Isolation, partial amino acid sequence, and immunohistqehemical ...

28 downloads 0 Views 2MB Size Report
... immunohistqehemical localization of a brain-specific calcium-binding protein ... Therefore, in this reportwe outline procedures for the complete purification of ... to human cerebellar calbindin-28kDa was kindlyprovided by. K. G. Baimbridge ...
Proc. Nati. Acad. Sci. USA Vol. 86, pp. 10139-10143, December 1989 Neurobiology

Isolation, partial amino acid sequence, and immunohistqehemical localization of a brain-specific calcium-binding protein (protein 10/calretinin/calbindin-28kDa)

Lois WINSKY*t, HIROYASU NAKATA*, BRIAN M. MARTINt, AND DAVID M. JACOBOWITZ* *Laboratory of Clinical Science and

tClinical Neuroscience Branch, National Institute of Mental Health,

Building 10, Room

3D-48, Bethesda, MD 20892

Communicated by Ann M. Graybiel, September 1, 1989 (received for review June 23, 1989)

and sufficient EDTA was added to give 1 mM EDTA. After a 1-hr incubation, the fractions were desalted using Sephadex G-25 columns (PD10, Pharmacia) previously equilibrated with 1 mM EDTA/20 mM Tris/20 mM NaCl, pH 8.0, and were applied on a second DEAE-cellulose (1 x 3.5 cm) column. The column was washed extensively with column buffer followed by buffer containing 50 mM NaCl. Protein 10 was then eluted with buffer containing 100 mM NaCl. The 100 mM NaCl eluate was concentrated (Amicon) and applied to a gel-filtration column (TSK 3000 SW, Beckman) on FPLC (70 gg/100 gl, 0.4 ml/min) using a buffer consisting of 50 mM sodium phosphate (pH 6.5) and 150 mM NaCl. Peak fractions were collected, and the presence of protein 10 was assessed by gel electrophoresis. Two-Dimensional Gel Electrophoresis, Calcium Binding, and Development of Immunoblots. Two-dimensional gel electrophoresis was performed according to the methods of O'Farrell (9) with details as described by Heydorn and Jacobowitz (10). Proteins were transferred onto nitrocellulose paper (0.45-gm pore size), and immunoblots were developed as described by Towbin et al. (11). Calcium binding of protein 10 was accomplished on nitrocellulose blots of two-dimensional gels as described (8, 12) with 100 gCi (1 Ci =-37QGBq) of 45CaC12 (New England Nuclear, 56.5 mCi/mg). Autoradiograms of blots were developed following exposure of XAR2 (Kodak) film for 2-4 days. The polyclonal antibody to human cerebellar calbindin-28kDa was kindly provided by K. G. Baimbridge (University of British Columbia, Vancouver). The use of this antisera has been reported (3). Proteolytic Digestion and Partial Amino Acid Sequence for Protein 10. Sequence grade V8 protease, Asp-N protease, and trypsin were used essentially as described by the manufacturer (Boehringer-Mannheim) at a substrate-to-enzyme ratio of 50:1. The peptides generated by each individual enzyme digest were chromatographed on a Vydac C4 narrowbore HPLC column using a Beckman System Gold liquid chromatography system equipped with a model 167 variablewavelength detector monitoring at both 225 and 280 nm. The chromatographic conditions were flow rate, 1 ml/min; linear gradient from 0 to 60% B over 60 min; buffer A, 0.12% trifluoroacetic acid in H20; buffer B, 0.1% trifluoracetic acid in acetonitrile. Several peaks were subjected to amino acid sequence analysis by using an Applied Biosystem 470A gas-phase sequencer equipped with a model 120A on-line phenylthiohydantoin analyzer. Antibody Production and Immunohistochemistry. One adult male New Zealand White rabbit received multiple intradermal injections (13) of 0.33 ml of the protein 10 fraction from the TSK column (30 ,ug) emulsified in 0.66 ml of Freund's complete adjuvant (GIBCO) two times at monthly intervals. Antisera used in this report were obtained 6 weeks after the initial injection of antigen. Rats and guinea pigs were purfused, and tissue was fixed and processed for immunohistochemistry as

A calcium-binding protein (protein 10) havABSTRACT ing a molecular mass of 29 kDa and an isoelectric point of 5.3 was purified from guinea pig brain. The amino acid sequence of fragments from proteolytic digestion of protein 10 revealed an 86% sequence identity with a calcium-binding protein (calretinin) found in chicken retina. Polyclonal antibodies against protein 10 revealed a specific distribution ofthis protein within sensory neurons of auditory, visual, olfactory, nociceptive, and gustatory systems as well as other discrete neuronal circuits in rat and guinea pig brain, whereas no specific label was observed in any of several peripheral tissues examined.

In recent years a number of calcium-binding proteins have been localized to the nervous system and other excitatory cells (1). Calmodulin appears throughout the body, and studies of calmodulin action strongly suggest a critical role of this calcium-binding protein in the regulation of calciummediated changes in enzyme activity (2). Other calciumbinding proteins such as calbindin-28kDa, S-100, and parvalbumin are distributed within specific populations of cells in brain (3-6). However, the functions of these calcium-binding proteins within the brain remain to be elucidated. A recently described calcium-binding protein (protein 10) with a molecular mass of 29 kDa and an isoelectric point (pl) of 5.3 was shown to be prominent on two-dimensional polyacrylamide gels from micropunch samples of cochlear nuclei as compared with other auditory and some nonauditory regions of rabbit, rat, and guinea pig brain (7, 8). Results of these earlier reports clearly distinguished protein 10 from calbindin28kDa, which has a slightly lower molecular mass (27-28 kDa), is more acidic, and is specifically localized to different brain regions than protein 10, as indicated by twodimensional gel electrophoresis (7, 8). However, these previous results gave no suggestion as to the identity of protein 10. Therefore, in this report we outline procedures for the complete purification of protein 10, provide a partial amino acid sequence based on proteolytic fragments of protein 10, and describe the immunohistochemical distribution of this calcium-binding protein in guinea pig and rat brain.

MATERIALS AND METHODS Protein Purification. Protein 10 was partially purified from 15 g of guinea pig brain homogenate by precipitation in 60-80% ammonium sulfate as previously described (8). Approximately 40 mg of the precipitate was desalted into column buffer and applied to a DEAE-cellulose (DE-52, Whatman) column (1 x 7 cm, 0.3 ml/min) essentially as described (8), except for the addition of 100 gM CaCl2 to the column buffer (20 mM Tris/20 mM NaC1, pH 8.0) for a greater yield of protein 10 in the pass-through protein fraction. The pass-through and wash fractions were collected, The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

tTo whom reprint requests should be addressed. 10139

Proc. Natl. Acad. Sci. USA 86 (1989)

Neurobiology: Winsky et al.

10140 0.1

predominantly in the pass-through fraction following application onto the first DEAE-cellulose column. Incubation of this fraction in buffer containing 1 mM EDTA, instead of calcium, resulted in elution of protein 10 by 100 mM NaCl on the second DEAE-cellulose column without significant contamination by other proteins. Thus, protein eluted by 100 mM NaCi in the presence of 1 mM EDTA appeared as a single protein spot on two-dimensional gels, which migrated to the position of protein 10 when added to a tissue homogenate (data not shown). Application of this fraction to the TSK column resulted in a peak at a calculated molecular mass of 35 kDa, indicating that protein 10 exists predominantly as a monomer (Fig. 1). Amino-terminal sequence analysis of intact protein 10 failed to reveal any amino acid identification through 10 cycles of Edman degradation, suggesting that the protein was blocked at the amino terminus. Amino acid sequence analysis of individual peptides from the various enzyme digests revealed single unique amino acid identifications (Fig. 2). As indicated by the blocks around identical amino acids, there was a high degree of sequence homology (86%) between amino acids of protein 10 and another calcium-binding protein (calretinin) whose partial amino acid sequence has been deduced from the cDNA sequence of a clone from chicken retina (16). In contrast, sequence homology was only 60% between the protein 10 amino acid fragments and the sequence of the rat calbindin-28kDa (15). Antisera appeared specific for protein 10 on Western blots developed from two-dimensional gels of both guinea pig and rat cochlear nucleus or cerebellum. Fig. 3 presents silver stains (A and B) and immunoblots (C-F) developed from two-dimensional gels of guinea pig cochlear nucleus (A, C, and E) and cerebellum (B, D, and F). As shown, protein 10 antibody (1:3000) reacted with a single protein corresponding with the position and relative amounts of protein 10 (large arrows) on the silver-stained gels (Fig. 3, compare C and D with A and B). In contrast, and in agreement with previous reports (7, 8), the calbindin-28kDa antibody (1:5000) reacted

-

E

co ax

0.05 -

C

co

-DE

cn

.0

o-

3.0

0

-- T 6.0

X 9.0

l 11.0

Elution Volume (ml) FIG. 1. Elution profile of protein 10 on TSK. The 100 mM NaCl eluate from the second DEAE-cellulose column was concentrated and Two-dimensionalgel showing applied to a TSK column. the silver st aining of protein 10 (1 ,ug) from the fraction corresponding to the line. (Lower Inset) Autoradiogram of a blot showing 45Ca2+ binding of rprotein 10 from this same fraction. The molecular mass of proteins on ithe gel and the autoradiogram ranged from 100 to 10 kDa (top to bott()m), and isoelectric points ranged from 3 to 10 (left to right).

(Upperlnset)

conoforstim nos Pabelwereabsrforpedbyincubation

(4 Preabsorption controls for immunohisdescribed (14). label were performed by incubation of sections iin tochemica li 1 a solution consisting of protein 10 antisera (1:1000) and cuM purified pirotein 10 for 24 hr. Control sections were incubated under iderntical conditions except that no antigen was added.

RESULTS In agreennent with previous findings (8), protein 10 was precipitat4ed by 60-80% ammonium sulfate and appeared CR Prlo

30

20

10

CB

Met Ala Glu Ser His Leu Gln Ser Ser Leu Ile ThrAla Ser Gln Phe Phe|Glu Ile Trp Leu His Phe Asp Ala Asp GlyI Ala Pro[His Leu|His Leu Ala Asp ValSeriAla Ser Gln Phe Leu Asp Arg His Phe Asp Ala Asp Gly Asn Glu Ile Tp Lys His Phe Asp Ala Asp G ulAsn

-ValTrp

40

Gly TyrjeuGlu Gly Tyr Il Glu Gly Tyr ile Glu

60 s0 Arg Lys Lys Ala ly|Leu Glu------Leu-Ser Pro Glu Ser Ser Ser Asp Lys Glu Val Asp Ala Arg Lys Gly Thr Lys Arg Asp Ltuj Gly

CB CR PrlO

Gly Lys Glu Leu Glu Asn Phe Phe Gln Gl

CB CR PrlO

90 70 Va Asp Gin TrGly Gln Arg Asp Asp Gly Lys Ile Gly Ile ValG Leu Ala His Val Leu Pro Thr Glu Glu Asn Met Lys IThr [&Pro Glu Asni Glu Glu Phe t His Lys Tyr Asp Lys Asn Ala Asp Gly Lys Ile( Glu Met Ala Glu Leu Ala Gin IleLeu Me Gln Lys Tyr Asp Lys Asn Ser !Ep

CB CR PrlO

Gly Lys Glu Leu Gln Asn Leu Ile Gln Glu

GlyLy

u Leu Gln Ala u

W

100 C s 110 Gln G Phe etLsT TrpArgLysTyr Asp Thr AspHis |Ser Gl Phe Leu LeuLeuPheArg GnGnLeu s r Giu a Arg [Ser2 Trp Arg Arg Tyr Asp Thr [Phe LeuLeu Cys Phe Arg Gln His Vai Giy Ser Ser Ser Glu Phe Met MTp " jGin His Val Gly Ser Ser Ala

130

CB

Giu Thr Glu Glu Leu Lys Asn Phe Le Glu Ala Asn Glu Leu Lys Gly L

150 140 Val Asp Asp Thr Lys Leu Ala Glu Tyr ThrJAsp Tyr Aspj Glu Ala Lys Leu Gln Glu Tyr Thr Gln

CR

Lys Asp Leu Leu Glu Lys Ala Asn Lys Thr Ser Asp Leu Leu Lys Lys Ala Asn Arg Pro 160 Leu Met Leu Lys Leu Phe Asp Ser Asn Asn Asp Gly Lys Leu Gln Le Thr Giu Met Ser GMet | Met H Giy Asp Giy Lys LeuIGiy[ &hr le(J Arg Met|

CB CR PrlO

Leu Lys Phe Gln Giy Cys Gly Lys Glu Phe Asn Lys Ala Phe Glu Leu Tyr Asp Gln. Leu Lys Phe Gln Gly Met Lys Leu Ser Leu Lys Phe Gln Gly Met Lys Leu Thr Ser Glu Glu Phe Asn Ala Ile Phe Thr Phe Tyr Asp Lys

CR

CB

AspIMet 190

120

Phejfl Tyr

180

170

Ala Arg Leu Leu Pro Val Gln Glu Asn Phe Leu Arg Leu Leu Pro Val Gln Glu Asn Phe Leu Ser Arg Leu Leu Pro Val Gin Glu Asn Phe Leu

210

200 ..

FIG. 2. Comparison of the amino acid sequences of rat calbindin-28kDa (CB) (15), chicken calretinin (CR) (16), and proteolytic digestion fragments of protein 10 (PrlO). The amino acid sequence of protein 10 fragments (italicized) following digestion by Asp-N (peptides 16-32 and 168-177), V8 (peptides 64-73, 169-177, and 178-191), and trypsin (peptides 20-32, 97-110, and 188-202). Amino acid numbers correspond to those of calretinin (16). A gap has been introduced at amino acids 52-56 of calbindin-28kDa for maximum sequence identity. For the same reason, a Cys has been placed out of the calbindin-28kDa sequence after position %. One tryptic peptide sequence (Asn Glu Pro Ala lie Leu) may correspond to amino acids 76-83 of calretinin, whereas another sequence (Glu Met Asn Ile Gin Gin Leu Trp Asn Tyr) could not be matched with either the calbindin-28kDa or calretinin sequences to give greater than 40% sequence identity.

Neurobiology: Winsky et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

10141

Ii

94.067.043.0-

_

-

* -

.

.

v

.

,

4.1

.L B*

20.1-

s

_

PI

5.0

6.0

7.0

.

C

D

4-~w lo4

E

F

FIG. 3. Silver staining of proteins (A and B) and immunoblots (C-F) from two-dimensional gel samples of guinea pig cochlear nucleus (A, C, and E) and cerebellum (B, D, and F) containing 24 pug of protein each. Molecular mass (in kDa) and pl scales are the same for both silver stains (A and B) and blots (C-F) and are as indicated on the abscissa and ordinate, respectively (A and B). Protein 10 is indicated by large arrows; calbindin-28kDa is indicated by small arrows. (C and D) Reaction of immunoblots with protein 10 antibody (1:3000). (E and F) Reaction of immunoblots with calbindin-28kDa antibody (1:5000).

predominantly with a slightly lower molecular mass and pI protein, which appeared to be more abundant on silverstained gels of cerebellum (Fig. 3, compare E and F with A and B), although some cross-reactivity of this calbindin28kDa antibody with protein 10 was apparent (Fig. 3 E and F). Protein 10 antibody recognized pure protein 10 (0.8 ,Ug) at dilutions up to 1:100,000 at a molecular mass corresponding to calcium-binding activity but did not react with any proteins from samples of heart, duodenum, spleen, kidney, lung, liver, bladder, and testes (50 Ag each) on a onedimensional immunoblot (data not shown). Protein 10 antisera produced intense immunoreactivity of cell bodies, axons, and varicose fibers in numerous regions of the brain as observed in perfused, formalin-fixed sections of guinea pig and rat with optimal label at dilutions of 1:10001:8000 (Fig. 4). This label was completely abolished by preabsorption of antisera in 0.1-1.0 gM pure protein 10 (Fig. 4C). In agreement with results of immunoblots, no specific protein 10 antibody label was seen in any peripheral organs examined including heart, kidney, liver, lung, gut, testes, bladder, and adrenal gland. Within the brain, protein 10 antibody appeared to react with neurons within some components of every sensory system examined. In the auditory system, intense immunofluorescence was observed in the VIIIth nerve, cochlear nuclei (Fig. 4D), superior olivary complex, trapezoid bodies, and nuclei of the lateral lemniscus. Protein 10 antisera also revealed regions of abundant innervation in the vestibulocerebellar pathways (i.e., medial vestibular nucleus, floculus, nodule, and fastigial nucleus). In the cerebellum, label was apparent in rosettes, whereas the Purkinje cells remained unstained (Fig. 4C Left). In the retina, intense immunofluorescence was observed in gan-

glion cells, the optic nerve layer, bipolar cells of the inner nuclear layer, and in three terminal layers in the inner plexiform region (Fig. 4B). Within the brain, the optic nerve appeared to terminate as a dense plexus of immunoreactive fibers in the lateral geniculate body and superior colliculus. The nociceptive system was revealed by intensely immunofluorescent axons in the tractus spinalis of the trigeminal nerve and the substantia gelatinosa of the medulla oblongata and spinal cord. In addition, many intensely fluorescent cell bodies were noted in the trigeminal ganglion. The olfactory bulb contained intense fluorescence in the olfactory nerve and a variety of cells within the glomeruli and primary and accessory olfactory bulb. Sensory gustatory nerve fibers were labeled in the taste buds of the tongue. Protein 10 antisera also revealed a variety of other interesting cellular elements and neuronal circuits within the brain. In the cortex, immunohistochemical label revealed an extensive array of fine varicose fibers that appeared to emanate from cell bodies singly dispersed in all layers of the cortex (Fig. 4A). Many cells were also labeled by protein 10 antisera in the septal fimbrialis nucleus (Fig. 4E). Immunoreactivity was also apparent in several neuronal circuits such as connections between the habenula and interpreduncular nuclei by means of the fasciculus retroflexus and between the lateral mamillary nucleus and the dorsal thalamus by means of the mamillothalamic tract. Finally, the cells of the entire substantia nigra compacta (Fig. 4F), which project to the caudate putamen, also contained intense protein 10 immunoreactivity.

DISCUSSION The present findings have extended our previous results indicating greater amounts of protein 10 in the cochlear

10142

Neurobiology: Winsky et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

FIG. 4. Localization of immunoreactive nerves by protein 10 antisera (1:1000). (A) Guinea pig frontal cortex. (x 195.) (B) Rat retina. G, Ganglion cells; ON, optic nerve; ip, inner plexiform layer; in, inner nuclear layer-RC, rods and cones. (x430.) (C) Granule cell layer of guinea pig cerebellum. (Left) Overnight incubation with antisera. The large arrow points to a rosette; the small arrow points to an unstained Purkinje cell. (Right) Preabsorption of the rosette label on an adjacent section by 1 1LM purified protein 10. gl, granule cell layer; ml, molecular layer. (x540.) (D) Rat ventral cochlear nucleus (x205.) (E) Rat septum fimbrialis. (x205.) (F) Rat substantia nigra. snc, substantia nigra compacta; snr, substantia nigra reticulata. (x 140.)

nucleus as compared with other auditory and some nonauditory regions of rat, rabbit, and guinea pig brain (7, 8). Thus, the immunohistochemical and biochemical data indicate that protein 10 may represent an important calcium-binding protein within the major sensory systems as well as other discrete neuronal circuits of the brain and spinal cord. A comparison of amino acid sequences obtained from the proteolytic fragments of protein 10 with other known proteins revealed an 86% homology with a 29-kDa calcium-binding protein (calretinin) encoded by a cDNA clone from chicken retina (16) (Fig. 2). While a positive identification of protein 10 as calretinin must await the determination of complete amino acid sequences for both calretinin and protein 10, the information available in Fig. 2 strongly suggests that these may be identical proteins whose minor differences in amino acid composition are merely due to species variation. This conclusion is further supported by the similar distribution of protein 10 antibody label (Fig. 4 B and D) and calretinin in situ hybridization mRNA in retinal ganglion cells and auditory brain stem nuclei and an apparent absence of both proteins in peripheral tissues (16). The amino acid sequence data also confirmed our earlier results suggesting that protein 10 and calbindin-28kDa are different proteins (7, 8). Thus, comparison of the amino acid sequence of rat or bovine (brain) calbindin-28kDa (15, 17) with protein 10 from guinea pig revealed a 60%o homology. This is in sharp contrast with the 98% sequence homology of

amino acids between rat and bovine calbindin-28kDa (15, 17) but not unlike the homology of 47% between calretinin and the calbindin-28kDa (rat and bovine) amino acid sequences. Protein 10 and calbindin-28kDa may also be distinguished by the following: (i) Protein 10 is precipitated by 60-80% ammonium sulfate, whereas the majority of the calbindin-28kDa is precipitated by higher percentages of ammonium sulfate (18). (ii) Calbindin-28kDa and protein 10 are, for the most part, localized to different populations of cells within the brain (7, 8). For example, protein 10 is not a prominent protein on two-dimensional gels of cerebellum (Fig. 3), and antisera to protein 10 does not label the cerebellar Purkinje cells (Fig. 4C Left). In contrast, calbindin-28kDa may represent as much as 1.5% of total soluble protein in rat cerebellum where it has been localized to the Purkinje cells (3). (iii) Protein 10 does not appear to be localized to peripheral tissues (except some sensory organs, e.g., taste buds of the tongue), whereas calbindin-28kDa is present in kidney (18). The calcium-dependent elution of protein 10 was an important factor in the purification of protein 10 as outlined in this report. A similar method was employed by Maruyama et al. (18) for the purification of calbindin-28kDa from cerebellum and kidney. In that study, calbindin-28kDa was first eluted by 150 mM NaCl in the presence of 0.1 mM EGTA and then by 90 mM NaCl in the presence of 0.2 mM CaCl2 (18). Interestingly, another calcium-binding protein of 30 kDa was

Neurobiology: Winsky et al. eluted by 1.5 M NaCi in the column containing calcium and including proteins from cerebellum but was not visible during the purification of calbindin-28kDa from kidney (18). Possibly, this second protein is identical to protein 10 eluted by lower concentrations of NaCl under slightly different purification conditions [absence of 1 mM 2-mercaptoethanol used by Maruyama (18)]. The similarities between calbindin28kDa and protein 10 (i.e., 60% amino acid sequence homology and similar purification procedures) may be relevant in regard to reports showing cross-reactivity of some calbindin28kDa antisera with two calcium-binding proteins of 27-28 kDa and 29-30 kDa in brain but only the 27- to 28-kDa protein in peripheral tissue (19, 20). Thus, it is quite possible that protein 10 may be the equivalent of what was previously regarded by some investigators (19, 20) as a higher molecular mass calbindin. The results shown in Fig. 3 demonstrate that calbindin-28kDa antibodies may react with protein 10 on immunoblots. However, it should be noted that this crossreactivity between calbindin-28kDa antisera and protein 10 was not observed on immunoblots developed from rat or rabbit brain tissue. Thus, subtle species differences in protein 10 composition may alter the degree to which calbindin28kDa antibodies react with protein 10. In any case, the results presented in Fig. 3 strengthen concerns raised by Rogers (16) and others (21, 22) regarding the specificity of calbindin antibodies and suggest that some calbindin-28kDa antisera may also recognize protein 10 in immunohistochemical studies. The protein 10 antibody was shown in this report to label neurons in both guinea pig and rat. We have observed similar specific labeling in bovine, monkey, and human brain tissue, suggesting that protein 10, like several other calcium-binding proteins (1, 19, 20), may be highly conserved in mammals. The pattern of immunoreactivity produced by protein 10 antibody suggests that it may play an important role in sensory neurons, which is most likely related to its ability to bind calcium. Other cell groups in the brain (e.g., cells of the septal fimbrialis and nigral compacta), which are also labeled by protein 10 antibody, may have functional properties in common with sensory neurons. Excluding calmodulin, which appears throughout the brain and periphery, an interesting pattern appears to be emerging in the literature to indicate that different cell types in brain employ different calcium-binding proteins. For example, the S-100 proteins are predominantly found in glia while calbindin-28kDa and parvalbumin are prominent in the Purkinje cells of the cerebellum and in specific populations of hippocampal and cortical neurons (1, 3, 5, 6). A careful examination

Proc. Natl. Acad. Sci. USA 86 (1989)

10143

of the similarities and differences between these classes of cells containing different calcium-binding proteins will hopefully provide some insight as to the function of these proteins. We wish to thank Dr. K. G. Baimbridge for the generous gift of calbindin-28kDa antibody. We also thank Lisa Nguyen and Ben Hu for their assistance in the preparation of histochemical sections and Mrs. Lois Brown for the preparation of this manuscript. 1. Van Eldik, L. J., Zendegui, J. G., Marshak, D. R. & Watterson, D. M. (1982) in International Review of Cytology, eds. Bourne, G. H. & Danielli, J. F. (Academic, New York), Vol. 77, pp. 1-61. 2. Klee, C. B. (1980) in Protein Phosphorylation and Bio-Regulation, eds. Thomas, G., Podesta, E. J. & Gordon, J. (Karger, Basel), pp. 61-69. 3. Baimbridge, K. G., Miller, J. J. & Parkes, C. D. (1982) Brain Res. 239, 519-525. 4. Garcia-Sequra, L. M., Baetens, D., Roth, J., Norman, A. W. & Orci, L. (1984) Brain Res. 296, 75-86. 5. Heizmann, C. W. (1984) Experientia 40, 910-921. 6. Boyles, B. E., Kim, S. U., Lee, V. & Sung, S. C. (1986) Neuroscience 17, 857-865. 7. Winsky, L., Harvey, J. A., McMaster, S. E. & Jacobowitz, D. M. (1989) Brain Res. 493, 136-146. 8. Winsky, L., Nakata, H. & Jacobowitz, D. M. (1989) Neurochem. Int., in press. 9. O'Farrell, P. H. (1975) J. Biol. Chem. 250, 4007-4021. 10. Heydorn, W. E. & Jacobowitz, D. M. (1988) in Neurobiological Research, eds. Marangos, P. J., Campbell, I. & Cohen, R. M. (Academic, New York), Vol. 2, pp. 25-68. 11. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354. 12. Maruyama, K., Mikawa, T. & Ebashi, S. (1984) J. Biochem. 95,

511-519. 13. Vaitukaitis, J., Robbins, B., Nieschlag, E. & Ross, G. T. (1971) J. Clin. Endocrinol. 33, 988-991. 14. Jacobowitz, D. M. & O'Donohoe, T. L. (1978) Proc. Natl. Acad. Sci. USA 75, 6300-6304. 15. Yamakuni, T., Kuwano, R., Odani, S., Miki, N., Yamaguchi, K. & Takahashi, Y. (1987) J. Neurochem. 48, 1590-1596. 16. Rogers, J. H. (1987) J. Cell. Biol. 105, 1343-1353. 17. Takagi, T., Masanobu, N., Konishi, K., Maruyama, K. & Nonomura, Y. (1986) FEBS Lett. 201, 41-45. 18. Maruyama, K., Ebisawa, K. & Nonomura, Y. (1985) Anal. Biochem. 151, 1-6. 19. Pochet, R., Parmentier, M., Lawson, D. E. M. & Pasteels, J. L. (1985) Brain Res. 345, 251-256. 20. Parmentier, M., Ghysens, M., Rypens, F., Lawson, D. E. M., Pasteels, J. L. & Pochet, R. (1987) Gen. Comp. Endocrinol. 65, 399-407. 21. Sequier, J., Hunziker, W. & Richards, G. (1986) Neurosci. Lett. 86, 155-160. 22. Pasteels, B., Miki, N., Hatakenaka, S. & Pochet, R. (1987) Brain Res. 412, 107-113.