association of neuroactive peptides with the protein secretory pathway in identified neurons of...

THE JOURNAL OF COMPARATIVE NEUROLOGY 272~358-369 (1988)

Association of Neuroactive Peptides With the Protein Secretory Pathway in

Identified Neurons of Aplysia californica: Immunolocalization of SCPA and SCPB to the Contents of

Dense-Core Vesicles and the Trans Face of the Golgi Apparatus

WILLIAM REED, KLAUDIUSZ R. WEISS, PHILIP E. LLOYD, IRVING KUPFERMA", MARY CHEN, AND CRAIG H. BAILEY

Center for Neurobiology and Behavior, Columbia University and the New York State Psychiatric Institute, College of Physicians and Surgeons, New York, New York 10032

ABSTRACT The subcellular distribution of two molluscan neuropeptides, the small

cardioactive peptides A and B (SCPA and SCPB), has been determined in two identified Aplysia buccal ganglion neurons, B1 and B2. These neurons were previously shown to synthesize and release these neuropeptides. B1 and B2, identified by their size and location within the ganglion, were labeled by intrasomatic injection of an electron-dense particulate marker (ferritin or Imposil) permitting the unequivocal identification of their somata and proximal processes in thin sections. The somatic cytoplasm of both neurons had a conspicuous population of large dense-core vesicles along with a smaller number of compound vesicles and small lucent vesicles. All three vesicle types are found in the neurites within the neuropil and proximal axon in the esophageal nerve. Immunoreactivity was localized on the surface of thin sections by the indirect immunogold method. The primary antiserum was shown to recognize both SCPA and SCPB after the neuropeptides had been immobilized on protein-coated nitrocellulose membranes by means of glutaraldehyde, the primary fixative used to immobilize SCPA and SCPB in situ. SCP immunoreactivity was present in the lumens of the dense-core vesicles distributed throughout the cytoplasm of B1 and B2 and in dense-core regions of the Golgi apparatus in the somatic cytoplasm. Taken together with biochemical evidence that B1 and B2 synthesize and release SCPs, these data suggest that the neuropeptides are sequestered into the protein secretory pathway of B1 and B2, a distribution that supports the notion that the SCPs function physiologically as neurotransmitters or neuromodulators.

Key words: small cardioactive peptides, FMRFamide, neuropeptides, EM localization

Accepted December 11, 1987. matology and Immunology, 932 Faculty Laboratory Office Build- ing 231H. ChaDel Hill. NC 27514.

Address reprint requests to Dr. Craig Bailey, Center for Neuro- biology and Behavior, Columbia University and the New York State Psychiatric Institute, College of Physicians and Surgeons, 722 West 168th Street, New York, NY 10032.

William Reed is now at the University of North Carolina at Chapel Hill School of Medicine, Dept. of Medicine, Div. of Rheu-

0 1988 ALAN R. LISS, INC.

Philip Lloydis now at the University of Chicago, Dept. of Phar- macology and Physiological Sciences, 947 E. 58th Street, Chicago, IL 60637.

IMMUNOLOCALIZATION OF SCPA AND SCPB 359

The small cardioactive peptides A and B (SCPA and SCPB) are two closely related molluscan neuropeptides that exert potent modulatory effects both centrally (Abrams et al., '84) and peripherally (Lloyd et al., '84). The neuropeptides have been isolated from Aplysia californica and sequenced by Lloyd and collaborators (Morris et al., '82; Lloyd et al., '84). A complementary DNA clone encoding a precursor containing the primary sequence of both neuropeptides has been isolated (Mahon et al., '85). The precursor also codes for a leading hydrophobic sequence of amino acids, a putative signal sequence, that is presumed to direct the sequestration of the precursor into the membrane compartments of the protein secretory pathway in eucaryotic cells.

Biochemical fractionation followed by bioassay has shown that SCPA and SCPB are present in Aplysia central and peripheral nervous tissue (Lloyd et al., '84; Lloyd et al., '85b). Moreover, SCPB immunoreactivity has been localized to a specifiable subset of neuronal cell bodies and their central and peripheral processes by immunofluorescence microscopy by use of an antiserum raised to synthetic SCPB conjugated to bovine serum albumin (Lloyd et al., '85b; Mahon et al., '85). Particularly intense SCPB immunoreactivity was localized to two large identified neurons of the buccal ganglion, B1 and B2, which send major processes to the periphery by way of the esophageal nerve. B1 and B2 synthesize large quantities of SCPA and SCPB (Lloyd et al., '85b) and release these neuropeptides in a CaZf-dependent fashion in response to intracellular stimulation (Lloyd et al., '86).

In an earlier study employing immunoelectron microscopy, Kreiner et al. ('86) noted SCPB immunoreactivity associated with large somatic dense-core vesicles (DCV) in buccal ganglion neurons presumed to be B1 and B2. "he subcellular distribution of SCPA and SCPB in B1 and B2 has been examined here by immunoelectron microscopy combined with the intracellular injection of electron-dense compounds that permit the unequivocal identification of B1 and B2 somata and their processes. SCP immunoreactivity (iSCP) is localized to the contents of somatic dense-core vesicles as well as dense-core vesicles forming on the trans face of the Golgi apparatus. Dense-core vesicles containing iSCP are also found in the proximal processes of B1 and B2 in the buccal ganglion neuropil and in the peripherally targeted axons of the neurons. Taken together with the biochemical evidence, these data suggest that SCPA and SCPB are sequestered in secretory storage vesicles and other membrane-delimited compartments of the protein secretory pathway in B1 and B2.

MATERIALS AND METHODS Intrasomatic labeling

Ferritin (EM grade, obtained from Polysciences) and Im- posil (kindly provided under its trade name, NONEMIC, by Dr. R. Lynn of Schering Corp.) were dialyzed at 4°C for 5- 15 hours into a cytoplasmiclike buffer (350 mM KC1, 50 mM NaC1,2 mM MgC1, and 1 mM MOPS-NaOH pH 7.2) to remove saline from both and phenol from the Imposil (Granzow et al., '85). Buccal ganglia from 50 to 150 g animals were excised, pinned to Sylgard coated petri dishes containing artificial seawater and desheathed (i.e., the muscle and connective tissue overlying the ventral surface of the ganglion was carefully peeled away). B1 or B2 were identified by their size and position (Gardner, '71) in the ganglion and labeled by intrasomatic pressure injection of Imposil or ferritin according to the method originally de-

scribed by Eisenstadt et al. ('73). Ganglia were incubated 1-18 hours before they were fixed, permitting the spread of label into the injected cell's proximal processes; 1-2-hour incubations were done at room temperature in artificial seawater. Longer incubations were done at 14°C in hemo- lymph.

Electron microscopy After desheathing, intrasomatic injection, and incuba-

tion, all the ganglia were fixed in the primary fixative (6% glutaraldehyde, 0.54 M sucrose, and 0.2 M sym-collidine- HC1, pH 7.3) for 2 to 4 hours. Some of the ganglia were dehydrated in a graded series of methanol (50%, 70%, 90%) at -15"C, and embedded in LR White resin (from E. Ful- lam). The resin was cold cured at 0°C under vacuum. Alter- natively, ganglia were postfixed in 2% osmium tetraoxide in 0.05 M phosphate buffer (pH 7.2) for 20 minutes, dehydrated in a graded series of alcohol, and embedded in Embed 812 (from Polysciences). Glutaraldehyde and osmium tetraoxide were obtained from Polysciences.

Silver-gold sections were cut and mounted on bare 200- mesh nickel grids or 0.35% formvar coated, carbon stabi- lized, nickel slot grids. Sections were examined and photo- graphed on a JEOL 100 CX electron microscope operating at an accelerating voltage of 80 kV.

Peptide dot blotting Nitrocellulose membranes were coated with normal goat

serum (NGS, obtained from Cappel) as follows. Nitrocellu- lose paper (from Biorad Laboratories) was immersed in 4% NGS in Tris-buffered saline (TBS, 0.9% NaC1,20 mM Tris- HC1 pH 8.2) for 10 minutes, then removed, air dried, rinsed thoroughly in TBS, dried again, and stored at room temperature until used.

Ten nmoles of synthetic peptide (Peninsula Laboratories) was dissolved in blotting buffer (0.5% glutaraldehyde in 0.1 M glacial acetic acid) and diluted in a serial fashion with the same buffer. The peptide dilutions were blotted in 1 pl aliquots onto the NGS-coated nitrocellulose paper and air dried. The blots were incubated in blocking buffer (10% ethanolamine, from Sigma; 3% gelatin, from Norland Prod- ucts; 0.05 M Tris-HC1, pH 7.6) for 1-2 hours and rinsed in TBS. At this point, the membrane could be dried and stored at room temperature for subsequent immunostaining.

The blotted and blocked nitrocellulose membranes were probed for the presence of peptide during a 2-hour incubation in primary antibody (1:1600 anti-SCPB; 1:lOOO anti- Phe-Met-Arg-Phe-amide [FMRFa], obtained from Penin- sula Labs), followed by 30 minutes in a 1:40 dilution of goat antirabbit (IgG) secondary antibody (Cappel), and 1 hour in a 1:lOO dilution of peroxidase-antiperoxidase (PAP, from Cappel). The peroxidase was developed during a 30-second to 2-minute incubation in 0.05% diaminobenzidine, 0.6% hydrogen peroxide (DAB-HzOz). The primary, secondary, and PAP were diluted in 4% NGS-TBS, and DAB-HzO2 was dissolved in 0.9% NaC1,0.05 M Tris-HC1, pH 7.6.

Indirect immunolocalization Indirect immunogold localization on thin sections was

performed according to a method modified from Bendayan and Zollinger ('83). A grid bearing a section to be labeled was placed (section side down) onto a drop of saturated sodium metaperiodate (NaI04) for 75 minutes. After rins- ing in distilled-deionized water and blotting away excess fluid, the grid was placed for 15 minutes onto a drop of 8%

W. REED ET AL. 360

NGS-TBS, then dipped briefly in TBS, blotted, and placed onto a drop of primary antiserum diluted in 4% NGS-TBS (1:250 anti-SCPB; 1:200 anti-FMRFamide). After 12-16 hours at 4"C, the grid was rinsed in TBS, blotted, and placed for 30 minutes on a drop of a 1:3 dilution (in 4% NGS-TBS) of a suspension of goat antirabbit (IgG) secondary antibody adsorbed to 15 nm colloidal gold particles (SPI Supplies). Finally, the grid was rinsed in TBS, blotted, and set aside to dry before counterstaining with uranyl acetate (30 seconds) and Reynold's lead citrate (30 seconds).

Quantitation of immunolocalization Random, immunolabeled thin sections of B1 and B2 cell

somata were sampled by taking micrographs at regularly spaced intervals. Enlargements (98,000 X) of the micrographs were mounted on a Bioquant I1 digitizing tablet (R&M Biometrics, Inc., Nashville, TN) that is interfaced with an Apple IIe microcomputer running the Bioquant I1 morphometry program. The surface areas subtended by cell structures were determined by the Bioquant I1 software on the basis of a digitized tracing of the outline of the structure. The number of gold particles associated with the measured structures was counted and labeling was expressed as gold particles per unit surface area. Background labeling was estimated by measuring gold particle density over extracellular regions adjacent to the cell soma.

RESULTS Buccal ganglion neurons B1 and B2 were identified visu-

ally, by size and location within the ganglion, and by their electrophysiological properties. Before fixation and embedding, they were labeled by intrasomatic injection of a particulate electron-dense marker. This permitted unequivocal identification of the neurons and their processes in thin sections. Two iron-containing markers, Imposil and ferritin, were used. B1 was labeled with Imposil, a rod-shaped iron- dextran conjugate 25 nm long and 10 nm in diameter (Fig. la, inset). B2 was labeled with ferritin, a protein with a 5.5 nm spherical iron hydroxide core (Fig. lb, inset). The effect of intrasomatic labeling on the ultrastructure of B1 and B2 was evaluated by comparing injected somata to a few un- injected B1 and B2 that could be identified in their embed- ment with considerable certainty. Except for some minor fragmentation of the rough endoplasmic reticulum, somatic architecture appeared unaltered by intrasomatic labeling.

B1 and B2: a conspicuous population of dense-core vesicles

An overview of the somatic cytoplasm of B1 and B2 (Fig. 1) reveals ultrastructure that is typical of the somata of Aplysia neurons except for a conspicuous population of large, dense-core vesicles. The population of somatic dense- core vesicles in B1 and B2 is conspicuous in part because of the high electron density of the vesicle core, but also because their concentration (dense-core vesicles per unit cy- toplasmic volume) is relatively large. Indeed, B1 and B2 somata have a significantly higher concentration of somatic dense-core vesicles than the majority of neurons in the buccal ganglion. Within a given cell, the concentration of dense-core vesicles varies considerably from point to point. Although the difference between the concentrations of

dense-core vesicles in B1 and B2 in Figure 1 appears large, it is typical of the variation observed within an individual cell. B1 and B2 have been indistinguishable on the basis of qualitative and prelimary quantitative estimates of somatic dense-core vesicle concentration.

B1 and B2 have well-developed Golgi apparatuses (G, Fig. lA,B). Each Golgi apparatus is associated with a group of dense-core vesicles, and many Golgi apparatuses have local, dense-core focal dilations that are putative dense-core vesicle precursors (see Fig. 11).

Closer scrutiny of the somatic cytoplasm (Fig. 2) reveals the presence of other somatic vesicle populations. In addition to the dense-core vesicles, each cell assembles small, relatively electron-lucent vesicles (large arrowheads, Fig. 2a,b) as well as compound vesicles that have the appear- ance of a vesicle within a vesicle. The compound vesicles are similar to those described by Shkolnik and Schwartz ('80) in serotonergic neurons in Aplysia.

The size and shape of the dense-core vesicle profiles observed in thin section were estimated using the Bioquant I1 morphometry software which, given a digitized tracing of the vesicle perimeter, will determine the longest dimension of the profile, the orientation of the longest dimension, and a shape factor. The shape factor is 1 for circular profiles and decreases with increasing noncircularity.

The dense-core vesicles in B1 and B2 are bounded by a single unit membrane that is separated from the electron- dense core by a slender lucent zone that varies in width from 5 to 15 nm (Fig. 2). The vesicle profiles are often noncircular with long dimensions that vary between 60 and 150 nm. A histogram of vesicle profile long dimension is multimodal with a major peak around 88 nm, one or two major peaks around 110 nm, and a minor peak around 135 nm.

The shape factors computed for the measured profiles varied between 0.45 and 1, with 0.8 being the most proba- ble. Only 13% of the profiles measured were circular. More- over, the long dimensions of the profiles within a given section were of various orientations suggesting that the noncircularity of the profiles is not a compression artifact introduced during sectioning. Thus the dense-core vesicles in B1 and B2 are apparently nonspherical, because spherical vesicles should yield a high percentage of circular profiles. The nonspherical nature of the dense-core vesicle and the fact that its dimensions are greater than a section thickness greatly complicates estimation of dense-core vesicle size and shape from the vesicle profiles.

Fig. 1. Overview of the somatic cytoplasm of buccal neurons B1 (a) and B2 (b). Both B1 and B2 have a large nucleus, numerous mitochondria, large end stage lysosomes, rough endoplasmic reticulum, a Golgi apparatus (GI, and a heterogeneous vesicle population. The inscribed regions in (a) and (b) are shown in greater detail in the insets, revealing the presence of the intracellular label that confirms the identity of the cell. B1 was labeled with Imposil, a rod-shaped marker (a, inset) and B2 was labeled with ferritin, a spherical marker (b, inset). (a) and (b), 25,200~; scale, 0.5 pm; insets 114,000x; scale, 0.11 pm.


Figure I

362 W. REED ET AL.

SCPA ala - arg - pro - gly - tyr- leu - ala - phe - pro - arg - met - NH3

SCPB met-asn-tyr- leu - ala - phe - pro- arg- met - NH3

leu - ala - phe - pro - arg - met - NH3 SCPHeY.

FMRFa phe - met-arg - phe - NH3

Fig. 3. Peptide dot blot characterization of the SCP and FMRFamide dehyde-fixed SCPA, SCPB, and a six-amino acid peptide common to both antisera specificity. Various quantities (0-25 pmol) of the peptides, as indi- (SCP hexapeptide, see bottom panel), but does not bind FMRFa (top, left cated in the top panels, were immobilized with glutaraldehyde onto NGS- panel). The FMRFa antiserum has the opposite specificity (top, right panel). coated nitrocellulose membranes. The SCP antiserum recognizes glutaral- The sequences of the immobilized peptides are shown in bottom panels.

IMMUNOLOCALIZATION OF SCPA AND SCPs 363

Fig. 4. Indirect immunogold localization of SCP immunoreactivity within somatic cytoplasm of BUa) and B2(b). Preabsorption of the antiserum with

100 pm SCPB inhibited dense-core vesicle labeling (c) and (d). (a) and (c), 108,OOOX; scale, 0.19 pm; 6) and (d), 204,000~; scale, 0.1 pm.

364

SCPB antiserum recognizes glutaraldehyde- immobilized SCPA and SCPB

The small cardioactive peptides, immobilized by glutaraldehyde fixation in situ, were localized on thin sections by the indirect immunogold method (see Methods) by using a polyclonal antiserum raised to synthetic SCPB conjugated to BSA (Mahon et al., '85). The same antiserum had been employed previously in the immunofluorescent localization of SCPB immunoreactivity (iSCPB) in Aplysia central (Lloyd et al., '85b; Mahon et al., '85) and peripheral (Lloyd et al., '84) nervous system. The specificity of the antiserum was explored by peptide dot blot analysis (Fig. 3). Glutaralde- hyde immobilized SCPA, SCPB, and SCP-hexapeptide, a six- amino acid fragment common to the carboxy termini of both neuropeptides (Lloyd et al., '85a), immunoreacted with the SCPB antiserum, whereas immobilized FMRFa, an- other molluscan cardioexcitatory peptide found throughout the nervous system of Aplysia, was not immunoreactive (Fig. 3, left panel). A commercially available FMRFa antiserum, used as a control for the specificity of iSCP localization (Fig. 3, right panel), reacted with FMFtFa but not with any of the SCPs. This antiserum has also been used previously in the fluorescent localization of FMRFa immunoreactivity in Aplysia buccal ganglion and muscles (Lloyd et al., '87; Weiss et al., '87).

SCP immunoreactivity associated with the luminal contents of the large, dense-core vesicles in

B1 and B2 When thin sections were treated with the SCP rabbit

antiserum followed by goat anti-rabbit (IgG) adsorbed to 15 nm colloidal gold particles, a significant portion of gold label was associated with the population of dense-core vesicles in B1 and B2 (Fig. 4a,b). Compound and small luccent vesicles were not labeled (Fig. 4a). Preabsorption of the antiserum with 100 pM SCPB inhibited dense-core vesicle labeling (Fig. 4c,d) in a dose-dependent fashion (not shown).

Figure 5 shows a region of the somatic cytoplasm of B2 after fixation in glutaraldehyde alone and embedded in LR White, a polar embedding resin that improves antigen pres- ervation (Newman et al., '83). The large dense-core vesicles seen in conventionally prepared thin sections (Figs. 1,2,4) are easily identified and are heavily labeled after indirect immunogold localization of iSCP. Thus, the localization of iSCP to the somatic dense-core vesicles revealed on osmium-fixed ganglia embedded in a more hydrophobic resin (Embed 812) was confirmed by using alternative fixation and embedding procedures.

A comprehensive view of iSCP localization to the dense- core vesicles in B1 and B2 is presented in Figure 6. Immu- noreactivity was quantified by determining the concentration of gold particles associated with a given cell structure (gold particles per unit area of thin section subtended by the structure). Measurements were performed on five identified cells (3 B1 and 2 B2) treated with the antiserum alone (solid bars, Fig. 6) or antiserum preabsorbed with 100 pM synthetic SCPB (open bars, Fig. 6). The concentration of gold particles on structure-free extracellular regions of thin sections was taken as a measure of background; that is, gold label associated nonspecifically with the section (Fig. 6, extracellular space). Background was independent of the presence of preabsorbing antigen. Gold particle concentrations associated with nuclei and mitochondria were not above background in B1 and B2, whereas dense-core vesicle label was nearly ten times background (solid bars, Fig. 6).

W. REED ET AL.

Fig. 5. Immunogold localization of SCP immunoreactivity within B2's soma after alternative fixation and embedding in the polar resin LR white, in the absence of osmium fixation. 70,000~; scale, 0.1 pm.

Dense-core Nucleus Mttochondria Extracellular vesicles space

Fig. 6. Quantitative analysis of immunogold label in B1 and B2 somatic cytoplasm. The concentration of gold particles associated with densecore vesicles, nucleus and mitochondria in B1 and B2, as well as with the extracellular space surrounding the cells, was determined from thin sections probed with the SCP antiserum alone (solid bars) or preabsorbed with 10 p M SCPB (open bars). Dense-core vesicles are heavily labeled compared to the nucleus, mitochondria and extracellular space. As suggested by Figure 4c,d, preabsorption with SCPB strongly inhibits immunogold localization on the dense-core vesicles. Each value represents the average * SEM gold particle concentration observed in five cells (3 B1 and 2 B2). Gold particle concentration was defined as the number of particles associated with a particular region or structure per unit of thin section surface area subtended by the region or structure. Total areas measured for experimen- tals were: 3.7 pm2 (534 dense-core vesicles), 68.55 pm2 (nuclei), 3.2 pm2 (mitochondria), and 85.0 m2 (extracellular space). Total areas measured for controls were: 2.4 pm2 (331 dense-core vesicles), 5.16 pm2 (nucleus), 1.78 pm2 (mitochondria), and 27.7 pm2 (extracellular space).


the identification of the proximal processes of B1 and B2 in the neuropil (e.g., Fig. 8; see also Fig. 10). More distal processes were also identified, such as the major axons of B1 and B2 as they exit the buccal ganglion via the esophageal nerve (e.g., Fig. 9). In many cases, these processes contained dense-core vesicles that were morphologically similar to the dense-core vesicles in the cell soma. In every case, the dense-core vesicles in these identified processes were SCP-immunoreactive.

Localization of SCP immunoreactivity is cell specific

The specificity of the iSCP localization is further examined in Figure 10, which shows a selected region of neuropil in two serial sections. Among the neurites in this region is an identified process of B1 filled with iSCP dense-core vesicles (Bl, Fig. 1Oa) and above B1, an unidentified process (**, Fig. 10a) containing a population of dense-core vesicles that are morphologically similar to those in B1 below but are not SCP-immunoreactive. An adjacent section stained for FMRFa revealed that although the densecore vesicles in Bl’s process are SCP-immunoreactive they are not FMRFa-immunoreactive (Fig. lob). By contrast, the dense- core vesicles in the unidentified process have the converse immunoreactivity. The mutually exclusive localization of iSCP and iFMRFa demonstrates that the immunolocalization is highly cell specific. Several neurons in the buccal ganglion have been shown to contain both SCPs and FMRFa (Lloyd et al., ’78). However, most neurons, including both B1 and B2, are immunoreactive for either the SCPs or FMRFa. Moreover, this finding demonstrates that vesicles with similar morphology do not necessarily contain the same peptides.

Fig. 7. Localization of X p imunoreactivity to the contents ofthe dense. core vesicle. One surface of a thin section was probed for the presence of iSCP by using the indirect immunogold localization technique. The section was embedded and resectioned in a plane perpendicular to the original plane of section. The resectioned view in (b) shows that the lumenal contents of a dense-core vesicle were exposed on the probed surface of the original thin section and that the contents of the dense-core vesicle are immunoreactive for SCP. The pattern of labeling of the exposed contents is similar to the labeling seen in a different densecore vesicle shown for comparison in normal section in (a) above. A bracket depicts the thickness of the original thin section that was probed for iSCP and an open arrow indicates the Formvar support. (a) and (b), 192,000~; scale, 0.08 pm.

SCp immunoreactivity associated with densemcore Of the apparatus in B1 and B2

me popu~ation of dense-core vesicles in ~1 and B2 is presumably a distal denlent of the Protein secretory path- Way in these neurons. Consequently, the more proximal elements of this pathway, the Golgi apparatus and the rough endoplasmic reticulum (RER) were examined for the presence of iSCP. Gold label was consistently associated with the dense-core compartments of the Golgi apparatus,

fieabsorption ofthe antiserum with 100 CLM SCpB reduced especially the focal dense-core dilations on the trans face the gold particle concentration on dense-core vesicles to that appear by morphological criteria to be condensing background levels. The position of the gold label relative to dense-core vesicles (Fig. 11). the contents of &-+core vesicles was explored by re- The frequency and concentration Of the Golgi-associated embedding immunolabeled thin sections and resectioning label were qualitatively similar to the labeling of dense- in a plane perpendicular to the labeled section. The set- core vesicles, and were inhibited by preabsorption of the tioned profiles of the dense-core vesicles could easily be antiserum (not Shown). Other elements of the Proximal identified. The lumenal contents of immunolabeled dense- secretory Pathway in €31 and B2, the RER, and the eledron- core vesicles were invariably exposed on the treated surface lucent compartments ofthe apparatus were not con- of the section (e.g., Fig. 7b) suggesting that the SCP-like sistently SCP-immunoreactive.

DISCUSSION antigens are associated with the luminal contents of the dense-core vesicles. Unlabeled vesicles with exposed contents were present; however, increasing the SCP antiserum SCP immunoreactivity has been localized to a population concentration above the standard concentration (1:250) in- of large dense-core vesicles in B1 and B2, two identified creased the percentage of labeled vesicles, suggesting that Aplysia neurons that synthesize and secrete the neuropep- immunolabeling at the standard concentration was tides. Kreiner et al. (‘86) have also reported the association submaximal. of SCPR immunoreactivity with somatic dense-core vesi-

cles, using different fixation and embedding protocols. We have further demonstrated that iSCP is also present on the trans face of the Golgi atmaratus. and in dense-core vesicles

scp immunoreactivity present in dense-core in the proximal processes Of B1 and B2

Y I.

Intracellular injection of Imposil and ferritin permitted in the central and proximal peripheral processes of B1 and

366 W. REED ET AL.

Fig. 8. Indirect immunogold localization of SCP immunoreactivity to the dense-core vesicles in an identified B1 neurite in the neuropil. 40,000~; scale, 0.25 pm.

B2. Finally, we have presented evidence that the iSCP is confined to the lumenal side of the dense-core vesicle.

The neuronspecific feature of the iSCP localization (Fig. 10) and the dose-dependent inhibition of labeling by preabsorption of the antiserum with synthetic SCPB suggest that the localization is specific for SCP-like antigens. Consider- ing the biochemical evidence that B1 and B2 synthesize and release the SCPs, the localization presented here suggests that the SCPs are associated with the lumenal contents of dense-core vesicles in B1 and B2 and that the SCPs are transported into the neuropil and toward the periphery in dense-core vesicles.

The notion that the SCPs are associated with dense-core vesicles is further supported by a qualitative comparison of B1 and B2 to B15, an identified buccal ganglion motor neuron. B15 has recently been shown to synthesize the SCPs and assemble dense-core vesicles that are SCP immunoreactive, although they are morphologically distinct from the dense-core vesicles in B1 and B2 (Cropper et al., '86). Like most other buccal ganglion neurons that contain the SCPs, B15 has a significantly lower concentration of somatic dense-core vesicles than B1 and B2. This variation in the concentration of somatic dense-core vesicles suggests that the ratio of the rates of peptide synthesis and vesicle transport varies to a significant degree among neurons that synthesize the SCPs. Quantitative bioassay reveals that the SCP abundance in B15 somata is correspondingly lower

Fig. 9. Indirect immunogold localization of SCP immunoreactivity to the dense-core vesicles in Bl's axon in the esophageal nerve. 144,000~; scale, 0.1 pm.


Fig. 10. Demonstration of the cell specific nature of the localization of SCP immunoreactivity. Two serial sections through the buccal ganglion neuropil are shown in (a) and (b). Both sections include the same two dense- core vesicle-filled neurites, one belonging to B1 (B1, Imposil labeled) and the other belonging to an unidentified neuron (**I. One section (a) was probed with the anti-SCP antiserum and the other 6) with the anti-FMRFa

antiserum. Panel (a) demonstrates that under labeling conditions where the B1 neurite expresses SCP immunoreactivity, the adjacent unidentified neurite (**I is not labeled suggesting highly cell specific immunolocalization of SCP. Panel (b) demonstrates that the unidentified neurite (**) could be immunolabeled with the anti-FMRFa antiserum confirming the presence of a neuropeptide other than SCP in the unidentified neurite.

Fig. 11. Indirect immunogold localization of SCP immunoreactivity to the dense-core elements of a Golgi apparatus in B1. 84,000~; scale, 0.1 pm.

368 W. REED ET AL.

than that in B1 and B2 somata (Lloyd et al., '85b; Cropper SCPB have satisfied most of the criteria routinely accepted et al., '86). The qualitative correspondence between somatic as demonstrating that a substance is a neurotransmitter dense-core vesicle concentration and SCP abundance is con- (or modulator). For example, the peptides are synthesized sistent with iSCP localization. in B1 and B2, and are released in a voltage- and calcium-

The general pathway of protein secretion in eucaryotic dependent manner. Although localization of putative trans- cells is well established (Palade, '75). Secretory proteins, mitter peptides to the conventional protein secretory path- synthesized as precursors in the rough endoplasmic reticu- way is not routinely taken as a criterion establishing a lum and nuclear envelope, are modified, sorted, and pack- peptide as an authentic transmitter, the present findings aged into secretory vesicles as they pass through the Golgi strengthen this conclusion and complement the previous apparatus in a cis to trans direction (Farquhar and Palade, lines of evidence. '81; Rothman, '81) and are ultimately released at the cell surface via exocytosis. Exocytosis may occur in a constitu-

ACKNOWLEDGMENTS tive (continuous) or regulated (intermittent) fashion (Tar- takoff, '80; Kelly, '85). Secretory products whose release is regulated are typically concentrated 10- to 200-fold as they We thank Maya Frankfurt for very helpful advice con- are packaged on the trans face of the Golgi apparatus into cerning the development of the dot-blotting technique. This secretory storage vesicles with electron-dense cores (Kelly, work was supported by NIH grants MH37134, MH35564, '85). The somatic ultrastructure of Aplysia neurons B1 and MH3673330, RSDA MM00367330, and Scopes B and C of B2 suggests the presence of a typical regulated protein NIGHS grant GM23549. WR was the recipient of an NIH secretory pathway. In both B1 and B2 dense-core vesicles postdoctoral fellowship. Some of the animals were provided are juxtaposed with dense-core saccules on the trans face of by HHMI Mariculture Resource Facility at Woods Hole. the-Gol& apparatus (e.g., Figs. lb, 11) and, in some cases, these saccules include focal dense-core dilations. The association of SCP immunoreactivity with the dense-core vesicles (Fig. 7) and with the focal dense-core dilations of Golgi apparatus trans saccules (Fig. 11) suggests that SCPs are packaged into dense-core secretory storage vesicles on the trans face of the Golgi apparatus. Moreover, it supports the notion that SCPs are confined to the protein secretory pathway in B1 and B2. This is in agreement with the hypothesis (Mahon et al., '85) that the leading hydrophobic sequence in the SCP precursor is a signal sequence that directs sequestration of the precursor into the membrane or membrane-delimited compartments of the cell.

Neurosecretory products have been localized by others to the RER, nuclear envelope, and cis saccules of the Golgi apparatus in neurons of other species using different immunolocalization techniques (Broadwell et al., '79). In contrast to what was observed here, these methods have not demonstrated affiliation of secretory product with the trans face of the Golgi apparatus. Conversely, SCP immunorea- citivity was not detected in association with the RER, nuclear envelope, or the cis saccules of the Golgi apparatus probably for technical reasons. The absence of immunoreactivity may result from a predominance of unprocessed SCP precursor in these proximal compartments. It is not known whether the anti-SCP antibody employed in this study recognizes the glutaraldehyde-immobilized SCP precursor. A more likely explanation is that the concentration of SCPs and their precursor in the RER and cis Golgi saccules is significantly lower than in the trans Golgi saccules and in dense-core vesicles as has been demonstrated for secretory products whose release is regulated.

Intracellular labeling has allowed the identification of varicosities and neurites of B1 in the buccal ganglion neuropil revealing a striking accumulation of SCP-immunoreactive vesicles in Bl's central processes (Figs. 8,lO). Taken together with a report that exogenously applied SCPA has actions on buccal ganglion neurons (Sossin et al., '861, this observation suggests that in addition to peripheral action, the SCPs may modulate central neurons in the buccal ganglion. A concerted attempt to identify likely secretory sites for the SCP-immunoreactive dense-core vesicles in the neuropil by morphological criteria was not successful.

Studies of neurons B1 and B2 indicate that SCPA and

LITERATURE CITED Abrams, T.W., V.F. Castellucci, J.S. Camardo, E.R. Kandel, and P.E. Lloyd

(1984) Two endogenous neuropeptides modulate the gill and siphon withdrawal reflex in Aplysia by presynaptic facilitation involving c- AMP-dependent closure of a serotonin-sensitive potassium channel. F'roc. Natl. Acad. Sci. 81:7956-7960.

Bendayan, M., and M. Zollinger (1983) Ultrastructural localization of anti- genic sites on osmium-fixed tissues applying the protein A-gold technique. J. Histochem. Cytochem. 31:lOl-109.

Broadwell, R.D., C. Oliver, and M.W. Brightman (1979) Localization of neurophysin within organelles associated with protein synthesis and packaging in the hypothalamolneurohypophysial system: An immuno- cytochemical study. Proc. Natl. Acad. Sci. 76t5999-6003.

Cropper, E.C., P.E. Lloyd, W. Reed, R. Tenenbaum, I. Kupfermann, and K.R. Weiss (1987) Multiple neuropeptides in cholinergic motor neurons of Aplysia. Evidence for modulatory systems intrinsic to the motor circuit. Proc. Natl. Acad. Sci. USA (in press).

Eisenstadt, M., J.E. Goldman, E.R. Kandel, H. Koike, J. Koester, and J.H. Schwartz (1973) Intrasomatic injection of radioactive precursors for studying transmitter synthesis in identified neurons of Aplysia californica Proc. Natl. Acad. Sci. 7Ot3371-3375.

Farquhar, M.D., and G.E. Palade (1981) The Golgi apparatus (complex)- (1954-1981)-from artifact to center stage. J. Cell. Acad. Sci. 70r77s- 103s.

Gardner, D. (1971) Bilateral symmetry and interneuronal organization in the buccal ganglion of Aplysia Science 173t550-553.

Granzow, B., W.O. Friesen, and W.B. Kristan, Jr. (1985) Physiological and morphological analysis of synaptic transmission between leech motor neurons. J. Neurosci. 5t2035-2050.

Hokfelt, T., 0. Johansson, A. Ljungdahl, J.M. Lundberg, and M. Schultzberg (1980) Peptidergic neurons. Nature 284515-521.

Kelly, R.B. (1985) Pathways of protein secretion in eukaryotes. Science 230:25-32.

Kreiner, T., W. Sossin, and R.H. Scheller (1986) Localization of Aplysia neurosecretory peptides to multiple populations of dense core vesicles. J. Cell Biol. 102:769-782.

Lloyd, P.E. (1986) The small cardioactive peptides: A class of modulatory neuropeptides in Aplysia. Trends Neurosci. 9:428-431.

Lloyd, P.E., I. Kupfermann, and K.R. Weiss (1984) Evidence for parallel actions of a molluscan neuropeptide and serotonin in mediating arousal in Aplysia Proc. Natl. Acad. Sci. 81t2934-2937.

Lloyd, P.E., I. Kupfermann, and K.R. Weiss (1985a) Two endogenous neuropeptides (SCPA and SCPB) produce a CAMP-mediated stimulation of cardiac activity in Aplysia. J. Comp. Physiol. A 156:659-667.


Lloyd, P.E., Schacher, I. Kupfermann, and K.R. Weiss (1986) Release of neuropeptides during intracellular stimulation of single identified Aply- sia neurons in culture. Roc. Natl. Acad. Sci. USA (in press).

Lloyd, P.E., A.C. Mahon, I. Kupfermann, J.L. Cohen, R.H. Scheller, and K.R. Weiss (198513) Biochemical and immunocytological localization of molluscan small cardioactive peptides in the nervous system of Aplysia california. J. Neurosci. 51851-1861.

Lloyd, P.E., S. Schacher, I. Kupfermann, and K.R. Weiss (1986) Release of neuropeptides during intracellular stimulation of single identified ApZp sia neurons in culture. Proc. Natl. Acad. Sci. USA (in press).

Lloyd, P.E., M. Frankfurt, P. Stevens, I. Kupfermann, and K.R. Weiss (1987) Biochemical and immunocytological localization of the neuropeptides FMRF-amide, SCPA, SCPB to the neurons involved in the regulation of feeding in Aplysia. J. Neurosci. (in press).

Mahon, A.C., P.E. Lloyd, K.R. Weiss, I. Kupfermann, and R.H. Scheller (1985) The small cardioactive peptides A and B in Aplysia are derived from a common precursor molecule. Proc. Natl. Acad. Sci. USA 82:3925- 3929.

Morris, H.R., M. Panico, A. Karplus, P.E. Lloyd, and B. Riniker (1982) Identification by FAB-MS of the structure of a new cardioactive peptide

from Aplysia Nature 3OOt643-645. Newman, G.R., B. Jasani, and E.D. Williams (1983) A simple post-embed-

ding system for the rapid demonstration of tissue antigens under the electron microscope. Histochem. J. 15543-555,

Palade, G.E. (1975) Intracellular aspects of the process of protein synthesis. Science 189:347-358.

Rothman, J.E. (1981) The Golgi apparatus: Two organelles in tandem. Sci- ence 213:1212-1219.

Shkolnik, L.J., and J.H. Schwartz (1980) Genesis and maturation of serotonergic vesicles in identified giant cerebral neuron of Aplysia J. Neurc- physiol. 43:945-967.

Sossin, W., M. Kirk, and R.H. Scheller (1986) Peptidergic modulation of the neuronal circuitry controlling feeding. Soc. Neurosci. Abstr. 12:38.

Tartakoff, A.M. (1980) The Golgi complex: Crossroads for vesicular traffic. Int. Rev. Exp. Path. 22.227-251.

Weiss, K.R., P.E. Lloyd, E.C. Cropper, M. Frankfurt, and I. Kupfermann (1986) FMRFamide is present in the ARC muscle of Aplysia and de- presses its contractions. Soc. Neurosci. Abstr. 12:946.

association of neuroactive peptides with the protein secretory pathway in identified neurons of...

Documents