Printed in Great Britain · The fate of particles endocytosed by coelenterate endodermal cells has only recently been investigated. Previous investigators noted acid-phosphatase activity

J. Cell Sci. 64, 195-212 (1983) 195Printed in Great Britain © The Company of Biologists Limited 1983

ENDOCYTOSIS OF THE SYMBIOTICDINOFLAGELLATE SYMBIODINIUM

MICROADRIATICUM FREUDENTHAL BYENDODERMAL CELLS OF THE SCYPHISTOMAE OFCASSIOPEIA XAMACHANA AND RESISTANCE OFTHE ALGAE TO HOST DIGESTION

WILLIAM K. FITT AND ROBERT K. TRENCHDepartment of Biological Sciences and Marine Science Institute, University ofCalifornia, Santa Barbara, California 93106, U.SA.

SUMMARY

The ingestion and fate of four types of particles by endodermal cells of the scyphistomae ofCassiopeia xamachana were investigated by scanning and transmission electron microscopy. Fer-ritin was endocytosed pinocytotically by invagination of the plasmalemma. These small pinocytoticvesicles fuse with other similar vesicles to form larger ferritin-containing vacuoles, which eventuallyfuse with lysosomes. Such secondary lysosomes exhibit acid phosphatasc activity. The co-occurrence of acid phosphatase activity and ferritin in secondary lysosomes achieved maximumfrequency within 2 h of uptake of ferritin and was evident for at least 4 h following uptake. Artemiaparticles, live freshly isolated symbiotic algae (Symbiodinium microadriaticum), and heat-killed5. microadriaticum are phagocytosed by endodermal cells.

Ferritin-labelled lysosomes fused with food vacuoles containing particles of Artemia. Vacuolescontaining heat-killed S. microadriaticum also showed evidence of phago—lysosome fusion.S. microadriaticum in situ (i.e. in host cells) after 3 days exposure to the photosynthetic inhibitor,3-(3-4-dichlorophenyl)-l, 1 -dimethylurea, appeared degenerate, and were found in loose-fitting hostvacuoles, many in mid and apical portions of the host cell. More than 70% of these vacuoles withmoribund algae contained the ferritin label, indicating that lysosome fusion had occurred. Incontrast, live 5. microadriaticum in control animals were almost always found at the base of the hostcell in individual tight-fitting vacuoles with no evidence of lysosome fusion.

Live S. microadriaticum apparently escape host digestion by prohibiting the fusion of lysosomeswith the vacuole in which they reside. Vacuoles containing defunct algal symbionts, in contrast,were subject to lysosomal attack.

INTRODUCTION

Entry of a genetically foreign entity into a host cell during the establishment of asymbiotic association involves a number of steps or 'obstacles', which may be sum-marized as: coming together, uptake (endocytosis), resistance to or avoidance of hostcellular defences, and sequestration in the appropriate location in the host cell (cf.Muscatine, Cook, Pardy & Pool, 1975a; Trench, 1979, 1980a; Smith, 1980). In aprevious paper (Fitt & Trench, 1983) a number of mechanisms of 'coming together'of marine hosts and the symbiotic alga Symbiodinium microadriaticum were

196 W. K. Fitt and R. K. Trench

addressed experimentally. In the present study we place emphasis on the two follow-ing steps: namely, endocytosis and resistance to host digestive attack.

Previous studies on uptake of algal symbionts by coelenterate hosts have centred onthe Hydra-Chlorella symbiosis. Phagocytosis in Hydra is initiated by contact of algaewith the digestive cell surface (Cook, 19,80), followed by engulfment by microvilli orpseudopods that possess a diffuse glycocalyx (Cook, D'Elia & Muscatine, 1978).McNeil (1981) characterized two different modes whereby freshly isolated algae weretaken into digestive cells of hydra. Live algal cells were phagocytosed via either afunnel-shaped extension of the plasmalemma or a meshwork of microvilli. Heat-treated symbionts were only phagocytosed by the 'funnel' method. These twomethods of entry are in contrast to the 'multiple overlapping folds' mode found in theuptake oiArtemia (food) particles. Following phagocytosis of algae, the newly formedvacuoles contained only one alga per vacuole, with no evidence of the coated vesiclesor diffuse glycocalices that were observed on the unchallenged plasmalemma (Cooket al. 1978). Algae were detected within phagosomes in less than 1 min after injectioninto the coelenteron (Cook et al. 1978), illustrating the rapidity of phagocytosis.

In the only other intracellular algal symbiosis in which endocytosis has beenstudied, ingestion of Chlorella by Paramecium bursaria differed from algal uptakeinto coelenterates in that more than one alga can initially be taken into a vacuole.Sequestration of one alga per vacuole was the usual end result of phagocytosis of thelive symbionts (Karakashian & Karakashian, 1973).

There are few studies on the acquisition of algae in marine symbioses. Thephenomenon is important, in the light of the growing list of symbiotic species that donot pass their algae directly to their sexual progeny via the egg (Trench, 1980a, 1983;Fitt, 1982). These species must acquire symbionts from the ambient environment.Present evidence indicates that S. microadriaticum are phagocytosed by coelenterateendodermal cells (Fitt, 1982; Colley & Trench, 1983).

The fate of particles endocytosed by coelenterate endodermal cells has only recentlybeen investigated. Previous investigators noted acid-phosphatase activity at the apicalends of Hydra digestive cells and deduced that food particles in vacuoles in theseportions of the cell were being digested (Lentz, 1966). Algal symbionts, in contrast,are apparently able to resist digestion and they persist, grow and divide inside thedigestive cells. Studies with unicellular parasites and other algal-invertebrate associa-tions (see Muscatine et al. 19756, for a review) have shown several ways in which anintracellular symbiont may avoid digestion, survive and persist inside a host cell.These include escape from the phagosome to lie free in the host cell's cytoplasm (seeTrench, 19806), movement of the vacuole and algae to areas away from the digestivezones of the cell (e.g. see Weis, 1976; Cooper & Margulis, 1977; McNeil, Hohman& Muscatine, 1981), prevention of phago—lysosome fusion (e.g. see Hohman, McNeil& Muscatine, 1982; O'Brien, 1980; Karakashian & Rudzinska, 1981), and resistanceto or inactivation of lysosomal enzymes.

In this study we analysed the mechanism of ingestion, and the intracellular fate offood particles and of live and dead 5. microadriaticum within endodermal cells of thescyphistomae of Cassiopeia xamachana.

Endocytosis and persistence of zooxanthellae 197

MATERIALS AND METHODS

Maintenance of experimental animalsThe host animals used in all experiments were scyphistomae of Cassiopeia xamachana cloned

from asexual buds. Standard maintenance conditions for aposymbiotic animals were constant dark-ness at 25 ± 1 deg.C. They were fed Artemia nauplii 2-3 times per week. During incubation withalgae, food particles or ferritin, the animals were kept at light intensities of 20 /lEinsteins m~2 s~' at25 °C and were starved for 2 days before injection.

ParticlesFerritin (2X recrystallized, Sigma) was diluted approximately 8:1 with 0-45jJm Millipore-

filtered sea-water (MFSW) to a concentration of about lOmg/ml.Food particles were prepared by homogenizing Artemia nauplii in a glass tissue-grinder and

filtering the homogenate through a 0-22/Um Millipore filter.S. microadriaticum (clone no. 313) were isolated from 1- to 3-month-old C. xamachana jellyfish.

These algae were originally introduced to C. xamachana scyphistomae from cultures maintainedin ASP-8A (see Fitt, Chang & Trench, 1981). The algal symbionts were obtained by homogenizingjellyfish in a glass tissue-grinder; the resulting slurry was strained through four layers of cheeseclothto remove larger pieces of animal tissue, and then centrifuged at 433 £ and rinsed three times withMFSW. Algae were heat-killed by exposure to boiling water for 5 min.

All particles were introduced into the mouths of 48-h-starved aposymbiotic scyphistomae via amicropipette. Algae, Artemia and ferritin were incubated in scyphistomae for various periods oftime in sea-water containing the same concentration of particles as that injected into the coelenteron,before being rinsed with filtered sea-water and fixed for electron microscopy.

Time-course experiments of phagocytosis of algal cells into endodermal cells of scyphistomaewere done for periods extending from 5 min to 24 h using algal densities of KT'-IO8 algae per ml.After incubation for the prescribed period of time scyphistomae were transferred to MFSW, theirbases cut off with a razor-blade to remove any adhering algae and the non-phagocytosed algal cellsremaining were rinsed out of the coelenteron with micropipette-injected MFSW. The number ofalgae remaining was determined by homogenizing each scyphistoma in 0-2 ml MFSW in a tissue-grinder and taking the average of four replicate haemacytometer counts.

Electron microscopy

Scyphistomae were fixed in Karnovsky's (1965) fixative for 1 h, rinsed in 0-2M-cacodylate buffer(pH7-4), and post-fixed in cacodylate-buffered 1% osmium tetroxide. After dehydration inethanol, animals for transmission electron microscopy (TEM) were embedded in SpUrr's medium,sectioned and observed on a Siemens Elmiskop I electron microscope. Some sections were leftunstained to enhance unambiguous recognition of ferritin labelling. Other sections were stainedwith uranyl acetate and lead citrate. Animals for scanning electron microscopy (SEM) were critical-point dried, cut in half with a razor-blade, sputter-coated with gold/palladium, and observed withaJEOLLSM-2SEM.

Acid phosphatase was assayed after glutaraldehyde fixation by a modified Gomori technique(Barrett & Heath, 1977), using B-glycerophosphate as substrate, in acetate buffer (pH6-l) at 37 °Cfor 20 h. Animals were rinsed in buffer and post-fixed in 1 % osmium tetroxide. Control animalswere incubated without the substrate.

Assay for phago—lysosome fusion

Binding of secondary lysosomes containing ferritin to phagosomes was assayed by methods similarto those of Armstrong & Hart (1971). Ferritin (10 mg/ml) was injected into the mouths of 2-day-starved scyphistomae and the whole animal was allowed to incubate in sea-water containing the sameconcentration of ferritin. After 15 min the coelenteron of the scyphistomae and surrounding solutionwas flushed with MFSW three to four times over a 15-min period. The animals were then maintainedfor 1-5 h in MFSW to allow all residual ferritin to be taken into cells before algae or food particleswere injected into the coelenteron. Scyphistomae were fixed for electron microscopy 2h afterinjection with live or dead 5. microadriaticum or Artemia particles.


The kinetics of formation of secondary lysosomes containing ferritin were established by incuba-tion of scyphistomae in ferntin for IS min, rinsing in MFSW three to four times and fixing scyphis-tomae periodically between 2 min and 4h after the initial ferritin injection. Acid phosphatase wasassayed as previously described.

For some experiments scyphistomae already containing clone no. 313 S. microadriaticum wereplaced in 10~6M-DCMU (3-(3-4-dichlorophenyl)-l,l-dimethylurea) for 3 days before ferritin wasinjected as previously described. Ferritin was allowed to incubate for 4h, then washed out of thecoelenteron. After 1 h post-incubation in sea-water these scyphistomae were fixed and processed forelectron microscopy.

RESULTS

Ferritin-labelling of lysosomes

The endodermal cells of the scyphistomae of C. xamachana pinocytosed ferritininto small vacuoles (Fig. 1A), which appeared to fuse to form larger vacuoles (Fig.ID) . Lysosomes fused with vacuoles containing ferritin (Fig. 1B, C) so that the percen-tage of vacuoles containing both acid phosphatase and ferritin increased from less than5 % at 5 min to maximum detectable levels of 65-70 % 2 h after the injection offerritin (Fig. 2).

Phagocytosis o/Artemia particles

Food particles (Artemia) were engulfed phagocytically (Figs 3A, 4). Small vacuoles(Fig. 3B) containing particles were present in the apical portions of the endodermalcells within 5 min of injection. Scanning electron micrographs showed that Artemiaparticles were often found attached to the flagella of endodermal cells shortly afterinjection into the coelenteron (Fig. 4A) and phagocytosed in groups by overlappingpseudopodia (Fig. 4B). The fused pseudopods containing newly phagocytosed'packets' of Artemia particles (Fig. 4c) were prevalent in most endodermal cells afterlh(Fig . 4D).

When animals had been preincubated with ferritin to label secondary lysosomes,and then provided WithArtemia particles, lysosomes fused with phagosomes (Fig. 3c)and emptied their contents into the phagosomes (Fig. 3D).

EndocytosisofS. microadriaticum

Live and dead 5. microadriaticum appeared to be phagocytosed by mechanismsinvolving engulfment by extensions of the plasmalemma (Figs 5, 6, 7). Algae wereseen adjacent to host endodermal cells (Fig. 6A), often in contact with host cell flagella(Figs 5B, 6B), shortly after injection into the coelenteron. After initial contact with ahost cell (Figs 5A, 6B), pseudopods surrounded the alga (Figs 5c, 6c, 7A, B). Engulf-ment by microvilli was not seen. Live freshly isolated S. microadriaticum weresequestered in individual tight-fitting vacuoles and were commonly seen at the baseof endodermal cells after 1 day (Fig. 5D). Heat-killed freshly isolated S. microadria-ticum were usually seen in loose-fitting vacuoles (Fig. 7D, E) and only rarely near thebase of the host cell. Dead algae were not seen in macerations of scyphistomae 48 hafter injection. Neither coated vesicles nor diffuse glycocalices were seen duringuptake of 5. microadriaticum.


Fig. 1. Endocytosis and fate of ferritin particles in endodermal cells of scyphistomae ofC. xamachana. A. Pinocytosis of ferritin 5 min after injection into the coelenteron (c).Small vacuoles (small arrows) fuse to form larger vacuoles. Ferritin is also found betweenendodermal cells (large arrows). Unstained. X20500. B. Large secondary lysosomes (/) 2 hafter injection of ferritin. Unstained. X30800. c. Small secondary lysosomes (large arrow)formed by the fusion of a primary lysosome (small arrows) containing acid phosphataseand a vacuole containing ferritin (/) 5 min after injection of ferritin. Unstained. X63 600.D. Large phagosomes containing ferritin (/), after injection of ferritin. Stained. X14900.

200 W. K. Fitt and R. K. Trench100

Time (h)

Fig. 2. Percentage of vacuoles containing ferritin that also contain acid phosphatase label,expressed as a function of time after injection of ferritin. Each point represents data from100 vacuoles.

Phagocytosis of live freshly isolated algae is a rapid process. Fig. 8 shows that atconcentrations of 104 and 106 algae per ml phagocytosis saturates at 1000-2000 cellsper scyphistoma after only 30min. At 108 algae per ml uptake is slower, probablybecause of the clumping of cells characteristic of this clone of 5. microadriaticum athigh concentrations. Maximum levels of uptake reached 10 000 cells per scyphistoma.The actual concentration of algae in the coelenteron changed in an undetermined


Fig. 3. Endocytosis and fate of Artemia particles in endodermal cells of scyphistomae ofC. xamachana. Unstained, c, coelenteron. A. Phagocytosis of Artemia particles by cellularextensions of the plasmalemma (arrows) 5 min after injection of Artemia. X22700. B.Phagosomes containing one or more Artemia particles (arrows) 5 min after injection ofArtemia. X 28 000. c. Fusion (?) of a secondary lysosome (/) to a phagosome (p) containingArtemia particles 30 min after injection of Artemia. x 19 300. D. Ferritin label, indicatinglysosomal contents in phagosome containing Artemia particles (arrows) 2h after injectionof Artemia. X22 000.

fashion as the scyphistomae concentrated algae by ciliary action, which may explainsome of the variability in the data.

Intrinsic acid-phosphatase activity was found in 21—36% of live algal cells freshly

202 W. K. Fitt andR. K. Trench

Fig. 4. Scanning electron micrographs of endocytosis oi Artemta particles by endodermalcells of scyphistomae of Cassiopeia xamachana. A. Spherical Anemia particles (arrows)adjacent to and attached to flagella of endodermal cells (d) 5 min after injection oi Anemia.X16 700. B. Phagocytosis of a mucus-bound 'packet' of Artemta particles (p) bypseudopods (arrows) 10 min after injection oi Artemta. X16 600. c. Phagosome (arrows)in the apical portion of an endodermal cell enclosing Artemta particles (p) 5 min afterinjection oi Anemia. X3300. D. Bulbous-appearing endodermal cell surface after 60 minof phagocytosis of Artemia particles. X2200.

Fig. 5. Endocytosis and fate of live freshly isolated S. microadriaticum in endodermalcells of scyphistomae of C. xamachana. Unstained, A. Contact of host plasmalemma (p)with algal cell (s) 5 min after injection of algae into coelenteron. X 14000. B. Algal cell incontact with host-cell flagellum close to endodermal cell surface. Note root fibre offlagellum (arrow). X 11 000. c. Extension of plasmalemma (large arrows) around algal cellduring phagocytosis. Note host-cell flagellum (small arrow). X9400. D. Newlyphagocytosed algal cell in tight-fitting vacuole (arrows). Note nearby lysosome containingferritin (/) and lack of ferritin in phagosome. X14700.


Fig. 5

204 W. K. Fitt andR. K. Trench

isolated from medusae of C. xamachana (Table 1). After phagocytosis, slightlyhigher incidence (between 33 and 59 %) of acid-phosphatase activity was associatedwith the algae within the phagosome. These data suggest that between 12 and 23 %of the algae phagocytosed were unable to prevent lysosome fusion. Intrinsic acidphosphatase was apparently denatured in heat-killed algae, but was found associatedwith all heat-killed algae after phagocytosis, indicating that host lysosomes had fusedwith the phagosomes.

Similarly, Table 2 shows that while ferritin is seldom found in vacuoles containingnewly phagocytosed, live freshly isolated S. microadriaticwn (Fig. 5D), it commonlyoccurs in vacuoles with heat-killed algae (Fig. 7c, D, E). These observations suggestthat some property of the live alga inhibits lysosome fusion with algal phagosomes.However, perialgal material, which often coats the surface of freshly isolated S. micro-adriaticum (Fig. 7A, C), may also be responsible for this phenomenon (Trench,

Fig. 6. Scanning electron micrographs of phagocytosis of live freshly isolated 5. micro-adriaticwn by endodermal cells of scyphistomae of C. xamachana. A. Algal cells (s)adjacent to host endodermal cells 2min after injection of algae into host coelenteron.X1200. B. Contact of algal cell (s) with host endodermal cell 5 min after injection of algae.Note host-cell flagellum adjacent to alga (arrows). X6700. c. Pseudopods (arrows) of hostendodermal cell extending over algal cell 5 min after injection of algae. X3000.


Colley & Fitt, 1981; Colley & Trench, 1983). In order to distinguish between these twoalternatives, some symbiotic scyphistomae were treated with DCMU for 3 days whilecontrol animals were maintained without DCMU. Animals in DCMU for 3 days haddegenerate-appearing algae in loose-fitting vacuoles (Fig. 9A), usually containing

Fig. 7. Endocytosis and fate of heat-killed freshly isolated S. microadriaticum in endoder-mal cells of scyphistomae of C. xamachana. A. Contact and extension of pseudopods (largearrows) around an alga. Note microfilaments (small arrows) and lysosomes containingferritin (f). Unstained. X5800. B. Extension of pseudopods completely around an alga(large arrows). Note host-cell flagellum (small arrow). Unstained. X5300. c. Juxtaposedsecondary lysosome (/) with algal phagosome (p). Note contaminating membranes offreshly isolated alga (arrows). Unstained. X21 000. D. Ferritin label (arrows) inside hostvacuole (t>) adjacent to algal cell wall. Unstained. X19 200. E. Algal cell in loose-fittingvacuole (u). Note presence of ferritin label inside vacuole (/). Stained. X5100.


3 4Time (h)

Fig. 8. Kinetics of phagocytosis of live freshly isolated S. microadnaticum by endodermalcells of the scyphistomae of C. xamachana. ( • • ) 104cells/ml; ( • D) 10*cells/ml; (A • ) 108cells/ml. Mean values from five animals (± S.D.) are plotted.

ferritin (Table 2, Fig. 9B) . Control animals had healthy-appearing algae in tight-fitting vacuoles without ferritin.

DISCUSSION

The question whether algal symbionts are digested by their invertebrate hosts hasbeen debated for over 50 years (Boschma, 1924, 1925; Yonge & Nichols, 1931;Yonge, 1936; Mansour, 1946a,b,c). Observations of disintegrating algae in the host(Oschman, 1968; Karakashian, Karakashian & Rudzinska, 1968; Weiss, 1976; Steele& Goreau, 1977), including evidence of acid-phosphatase activity associated withdegenerating algae (Fankboner, 1971; Karakashian & Karakashian, 1973), have beeninterpreted as evidence of host digestion of their symbionts. However, results of thesetypes of experiments do not rule out algal autodegradation (Muscatine, 1973; Trench,1974, 1979), especially in the light of the recent discovery that S. microadnaticum has


Table 1. Percentage o/S. microadriaticum containing acid phosphatase label before(outside scyphistomae) and after (inside scyphistomae) phagocytosis by endodermal

cells of scyphistomae ofC. xamachana

% Of algae having acid phosphatase labelTime A

(h) Outside scyphistomae Inside scyphistomae

Live algae0-25 21-0(505) 33-3(9)0-5 35-4 (225) 58-8 (34)2-0 24-3 (169) 46-1 (39)

1440-0 - 44-0 (125)

Heat-killed algae2-0 0* (50) 100-0(20)

•Algal enzymes denatured.Time = h after injection into scyphistomae or after isolation. Numbers of algae observed are given

in parentheses.

Table 2. Percentage ofvacuoles in endodermal cells of scyphistomae o/C. xamachanacontaining both S. microadriaticum andferritin 1 h and 2 h after being injected withlive freshly isolated or heat-killed freshly isolated algae (A) and in previously infectedscyphistomae maintained in DCMU or sea-water for 3 days before being injected with

ferritin (B)

Percentage

l h 2h

AFreshly isolated 8-3 (72) 9-8 (51)Heat-killed 85-7 (35) 73-1 (78)

BSea-water - 6-0 (50)DCMU - 74-2 (31)

See the text for details. Numbers of algae observed are given in parentheses.

its own intrinsic acid phosphatases (Trench etal. 1981). Our results show only a slightincrease in the proportion of live algal cells labelled with acid phosphatase afterphagocytosis into digestive cells, suggesting that most of the enzyme found originatedfrom the alga and not the host.

Two different views of 'digestion' of S. microadriaticum may be derived fromobservations of 'dead' and 'live' algae in the host cell. If one assumes that the host is'farming' its algae (cf. Yonge, 1936; Steele & Goreau, 1977), then there must be ameans of selecting which algae are digested and which are left to reproduce. Anotherpossibility is that a live healthy alga inhibits normal host digestion by inhibiting


Fig. 9. Fate of 5. nticroadriaticum in endodermal cells of scyphistomae of C. xamachanaafter being treated with the photosynthetic inhibitor DCMU for 3 days. Unstained. A.Algal cell in loose-fitting vacuole (v) and surrounded by ferritin (arrows). X 13604. B.Ferritin (arrows) inside algal vacuole (y). X32 100.

phago—lysosome fusion and possibly by resisting enzymic attack, while a dead ordying alga is unable to inhibit the normal digestive process. From this point of view,the host treats the symbiosome like any other phagosome after the alga dies or fails tofunction properly; lysosomes bind to the vacuoles and the hydrolytic enzymesreleased help to break down the contents. Our data on binding of labelled primary(acid phosphatase) and secondary (ferritin) lysosomes to vacuoles containing algae areconsistent with this latter hypothesis. Algal hydrolytic enzymes may initiate and/orfacilitate this process, thereby adding an autodegradative component.

The actual molecular mechanism of inhibition of lysosomal fusion has not beendetermined in any intracellular system. Present research has focussed on secretion ofan enzyme that blocks fusion (cf. Lowrie, Jackett & Ratcliffe, 1975) or changes theproperties of the phagosome membrane to inhibit lysosome fusion. In some systemsanionic substances inhibit phago—lysosome formation, possibly through charge repul-sion (Hart & Young, 1979). Polycationic polyanions can overcome this inhibition.Hohmanef al. (1982) found that Hydra lysosomes bind to phagosomes containing liveChlorella pretreated with polylysine, whereas few instances of lysosome-binding tovacuoles containing untreated Chlorella were found. This is consistent with the theorythat a change in the membrane charge or chemistry, possibly induced by a livesymbiont, may alter lysosome fusion with the algal vacuole. Phagosome membranesin Amoeba (Bowers, 1980), Paramecium (Allen, 1976), and Tetrahymena (Kitajima


& Thompson, 1977) have been shown to differ in particle densities and chemicalcomposition (higher protein to phospholipid ratios and glycosphingolipid content)from the plasma membrane from which they were derived. The change apparentlytakes place within 15min of ingestion and is not evident during the early stages ofphagocytosis when the phagosome membrane is still part of the plasmalemma.Freeze-fracture evidence (Meier, Reisser & Wiessner, 1980) of differences betweenmembranes of perialgal and digestive vacuoles in Paramecium bursaria suggests thatlive algae do in fact change the chemical properties of the phagosome.

Hypotheses of the method of algal uptake by marine coelenterates (Muscatine etal. 19756; Taylor, 1973; Kinzie, 1974; Trench, 1979) have been based largely onelectron microscopy of Hydra digestive cells (Cook et al. 1978; McNeil, 1981). Wepresent here direct evidence showing that the method of entry of 5. microadriaticuminto a coelenterate host is by phagocytosis by endodermal cells (confirmed by Colley& Trench, 1983). Phagocytosis of live freshly isolated algae occurs in only a fractionof the endodermal cells of scyphistomae. The average number of endodermal cellsin a 1 mm diameter scyphistoma was calculated from SEM and TEM pictures to beabout 160 000. Therefore, between 1 and 7 % of these cells take up algae, dependingon the concentration injected. This is in contrast to endocytosis of Chlorella by 95 %of the endoderm cells of Hydra viridis (P. McNeil, personal communication). Therewas apparently no difference in the modes of uptake of live or dead S. micro-adriaticum as judged by profiles seen in transmission electron microscopy. S.microadriaticum reside in individual vacuoles after engulfment. Live algae werefound in tight-fitting vacuoles at the base of the host cell, away from apical areas ofmaximum lysosomal activity 24 h after injection, whereas dead algae were found inloose-fitting vacuoles and are not seen in macerations of scyphistomae 48 h afterinjection.

Transmission and scanning EM showed that the digestive cells of scyphistomae ofC. xamachana phagocytose Artemia in a fashion similar to other coelenterates using'pseudopods', 'overlapping folds' and 'ruffles' (Gauthier, 1963; Afzelius & Rosen,1965; McNeil, 1981). There have been no reports documenting morphological dif-ferences in phagocytosis of food particles by coelenterate digestive cells.

Endocytosis of ferritin occurs in the classical pinocytotic fashion, including mem-brane invagination, fusion of microvesicles into larger vesicles (Lentz, 1966) andsubsequent fusion to lysosomes. Ferritin is also found between some cells (Fig. 1A;cf. Lentz, 1966). Levels of acid phosphatase/ferritin double-labelling of digestivevacuoles after 2h (Fig. 2) are probably minimal estimates, since heavy acid-phosphatase labelling (i.e. see Fig. 1B) tends to obliterate the ferritin label and TEMphotographs may not show both heterogeneous labels in the same vacuole in the sameEM section.

In summary, the symbiotic dinoflagellate S. microadriaticum is phagocytosed byendodermal cells of the coelenterate C. xamachana. The algae may use threemechanisms to help escape host digestive attack. First, live healthy symbionts aretransported to the base of the host cell, away from most but not all lysosomalactivity. Secondly, live algae in vacuoles inhibit lysosomal fusion. Thirdly, our data

210 W.K. Fitt and R. K. Trench

do not exclude the possibility that 5. microadriaticum is resistant to lysosomalenzymes. So far we have not seen ferritin inside degenerate algal cells. None of thesemechanisms of resisting host digestion is found with dead symbionts, food particlesor ferritin.

We thank R. Gill and D. Pierce for technical assistance on the electron microscope and Drs N.Colley, S. Fisher, A. Gibor, L. Muscatine and F. Wilkerson for comments on the manuscript. Thiswork was supported in part by NSF gTant PCM 78-15209 and U.S. AID grant DPE-5542-G-SS-1085(toR.K.T.).

REFERENCES

AFZELIUS, B. A. & ROWEN, B. (1965). Nutritive phagocytosis in animal cells. An electronmicroscopic study of the gastroderm of the hydroid. Z. Zellforsch. mikrvsk. Anat. 67, 24—33.

ALLEN, R. D. (1976). Freeze-fracture evidence for intramembrane changes accompanying mem-brane recycling in Paramecium. Cytobiologie 12, 254—273.

ARMSTRONG, J. A. & HART, P. D. (1971). Response of cultured macrophages to Mycobacteriumtuberculosis, with observations on fusion of lysosomes with phagosomes. J. exp. Med. 134,713-740.

BARRETT, A. J. & HEATH, M. F. (1977). Lysosomal enzymes. In Lysosomes, A Laboratory Hand-book (ed. J. T. Dingle), pp. 19-145. New York: Elsevier.

BOSCHMA, H. (1924). On the food of madreporaria. Pmc. K. ned. Akad. Wet. 27, 13-23.BOSCHMA, H. (1925). The nature of the association between anthozoa and zooxanthellae. Proc.

natn. Acad. Set. U.SA. 11, 65-67.BOWERS, B. (1980). A morphological study of plasma and phagosome membranes during endo-

cytosis in Acanthamoeba. J. Cell Biol. 84, 246-260.COLLEY, N. J. & TRENCH, R. K. (1983). Selectivity in phagocytosis and persistence of symbiotic

algae by the scyphistomae stage of the jellyfish Cassiopeia xamachana. Pmc. R. Soc. Land. B, 215(in press).

COOK, C. B. (1980). Infection of invertebrates with algae. In Cellular Interaction in Symbiosis andParasitism (ed. C. B. Cook, P. W. Pappas & E. D. Rudolf), pp. 47-73. Columbus: Ohio StateUniversity Press.

COOK, C. B., D'ELIA, C. F. & MUSCATINE, L. (1978). Endocytic mechanisms of the digestive cellsof Hydra viridis. I. Morphological aspects. Cytobios. 23, 17-31.

COOPER, G. & MARGULIS, L. (1977). Delay in migration of symbiotic algae in Hydra viridis byinhibitors of microtubule protein polymerization. Cytobios. 19, 7-19.

FANKBONER, P. V. (1971). Intracellular digestion of symbiotic zooxanthellae by host amoebocytesin giant clams (Bivalvia:Tridacnidae), with a note on the nutritional role of the hypertrophiedsiphonal epidermis. Biol. Bull. mar. biol. Lab., Woods Hole 141, 222-234.

FITT, W. K. (1982). Mechanisms of Infection of Invertebrate Hosts with the SymbioticDinoflagellate Symbiodinium microadriaticum Freudenthal. Ph.D. thesis, University of Califor-nia, Santa Barbara.

FITT, W. K., CHANG, S. S. & TRENCH, R. K. (1981). Motility patterns of different strains of thesymbiotic dinoflagellate Symbiodinium (=Gymnodinium) microaddriaticum (Freudenthal) inculture. Bull. mar. Sd. GulfCaribb. 31, 436-443.

FITT, W. K. & TRENCH, R. K. (1983). Infection of coelenterate hosts with the symbioticdinoflagellate Symbiodinium microadriaticum. In Endocytobiology (ed. W. Schwemmler & H. E.A. Schenk), (in press).

GAUTHIER, G. F. (1963). Cytological studies on the gastroderm of Hydra. J. exp. Zool. 152, 13-39.HART, P. D. & YOUNG, M. R. (1979). The effect of inhibitors and enhancers of phagosome-

lysosome fusion in cultured macrophages on the phagosome membranes of ingested yeast. ExplCell Res. 118, 365-375.

HOHMAN, T. C , MCNEIL, P. L. & MUSCATINE, L. (1982). Phagosome-lysosome fusion inhibitedby algal symbionts of Hydra viridis. J. Cell Biol. 94, 56-63.


KARAKASHIAN, M. W. & KARAKASHIAN, S. J. (1973). Intracellular digestion and symbiosis inParameciwn bursaria. Expl Cell Res. 81, 111-119.

KARAKASHIAN, S. J., KARAKASHIAN, M. W. & RUDZINSKA, M. A. (1968). Electron microscopicobservations on the symbiosis of Paramecium bursaria and its intracellular algae. J'. Protozool.15, 113-128.

KARAKASHIAN, S. J. & RUDZINSKA, M. A. (1981). Inhibition of lysosomal fusion with symbiont-containing vacuoles in Paramecium bursaria. Expl Cell Res. 131, 387-393.

KARNOVSKY, M. J. (1965). A formaldehyde-glutaraldehyde fixative of high osmolality for use inelectron microscopy. J. Cell Biol. 11, 137-138A.

KINZIE, R. A. I l l (1974). Experimental infection of aposymbiotic gorgonian polyps withzooxanthellae. J. exp. mar. Biol. Ecol. 15, 335-345.

KITAJIMA, Y. & THOMPSON, G. A. JR (1977). Differentiation of food vacuolar membranes duringendocytosis in Tetrahymena. J. Cell Biol. 75, 436-445.

LENTZ, T. L. (1966). The Cell Biology of Hydra. Amsterdam: North Holland.LOWRIE, D. B., JACKETT, P. S. & RATCLIFF, N. A. (1975). Mycobacterium microti may protect

itself from intracellular destruction by releasing cyclic AMP into phagosomes. Nature, Land. 254,600-602.

MANSOUR, K. (1946a). Source and fate of the zooxantthellae of the visceral mass of Tridacnaelongata. Nature, Land. 158, 130.

MANSOUR, K. (19466). Communication between the dorsal edge of the mantle and the stomach ofTridacna. Nature, Land. 157, 844.

MANSOUR, K. (1946C). The zooxanthellae, morphological peculiarities and food and feeding habitsof the Tridacnidae with reference to other lamellibranchs. Proc. Egyptian Acad. Set. 1, 1—11.

MCNEIL, P. L. (1981). Mechanisms of nutritive endocytosis. I. Phagocytic versatility and cellularrecognition in Chlorohydra digestive cells, a scanning electron microscope study. J. Cell Set. 49,311-339.

MCNEIL, P. L., HOHMAN, T. L. & MUSCATINE, L. (1981). Mechanisms of nutritive endocytosis.II. The effect of charged agents on phagocytic recognition by digestive cells. J . Cell Sci. 52,243-269.

MEIER, R., REISSER, W. & WIESSNER, W. (1980). Freeze-fracture evidence of differences betweenmembranes of perialgal and digestive vacuoles in Paramecium bursaria. Z. Naturf. 35,1107-1110.

MUSCATINE, L. (1973). Nutrition of corals. In Biology and Ecology of Coral Reefs (ed. O. A. Jones& R. Endean), vol. 2, pp. 77-115. New York: Academic Press.

MUSCATINE, L., COOK, C. B., PAKDY, R. L. & POOL, R. R. (1975a). Uptake, recognition, andmaintenance of symbiotic chlorella by Hydra viridis. Symp. Soc. exp. Biol. 29, 175-203.

MUSCATINE, L., POOL, R. R. & TRENCH, R. K. (19756). Symbiosis of algae and invertebrates:aspects of the symbiont surface and the host-symbiont interface. Trans. Am. microsc. Soc. 94,450-469.

O'BRIEN, T. L. (1980). The symbiotic association between intracellular zoochlorellae(Chlorophyceae) and the coelenterate Anthopleura xanthogrammica. J. exp. Zool. 211, 343-355.

OSCHMAN, J. L. (1966). Apparent digestion of algae and nematocysts in the gastrodermalphagocytes of Chorohydra viridissima. Am. Zool. 6, 320.

SMITH, D. C. (1980). Principles of the colonisation of cells by symbionts as illustrated by symbioticalgae. In Endocytobiology (ed. W. Schwemmler & H. E. A. Schenk), vol. 1, pp. 317-332. NewYork: Walter de Gruyter.

STEELE, R. D. & GOREAU, N. I. (1977). The breakdown of symbiotic zooxanthellae in the seaanemone Phyllactis (=Oulactis) flosculifera (Actinaria). J. Zool. 181, 421-437.

TAYLOR, D. L. (1973). The cellular interactions of algal-invertebrate symbiosis. Adv. mar. Biol.11, 1-56.

TRENCH, R. K. (1974). Nutritional potentials in Zoanthus sociatus (Coelenterata, Anthozoa).Helgoldnder wiss. Meeresunters. 26, 174—216.

TRENCH, R. K. (1979). The cell biology of plant-animal symbiosis. A. Rev. PI. Physiol. 30,485-532.

TRENCH, R. K. (1980a). Integrative mechanisms in mutualisticsymbioses. In Cellular Interactionsin Symbiosis and Parasitism (ed. C. B. Cook, P. W. Pappas & E. D. Rudolf), pp. 275-297.Columbus: Ohio State University Press.

212 W.K. Fitt and R. K. Trench

TRENCH, R. K. (19806). Uptake, retention and function of chloroplasts in animal cells. In Endo-cytobiology (ed. W. Schwemmler & H. E. A. Schenk), vol. 1, pp. 703-727. New York: Walterde Gruyter.

TRENCH, R. K. (1983). Dinoflagellates in non-parasitic symbioses. In The Biology ofDinoflagellates (ed. F. J. R. Taylor). Oxford: Blackwell Scientific (in press).

TRENCH, R. K., COLLEY, N. J. & Frrr , W. K. (1981). Recognition phenomena in symbiosesbetween marine invertebrates and "zooxanthellae"; uptake, sequestration, and persistence. Ber.dt. bot. Ges. 94, 568-577.

WEIS, D. S. (1976). Digestion of added homologous algae by chlorella-bearing Parameaum bur-saria.J. Protozool. 23, 527-529.

YONGE, C. M. (1936). Mode of life, feeding, digestion and symbiosis with zooxanthellae in theTridacnidae. Set. Rep. Gt Barrier Reef Exped. 1, 283-321.

YONGE, C. M. & NICHOLS, A. G. (1931). Studies on the physiology of corals. V. The effect ofstarvation in light and darkness on the relationship between corals and zooxanthellae. Set. Rep.Gt Barrier Reef Exped. 1, 177-211.

(Received 11 April 1983-Accepted 9 May 1983)

Documents

Printed in Great Britain · The fate of particles endocytosed by coelenterate endodermal cells has only recently been investigated. Previous investigators noted acid-phosphatase activity