ATP and CaImoduIin Dependent Actomyosin Aggregates Induced by Cytochalasin D in Goldfish Retinal Ganglion Cell Axons In Vitro Brian T. Edmonds and Edward Koenig* Department of Physiology, State University of New York at Buffalo, Buffalo, New York 14214, USA
SUMMARY Growing retinal ganglion cell (RGC) axons of the goldfish have mobile varicosities, which play a role in rapid bulk redistribution of axoplasm (Koenig, Kinsman, Repasky, and Sultz, 1985; Edmonds and Koenig, 1987). Varicosities contain a tubulo-vesicular SER embedded in an actin-containing cytomatrix (Koenig et al., 1985). Cytochalasin D (CD) induces the formation of focal cytoskeletal aggregates throughout preterminal axons and especially in varicosities. The aggregates are visible when labelled with fluoroscein isothiocyanate (F1TC)conjugated phalloidin. Double-labelling experiments show that Texas red-myosin or rhodamine isothiocyanate (RITC) -calmodulin immunofluorescence co-local-
izes with FITC-phalloidin-labelled aggregates. Formation of aggregates is blocked by calmidazolium, a calmodulin antagonist. A4xonmodels permeabilized with digitonin retain the capacity to form focal aggregates in response to CD, when ATP or adenosine-5’-O( 3-thiotriphosphate) (ATP-$3) is present in the permeabilization buffer, but not when 5’-adenylylimidodiphosphate (AMP-PNP) is present. The latter result indicates that formation of focal aggregates depends on ATP. The findings suggest that the formation of focal aggregates in immature axons is a manifestation of actomyosin interactions after free actin-filament ends are generated by CD treatment.
INTRODUCTION
teractions. At the electron microscopic level, varicosities contain an extensive array of tubulo-vesicular SER associated with a cytomatrix (Koenig et al., 1985). Varicosities of RGC axons are unusual in at least two respects. First, some of them are mobile and undergo rapid bidirectional translocation (Koenig et al., 1985; Edmonds and Koenig, 1987). Second, when participating in bulk redistribution of axoplasm, varicosities exhibit a striking structural plasticity in which they can increase in size by fusion with smaller mobile varicosities and nonprotruding phase-dense inclusions (PDIs), or they can decrease in size by giving off PDIs through fission. Varicosities and PDIs along the axon are referred to here collectively as axoplasmic varicose aggregates ( AVAs). Similar mobile AVAs have been reported in dorsal root and sympathetic ganglion cell axons of the chick (Hollenbeck and Bray, 1987). The present report is concerned with the formation of small discrete cytoskeletal foci in response to treatment with cytochalasin D (CD).
Although myosin is present in growing axons (Roisen, Inczedy-Marcsek, Hsu, and Yorke, 1978; Kuczmarski and Rosenbaum, 1979; Letourneau, 1981; Bridgman and Dailey, 1989), and presumably plays a role in tension development (Bray and White, 1988; Mitchison and Kirschner, 1988) related to motility of the growth cone (Smith, 1988): evidence for an active actomyosin contractile system in the growing axon is lacking. Retinal ganglion cell axons (RGC) of goldfish retinal explants (Landreth and Agranoff, 1976) exhibit randomly distributed varicosities. Because varicosities contain myosin (E. Koenig, unpublished), actin, and calmodulin (Koenig et al., 1985), they offer the possibility of probing for potential actomyosin inReceived December 12, 1989; accepted March 5, 1990 Journal of Neurobiology, Vol. 21, No. 4, pp. 555-566 (1990) 0 1990 John Wiley & Sons, Inc. CCC 0022-3034/90/O4055 5- 12$04.00 * To whom correspondence should be addressed.
555
556
Edmonds and Koenig
These focal cytoskeletal aggregates contain predominantly actin, myosin, and calmodulin, and are especially prevalent in varicosities of RGC axons. Their formation is inhibited in the intact preparation by a calmodulin inhibitor and in a permeabilized axon model by 5 '-adenylylimidodiphosphate (AMP-PNP), a nonhydroyzable AT'P analog. These findings suggest that the formation of focal aggregates may depend on an active actomyosin contractile mechanism analogous to that in smooth muscle (Sellers and Adelstein, 1987).
MATERIALS AND METHODS Biochemicals The following were purchased from Sigma: cytochalasin D, poly-L-lysine, 5-fluorodeoxyuridine. gentamycin sulfate, uridine, methyl cellulose, D-gluconic acid, vanadate-free adensoine-5 '-triphosphate ( ATP) , 5 '-adenylylimidodiphosphate ( AMP-PNP), digitonin, A23 187. and ionomycin. Adenosine-5'-0( 3-thiotriphosphate) ( ATP-yS), and calmidazolium were purchased from Boehringer Mannheim. The acetoxymcthyl ester of 1,2bis( O-aminophenoxy) ethane-N,N,-","-tetraacetic acid (BAPTA-AM) was from Molecular Probes.
Retinal Explant Preparation Goldfish retinal explants were prepared as described elsewhere (Koenig and Adams, 1982; Koenig et al., 1985) and were used for experimental observations after 3-5 days in culture. Briefly, 2-4 weeks after crushing the optic nerve, the retina was isolated, chopped into squares (0.65 X 0.65 m m ) , plated out onto polylysinecoated no. 1.5 circular coverslips, and cultured in L- 15 (Gibco) medium, supplemented with 10% fetal calf serum (Flow), 0.02 M Hepes, 0.1 m M 5-fluorodeoxyuridine, 0.1 mg/mL gentamycin sulfate, 0.2 m M uridine. and 0.6% methyl cellulose. Explants were cultured in humid air atmosphere at 27°C.
Video Microscopy For viewing, the circular coverslip containing retinal explants either was mounted in a Dvorak-Stotler chamber (Nicolson Precision Instruments), or was inverted over a 35 X 50 mm no. 2 coverslip supported by 0.5-1 m m thick spacers polymerized and trimmed from Silastic medical elastomer (Dow). Although both chambers permit the total exchange of bathing medium within 1 min, the latter allowed for more uniform Kohler illumination of axonal fields. The standard bathing medium was a modified Cortland physiological fish saline (Koenig and Adams, 1982)composed of(in m M ) : 132, NaCI; 5 , KCI; 1.6, MgCI2; 1.8, CaCl,; 5.5, glucose; 20 Hepes. adjusted to pH 7.2 with Tris. All experiments were conducted at 22-24°C. Axonal fields were viewed
under phase-contrast microscopy (Olympus BHS microscope) with a X 100 oil immersion planapochromat objective (Zeiss N.A. = 1.25) combined with an achromat condenser (Qlympus, N.A. = 1.4) oiled to the bottom coverslip. The microscope stage was isolated from external vibration by a Vibraplane air-suspension table top platform (Kinetic Systems, Inc.). The phase image was displayed on a video monitor (Sanyo) using a DAGE NC-67M video camera with a Newicon tube (DAGE-MTI, Inc.) mounted on a trinocular head of the microscope. Experiments were recorded in a time-lapse mode with a video recorder (TLC 2001, GYYR, Inc.), where time was compressed by a factor of 12. Still photographs were taken from the monitor screen using a Polaroid CU-5 Land camcra type 665 positive/negative Polaroid film. or directly through the microscope with an attached 35 m m photomicrographic system (Olympus PM-1OAD) on Kodak panatomic X film (ASA 32) and developed in D-76.
Fluorescence Microscopy Preparations treated with fluorophore-conjugated reagents were viewed with epi-illumination provided by a 100 W Hg light source. Generally, a X 100 oil immersion planapochromat objective (Zeiss; N.A. = 1.25) was used. The following filter sets (Olympus) were used: fluorescein isothiocyanate (FITC)-conjugated phalloidin; 490 n m band pass excitation filter; secondary antibodies conjugated with rhodamine isothiocyanate (RITC), or Texas Red, 545 nm pass excitation filter, combined with a 580 nm dichroic mirror and 590 nm long band pass barrier filter. Photomicrographs of fluorescent specimens were taken with an attached 35 mm photomicrography system (Qlympus) with either Kodak Tri-X Pan Professional, or T-Max (ASA 400) film, developed in D-76, or T-Max developer, respectively.
Treatment with Cytochalasin D Cytochalasin D was reconstituted in dimethylsulfoxide (DMSO) as a 2 m M stock solution and stored at 20°C in aliquots. Explants were rinsed twice with Cortland saline. and then incubated in CD for 10 min at concentrations indicated in the text. Final DMSO concentration was ~ 0 . 5 % .
Phalloidin- and Immuno-cytochemistry Explants were rinsed in modified PHEM buffer (Schliwa and Blerkom, 1981 ), which consisted of (in m M ) 100, Pipes; 25, Hepes; 15, EGTA; 2, MgC12; pH 6.9 with KOH; and then were exposed for 20 min to a lysis-fixation buffer containing 2% fresh paraformaldehyde, 0.2% glutaraldehyde, 0.5% Triton X-100 in PHEM buffer. For double-labelling experiments, fixation was preceded by lysis for 2 min with 0.15% Triton in PHEM buffer in order to improve the resolution of the phalloidin-stained aggregates. This lysis step appeared to remove much of the soluble immunoreactive
Actomyosin Aggreguies in Growing Axnns calmodulin from axoplasm not associated with the CDinduced aggregates, but did not do so in the case of myosin. After fixation, the explants were washed three times for 15 rnin with TM buffer (in mM: 130, NaCl; 10, Tris HCI; 5, KCl; 5 , MgCI2; 1, EGTA; pH 7.2) containing 1% sodium borohydride, and then washed three times for 10 rnin in TM buffer alone. Thc first wash after fixation contained 50 mMNH4CI and 1% bovine serum albumin (BSA ). Incubation with FITC-conjugated phalloidin ( 1:20; Molecular Probes) was for 30 min. Incubation with primary antisera was for 1 h in TM buffer at the following dilutions: anti-human platelet myosin (1:lOO; a gift from Dr. K. Fujiwara), antismooth muscle actin (1:200; a gift from Dr. U . Groschel-Stewartj , or anti-calmodulin ( 1: 100; CAABCO). Following three 5-min washes in TM buffer, antiseratreated coverslips were then incubated with Texas Redconjugated goat antirabbit IgG ( 1 :2000: E-Y Labs j . or in the case of calmodulin, RITC-conjugated rabbit antisheep IgG ( 1:500; Cappel Labs). Phalloidin- and antiserum-treated coverslips then were rinsed three times for 5 rnin with TM buffer, rinsed in distilled water and mounted in Elvanol (Dupont).
Digitonin Permeabilization of Axonal Fields During permeabilization, it was essential to exchange solutions in the chamber very slowly in order to prevent dispersion of axonal residue. Explants were washed twice with Cortland saline, and then with PHEM buffer. The permeabilization buffer was (in m M ) : 0.00 1% digitonin; 110, Pipes; 25, Hepes; and 2, MgC12: ( p H 6.9 with KOH). After the preliminary washes, the explants were incubated in permeabilization buffer containing in addition 5 mMNaN, and 1 ATP-yS for 30 min, followed by incubation for 30 min in either 1 m M AMP-PNP, 250 pLM Ca2+and 10 p M CD, or 1 ATP, 250 pM Ca2' and 10 pLMCD. This was followed by phalloidin staining (see above)
Determination of Extracellular-Free Calcium Free [ Ca2+]in the bathing medium was determined by the method of Bers ( I982 j , or by a computer program designed to compensate for pH, [Mg2+], and ionic strength (courtesy of Dr. Zahur Ahmed j. In either case, the free-Ca2+ activity was confirmed with a calciumsensitive electrode (Radiometer). In those experiments concerned with depleting intracellular calcium, explants were incubated in Ca2+-free,Cortland saline containing 1 m M EGTA and 5 p M ionomycin, a calcium ionophore (Liu and Hermann, 1978), and the membrane permeant Ca2+chelator, BAPTA-AM (20 gtM) (Tsien, 1980), for 1 h at 27°C prior to permeabilization.
557
RESULTS FITC-conjugated phalloidin, an F-actin specific probe (Barak, Yocum and Webb, 1981), exhibits a strong signal predominantly in growth cones [Fig. 1 (a-d)] , and appears only as scattered. occasional fluorescent streaks and puncta in preterminal RGC axonal fields [Fig. 1 ( f ) ] . Most axonal varicosities exhibit a weak, diffuse and short-lived phalloidin fluorescence, which is difficult to capture photomicrographically. On the other hand, actin immunostaining, as shown in Fig. lg, and h and reported previously (Koenig et al., 1985), is diffuse and strong, in varicosities. Although the more intense immunofluorescence reflects staining of both Gand F-actin, the apparent paucity of phalloidin fluorescence does not mean a lack of F-actin necessarily. Rather, given the amorphous cytomatrix visible in Varicosities at the electron microscopic level (Koenig et al., 1985), it suggests that phalloidin shows up F-actin best when the latter is organized into highly condensed structures, such as actin bundles and dense meshwork found in growth cones, as opposed to actin organized in a loose meshwork, more characteristic of that in varicosities. Differences between phalloidin labelling and immunostaining of actin have been noted also very recently in growing distal sympathetic ganglion cell axons (Bridgman and Dailey. 1989). Ten minutes after exposure of axonal fields to cytochalasin D, the distribution of phalloidin fluorescence is altered markedly. Numerous actincontaining foci appear scattered throughout the length of individual axons [Fig. 2(b,d)]. These focal aggregates stain very strongly with phalloidin. Although focal aggregates appear randomly distributed, regional differences are apparent from inspection because varicosities appear to contain a higher density [Fig. 2 ( d ) ] when compared to intervening axonal segments. Focal aggregates formed in response to CD treatment are not visible in phase-contrast microscopy [Fig. 2 (a,c)] and can be detected only at the light microscope level with fluorescent probes. Axonal fields incubated for 10 rnin in CD ( 1 p M ) , and then allowed to recover for 5 h in a CD-free medium do not exhibit phalloidin-stained foci. Such observations indicate that CD induces F-actin-containing aggregates to form de now and that the process is reversible. Most experiments described here were performed with 10 pA4 CD, which also was found to be reversible; however, foci could be obtained even with 0.1 pM. Up to 1% of DMSO alone did not induce the formation of foci. When viewed with time-lapse, phase-contrast
558
Edmonds and Koenig
Figure 1 Distribution of actin in RGC axons as revealed by either FITC-phalloidin fluorescence or RITC-actin immunofluorescence. (a, c, e ) Phase-contrast, and (b, d, f ) fluorescence micrographs of areas labelled with FITC-conjugated phalloidin. (b, d ) Growth cones exhibit a very strong fluorescence, ( f ) whereas preterminal axons generally show scanty phalloidin fluorescence. Actin immunofluorescence ( 6 ) ofgrowth cones and ( h ) varicosities is strong and distributed diffusely. Bar: 10 pm.
Actomyosin Aggregates in Growing Axons
559
Figure 2 Formation of focal aggregates labelled by FIX-phalloidin. (a, c ) Phase-contrast and (b, d ) corresponding fluorescence micrographs showing discrete FITC-phalloidin labelled focal aggregates scattered throughout preterminal axons following treatment with cytochalasin D ( 10 p M ) . Fluorescent focal aggregates appear in higher density (d) in Varicosities. Note that focal aggregates are not visible in phase-contrast microscopy. Bars: a,b, 10 pm;c,d, 5 pm.
videomicroscopy, CD did not appear to perturb visible-particle transport activity because the bidirectional movements of organelles continued unabated (Edmonds and Koenig, unpublished). Growth cones were the only structures that showed a sensitivity to CD treatment. Generally, all growth-cone motility ceased upon addition of CD, concomitant with the collapse of filopodial organization. The morphological changes of ,growth cones were consistent with those reported lby Marsh and Letourneau ( 19841, but the overall morphology of preterminal axons otherwise remained unchanged. T o investigate whether cytoskeletal components and/or regulatory proteins other than actin might be involved in the formation of focal aggregates, an immunocytochemical survey was undertaken to evaluate myosin and calmodulin distributions in RGC axons after CD treatment. Myosin normally appears in RGC axons as a diffuse immunofluorescence. which is unform and strongest iiri varicosities of untreated axons [Fig. 3 ( b,d)] .
Although myosin immunofluorescence is still somewhat diffuse following exposure to CD, the immunofluorescence in varicosities is nonuniform in distribution [Fig. 3 (g,i)] with the strongest immunofluorescence associated with phalloidinstained aggregates [Fig. 3 (f,i)]. This correlation suggests that myosin becomes redistributed to some extent during the CD treatment to become associated with F-actin-containing foci. Calmodulin, as shown previously (Koenig et al., 1985), is associated with varicosities and growth cones preferentially [Fig. 4 (a,b)] . Following CD treatment, residual calmodulin immunofluorescence [Fig. 4(g,h)] after the lysis/fixation procedure was co-distributed largely with phalloidin-labelled foci [Fig. 4( e, f ) ] . The apparent co-localization of myosin and calmodulin with F-actin-containing foci raises the possibility that the formation of the latter may be mediated by an actom yosin interaction. Because calmodulin antagonists have been shown to interfere with smooth muscle contraction (Kerrick,
560
Edmond.7 and Koenig
Figure 3 Apparent co-distribution of myosin with cytochalasin D-induced focal aggregates. Texas red-myosin immunofluorescence (b, d ) in untreated RGC axons and (g, j) in axons after cytochalasin D treatment, and in those (f, i ) which have been counterstained with FITC-phalloidin also. (a, c, e, h ) Phase-contrast, and (b, d, f-g, i-j ) corresponding fluorescence micrographs. Myosin immunofluorescence is normally diffuse and uniform in distribution, (b, d ) with varicosities exhibiting the strongest fluorescence. After cytochalasin D ( 10 p M ) , (g, j ) myosin immunofluorescence becomes distributed nonuniformly such as to exhibit enhanced labelling of focal aggregates ( f, i). which are labelled also strongly by FITC-phalloidin fluorescence (arrows). Bar: 10 pm.
Actomyosin Aggregates in Growing A.xons
Hoar, and Cassidy, 1980), calmidazolium, a potent calmodulin inhibitor (Gietzen, Wurthrich, and Bader, 1982; Van Belle, 1981), was used to test this possibility. Initial experiments showed that calmidazolium treatment of RGC axons caused a rapid and irreversible solation of the cytomatrix in the growth cone and varicosities. in which visible organelles appeared in random motion. Experiments later indicated that the solation may have been caused by a chloride-mediated osmotic swelling because removal of chloride from the bathing medium eliminated it (B. Edmonds, unpublished). When chloride was replaced on a molar basis by gluconic acid in the Cortland medium, both morphology and transport activity of axons were normal in the modified medium. Pretreatment of axonal fields for 30 min with calmidazolium ( 1 p M ) greatly reduced the number of aggregates that formed following subsequent treatment with CD [Fig. 5 ( d ) 1, while calmidazolium treatment after CD did not [Fig. 5 (b)]. Calmidazolium alone, at this concentration, had no detectable effect on axonal morpholog) or organelle transport. These results suggest that calmodulin may be important in the formation of focal cytoskeletal aggregates. As skinned muscle fibers (Cassidy, Hoar, and Kemck, 1979; Hoar, Kemck, and Cassidy, 1979; Kenick, Hoar, Cassidy, Bolles, and Haleucik, I98 1 ) and some permeabilized cell models (Holzapfcl et al., 1983; Kreis and Birchmeier, 1980; Masuda, Owaribe, Hayashi, and Hatano, 1983) havc the capacity to contract when exposed to suitable concentration of Ca2’ and ATP, a potential actomyosin involvement in the formation of CD-induced focal aggregates was probed in a permeabilized axonal model. Although immature RGC axons are structurally labile, axonal fields treated with 0.001% digitonin in a modified PHLM buffer (see Methods), yielded a reasonably stabilized preparation that stained with phalloidin after exposure for I min, indicating that the permeability barrier to the peptide was breached by this treatment. When CD was introduced during permeabilization, or a few minutes after, aggregate formation occurred in the absence of added ATP and Ca2+. After axons were maintained in permeabilization buffer for 30 min in the absence of ATP, CD failed to induce the formation of aggregates consistently, even if ATP and Ca2+were included. However, the inclusion of 1 m M ATP-TS, a thiophosphate ATP analog, in the permeabilization buffer significantly improved reproducibility of the Subsequent CD-induced formation of aggregates. It is possible
561
that this may have been due to a stable activation of myosin by thiophosphorylation of myosin (see Discussion ) . In experiments testing for ATP dependence of aggregate formation, 5 m M NaN3 was included in an ATP-yS-containing permeabilization buffer. After 30 min in this medium. CD was added with 1 m M ATP, resulting in the formation of aggregates [Fig. 6 (b)];however, CD added with 1 m M AMP-PNP, a nonhydrolyzable ATP analog, did not result in the formation of aggregates [Fig. 6(d)]. These experiments indicated that CD-induced aggregate formation was dependent on available ATP. Smooth muscle contraction (Sellers and Adelstein, 1987) and contraction of some permeabilized cell models (Holzapfel et al., 1983; Masuda et al., 1984; and Masuda. Owaribe, and Hatano, 1983) appear to have a requirement for calcium. A calcium dependency in CD-induced formation of aggregates was tested by varying [Ca”],. Permeabilization buffers with known free- [ Ca” ] , ranging from 0.1 pM to 250 pM were prepared (see Methods). In addition, some axonal fields were incubated for 1 h prior to permeabilization with BAPTA-AM ( 10 p M ), and ionomycin ( 5 p M ) , a calcium ionophore, in a Ca2+-free/EGTA-buffered medium to deplete [ Ca2+Il(see Methods). CD-induced aggregate formation was not affected by any of the conditions designed to modify [ Ca2+],including those involving a presumptive depletion of intracellular calcium. DISCUSSION
When goldfish axons regenerating in vitro are treated with cytochalasin D, numerous discrcte focal aggregates form, especially in varicosities, which stain strongly with phalloidin. As CD is known to interfere with actin polymerization (see below), it seems likely that CD induces a reorganization /condensation of preformed F-actin in varicosities, notwithstanding the fact that the latter normally stain very weakly with phalloidin. One possible explanation for weak phalloidin staining in varicosities may relate to a looser actin-based cytoskelctal meshwork compared to that present in growth cones (Bridgman and Dailey, 1989). Actin aggregate formation in response to cytochalasin treatment has been reported in plants (Palevitz. 1988), and algae (Menzel and Schliwa, 1986), as well as in animal cells (Bitch and Allen, 1981; Godman, Woda, and Kolberg, 1980;
562
Edinonh und Koenig
Figure 4 Apparent co-distribution of calmodulin with cytochalasin D-induced focal aggregates. Distribution of RITC-calmodulin immunofluorescence (b) in untreated RGC axons and (g, h) in axons after cytochalasin D treatment and in those (e, f ) which have been
Actomyosin Aggregates in Growing Axons
563
Figure 5 Blockade of cytochalsin-D-induced formation of focal aggregates after antagonism of calmodulin. ( a ) Phase-contrast and (b) corresponding FITC-phalloidin fluorescence micrographs of an axonal fascicle pretreated with cytochalasin D (10 p M ) and followed by calmidazolium. (c) Phase-contrast and ( d ) FITC-phalloidin fluoresccnce micrographs of axonal fascicles pretreated for 30 rnin with calmidazolium ( 1 pm) and followed with cytochalasin D (10 phi'). Bar: 10 pm.
Schliwa. 1982; Weber. Rathke. Osborn. and Franke, 1976), including neurons (Marsh and Letourneau, 1984; Letourneau and Shattuck, 1989). Typically, cytochalasin B (CB) or CD was used in the latter studies. Although cytochalasins, as a group, bind actin and interfere with its polymerization, some are more specific. For example, CD, which does not affect membrane glucose transport, is more specific for actin than is CB, which does. and its efficacy in inhibiting actin polymerization in vitru is about 10 times greater (Cooper, 1987). The process of aggregate formation induced by CD in RGC axons appears to require ATP as indi-
cated by experiments in which aggregate formation was blocked in digitonin-permeabilized axon models by AMP-PNP, a nonhydrolyzable ATP analog. These results are in agreement with those reported in other systems (Miranda, Godman, Deitch, and Tannenbaum, 1974; Godman et al., 1980; Schliwa, 1982). In addition to actin, there is also a redistribution of myosin and calmodulin in response to CD treatment, such that these proteins appear to co-localize with phalloidin-stained aggregates. The co-localization of actin. myosin, and calmodulin in aggregates suggests that CD-induced for-
counterstained with FITC-phalloidin also. (a, c, d ) Phase-contrast and (b, e-h) corresponding fluorescence micrographs. Typically, ( b1 RITC-calmodulin immunofluorescence is associated largely with varicosities and growth cones where it is distributed diffuscly. After treatment with cytochalasin D ( 10 pLM) and washout during the lysis/fixation procedure (see Methods), ( g , h ) the calmodulin immunofluorescence is focal and ( e , f ) co-distributed with FITC-phalloidin fluorescence. Bar: 10 pm.
564
Edmnnds and Koenig
Figure 6 Formation of focal aggregates in permeabilized RGC axons in the presence of ATP but not in presence of AMP-PNP. Axons were permeabilixd with diptonin (0.001%) in PHEM buffer (see Methods). ( a ) Phase-contrast and ( b ) FITC-phalloidin fluorescence micrographs after treatment with cytochalasin D (10 p M ) , ATP ( I mhf),and calcium ( 10 &if). (c) Phase-contrast and ( d ) FITC-phalloidinfluorescence micrographs after treatment with cytochalasin D (10 p M ) , AMP-PNP ( 1 mM), and calcium ( 10 pLhf). Bar: 10 pm.
mation of the latter may reflect an active contractile event. The requirement for ATP also supports this inference. In addition, the inhibition by calmidazolium, a potent calmodulin inhibitor (Van Belle, 198 1 ) and the improved reproducibility when ATP-yS is included during permeabilization of axons are also consistent with the possibility of a calmodulin-dependent activation of myosin lightchain kinase (MLCK) anteceding the formation of aggregates. MLCK exhibits a similar diffuse distribution in varicosities as that of myosin in RGC axons (B. T. Edmonds, unpublished). Calmodulin antagonists block contraction in slunned smooth muscle fibers (Kerrick et al., 1 98 1 ) and in permeabilized Swiss 3T3 cells (Holzapfel et al., 1983). Moreover, contraction in the latter cells is inhibited by MLCK antibodies (Holzapfel et al., 1983). MLCK in gizzard muscle (Sherry, Gorecka, Aksoy, Dabrowska, and Hartshorne, 1978) and in rabbit smooth muscle (Cassidy et al., 1979) can utilize ATP-yS to thiophosphorylate the myosin light chain, which effectively maintains myosin in an activated state, as dethtophosphorylation is not cata-
lyzed readily by the phosphatase (Cassidy et al., 1979). In some examples of permeabilized cell models, contractions exhibit a strict requirement for calcium (Hoar et al., 1979; Holzapfel et al., 1983; Kemck et al., 1980: Masuda et al., 1983). CD-induced aggregate formation in RGC axons occurred with all calcium concentrations tested, including conditions designed to deplete intracellular calcium. Although depletion of intracellular Ca2+could not be verified, EGTA ( > 1 m M ) in the permeabilizing buffer in combination with the intracellular-Ca2+ chelator, BAPTA-AM, should have reduced free [Ca”], to a very low level. Thesc rcsults are not consistent with experiments in which blockade was produced by calmodulin antagonism. One possible explanation is that the contractile machinery is already in an activated state after pcrmeabilization. There is some evidence that the method of pcrmeabihzdtion can affect the degree of calcium sensitivity of contraction in cell models. For example, whereas mouse 3T3 cells undergo contractions with calcium concen-
Artomyosin Aggregates in Growing Axons
trations > 1 pM after Triton treatment, contractions in glycerinated 3T3 cells exhibit no calcium dependency (Masuda et al.. 1983). Electronmicroscopic inspection of RGC axons reveals that varicosities contain abundant arrays of tubulo-vesicular profiles, embedded in an amorphous cytomatrix (Koenig and Adams, 1982; Koenig et al., 1985). A weak phalloidin signal is consistent with the view that actin is organized probably as a loose meshwork (see above). Myosin immunofluorescence is distributed diffusely within varicosities. Whether actomyosin interactions may contribute normally to tension of the cytomatrix in varicositics is not known, but it is a possibility. In this context, it is noteworthy that varicosities of RGC axons, in a calcium-free Cortland bathing medium containing the calcium ionophore A23 187, undergo a reduction in volume of about 20% when [Ca2+], is changed from 0 to nominally 0.0 I p A 4 (E. Edmonds, unpublished), The reduction in volume suggests that a contraction may have occurred, although a loss of volume through a calcium-dependent efflux of K (Edmonds and Koenig, 1990) cannot be ruled out. The mechanisms underlying the formation of aggregates in response to CD treatment are uncertain. The evidence suggests that CD caps the fast growing, barbed ends of actin filaments, and that it can compete with capping proteins (Cooper, 1987). As pointed out by Cooper (1987), CD could displace capping proteins that hold actin filaments in place, and thereby generate free filament ends. If there is a basal level of isometric tension mediated by actomyosin interactions within the cytomatrix of varicosities, then the production of "free" actin filament ends by CD would generate isotonic contractions that would appear as focal cytoskeletal condensations (i.e.. aggregates). This view is consistent with a model proposed by Schliwa and van Rlerkom ( 198 1 ) in which cytochalasin-induced aggregates form in a manner similar to contractions in smooth muscle +
We thank Mrs. Sarah Finnegan Sloan for her technical assistance in certain aspects of thc work. This research was supported in part by grants NS21843 from NTNCDS and BRSG SO7 RK05600-26 awarded by the Biomedical Research Support and Grant Program, Division of Research Resources. NTH.
REFERENCES BARAK,L. s.,YOCUM,R. R., and WEBB,W. W. ( 198 1 ). In vivo staining of cytoskcletal actin by autointernalization of nontoxic concentrations of nitrobenzoxadiazole-phallacidin. J . Cell Biol. 89:368-372.
565
BERS.D. M. (1982). A simple method for the accurate determination of free [ Ca2' ] in Ca-EGTA solutions. Am. .I. Physiol. (Cdl Physiol. 1I ) 242:C404-C408. BRAY,D., and WIIITE.J . G. (1988). Cortical flow in animal cells. Science 239983-888. BRIDGME.N. P. C., and DAILEY, M. E. ( 1989). The organization of myosin and actin in rapid frozen nerve growth cones. J. Cell Biol. 108:95-109. BKITCH,M., and ALLEN,T. D. (1981). The effects of cytochalasin B on the cytoplasmic contractile network revealed by whole-cell transmission electron microscopy. L k p . Cell Res. 131:161-172. CASSIDY,P., H0.4R, P. E.. and KERRICK.W. G . L. ( 1979). Irreversible thiophosphorylation and activation of tension in functionally skinned rabbit ileum strips by 3'S-ATP-yS. J. Biol. Ckem. 254:l 114811153. COOPER, J. A. (1987). Effects of cytochalasin and phalloidin on actin. J. Cell Bid. 1051473-1478. EDMONDS,B. T., and KOENIG,E. ( 1087). Powering of bulk transport (varicosities j and differential sensitivities of directional transport in growing axons. Bruin Rex 406:288-293. EDMONDS,€3. T., and KOENIG, E. (1990). Calcium-dependent volume reduction in regenerating ganglion cell axons in v i m . J. hieurosci. Rex In press. GIETZEN, K., WUKTHRICH. A.. and BADER,H. ( 1982). R24571: a new powerful inhibitor of red blood cell Ca++-transport ATPase and of calmodulin-regulated functions. Hiochem. Eliophys. Re.y. Comm. 101:418425. GODMAN. G., WODA, B., and KOLBERG,R. (1980). Redistribution of contractile and cytoskeletal components induced by cytochalasin. I. In Hmf cells, a nontransformed fibroblastoid line. Eur. J . Cell B i d . 22:733-744. HOLLCNHECK. P. J., and BRAY, D. (1987). Rapidly transported organelles containing membrane and cytoskeletal components: their relation to axonal growth. .I. Cell Biol. 105:2827-2835. HOAR,P. E., KEKRICK, W. G. L., and CASSIDY,P. S. ( 1979). Chicken gizzard: relation between calciumactivated phosphorylation and contraction. Science 2041503-506. HOLZAPFEL, G., WEHLAND, J., and WEBER,K. ( 1983j. Calcium control of actin-myosin based contraction in Triton models of mouse 3T3 fibroblasts is mediated by myosin light chain kinase (MLCK)-calmodulin complex. Exp. Cell Rcs. 148: I 17- 126. KERRICK, W. G. L., HOAR,P. E. and CASSIDY,P. S. ( 1980). Calcium-activated tension: the role of myosin light chain phosphorylation. Fed. Proc. 39: 15581563. KERRICK,W. G. L., HOAR, P. E.. CASSIDY.P. S., BOLI,ES, L., and HALEUCIK, D. A. (1981). Calciumregulatory mechanisms: functional classification using skinned fibers. J. Gen. Ph,ysiol. 77:177-190. KOENIG.E., and ADAMS,P. (1982). Local protein syn-
566
Edmonds and Kornig
thesizing activity in axonal fields regenerating in vitro. J. Neurochem. 39:386-400. KOENIG,E., KINSMAN, S., RFPASKY,E., and SuLiz, L. (1985). Rapid motility of motile varicosities and inclusions containing spectrin, actin and calmodulin in regenerating axons in vitro. J. Neuro.sci. 5 7 15-729. W. ( 1980). Studies on the KREIS,T., and BIRCHMEIER, organization and localization of actin and myosin in neurons. J. Cell Biol. 8:356-37 I . KUCZMARSKI, E. R., and ROSENBAUM, J. L. (1979). Studies on the organization and localization of actin and myosin in neurons. J . Cell Biol. 80:356-37 I . LANDRETH, G. E.. and AGRANOFF,B. W. (1976). Explant culture of adult goldfish retina: effect of prior optic nerve crush. Brain Rex 118:299-303. LETOURNEAU, P. C. ( 1981 ), Immunocytochemical evidence for colocalization in neurite growth cones of actin and myosin and their relationship to cell-substratum adhesions. Dev. B i d 85: 1 13- 122. LETOURNEAU, P. C., and SHATTLJCK, T. A. (1989). Distribution and possible interactions of actin-associated proteins and cell adhesion molecules of nerve growth cones. Development 105505-5 19. LIU,C., and HERMANN, T. E. ( 1978). Characterization of ionomycin as a calcium ionophore. J . Biol. C'hem. 25335892-5894. MARSH,L., and LETOURNEAU, P. C. ( 1984). Growth of neuntes without filopodidl or lamcllipodial activity in the presence of cytochalasin B. J. Cell B i d 99:20412047. MASIDA, H.. OWARIBE. K.: and HATANO,s. (1983). Contraction of Triton-treated culture cells. A calcium-sensitive contractile model. E.q. Cell Res. 143:79-90. MASVDA, H., OWARIBE, HAYASHI, H.; and HATANO,S. ( 1983). Ca++-dependentcontraction of human lung fibroblasts treated with Triton X-100: a role ofCa++calmodulin dependent phosphorylation of myosin 20.000 Dalton light chain. Cell Motil. 4:3 15-33 I . MENZEL,D., and SCHLIWA, M. (1986). Motility in the siphonaceius green alga Bryopsit IT. Chlorplast movement requires organized arrays of both microtubules and actin filaments. Eur. J . Cell Biol. 40:286-295.
MIRAYDA,A., GODMAN,G., DEITCH,A., and TANNENBAUM, S . (1974). Action of cytochalasin D on cells of established lines, 1. Early events. J. Cell Biol. 61:481-500. MITCHISON, T., and KIRSCHNER, M. ( 1988). Cytoskeleta1 dynamics and nerve growth. IVeuron 1:76 1-772. B. A. ( 1988). Cytochalasin-induced reorgaPALEVITZ, nization of actin in Allum root cells. CMc'9:283-298. M., HXJ, L., and ROISEN,F., INCZEDY-MARCSEK, YORKE,W. (1978). Myosin: immunofluorescent localization in neuronal and glial cultures. Science 199:1445-1448. SCHLIWA. M. ( 1982). Action ofcytochalasin D on cytoskeletal networks. J. CPll Bid. 92:79-9 I. SCHLIWA, M., and V A BLEKKOM, ~ J. ( 1981 ). Structural interaction of cytoskeletal components. J. Cell BioE. 90:222-235. SELIERS. J. R., and ADELSI'EI~, R. S. (1987). Regulation of contractile activity. In: The Enzymes. Vol. XVIII B. P. D. Buyer and E. G. Krebs, Eds.. Academic Press, New York, pp. 38 1-4 18. SHERRY. J. M. F., GORECKA, A.?AKSOY,M. O., DABROWSKA, K.,and HAKTSHOKNE, D. J. ( 1978). Roles of calcium and phosphorylation in regulation of the activity of guzard myosin. Biochem. 17:44 1 1-44 18. SMITH. S. J. (1988). Neuronal cytomechanics: the actin-based motility of' growth cones. Science 242708-7 1 5 . TSIEN,R . Y. (1980). New calcium indicators and buffers with high selectivity against magnesium and protons: design. synthesis, and properties of prototype structures. Biochern. I9:2396-2404. VANBELLE,H. (1981). R24571: a potent inhibitor of calmodulin-activated enzymes. Cell Culciurn 2:483494. P. C., OSBORN, M., and FRANKE. WEBER,K., RATHKE, W. W. (1976). Distribution of actin and tubulin in cells and in glycerinated cell models after treatment with cytochalasin B (CB). Exp. Cell Rex 102:285297.
SUMMARY Growing retinal ganglion cell (RGC) axons of the goldfish have mobile varicosities, which play a role in rapid bulk redistribution of axoplasm (Koenig, Kinsman, Repasky, and Sultz, 1985; Edmonds and Koenig, 1987). Varicosities contain a tubulo-vesicular SER embedded in an actin-containing cytomatrix (Koenig et al., 1985). Cytochalasin D (CD) induces the formation of focal cytoskeletal aggregates throughout preterminal axons and especially in varicosities. The aggregates are visible when labelled with fluoroscein isothiocyanate (F1TC)conjugated phalloidin. Double-labelling experiments show that Texas red-myosin or rhodamine isothiocyanate (RITC) -calmodulin immunofluorescence co-local-
izes with FITC-phalloidin-labelled aggregates. Formation of aggregates is blocked by calmidazolium, a calmodulin antagonist. A4xonmodels permeabilized with digitonin retain the capacity to form focal aggregates in response to CD, when ATP or adenosine-5’-O( 3-thiotriphosphate) (ATP-$3) is present in the permeabilization buffer, but not when 5’-adenylylimidodiphosphate (AMP-PNP) is present. The latter result indicates that formation of focal aggregates depends on ATP. The findings suggest that the formation of focal aggregates in immature axons is a manifestation of actomyosin interactions after free actin-filament ends are generated by CD treatment.
INTRODUCTION
teractions. At the electron microscopic level, varicosities contain an extensive array of tubulo-vesicular SER associated with a cytomatrix (Koenig et al., 1985). Varicosities of RGC axons are unusual in at least two respects. First, some of them are mobile and undergo rapid bidirectional translocation (Koenig et al., 1985; Edmonds and Koenig, 1987). Second, when participating in bulk redistribution of axoplasm, varicosities exhibit a striking structural plasticity in which they can increase in size by fusion with smaller mobile varicosities and nonprotruding phase-dense inclusions (PDIs), or they can decrease in size by giving off PDIs through fission. Varicosities and PDIs along the axon are referred to here collectively as axoplasmic varicose aggregates ( AVAs). Similar mobile AVAs have been reported in dorsal root and sympathetic ganglion cell axons of the chick (Hollenbeck and Bray, 1987). The present report is concerned with the formation of small discrete cytoskeletal foci in response to treatment with cytochalasin D (CD).
Although myosin is present in growing axons (Roisen, Inczedy-Marcsek, Hsu, and Yorke, 1978; Kuczmarski and Rosenbaum, 1979; Letourneau, 1981; Bridgman and Dailey, 1989), and presumably plays a role in tension development (Bray and White, 1988; Mitchison and Kirschner, 1988) related to motility of the growth cone (Smith, 1988): evidence for an active actomyosin contractile system in the growing axon is lacking. Retinal ganglion cell axons (RGC) of goldfish retinal explants (Landreth and Agranoff, 1976) exhibit randomly distributed varicosities. Because varicosities contain myosin (E. Koenig, unpublished), actin, and calmodulin (Koenig et al., 1985), they offer the possibility of probing for potential actomyosin inReceived December 12, 1989; accepted March 5, 1990 Journal of Neurobiology, Vol. 21, No. 4, pp. 555-566 (1990) 0 1990 John Wiley & Sons, Inc. CCC 0022-3034/90/O4055 5- 12$04.00 * To whom correspondence should be addressed.
555
556
Edmonds and Koenig
These focal cytoskeletal aggregates contain predominantly actin, myosin, and calmodulin, and are especially prevalent in varicosities of RGC axons. Their formation is inhibited in the intact preparation by a calmodulin inhibitor and in a permeabilized axon model by 5 '-adenylylimidodiphosphate (AMP-PNP), a nonhydroyzable AT'P analog. These findings suggest that the formation of focal aggregates may depend on an active actomyosin contractile mechanism analogous to that in smooth muscle (Sellers and Adelstein, 1987).
MATERIALS AND METHODS Biochemicals The following were purchased from Sigma: cytochalasin D, poly-L-lysine, 5-fluorodeoxyuridine. gentamycin sulfate, uridine, methyl cellulose, D-gluconic acid, vanadate-free adensoine-5 '-triphosphate ( ATP) , 5 '-adenylylimidodiphosphate ( AMP-PNP), digitonin, A23 187. and ionomycin. Adenosine-5'-0( 3-thiotriphosphate) ( ATP-yS), and calmidazolium were purchased from Boehringer Mannheim. The acetoxymcthyl ester of 1,2bis( O-aminophenoxy) ethane-N,N,-","-tetraacetic acid (BAPTA-AM) was from Molecular Probes.
Retinal Explant Preparation Goldfish retinal explants were prepared as described elsewhere (Koenig and Adams, 1982; Koenig et al., 1985) and were used for experimental observations after 3-5 days in culture. Briefly, 2-4 weeks after crushing the optic nerve, the retina was isolated, chopped into squares (0.65 X 0.65 m m ) , plated out onto polylysinecoated no. 1.5 circular coverslips, and cultured in L- 15 (Gibco) medium, supplemented with 10% fetal calf serum (Flow), 0.02 M Hepes, 0.1 m M 5-fluorodeoxyuridine, 0.1 mg/mL gentamycin sulfate, 0.2 m M uridine. and 0.6% methyl cellulose. Explants were cultured in humid air atmosphere at 27°C.
Video Microscopy For viewing, the circular coverslip containing retinal explants either was mounted in a Dvorak-Stotler chamber (Nicolson Precision Instruments), or was inverted over a 35 X 50 mm no. 2 coverslip supported by 0.5-1 m m thick spacers polymerized and trimmed from Silastic medical elastomer (Dow). Although both chambers permit the total exchange of bathing medium within 1 min, the latter allowed for more uniform Kohler illumination of axonal fields. The standard bathing medium was a modified Cortland physiological fish saline (Koenig and Adams, 1982)composed of(in m M ) : 132, NaCI; 5 , KCI; 1.6, MgCI2; 1.8, CaCl,; 5.5, glucose; 20 Hepes. adjusted to pH 7.2 with Tris. All experiments were conducted at 22-24°C. Axonal fields were viewed
under phase-contrast microscopy (Olympus BHS microscope) with a X 100 oil immersion planapochromat objective (Zeiss N.A. = 1.25) combined with an achromat condenser (Qlympus, N.A. = 1.4) oiled to the bottom coverslip. The microscope stage was isolated from external vibration by a Vibraplane air-suspension table top platform (Kinetic Systems, Inc.). The phase image was displayed on a video monitor (Sanyo) using a DAGE NC-67M video camera with a Newicon tube (DAGE-MTI, Inc.) mounted on a trinocular head of the microscope. Experiments were recorded in a time-lapse mode with a video recorder (TLC 2001, GYYR, Inc.), where time was compressed by a factor of 12. Still photographs were taken from the monitor screen using a Polaroid CU-5 Land camcra type 665 positive/negative Polaroid film. or directly through the microscope with an attached 35 m m photomicrographic system (Olympus PM-1OAD) on Kodak panatomic X film (ASA 32) and developed in D-76.
Fluorescence Microscopy Preparations treated with fluorophore-conjugated reagents were viewed with epi-illumination provided by a 100 W Hg light source. Generally, a X 100 oil immersion planapochromat objective (Zeiss; N.A. = 1.25) was used. The following filter sets (Olympus) were used: fluorescein isothiocyanate (FITC)-conjugated phalloidin; 490 n m band pass excitation filter; secondary antibodies conjugated with rhodamine isothiocyanate (RITC), or Texas Red, 545 nm pass excitation filter, combined with a 580 nm dichroic mirror and 590 nm long band pass barrier filter. Photomicrographs of fluorescent specimens were taken with an attached 35 mm photomicrography system (Qlympus) with either Kodak Tri-X Pan Professional, or T-Max (ASA 400) film, developed in D-76, or T-Max developer, respectively.
Treatment with Cytochalasin D Cytochalasin D was reconstituted in dimethylsulfoxide (DMSO) as a 2 m M stock solution and stored at 20°C in aliquots. Explants were rinsed twice with Cortland saline. and then incubated in CD for 10 min at concentrations indicated in the text. Final DMSO concentration was ~ 0 . 5 % .
Phalloidin- and Immuno-cytochemistry Explants were rinsed in modified PHEM buffer (Schliwa and Blerkom, 1981 ), which consisted of (in m M ) 100, Pipes; 25, Hepes; 15, EGTA; 2, MgC12; pH 6.9 with KOH; and then were exposed for 20 min to a lysis-fixation buffer containing 2% fresh paraformaldehyde, 0.2% glutaraldehyde, 0.5% Triton X-100 in PHEM buffer. For double-labelling experiments, fixation was preceded by lysis for 2 min with 0.15% Triton in PHEM buffer in order to improve the resolution of the phalloidin-stained aggregates. This lysis step appeared to remove much of the soluble immunoreactive
Actomyosin Aggreguies in Growing Axnns calmodulin from axoplasm not associated with the CDinduced aggregates, but did not do so in the case of myosin. After fixation, the explants were washed three times for 15 rnin with TM buffer (in mM: 130, NaCl; 10, Tris HCI; 5, KCl; 5 , MgCI2; 1, EGTA; pH 7.2) containing 1% sodium borohydride, and then washed three times for 10 rnin in TM buffer alone. Thc first wash after fixation contained 50 mMNH4CI and 1% bovine serum albumin (BSA ). Incubation with FITC-conjugated phalloidin ( 1:20; Molecular Probes) was for 30 min. Incubation with primary antisera was for 1 h in TM buffer at the following dilutions: anti-human platelet myosin (1:lOO; a gift from Dr. K. Fujiwara), antismooth muscle actin (1:200; a gift from Dr. U . Groschel-Stewartj , or anti-calmodulin ( 1: 100; CAABCO). Following three 5-min washes in TM buffer, antiseratreated coverslips were then incubated with Texas Redconjugated goat antirabbit IgG ( 1 :2000: E-Y Labs j . or in the case of calmodulin, RITC-conjugated rabbit antisheep IgG ( 1:500; Cappel Labs). Phalloidin- and antiserum-treated coverslips then were rinsed three times for 5 rnin with TM buffer, rinsed in distilled water and mounted in Elvanol (Dupont).
Digitonin Permeabilization of Axonal Fields During permeabilization, it was essential to exchange solutions in the chamber very slowly in order to prevent dispersion of axonal residue. Explants were washed twice with Cortland saline, and then with PHEM buffer. The permeabilization buffer was (in m M ) : 0.00 1% digitonin; 110, Pipes; 25, Hepes; and 2, MgC12: ( p H 6.9 with KOH). After the preliminary washes, the explants were incubated in permeabilization buffer containing in addition 5 mMNaN, and 1 ATP-yS for 30 min, followed by incubation for 30 min in either 1 m M AMP-PNP, 250 pLM Ca2+and 10 p M CD, or 1 ATP, 250 pM Ca2' and 10 pLMCD. This was followed by phalloidin staining (see above)
Determination of Extracellular-Free Calcium Free [ Ca2+]in the bathing medium was determined by the method of Bers ( I982 j , or by a computer program designed to compensate for pH, [Mg2+], and ionic strength (courtesy of Dr. Zahur Ahmed j. In either case, the free-Ca2+ activity was confirmed with a calciumsensitive electrode (Radiometer). In those experiments concerned with depleting intracellular calcium, explants were incubated in Ca2+-free,Cortland saline containing 1 m M EGTA and 5 p M ionomycin, a calcium ionophore (Liu and Hermann, 1978), and the membrane permeant Ca2+chelator, BAPTA-AM (20 gtM) (Tsien, 1980), for 1 h at 27°C prior to permeabilization.
557
RESULTS FITC-conjugated phalloidin, an F-actin specific probe (Barak, Yocum and Webb, 1981), exhibits a strong signal predominantly in growth cones [Fig. 1 (a-d)] , and appears only as scattered. occasional fluorescent streaks and puncta in preterminal RGC axonal fields [Fig. 1 ( f ) ] . Most axonal varicosities exhibit a weak, diffuse and short-lived phalloidin fluorescence, which is difficult to capture photomicrographically. On the other hand, actin immunostaining, as shown in Fig. lg, and h and reported previously (Koenig et al., 1985), is diffuse and strong, in varicosities. Although the more intense immunofluorescence reflects staining of both Gand F-actin, the apparent paucity of phalloidin fluorescence does not mean a lack of F-actin necessarily. Rather, given the amorphous cytomatrix visible in Varicosities at the electron microscopic level (Koenig et al., 1985), it suggests that phalloidin shows up F-actin best when the latter is organized into highly condensed structures, such as actin bundles and dense meshwork found in growth cones, as opposed to actin organized in a loose meshwork, more characteristic of that in varicosities. Differences between phalloidin labelling and immunostaining of actin have been noted also very recently in growing distal sympathetic ganglion cell axons (Bridgman and Dailey. 1989). Ten minutes after exposure of axonal fields to cytochalasin D, the distribution of phalloidin fluorescence is altered markedly. Numerous actincontaining foci appear scattered throughout the length of individual axons [Fig. 2(b,d)]. These focal aggregates stain very strongly with phalloidin. Although focal aggregates appear randomly distributed, regional differences are apparent from inspection because varicosities appear to contain a higher density [Fig. 2 ( d ) ] when compared to intervening axonal segments. Focal aggregates formed in response to CD treatment are not visible in phase-contrast microscopy [Fig. 2 (a,c)] and can be detected only at the light microscope level with fluorescent probes. Axonal fields incubated for 10 rnin in CD ( 1 p M ) , and then allowed to recover for 5 h in a CD-free medium do not exhibit phalloidin-stained foci. Such observations indicate that CD induces F-actin-containing aggregates to form de now and that the process is reversible. Most experiments described here were performed with 10 pA4 CD, which also was found to be reversible; however, foci could be obtained even with 0.1 pM. Up to 1% of DMSO alone did not induce the formation of foci. When viewed with time-lapse, phase-contrast
558
Edmonds and Koenig
Figure 1 Distribution of actin in RGC axons as revealed by either FITC-phalloidin fluorescence or RITC-actin immunofluorescence. (a, c, e ) Phase-contrast, and (b, d, f ) fluorescence micrographs of areas labelled with FITC-conjugated phalloidin. (b, d ) Growth cones exhibit a very strong fluorescence, ( f ) whereas preterminal axons generally show scanty phalloidin fluorescence. Actin immunofluorescence ( 6 ) ofgrowth cones and ( h ) varicosities is strong and distributed diffusely. Bar: 10 pm.
Actomyosin Aggregates in Growing Axons
559
Figure 2 Formation of focal aggregates labelled by FIX-phalloidin. (a, c ) Phase-contrast and (b, d ) corresponding fluorescence micrographs showing discrete FITC-phalloidin labelled focal aggregates scattered throughout preterminal axons following treatment with cytochalasin D ( 10 p M ) . Fluorescent focal aggregates appear in higher density (d) in Varicosities. Note that focal aggregates are not visible in phase-contrast microscopy. Bars: a,b, 10 pm;c,d, 5 pm.
videomicroscopy, CD did not appear to perturb visible-particle transport activity because the bidirectional movements of organelles continued unabated (Edmonds and Koenig, unpublished). Growth cones were the only structures that showed a sensitivity to CD treatment. Generally, all growth-cone motility ceased upon addition of CD, concomitant with the collapse of filopodial organization. The morphological changes of ,growth cones were consistent with those reported lby Marsh and Letourneau ( 19841, but the overall morphology of preterminal axons otherwise remained unchanged. T o investigate whether cytoskeletal components and/or regulatory proteins other than actin might be involved in the formation of focal aggregates, an immunocytochemical survey was undertaken to evaluate myosin and calmodulin distributions in RGC axons after CD treatment. Myosin normally appears in RGC axons as a diffuse immunofluorescence. which is unform and strongest iiri varicosities of untreated axons [Fig. 3 ( b,d)] .
Although myosin immunofluorescence is still somewhat diffuse following exposure to CD, the immunofluorescence in varicosities is nonuniform in distribution [Fig. 3 (g,i)] with the strongest immunofluorescence associated with phalloidinstained aggregates [Fig. 3 (f,i)]. This correlation suggests that myosin becomes redistributed to some extent during the CD treatment to become associated with F-actin-containing foci. Calmodulin, as shown previously (Koenig et al., 1985), is associated with varicosities and growth cones preferentially [Fig. 4 (a,b)] . Following CD treatment, residual calmodulin immunofluorescence [Fig. 4(g,h)] after the lysis/fixation procedure was co-distributed largely with phalloidin-labelled foci [Fig. 4( e, f ) ] . The apparent co-localization of myosin and calmodulin with F-actin-containing foci raises the possibility that the formation of the latter may be mediated by an actom yosin interaction. Because calmodulin antagonists have been shown to interfere with smooth muscle contraction (Kerrick,
560
Edmond.7 and Koenig
Figure 3 Apparent co-distribution of myosin with cytochalasin D-induced focal aggregates. Texas red-myosin immunofluorescence (b, d ) in untreated RGC axons and (g, j) in axons after cytochalasin D treatment, and in those (f, i ) which have been counterstained with FITC-phalloidin also. (a, c, e, h ) Phase-contrast, and (b, d, f-g, i-j ) corresponding fluorescence micrographs. Myosin immunofluorescence is normally diffuse and uniform in distribution, (b, d ) with varicosities exhibiting the strongest fluorescence. After cytochalasin D ( 10 p M ) , (g, j ) myosin immunofluorescence becomes distributed nonuniformly such as to exhibit enhanced labelling of focal aggregates ( f, i). which are labelled also strongly by FITC-phalloidin fluorescence (arrows). Bar: 10 pm.
Actomyosin Aggregates in Growing A.xons
Hoar, and Cassidy, 1980), calmidazolium, a potent calmodulin inhibitor (Gietzen, Wurthrich, and Bader, 1982; Van Belle, 1981), was used to test this possibility. Initial experiments showed that calmidazolium treatment of RGC axons caused a rapid and irreversible solation of the cytomatrix in the growth cone and varicosities. in which visible organelles appeared in random motion. Experiments later indicated that the solation may have been caused by a chloride-mediated osmotic swelling because removal of chloride from the bathing medium eliminated it (B. Edmonds, unpublished). When chloride was replaced on a molar basis by gluconic acid in the Cortland medium, both morphology and transport activity of axons were normal in the modified medium. Pretreatment of axonal fields for 30 min with calmidazolium ( 1 p M ) greatly reduced the number of aggregates that formed following subsequent treatment with CD [Fig. 5 ( d ) 1, while calmidazolium treatment after CD did not [Fig. 5 (b)]. Calmidazolium alone, at this concentration, had no detectable effect on axonal morpholog) or organelle transport. These results suggest that calmodulin may be important in the formation of focal cytoskeletal aggregates. As skinned muscle fibers (Cassidy, Hoar, and Kemck, 1979; Hoar, Kemck, and Cassidy, 1979; Kenick, Hoar, Cassidy, Bolles, and Haleucik, I98 1 ) and some permeabilized cell models (Holzapfcl et al., 1983; Kreis and Birchmeier, 1980; Masuda, Owaribe, Hayashi, and Hatano, 1983) havc the capacity to contract when exposed to suitable concentration of Ca2’ and ATP, a potential actomyosin involvement in the formation of CD-induced focal aggregates was probed in a permeabilized axonal model. Although immature RGC axons are structurally labile, axonal fields treated with 0.001% digitonin in a modified PHLM buffer (see Methods), yielded a reasonably stabilized preparation that stained with phalloidin after exposure for I min, indicating that the permeability barrier to the peptide was breached by this treatment. When CD was introduced during permeabilization, or a few minutes after, aggregate formation occurred in the absence of added ATP and Ca2+. After axons were maintained in permeabilization buffer for 30 min in the absence of ATP, CD failed to induce the formation of aggregates consistently, even if ATP and Ca2+were included. However, the inclusion of 1 m M ATP-TS, a thiophosphate ATP analog, in the permeabilization buffer significantly improved reproducibility of the Subsequent CD-induced formation of aggregates. It is possible
561
that this may have been due to a stable activation of myosin by thiophosphorylation of myosin (see Discussion ) . In experiments testing for ATP dependence of aggregate formation, 5 m M NaN3 was included in an ATP-yS-containing permeabilization buffer. After 30 min in this medium. CD was added with 1 m M ATP, resulting in the formation of aggregates [Fig. 6 (b)];however, CD added with 1 m M AMP-PNP, a nonhydrolyzable ATP analog, did not result in the formation of aggregates [Fig. 6(d)]. These experiments indicated that CD-induced aggregate formation was dependent on available ATP. Smooth muscle contraction (Sellers and Adelstein, 1987) and contraction of some permeabilized cell models (Holzapfel et al., 1983; Masuda et al., 1984; and Masuda. Owaribe, and Hatano, 1983) appear to have a requirement for calcium. A calcium dependency in CD-induced formation of aggregates was tested by varying [Ca”],. Permeabilization buffers with known free- [ Ca” ] , ranging from 0.1 pM to 250 pM were prepared (see Methods). In addition, some axonal fields were incubated for 1 h prior to permeabilization with BAPTA-AM ( 10 p M ), and ionomycin ( 5 p M ) , a calcium ionophore, in a Ca2+-free/EGTA-buffered medium to deplete [ Ca2+Il(see Methods). CD-induced aggregate formation was not affected by any of the conditions designed to modify [ Ca2+],including those involving a presumptive depletion of intracellular calcium. DISCUSSION
When goldfish axons regenerating in vitro are treated with cytochalasin D, numerous discrcte focal aggregates form, especially in varicosities, which stain strongly with phalloidin. As CD is known to interfere with actin polymerization (see below), it seems likely that CD induces a reorganization /condensation of preformed F-actin in varicosities, notwithstanding the fact that the latter normally stain very weakly with phalloidin. One possible explanation for weak phalloidin staining in varicosities may relate to a looser actin-based cytoskelctal meshwork compared to that present in growth cones (Bridgman and Dailey, 1989). Actin aggregate formation in response to cytochalasin treatment has been reported in plants (Palevitz. 1988), and algae (Menzel and Schliwa, 1986), as well as in animal cells (Bitch and Allen, 1981; Godman, Woda, and Kolberg, 1980;
562
Edinonh und Koenig
Figure 4 Apparent co-distribution of calmodulin with cytochalasin D-induced focal aggregates. Distribution of RITC-calmodulin immunofluorescence (b) in untreated RGC axons and (g, h) in axons after cytochalasin D treatment and in those (e, f ) which have been
Actomyosin Aggregates in Growing Axons
563
Figure 5 Blockade of cytochalsin-D-induced formation of focal aggregates after antagonism of calmodulin. ( a ) Phase-contrast and (b) corresponding FITC-phalloidin fluorescence micrographs of an axonal fascicle pretreated with cytochalasin D (10 p M ) and followed by calmidazolium. (c) Phase-contrast and ( d ) FITC-phalloidin fluoresccnce micrographs of axonal fascicles pretreated for 30 rnin with calmidazolium ( 1 pm) and followed with cytochalasin D (10 phi'). Bar: 10 pm.
Schliwa. 1982; Weber. Rathke. Osborn. and Franke, 1976), including neurons (Marsh and Letourneau, 1984; Letourneau and Shattuck, 1989). Typically, cytochalasin B (CB) or CD was used in the latter studies. Although cytochalasins, as a group, bind actin and interfere with its polymerization, some are more specific. For example, CD, which does not affect membrane glucose transport, is more specific for actin than is CB, which does. and its efficacy in inhibiting actin polymerization in vitru is about 10 times greater (Cooper, 1987). The process of aggregate formation induced by CD in RGC axons appears to require ATP as indi-
cated by experiments in which aggregate formation was blocked in digitonin-permeabilized axon models by AMP-PNP, a nonhydrolyzable ATP analog. These results are in agreement with those reported in other systems (Miranda, Godman, Deitch, and Tannenbaum, 1974; Godman et al., 1980; Schliwa, 1982). In addition to actin, there is also a redistribution of myosin and calmodulin in response to CD treatment, such that these proteins appear to co-localize with phalloidin-stained aggregates. The co-localization of actin. myosin, and calmodulin in aggregates suggests that CD-induced for-
counterstained with FITC-phalloidin also. (a, c, d ) Phase-contrast and (b, e-h) corresponding fluorescence micrographs. Typically, ( b1 RITC-calmodulin immunofluorescence is associated largely with varicosities and growth cones where it is distributed diffuscly. After treatment with cytochalasin D ( 10 pLM) and washout during the lysis/fixation procedure (see Methods), ( g , h ) the calmodulin immunofluorescence is focal and ( e , f ) co-distributed with FITC-phalloidin fluorescence. Bar: 10 pm.
564
Edmnnds and Koenig
Figure 6 Formation of focal aggregates in permeabilized RGC axons in the presence of ATP but not in presence of AMP-PNP. Axons were permeabilixd with diptonin (0.001%) in PHEM buffer (see Methods). ( a ) Phase-contrast and ( b ) FITC-phalloidin fluorescence micrographs after treatment with cytochalasin D (10 p M ) , ATP ( I mhf),and calcium ( 10 &if). (c) Phase-contrast and ( d ) FITC-phalloidinfluorescence micrographs after treatment with cytochalasin D (10 p M ) , AMP-PNP ( 1 mM), and calcium ( 10 pLhf). Bar: 10 pm.
mation of the latter may reflect an active contractile event. The requirement for ATP also supports this inference. In addition, the inhibition by calmidazolium, a potent calmodulin inhibitor (Van Belle, 198 1 ) and the improved reproducibility when ATP-yS is included during permeabilization of axons are also consistent with the possibility of a calmodulin-dependent activation of myosin lightchain kinase (MLCK) anteceding the formation of aggregates. MLCK exhibits a similar diffuse distribution in varicosities as that of myosin in RGC axons (B. T. Edmonds, unpublished). Calmodulin antagonists block contraction in slunned smooth muscle fibers (Kerrick et al., 1 98 1 ) and in permeabilized Swiss 3T3 cells (Holzapfel et al., 1983). Moreover, contraction in the latter cells is inhibited by MLCK antibodies (Holzapfel et al., 1983). MLCK in gizzard muscle (Sherry, Gorecka, Aksoy, Dabrowska, and Hartshorne, 1978) and in rabbit smooth muscle (Cassidy et al., 1979) can utilize ATP-yS to thiophosphorylate the myosin light chain, which effectively maintains myosin in an activated state, as dethtophosphorylation is not cata-
lyzed readily by the phosphatase (Cassidy et al., 1979). In some examples of permeabilized cell models, contractions exhibit a strict requirement for calcium (Hoar et al., 1979; Holzapfel et al., 1983; Kemck et al., 1980: Masuda et al., 1983). CD-induced aggregate formation in RGC axons occurred with all calcium concentrations tested, including conditions designed to deplete intracellular calcium. Although depletion of intracellular Ca2+could not be verified, EGTA ( > 1 m M ) in the permeabilizing buffer in combination with the intracellular-Ca2+ chelator, BAPTA-AM, should have reduced free [Ca”], to a very low level. Thesc rcsults are not consistent with experiments in which blockade was produced by calmodulin antagonism. One possible explanation is that the contractile machinery is already in an activated state after pcrmeabilization. There is some evidence that the method of pcrmeabihzdtion can affect the degree of calcium sensitivity of contraction in cell models. For example, whereas mouse 3T3 cells undergo contractions with calcium concen-
Artomyosin Aggregates in Growing Axons
trations > 1 pM after Triton treatment, contractions in glycerinated 3T3 cells exhibit no calcium dependency (Masuda et al.. 1983). Electronmicroscopic inspection of RGC axons reveals that varicosities contain abundant arrays of tubulo-vesicular profiles, embedded in an amorphous cytomatrix (Koenig and Adams, 1982; Koenig et al., 1985). A weak phalloidin signal is consistent with the view that actin is organized probably as a loose meshwork (see above). Myosin immunofluorescence is distributed diffusely within varicosities. Whether actomyosin interactions may contribute normally to tension of the cytomatrix in varicositics is not known, but it is a possibility. In this context, it is noteworthy that varicosities of RGC axons, in a calcium-free Cortland bathing medium containing the calcium ionophore A23 187, undergo a reduction in volume of about 20% when [Ca2+], is changed from 0 to nominally 0.0 I p A 4 (E. Edmonds, unpublished), The reduction in volume suggests that a contraction may have occurred, although a loss of volume through a calcium-dependent efflux of K (Edmonds and Koenig, 1990) cannot be ruled out. The mechanisms underlying the formation of aggregates in response to CD treatment are uncertain. The evidence suggests that CD caps the fast growing, barbed ends of actin filaments, and that it can compete with capping proteins (Cooper, 1987). As pointed out by Cooper (1987), CD could displace capping proteins that hold actin filaments in place, and thereby generate free filament ends. If there is a basal level of isometric tension mediated by actomyosin interactions within the cytomatrix of varicosities, then the production of "free" actin filament ends by CD would generate isotonic contractions that would appear as focal cytoskeletal condensations (i.e.. aggregates). This view is consistent with a model proposed by Schliwa and van Rlerkom ( 198 1 ) in which cytochalasin-induced aggregates form in a manner similar to contractions in smooth muscle +
We thank Mrs. Sarah Finnegan Sloan for her technical assistance in certain aspects of thc work. This research was supported in part by grants NS21843 from NTNCDS and BRSG SO7 RK05600-26 awarded by the Biomedical Research Support and Grant Program, Division of Research Resources. NTH.
REFERENCES BARAK,L. s.,YOCUM,R. R., and WEBB,W. W. ( 198 1 ). In vivo staining of cytoskcletal actin by autointernalization of nontoxic concentrations of nitrobenzoxadiazole-phallacidin. J . Cell Biol. 89:368-372.
565
BERS.D. M. (1982). A simple method for the accurate determination of free [ Ca2' ] in Ca-EGTA solutions. Am. .I. Physiol. (Cdl Physiol. 1I ) 242:C404-C408. BRAY,D., and WIIITE.J . G. (1988). Cortical flow in animal cells. Science 239983-888. BRIDGME.N. P. C., and DAILEY, M. E. ( 1989). The organization of myosin and actin in rapid frozen nerve growth cones. J. Cell Biol. 108:95-109. BKITCH,M., and ALLEN,T. D. (1981). The effects of cytochalasin B on the cytoplasmic contractile network revealed by whole-cell transmission electron microscopy. L k p . Cell Res. 131:161-172. CASSIDY,P., H0.4R, P. E.. and KERRICK.W. G . L. ( 1979). Irreversible thiophosphorylation and activation of tension in functionally skinned rabbit ileum strips by 3'S-ATP-yS. J. Biol. Ckem. 254:l 114811153. COOPER, J. A. (1987). Effects of cytochalasin and phalloidin on actin. J. Cell Bid. 1051473-1478. EDMONDS,B. T., and KOENIG,E. ( 1087). Powering of bulk transport (varicosities j and differential sensitivities of directional transport in growing axons. Bruin Rex 406:288-293. EDMONDS,€3. T., and KOENIG, E. (1990). Calcium-dependent volume reduction in regenerating ganglion cell axons in v i m . J. hieurosci. Rex In press. GIETZEN, K., WUKTHRICH. A.. and BADER,H. ( 1982). R24571: a new powerful inhibitor of red blood cell Ca++-transport ATPase and of calmodulin-regulated functions. Hiochem. Eliophys. Re.y. Comm. 101:418425. GODMAN. G., WODA, B., and KOLBERG,R. (1980). Redistribution of contractile and cytoskeletal components induced by cytochalasin. I. In Hmf cells, a nontransformed fibroblastoid line. Eur. J . Cell B i d . 22:733-744. HOLLCNHECK. P. J., and BRAY, D. (1987). Rapidly transported organelles containing membrane and cytoskeletal components: their relation to axonal growth. .I. Cell Biol. 105:2827-2835. HOAR,P. E., KEKRICK, W. G. L., and CASSIDY,P. S. ( 1979). Chicken gizzard: relation between calciumactivated phosphorylation and contraction. Science 2041503-506. HOLZAPFEL, G., WEHLAND, J., and WEBER,K. ( 1983j. Calcium control of actin-myosin based contraction in Triton models of mouse 3T3 fibroblasts is mediated by myosin light chain kinase (MLCK)-calmodulin complex. Exp. Cell Rcs. 148: I 17- 126. KERRICK, W. G. L., HOAR,P. E. and CASSIDY,P. S. ( 1980). Calcium-activated tension: the role of myosin light chain phosphorylation. Fed. Proc. 39: 15581563. KERRICK,W. G. L., HOAR, P. E.. CASSIDY.P. S., BOLI,ES, L., and HALEUCIK, D. A. (1981). Calciumregulatory mechanisms: functional classification using skinned fibers. J. Gen. Ph,ysiol. 77:177-190. KOENIG.E., and ADAMS,P. (1982). Local protein syn-
566
Edmonds and Kornig
thesizing activity in axonal fields regenerating in vitro. J. Neurochem. 39:386-400. KOENIG,E., KINSMAN, S., RFPASKY,E., and SuLiz, L. (1985). Rapid motility of motile varicosities and inclusions containing spectrin, actin and calmodulin in regenerating axons in vitro. J. Neuro.sci. 5 7 15-729. W. ( 1980). Studies on the KREIS,T., and BIRCHMEIER, organization and localization of actin and myosin in neurons. J. Cell Biol. 8:356-37 I . KUCZMARSKI, E. R., and ROSENBAUM, J. L. (1979). Studies on the organization and localization of actin and myosin in neurons. J . Cell Biol. 80:356-37 I . LANDRETH, G. E.. and AGRANOFF,B. W. (1976). Explant culture of adult goldfish retina: effect of prior optic nerve crush. Brain Rex 118:299-303. LETOURNEAU, P. C. ( 1981 ), Immunocytochemical evidence for colocalization in neurite growth cones of actin and myosin and their relationship to cell-substratum adhesions. Dev. B i d 85: 1 13- 122. LETOURNEAU, P. C., and SHATTLJCK, T. A. (1989). Distribution and possible interactions of actin-associated proteins and cell adhesion molecules of nerve growth cones. Development 105505-5 19. LIU,C., and HERMANN, T. E. ( 1978). Characterization of ionomycin as a calcium ionophore. J . Biol. C'hem. 25335892-5894. MARSH,L., and LETOURNEAU, P. C. ( 1984). Growth of neuntes without filopodidl or lamcllipodial activity in the presence of cytochalasin B. J. Cell B i d 99:20412047. MASIDA, H.. OWARIBE. K.: and HATANO,s. (1983). Contraction of Triton-treated culture cells. A calcium-sensitive contractile model. E.q. Cell Res. 143:79-90. MASVDA, H., OWARIBE, HAYASHI, H.; and HATANO,S. ( 1983). Ca++-dependentcontraction of human lung fibroblasts treated with Triton X-100: a role ofCa++calmodulin dependent phosphorylation of myosin 20.000 Dalton light chain. Cell Motil. 4:3 15-33 I . MENZEL,D., and SCHLIWA, M. (1986). Motility in the siphonaceius green alga Bryopsit IT. Chlorplast movement requires organized arrays of both microtubules and actin filaments. Eur. J . Cell Biol. 40:286-295.
MIRAYDA,A., GODMAN,G., DEITCH,A., and TANNENBAUM, S . (1974). Action of cytochalasin D on cells of established lines, 1. Early events. J. Cell Biol. 61:481-500. MITCHISON, T., and KIRSCHNER, M. ( 1988). Cytoskeleta1 dynamics and nerve growth. IVeuron 1:76 1-772. B. A. ( 1988). Cytochalasin-induced reorgaPALEVITZ, nization of actin in Allum root cells. CMc'9:283-298. M., HXJ, L., and ROISEN,F., INCZEDY-MARCSEK, YORKE,W. (1978). Myosin: immunofluorescent localization in neuronal and glial cultures. Science 199:1445-1448. SCHLIWA. M. ( 1982). Action ofcytochalasin D on cytoskeletal networks. J. CPll Bid. 92:79-9 I. SCHLIWA, M., and V A BLEKKOM, ~ J. ( 1981 ). Structural interaction of cytoskeletal components. J. Cell BioE. 90:222-235. SELIERS. J. R., and ADELSI'EI~, R. S. (1987). Regulation of contractile activity. In: The Enzymes. Vol. XVIII B. P. D. Buyer and E. G. Krebs, Eds.. Academic Press, New York, pp. 38 1-4 18. SHERRY. J. M. F., GORECKA, A.?AKSOY,M. O., DABROWSKA, K.,and HAKTSHOKNE, D. J. ( 1978). Roles of calcium and phosphorylation in regulation of the activity of guzard myosin. Biochem. 17:44 1 1-44 18. SMITH. S. J. (1988). Neuronal cytomechanics: the actin-based motility of' growth cones. Science 242708-7 1 5 . TSIEN,R . Y. (1980). New calcium indicators and buffers with high selectivity against magnesium and protons: design. synthesis, and properties of prototype structures. Biochern. I9:2396-2404. VANBELLE,H. (1981). R24571: a potent inhibitor of calmodulin-activated enzymes. Cell Culciurn 2:483494. P. C., OSBORN, M., and FRANKE. WEBER,K., RATHKE, W. W. (1976). Distribution of actin and tubulin in cells and in glycerinated cell models after treatment with cytochalasin B (CB). Exp. Cell Rex 102:285297.
