Detection of Sodium Channel Messenger RNA in the Stellate Ganglion, Optic Lobe, and Mantle Muscle of the Squid Loligo Opalescens ABSTRACT Different types of Na channels may be distinguished not only between species but also within individual organisms. Investigations into the differences between Na channel isoforms may be useful in elucidating mechanisms by which Na channels carry out their specific functional roles within cells. In this study, two RNA probes transcribed from segments of CDNA encoding a putative squid giant axon Na channel (GFLNI) were employed in in situ hybridizations and RNase protection assays to detec Na channel messenger RNAs in the stellate ganglion, optic lobe and mantle muscle of the squid Loligo opalescens. The results of this study indicate the presence of additional Na channel mRNAs with partial sequence homology to GFLNI mRNA in squid stellate ganglion and optic lobe, and provide evidence for the expression of Na channel mRNA in mantle muscle cells of L. opalescens. INTRODUCTION Na channels-the voltage-dependent ionic channels responsible for generation of action potentials-display remarkably similar Na currents across different phyla (Shaller et al, 1992). Despite this conservation in functional properties, important differences in toxin sensitivity, antibody labeling and primary structure have served to distinguish different types of Na channels not only between species but also within individual organisms (Shaller et al, 1992). Putative Na channels cloned and sequenced from the optic lobe (SOSR22/2-7; Sato & Matsumoto, 1992) and giant axon (GFLNI; Rosenthal & Gilly, 1993) of two presumably closely related species of squid (Loligo bleekeri and Loligo opalescens, respectively), for instance, display sequence differences in functionally important regions of the channel protein and little overall amino acid sequence identity (only 32%; Rosenthal & Gilly, 1993). Indeed, GELNI demonstrates higher identity with Na channels found in vertebrate tissues such as that isolated from electric eel electroplax (43%; Rosenthal & Gilly. 1993). Little is presently known about multiple Na channel subtypes in squid, but distinct Na channel species are likely to exist in different tissues. Thus, GFLNI is expressed only at extremely low levels in the optic lobe of L. opalescens (Rosenthal & Gilly, 1993), a complex region of the nervous system undoubtedly rich in Na channels. In addition, voltage-clamp expériments on L. opalescens in our laboratory have indicated the presence of Na channels with distinct kinetic properties in several types of non-giant-axon-forming squid neurons and in mantle muscle cells. These Na channels may have a different molecular identity from the GFLNI-type channels in the giant axon of the same species. Further investigation into molecular differences between these Na channels will be useful in elucidating the mechanisms by which Na channels fulfill their specific functional roles in cells of many species. In this study, two RNA probes transcribed from segments of CDNA encoding GFLNI were employed in in situ hybridizations and RNase protection assays to test for the presence of GFLNI-related Na channels mRNAs in the stellate ganglion sincluding giant fiber lobe (GFL)I, optic lobe and mantle muscle of L. opalescens. The results of these experiments indicate the existence of additional Na channels with partial sequence homology to GFLNI in the stellate ganglion and optic lobe of L. opalescens. and provide additional evidence for the presence of a similar type of Na channel mRNA in mantle muscle cells in this particular species of squid. MATERIALS AND METHODS Synthesis of RNA probes. CDNA encoding peptides 484-576 of the second half of interdomain I-Il of GFLNI was released from pNCSETIO (a vector created by Josh Rosenthal in our laboratory for fusion protein generation in the development of Na channel antibodies) with Hind III and Eco RI. This CDNA was then ligated into Bluescript II KS- phagemid (double digested with Hind III and Eco RI) to create pHMSI. pNZS¬ (encoding the 3'end of GFLNI mRNA; see Fig. 1A) and pNC5' (encoding the 5'end of GFLNI mRNA; see Fig. 1A) have been described previously (Rosenthal & Gilly, 1993). 358-labeled antisense HMSI (containing 282 nt coding and 69 nt vector sequence), antisense NZ5- (containing 135 nt coding, 102 nt untranslated and 30 nt vector sequence) and sense NCS (containing 161 nt untranslated and 143 nt vector sequence) RNA probes for in situ hybridizations were obtained from pHMSI, pNZS- and pNCS respectively. Each was transcribed (after linearization with the appropriate restriction enzymes; see Fig. 1A) with T7 RNA polymerase in the presence of 100 uCi of 13581 UTP by use of a kit (Riboprobe system; Promega). 32P labeled HMSI and NZ5- probes for RNase protection assay were similarly transcribed from these plasmids in the presence of 125 uG of 32PJ UTP. In situ hybridizations. Adult stellate ganglion (including GFL), adult optic lobe and 4-month-old mantle muscle tissues were dissected from live squid (collected from Monterey Bay) and fixed for 1 hour in fresh, ice cold 4% paraformaldehyde in artificial sea water (ASW). Fixed tissue samples were washed twice for 5 minutes in ASW, once for 5 minutes in a 1:1 dilution of ASW and PBS (pH 7.0), and once for 5 minutes in PBS (pH 7.0). Samples were soaked twice for 40 minutes in 20% sucrose in PBS at 4°C (to cryoprotect tissue from ice crystal damage during sectioning), frozen in Polyfreeze (Polysciences. Inc.) using dry ice, and stored at -70'C. Tissue was sectioned at 10 microns with a cryostat and affixed to silane coated glass slides (Sigma). In situ hybridizations with 358-labeled RNA probes were performed on the tissue sections using the protocol outlined by Simmons et al (1989). Following the hybridization procedures, slides were dipped in NBT-2 emulsion (melted and diluted 1:1 with dH20), allowed to dry vertically at room temperature for 3 hours, stored at 4’C in light-proof containers with desiccant for 12 days, and developed in Kodak D-19 for 4 minutes, dH20 stop bath for 15 seconds, and Kodak film strength rapid fixer for 4 minutes. Slides were then rinsed in running tap water for 1 hour, stained in 0.5% cresyl violet and mounted with Permount (Fisher Scientific). For each tissue type, one slide was stained with Multiple Stain (Toluidine Blue and Basic Fuchsin; Polysciences, Inc.) and mounted with Permount immediately after sectioning for comparison with the in situ hybridized slides. RNA extraction. Stellate ganglion (including GFL), optic lobe, mantle muscle and gill tissues were dissected from live adult squid (collected from Monterey Bay) and immediately frozen in liquid nitrogen and stored at 70C. Approximately equal amounts of each tissue were homogenized with à motorized tissue homogenizer (Brinkman, Inc.) into 2 ml of guanidine thyocyanate solution (2.6 M guanidine thyocyanate, 9.2 mM n¬ laurel sarcosine, and 12.5 mM sodium citrate). 1.5 ml of homogenate was layered over 550 ml of 5.7 M cesium chloride and centrifuged in a Beckman RP 55-S Swinging Bucket Rotor for 3 hours at 55,000 rpm and 23 C. The RNA pellet was resuspended in 100 ul of TE-1% SDS, extracted with phenol/chloroform, and ethanol precipitated. RNA concentrations were determined using a spectrophotometer, and a small aliquot of RNA from each tissue type was run on a 1% agarose gel containing ethidium bromide to ensure there was no degradation. RNAse protection assay. 105 cpm of 32P-labeled antisense HMSI probe were hybridized to 10 ug of each type of RNA in 3 ul of hybridization solution (4 M Nacl, 400 mM PIPES, and 10 mM EDTA) and 24 ul of deionized 100% formamide. In addition, 105 cpm of HMSI probe were hybridized to 10 ul of 10 mg/ml tRNA as a negative control. All hybridization mixtures were heated to 85C for 5 minutes and then incubated at 48°C overnight. RNase digestions were carried out by addition of 300 ul of RNase solution (300 mM NaCl, 5 mM EDTA, 10 mM Tris, 10 ug/ml RNase A, and 0.4 ug/ml RNase TI) to each sample and incubation at room temperature for 1 hour. The reaction was stopped by addition of 20 ul of 10% SDS and 2 ul of proteinase K (25 mg/ml) and incubation for 15 minutes at 37°C. Each sample was then phenol/chloroform extracted, ethanol precipitated using 25 ug of tRNA as a carrier, resuspended in S ul of loading buffer (80% deionized formamide, IXTBE, and 0.1% bromothymol blue), and run on a 6% denaturing acrylamide gel. 1000 cpm of undigested HMSI and NZ5- probes were also run as markers. Kodak XAR-S film using an intensifying screen was exposed to the gel overnight at -70°C. RESULTS In situ hybridizations. Stellate ganglion (including GFL) is a known site of GFLNI mRNA production (Rosenthal & Gilly, 1993) and was thus selected as a positive control for both probes. In the stellate ganglion sections, localization of probe was expected to be noted only in regions containing cell bodies and not in regions consisting of neuropil (Fig. 24). For both antisense probes (NZ5- and HMSI), this was seen to be the case (Fig. 2B-C). Sections hybridized with sense NC5' probe displayed very little background labeling (Fig. 2D). Although both probes had a similar labeling pattern in the stellate ganglion, an important difference in labeling between the two antisense probes was apparent. Whereas NZ5- labeled GFL and non-GFL cell bodies quite evenly (Fig. 2B), HMSI produced much stronger labeling in the non- GFL portion of the ganglion (Fig. 20). This suggests that HMSI may detect additional Na channel mRNAs not recognized by NZS¬ Probe-specific patterns of localized labeling were also evident in the optic lobe with both antisense probes. In contrast, only very slight and evenly distributed labeling by sense NC5' could be noted (Fig. 3B). NZS- was observed to exclusively label very few cells along the inner granular layer (Fig. 30). HMSI displayed a similar labeling pattern in the inner granular layer, but also reliably labeled cells in the outer granular laver and in the medulla (Fig. 3D-E). Thus, HMSI appears to label more cell types than does NZ5- in the optic lobe as well as in the stellate ganglion. In transverse sections of mantle muscle, labeling above background of the glass slide was not observed with sense NC5' (Fig. 44). In contrast. both NZ5- and HMSI exhibited labeling in the tissue sections above this background level, although labeling appeared to be evenly distributed and was not localized to any specific regions of the tissue (Fig. 4B-C). RNase Protection Assay. RNase protection assays were performed in which HMSI probe was hybridized with stellate ganglion (including GFL) optic lobe, gill, or mantle muscle RNA, and with control tRNA, and then digested with RNase. Both tRNA and gill (which is known to lack Na channels) RNA served as negative controls; a band was not noted for either of these. Three bands were observed for stellate ganglion; these included an apparently fully protected band of 282 nt, and two partially protected bands estimated to be 255 nt and 210 nt (Fig. 5). Bands were not noted for either optic lobe or mantle muscles (even after additional 3 day exposure; data not shown) These hybridization results suggest that, in addition to an mRNA with complete sequence identity to GFLNI mRNA in the second half of interdomain I-Il (indicated by the fully protected band), at least one additional mRNA species is expressed in the stellate ganglion that is highly homologous (but not identical) to GFLNI. DISCUSSION The most important finding in the present study is that the HMSI probe detects several types of putative Na channel mRNAs in stellate ganglion and optic lobe tissues. This is indicated by the less specific labeling of HMSI probe in in situ hybridizations (relative to that of NZS- and by the appearance of several partially protected bands in RNase protection assays employing HMSI probe. RNase protection assays conducted by Rosenthal & Gilly (1993) on stellate ganglion and giant fiber lobe RNA using antisense NZ5- probe detected only a single presumably fully protected band for either tissue. The highly specific nature of the NZ5- probe is not surprising, since it contains a large portion of 3 untranslated sequence for GFLNI. In contrast, HMSI probe consists entirely of coding sequence and may thus be less specific than NZ5- and capable of detecting Na channel mRNAs which are closely related to GFLNI in the interdomain I-Il region. This explanation would account for the differing results in the in situ hybridizations and RNase protection assays with the two probes. As noted in Figure 1B, HMSI probe corresponds to the second half of interdomain I-Il of the Na channel protein, whereas NZS- probe relates to the C-terminal end. Both of these regions lie in the cytoplasmic portions of the protein (Fig. 1B). Comparisons of protein sequences in a variety of isolated Na channels have shown that the cytoplasmic and extracellular portions tend to be the most variable regions of the protein (Schaller et al. 1992). The degree of variability in the interdomain I-Il region for mRNAs encoding squid Na channels requires further investigation, but the RNase protection assay results imply that one or two species must be identical to GFLNI over a sizable portion of this region. Na channel isoforms containing related but not identical amino acid sequences in interdomain I-Il could arise from transcription of separate genes, alternative splicing of mRNAs, or posttranslational modifications of the proteins themselves. In studies conducted on related Na channel mRNAs identified in rat brain and rat muscle, several forms were found to be the consequence of alternative splicing in a region corresponding to the second half of the interdomain I-Il portion of Na channel protein (Shaller et al, 1992). Alternative splicing in the first cytoplasmic loop of a Na channel mRNA has also been noted in Drosophila (Lougney et al, 1989). In addition to being the site of structural diversity, the interdomain l-Il loop is also the site of potential functional regulation. CAMP¬ dependent phosphorylation at multiple sites present in this region (Rossie et al, 1987) reduces peak Na currents (Li et al, 1993). The possibility thus emerges that alternative splicing in this region could selectively express sites that act to regulate functional aspects of channels and thereby cellular excitability. In addition, it has been postulated that variations in cytoplasmic domains such as the interdomain I-Il loop could play a role in specifying the localization of Na channels to different areas within cells (Schaller et al, 1992). The Na channels related to GFLNI found in this study in the L. opalescens stellate ganglion and optic lobe show considerable sequence homology in interdomain I-Il and may also result from RNA splicing in this region. Again, changes in the amino acid sequence in this cytoplasmic loop may endow Na channels with specialized properties within each tissue type. Clearly attributing a specific role to each Na channel isoform based on the results of this study is impossible. It is likely, however, that at least one of the putative Na channel mRNAs detected in the stellate ganglion and giant fiber lobe using GFLNI related probes corresponds to the much studied Na channel of the giant axon. The larger cells in the inner and outer granular layer regions of the optic lobe have distinctive morphologies and patterns of positioning and are thought to be involved in early-stage sensory integration within the optic lobe, whereas the more poorly characterized, smaller cells in the medulla carry out additional integration and transmit outputs to other portions of the nervous system (Young, 1974). Distinct labeling was observed with both antisense probes only in the inner granular layer of the optic lobe in large cells spaced quite far apart. This result indicates that a putative Na channel mRNA identical to GFLNI may be present in these cells. This conclusion is further supported by the RNase protection assay results of Rosenthal & Gilly (1993) noting a single fully protected band of extremely low abundance in optic lobe RNA using NZ5- probe. Additional Na channel mRNAs in optic lobe detected by HMSI probe appear to be more abundant and may relate to the SOSR22/2-7 Na channel CDNA isolated in optic lobe of a similar species of squid (Sato & Matsumoto, 1992). The absence of bands relating to optic lobe Na channel mRNAs in the RNase protection assay using HMSI probe can be attributed to a low ratio of Na channel mRNA species to total RNA. A longer exposure time in the RNase protection assay is most likely necessary to detect signal for Na channel mRNA in this tissue. In in situ hybridizations carried out on mantle muscle tissue, both antisense probes demonstrated labeling above the background level produced by the sense probe. This observation suggests that Na channel mRNA related to GFLNI mRNA may be expressed in mantle muscle cells. Localization of labeling to one of the two zones of identified circular muscle fiber types of mantle muscle (Bone et al, 1981) was not noted, however, and only diffuse labeling by either antisense probe in all portions of tissue could be seen. This may be due to poor resolution in the detection of 32P signal using emulsion coating of the slides. Additional in situ hybridizations of mantle muscle employing a higher resolution technique may be necessary to detect localization of probe within mantle muscle tissue. As with optic lobe, the abundance of Na channel mRNA in mantle muscle tissue is probably very low, and lengthy exposures are most likely required in order to detect the presence of Na channel mRNA in mantle muscle tissue using the RNase protection assay. ACKNOWLEDGMENTS 1 would like to express my sincere gratitude to Professor Gilly for his encouragement and guidance in focusing my ideas in my project throughout the entire quarter. I owe an especially big thank you to Taylor Liu for unselfishly devoting numerous hours helping me step by step through techniques and for always providing great company and being an invaluable source of advice. I am greatly indebted to Marie Perri for also giving up much of her time helping me with the RNase protection assays and for her cheerful companionship in the lab. I would like to thank Matt McFarlane and Josh Rosenthal for their great senses of humor and always ready willingness to help me out. Lastly, I would like to thank Simone, Tony, Zora, Adam and Don for helping make my experience in the Gilly lab à fun and memorable one. LITERATURE CITED Bone, Q., Pulsford, A., & Chubb, A.D. (1981) Squid mantle muscle. J. mar biol Ass. U.K. 61, 327-342. Li, M., West, J.W., Numann, R., Murphy, B.J., Scheuer, T., & Catterall, W.A. (1993) Convergent regulation of sodium channels by protein kinase C and CAMP-dependent protein kinase. Science 261, 1439-1442. Lougney, K., Kreber, R., & Ganetzky, B. (1989) Molecular analysis of the para locus, à sodium channel gene in Drosophila. Cell 5 8, 1143-1154. Rosenthal, J., & Gilly, W. (1993) Amino acid sequence of a putative sodium channel expressed in the giant axon of the squid Loligo opalescens. Proc. Natl. Acad. Sci. USA 90, 10026-100030. Rossie, S., Gordon, D., & Catterall, W.A. (1987) Identification of an intracellular domain of the sodium channel having multiple CAMP¬ dependent phosphorylation sites. J. Biol. Chem. 262, 17530-17535. Sato, C., & Matsumoto, G. (1992) Primary structure of squid sodium channel deduced from the complementary DNA sequence. Biochem. Biophys. Res. Commun. 186, 61-68. Shaller, K.L., Krezmien, D.M., Mckenna, N.M., & Caldwell, J.H. (1992) Alternatively spliced sodium channel transcripts in brain and muscle. J. Neurosci. 11, 918-927. Simmons, D.M., Arriza, J.L., & Swanson, L.W. (1989) A complete protocol for in situ hybridization of messenger RNAs in brain and other tissues with radiolabeled single stranded RNA probe. JHistotech. 1 2(3), 169- 181. Young, T.Z. (1974) The central nervous system of Loligo. Phil. Trans. of Roy. Soc. of London 267, 268. LEGEND TO FIGURES Figure 1. A. GFLNI and related CDNAs employed in synthesizing probe The approximate locations of Nde I, Hind III and Nco I restriction enzyme sites used in linearizing pNZ5-, pHMS1, and pNCS'are shown. Figures are not drawn to scale. B. Schematic representation of the regions of GFLNI to which each probe corresponds. Note that all three probes relate to cytoplasmic portions of the protein. Figure 2. All photos are taken at 4X magnification. A. 10 micron thick section of adult L. opalescens stellate ganglion stained with Multiple Stain (Toluidine Blue and Basic Fuchsin; Polysciences, Inc.). The section consists of both the giant fiber lobe (I) and stellate ganglion (11). B. Adult stellate ganglion section hybridized with NZS¬ antisense probe. Diffuse staining in the cell body regions of the giant fiber lobe and stellate ganglion (*) may be noted. C. Adult stellate ganglion section hybridized with HMSI antisense probe. Diffuse staining in the giant fiber lobe (1) and heavy staining in the stellate ganglion (II) are seen. D. Adult stellate ganglion section hybridized with NC5' sense (negative control) probe; labeling is not apparent. Figure 3. Photos A, B, C and D are taken at 10X magnification; photo E at 20X magnification. A. 10 micron thick section of adult L. opalescens optic lobe stained with Multiple Stain (Toluidine Blue and Basic Fuchsin; Polysciences, Inc.). Dark stained regions include the outer granular layer (1), inner granular layer (II), and medulla (III). B. Adult optic lobe section hybridized with NC5' sense probe only diffuse background labeling can be seen. C. Adult optic lobe section hybridized with NZ5- antisense probe; the probe labeled localized regions along the inner granular layer (*). D. Adult optic lobe section hybridized with HMSI antisense probe. Localized labeling is apparent along the inner granular layer (1), outer granular layer (II), and in the medulla (III). E. Magnified view of localized labeling by HMSI probe in the medulla of an adult optic lobe section. Figure 4. All photos are taken at 40X magnification. A. 4-month-old mantle muscle section (transverse) hybridized with NC5' sense probe. Labeling is not observed above the background labeling of the glass slide. B. 4-month-old mantle muscle section (transverse hybridized with NZ5- antisense probe. Diffuse labeling in the tissue is notable above background labeling. C. 4-month-old mantle muscle section (transverse) hybridized with HMSI antisense probe. Diffuse labeling in the tissue above background levels is apparent. Figure 5. Results from RNase protection assays on extracted RNAs using 52P-labeled HMSI probe (overnight exposure). Samples run include undigested HMSI probe (A), undigested NZ5- probe (B), negative control tRNA (C), stellate ganglion RNA (D), mantle muscle RNA (E) optic lobe RNA (F), and negative control gill RNA. Bands are noted only for undigested HMSI probe, undigested NZ5- probe, and for stellate ganglion RNA. The three bands corresponding to stellate ganglion consist of a presumably fully protected band (282 nt) and two partially protected bands (approximately 255 nt and 210 nt). The lower mobility of the fully protected band relative to the undigested HMSI probe is due to removal of 69 nt of vector sequence in the probe. Figure 1 GFLNI — WA NZ5 — HI HMSI Ncol NCS Vuntranslated sequence coding sequence — vector sequence B. — — oufside inside 5 PHRS! Ozawa/Na Channels VA NdeI EcoRI —17 0 + —— +- N25- 16 Figure 2 A. Ozawa/Na Channels Figure 2 continued C. Ozawa/Na Channels Figure 3 A. Ozawa/Na Channels III B. Figure 3 continued Ozawa/Na Channels 20 Figure 3 continued E. Ozawa/Na Channels Figure 4 A. Ozawa/Na Channels 22 Figure 4 continued C. Ozawa/Na Channels Figure 5 Özawa/Na Channels EF