Unconfined lateral diffusion and an estimate of pericellular matrix viscosity revealed by measuring the mobility of gold-tagged lipids.

Cell Culture

Primary cultures of keratucytes were prepared from freshly removed goldfishscales. The scaleswere rinsedthree times for 15mineach inHam's ~12 Nutrient Mixture with 25 mM Hepes (SigmaChemical Co., St. Louis, MO) plus 10% heat-inactivatedFBS (TCM+) containinapeuir tomycin and fu%al,one.The scales were placed singly, cellside down, on 22 mm2, No. 1, cleaned (as described in Lee r el., 1991)glass coverslips in 35-ram petri dishes in a moist chamber. Each scale was covered with a 12-mmroundcovendip.After 15rain, 1mi OfTCM+ was ~__~_~.The scale culture8were kept at toomtenkeeramreinthe moistchamber overnight. To dissociate sheetsofcells which hadmigrated offthe scele, tho scalecultures were treated with PBS containing1~)mM Mg2+(no Ca2+)for 15 min and then were returned to TCM+. Fmh scale fibmbinsts were obtained by pas- sage ofthe fish scele cultures a~'r llbroblasts had ovezgrovmthe other ceil types (in 1-2 wks). Fish scale fibroblam were used in their third to fifth paa~ge and 2 daRer plating. PtK1 ceils were ~ in TCM+ with g e a t ~ and C3H-10TI/2ceils were grown in Eagle's basal medium plus 25 mM Hepes (Sigma Chemical Co.) with 10% heat- inactivated FBS and ~ . PtKI and C3H-10TI/2 cells 1.Abbrm,im~msusedintMspaper, anti-Fl, anfi-fluorescein; D, diffusion coefficient; FI-PE, fluorescein phusphafidylOh~nalamine; MSD, mean square displacement. were maintained at 37~ in a 7.5% CO2 incubator and were also used the second day after plating on cleaned 22 mm2, No. 1 glass coverslips.

Labeling of Cells

The cells were rinsed with TCM (no serum) for I to 2 min before adding 1 to 2 ~tg/mlFI-PE (N-[5-fluoresceinthiocerbamoyl]dipalmitoyl-L-alpha- phosphatidylethunolamine),Lot 9A (Molecular Probes, Inc., Eugene, OR) in TCM (prepared by adding 1to 2/A of 2 mg/mi stock solution in ethanol to 2 ml TCM while vortexing). The cells were incubated for 10 min with the FI-PE in a moist chamber in the dark. FOr all labeling steps, the fish cultures were maintained at room temperature while PtK and C3H ceils were maintained at 37~ The cells were rinsed three times 2 rain with TCM+, and then the coverslips were inverted over 100/tl of 30-nm gold anti-fluorescein(Fl) in TCM+ on a glass slide with two spacers cut from No, 2 glass coversilps. The gold anti-Flwas prepared and stored as previ- ously described (Lee et al., 1991).The goldanti-Flwas washedthree times in TCM+ before putting on the cells. After a 20-rain incubation, unbound gold anti-Flwas washed out with 200/~1of TCM+, and the chamber was sealed with VALAP (1:1:2by weight vaseline, lanolin, and paraffin). FOr FRAP measurements, all ceils were labeled at room temperature using a four- to ten-fold higher concentration of FI-PE.

Enzyme Treatment

For the heparinase treatment of keratocytes, 5 U of heparinase I from Flavobacteriumheparinum(Sigma Chemical Co.) in PBS with Ca2+ and Mg2+was put on the cells for 20 minat room temperature. The heparinase was removedand FI-PE was added as above. For the PtK ceils, heparinase (F/avobacter/um) from Fluka Chemika-Biochemika (l~nkonkoma, NY) was used at approximatelythe same strength as for the keratocytes and the enzyme treatment was at 37~ Particle Exclusion Assay This assay is used to delineate the periceilular matrix or cell coat of living ceils (Goldbergand Toole, 1984; Knudson and Knudson, 1991).Horse red blood ceils (big Scientific,Bufa]o Grove,IL) werewashedin PBSandthen fixed in 4% formaldehyde for 2 h. The erythrocytes were washed exten- sively in PBS before use. 100 pl of 1 * 10-s cells/ml in PBS with 0.1% BSAwas placed on a slide with spacers. A coverslipwith cells was inverted overthe redblood cells, andthechamber wassealedwith VALAP.The slide was placed on the microscope and the erythrocy~s were allowedto settle for l0 rain before observation and photography.

Light Microscopy and Image Analysis

Theinstrumentation andprocedure havepreviously beendescribed indetail (Lee et al., 1991). Briefly, we used an Axiovert microscope (Carl Zeiss, Oberhocben, Germany) equipped with a 1.4 NA oil immersion condenser anda 1.3NA, 100* Plan-Neofluarobjective. A HamanmmuC2400 newvi- concamera wasconnectedto the microscope with a 4* adapter. Brightfield microscopy was used for gold particle ima~in~~ differentialinterference contrast microscopy was used to see celluiar details. An Imnge-Iprocessor (UniversalImagingCorp., WestChester, PA)wasused to digitally enhance, record, andanalyze imnges.Imageswere stored on an optical memory disk recorder (Panasonic2028F), usually in a series of 200 imagesat video rate (30 frames/s). Images were retrieved and farther processed to improvethe distiuction betwe~ the gold and the background before locating the cen- troids ofeach goldparticle ineach im~e, The a~uracy oflucatin8 theposi- tion of the centmid was limited to pixel size which was 50-60 nm at the masnificationused. A tracki-~program (Lee et al., 1991)wasused to com- pete the trajectories by locatingthe nearest centmidin each succeedingim- age. Mean square displacements (MSD) and lateral diff3mioncoe/ficie~ts (D) were competed as previously described (Lee et al., 1991). For the crawling ceils, velocity corrections were made on the MSD according to the formula: MSD ffi 4Dr + v2t2 (1) where v is the velocity of the crawling cell. The velocity was determined from the centroid of the nuclear area at the be~'nnine~ endingtime points. Alternatively, when the whole nucleus was not visible (some ceils wouldnotwholly fitonthemonitoratthemagnificationused), thedisplace- meritoftheleading edgeofthe nucleus wasmeasured. Thevelocityfor each cell was applied to the corresponding trajectories for that cell. Thiscorrec- The Journal of Cell Biology, Volume 120, 1993 26 tion had little to no effect on the value of D for each trajectory because D was computed from the slope of the MSD plot overthe shortertimeintervals where the plot was linear (Lee et ai., 1991). The corrections occasionally generated negative numbers for the MSD for the longertime intervals;these numbers were deleted when averaging the MSD's across trajectories and overall had little effect on the shape of the MSD plot and no effect on the value of Do. FRAP afterPhotobleaching FRAP was used to obtainthe D ofthe F1-PE in the cell membranesindepen- dent of the gold tag. The instrumentationand procedure for FRAP has been previously described (Lee et al., 1991;Zhang et ai., 1991). An oil immer- sion, 1.3 NA, 40x objective was used to bleach a 2.2-/~m diana (1/e2 width) spot. ComputerSimulations of CorralledDiffusion The method for simulating two-dimensional random diffusion (Lee et al., 1991) was modified by creating a circular,impenetrable boundary around the domain. The size of the domain could be varied by adjusting the di- ameter of the boundary. The particle started at a random location within the circle and movedby steps in randomdirections. The step size was deter- mined by the diffusion coefficient. The boundary was a non-reflective hat- tier. Thus a particle moving within the boundary would not be able to take a step which moved it over the boundary, but the clock would be in- cremented. No motion would be allowed until a step occurred in a direction which would movetheparticle to a newpositioninside theboundary region. EM Whole Mount. Cells were plated on formvar-coated gold locater grids at- tached to coverslips essentially as described (vanden Pol, 1989). The cells wore labeled as above and then fixed for 10 rain in 3% ghitaraldebyde in PBS containing 5 mM Mg2+ at room temperature. Video micrographs were made of selected cells. (There was no movement of the gold on the fixed cells.) With the grids still attached to the coverslips, the cells were postfixed in 0.5% OsO4 0.8% K3Fe(CN)6 in PBS plus Mg2+ for 5 rain at room temperature. After dehydration through increasing concentrations of ethanol and critical point drying, the grids wore removed from the cover- slips and the cells were examined on a JEOL transmission electron micro- scope. (A better method needs to be devised for removing the grids for the coverslips as this frequently resulted in cells with breaks in the cytoplasm.) Sectioned Cells. Cells wore grown in Lab-tek chambers with Permanox slides and labeled as above. After fixation in 2% glutaraldehyde containing 1% Alcian blue (Eastman Kodak, Co., Rochester, NY; Behnke and Zelander, 1970) in 0.1 M cacodylate buffer, pH 7, for 20 min at room tem- perature, the microchamber was removed from the slide containing the cells. The slide was inverted in buffer in a petri dish so that the cells would not be exposed to air or surface tensionduring subsequentsolutionchanges. The cells were postfixed in 1% OsO4 in cacodylate buffer for 60 rain at room temperature. After embedding in epoxy resin, individual cells were selected and then sectioned perpendicular to the substrate. 60- to 70-nm sections were stained in 4% aqueous uranyl acetate for 20 rain followed by 0.4% lead citrate for 8 min.