Regulated expression of vimentin cDNA in cells in the presence and absence of a preexisting vimentin filament network.

Isolation of a Mouse Vimentin cDNA Total cellular

RNA was isolated from mouse L-929 cells bythe guanidinium thi0cyanate procedure of Han et al. (1987) and poly A+ RNA selected by oligo (dT)-cellulose chromatography. This RNA was used for DNA synthe- sis using a commercial eDNA synthesis kit (Pharmacia Fine Chemicals, Piscataway, NJ) to produce a recombinantlibrary in the phagemid lambda- ZAP (Stratagene Corp., La Jolla, CA). The library was screened by replica filter hybridization using a 32p-labeled hamster vimentin eDNA, pVim-1 (Quax-Jeuken et al., 1983) (girl of Dr. H. Bloemendal). A number of posi- tive recombinants were obtained and two were selected for further study. Both were found to contain inserts of '~1.8kb that hybridized with the pV'lm-1 probe (data not shown). When produced as the excised plasmid form (Bluescript SK-), the plasmidspFB4.2 and pFBT.2 yielded restriction maps indicating identical vimentininserts in opposite orientations. Both plasmids were linearizedwith Xba I and the insertstranscribedin vitrousing T7 RNA polymerase. The resulting RNA was then translated in a reticulocyte lysate system containing [35S]methionine followed by analysis of the 35S-labeled proteins by SDS-PAGE (Laemmli, t970). Molecular weight markers from a commercial kit were used as a reference (SDS-6H; Sigma Chemical Co., St. Louis, MO). As shown in Fig. 1, pFB4.2 yielded a single labeled 57-kD protein species which comigrated with purified mouse vimentin, whereas pFB7.2 yielded only background reticulocyte proteins. A comparison of the restrictionmaps ofpFB4.2 and pFB7.2 (not shown) withthe hamstervimen- tin sequence (Quax et al., 1983), and the in vitro transcription/translation data indicate that pFB4.2 containeda full-length vimentincDNA sequence in the sense orientation with respect to the T7 promoter.

Construction of Regulatable

Vimentin cDNA and Glucocorticoid Receptor-Neomycin Resistance Expression Vectors

The plasmid pFB4.2 was digested with Eco RI and the excised 1.8-kb vi- mentin eDNA insert was purified. The ends were made flush with T4 DNA polymerase, Barn HI linkers were added, and the cDNA was inserted into the Barn HI site of the ptasmid pSP64-MMTV. This plasmid is a pSP64- based plasmid containing the promoter and hormone response element of The Journal of Cell Biology, Volume 111, 1990 554 mouse mammary tumor virus and the polyadenylation signal of murine sar- coma virus (gift of E van der Hoorn). A plasmid containing the 1.8-kb in- sert in the sense orientation, was designated pSP64-MMTV-VimS (Fig. 2). For use in cotransfectionexperiments, a plasmid designated pRGRN was constructed, pRGRN contains the rat glucocorticoid receptor eDNA and the neomycinphosphotransferase gene as separate transcription units, each under the control 'of the Rous sarcoma virus promoter. To construct pRGRN, the Barn HI site of the rat glucocorticoid receptor expression vec- tor pRSVGR (Miesfeld et al., 1986) (gift of K. Yamamoto) was converted to an Mlu I site. The RSV promoter and glucocorticoid receptor gene were excised from the resulting vector by digestion with Mlu I. This MIu I frag- ment was inserted into the plasmid pRSVneo (gift of B. Howard) at the unique Mlu I site to create pRGRN.

Cell Culture and DNA Transfection

HeLa cells (ATCC CCL 2.2) and SW-13 cells (ATCC CCL 105) were ob- tained from the American Type Culture Collection (Rockville, MD). MCF-7 cells were obtained from Dr. D. Edwards. Cells were grown in monolayer culture in a 1:1 mixture of Ham's Fl2:Dulbecco's MEM containing 5% FBS. SW-13 cells express vimentin in a mosaic pattern (Hedberg and Chen, 1986). SW-13 ceils were subcloned as described by Hedberg and Cben (1986). 50 subelones were obtained and examined for human vimentin con- tent by immunofluorescencemicroscopy. 25 of the subclones were found to be essentiallyvimentin negative, 22 were mosaics with a significant number of positive and negative cells, and 3 subclones were obtained that appeared to contain essentiallyonly cells with prominent vimentin filament networks. A subelone of the vimentin-positive cells, designated SW-13/cl.1 vim+ and vimentin negative cells, designated SW-13/cl.2 vim- were selected for fur- ther study. Stable cell lines were obtained by cotransfecting 0.6--0.8 x 106 cells in 10-cm dishes with 40/~g pSP64-MMTV-VimS and 1.0 #g pRGRN by cal- cium phosphate precipitation ;(Graham and van der Eb, 1973) and 15% glycerol shock (Parker and Strak, 1979). Stable transfectants were selected in medium containing 400 #g/mi G-41g. Individual colonies were recovered and examined for mouse vimentin filament content in the presence and ab- sence of l0-7 M dexamethasone by indirect immanefluorescence. After 2-3 mo of continuous culture in 400 #g/ml G-418, stable cell lines were then maintained in medium containing 200 t~g/ml G-418. Although prelimi- nary experiments indicated that G-418 had no effect, all experiments were conducted in the absence of this antibiotic. Transient transfections were per- formed under similar conditions without cotransfection with pRGRN or G-418 selection.

Indirect Immunofluorescence

Cells were plated on sterile glass cover slips. The cells were rinsed briefly in PBS and then fixed in 70:30 acetone/methanol (voi/vol) at -20C for 10 rain. The cover slips were rinsed in PBS, and processed for indirect im- munofluorescence as described by Franke et al. (1978). Rabbit anti- vimentin (Moscinski and Evans, 1987), and monoclonal anti-vimentin (V-9; Boehringer Mannheim Biochemicals, Indianapolis, IN) were used as pri- mary antibodies. The rabbit anti-vimentin serum (diluted 1:100) visualized intermediate filaments in mouse cells, but did not detect filaments in a vari- ety of vimenfin containing human cells. Conversely, the commercial mono- clonal anti-vimentin (2 ~,g/ml) visualized intermediate filaments in human cells but did not detect filaments in vimentin-containing mouse L-929 or 3T3 cells. Fluorescein-conjugated anti-rabbit and lissamine-rhodamine- conjugated anti-mouse antisera (Boehringer Mannbeim Biochemicals) were used as second antibodies, respectively. All antibodies were diluted in PBS containing 1% ovalbumin and 1% normal goat serum. The cover slips were mounted in Aqua-mount (Lerner Laboratories, Pittsburgh, PA) and viewed on an Olympus microscope equipped with epifluorescence optics. Photo- graphic exposures were made for 15-20 s using Kodak T-Max 400 film and the film processed with an exposure index of 1,200 using Kodak HC-IIO developer. For all figures, the photographs represent identical exposures and photographic processing conditions. Preparation andAnalysis ofrsS]labeled Triton-insoluble Cytoskeletons Cells were labeled with 25-50 ~Ci/ml [35S]methionine for 2 h in methio- nine-free medium containing 1% FBS. Triton-insoluble cytoskeletons were prepared by the method of Zackroff and Goldman (1979) as previously de- scribed (Evans, 1984). In experiments with SW-13cells, DNase I was de- leted from the Triton-KC1 solution to preserve filamentous actin as an inter- hal marker. Triton-insolublecytoskeletonswere analyzed by one (Laemmli, 1970) and two-dimensional gel electrophoresis (O'Farrell, 1975). All sec- ond dimension SDS-PAGE contained 7.5 % acrylamide. After electmphore- sis, gels were stained with Coomnssie blue and destained as previously de- scribed (Evans, 1984). The gels were dried and autoradiographed on Kodak XAR x-ray film at -70C. In some experiments, [35S]labeled proteins were directly qnantitated in dried gels using an imaging scanner (System 200; Bioscan, Inc., Washington, DC). The gels were scanned for 20 min per lane using the tm>-dimensional analysis program. Background was subtracted using a nonradioactive portion of the gel. The [ass] radioactivity for each gel lane and within individual protein peaks was determined using the man- ual mode.