Mechanisms of nutritive endocytosis i. phagocytic versatility and cellular recognition in Chlorohydra digestive cells, a scanning electron microscope study

During phagocytosis particles are enveloped by a radically restructured domain of plasma membrane. A vital part, therefore, of any mechanistic understanding of phagoctyosis is the precise determination of phagocytic morphology. Scanning electron microscopy (SEM), with its unrivaled 3-dimensional images of cellular surface, has often been used in studies of phagocytizing mammalian cells (Tizard & Holmes, 1973; Kaplan, Guidenack & Seljelid, 1975; Polliak & Gordon, 1975; Walters, Papadimitriou & Robertson, 1976; Jones, Minick & Young, 1977; Kaplan, 1977; Orenstein & Shelton, 1977). The most detailed of these, by Kaplan (1977), and Orenstein & Shelton (1977), showed that several morphologically distinct modes may govern particle uptake, and that morphologically, therefore, the macrophage can be characterized as a versatile phagocyte. Much work, primarily biochemical, has focused on the mechanism of phagocytic recognition in macrophages (Silverstein, Steinman & Cohn, 1977; Stossel, 1978).

By contrast, little is known of how nutritive cells phagocytize (Afzelius & Rosen, 1965; Weisman & Korn, 1967; Rabinovitch & Stefano, 1971; Korn, 1975; Bowers, 1977; Chapman-andresen, 1977). It is not known if, like the macrophage, a nutritive phagocyte utilizes a morphologically versatile phagocytic repertoire, or by what means it recognizes different particles.

Digestive cells of Chlorohydra viridissima (Cnidaria: Hydridae) possess several advantages for the detailed SEM characterization of phagocytic versatility and recognition. Chlorohydra depends on endocytosis as its principle means of obtaining external nutrients (Lenhoff, 1974). Its digestive cells might, for this reason, be expected to possess a versatile phagocytic apparatus, capable of taking up a variety of different particle types. All experiments can be carried out in vivo, avoiding 2 complications inherent in using mammalian macrophages; a requirement for serum factors (Silverstein et al. 1977), and the preliminary in vitro flattening of experimental cells on an artificial substratum. Since the digestive cell is located in the endodermal layer of Chlorohydra, the phagocytically active surface is protected from mechanical damage during routine microscopal preparation; it is not exposed, by longitudinal splitting of the animal, until after critical-point drying. Finally, the digestive cell in vivo internalizes a wide range of particles at rates which can be readily quantified (Muscatine, Cook, Pardy & Pool, 1975; Pool & Muscatine, 1980); comparisons between quantitative and morphological aspects of phagocytosis are therefore easily made.

The object of this paper is to determine whether digestive cells display phagocytic versatility and recognition and, by electron microscopy, to determine the detailed morphological characteristics of the versatility, and to describe morphological correlates of the phagocytic recognition process.

Chlorohydra viridissima (Florida strain) were maintained at 18 °C in ‘M’ solution (Muscatine & Lenhoff, 1965) on a 12-h light/12-h dark photoperiod (2·1 × 10⁻⁵ einsteins m⁻²s⁻¹; Cool-white, Sylvania, fluorescent tubes in a Precision Science Incubator). Aposymbiotic C. viridissima were raised from algae-free embryos. Animals in mass culture were fed Artemia nauplii daily; those used in experiments were starved for 72 h before use.

Freshly isolated symbiotic algae (FIS) were obtained from Chlorohydra as described by Muscatine, Pool & Trench (1975) and used in experiments within 2 h. Heat-treated symbionts (HTS) were prepared by heating FIS in a boiling water bath for 5 min, and then washing (resuspension in 10 ml M solution followed by centrifugation at 2500 g for 5 min) 5 times in M solution. Heat-treatment did not significantly change the FIS size or shape (light and SEM observations). The free living algae (FA), Chlorella vulgaris (strain 3978), cultured in modified Loefer’s medium (Karakashian, 1963), were washed 10 times with M solution prior to use. Latex spheres (LS) (5·7 μm; polyvinyltoluene, Sigma) were dialysed overnight in M solution before injection into aposymbiotic gut. For Artemia particles (0·45-μm AP) whole shrimp were centrifuged (2500 g, 2 min) to give a 2-ml wet-packed volume, ground in 10 ml M solution on ice, with 12 up -and-down strokes of a Teflon pestle (Thomas, no. 531), centrifuged (2500g, 5 min) to remove larger debris and then filtered twice through a 0·45-μm Millipore filter. These 0·45-μm AP were stored on ice and used within 2 h. Particle diameters were estimated from light microscope measurements of over too specimens, they were (μm±S.E.): LS, 5·8±0·03; FIS, 3·510·01 (long axis); HTS, 3·60±0·02 (long axis); FA, 3·86±0·02.

Chlorohydra digestive cells were phagocytotically challenged by microinjection (Pardy & Muscatine, 1973) of particles into 72-h starved aposymbiotic hydra gut. Prior to injection, FIS, HTS and LS pellets (all wet-packed by centrifugation at 2500 g, for 5 min) were suspended 1 :1 (v:v) in M solution; 0·45-AP were injected without dilution. The final concentrations of the injected particles, as determined by haemocytometer counts, were (particles/μ l ± S.E.M.): FIS, 4·7 (±7)× 10⁷; HTS, 4·7×10⁷; FA, 3·5×10⁷; LS 2·2×10’. These concentrations exceeded by at least 5-fold the saturating doses for injected Chlorohydra described by Pool & Muscatine (1980).

Rates of FIS, HTS, FA and LS uptake were measured as described by Pool (1979). This method gave the number of particles interiorized per digestive cell. However, since it was desired to compare the uptake of particles of very different sizes (such as the s L.S and 3·75-μm maximum diameter FIS), data were converted to surface area of membrane interiorized per digestive cell using the simple formula:

Surface area (μm²) membrane interiorized per digestive cell = number particles per digestive cell × surface area particle.

It was assumed for this calculation that the surface area of the particle equalled the surface area of the digestive-cell membrane required for interiorization of that particle. Algal cells (FIS, HTS and FA) were assumed to approximate the shape of an oblate spheroid and LS that of a sphere, and appropriate equations (CRC Handbook of Chemistry and Physics, 1964) were then used to calculate the surface areas of the particles.

Animals were relaxed in 1% urethane in M solution for 2 min at 21 °C. Urethane was then removed, and rapidly replaced by 1% formaldehyde (freshly generated from paraformaldehyde) and 2% glutaraldehyde in 0·05 M-cacodylate buffer (pH 7·4). Following 1 h of this primary fixation, specimens were rinsed in 0·05 M-cacodylate buffer, post-fixed at 4 °C for 1 h in 1% OsO₄, rapidly dehydrated in a graded series of ethanol, critical-point dried, and mounted on double-stick tape (‘Scotch’ brand). Each animal was then split down its longitudinal axis, as described by Westfall & Townsend (1977), using a new double-edged razor blade (Gillette ‘Blue Blade’), and carefully opened to reveal the gut surface. To improve SEM visualization of cell surfaces, excess challenge particles were removed by air puffs directed at the split animals. Finally, specimens were sputter-coated with carbon and gold palladium (Hummer sputter-coater), and examined at 10 kV on an ETEC scanning electron microscope at angles of 0 to 60°.

Animals were prepared for TEM by one of two separate fixation techniques. In the first, animals were fixed in the formaldehyde/glutaraldehyde primary fixative already described for SEM. In another, the primary fixative was a Schiff’s base, used as described by Hauser (1978) in 0·05 M-cacodylate buffer. Schiff’s base fixation enhanced TEM visualization of a fuzzy material on digestive-cell and FIS surfaces. Animals fixed by either technique were post-fixed in OsO₄ (1%, 1 h, 4 °C), dehydrated in a graded acetone or ethanol series, and then embedded in Spurr (1969) plastic. Silver and grey sections, cut on an LKB (model III) ultramicrotome and stained in uranyl acetate, were examined on a Philips 200 transmission electron microscope.

Challenge with LS and FIS resulted in phagocytic internalization of significantly more (Student’s Z-test, P < 0·05) plasma membrane at each time-interval measured, than did either HTS or FA challenge (Fig. 1). This discriminatory ability shows that the digestive cell is capable of phagocytic recognition. Most of the phagocytizing preparations used in this study were fixed at between 5 and 30 min post-challenge, when, as shown in Fig. 1, uptake proceeds most rapidly.

The appearance of digestive-cell surfaces before phagocytosis are shown in Fig. 2 A. Each of the 3 gut regions is composed of cells having a characteristic surface morphology. Observations reported in this study have been confined to the gastric region, the middle and most phagocytically active of the three. In the gastric region the distal surface of the unchallenged digestive cell (Fig. 2B) is covered sparsely with microvilli and small folds, and is approximately hemispherical in shape.

Urethane treatment greatly enhanced the reproducable fixation of hydra in an extended (relaxed) state. SEM revealed no apparent morphological differences between digestive, cell-surface morphology in urethane-treated, and untreated control animals. Benos, Kirk, Barba & Goldner (1977) similarly found no effect of urethane in a TEM study of hydra.

To ascertain whether digestive cells exhibited phagocytic versatility, and/or morphological correlates of the phagocytic recognition process described above, animals were challenged with various particles (FIS, HTS, LS, FA and 0·45-μm AP), fixed at intervals of 0–60 min thereafter, and prepared for SEM (and TEM). For each particle type, numerous images of phagocytosis (25–50) from at least 2 separate experiments were recorded. These images could be so arranged as to describe 3 distinct phagocytic modes, each of which is detailed in the following sections.

FIS challenge evokes the rapid (15 min post-injection) formation of a dense cover of microvilli † - on digestive-cell surfaces (Fig. 3 A). Fig. 4A-D demonstrates that such microvilli participate in FIS uptake. In this ‘microvillar’ uptake mode, initial contact of digestive-cell and FIS surfaces is made by shorter microvilli (< 1 μm in length). A few, or even single microvilli may initiate envelopment of the FIS, and in so doing these may sometimes extend for 2–3 μm over the FIS circumference (Fig. 4A). Although some enveloping microvilli are closely applied throughout their course over the alga, others make algal contacts discontinuously, with intervals of intimate association matched by areas of relative separation (Fig. 8). In either case, initial contacts are followed by microvillar advance over the alga (Fig. 4A, B, c), forming an enclosing network of separate, and sometimes overlapping, microvilli (Fig. 4B). The final stages of particle enclosure by microvilli are: the conversion of an overlapping into an interlocking network of microvilli (Fig. 4 c); the appearance of a continuous surface, with, however, its formerly tubular nature still evident (Fig. 4D); and finally, the disappearance of all signs of its formerly tubular nature, leaving a rough, but continuous surface around the newly enclosed FIS (Fig. 4E, F). In the funnel uptake mode described below, a broad, continuous surface (funnel) envelops the FIS, but at no stage does it show the rough, highly textured aspect seen in the final stages of enclosure by microvilli.

TEM reveals numerous microvillar profiles surrounding the alga sectioned tangentially to the phagocyte surface (Fig. 5 A). Cytoplasmic constituents, including the discoidal coated vesicles (Slautterback, 1967) known to mediate ferritin uptake in hydra, are excluded from such microvilli (Fig. 5B, c). Similarly, Cook, D’Elia & Muscatine (1979) have reported on the lack of involvement of discoidal coated vesicles in FIS uptake. The often intimate association of enveloping microvilli and FIS surfaces is readily visualized using TEM (Fig. 5 c). Contacts between adjacent microvillar profiles are frequent, and may represent multiple membrane fusions completing the enclosure of FIS (Fig. 5B, c). Microfilaments were not discerned within the microvilli of hydra (Fig. 5 c), but these have been difficult, in general, to preserve in the digestive cell (Cook et al. 1979).

A second mode of FIS uptake, the ‘funnel’ phagocytic mode, was readily distinguished using SEM. In this mode symbionts are enclosed by a continuous, smooth-surfaced, tightly fitting, funnel-like extension of the phagocyte surface (Fig. 6A-E), rather than by an enveloping network of microvilli. Animal-algal surface interactions in the funnel uptake mode are initiated by broad, non-microvillar surface folds (Fig. 6A). While 2 or more of these broad extensions may initiate FIS engulfment by the funnel (Fig. 6A), later stages (1/4 or more of the FIS enclosed) always show a single continuous, cylindrical enveloping surface (Fig. 6B-E). Enclosure proceeds as this smooth, cylindrical surface is extended over the symbiont (Fig. 6B-E). Intimate circumferential contact is generally maintained between the leading edges of the advancing surface and the partly enclosed symbiont. Not infrequently, one of the numerous gastrodermal flagella adhere to the partly enveloped FIS (6B, C). Whether this is a functional or simply fortuitous association remains an interesting, but unanswered, question. Enclosure is completed by membrane fusion at a single limiting aperture, on the tip of the nearly closed funnel (Fig. 6E). The completely engulfed symbiont is then drawn into the gastrodermal cytoplasm. Enclosure by the funnel uptake mode was sometimes (57% of 30 total funnel recordings), but not always, characteristic of a relatively smooth-surfaced (few microvilli and folds) digestive cell. The interior surface of empty funnels often displayed tubular protrusions that were especially concentrated in the funnel base (Fig. 7A, B); they did not, as did the leading funnel edge, conform in detail to the smoothly oval FIS surface.

That a single digestive cell may respond to FIS with either microvillar or funnel phagocytic modes was strikingly demonstrated by the simultaneous envelopment of a single FIS by each of these 2 modes (Fig. 8): a smooth continuous funnel surface at one end of the oval FIS, and multiple microvilli at the other.

TEM clearly shows that intimate animal-algal contact is maintained at leading edges of advancing funnels (Fig. 9A-D). In contrast, other regions of enveloping funnels often show a marked non-conformity of algal and internal funnel surfaces (Fig. 9A, B). Fortuitous micrographs of serial (non-consecutive) thin sections show a putative membrane-fusion event at, as was proposed earlier, the tip of a nearly closed funnel (Fig. 90).

A final relevant feature of FIS uptake is a ‘fuzzy material’, a descriptive term used here for amorphous material seen on particle and digestive-cell surfaces in both SEM and TEM micrographs. Fuzzy material appears on the surfaces of these algae after their injection into the gut of Chlorohydra (Fig. 3 A). The presence of this fuzzy material often made sharp SEM visualization of FIS uptake difficult, especially of those smaller projections involved in the microvillar uptake mode; fuzzy material, when present on both FIS and digestive-cell surfaces, makes the SEM images of these surfaces appear slightly out of focus (see, e.g. Fig. 4F). Thus, while the majority of FIS had a fuzzy coating, many of the SEM (but not TEM) micrographs selected for presentation here show little evidence of this fuzzy material on their visible surfaces (see, e.g., Fig. 4A, B), although it may have been evident by TEM.

The presence of fuzzy material was analysed by applying a simple, semi-quantitative stereological analysis to randomly selected, high-resolution SEM micrographs of particle surfaces. Table 1 shows that 26·4 ±7·3% of the uningested FIS surface is covered with patches of fuzzy material. This, however, may be an underestimate since the higher resolution of TEM revealed a ubiquitous but very fine coat of fuzzy material on the thin-sectioned FIS surface (Figs. 5, 9, 10).

FIS do not demonstrate fuzzy material before injection into the gut of hydra (Fig. 3B). TEM was used to characterize further the surface and intracellular distribution of fuzzy material. Such analysis clearly showed that fuzzy material is found also on digestive-cell surfaces, including phagocytically active microvillar and funnel surfaces (Figs. 5B, 9 A, 10 A). TEM of the completed but newly formed phagosome reveals that fuzzy material may also be found in that space bounded by phagosomal membrane and external FIS surface (Fig. 10B). In other phagosomes fuzzy material is found predominantly on the internal phagosomal surface (Fig. 10c). Less recently formed phagosomes, identified by their more basal position in the digestive cell (Muscatine et al. 1975), did not contain fuzzy material. Careful study of recently formed phagosomes indicates that small vesicles, some perhaps containing the fuzzy material, are associated with the recently formed phagosome (Fig. 10A, B, C). Much larger intracellular vesicles were often found in close proximity to the recently enclosed symbiont, and these also frequently contained presumptive fuzzy material (Fig. 10A).

LS, like FIS, were phagocytized to a relatively large extent. Like FIS, LS had a spotty coating of fuzzy material on 13·9 ± 0·4% of their surfaces after injection into the gut of Chlorohydra (Table 1). LS were also engulfed by the funnel uptake mode, with intimate contact of advancing funnel surface with particle surface (Fig. 11 A, B). Unlike FIS, however, LS were never observed to be enveloped by microvilli, nor did challenge with LS elicit microvilli formation on the digestive-cell surfaces.

To determine what morphological features might be associated with phagocytic recognition, those particles taken up at reduced rates (HTS, FA) were also examined with SEM. Several features distinguished the uptake of FIS from that of HTS. First, HTS injected into Chlorohydra were not blemished with the fuzzy material (Fig. 12A, B, c), only 0·7± 0·4% of their surface had any (Table 1). TEM similarly fails to reveal any fuzzy material on the HTS or HTS-challenged digestive-cell surfaces (Fig. 13). And, in thin-section phagosomes, no fuzzy material filled the space bounded by animal phagosomal membrane and algal cell wall in the phagosome. Secondly, HTS were never enveloped by microvilli, nor did HTS evoke the formation of microvilli on digestive-cell surfaces. Where uptake events were observed, enclosure by broad, funnel-like extensions of digestive-cell surface prevailed (Fig. 12 A, B, c). A third distinguishing feature of HTS phagocytosis was the frequently less intimate contact (than for FIS uptake) of advancing animal surface with HTS surfaces (Fig. 12B, c); HTS were enveloped by what is therefore called a modified funnel uptake mode.

The surface events that governed the uptake of FA, another particle whose extent of uptake by digestive cells is reduced, were examined. FA, like HTS, had little (Table 1) fuzzy material on their surfaces after injection into the gut of Chlorohydra (Fig. 14). FA were also engulfed by the modified funnel uptake mode, with its typically less-intimate algal-animal contact at the leading edge of the enveloping funnel surface (Fig. 14). Finally, completing the resemblance to HTS uptake, FA were never enveloped by microvilli, and did not stimulate the formation of microvilli on digestive-cell surfaces.

The fifth and final particle whose uptake was studied with SEM was the 0·45-μm (maximum diameter) AP. In earlier experiments Chlorohydra were prepared for SEM after feeding of whole Artemia nauplii, but phagocytosis under these conditions was difficult to visualize. Mucus and large, irregularly shaped particles often masked the phagocytically active digestive cells. When Artemia were homogenized, and the homogenate passed through a 0·45-μm filter (see Materials and methods), the particles I have called 0·45-μm AP were produced. When these particles were injected into Chlorohydra gut their uptake was readily visualized. In animals fixed 1–2 min after injection, particles of 0·45-μm diameter were found, as would be expected, adhering to gastrodermal and flagellar surfaces. If fixation was delayed until 10–15 min after injection, much larger (6·25-μm) particles predominated (referred to hereafter as 6·25-μm AP). These larger 6·25-μM AP were, with few exceptions, associated with the tips of digestive-cell flagella (Fig. 15 A, B, c, D, F). It is tempting to suggest that following their injection, 0·45-μM AP were compacted into the much larger 6·25-μM AP, and furthermore, that such compaction was mediated, in part at least, by flagella. The validity of these suggestions is currently being tested, since they would document a novel role for that versatile organelle, the flagellum (Satir, 1976).

While it was not possible to quantify the rate of 6·25-μM AP uptake, the phagocytic morphology of 6·25-μM AP was found to be distinct from that of the other 4 particles studied and so defined a third phagocytic mode - uptake by multiple folds (Fig. 15A-F). In this mode, particle envelopment is initiated by multiple, broad folds of the digestive-cell surface (Fig. 15 A). Although these folds conform to the spherical outline of the 6·25-μm AP, they do so without establishing that intimate, circumferential association of particle and leading funnel edge, which typified FIS and LS uptake by funnels (Fig. 15 c, D). It is possible that such enveloping folds arose from the similar, but less-broad folds that frequently texture phagocytically uninvolved surfaces (Fig. 15B). Envelopment proceeds by the movement of 2 or sometimes 3 folds over the 6·25-μm AP surface, again without the intimate contact typical of the funnel uptake mode (Fig. 14C-F). As multiple folds move over 6·25-μm AP they clearly retain their separate identities; fusion of advancing folds is not apparent until some point after the particle is covered by enveloping surface, at which time it is possible that the lateral edges of such fully advanced folds then ‘zipper’ up towards their distal tips (Fig. 15F). Particle interiorization is presumably completed by withdrawal of the completed phagosome into the cytoplasm of the digestive cell.

This study shows that the Chlorohydra digestive cell is a versatile nutritive phagocyte, capable of enveloping particles by at least 3 morphologically distinct phagocytic modes. Summarized schematically in Fig. 16 these are: I, Microvillar; II, Funnel; and III, Multiple-fold modes. Sequences of the latter 2 modes have also been described in phagocytosis by peritoneal macrophages in vitro and in situ (Orenstein & Shelton, 1977). Particle envelopment by microvilli, however, seems unique to Chlorohydra digestive cells. Interestingly, Van Oss (1978) has recently suggested, based solely on physical considerations of interacting particle and phagocyte surfaces, that those surface extensions with smaller radii of curvature, such as micro villi, would experience less resistance to the enclosure of certain particles. There may, in other words, be some physical advantage to utilizing microvilli for particle envelopment.

What seems to be a diverse array of phagocytic morphologies have been described in macrophages (Goodall & Thomspon, 1971 ; Tizard & Holmes, 1973 ; Kaplan et al. 1975; Polliak & Gordon, 1975; Walters et al. 1976; Jones et al. 1977; Kaplan, 1977; Orenstein & Shelton, 1977; Seitz, Weiblen & Claviez, 1977). However, since the preparations used varied from one report to another, the observed differences may have been artificial. Also, while most studies have proposed a specific phagocytic morphology, few have presented convincing representative sequences in justification. Without these one cannot unambiguously conclude that a cell displays 2 or more uptake modes; different stages of a single mode might be mistaken for 2 separate modes, or a fortuitous association of particle and phagocyte surfaces mistaken as representing a new mode. A macrophage study (Orenstein & Shelton, 1977), and this report on the Chlorohydra digestive cell, have both used critical-point drying (Bodye, Bailey, Jones & Tamarin, 1977) to provide the desired representative sequences of uptake modes. Here, 3 uptake modes were demonstrable in a single cell-type, consistently fixed and prepared by the same techniques. Phagocytic versatility cannot therefore be discounted as an artifact of SEM preparation. Rather, it is an interesting and mechanistically unexplained phagocytic property, now documented in 2 evolutionary and functionally very different phagocytes, the mammalian macrophage and the digestive cell of the cnidarian, Chlorohydra.

Where phagocytic versatility is observed, 2 or more phagocytic mechanisms are suggested. Thus, in addition to identifying putative generators of motile force (Hartwig, Davies & Stossel, 1977; Stossel, 1978; Stendahl, Hartwig, Brotschi & Stossel, 1980), and surface-orienting interactions of particle and phagocyte (Griffin, Griffin & Silverstein, 1976; Michl & Silverstein, 1978), a mechanistic explanation of phagocytosis must account also for the cellular skeleton upon which surfacetransforming motive force acts, and must account precisely for the membrane movements thus produced (phagocytic morphology). The funnel and multiple-fold modes closely resemble each other morphologically, and may therefore be slight variations on a single mechanistic theme. The ‘zipper’ model of phagocytosis, as proposed for the macrophage (Michl & Silverstein, 1978), although compatible with the morphology of the funnel uptake mode of Chlorohydra, requires circumferential contact of particle and phagocyte surfaces. This criterion is not fulfilled by multiple-fold and especially microvillar uptake modes. The microvillar uptake mode may, in fact, be driven by a distinctly different cellular and molecular apparatus. Additional information as to how microvilli are formed, and how and when they are extended, will be essential for a mechanistic understanding of phagocytic engulf-ment by the microvillar mode.

The demonstration of phagocytic versatility raises the question of what factor(s) evokes a particular uptake mode? Where several uptake modes have been observed in macrophages, one study has suggested that particle surface (Kaplan, 1977), and another, that responding phagocyte surface (as it is affected by culture history) (Orenstein & Shelton, 1977), determines the uptake mode. The experimental design used here suggests that evocation of a particular uptake mode is a function of the nature (surface, chemical?) of the particle challenge: identically prepared digestive cells enclose FIS by the microvillar and funnel uptake modes; HTS by the loosefitting funnel surface only; LS by the funnel uptake mode only; and 6·25-μm AP by the multiple-fold mode only. Yet it is possible that 2 distinct populations of digestive cells existed in the gastric region of the gut of Chlorohydra, each using a different mode. Inconsistent with this possibility, however, was the observation of a single digestive cell utilizing both microvillar and funnel uptake modes simultaneously. In addition, functional or morphological heterogeneity has not been reported in the digestive cells of the gastric region.

It remains for future work to identify those specific chemical and physical attributes that are responsible for specific evocation of the 3 uptake modes. This study does show, however, that initiation of each of the 3 phagocytic modes requires, at most, only limited particle to digestive-cell surface interaction. In the funnel uptake mode, for example, FIS typically made contacts with the phagocyte surface at the rounded apex of their oval surfaces (Figs. 4, 6) and in the multiple-fold mode intimate contact between particle (6·25-μM AP) and phagocyte surfaces was never extensive (Fig. 15). Surface properties are only one aspect of the phagocytic challenge to which digestive cells might respond with a particular mode. Particle size, and/or the production of a soluble chemical messenger, which might be exuded from particles and affect digestive cells at a distance, cannot be ruled out. AP (6·25-μm) were, for example, of greater diameter on average than FIS, and FIS are known to exude various carbohydrates and amino acids (Muscatine, 1965).

The present study provides 2 types of evidence for the existence of phagocytic recognition in digestive cells. The first is manifested in the different uptake rates for the various particles tested. While this type of phagocytic recognition has been documented previously in the digestive cell (Pardy & Muscatine, 1973 ; Pool, 1979), it should be noted here that the degree of specificity involved may be low (McNeil, Hohman & Muscatine, unpublished data). The second, on the other hand, is manifested morphologically, in the selection of a particular uptake mode by the digestive cell for a particular particle type: thus only FIS are enveloped by the microvillar mode, FIS and LS by the tightly fitting funnel mode, HTS and FA by the loosely fitting (relative to FIS and LS) funnel mode, and only 6·25-μm AP by the multiplefold mode. Further work will be required to establish the degree of specificity involved and the relationship, if any, between these 2 differently defined processes of phagocytic recognition in digestive cells.

There is a possible functional correlation between the extent of the uptake of a particle and several morphological features of the variously challenged digestive-cell and particle surfaces (Table 2). In the first place, the avidly phagocytized FIS evoke an uptake mode, the microvillar mode, which HTS do not. It is thus possible to suggest that differences in the extent of phagocytic uptake (phagocytic recognition) do not necessarily result from quantitative differences in the particle-phagocyte interactions of a single uptake mode. Rather, the uptake rate may result in part from a qualitative switch in uptake mode, made as a response to a particular particle type (such as the evocation of the microvillar uptake mode in the digestive cell challenged with FIS).

The morphology of the uptake of HTS and FA resembled that of the funnel mode as described for FIS. They were not, however, identical; there was a less intimate association of alga (HTS and FA) and the enveloping animal surface than for the FIS. It is tempting to speculate that the reduced uptake of these 2 particles, and its morphological correlate of less intimate particle-phagocyte interaction, may be due partly to a weaker algal-animal surface interaction. It is also possible that shrinkage differences have simply introduced an artifact of SEM preparation into the modes of HTS and FA uptake.

A third surface phenomenon, which may be correlated with extent of uptake by digestive cells, is the appearance of a fuzzy material on FIS and LS surfaces. Because fuzzy material was observed on FIS and LS surfaces only, the 2 particles avidly phagocytized by digestive cells, it may function to enhance the uptake of those particles it coats. In studies on amoeba (Korn & Weisman, 1967), a surface coating has been described, on latex spheres undergoing phagocytosis and in the phagosomes of recently engulfed spheres. Similarly, in studies on the macrophage (Eguchi, Sannes & Spicer, 1979), cytochemical staining has been used to reveal a complex carbohydrate coating on latex spheres and challenged macrophage surfaces, and in the newly formed phagosomes. Neither study noted any correlation between surface coat and phagocytic recognition, as suggested here. Further work on the relation of surface coats to phagocytosis is clearly indicated.

Since the fuzzy material was not seen on uningested FIS, but did appear on LS (which cannot have produced it themselves) and also on the digestive-cell surfaces following a LS or FIS challenge, it follows that the fuzzy coating is a product of animal origin, not a particle contaminant. TEM observations are compatible with the following intracellular pathway for fuzzy material (Fig. 17): from its initial location on the FIS, LS and challenged digestive-cell surfaces, the fuzzy material enters into the newly formed phagosome; it is then sequestered onto the internal phagosomal membrane surface, and finally into discrete outpocketings of phagosomal membrane. These outpocketings pinch off, leaving the phagosome to fuse perhaps with much larger vesicles nearby, which in thin section also appear to contain fuzzy material. Recycling of the fuzzy material for future phagocytic efforts might then occur by release of fuzzy material from these larger intracellular vesicles into the gut of Chlorohydra. The intracellular pathway suggested here is, of course, specula-tory, since it is based purely on the static evidence of TEM micrographs.

In addition to its characterization of phagocytic morphology, this SEM study introduces 2 previously unsuspected and novel cellular phenomena: (1) the rapid formation of microvilli on digestive-cell surfaces (a fourth correlate of phagocytic recognition in the digestive cell), induced by the presence of a phagocytically attractive challenge (FIS), and (2) a compaction process, in the gut of hydra, of smaller particles (0·45-μm SE) into much larger particles (6·25-μm SEB) aod evideoce that flagella probably play a prominent role in compaction. Investigations of these and other intriguing aspects of cellular behaviour by nutritive phagocytes are currently underway.

Thanks are due to Dr L. Muscatine for his advice, support and encouragement; to Dr J. Berliner for her generous allowance of advice and of time on the scanning electron microscope ; to Ms J. Baumer, Ms S. Beydler, and Mr H. Kabe for their excellent technical help; and to Mr R. Marian, Mr T. Hohman and Dr F. S. Sjostrand for their critical readings of this manuscript. The research was supported, in part, by grants from the National Science Foundation (PCM-75-03380 and PCM-7827380) and a UCLA Biomedical Research Support Grant (USPHS 5-507 RR 07009-14), all to Dr L. Muscatine. This work is submitted in partial fulfilment of the Ph.D. in biology at the University of California, Los Angeles.

^† The term ‘microvilli’ is used here to refer to all roughly tubular, 0·25-5μm long projections of the digestive-cell surface. While it is true that these projections are highly pleomorphic, ‘microvilli’ remains as the term of general usage that most simply and accurately reflects the observed morphology.