and Horwitz, 1987; Burridge et al. 1988; Hynes, 1987;. Ruoslahti, 1988). ..... antibody; Dr Irving Goldschneider, and Mr Jerome Murphy for the FACS analysis; Mr ...
Characterization of Chinese hamster ovary cells with impaired spreading properties on fibronectin L. B. JOSEPH 1 ' 2 '*, D. L. KREUTZER2 and M. L. TANZER1 Departments of ^BioStructure and Function, and ^Pathology, University of Connecticut Health Center, Farmington, Connecticut 06032, USA * Author for correspondence at Department of Orthopaedic Surgery, Thomas Jefferson University, 1015 Walnut Street, Philadelphia, Pennsylvania 19107, USA
Summary
The development of receptor-defective or -deficient mutants can be applied to the investigation of cellmatrix interactions including cell adherence and spreading. In the present study we developed a series of ethyl methyl sulfonate (EMS)-induced Chinese hamster ovary (CHO) cell mutants, which adhere to fibronectin but have impaired spreading characteristics. Using morphometric analysis, a significant suppression in the degree of cell spreading between the wild-type and the mutant cells (P are important in cell adhesion and cell spreading, we carried out a comparative immunochemical analysis, using a monoclonal antibody to the Pi subunit of integrin (7E2). Western blot analysis of cell extracts and cell membranes indicated that both wild-type and mutant cells expressed the a and
Introduction Extracellular components, such as fibronectin, appear to play a crucial role in the maintenance of cell function during a variety of events including embryogenesis and wound healing (Hynes, 1987; Reddi, 1984; Ruoslahti, 1988). Cell—matrix interactions seen during wound healing and embryogenesis clearly involve a variety of factors and interactions that are initiated by cell adhesion and spreading (Bronner-Fraser, 1985; Lackie, 1986: Thiery et al. 1985). Cellular adhesion can involve non-specific interactions, e.g. ionic or hydrophobic, or specific surface receptors (e.g. fibronectin receptor with fibronectin) (Pytela et al. 1985; Ruoslahti and Pierschbacher, 1987). Cell adhesion has been investigated at the cellular and the molecular level using transformed or mutant cell lines (Dahl and Grabel, 1989; Esko et al. 1985; Harper and Juliano, 1980; Hirst et al. 1986; LeBaron et al. 1988). Kao and Puck (1968), using Chinese hamster ovary (CHO) cells, found they could isolate stable mutants of these cells Journal of Cell Science 96, 519-526 (1990) Printed in Great Britain © The Company of Biologists Limited 1990
^ subunits of the fibronectin receptor; the mutant cells displayed reduced levels of the subunit. Immunohistochemical analysis indicated that, despite the presence of the receptor in both cell types, their patterns of localization and aggregation were different. The wild-type cells showed a needle-like distribution of the receptor, in contrast to the clumped appearance in the mutants. Furthermore, FACS analysis of antibody-labelled cells showed that the mutant cells had threefold less labelling of the fibronectin receptor by the 7E2 antibody than the wildtype cells. These data suggest that, in the mutant cells, the fibronectin receptor is present, but it is less abundant at the cell surface. This decrease in surface fibronectin receptors may be responsible for decreased cell spreading of the mutant cells on the fibronectin substratum. Thus, the linkage between adhesion and spreading potentially may be explored by studying these mutant cells.
Key words: cell spreading, Chinese hamster ovary cells, fibronectin receptor, fluorescence-activated cell sorter, integrins.
using ethyl methyl sulfonate (EMS). Using EMS, Esko et al. (1985) developed a series of glycosyl-transferase CHO cell mutants that, due to differences in surface glycosylaminoglycans, had varying degrees of adhesion to extracellular matrices. Harper and Juliano (1980), also using EMS with CHO cells, isolated and characterized a group of cell variants that would not adhere to fibronectin-coated substrata. The addition of dibutyryl cyclic AMP (dbcAMP) to these CHO cell variants corrected for altered type I cAMP-dependent protein kinase defects, and caused behavioral reversion of these putative adhesion mutants, suggesting the involvement of a kinase in cellular adhesion and spreading (Cheung and Juliano, 1985, 1987). Cyclic AMP-dependent protein kinases have been shown to regulate cell adhesion, cell spreading and cell aggregation (Cheung and Juliano, 1985, 1987; Maher and Pasquale, 1988; Nairn et al. 1987). Although these studies and others have begun to elucidate the details of cell adhesion, less is known about the relationship between cell adhesion and cell spreading. Recently, proteoglycans, 519
integrins, microfilaments and their accessory proteins have been implicated in the process of cell spreading by studies that disrupt these components and thereby alter cell spreading (Akiyama et al. 1989; for reviews see Buck and Horwitz, 1987; Burridge et al. 1988; Hynes, 1987; Ruoslahti, 1988). However, the molecular basis of these interactions within the cell is still being unraveled. In the present study we describe a novel approach for the characterization of a mutant CHO cell line having a functional spreading defect on fibronectin. Both, wild-type and mutant CHO cells were compared in suspension using fluorescence-activated cell sorter (FACS) analysis to describe cell size and the surface labelling of the fibronectin receptor. This innovative approach was used in conjunction with the more traditional method of immunofluorescent staining of CHO cells grown on fibronectin substrata. The two cell lines were also compared for their ability to spread on fibronectin, their formation of stress fibers, and their cellular expression and localization of the fibronectin receptor. These data suggest that the behavioral differences observed could be due, in part, to a decrease in the number of surface fibronectin receptors in the mutant cells, which may be the result of altered receptor processing and insertion into the mutant cells' plasma membrane.
Materials and methods Cell culture Chinese hamster ovary (CHO) cells designated Kx isolate were purchased from the American Type Culture Collection (CCL 61). CHO Ki were maintained in F-12 culture medium supplemented with 10% fetal bovine serum (FBS) in a 37 °C humidified, 5% CO2/95 % air environment (Ham, 1965). The mutant CHO cells were maintained in F-12 medium supplemented with 10% defibronectinized FBS. Fibronectin was removed from the FBS by gelatin affinity chromatography (Ruoslahti et al. 1982). FBS was analyzed by Western blotting and an enzyme-linked immunosorbence assay (BLISA) to determine the residual amount of fibronectin. The fibronectin concentration in the FBS was below detectable limits (10 ng). Both the mutant and the wild-type cells were grown for 1 week in suspension, in alpha (aO-Medium (1:1 (v/v) cr-MEM:RPMI 1640 supplemented with 10% FBS; Juliano and Behar-Bannelier, 1975), before experimentation. The wildtype and mutant cells (clone E4) when grown in suspension have a population doubling time of approximately 26 hours. Fibronectin isolation from human serum Fibronectin was prepared from outdated human plasma (Red Cross Blood Bank), which was spun at 2000# for 30min at 4°C then filtered (0.45 /an) to remove debris. The plasma was clotted with 0.1M CaCl2 and passed through a pre-equilibrated gelatinlinked Sepharose 4B (Pharmacia) affinity column. The column was eluted with 6 M urea plus 50 mM Tris-HCl (pH6.5) and the elute was dialyzed against PBS (phosphate-buffered saline) to remove the urea (Ruoslahti et al. 1982). A sample of the fibronectin was run on a 7.5 % polyacrylamide gel, transferred to nitrocellulose (Vangard) and probed with anti-human fibronectin (Cappel). A parallel polyacrylamide gel was run and stained with Coomassie Brilliant Blue R, to determine if the fibronectin was pure. Fibronectin coating Lab-Tek multichamber slides (Miles) were washed and coated with 15/igml" 1 human fibronectin in PBS and allowed to form a dry film (overnight, 37 °C) (Ruoslahti et al. 1982). Both bacteriological plastic Petri dishes and tissue culture plastic flasks were coated at a concentration of lS/igml" 1 fibronectin in 0.05 M sodium bicarbonate-carbonate buffer (pH9.7) overnight at 4°C 520
L. B. Joseph et al.
(Form et al. 1986). Both dry and wet films were equilibrated with PBS and then with culture medium just prior to use. Mutagenesis Suspension cultures of CHO-Kj were mutagenized according to the method of Kao and Puck (1968) as modified by Harper and Juliano (1980). Cells were grown in cr-Medium until a density of lxlO 5 cells ml" 1 was achieved. EMS (Kodak) at 50 fig ml" 1 was then added for 24 h to allow the cells to undergo one population doubling. After mutagenesis cells were washed with PBS to remove the EMS, and cell viability was measured using Trypan Blue exclusion. Cells were resuspended in o--Medium and the mutations were allowed to fix for four population doublings (4 days) before screening of the mutant cells. Screening Since cyclic AMP-dependent protein kinases have been shown to regulate cell adhesion, cell spreading and cell aggregation (Cheung and Juliano, 1985; Cheung et al. 1987; Maher and Pasquale, 1988; Nairn et al. 1987), dbcAMP was added to correct for altered type I cyclic AMP-dependent protein kinase defects (Cheung and Juliano, 1985,1987). This allowed us to separate the wild-type cells from the putative mutant cells. The EMS-treated cells (lxlO 5 cells ml" 1 ) were exposed to dbcAMP (500 AIM, final concentration) plus methyl xanthine, a phosphodiesterase inhibitor (MIX 200/IM, final concentration) for 18h in a 1:1, ex MEM:RPMI 1640 medium supplemented with 1% FBS and 4% BSA (bovine serum albumin) (Cheung and Juliano, 1985). Cells were plated onto fibronectin-coated (15 fig ml" 1 ) flasks (Form et al. 1986) and after 2h the loosely adherent and non-adherent cells were saved separately and grown in suspension to a density of lxlO 5 cells ml" 1 . The loosely adherent cells were rinsed with cold PBS and the flask was gently tapped. The cells plus PBS were triturated 10 times against the bottom of the flask to remove the loosely adherent cells. The cells were pelleted at 900 £ and replated into the cr-Medium. Both the loosely adherent and nonadherent cells were rescreened in suspension with dbcAMP/ methyl xanthine and plated onto fibronectin. Screening was repeated a total of four times. Clones were isolated, expanded and either frozen in liquid nitrogen for future studies or rescreened before experimentation. Cells were rescreened every 6 months for maintenance of phenotypic behavior. Spreading assay To determine whether the mutant and wild-type cells were able to achieve the same surface area, 10000 cells were plated on fibronectin-coated slides (Ruoslahti et al. 1982) in serum-free Ham's F-12 for 1 and 2h incubations under standard growth conditions. Cells were fixed according to Dejana et al. (1987) and stained with 1 % Toluidine Blue in formaldehyde (Ruoslahti et al. 1982). One hundred non-overlapping cells were drawn using Nomarski imaging at x650 magnification. The camera lucida drawings were digitized on a Numonics digitizing tablet interfaced to a LSI-11/73 computer. Area, perimeter, diameter and shape were calculated for each cell (Schneiderman et al. 1988). Prior to comparing cell areas using Student's i-test, a bar graph was generated to determine the degree of skewness. The data were found to be skewed, so a natural log transformation was done, followed by Student's £-test using Stat View 512+ (Feldman and Gagnon, 1986). Fluorescence-activated cell sorter (FACS) analysis Fluorescence-activated cell sorter analysis was used to determine: (1) whether the parental and mutant cells have the same physical parameters; and (2) are there any differences in the surface expression of fibronectin receptors between the wild-type and mutant cells? Morphometric analysis of CHO cells Suspension cultures of both the mutant and wild-type cells were washed with PBS, fixed with 10% buffered formalin, and then stored at 4°C until processing. The side and back light scatter
were determined from both fixed and unfixed cells. Low-angle light scatter was used to describe cellular shape and volume, and 90° light scatter was used to describe relative size (Shapiro, 1985). Surface labelling of CHO cells with a monoclonal antibody to the fibronectin receptor Suspension cultures of both the mutant and wild-type cells were washed with PBS, and blocked with 5 % goat serum for 2 h. Cells were washed, then incubated with either a 1:1000 dilution of the monoclonal antibody 7E2 (which recognizes the integrin Pi subunit, a kind gift from R. L. Juliano (Brown and Juliano, 1988)) or IgG K3 (mouse myeloma protein, Bionetics) for 2 h. Visualization of the primary antibody was done using FITC-linked goat anti-mouse IgG (Cappel). The cells were fixed with 10% buffered paraformaldehyde and stored at 4°C until analysis. Cells were processed at 4°C and all buffers contained 0.1% sodium azide. Sodium azide was added to inhibit the capping of the antigen. The cells were analyzed using a Becton-Dickinson FACS Analyzer and the data were collected using a Becton-Dickinson Consort 30 mini-computer. The fluorescence-generated histograms were analyzed using Kolmogorov-Smirnov statistics, a non-parametric test, to determine the difference between the two populations of cells at a confidence level of 99.9 % (Pcrit, then the populations of cells were considered to be different.
added for 30 min at room temperature. Then rhodamine/ phalloidin (Molecular Probes) was added to the secondary antibody at a 1:100 dilution for 30 min at room temperature or, alternatively, the cells were incubated with rhodamine/ phalloidin immediately after fixation and permeabilization. The cells were washed in 1 % goat serum, dipped in 2 M urea for 1 min and then soaked in PBS. Coverslips were applied with Fluoromount-8 (Fisher Scientific) plus n-propyl gallate and allowed to harden at 4°C. Cells were viewed with a Nikon Optiphot microscope and photographed using Kodak Tech Pan (ASA 125).
Results Mutagenesis and cell screening Using the screening protocol outlined in Materials and methods, we successfully isolated 40 clones of CHO Ki cells. These cell clones were selected for their ability to adhere to but not spread on fibronectin. A single clone, designated E 4 , was expanded and utilized for all studies described below. Morphometric analysis of cell spreading Cell area was evaluated using Nomarski imaging to determine whether the wild-type and mutant cells achieved the same degree of spreading on fibronectin. Fig. 1 is a graphic representation of the areas covered by the wild-type and the mutant cells 1 and 2 h post-plating on fibronectin. At 1 h post-plating there was a significant difference between the wild-type and mutant cells, which was maintained at 2 h (Student's t-test, P
