Search

Tropomyosin (TM) is a regulatory protein of actomyosin system. Muscle type-specific expression of TM isoforms is generated from different genes and by alternative splicing. beta-TM isoforms in chicken skeletal and smooth muscles are encoded by a single gene and transcribed from the same promoter. We previously reported a smooth muscle cell (SMC) phenotype-dependent change in beta-TM expression (Kashiwada, K., Nishida, W., Hayashi, K., Ozawa, K., Yamanaka, Y., Saga, H., Yamashita, T., Tohyama, M., Shimada, S., Sato, K., and Sobue, K. (1997) J. Biol. Chem. 272, 15396-15404), and identified beta-TM as an SMC-differentiation marker. Here, we characterized the transcriptional machinery of the beta-TM gene in SMCs. Promoter and gel mobility shift analyses revealed an obligatory role for serum response factor and its interaction with the CArG box sequence in the SMC-specific transcription of the beta-TM gene in differentiated SMCs. We further isolated a novel homologue of the Barx homeoprotein family, Barx1b, from chicken gizzard. Barx1b was exclusively localized to SMCs of the upper digestive organs and their attached arteries and to craniofacial structures. Serum response factor and Barx1b bound each other directly, coordinately transactivated the beta-TM gene in differentiated SMCs and heterologous cells, and formed a ternary complex with a CArG probe. Taken together, these results suggest that SRF and Barx1b are coordinately involved in the SMC-specific transcription of the beta-TM gene in the upper digestive organs and their attached arteries.

The reduction of flavin in hepatic NADH-cytochrome b5 reductase by the hydrated electron (eaq-) was investigated by pulse radiolysis. The eaq- reduced the flavin of NADH-cytochrome b5 reductase to form the red semiquinone between pH 5 and 9. The spectrum of the red semiquinone differs from that of enzyme reduced by dithionite in the presence of NAD+. After the first phase of the reduction, conversion of the red to blue semiquinone was observed at acidic pH. Resulting products are the blue (neutral) or red (anionic) semiquinone or a mixture of the two forms. The pK value for this flavin radical was approximately 6.3. Subsequently, the semiquinone form reacted by dismutation to form the oxidized and the fully reduced forms of the enzyme with a rate constant of 1 x 10(3) M-1 s-1 at pH 7.1. In the presence of NAD+, eaq- reacted with NAD+ to yield NAD(.). Subsequently, NAD. transferred an electron to NAD+-bound oxidized enzyme to form the blue and red semiquinone or mixture of the two forms of the enzyme, where pK value of this flavin radical was approximately 6.3. The blue semiquinone obtained at acidic pH was found to convert to the red semiquinone with a first order rate constant of 90 s-1, where the rates were not affected by pH or the concentration of NAD+. The final product is NAD+-bound red semiquinone of the enzyme.

The expressional regulation of chicken alpha1 integrin in smooth muscle cells was studied. The alpha1 integrin mRNA was expressed developmentally and was distributed dominantly in vascular and visceral smooth muscles in chick embryos. In a primary culture of smooth muscle cells, alpha1 integrin expression was dramatically down-regulated during serum-induced dedifferentiation. Promoter analyses revealed that the 5'-upstream region (-516 to +281) was sufficient for transcriptional activation in differentiated smooth muscle cells but not in dedifferentiated smooth muscle cells or chick embryo fibroblasts. Like other alpha integrin promoters, the promoter region of the alpha1 integrin gene lacks TATA and CCAAT boxes and contains binding sites for AP1 and AP2. The essential difference from other alpha integrin promoters is the presence of a CArG box-like motif. Deletion and site-directed mutation analyses revealed that the CArG box-like motif was an essential cis-element for transcriptional activation in differentiated smooth muscle cells, whereas the binding sites for AP1 and AP2 were not. Using specific antibodies, a nuclear protein factor specifically bound to the CArG box-like motif was identified as serum response factor. These results indicate that alpha1 integrin expression in smooth muscle cells is regulated transcriptionally in a phenotype-dependent manner and that serum response factor binding plays a crucial role in this regulation.

We have cloned a novel matrix metalloproteinase (MMP) from cultured chicken embryo fibroblasts. The cDNA-derived protein sequence contains 608 amino acids including a C-terminal hydrophobic transmembrane domain of 24 amino acids and a cytoplasmic domain of 20 amino acids. This chicken MMP is 72% similar to a recently described membrane-type MMP (MT-MMP) from human placenta (Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E., and Seiki, M. (1994) Nature 370, 61-65). Accordingly, we name this novel MMP chicken MT-MMP. As shown by Northern blotting, two MT-MMP mRNAs of 6 and 10 kilobases are constitutively expressed but only modestly regulated by growth factors and cytokines in cultured chicken embryo fibroblasts. Both mRNAs are abundant in the head and body of 8- and 9-day-old chicken embryos. As shown by in situ mRNA hybridization, MT-MMP is expressed in embryonic neural tube, spinal ganglia, and respiratory epithelium, as well as in developing cartilage and muscle. Using reverse transcription-polymerase chain reaction, we have found MT-MMP mRNA in 2-day-old chicken embryos and extraembryonic membranes. In addition, a strong correlation was observed between the mRNA expression of MT-MMP and 72-kDa type IV collagenase. Collectively, the early MT-MMP mRNA expression and its co-localization in several tissues with 72-kDa type IV collagenase mRNA suggest that the MT-MMP plays an important role in early development.

The phospholipase A2 (PLA2) inhibitor PLIbeta, purified from the blood plasma of Chinese mamushi snake (Agkistrodon blomhoffii siniticus), is a 160-kDa trimer with three 50-kDa subunits; and it inhibits specifically the enzymatic activity of the basic PLA2 from its own venom (Ohkura, N., Okuhara, H., Inoue, S., Ikeda, K., and Hayashi, K. (1997) Biochem. J. 325, 527-531). In the present study, the 50-kDa subunit was found to be glycosylated with N-linked carbohydrate, and enzymatic deglycosylation decreased the molecular mass of the 50-kDa subunit to 39-kDa. One 160-kDa trimer of PLIbeta was found to form a stable complex with three basic PLA2 molecules, indicating that one basic PLA2 molecule would bind stoichiometrically to one subunit of PLIbeta. A cDNA encoding PLIbeta was isolated from a Chinese mamushi liver cDNA library by use of a probe prepared by a polymerase chain reaction on the basis of the partially determined amino acid sequence of the subunit. The cDNA contained an open reading frame encoding a 23-residue signal sequence followed by a 308-residue protein, which contained the sequences of all the peptides derived by lysyl endopeptidase digestion of the subunit. The molecular mass of the mature protein was calculated to be 34,594 Da, and the deduced amino acid sequence contained four potential N-glycosylation sites. The sequence of PLIbeta showed no significant homology with that of the known PLA2 inhibitors. But, interestingly, it exhibited 33% identity with that of human leucine-rich alpha2-glycoprotein, a serum protein of unknown function. The most striking feature of the sequence is that it contained nine leucine-rich repeats (LRRs), each of 24 amino acid residues and thus encompassing over two-thirds of the molecule. LRRs in PLIbeta might be responsible for the specific binding to basic PLA2, since LRRs are considered as the motifs involved in protein-protein interactions.

Isoform diversity of tropomyosin is generated from the limited genes by a combination of differential transcription and alternative splicing. In the case of the alpha-tropomyosin (alpha-TM) gene, exon 2a rather than exon 2b is specifically spliced in alpha-TM-SM mRNA, which is one of the major tropomyosin isoforms in smooth muscle cells. Here we demonstrate that expressions of alpha-tropomyosin and caldesmon isoforms are coordinately regulated in association with phenotypic modulation of smooth muscle cells. Molecular cloning and Western and Northern blottings have revealed that in addition to the down-regulation of beta-TM-SM, alpha-TM-SM converted to alpha-TM-F1 and alpha-TM-F2 by a selectional change from exon 2a to exon 2b during dedifferentiation of smooth muscle cells in culture. Simultaneously, a change of caldesmon isoforms from high Mr type to low Mr type was also observed by alternative selection between exons 3b and 4 in the caldesmon gene during this process. In contrast, cultured smooth muscle cells maintaining a differentiated phenotype continued to express alpha-TM-SM, beta-TM-SM, and high Mr caldesmon. In situ hybridization revealed specific coexpression of alpha-TM-SM and high Mr caldesmon in smooth muscle in developing embryos. These results suggest a common splicing mechanism for phenotype-dependent expression of tropomyosin and caldesmon isoforms in both visceral and vascular smooth muscle cells.

The amino acid sequences of beta-glucosidases from Cellvibrio gilvus and Agrobacterium tumefaciens show about 40% similarity. The pH/temperature optima and stabilities and substrate specificities of the two enzymes are quite different. C. gilvus beta-glucosidase exhibits an optimum pH of 6.2-6.4 and temperature of 35 degrees C, whereas the corresponding values for A. tumefaciens are 7.2-7.4 and 60 degrees C, respectively. The substrate specificity of A. tumefaciens enzyme toward different aryl glycosides is broader than C. gilvus enzyme. To analyze these properties further, three chimeric beta-glucosidases were constructed by substituting segments from the C-terminal homologous region of C. gilvus beta-glucosidase gene with that of A. tumefaciens. The chimeric enzymes were characterized with respect to pH/temperature activity and stability and substrate specificity. Chimeric enzymes exhibited chromatographic behavior similar to that of C. gilvus enzyme. However, enzymatic properties of chimeras were admixtures of those of the two parents. The chimeric enzymes were optimally active at 45-50 degrees C and pH 6.6-7.0. Km values of chimeric enzymes for the various saccharides were admixtures of both parental enzymes. These results suggest that the two domains of C. gilvus and A. tumefaciens enzymes probably can fold independently. The homologous C-terminal region in beta-glucosidase appears to play an important role in determining enzyme characteristics. Changes in the properties on substitution of segments in this region might be related to the enzyme specificity, and beta-glucosidases with improved properties can be prepared by manipulating this region.

Caldesmon, which plays a vital role in the actomyosin system, is distributed in smooth muscle and non-muscle cells, and its isoformal interconversion between a high M(r) form and low M(r) form is a favorable molecular event for studying phenotypic modulation of smooth muscle cells. Genomic analysis reveals two promoters, of which the gizzard-type promoter displays much higher activity than the brain-type promoter. Here, we have characterized transcriptional regulation of the gizzard-type promoter. Transient transfection assays in chick gizzard smooth muscle cells, chick embryo fibroblasts, mouse skeletal muscle cell line (C2C12), and HeLa cells revealed that the promoter activity was high in smooth muscle cells and fibroblasts, but was extremely low in other cells. Cell type-specific promoter activity depended on an element, CArG1, containing a unique CArG box-like motif (CCAAAAAAGG) at -315, while multiple E boxes were not directly involved in this event. Gel shift assays showed the specific interaction between the CArG1 and nuclear protein factors in smooth muscle cells and fibroblasts. These results suggest that the CArG1 is an essential cis-element for cell type-specific expression of caldesmon and that the function of CArG1 might be controlled under phenotypic modulation of smooth muscle cells.

Mouse heart cells in culture synthesized and secreted a nerve growth factor (NGF) which cross-reacted with mouse submaxillary gland beta NGF. The synthesis of NGF by heart cells was greatly reduced by addition of actinomycin D or cycloheximide but not by cytosine arabinoside, suggesting that the synthesis of NGF by heart cells required DNA transcription but not DNA replication. Heart NGF and beta NGF were virtually equivalent with respect to the neurite outgrowth stimulating activity and immunologically indistinguishable. The molecular weight of heart NGF estimated by the gel filtration method was identical with that of beta NGF. Isoelectric point of heart NGF was about 9.5, like beta NGF. These results suggest that NGF synthesized and secreted by mouse heart cells is an identical molecule to beta NGF of the mouse submaxillary gland.

L-M cells, a mouse fibroblast cell line, synthesized and secreted a nerve growth factor (NGF). The neurite outgrowth stimulatory activity, immunoreactivity, molecular weight, and isoelectric point of L-M cell NGF were identical to those of beta-NGF of the mouse submaxillary gland. Treatment of the cells with either norepinephrine or epinephrine in the range of 0.05-0.2 mM for 24 h resulted in a 3-20-fold increase in NGF content in the medium of the L-M cells. The NGF of epinephrine-treated cell was identical to that of control cell. The stimulation of the increase in NGF content was observed after a 4-h lag time. The rate of incorporation of [3H]leucine into trichloroacetic acid-insoluble materials was essentially unchanged during the treatment. These results suggested that norepinephrine and epinephrine stimulated the de novo synthesis and secretion of NGF protein. Evidence is also presented to indicate that the effects of the drugs are due to the catechol part of the molecule and not mediated by adrenergic receptors.

The reduction of bovine heart cytochrome oxidase by the 1-methylnicotinamide (MNA) radical was investigated by the use of pulse radiolysis. With the decay of the MNA radical, the absorption at 445 and 605 nm, a characteristic to ferrous heme a of the oxidase, increased. The kinetic difference spectrum obtained was similar to that of the fully reduced minus the fully oxidized form of the oxidase, and was not different from that obtained in the reaction of the MNA radical with the mixed valence CO complex of the oxidase, where heme a3is the CO-bound reduced form with heme a oxidized. This suggests that the absorption changes at 445 and 605 nm arise from the reduction of heme a, not heme a3. In order to elucidate the contribution of “visible” copper in this reaction, the absorption of the oxidase in the near-infrared region was measured. A decrease of the 830 nm band due to the reduction of visible copper was detected with a half-life of 5 µs. This absorption change obeyed pseudo-first order kinetics and its rate constant increased with the concentration of the oxidase. This suggests that the absorption change at 830 nm is followed by a bimolecular reaction of the MNA radical with visible copper of the oxidase. After the first phase of the reduction, the return of the 830 nm band corresponding to oxidation of the copper was observed with a half-life of 100 µs. Concomitantly, the absorption at 605 and 445 nm due to the reduction of heme a increased. The rates of oxidation of the copper were identical to those of the reduction of heme a and independent of the oxidase concentration. This suggests that the MNA radical reacts with visible copper of the oxidase with a second order rate constant of 1.5 × 109m−1s−1and subsequently the electron flows to heme a by intramolecular electron migration with a first order rate constant of 1.8 × 104s−1. An activation energy of the intramolecular electron transfer was calculated to be 2.8 kcal/mol in the range 4–33 °C.

We have previously characterized a 95-kDa plasma membrane receptor for low and very low density lipoproteins in chicken oocytes (George, R., Barber, D. L., and Schneider, W. J. (1987) J. Biol. Chem.262, 16838–16847). We now report that somatic cells of chickens, such as fibroblasts, express a different receptor for these lipoproteins. This receptor has a Mrof 130,000 and is part of a regulatory system for cholesterol homeostasis analogous to the low density lipoprotein receptor pathway in mammalian cells. Oocytes produce only the 95-kDa receptor, while fibroblasts synthesize exclusively the 130-kDa receptor. In addition to their different Mrvalues, another distinctive feature of the two proteins was revealed by ligand blotting experiments: the oocyte receptor bound rabbit β-VLDL (a class of apolipoprotein-B and -E containing lipoprotein particles), whereas the fibroblast receptor did not. Furthermore, polyclonal rabbit antibodies that recognize the oocyte 95-kDa receptor failed to cross-react with the 130-kDa protein on fibroblasts [corrected]. We suggest that different receptors have evolved in the chicken in order to facilitate the deposition of lipids into oocytes (i.e.yolk formation) with concomitant maintenance of cholesterol homeostasis in extraoocytic tissues.

The reaction of superoxide radical (O2-) with the heme protein indoleamine 2,3-dioxygenase has been investigated by the use of pulse radiolysis. In the absence of the substrate tryptophan (Trp), the ferric enzyme reacted quantitatively with (O2-) to form the oxygenated enzyme. The rate constant for the reaction (8.0 × 106M−1s−1at pH 7.0) increased with a decrease in pH. In the presence of low concentrations of L-Trp (∼ 50 µM), under which the catalytic site of the ferric enzyme is >99% Trp-free at pH 7.0, the only spectral species observed upon (O2-) binding was L-Trp-bound oxygenated enzyme, the ternary complex. This suggests that under the conditions employed (O2-) binds first to the ferric enzyme to form the oxygenated enzyme and is followed by rapid binding of L-Trp. It was also found that absorbance changes (Δ A) for the enzyme after the pulse were significantly decreased when an increased L-Trp concentration was employed. A 50% decrease in Δ A was caused with ∼50 μML-Trp at pH 7.0. Similar results were also observed with other indole derivatives with decreasing Δ A values in the order of indole, 3-indoleethanol, α-methyl-DL-Trp, and D-Trp. These results suggest that there exists a binding site for these compounds in the dioxygenase different from the catalytic site for Trp and, most significantly, that binding of Trp to the effector binding site of the ferric enzyme markedly inhibits its reaction with (O2-).

Densitometric scans of electron micrographs of type I collagen segment long spacing crystallites stained with uranyl acetate or phosphotungstic acid and uranyl acetate have been correlated with computer-synthesized scans derived from the sequence of the alpha 1(I) and alpha 2 chains. Three models that differ in the location of the alpha 2 chain were used in the computer synthesis; Models A, B, and C have the alpha 2 chain in the A, B, and C chain positions, respectively. For all 13 experimental scans, the order of decreasing correlation with the models was found to be A,B,C. The probability of getting the same order of decreasing correlation all 13 times is 6/6(13). It was also determined at the 0.99 confidence level that the mean of the differences in the correlation coefficients among the models is greater than 0, supporting the conclusion that sequence-derived models best fit the experimental data when the alpha 2 chain is in the A position. Our results also agree with recent studies that show that uranyl ions bind to both positively and negatively charged residues on collagen type I.

We have studied the time course of the absorption of bovine liver catalase after pulse radiolysis with oxygen saturation in the presence and absence of superoxide dismutase. In the absence of superoxide dismutase, catalase produced Compound I and another species. The formation of Compound I is due to the reaction of ferric catalase with hydrogen peroxide, which is generated by the disproportionation of the superoxide anion (O-2). The kinetic difference spectrum showed that the other species was neither Compound I nor II. In the presence of superoxide dismutase, the formation of this species was found to be inhibited, whereas that of Compound I was little affected. This suggests that this species is formed by the reaction of ferric catalase with O-2 and is probably the oxy form of this enzyme (Compound III). The rate constant for the reaction of O-2 and ferric catalase increased with a decrease in pH (cf. 4.5 X 10(4) M-1 s-1 at pH 9 and 4.6 X 10(6) M-1 s-1 at pH 5.). The pH dependence of the rate constant can be explained by assuming that HO2 reacts with this enzyme more rapidly than O-2.

We have reported the localization on yeast microsomes for a modified P450c27 (mic-P450c27) that contains the microsomal targeting signal of bovine P450c17 in front of the mature form of rat mitochondrial P450c27 (Sakaki, T., Akiyoshi-Shibata, M., Yabusaki, Y., and Ohkawa, H. (1992) J. Biol. Chem. 267, 16497-16502). In this study, we found that mic-P450c27 could be reduced by NADPH in the yeast microsomes without supplement of its physiological redox partners, adrenodoxin and NADPH-adrenodoxin reductase. In order to elucidate the direct electron transfer from NADPH-P450 reductase to mic-P450c27, we carried out simultaneous expression of mic-P450c27 and yeast P450 reductase. The reduction rate of mic-P450c27 was increased by overproduction of yeast P450 reductase, roughly in proportion to the reductase content in the microsomes. In addition, we constructed a fused enzyme between mic-P450c27 and yeast P450 reductase. The reduction rate of heme iron in the fused enzyme was too rapid to be measured. These recombinant yeast microsomes showed a notable 27-hydroxylation activity toward 5beta-cholestane-3alpha,7alpha, 12alpha-triol in the absence of adrenodoxin and adrenodoxin reductase. Finally, we purified mic-P450c27 from the recombinant yeast microsomes and reconstituted the hydroxylation system in liposomal membranes using the purified mic-P450c27 and yeast NADPH-P450 reductase. Mic-P450c27 was reduced by NADPH and showed its monooxygenase activity on the reconstituted system. Therefore, yeast NADPH-P450 reductase alone was found to transfer two electrons from NADPH to mic-P450c27. These results clearly show that mic-P450c27 not only localizes on the microsomes but also functions as a microsomal cytochrome P450 that accepts electrons from NADPH-P450 reductase.

Two different Mr forms of caldesmon as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Mr values in the range of 120,000-150,000, h-caldesmon and 70,000-80,000, l-caldesmon) have been already identified. h-Caldesmon is predominantly expressed in smooth muscle cells, whereas l-caldesmon widely distributes in non-muscle cells. Most recently, the molecular cloning of h-caldesmon has been reported (Hayashi, K., Kanda, K., Kimizuka, F., Kato, I., and Sobue, K. (1989) Biochem. Biophys. Res. Commun. 164, 503-511; Bryan, J., Imai, M., Lee, R., Moore, P., Cook, R. G., and Lin, W-G. (1989) J. Biol. Chem. 264, 13873-13879). The calculated Mr of this protein from its primary structure is 88,743. Here, the nucleotide and deduced amino acid sequences of l-caldesmon have been determined by cloning and sequencing the cDNA from chick brain and compared with those of h-caldesmon. The l-caldesmon cDNA encodes a sequence of 517 amino acids with the calculated Mr of 58,844. Two isoforms of caldesmon conserve the completely identical sequences in the NH2- and COOH-terminal domains except for the insertion of Ala-508 in l-caldesmon. Interestingly, the central repeating sequence of h-caldesmon (residues 201-447) is deleted in the l-caldesmon molecule. The short NH2-terminals of two caldesmons individually show the unique sequences. The results of Northern and Southern blot analyses suggest that two mRNAs (4.8 and 4.1 kilo-bases) coding for caldesmon isoforms may be generated from a single gene by alternative splicing. Using a series of truncated caldesmons expressed in Escherichia coli, the common calmodulin-, tropomyosin-, and actin-binding sites and the minimum regulatory domains, which are involved in the Ca2(+)-dependent regulation of actin-myosin interaction, have been identified within the limited consensus sequences (residues 381-433 for l-caldesmon and residues 636-688 for h-caldesmon).

Two systems were used to generate collagen fibrils in vitro by enzymic cleavage of intermediates in the conversion of procollagen to collagen. In one system fibrils were generated by using procollagen NH2-terminal proteinase to cleave pNcollagen, the intermediate which contains the NH2-terminal but not the COOH-terminal propeptides found in procollagen. When pNcollagen was incubated with procollagen NH2-terminal proteinase, the NH2-terminal propeptides were enzymically cleaved from the protein, and there was an increase in the turbidity of the solution over and above the turbidity observed with pNcollagen alone. Electron microscope examination of the samples demonstrated that the increase in turbidity was associated with the assembly of collagen fibrils. The fibrils had a mean diameter of 104 nm +/- 51.7 S.D. or about the same as fibrils formed from pNcollagen alone. However, the fibrils formed by enzymic cleavage of pNcollagen had a more distinct gap-overlap pattern and they appeared to be more tightly packed than fibrils of pNcollagen. Varying the concentration of enzyme varied both the rate of enzymic cleavage of the pNcollagen and the rate of fibril assembly, but there was no consistent effect on the diameter or morphology of the fibrils. In the second system, fibrils were generated with a recently described procedure (Miyahara, M., Njieha, F. K., and Prockop, D. J. (1982) J. Biol. Chem. 257, 8442-8448) in which procollagen COOH-terminal proteinase is used to cleave pCcollagen, the intermediate containing the COOH-terminal but not the NH2-terminal propeptides found in procollagen. When incubated with procollagen COOH-terminal proteinase, the COOH-terminal propeptides were cleaved and collagen fibrils assembled. The collagen fibrils were unusually thick with a mean diameter of 1184 nm +/- 291 S.D. The large diameters of the fibrils made it possible to demonstrate by scanning electron microscopy that each fibril was comprised of a bundle of subfibrils packed into a right-handed helix. The fibrils frequently had branch points which appeared to consist of subfibrils which separated from the main axis of the structure. Also, the surface of the fibrils was scalloped at 270- to 300-nm intervals, suggesting that some of the collagen molecules on the surface were in a 4D staggered array. The results suggested the hypothesis that the order in which the NH2-terminal and COOH-terminal propeptides are cleaved in the conversion of procollagen to collagen may provide a mechanism for controlling the diameter, or both the diameter and morphology, of collagen fibrils.

Under conventional culture conditions, smooth muscle cells display their phenotypic modulation from a differentiated to a dedifferentiated state. Here, we established a primary culture system of smooth muscle cells maintaining a differentiated phenotype, as characterized by expression of smooth muscle-specific marker genes such as h-caldesmon and calponin, cell morphology, and ligand-induced contractility. Laminin retarded the progression of dedifferentiation of smooth muscle cells. Insulin-like growth factors (IGF-I and IGF-II) and insulin markedly prolonged the differentiated phenotype, with IGF-I being the more potent. In contrast, serum, epidermal growth factor, transforming growth factors, and platelet-derived growth factors potently induced dedifferentiation compared with angiotensin II, arginine-vasopressin, and basic fibroblast growth factor. Using the present culture system, we investigated signaling pathways regulating a phenotype of smooth muscle cells. In cultured cells, IGF-I specifically activated phosphatidylinositol 3-kinase (PI3-kinase) and its downstream target, protein kinase B, but not mitogen-activated protein kinases. Specific inhibitors of PI3-kinase (wortmannin and LY294002) induced dedifferentiation of smooth muscle cells even when they were cultured on laminin under IGF-I-stimulated conditions. The sole effect of laminin to retard the dedifferentiation was completely blocked by anti-IGF-I antibody, and laminin promoted the endogenous expression of IGF-I in cultured cells. The reduced promoter activity of the caldesmon gene induced by platelet-derived growth factor BB was overcome by the forced expression of the constitutive active form of PI3-kinase p110alpha catalytic subunit. These findings suggest that an IGF-I signaling pathway through PI3-kinase plays a critical role in maintaining a differentiated phenotype of smooth muscle cells.