Abstract Marine Conus snails produce a wide variety of neurotoxins ("conotoxins") that specifically bind to different ion channels and neuronal receptors. Much research has revolved around the physiological effects of the conotoxins without focus on how these snails produce their venom. For example, the temperate snail Conus californicus produces a novel conotoxin that targets voltage-sensitive sodium channels in cephalopod, but not gastropod neurons. The predicted molecular weight from the CDNA sequence encoding this peptide is 4680 Da. Numerous dense granules are present in the venom duct of Conus californicus, where venom production occurs. In an attempt to isolate the granules, as a first step towards biochemical characterization, crude venom was run on a Ficoll-sucrose gradient. Inspection of these granules with light microscopy and scanning electron microscopy confirmed the presence of dense granules of approximately 2-6 um in diameter. Electrophoresis of the SDS solubilized granules revealed a prominent 5 kDa band on SDS-polyacrylamide tricine gels that matches a band in "milked" venom and duct venom. This 5 kDa granule peptide was not solubilized by nonionic detergent (NP40) suggesting that it is packed within a tightly associated granule complex. Observation with light microscopy indicated that application of heat disintegrated the granules and released soluble, bio-active peptide into the medium. Preliminary whole cell voltage-clamp recordings of disassociated cells from the stellate ganglion of Sepia officinalis showed that the granule contents block voltage-gated sodium channels. In conclusion, it appears that these dense granules contain a peptide resembling mature milked venom peptide in both mass and function. Introduction Lurking within the waters of Monterey Bay, Conus californicus utilizes an assortment of peptide toxins to catch its prey of polychaetes, snails, a small species of octopus, and small fish. Conus californicus is just one of an estimated 500 Conus species that produces “conotoxins" to capture prey. Conotoxins are produced in the venom duct of the Conus snail and stored in modified radular teeth. The predatory snail harpoons its victims with a barbed tooth and injects up to 30 ul of venom. (Bingham, et. al. 1996) Different Conus species have evolved neurotoxins, which specialize in swiftly immobilizing their prey. (Duda and Palumbi 1999) These peptide toxins are now valued for their potential medicinal role in targeting specific families of ion channels and neuronal receptors. The peptide toxins of the Conus snails are generally 10-30 amino acids long. (Ramilo et. al. 1992) Some of the conotoxin families include o-conotoxin that targets voltage-sensitive calcium channels and a-conotoxin that affects nicotinic acetylcholine receptors. This research examines a novel family of conotoxin from the "milked" venom of Conus californicus, which is structurally unrelated to previously classified Conus peptides. This novel conotoxin has a predicted molecular weight of 4680 Da from its CDNA sequence and selectively targets voltage-sensitive sodium channels in cephalopod, but not gastropod neurons. (Based on unpublished data by Joseph Schulz) Although the physiological effects of these various conotoxin families have been examined extensively, little research has been devoted towards understanding how the Conus snails produce their complex venoms. Production of Conus venom occurs in four general steps: precursor synthesis, processing, packaging, and venom delivery. (Maguire and Kwan, 1992) The first three steps occur within the venom duct of the Conus snail. Visual inspection of the venom duct shows that it contains numerous dense granules of various sizes. However, the purpose of the granules remains a mystery. The granules may be involved in packaging of immature venom, synthesis of venom to a certain phase, transportation of mature venom to the radular teeth, or even transport of enzymes necessary for venom production. In order to ascertain the function of the granules, the first step is to identify their contents. The purpose of this research was to perform a biochemical analysis of the Conus californicus granules in order to determine if the granules contain material similar to mature venom peptides in either duct derived or milked venom. Materials and Methods Dissections Conus californicus were obtained from the Monterey Bay and kept in tanks. Specimens were anaesthetized in 2% magnesium chloride in seawater before dissections. The venom duct and venom bulb were removed and the crude duct venom was squeezed out. Samples were microcentrifuged for 15 minutes at 14,000 x g, and supernatant was removed. (Fig. 1) The pellet was resuspended in a small quantity of seawater. Samples were immediately frozen in liquid nitrogen. Granule Isolation using Sucrose Discontinuous Velocity Sedimentation The gradient was set inside a 2 ml ultracentrifuge tube. The top layer was composed of 0.5 ml of seawater. The second layer consisted of 0.5 ml of 0.25 M sucrose in seawater. The bottom layer was made from 1 M sucrose in seawater. The duct venom pellet was resuspended and added to the top of the gradient. The gradient was then spun at 100,000x g at 4 C for 45 minutes. Each layer was removed at stored at 4 C. All natural seawater was buffered with Hepes and had a pH of 7 to 8. Granule Isolation using Ficoll-Sucrose Discontinuous Velocity Sedimentation The Ficoll-sucrose gradient was a modification of the protocol described by R.S. Cameron and J.D. Castle (1984). The gradient was prepared in a 2 ml ultracentrifuge tube. (Fig.2) The top layer consisted of 0.2 ml of 0.8 M sucrose in seawater. A second layer of 0.5ml of 1.58 M sucrose in seawater followed this. The next layer was composed of 0.75 ml of 1.55M sucrose and 4% Ficoll in seawater. A final layer was prepared with 0.5 ml of 2M sucrose and 4% Ficoll in seawater. The resuspended duct venom pellet was added to the gradient and spun at 100,000x g for 45 minutes. Each layer was subsequently removed. The Ficoll pellet was resuspended in 0.4 ml of seawater and spun in the microfuge for 15 minutes. The Ficoll pellet was then resuspended in a final volume of 30 uL. Seawater was Hepes buffered and possessed a pH of 7.8. All samples were frozen at -80'C. Solubilization of Granules in NP40 Granule samples from the sucrose gradient or the Ficoll-sucrose gradient were solubilized for twenty minutes in 1% Nonidet P40 in seawater. The samples were then microcentrifuged to produce two layers, which were subsequently separated and stored at —4 C. SDS-PAGE Gels Gels were set up using the Bio-Rad Mini Protean II Electrophoresis gel system. Samples were dissolved in 2X Laemmli sample buffer and heated to 95°C for 6 minutes. Gels were run with 16mA/gel for approximately 1.5 hours or until the dye front reached the end of the gel. 10% and 5% SDS-PAGE gels were only run for samples from the sucrose gradient. Tricine SDS-PAGE Gels As described by Judd in the Protein Protocol's Handbook, the separating gel was composed of 1.34 ml water, 2 ml separating gel buffer, 2 ml of IX crosslinker separating/spacer gel acrylamide, 640 uL glycerol, 2 uL TEMED, and 100 uL 10% ammonium persulfate. The recipe for the spacer gel was 1.38 ml water, 1 ml separating/spacer gel buffer, 0.6 ml IX crosslinker separating/spacer gel acrylamide, 1 uL TEMED, and 10 uL 10% ammonium persulfate. The stacking gel recipe was composed of 2.06 ml water, 0.38 ml 1.OM Tris buffer, 0.5 ml stacking gel acrylamide, 12 uL O.SM EDTA, 1.5 uL TEMED, and 30 uL 10% ammonium persulfate. Samples were dissolved in 2X sample buffer and heated to 95'C for 6 minutes. 1OX cathode running buffer was made of IM Trizma base, IM tricine, and 1% SDS and had a pH of 8.25. The 1OX stock of anode buffer was composed of 2M Trizma base with a pH of 8.9. Tricine gels were run with 16mA/gel for approximately 3 hours when the dye front reached the end of the gel. Staining the Gels SDS polyacrylamide gels were silver stained according to Morrissey (1981). Gels were prefixed for approximately 1 day in 50% methanol and 10% glacial acetic acid followed by 5% methanol and 7% glacial acetic acid for 30 minutes. Gels were then fixed in 10% glutaraldehyde for 30 minutes and then soaked in water for over 1 hour. This was followed by soaking the gels for 30 minutes in 5ug/ml dithiothreitol and then 30 minutes in 0.1% silver nitrate. The gels were developed in 50 uL of 37% formaldehyde in 100 ml 3% NacOz. Staining was terminated by adding 1% glacial acetic acid. Gentle agitation occurred during each of the above steps. Gels were stored in 0.03% NacO, Scanning Electron Microscopy Granules were fixed in 2% EM-grade glutaraldehyde in seawater. Dehydration occurred through a graded series of 20%, 50%, 70%, 80%, 95%, and 100% ethanol concentrations. Approximately 1 ml of hexamethyldisalizane was added to dry the granules. The granule samples were then mounted on carbon tape and sputter coated with gold. Images were taken with a Hitachi S-450 Scanning Electron Microscope operating at 15 KV. Verification of Release of Granule Contents A lul sample of the Ficoll gradient granules was heated for 5 minutes at 95'C. Unheated and heated lul granule samples were observed with oil emersion light¬ microscopy at 60X. Whole Cell Voltage-Clamp Recordings Whole cell voltage-clamp was performed on disassociated cells from the stellate ganglion of Sepia officinalis. The external recording solution was composed of 480 ml Nacl, 10 ml CaCh, 20 ml MgCl, 20 ml MgSO4, and 10 ml Hepes at a pH of 8.0. The supernatant from the duct venom was composed of 3 C. californicus specimens in 0.4 ml of seawater. The duct venom supernatant was heated for 5 minutes at 95°C. 1/100 and 1/500 dilutions of duct venom supernatant were applied to the cells. In order to release the granule contents, the granules from the Ficoll gradient pellet were also heated at 95'C for 5 minutes. A 1/500 or 1/100 dilution of the Ficoll pellet was applied to disassociated cells from the stellate ganglion. All recordings were done at 12-13'C using custom software. Results Light Microscopy and Scanning Electron Microscopy The contents of the Ficoll-sucrose pellet were observed with a light microscope as shown in Figure 3. The granules appeared abundant, dense, and varied in size. They generally clumped together as exemplified by Fig. 3B. SEM images also revealed dense granules ranging from 2-6um in diameter. (Figs. 4A,B,C, & D) Most of the granules had a convoluted surface structure (Figs. 4A, B, and C), which is probably due to the dehydration process. Only two out of approximately 100 granules appeared non- shrunken. (Fig. 4D) SDS PAGE Gel Results of Sucrose Discontinuous Velocity Sedimentation Centrifugation of the sucrose gradient produced a supernatant of mixed sucrose layers and a pellet. During an initial attempt to analyze the results of the sucrose discontinuous gradient, 10% and 5% SDS PAGE gels were run. As shown in Figure 5. all duct venom components were very complicated and many bands of similar molecular weights appeared in lanes 2-6. The material from the sucrose gradient's supernatant (lane 2) appeared to be composed of similar material as the supernatant from the whole duct venom of C. californicus (lane 3). For example, both lanes 2 and 3 possessed a prominent band at approximately 25 kDa and a 49 kDa band. However, taking the unbalanced sample loads into consideration, both bands also appeared in the sucrose gradient's "granule" pellet (lane 4), whole duct venom (lane 5), and duct venom pellet (lane 6) as well. Overall, it appeared that the majority of the bands in the sucrose gradient pellet were also present in the gradient's supernatant. The tricine gel (Fig. 6) showed a prominent 5 kDa band in the "granule" pellet from the sucrose gradient. Lane 4 from Figure 6 shows this 5 kDa band, which matches a band of the same molecular weight in the whole duct venom (lanes 3 & 9) and in milked" venom from C. californicus (lane 10). This 5 kDa band did not appear in the supernatant from the sucrose gradient (lane 8). SDS PAGE Tricine Gel Results of Ficoll-Sucrose Discontinuent Velocity Sedimentation Ultracentrifugation resulted in a top layer, a thin white band, a middle layer, and a pellet. (Fig. 2B) The tricine gel (Fig. 6) produced a prominent 5 kDa band from the Ficoll gradient pellet (lane 5) that matched the sucrose gradient pellet band (lane 4) in both weight and intensity. Many of the bands from the sucrose gradient pellet (lane 4), such as the thick band weighing less than 1.4 kDa, were not present in the Ficoll gradient pellet (lane 5). The tricine gel in Figure 7 compared the bands from the different Ficoll gradient layers as described in Figure 2. Unfortunately, Figure 7 does not adequately represent these results because this gel did not run uniformly. As a result, it is difficult to match the molecular weight standard bands (lane 1) directly to the bands in the other lanes without considering the entire gel. Careful inspection of this tricine gel showed a 5 kDa band present in the Ficoll gradient pellet (lane 3), the whole duct venom (lane 2), and in the white band (lane 5) from the Ficoll gradient. The 5 kDa band did not appear in the middle layer (lane 4) or top layer of the Ficoll gradient (lane 6). Molecular weights were calculated (Fig. 8) from the distances the bands traveled in the IX tricine gel shown in Figure 7. Solubilization of Granules in NP40 After centrifugation of the granule pellet from the sucrose gradient in non-ionic detergent, the NP40 pellet still showed a strong band of 5 kDa (lane 3 in Fig. 9). There was also a less prominent band of 5 kDa in the NP40 supernatant (lane 4). Except for the 5 kDa band, it appeared that most of the bands were different in the NP40 supernatant and pellet. Combining the bands from NP40 pellet and NP40 supernatant produced the original granule pellet bands from the sucrose gradient (Fig. 2), which did not resolve well in the figure. The solubilized Ficoll gradient pellet produced similar results. However, the bands were much lighter and difficult to scan onto a computer. Whole Voltage-Clamp Recordings After applying heat to the Ficoll-sucrose gradient pellet, there were very few intact granules present compared to the control. For example, within a minute’s limit an average of 11 intact granules were counted in the control slide and only 3-4 granules were found in the heated sample. Most of the heated granules appeared broken up and there was more free material present in the background. Preliminary whole cell voltage-clamp recordings demonstrated that whole duct venom from Conus californicus blocked inactivating Na current. (Fig. 10A) A 1:500 dilution of duct venom partially blocked inward Na- current, and a 1:100 dilution appeared to nearly eliminate inactivating Na current. A 1:500 dilution of granule contents had little effect on inactivating Na current, but a 1:100 dilution blocked approximately 1/2 of the inward Na current. (Fig. 10B) In contrast, a 1:500 dilution of the granule contents blocked fast-inactivating inward current. (Fig. 10C) Both the duct venom and the granule contents were less effective on the non-inactivating, inward tail currents. These latter currents are carried primarily by Ca. Discussion Much time was devoted towards developing an adequate method of purifying the granules from the duct venom. An initial attempt was made using a sucrose discontinuous gradient. However, massive components, such as epithelial cells inadvertently obtained from the dissection process, appeared to contaminate the granule pellet. In an attempt to further separate these undesirable components from the granules, 10 a Ficoll-sucrose discontinuous gradient was employed since it possessed additional dense layers. Only very heavy duct venom components, such as dense granules, could possibly make their way to the very bottom of the Ficoll-sucrose gradient. This granule purification process appeared more efficient because it eliminated excess bands found in the lanes containing the sucrose gradient pellet. Because each gel varied slightly in shape and how the samples ran, it was difficult to align the bands without surveying the entire gel. Results indicate that the granules possess a peptide with a molecular weight of approximately 5 kDa, which corresponds to the predicted molecular weight from the novel C. californicus peptide’s CDNA sequence of 4680 Da. However, there is a discrepancy in the exact molecular weight of the C. californicus novel peptide. Mass spectroscopy suggests that the novel peptide is 3120 Da. Possible ionization or fragmentation problems with the mass spectroscopy may have caused the lower molecular weight results. Unfortunately, this novel peptide was not run on a gel so it is not possible to directly compare bands. Another concern is that Conus venoms change while the animal remains in captivity and this may cause variances in molecular weight predictions. Overall, the tricine gels indicted that the granules contain a prominent peptide of approximately 5 kDa, which matches a band in whole duct venom and in milked venom. The 5 kDa band is also present in the thin, second layer from the Ficoll-sucrose discontinuous gradient (Fig. 2B). Visual inspection with light microscopy revealed minute, round bodies in this thin layer. These two clues suggest that this layer contains tiny granules. Physiological experiments were not performed with this material and it is unknown if the peptides from these "granules" contain active peptide toxins. In contrast to the complicated material in the Ficoll-sucrose gradient pellet, the second band layer contains only a few components (lane 5 in Fig. 7). If the "granule" contents from the thin band layer are biologically active, they may supply an excellent source of purified Conus peptide. As shown in Figure 10D, the novel C. californicus peptide selectively blocks fast- inactivating Na current in disassociated cells from the stellate ganglion of Sepia officinalis. (Based on unpublished data by Gilly). Preliminary whole-cell voltage clamp experiments suggest that the granule contents block both fast- and slower- inactivating Na currents. Although little work has been done to investigate the novel peptide's ability to block inactivating inward current, it appears that the granule contents contain the novel peptide in some form. For example, the granules may possess an unmodified, less discriminating form of the novel peptide that blocks various types of inactivating Na channels. A concern about these experiments is that it was difficult to discern the exact peptide concentrations in duct venom and the granule contents. An estimation was made by assessing the different volumes of the Ficoll gradient layers obtained from a known amount of duct venom. These physiology experiments suggest that duct venom and granule contents reduce the current amplitude to different degrees. (Fig. 10A,B) A discrepancy in concentrations of the peptide toxin may cause this effect. In order to confirm the results of these experiments, further physiology experiments need to be conducted after verification of peptide concentrations within each sample. Future experiments should also explore the kinetics of the channels affected by these peptide toxins. 12 In conclusion, the results from the biochemical and physiological experiments suggest that the granules contain a putative Na channel-blocking peptide, which is a component of both duct and milked venom. Thus, the venom duct granules may provide a significant, but overlooked, source of potent peptide neurotoxins. This research focused on identifying the 5 kDa peptide in the granules. However, the granules contain numerous other components, which need further analysis. Additional research might also focus on why the peptides must be enclosed within a tightly associated complex that is not soluble in non-ionic detergent. Overall, this study is just the first step in understanding the contents and function of the Conus duct venom granules. Acknowledgements 1 thank Professor Gilly for his guidance, encouragement, and Conus californicus specimens. I would also like to thank him for allowing me to have unlimited access to his lab over the past few months. I want to give an extra special thanks to Joseph Schulz for showing me these biochemical techniques and for his never-ending help and patience throughout the quarter. I would also like to acknowledge all of the members of Gilly's lab for answering my numerous questions. Finally, I want to thank Chris Patton for helping me produce the SEM images of the granules. References 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 ionspray mass spectrometry. pp. 13-27. in Biomedical Aspects of Marine Pharmacology. Cameron, R.S. and J.D. Castle. 1984. Isolation and compositional analysis of secretion granules and their membrane subfraction from the rat parotid gland. J. Membrane Biol. 79:127-144. Duda, Jr. T.F. and S.R. Palumbi. 1999. Molecular genetics of ecological diversification: duplication and rapid evolution of toxin genes of the venomous gastropod Conus. Proc. Natl. Acad. Sci. USA. 96:6820-6823. Gonoi, T., Y. Ohizumi, J. Kobayashi, H. Nakamura, and W.A. Catterall. 1987. Actions of a polypeptide toxin from the marine snail Conus striatus on voltage-sensitive sodium channels. Molecular Pharmacology. 32:691-698. Hnegardner, R.T. 1958. The venom apparatus of the cone shell. Hawaii Medical J. :533-6. Judd, R.C. SDS-polyacrylamide gel electrophoresis of peptides. pp. 101-107 in J.M. Walker, ed. The Protein Protocols Handbook, Humana Press Inc., Totowa, NJ. Klafki, H.W., J. Wiltfang, and M. Staufenbiel. 1996. Electrophoretic separation of BA4 peptides (1-40) and (1-42). Anal. Biochem. 237:24-29. Lavrenko, P.N., O.I. Mikriukova, and O.V. Okatova. 1987. On the separation ability of various ficoll gradient solutions in zonal centrifugation. Anal. Biochem. 166:287- 297. Maguire, D. and J. Kwan. 1992. Coneshell venoms-synthesis and packaging. pp. 11-18 in D. Watters, M. Lavin., D. Maguire, and J. Pearn, eds. Toxins and Targets, Harwood Academic Publishers. Mclntosh, J.M., A. Hasson, M.E. Spira, W.R. Gray, W. Li, M. Marsh, D.R. Hillyard, and B.M. Olivera. 1995. A new family of conotoxins that blocks voltage-gated sodium channels. J. Biol. Chem. 270:16796-16802. Morrissey, J.H. 1981. Silver stain for proteins in polyacrylamide gels: a modified procedure with enhanced uniform sensitivity. Anal. Biochem. 117:307-310. Ramilo, C.A., G.C. Zafaralla, L. Nadasdi, L.G. Hammerland, D. Yoshikami, W.R. Gray, R. Kristipati, J. Ramachandran, G. Miljanich, B.M. Olivera, and L.J. Cruz. 1992. Novel a- and o-conotoxins from Conus striatus venom. Biochem. 31:9919-9926. Yanagawa, Y., T. Abe, M. Satake, S. Odani, J. Suzuki, and K. Ishikawa. 1988. A novel sodium channel inhibitor from Conus geographus: purification, structure, and pharmacological properties. American Chem. Soc. 27:6256-6262. Figures Fig.1 A flow chart of the general protocol used in an attempt to purify the granules from Conus californicus. This chart focuses on the Ficoll-sucrose discontinuous, velocity sedimentation process because it appeared more efficient than the original sucrose gradient. The progression of the granules from the intact specimen to the Ficoll-sucrose pellet is emphasized by the red words. Fig. 2A Schematic representation of the Ficoll-sucrose gradient, which was based on a modified protocol by Cameron and Castle, 1984. The concentrations of sucrose and Ficoll are shown in the gradient. The volume utilized for each layer is shown on the outside. Fig.2B Schematic representation of the Ficoll-sucrose gradient after ultracentrifugation. There was a thin, white band under the top layer, which may contain very small granules. The Ficoll pellet, shown at the bottom of the tube, contained large, dense granules. Fig.3A Microscopic view at 60X of the granules in seawater. The granules are from the pellet of the Ficoll-sucrose gradient. They appear numerous and of many different sizes. Fig.3B Microscopic view at 60X of a clump of granules from the same sample. Many of the granules appeared bunched together under the microscope since it was difficult to suspend the pellet in seawater. Fig. 4A Scanning electron micrograph of several granules that are 3-4 um in diameter. Fig.4B Close up of the granule at the far left of Fig. 4A The surface appears very convoluted and misshapen due to the ethanol step-dehydration process necessary to fix the granules for scanning electron microscopy. Fig. 4C Micrograph of a venom duct granule. The granule appears very dense and is 6 um in diameter. Fig.4D Intact granules that withstood the dehydration process. Out of approximately 100 granules, only these two remained intact. The granules appear very dense with a smooth surface. Fig. Initial 10% SDS-PAGE acrylamide gel with results from sucrose discontinuous velocity sedimentation. Molecular weights (kDa) are listed to the left and the description of the sample is listed below each lane. The top arrow indicates the 49 kDa bands and the bottom arrow denotes the 25 kDa bands. This initial gel was used to determine the effectiveness of the sucrose gradient in isolating C. californicus granules and to approximate sample concentrations for future gels. Fig.6 SDS-Polyacrylamide tricine gel comparing results from the sucrose and Ficoll- sucrose discontinuent gradients. Both arrows denote the 5 kDa band. The gel curved down towards the right, so the 5 kDa band appears lower on the right side of the gel than on the left side. The 5 kDa band is present in lanes 2-5, 9, and 10. SDS-Polyacrylamide tricine gel comparing the components of the different Fig.7 fractions from the Ficoll-sucrose discontinuous gradient. The arrow denotes the 5 kDa band, which appears in lanes 2,3, and 5. Once again, the gel was curved inward and it is difficult to compare bands to the molecular weight standard without assessing the entire gel. Graph of the log of the molecular weight vs. distanced a band traveled in the Fig.8 SDS-Polyacrylamide tricine gel shown in Figure 7. This graph was used to compute the molecular weight of the granule band as 5 kDa. Fig.9 Tricine gel comparing the NP40 soluble and insoluble material within the pellet from the sucrose gradient. The arrow highlights the 5 kDa band. Fig. 10A Whole cell voltage-clamp recordings of disassociated cells from the stellate ganglion of Sepia officinalis. Shows the effects of 1:100 and 1:500 dilutions of Conus californicus whole duct venom on inactivating inward current. Voltage step and pulse durations are listed at the bottom. Fig. IOB This second trace illustrates the similar effects of 1:100 and 1:500 dilutions of the granule contents on inactivating inward current. Fig. 10C Recording demonstrating the effects of the 1:500 dilution of the granule contents on fast-inactivating inward current. Voltage step and pulse duration listed below the trace. Fig. 10D Recording illustrating the effects of the putative novel peptide from C. californicus on fast-inactivating inward current. Recording based on work by Wm. F. Gilly. 18 0 4 2 2 L — 0 + 5 L 0 O 0 L 0 L o L 7 0 L 6 5 . Q L 5 50 9 o5 8 8 kaa L 7 O () o 5o 80 L 5 5 5 8 0 L 1 0 0 L 60 ed ( 8 10