SUMMARY
Barnacles, like many marine invertebrates, cause serious biofouling to marine industrial constructions and hulls of vessels as they attach themselves to such surfaces. Precise biochemical understanding of the underwater adhesion to surfaces requires a detailed characterization of the biology of the control of barnacle cement secretion and the proteins that make up the cement. In this study, we have investigated cement secretion by cyprid larvae of Balanus improvisus (D.) and the morphology of their cement glands. We studied the cement protein organization within cement granules and categorized the granules into four different types according to their size and morphology,before and after stimulation of secretion. In addition, we followed the exocytotic process of cement secretion in vivo and discovered that granules undergo a dramatic swelling during secretion. Such swelling might be due to an increased osmotic activity of granule contents, following a process of hydration. We hypothesize that this hydration is essential for exocytotic secretion and conclude that cement protein exocytosis is a more complex process than previously thought and is similar to exocytotic secretion in vertebrate systems, such as histamine secretion from mast cells and exocrine secretion in the salivary gland and the pancreas.
Introduction
Sessile marine invertebrates, such as barnacles, mussels and tube worms,that populate wave-swept shallow waters attach themselves to almost any kind of surface to colonize. These type of animals use various unique strategies to`glue' themselves to hard surfaces found in these waters, including man-made marine installations and ships' hulls. An estimated 2500–3000 species of marine organisms are known to cause significant biofouling worldwide. This type of marine biofouling is a tremendous economic burden on marine industries such as shipping and oil-drilling structures. This is especially true of the many species of barnacles found in the oceans globally, and thus they are exceptional models to investigate the nature of the adhesive and the biology of the secretory process that support settling of these animals.
The life cycle of Balanus improvisus has seven planktonic larval stages before it metamorphoses into a sessile organism. Between the sixth and seventh transitions, the larva transforms from a nauplii larva into a cyprid larva. The cyprid larva searches for a suitable surface to attach itself to and to metamorphose into a sessile reproducing animal. The prerequisite for settling and metamorphosis is the ability to produce, store and secrete the adhesive proteins once the cyprid identifies an appropriate surface to adhere. The adhesive-secreting cells are located within a pair of cement glands, which are connected by cement ducts that widen into muscular cement sacs, the presumed temporary storage location during cement secretion. Cement ducts connect the sacs to an antenna, which is composed of four segments. The cement duct extends into the third segment, the adhesive discs. The adhesive is secreted through the discs, and the cyprid larva is able to attach itself to the surface and begin metamorphosis(Harrison and Sandeman, 1999; Nott and Foster, 1969).
Our approach is to understand the biology of barnacle cement secretion in detail so that new techniques could be developed to control their settling on to man-made marine surfaces. Current methods of control mainly use biocide doped paints on surfaces, and such biocides leach in significant quantities to cause serious toxicity to the marine environment. New approaches are necessary to devise more environmentally benign modes of control of barnacle-induced biofouling (de Nys and Steinberg,2002; Omae, 2003). Understanding the neural control of cement secretion is expected to provide the lead to develop novel substances that inhibit cement secretion. In the present study, we have focused on the morphology of the cyprid cement glands and the physiology of adhesive secretion.
Materials and methods
Rearing of barnacle cyprid larvae
Cyprid larvae of Balanus improvisus (Darwin 1854) were reared in a laboratory culture as described by(Berntsson et al., 2000). In short, adult barnacles were allowed to settle on Plexiglas panels placed in the sea off the west coast of Sweden in the vicinity of TjärnöBiological Laboratory (58°53′ N, 11°8′ E). The panels were brought to the laboratory and placed in buckets with running seawater. These animals were used as brood stock in all laboratory experiments. Adult B. improvisus will spawn throughout the year when regularly fed with nauplii of Artemia sp. and the prymnesiophyte Isocrysis galbana. When kept at 27–28°C, the development to cyprid larvae takes 6–7 days, and the newly moulted cyprid larvae were filtered through a stack of sieves and washed to remove algae and detritus. The cyprids were then transported in 25 cm3 flasks to the laboratory, stored at 4°C and usually used within a week.
All experiments were performed in filtered seawater (FSW, 0.22 μm), and drugs used were purchased from Sigma. The drugs were diluted in FSW to desired concentrations.
Preparation for light microscopy and transmission electron microscopy(TEM)
The cyprids were placed in FSW containing no dopamine, or 1 mmol l–1 dopamine for a time series of 2, 4 and 10 min treatment. The cyprids were thereafter immediately transferred over to the fixative.
The cyprids were fixed according to Harrison and Sandeman(Harrison and Sandeman 1999). Cyprids were chilled to 4°C for 30 min and transferred to fixation solution containing 2.5% glutaraldehyde and 2% formalin in FSW, pH 8.2. The cyprids were then microwaved in an ice bath up to 37°C and thereafter placed in fixative at 4°C overnight. Several washings were then performed in FSW over 1 h, followed by treatment with 2% osmium tetroxide for 30 min and then 2% uranyl acetate for 20 min. Dehydration was thereafter performed in an ethanol series of 50%, 70% and 90%. Finally, propylenoxide and infiltration of Agar 100 resin (Epon) was performed overnight in 4°C. Samples were embedded in gelatine capsules in Agar 100 and polymerized at 60°C for 2 days.
For light microscopy, the blocks were sectioned into 2 μm-thick sections and stained with 1% Toluidine Blue and Pyronin G for 1 min at 60°C. The sections were then mounted in Pertex and examined in a Nikon Optiphot microscope connected to a Nikon DXM 1200 digital camera. Images were acquired using the ACT-1 software (Nikon Microscopes, Europe BV, Badhoevedorp, The Netherlands). Contrast levels were then adjusted in Adobe Photoshop CS (Adobe Systems, Inc., Mountain View, CA, USA).
Confocal microscopy
Cyprid larvae are transparent animals and thus are easily observed under a confocal microscope to study the secretory process in situ. We found that Acridine Orange accumulates within secretory vesicles in the cement glands of cyprids and could be used as a flourophore to study the exocytotic secretion. This is similar to secretory granules of pancreatic β-cells due to the acidic intragranular environment with respect to the cytosol(Pace and Sachs, 1982). The excitation wavelength was 490 nm and fluorescence emission was collected at 519 nm. We incubated cyprid larvae in FSW containing 10 mmol l–1 of Acridine Orange for 1 h and washed them several times in FSW prior to microscopy. The cyprids seemed unaffected by the Acridine Orange loading.
The larvae were mounted on a cover glass using Kwik Sil silicon glue (World Precision Instruments, Herts, UK) at the caudal end, while the anterior end remained free. They were kept in FSW in darkness for 1 h. A perfusion chamber was built up by using Tack It™ (Faber-Castell, Germany) on a glass slide, and the cover glass with the cyprid was inverted and attached to Tack It, thereby creating a perfusion chamber. Larvae were imaged in a Bio-Rad MRC 1024 laser scanning confocal microscope equipped with a Krypton/Argon laser. Images were acquired using Lasersharp 2000 (Bio-Rad, Hercules, CA, USA) at a magnification of 800× and were analyzed using Adobe Photoshop CS.
Differential interference contrast (DIC) microscopy
Cyprid larvae were immobilized in 1% low gelling temperature agarose, type VII (Sigma A-9045, gelling temperature below 25°C) dissolved in FSW. Agar was poured over the larvae and they were placed inside a refrigerator for 5 min for polymerization. Animals were then examined under an inverted microscope and imaged. Images were acquired using a CCD camera (Pulnix Corp.,San Jose, CA, USA), and the microscope and camera systems were controlled by Synapse (Synergy Research Inc., Silver Spring, MD, USA) as described earlier(Simpson et al., 1997). Cyprids immobilized in agarose appeared unharmed and tried to move within the gel during microscopic observation.
Estimation of secretory granule number and size
In order to measure the changes in secretory granule type and number during the secretory event, we stimulated cement secretion in intact larvae using dopamine (1 mmol l–1). Larvae were exposed to dopamine for 0,2, 4 or 10 min and were immediately fixed as described earlier. Sections were cut and were stained with Toluidine Blue for light microscopic observation and granule counting. Statistical comparison was carried out using analysis of variance (ANOVA, P<0.0001). Newman–Keuls multiple comparison test was used as a post-hoc test with an α-value of 0.05.
For secretory vesicle size measurements, we used Easy Image Measurement 2000 (Bergström Instrument AB, Solna, Sweden), and the cross-sectional area of the granules in six different sections showing all four types of granules was measured. All granules of type 2, 3 and 4 within a section were measured. Due to the presence of large numbers of type 1 granules, the cross-sectional area of all granules within 2–3 cells was measured. The result was evaluated by Kruskall–Wallis test (P<0.0001) and Dunn's multiple comparison as post-hoc test with an α-value of 0.05. ANOVA evaluations were done using Graph Pad Software (GraphPad Software,Inc., San Diego, CA, USA).
The cement gland, as it appears in a living cyprid under differential interference contrast (DIC) optics. The individual granules can be seen as bumps on the cement gland surface. At the apical end, the cement duct, which widens to form the muscular sac, can be seen. CG, cement gland; CD, cement duct; MS, muscular sac; CE, compound eye.
The cement gland, as it appears in a living cyprid under differential interference contrast (DIC) optics. The individual granules can be seen as bumps on the cement gland surface. At the apical end, the cement duct, which widens to form the muscular sac, can be seen. CG, cement gland; CD, cement duct; MS, muscular sac; CE, compound eye.
Results
The cyprid larva
The cyprid larva is the final planktonic larval stage in the metamorphosis of barnacles. In the case of B. improvisus, they are approximately 500 μm in length (Berntsson et al.,2000) and live for up to a month. The cement gland is easy to recognize under a microscope, and Fig. 1 shows the gland and the cement duct, which widens into a muscular sac. The muscular sac is connected by another cement duct to the antennae. The individual cement granules within the gland are large enough to be seen as individual bumps on the gland surface.
Characterization of cement granules
Light microscopic observation of Toluidine Blue-stained sections of the gland revealed that control untreated glands contained columnar cells filled with secretory vesicles stained blue. When stimulated with dopamine (1 mmol l–1), however, four different types of secretory granules can be recognized within the glands on the basis of morphological appearance. Fig. 2 shows a typical micrograph depicting all the four types of secretory granules following stimulation of a cyprid for 10 min with dopamine. The most abundant are the dense, dark granules stained blue (type 1). The granules that appear light blue represent type 2 vesicles. Granules that appear to be `moth eaten' are type 3 granules. The fourth type of granules are partially or completely empty and appear as vacuoles. Fig. 3summarizes the differences between the size of the different granule types. Overall, type 1 granules were the smallest, and dopamine stimulation seemed to induce granule swelling followed by varying degrees of emptying of granule contents, finally resulting in completely empty vesicles that appear as vacuoles. Control, unstimulated glands contained mostly type 1 granules, and only occasionally were type 2 and 3 vesicles found in some cells. This observation supports our view that the loss of granule contents and vacuole formation were not experimental artefacts caused by fixatives.
Appearance of the different types of secretory granules after stimulation of cyprids with dopamine (1 mmol l–1) for 10 min. Cyprids were aldehyde fixed, sectioned and stained with Toluidine Blue. The different types of granules are labelled (1–4). The cement duct, with dissolved proteins, is seen as an extension away from the cement gland towards the muscular sac.
Appearance of the different types of secretory granules after stimulation of cyprids with dopamine (1 mmol l–1) for 10 min. Cyprids were aldehyde fixed, sectioned and stained with Toluidine Blue. The different types of granules are labelled (1–4). The cement duct, with dissolved proteins, is seen as an extension away from the cement gland towards the muscular sac.
The different types of granules differ in their overall size. The surface area of the different types of granules was measured by Easy Image Measurements 2000. Overall, 195 of type 1, 113 of type 2, 63 of type 3 and 51 of type 4 were measured, and the sizes were compared using statistical tests. The different types of granules were vastly different in their size except that type 2 granules were not significantly different from type 3 granules.
The different types of granules differ in their overall size. The surface area of the different types of granules was measured by Easy Image Measurements 2000. Overall, 195 of type 1, 113 of type 2, 63 of type 3 and 51 of type 4 were measured, and the sizes were compared using statistical tests. The different types of granules were vastly different in their size except that type 2 granules were not significantly different from type 3 granules.
Electron microscopy of cement glands in (A) control, unstimulated cyprids and (B) cyprids stimulated with 1 mmol l–1 dopamine for 10 min. The different types of granules are labelled in B and can be compared with Fig. 2. In control,unstimulated animals, most of the granules are type 1, but all the four types of granules are visible in stimulated cement glands (B). Note also that the different types of granules are within the same cell in B. Scale bar: 3.8μm in A; 2.5 μm in B.
Electron microscopy of cement glands in (A) control, unstimulated cyprids and (B) cyprids stimulated with 1 mmol l–1 dopamine for 10 min. The different types of granules are labelled in B and can be compared with Fig. 2. In control,unstimulated animals, most of the granules are type 1, but all the four types of granules are visible in stimulated cement glands (B). Note also that the different types of granules are within the same cell in B. Scale bar: 3.8μm in A; 2.5 μm in B.
Secretory granule ultrastructure was examined using TEM in controls(Fig. 4A) and in animals stimulated with dopamine for 10 min (Fig. 4B). Cement glands of unstimulated control cyprids had mostly type 1 granules, which were small and their interior appeared dense with a variegated crystal-like structure. Type 1 granules in the stimulated glands were identical in structure to those in unstimulated glands. Stimulated glands contained granules whose contents showed varying morphologies categorized into the four different types (see Fig. 2 for comparison). Fig. 5A shows dense type 1 granules with an intragranular pattern of an organized electron-dense core. This appearance might indicate that the cement proteins are stored in very high concentrations within the vesicles. Most endocrine and neurosecretory vesicles in mammalian cell systems are known to contain secretory material in concentrations as high as 0.2 mol l–1 (Dreifuss,1975). Fig. 5Bshows type 2 granules that appear swollen, filled with smooth, amorphous contents that differ in appearance from the organized regular pattern seen in dense-core type 1 granules. Type 3 granules appeared partially empty,containing empty holes that gave a moth-eaten appearance(Fig. 5C) reflecting their appearance under the light microscope. Closer examination shows that the empty holes contain disorganized proteinaceous material with a filamentous structure interspersed within the smooth amorphous material similar to type 2 granules. Finally, we classified the empty vacuole-like granules as post-exocytotic type 4 granules devoid of contents. Fig. 5D shows an example of a vacuole, and in this example it is seen together with type 1 granules. Thus, different types of granules occasionally occur within the same cell. Type 4 granules (vacuoles) are membrane-bound areas where the cement proteins have been almost completely lost, and only some remnants are retained. These structures appear similar to degranulation sacs seen in histamine-secreting mast cells following maximal stimulation(Cho et al., 2002b; Crivellato et al., 2002a). Overall, these morphological characteristics of secretory granules in the cement gland of cyprids appear to be similar to secretory vesicles in other secretory systems during exocytosis such as the entero-endocrine cells(Crivellato et al., 2002b), the different cells of the immune system(Crivellato et al., 2003) and the pancreatic acinar cells (Raraty et al., 2000) (for a review, see Jena, 2005).
Visualization of secretion in vivo
Acridine Orange accumulated into cement granules in live cyprids served as a fluorescent marker for following cement secretion-dependent changes in granule morphology and loss of contents. Stimulation of cyprids immobilized in Kwik Sil with dopamine resulted in a gradual loss of fluorescence in vesicles and appearance of large dark vacuoles. Fig. 6 demonstrates this observation(Fig. 6A, control; Fig. 6B, after 15 min in 1 mmol l–1 dopamine). Dopamine at concentrations between 100 μmol l–1 and 1 mmol l–1 induced massive vacuole formation in stimulated cyprids. Noradrenaline (1 mmol l–1)induced vacuole formation about half of the time (50% of trials), similar to previously published observations (Okano et al., 1996). We tested a number of biogenic amines using this technique, namely serotonin, histamine, octopamine, tyramine and melatonin,all at a concentration of 1 mmol l–1 (data not shown). Unlike dopamine and noradrenaline, none of the other amines tested stimulated cement secretion (vacuole formation). In control experiments, we tested spontaneous vacuole formation in the absence of stimulation. Ten animals were incubated in filtered seawater for 2 h and none of them showed any signs of vacuole formation during that time. Thus, vacuoles do not form spontaneously in the absence of stimulation in cement glands of cyprids kept in FSW. The loss of fluorescence and vacuole formation is probably due to a decrease in concentration of Acridine Orange within vesicles as they swell and a loss of dye as secretion of vesicle contents proceeds.
Electron microscopy of secretory vesicle types. (A) Unstimulated cement gland where most of the cement granules appear densely packed with secretory material. Granule contents appear to have a distinct organization. Scale bar,0.6 μm. (B) Type 2 granules appear larger and their contents appear amorphous and lack the organization observed in the dense-core vesicles seen in A. Scale bar, 0.6 μm. (C) Type 3 granules appear similar to type 2,except have a `moth-eaten' appearance with clear spaces or hydration channels due to partial loss of contents. Scale bar, 0.25 μm. (D) Type 4 vesicles appear like vacuoles with a reticulated appearance. Note that the reticulated granules appear within the same cell as the dense-core type 1 granules. Scale bar, 0.4 μm.
Electron microscopy of secretory vesicle types. (A) Unstimulated cement gland where most of the cement granules appear densely packed with secretory material. Granule contents appear to have a distinct organization. Scale bar,0.6 μm. (B) Type 2 granules appear larger and their contents appear amorphous and lack the organization observed in the dense-core vesicles seen in A. Scale bar, 0.6 μm. (C) Type 3 granules appear similar to type 2,except have a `moth-eaten' appearance with clear spaces or hydration channels due to partial loss of contents. Scale bar, 0.25 μm. (D) Type 4 vesicles appear like vacuoles with a reticulated appearance. Note that the reticulated granules appear within the same cell as the dense-core type 1 granules. Scale bar, 0.4 μm.
In another set of experiments, we investigated dopamine-induced cement secretion in intact live cyprids immobilized in low-melting-point agarose using DIC microscopy. Similar to our observation of loss of Acridine Orange fluorescence, we were able to monitor changes in interference contrast followed by vacuole formation within cement glands in all the animals that we examined (Fig. 7). We counted the appearance of the different types of granules at various times during dopamine (1 mmol l–1) stimulation. Fig. 8 shows the change in number of different types of granules during dopamine stimulation over time. Similar to the results from Toluidine Blue-stained sections, dense-core, type 1 granules reduced in number as stimulation proceeded, and the number of vacuoles increased. Thus, a number of different microscopy techniques –light and electron microscopy of fixed and stained sections, confocal microscopy of Acridine Orange-loaded cyprids and DIC microscopy of live cyprids – all show stimulation-dependent changes in granule morphology and granule emptying. While the loss of intragranular material is easily observed, the mechanism through which such loss occurs is not revealed by these experiments. The morphological data, however, suggest that sequential exocytosis might account for the loss of granule contents (see Figs 2, 9). The emergence of vacuoles is seen as early as 2 min after onset of stimulation, and significant loss of granules is observed after 10 min, when approximately half of the dense-core granules have been secreted. Throughout the duration of stimulation, there is no difference in the number of type 2 and type 3 granules. This observation might suggest that the granules undergo swelling upon stimulation and sequentially undergo exocytosis.
In some early experiments, Walker suggested that there are two types of cells within the cement gland (Walker,1971), α and β-cells, based on histological criteria and the morphological appearance of the cement granules. In our hands, within B. improvisus cyprid cement glands, different types of granules could be observed within the same cell (Figs 2, 4B). While the possibility exists that different types of cells might exist within the cement gland, it appears that different types of granules occur within the same cell. We conclude that the difference in appearance of the granules is more likely linked to the secretory process.
Discussion
In this study, we used several microscopy techniques to visualize the secretory process in the cement glands of B. improvisus cyprids. This analysis allowed us to characterize secretory vesicle morphology and dynamics during stimulated secretion. In addition, the data are consistent with an exocytotic secretory process for cement release by the cyprids. Upon stimulation of cement secretion in intact cyprids, secretory vesicles undergo a series of changes, starting with an initial swelling to complete emptying of contents. DIC microscopy allowed direct visualization of exocytotic secretion in vivo.
Confocal microscopy of an Acridine Orange-stained living cyprid cement gland. The larva was immobilized on a cover slip using Kwik Sil and imaged.(A) Unstimulated cement gland within the living cyprid. (B) The cyprid was stimulated with dopamine (1 mmol l–1) and imaged 15 min later. Note the brightly stained secretory vesicles in the control gland (A)and their loss and the appearance of large vacuoles after stimulation.
Confocal microscopy of an Acridine Orange-stained living cyprid cement gland. The larva was immobilized on a cover slip using Kwik Sil and imaged.(A) Unstimulated cement gland within the living cyprid. (B) The cyprid was stimulated with dopamine (1 mmol l–1) and imaged 15 min later. Note the brightly stained secretory vesicles in the control gland (A)and their loss and the appearance of large vacuoles after stimulation.
The observation that secretory granules undergo a series of morphological changes upon stimulation of secretion is significant and may be somewhat similar to two previous reports on barnacle cement secretion(Okano et al., 1996; Walker, 1971). Walker described cement gland morphology in both free-swimming as well as settled Balanus amphitrite cyprids. He discovered that the cement gland in non-settled animals contains a large number of dense granules with a diameter of approximately 3–5 μm, inside columnar cells, and called these cells α-cells. He also described another type of cell, β-cells,with granules that had a reticulated appearance. In addition, Walker observed vacuoles in the apical part of α-cells, where material had been discharged within glands from settled cyprids. We found that in unstimulated cyprids, the cement glands contain small, dense-core secretory granules, which begin to swell within minutes of stimulation with dopamine. Swollen granules gradually lose their contents, possibly due to secretion, to finally become empty vacuole-like organelles. The appearance of vacuoles indeed occurred near the median central groove of the cement glands, similar to the finding by Walker. In addition, like Walker, we occasionally found granules with type 2 or type 3 appearance in control, unstimulated cyprid cement glands. In cement glands from dopamine-stimulated cyprids, we always found the three different types of granules as well as vacuoles within the same cells (Figs 2, 4B, 9B,C).
This finding might question the previous assertion that different types of cells exist in cyprid cement glands containing different types of granules possibly differing in their protein compositions(Walker, 1971). The fact that Walker found vacuoles and reticulated granules in unsettled cyprids might indicate that the cyprid larvae, as well as the recently settled cyprids,might have undergone partial secretion of cement, associated with larval settlement. Vacuole formation suggests secretion has occurred, possibly in an attempt at settlement by the cyprid (see later). Similar vacuole-like structures have been seen in other secretory systems where secretory cells contain dense-core secretory vesicles, e.g. pancreatic acinar cells(Cho et al., 2002b; Raraty et al., 2000),gastrointestinal epithelia (Crivellato et al., 2002b; Kuver et al.,2000) and mast cells(Crivellato et al., 2002a). Similar observations of exocytosis and membrane retrieval in invertebrate neurosecretion were made in the crab sinus gland over two decades ago(May and Golding, 1983; Morris and Nordmann,1980).
The current dogma in the case of cyprid settling is that cement secretion occurs in an all-or-none fashion. The observation of granules with partial loss of contents in normal unstimulated cyprid cement glands might suggest that some secretion can occur in the absence of settling in a form of piecemeal degranulation (Aravanis et al.,2003; Crivellato et al.,2003). Secondly, it is possible that only some of the cement-secreting cells participate in secretion during a settling attempt,sparing other cells in order to provide the cyprid with the possibility of multiple attempts to settle. Finally, the conclusion that the different types of granules found in stimulated cement glands indeed represent granules that have undergone varying degrees of loss of contents calls into question the idea that vesicles with different intragranular composition might exist in order to support cement curing. While it is entirely possible that several different proteins make up the secreted cement, it is not clear if they are derived from multiple types of secretory granules. Not enough is known of the composition of the vesicle contents, nor the cement, to conclude if the barnacle adhesive is a single- or multi-component glue. Dense-core secretory vesicles in general are known to contain multiple protein and other components that form a complex mixture that is held at extremely high concentrations within the vesicles (Lagercrantz,1976; Uvnäs et al.,1970; Winkler,1976). It is quite likely that all the necessary components that make up the barnacle adhesive are stored within the vesicles such that, upon secretion, interaction with seawater or solid surfaces with the right chemical features results in glue hardening. More detailed information of the chemistry of the cement or the granule contents is necessary in order to precisely understand the adhesive curing mechanism.
Visualization of cyprid cement secretion under differential interference contrast (DIC) optics. A cyprid larva was immobilized in agarose and observed in a coverslip chamber using an inverted microscope using DIC optics. The montage shows a series of images separated by 10 s intervals. Images are arranged starting at the top and going left to right. Note the appearance of a vacuole that seems to grow larger with time (arrow). Note the increase in size of the cement sac between frames 1 and 12. See supplemental material for a movie sequence of the original data at http://vivaldi.zool.gu.se/film/Movie-2-sm.movand http://vivaldi.zool.gu.se/film/Movie-2-sm.avi.
Visualization of cyprid cement secretion under differential interference contrast (DIC) optics. A cyprid larva was immobilized in agarose and observed in a coverslip chamber using an inverted microscope using DIC optics. The montage shows a series of images separated by 10 s intervals. Images are arranged starting at the top and going left to right. Note the appearance of a vacuole that seems to grow larger with time (arrow). Note the increase in size of the cement sac between frames 1 and 12. See supplemental material for a movie sequence of the original data at http://vivaldi.zool.gu.se/film/Movie-2-sm.movand http://vivaldi.zool.gu.se/film/Movie-2-sm.avi.
We confirm previously published observations that dopamine and, to a lesser extent, noradrenaline stimulate cement secretion from cyprid larvae(Okano et al., 1996). The major difference between the previously published work and the present study is that Okano's experiments were carried out using isolated cement gland preparations in vitro while our study is performed in intact cyprids in vivo. Perhaps, for that reason, our experiments required a higher concentration of dopamine to achieve stimulation. The higher concentration of dopamine might induce a massive cement secretion and, consequently, more granule swelling and vacuole formation(Kelly et al., 2004). Another difference could be that the extracellular medium used in Okano's studies did not support granule swelling like the extracellular fluid in the living cyprid kept in seawater. Granule swelling is known to be dependent on medium pH,Ca2+ (Espinosa et al.,2002) and ionic strength(Finkelstein et al., 1986; Nanavati and Fernandez, 1993),which may differ in the two experimental conditions. In zymogen granules of the pancreas, the mechanism of granule swelling during exocytosis was found to be regulated by a GTP-mediated process involving Gαi3 and aquaporin (Cho et al.,2002a).
Dopamine stimulation causes secretory vesicle loss and appearance of vacuoles. Cyprids were fixed at different times during exposure to dopamine (1 mmol l–1) and sectioned. Sections were stained with Toluidine Blue, and the different types of vesicles in the stained sections were counted under the microscope. Note that dense-core granules (type 1) reduce in number over time, with a proportionate increase in the number of vacuoles (type 4)after 10 min of dopamine treatment.
Dopamine stimulation causes secretory vesicle loss and appearance of vacuoles. Cyprids were fixed at different times during exposure to dopamine (1 mmol l–1) and sectioned. Sections were stained with Toluidine Blue, and the different types of vesicles in the stained sections were counted under the microscope. Note that dense-core granules (type 1) reduce in number over time, with a proportionate increase in the number of vacuoles (type 4)after 10 min of dopamine treatment.
Light microscopy of dopamine-stimulated cement glands. Cement glands were fixed at various times during dopamine exposure, and sections were cut. Toluidine Blue-stained sections were examined under the microscope. Note the absence of vacuoles in the control gland (A) and the appearance of vacuoles after 4 min exposure to dopamine (B), which increase in number after 10 min exposure (C).
Light microscopy of dopamine-stimulated cement glands. Cement glands were fixed at various times during dopamine exposure, and sections were cut. Toluidine Blue-stained sections were examined under the microscope. Note the absence of vacuoles in the control gland (A) and the appearance of vacuoles after 4 min exposure to dopamine (B), which increase in number after 10 min exposure (C).
As in most secretory systems, the cement proteins are stored within the vesicles at extremely high concentrations in a colloidal complex with numerous components such that the osmotic pressure inside vesicles is low(Kreuger et al., 1989). The hydration forces on the granule will be expected to be controlled by the colloid osmotic pressure within the vesicles and would be modulated by changes in the osmolality of the surrounding medium(Whitaker and Zimmerberg,1987). The intragranular complex needs to dissociate during exocytosis for secretion to occur. In some granules, a phase boundary between compact electron-dense material and less-compact amorphous material within vesicles has been observed, together with loss of material (see Walker, 1971 for comparison). Granule swelling following hydration of the granule contents might therefore be essential for secretion to occur. The structure within the type 3 granules,appearing as hydration channels, might be due to such changes, and similar structures have been observed in other dense-core granules such as the sea urchin egg cortical granules and mast cell granules(Dvorak and Morgan, 2000; Whalley et al., 2000).
Exocytosis is an ubiquitous event in biology and several hypotheses have been suggested as possible molecular mechanisms. It seems likely that the detailed molecular mechanisms of exocytosis might differ in different secretory systems depending upon the physiologically required speed of the exocytotic secretory event. It seems possible that in some cellular systems a single granule might undergo multiple fusion–secretion cycles (for reviews, see Burgoyne and Morgan,2003; Lindau and Alvarez de Toledo, 2003), leading to partial secretion of vesicle contents or a pulsatile form of secretion. Similar `kiss-and-run' type partial exocytosis has been observed in synaptic transmitter release(Aravanis et al., 2003). During exocytosis, the granule membrane transiently becomes part of the cell membrane and then is selectively retrieved, and it is possible that rapid cycling between a fusion state and a non-fusion state may occur(Burgoyne and Morgan, 2003; Schneider, 2001). The duration and diameter of the fusion pore opening will regulate how much granular material is secreted for each secretion cycle(Tsuboi and Rutter, 2003; Tabares et al., 2003). Thus,it is possible that the empty vacuoles, like the type 4 granules we observe,could result from multiple rounds of partial exocytosis, or complete exocytosis and total emptying.
In conclusion, the data we present here support the suggestion that the dense-core secretory vesicles within cement glands in cyprid larvae are secreted through a process of exocytosis. Exocytotic secretion of barnacle cyprid cement resembles the secretory event observed in many mammalian cell systems, including pulsatile or partial secretion of granule contents. In addition, we observed four different types of cement granules in dopamine-stimulated cyprids. With the exception of type 1 granules, all others appear swollen with partial or complete loss of granule contents and might represent vesicles that have undergone partial or complete loss of contents. Thus, the cement gland in barnacles appears to be a precisely regulated exocrine organ that is more complex in its organization and regulation than previously thought.
Acknowledgements
We would like to acknowledge Ulla Svedin and Elisabeth Norström for their outstanding technical skills. Financial support was provided by the MISTRA program Marine Paint, Stiftelsen Konung Carl XVI Gustafs 50-årsfond and Carl Tryggers Foundation.