| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Cell and Animal Biology
Confocal Microscopy Unit, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem, Israel
1Correspondence: Department of Cell and Animal Biology, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel. E-mail: dsher{at}pob.huji.ac.il or zlotkin{at}vms.huji.ac.il
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: chemical ecology • toxin • digestion • cnidaria
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Many PFPs are considered to be toxins; they are common in pathogenic bacteria and are found extensively in animal venoms. Similar proteins are produced by the immune system to lyse potential pathogens (e.g., perforin, complement c9; refs. 2
, 3
). In several specific cases, PFPs have other physiological roles not directly involved in cell lysis, such as formation of intracellular chloride channels (4)
and induction of apoptosis (5)
.
Sessile cnidarians such as sea anemones and hydra produce many different PFPs (6
7
8
9
10
11
12)
. Many of these PFPs are toxic on injection to test organisms, and several have been shown to originate from the cnidarian stinging cells, the nematocytes (7
8
9
, 13)
. Indeed, PFPs have been suggested to be involved in the life-threatening stings of the Portuguese Man-O-War (Physalia physalis) and of box jellyfish (8
, 9
, 14)
.
Recently, we isolated and characterized a novel family of cytolytic and paralytic toxins, Hydralysins (Hlns), from the green hydra, Chlorohydra viridissima (12)
. Hlns differ from all previously studied cnidarian cytolytic toxins, revealing structural and functional similarities with a diverse group of PFPs from bacteria, plants, and fungi (10)
. Large amounts of Hlns are produced by Chlorohydra, with protein levels corresponding to 
0.3% of the total dry weight of the animal (12)
. Surprisingly, Hlns are not found in the nematocytes, suggesting that, unlike other cnidarian PFP toxins, they are not involved in prey capture (12)
.
The uniqueness of Hlns, particularly their non-nematocystic occurrence, motivated us to clarify their biological role. Toxicity in cnidarians that does not originate from nematocysts has been observed in various species of box jellyfish (8
, 15
16
17)
and a sea anemone (18)
. Thus, insights gained from the study of Hlns may be relevant to cnidarian organisms in general. Here we show that Hlns are used by Chlorohydra for preliminary disintegration of their prey within their open gastrovascular cavity (GVC) (19
, 20)
, thus allowing a soft-bodied organism to feed on hard, cuticle-covered prey. This reveals a novel biological role for PFPs, highlights the pharmacological flexibility and ecological versatility of such proteins, and suggests that novel biological insights can be gained by studying PFPs and other bioactive compounds within their natural biological context.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Maintenance of Chlorohydra
Chlorohydra viridissima were originally obtained from the Volcani Center for Agricultural Research, Bet-Dagan, Israel. They were maintained in glass dishes in M-medium (21)
at 20°C with a 12:12 light:dark cycle. The Chlrohydra were fed three times a week with freshly hatched nauplii of Artemia salina. Unless otherwise mentioned, the animals were starved for 3 days prior to experimentation.
In situ hybridization
In situ hybridization was performed using a modification of the protocol reported by Martinez et al. (22)
, available online at http://www.ucihs.uci.edu/biochem/steele/insituprotocol.doc. A 735 bp DIG-labeled probe corresponding to the full coding region of Hln-2 was used, and identical results were obtained with a smaller, gel-purified 424 bp probe (not shown). NBT-BCIP (Roche Applied Sciences, Nutley, NJ, USA) was used as the color substrate. A detailed description of the protocol is available as Supplemental Methods online.
For sectioning of Chlorohydra after whole-mount in situ hybridization, the animals were washed extensively in PBST, infiltrated with 30% sucrose in PBST, immersed in Tissue-Tek medium, and frozen on dry ice or liquid nitrogen. Sections 7 µm thick were taken using a Leica CM1850 cryostat (Wetzlar, Germany) and mounted in glycerol without dehydration and tissue clearing, as this can cause loss of the color precipitate. Micrographs were taken with a Zeiss Axioskop-2 microscope (Carl Zeiss, Jena, Germany) equipped with an Olympus DP10 CCD camera (Tokyo, Japan).
Immunohistochemistry
Recombinant Hln-2 was expressed and purified as described in ref. 10
, coupled to TiterMax Gold Adjuvant, and used to raise antisera in rats (Anilab Biological Services, Tal Shachar, Israel). The antiserum was tested for specificity as shown in Supplemental Fig. 1. A detailed protocol for the immunohistochemistry and confocal microscopy is provided as a supplemental protocol online.
Subcellular fractionation of Chlorohydra and Artemia
Subcellular fractionation was performed according to Gordon et al. (23)
, with an additional step of ultracentrifugation for 2 h at 100,000 g to isolate the "microsomal" fraction, which contains mainly small Golgi- and ER-derived vesicles (i.e., secreted proteins; ref. 24
). A detailed protocol is provided as a supplemental protocol online.
For experiments to test the binding of Hln to these fractions, all samples were washed once more in Mannitol buffer, then washed three times after incubation in ice-cold PBS to remove unbound Hln before detection of bound Hln using Western blot.
Measuring the effect of Hln on Artemia
Three-day-old nauplii of Artemia salina were washed extensively in either PBS (for the experiments depicted in Fig. 4A
) or M-medium (see Fig. 4C-F
), transferred to a 96-well microtiter dish containing recombinant Hln-2 at the appropriate concentration, and left at 18°C for 2 or 24 h (see Fig. 4A
) or for 22 h (see Fig. 4C-F
). To count live Artemia, they were transferred to a 1.5 ml Eppendorf tube and left for several minutes. Under these conditions, the dead or immobile Artemia settle to the bottom of the tube whereas the live and mobile ones continue swimming. The live Artemia were aspirated using a Pasteur pipette and the dead ones were left in the same tube. When no live Artemia remained in the test tube, both samples were fixed in 4% paraformaldehyde overnight, mounted on microscope slides, and counted or photographed. To detect pore formation, an equal volume of 0.4% trypan blue solution was added to the Artemia 10 min before transferring them to the Eppendorf tubes as described above.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). Low levels of expression were occasionally seen in the tentacles. In budding Chlorohydra, Hln expression was concomitant with the development of the buds’ independent GVC: during the initial stages of budding, expression of Hlns could be seen in the bud as a continuation of the expression in the gastral region of the adult, with regions of expression becoming separate as the bud matured (Fig. 1B, C
). No background was observed in control experiments using sense probe (Fig. 1D
).
|
To detect which cell type expresses Hlns, we sectioned Chlorohydra following whole-mount in situ hybridization. These sections revealed that Hlns are expressed exclusively in the endodermal layer surrounding the GVC (Fig. 1E, F
). In the center of the gastral region, Hln expression was limited to a clearly defined band at the base of the endoderm, while toward the hypostome and base of the animal the expression widened to encompass the entire endodermal layer (arrows in Fig. 1E, F
). The expression of Hlns at the base of the endoderm suggests that, in the midgastral region, the digestive cells are those expressing the protein, since the other major cell type (gland cells) are found mainly in the apical part of the endoderm (compare with Fig. 2C
in ref. 25
; see also refs. 26
, 27
).
|
Cnidarians are famous for their tight symbiosis with photosynthetic algae, which, in Chlorohydra, are found mainly in the basal part of the endodermal digestive cells (28
, 29)
. Since the expression of Hlns coincided with this region, we wanted to ascertain whether Hlns are expressed by Chlorohydra itself or by the symbiotic algae. To differentiate between these possibilities, we bleached Chlorohydra, eliminating their symbiotic algae by treatment with the photosynthetic inhibitor 3-(3,4-dichlorophenyl)-1,1-dimethyl urea (30)
. Hln expression in Chlorohydra maintained in an aposymbiotic state for several weeks was the same as in normal, symbiotic animals (Fig. 1G
). Thus, Hlns are expressed by Chlorohydra and not by their symbiotic algae, and do not seem to be specifically involved in the symbiotic relationship.
Hln proteins are found in both the apical and basal cytosol of endodermal digestive cells
Detection of Hln transcripts specifically at the base of the endodermal digestive cells prompted us to determine whether the mature protein is found at the same place, which would suggest that this subcellular region is where Hlns fulfill their biological role. To this end, we performed immunohistochemistry on cryostat sections of Chlorohydra, using antisera raised in rats against Hln-2 (Supplemental Fig. 1). Hln proteins can be detected in the endodermal cells surrounding the GVC, with little or no protein observed in the mouth region and the tentacles (Fig. 2
: grayscale images for the color-blind in Supplemental Fig. 2
). In contrast to the in situ hybridization, in which Hln transcription was weaker in the base of Chlorohydra than in the gastral region (Fig. 1A
), no such difference in the mature protein was observed using immunohistochemistry (Fig. 2A
and data not shown). This may be due to migration of endodermal cells following expression and/or accumulation of Hln from the gastral region toward the base of the animal (31)
.
Within the endodermal digestive cells, Hlns were detected not only at the basal part of the cells, but also in the apical part facing the GVC (Fig. 2B, C
). The distribution of Hlns within the cells was not homogeneous, but was localized in specific cellular regions or large organelles. Immuno-electron microscopy revealed that, within the digestive cells, Hlns are found in irregular, condensed structures of 5–7 µm, which usually are not surrounded by membranes (Supplemental Fig. 3). At the apical part of the cells, these structures resemble digestive vacuoles, whereas at the basal part the Hln-containing structures are morphologically simpler. No Hlns could be detected in gland cells. Localization of Hlns to a cytosolic compartment would be expected, since the primary sequence of the proteins does not contain any signal sequence suggestive of secretion through the ER-Golgi pathway (12)
. In support of this conclusion, we separated Chlorohydra into subcellular fractions and used Western blot to assess in which cellular compartment (or compartments) Hlns are found. As can be seen in Fig. 2D
, the vast majority of Hln proteins was found in the cytosolic fraction. Under the conditions of the experiment, Hln was not found on the plasma membrane; thus, this PFP is not an integral constituent of Chlorohydra membranes.
Hlns are secreted into the GVC and onto the prey during feeding
The observation that Hln transcription takes place mainly at the base of the endodermal digestive cells, whereas the protein product is also found in the apical part of these cells, suggested to us the possibility that this translocation may be a first step prior to secretion of the protein into the GVC. In addition, some of the apical bodies containing Hlns were similar to digestive vacuoles, possibly containing remnants of prey. We thus followed the dynamics of Hlns during feeding of the Chlorohydra with nauplii of Artemia salina, their common laboratory crustacean prey. These results are presented in Fig. 3
and as grayscale images in Supplemental Figs. 4 and 5.
|
Within 15 min of prey capture by Chlorohydra, we observed massive translocation of Hlns toward the apical part of the endoderm, followed by their release into the GVC (Fig. 3C-E
). One hour after prey capture, Hln proteins could no longer be observed in the GVC itself whereas large amounts could be seen within the prey tissue (Fig. 3F-H
). The prey was processed within the GVC for 8 h, after which the undigested remnants were ejected through the mouth; Hlns could be detected on the ingested prey throughout this period (not shown).
Three specific points are worthy of additional consideration. First, Hlns could be detected as bright, localized fluorescent points within the GVC 15 min after feeding (Fig. 3D, E
). A closer examination of the tissue morphology of the same micrographs revealed that in many cases Hlns found inside the GVC were enclosed in large membrane-bound structures resembling cellular debris (Supplemental Fig. 6A). This cellular debris was absent from the GVC in unfed animals or at other times after feeding; it was not seen in other parts of the animal and was reproduced in at least four independent experiments using different fixation and embedding methods. This suggests the Hln-containing cellular debris to be a bona fide phenomenon, and not an artifact caused by poor preservation of the tissue. Ultrastructural analysis of the endoderm of Chlorohydra using transmission electron microscopy (TEM) revealed the presence of large heterogeneous cellular debris within the GVC 15 min after feeding (Supplemental Fig. 6). It is noteworthy that some of these membrane-bound cellular debris contained symbiotic algae, revealing that they originate from Chlorohydra digestive cells and not from the disintegrating prey tissue (not shown). These results are consistent with apocrine secretion of Hln from the endodermal digestive cells into the GVC (32)
.
Second, 1 h after feeding, Hlns could be detected not only on the surface of the prey within the GVC, but throughout the tissue of the prey. This is especially evident in Fig. 3H
, which shows a cross section of an Artemia inside Chlorohydra: here, Hlns can be seen not only on the apical side of the epithelium lining the gut of the Artemia, but also between these cells as well as on cells in other parts of the prey animal. Thus, Hlns are able to penetrate the external cuticle of the crustacean prey and enter into the tissue. In addition, in Fig. 3G, H
, Hlns seem to be localized on the membranes of Artemia cells.
Finally, it is noteworthy that many studies have suggested a role for gland cells, and not digestive cells, as the main cell secreting digestive enzymes into the GVC during feeding. However, given the clear difference between the expression pattern of Hlns and that of a gene expressed by gland cells (25)
, and the lack of Hln detected in TEMs of gland cells, the most probable scenario is that Hln is expressed and secreted by digestive cells, then retaken by these cells from the GVC as they engulf and phagocytose the prey debris. This is in agreement with other studies revealing that a putative digestive enzyme, an endoglycoceramidase, is expressed by digestive cells and secreted into the GVC during feeding (33)
Taken together, these results show that Hlns are secreted into the GVC and onto the prey cell membranes during the initial stages of prey digestion.
Hln-2 binds, lyses, and disrupts crustacean prey of Chlorohydra
A paralytic PFP that is secreted onto the prey during feeding could be involved in the process of digestion in one of two possible ways. First, the venom of predators is aimed to immobilize the prey and not necessarily kill it immediately (11
, 34)
; paralytic-neurotoxic proteins such as Hlns (10
, 12)
may be secreted into the GVC to keep prey paralyzed and subdued until it is digested. Alternatively, Hlns may actively participate in digestion or disintegration of prey either through their pore-forming, cytolytic capability or through a so far undiscovered enzymatic activity.
To differentiate between the two possibilities, we tested what effect Hlns would have when applied externally to Artemia salina. As shown in Fig. 4
A, recombinant Hln-2 did reveal a time- and dose-dependent toxic effect on Artemia. However, this effect was only evident after several hours, suggesting Hlns do not induce fast neurotoxic paralysis of Artemia and therefore cannot be used to rapidly subdue an active organism. In addition, we observed that the paralytic effect caused when an Artemia is caught by the tentacles of Chlorohydra is irreversible. Of 90 Artemia caught by Chlorohydra tentacles and separated from these tentacles before engulfment, only one was seen moving several hours later. Thus, it seems that Hlns cannot be used to maintain paralysis of the prey, nor are they needed by Chlorohydra for this role since, at least for prey in the general size category of Artemia nauplii, the nematocyst venom causes irreversible paralysis and death (35)
.
Although Hln-2 did not rapidly paralyze Artemia nauplii when applied to the water in which they live, it clearly caused the disintegration of prey animals into tissue fragments on extended periods of incubation (22–24 h, Fig. 4C-F
). This effect was probably mediated by the pore-forming effect of Hln, as seen by uptake of the nuclear dye trypan blue. (Fig. 4C
; ref. 12
). When recombinant Hln was incubated with isolated Artemia membrane fractions, the majority of the bound protein was observed as a high molecular weight species (Fig. 4B
), possibly corresponding to oligomers, which form the functional pore of PFPs (10)
. We used 1 µM of Hln-2 for these experiments, since this is in the range of concentrations expected to be found within the GVC during feeding based on the volume of a fed Chlorohydra (0.67±0.13 µl), the volume of the GVC (50% of the total volume of the Chlorohydra 1 h after feeding as estimated from the cryostat sections used for immunohistochemistry), the amount of Hln produced by each animal (50 ng; ref. 12
), and the reduction of the Hln-derived body toxicity after feeding (50%, not shown). This was also in accordance with the hemolytic activity detected in isolated fluid from the GVC (corresponding to 0.4 µM Hln). Thus, prolonged (22 h) in vitro conditions are able to mimic a process that in vivo takes 8 h. It may be assumed that, within the GVC, additional means are used to intensify the tissue-disrupting effect of Hlns (see Discussion).
Taken together, these results suggest that Hlns probably do not have a paralytic/neurotoxic role in Chlorohydra. Rather, they are able to partially disintegrate the tissue of the prey at biologically relevant concentrations found within the GVC. The possibility that the tissue-disrupting effect of Hlns is due to some enzymatic activity of the protein is unlikely, since it has no detectable protease, acid or basic phosphatase, or DNase activity (not shown). Thus, the tissue-disrupting effect is likely the result of osmotically driven lysis due to the formation of pores in prey membranes.
Hydralysins cannot bind Chlorohydra membranes
One problem faced by all organisms is how to protect themselves from the activity of their own digestive enzymes. Unlike many digestive enzymes, which by definition are nonselective, PFPs like Hlns are often highly specific, recognizing their target membranes through a specific receptor (ref. 10
and references therein). Hlns have been shown to be cytosolic in Chlorohydra cells under normal conditions (Fig. 2D
). However, this does not necessarily mean that Hlns are unable to bind Chlorohydra membranes when secreted from the cells into the GVC. To test whether Hlns are able to interact with Chlorohydra membranes, we incubated various subcellular fractions from Chlorohydra with exogenously added recombinant Hln-2 and used Western blot to visualize bound protein. Hln could not bind Chlorohydra membranes, but could be clearly seen bound to equal amounts of Artemia membranes (Fig. 4B
). Thus, Chlorohydra is protected from the activity of its own pore-forming, cytolytic protein, probably due to the lack of a specific receptor recognized by Hlns.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 36)
. Having caught their prey, however, Chlorohydra face further difficulties in digestion. First, many of the organisms that serve as Chlorohydra’s prey, such as small crustaceans, annelids, and juvenile fish (37
, 38)
, are covered with a hard exterior: a cuticle or scales. Most predators use mechanical means for the preliminary disintegration or mastication of such prey, such as teeth, gizzards, or other hard components of their digestive tract. The soft-bodied Chlorohydra, which is composed of only two epithelial cell layers, does not possess such morphological components and can only utilize contractions of its soft body to help shear the prey (39)
. Neither does the nematocyst venom serve for preliminary digestion of the prey (refs. 11
, 35
and our unpublished observations), as is the case with viperid snakes (40)
. In addition, in Chlorohydra and related hydroids, the GVC is involved not only in prey digestion, but also serves as a circulatory system, takes part in gas exchange and excretion, and forms the hydrostatic skeleton of the animal (19
, 20)
. These multiple roles necessitate periodical exchange of water with the environment (19
, 20)
, limiting the ability of the animal to control the content of the gut and making it in essence a direct branch of the outside environment. Therefore, green hydra, as well as other hydra species, do not have the morphological traits needed for mechanical dissociation of the prey, and need to use gastrodermally derived chemistry to perform this role. This chemistry must be able to work in the open environment of the GVC, which may not be appropriate for efficient use of chemical-enzymatic digestion.
One manner in which Chlorohydra and other soft-bodied cnidarians digest hard prey is by performing most of the digestion of the food within the endodermal digestive cells. After ingestion by Chlorohydra and other hydra species, the prey is disintegrated into small particles, which are then circulated throughout the GVC by the beating of the endodermal cilia and are actively phagocytosed by the endodermal digestive cells (20
, 29
, 38
, 41
, 42)
. This is in contrast to most other animals, where food is extracellularly digested into small molecules such as amino acids and simple sugars within the gut lumen; only these small molecules are absorbed by cells lining the intestine (20
, 43)
. Thus, in hydra species, food digestion has been transferred from the large open GVC into small, controllable compartments within the cells. However, the problem remains of how to perform the preliminary stage of extracellular prey disintegration into particles small enough to be phaocytosed.
We suggest that to induce rapid disintegration of the prey within the GVC, the green hydra utilizes the pore-forming capability of Hlns to take advantage of the fact that it lives in a freshwater environment—a hypo-osmotic environment compared to that found within the tissue of most organisms (19)
. Hlns are expressed in large amounts in the digestive cells surrounding the GVC and are secreted onto the ingested prey during the initial stages of digestion. At biologically relevant concentrations, this causes the membrane integrity of some prey cells to be compromised, leading to lysis of these cells. This lysis fulfills two main roles: first, it releases the content of the lysed cells into the GVC, either in the form of water-soluble macromolecules or as small cellular debris, which can then be ingested by the endodermal digestive cells through pinocytosis or phagocytosis (20
, 29
, 38
, 41
, 42)
. Second, Hlns are not found alone within the GVC, but are accompanied by other digestive enzymes (33
, 38)
; the lysis of part of the prey cells loosens the connections between the various cells of the prey tissue, thus permeating the tissue to both other Hln molecules and digestive enzymes. This enhances and accelerates the process beyond the rate that can be demonstrated with Hlns alone. While it is probable that the disrupting effect is enhanced by the hypo-osmotic (freshwater) environment of Chlorohydra, such a mechanism could conceivably also work in marine cnidarians, albeit perhaps less efficiently. In support of this, the occurrence of a hemolytic protein, coelenterolysin, has been reported in the GVC of a sea anemone (44)
.
Our results show for the first time that a multicellular predatory organism can utilize a pore-forming protein as part of an osmotically driven mechanism for food digestion. Such a mechanism has been suggested only for single-cell organisms (45
, 46)
or for parasitic organisms feeding directly on cells in suspension (47
, 48)
. Many additional roles have been proposed for putative PFPs in cnidarians, including immunity (49)
and regulation of development (50)
. While our results demonstrate the role of Hlns in digestion by Chlorohydra, it is possible that this family of proteins may also fulfill other roles, such as inhibition of bacterial growth within the GVC. The ability of PFPs to permeate membranes in a time- and location-specific manner suggests that these pharmacologically flexible proteins can fulfill numerous biological roles, highlighting the utility of studying these proteins within their natural biological context.
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 28, 2007. Accepted for publication July 5, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Schlesinger, E. Zlotkin, E. Kramarsky-Winter, and Y Loya Cnidarian internal stinging mechanism Proc R Soc B, March 22, 2009; 276(1659): 1063 - 1067. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
