Sensitive immunoassay for rat parvalbumin: tissue distribution and developmental changes

Biochimica et BiophysicaActa, 1075( 1991) 68-74 © 1991ElsevierScience PublishersB.V. 0304-4165/91/$03.50 ADONIS 0304416591002190

BBAGEN 23579

Sensitive immunoassay for rat parvalbumin: tissue distribution and developmental changes Y u t a k a I n a g u m a , N a o m i K u r o b e , Ha~-uo S h i n o h a r a a n d K a n e f u s a K a t o Department of Biochemistry, Institute for Det'elopmentalResearch, Aichi Prefectural Colony, Aichi (Japan)

(Received 26 March 1991)

Key words: Parvalbumin;Immunoassay;Muscle;Nervoustissue A sensitive enzyme immunoassay for measurements of rat parvalbumin was established using antibodies raised in rabbits with parvalbumin purified from skeletal muscles. Antibodies in the antiserum were purified with a parvalbumin-coupled Sepharose column. The sam'wich-type immunoassay system for parvalbumin was composed of polystyret~e balls with immobilized purified antibodies and the same antibodies labeled with/].n.galactosidase from Escherichia coli. The assay was highly sensitive and the m i n i m u m detection limit was 1 pg p a r v a l b u m i n / t u b e . The assay did not cross-react with other calcium binding proteins, including h u m a n S-100a 0 and S-100b proteins, rat 28-kDa calbindin-D, and bovine calmodnlin. High concentrations of parvalbumin were observed in the skeletal muscles, especially in those composed of fast-twitch fibers, and in the diaphragm and tongue, but not in heart muscle. A relatively high concentration was estimated in the central nervous tissue. Parvalbumin was detected in the cerebral cortex and cerebellum of gestational IS-day fetuses. However, the levels of parvalbumin in the muscle tissues and central nervous tissue were very low in rats before I week of age. Thereafter, they increased sharply, reaching the adult levels by 5 weeks in most of the tissues. Parvalbumin concentrations in adult rat soleus muscle increased < 20.fold within 10 days after transection of the ipsilateral sciatic nerve, while the concentrations in the extensor digitorom Iongus muscle did not change in the same period.

Introduction Parvalbumin (M r 12000) is a Ca z÷ binding protein which possesses two active EF-hand-type calcium-binding domains in the molecule predicted from its e D N A It]. However, the biological significance of parvalbumin is still unclear, lmmunohistochemical studies revealed that parvalbumin is located in some GABAergic neurons in the hippocampus [2], visual cortex [3], geniculate nucleus [4], and horizontal cells of the retina [5]. Kawaguchi et al. [6], demonstrated that the fast spiking

Abbreviations: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis;buffer A, II.01 M sodium phosphate buffer (pH 7.(I) containing O.l M NaCI, 1 mM MgCIz, 0.1% bovine serum albumin and II.l% NAN3;buffer B. 0.01 M sodium phosphate buffer (pH 7.0) containing 0.3 M NaCI, 0.1% bovine serum albumin, 0.5% proteinase-treated gelatin and 0.1% NaN3. Correspondence: K. Kato, Department of Biochemistry,Institute for Developmental Research, Aichi Prefectural Colony, Kamiya, Kasugai, Aichi 48(}-03,Japan.

cells in the CA~ region contained parvalbumin. In skeletal muscles, type lib muscle fibers of the fasttwitch muscle were strongly immunopositive [7]. Heizmann et al. [8] showed a correlation of parvalbumin concentrations with relaxation speed in mammalian muscles, it is also suggested that parvalbumin is involved in steroidgenesis of Leydig cells in the testis [9]. For the quantitation of parvalbumin in biological samples, immunological methods including radioimmunoassay [9,10] and enzyme immunoassay [ 11,12] have been reported. The minimum detection limit of these assays ranged widely from 30 pg to 100 ng. In order to elucidate the biological role(s) of EF-hand-type calcium-binding proteins, we have established highly sensitive immunoassay methods for the measurement of S-10Oot and S-100fl [13,14], and calmodulin [15]. Here, we developed a similarly sensitive enzyme immunoassay method which can detect an amount as small as 1 pg of rat parvalbumin. By using this sensitive method, the tissue distribution and developmental changes of parvalbumin in rat central nervous tissue and muscle tissue were determined.

69

Materials and Methods

sensitivity of the assay and to protect the assay from sample interference, the assay system was prepared with the antibody fragments without Fc portion [17].

PuriJTcation of rat parvalbumin Parvalbumin was purified from frozen rat skeletal muscles according tc the method described by Endo et al. [10[ with modifications. In brieL the DEAE Sephadex fraction was further chromatographed on a column of Sephadex G-100 to eliminate the contalnihated proteins. As shown in Fig. 1, on sodium dodecyl sulfate polyacrylam~,de slab gel electr')phoresis (SDSPAGE), the final preparation showed a single band at a position corresponding to the molecular weight of 12000.

Antibodies The antisera were raised in Japanese white rabbits by the subcutaneous injection of purified parvalbumin (1 mg/rabbit) emulsified with an equal volume of Freund's complete adjuvant. The immunization was repeated four times every 2 weeks. Then, the rabbits were bled once a month 1 week after the booster immunization. The antibody lgG was purified from antisera by the use of purified parvalbumin-coupled Sepharose 4B (about 10 mg of parvaloumin were coupled with 2 g of CNBr-activated Sepharose 4B from Pharmacia Fine Chemicals, Tokyo) in the same manner as described previously [16]. About 40 mg of the antibody lgG were obtained per 100 ml of the antiserum. The specificity of the antibodies thus prepared was tested by the immunoblotting test with a crude extract of rat leg muscles. As shown in Fig. lB, the extract transferred to a nylon membrane showed a single immunoreaetive band at a position corresponding to that of parvalbumin. The purified antibody lgG was digested with pepsin from porcine stomach mucosa (Sigma Chemicals St. Louis, U.S.A.) to obtain the F(ab') 2 fragment of antibody. In order to increase the

Priciple of the present im,ta¢noassay The present method is based on the priciple of sandwich enzyme immunoassay. In brief, parvalbumin in the sample reacts with the solid-phase antibody through one of the antigenic sites on the parvalbumin molecules. After washing out the unreacted material in the sample, galactosidase-labeled anti-parvalbumin antibody is allowed to react with the other antigenic sites of parvalbumin molecules bound to the solid-phase. Unbound labeled antibedy is washed out. Finally, the enzyme activity bound on the solid-phase is measures by incubation with the fluo~'ogenic substrate 4-methylumbellifcryl-/3-D-galactoside.

Immunoassay reagents The assay system consisted of polystyrene balls with immobilized antibody F(ab')_, fragments and the same antibody Fab' fragments labeled with B-D-galactosidase from Escherichia coil The antibody F(ab')_, fragments were immobilized noncovalently on polystyrene balls (3.2 m m in diameter, i m m u n o Chemicals, Okayama, Japan) as described previously [18]. To stabilize the assay, the balls with antibodies were kept at 4°C for at least overnight in 0.01 M sodium phosphate buffer (pH 7.0) containing 0.1 M NaCI, 5 m M MgCI z, 0.1% bovine serum albumin (demineralized, Organon Teknika, The Netherlands), and 0.1% NaN 3 (buffer A). Antibody Fab' fragments, obtained by the reduction of the F(ab')2 with 2-mercaptoethylamine, were coupled with /3-o-galactosidase (Boehringer Mannheim GmbH, F.R.G.) by the use of o-phenylenedimaleimide (Aldrich Chemicals, U.S.A.) as described previously [19,20], and the amount of the labeled antibodies was defined as the unit of galaetosidase activity (one unit = 1 v.mol p r o d u c t / m i n under the conditions described below). The antibody Fab' fragments were also coupled with horseradish peroxidase by the use of Nsuccinimidyl 4-(N-maleimidomethyl) cyclohexane-lcarboxylate (Zieben Chemicals, Tokyo, Japan) as described by Yoshitake et al. [21], and the conjugate was used for the immunoblotting test.

lmmunoassay procedures ,= ..

2

3

2

3

Fig. 1. SDS-PAGE of the purified rat parvalbuminand an extract of rat leg muscle (Ah and immunoblotswith purified anti-parvalbumin (B). Lane I, standard proteins with molecular mass in kDa; lane 2, 1.5 V,gof purified rat parvalbumin"lane 3, a extract of rat leg muscle containing 1.5 #g immunoreactiveparvalbuminwith 85 v,g proteins. A, polyacrylamidegel stained with Coomassie blue; B, nylon membrane stained with purified antibody Fab' labeled with peroxidase.

The procedures are the same as those described for the assay of S-100 proteins [13,14]. In brief, a piece of polystyrene hall with immobilized antibodies was incubated with 10/xl each of the standard parvalbumin or samples at 30°C for 5 h with shaking in a final volume ot 0.S ml with buffer B (0.01 M sodium phosphate buffer (pH 7.0) containing 0.3 M NaCI, 0.1% bovine serum albumin, 0.5% proteinase-treated gelatin [22],

7O and 0.1% NAN3). After washing, each ball was incubated at 5°C overnight in a fresh test tube containing 0.2 ml buffer A with 1 m U of the galactosidase-labeled antibody Fab'. The galactosidase activity bound on the ball was assayed with a fluorogenic substrate, 4-methylumbeUiferyl-fl-t~-galactoside (Sigma) as described previously [18].

described by Schaffner and Weissmann [26]. H u m a n S-100a 0 [27] and S-100b [13] proteins were purified as described previously. Rat 28-kDa calbindin-D was purified from frozen kidneys according to the method described Hitchman et al. [28]. Bovine calmodulin was obtained from A m a n o Pharmaceuticals, Nagoya, Japan. Results

Tissue extracts Male Wistar rats were used for most of the present study. The fetuses were obtained from female Wistar rats after mating, and the date of plug detection was designated gestational 0 day. Blood samples were obtained with a heparinized syringe from the abdominal aorta under diethyl ether anesthesia. Blood plasma, lymphocyte fractions, and erythrocytes were separated by centrifugation using Ficoll-Paque (Pharmacia AB, Sweden). Denervated rat leg muscles were prepared by transecting the unilateral sciatic nerve trunk at the middle part of the thigh under anesthesia with pentobarbital. The tissues were sampled immediately after decapitation, and stored at - 2 0 ° C for a few days until analysis. The frozen tissue was homogenized at 0°C in a glass homogenizer with a 10 vol. (v/w) of 50 m M Tris-HCI buffer (pH 7.5), containing 5 m M EDTA. The homogenate was sonicated at 0°C for 1 min, and centrifuged at 1 4 0 0 0 x g for 30 min. The supernatant fraction was used for the assay of parvalbumin and protein concentrations. Electrophoresis and immunoblot SDS-PAGE was performed by the use of a Tricine electrophoresis system with a 10% acrylamide gel containing 3% bisacrylamide as described by Sch[igger and yon Jagow [23]. Proteins separated in the gel were visualized with Coomassie blue. Immunoblots were performed with anti-parvalbumin Fab' labeled with horseradish peroxidase. In brief, proteins in the polyacrylamide gel were transferred to a nylon membrane (NYTRAN, Schleicher & Schuell, Dassel, F.R.G.). The membrane with proteins was fixed in the phosphatebuffered saline containing 2% glutaraldehyde. After being washed with the phosphate-buffered saline, the membrane was incubated at room temperature successively for 30 rain with buffer B and for 1 h with the peroxidase-labeled anti-parvalbumin Fab' (0.06 p,g/ml) in the same buffer. The peroxidase activity on the sheet was visualized with 3,3'-diaminobenzidine and H202 as described previously [24]. Other methods Protein concentrations of tissue extracts were estimated with Bio-Rad protein assay (Bio-Rad, Richrpond, CA, U.S.A.), which utilizes the principle of protein-dye binding [25]. Concentrations in p,g of the purified parvalbumin were determined with the method

Detection limit, specificity, and precision of the immunoassay for rat parcalbumin A standard curve for the assay of rat parvalbumin and its cross-reactivity with other caicium-binding proteins is shown in Fig. 2. The assay was highly sensitive, and the minimum detection limit, defined as the lowest amount of parvalbumin giving a galactosidase activity significantly different from that of the 0 standard at 0.99 confidence, was < 1 p g / a s s a y tube. The assay was specific to parvalbumin, showing no reactivity with other EF-hand-type calcium-binding proteins, including h u m a n S-100a 0 ( a a form) and S100b (/3/3 form), bovine calmodulin and rat 28-kDa calbindin-D (Fig. 2). In order to confirm the proportionality of the parvalbumin assay with tissue extracts and to determine the optimal sample volume of each extract, several tissue extracts were diluted serially with buffer A, and the diluted extracts were subjected to the immunoassay. As shown in Fig. 3, the values estimated in each

i ~=

i+ o

o.1

1 lO leo Parvalbumln (p9)

Fig. 2. Standard curve of the assay for parvalbumin and its cross-reactivity with other calcium-bindingproteins. Indicated amounts of rat pawalbumin (o), human S-100ao (z~), human S-100b (v). rat 28-kDa ealbindin-D(o). or bovine calmodulin([2) were subjected to the immunoassayin duplicate as described in the text. /3-D-Galactosidase activity bound on the polystyrene ball is expressed as the fluorescenceintensityof 4-methylumbelliferoneproduced in., 20-rain reaction at 30°C with 0.1 mM 4-methylumbelliferyI-FD-galactoside. In the fluorescenceintensityscale, to00 equals 1 pM 4-methylumbelliferone.

TABLE I

D~'lt'rtmnaliorl of parvalbumin in various rat tixsttt's

i°:::I

Extract added

(nil

/

Fig. 3. Effect of sample dilution on the assay of parvalbumin in the crude extract. The soluble fraction of Ill':;- (w/v) homngenate from rat vastus lateralis muscle (|L diaphragm (2}, tongue ¢3). cerebral cortex (4-), or bladder (5} was diluted differently, and the samples containing the indicated ',plumes of original cxttat.i ~ere r,ubiccted to the immnnoassay. Each point is the mean of a duplicate assay.

extract were proportional to the s a m p l e volume employed in the assay. T h e accuracy of the assay was e x a m i n e d by assaying five rat p l a s m a s a m p l e s with or without 300 pg stand a r d parvalbumin. The analytical recovery of a d d e d p a r v a l b u m i n was 92 ± 8%. The precision of the assay was tested by assaying five s a m p l e s of cerebral cortex extract 20 times in one assay (within series) or the same s a m p l e s in d u p l i c a t e in ten consecutive assays (between series). The coefficients of variation were 4.9 to 6%, a n d 5 to 9.1%, respectively. T h e s e results indtcated that the present assay was applicable to the assay of p a r v a l b u m i n in the crude extract.

Distribution of immunoreactit'e part'albumin in rat tissues The c o n c e n t r a t i o n s of parvalbumin in various tissues w e r e d e t e r m i n e d in s a m p l e s o b t a i n e d from adult male rats. Parvalbumin was d e t e c t e d in most tissues examined. The results are s u m m a r i z e d in T a b l e 1. In the central nervous system, c e r e b e l l u m c o n t a i n e d about 2 / x g / m g protein of parvalbumin, and the conc e n t r a t i o n s in the b r a i n s t e m a n d cerebral cortex were about 0.6 and 0.2 / z g / m g protein, respectively. The h i p p o c a m p u s c o n t a i n e d less than 0.15 /~g parvaibum i n / m g protein. T h e vastus lateralis muscle, composed mostly of fast-twitch fibers, c o n t a i n e d an extremely high level ( 3 0 - 4 0 p . g / m g protein} of parvalbu rain. In contrast, p a r v a l b u m i n in the slow-twitc~ :,~leus muscle was e s t i m a t e d to be less than 0.2 ~ g / n , ~ protein. T h e d i a p h r a g m and t o n g u e also possessed relatively high levels of parvalbumin. Howe~er, the heart muscle and s m o o t h muscles in the intestine and bladd e r s h o w e d low p a r v a l b u m i n concentrations. Significant a m o u n t s of p a r v a l b u m i n were found in the adipose and cartilaginous (xiphoid) tissues (about 0.l / * g / m g protein). Parvalbumin was d e t e c t e d in rat blood p l a s m a (n = 5, 6.8 ± 1.9 n g / m l ) , but not in the erythrocyte a n d lymphocyte fractions.

Tissues Cerebral cortex Flippocampus Cerebellum Brainstem Lung |leart Liver Spleen Kid.".'."..' Intestine Bladder Adrenal Testis Epididymal adipose Xiphoid Tongue Diaphragm ~,~l,:usmuscle Va:a,.,~ lateralis muscle Blood plasma

Concentrations of parvalbumin (ng/mg soluble protein) 212 _+ 32 " 136 17 2080 --¢- 206 639 + 56 8.45 + L0S (}.S9-+ 0.24 0.48 _+ (I.24 1.49_+ 0.29 11.62--+ 0.13 8.76-+ 3.69 26.6 _+ 6.54 7.25_+ 1.47 4.27-+ 0.21 I I I _+ 39 ll)O + 26 I~,90 + 118 6940 4r | It3() 69.1 + 20.6 33000 _+2210 6.82_+ 1.86 t,

" Mean+S.D. of five male Wistar ra,~. h ng/ml.

Developmental changes of parvalbumin in the central herr'pus tissue Parvalbumin c o n c e n t r a t i o n s were d e t e r m i n e d in rat central nervous tissues from fetus ( 1 3 - 1 9 gestational days) to adult rats of 9 w e e k s of age. Parvalbumin was d e t e c t e d in the cerebral cortex a n d c e r e b e l l u m of 13to 15-day-old fetuses (n = 5; 0.031 + 0.004 n g / m g protein, and 2.9 ± 1.0 r i g / m s protein, respectively) and 17to 19-day-old rat fetuses (n = 5; 0.045 4- 0.008 n g / m g protein, and 5.7 ± 0.4 r i g / m s protein, respectively). The c e r e b e l l u m and b r a i n s t e m from newborn rats of 0 to 3 days of age c o n t a i n e d 7 - 1 0 n g / m g protein of parvalbumin. Thereafter, the parvalbumin in the two regions increased sharply, reaching p l a t e a u levels at 5

i A g e in weeks

Fig. 4. Developmental profiles of parvalbumin in rat cerebral cortex (o). cerebellum (el, and brainstem ( • ). Fach point is the mean 5: S.D. of five rats.

[]

[•

12

1.2

o.8

~4 0

o:

2

4 g Age in weeks

g

10

Fig. 5. Developmental profiles of parvalbumin in rat tongue ( [ ] ) and diaphragm ( l ) . Each point is the mean+_ S.D. of five rats.

weeks of age (Fig. 4). Parvalbumin levels in the cerebral cortex from rats 0 to 3 days old were 0.07-0.09 n g / m g protein, and at 6 days of age, they were 0.25 + 0.08 n g / m g protein, being much lower than those in the cerebellum and brainstem. However, thereafter, the concentration increased as in the brainstem and cerebellum, reaching the adult level at 5 weeks of age (Fig. 4).

Det'elopmental changes or part,albumin in the muscle tissues The developmental changes in the tongue and diaphragm are shown in Fig. 5. The concentrations in the tongue showed relatively high values (about 200 n g / m g protein) even in the newborn rats, and similar levels were maintained for 2 weeks after birth. Parvalbumin in the tongue then increased sharply for 2 weeks before reaching a plateau level. Parvalbumin in the diaphragm showed a similar profile: a relatively constant level ( < 200 n g / m g protein) for 2 weeks after birth, followed by a linear increase. However, the parvalbumin in the rat diaphragm continued increasing for up to 9 weeks after birth.

AI .~

~E

"f

4o

o 4 Denelvation period (days)

8

Fig. 7. Effect of denervation on parvalbumin concentrations in rat

soleus muscle (A) and extensor digitorum Iongusmuscle (B). At the indicated day after unilateral transection of the sciatic nerve trunk, rats were killed and palvalbumin concentrations in the ipsilateral muscles ( • . o) and contralateral control muscles ([]. ©) were measured. Each point is the mean + S.D. of five rats. Fig. 6 shows the developmental changes of the parvalbumin in the soleus, rectus femoris, and extensor digitorum Iongus muscles. Concentrations of parvalburain in the soleus, rectus femoris, and extensor digitorum longus muscles from rats 1 week old had average values of 7.2, 287, and 73 n g / m g protein, respectively, indicating that the low concentrations in the slow-twitch soleus muscle were already evident even at this stage. However, developmental increase of parvalbumin in the soleus and rectus femoris muscles showed similar patterns, reaching plateau levels at 4 weeks of age. The change in the concentration of parvalbumin in the extensor digitorum longus muscle was apparently different, showing the maximum level at 3 weeks of age, and thereafter the concentration decreased.

Effect of denert'ation on the part.albumin concentrations of the soleus muscle and extensor digitorum Iongus muscle

:°t8 o o4

I[-

o

,~

.....

Age in weeks Fig. 6. Developmental profiles of parvalbumin in rat rectus femoris muscle ( a ), extensor digitorum longus muscle (o), and soleus muscle

(U). Each point is the mean + S.D. of five rats.

The unilateral sciatic nerve trunk of 10-week-old rats was transected. After 3, 6 and 10 days, rats were killed and the ipsilateral soleus muscle and extensor digitorum muscle were sampled. The control muscles were sampled from the contralateral intact leg. As shown in Fig. 7, the parvalbumin concentrations in the extensor digitorum Iongus muscle did not change within 10 days of denervation, showing constant values similar to the controls. In contrast, the concentrations in the soleus muscle increased almost linearly, reaching 1.8 + 0.06 p . g / m g protein after 10 days, < 40-fold higher levels than those (0.039 4-_0.010 p . g / m g protein) in the contralateral control soleus muscle.

Discussion For measurement of parvalbumin in the extract, immunoassay methods have been developed, and the

minimum detection limit of these methods was 30 pg to 100 ng [9-12]. The assay method we established here is a sandwich-type immunoassay, which uses purified and monospecific antibodies. The assay was highly sensitive, as that for S-100 proteins [13,14], and the minimum detection limit was < 1 p g / t e s t tube. The assay was specific to parvalbumin, showing no cross-reactivity with other Ca2+-binding proteins, such as S-100o~ and S-100/3 proteins, 28-kDa calbindin-D, and calmodulin, which are thought to be evolutionarily related to parvalbumin. However, it is reported that antibodies raised with rat parvalbumin reacted with parvalbumin from other animal species [29]. In fact, the present assay system could measure the concentrations of parvalburain in human serum and cerebrospinal fluid (not shown). By employing the present method, it was confirmed that parvalbumin was present mainly in the skeletal muscle and nervous tissues in the rat. The highest concentration was observed in the skeletal muscles composed of fast-twitch fibers. In contrast, the concentration in the slow-twitch soleus muscle was less than 1% that of the fast-twitch muscles, such as the vastus lateralis muscle, rectus femoris muscle, and extensor dig;forum Iongus muscle. Among the central nervous ti' :cs, the cerebellum contained the highest level of parvalbumin, and the concentration in the cerebral cortex or hippocampus was about 10% that of the cerebellum. Endo et al. [30] reported that parvalbumin was not detectable immunohistochemically and immunochemically in the newborn rat cerebellum. However, with the present sensitive method, parvalbumin was measurable not only in the cerebellum hut also in the cerebral cortex of rat fetuses ot the late gestational stage (13-15 days). Parvalbumin in the rat central nervous system, however, showed a low level ( < 20 n g / m g protein) until I week after birth, and then it increased linealy as described by Endo et al. [30]. The developmental changes in the concentrations of parvalbumin in rat skeletal muscles determined in the present study were essentially the same as those reported previously in the rabbit [11] and mouse [12]. Berehtold and Means [31] reported that the mRNA for parvalbumin began to increase at the 5th day of age in the gastrocnemius muscle, and reached a 15- to 20-fold higher level by the 20th day after birth. These reports are compatible with the present results. Leberer and Pette [11] reported that the denervation did not affect the parvalbumin content of the adult rabbit soleus muscle, but it decreased after a delay of about 2 weeks in the parvalbumin content of fast-twitch tibialis anterior muscle of the same animal, and the parvalbumin concentration of the tibialis anterior muscle was similar to that normally found in the soleus muscle after 75 days of denervation. The changes in

parvalbumin concentrations during a short period of denervation in the rat soleus and extensor digitorum Iongus muscles observed in the present study were apparently different from those observed in rabbits. It still remains to be clarified for how long parvalbumin in the denervated rat soleus muscle keeps increasing its concentration. It has been suggested immunohistochemically that parvalbumin exists in the organs as well as in muscles and nervous tissues. In the present study, significant amounts of parvalbumin were detected in adipose tissue and xiphoid (a cartilaginous tissue). Detectable amounts of parvalbumin were present in most of the tissues examined, except for blood cells. These results might provide a clue to the understandiBg of the physiological function(s) of parvalbumin. Acknowledgments This work was supported in part by a Grant-in-Aid for Science Research on Priority Areas (Molecular Basis of Neural Connection) and a fund for basic experiments oriented to space station utilization, administered by the Ministry of Education, Science and Culture of Japan. References 1 Berchtold, M.W., Epstein, P., Beaude(, A., Payne, M,W., Heizmann, C,W. and Means, A.R. (1987) J. Biol. Chem. 262. 8696S701. 2 Kosaka, T., Katsumaru,H., Hama, K., Wu, J.-Y. and Heizmann, C.W. (19871 Brain Res. 419, 119-130. 3 Kosaka, T., Heizmann, C.W., Tateishi, K., Hamaok=, Y. and Hama, K. (1987) Brain Res. 409, 403-408. 4 Stichel, C.C., Singer, W. and Heizmann, C.W, (1988) J, C0mp. Neurol. 268, 29-37. s Rohrenbeck,J., Singer,W. and Heizmann,C.W. (1987)Neuroscl. Lett. 77, !~;5-260, 6 KawaguchiY., Katsumaru,H., Kosaka,T., Heizmann, C.W, and Hama. K. (1987) Brain Res. 416, 369-374. 7 Cello, M.R. and Heizmann, C.W, (1982) Nature 297, 504-506. 8 Heizmann, C.W., Berchtold, M,W. and Rowlerson,A.M. (1982) Proc. Natl. Acad. Sci. USA 79, 7243-7247, 9 Kiigi, U., Berchtold, M.W, and Heizmann, C.W. (1987) J. Biol. Chem. 262, 7314-7320. l0 Endo, T.. Takazav,'a, K. and Onaya, T. (1985) Endocrinology117, 527-531. 11 Leberer, E. and Pette, D. t1986) Biochem,J. ~t5, 67-73. 12 Sano. M.. Yokota, T.. Endo, T, and Tsukagoshi, H. (1990) J. Neurol. Sci. 97, 261-272. 13 Kimura.S., Kato, K., Semba, R. and Isobe,T, t 1984)Neurochem, Int. 6. 513-5|8. 14 Kato. K.. Suzuki. F.. Kurobe. N., Okajima, N., Nagaya, M. and Yamanaka, T. (1990) J. Mol. Neurosci. 2, 109-113. 15 Kitajima, S.. Seto-Ohshima.A., Sano, M. and Kato, K. (1983) J. Biochem. 9a 550-564. 16 Kato, K., Suzuki, F. and Semba, R. (1981) J. Neurochem. 37, 998-1005. 17 Kalo, K., Umeda, Y , , Suzuki. F. and Kosaka. A, (1979~ J. Appl. Biochem. 1,479-488.

18 Kato, K., Hamaguchi, Y., Okawa, S., Ishikawa, E., Kobayashi, K. and Katunuma, N. (1977) J. Biochem. 81,1557-1566. 19 Kato. K., Fukui, H., Hamaguchi, Y. and Ishikawa, E. (1976) J. Immunol. 116, 1554-1560. 20 Kato, K. (1983) Methods Enzymol. 92, 345-359. 21 Yoshitake, S., lmagawa, M., Ishikawa, E., Niitsu, Y., Urushizaki, I., Nishiura, M., Kanazawa, R., Kurosaki, H., Tachibana. S., Nakazawa, N. and Ogawa H. (1982) J. Bioehem. 92, 1413-1424. 22 Kato K., Umeda, Y., Suzuki, F. and Kosaka, A. (1980) Clin. Chim. Acta 120, 261-265. 23 Seh~igger, H. and yon Jagow, G. (1987) Anal. Bioehem. 166, 368-379. 24 Haimoto, H., Kurobe, N., Hosoda, S. and Kato, K. (1989) Clin. Chim. Acta 181, 27-36.

25 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. 26 Schaffner, W. and Weissmann, C. (1973) Anal. Biochem. 56, 502-514. 27 Kato, K., "£imura, S., Haimoto, H. and Suzuki, F. (1986) J. Neuroehem. 46,1555-1560. 28 Hitehman, A.J.W., Kerr, M.-K. and Harrison, J.E. (1973) Arch. Biochem. Biophys. 155, 221-222. 29 Berchtold, M.W., Celio, M.R. and Heizmann, C.W. (1985) J. Neurochem. 45, 235-239. 30 Endo, T., Kobayashi S. and Onaya, T. (1985) Neurosci. Lett. 60, 279-282. 31 Berchtold, M.W. and Means, A.R. (1985) Proe. Natl. Aead. Sei. USA 82, 1414-1418.