Basal lamina inhibition suppresses synthesis of calcium-dependent proteins associated with mammary epithelial cell spreading

Spreading of mouse mammary epithelial cells on collagen gels is closely correlated with the synthesis of a group of putative calcium-binding proteins (CBP) (Braslau et al., Exp cell res 155 (1984) 213). Collagen synthesis was shown to occur during cell spreading, while omission of serum prevented cell spreading and the synthesis of collagen. The proline analogues cis-hydroxyproline and L-azetidine-2-carboxylic acid were shown to inhibit epithelial cell spreading and to suppress the collagen synthesis that occurs during serum-supported cell spreading. Inhibition of collagen synthesis resulted in the inhibition of CBP synthesis associated with cell spreading. In contrast, the collagen cross-linking inhibitor B-aminopropionitrile did not inhibit cell spreading nor did it suppress collagen synthesis; CBP synthesis was also normal during treatment with this inhibitor. Thus, mammary epithelial cell spreading on collagen gels and CBP synthesis can both be suppressed by inhibition of collagen synthesis indicating that they may be integrated in some manner. It is suggested that inhibition of cell spreading during inhibition of collagen synthesis results from failure to assemble a normal basal lamina; this may in turn signal suppression of CBP synthesis.     

Collagen synthesis and deposition during mammary epithelial cell spreading on collagen gels

Mouse mammary epithelial cultivated on collagen gels demonstrate active spreading as the cells form monolayers. In this novel system, initiation of cell spreading is preceded by de novo synthesis of type IV collagen. The newly synthesized collagen is partitioned such that after 48 hr, approximately 24% is found in the culture medium, 35% is intracellular, and 41% is deposited in the extracellular matrix of the developing epithelium. Cultures deprived of serum failed to spread and to synthesize collagen. Proline analogues were shown to inhibit cell spreading and to suppress collagen synthesis in a dose-dependent manner. Cytochalasin D inhibition of F-actin elongation was shown to prevent cell spreading but not to suppress total collagen synthesis. During cytochalasin D treatment, inhibition of cell spreading was shown to result from failure to deposit or to maintain deposited collagen in the epithelium extracellular matrix. The data indicate that synthesis and extracellular deposition of a major basal lamina component (viz. type IV collagen) must precede and then accompany epithelial cell spreading in collagen gel culture. It is suggested that the microfilament apparatus, through some hypothetical integral membrane protein, can anchor extracellular type IV collagen, which then provides a necessary condition for cell spreading.     

Synthesis of novel calcium-dependent proteins associated with mammary epithelial cell migration and differentiation

A class of proteins from mouse mammary epithelial cells has been isolated which, like the calcium-binding protein calmodulin (CaM), binds to phenothiazine in a calcium-dependent manner. These proteins do not bind to phenothiazine through binding to CaM; we infer that they are calcium-binding proteins, and that they may be related to the similarly isolated 'calcimedins' of Moore, P D & Dedman, J, J biol chem 257 (1982) 9663 [8]. In primary cultures of mouse mammary cells on collagen gels, synthesis of certain of these proteins is associated with the spreading of cells to form monolayers; failure of cells to spread and differentiate, through omission of serum from culture medium, results in the inhibition of calcium-binding protein synthesis, with the exception of CaM and a 15 kD species. The CaM/15 kD pair are prominent during all phases of culture, and are secreted during the secretory differentiation phase of culture (floating gels). We propose that these calcium-binding proteins play a specific role in the motility of mammary epithelial cells and that they may also be involved in mammary secretory differentiation.     

Mammary gland Ca2+-binding (-dependent) proteins: Identification as calelectrins and calpactin I/p36

Calcium-binding (-dependent) proteins (CBPs) associated with the spreading of mammary epithelial cell cultures have been identified as various calelectrins and calpactins (p36). In immunoblot analysis, the CBPs of 30–36 kD and 68–70 KD variously react with different calelectrin and calpactin I monomer/p36 antisera. The same immunoreactive proteins were shown to be present in virgin mammary glands and collagen gel mouse mammary epithelial cell cultures. The mammary CBPs show extensive immunochemical relatedness; however, they fail to show cross-reaction with antiserum to calpactin II (lipocortin) antiserum. These immunoreactive CBPs comigrate in electrophoresis with ³⁵S-methionine-labeled CBPs isolated from mammary epithelial cell cultures. Unlike calmodulin, the mammary CBPs that correspond to calelectrins and calpactin I monomer/p36 are not stable to thermal denaturation. The mammary CBPs bind to epithelial cell membranes in a Ca²⁺-dependent manner and are differentially released from ruptured cells, compared with calmodulin, suggesting subcellular localization. Phenothiazineagarose and phenylagarose are equivalent in their ability to bind the mammary CBPs. Thus, mammary gland CBPs of 30–36 kD and 68–70 kD have been shown to be related or equivalent to the calelectrins and to calpactin I monomer/p36. Since these proteins are known to bind Ca²⁺, we conclude that the mammary gland CBPs are also Ca²⁺-binding proteins. The mammary gland CBPs are immunologically related and probably represent members of a larger family of related proteins.     

Developmental regulation of calcium-binding proteins (calelectrins and calpactin I) in mammary glands

We recently showed that mammary glands contain a novel class of calcium-binding proteins (CBPs) that bind to membranes in a calcium-dependent manner. We have also established that these mammary CBPs are equivalent to the calelectrins and calpactin I/p36. Since it has been suggested that these proteins might be involved in exocytosis, we examined mammary glands for these CBPs during secretory differentiation. Immunohistochemical examination showed glands from virgin animals to be rich in calelectrins and calpactin I/p36, while glands from lactating animals contained little immunoreactive material. In addition, silver-staining and immunoblot estimation of the CBPs in lysates from collagenase harvested secretory epithelia showed these proteins to be significantly reduced compared to nonsecretory epithelia. Close examination of the CBP immunoreactive cells of the mammary gland shows that ductal cells are prominent in their staining and that the immunoreactive material is associated with the cell surface. Also, in juvenile glands the myoepithelial stem cells (cap cells) of the elongating end bud are devoid of the CBPs. In contrast to the in vivo data, epithelia cultivated on collagen gels demonstrate comparable levels of the CBPs in both nonsecretory and secretory monolayers. The in vivo data indicate that the CBPs are developmentally regulated during mammary gland differentiation such that secretory epithelia are essentially devoid of these novel proteins. Furthermore, a role for calelectrin and calpactin I/p36 in exocytotic casein secretion is questioned.     

Development of a high-throughput screening system for the compounds that inhibit collagen–protein interactions

Collagen-binding proteins (CBPs) play important roles in various physiological events. Some CBPs are regarded as targets for drug development; for example, platelet glycoprotein VI (GPVI) and heat shock protein 47 (HSP47) are promising targets for the development of novel antiplatelet and antifibrotic drugs, respectively. However, no systematic screening method to search compounds that inhibit collagen–CBP interactions have been developed, and only a few CBP inhibitors have been reported to date. In this study, a facile turbidimetric multiwell plate assay was developed to evaluate inhibitors of CBPs. The assay is based on the finding that CBPs retard spontaneous collagen fibril formation in vitro and that fibril formation is restored in the presence of compounds that interfere with the collagen–CBP interactions. Using the same platform, the assay was performed in various combinations of fibril-forming collagen types and CBPs. This homogeneous assay is simple, convenient, and suitable as an automated high-throughput screening system.     

Influence of cytochalasin d-induced changes in cell shape on proteoglycan synthesis by cultured articular chondrocytes

There is growing evidence that cell shape regulates both proliferation and differentiated gene expression in a variety of cell types. We have explored the relationship between the morphology of articular chondrocytes in culture and the amount and type of proteoglycan they synthesize, using cytochalasin D to induce reversible cell rounding. When chondrocytes were prevented from spreading or when spread cells were induced to round up, 35SO4 incorporation into proteoglycan was stimulated. Incorporation into the cell layer was stimulated more than into the medium. When the cells were allowed to respread by removing cytochalasin D, proteoglycan synthesis returned to control levels. Cytochalasin D-induced stimulation of 35SO4 incorporation reflected an increase in core protein synthesis rather than lengthening of glycosaminoglycan chains, because [3H]serine incorporation into core protein was also stimulated. The observed stimulation of proteoglycan synthesis was not due to an overall stimulation of protein synthesis, to inhibition of DNA synthesis, or to accumulation of cells in one phase of the cell cycle. Cytochalasin D-treatment of cells in suspension caused no further stimulation of 35SO4 incorporation, suggesting that the observed effects were due to cell rounding rather than exposure to cytochalasin D per se; nevertheless, we cannot completely rule out other, nonspecific, effects of the drug. Fibroblasts and chondrocytes that had been passaged to stimulate dedifferentiation did not incorporate more 35SO4 when treated with cytochalasin D, suggesting that increased proteoglycan synthesis in response to rounding may itself be a differentiated property of chondrocytes.     

Dictyostelium calcium-binding protein 4a interacts with nucleomorphin, a BRCT-domain protein that regulates nuclear number

Nucleomorphin from Dictyostelium discoideum is a nuclear calmodulin-binding protein that is a member of the BRCT-domain containing cell cycle checkpoint proteins. Two differentially expressed isoforms, NumA and NumB, share an extensive acidic domain (DEED) that when deleted produces highly multinucleated cells. We performed a yeast two-hybrid screen of a Dictyostelium cDNA library using NumA as bait. Here we show that nucleomorphin interacts with calcium-binding protein 4a (CBP4a) in a Ca(2+)-dependent manner. Further deletion analysis suggests this interaction requires residues found within the DEED domain. NumA and CBP4a mRNAs are expressed at the same stages of development. CBP4a belongs to a large family of Dictyostelium CBPs, for which no cellular or developmental functions had previously been determined. Since the interaction of CBP4a with nucleomorphin requires the DEED domain, this suggests that CBP4a may respond to Ca(2+)-signalling through modulating factors that might function in concert to regulate nuclear number.     

Purification and characterization of chitin-binding proteins from the hemolymph of sweet potato hornworm, Agrius convolvuli

Three chitin-binding proteins (CBPs: CBP9, CBP15, CBP66) were identified from the larval hemolymph of sweet potato hornworm, Agrius convolvuli.Two (CBP9 and CBP15) of them have been isolated and purified by gel filtration (Superdex HR 75), cation-exchange chromatography (Mono S), and reverse-phase chromatography (μRPC PC 2.1/3). In experiments to detect CBPs in hemolymph, we examined whether ionic strength and existence of bovine serum albumin in the incubation solution influenced binding affinity of CBPs to chitin. The N-terminal sequences of three CBPs were determined by the automated Edman degradation and showed the sequence homology in basic local alignment search tool search. CBP15 and CBP66 were quite similar to lysozymes and bovine serum albumins, respectively. In contrast, CBP9 is not similar to any other known protein, as judged from databank comparisons. Therefore, we concluded that CBP9 is a novel protein with binding capacity to chitin that is a component of the fungal cell wall. CBP9 has no antibacterial activity against Escherichia coli and Micrococcus luteus, and also showed negative response in hemagglutination assay. CBP9 is confirmed as a monomer with a molecular mass of 9.14 kDa by electron spray ionization and matrix-assisted laser desorption ionization mass spectrometry.     

Presence of several cellulose-binding proteins in culture supernatant and cell lysate of Eubacterium cellulosolvens 5

Attempts were made to separate and characterize cellulose-binding proteins (CBPs) from both the culture supernatant and cell lysate of Eubacterium cellulosolvens 5. Once the CBPs were bound to Avicel cellulose, they were then effectively eluted with the solution containing 3.2 or 5% sodium dodecyl sulfate (SDS), but not eluted with the solution containing various kinds of carbohydrates and reagents. Namely, CBPs in both the culture supernatant and cell lysate of the bacterium bound tightly and strongly to cellulose. The SDS-polyacrylamide gel electrophoresis (SDS-PAGE) of the eluted CBPs indicated that the CBPs contained the two major proteins having the molecular weights of approximately 160 and 84 kilodaltons (kDa) and one sub-major protein having a molecular weight of approximately 140 kDa. Zymogram analysis after the SDS-PAGE of the eluted CBPs showed that two proteins exhibited the highest levels of carboxymethyl cellulase (CMCase) activity corresponding to the molecular weights of approximately 160 and 90 kDa. A major protein having the molecular weight of approximately 160 kDa exhibited a distinct CMCase activity and was designated as CBPE1. Western immunoblot analysis indicated that the proteins prepared from 16 representative strains of rumen bacteria did not cross-react with rabbit antiserum raised against CBPE1. Thus, CBPE1 may be a unique CBP that plays an important role in the adhesion of the bacterium to cellulose.     

Cytochalasin releases mRNA from the cytoskeletal framework and inhibits protein synthesis.

Cytochalasin D was shown to be a reversible inhibitor of protein synthesis in HeLa cells. The inhibition was detectable at drug levels typically used to perturb cell structure and increased in a dose-dependent manner. The drug also released mRNA from the cytoskeletal framework in direct proportion to the inhibition of protein synthesis. The released mRNA was unaltered in its translatability as measured in vitro but was no longer translated in the cytochalasin-treated HeLa cells. The residual protein synthesis occurred on polyribosomes that were reduced in amount but displayed a normal sedimentation distribution. The results support the hypothesis that mRNA binding to the cytoskeletal framework is necessary although not sufficient for translation. Analysis of the cytoskeletal framework, which binds the polyribosomes, revealed no alterations in composition or amount of protein as a result of treatment with cytochalasin D. Electron microscopy with embedment-free sections shows the framework in great detail. The micrographs revealed the profound reorganization effected by the drug but did not indicate substantial disaggregation of the cytoskeletal elements.     

Specific signaling cascades involved in cell spreading during healing of micro-wounded gastric epithelial monolayers

Mechanisms that specifically modulate cell spreading and/or cell migration following epithelial wounding are poorly understood. Using micro-wounded human gastric epithelial monolayers, we show herein that EGF and TGFalpha maximally increase spreading of epithelial sheets under a cell proliferation-independent mechanism. Treatment of confluent HGE-17 cells with the phosphatidylinositol 3-kinase inhibitor, LY294002, and the epidermal growth factor receptor inhibitor, PD153035, strongly reduced basal and TGFalpha-stimulated cell spreading. While pharmacological inhibition of pp60src-kinase activity also attenuated basal epithelial spreading, addition of the mTOR/p70S6K inhibitor rapamycin or a specific siRNA targeting ILK sequence did not affect the kinetic rates of wound closure. Epithelial wound healing was initiated by actin purse-string contraction followed by lamellae formation. Conversely, disruption of actin and tubulin stability with cytochalasin D and nocodazole, respectively, inhibited epithelial sheet spreading. Finally, antibodies directed against the alpha3 integrin subunit, but not against the alpha6 or alpha2 subunits, attenuated epithelial sheet spreading as well as lamellae formation. In conclusion, the current investigation establishes that EGF/TGFalpha and the alpha3beta1 integrin, pp60c-src, EGFR and PI3K pathways are mainly associated with the cell spreading of the restitution process during healing of human gastric epithelial wounds.     

Role of the cytoskeleton in laminin induced mammary gene expression

The differentiation of rat mammary epithelial cells is characterized both by morphologic changes and by the expression of a group of milk protein genes. We have previously shown that by culturing these cells on the basement membrane glycoprotein laminin, the synthesis of the milk proteins, transferrin, alpha-casein, and alpha-lactalbumin is induced. In order to determine if this effect is mediated through the cytoskeleton, we have treated these cells with cytochalasin D and colchicine. Treatment with cytochalasin D or colchicine for 24 h inhibits the accumulation of alpha-casein, transferrin, and alpha-lactalbumin without significant effect on general protein synthesis. Pulse chase studies show that cytochalasin D does not alter the intracellular turnover of alpha-casein or transferrin. Additionally, treatment with cytochalasin D causes an early (within 1 h) increase in secretion of alpha-casein and transferrin suggesting that the actin cytoskeleton provides a meshwork for secretory vesicles. The disruption of this network enhances the secretion of preformed proteins. However, long term (24 h) treatment with cytochalasin D inhibits synthesis of these milk proteins. Northern blot analysis indicates that treatment with cytochalasin D or colchicine inhibits the laminin induced increase in alpha-casein, alpha-lactalbumin, and transferrin mRNAs. These studies indicate that the major effect of the cytoskeleton on laminin induced milk protein gene expression occurs at the level of accumulation of mRNAs for these proteins. We conclude that the expression of laminin induced milk protein gene expression in primary rat mammary cultures depends on the integrity of the actin and microtubule cytoskeleton.     

Regulation of the catalytic activity of PTP1B: roles for cell adhesion, tyrosine residue 66, and proline residues 309 and 310

The reversible phosphorylation of proteins on tyrosine residues is fundamental to a variety of intracellular signaling pathways and is controlled by the actions of protein tyrosine kinases (PTKs) and protein tyrosine phosphatases (PTPs). While much progress has been made in understanding the regulation of PTKs, there is still relatively little known concerning the regulation of PTPs. Using immune complex phosphatase assays, we demonstrated that the enzymatic activity of the nonreceptor type PTP, PTP1B, is regulated by cell adhesion. Placing primary human foreskin fibroblasts (HFFs) in suspension leads to a distinct increase in PTP1B activity, whereas the readhesion of suspended HFFs onto fibronectin or collagen I inhibited activity. To gain insight into the mechanisms involved, we analyzed recombinant forms of PTP1B mutated at potential regulatory sites. Our results indicated that tyrosine residue 66 is essential for maintaining activity at 37 degrees C. We also found that the C-terminal region of PTP1B and localization to the endoplasmic reticulum are not required for the inhibition of activity by cell adhesion. However, analysis of PA-PTP1B, in which alanines are substituted for prolines 309 and 310, revealed an important role for these residues as the catalytic activity of this mutant did not decrease following readhesion onto collagen I. Since the binding of p130cas and Src to PTP1B is dependent upon these proline residues, we assayed the regulation of PTP1B in mouse embryo fibroblasts deficient in these proteins. We found that neither p130cas nor Src is required for the inhibition of PTP1B activity by adhesion to extracellular matrix proteins. Additionally, pretreatment with cytochalasin D did not prevent the reduction of PTP1B activity when cells adhered to collagen I, indicating that cell spreading is not required for this regulation. The control of the catalytic activity of PTP1B by cell adhesion demonstrated in this study is likely to have important implications for growth factor and insulin signaling.     

Identification of a 52-kD calmodulin-binding protein associated with the mitotic spindle apparatus in mammalian cells

A pool of 10 calmodulin-binding proteins (CBPs) was isolated from Chinese hamster ovary (CHO) cells via calmodulin (CaM)-Sepharose affinity chromatography. One of these ten isolated CBPs with a molecular mass of 52 kD was also found to be present in isolated CHO cell mitotic spindles. Affinity-purified antibodies generated against this pool of isolated CBPs recognize a single 52-kD protein in isolated CHO cell mitotic spindles by immunoblot analysis. Immunofluorescence examination of CHO, 3T3, NRK, PTK-2, and HeLa cells resulted in a distinct pattern of mitotic spindle fluorescence. The localization pattern of this 52-kD CBP directly parallels that of CaM in the spindle apparatus throughout the various stages of mitosis. Interestingly, there was no association of this 52-kD CBP with cytoplasmic microtubules. As is the case with CaM, the localization pattern of the 52-kD CBP in interphase cells is diffuse within the cytoplasm and is not associated with any discrete, cellular structures. This 52-kD CBP appears to represent the first mitotic spindle-specific calmodulin-binding protein identified and represents an initial step toward the ultimate determination of CaM function in the mitotic spindle apparatus.     

A tissue culture system for studies of the regulation of ether-linked and ester-linked aliphatic moieties in glycerolipids

A protein has been detected in chick brain which is immunologically identical to the vitamin D-induced calcium-binding protein found in intestinal tissue. It is present in highest concentration in the cerebellum and at lower levels in the remainder of the brain. Brain calcium-binding protein has the same molecular weight and electrophoretic properties as intestinal calcium-binding protein. Although the synthesis of intestinal calcium-binding protein is totally dependent upon a source of vitamin D, this has not yet been shown for brain calcium-binding protein. The total calcium-binding protein content of the cerebellum of chicks fed a vitamin D-free diet continued to increase during growth from 1 to 5 weeks, and is not responsive to exogenous vitamin D or 1,25-dihydroxyvitamin D.     

Chitin binding proteins act synergistically with chitinases in Serratia proteamaculans 568

Genome sequence of Serratia proteamaculans 568 revealed the presence of three family 33 chitin binding proteins (CBPs). The three Sp CBPs (Sp CBP21, Sp CBP28 and Sp CBP50) were heterologously expressed and purified. Sp CBP21 and Sp CBP50 showed binding preference to β-chitin, while Sp CBP28 did not bind to chitin and cellulose substrates. Both Sp CBP21 and Sp CBP50 were synergistic with four chitinases from S. proteamaculans 568 (Sp ChiA, Sp ChiB, Sp ChiC and Sp ChiD) in degradation of α- and β-chitin, especially in the presence of external electron donor (reduced glutathione). Sp ChiD benefited most from Sp CBP21 or Sp CBP50 on α-chitin, while Sp ChiB and Sp ChiD had major advantage with these Sp CBPs on β-chitin. Dose responsive studies indicated that both the Sp CBPs exhibit synergism ≥ 0.2 μM. The addition of both Sp CBP21 and Sp CBP50 in different ratios to a synergistic mixture did not significantly increase the activity. Highly conserved polar residues, important in binding and activity of CBP21 from S. marcescens (Sm CBP21), were present in Sp CBP21 and Sp CBP50, while Sp CBP28 had only one such polar residue. The inability of Sp CBP28 to bind to the test substrates could be attributed to the absence of important polar residues.     

Differential properties of attachment of human fibroblasts to various extracellular matrix proteins

Human diploid fibroblasts (TIG-3) were shown to attach and spread onto substrata coated with collagen, fibronectin, laminin and vitronectin. The cell attachment to these proteins required divalent cations. Mg2+ stimulated the cell attachment to all the proteins, while Ca2+ alone was not effective for the attachment to collagen and laminin. A mild trypsin treatment had prevented cells from attaching to the laminin, while it had no effect on the attachment to the other proteins. The fibronectin fragment, which retained cell binding activity, inhibited the cells from attaching and spreading onto fibronectin, but it did not cause any inhibition on the other proteins. The synthetic peptide GRGDSP inhibited the cells from attaching and spreading onto fibronectin and vitronectin, while it did not cause any inhibition on collagen and laminin. In attempts to isolate distinct receptors for these proteins, we were able to purify proteins very similar to the fibronectin and vitronectin receptors of human placenta. Based on the differential properties of the attachment of TIG-3 cells to these proteins and biochemical data, we indicate that human diploid fibroblasts have distinctive binding sites (receptors) for collagen, fibronectin, laminin and vitronectin.     

Endogenous lectins from cultured cells. Isolation and characterization of carbohydrate-binding proteins from 3T3 fibroblasts

Extracts of cultured 3T3 fibroblasts, obtained by homogenization and Triton X-100 solubilization, were fractionated on Sepharose columns covalently derivatized with asialofetuin. Three distinct carbohydrate-binding proteins (CBPs) were purified from the material bound to the affinity column: CBP35 (Mr = 35,000), CBP16 (Mr = 16,000), and CBP13.5 (Mr = 13,500). These CBPs were similar in several key properties. (a) They showed agglutination activity when assayed with rabbit erythrocytes; (b) they all appear to specifically recognize galactose-containing glycoconjugates; (c) they have low isoelectric points, pI 4.5-4.7; (d) their binding activities are rapidly lost in the absence of beta-mercaptoethanol; (e) the CBPs do not interact with each other, and the fractionated proteins can bind to asialofetuin independent of associated polypeptides; and (f) none of the proteins tend to self-associate to form oligomers of identical subunits. Comparisons of these and other properties of the CBPs suggest that CBP16 and CBP13.5 may be the murine counterparts of lactose-specific lectins previously identified in electric eel and in several bovine and avian tissues. In contrast, it appears that CBP35 represents a newly identified protein capable of binding to galactose-containing carbohydrates.     

p130cas but not paxillin is essential for Caco-2 intestinal epithelial cell spreading and migration on collagen IV

We have previously observed that collagen IV regulates Caco-2 intestinal epithelial cell spreading and migration via Src kinase and stimulates Src-dependent tyrosine phosphorylation of p130cas. We observed that collagen IV also stimulated Src-dependent phosphorylation of both paxillin Tyr31 and paxillin Tyr118. Caco-2 transfection with paxillin or p130cas siRNAs inhibited expression of these proteins by more than 90% for at least 5 days after transfection. Although p130cas siRNA inhibited cell spreading on collagen IV by 33%, three different paxillin siRNAs did not inhibit cell spreading. p130cas siRNA did not affect Src Tyr416 or Src Tyr527 phosphorylation, FAK Tyr397 phosphorylation, or Src-dependent phosphorylation of FAK Tyr925, suggesting that p130cas did not inhibit cell spreading by altering FAK or Src activity. Rat p130cas expression after siRNA knock-out of endogenous human p130cas in Caco-2 cells reduced cell spreading inhibition by 71%. In contrast, expression of rat p130cas from which the Src-phosphorylated substrate domain was deleted did not rescue siRNA inhibition of cell spreading. Combined treatment with siRNAs to Crk and CrkL, which bind to the p130cas substrate domain, inhibited cell spreading by 54%. Both p130cas siRNA and the combined Crk/CrkL siRNAs strongly inhibited (52 and 46% inhibition, respectively) Caco-2 sheet migration on collagen IV and noticeably inhibited lamellipodial extension, whereas paxillin siRNA only inhibited migration by 18% and did not noticeably affect lamellipodial extension. These results suggest that Src may regulate Caco-2 migration on collagen IV via both p130cas and paxillin but that Src phosphorylation of p130cas is more important for this process.