A contractile cytoskeletal network of Paramecium: the infraciliary lattice

NICOLE GARREAU DE LOUBRESSE 1 *, GUY KERYER 2 , BERNARD VIGUES 3 and JANINE BEISSON 2 'Centre lie liintogie Celhilaire, CXRS, 94205, Iviy-snr-Seine, France H'enlre cle Ge'ne'liqne Molecnlane, CXRS, 91190 Gif-sttr-Yvette, France 3 Gmnf>e do'/.oolofiicel I'mlistohgie, L'nnersile tie Clennnnl II, 63170 Anbiere, France * Author for correspondence

Summary Among the various cytoskeletal networks and structures that compose the highly organized cortex of Paramecium, the infraciliary lattice (ICL) is the innermost one and marks the cortical boundary. It is made up of filamentous bundles running around the proximal ends of basal bodies. Immunofluorescence studies, using two different antibodies specifically labelling the network, allowed us to describe the disorganization-reorganization cycle of the ICL during division, and observations by immunoelectron microscopy provided new information on its ultrastructure. A partial purification of large frag-

ments of a contracted form of the network was obtained. Among the proteins from ICL-enriched fractions, two 23 to 24(XlO3)Mr polypeptides were shown to cross-react with a polyclonal antibody raised against 22 to 23(xlO 3 )M r Ca2+-binding proteins from the microfibrillar ectoplasmicendoplasmic boundary of Isotricha and to be selectively solubilized in the presence of EGTA. The possibility that the ICL may be a new example of a non-actin contractile system found in lower eukaryotes is discussed.

Introduction

onal meshes around the proximal ends of basal bodies. Apart from these ultrastructural data, nothing was known of the behaviour during division, the chemical nature or the function of these networks. An approach to the biochemical and functional characterization of the outer and infraciliary lattices was provided by specific immunological probes. Among a batter}' of monoclonal antibodies raised against cortical structures from the ciliated cells of quail oviduct (Klotz et al. 1984), two were found to cross-react with Paramecium. The first one, CC212, labelled the outer lattice and permitted us to follow its behaviour during division and to identify a 130K (K = 103-/17r) polypeptide as a major component of the structure (Cohen el al. 1987). The second antibody, 3D 12, labelled the infraciliary lattice. Furthermore, a polyclonal antibody raised against a microfilamentous system (ectoplasmic-endoplasmic boundary or EEB) of the ciliated protozoon Isotricha and shown to react with a 22-23K Ca 2+ -binding polypeptide (Vigues et al. 1984; Vigues & Grolie"re, 1985), was found by Vigues (unpublished observations) to mark the infraciliary lattice of Paramecium.

The cortical cytoskeleton of Paramecium comprises two sets of inter-mingled components. The first set is mainly composed of the basal bodies and the various structures that arise from them or are closely apposed to them (cilia, kinetodesmal fibres, microtubular ribbons, epiplasm, striated bands); their arrangement forms complex repeated units, the cortical units, aligned in parallel longitudinal rows. The second set comprises two fibrous systems, which form continuous networks spanning the whole cell surface and whose meshes run around the cortical units. The organization of these two networks, first visualized in light microscopy by silver staining (Von Gelei, 1937; Parducz, 1962), was studied by electron microscopy (Pitelka, 1965, 1969; Jurand & Selman, 1969; Allen, 1971; Ehret & MacArdle, 1974). As has been well established, especially by Allen (1971), the outermost network (outer lattice) runs along the top of the surface ridges, right beneath the plasma membrane, and delineates the cortical units, while the innermost network, the infraciliary lattice, forms irregular polygJournal of Cell Science 90, 351-364 (1988) Printed in Great Britain © The Company of Biologists Limited 1988

Key words: cytoskeleton, Paramecium, immunocytochemistry, contractility.

351

The use of both antibodies, 3D12 and anti-EEB, allowed us to study by immunofluorescence, at the whole-cell level, the dynamics of the infraciliary lattice throughout the cell cycle. Electron-microscope studies using the immunogold technique confirmed the localization of antibody labelling on the infraciliary lattice and revealed new aspects of its ultrastructure. Finally, we have been able to obtain a partial purification of the infraciliary lattice and characterize a doublet of 23-24K polypcptidcs as the major component of the network. We show that the whole network displays contractile properties and that its assembly is Ca 2+ dependent. The possibility that the 23-24K polypeptides may belong to the class of spasmins (Amos, 1986), that is, of proteins whose contraction is directly induced by Ca 2+ -binding, is discussed.

Materials and methods Strains and culture conditions The cells used in the experiments were the stock d4-2 of I'aramecium letraurelia (Sonneborn, 1975). Cells were grown at 27°C according to usual procedures (Sonneborn, 1970) in grass infusion bacterized the day before use with Aembacter aemgenes and supplemented with 04jigml~' fisitosterol. Antibodies We used the monoclonal antibody 3D 12 raised against cytoskeletal proteins of quail oviduct ciliated cells by immunization of mice with a crude preparation of basal bodies and associated material (Klotze/ al. 1984, 1986). We also utilized a polyclonal antibody raised against proteins associated with the ectoplasmic-endoplasmic boundary (EEB) of the ciliated protozoon Isotricha. This rabbit antiserum was obtained after injection of the material from two closely migrating bands (22-23K) separated by SDS-PAGE (polyacrylamide gel electrophoresis) of isolated cortices of Isotricha (Yigues et at. 1984). Iminiinocytocheniical pivcedures In the previous studies of cytoskeletal structures of Paramecium (Cohen et al. 1982, 1984, 1987) we generally used the PHEM buffer of Schliwa & Van Blerkom (1981) for cell permeabilization and further immunological processing. However, the infraciliar lattice (ICL) was poorly preserved in this buffer. The following phosphate buffer (PB) was found satisfactory: 10mM-KH 2 PO 4 /Na 2 HPO 4 (pH7-2), 10mMEGTA, 2mM-MgCI2. Inwiiniofliiorescence micmscupy. Two distinct protocols, one without and the other with fixation, were used and gave identical results. (I) Living cells were cither first permeabilized in PB containing 0'5 % Triton X-100 and then transferred to the antibody solution, or simultaneously permeabilized and labelled by transfer into a mixture containing both the antibody and the detergent in PB. In the latter method, which was principally used, the hybridoma supernatant was 352

A', (jarreau de Loubresse et al.

diluted 1/2 in 1 % Triton X-100 in PB while the antiserum was diluted 1/200 or 1/300 in the same buffer containing 0-5% detergent. The cells were incubated for 45-60 min at room temperature, in the presence of 3 % bovine serum albumin (BSA) and 0 1 % Tween 20. Then cells were rinsed three times for 10 min in PB containing BSA and Tween 20 and incubated for 60 min in a FITC (fluorescein isothiocyanate)-antimouse Ig antibody or a FITC-anti-rabbit Ig antibody (from Pasteur Institute) diluted 1/200 in the washing solution. Three additional rinses followed this incubation. (2) After permeabilization for 5-10 min in PB containing 1% Triton X-100, cells were fixed for 60min at room temperature in 2 % formaldehyde in PB. Then the cells were rinsed and further processed either in PB or in phosphatebuffered saline (PBS) supplemented with 3 % BSA and 0-1 % Tween 20. Cells were mounted in Citifluor (Citifluor Ltd, London), observed under a Zeiss epifluorescent photomicroscope and photographed using Agfapan 100 film. Immunoelectron nucmscopv. The protocol without fixation was principally used to avoid the damaging effects of successive centnfugations needed when cells are processed en masse. Cells were treated as for immunofluorescence microscopy except that the fluorescent antibody was replaced by either a colloidal gold-labelled anti-mouse Ig antibody (GAM G5 from Janssen Chimical) or anti-rabbit Ig antibody (GAR G5), diluted 1/5 in PB containing 3 % BSA and 0-1 % Tween 20). For control experiments, the first antibody was omitted and the cells, processed as above, were reacted only with the gold-labelled second antibody. Samples were then processed according to the general procedures for electron microscopy except that thin sections were contrasted with ethanolic uranyl acetate only. Electron microscopy General pivcedures. Whole-cell pellets were fixed in 2 % glutaraldehyde in 005 M-cacodylate buffer (pH 7-2) for 90 min at 4°C. After washing in the same buffer, the samples were post-fixed in 1 % osmium tetroxide in 0-05 M-cacodylate buffer, for 60 min at 4°C. Specific procedures. Fractions enriched in infraciliary lattice (see below) were fixed in 4 % glutaraldehyde in 10mMT r i s H C l (pH 7 3 ) for 90 min at room temperature. After extensive washing in a 10mM-Tris'HO, 0-25 mM-sucrose solution, the samples were fixed in 1 % osmium tetroxide in T n s H C l buffer, for 60 min, at room temperature. In both procedures, the post-fixed samples were dehydrated by passages through a series of ethanol and propylene oxide baths and embedded in Epon. Thin sections were contrasted with ethanolic uranyl acetate and lead citrate, then examined with a Philips EM 201 electron microscope. Preparation offractions enriched in infraciliar}' lattice Batches of 21 of culture ( = 8 x 10°cells), grown at 28°C and then kept as a stationary-phase population at 18°C for 2—4 days, were centrifuged at low speed and the pellets resuspended in 400ml lOmM-Tns-HCl (pH8-5), and washed twice in this buffer. The cell pellet was then injected, with constant stirring, into 150 ml of the following lysis buffer

(buffer A): 1 mM-Tris-HC1 (pH8-5), 0-25 M-sucrose, 1 mMCaCl2, 0-lmM-MgClz, 1-5% Triton X-100, 2xl(T s M-£mercaptoethanol and 50^gml~' heparin. After 1 min, urea was added, to a final concentration of 1-SM, and the suspension was centnfuged for 10 min, at 4°C, at 1500 revs min" 1 in a Sorvall SS-34 rotor. The pellets were resuspended in buffer B: 10 mM-Tris- HC1, 0-25 M-sucrose with protease inhibitors, 0-SmM-PMSF (phenylmethylsulphonyl fluoride) and 1 mM-leupeptin, and washed twice in the same buffer. The final pellets, markedly enriched in whole infraciliary lattices or large fragments of the network, were either boiled for 3 min in SDS to a final concentration of 2 % for SDS-PAGE analysis, or fixed for electron microscopy (see above). In other series of experiments, the infraciliary lattice (ICL)-enriched fraction was resuspended in 2ml of buffer B and divided into two samples, to one of which EGTA, to a final concentration of lOmM, was added. The EGTA and control samples were then centrifuged at 10000|f, for 10 min, and the respective pellets and supernatants were recovered. The pellets were immediately boiled in SDS as indicated above, while the supernatants were dialysed for 24 h at 4°C against 1 M-ammonium acetate, concentrated by freeze-drying and then boiled in SDS.

Organization and dynamics of the infraciliaiy lattice Fig. 1A-B shows the interphase pattern of the network as viewed by immunofluorescence microscopy using the anti-EEB antiserum. Identical results were obtained with the monoclonal antibody 3D12. On most of the cell surface, the network displays polygonal meshes of irregular shape except in two zones: (1) in the left anterior ventral field (Fig. 1A, arrow) where quadrangular meshes are precisely adjusted to the shape of cortical units; (2) at the poles where smaller meshes correspond to more tightly packed basal bodies. On the anterior third of the dorsal surface (Fig. IB), the ICL shows a distinctive arrangement, which is not correlated with any known difference in the organization of basal body rows. Both antibodies also label, by immunofluorescence and immunogold staining (data not shown), a thick bundle of microfibrils, the cytostomal cord (Allen, 1974), which follows the contour of the tip of the buccal cavity. It remains to be determined whether the material decorated in the gullet is continuous with the cortical lattice or constitutes a distinct array.

Polvaciylamide gel electrophoresis and immunoblotting pmcedures

During division, the network undergoes a progressive disorganization followed by a reorganization. The major steps of these changes are illustrated in Fig. 1C-H and their progress is shown schematically in Fig. 2. The disorganization becomes apparent by stages 3-4 of cell division (Fig. 2), i.e. when the old and neoformed gullets start moving apart and elongation of the dividing cell begins. The disorganization proceeds by a discrete disassembly of some polygonal meshes, especially along the right side of the anterior suture and beneath the fission line (Fig. 1C,D). Progressively, more and more meshes disappear (stages 5-6). This leads, by stage 7, to a mainly discontinuous and irregular pattern. The disorganization is most pronounced around the gullets and in the equatorial part of the dividing cell, on both sides of the fission furrow, but does not affect the polar regions. Before cytokinesis is completed (Fig. 1E,F; stage 8), mesh reassembly begins along the anterior and posterior parts of the dividing cell. The reconstitution of the interphase pattern becomes visible in the newly separated fission products (Fig. 1G,H). As summarized in Fig. 2, it is of interest to note that the disorganization progresses over the cell surface as a wave originating from the gullet region. The wave extends along the fission furrow all around the cell: it affects first the posterior fission product and finally spreads anteriorly and posteriorly. In order to determine whether the observed disorganization was due to the permeabilization treatment required for the immunofluorescence study, thin section of intact cells were examined. Fig. 3 compares similar tangential sections of interphase (A) and dividing (B) cells. In interphase cells, as well shown by

Polyacrylamide gel electrophoresis was carried out according to Laemmh (1970) on 12% continuous mini-gels. The samples were loaded as a tandem repeat and one series was stained with Coomassie Blue R-250, while the other was blotted onto nitrocellulose filters according to Towbin et al. (1979) with slight modifications. The filters were reacted for 2h at 37°C with the first antibody, washed and reacted with a peroxidase-conjugated anti-mouse or anti-rabbit antibody (Institut Pasteur, France). The immunoreactive bands were revealed by 4-chloro-l-naphthol.

Results The complex surface pattern of Paramecium results from the juxtaposition of about 4000 unit territories, the cortical units, all similarly organized around one or two basal bodies. These units are aligned in longitudinal parallel rows and their overall arrangement, sustained by several fibrous networks, is polarized and asymmetrical, with three major singular points on the ventral surface; the gullet (feeding apparatus) and the anterior and posterior sutures. While on the dorsal surface all rows run straight and merge at the cell poles, on the ventral surface the basal body rows display a curvature around the gullet (more pronounced on the cell's left side) and merge along the sutures. All the repeated cortical structures precisely follow the basal body arrangement and so does the outer lattice, whose meshes delineate the cortical units (Cohen et al. 1987). In contrast, and as first described by Von Gelei (1937), Parducz (1962) and Pitelka (1969), the infraciliary lattice forms irregular polygonal meshes, which do not exactly follow the cortical unit arrangement.

Contractile network in Paramecium

353

Fig. 1. Immunofluorescence labelling of the infraciliary lattice, using the anti-EEB antibody, throughout the cell cycle. A,B. Interphase organization. A, ventral face; B, dorsal face of the same cell. The network shows regular meshes in the anterior part of the ventral face, especially in the left-hand field (A, arrow) and in the anterior part of the dorsal face (B, arrow). In the rest of the cell, the lattice displays irregular meshes, as, Anterior suture; g, gullet; c, cytoproct. C,D. Early stage of division: C, ventral face; D, dorsal face of the same cell. The old (og) and new (ng) gullets begin to move apart and the equatorial fission line (fl) is more visible on the ventral face than on the dorsal face. The disorganization of the network appears posteriorly to the fission furrow all around the cell, around the gullets and along the anterior cell suture. In the other regions, a few links are missing, although the pattern still retains its overall organization. E,F. End-division

354

N. Gatreau de Loubresse et al.

stage: E, ventral face; F, dorsal face of the same cell. Over a wide zone extending on both sides of the cleavage furrow, the network is very disorganized. On the ventral face, most meshes are incomplete or reduced to short elements, which appear aligned in the anterior part of the posterior fission product (arrowhead). No trace of the network is seen in the zone that extends between the gullets (small arrows). At this stage, note the beginning of the pattern reorganization (large arrow) starting from the anterior pole to the gullet; this onset of reconstitution is more extensive in the dorsal face where the disorganization of the network is only observed in the equatorial zone. G,H. A recently divided cell; G, ventral face; H, dorsal face of the same cell, an anterior fission product. The meshes have mostly recovered their typical disposition in the anterior region, especially in the left ventral field (G, arrow). This reconstitution is not yet completed in the mid-regions. Bar, lOjitm; A-D, X805; E-H, x831.

Contractile network in Paramecium

355

int (5h)

1

2

3

4

5

6

7

8

sep

0m

5m

7m

9m

72m

74m

76 m

18m

22m

a

Fig. 2. Schematic representation of infraciliary lattice dynamics in relation to the main morphological events during the cell cycle. This mapping of ICL disorganization and reorganization uses the same symbolism as that devised by Iftode el al. (unpublished data) for mapping of reorganization of various cortical structures throughout division. I. Sequence of cellular and nuclear events. The elongation and the cleavage of the dividing cell are shown in relation to nuclear and oral apparatus events: the moving of macronucleus (ma) and its elongation, the division of micronuclei (mi), the gullet (g) duplication followed by the separation (sep) of old (og) and new gullets (ng). The average time elapsing from stage 1 until each succeeding division stage are indicated below the scheme according to data from Tucker et al. (1980). int, interphase; cs, cell suture; ms, mitotic spindle; ss, separation spindle. II. Successive steps of the disassembly and the reassembly of the infraciliary lattice. Hatched areas, ICL disorganization; filled area, total absence of ICL meshes; cross-hatched areas, ICL reorganization. I la, ventral face; l i b , dorsal face. The first sign of mesh disassembly is seen by stage 3, when the macronucleus begins to elongate and the gullets separate. The wave of disruption, starting from the gullet region, proceeds along the right side of the anterior suture and beneath the fission line, initially in the left part of the ventral and dorsal faces. By stage 4, the disorganized region forms an equatorial ring. Then, the wave of disassembly spreads on both sides of the fission furrow and anterior suture (stages 5 and 6) towards the anterior and posterior poles. On the ventral face, a zone devoid of ICL meshes develops between the gullets as they move apart. By stage 7, the disorganization has reached its maximum extension but the polar regions remain untouched. Reassembly begins near the end of division (stage 8), apparently spreading from both poles and progresses faster on the dorsal face. In the first separated cells reassembly is not yet complete. Allen (1971) and Ehret & MacArdle (1974), the network is made up of an anastomosing system of filamentous bundles forming the reticulate pattern mentioned above. A triangular electron-opaque struc356

t\r. Gaireau cle Loubresse et al.

ture, called a "post" by Allen (1971) and localized within a small area in a tangential view of the cortical surface, is present at the centre of each branching locus (Fig. 3A, arrows). In contrast, in dividing cells, the

Fig. 3. Tangential thin sections through the cortex showing the differences in the organization of the ICL in an interphase cell (A) and in a cell at an early stage of division (B) (corresponding to that of Fig. 2, stages 5-6). As previously documented (see, in particular, Allen, 1971), the polygonal meshes surrounding basal bodies (bb) and trichocvst (/) tips are composed of branched filamentous bundles of rather rigid aspect (A). Note at the branching points (arrows) the regular presence of a dense triangular structure called a "post" (Allen, 1971). In a dividing cell (B), the meshes are incomplete. The filamentous bundles are thicker and somewhat wavy. At the branching points, the post is either absent (arrowheads) or replaced by two dark points (arrowheads), suggesting a duplication of the structure. New branches seem to originate from this specific site; kf, kinetodesmal fibre; in, mitochondria. Bar, 0-2^jm; X42000.

Contractile network in Paramecium

357

meshes, generally interrupted, are thicker and less straight. At the branching points, the posts are no longer visible (Fig. 3B, arrowheads) or they appear to have duplicated (Fig. 3B, arrows). The data on untreated whole cells therefore confirm the immunofluorescence observations. Furthermore, they indicate that the disorganization of the overall pattern is accompanied by important changes in the ultrastructure of the filamentous bundles. Finally, the apparent duplication of some posts suggests that they might be organizing centres for the reassembly of the network.

Ultrastructure of the infraciliary lattice As the permeabilization treatment cleared the cytoplasmic matrix surrounding the network, new aspects of its ultrastructure were uncovered by immunoelectron microscopy, using both the monoclonal 3D 12 (Fig. 4) and the anti-EEB antiserum (Fig. 5) demonstrated by the immunogold technique. Figs 4 and 5 show that the gold particles are located on the filamentous bundles; however, their distribution along the network is different. As observed by immunofluorescence, 3D 12 decorates the microtubules also (except in the ciliary axonemes), while the anti-EEB specifically labels the infraciliary lattice. With the antibody 3D12, whose pattern of decoration in immunofluorescence microscopy (Fig. 4A) can be compared with that of the anti-EEB (Fig. 1A), the gold particles are spread more or less evenly over the filamentous bundles. In these preparations, when cells were incubated in the simultaneous presence of the permeabilization buffer and antibody (treatment likely to yield better preservation of the structure), a previously unknown structural feature of the network was observed. Fig. 4B—D shows that the longitudinal filaments arranged to form the wall of close-packed tubules and cross-striations at regular intervals are visible. A hexagonal arrangement of the packed filaments is apparent in cross-section (Fig. 4E). In preparations labelled by the anti-EEB, the tubular organization of the ICL was less well preserved. The gold particles seem arranged in a helix around the filamentous bundles (Fig. 5B) and thus reveal another aspect of their complex organization. Partial purification of the infraciliary lattice When Paramecium is lysed by Triton X-100 in a low ionic strength buffer (1 mM-Tris - HC1), the cortex breaks open and swells while the infraciliary lattice, as a whole or in halves, contracts away from the cortex. Fig. 6A illustrates the phenomenon. A large part of one ICL from such a crude preparation is shown in Fig. 6B. However, the network tends to remain attached to 358

N. Gatreau de Loubresse et al.

the cortex at the polar regions and around the gullet; these resistance points are likely to account for its frequent breaking into two halves. Attempts to purify the structure were undertaken and the procedure described in Materials and methods led, quite reproducibly, to fractions notably enriched in ICL but still contaminated with small pieces of cortex, whole gullets and numerous trichocysts. Limited attempts to obtain a better purification on sucrose or on Percoll gradients were unsuccessful. Thin sections of the ICL-enriched fractions are presented in Fig. 6C-E. Remnants of cortical structures, still attached to the network, are visible. The partly-tangential section (Fig. 6E) shows that the general organization in irregular polygonal meshes is retained. However, the mean 'diameter' of the meshes is about half that observed in situ (0-4jiim compared to 0*9/urn). The reduction of mesh size is illustrated by Fig. 7. In the ICL preparation, still attached basal bodies are squeezed within the contracted ICL meshes, which are correspondingly much thicker than in situ. This demonstrates that, at least under these conditions, the network as a whole displays contactile properties. Figs 6D and 7B indicate that ICL contraction, at least as observed in vitro, is accompanied by a marked change in ultrastructure, which suggests a helical compaction (Fig. 6E).

Biochemical characterization of the i?ifraciliary lattice components ICL-enriched fractions were run on SDS—polyacrylamide gels and different preparations yielded the same electrophoretic pattern, illustrated in Fig. 8A (lane a). Immunoblots of such gels incubated with the monoclonal antibody 3D 12 gave no reaction or inconsistent ones. In contrast, when the blots were reacted with the anti-EEB antiserum, they consistently showed a doublet of 23-24K polypeptides, which were the only ones recognized by the antibody. As addition of EGTA to ICL preparations resulted in the total disappearance of the network in phase-contrast microscopy, we examined whether the immunoreactive components were indeed solubilized. The results are presented in Fig. 8. In Fig. 8A, the electrophoretograms indicate that the 23-24K immunoreactive polypeptides (star) present in the initial fraction (here 'control pellet', lane a), are strongly decreased in the EGTA-treated sample (lane b), while all other polypeptides are still present. In contrast, lanes c and d show that the two polypeptides are the major ones in the supernatant of the EGTAtreated fractions (lane d), while practically nothing is

Fig. 4. Infraciliary lattice labelling by 3D 12. A. In immunofluorescence microscopy the pattern of the network is identical to that observed with the anti-EEB (cf. Fig. 1A). B-E. Immunogold staining. B. In tangential section, the filamentous bundles are organized into tubular structures crossed by other filamentous components (arrows). The triangular shape of posts (double arrows) at the branching locus is well preserved. C. In longitudinal section the tubular arrangement of the filaments is also visible; the gold particles are located along these 'tubules' but also found on microtubular structures such as internal microtubules (iml) or microtubular ribbons (mlr). The labelling of microtubules is significant as far as it is also observed in immunofluorescence microscopy (not shown), ep, Epiplasm; kf, kinetodesmal fibre. D,E. Higher magnification of the filamentous bundles in longitudinal section (D) and cross-section (E). Note the typical tubular organization (D, arrow), the hexagonal organization of the tubules (E) and the peripheral location of the gold particles. A. Bar, X680; B,C, bar, 0-2ftm, X68000; E,F, bar, 0 1 Jim, X 175 000. detected in the control supernatant (lane c). The immunoblot (Fig. 8B) reveals the immunoreactive polypeptides only in the control pellet and in the

EGTA supernatant. Altogether these results suggest that the 23-24K polypeptides are major components of the ICL. Contractile network in Paramecium

359

5A

'Mb

B

Fig. 5. Immunogold staining of an interphase cell using the anti-EEB antibody. A. Tangential section through the cortex showing the specific decoration of the ICL. The other cytoskeletal structures, basal bodies (bb), cilia (r), epiplasm (ep), kinetodesmal fibres (kf) and microtubular ribbons (mtr), do not show any significant labelling. B. An enlargement of the boxed area in A, revealing the helical arrangement of gold particles around the filamentous bundles. The tubular organization observed in Fig. 4B-D is less well preserved but visible (arrow). A. Bar, 0-2|Um, X50000; B, bar, 0-1 //m, X175OOO.

Discussion Immunocytochemical and ultrastructural studies of the infraciliary lattice of Pararneciimi were carried out both in situ and on partly purified preparations of the network. Our results present two major points of interest concerning, respectively, the organization and dynamics of the lattice and its possible contractile function. Organisation and morphogenesis of the ICL While earlier ultrastructural observations derived from classical fixation of whole cells (Pitelka, 1965, 1969; Jurand & Selman, 1969; Allen, 1971; Ehret & MacArdle, 1974) only showed that the ICL was made up of bundles of filaments, our immunoelectron-microscopy studies demonstrate a more complex organization. First, the filaments are organized in adjacent tubules that, in cross-section, appear as closely packed hexagons. This tight organization may explain why immunogold labelling by the antibody 3D 12 is restricted to the periphery of the bundles. Second, the decoration observed by the anti-EEB antiserum follows a peripheral spiral, revealing additional structural components 360

A'. Caireau de Loubresse et al.

that might tie together the fibrils and, as will be discussed below, play a role in the contractility of the network. Altogether, this highly structured network might be mechanically resistant and play a role in maintaining cell shape, as suggested by Sibley & Hanson (1974). This ultrastructural complexity renders the dynamics of the network all the more interesting. At the whole-cell level, in immunofluorescence, the ICL appears during interphase as a continuous meshwork underlying the whole cortex. During division, the precise architecture of the meshes is disorganized over most of the cell surface and then progressively reorganized. This disassembly—reassembly cycle contrasts with the behaviour of other cortical filamentous structures in Paramecium and in particular with that of the other continuous network, the outer lattice, which grows during division without disruption of the preexisting meshes, by longitudinal elongation (Cohen et al. 1987). Our data provide some information: first, on the geography of the disorganization - that is, on its spatiotemporal pattern at the whole-cell level - and second, on the mechanisms of its reorganization. As is well shown at a late stage of division

o Fig. 6. Partial purification of the infraciliary lattice. A,B. Phase-contrast images of parts of ICL released by cell lysis under low ionic strength conditions (see Materials and methods). A. A contracted fragment of the network (arrow) is released from the lysed cell but still attached to one of its poles. B. At higher magnification, a large piece of ICL is seen to retain its polygonal organization. C-E. Thin sections of semi-purified ICL preparations. C. This low-magnification view shows the average size and shape of the ICL fragments in the preparation. D. In this enlargement of the area boxed in A, the organization of the ICL bundles displays a 'periodic' structure suggesting a hehcoidal contraction. E. The extent of contamination by cortical structures still bound to the ICL is apparent. The partly tangential section confirms that the general organization of the network is preserved and shows a reduced mean mesh size (see the text and also Fig. 7). A. Bar, S^m, X1380; B, bar, Sftrn, X21S0; C, bar, 1/