Fluoroaluminate stimulates phosphorylation of p130 Cas and Fak and increases attachment and...

Fluoroaluminate Stimulates Phosphorylation of p130 Cas andFak and Increases Attachment and Spreading ofPreosteoblastic MC3T3-E1 CellsF. FREITAS, M. JESCHKE, I. MAJSTOROVIC, D. R. MUELLER, P. SCHINDLER, H. VOSHOL,J. VAN OOSTRUM, and M. SUSA

Research Bone Metabolism and Functional Genomics, Novartis Pharma AG, Basel, Switzerland

Fluoroaluminate is a G-protein activator, it stimulates osteo-blastic cells in culture, and is a bone-forming agent in vivo.To elucidate the mechanisms of G-protein-mediated action offluoroaluminate in osteoblasts, we studied protein tyrosinephosphorylation in the preosteoblastic cell line MC3T3-E1.Fluoroaluminate, lysophosphatidic acid (LPA; an agonist forG-protein-coupled receptor), or adhesion to type I collagenall stimulated phosphorylation of a similar set of proteins,including p130, p120, p110 (previously identified as proline-rich tyrosine kinase 2, Pyk2), and p70. The phosphorylationof these proteins was sensitive to an Src inhibitor, but not toa Gi-protein inactivator, pertussis toxin. By purification/mass spectrometry and by immunodepletion, p130 proteinwas identified as p130 Cas (Crk-associated protein), a Srcsubstrate and a protein involved in signaling by cell-adhesionreceptors, integrins. Phosphorylation of immunoprecipitatedp130 Cas increased upon stimulation with fluoroaluminateand with agonists of G-protein-coupled receptors, but notwith growth factors. By immunodepletion, the p120 proteinwas identified as focal adhesion kinase, Fak. The addition offluoroaluminate during cell attachment to type I collagenfurther stimulated phosphorylation of p130 Cas and of Fak.Simultaneously, fluoroaluminate increased the number ofattached MC3T3-E1 cells and their spreading. These novelaspects of fluoroaluminate action in cell culture may beimportant for the bone-forming action of fluoroaluminate invivo. (Bone 30:99–108; 2002) © 2002 by Elsevier ScienceInc. All rights reserved.

Key Words: Osteoblast; Fluoroaluminate; p130 Cas; Fak; Ad-hesion; Bone.

Introduction

Osteoporosis is a metabolic bone disease characterized by lowbone mass and microarchitectural deterioration of the bonetissue, which results in an increased fracture rate.34 Currenttherapy for osteoporosis focuses on increasing bone mass duringdevelopment or on suppressing bone resorption after the fifth

decade of life; however, the most efficacious agents are expectedto be bone anabolics, drugs that stimulate bone formation. Potent,well-tolerated, bone-selective anabolic agents that would beconvenient to use in a large population of elderly osteoporoticpatients are not yet available.

To determine new targets for bone anabolic drugs in osteo-blasts, we have previously studied signal transduction mecha-nisms of fluoride, a known bone anabolic agent.

Fluoride in complex with aluminum (fluoroaluminate) caninduce bone formation in animal models, which mimics theaction of fluoride in humans that receive aluminum via diet or inmine workers suffering from osteosclerosis after exposure toaluminum-fluoride complexes.2 Fluoroaluminate is known as aG-protein activator, and our group and others have shown thatfluoroaluminate can activate Gi-dependent signaling to mitogen-activated protein (MAP) kinase family members Erk1/2, and canalso increase Gi-independent tyrosine phosphorylation of severalintracellular proteins.2,3,31 Recently, we reported that two proteintyrosine kinases activated by fluoroaluminate in osteoblasticMC3T3-E1 cells are the cytoplasmic tyrosine kinases Src andPyk2.15

Identification of tyrosine-phosphorylated proteins in flu-oroaluminate-treated osteoblastic cells may provide new insightsinto the mechanism of action of fluoroaluminate. In particular, itsbiological effect on osteoblasts has so far remained elusive.Although in some cases fluoride given as NaF can stimulateosteoblast proliferation and differentiation, there is no generalagreement on how fluoride or fluoroaluminate act.29 In this studywe identified major tyrosine-phosphorylated proteins whosephosphorylation was stimulated by fluoroaluminate. Based onsimilarities with integrin signaling, we examined the interactionbetween fluoroaluminate and cell adhesion, both at the level ofintracellular signaling and level of cellular function.

Materials and Methods

Materials

NaF was purchased from Fluka and AlCl3 from Sigma. Flu-oroaluminate solution was prepared immediately before the ex-periment as described previously.31 Lysophosphatidic acid(LPA), pertussis toxin (PTX), and protein G-Sepharose 4B fastflow were from Sigma. Antibodies were obtained as follows:antiphosphotyrosine monoclonal antibody (MAb) 4G10 (UBI)and anti-p130 Cas MAb (Transduction Laboratories); anti-p130Crk-associated substrate (Cas) polyclonal antibody (PAb) C-30

Address for correspondence and reprints: Dr. Mira Susa, NovartisPharma Research, Arthritis and Bone Metabolism Therapeutic Area,WKL-125 9.12, P.O. Box, CH-4002 Basel, Switzerland. E-mail:[email protected]

Bone Vol. 30, No. 1January 2002:99–108

99© 2002 by Elsevier Science Inc. 8756-3282/02/$22.00All rights reserved. PII S8756-3282(01)00625-1

(Santa Cruz Biotechnologies); anti-focal adhesion kinase (anti-Fak) monoclonal antibodies (UBI, MAb 2A7 and TransductionLaboratories); and horseradish peroxidase-conjugated secondaryantibodies (Cappel). Src family inhibitor CGP77675 was synthe-sized in the Chemistry Research Laboratories at Novartis, and itscharacterization has been described previously.19

Cell Culture

The mouse osteoblastic cell line MC3T3-E1 was cultured in�-MEM supplemented with 10% fetal calf serum (FCS; Gibco)as described earlier.31 Prior to stimulation, cell cultures weredeprived of growth factors for 24 h in the absence of serum.

Cell Extraction

MC3T3-E1 cells were washed with ice-cold phosphate-bufferedsaline (PBS) and cellular extracts were prepared using NP-40lysis and cleared by centrifugation in a microcentrifuge (14,000rpm at 4°C for 5 min) buffer as described elsewhere.31 Proteinconcentrations were determined using a Bradford protein assay(Bio-Rad).

Western Blotting, Immunoprecipitation, and Immunodepletion

These techniques were done as our group has previously de-scribed in detail.19

Buffers Used for p130 Purification

Heparin-sepharose buffer consisted of NP40 lysis buffer supple-mented with 100 nmol/L NaCl and 0.05 mmol/L sodium vana-date; FFQ buffer was 20 mmol/L Tris (pH 7.7), 0.1% NP40, 10mmol/L �-glycerophosphate, 10 mmol/L NaF, 10 mmol/L di-thiothreitol (DTT), 5 mmol/L ethylene-diamine tetraacetic acid(EDTA), 1 mmol/L EGTA, 1 mmol/L benzamidine, 0.1 mmol/LPMSF, and 0.1 mmol/L sodium vanadate; and 4G10-sepharoseelution buffer was 20 mmol/L glycine (pH 2.3).

Purification of p130

MC3T3-E1 osteoblastic cells were grown on 15 cm plates to90% confluency and deprived of serum for 24 h. Upon stimula-tion with fluoroaluminate (10 mmol/L NaF/10 �mol/L AlCl3) for40 min, cells were washed in PBS and harvested in NP40 lysisbuffer. Cell extracts obtained from 100 plates were pooled andcleared by centrifugation. About 200 mg of protein in 100 mL oflysis buffer were adjusted to 100 mmol/L NaCl and subjected toaffinity chromatography using heparin-sepharose (Pharmacia), astep that separated Fak from p130 (see Figure 2A, lanes 1 and 2).The flow-through from HeparinSepharose (protein concentration1.5 mg/mL, total protein 150 mg) was subjected to FFQ anion-exchange chromatography to concentrate the sample (Figure 2A,lane 3). The proteins were eluted from the FFQ column by abatch procedure with 500 mmol/L NaCl (12 mL, 42 mg protein).Finally, the FFQ eluate pool was subjected to immunoaffinitychromatography using immobilized antiphosphotyrosine anti-bodies (4G10Sepharose, UBI) (Figure 2A, lanes 4–6). Boundproteins were eluted in 1 mL fractions with glycine buffer (pH2.3) and immediately neutralized with a 1/20 volume of 1 mol/LTris (pH 8.0). A portion of each eluate (1/100) was analyzed byWestern blotting for the presence of p130. The eluted fractionswere individually dialyzed against ammonium bicarbonate buffercontaining 0.025% sodium dodecylsulfate (SDS) and 20 mmol/L�-mercaptoethanol and subsequently concentrated on a Speed-

Vac to a final volume of 50 �L. The fraction containing about95% of the eluted phosphoproteins was separated by SDS-polyacrylamide gel electrophoresis (PAGE) (10% polyacryl-amide, Bio-Rad Minigel, 1 mm). Gels were stained using a silvernitrate method27 and a 130 kDa protein was cut out of the gel.

Mass Spectrometry

Proteins were digested in the gel with sequence-grade trypsin(Boehringer Mannheim) as described elsewhere.27 Reflectron-positive matrix-assisted laser desorption ionization (MALDI)mass spectra were recorded on a mass spectrometer (VoyagerElite, PerSeptive Biosystems, Framingham, MA) at 20 kV ac-celerating potential in delayed extraction mode using standardsettings for delay times and grid voltages. Database searches(GenePept) were performed with an expected accuracy of 50ppm for masses of tryptic peptides and with a mass windowbetween 50 and 200 kDa for the parent protein. The minimumnumbers of peptides required for a match was set to eight in theprimary search, and to six for a second-pass search with theremaining masses. For nanoelectrospray tandem mass spectrom-etry (MSMS) the samples were dissolved in 5% formic acidaqueous solution, and 12 �L (corresponding to 65% of totalsample) of the sample was concentrated and desalted (borosili-cate capillary, Protein Analysis Co., Odense, Denmark) andpacked with �100 nL of Poros R2 sorbent (PerSeptive Biosys-tems). After a washing step with 20 �L of 5% formic acid, thepeptide mixture was eluted directly into the spraying needle(gold-coated borosilicate capillary, Protein Analysis Co.) by theaddition of 2 �L of 5% formic acid in 50/45 CH3OH/H2O.Application of a voltage difference between the needle tip andthe orifice (1.5 mm distance) of the mass spectrometer (API III,PE-Sciex, Toronto, ON, Canada) began the spraying process. Afull mass spectrum was acquired over a wide mass range (m/z350–1500) and subsequently selected parent ions were massselected and fragmented by collision-induced dissociation withargon (collision energy 96 eV, collision gas thickness 300 � 1014

molecules/cm2) according to Wilm and Mann.35

Cell Adhesion Assays

Assays were done essentially as described previously.32 Briefly,the cells were deprived of serum, trypsinized, and seeded ontowells coated with type I collagen. For Western blot analysis, theprotein was extracted after various incubation times. For quan-tification of cell attachment and spreading, the cells were fixed,and stained with crystal violet. Subsequently, the cells werecounted, or total cell area in the well was measured with theLeica imaging system (Leica Microsystems, Glattbrugg, Swit-zerland).

Statistical Analysis

Where appropriate, statistical analysis was performed using one-way analysis of variance (ANOVA) with the GRAPHPAD PRISM

program.

Results

Tyrosine Phosphorylated Proteins Induced byFluoroaluminate, LPA, or Adhesion to Type I Collagen

Previously we have reported the phosphorylation of p130, p120,and p70 in MC3T3-E1 cells stimulated with the G-proteinactivator, fluoroaluminate.31 In addition, we showed that aweakly phosphorylated band at p110 is identical to proline-rich

100 F. Freitas et al. Bone Vol. 30, No. 1Fluoroaluminate, p130 Cas, and Fak in osteoblasts January 2002:99–108

tyrosine kinase 2 (Pyk2). This pattern of protein phosphorylationis not identical to that induced by platelet-derived growth factor(PDGF), which has been shown to activate a receptor tyrosinekinase.31 Here we examined the similarity of a fluoroaluminate-induced tyrosine phosphorylation pattern with that induced bytwo other types of stimuli: (i) LPA, acting on a G-protein-coupled receptor; and (ii) cell adhesion to type I collagen, actingon integrin receptor �2�1. As shown in Figure 1A, similar tofluoroaluminate, LPA induced phosphorylation of p130, p120,and p70. Because signaling by LPA was weaker than by flu-oroaluminate, this may be the reason why Pyk2 phosphorylationwas not detected by LPA.

We also examined tyrosine phosphorylation in MC3T3-E1cells upon adhesion to type I collagen, the most abundant proteinin the bone extracellular matrix. Except for a weak band at about180 kDa, nonadhering cells had very low, undetectable levels ofprotein tyrosine phosphorylation (Figure 1B). Adhesion to type Icollagen for 30–60 min induced a strong, time-dependent phos-phorylation of a triplet of proteins at 120–130 kDa (tentativelyp130, p125, and p120) and a protein p70. This pattern is similarto the fluoroaluminate-induced pattern, except for a middle band

in a triplet, which is absent with fluoroaluminate (Figure 1A–C).We concluded that the G-protein activators fluoroaluminate andLPA, as well as cell adhesion to type I collagen induced a verysimilar set of tyrosine-phosphorylated proteins in preosteoblasticMC3T3-E1 cells.

Effect of Src Inhibitor and Perfussis Toxin on TyrosinePhosphorylation Induced by Fluoroaluminate

To understand which signaling pathway contributes to the de-scribed increase in tyrosine phosphorylation, we employed twoinhibitors: CGP77675, a Src inhibitor; and pertussis toxin, aninhibitor of the Gi class of � subunits of heterotrimeric Gproteins. Treatment with Src inhibitor induced a dose-dependantreduction in tyrosine phosphorylation of p130, p120, and p70bands (Figure 1C). The IC50 value for p130 was about 0.28�mol/L which is very close to the value reported for inhibition ofcellular Src activity (0.2 �mol/L).19 By contrast, pertussis toxindid not affect tyrosine phosphorylation of p130 and p120 (Figure1D), in agreement with our previous data.31 Thus, tyrosine

Figure 1. Induction of tyrosine phosphorylation of 120 and 130 kDa proteins in MC3T3-E1 cells by LPA, fluoroaluminate (FAl), and cell adhesionto type I collagen, and the effects of Src inhibitor and pertussis toxin. MC3T3-E1 cells were grown to confluence, deprived from serum, and stimulatedas indicated. Cellular protein was extracted and equal amounts of protein were analyzed with antiphosphotyrosine Western blotting (antibody 4G10).(A) Preconfluent cultures were serum-starved for 24 h and stimulated with fluoroaluminate (FAl: 10 mmol/L NaF/10 �mol/L AlCl3) or LPA (25�mol/L) for 20 min or left untreated. The positions of p120 and p130 are indicated on the right. The identity of Pyk2 was established byimmunodepletion in our previous study.15 The positions of molecular weight markers are indicated on the left (Bio-Rad prestained marker, high range).(B) Serum-starved cells were trypsinized and replated on plastic or type I collagen-coated plates (10 �g/cm2) for the indicated times. (C) Prior tostimulation with fluoroaluminate as in (A), the cells were pretreated with the indicated concentrations of Src inhibitor CGP77675 for 30 min. (D) Priorto stimulation with the indicated concentrations of fluoroaluminate, the cells were pretreated with pertussis toxin (PTX) for 24 h.

101Bone Vol. 30, No. 1 F. Freitas et al.January 2002:99–108 Fluoroaluminate, p130 Cas, and Fak in osteoblasts

phosphorylation by fluoroaluminate is dependent on Src, but noton Gi proteins.

p130 Is Identical to p130 Cas

To understand the molecular basis of fluoroaluminate action inosteoblasts, we chose to identify p130 tyrosine-phosphorylatedprotein. This protein was of particular interest, because it hasbeen shown to be strongly phosphorylated in MC3T3-E1 preos-teoblastic cells, but only weakly in fibroblastic NIH3T3 cells.31

Thus, its identification could help us understand the specificity ofsignaling by fluoroaluminate in osteoblastic cells. We first at-tempted to identify p130 by performing Western blots with anumber of antibodies corresponding to the known tyrosine-phosphorylated proteins of about p130 kDa. These included p110AFAP, p116 phosphatidylinositol kinase �, p116 vinculin, p120Cas, p130 Cas, p125 Cbl, p120-inducible nitric oxide synthase,p120 JAK-3, p120 ras-GAP, p130 JAK-2, p135 JAK-1, and p148phospholipase C�. None of these proteins, however, comigratedexactly with the p130 tyrosine-phosphorylated band. Thus, for itsidentification, we purified the protein and analyzed it by massspectrometric techniques (Figure 2). The purified p130 proteinband was excised out of the gel (Fig. 2A), in-gel reduced,alkylated, and digested with trypsin, and the resulting peptideswere analyzed by the MALDI/peptide map approach. Searchingthe GenePept database with a list of peptide masses obtainedfrom MALDI analysis revealed, as best matches, two isoforms ofCrk-associated substrate (p130 Cas and Cas-b, U28151 andU48853 in the GenePept database; Figure 2B). The matchingpeptides covered 28% of the p130 Cas sequence, which is areliable match in the molecular mass range of 100 kDa (Figure2C). Nanoelectrospray tandem mass spectrometry of a selectedtryptic peptide provided partial but sufficient sequence informa-tion to confirm the identification of p130 as Cas (identifiedamino acids shown in italics; Figure 2C).

The reason for the initial lack of identification of p130 Cas byblotting was a slower migration of the phosphorylated protein(detected by antiphosphotyrosine antibodies), as compared withthe bulk of nonphosphorylated protein (detected by anti-Casantibodies). A similar phenomenon was observed for variousother phosphoproteins.

p120 Is Identical to Focal Adhesion Kinase (Fak)

To identify other tyrosine-phosphorylated proteins in the 120–130 kDa region in fluoroaluminate-stimulated MC3T3-E1 celllysates, we performed immunodepletion experiments with anti-Fak antibodies. Fak was chosen as a candidate protein because:(a) it has molecular weight of 125 kDa; (b) cell adhesion inducesFak phosphorylation in MC3T3-E1 cells32 (as fluoroaluminateand adhesion to type I collagen induce tyrosine phosphorylationof a similar set of proteins, it was possible that one of them couldbe Fak); and (c) p130 protein was identical to p130 Cas, whichis involved in integrin signaling together with Fak.24 Thus, otherphosphorylated proteins may also be related to integrin signaling,and Fak is known to participate in integrin signaling.

Fak protein was depleted from the lysates after two rounds ofimmunoprecipitation (Figure 3A, bottom panel). In parallel, thep120 tyrosine-phosphorylated band was removed from the lysate(Figure 3A, upper panel). The p130 band (identical to p130 Cas)remained in the lysate. Thus, immunodepletion clearly demon-strated that p120 protein was identical to Fak.

p130 Cas Phosphorylation Is Increased by Fluoroaluminate,But Not by Insulin and Epidermal Growth Factor (EGF)

Having identified p120 as Fak by immunodepletion, we sought toconfirm the identity of p130 as Cas by the same method. Thus,two rounds of immunoprecipitation were done with anti-Casantibodies. The initial lysate, together with supernatants afterimmunoprecipitation, was analyzed by Western blotting withantiphosphotyrosine and anti-Cas antibodies (Figure 3B). p130Cas protein was depleted from the lysates after two rounds ofimmunoprecipitation (Figure 3B, lower panel). In parallel, thep130 tyrosine-phosphorylated band was almost completely re-moved from the lysate (Figure 3B, upper panel). The p120 band(identical to Fak) remained in the lysate. Thus, immunodepletionconfirmed that the major part of the p130 protein was identical top130 Cas.

Another way to show that p130 Cas behaves as a p130tyrosine-phosphorylated band is to examine p130 Cas tyrosinephosphorylation in anti-Cas immunoprecipitates. Similar to flu-oroaluminate, several agonists of G-protein-coupled receptorsinduced p130 phosphorylation (Table 1). In contrast to flu-oroaluminate, growth factors are not potent inducers of p130-band phosphorylation in MC3T3-E1 cells (Table 1 and ref. 31).Thus, we compared the effects of fluoroaluminate, insulin, andEGF on p130 Cas phosphorylation in immunoprecipitates (Fig-ure 3C). The same amounts of Cas were immunoprecipitated inall samples (Figure 3C, lower panel). p130 Cas was phosphory-lated in basal conditions (Figure 3C, upper panel). However,p130 Cas phosphorylation increased only in fluoroaluminate-treated samples, but not in samples treated with insulin and EGF(Figure 3C, upper panel). These results show that p130 Cas hasall the features of the p130 tyrosine-phosphorylated band influoroaluminate-stimulated lysates from MC3T3-E1 cells. Fur-thermore, there is a selectivity in induction of p130 Cas phos-phorylation by fluoroaluminate and G-protein-coupled receptoragonists over growth factors utilizing receptor tyrosine kinasereceptors. This is in contrast to the role of Fak, which is equallyinduced by fluoroaluminate and growth factors,31 and whichintegrates signaling by integrins and growth factor receptors.28

Fluoroaluminate Enhances Tyrosine Phosphorylation of p130Cas and Fak by Type I Collagen

Fluoroaluminate induces tyrosine phosphorylation of p130 Casand Fak, two proteins known to be phosphorylated in signalingby cell adhesion and known Src substrates in fibroblasticcells.9,21,22,32 In MC3T3-E1 cells, Fak was shown to be increas-ingly phosphorylated by adhesion to type I collagen.32 p130 Caspromotes cell spreading and mediates Fak-promoted cell migra-tion,1,12 and dephosphorylation of p130 Cas leads to the inhibi-tion of cell spreading.21

Because both fluoroaluminate and integrins increase tyrosinephosphorylation of p130 Cas and Fak, the question arises as towhether they can cooperate. To test this, we analyzed proteintyrosine phosphorylation in MC3T3-E1 cells that were firstdetached and then allowed to attach to type I collagen in theabsence or presence of fluoroaluminate (2 mmol/L NaF and 10�mol/L AlCl3; Figure 4A). As shown in Figure 4A, at allconcentrations of type I collagen used for coating of plastic(0.3–10 �g/cm2), we detected enhancement by fluoroaluminateof tyrosine phosphorylation of p130 and p120 proteins, previ-ously identified as p130 Cas and Fak (Figures 2 and 3). Thephosphorylation of other, weakly phosphorylated proteins (p180,p100, p50) was not greatly affected. We conclude that fluoroalu-minate can enhance phosphorylation of p130 Cas and Fak duringcell adhesion to type I collagen.


Effect of Fluoroaluminate on the Attachment and Spreading ofMC3T3-E1 Cells

After showing cooperation of fluoroaluminate and cell adhesionat the molecular level, we examined whether they could coop-erate at the level of cell function. For this purpose, we tested theeffect of addition of fluoroaluminate (1 and 2 mmol/L NaF, 10�mol/L AlCl3) to the cells during their attachment to type Icollagen substratum (Figure 4B). The enhancement of tyrosinephosphorylation of Fak and Cas by fluoroaluminate treatmentduring cell adhesion was also done with 2 mmol/L NaF and 10

�mol/L AlCl3 (Figure 4A). It was not possible to use higherconcentrations of NaF, because they caused cell toxicity in thistype of experiment, whereas lower concentrations were ineffec-tive. After 1 or 2 h of adhesion, MC3T3-E1 cells were fixed,stained with crystal violet, and microphotographed. After 1 h, thecells appeared to spread more in the presence of fluoroaluminate(2 mmol/L NaF/10 �mol/L AlCl3) (Figure 4Ba–c). After 2 h,more cells attached and spread on the type I collagen, and cellspreading appeared to be enhanced by treatment with fluoroalu-minate (Figure 4Bd–f).

Figure 2. Purification of p130 from fluoroaluminate-treated MC3T3-E1 cells and its identification as Cas by mass spectrometry. (A) Purification ofp130 from fluoroaluminate-stimulated MC3T3-E1 cells. Fractions during purification were analyzed by Coomassie staining (left panel) and byantiphosphotyrosine Western blotting (right panel). Thirty micrograms of protein (lanes 1–4) or 0.5% of the eluted fractions (lanes 5 and 6) were loaded.Lanes: 1—cell lysates; 2— HeparinSepharose flow-through; 3—FFQ anion-exchange eluate; 4—4G10Sepharose flow-through; 5 and 6—two elutedfractions from 4G10Sepharose. M, molecular weight markers (Bio-Rad, unstained, broad-range). (B) Search result of the GenePept database with p130tryptic peptides. The p130 monoisotopic tryptic peptide masses obtained from two combined MALDI mass spectra were used. The search resultsidentifying p130 Cas protein are shown in bold. (C) Matching peptides in p130 Cas. Fifteen matching peptides are shown within the mouse p130 Cassequence (bold and underlined). A second peptide (starting with ALY, larger font size) immediately follows the first peptide (starting with MKY), whichmakes them appear as one. The second peptide was submitted to nanoelectrospray MSMS analysis and the resulting fragment ions identified four aminoacid residues within this peptide (Y, A, E, and S, marked in italics).


To obtain quantitative data for both cell attachment andspreading, we performed cell counting and the measurement oftotal cell area with an imaging system (Figure 5). The number ofattached cells increased after 1 h, reaching significance with 2mmol/L NaF (Figure 5A). The total cell area increased after 1and 2 h, reaching significance with 2 mmol/L NaF at both time

points (Figure 5B). The area per cell (total cell area per welldivided by cell number per well) showed a highly significantincrease at 2 h with 2 mmol/L NaF (Figure 5C). These resultsindicated that the effect of fluoroaluminate on cell attachmentwas more pronounced at early timepoints (1 h), whereas later (2h) the effect on cell spreading was predominant.

Discussion

Fluoride has a long history as a bone anabolic and as a protectingagent for teeth. It has been described as the most effective agentin increasing cancellous bone mass with a unique ability toincrease bone formation without coupling to bone resorption.17

Nevertheless, it has never been generally accepted as a therapeu-tic agent for osteoporosis. This is mainly due to its narrowtherapeutic window, because higher doses induce the formationof low-quality bone and several other adverse effects.17,29 Inaddition, the molecular target for fluoride action has remainedelusive, which has hampered understanding of the mechanismsof fluoride action in cells.

Recently, we and others have explored the idea that fluorideacts in tandem with aluminum, which is readily available in traceamounts in food, water, and tissues. In vivo, a combination offluoride and aluminum salts (fluoroaluminate) leads to increasedbone formation in rodents, a feature not reproduced with fluoridealone.2 In humans, administration of fluoride alone is sufficientto induce effects on bone, most likely due to the higher avail-ability of aluminum in human food as compared with food forlaboratory animals. In cells, fluoroaluminate stimulates hetero-trimeric G proteins via binding to the G� subunits in a complexwith guanosine diphosphate (GDP). In contrast to the commonview of fluoroaluminate as a nonspecific activator of G proteins,we showed that, in osteoblastic cells, fluoroaluminate preferen-tially activates Gi subunits (inhibiting adenylate cyclase) over Gssubunits (stimulating adenylate cyclase), thus leading to a netdecrease of cAMP and activation of MAP kinase Erk.31 Inagreement with these data, others have recently detected afluoroaluminate-induced increase in Shc phosphorylation and

Table 1. Induction of p130 tyrosine phosphorylation in MC3T3-E1 cells

Agonist/signaling Concentration p130 PTyr

G-protein-coupledThrombin 1 U/mL �Bradykinin 1 �mol/L �Endothelin-1 50 nmol/L �LPA 2.5–25 �mol/L �PGE2 20 nmol/L �/�FAl 10 mmol/L F�/

10 �mol/L Al3��

FAl � 1% FCS ��Receptor Tyr kinase

EGF 10–100 nmol/L �PDGF 30 ng/mL �Insulin 100 nmol/L �

Serum-deprived MC3T3-E1 cells were stimulated with the indicatedagonists and growth factors for 5–40 min, and the proteins were ex-tracted and analyzed by antiphosphotyrosine western blotting with 4G10monoclonal antibodies. The intensity of p130 tyrosine phosphorylation(p130 PTyr) was evaluated visually. The results are representative of atleast two independent experiments. Where stimulation was lacking, thepresence of the corresponding receptor was confirmed by other tech-niques.KEY: EGF, epidermal growth factor; FAl, fluoroaluminate; LPA, lyso-phosphatidic acid; PDGF, platelet-derived growth factor; PGE2, prosta-glandin E2; PTyr, tyrosine phosphorylation.

Figure 3. Identification of p120 and p130 as Fak and Cas by immu-nodepletion and selective phosphorylation of immunoprecipitated Cas.The lysates from fluoroaluminate-stimulated MC3T3-E1 cells were sub-jected to two rounds of immunoprecipitation with anti-Fak (A) oranti-Cas (B) antibodies. The initial lysate (L) and supernatants after thefirst or second immunoprecipitation (immunodepletion [ID], lanes 1 and2) were analyzed by Western blotting with antiphosphotyrosine antibod-ies [(A) and (B), upper panels] or with anti-Fak and anti-Cas antibodies[(A) and (B), lower panels]. (C) Confluent MC3T3-E1 cultures wereserum-starved for 24 h and stimulated with insulin (Ins, 220 nmol/L) for5 min, EGF (10 nmol/L) for 5 min, and fluoroaluminate (FAl, 10 mmol/LNaF/10 �mol/L AlCl3) for 40 min. P130 Cas was immunoprecipitated(IP) from equal amounts of cell extracts (200 �g protein) using 2 �g ofp130 Cas-specific monoclonal antibody. The precipitated proteins wereanalyzed by Western blotting with antiphosphotyrosine antibodies [(C),upper panel] and with anti-Cas antibodies [(C), lower panel]. Thepositions of p120 Fak and p130 Cas are marked with arrows.


Shc-Grb2 association, which are events known to lead to Erkactivation.5 Furthermore, they detected a decrease in phosphor-ylation of transcription factor CREB (cAMP-response elementbinding protein), in accord with a previously detected decrease incAMP.

Formally, it cannot be excluded that fluoroaluminate acts on

a target other than G proteins. As this study has shown theinfluence of fluoroaluminate on cell adhesion, one might envis-age that fluoroaluminate affects the structural integrity of type Icollagen or that it increases integrin aggregation. Anions, such assulfate, have been reported to affect hydrogen bonding in colla-gen peptides, but only at high concentrations (�500 mmol/L)

Figure 4. Effect of fluoroaluminate on Fak and Cas phosphorylation and on MC3T3-E1 cell adhesion during attachment to type I collagen. (A)MC3T3-E1 cells were deprived from serum, trypsinized, and plated onto wells coated with indicated concentrations of type I collagen. The control wellswithout collagen coating (0) were blocked with bovine serum albumin (BSA). The cells were allowed to attach for 1 h in the absence or presence offluoroaluminate (FAl: 2 mmol/L NaF/10 �mol/L AlCl3). The cellular protein was extracted and equal amounts of protein were analyzed by Westernblotting with antiphosphotyrosine antibodies. The positions of molecular weight markers are indicated on the left, and the positions of p130 Cas andFak are shown on the right. (B) MC3T3-E1 cells were deprived from serum, trypsinized, and plated onto wells coated with 10 �g/cm2 of type I collagen.The cells were allowed to attach during 1 or 2h in the absence or presence of the indicated concentrations of fluoroaluminate (FAl). The cells were fixed,stained, and microphotographed.


and when peptides were in solution.36 The highest NaF concen-tration used in our experiments was 10 mmol/L and the concen-tration of the fluoroaluminate complex was limited by the con-centration of AlCl3 (10 �mol/L). Furthermore, collagen wasimmobilized on plastic and was not in solution. The crystalstructure of collagen peptide and integrin �2�1 domain I doesnot suggest any modulation by anions, but only by divalentcations.7 By contrast, regulation of heterotrimeric G proteins byfluoroalumnate is proven and accepted (reviewed in refs. 4 and

31). Thus, we propose that involvement of G proteins is the mostlikely mechanism for explaining the effects of fluoroaluminateon cell adhesion.

Gi stimulation is not the only type of signaling induced byfluoroaluminate. Prominent tyrosine phosphorylation of severalproteins can be detected as well.31 In contrast to Erk phosphor-ylation, the phosphorylation of these proteins was independent ofpertussis toxin, an inhibitor of Gi subunits. In agreement with thenotion that Erk activation and main protein tyrosine phosphory-lation are on distinct pathways, Erk activity is unaffected bycytochalasin D, whereas the same cytoskeleton-disrupting agentwas shown to reduce phosphorylation of the cytoskeleton-asso-ciated proteins Fak and paxillin.3 Two nonreceptor tyrosinekinases, Src and Pyk2, are activated and associated with eachother upon fluoroaluminate treatment.15 Although Fak was ex-pressed at higher levels than Pyk2 in MC3T3-E1 cells, it asso-ciated only poorly with Src. In this work, we identified two majorfluoroaluminate-induced tyrosine-phosphorylated proteins: (a)tyrosine kinase substrate and an adapter protein p130 Cas and (b)Fak, a tyrosine kinase related to Pyk2. p130 Cas is a knownsubstrate of v- and c-Src,10,11,22,26,33 in accordance with our dataon the activation of Src by fluoroaluminate. It has been shownpreviously that Src, Fak, and p130 Cas are all members ofintegrin-induced focal adhesion complexes; that is, Fak and p130Cas associate following adhesion, and p130 Cas mediates cellmigration and spreading in fibroblasts.1,12,23,25 In MC3T3-E1cells, we showed that phosphorylation of p130 Cas and Fak issensitive to a Src inhibitor and not to pertussis toxin (Figure1C,D), indicating that G protein distinct from Gi must be in-volved in the process. The most likely candidate that mediatesactivation of Src and Pyk2, and phosphorylation of p130 Cas andFak is a subunit from the G12 class, because it was shown thatphosphorylation of Fak, p130 Cas, and paxillin is directly in-duced by activated G�12/13 proteins.20 Because all four classesof G� subunits are present in MC3T3-E1 cells (our unpublisheddata), it is possible that G12/13 transduce the signal fromfluoroaluminate to Fak and Cas.

The results examined so far indicate that both fluoroalumi-nate-induced protein kinases and the main tyrosine-phosphor-ylated proteins belong to the class of molecules implicated inintegrin signaling.23 To test whether fluoroaluminate can alsoinfluence integrin function, we performed adhesion experimentsto type I collagen in the presence of fluoroaluminate. The resultsindicate that fluoroaluminate dose-dependently and significantlyincreased both the number of attached cells and their spreading.The narrow effective concentration range in these experiments(1–2 mmol/L NaF) correlated with a narrow therapeutic windowactive in vivo.17 These data establish a link between signalingand the cellular functions affected by fluoroaluminate. The in-volvement of Src in this process is consistent with the role of Srcfamily kinases in integrin signaling and for turnover of focaladhesion.8,18,30 In conclusion, fluoroaluminate (most likely via Gproteins) increases kinase activity and association of Src andPyk2, and phosphorylation of Fak and p130 Cas. This results inan enhancement of integrin function (attachment, spreading),either by stimulating existing downstream integrin signaling orby influencing the integrin aggregation state from inside of thecells.6

In vivo, increased preosteoblast attachment and spreadingmay result in a faster recruitment of osteoblast precursors, areduction in their apoptosis rate, and increased differentiation, assuggested by recent work in cell culture.9,16 The bone phenotypeof mice with impaired function of integrin �1 subunit (collagentype I receptor subunit) in osteoblasts supports the notion thatsignaling by osteoblast adhesion is crucial for bone formation.37

The embryonic lethality of p130 Cas-deficient and Fak-deficient

Figure 5. Quantification of MC3T3-E1 cell attachment and spreading tocollagen type I in the absence or presence of fluoroaluminate. Theadhesion assays with MC3T3-E1 cells were performed as described inthe legend to Figure 4. Quantification included counting the number ofattached cells per 1 mm2 field (A); measuring the total cell area per wellwith Leica imaging system (B); and calculating the area per cell (spread-ing) from total cell area divided by the number of attached cells (C).Statistically significant differences are indicated: *p � 0.05; **p � 0.01.


mice makes it impossible to draw conclusions on the roles ofthese two proteins in osteoblasts in adult animals.13,14 Furtherstudies with transgenic/conditional knockout mice are necessaryto clarify whether one or both of these two proteins (p130 Casand Fak) are required for the action of fluoroaluminate on bonein vivo. In vitro, it remains to be clarified how p130 Cas and Fakmediate signaling by fluoroaluminate in osteoblasts. Further-more, the molecular basis for a selective signaling by fluoroalu-minate in osteoblastic cells versus other cell types need to beelucidated. Because there are no osteoblast-specific primarytargets (G� proteins) for fluoroaluminate in osteoblasts, theselectivity must be achieved downstream in the signaling path-way. A possibility exists that p130 Cas and Fak contribute to thisspecific signaling. In conclusion, the unexpected connectionbetween fluoroaluminate and osteoblast adhesion to type I col-lagen provides a basis for understanding the action of fluoroalu-minate in defined molecular terms.

Acknowledgments: The authors thank D. Rohner, N.-H. Luong-Nguyen,S. Ewert, M. Schaeublin, and M. Coulot for technical assistance in thevarious aspects of this work. We acknowledge Dr. P. Dennis for advicein purification, Dr. G. Standke for discussions, Dr. M. Glatt for help instatistical analysis, and Dr. J. Green for a critical reading of the manu-script. We thank Dr. C. Ruegg for advice on the adhesion assays.

References

1. Cary, L. A., Han, D. C., Polte, T. R., Hanks, S. K., and Guan, J. L.Identification of p130Cas as a mediator of focal adhesion kinase-promoted cellmigration. J Cell Biol 140:211–221; 1998.

2. Caverzasio, J., Imai, T., Ammann, P., Burgener, D., and Bonjour, J. P.Aluminum potentiates the effect of fluoride on tyrosine phosphorylation andosteoblast replication in vitro and bone mass in vivo. J Bone Miner Res11:46–55; 1996.

3. Caverzasio, J., Palmer, G., Suzuki, A., and Bonjour, J. P. Mechanism of themitogenic effect of fluoride on osteoblast-like cells: Evidences for a G protein-dependent tyrosine phosphorylation process. J Bone Miner Res 12:1975–1983;1997.

4. Chabre, M. Aluminofluoride and beryllofluoride complexes: A new phosphateanalogs in enzymology. Trends Biochem Sci 15:6–10; 1990.

5. Chae, H. J., Chae, S. W., Kang, J. S., Kim, D. E., and Kim, H. R. Mechanismof mitogenic effect of fluoride on fetal rat osteoblastic cells: Evidence for Shc,Grb2 and P-CREB-dependent pathways. Res Commun Mol Pathol Pharmacol105:185–199; 1999.

6. Chrzanowska-Wodnicka, M. and Burridge, K. Rho-stimulated contractilitydrives the formation of stress fibers and focal adhesions. J Cell Biol 133:1403–1415; 1996.

7. Emsley, J., Knight, C. G., Farndale, R. W., Barnes, M. J., and Liddington, R. C.Structural basis of collagen recognition by integrin alpha2beta1. Cell 101:47–56; 2000.

8. Fincham, V. J. and Frame, M. C. The catalytic activity of Src is dispensable fortranslocation to focal adhesions but controls the turnover of these structuresduring cell motility. Eur Med Biol Org J 1998 17:81–92; 1998.

9. Globus, R. K., Doty, S. B., Lull, J. C., Holmuhamedov, M. J., and Damsky,C. H. Fibronectin is a survival factor for differentiated osteoblasts. J Cell Sci111:1385–93; 1998.

10. Hamasaki, K., Mimura, T., Morino, N., Furuya, H., Nakamoto, T., Aizawa, S.,Morimoto, C., Yazaki, Y., Hirai, H., and Nojima, Y. Src kinase plays anessential role in integrin-mediated tyrosine phosphorylation of Crk-associatedsubstrate p130 Cas, Biochem Biophys Res Commun 222:338–343; 1996.

11. Harte, M. T., Hildebrand, J. D., Burnham, M. R., Bouton, A. H., and Parsons,J. T. p130 Cas, a substrate associated with v-Src and v-Crk, localizes to focaladhesions and binds to focal adhesion kinase. J Biol Chem 271:13649–13655;1996.

12. Honda, H., Nakamoto, T., Sakai, R., and Hirai, H. p130(Cas), an assemblingmolecule of actin filaments, promotes cell movement, cell migration, and cellspreading in fibroblasts. Biochem Biophys Res Commun 262:25–30; 1999.

13. Honda, H., Oda, H., Nakamoto, T., Honda, Z.-I., Sakai, R., Suzuki, T., Saito,T., Nakamura, K., Nakao, K., Ishikawa, T., Katsuki, M., Yazaki, Y., and Hirai,

H. Cardiovascular anomaly, impaired actin bundling and resistance to Src-induced transformation in mice lacking p130 Cas. Nature Genet 19:361–365;1998.

14. Ilic, D., Furuta, Y., Takeda, N., Kanazawa, S., Ikawa, Y., Yamamoto, T., andAizawa, S. Reduced cell motility and enhanced focal adhesion formation incells from FAK-deficient mice. Nature 377:539–544; 1995.

15. Jeschke, M., Standke, G. J. R., and Susa, M. Fluoroaluminate induces activa-tion and association of Src and Pyk2 tyrosine kinases in osteoblasticMC3T3-E1 cells. J Biol Chem 273:11354–11361; 1998.

16. Jikko, A., Harris, S. E., Chen, D., Mendrick, D. L., and Damsky, C. H.Collagen integrin receptors regulate early osteoblast differentiation induced byBMP-2. J Bone Miner Res 14:1075–1083; 1999.

17. Kleerekoper, M. Fluoride and the skeleton. In: Bilezikian, J. P., Raisz, L. G.,and Rodan, G. A., Eds. Principles of Bone Biology. San Diego, CA: Academic;1996; 1053–1062.

18. Klinghoffer, M. A., Sachsenmaier, C., Cooper, J. A., and Soriano, P. Src familykinases are required for integrin but not PDGFR signal transduction. Eur MedBiol Org J 18:2459–2471; 1999.

19. Missbach, M., Jeschke, M., Feyen, J., Muller, K., Glatt, M., Green, J., and Susa,M. A novel inhibitor of the tyrosine kinase Src suppresses phosphorylation ofits major cellular substrates and reduces bone resorption in vitro and in rodentmodels in vivo. Bone 24:437–449; 1999.

20. Needham, L. K. and Rozengurt, E. J. G�12 and G�13 stimulate Rho-dependenttyrosine phosphorylation of focal adhesion kinase, paxillin, and p130 Crk-associated substrate. J Biol Chem 273:14626–14632; 1998.

21. Noguchi, T., Tsuda, M., Takeda, H., Takada, T., Inagaki, K., Yamao, T.,Fukunaga, K., Matozaki, T., and Kasuga, M. Inhibition of cell growth andspreading by stomach cancer-associated protein-tyrosine phosphatase-1(SAP-1) through dephosphorylation of p130cas. J Biol Chem 276:15216–15224; 2001.

22. Nojima, Y., Morino, N., Mimura, T., Hamasaki, K., Furuya, H., Sakai, R., Sato,T., Tachibana, K., Morimoto, C., Yazaki, Y., and Hirai, H. Integrin-mediatedcell adhesion promotes tyrosine phosphorylation of p130Cas, a Src homology3-containing molecule having multiple Src homology 2-binding motifs. J BiolChem 270:15398–15402; 1995.

23. Parsons, J. T. Integrin-mediated signalling: Regulation by protein tyrosinekinases and small GTP-binding proteins. Curr Opin Cell Biol 8:146–152; 1996.

24. Polte, T. R. and Hanks, S. K. Interaction between focal adhesion kinase andCrk-associated tyrosine kinase substrate p130 Cas. Proc Natl Acad Sci USA92:10678–10682; 1995.

25. Polte, T. R. and Hanks, S. K. Complexes of focal adhesion kinase (FAK) andCrk-associated substrate (p130(Cas)) are elevated in cytoskeleton-associatedfractions following adhesion and Src transformation. Requirements for Srckinase activity and FAK proline-rich motifs: J Biol Chem 272:5501–5509;1997.

26. Sakai, R., Iwamatsu, A., Hirano, N., Ogawa, S., Tanaka, T., Mano, H., Yazaki,Y., and Hirai, H. A novel signaling molecule, p130, forms stable complexes invivo with v-Crk and v-Src in a tyrosine phosphorylation-dependent manner.Eur Med Biol Org J 13:3748–3756; 1994.

27. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. Mass spectrometricsequencing of proteins silver-stained polyacrylamide gels. Anal Chem 68:850–858; 1996.

28. Sieg, D. J., Hauck, C. R., Ilic, D., Klingbeil, C. K., Schaefer, E., Damsky, C. H.,and Schlaepfer, D. D. FAK integrates growth-factor and integrin signals topromote cell migration. Nat Cell Biol 2:249–256; 2000.

29. Susa, M. Heterotrimeric G proteins as fluoride targets in bone [review]. Int JMol Med 3:115–126; 1999.

30. Susa, M., Missbach, M., and Green, J. Src inhibitors: Drugs for the treatmentof osteoporosis, cancer or both? Trends Pharmacol Sci 21:489–495; 2000.

31. Susa, M., Standke, G. J. R., Jeschke, M., and Rohner, D. Fluoroaluminateinduces pertussis toxin-sensitive protein phosphorylation: Differences inMC3T3-E1 osteoblastic and NIH3T3 fibroblastic cells. Biochem Biophys ResCommun 235:680–684; 1997.

32. Takeuchi, Y., Suzawa, M., Kikuchi, T., Nishida, E., Fujita, T., and Matsumoto,T. Differentiation and transforming growth factor-beta receptor down-regula-tion by collagen-�2�1 integrin interaction is mediated by focal adhesion kinaseand its downstream signals in murine osteoblastic cells. J Biol Chem 272:29309–29316; 1997.

33. Vuori, K. and Ruoslahti, E. Tyrosine phosphorylation of p130Cas and cortactinaccompanies integrin-mediated cell adhesion to extracellular matrix. J BiolChem 270:22259–22262; 1995.

34. Wasnich, R. D. Epidemiology of osteoporosis. In: Favus, M. J., Ed. Primer on


the Metabolic Bone Diseases and Disorders of Mineral Metabolism, 3rd Ed.Philadelphia, PA: Lippincott-Raven; 1996; 249–251.

35. Wilm, M. and Mann, M. Analytical properties of the nanoelectrospray ionsource. Anal Chem 68:1–8; 1996

36. Zanaboni, G., Rossi, A., Onana, A. M., and Tenni, R. Stability and networks ofhydrogen bonds of the collagen triple helical structure: Influence of pH andchaotropic nature of three anions. Matrix Biol 19:511–520; 2000.

37. Zimmerman, D., Jin, F., Leboy, P., Hardy, S., and Damsky, C. Impaired bone

formation in transgenic mice resulting from altered integrin function in osteo-blasts. Dev Biol 220:2–15; 2000.

Date Received: January 22, 2001Date Revised: September 5, 2001Date Accepted: September 5, 2001


Fluoroaluminate stimulates phosphorylation of p130 Cas and Fak and increases attachment and...

Documents

Transcript of Fluoroaluminate stimulates phosphorylation of p130 Cas and Fak and increases attachment and...