Mapping cell-glass contacts of Dictyostelium amoebae by total internal

reflection aqueous fluorescence overcomes a basic ambiguity of

interference reflection microscopy

I. TODD, J. S. MELLOR and D. GINGELL

Department of Anatomy and Developmental Biology, University College and Middlesex School of Medicine, Cleveland St,Umdon \YIIJ6DB, UK

Summary

The widespread ability of eukaryotic cells toproduce thin cytoplasmic sheets or lamellae 100-200 nm thick can give rise to uncertainties in theinterpretation of interference reflection mi-croscopy (IRM) images when cell-substratumtopography is the key interest. If allowed tospread upon a poly-L-lysine-coated surface,Dictyostelium discoideum amoebae typicallyform ultrathin lamellae of — lOOnm thickness bycytoplasmic retraction. Whereas the cell body isgrey, the lamellae appear very dark under IRMoptics. These dark areas could be misinterpretedas stemming from a closer cell-substratum appo-sition beneath the lamellae than the cell body.This ambiguity can be avoided if the technique oftotal internal reflection aqueous fluorescence

(TIRAF) is used in conjunction with a high refrac-tive index glass (n = 1-83) as substratum. Contri-butions to the image generated by thin cytoplasmand also variable cytoplasmic refractive indexare thereby minimized due to the extremely shortrange of the 'illuminating' evanescent wave.From our comparative IRM and TIRAF study ofthe ultrathin lamellae of Dictyostelium amoebaeit is concluded that the cell-glass gap is relativelyuniform beneath the entire cell. We briefly discussthe sensitivity of several cell types to TIRAF, thegeneration of ultrathin lamellae and the nature ofthe cell-glass gap.

Key words: IRM, TIRAF, TIRF, fluorescence, ultrathinlamellae, cell-substratum contact, Diclvosleliuni.

Introduction

It is well known that in certain circumstances cells canform very thin cytoplasmic sheets. The formation oflamellipodia, some 100-200 nm thick, at the advancingedge of locomoting fibroblasts in culture was firstrecognized by Abercrombie et ctl. (1971) using trans-mission electron microscopy (TEM), and further stud-ied with interference reflection microscopy (IRM) byIzzard & Lochner (1980). Phagocytic engulfment canalso involve enveloping sheets of cytoplasm 200 nm orless in thickness (Zenian*?/ cd. 1979; McNeil, 1981). Inthese situations the lamellipodia apparently play a rolein extending the cell periphery and probably in ad-hesion as well.

Alternatively, after spreading upon a poly-L-lysine-coated coverslip Dictyostelium amoebae initially gener-ate extensive, peripheral, thin cytoplasmic sheets by aJournal of Cell Science 89, 107-114 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

process of centripetal thinning (Gingell & Vince,1982a). These ultrathin lamellae, which persist forsome 10—20 min before active locomotion starts, appearblack under IRM optics, in contrast to the mass of thecell body, which is mid-grey. Using quantitative IRMtheory (Gingell & Todd, 1979; Gingell et al. 1981),Gingell & Vince (1982«) argued that the black seen inIRM is generated when the cytoplasm thins down toaround 100nm, an interpretation supported by theirscanning electron microscope (SEM) images. Mow-ever, it is not widely appreciated that using IRM aloneit is difficult, if not impossible, to differentiate betweenthe contributions to the final image arising from thethickness of such lamellae and their separation from thesubstratum.

In the present paper we have used a fluorescencetechnique to show how this difficulty can be unambigu-ously resolved. Suppose cells are spread in an aqueous

107

medium on a glass block. If a laser beam is directed intothe block such that it is incident on the glass-waterinterface at an angle exceeding the critical angle nobeam is transmitted into the water, but a localizeddisturbance called an evanescent leave is generated inthe water adjacent to the glass. This wave dies offexponentially within —30-100 nm of the glass, theexact value depending on the refractive index stepacross the interface as well as the angle of incidence andthe wavelength of the laser beam. Any fluorochromemolecules located within this zone, but not beyond it,will therefore fluoresce, giving TIRF (total internalreflection fluorescence) images. Axelrod (1981) was thefirst to use this method in cell biology. He used afluorochrome that adsorbed to cell surfaces and ob-tained bright TIRF images of cell-glass contact zonesagainst a black background. Our 'extracellular volumemarker' variant of this (Gingell et al. 1985) utilizes anaqueous fluorochrome, which does not adsorb signifi-cantly to either the cell or to the glass and does notpenetrate into healthy cells. The cell contacts displacefluorochrome molecules and thus appear as darkpatches on a brightly fluorescent background. We haverecently shown how the cell-glass apposition distancecan be calculated from fluorescence measurementswhen extracellular volume marker TIRF is used(Gingell et al. 1987). For brevity and clarity we nowuse the acronym TIRAF (total internal reflectionaqueous fluorescence) for our technique.

Using TIRAF we have now shown that our earlierinterpretation of IRM images was indeed correct.Furthermore, we have been able to demonstrate thatthe depth of the aqueous gap separating the lipidbilayer from the glass is constant beneath the lamellaeand is the same beneath the relatively thick cell body.This direct result, which has not been obtained onliving cells by any other technique, has implications forthe mechanics of cell adhesion.

Materials and methods

Total internal reflection fluorescence, video imaging andrecording equipment were used as described by Gingell et al.(1985). This is based on a Zeiss UEM microscope (Carl ZeissLtd, Welwyn Garden City, Herts, UK), a Lexel 85-1 argonion laser operating at 488 nm (Lamda Photometries Ltd,Harpendcn, Herts, UK) and an Isocon low-light TV camera(English Electric Valve Ltd, Chelmsford, UK). All refractiveindices quoted are for 486-1 nm or 488 nm.

Dictyostelium discoideum amoebae of the axenic strain Ax2were cultured and harvested as described by Gingell & Vince(1982rt). Instead of washing by repeated centrifugation, cellswere allowed to settle on glass slides and exposed to gentlyrunning 20 mM-NaCI solution. They were then washed off bymore vigorous pipetting and collected in tubes to be main-tained on ice.

Glass coverslips (Chance Propper, from E. Lamb & Co.Ltd, London, UK) were degreased and exposed to 4 % IIF in40% HNO3 for several seconds, before rinsing in copiousdistilled water. The latter was prepared in an all-glass still asdescribed by Owens & Gingell (1987). Coverslips were left toequilibrate overnight in distilled water before use. Cleancoverslips were treated with a 0-1 % poly-L-lysine hydrobro-mide (/Wr = 30000-70000; Sigma Chemical Co. Ltd, South-ampton, Hants, UK) solution for 1 min, then rinsed well indistilled water before use. Adequate polycation adsorptionwas shown by the strong adhesion of aldehyde-treated redblood cells (Todd & Gingell, 1980) in distilled water. Noadhesion occurs on clean untreated glass.

The poly-I.-lysine-treated coverslips were optically matedto the upper glass plate of the prism assembly (see fig. 1 ofGingell et al. 1985) with standard Zeiss immersion oil. Thee-fold characteristic decay depth (s) of the evanescent wavefor standard coverslip glass under the conditions used is60 nm. For observations at low penetration (s = 35 nm)polished glass plates approximately 1 mm thick were cut(Ealing Optical Works, London, UK) from a high refractiveindex glass (/; = 183) kindly supplied by Mr R. Dunning ofPilkington Brothers, Ormskirk, Lancashire, UK. The platesdisplayed very low optical scatter when mated to the topsurface of a prism, cut from the same glass, using Cargilleimmersion liquid Ml (n = 1-765; Cargille Laboratories Ltd,NJ, USA). The glass plates were cleaned in the same manneras the coverslips except that HF was omitted, since we foundthat it damaged the scatter-free finish. In this case HNO3 wasfollowed by a 30-45 min wash in 10% Decon 90 (DeconLaboratories Ltd, Hove, UK) with sonication. Poly-L-lysinecoating was done as for the coverslips.

A 63 X water-immersion objective (Carl Zeiss) was focusedonto the interface under IRM optics. The amoebae, whichrapidly settled and spread, were then examined under IRMand TIRAF in turn. Because of the extreme susceptibility ofthe lamellae to TIRAF conditions (see Discussion) weworked at very low light levels, necessitating high cameragain. The resulting noisy images were processed in anArlunya TV integrating framestore (Agar Scientific Ltd,Stanstead, UK). Integrated images were photographed di-rectly from the monitor screen, using a Nikon F3 camera andIlford FP4 film. All microscopic observations were made in20mM-NaCl containing 10~3-10~4M-fluorescein in the formof the sodium salt (Sigma) or fluoresceinated dextran (FD4,MT = 4100; Sigma) at 20°C. These extracellular markers giveessentially identical images with chick fibroblasts (Gingell etal. 1985).

For electron microscopy, grids were coated with a film ofParlodion (Agar Scientific) prior to 0 ' l % poly-L-lysinesolution. Cells were seeded directly onto the grids in a smallchamber. A similar preparation on a coverslip was used tomonitor the cell responses. After 17 min grids were trans-ferred to 1/4 strength Karnovsky's fixative and left for 1 h.They were then rinsed in 0 1 M-cacodylate buffer, pH 7-4,stained in uranyl acetate, rinsed in buffer and air-dried(ethanol-treated grids lost their films). Specimens wereexamined in a Philips 300 instrument.

108 /. Todd et al.

Results and discussion

Comparison of fluorescence and interference imagesDestructive interference creating blackness in IRMimages of cells arises between reflections from either:(1) the lower plasmalcmma and the glass surface wherethe two are in close apposition; or (2) the upper andlower cell surfaces where thin sheets of cytoplasm haveformed (Gingell, 1981; Gingell & Vince", 1982a). Theexistence of these two mechanisms can give rise toambiguities in the interpretation of IRM images.

The fact that Dictyostelium amoebae form extensivethin sheets of cytoplasm on a poly-L-lysine substratumis shown in Fig. 1. The upper TEM image shows a cellthat has developed a broad lamella; the lower showsanother in the process of withdrawing cytoplasm fromthe surrounding lamella. In these whole mounts thelow electron density of the peripheral areas clearlyindicates that they are far thinner than the cell body.However, the thickness of the lamellae and theircontact relationships with the substratum cannot beascertained from such preparations. We shall showfrom our optical study that the distinctive IRM imagesof the cell peripheries do indeed stem entirely fromultrathin lamellar cytoplasm and that the cell-glassdistance is the same beneath the cell body and thelamellae.

In Fig. 2A,B the lamella is being retracted, produc-ing an arborescent periphery, similar to that seen inFig. IB. Under IRM (Figs 2A, 3C) the lamellae lookmuch darker than the cell bodies due to destructiveinterference across the thin, almost parallel, sheet ofcytoplasm. Under TIRAF, on low refractive indexglass (Figs 2B, 3D) the lamellae appear brighter thanthe cell bodies because the evanescent wave partiallypenetrates the lamellae and stimulates a limited amountof fluorescence from the medium beyond it. A similarcontrast reversal between IRM and TIRAF is seen inFig. 3A,B, where two adjacent regions on the upperleft of the cell (arrowed) have also thinned intolamellae. Since under TIRAF these areas appearsubstantially brighter than the cell body, which iscertainly thicker than 1 fim, the paired images areconsistent with the existence of very thin eytoplasmiclamellae.

Beneath the cell body in Fig. 2B a small brightlyfluorescent area can be seen where the cell has liftedaway from the glass (vertical arrow), while the corre-sponding IRM image (Fig. 2A) shows pale concentricfringes. Similar lifting is apparent in Fig. 3A,B (out-lined arrows). Here the close proximity of tubules ofthe contractile vacuolar apparatus (white arrowheads inFig. 3A), which are seen as moving black threadsunder IRM (Gingell & Vince, 1982a), suggests that theimage may be associated with expulsion of water from a

contractile vacuole. This has been seen under IRM tolift the cell transiently at the point of discharge.

Within the lamella of the cell shown in Fig. 2 is alocal thickening of the cytoplasm (horizontal arrow),which is much paler under IRM (Fig. 2A) than thesurrounding thin lamella (Gingell, 1981). UnderTIRAF (Fig. 2B) this thickening is as dark as the cellbody, because it is too thick for the evanescent wave topenetrate. The lower cell of Fig. 2 has started loco-motion and is trailing a lamella that has broken up intocellular debris; this too shows a localized region(horizontal arrow) where the lamella has thickened. Ifthe dark zone (IRM) were due to destructive inter-ference across a very small water gap beneath a part ofthe cell that is greater than several tenths of a micro-metre thick, the paler zone would have to represent lift-off. In this case the TIRAF image would have to bebrighter at this point, whereas it is darker. Thuseytoplasmic thickening is responsible, and from this itis virtually certain that ultrathin lamellae cause the darkIRM images.

These observations, taken alone, do not excludeminor optical perturbations from two sources. Inprevious work using immersion refractometry it wasfound that the refractive index of the lamella ofDictyostelium is slightly higher than the average for thewhole cell (Gingell & Vince, 1982fl). According to thecalculations of Gingell et at. (1987) this could have asmall influence on the TIRAF image, tending tobrighten the lamella of a cell attached to coverslip glass(H = 1-528) by some 2%. Second, while there is noroom for doubt from the arguments given above that athin lamella is responsible for the images in Figs 2 and3 it is conceivable that the lamella-glass separationdiffers slightly from the cell body-glass separation aswell. This point has been settled by observing theTIRAF image on glass of high refractive index(n = l-83), where the peripheral lamellae were found tobe indistinguishable from the cell body (Fig. 4A,B).On high-index glass the wave decay characteristic canbe made so low (Fig. 5) that the eytoplasmic thicknessover all parts of the cell, including the lamella, con-siderably exceeds it (Gingell et al. 1987). We usedz = 35 nm, in which case the wave energy falls to 5 % ofthe interracial value at a distance of 105 nm. Anadditional advantage is that the increment in lamellarbrightness due to a slightly higher eytoplasmic refrac-tive index is reduced to 0-5%. Consequently, anydifferences in contrast seen under these conditionsmust be due to variations in the thickness of the zonebetween the plasmalemma and the glass. The fact thatTIRAF on high-index glass shows uniform apposition(Fig. 4A,B) where IRM shows a cell with an extensivedark periphery, proves that the dark zone seen in IRMis due solely to destructive interference across theultrathin lamella.

Contacts beneath ultrathin lamellae 109

1A

B

Fig. 1. TEM of Dictyostelium amoebae on Parlodion treated with poly-L-lysine. A. Cell showing a newly formed thinlamella of uniform electron density. B. Cell in which the lamella is being withdrawn, leaving behind remnants, again ofuniform electron density. Bar, 2-S^tm.

110 l.Toddetal.

Fig. 2. Dictyostelium amoebae spread on low refractive index glass {£ = 60 nm) treated with poly-L-lysine. (For explanationof arrows see text.) A. Cell as seen by IRM showing a dark cell periphery. B. Same cell as in A seen by TIRAF where theperiphery is now paler than the cell body. C. IRM view of a locomoting cell showing a dark area at the rear. D. Same cellas in C by TIRAF. Bar, 5jum.

Mode of adhesion

We have shown that IRM images of thin cytoplasmiclamellae can be a hidden trap for the unwary becausethey are difficult to distinguish from zones of very closecontact. TIRAF observations on high-index glassavoid this difficulty and determine only the size of thecell—glass gap. However, TIRAF on ordinary coverslipglass may give ambiguous images of ultrathin lamellaebecause of the unavoidably longer range of the evan-escent wave, even if a high angle of incidence is used.Providing the fluorescent marker molecules are able to

diffuse beneath the cell (large molecules can be steri-cally excluded; Gingell et al. 1985) a dark image withTIRAF must imply a 'small' plasmalemma—glass gap.In Dictyostelium it appears to be uniform in thickness,without visible specialized adhesion sites. The dimen-sions of this zone are, however, responsive to changesin the ionic strength of the medium in a reversible way(Gingell & Vince, 19826), thus implicating Debycelectrostatic repulsion between fixed charges. The factthat amoebae attached to glass in 20mM-NaCl cannotbe removed by a shearing force of 6 N m~2 (Owens,Gingell & Trommler, unpublished data) shows that

Contacts beneath ultrathin lamellae 111

3A B

Fig. 3. Dictyostelium amoebae on low refractive index glass. A,C. As seen by IRM; B,D, as seen by TIRAF. Bar, 5/.mi.

they are not attached by virtue of a force balance in acolloid secondary minimum (Parsegian & Gingell,1980), but that the molecules of the glycocalyx span thewater gap and their ends are adsorbed directly to theglass. The increase in this gap seen in distilled water isevidently caused by electrostatic swelling of the poly-anionic matrix of the glycocalyx. A similar conclusionhas been reached regarding the adhesion of red bloodcells to glass (Trommler et al. 1985). We are currentlyusing TIRAF to investigate the kinetics of diffusion offluorescently labelled markers through the glycocalyx,in a direction parallel to the lipid bilayer, and tomeasure the cell-glass separation.

TIRAF can damage cells

We are also investigating the sensitivity of cells toTIRAF. Under most circumstances, with chick limbbud or 3T3 fibroblasts, for example, illumination atincident power levels, which give acceptable images byeye, have no discernible effects upon the cells. How-ever, after prolonged exposure to high-incident laserpower (4-8luJlum~2s~l) a rapid loss of contrast isoften observed, indicating permeabilization of theplasmalemma and entry of fluorescein into the cyto-plasm. Since the use of lower laser power and/orpulsed observation obviates this effect we do notnormally regard it as a serious limitation. In contrast,

112 I.Toddetal.

200

Fig. 4. IRM (A) and TIRAF (B) images of the sameamoeba generating an ultrathin lamella on a poly-L-lysinesubstratum adsorbed onto high refractive index glass(i = 35 nm). The black periphery of the IRM imageindicates that the cytoplasm has thinned to some 100 nm inthese areas whilst the interference fringes (arrowed) implythat this part of the cell is between 0-1 and 1 fitn thick.Apart from two small brighter spots (circled), whichcorrespond in both images and suggest areas of greaterseparation, TIRAF demonstrates the uniformity of thecell—glass gap beneath the entire cell. Laser illuminationwas kept sufficiently brief so that cytoplasmic flow into thelamella was not stimulated. Bar, 5jum.

we have found Dictyostelium amoebae to be exceed-ingly sensitive to TIRAF. Cells can be observed undersimultaneous IRM and TIRAF in such a way that IRMmakes the major contribution to the combined image:within 30 s the dark peripheral lamellae increase to arelative irradiance characteristic of the cell body. WithTIRAF alone, under conditions where the penetrationdepth of the evanescent wave is relatively high(£ = 60nm), the lamellae first appear pale, as inFig. 2B, but then rapidly darken. These changes,which do not involve permeabilization of the cells tofluorescein, are caused by cytoplasmic flow from thecell bodies into the lamellae, so creating the localizedthickenings referred to in the analysis of Figs 2, 3.When cells were exposed to TIR without extracellularfluorochrome, no thickening took place, but within halfa minute of adding fluorescein the dark IRM image waslost. We believe this result implicates the products ofphotolysis of fluorescein (probably acting at the plas-malemma), but as neither 25 mM-thiourea (an OH freeradical scavenger) nor superoxide dismutase (whichscavenges superoxide) at 500^tgml~' appear to reducethe sensitivity of the amoebae, we do not yet know theiridentity. As a final point, our experience with human

150

auCo

100

50

040 50 60 70 80 90Angle of incidence (deg.)

Fig. 5. Curves showing thecharacteristic penetrationdepth (i , nm) of evanescentwaves as a function of theangle of incidence. Curve/-I isfor normal low refractiveindex coverslip glass (1-528);and curve B is for high indexglass (1-830). In both casesthe refractive index of wateris 1-339. All refractive indicesare for a laser wavelength of488 nm.

monocytes and TIRAF has indicated that they toogenerate ultrathin lamellae by retraction but theirsensitivity to TIRAF appears to be intermediate be-tween that of Dictyostelium and fibroblasts.

Ultrathin lamellae

The lamellar response of amoebae closely resemblesthat of polymorphonuclear leucocytes (see Boyles &Bainton, 1979; Keller et al. 1979, 1983) and mono-nuclear leucocytes (Mellor, Gingell & Todd, unpub-lished data). It is therefore a common cellular re-sponse, not just an obscure and irrelevant feature of aprimitive soil amoeba. The ultrastructural basis oflamella formation and withdrawal is little understood,but related studies on macrophages (Trotter, 1981; andfor a review, 1985; Hartwig & Shevlin, 1986) andpolymorphonuclear leucocytes (Malech et al. 1977;Boyles & Bainton, 1979) have shown that their lamellaehave an extensive network of actin filaments in associ-ation with the adherent membrane. Dictyosteliumamoebae form ultrathin lamellae on hydrophobic sur-faces of methylated glass and polystyrene (Gingell &Vince, 1982*?) as well as on hydrophilic surfaces,including aminated glass, poly-L-lysine and cleanuntreated glass at high ionic strength (Owens, Bailey &Gingell, unpublished data). These diverse surfaces allappear to cause strong adhesion, but we are not yet in aposition to test whether this is the common trigger.

References

ABERCROMBIE, M., HEAYSMAN, J. E. M. & PECRUM, S. M.

(1971). The locomotion of fibroblasts in culture.IV. Electron microscopy of the leading lamella. lixpl CellRes. 67, 359-367.

AXELROD, D. (1981). Cell-substrate contacts illuminatedby total internal reflection fluorescence..J. Cell Biol. 89,141-145.

BOYLES, J. & BAINTON, D. F. (1979). Changing patterns ofplasma membrane-associated filaments during the initial

Contacts beneath ultrathin lamellae 113

phases of polymorphonuclear leucocyte adherence.J. Cell Biol. 82, 347-368.

GlNGELL, D. (1981). The interpretation of interferencereflection images of spread cells: significant contributionsfrom thin cytoplasm. J . Cell Sci. 49, 237-248.

GINGELL, D., HEAVENS, O. S. & MELLOR, J. S. (1987).

General electromagnetic theory of total internal reflectionfluorescence; the quantitative basis for mappingcell-substratum topography. J. Cell Sci. 87, 677-693.

GlNGELL, D. & TODD, I. (1979). Interference reflectionmicroscopy. A quantitative theory for imageinterpretation and its application to cell-substratumseparation measurement. Biophys. J. 26, 507-526.

GINGELL, D., TODD, I. & BAILEY, J. (1985). Topography

of cell-glass apposition revealed by total internalreflection fluorescence of volume markers. J. Cell Biol.100, 1334-1338.

GINGELL, D., TODD, I. & HEAVENS, O. S. (1981).

Quantitative interference microscopy: effect of aperture.Opt ice, Ada 29, 901-908.

GINGELL, D. & VINCE, S. (1982a)- Substratum wettabilityand charge influence the spreading of Dictyosleliumamoebae and the formation of ultrathin cytoplasmiclamellae. J . Cell Sci. 54, 255-285.

GINGELL, D. & VINCE, S. (19826). Cell-glass separationdepends on salt concentration and valency:measurements on Dictyoslelium amoebae by finiteaperture interferometry. jf. Cell Sci. 54, 299-310.

HARTWIG, J. H. & SHEVLIN, P. (1986). The architecture ofactin filaments and the ultrastructural location of actin-binding protein in the periphery of lung macrophages.y. Cell Biol. 103, 1007-1020.

IZZARD, C. S. & LOCHNER, L. R. (1980). Formation ofcell-to-substrate contacts during fibroblast motility: aninterference-reflexion study. J. Cell Sci. 42, 81 — 116.

KELLER, II. U., BARANDUM, S., KISTLER, P. & PLOEM,

J. S. (1979). Locomotion and adhesion of neutrophilgranulocytes. Effects of albumin, fibrinogens and gamma

globulins studied by reflection microscopy. Expl CellRes. 122, 351-362.

KELLER, H. U., ZIMMERMAN, A. & COTTIER, H. (1983).

Crawling-like movements, adhesion to solid substrataand cheniokinesis of neutrophil granulocytes. y. Cell Sci.64, 89-106.

MALECH, H. L., ROOT, R. K. & GALLIN, J . I . (1977).

Structural analysis of human neutrophil migration.Centriole, microtubule and microfilament orientation andfunction during chemotaxis. y. Cell Biol. 75, 666-693.

MCNEIL, P. L. (1981). Mechanisms of nutrititiveendocytosis. I. Phagocytic versatility and cellularrecognition in Chlorohydra digestive cells, a scanning andelectron microscope study. J . Cell Sci. 49, 311-341.

OWENS, N. & GINGELL, D. (1987). Surface properties of along-chain 10:12 diynoic acid monolayer at air-liquidand solid-liquid interfaces. Thin Solid Films (in press).

PARSEGIAN, V. A. & GINGELL, D. (1980). Red blood celladhesion. III. Analysis of forces. J . Cell Sci. 41,151-157.

TODD, I. & GINGELL, D. (1980). Red blood cell adhesion.I. Determination of the ionic conditions for adhesion toan oil-water interface. J . Cell Sci. 41, 125-133.

TROMMLER, A., GINGELL, D. & WOLF, H. (1985). Red

blood cells experience electrostatic repulsion but makemolecular adhesions with glass. Biophys. J. 48, 835-841.

TROTTER, J. A. (1981). The organization of actin inspreading macrophages is linked to the substratum viatransmembrane connections. Expl Cell Res. 132,235-248.

TROTTER, J. A. (1985). In Research Monographs in Celland Tissue Physiology, vol. 11 (ed. R. T. Dean & W.Jessup), pp. 285-305. Amsterdam, New York, Oxford:Elsevier.

ZENIAN, A., ROWLES, P. & GINGELL, D. (1979). Scanningelectron-microscopic study of the uptake of Leishmaiiiaparasites by macrophages. y. Cell Sci. 39, 187-199.

(Received 21 August 1987 - Accepted J October 1987)

114 /. Todd et al.