Abstract Like other members of its genus, Conus californicus has a highly evolved venom apparatus with which it injects paralyzing venom into its prey through a modified radula tooth. Although much research is being conducted on the venom itself, relatively little is known about the biology behind the snail’s ability to inject the venom. The venom duct, radula sac and teeth were characterized in terms of anatomy, physiology and chemistry. The passageway between the venom duct and proboscis is a complex branching structure lined by a unique secretory cell type. Evidence from a combination of all methods used suggests that there are similar active peptides in the teeth as in the duct, leading to the possibility that the teeth are pre-loaded with venom. Introduction Snails of the genus Conus have long been of interest, both for their unusual shells and unique predatory hunting strategies, which employ a battery of peptide neurotoxins to paralyze their prey. A typical gastropod has a row of chitinous radular teeth that it uses for scraping food. The cones have modified their teeth into a combination of miniature harpoon and hypodermic needle that is used one at a time to inject venom into the prey. Each species has a unique tooth structure with varying size, shape, and number of barbs, and the differences in the teeth constitutes a basis for phylogenic classifications of Conus (Kohn et al., 1999 Although there are several hundred tropical species, Conus californicus is the only temperate species to inhabit the ocean off of the Pacific coast from the Farralone Islands, California to Cabo San Lucas, Baja California (Hanna and Strong, 1949). C. californicus, like all Conus species, is predatory, but it is unusually omnivorous compared to many typical tropical species, many of which have highly specific feeding habits (Kohn, 1966; Nybakken, 1970). In all cones, the venom apparatus consists of a muscular bulb connected to a long, highly convoluted venom duct (Fig.1). The teeth are made in the radula sac, which consists of a short and long arm. The radula sac enters the proboscis slightly anterior from the venom duct, i.e. towards the tip of the proboscis (Fig.1C). The teeth are formed in the long arm of the radula sac, and mature teeth are thought to be stored in the short arm until use (Marsh, 1977). The teeth of C. californicus are about 1 mm long with a basal and an apical opening and five barbs (Kohn et al. 1999). When resting the cone holds its proboscis inside of its mouth, but when prey is sensed, through chemoreceptors in its osphradium, (Spengler and Kohn; 1995), it extends the proboscis and begins to search for food. When the cone finds its prey, it will sting forcefully by inserting the tooth and injecting the venom. The anatomy of the venom apparatus has been characterized for several species. and there is a general model for how the venom is made and injected through the tooth (Endean and Duchemin, 1967; Songdahl, 1973; Hinegardner, 1958, Freeman et al., 1974). Specific peptide components of the venom are thought to be made as pro-peptides in the posterior region of the venom duct near the muscular bulb and then undergo posttranslational modifications to form mature venom peptides as they move down the duct (Bingham et al. 1996). It has been hypothesized that the muscular bulb provides the force to move the venom through the duct into the proboscis (Olivera, 1997). This idea has been questioned, however, because the muscular bulb does not appear to be capable of providing sufficient pressure to inject the venom (Songdahl, 1973). It is more likely that the proboscis itself provides the force to inject the venom (Kohn, et al. 1999). The hollow tooth is presumably held at the tip of the proboscis by a sphincter muscle, and in some species it has been reported that the tip of the tooth can be seen before the sting (Kohn 1956). In some cases the cone continues to hold onto the tooth after the injection and then uses it to pull the prey into its mouth. The main contemporary thrust of Conus research is focused on purification and characterization of specific peptides in the venom. These efforts have primarily employed tropical Conus species, because they are large and their venoms, especially those of the piscivorous species, are lethal to mammals including humans (Kohn,1998). Several Conus peptides are of commercial pharmaceutical interest at the present time (Well, 1998) This study provides a description of the general anatomy of the venom apparatus in C. californicus with the specific aim of elucidating the anatomical pathways by which the venom and teeth enter into the proboscis. The study also examined the composition of both the venom duct and the teeth using anatomical, physiological and chemical means. Material and Methods Histology Specimens of C. californicus were collected from Monterey Bay. Some were kept in an aquarium with water at ambient temperature, and others were held at room temperature. Three animals were relaxed in 2% magnesium chloride in order to dissect out the venom apparatus. Tissue samples were immediately fixed in 2% glutaraldehyde, 80% filtered sea water and 100 mM HEPES, dehydrated in a graded ethanol series and embedded in Spurs resin. Only two of the three samples infiltrated well, and sections 0.5u to 2 u thick were cut on an ultra-microtome with glass knives, stained with methylene blue in 0.1 N NaÖH, and examined on an Olympus BH-2 microscope. Physiology Individual teeth were manually removed from the short arm of the radula sac in filtered seawater. They were then broken up using a hand homogenizer in 50 uL of filtered seawater, and added to 0.2 ml of external recording solution (see below) yielding a final concentration of 0.022 teeth per uL of solution. Venom from the duct was manually extruded onto a piece of parafilm, mixed with 0.5 ml of external recording solution in a 1.Sml microcentrifuge tube, homogenized by quickly vortexing for 20 seconds and heated to 95° C for 5 minutes. Äfter a brief centrifugation (4 min at 13,000 rpm), the supernatant was diluted 1:5 with external recording solution for testing activity. Voltage dependent ionic currents were recorded form squid giant fiber lobe neurons using conventional single-electrode, whole-cell patch clamp. Linear (i.e. not voltage-dependent) ionic and capacity currents were removed online with a standard P/4 technique. The holding potential was-80 mV. Temperature is indicated in the legend to Fig. 7. External and internal pipette solutions were designed to eliminate K+ currents by eliminating K from both solutions. The external solution was composed of 480 mM Nacl, 10 mM CaClz, 20 mM MgCl, 20 mM MgSO4, and 10 mM Hepes at pH 8. The internal solution was composed of 100 mM sodium glutamate, 50 mM NaF, 50 mM Nacl, 300 mM tetramethylammonium glutamate, 10 mM NazEGTA, 25 mM tetraethylammonium chloride, and 10 mM Hepes at pH 7.8. Matrix-assisted laser desorption-ionization time of flight mass spectrometry (MALDI TOF The venom apparatus was dissected out and shipped to University of Illinois on ice. Individual portions of the tissues were dissected in isotonic NaCl, and placed in 2,5dihydroxybenzoic acid (DHB) on a MALDI target. Mass spectra were obtained using ’Physiology experiments were done by William F. Gilly at Hopkins Marine Station, Stanford University, MALDI TOF MS was done by Jonathan Sweedler at the University of Illinois at Urbana-Champaign. two mass spectrometers; a Voyager Elite and a Voyager DE STR (PE Biosystems, Framingham, MA) equipped with delayed ion extraction. A pulsed nitrogen laser (337nm) was used as the desorption/ionization source, and positive-ion mass spectra were produced using both linear and reflectron mode. Mass calibration was preformed externally using either bovine insulin (SIGMA, St. Louis, MO) or a previously calibrated spectrum obtained from Aplysia bag cells. Results The Venom Duct The venom duct is a thin convoluted tube connecting the muscular bulb and the proboscis. Although the entire duct was not analyzed completely, it was examined in several places as indicated in Fig. 20. As the duct leaves the muscular bulb (Fig. 2A) it is an orange pink color, about 250u in diameter, and the wall is composed of 3 distinct layers. There is an outer epithelium overlaying a muscular layer. A layer of relatively large cells (60u wide x 40u tall) surrounds the lumen, which consist of some granular material as well some diffuse non-cellular content (Fig. 2A). About one third of the way to the proboscis the duct grows to about 375u in diameter, changes color to creamy white, and also changes morphologically (Fig. 2B). There is still an outer epithelium, although it is much thinner, and this covers an extremely thin muscular layer (Fig. 2B). The cells of the inner layer are much larger, about 75u tall and 60u wide, and are densely packed with granules of different kinds, There is still a central lumen, and this is packed with granules. It is difficult to unambiguously visualize the border between these cells, and between these cells and the lumen. The anterior end of the venom duct looks similar to the middle section, except it has grown to 475u in diameter. At the site where the venom duct enters the proboscis musculature it narrows to 125u in diameter and is lined by cells about 25u tall with large basal nuclei (Fig. 2D). Cells with visible boundaries that contain granules can be seen, and these granules often appear to be segregated into different types within individual cells. In addition, a very distinct type of cell appears in this region that is filled with tiny granules stained dark purple by methylene blue. These cells surround the duct as it passes through the proboscis (Fig. 3-4). As the duct crosses the proboscis, the lumen becomes very narrow, about 20u wide, and is mostly filled with small granules that do not stain with methylene blue. Light and dark blue granules from the lumen of the duct outside of the proboscis can be seen entering into the passageway but seem to disappear where the lumen branches (Fig. 3D,E). The white granules continue in the channel, however. The passageway branches about 250u into the proboscis (Fig. 3E, F) and the two channels cross an ambiguous area before they exit the proboscis between folds of the epithelium that lines the proboscis lumen (Fig. 4). These epithelial cells have well defined basal nuclei. In the area where the venom duct exits, these folds extend further into the proboscis wall (Fig. 4B), and the epithelium appears to merge here with the purple cell layer that comprises the wall of the venom duct. There is a central finger that has purple cells lining both sides, and two flanking fingers where purple cells only line the side facing the duct. The Radula Sac The short arm of the radula sac consists of an outer epithelial layer about 7.5u thick, a layer of connective tissue and an inner epithelial layer about 125u thick. The ligament sac takes up a large part of the short arm. Another large portion is filled with cells about 25u long, with basal or central nuclei. These cells have diffuse contents that stain light blue with methylene blue (Fig. 5C), and form a thick layer that folds onto itself. The teeth in the short arm are usually in close contact with this cell layer. Some of the teeth are positioned in random directions, but a majority of them face with their points towards the proboscis, and all of the teeth entering the proboscis point this way (Fig. 5A). The space around the teeth contains a mixture of granular, cellular and unidentifiable material, some of which appears to be similar to material in the venom duct (Fig. 5E, F). Similar material also fills a majority of the teeth, especially those entering the proboscis (Fig. 5A). The teeth are enveloped by muscle as they travel through the proboscis. As the connection forms, the sac and the proboscis are in close contact on many sides, and folds of the proboscis wrap around the sac. The teeth enter into the proboscis tip first, and more then one tooth in the passageway can be seen (Fig. 5A) The long arm of the sac was studied not in detail, although individual teeth were examined. Teeth at the blind end were less developed, and generally had less discrete matter in their lumen than teeth from the short arm (Fig. 6). Physiology Contents of the venom duct and radula sac were also characterized in terms of their neurotoxic activities. Previous work on in C. californicus has indicated the existence of a peptide that specifically inhibits voltage-gated sodium channels in cephalopods and not in gastropods (Bingham et al. 2000). Duct venom isolated without any chemical extraction in the present study also inhibits sodium channel activity in squid neurons (Fig. 7B). Fig 7A shows that both outward (top panel) and inward (middle panel) sodium currents were inhibited by the tooth venom. Mass Spectrometry Results from the reflectron MALDI mass spectra, which requires more concentrated samples and greater instrumental optimization, but provides greater accuracy, are shown in Fig 8. Linear mode operation gave similar results, but these data are not shown. Region 1, the most posterior region examined (level 1 in schematic of Fig. 8), shows only one peptide, with a molecular weight of 2993-2995, that occurs elsewhere in the duct and none that co-occur in the radula sac (Fig. 8A,B,C). Region 2 and region 3, the mid and anterior regions of the duct, have many peptides in common (Fig. 8B,C). They also share at least three peptides with those found in the tooth (Fig. 8B,C,D). Discussion The Venom Duct The anatomy and chemistry of the venom duct of C. californicus is in general agreement with previous studies of other Conus species (Hermitte, 1946; Endean and Duchemin 1967; Bingham et al. 1996). The posterior or muscular bulb end of the duct has a distinct morphology, the peptides in this region are larger than those in more anterior regions, and is not as active physiologically. This supports the current hypothesis that a pro-peptides in the venom is manufactured in the posterior region and then undergo post transitional modifications as they moves towards the anterior region of the duct (Olivera, 1997; Endean and Duchemin, 1967,) How the venom gets from the duct where it is made to the tip of the proboscis has not been discovered. None of the current literature describes the passageway from the duct into the lumen of the proboscis. The most prominent feature of this region is the appearance of cells containing granules that stained purple. These cells appear in the wall of the duct just before it enters the proboscis and line the passageway for the venom through the proboscis wall. Although the purple cells appear to be secretory, the identity of the putative product is unknown. There are several possibilities. The cells may simply secrete some sort of mucus, which might act as a carrier for the venom flow. Another possibility is that some enzymatic material is secreted that could activate the venom peptides. Secreted enzymes could also possibly help break down tissue at the injection site and allows the venom to exert its effects faster or more efficiently. 10 A final possibility is that these purple cells could be secreting a unique venomous peptide not found elsewhere in the system. Bingham et al. (1996) suggested the existence of such a unique peptide in milked venom samples of C. striatus, which they felt was not the result of a posttranslational modification of a peptide in the venom duct. The location of these purple cells puts them in a position to be the source of such peptides. Further studies on these purple granules need to be done in order to determine what they are secreting and what role they play in the venom apparatus. The branching of the passageway also raises some questions. It is difficult to follow any one branch as it travels through the proboscis, leaving a gap in the understanding of the passageway’s structure. However, significant insight has been gained. The passageways after branching are very small, only about 1Ou across at its widest point, yet this appearance of the passageway lasts through more than 100u of sections. This suggests the passage is not a cylindrical duct, but rather a flat channel that is much bigger in one dimension than the other. How many of these channel there are is still uncertain. The branching also could suggest that some sorting of granules goes on in the passageway to ensure that only mature peptides are passed into the proboscis lumen. The venom that the snail injects is much simpler than the venom found in the duct in regards to the number of peptides. Perhaps the complex nature of the passageway acts like a sieve to permit sorting of venom components on a basis of granular size or some other feature of individual peptides. The area between the branching point of the passageway branches and the proboscis lumen is also nebulous. In my sections a large cavity-like area appears in this gap, but it is unclear whether this is real cavity or an affect of the fixation process (Fig. 4). Small channels can be seen entering into the passageway around a finger of the proboscis lumen and it is presumed that the venom exits through these (Fig. 4 Further studies need to be done in order to better determine the nature of this complex passageway. Nevertheless, it is clear that the passageway through the proboscis is not a straight large diameter tube. The Radula Sac and Teeth The literature to date characterizes the teeth as being hollow and empty when they are ready for insertion to prey. (Kohn, 1998; Marsh 1977). Marsh (1971) reported that the teeth in the short arm of the radula sac for several Australian species of Conus were free from cellular debris, and speculated that the salivary glad duct could be responsible for this clearing. The teeth of C. californicus, however, are full of different types of granules and cellular debris. This is true for teeth in both arms of the sac, and especially for the teeth entering into the proboscis. The content of the teeth looks very similar to the content of the sac outside of the teeth. In some cases, as in Fig. 5D, clusters of granules that stain solid blue with methylene blue are found in the wall of the radula sac, and similar individual granules inside teeth as well. These granules are similar to ones in the venom duct. Although the exact nature of these individual granules is unknown. similarities in their morphology suggest that they may contain similar peptides. Two other lines of evidence support this idea. First, examination of an individual tooth using MALDI revealed that some of its peptides were indistinguishable from those 12 in the venom duct. The peaks in the 3131-3133 MW range in Fig. 8, are of particular interest. Bingham et al. (2000) have characterized a venom duct peptide of this molecular weight that is thought to be the peptide with sodium-channel blocking activity. Physiological tests also indicate the presence of a specific sodium-channel blocker in material isolated from both the venom duct and the teeth (Fig. 7). Although the effect of the tooth venom is not as strong as the duct venom samples, it nevertheless exists. The discrepancy is probably only due to the fact that the teeth sample was extracted from 5-6 teeth with an estimated total volume of several nL, where as the venom duct sample was prepared from a whole duct with an estimated volume of several mL. Further experiments are needed to quantitatively compare the potency of material from the two source Results from this study provide strong evidence that the radula teeth of C. californicus contain some of the same peptides as those normally found in the venom duct. Furthermore, one of these peptides acts to inhibit sodium channel activity. These data are thus contrary to the widely held view that the teeth are completely empty tube and act only as a passive delivery device. Whether or not these data are relevant to other Conus species remains to be established. It is also unclear what role the seemingly small amounts of toxins stored in the teeth might actually play in the hunting process. Additional studies are needed to structurally identify the peptides in the teeth, to determine the biological activity of these peptides and to elucidate the source of these peptides and the mean by which they are packed into the teeth. 13 Acknowledgements 1 would like to thank William F. Gilly for being my advisor on this project and being so patient through this entire learning experience, and for providing me with physiology data for this project. I would also like to thank Jonathan Sweedler and his group at the University of Illinois for doing the MALDI work. Many thanks go out to Lisa Walling, and Gina Kang, and special thanks to Jim Watanabe for providing Cone specimens. My gratitude is also extended to the following people: Jon-Paul Bingham, Barbara Block, Chris Patton and William C. Pitts and the entire Gilly lab for their support. Literature Sited Bingham, J. P., A. Jones, R. J. Lewis, P. R. Andrews, and P. F. Alewood. 1996.Conus Venom Peptides (Conopeptides): Inter-Species, Intra-Species and Within Individual Variation Revealed by lonspray Mass Spectrometry. pp. 13-27. in Biomedical Aspects of Marine Pharmacology. Bingham, J. P., A. Burlingame, E. Moczydlowski, and Wm. F. Gilly. 2000. A new highly selective conotoxin from Conus californicus that targets voltage-gated neuronal Na' channels of squid. Abstracts of Papers Presented at the 54th Annual Meeting of the Society of General Physiologists. Endean, R. and C. Duchemin. 1967. The venom apparatus of Conus magus. Toxicon. 4:275-284. Freeman, S. E., R. J. Turner, S.R. Silva. 1974. The venom and venom apparatus of the marine gastropod Conus striatus Linne. Toxicon. 12:587-592. Hanna G. D., and A. M. Strong. 1949. West American Conus. Proceedings of the California Academy of Science, 4th series. 26: 247-322. Hermitte, L. C. D. 1946. Communications: Venomous marine molluscs of the genus Conus. Transactions of the Royal Society of Tropical Medicine and Hygiene. 39: 485-512. Hinegardener, R. T. 1958. The venom apparatus of the cone shell. Hawaii Medical Journal 7:533-536 Kohn, A. J. 1956. Picivorous gastropods of the genus Conus. PNAS. 42:168-172 Kohn, A. J. 1966. Food specialization in Conus in Hawaii and California. Ecology. 47:1041-1043. Kohn, A. J. 1998. Family Conidae. pp in Beesley, P. L., Ross, G. J. B., and Wells, A. (eds) Mollusca: The Southern Synthesis. Fauna of Australia. Vol. 5 CSIRO Publishing: Melbourne. Kohn, A. J., M. Nishi, and B. Pernet. 1999. Snail spears and scimitars: A character analysis of Conus radular teeth. J. Moll. Stud. 65:461-481. Marsh, H. 1971. The foregut glands of vermivorous Cone Shells. Australian Journal of Zoology. 19:313-326. Marsh, H. 1977. The radular apparatus of Conus. J. moll. stud. 43:1-11 Nybakken, J. 1970. Correlation of radula tooth structure and food habits of three vermivorous Species of Conus. The Veliger. 12:316-318. Olivera, B. M. 1997. Conus venom peptides, receptor and ion channel targets, and drug design: 50 million years of neuropharmacology. Molecular Biology of the Cell. 8:2101-2109. Songdahl, J. H., 1973. The venom and venom apparatus of the Atlantic cone, Conus spurius atlanticus (Clench). Bulletin of Marine Science. 23:600-612 Spengler, H. A., and A. J. Kohn. 1995. Comparative external morphology of the Conus osphradium. Journal of Zoology (London). 235:439-453 Wells, W. A., 1998. The snail companies. Chemistry and Biology. :R235-R236 0 Fig. 1A. Schematic representation of venom apparatus in C. californicus. This figure, which is not to scale, shows the main parts of the venom apparatus including the siphon, proboscis, muscular bulb, venom duct and radula sac. Fig. 1B. Complete venom apparatus as it looks in a snail . The venom duct changes color at about the same region it would in a live snail. Fig. 1C. Venom duct and radula sac connection to proboscis, showing that the venom duct enters posterior to the radula sac Abbreviations: es, esophagus; m, mouth; mb, muscular bulb; nr, nerve ring; p, proboscis; rs, radula sac; sg, salivary gland; sgd, salivary gland duct; t, radula teeth. Fig. 2 Micrographs of cross sections through the venom duct from the regions indicated in panel C. Schematic representations, not to scale of two of the regions are shown in panel 2A1 and 2B1. Scale bar is 50u throughout. Fig. 2A. Posterior or muscular bulb end of the venom duct, showing inner epithelial cells surrounding a central lumen that contains some granules Mid-region of venom duct after color change. Inner epithelial cells are Fig. 2B. larger and packed with granules. The central lumen also has more granules. Fig. 2C. Diagram illustrating the different levels of the venom duct analyzed. Fig. 2D. Venom duct as it approaches the proboscis. Cells with larger basal nuclei appear. Discrete bundles of granules can be seen in the inner epithelial cells. Blue granules are localized in cells at the top; clear granules are in cells at the bottom-left. Fig. 3. Longitudinal sections through the venom duct as in the proboscis wall. Scale bars are 50u throughout. Fig. 3A Venom duct as it starts to enter proboscis wall. The purple cells first appear here. Fig. 3B. Magnified view of A. Fig. 3C. The venom duct is pushing through proboscis wall with the purple cells lining passage. Fig. 3D. Venom granules in channel in the proboscis wall. A narrow passage for venom from the duct to enter into the proboscis is seen here. Fig. 3E. & F Two views of the channel branching. 3E shows the branches disappearing on the right. The large layer of purple cells surround the central channel is shown. 3E also shows that the blue granules which are seen before the branching seem to disappear, but the clear granules continue. Fig. 4. Longitudinal sections through the venom duct crossing the proboscis wall and entering proboscis lumen. Scale bar is 50u throughout. Overall view of venom duct crossing the proboscis. The whole proboscis Fig. 4A. wall is shown. Magnified view of channel exiting proboscis wall into proboscis lumen Fig. 4B. There are two exit channels visible. The proboscis lumen epithelium merges with the venom duct epithelium in this region. Fig. 4C. Fig. 5. Fig. 5A & B. Fig. 5C&D Fig. 5E& F. Fig. 6. Fig. 6A. Fig. 6B. Fig. 6C. Fig. 7. Fig. 8. Schematic diagram of how the venom duct crosses the proboscis. Not to scale. Teeth in the radula sac. Three different views of teeth in the radula sac are shown. Each panel on the right is a magnified image of the panel of the left. Scale bar is 50u throughout. Teeth entering proboscis wall from radula sac. The teeth are all pointing towards the proboscis lumen and are all filled with material. Comparison of radula sac content with tooth content. The basal opening can be seen on two teeth in the short arm of the radula sac. Material inside and outside teeth appears very similar Blue granules in clusters can be seen in the radula sac wall and individual granules can be seen in the tooth. These granules look similar to those found in the venom duct, (see Fig. 2). Teeth from different parts of the radula sac. Line drawing shows approximately where teeth were taken from. Scale bar on all is 50u. Living tooth taken from the short arm of the radula sac. It is more fully developed, larger, and the lumen is packed with granular material in it. Tooth from long arm fixed in gluteraldehyde, placed into agar to hold it, embedded in Spurs resin, and stained with metheylene blue. The ligament can be seen off the basal end. The lumen is not as full of material. Living immature tooth from extreme end of long arm. The ligament sac is coming off right end. The tooth is not fully developed. Results of the tooth venom and venom duct extract experiments. The tot panels shows the effect on a inward current, the middle panels shows the effect on an outward currents and the bottom panels show the voltage applied. The tooth venom experiment (panel A), showing that 0.002 tooth units inhibits the sodium channel. The venom duct extracts (panel B) show similar effects. The temperature of each experiment is indicated at the bottom. Results from the reflectron MALDI MS. Intensity is along the y-axis and mass per charge is along the x-axis. Marked peaks of indicated molecular weight are the same peptide in different samples. The sample locations are shown at right. Fig.1 S siphon proboscis S mouth mb sgd es radula sac venom duct — B vd nr Sg A muscular —esophogus foot 5.. rS C Fig2 A puisculature epithelium e 4 lumen innerlg granules epithelium B musculature inner epithelium granules lumen epithelium mb A2 2. 25 80 56 0 * - 9 E A 4 S S- D Z — 1u Og82 v398 — 8668 Og82 8298 1818 2818 2 1u 8662 3982 992 9668 8818 —.