Proc. Nati. Acad. Sci. USAVol. 78, No. 6, pp. 3624-3628, June 1981Cell Biology

Novel role of calcium in exocytosis: Mechanism of nematocystdischarge as shown by x-ray microanalysis

(sea anemone/nematocyte excitation/calcium-binding protein/stimulus-secretion coupling)

ROGER LUBBOCK, BRIJ L. GUPrA, AND T. A. HALLDepartment of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, England

Communicated by Sir Vincent Wigglesworth, February 20, 1981

ABSTRACT Mature nematocysts of the sea anemones Rho-dactis rhodostoma and Anthopleura elegantissima contain a fluidthat has a high concentration of solutes and is extraordinarily richin calcium (ca. 500-0 mmol/kg wet weight); this contrasts withthe surrounding cytoplasm which is rich in potassium but poor incalcium. The undischarged capsule is surrounded by a membranethat probably acts as a selective permeability barrier between thecytoplasm and the nematocyst fluid. During discharge the ne-matocyst moves to the surface of the nematocyte and comes intocontact with the external sea water medium. Calcium, which maybe bound to proteins in the undischarged state, is rapidly lost fromthe fluid; at the same time, sea water enters the capsule. In vitroexperiments have already shown that calcium loss increases theosmotic pressure of the capsular fluid, causing an influx of waterfrom the external medium; this influx appears to increase the hy-drostatic pressure inside the capsule to the point that the threadeverts explosively.

Nematocysts are stinging organelles found in cnidarians; eachone consists of a closed capsule which contains an inverted har-poon-like thread that can be discharged explosively. Althoughthe structural events during eversion of the thread have beendescribed in detail (1), little is known about the mechanism ofdischarge (2-5). Recently, Lubbock and Amos (6) found thatdischarge could be initiated in vitro by the removal of boundcalcium from the fluid inside the nematocyst; calcium removalincreased the osmotic pressure of the nematocyst fluid, thusproducing an influx of water from the external medium and anincrease in the internal hydrostatic pressure of the capsule. Itwas notable that alterations in the permeability of the capsulewall to water did not seem to be involved in controlling dis-charge in vitro. At that time, the relevance of the in vitro ex-periments to nematocyst discharge in vivo was not clearly es-tablished because data were not available on the ionic changesoccurring under natural conditions.

In the present study, we examined, by x-ray microanalysisof frozen sections, the elemental composition of both excitedand unexcited nematocytes in the sea anemones Rhodactis rho-dostoma and Anthopleura elegantissima. Our results enable usto present a generalized model of the events that take placeduring excitation.

MATERIALS AND METHODSMesenterial filaments from R. rhodostoma were placed in syn-thetic seawater (Tropic Marin Neu; specific gravity, 1.021) con-taining 16.8% (wt/wt) dextran (250,000 daltors; Sigma, clinicalgrade). To quench freeze, filaments with droplets of dextran/sea water were placed on the truncated conical tips of copperpins and rapidly immersed in a well-stirred slush of Freon-13

(monochlorodifluoromethane, m.p. -181'C) (7-9). In certaincases, filaments mounted on copper pins were electrically stim-ulated <3 sec before instant quenching in order to induce ne-matocyst discharge. Acrorhagial tissue was obtained from A.elegantissima by removing inflated acrorhagi with forceps andplaced in sea water containing 20% (wt/wt) dextran and rubid-ium chloride (final concentration, 80 mmol/kg wet weight).Nematocyst discharge was induced by touching the tip of aninflated acrohagus with a small piece of allogeneic column tis-sue. Acrorhagi were quenched as described above. All frozensamples were removed from Freon and stored in liquid nitrogenuntil required.

For electron microprobe x-ray analysis, frozen sections 1-3Aum thick were cut at -60 to -80'C in a cryomicrotome (SLEE,London) with steel knives, mounted, and examined in a JEOLJXA-50A microprobe analyzer. This instrument was fitted witha low-temperature transmission stage, an anticontaminationcap, two x-ray diffracting spectrometers (used for sodium andcalcium), a Kevex x-ray energy spectrometer, and a Link Sys-tems multichannel analyzer backed with a computer andQUANTEM/FLS program (on loan from Link Systems, En-gland) for deconvolution and background subtraction by least-squares fitting ofprefiltered spectra (10). Quantification of x-raydatawas performed by the continuum method (11) with dextran/sea water as the peripheral reference standard. Sections wereanalyzed first in a hydrated state and second after dehydrationwithin the microanalyzer. Further details of the technique,quantification procedures, and criteria to assess full hydrationof sections have been discussed elsewhere (7-9, 12, 13).

For conventional transmission electron microscopy, materialwas prepared by standard procedures of glutaraldehyde/os-mium fixation, dehydrated in ethanol/and embedded in Aral-dite. Thin sections stained with uranyl acetate and lead citratewere examined in a Philips EM 300 electron microscope.

RESULTSThe concentrations of elements detected in nematocytes of R.rhodostoma are shown in Table 1. The cytoplasm of unexcitednematocytes was found to contain relatively large amounts ofpotassium, chloride, and sulfur; sodium, magnesium, phospho-rus, and calcium were also present but at lower concentrations.The composition of the cytoplasm contrasted strongly with thatof the fluid inside undischarged- nematocysts (Figs. 1-3) (ne-matocysts classified as holotrichous isorhizas; cf. ref. 5), the lat-ter being particularly rich in calcium (606 mmol/kg wet weight)but poor in other elements detected. The nematocyst fluid con-tained about 32% dry-matter; irrdicating an excess ofalmost 30%over sea water; this agrees with results obtained by interferencemicroscopy (6). The wall ofundischarged nematocysts, probablypermeated with some fluid, contained 380 mmol ofcalcium perkg wet weight as well as moderate amounts of sodium, chloride,and sulfur; previous microprobe studies ofair-dried empty cap-

3624

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

Proc. Natl. Acad. Sci. USA 78 (1981) 3625

Table 1. Concentrations (mmol/kg, wet weight) of elements, expressed as mean ± SEM, in nematocytes ofR. rhodostoma

n dry mass Na Mg P S Cl K Ca

Sea water (standard) - 19 328 38 - 20 382 7 7Cytoplasm ofunexcitednematocyte 34 18 ± 2 43 ± 23 32 ± 7 38 ± 4 135 ± 5 106 ± 4 108 ± 6 9 ± 3

Capsule wall ofundischargednematocyst 5 44 ± 6 72 ± 7 10 ± 1 3 ± 1 95 ± 5 69 ± 4 28 ± 6 381 ± 44

Filament ofundischargednematocyst 14 36 ± 4 101 ± 28 17 ± 4 2 ± 2 136 ± 10 13 ± 4 14 ± 3 212 ± 9

Fluid insideundischargednematocyst 43 32 ± 2 55 ± 14 19 ± 5 7 ± 3 54 ± 10 19 ± 5 11 ± 2 606 ± 19

Fluid insidedischargingnematocyst 14 20 ± 2 194 ± 20 18±4 - 18 ± 2 105 ± 7 7 ± 1 189 ± 37

Fluid surroundingdischargingnematocyst 10 14 ± 2 234 ± 34 24 ± 6 2 ± 4 19 ± 2 389 ± 24 11 ± 4 531 ± 22

Fluid inside formingnematocyst 7 18 ± 4 34 ± 4 4 ± 2 33 ± 12 137 ± 13 119 ± 12 121 ± 11 13 ± 2

sules (14) have shown sulfur to be an important constituent ofthe capsule wall. The thread, probably including some fluid,contained 212 mmol of calcium per kg wet weight as well assignificant quantities of sodium, sulfur, and zinc; zinc is notshown in Table 1 because it was not quantified accurately, butit was present in the thread at about 50-100 mmol/kg wetweight. High calcium concentrations seemed to be restrictedto mature nematocysts; the fluid contents ofnematocysts in theprocess of formation (Fig. 1; Table 1) had an elemental com-position similar to that of the surrounding cytoplasm.

Prior to discharge, the nematocyst is relatively deeplyembedded in the nematocyte and can be seen in transmissionelectron micrographs to be surrounded by a unit membrane justexterior to the capsule wall. At an early stage in the dischargeprocess the capsule seems to move toward the plasma mem-brane so that its tip is exposed to the external sea water medium;the rest of the capsule remains in the nematocyte but becomessurrounded by a narrow layer of fluid (Fig. 1 Inset). Eversionof the thread was only seen to begin after the capsule had comeinto contact with sea water.

As can be seen from Table 1, there is a marked change in thecomposition ofthe fluid inside the nematocyst during discharge.Previous work (6) on isolated capsules has shown that the fluidinside the nematocyst is progressively diluted until it has a re-fractive index similar to that of sea water. A decrease in the drymass of the fluid inside discharging capsules was also noted inthe present experiments; the extent of this decrease was prob-ably not significantly affected by the presence of dextran(250,000 daltons) in the outside medium because the capsulewall is not permeable to substances above about 1000 daltons(6). Analysis of the major changes in elemental compositionduring discharge was carried out by plotting % dry mass of thefluid inside the nematocyst against the concentrations of cal-cium, sodium, and chloride (Fig. 4a). The results indicate thatat the beginning ofdischarge there is a massive efflux ofcalciumfrom the nematocyst fluid that is disproportionate to the loss ofdry mass: a 50% decrease in calcium concentration, for example,is accompanied by a decrease of about 15% in dry mass. As thecalcium is being lost, there is an influx of sodium and chloridebut not of potassium; this indicates that, during discharge invivo, it is sea water that enters the capsule and not the cytosolfrom the nematocyte.

As pointed out above, the discharging capsule becomes sur-rounded by a narrow laver of fluid (Fig. 1 Inset); this fluid con-tains substantial quantities of sodium, chloride, and calcium(Table 1). Furthermore, the calcium concentration in the sur-rounding fluid frequently exceeds that of the internal fluid of

the same nematocyst by more than 100-200 mmol/kg wetweight. The elemental composition of the surrounding fluidsuggests that it may be sea water into which calcium has escapedfrom the capsule.

The concentrations of elements detected in nematocytes ofA. elegantissima are shown in Table 2. The cytoplasm of unex-cited nematocytes contained relatively large amounts of potas-sium and chloride; sodium, magnesium, phosphorus, sulfur,and calcium also were present but at lower concentrations. Therelatively small size of the nematocysts (Fig. 5) (nematocystsclassified as holotrichous isorhizas; cf. ref. 5) prevented properresolution of the various components, and it is likely that mea-

FIG. 1. Scanning transmission electron micrograph offrozen driedsection (1-2 ,um thick) of mesenterial filament from R. rhodostoma,showing external sea water medium (s), a mature undischarged ne-matocyst (n) (dark structures within the nematocyst are sections ofthecoiled thread), and a forming nematocyst (f). (Inset) Discharging ne-matocysts surrounded by a fluid layer (1). (Scale bar = 20 num.)

Cell Biology: Lubbock et al -

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

3626 Cell Biology: Lubbock et al.

kg wet weight); magnesium was more concentrated than in thecytoplasm whereas potassium and chloride again were very low(see Table 1). Rubidium, included as a potassium analogue inthe sea water, was found to enter the cytoplasm; the level ofrubidium within the nematocyst was similar to that in the cy-toplasm, showing that exchange of these ions must occur be-tween the cytoplasm and the nematocyst fluid. The very highlevel of calcium in undischarged, mature nematocysts was ab-

600

500

400.0

3-4

to

zuuI

a

0

0 a0

0_

0

.0

0

30 25 20 15

FIG. 2. (a) Section of mesenterial filament from R. rhodostomashowing mature undischarged nematocysts; specimen prepared andphotographed as in Fig. 1. (b) Calcium K., x-ray intensity distributionover the same field, collected with 1-nA current over 500 sec. (Scalebar= 20 gm.)

surements of the fluid contents were in some cases affected bythe unseen presence of thread, capsule wall, or even overlap-ping cytoplasm. Nevertheless, it is clear from Table 2 that theundischarged nematocysts contained a rather high proportionof dry mass and a very high calcium concentration (542 mmol/

a | | ' {1000

Na Mg I P Cl K Ca Ca Q(K.l) (Ken) ~ 500

b

1 2 3 4 5keV

FIG. 3. X-ray energy spectra from the internal fluid of a mature,undischarged nematocyst (a) and the adjacent cytoplasm within thenematocyte (b) ofR. rhodostoma.

40014be)-)a)-633 300beI--0E

b

a

0 1dD

0

S

0

0

50 40 30 20% dry mass

10 0

FIG. 4. Changes in the levels of calcium (solid circles, solid line),sodium (squares, broken line), and chloride (open circles, dotted line)in the internal fluid of discharging nematocysts from R. rhodostoma(a) andA. elegantissima (b). Each point is based on measurements froma single discharging nematocyst, except that the ringed points repre-sent the condition before discharge and are based on 43 nematocystsin R. rhodostoma and 50 nematocysts in A. elegantissima.

Proc. Natl. Acad. Sci. USA 78 (1981)

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

Proc. Natl. Acad. Sci. USA 78 (1981) 3627

Table 2. Concentrations (mmol/kg wet weight) of elements, expressed as mean ± SEM, in nematocytes ofA. elegantissima% dry

n mass Na Mg P S Cl K Ca RbSea water (standard) - 23.4 314 36 - 19 446 7 7 80Cytoplasm ofunexcited nematocyte 24 19 ± 2 52 ± 15 14 ± 2 67 ± 2 70 ± 2 152 ± 2 138 ± 2 18 ± 2 19 ± 2

Contents of undischargednematocyst 50 44 ± 4 44 ± 7 73 ± 6 72 ± 2 88 ± 2 26 ± 3 22 ± 3 542 ± 20 15 ± 5

Contents of dischargingnematocyst 13 27 ± 2 272 ± 19 63 ± 8 43 ± 7 52 ± 7 250 ± 18 17 ± 3 90 ± 18 57 ± 4

Contents of formingnematocyst 9 34 ± 3 45 ± 15 63 ± 6 40 ± 5 183 ± 10 236 ± 12 75 ± 2 2 ± 2 14 ± 2

sent from those in the process offormation; the latter containedhigh levels ofchloride and sulfur as well as moderate quantitiesof potassium and magnesium.

During discharge, the elemental composition of A. elegan-tissima's nematocysts also changes markedly (Table 2). Thereis a massive initial efflux of calcium from the nematocyst fluidwhich, as in R. rhodostoma, is disproportionate to the loss ofdry mass (Fig. 4b): for example, a 70% decrease in the calciumlevel is accompanied by an approximately 10% decrease in drymass. As can be seen from Table 2, the efflux of calcium is ac-companied by an influx of sodium, chloride, and rubidium butnot potassium. This indicates that, as in R. rhodostonw, ionsand fluid entering the discharging capsule are coming from thesea water medium rather than the cytoplasm.

DISCUSSIONThe results from both R. rhodostoma and A. elegantissima in-dicate that the fluid contained within a mature but undischargednematocyst is ofa fundamentally different composition from thecytoplasm of the surrounding cell: the cytoplasm is rich in po-tassium but poor in calcium, whereas the nematocyst fluid ispoor in potassium but very rich in calcium. High calcium con-centrations seem to be found only in mature nematocysts, thosein the process of formation having levels of calcium similar tothose of the cytoplasm. The principal permeability barrier be-tween the cytoplasm and the nematocyst is probably the unitmembrane that surrounds the capsule wall rather than the cap-sule wall itself, the latter being permeable in vitro to chargedsubstances of less than about 1000 daltons (6). The membrane

*

- dAft..ewelAAf

~~~~Aew

* ,5M,

FIG. 5.Sanning

transmission electron micrograph of a frozen-dried section (1-2 ,um thick) of a discharging acrorhagus from A. ele-gantisima, showing numerous nematocysts discharging from the ac-rorhagus (left) into a piece of allogeneic column tissue (right). (Scalebar = 10 ian.)

appears to be selectively permeable because rubidium can enterundischarged nematocysts from the cytoplasm. At the begin-ning of discharge, the capsule moves toward the surface of thehost cell and comes into contact with the external sea watermedium. The fate of the membrane surrounding the capsulewas not observed, although it seems likely that it fuses with theplasma membrane as in other exocytotic processes (15). Oncethe discharging capsule is in contact with the sea water, thereis a massive loss ofcalcium from the internal fluid. Calcium losshas been shown in vitro (6) to increase the osmotic pressure ofthe nematocyst fluid, thus causing an influx ofwater and an in-crease in the internal hydrostatic pressure of the capsule; ul-timately, the thread everts explosively.The molecular events occurring during the liberation of cal-

cium from the nematocyst are of considerable interest. In theundischarged state, the internal fluid ofthe nematocyst containsa substantial dry mass (about 32% in R. rhodostoma and 44%in A. elegantissima); data from other anemone species suggeststhat this dry mass may be primarily attributable to dissolvedproteins (16, 17). The very large concentration ofcalcium insidethe nematocyst is clearly not balanced by possible anionic ele-ments detected with the microprobe; this indicates that thecalcium may be associated either with hydroxyl ions or withorganic anions. In this respect it is interesting to note that pro-teins in the nematocyst fluid are unusually rich in acidic aminoacid residues, in particular glutamic acid and aspartic acid (16,17); calcium-binding proteins in other cells typically have highlevels of aspartic, glutamic, or y-carboxyglutamic acid residueswhich may be involved in the formation ofhigh-affinity calcium-binding sites (18). This suggests that calcium within the nemato-cyst may be associated with glutamate or aspartate residues ofdissolved proteins. The mechanism by which the osmotic pres-sure increases when calcium is lost (6) is not clear at present;one interesting possibility is that, in the resting state, dissolvedproteins are bound by calcium into aggregates and that the re-moval of calcium causes dissociation of the aggregates into agreater number of smaller molecular particles.The rapid loss of calcium from the nematocyst at the begin-

ning ofdischarge is accompanied by only a small decrease in thedry mass of the capsular fluid. If the anions associated with cal-cium in the undischarged state are part of the bulk of the drymass in the fluid, then clearly the loss of calcium must producea marked cation deficiency. Although the sodium concentrationin discharging capsules sometimes exceeds the chloride con-centration, it seems clear that any net entry of cations from thesea is insufficient to balance the loss of calcium. This suggeststhat calcium within the capsule may be replaced by either or-ganic cations or protons, thus maintaining the charge equilib-rium. In R. rhodostoma there appears to be a narrow layer offluid between the discharging capsule and the cytoplasm, intowhich the calcium passes. This fluid is partly derived from seawater but contains more chloride than sodium. The chloride

Cell Biology: Lubbock et al -

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

3628 Cell Biology: Lubbock et al.

excess is quite insufficient, however, to balance the large cal-cium content and it seems likely that again either organic anionsor hydroxyl ions are also present.

Discharge does not appear to be attributable simply to thenematocyst coming into contact with sea water. Unpublishedexperiments show that isolated nematocysts brought into con-tact with sea water frequently do not discharge immediatelyunless a calcium chelating agent such as citrate is added to themedium (cf. ref. 6). This suggests that two simultaneous pro-cesses may be involved in vivo, the first being the translocationof the nematocyst into contact with the external medium bymembrane fusion (15) and the second being the release of sub-stances that unbind calcium.The above results together with those ofother recent studies

suggest the following tentative model of the events occurringduring nematocyte excitation. Specific cell surface receptors onthe nematocyte or nearby cells are excited by contact with anappropriate substrate (19, 20), producing electrical activity (21).The nematocyst capsule moves to the surface ofthe nematocyteand comes into contact with sea water; the membrane that sur-rounds the undischarged nematocyst presumably fuses with theplasma membrane as in other exocytotic processes (15). Possiblyat the same time, substances that unbind calcium are releasedfrom the nematocyte cytoplasm. Calcium, which is present ata very high concentration in the nematocyst fluid and may bebound to proteins, is rapidly lost from the capsule. The calciumloss increases the osmotic pressure ofthe capsular fluid (6), pro-ducing an influx of sea water. How the increase in osmotic pres-sure is produced is not clear, but it could involve the dissociationof protein aggregates bound by calcium. The water influx raisesthe internal hydrostatic pressure ofthe capsule to the point thatthe thread discharges explosively. Once the thread is fullyeverted, the hydrostatic pressure serves to inject the internalfluid, which is venomous (5), through the thread into the target.

In conclusion, it seems clear that calcium plays an importantrole in nematocyst discharge in vivo. A question ofconsiderableinterest is whether analogous discharge mechanisms exist in anyother exocytotic processes; mucus granules in vertebrate gobletcells, for example, contain fairly high concentrations of calciumprior to release (8). The nematocyst is a specialized product ofthe Golgi apparatus (22), and it perhaps would not be surprisingif its explosive discharge mechanism was modified from otherless-striking secretion processes.

We thank Dr. W. B. Amos for commenting on the manuscript andMrs. M. Bray, Mr. A. J. Burgess, and Mrs. J. Wardle for technical as-sistance. The Biological Microprobe Laboratory in the Department wasestablished with a grant from the British Science Research Council.T.A.H. is a Leverhulme Trust Research Fellow.

1. Skaer, R. J. & Picken, L. E. R. (1965) Philos. Trans. R. Soc. Lon-don Ser. B 250, 131-164.

2. Cormier, S. M. & Hessinger, D. A. (1980) Tissue Cell 12,713-721.

3. Carr6, D. (1980) Eur. J. Cell Biol. 20, 265-271.4. Robson, E. A. (1973) Publ. Seto Mar. Biol. Lab. 20, 653-673.5. Mariscal, R. N. (1974) in Coelenterate Biology: Reviews and New

Perspectives, eds. Muscatine, L. & Lenhoff, H. (Academic, Lon-don), pp. 129-178.

6. Lubbock, R. & Amos, W. B. (1981) Nature (London) 290, 500-501.

7. Gupta, B. L., Hall, T. A. & Moreton, R. B. (1977) in Transportof Ions and Water in Animals, eds. Gupta, B. L., Moreton, R.B., Oschman, J. L. & Wall, B. J. (Academic, London), pp.83-143.

8. Gupta, B. L. & Hall, T. A. (1978) Ann. N.Y. Acad. Sci. 307,28-51.

9. Gupta, B. L. & Hall, T. A. (1979) Fed. Proc. Fed. Am. Soc. Exp.Biol. 38, 144-153.

10. Statham, P. J. (1977) Anal. Chem. 49, 2149.11. Hall, T. A. (1971) in Physical Techniques in Biological Research

1A, ed. Oster, G. (Academic, New York), 2nd Ed., pp. 157-275.12. Gupta, B. L. (1979) in Microbeam Analysis in Biology, eds.

Lechene, C. & Warner, R. (Academic, New York), pp. 375-408.13. Gupta, B. L., Wall, B. J., Oschman, J. L. & Hall, T. A. (1980)J.

Exp. Biol. 88, 2147.14. Mariscal, R. N. (1980) in Developmental and Cellular Biology of

Coelenterates, eds. Tardent, P. & Tardent, R. (Elsevier/NorthHolland, Amsterdam), pp. 337-342.

15. Satir, B. (1974) Symp. Soc. Exp. Biol. 28, 399-418.16. Blanquet, R. (1968) Comp. Biochem. Physiol. 25, 893-902.17. Goodwin, M. H. & Telford, M. (1971) Biol. Bull. (Woods Hole,

Mass.) 140, 389-399.18. Williams, R. P. J. (1977) in Calcium-Binding Proteins and Cal-

cium Function, ed. Wassermann, R. H., Corradino, R. A., Car-afoli, E., Kretsinger, R. H., Maclennan, R. H. & Siegel, F. L.(North Holland, New York), pp. 3-12.

19. Lubbock, R. (1979) J. Exp. Biol. 83, 283-292.20. Lubbock, R. (1980) Proc. Natl. Acad. Sci. USA 77, 6667-6669.21. Lubbock, R. & Shelton, G. A. B. (1981) Nature (London) 289,

59-60.22. Skaer, R. J. (1973)J. Cell Sci. 13, 371-393.

Proc. Natl. Acad - Sci - USA 78 (1981)

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0