Immunological identification of two proteoglycan fragments derived from neurocan, a brain-specific chondroitin sulfate proteoglycan

Pergamon

0197-0186(94)00085-9

Neurochem Int Vol 25, No 5, pp 425-431,1994 Copyright © 1994ElsevierSoeneeLtd Pnntedin Great Britain All rightsreserved 0197-0186/94$7 00+0.00

IMMUNOLOGICAL IDENTIFICATION OF TWO PROTEOGLYCAN FRAGMENTS DERIVED FROM NEUROCAN, A BRAIN-SPECIFIC CHONDROITIN SULFATE PROTEOGLYCAN FUMIKO MATSUI*, EIJI WATANABE a n d ATSUHIKO OOHIRA Department of Pennatoiogy, Instatute for Developmental Research, Kasugm, Aichi 480-03, Japan (Received 13 April 1994, accepted 18 May 1994)

Abstraet--Nenrocan Is a brain-umque chondroRin sulfate proteoglycan (CSPG) whose expression and proteolytac cleavage are developmentally regulated. One of the proteolytic products (C-terminal hal0 ts known to be a CSPG with a 150 kDa core glycoprotem (CSPG-150). To identify the N-terminal half of neurocan, we raised an anti-nenrocan polyclonal antibody (PAb 291) using a synthetic peptade whose amino acid sequence matched a part of the N-terrmnal half of neurocan. Western blots showed that PAb 291 recognized two CSPGs, one with a 220 kDa core glycoprotein (CSPG-220, namely neurecan) and one with a 130 kDa core glycoprotem (CSPG-130) isolated from young rat brains. CSPG-130 was co-purified along with CSPG-220 by PAb 291-immunoaflimtycolumn chromatography. The amino acid sequence of the N-terminus of the immunopurified CSPG-130 was exactly the same as the N-terminal sequence of CSPG-220 These results suggest that not only the C-terminal half (CSPG-150) but also the N-terminal half (CSPG-130) of CSPG-220 exists m a CSPG form in rat brain. Using PAb 291 and moneclonal antibody 1G2 (MAb IG2) which recogmzes CSPG-150 in addition to CSPG-220, we found that the contents of CSPG-130 and CSPG-150 in the rat brain reached maximum levels around the ttme of birth. Both CSPG-130 and 150 were observed, while CSPG-220 was hardly detectable in extracts from the adult rat brain. Immunohistochemical investigation showed that the PAb 291 antagenhad a similar distribution pattern to the MAb IG2 antigen. PAb 291 and MAb 1G2 are proposed to be useful tools for studying the mechamsm and biological significanceof the cleavage of neurocan.

hybridized with a single band at ~ 7.5 kb on Northern blots of m R N A from 4-day-old and adult rat brain. We previously isolated chondroitin sulfate proteoglycans (CSPG mixture) with different molecular sized core glycoproteins (250, 220, 150, 130 and 93 kDa) from the PBS soluble fraction of rat brain (termed CSPG-250, 220, 150, 130 and 93, Oohira et al., 1988). We raised a monoclonal antibody (MAb I G2) which recognizes both CSPG-220 and CSPG150 and showed that CSPG-220 corresponds to the juvenile type and CSPG-150 corresponds to the adult type of neurocan by analyzing the amino acid sequences of N-termini of their core glycoproteins (Oohira et al., 1994). The developmentally regulated cleavage of neurocan must be of some biological significance. Thus, the function of both CSPG-220 and its proteolytic derivatives should be investigated, along with the mechanism of cleavage at the specific site of the core protein. However, because there was no immuno*Author to whom all correspondence should be addressed probe capable of recognizing only the N-terminal half 425

Proteoglycans are one of the major constituents of the extracellular matrix of various animal tissues including neural tissues. They have been suggested to play important roles in various cellular processes such as cell proliferation, neurite formation and cell adhesion (for reviews, see Maeda and Oohira, 1991 ; Margolis and Margolis, 1993) In the rat brain, 25 putatwe proteogiycan core proteins have been identified (Herndon and Lander, 1990). Neurocan is a CSPG species unique to the brain whose complete coding sequence was reported by Rauch et al. (1992). The molecular sizes of the core glycoproteins of the juvenile and adult types of neurocan are 245 and 150 kDa, respectively, the latter being the C-terminal half of the former. The adult type of neurocan is considered to be a product of a proteolytic processing of the juvenile type because a riboprobe corresponding to a region of neurocan

426

l

[ MIK() M A I S [ I

o f n e u r o c a n , zdentzficaUon a n d purification o f th~s p r o t e o l y t l c cleavage p r o d u c t have n o t 3.el been perf o r m e d In the p r e s e n t stud.L we rinsed a polyclonal a n h b o d y ( P A b 291) w h i c h r e c o g m z e s a p e p U d e w h o s e a m i n o acid s e q u e n c e is c o n t a i n e d w~thln the N - t e r minal h a l f o f n e u r o c a n U s i n g P A b 291, we f o u n d t h a t C S P G - 1 3 0 c o r r e s p o n d e d to the N - t e r m i n a l h a l f o f n e u r o c a n P A b 291 was used for the purtficaUon o f C S P G - 1 3 0 by ~ m m u n o a d s o r p t t o n c o l u m n c h r o m a t o graphy and m a subsequent lmmunocytochemlcal study EXPERIMENTAL PROCEDURES

Produ~ non o/antthodte~ An antl-neurocan monoclonal antibody (MAb 1G2) was produced as described prevxously (Oohlra et al, 1994) and ascltes contalmng M A b IG2 were produced m crossbred FI progeny of BALB/c and C3H strain mice (SLC Inc. Shlzuoka, Japan) An antl-neurocan polyclonal anubody (PAb 291) was rinsed using a synthetic peptlde, K G L N G R H F Q Q Q G P E D Q (amino acids 483-498, see Fig 1), m Japanese white rabbits Multiple antlgemc peptlde (MAP) was synthesized according to the method of Posnett and Tam (1989) Rabbits were lmmumzed by subcutaneous najecUons with 0 5 nag of MAP per ,mmumzauon The ant*gen was admimstered w~th Freund's complete adjuvant for the first rejection and w,th Freund's incomplete adjuvant for booster mjectmns The mjecuon was performed once a month and rabbits were bled 10 days after rejection Tissue preparanon /or 14e.~te~n hlottmq Cerebral t~ssue (100 mg wet we*ght) ot 10-day-old Sprague-Dawley rats (SLC lnc, Sh~zuoka, Japan) was homogemzed m 300/tl of ~ce-cold PBS containing 20 m M EDTA. l0 mM N-ethylmale,m~de (NEM) and 2 mM phenylmethylsulfonyl fluoride (PMSF) with a tight-fitting glass

Amino acid No.

0 [__

200 I

400 J

i~"~\~

Teflon homogemzei An ahquot (100 itl) ol Ihc homogenatc ~ a s t m x e d ~lth 400/~1 of 2% SDS 50raM l)ls H(I. plf 7 5, containing 20 mM EDTA. 10 mM NEM and 2 mM PMSF and the mixture wa~ immediately boded |ol q mm (brain homogenate) An ahquot ( 100 id) of the ht)mogena(c was centrifuged at 6000 g for 20 mm at 4 ( Fhe resultant supernatant was mixed with 400 Id of 2 ° o SDS qo mM Tl is HCI. pH 7 5, containing the protease inhibitor., and the mixture was immediately boiled l'oi 5 mm (blare PBSextract) To an ahquot of the brain PBS-extrd~.t. 3 ~ol ol 95% ethanol containing 1 3°,0 pota~,slum m_etatc ~ele added and the precipflated materml x~.ts digested v,~th protea..cfree chondroitma..,e ABC ( E( 4 2 2 4. Selkag,lktt Kog,,,o (,) Yokyo) To ahquots (13 5 ,ul/lane) of the brain homogenate, the brain PBS-e~tracts and the brain PBS-extract d~gested '~ lth chondromnase ABC, 3 vol of 95% ethanol containing 1 3". potassium acetate were added and the precipitated materials were re~olved by S D S - P A G E on a 3°,, stacking gel and a 6°'0 separating gel Alter electrophoresis, materials m SDS gels were electrotransferred onto polyvmyhdene dlfluonde (PVDF) membranes and lmmunoreactlve materials were detected by staining with MAb IG2 or PAb 2~1 using a Vectastam ABC Kit (Vector L a b , Burhngame, CA. L S /k ) The temporal expressmn patterns of CSPG-220, -150 and -130 were mvesUgated by immunoblottlng using brain PBSextracts digested with chondro~tmase ABC according to thc method which we described pre~musly (Oohlra t ~ al 1994) The intensity of lmmunostammg was quantified ()u ,t Ma,.lntosh computer using the NIH Image program l~o[atton o/proteoqlycan s bl' tmmunoalflmt ~ chromatoqt aphl Immunoglobuhn (IgG) was purified l~om mouse ascltes and rabNt serum by chromatography using Protein A agarose from a Affi-Gel Protein A MAPS kit (Blo-Rad, R~chmond, CA, U S A ) lmmunoglobuhn thus obtained ~as conjugated to AN-Gel H7 gel (BIo-Rad) according to the method recommended by the supplier Tablc t ,hows the amount~ of antibodies conjugated to the gel and antigen,, obtamed The PBS-soluble CSPG mixture was purified trom the

600 [

S)gnal Ig-hkeTandem pep~ , ~ ~ juvenile type

('l ~l/

800 )

Chondroltln sulfaten domai

10o0 I

12o0 1257 _[___J

EGF- Lectm- Complement like'N Ill.like , lregulatoryike

)] KGLNGRHFQQQGPE;Q

483

a u,, pe

498

.

llNI

t

639

130K

1 50K

Fig 1 Structure ofjuvemle and adult types ofneurocan The amino acid sequence of the peptlde used to produce PAb 291 ts shown under the figure of the juvemle type of neurocan

427

Immunological identification of two neurocan fragments Table 1. Conjugation of annbodses (Ab) to Afli-GeiHz gel and purification of proteoglycans by immunoadsorptaon

Ab MAb 1(32 PAb 291

Ab mtxed with gel (rag)*

Gel rmxed with Ab (ml)

Ab conjugated with gel (rag)

CSPG applied to gel (nmol UA)t

CSPGs adsorbed (nmol UA)

7 5.9

2 2

21 3.1

400--800 100-200

20-90 3-5

*The mount of protein was determined using a Bio-Rad Protein Assay Kit (Bio-Rad, Rw,hmond, CA, U.S.A.). ~"Hexuronate (destgnated as UA) was determined by the method of B~tterand Mmr (1962) brains of 10-day-old Sprague-Dawley rats as described previously (Oohira et aL, 1985) and dissolved in 4 ml of PBS containin~ 0.5 M NaCI, 0.1% 3-[(3-cholamidopropyl) dimethylammonio]-l-pr~onate (CHAPS, Katayama Chemical Industries Co., Ltd., Oulm, Japan), 2 mM EDTA, 1 mM NEM and 0.2 ram PMSF (solution-A). The CSPG mixture solution was added to 2 ml of antibody-conjugated gel, mixed gently overnight at 4°C and then poured into columns. After flow-through fractions were eluted, the column containing MAb IG2-conjugated gel was washed according to the mothod of Zoller and Matzku (1976) with modifications. Briefly, the column was successively washed wtth 10 ml of solution-A and with 20 ml of 2 M NaC1 contaimug 0.1 M NaOH-giycine, pH 10.0, 0.1% CHAPS, 2 mM EDTA and 1 mM NEM (solution-B). The column containing PAb 291-¢,onjugated gel was washed only with 10 ml of solution-A because a pomon of the antigens which bound to PAb 291 was eluted with solution-B. Matenals bound to the gel were eluted from the column with 20 ml of 3 M KSCN containing 0.1% CHAPS, 2 mM EDTA and 1 mM NEM. The flow rates were 10 ml/h for flow-through and elution and 20 ml/h for washing An aliquot of each fraction was assayed for immunoreactivity with MAb 1G2 or PAb 291 by dot-blot analysts. The fracttons containing proteoglycans (6-7 ml) were concentrated to < 1 ml voth a Centriflo CF50 (Amicon, Lexington, KY, U.S.A.) and 3 vol of 95% ethanol containmg 1.3% potassium acetate were added to the concentrated sohiuon to precipitate proteoglycans at 0°C. After 1 h, the precipitate was collected by centrifugauon at 13,000 g for 30 win at - 5°C. After washing the preopitate voth 70% ethanol- 1% potasstum acetate, proteogiycans were resolved by SDSPAGE using a 3% staclung gel and a 6% separating gel, both before and after dtgestion with chondroltinase ABC as desenbed above.

Amino actd sequence analysts For amino acLd sequence analysis, core glycoproteins were prepared from purified proteoglycan preparations by chondroitmase ABC digestion. They were first separated by SDSPAGE and then transferred to PVDF protein sequencing membranes as described prevaously (Oohira et al, 1994) Amino acid sequencing was performed using a Shimadzu PPSQ-10 protein sequencer. lmmunohzstochemistry Seven-day-old Sprague-Dawley rats were deeply anesthetized wtth pentobarbital and perfused transcardially wtth 0.05 M PBS, pH 7.4, followed by a fixative contmning 4% paraformaldehyde. After perfusion, brains were removed and postfixed in the same fixauve for 2 h at room temperature Brains were cut voth a Vibratome (Model G, Oxford, England) at a thickness of 50/an into frontal secUons. Cort-

ices removed from other brains were flattened by holding between two microscope slides, stored in the same fixative for at least I week at 4°C and cut parallel to the pml surface with a Vibratome at a thickness of 50/an. The sections were incubated at room temperature sequentially in the following solutions: (1) 3% H2Oz/Tris-buffered sahne (TBS), pH 7.4, for 10 rain; (2) 2% bovine serum albumin/4% horse serumfFBS for 30 rain; (3) rabbit serum contaimng PAb 291 (1:200 dilution with TBS containing 0.1% bovine serum albumin) or in culture supematant contmning MAb 1G2 for 2 h; (4) biotmylated anti-rabbit IgG solution for PAb 291 or anti-mouse IgG solution for MAb IG2 for 60 min, (5) avidin-biotin-peroxidase complex solution (the Vectastam ABC kit) for 30 min ; and (6) 0.1% diaminobenzldine/0.02% hydrogen peroxidefrBS. As controls, some sections were stained with MAb 1G2 or with PAb 291 which had been preancubated with an excess of the brain CSPG preparatton No stgnificant staining was apparent m these specimens.

RESULTS

Production and characterization o f antibodies M A b 1D1 (Rauch et al., 1991) and M A b 1G2 (Oohira et al., 1994) recognized the C-terminal half of neurocan. T o obtain an anUbody which recognizes the N-terminal half of this molecule, we raised a polyclonal antibody (PAb 291) against a synthetic peptide corresponding to the amino acid sequence from residues 483 to 498 of neurocan according to the amino acid sequence deduced from the nucleotide sequence of the e D N A reported by Rauch et al. (1992) (see Fig. 1). Fig. 2 shows the specificity of P A b 291 and M A b 1G2. Both in the brain homogenate (lanes 4 and 8) and m the brain PBS-extracts (lanes 3 and 7), P A b 291 recognized a very diffuse band with the same mobility as that recognized by M A b 1G2. M A b 1(32 recognized both 220 and 150 k D a core glycoproteins in the chondroitinasc-digested sample (lane 2), as we previously described (Oohira et al., 1994). P A b 291 recogmzed not only the 220 k D a core glycoprotein but also that with a molecular weight of 130 k D a in both the chondroitmase-dlgested C S P G mixture (lane 5) and in the chondroitinase-digested PBS-cxtract (lane 6). Since P A b 291 was raised against a peptide sequence included in the N-terminal half of neurocan,

428

l--t[MIK~)

M

~ rst,i c: a/

220K

150K 130K

1

2

3

4

5

6

7

"1'

Fig 2 Characterization of PAb 291 and MAb IG2 with samples obtained from 10-da)-old rat brain The PBS-soluble blare CSPG mixture of 10-day-old rats digested Vclth chondromnase ABC (lane I and 5~ th~ brain homogenate (lanes 4 and 8), the brain PBS-extracts (lanes 3 and 7) and the brain PBS-extract &gestcd with chondromnase ABC (laneb 2 and 6) were separated by SDS PAGE After electrophores,s, mateHal~ in SDS gels were electrotransferred onto PVDF membranes In lane 1, protein bands were v~suahzed b~ staining the membrane w~th Coomass~e brdhant blue R-250 lmmunoreacu~e matermls ~ere detected b~ staining with MAb IG2 (lanes 2 to 4) or PAb 291 danes 5 to 8) this result suggests that CSPG-130 corresponds to the N-terminal half of thts molecule

Pur~catton oJ anttgens h)' tmrnunoadsorptton column The antigens recogmzed by either M A b 1G2 or PAb 291 were purified from the m~xture of CSPG-250, 220, 150, 130 and 93 (Fig 3, lane 1) by lmmunoadsorpUon column chromatography as described m Experimental Procedures. The eluate from the M A b IG2-column contained 20-90 nmol of uromc acid and that from the PAb 291-column contained 3-5 nmol of uromc acid (Table 1) Fig 3 shows an electrophoretogram of proteoglycans before and after purification Chondroltm sulfate proteoglycans with 220 and 150 k D a core glycoprotelns were eluted from the M A b 1G2column (lane 2), while those with 220 kDa and 130 kDa core glycoprotems were eluted from the P A b 291 column (lane 4) Partml (6-7 residues) amino acid sequences of the N-terminus of the 130 k D a core glycoprotems before and after purificatmn through the ~mmunoadsorptmn

column were D Q D T Q D , eqmvalent to the N-ternnnal sequence of CSPG-220, 1 e the juvemle typc of neurocan Together with the observatmn that the N-terminal amino acid sequences of the 150 kDa core glycoprotem and the adult type of neurocan were identical (Oohlra et al. 1994), these findings show that the juvemle type of neurocan with the 220 kDa core glycoprotem (CSPG-220) is cleaved into the Nterminal half with 130 kDa core glycoprotem (CSPG130) and the C-terminal half with 150 kDa core glycoprotem (CSPG- 150)

Temporal expression pattern.s ol CSPG-220, 150 and 130 To determine the temporal expressmn patterns of CSPG-220, 150 and 130, PBS-extracts of rat cerebrum at various developmental stages from embryonic day 14 (El4) to 2 years old were digested w~th protease free-chondromnase A B C and processed for immunoblottmg using M A b IG2 [Fig. 4(A)] or PAb 291 [Fig 4(B)] The intensity o f l m m u n o s t a m l n g is shown

Immunological identification of two neurocan fragments

429

250K 220K

150K 130K

1

2

3

4

5

Fig. 3. Purificataon of antigens using immunoadsorptmn columns. The PBS-soluble CSPG mLxturepurified from 10-day-old rat brain was applied to a MAb 1G2-column (lanes 2 and 3) or PAb 291-column (lanes 4 and 5). The eluates were separated by SDS-PAGE before (lanes 3 and 5) and after (lanes 2 and 4) chondroitinase ABC digestion. An aliquot of PBS-soluble CSPG rmxture was also digested with chondroitinase ABC and separated by SDS-PAGE (lane 1). Protein bands were visualized by silver staimng in Fig. 4(C and D). The amounts of CSPG-220, 150 and 130 reached maximum levels around birth. Although CSPG-220 was hardly detectable in the mature brain, considerable amounts of CSPG-150 and 130 were observed even in the 2-year-old rat brain. In addition, PAb 291 reacted with a 110 kDa protein band in mature brain extracts. Thus, CSPG-130 may be further cleaved into a proteoglycan fragment with a 110 k D a core glycoprotein and (a) small peptide(s) in mature brains. Distribution o f neurocan in the cerebrum We previously reported the localization of M A b IG2 antigen in the developing rat cerebrum (Oohira et al., 1994). In the present study, the localization of PAb 291 antigen was compared with that o f M A b 1G2 antigen in the cerebral cortex of 7-day-old SpragueDawley rats. The results show that both antigens had a similar distribution, with the superficial layers of the cortex being lmmunolabeled more strongly than the

inner layers with both antisera [Fig. 5(A and B)], while the barrel hollows in layer IV were free ofimmunoreactivity [Fig. 5(A, B, C and D)]. Immunostaining with M A b IG2 showed stronger contrast than with PAb 291, possibly indicating that CSPG-130 is distributed more diffusely than CSPG-150. DISCUSSION

In the present study, we identified and purified the N-terminal proteolytic cleavage product (CSPG-130) of neurocan (CSPG-220) using PAb 291. CSPG-130 existed in CSPG form in rat brain along with the Cterminal half (CSPG-150) of neuroean, CSPG-130 and CSPG-150 are considered to be proteolytic products of CSPG-220 because a riboprobe corresponding to a region of neurocan c D N A hybridized with a single band at ~ 7.5 kb on Northern blots of m R N A from both 4-day-old and adult rat brain (Rauch et al., 1992). From the amino acid sequence of the N-ter-

i-"lrSllk(i MAq",UI l'/t/[

43O A

B E 14

P 16 18

0

E 5

10 20 30 90

730

14 16

18

P 0

5

10

20 30

90

730

220K 150K 130K 110K

C

D Prenatal

I

o~)

:

1 5 --

Postnatal

.i"_~.-~' ~ . I \

I I

14 16 18

10 20 30

Age (days)

Postnatal

E o I o

~[ I L t_ ;~_i_~/ ] 0

Prenatal

"

90

" 730

o 14 16 18

0

10 20 30

'

90

730

Age (days)

Fig 4 Developmental changes m the relative a m o u n t s of 220, 150 and 130 k D a core glycoprotems of neurocan PBS extracts were prepared from homogenates (30 pg protein) of rat cerebra at various developmental stages from E l 4 to postnatal day 730 (P730) and were digested with chondrmtmase A B C as described in Experimental Procedures The digests were resolved by S D S - P A G E and immunoblotted with M A b 1G2 (A) or PAb 291 (B) The intensities of the lmmunolabeled protein bands shown in A were quantified by densatometry, and the relative a m o u n t s of 220 k D a ( O ) and 150 k D a ( U ) core glycoprotems are shown in C The relative a m o u n t s of 220 k D a ( O ) and 130 k D a ( • ) core glycoproteins were determined by quantifying the intensities of the lmmunolabeled protein bands shown in B by densltometry The data were normalized to the level of the 220 k D a component on PI0

Fig 5 I m m u n o s t a m m g of 7-day-old rat cerebrum with M A b 1G2 (A and C) and PAb 291 (B and D) A and B, frontal sections of cortex, C and D, tangential sections at layer-IV of the flattened cortex Scale bar = 200 # m

Immunological identification of two neurocan fragments minus of CSPG-150, neurocan is supposed to be cleaved between amino acid residues 638 and 639 of the amino acid sequence deduced from the nucleotide sequence of the eDNA reported by Rauch et al. (1992). Both CSPG-130 and CSPG-150 exist in proteoglycan forms in the brain and the cleavage of neurocan is regulated developmentally. Both these pieces of evidence suggest that the cleavage of neurocan may have some biological significance in relation to brain development. Grumet et al. (1993) reported that the adult type of neurocan (CSPG- 150) is less effective than the juvenile type (CSPG-220) not only in inhibiting binding to neurons of Ng-CAM but also in evoking the aggregation of neurons. The cleavage of neurocan may result in the adjusting of the extent of cell-cell or cellsubstratum interactions. Our preliminary investigation using lectins indicated that the core proteins of CSPG-220, 150 and 130 are giyeosylated with various sugar chains and that there are some differences m glycosylation between these core giyeoproteins (unpublished data). Thus, the differences in effectiveness between CSPG220 and CSPG-150 may, in part, be attributable to the differential giycosylation of these core glycoproteins. The structure and functions of the sugar chains on these molecules remain to be determined along with the mechanism of developmental regulation of the enzyme(s) which splits neurocan into CSPG-130 and CSPG-150. PAb 291 and MAb 1G2 are proposed to be useful tools for studying the mechamsm and biological significance of the cleavage of neurocan. Acknowledgements--We thank Dr T. Asano for helpful

suggestions to produce lmmunoadsorption columns and skillful assistance for determining N-termanal amino acid sequences. We also thank to Dr R. Katoh-Semba for helpful suggestions to produce PAb 291. This work was supported

431

in part by a Grant-ln-Atd for SeienUfic Research (No. 03454121) from the Ministry of Education, Scienceand Culture of Japan. REFERENCES

Bitter T and Muir H M. (1962) Modified uromc acad carbazole reaction Analyt Biochem 4, 330--334. Grumet M., Flaccus A. and Margolis R. U (1993) Functional charactenzaUon of chondroitin sulfate proteoglycans of brain' interactions with neurons and neural cell adhesion molecules J. Cell Biol. 120, 815-824 Herndon M. E and Lander A. D. (1990) A diverse set of developmentally regulated proteoglycans is expressed m the rat central nervous system. Neuron 4, 949-961. Maeda N. and Oohlra A. (1991) Proteoglycan and development of the brain : functions of brain chondromn sulfate proteoglycan. Trends Glycosct Glycotechnol. 3, 28-35, Margohs R. K and Margohs R. U. (1993) Nervous tissue proteoglycans. Experimentla 49, 429-446. Oohira A Matsm F , Matsuda M., Takida Y and Kubokl Y. (1988) Occurrence of three distinct molecular species of chondromn sulfate proteoglycan in the developing rat brain. J. biol. Chem. 263, 10,240-10,246. Oohira A, Matstu F., Watanabe E, Kushima Y and Maeda N. (1994) Developmentallyregulated expression of a brain specific species of chondrmtm sulfate proteoglycan, neurocan, identified with a monoclonal antibody 1G2 in the rat cerebrum. Neuroscience 60, 145-157 Posnett D. N and Tam J. P. (1989) Multiple antigenic pepUde method for produong antapeptide site-specific antibodies. Meth Enzym 178, 739-746. Ranch U , Gao P., Janetzko A., Flaccus A., Hilgenberg L, Tekotte H., Margohs R K. and Margohs R. U (1991) Isolation and characterization of developmentally regulated chondroitm sulfate and chondroitm/keratan sulfate proteoglycans of brain identified with monoclonal antibodies J. biol. Chem 266, 14,785-14,801 Ranch U., Karthikeyan L., Maurel P., Margohs R. K and Margolis R U (1992) Cloning and primary structure of neurocan, a developmentallyregulated, aggregating chondroitin sulfate proteoglycan of brain. J. biol. Chem. 267, 19,536-19,547. Zoller M. and Matzku S. (1976) Antigen and antabody punfication by lmmunoadsorpUon: elinunation of non-blospecificallybound proteins. J Immun Meth. 11, 287-295.