Assembly of 2.5 nm filaments from giardin, a protein associated with cytoskeletal microtubules in Giardia

0 Present address: Department of Physiology, University College , Gower Street, London 1 The Department of Zoology, The University of Hull , Hull HU6 7RX, England MICROTUBULES IN GIARDIA RICHARD CROSSLEY* AND DAVID HOLBERTON SUMMARY The giardins are a family of ~30 000 M, structural proteins found in microribbons attached to microtubules in the disc cytoskeleton of Giardia. After examining the solubility of giardins in various agents, a method has been developed to extract these polypeptides and subsequently precipitate them selectively. The giardin chains are soluble in 10 mM-HEPES/EDTA buffer at high pH and low ionic strength, but become insoluble in 10 mM-MES/EDTA buffer at pH 6-7 when the ionic strength is raised above 50 mM salt. By dialysing giardin extracts in turn against dissociating and reassembly buffers, the purification is obtained of a subset of giardin chains identified by sodium dodecyl sulphate/polyacrylamide gel electrophoresis as the cytoskeleton bands 14a, 14b and IS. The structures forming under assembly conditions are all composed of fine filaments, 2-3 nm in diameter. Filaments after the first cycle of assembly are found in bundles, narrow ribbons of two or three filaments, and large ordered tactoids. Assembly after a second cycle of solubilization yields a more uniform population of long ribbons. Both the tactoids and the second cycle ribbons are transversely banded at the 15 nm interval characteristic of microribbons in the cytoskeleton. Filaments in the tactoids are precisely placed at a centre-to-centre separation of 2-5 nm. Other structural features of the tactoids are discussed in relation to the association behaviour and possible dimensions of the giardin molecular subunit. ASSEMBLY OF 2-5 nm FILAMENTS FROM The refining of in vitro models of tubulin assembly has drawn attention to the additional proteins (microtubule-associated proteins or MAPs) that are isolated with microtubules (Berkowitz, Katagiri, Binder & Williams, 1977; Murphy, Vallee & 1983). Various roles have been proposed for MAPs, i.e. in microtubule assembly proteins in cells have restricted the detailed study of their chemistry and structure. For the specialized microtubules of the flagellar axoneme, recent studies of the dissociation of doublets have identified residue filaments that have chemical and morphological similarities to some intermediate filaments (Linck & Langevin, 1982). It has been suggested that these filaments, composed of a set of proteins ('tetkins'), may be integrated into the microtubule wall in place of some tubulin protofilaments (Linck, 1982). Our own studies have concerned microtubules in the disc cytoskeleton of Giardia, where large insoluble microribbons are joined to specific protofilaments (Holberton, 1981). Because of the ease with which these organelles can be isolated without contamination, it has been possible to characterize microribbon proteins in some detail (Crossley & Holberton, 1983a). Microribbons contain a group of polypeptides with chain molecular weight (Mr) values close to 30000. Two of these are prominent and usually appear on electrophoresis gels as a close-spaced double band (band 14); we have called these giardins. Other polypeptides in the set make up the more variable accompanying bands (bands 13 and 15) that normally stain more weakly. Some earlier experiments showed that ribbon proteins shared with flagellar tektins the propensity to form insoluble 2-3 nm diameter filaments (Crossley & Holberton, 19836). Fractions from cytoskeletons prepared in 0-5% sodium lauryl sarcosinate (Sarkosyl) contained soluble giardins, but the proportions of the different ~ 30000 M, chains in these samples were not determined. Salting out with potassium or magnesium salts caused the assembly of some 2-5nm filaments, which were visible amongst polymorphic aggregates because of their tendency to form thin sheets. The composition of the same precipitates was heterogeneous; there was a high overall content of giardin, although other proteins, notably tubulin, had co-precipitated. On the microgels run from this material, the different giardin chains could not be distinguished adequately. Although it was likely that 2-5 nm filaments were composed of giardin, it was not clear which chains were involved in filament formation. In this paper we report the results of a new method of purifying 2 5 nm filaments by cycles of assembly and disassembly, which enables us to identify filament proteins unambiguously. Trophozoites of an axenic strain of Giardia lamblia were cultured in T P S - 1 medium (Visvesvera, 1980) as described previously (Crossley & Holberton, 1983a). Protein extraction experiments For preliminary experiments to compare various dissociating procedures, cells were harvested and washed in TEDAMP buffer, and cytoskeletons prepared by demembranation in TED AMP + Triton as before (Crossley & Holberton, 1983a ). Samples of buffer-washed cytoskeletons with a protein concentration between 0-4 and 0-6 mg ml"1 were extracted in media containing either a chaotropic salt or Sarkosyl at appropriate concentrations. Samples also contained 50 fig ml"1 phenylmethylsulphonyl fluoride (PMSF) and 0-1 mM-benzethonium chloride as a control against endogenous proteolysis during isolation and extraction. After extraction at 0-4 C each sample was pelleted at 48 000 g for 15 min and the soluble protein in portions of the supernatant was assayed by a microbiuret (Goa, 1953) or modified Lowry (Schacterle & Pollack, 1973) procedure. The following solubilizing buffers were used, (a) Tris/EDTA buffer (pH 7-6): 2mM-Tris. HC1, 0-2 mMethylenediaminetetra-acetate (EDTA), 0-5mM-dithiothreitol ( D T T ) . Samples were dialysed against a 100-to 500-fold volume excess of buffer for 48 h. (b) High salt buffer: 0 - 6 M - K C 1 , 10mMTris, 0-1 mM-EDTA, O-5mM-DTT (pH 7-8), dialysed for 18h against 100-250 vol. (Mabuchi, Shimizu & Mabuchi, 1976). ( C ) 0 - 6 M - K I , 10 mM-Tris, 0-1 mM-EDTA (pH 7-8) extracted for 5 min before pelleting, (d) KSCN at concentrations in the range 0-05-0-4 M was added as a 9 vol. excess to 100 /xl samples of cytoskeletons in 10mM-Tris-phosphate buffer, 0-lmM-DTT (pH 7-2) and extracted on ice for 1 h. (e) Extracted for 1 h in Sarkosyl made up in 10mM-Trisphosphate, 0-lmM-DTT (pH 7-8) in the range 0-0125-0-3%. (c), (d), (e), after Linck (1976). To assess extraction of different protein species, gels were prepared from the insoluble residues and the staining of the corresponding bands was determined by densitometry (Holberton & Ward, 1981). Two buffers were used in reassembly experiments: dissociating buffer was 10mM-HEPES, 5 mM-EDTA, 0-05mM-DTT (pH 8-6-8-7); reassembly buffer was 10mM-MES, 2mM-EDTA, 0-5 mM-DTT (pH 6-7) with between 50 mM and 200mM-KCl. Cells were harvested from 41 of 4day cultures and washed three times by suspension and pelleting in 200 ml of cold reassembly buffer with 150mM-KCl and 1 mM-ATP. Cells were spun down each time at 4 (...truncated)