Brain Research, 545 (1991) 151-163 © 1991 Elsevier Science Publishers B.V. (Biomedical Division) 0006-8993/91/$03.50 ADONIS 0006899391164728

151

BRES 16472

Monoclonal antibodies identify two novel proteins associated with vasopressin secretory granules of the rat neurohypophysis* Charles M. Paden and Sharon J. Hapner Department of Biology, Montana State University, Bozeman, MT 59717 (U.S.A.) (Accepted 23 October 1990) Key words: MagnoceUular; Supraoptic nucleus; Paraventricular nucleus; Suprachiasmatic nucleus; VPGP38; VPGP68; Immunocytochemistry; Vasopressin

lmmunocytochemical and immunoblotting techniques have been used to characterize the antigens recognized by two monoclonal antibodies (MAbs C6 and D5) produced against dissociated cells from punches of neonatal supraoptic (SON) and paraventricular (PVN) hypothalamic nuclei of the rat. Peroxidase immunocytochemistry revealed that both MAbs label magnocellular perikarya in the adult and neonatal SON and PVN as well as smaller neurons in the suprachiasmatic nucleus. Axons of the hypothalamo-neurohypophysial tract are also immunolabeled within the hypothalamus and zona interna of the median eminence, and C6 and D5 each bind specifically to both the adult and neonatal neurohypophysis. Dual-label immunofluorescence experiments employing C6 or D5 simultaneously with rabbit antisera specific for either oxytoxin, neurophysin or vasopressin neurophysin revealed that C6 binds only to vasopressinergic magnocellular perikarya in the SON, while D5 labels both vasopressinergic and a small subset of oxytocinergic magnocellular neurons. Post-embedding immunogold analysis of MAb binding to the neurohypophysis at the ultrastructural level showed that both C6 and D5 recognize antigens associated with large dense core neurosecretory granules in a subset of neurosecretory axons. Initial biochemical characterization of the antigens recognized by C6 and D5 was performed using SDS-PAGE and Western immunoblotting. MAbs C6 and D5 label single protein bands with apparent molecular weights of 38 and 68 kDa, resp., in blots of reduced extracts from the adult neurointermediate lobe. No cross-reactivity between C6 and D5 and the neurophysins was apparent, nor did anti-neurophysin sera recognize the bands identified by C6 and D5. We have therefore designated these novel antigens as VPGP38 and VPGP68 for VasoPressin Granule Proteins. INTRODUCTION T h e h y p o t h a l a m i c magnocellular n e u r o s e c r e t o r y system ( M N S ) has served as the p r e m i e r m o d e l for studies of central peptidergic neurons since the original descriptions of neurosecretion 38, and there exists a vast and rapidly growing literature concerning virtually every aspect of the biology of oxytocinergic and vasopressinergic neurons. It is clear, however, from the ongoing discoveries of o t h e r peptides co-localized within the M N S 10'12'14'18'23'34'36'39'44'51'53 that much remains to be l e a r n e d regarding the total c o m p l e m e n t of cell-typespecific molecules expressed by magnocellular neurons. We are approaching this question by raising monoclonal antibodies ( M A b s ) against immunogens enriched in magnoceilular neurons or their m e m b r a n e s using the neonatal tolerization p a r a d i g m of Hockfield ~1. The MNS is an advantageous system for this approach because of the anatomical concentration of both the neuronal p e r i k a r y a in the supraoptic (SON) and paraventricular

(PVN) hypothalamic nuclei and their axons in the neurohypophysis. We r e p o r t here the isolation of M A b s raised against dissociated n e o n a t a l cells of the S O N and P V N which recognize two novel proteins expressed primarily by vasopressinergic neurons of the MNS. These novel proteins a p p e a r to be associated with large dense core vesicles ( L D C V ) in the neurohypophysis, which are traditionally referred to as neurosecretory granules, or N S G . R e c e n t advances have fostered a growing appreciation of the molecular complexity of secretory vesicles, b o t h small and large 7"42'48, and isolating the c o m p o n e n t s of vesicles and determining their functions are central to u n d e r s t a n d i n g the control mechanisms of intercellular signaling. T h e production of M A b s against vesicle c o m p o n e n t s has p r o v e n to be a powerful tool in this e n d e a v o r and has led to the discovery of several novel vesicle proteins with varying distributions across different secretory cell types 3'5'24' 31,32,41,48,52 H o w e v e r , analysis of vesicle constituents suggests that many vesicle-associated proteins remain to

* Some of the results reported here have been previously described in abstract form: Hapner, S.J., Marshall, P.E. and Paden, C.M., Monoclonal antibodies which recognize subsets of hypothalamic magnocellular neurosecretory neurons in the rat, Soc. Neurosci. Abstr., 15 (1989) 1087. Correspondence: C.M. Paden, Department of Biology, Montana State University, Bozeman, MT 59717, U.S.A.

152 be described. For example, gel electrophoresis of extracts of bovine NSG indicates that at least 25 proteins are associated with the membrane fraction alone 5. The data presented here indicate that the two proteins identified by our MAbs are distinct from any of the known constituents of vasopressinergic NSG, and furthermore that their distribution is highly restricted compared to that of other vesicle-associated proteins which have been identified using MAbs (see above). We have therefore designated them VPGP38 and VPGP68 for VasoPressin Granule Proteins with apparent molecular weights of 38 and 68 kDa. Their potential relationship to various classes of NSG constituent proteins is discussed below.

MATERIALS AND METHODS

Production of monoclonal antibodies The neonatal tolerization and hyperimmunization paradigm of Hockfield 11 was employed in an attempt to generate MAbs recognizing cell-specific antigens expressed during development of magnocellular neurosecretory neurons. Newborn BALBc/BYJ mice were tolerized with a series of 5 i.p. injections between postnatal days 1 and 9, each consisting of mechanically dissociated cerebral cortical tissue from 2-day-old rat pups sufficient to provide 280-300 /~g protein. Commencing at 50 days of age, the tolerized mice were immunized 3 times at 5-day intervals with mechanically dissociated tissue from punches of neonatal rat supraoptic (SON) and paraventricular (PVN) hypothalamic nuclei (Holtzman rats bred in the MSU Animal Research Center from stocks obtained from Charles River were used throughout). Each immunization consisted of two injections of 350 pg protein in Freund's complete adjuvant, 0.1 ml i.p. and 0.05 ml in each hind footpad. Splenocytes were harvested one week after the last immunization for fusion with P-3/X63-Ag8.6 53 myeloma cells according to the protocol of Lane 17, and hybridomas were selected using standard methods 9. Hybridoma supernatants were screened for content of antibodies which selectively bound to the neurohypophysis of either adult or neonatal rats by fluorescence immunocytochemistry on cryostat sections of fixed pituitary (see below for methods of tissue preparation). Hybridomas of interest were cloned twice by limiting dilution to obtain stable monoclonal cell lines. Both MAbs D5 and C6 described in this paper were determined to belong to the IgM class using a commercial isotyping kit (Hyclone).

Peroxidase immunocytochemistry Adult rats were transcardially perfused under Metofane (Pitman-

Moore) anesthesia with 500 ml of 4% paraformaldehyde + 0.1% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4. Tissues were immersion post-fixed for 2 h, then either dehydrated and embedded in paraffin or stored at 4 °C in PBS containing 15% sucrose until sectioned on a Vibratome or frozen for cryosectioning. Tissue sections were pretreated with 0.1 M NaIO 4 and 0.05% NaBH4, each for 10 min, to eliminate unreacted aldehyde 37, then incubated with MAbs in PBS with 0.1% Tween 20 for 24-48 h at 4 °C, rinsed twice with PBS, incubated with affinity purified peroxidase conjugated goat anti-mouse Ig's (Cappel) at 1:100 for 1-3 h at room temperature (T), and rinsed twice with PBS. Peroxidase was visualized using diaminobenzidine (Sigma) as chromogen and H202 was generated by the glucose oxidase method without cobalt intensification13.

Dual-label fluorescence immunocytochemistry The degree of co-localization between MAbs D5 or C6 and vasopressin or oxytocin was determined by simultaneously incubating the same paraffin section of adult SON with rabbit antisera specific for either vasopressin neurophysin (VPNP) or oxytocin neurophysin (OTNP) and MAb D5 or C6. The anti-VPNP (used at 1:20,000 dilution) and anti-OTNP (used at 1:2500) sera were produced against synthetic peptides of the carboxy-terminal amino acids of the rat sequence and are completely specific for the two neurophysins by both immunocytochemical and RIA analysis (antisera and specificity data kindly provided by Dr. Alan Robinson, personal communication). Binding of anti-OTNP or VPNP to sections was visualized using FITC-conjugated goat anti-rabbit IgG (Hyclone), while binding of MAbs was visualized using biotinylated goat-anti mouse Ig's followed by phycoerythrin conjugated to avidin (avidin-PE; Becton-Dickenson). Dual-label fluorescence photomicroscopy was performed using a Nikon Labophot with epifiuorescence and the B2 (FITC) and G (PE) dichroic mirror-filter blocks. A narrow band pass 520-560 nm interference emission filter was employed in the B2 cube to eliminate spillover from PE. Fields containing labeled neurons were photographed with each filter block using a Nikon HFX-II automatic camera system with Kodak Tmax 400 film at ASA 1600, and positive cells identified visually on projected images of the negatives.

EM immunocytochemistry Adult rats were perfused as described above for light microscopic immunocytochemistry except that the fixative was wanned to 37 °C and perfusion carried out for 20 min at a gradually decreasing flow rate until a total volume of 500 mi had been passed. The neurohypophysis and attached intermediate lobe were removed and post-fixed by immersion in the same fixative for 1 h at room T, then cut into blocks approx. 0.5 mm On a side, rinsed extensively in 0.1 M phosphate buffer, pH 7.4, and dehydrated through 70% ethanol. The tissue was transferred directly from 70% ethanol into LR White resin (medium grade) for 5 rain at room T, then placed in a gelatin capsule containing fresh resin and hardened at 60 °C for 24 h. This very brief infiltration period (similar to that employed by van den Pol 4a) resulted in better preservation of axolemma and other

Fig. 1. Immunoperoxidase staining of coronal sections of rat pituitary and hypothalamus with MAbs C6 and DS. Similar patterns of immunostaining were obtained with both MAbs. A: C6 on a 6 p m paraffin section of adult pituitary; heavy punctate staining is apparent throughout the neural lobe (nl), while the intermediate lobe (il) and anterior lobe (al) are negative. B: D5 on adult pituitary; the pattern of staining is similar to that seen with C6. C: perikarya and processes in the lateral magnocellular subnucleus of the PVN labeled with D5; 40 /,m Vibratome section, medial to the left. D: magnocellular perikarya and processes in the SON labeled with C6. Numerous unstained perikarya are apparent in the more dorsal part of the nucleus; 40/tm Vibratome section, optic tract (ot) to the left. E: smaller neuronal perikarya and processes within the SCN labeled with C6; 40/~m Vibratome section, third ventricle (v) is on the upper right. F: accessory magnocellular neurons lying between the SON and PVN labeled with C6; 6 pm paraffin section. G: darkfield photomicrograph of two magnocellular perikarya (arrowheads) ventral to the optic tract (ot) at the level of the retrochiasmatic SON labeled with D5. Perikarya were overexposed in order to reveal numerous stained axons of the hypothalamo-neurohypophysial tract which are visible dorsal to the optic tract coursing towards the median eminence (to the right; not shown); 10/~m paraffin section. H: labeling of axons in cross-section by D5 in the median eminence. Note the presence of scattered immunoreactive axons in the zona externa (arrowheads) in addition to the intense labeling of the hypothalamo-neurohypophysial tract in the zona interna; 6/.tm paraffin section, v = third ventricle. Scale bars: (A,B) 200/~m; (C,D,F,G,H) 100 ~tm; (E) 50 ~m.

O

11

E

154 membranes than standard embedding protocols in our hands. Postfixation in osmium improved ultrastructural preservation, but greatly reduced immunolabeling of the tissue. Silver ultrathin sections were placed on 300 mesh Formvar/carbon-coated nickel grids (Ted Pella) and processed for post-embedding immunocytochemistry. Grids were pretreated by floating on drops of normal saline buffered with 0.1 M phosphate, pH 7.4, containing 5% normal goat serum, 3% BSA (Sigma Fraction V) and 0.1% Tween 20. This buffer was used throughout. Grids were then transferred directly to 50/zl drops of hybridoma supernatants diluted 1:1 with buffer in a moistened chamber prepared by placing Parafilm over the 16 central wells of a 24-welltissue culture dish and filling the side wells with water. After sealing the lid with Parafllm, grids were incubated for 24 h at 4 °(2. They were then rinsed 3 × 15 min by floating on 0.5 ml of buffer in a 24-well plate placed on a rocker, incubated in a 1:10 dilution of goat anti-mouse Ig's conjugated to 10 nm gold (Janssen) for 2 h at room temperature and rinsed as before. Binding of the second antibody was stabilized by floating the grid for 1 rain on 0.5% glutaraldehyde in distilled water, followed by 2 × 10 min water rinses and post-staining with aqueous uranyl acetate (40 min at 40 °C) and lead citrate (10 min at 20 °C) in a LKB Ultrostainer. TEM was performed using a Zeiss EM10C at 40 kV.

Gel electrophoresis and immunoblotting Tissue homogenates were prepared following transcardiac perfusion of anesthetized adult Holtzman rats with approx. 200 ml of ice-cold PBS containing 50 /~g/ml phenylmethylsulfonylfluoride (PMSF; Sigma) to inhibit proteolysis. Tissues of interest were rapidly homogenized in SDS-PAGE sample buffer (2% SDS and 10% glycerol in 0.625 M Tris-HCl, pH 6.8, with or without 5% fl-mercaptoethanol) and stored at -20 °C until solubilized by boiling for 5 rain. SDS-PAGE was run on 12.5% slab gels as described by Laemmli15. Proteins were then renatured in situ and electrophoretically transferred at pH 1035to Immobilon which was then placed in blocking buffer consisting of 0.5 N NaCI plus 0.05% Tween 20 and 5% BSA (Sigma fraction V) in 0.05 M Tris-HCl, pH 7.4, for 1 h at 37 °C. Following incubation with primary antibodies overnight at room temperature, binding was revealed using peroxidaseconjugated second antibodies (Cappel) and 4-chloro-l-naphthol (Sigma) as chromogen. Polyclonal rabbit anti-rat neurophysins sera #4 was generously provided by Dr. Alan Robinson. This sera labels both OTNP and VPNP as well as propressophysin and intermediate cleavage products.

RESULTS

Immunocytochemistry of MAb C6 and D5 binding in brain and pituitary M A b s C6 and D5 were initially selected for further characterization because of their robust and specific binding to frozen sections of the adult neural lobe (NL). Similar results were obtained using paraffin sections, and these are illustrated in Fig. 1. Fig. 1A shows the highly specific labeling of the adult NL obtained with M A b C6 (which recognizes VPGP38; see below). The entire cross-section of the NL is labeled. No staining was observed in the intermediate or anterior lobes when paraffin or frozen sections were employed, and paraffin sections of the adrenal gland were also negative (not shown). The pattern of i m m u n o l a b e l i n g with M A b D5 (which recognizes VPGP68; see below) in the adult pituitary is identical to that of C6 (Fig. 1B). Preadsorption at 4 °C for 12 h with up to 50/zM oxytocin (Sigma) or arginine vasopressin (Calbiochem) had no effect on immunolabeling of the NL with M A b s C6 or D5. Immunocytochemical surveys of C6 and D5 binding to coronal sections of the adult telencephalon and diencephalon were performed using frozen, paraffin and thick sections. Similar results were obtained with both M A b s on each type of tissue section. I m m u n o r e a c t i v e n e u r o n a l perikarya were found in the supraoptic (SON), paraventricular (PVN) and suprachiasmatic (SCN) hypothalamic nuclei, as well as in the magnocellular accessory nuclei (such as the nucleus circularis) and scattered between the PVN and SON (Fig. 1 C - G ) . N u m e r o u s immunoreactive perikarya within both the lateral and medial magnocellular subnuclei of the PVN (the PVL and PVM as described by A r m s t r o n g et al. 1) were observed, but those in the PVM were more faintly labeled with either M A b and were best visualized in thick sections. Labeled

5 Fig. 2. Immunoperoxidase staining of 10/~m coronal frozen sections of neonatal rat pituitary and SON with MAbs D5 and C6. A: pituitary labeled with D5, showing reaction product throughout the neural lobe (nl) and the absence of staining in the intermediate (il) and anterior (al) lobes; similar results were obtained with C6. Scale bar = 200/zm. B: neurons within the SON labeled with C6; similar results were obtained with D5. Medial is to the left. Scale bar = 100 gm.

155

I i

i

Fig. 3. Immunofluorescence staining of coronal paraffin sections of the adult SON dual-labeled with anti-VPNP or anti-OTNP and MAb C6. A: vasopressinergic neurons stained with anti-VPNP and FITC conjugated second antibody. B: the same field showing cells stained with C6 and avidin-PE; note the presence of numerous dual-labeled perikarya and the positive correlation between the fluorescence intensities of the two fluorophores. C: oxytocinergic neurons stained with anti-OTNP and FITC conjugated second antibody. D: the same field showing cells stained with C6 and avidin-PE (arrowheads). No oxytocinergic neurons are C6 positive, and vice versa. Note the low level of spillover between the two fluorophores, ot = optic tract. The magnification is the same for all micrographs; scale bar = 25/~m.

perikarya were present throughout the principal and retrochiasmatic divisions of the SON (Fig. 1D,G). Beaded axons of the hypothalamo-neurohypophysial tract were also labeled by both M A b s , although more faintly than perikarya, both within the hypothalamus (Fig. 1G) and within the zona interna of the median eminence (Fig. 1H). Scattered axons within the zona externa were also immunoreactive (Fig. 1H), suggesting that some parvocellular neurosecretory n e u r o n s may also

express the VPGP38 and V P G P 6 8 proteins. However, staining of parvocellular n e u r o n s in the PVN was very faint if present at all. No immunoreactivity was observed with either M A b outside of the hypothalamus on coronal sections of whole brain taken at least every 100 /~m starting at the level of the nucleus of the diagonal band and continuing to the mammillary bodies. In summary, the observed distribution of both C6 and D5 i m m u n o r e activity is consistent with the conclusion that both M A b s

156

Fig. 4. Immunofluorescence staining of coronal paraffin sections of the adult SON dual-labeled with anti-VPNP or anti-OTNP and M A b D5. A: vasopressinergic neurons stained with anti-VPNP and FITC conjugated second antibody. B: the same field showing cells stained with D5 and avidin-PE; note the predominance of dual-labeled perikarya and the positive correlation between the fluorescence intensities of the two fluorophores, ot = optic tract. C: oxytocinergic neurons stained with anti-OTNP and FITC conjugated second antibody. D: the same field showing cells stained with D5 and avidin-PE. A single dual-labeled oxytocinergic neuron is visible (arrowheads), while all other oxytocinergic neurons are negative for D5 immunoreactivity and vice versa. Approximately 7% of D5 reactive magnocellular neurons were observed to be oxytocinergic (see text). Scale bars = (A,B) 50 pm; (C,D) 25 pm.

Fig. 5. Electron micrographs illustrating immunogold localization of MAbs C6 and D5 binding to ultrathin sections of adult rat neural lobe embedded in LR White resin. A,B: gold particles are localized over NSG in a subset of neurosecretory axons on sections incubated with D5. Note the absence of gold particles over unlabeled axons (right in A; center in B). C,D: intense labeling of NSG, again in a subset of neurosecretory axons, is apparent on sections incubated with C6. A few scattered gold particles representing background labeling are visible over otherwise unlabeled axons (center in C, lower left in D). While preservation of ultrastructure was compromised by the use of conditions which optimized antigenicity (0.1% glutaraldehyde, no osmium post-fixation, and LR White resin), remnants of both axolemma (in A and D) and mitochondria (m; in B and C) are visible. No labeling of these structures or of other organelles such as cell nuclei or lipid inclusions in pituicytes was observed with either MAb. Scale bars = 0.5/~m.

158 are binding predominantly to vasopressinergic neurons (see below). Both C6 and D5 specifically labeled the NL of 2-day-old rats, as shown in Fig. 2A for MAb D5. Binding of MAbs C6 and D5 to P2 hypothalamus was also examined, and neurons within the immature SON and PVN were positively labeled (as illustrated in Fig. 2B for MAb C6 in the SON).

lmmunocytochemical co-localization of MAb C6 and D5 binding with vasopressin and oxytocin neurophysins in the SON and PVN Coronal paraffin sections of adult SON and PVN were examined using dual-label immunofluorescence in order to determine whether MAbs C6 and D5 bind preferentially to either vasopressinergic or oxytocinergic neurons. The dual-label paradigm which was utilized (see Methods) made it possible to visualize labeling with rabbit anti-vasopressin neurophysin (anti-VPNP) or anti-oxytocin neurophysin (anti-OTNP) on the same section with that of MAbs C6 or D5 with virtually no visible spillover between the two emissions. The low level of spillover in either direction is apparent in Figs. 3C,D and 4C,D. On sections incubated with both anti-VPNP and MAb C6, 120 cells were judged to be immunoreactive for MAb C6. Of these, 117 were clearly labeled by anti-VPNP as well (Fig. 3A,B). In contrast, of 147 cells labeled by MAb C6 on sections co-incubated with anti-OTNP, none were judged to be positive for OTNP (Fig. 3C,D). The great majority of the neurons analyzed were in the SON because those within the PVN were more faintly labeled, making it more difficult to unambiguously identify positive cells. In general, the magnitude of VPNP and C6 fluorescence appeared to be positively correlated within magnocellular neurons of both nuclei. The fainter VPNP positive cells (presumably those with only a small portion of the perikarya within the plane of section) often appeared to be negative for C6, but this result cannot be interpreted to mean that only a subset of VPNP containing cells are C6-positive because the fluorescence labeling of MAb binding was always fainter than that of the anti-neurophysins sera. Thus these results are consistent with the conclusion that the VPGP38 protein (see below) recognized by MAb C6 may be exclusively localized in vasopressinergic neurons. Very similar results were obtained with MAb D5, except that a small number of D5-positive cells in the SON contained OTNP. A total of 126 magnocellular neurons were judged to be unambiguously D5-positive in sections co-incubated with anti-VPNP, and 118 of these were also unambiguously positive for VPNP (Fig. 4A,B). However, of 158 unambiguous D5-positive cells in sections co-incubated with anti-OTNP, 11 which clearly

contained OTNP were also observed within the SON (Fig. 4C,D). As with MAb C6, neurons which showed only faint anti-VPNP labeling often appeared to be negative for D5, but this is probably an artifact of differences in FITC and PE fluorescence intensity. Thus it is possible that all vasopressinergic neurons express the VPGP68 protein recognized by MAb D5 (see below). In contrast, the absence of D5 labeling in numerous neurons with intense anti-OTNP fluorescence (Fig. 4C,D) suggests that only a small subset of oxytocinergic neurons contains the VPGP68 protein.

Ultrastructural localization of C6 and D5 binding in the neurohypophysis Both MAbs yielded robust labeling on ultrathin sections of non-osmicated adult NL when post-embedding immunocytochemistry was performed using 10 nm colloidal gold conjugated goat-anti mouse Ig's on tissue embedded in the hydrophilic resin LR White. In both cases, gold particles were almost exclusively localized over neurosecretory granules in a subpopulation of axons (Fig. 5). Although much ultrastructural detail is lost when neural tissue is embedded in LR White, it is possible to identify remnants of axolemma and organelles such as mitochondria and cell nuclei, none of which were labeled by MAbs C6 or D5. These results are consistent with the hypothesis that the 38 and 68 kDa proteins recognized by MAbs C6 and D5 in the neurointermediate lobe (see below) are primarily associated with neurosecretory granules of vasopressinergic neurons. Unfortunately, preservation of the SON embedded in LR White was inadequate to permit ultrastructural localization of MAb binding within magnocellular perikarya (not shown). Initial experiments employing various etching protocols on Epon-embedded tissue resulted in improved ultrastructure but complete loss of MAb binding in the NL.

Identification of antigens by gel electrophoresis and immunoblotting MAbs C6 and D5 each bind to a single band on immunoblots made from SDS gels of reduced neurointermediate lobe proteins from adult rats, with apparent molecular weights of 38 and 68 kDa, resp. (Fig. 6A). Also shown for comparison are bands labeled by rabbit anti-rat neurophysins serum (gift of Dr. Alan Robinson) and MAb A3, which was produced in the same fusion as C6 and D5 but apparently recognizes oxytocin neurophysin (Hapner and Paden, unpublished results). While the antigenic bands recognized by C6 and D5 cannot be unequivocally identified in the amido black stained protein lane from the same blot, it appears that they do not represent major components of the extract.

159

1

2

3

4

5

Fig. 6. Western immunobiots of proteins separated by SDS-PAGE showing labeling of different bands by MAbs C6 and D5 and controls. A: extracts of adult rat neurointermediate lobe run under reducing conditions and labeled as follows: lane 1, polyclonal rabbit anti-rat neurophysins; lane 2, MAb D5; lane 3, MAb C6; lane 4, MAb A3; lane 5, amido black protein stain. MAbs C6 and D5 consistently label single bands of apparent tool. wts. 38 and 68 kDa, resp., in neurointermediate lobe extracts, and do not cross-react with the neurophysins or other products of the propressophysin gene (which are stained in lane 1). Staining of oxytocin neurophysin by MAb A3 (produced in the same fusion) is shown for comparative purposes. B: immunoblots of various tissue extracts labeled with MAb D5, as follows: lane 1, reduced neurointermediate lobe; lane 2, reduced anterior pituitary; lane 3, reduced hypothalamus; lane 4, non-reduced hypothalamus; lane 5, tool. wt. standards (same for both A and B). The same 68 kDa band is labeled in all tissues; these are slightly distorted by the high lipid content in hypothalamic extracts. The fainter band of approx. 22 kDa in lane 2 was reliably present on blots of reduced extracts of anterior lobe, while the faint band of approx. 29 kDa visible in lane 4 was not reliably present on blots of hypothalamic extracts. MAb C6 did not label any bands on blots of tissues other than neurointermediate lobe (not shown), and neither C6 nor D5 labeled any bands on blots of reduced cerebral cortical extracts (not shown).

Labeling of i m m u n o b l o t s prepared from non-reduced n e u r o i n t e r m e d i a t e lobe proteins as well as from reduced and/or n o n - r e d u c e d extracts of hypothalamus, cerebral cortex, and anterior pituitary was also examined. M A b C6 failed to label any bands in these experiments (not shown), but the comparatively faint band produced by C6 in n e u r o i n t e r m e d i a t e lobe extracts suggests that less concentrated antigens could be present in hypothalamus but fall below the limits of detection. In contrast, M A b D5 recognized the same 68 kDa band u n d e r both reducing and non-reducing conditions in neurointermediate lobe and hypothalamus, and u n d e r reducing conditions in anterior pituitary (Fig. 6B). In addition, a lower mol. wt. b a n d around 22 kDa was reliably present on blots of reduced anterior lobe. The faint b a n d of approx. 29 kDa visible in non-reduced hypothalamus was

not reliably present and may represent a breakdown product. This b a n d was also absent from n o n - r e d u c e d n e u r o i n t e r m e d i a t e lobe (not shown). M A b D5 did not label any bands on blots of cerebral cortical proteins (not shown). DISCUSSION The distribution of i m m u n o l a b e l e d n e u r o n a l perikarya and processes within the hypothalamus, median eminence and NL is consistent with the conclusion derived from dual-label fluorescence studies that M A b s C6 and D5 bind exclusively to vasopressinergic neurons, with the exception of a small population of magnoceUular oxytocinergic n e u r o n s which are also labeled by M A b D5. Both C6 and D5 label smaller vasopressinergic n e u r o n s

160 of the SCN as well as magnocellular vasopressinergic neurons in the SON and PVN, indicating that the antigens recognized by these MAbs are not restricted to the magnocellular neurosecretory system. However, it is not possible to determine from the present results whether these antigens are present in all vasopressinergic neurons. Magnocellular neurons which were more faintly labeled with anti-VPNP often appeared negative for C6 and D5 binding in fluorescence studies, no staining of vasopressinergic neurons in other areas 4'4°'45 was observed, and staining of parvocellular vasopressinergic neurons in the PVN was equivocal on immunoperoxidase labeled sections. However, light staining of the zona externa of the median eminence where the axons of parvocellular neurons terminate was present (Fig. 1H). Thus it will be necessary to perform further experiments (for example in adrenalectomized or colchicine-treated animals) to search for these antigens in additional vasopressinergic neurons. The specific binding of MAbs C6 and D5 to large dense-core neurosecretory granules in ultrathin sections of the NL, taken in conjunction with their recognition of single protein bands on immunoblots of reduced neurointermediate lobe proteins separated by SDS-PAGE, have led us to designate the antigens which they recognize as VPGP38 and VPGP68 for VasoPressin Granule Proteins of apparent mol. wts. 38 and 68 kDa, resp. To the best of our knowledge, no proteins of similar size and distribution have been previously described. The distribution of VPGP38 and VPGP68 in brain and pituitary as determined by SDS-PAGE and immunoblotting is generally, but not completely, consistent with their distribution as determined by immunocytochemical methods. The VPGP38 band was present only on blots of reduced extracts of the neurointermediate lobe and was absent from blots of cerebral cortical and anterior lobe extracts, consistent with immunocytochemical observations, but it was also absent from blots of hypothalamic extracts where immunocytochemical results suggest that it should be present. The failure to label the VPGP38 band in blots of hypothalamic extracts may simply be the result of the lower concentration of the antigen in homogenates of medial-basal hypothalamus compared to neurointermediate lobe, in which case it may prove possible to identify VPGP38 on blots of proteins extracted from punches of the SON in future experiments. The lack of a band on blots of non-reduced extracts of neurointermediate lobe suggests that the VPGP38 protein may be associated with proteins or other cellular constituents which reduce its solubility and therefore restrict its entry into the gel under non-reducing conditions. In contrast, the VPGP68 band was present on blots of both reduced and non-reduced extracts of the neu-

rointermediate lobe and hypothalamus, although it was fainter on the hypothalamic blots, as expected. VPGP68 was not detected on blots of reduced cerebral cortical proteins, consistent with immunocytochemical results, but was present on blots of reduced extracts of anterior pituitary, while no positive staining was observed on paraffin or frozen sections of anterior lobe. A fainter immunoreactive band of approximate mol. wt. 22 kDa, which was absent from neurointermediate lobe, was consistently observed along with VPGP68 in blots of anterior lobe extracts. The presence of the smaller antigen suggests that VPGP68 may be more labile in the anterior lobe, which could affect its stability during tissue processing for immunocytochemistry. Different fixatives and tissue-processing techniques must be employed in future experiments to determine if immunoreactive cells can be found within the anterior lobe. The presence of VPGP68 in the anterior lobe is not necessarily inconsistent with its co-localization with vasopressin, since several groups have reported evidence for synthesis of vasopressin and neurophysins within the anterior lobe of the rat 6'49'5°. There are several classes of proteins believed to be associated with NSG to which VPGP38 and VPGP68 could belong. Among these are peptide-processing enzymes, vesicle membrane proteins, neuropeptide precursors or prohormones, and secreted proteins of unknown function such as secretogranins or chromogranins. Intragranular peptide processing enzymes are thought to be present in amounts as little as one or two molecules per granule (Y.P. Loh, personal communication, and ref. 22), making it unlikely that antibodies recognizing them would yield the robust immunocytochemical and immunobiot staining which we have observed. Two previously described proteins which are associated with synaptic vesicle membranes, p38 or synaptophysin and p658" 24,31 33, are similar in size to VPGP38 and VPGP68. However, it is unlikely that these are identical to VPGP38 and VPGP68 for several reasons. Both synaptophysin and p65 are distributed widely in both brain and other tissues such as the adrenal medulla 2~'24, whereas VPGP38 and VPGP68 are restricted almost exclusively to vasopressinergic neurons. In addition, there is conflicting evidence regarding the association of synaptophysin and p65 with NSG. On the one hand, antibodies to synaptophysin decorated only microvesicles and not NSG in EM immunocytochemical studies of bovine neural lobe homogenates 2s'29, and both synaptophysin and p65 were found to be associated with the microvesicle fraction obtained from sucrose gradients of neural lobe extracts 28. On the other hand, antibodies to both synaptophysin and p65 have been successfully utilized to immunoisolate LDCV from bovine adrenal medulla and rat pheochromocytoma cells 2~, and both proteins are associated with

161 coated vesicles from bovine brain as well 33. These divergent results could indicate that NSG are distinct from other L D C V and do not contain synaptophysin and p65, or alternatively, that the amount of these proteins present is very low 28. Consistent with this possibility, only 0.6-1.2 anti-p65 antibody molecules were estimated to bind on average to each immunolabeled coated vesicle from brain 33. Neither of these interpretations is consistent with the high density of immunogold labeling of NSG with MAbs C6 and D5, which suggests that the number of VPGP38 and VPGP68 molecules per granule is much greater than is the case for synaptophysin or p65. Additional biochemical and ultrastructural experiments are required to determine if VPGP38 and VPGP68 are associated with the NSG membrane. The EM immunocytochemical results obtained to date are inconclusive in this regard. While gold particles appear to be distributed across the entire profile of NSG in LR White sections (Fig. 5), this could be an artifact resulting from the poor preservation of membranes which occurs with this protocol. Preliminary immunocytochemical trials employing slam frozen NL have revealed that gold particles are preferentially localized over NSG membranes with both MAbs C6 and D5 when tissue is prepared by freeze-substitution in acetone with 2% osmium, but the density of labeling is very low. The amount of labeling increases substantially when the osmium concentration is reduced, but the membranes are less well preserved and the gold particles then appear over the core of the NSG as well (Dr. William Armstrong, personal communication). Since VPGP38 and VPGP68 may well prove to be intragranular proteins, their robust labeling in immunocytochemical experiments raises the question of possible relation to the principal known constituents of vasopressinergic NSG, the propressophysin gene products. Evidence supporting the existence of "big" vasopressin/ neurophysin prohormone molecules of approx. 70-80 kDa in size has been reported in several mammals including the rat 2A9'2°'27'3°. For example, Lauber et al. described limited trypsin digestion of a 80 kDa molecule extracted from bovine NL which yielded a 68 kDa fragment with fl-endorphin and A C T H immunoreactivities and a 10 kDa fragment homologous with neurophysin 2°. Little has been published in this area since the successful determination of the prepropressophysin gene REFERENCES 1 Armstrong, W.E., Warach, S., Hatton, G.I. and McNeill, T.H., Subnuclei in the rat hypothalamic paraventricular nucleus: a cytoarchitectural, horseradish peroxidase and immunocytochemical analysis, Neuroscience, 5 (1980) 1931-1958. 2 Beguin, P., Nicolas, P., Boussena, H., Fahy, C. and Cohen, P.,

sequence t6 and the demonstration that it encoded a much smaller prohormone. Nevertheless, the possibility remains that an alternative pathway for production of vasopressin could exist. The principal evidence arguing against the possibility that VPGP38 or VPGP68 are part of such a pathway is the absence of labeling of either the neurophysins by MAbs C6 and D5 on blots of neurointermediate lobe extracts, or of VPGP38 and VPGP68 by polyclonal rabbit anti-rat neurophysin sera. The polyclonal rabbit anti-rat VPNP and anti-rat OTNP sera used in our dual-label fluorescence studies also failed to crossreact with VPGP38 and VPGP68 on immunoblots (Hapner and Paden, unpublished results). In addition, preadsorption with oxytocin and vasopressin failed to block staining of the NL by MAbs C6 and D5. Finally, the presence of VPGP68 in a subset of oxytocinergic neurons suggests that it is not related to a vasopressin prohormone, since it has been demonstrated that magnocellular neurons are exclusively oxytocinergic or vasopressinergic26'46'47. Numerous other peptides have been co-localized within magnocellular vasopressinergic neurons, including dynorphin 51, leu-enkephalin 23, angiotensin II 14'34, and galanin 1°'25"36"39. Peptides related to the chromogranin/ secretogranin family have also been found in the neural lobe 12'18, and new reports of additional neuropeptides within the hypothalamo-neurohypophysial system continue to appear 44'53. However, in no case is the apparent distribution of immunoreactivity outside of the MNS the same for these peptides as for VPGP38 or VPGP68, nor have we found any reports of other peptides which share the unusual property of VPGP68 of being found in most or all vasopressinergic cells and in a subset of oxytocinergic cells. Still, immunocytochemical data are certainly not conclusive in themselves, and the possible relationship between VPGP38 and VPGP68 and other neuropeptides must be addressed in further experiments. Regardless of the class of proteins to which VPGP38 and VPGP68 may belong, the ultimate goal of future research will be to elucidate their physiological roles in the neurosecretory process. Acknowledgments. We wish to thank Peggie Bensch and Jessica Davis for outstanding technical assistance, and Prof. Alan Robinson for the generous gift of anti-neurophysin sera (prepared under NIH Grant AM16166). This research was supported by NIH Grant NS23642, NSF EPSCoR Grant RII-8921978, and Research Career Development Award NS01318 to CMP.

Characterization of the 80,000 molecular weight form of neurophysin isolated from bovine neurohypophysis, J. Biol. Chem., 256 (1981) 9289-9294. 3 Buckley, K. and Kelly, R.B., Identification of a transmembrane glycoprotein specific for secretory vesicles on neural and endocrine cells, J. Cell Biol., 100 (1985) 1284-1294. 4 Caffe, A.R. and Van Leeuwen, F.W., Vasopressin-immunore-

162

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

active cells in the dorsomedial hypothalamic region, medial amygdaloid nucleus and locus coeruleus of the rat, Cell tissue Res., 233 (1983) 23-33. Caorsi, C.E. and Gonz~ilez, C.B., Antibodies against specific membrane proteins of neurosecretory granules isolated from bovine neural lobes, Neurosci. Lett., 98 (1989) 241-246. Chateau, M., Marchetti, J., Burlet, A. and Boulange, M., Evidence of vasopressin in adenohypophysis. Research into its role in corticotroph activity, Neuroendocrinology, 28 (1979) 25-35. De Camilli, P. and Navone, E, Regulated secretory pathways of neurons and their relation to the regulated secretory pathway of endocrine cells, Ann. N.Y. Acad. Sci., 493 (1987) 461-479. DeVoto, S.H. and Barnstable, C.J., SVP38: a synaptic vesicle protein whose appearance correlates closely with synaptogenesis in the rat nervous system, Ann. N.Y. Acad. Sci., 493 (1987) 493-496. Galfre, G. and Milstein, C., Preparation of monoclonal antibodies: strategies and procedures. In J.J. Langone and H.U. Uunakis (Eds.), Methods in Enzymology, Vol. 73, part B, Academic Press, New York, 1981, pp. 3-46. Gaymann, W. and Martin, R., Immunoreactive galanin-like material in magnocellular hypothalamo-neurohypophysial neutones of the rat, Cell Tissue Res., 255 (1989) 139-147. Hockfield, S., A Mab to a unique cerebellar neuron generated by immunosuppression and rapid immunization, Science, 237 (1987) 67-70. Hutton, J.C., Peshavaria, M., Johnston, C.E, Ravazzola, M. and Orci, L., Immunolocalization of betagranin: a chromogranin A-related protein of the pancreatic B-cell, Endocrinology, 122 (1988) 1014-1020. ltoh, K., Konishi, A., Nomura, S., Mizuno, N., Nakamura, Y. and Sugimoto, T., Application of coupled oxidase reaction to electron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method, Brain Research, 175 (1979) 341-346. Kilcoyne, M., Hoffman, D.L. and Zimmerman, E.A., Immunocytochemical localization of angiotensin II and vasopressin in rat hypothalamus: evidence for production in the same neuron, Clin. Sci., 59 (1980) 57s-60s. Laemmli, U.K., Cleavage of structural proteins during the assembly of the head of Bacteriophage T, Nature, 227 (1970) 680-685. Land, H., Schutz, G., Schmale, H. and Richter, D., Nucleotide sequence of cloned cDNA encoding bovine arginine vasopressinneurophysin II precursor, Nature, 295 (1982) 299-303. Lane, R.D., A short-duration polyethylene glycol fusion technique for increasing production of monoclonal antibody-secreting hybridomas, J. lmmunol. Meth., 81 (1985) 223-228. Lassmann, H., Hagn, C., Fischer-Colbrie, R. and Winkler, H., Presence of chromogranin A, B and C in bovine endocrine and nervous tissues: a comparative immunohistochemical study, Histochem. J., 18 (1986) 380-386. Lauber, M., Camier, M. and Cohen, P., Immunological and biochemical characterization of distinct high molecular weight forms of neurophysin and somatostatin in mouse hypothalamus extracts, FEBS Lett., 97 (1979) 343-347. Lauber, M., Nicolas, P., Boussetta, H., Fahy, C., Beguin, P., Camier, M., Vaudry, H. and Cohen, P., The Mr 80,000 common forms of neurophysin and vasopressin from bovine neurohypophysis have corticotropin- and beta-endorphin-like sequences and liberate by proteolysis biologically active corticotropin, Proc. Natl. Acad. Sci. U.S.A, 78 (1981) 6086-6090. Lowe, A.W., Madeddu, L. and Kelly, R.B., Endocrine secretory granules and neuronal synaptic vesicles have three integral membrane proteins in common, J. Cell Biol., 106 (1988) 51-59. Mains, R.E., Cullen, E.I., May, V. and Eipper, B.A., The role of secretory granules in peptide biosynthesis, Ann. N.Y. Acad. Sci., 493 (1987) 278-291. Martin, R. and Voigt, K.H., Enkephalins co-exist with oxytocin

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38 39

40

41

42

and vasopressin in nerve terminals of rat neurohypophysis, Nature, 289 (1981) 502-504. Matthew, W.D., Tsavaler, L. and Reichardt, L.F., Identification of a synaptic vesicle-specific membrane protein with a wide distribution in neuronal and neurosecretory tissue, J. Cell Biol., 91 (1981) 257-269. Melander, T., H6kfelt, T., Rokaeus, A., Cuello, A.C., Oertel, W.H., Verhofstad, A. and Goldstein, M., Coexistence of galanin-like immunoreactivity with catecholamines, 5-hydroxytryptamine, GABA and neuropeptides in the rat CNS, J. Neurosci., 6 (1986) 3640-3654. Mohr, E., Bahnsen, U., Kiessling, C. and Richter, D., Expression of the vasopressin and oxytocin genes in rats occurs in mutually exclusive sets of hypothalamic neurons, FEBS Lett., 242 (1988) 144-148. Moore, G.J. and Rosenior, J.C., Characterization of the "giant precursors: (70-80K) of vasopressin and oxytocin in the rat hypothalamus, Prog. Brain Res., 60 (1983) 253-256. Navone, E, Di Gioia, G., Jahn, R., Browning, M., Greengard, P. and De Camilli, P., Microvesicles of the neurohypophysis are biochemically related to small synaptic vesicles of presynaptic nerve terminals, J. Cell Biol., 109 (1989) 3425-3433. Navone, E, Jahn, R., Di Gioia, G., Stukenbrok, H., Greengard, P. and De Camilli, P., Protein P38: an integral membrane protein specific for small vesicles of neurons and neuroendocrine cells, J. Cell Biol., 103 (1986) 2511-2527. Nicolas, P., Camier, M., Lauber, M., Masse, M.J., Mohring, J. and Cohen, P., Immunological identification of high molecular weight forms common to bovine neurophysin and vasopressin, Proc. Natl. Acad. Sci. U.S.A, 77 (1980) 2587-2591. Obata, D., Nishiye, H., Fujita, S.C., Shirao, T., Inoue, H. and Uchizono, K., Identification of a synaptic vesicle-specific 38,000-dalton protein by monoclonal antibodies, Brain Research, 375 (1986) 37-48. Obata, K., Kojima, N., Nishiye, H., Inoue, H., Shirao, T., Fujita, S.C. and Uchizono, K., Four synaptic vesicle-specific proteins: identification by monoclonal antibodies and distribution in the nervous tissue and the adrenal medulla, Brain Research, 404 (1987) 169-179. Pfeffer, S.R. and Kelly, R.B., The subpopulation of brain coated vesicles that carries synaptic vesicle proteins contains two unique polypeptides, Cell, 40 (1985) 949-957. Phillips, M.I., Weyhenmeyer, J., Felix, J., Ganten, D. and Hoffman, W.E., Evidence for an endogenous brain reninangiotensin system, Fed. Proc., 38 (1979) 2260-2266. Rosenbaum, L C . , Nilaver, G., Hagman, H.M. and Neuwelt, E.A., Detection of low molecular weight polypeptides on nitrocellulose with monoclonal antibodies, Anal. Biochem., 183 (1989) 250-257. ROkaeus, A., Young, W.S., III and Mezey, 1~., Galanin coexists with vasopressin in the normal rat hypothalamus and galanin's synthesis is increased in the Brattleboro (diabetes insipidus) rat, Neurosci. Lett., 90 (1988) 45-50. Schachner, M., Hedley-Whyte, E.T., Hsu, D.W., Schoonmaker, G. and Bignami, A., Ultrastructural localization of glial fibrillary acidic protein in mouse cerebellum by immunoperoxidase labelling, J. Cell Biol., 75 (1977) 67-73. Scharrer, E. and Scharrer, B., Hormones produced in neurosecretory cells, Recent Prog. Horm. Res., 10 (1954) 183-240. Skofitsch, G., Jacobowitz, D.M., Amann, R. and Lembeck, E, Galanin and vasopressin coexist in the rat hypothalamo-neurohypophyseal system, Neuroendocrinology, 49 (1989) 419-427. Sofroniew, M.V. and Weindl, A., Extrahypothalamic neurophysin-containing perikarya, fiber pathways and fiber clusters in the rat brain, Endocrinology, 102 (1978) 334-337. Trimble, W.S., Cowan, D.M. and Scheller, R.H., VAMP-l: a synaptic vesicle-associated integral membrane protein, Proc. Natl. Acad. Sci. U.S.A, 85 (1988) 4538-4542. Trimble, W.S. and Scheller, R.H., Molecular biology of synaptic vesicle-associated proteins, Trends Neurosci., 11 (1988) 241-242.

163 43 van den Poi, A.N., Dual ultrastructural localization of two neurotransmitter-related antigens: colloidal gold-labeled neurophysin-immunoreactive supraoptic neurons receive peroxidase-labeled glutamate decarboxylase- or gold-labeled GABAimmunoreactive synapses, J. Neurosci., 5 (1985) 2940-2954. 44 van den Pol, A.N., Decavel, C., Levi, A. and Paterson, B., Hypothalamic expression of a novel gene product, VGF: immunocytochemical analysis, J. Neurosci., 9 (1989) 4122-4137. 45 van Leeuwen, F. and Caffe, R., Vasopressin-immunoreactive cell bodies in the bed nucleus of the stria terminals of the rat, Cell Tissue Res., 228 (1983) 525-534. 46 Vandesande, E and Dieriekx, K., Identification of the vasopressin producing and of the oxytocin producing neurons in the hypothalamic magnocellular neurosecretory system of the rat, Cell Tissue Res., 164 (1975) 153-162. 47 Vandesande, E and Dierickx, K., The activated hypothalamic magnocellular neurosecretory system and the one neuron - - one neurohypophysial hormone concept, Cell Tissue Res., 200 (1979) 29-33. 48 Walker, J.H. and Agoston, D.V., The synaptic vesicle and the

cytoskeleton, Biochem. J., 247 (1987) 249-258. 49 Watkins, W.B., Presence of neurophysin-like material in the pituitary corticotrophs and melanotrophs and cells of the arcuate nucleus of the rat as revealed by immunocytochemistry, Neuropeptides, 3 (1983) 477-492. 50 Watkins, W.B., Non-coordinate localization of corticotroph neurophysin and B-endorphin in the anterior pituitary gland of the rat, Neurosci. Lett., 69 (1986) 89-93. 51 Watson, S.J., Akil, H., Fischli, W., Goldstein, A., Zimmerman, E.A., Nilaver, G. and Greidanus, V., Dynorphin and vasopressin: common localization in magnocellular neurons, Science, 216 (1982) 85-87. 52 Wiedenmann, B. and Franke, W.W., Identification and localization of synaptophysin, an integral membrane glycoprotein of Mr 38,000 characteristic of presynaptic vesicles, Cell, 41 (1985) 1017-1028. 53 Yoshizawa, T., Shinmi, O., Giaid, A., Yanagisawa, M., Gibson, S.J., Kimura, S., Uchiyama, Y., Polak, J.M., Masaki, T. and Kanazawa, I., Endothelin: a novel peptide in the posterior pituitary system, Science, 247 (1990) 462-464.