Scanning electron microscopy of focal contacts on the substratum

attachment surface of fibroblasts adherent to fibronectin

IRWIN I. SINGER, DIANA M. KAZAZIS and SOLOMON SCOTT

Department of Biochemical and Molecular Pathology, Merck Co., Inc., Merck, Sharp &f Dohme Research Labs, Rahway, Neto Jersey 07065, USA

Summary

We have examined the cell-to-substratum attach-ment surface of hamster fibroblasts with scanningEM, and describe the surface ultrastructure of focalcontacts and microspikes during cellular attach-ment and spreading on fibronectin. Nil 8 fibroblastswere seeded onto fibronectin-coated glass cover-slips in serum-free medium, fixed, and the fibro-blast-fibronectin monolayer was separated fromthe glass and inverted for scanning electron micro-scopic (EM) analysis. Focal contact developmentwas detected by interference reflection microscopyand correlated with the immunofluorescence micro-scopic distribution of fibronectin receptor antigens.The cell undersurface appeared smooth and fea-tureless at 0-5 h when focal contacts were undetect-able and fibronectin receptors were distributeddiffusely. By 1-2 h, undersurface membrane im-pressions of focal contacts were detected with scan-

ning EM; their size, shape and distributionmatched that of focal contacts seen with inter-ference reflection microscopy (ERM). These con-tacts had smooth external surfaces and were oftenarranged in chevron-shaped complexes. However,at 4-6 h, the surface texture of focal contactsbecame fibrous and the contact periphery wasdelineated with the orifices of membrane-associ-ated vesicles. Development of this filamentous sub-structure is correlated with the maximum concen-tration of fibronectin receptors and fibronectin atfocal contacts, suggesting that these molecules areinvolved in the maturation and stabilization offocal contacts.

Key words: substratum adhesion surface, focal contacts,scanning EM, fibronectin receptors.

Introduction

Focal contacts are tenacious adhesion sites that bind thelower fibroblast surface to the substratum in vitro (Izzard& Lochner, 1976, 1980). These contacts are situated atthe membrane insertion sites of actin microfilamentbundles (Abercrombie & Dunn, 1975; Heath & Dunn,1978; Wehland et al. 1979), and consequently containseveral actin-binding proteins at their cytoplasmic face:alpha-actinin (Lazarides & Burridge, 1975; Wehland etal. 1979), vinculin (Geiger, 1979; Burridge & Feramisco,1980), and talin (Burridge & Connell, 1983). Sincefibroblasts adhere to components of the extracellularmatrix, it is not surprising that fibronectin (Singer &Paradiso, 1981; Singer, 1982; Singer et al. 1987a) andheparan sulfate proteoglycan (Woods et al. 1984; Singeret al. 19876) have been located at the extracellular faces offocal contacts in stationary cells. Accordingly, transmem-brane receptors for extracellular matrix proteins such asfibronectin (Damsky et al. 1985; Giancotti et al. 1986;Grinnell, 1986; Singer et al. 1988) and vitronectin(Dejana et al. 1988; Singer et al. 1988) are also located at

Journal of Cell Science 93, 147-154 (1989)Printed in Great Britain © The Company of Biologists Limited 1989

focal contacts. Since the fibronectin receptor binds bothfibronectin and talin at different binding sites (Horwitz etal. 1986), it probably plays a critical role in linking theextracellular matrix to the cytoskeleton at focal contacts.Ultrastructural evidence also demonstrated a close one-on-one transmembranous connection between fibronec-tin fibers and actin microfilaments at the fibronexus(Singer, 1979a), a structure located at specialized focalcontacts (Singer & Paradiso, 1981; Singer, 1982, 1989),also termed extracellular matrix contacts (Chen et al.1985).

Focal contacts have been observed with a variety ofspecialized microscopic methods. Under interferencereflection microscopy (IRM), these contacts appear blackand therefore are the site of closest approach to thesubstratum (Izzard & Lochner, 1976, 1980). More re-cently, biophysical techniques such as antibody exclusionmicroscopy (Grinnell, 1980) and total internal reflectionfluorescence microscopy (Axelrod, 1981; Gingell et al.1985) have been used to detect focal contacts with greateraccuracy and contrast than under IRM. Transmissionelectron microscopy (EM) has demonstrated a dense

147

membrane-associated plaque associated with focal con-tacts, and immunofluorescence microscopy (IFM) hascontributed much biochemical information about theircomposition (see above; Burridge, 1986). However, apartfrom early transmission EM observations on replicas ofthe fibroblast's underside (Revel & Wolken, 1973), littledetailed evidence is available about the structure of thefocal contact at its outer membranous surface. In thispaper we report the high-resolution scanning EM visual-ization of the lower cell surface and development of theouter surface structure of focal contacts in fibroblastscultured on fibronectin substrata.

Materials and methods

Cell cultureNil 8 cells, a normal hamster fibroblast line, were cultivated inDulbecco's modified Eagle's minimum essential medium(DME) supplemented with 5% fetal bovine serum, 2mM-glutamine, and 50j<gml~' gentamycin as described (Singer etal. 1987a). For microscopic analysis, cultures were dissociatedwith trypsin in Hanks' balanced salts solution containing Ca2+

and Mg^+ and resuspended in serum-free DME containingl m g m l " heat-denatured bovine serum albumin. Cells wereseeded into Costar multiplate wells (1X105 cells/well, no. 3524,Costar, Cambridge, MA) containing 12 mm diameter glasscoverslips coated with a 5j(gml~ solution of purified humanplasma fibronectin (FN) in 0-1 M-NaHCOj. The cultures wereplaced in a humidified CO2 incubator and preserved aftervarious intervals at 37 °C as described below.

Immunofluorescence and interference reflectionmicroscopyCells were fixed with 3'5 % parafonnaldehyde in 0 1 M-sucrose,0-1 M-cacodylate buffer (pH7-2), and 4SmM-CaCl2 for 30minat 23 °C. Following gentle washing in buffer, the monolayerswere permeabilized with 0-1% Triton X-100, reduced withSmgnil"1 NaBH4inO-lM-Tris-HCl, pH7-2, and treated witha solution of non-fat dry milk to minimize background stainingas described (Singer et al. 1987a). Fibronectin receptors at thecell attachment surface were localized with affinity-purifiedrabbit antibodies that specifically recognize its alpha subunit asdemonstrated by immunoblotting experiments (Singer et al.1988). Cells were prepared for consecutive IFM and IRM asdescribed (Singer et al. 1987a). Briefly, cultures were incubatedwith a concentrated solution of the fibronectin receptor anti-body, washed in 0 1 M-Tris-HCl (pH72) , stained with fluor-escein isothiocyanate-conjugated affinity-purified goat anti-rab-bit IgG (Boehringer-Mannheim, Indianapolis, IN), andmounted on slides in a solution of 1 0 % //-propyl gallate, 1-0%dimethyl sulfoxide, and 0-1% NaN3 in 0-1 M-Tris-HCl(pH7-8). Cells were analyzed and photographed with a ZeissPhotomicroscope III (Carl Zeiss, Thornwood, NY) equippedfor IFM and IRM using a 63X antiflex neofluor objective lens(numerical aperture, 1-25; Singer et al. 1987a).

Scanning electron micmscopy of the loiver cell surfaceCultures were also established simultaneously on glass cover-slips for observation of the ventral substratum attachmentsurface with intermediate voltage scanning EM. They werepreserved with the fixative described above containing 1-0%glutaraldehyde in addition. Sectors of the cell monolayer werescored with a diamond pencil and released by careful floatingonto a 10% aqueous solution of HF. After extensive washing

with double-distilled water, the specimens were picked up fromabove using a second glow-cleaned coverglass, so that the lowercell surface was exposed for scanning EM study. These sampleswere then dehydrated with ethanol, infiltrated with amylacetate, critical point dried from CO2l and sputter-coated with athin layer of gold. Scanning EM was performed with a JEOL200 CX TEMSCAN microscope equipped with an ASID-3Dattachment operating at 200 kV.

Results

Development of membrane-associated complexes at thecell undersurfaceUsing the novel methodology described here, the surfaceultrastructure of the lower cell membrane could beanalyzed by scanning EM during cell attachment andspreading on fibronectin. In this way, the outer surfacemorphology of focal contacts was described during thecourse of their development. Although the upper mem-brane of Nil 8 fibroblasts plated on a fibronectin-coatedsurface appeared very active in terms of the presence ofnumerous microvilli above the nucleus and a prominentperipheral ruffling membrane (Fig. 1A), the lower sur-face of the cell was relatively smooth and quiescent at0-5 h (Fig. 1B,C). In this type of scanning EM prep-aration, the lower surface of the cell is covered by a thinfilm of material, which presumably is adsorbed fibronec-tin that separated from the coverglass due to HF treat-ment. Clefts were sometimes seen in this film at themargins of attached cells (Fig. 1C, arrowheads). No focalcontacts appeared at this time under IRM (Fig. ID).However, by 1-2 h in culture, the cells became flattened(Fig. 2A) and exhibited prominent focal contacts some-times arranged in V-shaped complexes (Fig. 2B). Whenthe undersurface of these cells was viewed with 'thescanning EM 1 h after seeding, peripheral microspikeswith rod-like cores were seen projecting towards thenucleus (Fig. 2C), and these structures correspond to themicrospikes detected with IRM (Fig. 2B, arrows). Theirfilamentous cores appeared closely apposed to the cellsurface membrane beneath microspikes (Fig. 2C, upperarrowhead). In addition, chevron-shaped linear struc-tures were seen with the scanning EM on the lowersurface of the cell (Fig. 2D), which closely resembled theV-shaped complexes of focal contacts evident with IRM(Fig. 2B). By 2h in culture, the ventral cell surfacedeveloped numerous linear structures that appearedsimilar to focal contacts and associated microfilamentbundles in shape and distribution (Fig. 2E). Regions ofthe lower cell membrane apparently interacted withstructural elements of focal contacts, producing a smoothsurface appearance along the length of the contact site,while intervening sectors of the plasma membrane werestudded with the openings of membrane-associated ves-icles (Fig. 2F). There was also a striking structuralsimilarity between the substratum attachment mem-branes of microspikes and focal contacts (compareFig. 2C with F). In 6h cultures viewed with IRM,peripheral focal contacts tended to be laterally apposedand appeared related to persistent V-shaped complexes ofcontacts (Fig. 3A). Linear structures with less-intense

148 /. /. Singer et al.

Fig. 1. Scanning electron micrographs of Nil 8 fibroblasts fixed 0-5 h after seeding onto a fibronectin-coated substratum in theabsence of serum. A. The upper cell surface exhibits numerous microvilli and ruffles above the nucleus, and the spreadingperipheral lamella contains a ruffling membrane at its periphery (arrow); X3000. B. Lower cell surface viewed from a lateralposition is relatively smooth compared to the dorsal surface; arrow depicts edge of the spreading lamella; X4000. C. Cellularundersurface viewed from above appears featureless, except for the pennuclear region, which seems to bulge outward (arrows);arrowheads indicate a cleft between the film of fibronectin stripped from the coverslip and the cell edge; X3000. D. IRMphotomicrograph of a similar Nil 8 fibroblast. The attachment surface shows irregular close contacts but no focal contacts;X1450

interference extended from the centripetal termini offocal contacts (Fig. 3A), and seemed to correspond tomicrofilament bundles. Scanning EM of the lower cellsurface revealed a set of membranous structures corre-sponding to the contacts detected with IRM at this time(Fig. 3B-D). A close morphological correspondenceexisted among chevron-shaped complexes of contacts,laterally apposed contacts, and individual focal contactsassociated with microfilament bundles seen with bothIRM and scanning EM (Fig. 3A-C). Unlike the smooth-surfaced contacts seen on the ventral plasmalemma at 2 h,the contacts found on 6h cells exhibited a fibroussubstructure (Fig. 3D). Also, orifices of membrane-associated vesicles delineated the margins of these con-tacts (Fig. 3D, arrowheads). The quantitative develop-ment of membrane-associated contacts detected withscanning EM on the substratum attachment surfaces ofNil 8 fibroblasts is shown in Table 1.

Time-course of fibronectin receptor accumulation infocal contactsWe stained inverted Nil 8 cell preparations with fibronec-tin receptor antibodies and 5 nm colloidal gold-conju-gated secondary antibodies to study scanning EM pat-terns of fibronectin receptor (FNR) accumulation at focalcontacts during their development on the lower cellsurface. While abundant receptor staining was detected atthe upper cell surface (Singer et at. 1988), our immuno-probes did not gain access to the lower cell membrane ofthe inverted preparations, even following extensive diges-tion with trypsin or pepsin. The accumulation of FNRantigens at developing focal contacts was therefore stud-ied with correlated IFM and IRM (Fig. 4). At 0-5 h,there was no focal concentration of FNR at the lower cellmembrane (Fig. 4A), while FNR did form conspicuouspatches at focal contacts and on microspikes 1—2 h later(Fig. 4B,C). Maximum concentration of FNR at focalcontacts occurred at 4—6h; these FNR aggregates ap-

SEM of focal contacts on the loner cell suiface 149

Fig. 2. Attachment surfaces of Nil 8 fibroblasts cultured on fibronectin-coated coverslips in serum-free media for l -2h .A. Scanning electron micrograph of the upper cell membrane of a cell fixed 2h following seeding. This surface, including thesupranuclear area («), is considerably flattened, although microvilli (arrowheads) are detectable; X3000. B. IRMphotomicrograph a Nil 8 cell similar to that shown in A. Numerous elongated focal contacts (arrowheads) are visible, and theyappear to be arranged into V-shaped complexes (large arrow). Microspikes at the edges of lamellipodia (small arrows) containrod-like cores that extend centripetally; X1780. C-F. SEM views of the lower cell surface. C. At 1 h, peripheral microspikesexhibit rod-like cores (arrowheads) that project towards the cell center, and are closely apposed to the ventral surfacemembrane; X3000. D. Several V-shaped structures (arrows), which correspond to the chevron-like assemblies of focal contactsobserved under IRM, are also seen at 1 h; X3000. E. Many structures resembling focal contacts (arrows) and associatedmicrofilament bundles (arrowheads) appear on the lower cell surface at 2h; X3000. F. Higher magnification of a portion of E.Regions of interaction between the adhesive membrane and focal contacts (long arrows) or microfilament bundles (short arrows)appear smooth, while the intervening membrane surfaces display numerous openings of membrane-associated vesicles(arrowheads). X10000.

peared linear and became laterally apposed or organizedinto V-shaped complexes (Fig. 4D,E). FNR distributionconsistently coincided with focal contacts identified byIRM (Fig. 4F,G).

150 I. I. Singer et al.

Discussion

We have developed a novel scanning EM method to studythe external surfaces of focal contacts developing on the

Fig. 3. Focal contacts on the substratum attachment surfaces of Nil 8 fibroblasts cultured on fibronectin-coated coverslips for6h. A. With IRM, peripheral focal contacts appear to have become apposed laterally (pairs of arrowheads), and V-shapedcomplexes (long arrows) persist. Linear structures with less-intense interference (short arrows) are seen centripetally, and aresometimes continuous with focal contacts (upper short arrow); X1450. B. SEM of the lower surface membrane at the edge ofthe cell. Several linear structures (arrowheads) correspond to the focal contacts observed in A in size, shape and distribution;these contacts are also apposed laterally (pair of arrowheads). The centripetal ends of some contacts are sometimes continuouswith thin linear filamentous structures (arrows) that may be related to the lighter fibers depicted by the short arrows in A;X3600. C. SEM of the cell undersurface at the leading lamella. Several V-shaped assemblies (V) composed of broad focalcontact-like zones (arrowheads) continuous with thinner microfilament-like structures (arrows) are evident. Laterally apposedfocal contacts (pair of arrowheads) are also present. Some microspikes are aligned with radial bundles that interact with thelower membrane (crossed arrow); X3000. D. Higher magnification of the lower chevron-like assembly depicted in C. The focalcontact {f) is composed of filamentous elements (long arrow) continuous with sub-fibers of the microfilament bundle (shortarrow). The margins of this assembly are delineated by openings of membrane-associated vesicles (arrowheads); X15 000.

Table 1. Scanning EM appearance of focal contact-like structures (FCs) on the undersurface of hamster

fibroblasts

Time (h) % Smooth cells % Cells with FCs

0-51-02-04-0

1004350

05795

100

Nil 8 fibroblasts were cultured on fibronectin-coated coverglasses,fixed, and their ventral surfaces were prepared for scanning EManalysis as described in Materials and methods. For each time point,the undersurfaces of 100 cells were scored as either lacking (smooth)or containing structures with the size, shape and distribution of focalcontacts.

lower (adherent) cell membrane during Nil 8 fibroblastattachment and spreading upon fibronectin. The struc-ture and distribution of focal contacts was determinedusing IRM of similar cells cultured in parallel. While theupper surface of the cell appeared active just after seeding(i.e. showing the presence of microvilli and ruffles), theundersurface was rather smooth, resembling in scanningEM the 'foot' of a gastropod. This appearance correlatedwith an absence of focal contacts under IRM, and a lackof FNR patching. However, at later culture periods, thelower surface membrane appeared to interact with inter-nal structural elements of focal contacts, chevron-shapedcomplexes of focal contacts and microfilament bundles,and peripheral microspikes. In 1-2 h cells, the membranebeneath focal contacts appeared very smooth, and thissmoothness may be due to the close apposition of the cellmembrane to the substratum that occurs at these sites

SEM of focal contacts on the lower cell suiface 151

Fig. 4. Time-course of fibronectin receptor (FNR) accumulation at focal contacts of Nil 8 fibroblasts cultured on FN substrata.Cells in serum-free medium were seeded onto FN-coated coverslips, fixed, permeabilized, stained with antibodies specific forthe FNR alpha subunit, and analyzed by immunofluorescence microscopy (IFM) as described in the Materials and methods.A. Only muted diffuse FNR immunofiuorescence is detected at 0 5 h. B. (1 h), C (2h). Ovoid patches of FNR antigens haveaccumulated in structures resembling focal contacts at the edge of the cell (arrowheads), and at microspikes (arrows). D. (4h),E (6h). FNR-positive focal contacts appear more intensely labeled, circumscribed, and elongated than earlier. Contacts at thecellular periphery (arrowheads) are arranged in V-shaped aggregates (D, upper arrowhead) or are laterally opposed (D, lowerarrowhead); they are also found beneath the cell center (arrows). F. 4h IFM; G, corresponding IRM photomicrograph;matching arrowheads depict FNR antigens localized at peripheral focal contacts. X870.

(Abercrombie & Dunn, 1975; Izzard & Lochner, 1976).In contrast, the lower membrane surface between focalcontacts was studded with the orifices of membrane-associated vesicles (Fig. 2F). Linear arrays of thesevesicles often delineated the margins of focal contacts inscanning electron micrographs of the Nil 8 fibroblastundersurface (Fig. 3D); this observation was corrobor-ated by transmission EM of the Nil 8 substratum-binding

membrane (fig. 2 of Singer, 1989). Similar assemblagesof membrane-associated vesicles and microfilamentbundles were observed in freeze-fracture replicas of theventral surfaces of human fibroblasts (Singer, 19796).The plasmalemma beneath microfilament cores of micro-spikes displayed a smooth morphology strikingly similarto that seen at the focal contact surface, suggesting thatthis membrane is also closely apposed to the substratum

152 /. /. Singer et al.

along a portion of its actin core. We believe that thefunctional basis of the similarity is best explained by theobservation that F-actin-containing microspikes are theapparent precursors of focal contacts (DePasquale &Izzard, 1987; Izzard, 1988; Rinnerthaler et al. 1988).

By 6 h in culture, the surface membranes of the focalcontacts became fibrillar, thus supporting the observationof Bailey & Gingell (1988) that fine linear filamentousdetails may be resolved in IRM images of focal contacts.The appearance of these striae on the focal contactmembrane may be caused by the condensation of corticalactin filament bundles just inside the contact membrane(Heath & Dunn, 1978; Wehland et al. 1979). That a veryclose association exists between actin microfilaments andthe inner surface of the focal contact membrane is borneout by higher-resolution transmission EM studies ofultrathin sections (Abercrombie et al. 1971; Singer,1979a) and replicas of the internal plasma membranesurface (Svitinka et al. 1984). The focal contact mem-brane may become 'tented' over closely underlying actinbundles, thus producing a filamentous appearance on theouter surface of the contact. In addition, polymerizationof fibronectin-containing extracellular matrix fibers onthe lower cell membrane may contribute to the appear-ance of filaments at the focal contact surface. The lattersuggestion is supported by the existence of fibronexusesat focal contacts (Singer, 1979a, 1982), and the detectionof numerous fibronectin fibers beneath Nil 8 focalcontacts by 6h (Singer et al. 1987a,b) when filamentswere observed at focal contact surfaces in the presentstudy. While unproven, the alignment of numerousmembrane-associated vesicles at the borders of focalcontacts suggests that they may provide a conduit forfibronectin deposition into the contact.

The essential features of our method of viewing thecellular undersurface with scanning EM are as follows.Nil 8 fibroblasts were fixed after various intervals ofculture upon fibronectin-coated glass coverslips in serum-free medium. The cells including their fibronectin sub-stratum were then separated from the coverglass bytreatment with 10% HF, floated on water, picked upfrom above, and inverted for scanning EM study. Higherconcentrations of HF were avoided because, although afaster separation of the monolayer from the glass oc-curred, we wished to avoid possible damage to the cells.The resultant monolayer preparation contained cellswhose lower surface was covered by the thin fibronectinlayer. This coating was visible in the scanning EM usinglower accelerating potentials (20-30 kV), but it wasdifficult to resolve clearly the structures that we observedon the ventral cell surface at 200 kV under these con-ditions. Similarly, Revel & Wolken (1973) used trans-mission EM to study platinum-carbon replicas of theundersurfaces of fibroblasts cultured in serum-containingmedium. Their study provided the first definitive lo-cation of substratum contact sites at the tips of microfila-ment bundles, but the resolution of the method wasinsufficient to visualize filamentous subcomponents ofthe contacts and membrane-associated vesicles. Thus,access to an intermediate voltage scanning EM is prob-ably essential for the higher-resolution microscopy

reported here. A real benefit of the fibronectin coating isthat it attaches to the overlying cells in their originalorientation and thus provides a convenient mechanismfor inverting them; it also probably stabilizes delicatestructures at the cell undersurface. An unavoidabledisadvantage of this fibronectin layer is its impervious-ness to antibodies. Incubating unseparated monolayerswith immunoprobes, or enzyme treatment of separatedpreparations (trypsin: 0-l- l-0mgml~ for 30min at37°C; pepsin: 0-2% in 0-02 M-HC1 for 16 h at 37°C) priorto immunostaining failed to permit labeling of the lowercell surface with fibronectin receptor antibodies. Wetherefore performed IFM in correlation with IRM onfixed and permeabilized cultures to monitor accumu-lation of fibronectin receptors at focal contacts.

Although fibronectin receptors have previously beenlocalized within focal contacts of mammalian fibroblasts(Giancotti et al. 1986; Grinnell, 1986; Singer et al.1988), the detailed time course of fibronectin receptoraccumulation at focal contacts has not been reported. Toavoid labeling of other related integrin molecules (Hynes,1987), we utilized antibodies that specifically recognizethe alpha subunits of fibronectin receptors (Singer et al.1988) for these studies. Fibronectin receptors were diffu-sely distributed on Nil fibroblasts in early cultures (0-5 h)that lacked focal contacts, but accumulated at micro-spikes and focal contacts in various stages of developmentfrom 1-6 h of culture on fibronectin-coated substrata inserum-free medium. We believe that fibronectin recep-tors that are initially distributed diffusely in the plane ofthe plasmalemma (Duband et al. 1986; Singer et al.1988) migrate into the focal contacts where they accumu-late. These observations also suggest that fibronectinreceptors bind to the extracellular fibronectin substratumduring early focal contact formation, and that thisbinding anchors these receptors within the contact, asrecently demonstrated by fluorescence recovery afterphotobleaching experiments (Duband et al. 1988). Oncepatching occurs, the linkage of fibronectin receptorswithin focal contacts may be stabilized further by attach-ment to actin-binding proteins such as talin (Horwitz etal. 1986) and perhaps vinculin (Dejana et al. 1988) at thecytoplasmic interface, and by fibronectin and heparansulfate proteoglycan, which accumulate in developingextracellular matrix fibers at the external surface of thefocal contact (Singer et al. 19876). The recent obser-vation that treatment of living 3T3 fibroblasts with Arg-Gly-Asp-Ser peptides induces rapid dissociation of vincu-lin and alpha-actinin from focal contacts (Stickel &Wang, 1988) strongly supports the suggestions that: (1)integrins (i.e. fibronectin and vitronectin receptors;Singer et al. 1988) stabilize focal contacts via the aboveexternal and intracellular interactions; and (2) Arg-Gly-Asp-Ser peptide antagonists of integrins cause cellulardetachment (Ruoslahti & Pierschbacher, 1987) by revers-ing these interactions and weakening focal contacts.High-resolution immuno-scanning EM studies of theaccumulation of integrins and their ligands at focalcontacts on the lower cell surface should add new detailsto our knowledge of how focal contacts develop and/ordestabilize during cellular locomotion. Also, the question

SEM of focal contacts on the lower cell surface 153

of whether membrane-associated vesicles play a role intranslocation of integrins and their ligands to the cellsurface needs to be addressed. Progress in these areas willdepend on the future development of methods to renderthe fibronectin layer that coats inverted cell preparationspermeable to immunoprobes.

References

ABERCROMBIE, M. & DUNN, G. A. (1975). Adhesion of fibroblasts tosubstratum during contact inhibition observed by interferencereflection microscopy. Expl Cell Res. 92, 57-62.

ABERCROMBIE, M., HEAYSMAN, J. E. M. & PEGRUM, S. M. (1971).

The locomotion of fibroblasts in culture. Expl Cell Res. 92, 57-62.AXELROD, D. (1981). Cell-substrate contacts illuminated by total

internal reflection fluorescence. J . Cell Biol. 89, 141-145.BAILEY, J. & GINGELL, D. (1988). Contacts of chick fibroblasts on

glass: results and limitations of quantitative interferometry. J. CellSci. 90, 215-224.

BURRIDGE, K. (1986). Substrate adhesions in normal andtransformed fibroblasts: organization and regulation of cytoskeletal,membrane and extracellular matrix components at focal contacts.Cancer Rev. 4, 18-78.

BURRIDGE, K. & CONNELL, L. (1983). A new protein of adhesionplaques and ruffling membranes. J. Cell Biol. 97, 359-367.

BURRIDGE, K. & FERAMISCO, J. R. (1980). Microinjection andlocalization of a 130K protein in living fibroblasts: a relationship toactin and fibronectin. Cell 19, 587-595.

CHEN, W.-T., HASEGAWA, E., HASEGAWA, T., WEINSTOCK, C. &

YAMADA, K. M. (1985). Development of cell surface linkagecomplexes in cultured fibroblasts. jf. Cell Biol. 100, 1103-1114.

DAMSKY, C. H., KNUDSEN, K. A., BRADLEY, D., BUCK, C. A. &

HoRwrrz, A. F. (1985). Distribution of the cell substratumattachment (CSAT) antigen on myogenic and fibroblastic cells inculture. J. Cell Biol. 100, 1528-1539.

DEJANA, E., COLELLA, S., CONFORTI, G., ABBADINI, M., GABOLI,

M. & MARCHISIO, P. C. (1988). Fibronectin and vitronectinregulate the organization of their respective arg-gly-asp adhesionreceptors in cultured human endothelial cells. J. Cell Biol. 107,1215-1223.

DEPASQUAL, J. A. & IZZARD, C. S. (1987). Evidence for an actin-containing cytoplasmic precursor of the focal contact and thetiming of incorporation of vinculin at the focal contact. J. Cell Biol.105, 2803-2809.

DUBAND, J.-L., NUCKOLLS, G. H., ISHIHARA, A., HASEGAWA, T.,YAMADA, K. M., THIERY, J. P. & JACOBSON, K. (1988).Fibronectin receptor exhibits high lateral mobility in embryoniclocomoting cells but is immobile in focal contacts and fibrillarstreaks in stationary cells. J. Cell Biol. 107, 1385-13%.

DUBAND, J.-L., ROCHER, S., CHEN, W.-T., YAMADA, K. M. &THIERY, J. P. (1986). Cell adhesion and migration in the earlyvertebrate embryo: location and possible role of the putativefibronectin receptor complex. J. Cell Biol. 102, 160-178.

GEIGER, B. (1979). A 130K protein from chicken gizzard: itslocalization at the termini of microfilament bundles in culturedchicken cells. Cell 18, 193-205.

GlANCOTTI, F. G., COMOGUO, P. M. & TARONE, G. (1986). A135,000 molecular weight plasma membrane glycoprotein involvedin fibronectin-mediated cell adhesion: immunofluorescencelocalization in normal and RSV-transformed fibroblasts. Expl CellRes. 163, 47-62.

GINGELL, D., TODD, I. & BAILEY, J. (1985). Topography of cell-glass apposition revealed by total internal reflection fluorescence ofvolume markers. J. Cell Biol. 100, 1334-1338.

GRINNELL, F. (1980). Visualization of cell-substrate adhesionplaques by antibody exclusion. Cell Biol. Int. Rep. 4, 1031-1036.

GRINNELL, F. (1986). Focal adhesion sites and the removal ofsubstratum bound fibronectin. J . Cell Biol. 103, 2697-2706.

HEATH, J. P. & DUNN, G. A. (1978). Cell to substratum contacts ofchick fibroblasts and their relation to the microfilament system. Acorrelated interference—reflexion and high voltage electron-microscope study. 7. Cell Sci. 29, 197-212.

HORWTTZ, A., DUGGAN, K., BUCK, C , BECKERLE, M. C. &BURRIDGE, K. (1986). Interaction of a plasma membranefibronectin receptor with talin: a transmembrane linkage. Nature,Land. 320, 531-533.

HYNES, R. O. (1987). Integrins: a family of cell surface receptors.Cell 48, 549-554.

IZZARD, C. S. (1988). A precursor of the focal contact in culturedfibroblasts. Cell Motil. 10, 137-142.

IZZARD, C. S. & LOCHNER, L. R. (1976). Cell-to-substrate contactsin living fibroblasts: an interference reflexion study with anevaluation of the technique. J. Cell Sci. 21, 129-159.

IZZARD, C. S. & LOCHNER, L. R. (1980). Formation of cell-to-substrate contacts during fibroblast motility: an interference-reflection study. 7. Cell Sci. 42, 81-116.

LAZARIDES, E. & BURRIDGE, K. (1975). Alpha-actinin:immunofluorescent localization of a muscle structural protein innonmuscle cells. Cell 6, 289-298.

REVEL, J. P. & WOLKEN, K. (1973). Electronmicroscopeinvestigations of the underside of cells in culture. Expl Cell Res.78, 1-14.

RINNERTHALER, G., GEIGER, B. & SMALL, J. V. (1988). Contact

formation during fibroblast locomotion: involvement of membraneruffles and microtubules. J. Cell Biol. 106, 747-760.

RUOSLAHTI, E. & PIERSCHBACHER, M. D. (1987). New perspectivesin cell adhesion: RGD and integrins. Science 238, 491-497.

SINGER, I. I. (1979a). The fibronexus: a transmembrane associationof fibronectin-containing fibers and bundles of 5 nm microfilamentsin hamster and human fibroblasts. Cell 16, 675-685.

SINGER, I . I . (19796). Microfilament bundles and the control ofpinocytotic vesicle distribution at the surfaces of normal andtransformed fibroblasts. Expl Cell Res. 112, 251-264.

SINGER, I. I. (1982). Association of fibronectin and vinculin withfocal contacts and stress fibers in stationary hamster fibroblasts. J.Cell Biol. 92, 398-408.

SINGER, I . I . (1989). Fibronectin-cytoskeleton relationships. InFibronectin (ed. D. F. Mosher), pp. 139-161. San Diego:Academic Press, Inc.

SINGER, I. I., KAWKA, D. W., SCOTT, S., MUMFORD, R. A. & LARK,

M. W. (1987a). The fibronectin cell attachment sequence arg-gly-asp-ser promotes focal contact formation during early fibroblastattachment and spreading. J. Cell Biol. 104, 573-584.

SINGER, I. I. & PARADISO, P. R. (1981). A transmembranerelationship between fibronectin and vinculin (130kD protein):serum modulation in normal and transformed hamster fibroblasts.Cell 24, 481-492.

SINGER, I. I., SCOTT, S., KAWKA, D. W. & HASSELL, J. R. (19876).

Extracellular matrix fibers containing fibronectin and basementmembrane heparan sulfate proteoglycan coalign with focal contactsand microfilament bundles in stationary fibroblasts. Expl Cell Res.173, 558-571.

SINGER, I. I., SCOTT, S., KAWKA, D. W., KAZAZIS, D. M., GAIUT,J. & RUOSLAHTI, E. (1988). Cell surface distribution of fibronectinand vitronectin receptors depends on substrate composition andextracellular matrix accumulation. J. Cell Biol. 106, 2171-2182.

STICKEL, S. K. & WANG, Y.-L. (1988). Synthetic peptide GRGDSinduces dissociation of alpha-actinin and vinculin from the sites offocal contacts. 7. Cell Biol. 107, 1231-1239.

SVTTKINA, T. M., SHEVELEV, A. A., BERSHADSKY, A. D. &

GELFAND, V. I. (1984). Cytoskeleton of mouse embryo fibroblasts.Electron microscopy of platinum replicas. Eur. J. Cell Biol. 34,64-74.

WEHLAND, J., OSBORN, M. & WEBER, K. (1979). Cell-to-substratum

contacts in living cells: a direct correlation between interference-reflexion and indirect-immunofluorescence microscopy usingantibodies against actin and alpha-actinin. 7 Cell Sci. 37, 257-273.

WOODS, A., HOOK, M., KJELLLEN, L., SMITH, C. G. & REES, D. A.

(1984). Relationship of heparan sulfate proteoglycans to thecytoskeleton and extracellular matrix of cultured fibroblasts.

{Received 11 November 1988 — Accepted, in revised form, 27 January1989)

154 /. /. Singer et al.