THE ALTERATION of MITOTIC EVENTS by IONOPHORE A23187 and CARBONYL CYANIDE N-CHLOROPHENYLHYDRAZONE

J. Cell Sci. 7S, 347-355 (1985) 347 Printed in Great Britain © The Company of Biologists Limited 1985

THE ALTERATION OF MITOTIC EVENTS BY IONOPHORE A23187 AND CARBONYL CYANIDE n-CHLOROPHENYLHYDRAZONE

MICHAEL L.ZIEGLER, JESSE E. SISKEN* AND SHARANJIT VEDBRATf Department of Medical Microbiology and Immunology, College of Medicine, University of Kentucky, Lexington, Kentucky 40536, U.SA.

SUMMARY A large quantity of published work indicates that calcium ions may be involved in the regulation of mitotic events and recent reports suggest that the onset of chromosome movement is dependent upon a transient increase in free cytosolic calcium ions. In this paper we examine the effects of two agents known to perturb intracellular calcium pools on mitosis in HeLa cells. These were the calcium-selective ionophore A23187 and carbonyl cyanide n-chlorophenylhydrazone (CCCP), which is a protonophoric inhibitor of oxidative phosphorylation. Owing to a stimulation of glycolysis, the latter agent does not decrease intracellular ATP in HeLa but does cause mitochondria to release calcium ions. Our data show that, at low concentrations, both agents prolong metaphase but differ in their effects on anaphase and cytokinesis. Studies with chlorotetracycline, a commonly used probe for membrane-associated calcium, verify that these agents do affect calcium pools under the conditions of our experiments. The data presented are consistent with the idea that increased cytosolic calcium levels can directly or indirectly affect mitotic events but, contrary to other suggestions, cause a prolongation of metaphase, i.e. they delay the onset of chromosome movement.

INTRODUCTION The role of calcium in cell division has been a subject of considerable interest for many years. A substantial quantity of published work suggests that calcium ions may play a regulatory role in mitotic events and that the availability of these ions is itself regulated, both temporally and spatially, during the division process (e.g. see Hepler & Palevitz, 1974; Harris, 1975; Rebhun, 1977; Silver, Cole & Cande, 1980; Sisken, 1980; Salmon & Segal, 1980; Wolniak, Hepler & Jackson, 1980, 1983; Kiehart, 1981; Izant, 1983). The exact role of these ions, their source(s) and the mechanisms by which they are regulated remain open questions. In previous experiments we showed that treatment of HeLa cells with low levels of nicotine can delay the onset of chromosome movement (i.e. prolong metaphase) and accelerate the rate of furrowing once it begins (VedBrat, Sisken & Anderson, 1979). Since nicotine stimulates muscle contraction and calcium-dependent stimulus—

• Author for correspondence. f Present address: Department of Microbiology, College of Physicians and Surgeons, Columbia University, New York, N.Y., U.S.A.

Key words: calcium ions, mitosis, ionophores, metaphase, HeLa cells. 348 M. L. Ziegler, J. E. Sisken and S. VedBrat secretion coupling by releasing calcium from intracellular pools (e.g. see Weiss, 1968; Tjalve & Papov, 1973), we suggested that nicotine might have similar effects on HeLa cells and that this release could be responsible for the changes in duration of mitotic events (VedBrat et al. 1979). In order to gain a better understanding of the nicotine effect and of the role of calcium in cell division, we have studied the effects of two other agents that release calcium ions (Ca2+) from intracellular sources: carbonyl cyanide n-chlorophenylhydrazone (CCCP), which causes release of Ca2+ from mitochondria and the calcium- selective ionophore A23187. Both cause changes in mitotic parameters in HeLa cells when used at non-toxic levels and both drastically reduce fluorescence in cells stained with the fluorescent chelate probe, chlorotetracycline (CTC), a commonly used probe for membrane-associated Ca2+ (e.g. see Caswell, 1979).

MATERIALS AND METHODS HeLa (GEY) cells (Microbiological Associates) were grown in plastic T-flasks in Eagle's minimal essential medium (MEM) with Hanks' salts supplemented with 10% calf serum and 2mM- glutamine. In some cases penicillin (100 units/ml) and neomycin (0-l /ig/ml) were also added to the medium. For time-lapse cinemicrography, cells were scraped or trypsinized from the inner surfaces of the flasks, injected into Rose chambers and allowed to grow for 16—48 h. The medium was then removed from the chambers and replaced with a warm sample of treatment or control medium. Control cultures were run in parallel with or in tandem to treated cultures. Time-lapse photography began immediately after the initiation of treatment at the rate of two frames per min for up to 33 h using methods previously described (Sisken, 1964). Frame-by-frame analyses of the films were done on an analytical projector (Photo Instrumentation Corp. Burbank, Calif.). Criteria for determining durations of mitotic stages were as follows. Metaphase was judged to begin when chromosomes were first observed to be aligned on the equator of the mitotic spindle as seen in lateral view, and to end when chromatids began anaphase movement. The time from beginning of chromatid movement to the initiation of cytokinesis was taken as an approximation of anaphase since it was previously shown that in cultured human amnion cells, cytokinesis began, under several different experimental conditions, when the chromatids had moved approximately 85 % of their final distance apart (Sisken, 1973). The beginning of cytokinesis was identified as that point at which dimpling of the cell surface could first be clearly seen and its completion was taken as that point when the separation of the two daughter cells, with the exception of the intercellular bridge, was completed. The data were analysed using a statistical model that takes into account both intra- and interexperiment variances (VedBrat et al. 1979). A23187 (obtained as a gift from Eli Lilly Co. or purchased from Calbiochem) was dissolved in dimethylsulphoxide (DMSO) and diluted to 10~4M in 85% DMSO, 15% distilled water. This stock was added to media to give final concentrations of 1-0 or 2'0//MA23187 and 0-85 or 1-7% DMSO, respectively. Separate control experiments indicated that these concentrations of DMSO had no effects on the mitotic parameters we measured. CCCP (Sigma) was dissolved in ethanol to give a 5 X 10~3 M stock solution and diluted appropriately in media. CTC (Sigma) was prepared fresh for each experiment and dissolved in distilled water to give a stock solution of 1 -0 mM. For staining with CTC, cells were incubated in Rose chambers for at least 24 h in complete culture medium. They were then rinsed twice and incubated for 1 h in MEM containing 10 /iM-CTC, 1 % calf serum, 20mM-HEPES buffer (pH7-l) and, for experimentals, either A23187 or CCCP. Fluorescence microscopy was performed with a Leitz microscope fitted with an epi-illumination system. A Leitz D filter cube was used along with a 400 nm narrow band pass filter in the excitation pathway to enhance the selectivity for the calcium chelate (see Fabiato & Fabiato, 1979). For doubling time measurements, cells were seeded in replicate plastic flasks with 5xlO4 cells/25 cm2 flask. Cells were grown for 4 days in the MEM described above or in MEM containing Calcium affectors and mitotic events 349 CCCP. At the end of this time, cells were trypsinized and counted with a haemocytometer. Doubling times were calculated from average cell counts from two flasks.

RESULTS A23187 is a divalent cation ionophore whose biological effects are related to its capacity to bind to cell membranes and allow calcium to diffuse through them along concentration gradients. The specific membranes most affected by this agent and the direction of flow of calcium is dependent upon cell type, A23187 and exogenous Ca2+ concentrations and the duration of treatment. For example, Babcock, Chen, Yip & Lardy (1979) found that at low concentrations, the agent released Ca2+ from internal stores and caused Ca2+ efflux from liver cells and bull sperm, while at higher concentrations it stimulated uptake of exogenous Ca +. Jensen & Rasmussen (1977), on the other hand, concluded that in human peripheral lymphocytes the initial effect was to stimulate uptake of Ca2+ through the plasma membrane followed by a time- dependent redistribution of the ionophore leading to an efflux of calcium from the cell. The effects of 1-0 and 2-0^iM-A23187 on the durations of metaphase, anaphase and cytokinesis in HeLa cells are presented in Table 1 and Fig. 1, which show that each phase of mitosis responded differently. Treatment with this ionophore caused an average increase of 24-33 % in the duration of metaphase and, in the same cells, a 21-33% decrease in the duration of cytokinesis. These effects were essentially

Table 1. The effects ofA23187 on the duration of mitotic stages Metaphase Anaphase Cytokinesis mean duration mean duration mean duration Drug (M) ±S.E.(min) ±S.E.(min) ±S.E.(min) None 109 25-4± 1-91 109 5-15±0-31 109 3-12 ±0-20 81 19-2 ±2-02 84 4-73 ±0-31 84 2-96 ±0-20 27 18-9 ±2-75 38 4-83 ±0-33 40 2-91 ±0-21 73 20-6 ±2-07 81 5-20 ±0-31 81 3-01 ±0-21 79 21-5 ±2-03 95 5-20 ±0-31 100 3-04 ±0-20 99 22-5 ±1-94 100 4-99 ±0-31 100 2-97 ± 0-20 Avg. 486 21-7 ±0-85 507 502 ±0-13 514 300 ±008

A23187 79 26-2 ±2-03 79 5-35 ±0-31 79 215 ±0-21 (1X1O"6M) 121 27-5 ±1-87 121 5-43 ±0-31 121 1-87 ±0-20 Avg. 210 26-9 ±1-38 200 5-39 ±0-22 200 2-00 ±0-15 P = 0-01 P = 0-18 P= 0-001

A23187 53 28-4 ±2-24 67 5-89 ±0-32 67 2-29 ±0-21 (2X10~6M) 45 28-0 ±2-34 50 5-33 ±0-32 50 2-35 ±0-21 40 30-2 ±2-42 50 5-05 ±0-32 50 2-45 ±0-21 Avg. 138 28-8 ±1-35 167 5-43 ±018 167 2-36 ±0-12

P = 0-001 P=0-l P = 0-001 350 M. L. Ziegler, J. E. Sisken and S. VedBrat constant throughout the 33 h of continuous treatment (Fig. 1A,B) and nearly identical to those produced by nicotine (VedBrat et al. 1979). Alterations in chromosome alignment or movement were not apparent in these studies or in those involving CCCP, which follow. Although the data in Table 1 indicate that A23187 might cause a small overall prolongation of anaphase, the differences, according to the conservative statistical analysis used in this work, were not significant. However, a time-course analysis of the data (Fig. lc) indicates that the effects of this ionophore on anaphase are somewhat complex. A concentration of 1-0/XM A23187 produced no time-dependent effects, while at 2-0 /XM A23187 caused a 24 % increase in anaphase duration during the first 8 h of treatment followed by a decrease to, or possibly below, control levels. Specific mitochondrial inhibitors such as the phenylhydrazone uncouplers of oxidative phosphorylation have been shown to cause mitochondria to lose calcium ions in other systems (Luthra & Olsen, 1976; BabcockeJ al. 1979) as well as in HeLa cells (Ikehara et al. 1984). We were, therefore, interested in the effects of CCCP on

40 Metaphase

—I 20

4 Cytokinesis -6

o I 2 ra 3 Q

Anaphase 1

_L 10 20 30 Duration of treatment (h)

Fig. 1. The effects of A23187 on metaphase, cytokinesis and anaphase as a function of the duration of treatment. (O O) Controls; (• •) 1-0/

Table 2. Effect ofCCCP on mitotic parameters Metaphase Anaphase Cytokinesis mean duration mean duration mean duration Drug (M) n ±S.E.(min) n ±S.E.(min) n ±S.E.(min) None 60 22-8 ±1-19 60 4-30 ±0-08 59 3-30 ±0-07

CCCP 20 37-6 ±2-9 20 5-85 ±0-14 21 4-12±0-15 (5xlO"7M) 33 34-5 ±3-3 33 5-46±0-13 33 3-89±0-ll Avg. S3 36-7 ±1-47 53 5-70 ±0-07 54 400 ±006

P = 0-001 P = 0-001 P = 0-001

CCCP 22 26-9 ±1-98 22 4-66 ±0-13 22 4-14±0-12 (5xlO~6M) 15 29-4 ±2-84 15 4-53 ±0-16 15 3-57 ±0-13 Avg. 37 27-9 ±1-17 37 4-59 ±007 37 3-82 ±006

P = 0-002 P= 0-002 P= 0-002 mitotic parameters of HeLa cells because it affects intracellular Ca2+ pools by a different mechanism from that of A23187. Table 2 shows that in the presence of 0-5 /XM-CCCP, metaphase duration was increased by 61 % (from 22-8 min for controls to 36-7 min). In addition, and in contrast to the effects of A23187, both cytokinesis and anaphase were somewhat prolonged. Higher concentrations of CCCP (5-0/ZM) consistently produced less prolongation of mitotic parameters though all phases were still significantly longer than in control cultures. Neither concentration was overtly toxic over a 4-day period, as doubling times in controls and cultures treated with 0-5 and 5-0//M-CCCP were 27-1, 27-2 and 27-7 h, respectively. To establish that these treatments actually alter intracellular Ca2+ pools under the conditions of our experiments, we examined their effects on the fluorescence of CTC. Both agents were found to reduce CTC fluorescence to very low levels. Control cells demonstrate fibrillar fluorescence due to mitochondria and diffuse fluorescence presumably resulting from the association of CTC with elements of the endoplasmic reticulum and perhaps other intracellular membranous components (Figs 2, 6). As seen in Figs 3 and 7, in the presence of A23187 very little of the fluorescenceremain s in either metaphase or interphase cells after 1 h of treatment. Similarly, both interphase and mitotic cells show very low levels of fluorescencewhe n treated with 0-5 /iM-CCCP and even less when treated with 5-0/XM-CCCP (not shown). That this loss of fluorescence is not simply due to a deficiency in ATP is indicated by the finding that treatment of cells with toxic levels of antimycin A (5-5 fXM), an inhibitor of electron transport, does not result in decreased CTC fluorescence (Figs 5, 9).

DISCUSSION Both A23187 and CCCP were observed to increase significantly the duration of metaphase (i.e. delay the onset of anaphase chromosome movement). Since they release Ca2+ from intracellular stores (e.g. see Luthra & Olsen, 1976; Babcock, Furst 352 M. L. Ziegler, J. E. Sisken and S. VedBrat

Figs 2-9 Calcium affectors and mitotic events 353 & Lardy, 1976; Babcock et al. 1979) and reduce CTC fluorescence (Figs 3, 4, 7, 8), the data are consistent with the idea that the prolongation of metaphase induced by these agents is mediated by altered cytosolic concentrations of Ca +. It is reasonable to suggest that increased Ca2+ levels could increase metaphase durations since microtubules that are major components of the mitotic spindle are known to be Ca2+-sensitive (Weisenberg, 1972). It has also been observed that A23187 can cause the breakdown of cytoplasmic (though not mitotic spindle) microtubules in other mammalian cells (e.g. see Fuller & Brinkley, 1976). However, to explain all of the effects of A23187 and CCCP on the basis of the direct action of increased cytosolic Ca2+ is probably too simplistic. The fact that these agents affect cytokinesis differently, A23187 shortening it while CCCP prolongs it somewhat, indicates that other intracellular changes induced by these agents are also affecting the dividing cell. Thus, while the effects on metaphase could result directly from increased Ca2+ levels, the possibility remains that other effects of one or both of these agents, perhaps secondary to calcium release, might account for some of the alterations in mitotic parameters. For example, an increase in cytosolic Ca + levels could cause an increase in intracellular pH due to loss of protons, as occurs in other systems (Steinhardt, Shen&Zucker, 1978; Tilney, Kiehart, Sardet & Tilney, 1978), and it could be this alteration that affects mitotic parameters. It is known that microtubules can be disassembled at alkaline pH (Regula et al. 1981) and that polymerization of microfilaments is enhanced by increases in pH (Spudich, Spudich &Amos, 1979). While the possibility that one or more effects of these agents might result from a reduction in intracellular ATP, this does not appear to be the case for CCCP, at least, since cells incubated in CCCP continue dividing for at least three doublings at control rates. Further, due to a stimulation of glycolysis, CCCP concentrations as high as 5 /iM do not reduce ATP levels in HeLa cells in the presence of glucose or glutamine (Ikehara et al. 1984), both of which were present in our medium. This is consistent with the observations that a related agent, £-trifluoromethoxyphenylhydrazone (FCCP), causes collapse of the mitochondrial membrane potential, release of calcium and disruption of microtubules in HeLa cells (Maro & Bormans, 1982). This was also not a simple result of ATP depletion, since neither azide nor oligomycin produced the same effect. We have no explanation for the reverse dose effect of CCCP seen in our studies. With respect to the role of Ca2+ in mitosis, the data in this and our previous paper (VedBrat et al. 1979) indicate that increased intracellular Ca2+ levels cause a prolongation of metaphase. However, the degree of prolongation, up to 33% for A23187 and 61 % for CCCP, is perhaps less than might have been expected given the

Figs 2-9. Fluorescence of cells exposed to 10^M-CTC as described in Materials and Methods. Exposures for photography and printing were the same for all photographs. Figs 2-5 are of metaphase cells as follows: Fig. 2, CTC alone; Fig. 3, CTC plus 1-O^M- A23187; Fig. 4, CTC plus 0-5 ^M-CCCP; Fig. 5, CTC plus 5-5 fiM-antimycin A. Figs 6-9 are of interphase cells as follows: Fig. 6, CTC alone; Fig. 7, CTC plus l-0/iM-A23187; Fig. 8, CTC plus 0-5/iM-CCCP; Fig. 9, CTC plus 5-5^M-antimycin A. X560. 354 M. L. Ziegler, J. E. Sisken and S. VedBrat profound reduction in CTC fluorescence caused by both agents. One possible explanation for this is that increased cytosolic Ca2+ levels caused by any release are buffered by other Ca regulatory systems such as a plasma membrane Ca +- dependent ATPase, which may pump much of the excess Ca2+ to the exterior of the cell. On the other hand, there are published data that suggest that we might have expected to see a decrease in metaphase durations. Izant (1983) has reported that injection of Ca2+ into Ptkl cells speeds up the metaphase/anaphase transition, i.e. shortens metaphase, and Wolniakei al. (1983) have reported a transient decrease in chlorotetracycline fluorescence (indicating a release of membrane-associated Ca ) in late metaphase, just before the onset of anaphase. The suggestion offered in both cases was that an increase in free Ca2+ might be involved in the trigger for anaphase movement. While one might reconcile all of the data by suggesting, for example, that a transient increase in Ca2+ levels might be stimulatory while a prolonged increase is inhibitory, more data are required to establish that increases in Ca2+ are general occurrences at metaphase/anaphase transitions and, if they are, that they are involved in the triggering of chromosome movement. The final point to be made concerns the fact that doubling times of cells were essentially unaffected by levels of CCCP that eliminate most of the CTC fluorescence in HeLa cells. The data suggest: (a) that most of the CTC fluorescence in these cells is mitochondrial; and (b) that this pool of Ca2+ plays no role in the progression of cells through the cell cycle.

The authors acknowledge the expert technical assistance of Mrs Munira Nasser and Mrs Sally D. Grasch. The work was supported in part by grant CA27399 from the National Cancer Institute, National Institutes of Health, by a grant from the University of Kentucky Tobacco and Health Research Institute and by a Biomedical Research Support grant to the College of Medicine, University of Kentucky.

REFERENCES BABCOCK, D. F., CHEN, J. L. J., YIP, B. P. & LARDY, H. A. (1979). Evidence for mitochondrial localization of the hormone responsive pool of Ca++ in isolated hepatocytes. J. biol. Chem. 254, 8117-8120. BABCOCK, D. F., FURST, N. L. & LARDY, H. A. (1976). Action of ionophore A23187 at the cellular level. Separation of effects at the plasma and mitochondrial membranes. J. biol. Chem. 251, 3881-3886. CASWELL, A. H. (1979). Methods of measuring intracellular calcium. Int. Rev. Cytol. 56, 145-181. FABIATO, A. A. & FABIATO, F. (1979). Use of chlorotetracycline fluorescence to demonstrate Ca2+- induced release of Ca2+ from the sarcoplasmic reticulum of skinned cardiac cells. Nature, Land. 281, 146-148. FULLER, G. M. & BRINKLEY, B. R. (1976). Structure and control of assembly of cytoplasmic microtubules in normal and transformed cells. J. supramolec. Struct. 5, 497-514. HARRIS, P. (1975). The role of membranes in the organization of the mitotic apparatus. Expl Cell Res. 94, 409-425. HEPLER, P. K. & PALEVTTZ, B. A. (1974). Microtubules and microfilaments. A. Rev. PL Physiol. 25, 309-362. IKEHARA, T., YAMAGUCHI, H., HOSOKAWA, K., SAKAJ, T. & MIYAMOTO, H. (1984). Rb+ influx in response to changes in energy generation: effect of the regulation of the ATP content of HeLa cells. J. cell. Physiol. 119, 273-282. Calcium affectors and mitotic events 355 IZANT, J. G. (1983). The role of calcium ions during mitosis. Calcium participates in the anaphase trigger. Chromosoma 88, 1-10. JENSEN, P. &RASMUSSEN, H. (1977). The effect of A23187 upon calcium metabolism in the human lymphocyte. Biochim. biophys. Ada 468, 146—156. KlEHART, D. P. (1981). Studies on the in vivo sensitivity of spindle microtubules to calcium ions and evidence for a vesicular calcium-sequestering system. J. Cell Biol. 88, 604—617. LUTHRA, R. & OLSON, M. (1976). Studies of mitochondrial calcium movements using chlorotetracycline. Biochim. biophys. Ada 440, 744-758. MARO, B. & BORMANS, M. (1982). Reorganization of HeLa cell cytoskeleton induced by an uncoupler of oxidative phosphorylation. Nature, Lond. 295, 334-336. REBHUN, L. I. (1977). Cyclic nucleotides, calcium and cell division. Int. Rev. Cytol. 49, 1-54. REGULA, C. S., PFEIFFER, J. R. & BERLIN, R. D. (1981). Microtubule assembly and disassembly at alkaline pH.J. Cell Biol. 89, 45-53. SALMON, E. D. & SEGAL, R. R. (1980). Calcium-labile mitotic spindles isolated from sea urchin eggs (Lytechinus variegatus).J. Cell Biol. 86, 355-365. SILVER, R. B., COLE, R. D. & CANDE, W. Z. (1980). Isolation of mitotic apparatus containing vesicles with calcium sequestration activity. Cell 19, 505-516. SISKEN, J. E. (1964). Methods for measuring the length of the mitotic cycle and the timing of DNA synthesis for mammalian cells in culture. InMethods in CellPhysiol. (ed. D. Prescott), vol. 1, pp. 387-401. New York: Academic Press. SlSKEN, J. E. (1973). The effects of />-DL-fluorophenylalanine on chromosome movement and cytokinesis of human amnion cells in culture. Chromosoma 44, 91-98. SlSKEN, J. E. (1980). The significance and regulation of calcium during mitotic events. InNuclear- CytoplasmicInteractions in the CellCycle (ed. G. Whitson), pp. 271-291. New York: Academic Press. SPUDICH, J. A., SPUDICH, A. & AMOS, L. (1979). Actin from the cortical layer of sea urchin eggs before and after fertilization. In Cell Motility: Molecules and Organization (ed. S. Hatano, H. Ishikowa & H. Sato), pp. 165-187. Baltimore: Union Park Press. STEINHARDT, R. A., SHEN, S. S. & ZUCKER, R. S. (1978). Direct evidence for ionic messengers in the two phases of metabolic derepression at fertilization of the sea urchin egg. In Cell Reproduction: In Honor of Daniel Mazia (ed. E. Dirksen, D. M. Prescott & C. F. Fox), pp. 415-424. New York: Academic Press. TILNEY, L. G., KIEHART, D. P., SARDET, C. & TILNEY, M. (1978). Polymerization of actin IV: Role of Ca++ and H+ in the assembly of actin in membrane fusion in acrosomal reaction of echinoderm sperm. J. Cell Biol. 77, 536-550. TJALVE, H. & PAPOV, D. (1973). Effect of nicotine and nicotine metabolites on insulin secretion from rabbit pancreas pieces. Endocrinology 92, 1343-1348. VEDBRAT, S. S., SISKEN, J. E. & ANDERSON, R. L. (1979). The effects of nicotine on cell division of HeLa cells. Eur.J. Cell Biol. 250-254. WEISENBERG, R. C. (1972). Microtubule formation in vitro in solutions containing low calcium concentrations. Science 177, 1104—1105. WEISS, G. B. (1968). On the site of action of nicotine on contraction in frog sartorius muscle. jf. Pharmacol, exp. Ther. 163, 45-53. WOLNIAK, S. M., HEPLER, P. K. & JACKSON, W. T. (1980). Detection of the membrane-calcium distribution during mitosis in Haemanthus endosperm with chlorotetracycline. J. Cell Biol. 87, 23-32. WOLNIAK, S. M., HEPLER, P. K. & JACKSON, W. T. (1983). Ionic changes in the mitotic apparatus at the metaphase/anaphase transition. J. Cell Biol. 96, 598-605.

{Received 4 September 1984 -Accepted 5 December 1984)