin affects cytoskeletal organization and cell morphology in

348 BIFULCO ET AL.

FRTL-5 cells. Our study expands the observation of Fenton et al. (1992) since it shows unequivocally that 1) in the sequence of events leading to cell rounding, after mevinolin challenge, depolymerization of stress fibers precedes fragmentation of cortical actin (however, only when the latter occurs, the polygonal, flat phenotype changes into a rounded up shape) and 2) microtubular integrity is a prerequisite for mevinolin to affect actin polymerization, thus suggesting that tubulin andlor microtubule associated proteins are the early sub- strates for post-translational processes involved in reg- ulation of cytoskeletal organization and cell morphol- OgY-

ACKNOWLEDGMENTS We wish to express our thanks to Dr. Jan Wolff for his

constant support and encouragement and for critically reviewing the manuscript.

This study was supported in part by the Progetto Finalizzato “Invecchiamento” of the Consiglio Nazion- ale delle Ricerche.

Chiara Laezza is a recipient of a fellowship from A.I.R.C., Associazione Italiana Ricerca sul Cancro.

This paper was presented in part, in abstract form, a t the 9th Meeting of the Italian Association for Cell Biol- ogy and Differentiation (Garbi et al., 1991).

LITERATURE CITED Alberts, A.W. (1988) Discovery, biochemistry and biology of lovasta-

tin. Am. J . Cardiol., 62rlOJ-15J. Alberts, A.W., Chen, J . , Kuron, G., Hunt, V., Huff, J., Hoffman, C.,

Rothock, J . , Lopez, M., Harris, E., Patchett, A,, Monaghan, R., Cur- rie, s., Stapley, E., Albers-Schonberg, G., Hensens, O., Hirshfield, J., Hoogsten, K., Liesch, J., and Springer, J. (1980) Mevinolin: A highly potent competitive inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A reductase and cholesterol-lowering agent. Proc. Natl. Acad. Sci. U.S.A., 77:39573961.

Ambesi-Impiombato, F.S., Parks, L.A.M., and Coon, H.G. (1980) Cul- ture of hormone-dependent functional epithelial cells from rat thy- roids. Proc. Natl. Acad. Sci. U.S.A., 77t3455-3459.

Ambesi-Impiombato, F.S., Picone, R., and Tramontano, D. (1982) In- fluence of hormones and serum on growth and differentiation of the thvroid cell strain FRTL. Cold Sorine Harbor Conf. Cell Prolifera- tiin, 9t483492.

Bifulco, M., Santillo, M., Tedesco, I., Zarrilli, R., Laezza, C., and Aloj, S.M. (1990) TSH modulates low density lipoprotein binding activity in FRTL-5 thyroid cells. J . Biol. Chem:, 265.:19336-19342.1

Birchmeier, C., Broek, D., and Wigler, M. (1985) Ras protein induce meiosis in xenopus oocytes. Cell, 43:615-621.

Burridge, K., Fath, K., Kelly, T., Nuckolls, G., and Turner, C. (1988) Focal adhesion: Transmembrane junction between the extracellular matrix and the cytoskeleton. Annu. Rev. Cell Biol., 4:487-525.

Fairbanks, K.P., Witte, L.D., and Goodman, D.S. (1984) Relationship between mevalonate and mitogenesis in human fibroblasts stimu- lated with platelet-derived growth factor. J. Biol. Chem., 259:1546 1551.

Fenton, R.G., Hsiang-fu Kung, Longo, D.L., and Smith, M.R. (1992) Regulation of intracellular actin polymerization by prenylated cel- lular proteins. J. Cell Biol., //7:347356.

Garbi, C., Zurzolo, C., Bifulco, M., and Nitsch, L. (1988) Synthesis of extracellular matrix glycoproteins by a differentiated thyroid epi- thelial cell line. J. Cell. Physiol., /35:3946.

Garbi, C., Colletta, G., Cirafici, A.M., Marchisio, P., and Nitsch, L. (1990) Transforming growth factor beta induces cytoskeleton and extracellular matrix modifications in FRTL-5 thyroid epithelial cells. Eur. J. Cell Biol., 53:281-289.

Garbi, C, Laezza, C., Aloj, S.M., and Bifulco, M. (1991) The role of mevalonate in determining the cytoskeleton organization in the thyroid cell line FRTL-5. Eur. J. Cell Biol., 55r36.

Goldstein, J.L., and Brown, M.S. (1990) Regulation of the mevalonate pathway. Nature, 343:425430.

Grieco, D., Beg, Z.H., Romano, A., Bifulco, M., and Aloj, S.M. (1990) Cell cycle progression and 3-hydroxy-3-methylglutaryl coenzime A reductase are regulated by thyrotropin in FRTL-5 thyroid cells. J. Biol. Chem., 265r19343-19350.

Habenicht, A.J.R., Glomset, J.A., and Ross, R. (1980) Relation of cho- lesterol and mevalonic acid to the cell cycle in smooth muscle and Swiss 3T3 cells stimulated to divide by platelet-derived growth fac- tor. J. Biol. Chem., 255:513&5140.

Hancock. J.F.. Magee. A.I.. Childs. J.E.. and Marshall. C.J. (1989) All ras proteins are polyisoprenylated but only some are palmitoylated. Cell, 57r1167-1177.

Kita, T., Brown, M.S.. and Goldstein, J.L. (1980) Feedback regulation of ‘3-hydroxy-3-methylglutaryl coenzyme A reductase in livers of mice treated with mevinolin, a competitive inhibitor of the reduc- tase. J. Clin. Invest., 66t1094-1100.

Maltese, W.A. (1984) Induction ofdifferentiation in murine neuroblas- toma cells by mevinolin, a competitive inhibitor of 3-hydroxy-3- methylglutarylcoenzyme A reductase. Biochem. Biophys. Res. Com- mun., 120t454460.

Maltese, W.A. (1990) Posttranslational modification of proteins by isoprenoids in mammalian cells. FASEB J., 4133194328.

Maltese, W.A., and Sheridan, K.M. (1985) Differentiation of neuro- blastoma cells induced by an inhibitor of mevalonate synthesis: Relation of neurite outgrowth and acetylcholinesterase activity to changes in cell proliferation and blocked isoprenoid synthesis. J. Cell. Physiol., 125t540-558.

Maltese, W.A., and Sheridan, K.M. (1987) Isoprenylated proteins in cultured cells: Subcellular distribution and changes related to al- tered morphology and growth arrest induced by mevalonate depri- vation. J. Cell. Physiol., 133t471481.

Meng-Sheng Q., Pitts, A.F., Winters, T.R., and Green, S.H. (1991) Isoprenylation is required for ras-induced but not for NGF-induced neuronal differentiation of PC12 cells. J . Cell Biol., 115:795-808.

Mohr, C., Just, I., Hall, A., and Aktories, K. (1990) Morphological alterations of Xenopus oocytes induced by valine-14p21rh” depend on isoprenylation and are inhibited by clostridium botulinium C3 ADP-ribosyltransferase. FEBS Lett., 275:168-172.

Niggli, V., and Burger, M.M. (1987) Interaction of the cytoskeleton with the plasma membrane. J. Membr. Biol., 100:97-121.

Quesney-Huneeus, V., Wiley, M.H., and Siperstein, M.D. (1979) Es- sential role for mevalonate synthesis in DNA replication. Proc. Natl. Acad Sci. U.S.A., 76.50565060.

Richert, N.D., Blithe, D.L., and Pastan, I. (1982) Properties of the src kinase purified from rous sarcoma virus-induced rat tumors. J. Biol. Cbem., 257t7143-7150.

Rinnerthaler, K., Geiger, B., and Small, J.V. (1988) Contact formation during fibroblast locomotion: Involvement of membrane ruffles and microtubules. J. Cell Biol., 106:747-760.

Schmiddt, R.A., Glomset, J.A., Wight, T.N., Habenicht, A.J.R., and Ross, R. (1982) A study of the influence of mevalonic acid and its metabolites on the morphology of swiss 3T3 cells. J. Cell Biol., 95r144-153.

Schmidt, R.A., Schneider, C.J., and Glomset J.A. (1984) Evidence for post-translational incorporation of a product of mevalonic acid into Swiss 3T3 cell proteins. J. Biol. Chem., 259:10175-10180.

Schmidt, H.P., Wagner, P., Pfeff, E., and Gallwitz, D. (1986) The ras related YPTl gene product in yeast: A GTP-binding protein that might be involved in microtubule organization. Cell, 47t401412.

Sefton, B.W., and Hunter, T. (1981) Vinculin: A cytoskeletal target of the transforming protein of rous sarcoma virus. Cell, 24:165-174.

Sinensky, M., and Logel, J. (1985) Defective macromolecule biosyn- thesis and cell-cycle progression in a mammalian cell starved for mevalonate. Proc. Natl. Acad. Sci. U.S.A., 82r3257-3261.

Tramontano, D., Avivi, A., Ambesi-Impiombato, F.S., Barak, L., Gei- ger, B., and Schlessinger, J. (1982) Thyrotropin induces changes in the morphology and the organization of microfilament structures in cultured thyroid cells. Exp. Cell Res., 137:269-275.

Valente, W.A., Vitti, P., Kohn, L.D., Brandi, M.L., Rotella, C., Toc- cafondi, R., Tramontano, D., Aloj, S.M., and Ambesi-Impiombato, F.S. (1983) The relationship of growth and adenylate cyclase activ- ity in cultured thryoid cells: Separate bioeffects of thyrotropin. En- docrinology, / 12:7 1-79.

Werth, D.K., and Pastan, I. (1984) Vinculin phosphorylation in re- sponse to calcium and phorbol esters in intact cells. J . Biol. Chem., 259352665270.