1531

Journal of Cell Science 108, 1531-1539 (1995)Printed in Great Britain © The Company of Biologists Limited 1995

Stationary organization of the actin cytoskeleton in Vallisneria: the role of

stable microfilaments at the end walls

Jung-Hwa Ryu, Shingo Takagi and Reiko Nagai*

Department of Biology, Faculty of Science, Osaka University, Machikaneyama 1-16, Toyonaka, Osaka, 560 Japan

*Author for correspondence

In mesophyll cells of the aquatic angiosperm Vallisneriagigantea, bundles of microfilaments (MFs) serve as tracksfor the rotational streaming of the cytoplasm, which occursalong the two longer side walls and the two shorter endwalls. The stationary organization of these bundles hasbeen shown to depend on the association of the bundleswith the plasma membrane at the end walls. To identify thesites of such association, the effects of cytochalasin B (CB)on the configuration of the bundles of MFs were examined.In the case of the side walls, MFs were completely disruptedafter treatment with CB at 100 µg/ml for 24 hours. Bycontrast, in the case of the end walls, a number of partiallydisrupted MFs remained even after 48 hours of treatment.After removal of CB, a completely normal arrangement ofbundles of MFs was once again evident within 24 hoursafter a rather complicated process of reassembly. Whenreassembly had been completed, the direction of cytoplas-mic streaming was reversed only in a small fraction of the

treated cells, suggesting that bundles of MFs are anchoredand stabilized at the end walls of each cell and that thepolarity of reorganized bundles and, therefore, thedirection of the cytoplasmic streaming is determined in amanner that depends on the original polarity of MFs thatremained in spite of the disruptive action of CB. Bycontrast, the direction of reinitiated cytoplasmic streamingwas reversed in 50% of cells in which the bundles of MFshad been completely disrupted by exogenously appliedtrypsin prior treatment with CB. The results confirm thatprotease-sensitive anchoring of microfilament bundles atthe end walls is crucial for maintenance of the unique, sta-tionary organization of the tracks for cytoplasmicstreaming in these cells.

Key words: actin, cytochalasin B, cytoplasmic streaming, end wall,microfilament, Vallisneria

SUMMARY

INTRODUCTION

In plant cells as in animal cells, the cytoskeleton, which iscomposed mainly of microtubules and microfilaments (MFs),plays a variety of pivotal roles. It is essential, for example, forcell division, morphogenesis and cell motility. The dynamicbehavior of cytoskeletal elements throughout the plant cellcycle and during developmental processes has been exten-sively examined and well documented (Gunning and Hardham,1982; Lloyd, 1982, 1991; Menzel, 1992). However, themechanism responsible for organization of stationary arrays ofMFs has not been investigated in detail.

In leaf cells of Vallisneria, an aquatic angiosperm, cyto-plasmic streaming is induced by external stimuli, such as irra-diation with light or application of various chemicals (Haupt,1959, 1982; Kamiya, 1959; Nagai, 1993; Seitz, 1979). Thecytoplasm streams rotationally along the four anticlinal walls:namely, the two longer side walls and the two shorter end walls(Fig. 1). The tracks for the cytoplasmic streaming are bundlesof MFs, which are mainly composed of fibrous (F-) actin(Yamaguchi and Nagai, 1981) and run in the ectoplasm parallelto the direction of streaming (Masuda et al., 1991; Takagi andNagai, 1983). The configuration and organization of the

bundles of MFs do not change regardless of the occurrence ofcytoplasmic streaming (Takagi and Nagai, 1983). In characeancells, in which the mechanism of cytoplasmic streaming hasbeen extensively investigated (Kamiya, 1981; Kuroda, 1990),the direction of cytoplasmic streaming has been revealed to bedetermined by the polarity of the F-actin that makes up thebundles of MFs (Kersey et al., 1976). The MFs in a givenbundle all have the same polarity (Palevitz et al., 1974; Palevitzand Hepler, 1975).

In mesophyll cells of V. asiatica, Ishigami and Nagai (1980)showed that cytochalasin B (CB) inhibits cytoplasmicstreaming and, moreover, that when the treatment with CB iscontinued for as long as 24 hours, the direction of the rotationalstreaming is reversed in about 50% of the treated cells after theremoval of CB. These findings indicate that MFs are disas-sembled completely in the presence of CB and that they canreassemble appropriately to regenerate tracks for streamingafter the removal of CB. The polarity of each re-formed bundlemay be determined by an exclusively stochastic process duringreassembly. Thus, the bundles of MFs in Vallisneria plantsappear to be rather sensitive to the action of CB.

Using single mesophyll cells of V. gigantea that had beenisolated by enzymatic digestion, Masuda et al. (1991) demon-

1532 J.-H. Ryu, S. Takagi and R. Nagai

Fig. 1. Schematic representation of the anatomy of a leaf ofVallisneria. A double-headed arrow indicates the longitudinal axis ofthe leaf. Arrowheads in each cell represent the direction ofcytoplasmic streaming, which occurs along the four anticlinal walls.

strated that hypertonic treatment induces abnormal patterns ofcytoplasmic streaming concomitantly with detachment of theplasma membrane specifically from one or both of the endwalls of the cell. While inhibitors of proteases, added to thesolution of enzymes used for isolation of the single cells, sup-pressed the disturbance of the tracks for streaming, exoge-nously applied proteases promoted this disturbance. Theseresults suggest that the bundles of MFs in the ectoplasm areanchored at specific sites, presumably at the end walls of thecell, via adhesion of the plasma membrane to the cell wall,with the probable involvement of some protease-sensitivefactor(s).

On the basis of the results obtained by Masuda et al. (1991)and by Ishigami and Nagai (1980), we postulated that we mightbe able to identify the sites at which the bundles of MFs areanchored by monitoring changes in the configuration of thebundles during treatment with CB. If the bundles of MFs areanchored and stabilized, in the presence of some putativecomponent(s), at the end walls, these bundles should be moreresistant than those at the side walls to long-term treatmentwith CB. To examine this possibility, we followed in detail theprocess of CB-induced disruption of the bundles of MFs at theside walls and at the end walls by staining MFs with fluores-cein isothiocyanate-conjugated (FITC-conjugated) phalloidin.If we assume that the bundles of MFs are stabilized at the endwalls, it seems plausible that they might play some role in thereconstruction of the normal organization of bundles of MFs.Therefore, we also examined the process of reassembly of thebundles of MFs after removal of CB. Furthermore, to confirmthe role of stabilized MFs, we examined the direction of rota-tional cytoplasmic streaming in each cell before and aftertreatment with CB. The results were compared with thoseobtained from cells, which were treated first with trypsin andthen with CB, in which the MFs at the end walls had been com-pletely disrupted.

MATERIALS AND METHODS

Plant materialVallisneria gigantea Graebner was purchased at a tropical-fish store

and cultured in water-filled buckets with soil at the bottom. Theculture was kept under a regime of 12 hours of light, at 2,000 lux,from fluorescent lamps (FL 20S-PG; National, Kadoma, Japan), and12 hours of darkness at 18 to 28°C.

Pretreatment of specimensTo obtain cells of nearly the same age, leaf segments of about l0 cmin length were consistently taken from a site about 40 cm from thebase of each leaf. Each segment was cut into smaller pieces of about1 cm in length. Each of these pieces was further cut along the lon-gitudinal or transverse axis of the leaf to expose the side walls orthe end walls of mesophyll cells, respectively (Fig. 1), for exami-nation of the bundles of MFs. The trimmed pieces of leaves wereincubated under the original light regime for 24 hours in artificialpond water (APW), which contained 0.05 mM KCl, 0.2 mM NaCl,0.1 mM Ca(NO3)2, 0.1 mM Mg(NO3)2, and 2 mM Pipes buffer atpH 7.0. For geometric reasons, one cannot observe both wallsof a cell at the same time. Therefore, each wall was observed sepa-rately.

Light microscopyCytoplasmic streaming and MFs were observed under a light micro-scope (Optiphoto-2; Nikon, Tokyo, Japan) equipped with differentialinterference contrast (DIC) optics and an epifluorescence illuminationsystem. When necessary, observations were recorded with a televi-sion camera (WV-1550; National) and stored on videotape with arecorder (HV-BS53; Mitsubishi Electric, Tokyo, Japan).

Staining of microfilamentsMFs were visualized by staining with FITC-phalloidin (MolecularProbes, Junction City, OR, USA), as described by Kakimoto andShibaoka (1987) with slight modifications (Masuda et al., 1991). Thespecimens were mixed with lysis buffer that contained 220 nM FITC-phalloidin, 1.0 mM NaCl, 2.5 mM KCl, 0.5 mM Mg(NO3)2, 0.02%(w/v) Triton X-100, 0.1% (w/v) p-phenylenediamine, 10 mM EGTA,100 mM potassium phosphate (pH 6.8), and protease inhibitors (100µg/ml leupeptin, 40 µg/ml (p-amidinophenyl)methanesulfonylfluoride hydrochloride, 80 µg/ml chymostatin, and 8 µg/ml pepstatin).The stained cells were examined after incubation for 30 minutes at25°C. A BA520-560 filter was used to eliminate autofluorescencefrom chlorophyll. Micrographs were taken with Kodak Tri-X pan film(ISO 400).

Treatment with chemicalsFor examination of the disruption of bundles of MFs, pretreatedpieces of leaves were treated with CB at 100 µg/ml for given times,as indicated in the text. After each treatment, specimens werewashed twice with fresh APW and then processed for visualizationof MFs. Specimens for a control experiment were subjected to thesame treatment in the absence of CB. For examination of thereassembly of the bundles of MFs, the pretreated pieces of leaveswere treated with CB at 100 µg/ml for 24 hours. In cases in whichthe end walls were also treated with trypsin, leaf pieces were treatedwith APW that contained trypsin (11,000 units/mg) at 30 µg/ml for30 minutes and then washed with APW prior to the treatment withCB for 24 hours. Then the specimens were washed several timesvigorously with APW and kept in fresh APW under the originallight/dark regime. MFs were visualized at the times indicated in thetext.

The direction of cytoplasmic streaming in each mesophyll cell wasrecorded before application of CB at 100 µg/ml or trypsin and CB.After a given period of treatment with CB, the drug was removed byvigorous washing with APW. The direction of the newly establishedcytoplasmic streaming was determined in the each cell. The ratio ofthe number of cells in which the direction of streaming had beenreversed to the total number of cells examined (Ntotal) was determinedand expressed as a percentage.

1533Actin cytoskeleton in Vallisneria

Fig. 2. Disruption of the bundles of MFs atthe side walls of mesophyll cells ofVallisneria during treatment with CB. (A) Before treatment. Arrays of bundles ofMFs parallel to the longitudinal axis of thecell are obvious. (B) After a 6 hour treatmentwith CB at 100 µg/ml. The bundles of MFshave been disrupted and small clusters ofshort MFs remain sporadically distributed.(C) After an 18 hour treatment. Disruption ofMFs seems to have proceeded further than inB. Note that short MFs remain at bothlongitudinal ends of the cell (arrows). (D) After a 24 hour treatment. The bundles ofMFs are no longer detectable. Short MFs stillremain at the longitudinal ends of the cell(arrows). Bar, 20 µm.

Fig. 3. Time course of changes in the configuration of the bundles ofMFs at the side walls during treatment with CB. Three typicalpatterns of MFs, types I to III (see the text for further details),visualized by staining with FITC-phalloidin, are representedschematically. The numbers of cells with MFs of each type (Nmf)were counted at given times after the start of treatment with CB at100 µg/ml. The ratio of Nmf to the total number of cells (Ntotal) thatexhibited a normal pattern of cytoplasmic streaming just before eachtreatment is plotted as a percentage (filled symbols; Ntotal= 636 to753). As controls, untreated cells were stained at the same timepoints as the treated cells (open circles; Ntotal= 660 to 714). Thepattern of MFs in untreated cells was type I.

N mf/N

tota

l×10

0

RESULTS

Disruption of bundles of MFs caused by CB at theside wallsIn control cells, which exhibited the normal pattern of cyto-plasmic streaming, several bundles of MFs were alignedparallel to one another to form an array at the side walls. Theentire array, as well as the individual bundles of MFs withinit, was oriented parallel to the longitudinal axis of the cell (Fig.2A; referred to as type I). After treatment with CB at 100 µg/mlfor 6 hours, the bundles of MFs were disrupted and only smallclusters of short MFs were detectable at numerous sites in thecytoplasm (Fig. 2B; type II). The number of clusters decreasedduring prolonged treatment with CB (Fig. 2C). Althoughalmost all the MFs at the side walls had disappeared after 24hours of treatment with CB (Fig. 2D; type III), short MFs con-sistently remained at both longitudinal ends of the cell (arrowsin Figs 2C,D). Fig. 3 shows the typical time course of thesechanges in the configuration of bundles of MFs. The numberof cells of type I rapidly decreased to almost 0% within 6 hoursof the start of treatment with CB. The number of cells of typeII increased to about 80% at 6 hours and then decreasedgradually during the next 12 hours. Finally, cells of type IIIbecame the most prominent after 24 hours of treatment with

1534 J.-H. Ryu, S. Takagi and R. Nagai

Fig. 4. Partial disruption of the bundles ofMFs at the end walls during treatment withCB. (A) Before treatment. Several bundles ofMFs run approximately parallel to oneanother. Note a fork-like structure (arrow) inthe vicinity of the margin of the end wall. (B) After a 6 hour treatment with CB at 100µg/ml. Some of the bundles of MFs havebecome disrupted and form circular or mesh-like arrays. A fork-like structure (arrow) isalso seen. (C) After a 24 hour treatment.Bundles of MFs are still visible, butfragmentation seems more pronounced thanin B. (D) 48 hours after treatment.Considerable numbers of short bundles ofMFs, which are randomly oriented, aredetectable. Bar, 20 µm.

Fig. 5. Changes in the number of cells that retain MFs at the endwalls during treatment with CB. The number of cells in which anytype of MF was detectable at the end walls (Nmf) was counted at agiven time after the start of treatment with CB at 100 µg/ml. Theratio of Nmf to Ntotal (62 to 75, see the legend to Fig. 3) was plottedas a percentage (filled circles). As controls, untreated cells wereexamined, as described in the legend to Fig. 3 (open circles; Ntotal=63 to 74).

N mf/N

tota

l×10

0

CB. We obtained very similar time courses for the disruptionof bundles of MFs in several experiments. It was clear that thebundles of MFs in the cytoplasmic layer that faced the sidewalls were completely disrupted by exposure to CB for 24hours. In addition, we noticed that short fragments of MFsremained in the vicinity of both junctions of each side wallwith the end walls. The sum of the number of cells with MFsof each type was always smaller than the total number of cellsthat exhibited a normal pattern of cytoplasmic streaming.About 10% of the cells were unstained by FITC-phalloidin.This result was probably due to insufficient permeation of theconjugate into cells.

Disruption of bundles of MFs caused by CB at theend wallsIn the control cells, which exhibited a normal pattern of cyto-plasmic streaming, several bundles of MFs could be seen thatwere approximately parallel to one another along the end wall(Fig. 4A). We noted that bundles appeared to separate intoseveral narrower bundles to form fork-like structures at thecorners of the cells (Fig. 4A,B, arrows). The bundles wereslightly disrupted after 6 hours of treatment with CB (Fig. 4B)and to a slightly greater extent after 24 hours (Fig. 4C). Furtherdisruption of the bundles proceeded during the next 24 hours.However, a considerable number of short bundles of MFs,which were randomly oriented, still remained in the cytoplasm(Fig. 4D). No typical pattern of fragmentation of the bundlesof MFs was observed at the end walls during such long-termtreatment with CB. The numbers of cells with any type of MF,

fragmented or not, were counted at given times after the startof treatment with CB. As shown in Fig. 5, the number of cellswith some type of MF at the end walls did not change through-out the treatment. We obtained very similar results in severalexperiments. These results suggested that the bundles of MFs

1535Actin cytoskeleton in Vallisneria

at the end walls were more resistant to CB than those at theside walls.

Reassembly of bundles of MFs after removal of CBAs described above, the bundles of MFs at the side walls dis-appeared completely after treatment with CB for 24 hours, withthe exception of short MFs in the vicinity of the end walls. Bycontrast, bundles of MFs at the end walls remained visible after48 hours of treatment, even though they were partiallydisrupted. Those MFs that remained after the long-termtreatment with CB might be assumed to provide the initialscaffold when the original organization of bundles is restored

Fig. 6. Reassembly of bundles of MFs at the side walls after removal of washed with APW. (A) 6 hours after the removal of CB. In a few cells, schloroplasts are detectable. (B) Enlarged view of a region in A. The arroaround the chloroplast. (C) After 12 hours. Fluorescent structures are lonseen on the left side of the cell. (D) The DIC image of C. An arrow indicentangled mass of MFs observed in C. (F,H) Other examples of reassemand thicker bundles of MFs (arrow), which run parallel to the longitudinthe entangled masses of MFs observed in F and H respectively. Bar, 20 µ

after removal of CB. To examine this possibility, we monitoredthe process of reassembly of MFs.

Leaf pieces were treated with CB at 100 µg/ml for 24 hoursand then washed with APW. Six hours after the removal of CB,few cells exhibited any fluorescence that indicated the presenceof reassembled MFs at the side walls. In a very few cells (Fig.6A,B), faint fluorescence that encircled the chloroplasts (Fig.6B, arrowhead) and represented short bundles of MFs (Fig. 6B,arrow) was detected sporadically. After 12 hours, the fluores-cent structures became longer and the fluorescence becamemore intense. Entangled masses of MFs of various shapes andsizes were often seen (Fig. 6C,E,F,G,H,I) in areas that were

CB. Cells were treated with CB at 100 µg/ml for 24 hours and thenhort bundles of MFs and faint fluorescence that encircles thew indicates a short bundle of MFs and the arrowhead indicates MFsger and fluorescence is more intense. An entangled mass of MFs isates a region rich in cytoplasm. (E) An enlarged photograph of thebled MFs and an entangled mass of MFs. A meshwork array of MFsal axis of the cell, have been re-formed. (G,I) Enlarged photographs ofm.

1536 J.-H. Ryu, S. Takagi and R. Nagai

Fig. 7. Reassembled bundlesof MFs at the side wall afterremoval of CB. Cells weretreated as described in thelegend to Fig. 6. (A) 24 hoursand (B) 48 hours after theremoval of CB. The normalpattern of bundles of MFs hasbeen regenerated and isindistinguishable from that inuntreated cells (e.g. Fig. 2A).Bar, 20 µm.

Fig. 8. Time course of the reassembly of MFs at the side walls afterremoval of CB. Cells were treated with CB at 100 µg/ml for 24 hoursand then washed with APW at time 0. Nmf and Ntotal (4,600 to 4,800)were determined as described in the legend to Fig. 5. Each pointrepresents the mean value of results from 40 leaf pieces with about100 cells in each. The vertical bars show the s.d. for each value.

N mf/N

tota

l×10

0

probably rich in cytoplasm (Fig. 6D, arrow). There were alsonetworks composed of thin bundles of MFs (Fig. 6C,F,H), aswell as longer and thicker bundles of MFs that ran parallel tothe longitudinal axis of each cell (Fig. 6H, arrow). After 24 to48 hours, almost normal patterns of bundles of MFs wereobserved in most cells (Fig. 7A,B). By this time, the varioustransiently observed configurations of MFs, described above,had completely disappeared. Fig. 8 shows the time course of

the increase in the number of cells in which MFs had reassem-bled after the removal of CB.

Cytoplasmic streaming could be induced by irradiation withlight about 12 hours after the removal of CB. Initially, localstreamlets occurred, the directions and patterns of which werevery unstable and changed with time. The directed streamingof the cytoplasm over long distances became obvious within 3hours of the start of irradiation with light. Complete recoveryof normal cytoplasmic streaming was first seen about 22 hoursor longer after the removal of CB.

Reassembly of bundles of MFs at the end walls in almost allcells was completed within 24 hours of the removal of CB (Fig.9A,B). The arrangement of the bundles was mostly indistin-guishable from that in control cells (see Fig. 4A). There wasless forking of bundles evident at the margin of some cells asshown in Fig. 9A. At this time, the pattern of cytoplasmicstreaming was also normal at the end walls.

The effects of CB on the direction of reinitiatedcytoplasmic streamingDuring our observations of the process of reassembly of thebundles of MFs at the side walls (Fig. 6), we found no imme-diately obvious evidence that MFs resistant to CB in thevicinity of the end walls might provide an initial scaffold forthe reconstruction of bundles. However, such a lack ofevidence could not be taken to rule out the possibility that thestabilized MFs might play some role in the reassembly ofbundles. Therefore, to confirm the involvement of stabilizedMFs in the determination of the polarity of reconstructed

Fig. 9. Reassembled bundles of MFs at theend walls after removal of CB. Cells weretreated as described in the legend to Fig. 6.24 hours after the removal of CB, bundlesof MFs, the arrangement of which isindistinguishable from that in untreatedcells, though less bifurcation of bundles isevident at the margins of the cell (A), havebeen reconstructed. Bar, 20 µm.

1537Actin cytoskeleton in Vallisneria

N x/N

tota

l×10

0

Fig. 10. Effects of CB on the direction of cytoplasmic streaming.Cells were treated as described in the legend to Fig. 6. The ratio ofthe number of cells in which the direction of cytoplasmic streamingwas reversed after the removal of CB (Nx) to the total number ofcells examined (Ntotal=77 to 311) is plotted as a percentage againstthe duration of treatment with CB at 100 µg/ml. Each pointrepresents the mean value from 4 to 11 leaf pieces with about 20 to30 cells in each. The vertical bars show the s.d. for each value. Datawere obtained 24 hours after rmoval of CB.

bundles of MFs, we examined the direction of cytoplasmicstreaming in each cell before and after treatment with CB.

As shown in Fig. 10, there was a reversal of the direction ofstreaming in 13% of cells upon the removal of CB after a 6hour treatment with CB at 100 µg/ml. This percentage of cellsgradually increased such that, after removal of CB subsequentto prolonged treatment with CB for 24 hours, 23% of cellsshowed reversal of the direction of streaming. After treatmentwith CB for 24 hours, the bundles of MFs at the side walls hadbeen completely disrupted (Fig. 2D), while the bundles at theend walls were only partially disrupted (Fig. 4C). Although thepercentage increased slightly to 32% after 48 hours oftreatment with CB and subsequent removal of CB, it neverreached 50% (see Discussion). Treatment with CB for morethan 48 hours had a lethal effect on the cells.

The effects of trypsin and CB on the reorganizationof MFs and reinitiated cytoplasmic streaming To remove completely the MFs at the end walls that appearedto be resistant to CB, we pretreated cells with an exogenousprotease, namely trypsin, before treatment with CB, sinceMasuda et al. (1991) had shown that exogenously appliedprotease disrupts the ordered arrangement of bundles of MFsin isolated mesophyll cells.

Fig. 11A shows an example of bundles of MFs at the endwall that had been incompletely but considerably disruptedafter pretreatment with trypsin for 30 minutes. As expected,the bundles were completely disrupted by subsequenttreatment with CB for 24 hours (Fig. 11B). Treatment of cellswith trypsin for 10 minutes or 20 minutes resulted in incom-plete disruption of MFs at the end walls even with subsequenttreatment with CB (data not shown).

To determine the effect of the complete disruption of thebundles of MFs at the end walls on the process of reorganiza-tion of the MFs, we monitored the reappearance of MFs atgiven times after the removal of CB. The patterns of arrange-ment of reassembled bundles of MFs could be categorized intofour types, as shown schematically in Table 1. In type 1, shortbundles of MFs often encircled the chloroplast. Some of themseemed to extend into the cytoplasmic matrix. The arrange-ment of MFs was very similar to that shown in Fig. 6A. Cells(Nmf) of type 1 decreased with increased duration of washingwith APW. In type 2, bundles of MFs were arranged almostparallel to each other but were still partially fragmented. Intype 3, the arrangement of the bundles of MFs was almostnormal, as shown in Fig. 4A. Nmf of type 3 increased withincreased duration of washing with APW. In type 4, a circulararrangement of bundles of MFs was seen. Nmf of type 4 wasconstant, 1-2%, throughout extensive washing with APW.About 20% of cells were unstained by FITC-phalloidin in suchexperiments. This fraction might contain cells that had beenpermeated insufficiently with FITC-phalloidin and/or cells thatdied during the experiment. Nmf of type 3 reached 41.2% after72 hours of washing. The value of 41.2% is small if wecompare it with that obtained after treatment with CB only.Moreover, much more time was required for the reappearanceof the normal pattern, while in cells treated with CB only, 48hours at the most were required.

Next, we counted the number of cells in which cytoplasmicstreaming had been reinitiated at the end walls 72 hours afterthe removal of trypsin and CB. The results are summarized inTable 2. The number of cells in which cytoplasmic streamingcould be observed with a normal pattern was 67% of Ntotal.This value coincides well with the results shown in Table 1, ifwe assume that the arrangement of bundles of MFs of type 2can provide tracks for resumed cytoplasmic streaming. Someof the leaf pieces in which we observed the normal pattern ofreinitiated cytoplasmic streaming at the end walls of cells werecarefully further trimmed with a razor blade so as to expose

Fig. 11. The bundles of MFs at the end wall.(A) After treatment with trypsin for 30minutes. The MFs are partially disrupted.(B) After treatment with trypsin for 30minutes and then with CB for 24 hours. TheMFs are completely disrupted. Bar, 20 µm.

1538 J.-H. Ryu, S. Takagi and R. Nagai

Table 1. Reorganization of the bundles at MFs in the end walls after removal of trypsin and CBArrangement of re-formed bundles of MFs

No MF Type 1 Type 2 Type 3 Type 4

Time afterremoval oftrypsin and CB(h) Nmf (Nmf/Ntotal × 100) Ntotal24 117 (23.3)* 234 (46.6) 109 (21.7) 38 (7.6) 4 (0.8) 50248 108 (19.3) 197 (35.1) 157 (28.0) 94 (16.8) 5 (0.9) 56172 122 (21.6) 79 (14.0) 122 (21.6) 233 (41.2) 10 (1.8) 566

Cells were first treated with trypsin for 30 minutes and then with CB for 24 hours.The numbers of cells with MFs of each type (Nmf) were counted at given times after the removal of CB.*Values are %.

Table 2. Patterns of reinitiated cytoplasmic streaming atthe end walls after removal of trypsin and CB

Normal NoSame* Reversed† Abnormal streaming

Experiment (%) (%) (%) (%) Ntotal

1 46.2 41.0 12.8 0.0 392 23.5 29.4 5.9 41.2 343 30.2 34.9 16.3 18.6 434 30.4 17.4 8.7 43.5 235 38.9 38.9 11.1 11.1 186 38.5 38.5 7.7 15.4 137 35.7 25.0 10.7 28.6 28

Average 34.8 32.2 10.5 22.6s.d. ±6.9 ±8.0 ±3.2 ±14.8

*Same indicates the percentage of cells in which the direction of reinitiatedstreaming was the same as that in cells before treatment with trypsin and CB.

†Reversed indicates the percentage of the cells with a reversed direction ofstreaming.

the side wall and/or the periclinal walls of cells. In most cases,we confirmed by light microscopy that a completely normalpattern of rotational cytoplasmic streaming had been recon-structed along the entire length of such cells. Therefore, itseems reasonable to conclude that, when the normal arrange-ment of the bundles of MFs has been reconstructed at one ofthe end walls of a cell, the complete reconstruction of thebundles of MFs at the other three anticlinal walls has beenaccomplished. About 10% of cells exhibited an abnormalpattern of cytoplasmic streaming, which might have been theresult of the arrangement of MFs of types 1 and 4. Only 23%of the cells examined did not show evidence of reinitiatedstreaming, suggesting that the effects of trypsin and CB aremostly limited to the MFs. In the case of cells with a normalpattern of cytoplasmic streaming, the percentage of cells inwhich the direction of streaming had been reversed was 48.1%,indicating that the bundles of MFs had been reconstructedindependently of the original polarity. Thus, we concluded thatthe contribution of MFs that had remained undisrupted whencells were treated with CB only had been eliminated.

DISCUSSION

The present study reveals that bundles of MFs at the side wallsdisappear via local fragmentation of the bundles of MFs during

a 24 hour treatment with CB (Fig. 2). By contrast, many of thebundles of MFs at the end walls remained even after a 48 hourtreatment (Fig. 4C). Thus, the bundles of MFs at the end wallswere much more resistant to the disruptive action of CB thanthose at the side walls. MFs at the corners of the cell, wherean end wall and a side wall meet, were also resistant to CB(Fig. 2C,D). The role of the stable MFs could be deduced fromthe effects of CB on the direction of reinitiated cytoplasmicstreaming. The number of cells in which the direction ofstreaming had been reversed after treatment with CB and itssubsequent removal was always less than 50% of the treatedcells, as far as we could determine (Fig. 10). If the value hadreached 50%, as it does in the case of mesophyll cells of V.asiatica (Ishigami and Nagai, 1980), it would be reasonable toassume that the bundles of MFs became completely disorgan-ized in the presence of CB, and that the reassembly of eachbundle occurred independently of the original polarity of MFs.This assumption was strongly supported by the result that thepercentage of cells in which the direction of reinitiated cyto-plasmic streaming had been reversed reached 48.1% (Table 2)when the bundles of MFs at the end walls had been completelydisrupted as a consequence of treatment with both trypsin andCB. From our results, it appears that the bundles of MFs in theectoplasm are anchored and stabilized at the end walls ofmesophyll cells of V. gigantea. In addition, we found that thestabilized bundles of MFs play a role in determining thepolarity of reorganized bundles of MFs after treatment withCB, and that trypsin impairs, at least partially, component(s)that may contribute to the anchoring of the MFs at the endwalls so that these MFs lose their resistance to the action ofCB. The effect of trypsin was demonstrated by the followingfindings. (1) The reconstruction of the bundles of MFs at theend walls required much more time after pretreatment withtrypsin and the removal of CB than that it did in cells treatedwith CB only. (2) In a considerable number of cells, thearrangement of the re-formed bundles of MFs remainedabnormal even 72 hours after the removal of CB (Table 1).

After removal of CB, the first signs of reassembly of MFsat the side walls were detected in the vicinity of the chloro-plasts (Fig. 6A,B). During the subsequent process of reassem-bly, we often observed entangled masses of MFs in regionsrich in cytoplasm, as well as fine networks of MFs throughoutthe cytoplasm (Fig. 6C,F,H). These transient configurationssuggest that the MFs polymerize rather randomly at numeroussites in the cytoplasm. There seems to be no specific site at theside walls that acts as a focus for the reorganization of MFs.

1539Actin cytoskeleton in Vallisneria

Nevertheless, the reconstruction of bundles of MFs occurredas if the bundles were arranged along the entire length of theoriginal bundles.

Re-establishment of the parallel arrangement of bundles ofMFs is not a simple process, and its molecular mechanism isunknown. However, the process may be composed of severalsteps, as follows. (1) Randomly oriented repolymerization ofglobular (G-) actins may occur, to generate F-actins in regionsrich in cytoplasm and, thus, probably rich in unpolymerizedactin. The ionic strength of the cytoplasm appears to be highenough for the polymerization of G-actin to F-actin (Macklon,1975; Pierce and Higinbotham, 1970; Tazawa, 1972). (2) Next,the annealing of short actin filaments occurs. Direct examina-tion by electron microscopy has confirmed that short actinfilaments can rapidly anneal in an end-to-end manner in vitroand that, during annealing, the structural polarity of newlyformed actin filaments is absolutely conserved (Murphy et al.,1988). (3) Bundling of actin filaments occurs. The pattern ofoptical diffraction of the bundles of MFs from V. gigantea isalmost identical to that of paracrystalline bundles of actinfilaments from muscle. There seem to be no cross-linkingmolecules between the MFs (Yamaguchi and Nagai, 1981).Paracrystalline bundles of actin filaments are known to be easilyformed in the presence of Mg2+ (Hanson, 1973). Therefore,once actin filaments have re-formed, it may be easy for thesefilaments to generate bundles. (4) Short bundles of MFs annealto one another. Although no supporting evidence exists in theliterature, to our knowledge, the patterns of staining of MFswith FITC-phalloidin during the reassembly of bundles (Fig. 6)suggest that shorter bundles of MFs become longer ones. Inaddition, when cells were irradiated with light about 12 hoursafter the removal of CB, the cytoplasm was initially translo-cated only a very short distance and the pattern of cytoplasmicstreaming was very unstable. The tracks for streaming appearedto become longer during the course of irradiation. Such obser-vations indicate that, although shorter bundles of MFs at theside walls had not yet stabilized at this time, they could providetracks for local streamlets. These local streamlets, in turn,would help the annealing of short bundles of MFs by increas-ing the chances that bundles would encounter one another. (5)The re-formed bundles become tethered to the stable bundles atthe end walls. The bundles of MFs that have re-formed at theside walls might be annealed to stable bundles at the end wallsthat had remained undisrupted during the long-term treatmentwith CB. Our finding, described above, that the direction ofreinitiated cytoplasmic streaming was determined by theoriginal direction (Fig. 10), tends to support this hypothesis.

In summary, our results strongly support the proposal byMasuda et al. (1991) that bundles of MFs, which serve as thetracks for rotational cytoplasmic streaming, are preferentiallyanchored and stabilized in the ectoplasm at the end walls, andthat the anchoring of bundles is crucial for the maintenance ofthe unique, stationary cytoskeletal organization of themesophyll cells of V. gigantea.

This work was supported in part by a Grant-in-Aid for ScientificResearch (no. 05454018) to R.N. from the Ministry of Education,Science and Culture, Japan.

REFERENCES

Gunning, B. E. S. and Hardham, A. R. (1982). Microtubules. Annu. Rev.Plant Physiol. 33, 651-698.

Hanson, J. (1973). Evidence from electron microscope studies on actinparacrystals concerning the origin of the cross-striation in the thin filamentsof vertebrate skeletal muscle. Proc. Roy. Soc. Lond., Ser. B 183, 39-58.

Haupt, W. (1959). Photodinese. Encyclop. Plant Physiol. 17, 388-398. Haupt, W. (1982). Light-mediated movement of chloroplasts. Annu. Rev. Plant

Physiol. 33, 205-233. Ishigami, M. and Nagai, R. (1980). Motile apparatus in Vallisneria leaf cells.

II. Effects of cytochalasin B and lead acetate on the rate and direction ofstreaming. Cell Struct. Funct. 5, 13-20.

Kakimoto, T. and Shibaoka, H. (1987). Actin filaments and microtubules inthe preprophase band and phragmoplast of tobacco cells. Protoplasma 140,151-156.

Kamiya, N. (1959). Protoplasmic streaming. Protoplamstologia 8/3a, (ed. L.V. Heilbrunn and F. Weber). Springer, Wien.

Kamiya, N. (1981). Physical and chemical basis of cytoplasmic streaming.Annu. Rev. Plant Physiol. 32, 205-236.

Kersey, Y. M., Hepler, P. K., Palevitz, B. A. and Wessells, N. K. (1976).Polarity of actin filaments in characean algae. Proc. Nat. Acad. Sci. USA 73,165-167.

Kuroda, K. (1990). Cytoplasmic streaming in plant cells. Int. Rev. Cytol. 121,267-307.

Lloyd, C. W. (1982). The Cytoskeleton in Plant Growth and Development.Academic Press Inc., London, New York.

Lloyd, C. W. (1991). The Cytoskeletal Basis of Plant Growth and Form.Academic Press Inc., London, New York.

Macklon, A. E. S. (1975). Cortical cell fluxes and transport to the stele inexcised root segments of Allum cepa L. Planta 122, 109-130.

Masuda, Y., Takagi, S. and Nagai, R. (1991). Protease-sensitive anchoring ofmicrofilament bundles provides tracks for cytoplasmic streaming inVallisneria. Protoplasma 162, 151-159.

Menzel, D. (1992). The Cytoskeleton of the Algae. CRC Press, Boca Raton, FL. Murphy, D. B., Gray, R. O., Grasser, W. A. and Pollard, T. D. (1988).

Direct demonstration of actin filament annealing in vitro. J. Cell Biol. 106,1947-1954.

Nagai, R. (1993). Regulation of intracellular movements in plant cells byenvironmental stimuli. Int. Rev. Cytol. 145, 251-310.

Palevitz, B. A., Ash, J. F. and Hepler, P. K. (1974). Actin in the green alga,Nitella. Proc. Nat. Acad. Sci. USA 71, 363-366.

Palevitz, B. A. and Hepler, P. K. (1975). Identification of actin in situ at theectoplasm-endoplasm interface of Nitella. Microfilament-chloroplastassociation. J. Cell Biol. 65, 29-38.

Pierce, W. S. and Higinbotham, N. (1970). Compartments and fluxes of K+,Na+ and Cl− in Avena coleoptile cells. Plant Physiol. 46, 666-673.

Seitz, K. (1979). Cytoplasmic streaming and cyclosis of chloroplasts.Encyclop. Plant Physiol. (n. s.) 7, 150-169. Springer-Verlag, Berlin.

Takagi, S. and Nagai, R (1983). Regulation of cytoplasmic streaming inVallisneria mesophyll cells. J. Cell Sci. 62, 385-405.

Takagi, S. and Nagai, R. (1985). Light-controlled cytoplasmic streaming inVallisneria mesophyll cells. Plant Cell Physiol. 26, 941-951.

Tazawa, M. (1972). Membrane characteristics as revealed by water and ionicrelations of algal cells. Protoplasma 75, 427-460.

Yamaguchi, Y. and Nagai, R. (1981). Motile apparatus in Vallisneria leafcells. I. Organization of microfilaments. J. Cell Sci. 48, 193-205.

(Received 14 April 1994 - Accepted 9 December 1994)