Journal of Cell Science 102, 19-30 (1992) 19 Printed in Great Britain © The Company of Biologists Limited 1992 Identification of the intermediate filament-associated protein gyronemin as filamin Implications for a novel mechanism of cytoskeletal interaction KEVIN D. BROWN* and LESTER I. BINDER! Department of Cell Biology, School of Medicine and Dentistry, University of Alabama at Birmingham, BHS-63l/University Station, Birmingham, AL 35294, USA *Present address: Department of Biological Chemistry, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, MD 21205, USA tAuthor for correspondence Summary In a previous paper, a monoclonal antibody (designated bovine gyronemin as the immunogen show this protein Ml.4) that recognized a 240 kDa polypeptide was to be associated with actin-containing stress fibers, characterized. This antibody stained the intermediate although our original Ml.4 antibody continued to be filaments in several cell lines, and biochemical charac- localized along vimentin filaments. Since two-dimen- teristics of the 240 kDa polypeptide led us to conclude sional electrophoretic analysis did not demonstrate a that it was a novel intermediate filament-associated difference in either relative molecular mass or iso- protein, which we termed gyronemin. Here we report electric point of this polypeptide when associated with that gyronemin is expressed in adult rat organs that either filamentous system, we conclude that filamin is a contain a substantial smooth muscle component. Taking bifunctional protein capable of associating with both the advantage of this observation, this protein was purified intermediate filament and actin cytoskeletal systems. from bovine uterine tissue and, by biochemical, immu- nological and amino acid sequence analysis, found to be homologous to the actin-associated protein filamin. Three novel monoclonal antibodies raised using purified Key words: intermediate filament, gyronemin, filamin. Introduction cytoskeletal element with which they associate. How- ever, some proteins have been shown to bind or be co- The eukaryotic cytoskeleton is largely composed of localized with more than one type of filament (for microtubules, microfilaments and intermediate fila- example, see Frappier et al., 1987; Griffith and Pollard, ments. Although they are biochemically distinct poly- 1982; Mangeat and Burridge, 1984), and therefore may meric systems, several observations indicate that they be involved, in part, in facilitating heterologous cross- are physically interconnected. For example, exposure linking of different cytoskeletal systems. Alternatively, of cells to microtubule-dismpting agents (e.g. colchi- differential functions of these proteins may be specified cine) leads not only to ablation of most of the via association with morphologically and biochemically cytoplasmic microtubule complex, but also to large distinct polymers. morphological changes in the intermediate filament This paper is an extension of our investigation of a cytoskeleton (Goldman, 1971; Goldman and Knipe, protein that we found to be associated with the 1972), suggesting an association between these two intermediate filament cytoskeleton in several mam- cytoskeletal systems. Additionally, ultrastructural malian cell lines (Brown and Binder, 1990). We analysis of cultured cells suggests that the microfilament determined that this protein, which we termed gyrone- and intermediate filament systems are interconnected min, possessed characteristics distinct from other (Green et al., 1986, 1987). Elements that serve to previously established intermediate filament-associated stabilize, attach or bridge homologous cytoskeletal proteins (IFAPs). Here we demonstrate that gyronemin elements, the cytoskeletal-associated proteins (i.e. is expressed in tissues that contain a substantial smooth MAPs, EFAPs, actin-associated proteins), have not muscle component, and we have taken advantage of generally been found to be promiscuous in regard to the this expression pattern to purify it from bovine uterus. 20 K. D. Brown and L. I. Binder Molecular, biochemical and immunological criteria indicated, gels were stained with Coomassie Blue (0.1% determined that gyronemin is identical to filamin, a Coomassie brilliant blue R/50% methanol/10% acetic acid, by well-characterized actin-cross-linking protein (for re- vol.) overnight and destained (50% methanol/10% acetic view, see Weihing, 1985). These observations serve to acid, v/v) for 4-5 h. Alternatively, gels were silver stained using the Gelcode staining system (Pierce). Prestained demonstrate that, in addition to binding actin filaments, molecular weight markers were purchased from BRL (Gaith- filamin possesses the ability to associate with the ersburg, MD). intermediate filament cytoskeleton. For double-label/two-dimensional electrophoresis, CV-1 cell extract was subjected to 2-D electrophoresis as outlined above. After running the first and second electrophoretic Materials and methods dimensions, the gels were transfered to nitrocellulose sheets and probed (overnight, room temperature) with an appropri- ate dilution of Ml.4. Subsequently, the sheets were probed Sample preparation 125 with I-labeled goat anti-mouse IgM (NEN-DuPont) for 2 h Adult Sprague-Dawley rats were anesthetized by intra- at room temperature, then rinsed, dried, and subjected to peritoneal injection of a Ketamine-Rompum mixture and autoradiography for 2 h at -80°C. Following this step, the subsequently killed by intra-cardiac perfusion with ice-cold blots were "stripped" by incubation in hot (100°C) borate- buffer (100 mM Tris-HCl, pH 7.4/6 mM EDTA). Upon buffered saline (BBS) (100 mM boric acid/50 mM sodium completion of perfusion, the indicated organs were dissected, borate/150 mM NaCl, pH 7.6) containing 2% SDS/5% /3- minced with a razor blade, and transferred to a Dounce mercaptoethanol, for 20 min. The blots were then extensively homogenizer. An appropriate amount of SDS-sample buffer rinsed in BBS, reblocked in BBS/5% non-fat dried milk and (75 mM Tris-HCl, pH 6.8/1% SDS/5% /3-mercaptoethanol/ incubated in an appropriate dilution of the monoclonal 10% glycerol) was added, and the organ samples were antibody G-10. The immunoblot was subsequently developed homogenized while immersed in a boiling water bath for 5 using goat anti-mouse y-chain peroxidase-conjugated second- min. Following boiling, the samples were cleared by centrifu- ary antibody. As a control, blots were probed with anti-mouse gation (15,000 g, 5 min), the supernatant was carefully /i-chain peroxidase-conjugated secondary antibody after strip- removed, and the protein concentration of each clarified ping. This control was performed to ensure that the outlined sample was determined using the method of Lowry et al. stripping process efficiently removed all Ml.4 from the blot. (1951) after precipitation of the protein by addition of perchloric acid and phosphotungstic acid to final concen- Purification of gyronemin trations of approximately 10% and 1%, respectively. Bovine serum albumin (BSA) was used as the standard. Two hundred grams of frozen bovine uterus (—20°C) were chopped into small pieces and homogenized for 5 min in a For two-dimensional electrophoretic analysis, a segment of Waring blendor in 300 ml of cold (4°C) buffer A (100 mM rat aorta was dissected and, after removal of the adventitia, PIPES-NaOH, pH 6.9/1 mM EGTA/l mM Mg2SO4). Follow- was solubilized as described above in a minimal volume of 2% ing clarification of the homogenate by centrifugation (50,000 SDS/5% /3-mercaptoethanol. Prior to electrophoresis, the g, 1 h, 4°C), the supernatant was harvested and adjusted to sample was diluted with ten times volume of 2-D sample 20% saturation with ammonium sulfate ((NH^SO.j) while buffer (9 M urea/4% Nonidet-P 40/2% /3-mercaptoetha- stiring in an ice-bath. The homogenate was then centrifuged nol/2% pH 3-10 Ampholytes (Bio-Rad)). (50,000g, 30 min, 4°C) and the resultant supernatant adjusted African Green Monkey kidney (CV-1) cells were cultured to 35% saturation with ammonium sulfate. The insoluble in Dulbecco's modified essential medium (DMEM) sup- protein fraction from this second fractionation was harvested plemented with 10% fetal calf serum in a humidified 5% CO2 by centrifugation (same as above) and resuspended in buffer atmosphere (37°C). For electrophoretic analysis, cells were A by Dounce homogenization. This protein fraction was grown in 100 mm Petri dishes until confluent. The cells were subsequently desalted by overnight dialysis against 200 rinsed briefly with PBS prior to the addition of a minimal volumes of buffer A at 4°C. volume of hot 2% SDS/5% j5-mercaptoethanol. Sub- sequently, the cell lysate was removed from the Petri dish and Following dialysis, the fraction was centrifuged (225,000 g, 1 h, 4°C) and the supernatant applied to a cellulose phosphate immersed in a boiling water bath for 5 min. The homogenate (Pll, Whatman) column with a bed volume of approximately was then briefly sonicated, cleared by centrifugation (15,000 45 ml equilibrated in buffer A. Nonadsorbed proteins were g, 10 min) and stored at -80°C. Prior to SDS-PAGE or iso- eluted in buffer A and the column was washed with 3 column electric focusing, the extract was diluted to a proper volumes of buffer A/50 mM NaCl. Gyronemin was eluted concentration in SDS-sample buffer or 2-D sample buffer, from the column by application of buffer A/150 mM NaCl. respectively. After cellulose phosphate chromatography, gyronemin- enriched fractions were pooled and applied to a hydroxyl- Electrophoresis and immunoblotting apetite (Bio-gel HTP, Bio-Rad) column (approx. bed volume One-dimensional electrophoresis (SDS-PAGE) was per- 10 ml) equilibrated in buffer A/150 mM NaCl. The column formed as described by Laemmli (1970) on 5 cm x 8.5 cm was rinsed with 3 column volumes of buffer A/150 mM vertical gels consisting of a 4% to 8% linear acrylamide NaCl/100 mM K2HPO4, pH 6.9. Gyronemin was eluted from gradient except where noted. Two-dimensional (2-D) electro- the column with buffer A/150 mM NaCI/200 mM K2HPO4, pH phoresis was performed using a modification of the procedure 6.9.