Cytoplasmic streaming in Chara: A cell model activated by atp and inhibited by cytochalasin b

The endoplasm of characean algae streams at 40 μm s⁻¹ or more. Bundles of microfilaments situated at the boundary between the stationary cortical cytoplasm and the flowing endoplasm are believed to have a role in the production of the motive force for streaming (Nagai & Rebhun, 1966; Pickett-Heaps, 1967; Bradley, 1973). Microtubules are not found in positions from which they could contribute to the motive force for streaming (Nagai & Rebhun, 1966) and depolymerizing them with colchicine does not inhibit the streaming (Pickett-Heaps, 1967).

Three types of study have contributed to our knowledge of the microfilaments and of the way in which streaming is driven. In light-microscope studies (Kamitsubo, 1966,1972), fibres have been observed whose position and orientation strongly suggests that they are the microfilament bundles. In areas where streaming was recovering from damage by centrifugation, a relationship was evident between the reformation of fibres and the resumption of streaming. Organelle movements closely followed the irregular path of the newly formed fibres and occurred only in close proximity to them. More recently it has been suggested (Allen & Allen, 1972; Allen, 1974) that the fibres seen by Kamitsubo -which are fixed, rigid structures -serve only a skeletal function as an anchor for other fibres which project into the endoplasm. Allen believes that streaming is the result of the propagation of waves of bending along these endoplasmic fibres.

A second type of study has used the inhibitor cytochalasin B. Many of the processes initially shown to be sensitive to the inhibitor were thought to be dependent on the functioning of microfilaments, and in many cases the microfilaments were disrupted by cytochalasin (Wessels et al. 1971). The inhibition of streaming in intact characean cells (Wessels et al. 1971 ; Williamson, 1972; Bradley, 1973) and of motility in isolated cytoplasmic fragments (Williamson, 1972) was consistent with the view that cytochalasin in some way affected microfilament functioning. However, many cases are now known in which cytochalasin inhibits specific membrane-transport systems (references in Pollard & Weihing, 1974). The desire to explain all effects of cytochalasin in terms of a single site of action has led to a tendency to regard the plasma membrane as the primary site of action for the inhibitor. The effects on microfilaments could then be due to resultant alterations in the ionic composition of the cytoplasm (Estensen, Rosenberg & Sheridan, 1971) or to the disruption of links between membranes and microfilaments (Spooner, 1973; Hepler & Palevitz, 1974).

The third approach has involved the demonstration that filaments from two characean species (Nitella flexilis:Palevitz, Ash & Hepler, 1974; Chara coralliner. Williamson, 1974) react with subfragments of muscle myosin to produce arrowhead filaments. This indicates strong similarities between the algal filaments and actin (see references in Williamson, 1974). Both papers furnished circumstantial evidence that the actin filaments were components of the microfilament bundles and definitive evidence is provided by the in situ decoration of the microfilaments of a glycerinated cell (unpublished, but quoted by Palevitz et al. 1974).

A further approach to studying problems of motility involves the use of cell models (see Arronet, 1973, for review). These are systems in which the structures responsible for motility are preserved in an organized but inactivated state while the cell is made permeable by chemical or mechanical disruption of the plasma membrane. The conditions controlling the operation of the motile cell components (energy source, ionic environment, etc.) can then be defined by experiments to restore motility to the model. Glycerinated muscle fibres which contract in the presence of ATP and Ca²⁺ ions are perhaps the most familiar example of a cell model. Such models are also of value in separating inhibitors of the in vivo motile process into those affecting directly the contractile elements, which also inhibit the model, and those inhibiting motility secondarily by, for example, interference with the energy supply. The latter type are not inhibitory to the operation of the model (see Arronet, 1973, p. 48). The only application of such methods to green plant cells seems to have been the glycerination of Acetabularia calyculus (Takata, 1961), where the addition of 1 mM ATP with 1 mM Ca or Mg salts was reported to restore transient streaming.

Several studies of streaming in characean cells have employed the technique of vacuolar perfusion (Kamiya & Tazawa, 1966; Tazawa & Kishimoto, 1964, 1968; Tazawa, 1968; Donaldson, 1972). In such experiments both ends of the cell are cut off in conditions which permit streaming to continue unimpaired. Experimental solutions can then be passed through the large central vacuole under the influence of a small pressure gradient. The flowing solution carries with it much of the endoplasm. In the present study of Chara, attention has been concentrated on that fraction of cytoplasm not washed out with the perfusing solution and which has been shown to constitute a cell model. Observations have been made on the role of the microfilament bundles in streaming, the nature and control of their interactions with endoplasmic organelles, and of the effects of cytochalasin B on such interactions. The observations are interpreted in the light of the presence in these cells of actin filaments (Williamson, 1974).

The apparatus (Fig. 1) was designed to allow the process of perfusion to be observed with an oil-immersion objective. The principles of the method were exactly similar to those of previous studies (references in the Introduction).

An internodal cell 40-70 mm in length was blotted and placed on a microscope slide. Its ends were sealed with grease into 2 glass rings (16 mm diameter, 9 mm tall), each with a groove in its base through which the cell could pass without damage. After about 60 s the rings were filled with perfusion solution and the central part of the cell covered with liquid paraffin. As found by Tazawa (1968), the organization of the cell was preserved more successfully with liquid paraffin than with an isotonic aqueous solution. Intact Chara cells continue to stream for several weeks when immersed in liquid paraffin (unpublished results). A coverslip supported at each comer with a small amount of grease was placed over the central portion of the cell.

After placing the assembled apparatus on the stage of the microscope, the levels of fluid in the 2 rings were equalized by eye. By inserting scissors into the rings, the 2 ends of the cell were successively removed. Any slight flows of perfusion fluid could be seen by observing the movements of the vacuolar bodies with a low-power objective. Solution was removed with a syringe from the appropriate ring until the vacuolar bodies were being swept with the endoplasm in the normal manner.

Sucrose was found in early experiments to prolong streaming against the flow of the perfusion solution and to improve the subsequent preservation of the chloroplasts. It was therefore used with K⁺ (added either as KC1 or as K₂EGTA-ethyleneglycol bis-tetra-acetic acid) as the main osmotic component of the solutions. These were approximately isotonic with 350 mM sucrose. The level of free Ca²⁺ ions was controlled with EGTA, which has a much higher affinity for Ca²⁺ than for Mg²⁺ and can therefore be used to give low, buffered levels of free Ca²⁺ ions (Portzehl, Caldwell & Ruegg, 1964). Solutions were prepared to give known levels of free Ca²⁺ and Mg²⁺ ions, allowing for the binding of both ions to EGTA and ATP. Solutions containing ATP were prepared to have the same level of free Ca²⁺ and Mg²⁺ ions as the ATP-free solution they replaced.

Cytochalasin B (Imperial Chemical Industries, Pharmaceutical Division) was dissolved at 5 mg ml⁻¹ in dimethyl sulphoxide.

Chara corallina was grown in the laboratory, rooted either in mud covered with artificial pond water (1 mM NaCl, 0 ·1 mM CaCl₂, 0 ·1 mM KC1) or in agar (Sandan, 1955) with a modified Forsberg medium II (Forsberg, 1965).

A Zeiss Universal Research Microscope with differential interference-contrast optics was used for all observations.

As found in previous applications of the perfusion technique, streaming continued without significant alteration after the ends of the cell had been removed. The cytoplasm remained as a thin sleeve around the large, central vacuole.

The applied pressure difference when solution was removed from one of the rings (about 7 mm of perfusion fluid) caused a rapid flow of perfusion solution through the vacuole which carried with it the bulk of the endoplasm. The direction of this flow was routinely arranged to be opposite to the direction of streaming in the area being observed. It has been observed previously when the bulk of the endoplasm is moving passively under the influence of centrifugation (Hayashi, 1957) or of perfusion (Tazawa, 1968; Donaldson, 1972) that some organelles just beneath the chloroplasts continue to move forwards. These movements were stopped only by the applications of much larger forces, greater than any applied in the present study. Using the improved optical conditions and the presence of considerable lengths of clear fibres (Fig. 3), it was possible in the present study to see that these persistent forward movements were closely associated with the surface of the fibres. (It should be noted that these are fibres of the type described by Kamitsubo, that is, situated just beneath the chloroplasts and showing no bending movements. No evidence of endoplasmic fibres of the type described by Allen or of the characteristic file of organelles indicating their propagation of bending waves has been found in this study.)

The forward-moving organelles seen in cells undergoing perfusion appeared to form a single file along the surface of the fibre. Their forward movements continued even when organelles to the side and beneath were being swept backwards. Members of the forward-moving file were from time to time swept away with the backward-flowing endoplasm. The association of the forward movements with the fibres was seen most clearly when a fibre was oriented obliquely to the main direction of streaming, usually as the result of an abrupt bend (Fig. 2). Organelle movements faithfully followed the deviations of such fibres.

Routinely the cell was first perfused with a solution of salts and sucrose lacking ATP. The exact composition of this solution had, within the limits tested, no major effects on events prior to reactivation (see below). The forward movements described above, lasting for periods of up to 60 s, continued until very little remained of the endoplasm flowing with the perfusion fluid. The forward-moving organelles then abruptly ceased moving and became anchored to the fibre along which they had been travelling (compare Figs. 3 and 4). In this inactive state (no added ATP), very few of these organelles could be dislodged from the fibres even by rapid and prolonged perfusion.

The ATP-free solution used for the initial perfusion was removed from the 2 glass rings and a solution identical but for the presence of 1 IDM Na₂ATP was added to one ring. (The addition was made so that the direction of flow of the ATP solution was the same as that of the initial perfusion.) ATP was routinely added 60 s after entry into the inactive state. The response of the preparation to ATP had 2 components: the resumption of organelle movements and the loosening of the linkages holding the organelles to the fibres.

The conditions affecting the velocity of the reactivated streaming will be discussed in detail below, but in vivo rates could be obtained immediately after reactivation under quite a range of the conditions tested. The direction of the reactivated streaming was always the same as the in vivo direction. The movements were intimately associated with the fibres; this was seen most clearly where gaps between chloroplasts were spanned by single fibres and where a fibre possessed an abrupt bend (Fig. 2). Organelle movements were confined to the fibre and followed any deviations in its track. Movements could involve single organelles or cytoplasmic fragments -that is, groups of organelles moving as a unit, not necessarily all in contact with the fibre. Under conditions giving high rates of streaming, organelles moved smoothly over distances greater than 100 μm. When the movements were slower, the progress of an organelle was often discontinuous, movements of a few to several tens of micrometres being interspersed with periods of Brownian motion near the fibre. The resumption of active movements was apparently dependent on renewed contact with the fibre. Simultaneous movements of organelles at different rates along the same fibre were observed. The maximum duration of motility was about 50 min and the velocity of the movements declined during this period. As in the intact cell, no movements of the fibres themselves were seen.

The second characteristic response to ATP was the release of organelles from the tight binding to the fibres which characterized the inactive state. Many organelles could be swept away even by very gently flowing solutions containing ATP. For this reason, ATP was introduced under a very small pressure difference and this was removed shortly after reactivation had occurred. The number of organelles undergoing movements associated with the fibres declined noticeably with time, more and more organelles being found free in the central vacuolar space. A consequence of this was that the fibres came to have an extremely clean appearance (compare Figs. 5 and 6; see also Fig. 2).

If ATP were included in the initial perfusion solution, no condition corresponding to the inactive state was observed. The few organelles remaining near the fibres continued moving and were easily swept away by the perfusion solution to leave the very clean type of fibre.

Very low rates of movement were obtained with any of the solutions tested at pH 6 ·0 (10 mM morpholinoethane sulphonic acid as buffer). At pH 8 ·0 (10 mM Tris buffer), no rates higher than 30μm s⁻¹ were obtained. The most thorough study of reactivation was therefore made at pH 7-0 (10 mM piperazine-N,N′-bis-2-ethane sulphonic acid). Here the full in vivo rate of streaming (50 μmS⁻¹) could be obtained subject to the following conditions: (i) The level of free Ca²⁺ ions should be 10⁻⁷ M or less. No evidence of a requirement for Ca²⁺ ions could be found, full rates of streaming being obtained even with 50 mM EGTA and no added calcium, 10⁻⁶ M Ca²⁺ produced an inhibition of about 20%; 10⁻⁶ M and above produced an inhibition of about 80%. Qualitatively, the effects of the various Ca²⁺ levels were similar at pH 8 ·0; all velocities were, however, lower than under equivalent conditions at pH 7 ·0. (ii) The level of free Mg²⁺ ions should be at least 1 mM, which would be in equilibrium with a MgATP level of 0 ·9 mM. 0 ·1 mM free Mg²⁺ (0 ·6 mM MgATP) was slightly inhibitory, the inhibition being almost total with no added Mg. (iii) The level of Cl⁻ ions should be 80 mM or less, no mM being strongly inhibitory.

Changing the K⁺ level between 35 and 170 mM (with compensating alterations in the sucrose concentration) had no significant effect on the velocity of the reactivated streaming. More extreme K⁺ levels were not investigated.

ADP and AMP. With 200 mM sucrose, 50 mM EGTA and 4 mM free Mg²⁺, streaming was reactivated with a velocity of 50 μm s⁻¹ by 1 mM ATP. Replacing the ATP with 1 mM ADP resulted in velocities not exceeding 11 μm s⁻¹. AMP (1 and 10 mM) produced neither movement nor dissociation of organelles from the fibres.

Pyrophosphate. Ten mM Na₄P₂O₇ was substituted for the adenine nucleotides tested above. No movements at all resulted, but the progressive release of organelles from tight binding to the fibres was evident. After some 10-15 min, this left extremely clean fibres similar to those seen after ATP treatment (see Figs. 2 and 6). A control experiment with 20 mM Na₂HPO₄ produced neither release nor movement of the organelles. One mM pyrophosphate did not give completely clean fibres.

Cytochalasin B. With the sucrose, EGTA and Mg²+ levels used to test ADP and AMP, cells were perfused with solutions lacking both ATP and cytochalasin. As soon as the inactive state had been reached, perfusion was continued with a solution identical but for the presence of cytochalasin B; 60 s after entry to the inactive state, a solution containing both cytochalasin and 1 mM ATP was perfused in the usual manner for reactivation. With 25 μg ml⁻¹ cytochalasin, both the release of the organelles and their movements were almost completely abolished. Movements did not usually exceed about 20 pm in total before stopping completely, and the fibres remained heavily coated with organelles; 10μg ml⁻¹ cytochalasin caused considerable but incomplete inhibition of both processes. The effects of 25 μg ml⁻¹ cytochalasin were partially reversed by subsequent perfusion with a solution containing an identical level (0·5%) of dimethyl sulphoxide. As the speeds obtainable on reactivation decline somewhat with time spent in the inactive state, the fact that full rates of streaming were not obtainable on washing out the cytochalasin may simply reflect the extra delay in such experiments.

Two observations in this study point to the generation of the motive force being intimately related to the microfilament bundles forming the fibres first described by Kamitsubo. (That the fibres in these particular cells consist of microfilaments has been confirmed by unpublished electron micrographs.) Firstly, in cells undergoing perfusion, organelles close to the fibres continue to move forwards while the rest of the endoplasm is moving backwards. In such a situation, forward movements must depend on a locally generated force and cannot be explained by forces generated elsewhere and transmitted by the viscosity of the endoplasm. Secondly, in the reactivated cell lacking most of its endoplasm, movements are very obviously associated with the microfilament bundles. This is most convincing with the movements of single organelles along the lengths of fibres between well spaced chloroplasts and where such fibres follow an irregular course with bends.

Two effects of restoring ATP to the Chara model were apparent: organelles formerly tightly bound to a fibre moved along it, but showed additionally an increased tendency to be released from contact with it. A plausible explanation of these results can be advanced in terms of actin filaments anchored in the cortex (Williamson, 1974) interacting with an as yet uncharacterized myosin-like component which can link the actin to endoplasmic organelles.

Thus in the inactive state produced by the absence of ATP, actin and myosin would be in rigor combination, linking the organelles tightly to the fibres. Now ATP has a dual effect on a muscle in rigor, causing by its binding to myosin the detachment of the cross-bridges and by its hydrolysis their cyclical interaction with actin to produce movement (see, for example, the paper of Reedy, Holmes & Tregear, 1965). In the Chara system the binding of ATP to detach linkages between the fibres and the endoplasmic organelles would tend to release the latter; the hydrolysis of ATP could power cyclical movements of the same linkages causing the organelles to move along the fibre just as thin and thick filaments move past each other in muscle.

A more critical test of the theory involves separating the effect caused by the binding of ATP from the effect caused by its hydrolysis. In muscle, this can be done by inhibiting the actomyosin ATPase by removing Ca²⁺ ions; this results in the released bridges remaining detached because the thin filaments are turned off (Reedy et al. 1965; Huxley, 1968). With the Chara model, this is ineffective as there is no evidence for a Ca²⁺ requirement. A similar effect can also be achieved in muscle by using a non-hydrolysable analogue of ATP; for example, inorganic pyrophosphate causes significant detachment of the myosin cross-bridges (Lymn & Huxley, 1972). This approach does separate the 2 effects in Chara, allowing detachment to occur in the absence of any movement. While pyrophosphate is a less-effective dissociating agent than ATP, inorganic phosphate is completely without effect. This indicates the significance of the pyrophosphate linkage irrespective of its capacity to be hydrolysed.

Until Chara myosin is identified and its subcellular location established, the theory cannot be fully tested. We do, however, know that rabbit myosin is dissociated from Chara actin by ATP (Williamson, 1974), so that it would be surprising if Chara myosin were not. The observation that all the arrowheads on a single bundle of actin filaments point in the same direction (Palevitz et al. 1974) is also consistent with this theory, as the direction of the arrowheads would specify the direction of myosin movement and therefore cause unidirectional organelle movements.

The theory is a particular application of the ‘active shear’ mechanisms previously discussed in general terms (Huxley, 1963, 1973; Wolpert, 1965; Jahn & Bovee, 1968). The essential point stressed by these authors was that widely different forms of motility could result from the assembly in different ways of proteins essentially similar to those of muscle. The present study demonstrates for the first time that motile organelles can link to the fibres in a way suggesting that a myosin-like protein is involved.

Organelles released from the fibres by ATP or pyrophosphate can disperse freely in the central vacuolar space of the perfused cell. There can thus be no continuous membrane of the tonoplast type. Indeed the tonoplast must be carried along with the perfusing solution in order to transmit the flow of that solution to the endoplasmic organelles. This lack of tonoplast presumably explains the sensitivity to ionic conditions shown by the reactivated streaming when compared to the lack of sensitivity shown by the streaming cytoplasm before full perfusion.

Of the various conditions controlling the reactivation of motility, perhaps the most interesting is that the Ca²⁺ level be kept at or below 10⁻⁷ M for maximal velocities to be achieved. This was found to be the case with wide variations in the K⁺, Mg²⁺ and sucrose levels, as well as at pH 8-o. This contrasts sharply with the requirements for actin-myosin interaction in muscle (see Weber & Murray, 1973) where interaction is blocked in the resting muscle by the low level of free Ca²⁺ ions (< 10⁻⁷ M). Contraction is then triggered by a rise in the level of free Ca²⁺ ions. Evidence concerning the regulation of actin-myosin interaction outside muscle is limited, much but not all of it pointing to regulation by Ca²⁺ ions (Pollard & Weihing, 1974). It may be noted that Ca²⁺ sensitivity is not intrinsic to actin-myosin interaction but is conferred by additional polypetides associated with either the actin or the myosin component (see Weber & Murray, 1973). Control might therefore be expected to show more variation than the basic mechanical process, and this might particularly be the case in comparing a continuously active system such as streaming in Chara, with systems showing on-off control such as muscle. Data either on other Chara model preparations (glycerol or detergent extracted) or on the Chara actomyosin ATPase itself will help to establish the significance of the present results.

No data are available on the level of free Ca²⁺ ions in plant cytoplasm. Total Ca²⁺ levels in the cytoplasm of Nitella transitions have been reported in the millimolar range (Spanswick & Williams, 1965) but, as in animal cells, much of this could be retained in membrane-bound organelles (endoplasmic reticulum, mitochondria, vacuoles). The need for millimolar levels of Ca²+ in the perfusion solution reported in earlier studies with N.flexilis (Tazawa & Kishimoto, 1964) reflects the level needed in the vacuole for prolonged, normal operation of the endoplasm. (Tazawa & Kishimoto used a gentle perfusion technique to leave the endoplasm intact and functional.) The much lower levels found in the present experiments are likely to be more relevant to the in vivo levels of free Ca²⁺ in the endoplasm itself.

The conditions under which the model was reactivated appear reasonably physiological. Thus characean cytoplasm is generally poor in Cl⁻ ions, which were found to inhibit motility at high concentrations, and rich in K⁺ ions, the main cation used in the reactivation experiments (see table 1 in MacRobbie, 1970, for a summary of ionic levels in characean cytoplasm). The total Mg²⁺ level in the flowing cytoplasm of these Chara cells was measured as 3· 6 mM (unpublished results). The ATP level in Nitella was measured as 0· 04 mM on a total cell volume basis (Hatano & Nakajima, 1963), which would be close to 1 mM if contained exclusively in a cytoplasmic compartment occupying about 5% of the total volume.

The low rates of movement supported by exogenous ADP may be due to its conversion to ATP by, for example, an adenylate kinase enzyme in the residual cytoplasm. The specificity of the presumptive myosin cannot be satisfactorily studied in the present system.

At 25 μtg ml⁻¹, cytochalasin B causes a rapid and complete inhibition of streaming by ATP. Applied externally to an intact Chara cell, such total inhibition of movements near the fibres requires periods longer than 60 min, even with 50 μg mb¹ (unpublished results). Thus the effectiveness of cytochalasin is reduced if it has to work from outside the cell, suggesting that the plasma membrane is a barrier to inhibition rather than its mediator. (See de Laat, Luchtel & Bluemink, 1973, for a similar conclusion in regard to egg cleavage.) Furthermore, inhibition is rapid and total even if the energy supply and ionic conditions around the filamentorganelle system are regulated as in these reactivation experiments. The observations strongly suggest that cytochalasin is exerting a fairly direct effect on the filamentorganelle system rather than one mediated by the plasma membrane or by changes in ionic environment.

Following treatment with ATP and cytochalasin, the organelles remain predominantly linked to the fibres, indicating that their ATP-induced dissociation has been blocked. Careful observation of Chara cells inhibited by externally applied cytochalasin has revealed that some immobile organelles become attached for long periods to the fibres. On subsequent perfusion, many of these organelles are not swept away with the perfusing solution, but few new ones are added to the fibres in the way that happens in the uninhibited cell on entry to the inactive state (unpublished results). It may be that cytochalasin both reduces the affinity of the organelles for the fibres and the ability of ATP to dissociate the fibre-organelle linkages. Few linkages would then form, but those that did would be more stable to dissociation by ATP. A more refined in vitro system will be needed to test these ideas.

The present results provide experimental evidence for the role of the microfilament bundles in streaming and the first evidence about the nature of their interaction with endoplasmic organelles. The existence of organelle-fibre linkages which are stable in the absence of ATP is expected only with the active shear mechanism described here and not with the mechanism dependent on wave propagation (Allen, 1974). This does not necessarily preclude the operation additionally of a wave propagation system, although theoretical treatments (Donaldson, 1972) favour force generation being limited to that narrow zone just beneath the chloroplasts where active shearing would take place. It will obviously be of considerable importance to establish whether actin and myosin are confined solely to this location in the cell, or whether they occur in functional combination throughout the endoplasm.

I thank Dr E. A. C. MacRobbie and Dr H. E. Huxley for many valuable discussions, and Churchill College for a Junior Research Fellowship, during the tenure of which this work was carried out.