ANALYTICAL
BIOCHEMISTRY
186,%%94
(1990)
Separation and Identification of 2’,3’-Cyclic Nucleotide 3’-Phosphodiesterase on Isoelectric Focusing Gels H&kan
Persson’
and Olle Corneliuson
Section of Neuroanatomy, Department of Anatomy, Gothenburg University, Box 33031,S-400 33 G6teborg Sweden
Received
October
5,1989
A method is presented for the separation and detection of the myelin marker enzyme 2’,3’-cyclic nucleotide 3’-phosphodiesterase on isoelectric focusing gels and by immunoblotting. The gel staining procedure is a modification of a method used to demonstrate enzyme activity on blots after sodium dodecyl sulfate-polyacrylamide gel electrophoresis and two-dimensional polyacrylamide gel electrophoresis. The results show that immunologically active 2’,3’-cyclic nucleotide 3’phosphodiesterase can be separated under equilibrium conditions on isoelectric focusing gels with an expanded alkaline pH range after solubilization in a mixture of nonionic/zwitterionic detergents and urea. Enzymatically active 2’,3’-cyclic nucleotide 3’phosphodiesterase focused as two closely spaced bands at pIaPP 8.1 and 8.8, respectively, while 2’,3’-cyclic nucleotide 3’-phosphodiesterase immunoreactivity was detected as four distinct bands at PI,,, 4.2,7.4,8.8, and 9.3 and a diffuse band at pIaPP 7.9-8.2. By two-dimensional separation these five bands showed molecular weights of about 43-47 kDa, i.e., corresponding to reported values for immunologically active 2’,3’-cyclic nucleotide 3’-phosphodiesterase. Since enzyme activity is associated with only two of the bands, nonspecific and artifactual banding due to, e.g., detergent micelle formation, is unlikely. o 1990 Academic press, IN.
Extraction of purified myelin with chloroform-methanol 2:l (v/v) yields an insoluble protein residue. When this fraction is subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)2 under rei To whom correspondence should be addressed. ’ Abbreviations used: CHAPS, 3-(3-cholamidopropyl)-dimethylammonio)-l-propanesulfonate; CNP, 2’,3’-cyclic nucleotide 3’-phosphodiesterase (immunologically active); CNPase, 2’,3’-cyclic nucleotide 3’.phosphodiesterase (enzymatically active); HRP, horseradish peroxidase; IEF, isoelectric focusing; MES, 4-morpholineethanesulfonic acid; NEPHGE, nonequilibrium pH gel electrophoresis; PAGE, polya&amide gel electrophoresis; PBS, phosphate-buffered saline; pl.,,, apparent isoelectric point; SDS, sodium dodecyl sulfate; TEMED, N,N,N’JV’-tetramethylethylenediamine. 90
ducing conditions, two proteins with apparent molecular weights between 43 and 50 kDa, depending on species, are prominent. These have been identified as subunits of an enzyme, 2’,3’-cyclic nucleotide 3’-phosphodiesterase (CNP; EC 3.1.4.37) which constitutes 4% of total myelin proteins (1,2). CNP is located almost exclusively in mammalian central nervous system, although low levels have been detected in other mammalian tissues and in bacteria (3). Sedimentation rate and activity analyses under reducing and nonreducing conditions have established a monomer-dimer relationship in CNP with catalytically active subunits (4). The previously employed methods for electrophoretic separation of CNP, i.e., SDS-PAGE and two-dimensional PAGE with nonequilibrium pH gel electrophoresis (NEPHGE) in the first dimension (5,6), have inferior separation characteristics as compared to isoelectric focusing (IEF) in analysis of other enzymes [see, e.g., (7,8)]. Therefore, an IEF procedure is presented for the analysis of CNP and CNPase activity with high resolution and under less denaturing conditions than previously used methods. The CNPase activity staining used is a modification of a method for Western blots (5). In the present work the immunologically identifiable CNP is referred to as CNP and the term CNPase is used for enzymatically active CNP. MATERIALS
AND
METHODS
Animals. Adult rabbits (New Zealand white) were used. The animals showed no signs of disease and had free accessto food and water. Chemicals. Acrylamide, N,N-methylenebisacrylamide, N,N,N’,N’-tetramethylethylenediamine (TEMED), 3 - ((3 - cholamidopropyl)dimethylammonio) - 1 - pro panesulfonate (CHAPS), EDTA, 4-morpholineethanesulfonic acid (MES), and 2’,3’-cyclic NADP were from Sigma (St Louis, MO). Ampholine 3.5-9.5, Repel-Silane, and Silane A-174 were from LKB Pharmacia (Bromma, Sweden). Glucose 6-phosphate and glucose6-phosphate dehydrogenase (grade II from yeast) were 0003-2697/90
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
ISOELECTRIC
FOCUSING
OF 2’,3’-CYCLIC
NUCLEOTIDE
3’-PHOSPHODIESTERASE
91
obtained from Boehringer-Mannheim (Lewes, East Sus- pensions were loaded in the stack gel sample wells and sex, U.K.). Gelatin (microbiological quality, article No. electrophoresed under standard conditions (13) in a 1.51.4070) and all other chemicals were from Kebo Lab AB mm-thick 12% acrylamide gel. (Spanga, Sweden). Nitrocellulose (0.22 pm) was from Myelin to be separated directly by SDS-PAGE was delipidated in ether-ethanol 3:2 (v/v) and solubilized in Schleicher & Schuell (Dassel, F.R.G.). Antisera. Goat anti-CNP antiserum was provided by SDS sample buffer at a concentration of 1 mg/ml. One hundred micrograms of protein was loaded in each samDr. T. J. Sprinkle. Horseradish peroxidase (HRP)-conple well and electrophoresed as above. jugated rabbit anti-goat antibodies were from Dakopatts A/S, Glostrup, Denmark. Gel staining. Reference IEF gels were fixed for 30 Subcellular fractionation. Myelin was isolated from min in 20% trichloroacetic acid, cleared for 5 min in derabbit spinal cord as described by Norton and Poduslo staining solution (ethanol-acetic acid-water 25:7.5:67.5, (9). Protein content was determined using bovine serum v/v/v), stained for 10 min at +6O”C in 0.1% Coomassie brilliant blue R-250 in destaining solution, and dealbumin as standard (10). The myelin was lyophilized stained overnight. Coomassie blue staining was used and stored desiccated at -70°C. since the chemicals used as glass repellents and adherIsoelectric focusing. Myelin samples to be electrofoents interfered with silver staining (see below). cused were prepared by partial delipidation of 1 mg lyStaining for CNPase activity was done by a modificaophilized myelin in ether-ethanol 3:2 (v/v) (11) and subtion of a previously published method (5). After the IEF sequent incubation for 15 min at room temperature in a glass homogenizer with 200 ~1 of a medium consisting of run, the gel was fixed in methanol-water 40:60 (v/v) for 1 h. The gel was then placed in water for 10 min and 7% of a mixture of Triton X-100 and CHAPS in a ratio soaked for 1 h in 50 mM MES buffer, pH 6.1, with 30 of 4:3 (w/w), 9 M urea, 1% dithiothreitol, and 2% amphomM MgClz, 0.1% Triton X-100, 1 mM EDTA, and 1 mM lines 3.5-9.5 with intermittent homogenization. dithiothreitol. The gels were then stained in a reaction IEF gels were prepared by dissolving 10.6 g urea and mixture consisting of 100 mM MES buffer, pH 6.1, con0.6 g CHAPS in a mixture of 2.33 ml 30% stock bisacryltaining 30 mM MgCl,, 0.1% Triton X-100,0.05 mM 2’,3’amide (28.38 g acrylamide and 1.62 g bis dissolved in 100 cyclic NADP, glucose-6-phosphate (2 mg/ml), nitroblue ml H,O), 8 ml 10% Triton X-100 stock solution, 1.25 ml tetrazolium (0.4 mg/ml), phenazine methosulfate (0.02 Ampholine 3.5-9.5, and 35 ~1 TEMED. The extra addimg/ml), and glucose-6-phosphate dehydrogenase (0.05 tion of TEMED together with appropriate electrode sounit/ml). A brownish-red color usually developed within lutions expanded the alkaline end of the gradient to over pH 10 (12). The solution was adjusted to 20 ml with HzO, 30 min. The reaction was terminated by immersing the deaerated, and mixed with 0.25 ml 1% ammonium per- gels in 10% acetic acid. SDS-PAGE gels were silver stained according to sulfate. Slab gels (180 ym thick, 6 X 10 cm) were molded Wray et al. (14). between glass plates treated with Silane A-174 and Repel-Silane. Electroimmunoblotting. SDS-PAGE gels were elecElectrofocusing was performed on an LKB Multiphor troblotted as described by Towbin et al. (15) except for at +lO”C. After 30 min prefocusing at 10 mA, constant the use of 0.22~pm instead of 0.45-pm nitrocellulose. IEF current, 200 pg protein was loaded onto each lane using gels were blotted as described by Johnson et al. (16) with 3 X lo-mm pieces of Whatman GF/D glass filter pieces minor modifications. Briefly, after the IEF run the gel as sample applicators. Control lanes with gel pieces with was equilibrated for 30 min in 50 ml of a solution of 3% blank solubilization mixture without myelin protein SDS in 25 mM Tris buffer, pH 6.8. The gel was then sample were run in parallel. The applicators were placed transferred to blotting buffer (0.192 M glycine and 25 mM 3 cm from the anode and removed after 1 h. After prefoTris, pH 8.3, with 20% (v/v) methanol) for 10 min and cusing, the samples were focused for 3 h at 120 V/cm electroblotted for 18 h at 150 mA in an LKB Transblot constant voltage (=3800 V h), with 0.04 M aspartate and unit with a 3-mm layer of 1% agarose gel equilibrated in 2% ampholines as analyte and 0.5 M NaOH, 0.05 M argi- blotting buffer inserted between the polyacrylamide gel nine, and 2% ampholines as catholyte. After the run, the and the nitrocellulose. pH gradient was measured with a surface electrode at After transfer, the nitrocellulose blots were cut into 4+25”C. mm strips and blocked overnight in phosphate-buffered SDS-PAGE. After focusing, the lo-mm-wide sam- saline (PBS) with 1.5% gelatin, 0.5% Tween-20, and 1% ple lane on the gel was cut into 8-mm-long pieces along normal rabbit serum at 4°C. After a 30-min rinse in PBS the pH gradient, yielding ten pieces from each lane. The the strips were incubated with primary antibody diluted gel pieces were placed in methanol-water 50:50 (v/v) for 1:2000 in PBS-0.5% Tween-20 for 30 min, rinsed 4 X 15 1 h and subsequently homogenized in SDS sample buffer min in PBS, and incubated with HRP-conjugated sec(1% SDS, 0.01% EDTA, 1.5% dithiothreitol, 9 M urea, ond antibody diluted 1:500 in PBS-0.5% Tween-20. 8% sucrose, and 0.001% bromphenol blue in 0.01 M TrisControl strips with PBS instead of diluted primary antiHCl, pH 8.0). After equilibration for 30 min the gel sus- body were run in parallel. After rinsing 4 X 15 min in
92
PERSSON
AND
CORNELIUSON
PBS, the strips were covered with substrate solution (1 part of a stock solution of 3 mg/ml4-chloro-l-naphthol in methanol and 3 parts PBS with 0.01% Hz02 added just before use). The reaction was terminated by rinsing with distilled water. IEF control strips for CNPase activity were made by running unblotted nitrocellulose strips through the standard staining procedure. IEF control CNP strips were made by incubating myelin blots as described but without primary antibody.
94
67
RESULTS
Delipidation of the myelin samples proved essential since experiments with focusing of undelipidated myelin sometimes totally abolished CNPase activity staining and instead resulted in a nonspecific staining of the lipids. However, this nonspecific lipid “staining” took considerably longer time to develop. It did not appear until after 3-4 h. The protein content of the myelin fraction was 2531%. Figure 1 shows the normal SDS-PAGE pattern of myelin proteins (lane A) and the corresponding nitrocellulose blot incubated with anti-CNP antibody (lane B). The antiserum stains a protein doublet with estimated relative molecular weights of 45-47 kDa corresponding to the CNP doublet on the silver-stained reference lane. In Fig. 2, lane A is a reference myelin lane stained with Coomassie brilliant blue. Lane B is a gel run in parallel and stained for CNPase activity, showing that enzyme activity is associated with two bands focusing at PI,,, 8.1 and PI,,, 8.8, respectively. The more alkaline of the two bands cofocuses with a protein focusing as a broad band with the same pl as the myelin proteolipid protein [see Fig. 3 and (16)]. Lane C shows the binding of CNP antibody. Four distinct bands at PI,,, 4.2, 7.4, 8.8, and 9.3 are seen. Furthermore, a diffuse band at PI,,, 7.9-8.2 is recognized by the antibody. CNPase activity thus coincides with the diffuse PI,,, 7.9-8.2 band and the pIapp8.8 band. Lane D is a control gel strip for CNPase, which is blank. Lane E is a control strip for CNP immunoreactivity where primary anti-CNP antibody was omitted. Figure 3 shows the results of submitting the focused proteins in the excised IEF gel pieces to SDS-PAGE. This method does not result in the familiar “spot” appearance of the O’Farrell type two-dimensional gels, but instead a continuum of bands at different heights result; for example, two spots with very similar molecular weight but situated say 8 mm apart can, with the described method, be “caught” within the same gel piece and thus become pooled in the same SDS-PAGE sample well after being homogenized in SDS sample buffer. However, in our hands the resolution and reproducibility of the method are superior as compared with preliminary experiments with “true” two-dimensional separation. Lanes a-j show the protein patterns after silver staining, lane a being the more anodal. Lane f represents the region of the IEF gel where the more acidic of the
43
30
20
FIG. 1.
SDS-PAGE of myelin proteins. Lane A shows the separation pattern of myelin after silver staining. Lane B shows the pattern obtained after incubation of a nitrocellulose blot of the proteins in lane A with anti-CNP antibodies and detection with HRP-conjugated second antibodies.
CNPase activity-stained bands focused, and lane h, the region where the more alkaline CNPase activity band focused. As can be seen in Fig. 2, the CNPase activity coincides with the appearance, just above 43 kDa, of the CNP doublet on the SDS-PAGE gel in Fig. 3, with the more acidic subunit being the smallest. However, both subunits are represented in the IEF gel in a far broader pH region than indicated by the enzyme activity stain. This is demonstrated by the immunoblot pattern, showing one band at ~1~~~4.2 (which can also be seen in Fig. 3, lane b, and four bands between PI,,, 7.4 and 9.3 which are also represented in Fig. 3, lanes f-i.
ISOELECTRIC
4
5
fi
FOCUSING
‘I
OF
8
2’,3’-CYCLIC
$3
10
PH
t FIG. 2. IEF of myelin proteins. Lane A shows the pattern obtained after Coomassie blue staining. Lane B shows a sample run in parallel and stained for CNPase as described in the text. Lane C shows the pattern obtained after incubation of a nitrocellulose blot of the proteins in lane A with anti-CNP antibody and detection with HRP-conjugated second antibody. Lane D shows a control gel strip without myelin sample, but incubated as in lane B. Lane E shows a control gel incubated as in lane C except that the primary anti-CNP antibody was omitted. CorrespondingpH gradient shown on top was measured with a Beckman surface electrode.
DISCUSSION
The biological significance and function of CNP are unknown. So are its biosynthesis and degradation pathways [see (18) for a current review on these topics]. Previous electrophoretic studies of CNP in gels have been limited to SDS-PAGE or two-dimensional PAGE with NEPHGE in the first dimension (5,6), presumably due to difficulties in solubilization of myelin proteins and in obtaining a sufficiently alkaline, stable pH gradient for IEF in slab gels. However, IEF in slabs gives considerably higher resolution than SDS-PAGE and two-dimensional PAGE (7,8), making it the method more often used for studying the isoforms of various enzymes (19,20), genetic polymorphism, and enzyme microheterogeneity (21,22) and for evaluating enzyme substrate specificity (23). The inhibition of alkaline phosphatase activity exerted by the chelating effect of ampholines is well documented (24), but metal cofactors have never been demonstrated to be of importance for CNPase activity. The reason for staining the IEF gel itself instead of a blot was that since the gel is only 0.18 mm thick, diffusion time for reaction products and background is reduced so much that blotting with accompanying losses and risk of SDS denaturation would not mean any advantages. Post-translational modifications has been demonstrated in all myelin proteins investigated in this regard (25). Since this processing consists of phosphorylation, methylation, glycosylation, etc., with no studies describing cleavage of peptide bonds, this implies that the net charge rather than the molecular weight would be changed during these processes. Thus, for further studies of the metabolism of CNP as well as studies of its genetic variations, possible polymorphism and microheterogeneity in normal animals as well as in the array of frequently used dysmyelinating murine (26), canine, and rabbit (6) mutants, IEF seems
NUCLEOTIDE
93
3’-PHOSPHODIESTERASE
the method of choice. As can be seen in Fig. 3, protein bands with molecular weights corresponding to the CNP doublet (45 and 47 kDa) are present in a far broader region than the CNPase activity staining in Fig. 2B. Since CNPase activity staining of two-dimensional gels (5) gives a pattern corresponding to Fig. 2B rather than to Fig. 3, an explanation would be the presence of inactive CNP subunits with slightly different surface charges not resolvable by SDS-PAGE. Since immunoreactivity and enzyme activity are not equally distributed on the IEF gels/blots the possibility that the focusing pattern is an artifact, as a result of unspecific interactions of CNP and ampholines or due to detergent micelle formation, is highly unlikely. As a rule, enzymes resolved on IEF gels give rise to multiple bands, and this fact has been exploited in a number of ways by many different workers [e.g., (27-29)]. However, substrate specificity and enzyme activity have remained problems, which have led to efforts to circumvent them by using immunocytochemistry. This has indeed proven to be a powerful tool which, in addition, permits studies of inactive precursors and their intracellular pathways and compartmentalization (30-32). Also, the greater sensitivity of the immunological methods as compared with enzyme histochemistry will most probably sharpen the clinical value of enzyme fractionation. Thus it is not surprising that the pattern produced by the immunoblots in the present work is more complex than the enzyme activity staining pattern. The CNP separation pattern and reported PI,,, of active enzyme in the present work, 8.1 and 8.8, adds to and partly explains the previous confusion [i.e., 8.6 (33), ap-
94 67
43
30
20
FIG. 3. Silver-stained SDS-PAGE of excised gel pieces from a lane of myelin proteins as shown in Fig. 2, lane A. Each piece was approx 8 mm and lanes a-j cover the IEF gel pH gradient from left (= most acidic gel piece) to right (= most alkaline gel piece). The figures to the right indicate molecular weights as determined with protein standards.
94
PERSSON
AND
prox 9 (34,35), approx 9.7 (36), and 10.33 (37)]. Maybe the more alkaline CNP subspecies, inactive when stained for activity as described here and in (5), retain their activity when treated and detected otherwise. Particularly, this could be an additional factor to take into account besides contamination by myelin membranes when considering the problem of the high variability obtained by different laboratories in measuring CNPase activity in oligodendrocyte preparations [reviewed in (38)]. Calcium-dependent protein kinases, apparently endogenous to the myelin membrane, and cyclic AMP-dependent protein kinases have been shown to actively phosphorylate the larger subunit of CNP both in vitro (39) and in uiuo (5,40,41). Thus, one explanation for the discrepancy in the number of bands seen on SDS-PAGE and IEF gels could be charge heterogeneity due to different degrees of phosphorylation. A number of nervous system- and myelin-affecting pathological conditions are accompanied by changes in CNPase activity [reviewed in (IS)]. Studies are underway in our laboratory to investigate the separation pattern during experimental allergic encephalitis (a myelinaffecting pathological condition used as an “animal model disease” for multiple sclerosis) to correlate changes in separation pattern with the clinical course of the disease. The results are to be published in a forthcoming study. ACKNOWLEDGMENTS This work was supported by grants from the Swedish Medical Research Council (Project 03157) and Anna Ahrenbergs stiftelse. The authors thank Rita Grander and Marieanne Eriksson for excellent technical assistance. We are grateful to Dr. Terry Joe Sprinkle for generously providing the antiserum against CNP and Professor C-H. Berthold for always being an enthusiast and a support.
CORNELIUSON 10. Peterson, G. L. (1983) in Methods in Enzymology (Hirs, C. H. W., and Timasheff, S. N., Eds.), Vol. 91, pp. 95-119, Academic Press, New York. 11. Greenfield, S., Norton, W. T., and Morell, P. (1971) J. Neurochem. l&2119-2128. 12. Yao-Jun, G., andBishop, R. (1982) J. Chromatogr. 234,459-462. 13. Laemmli, U. K. (1970) Nature (London) 227,680-685. 14. Wray, W., Boulikas, T., Wray, V. P., and Hancock, R. (1981) Anal. Biochem. 118,197-203. 15. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. N&l. Acad. Sci. USA 76,4350-4354. 16. Johnson, T. K., Leonard Yuen, K. C., Denell, R. E., and Consigli, R. A. (1983) Anal. Biochem. 133,X26-131. 17. Draper, M., Lees, M. B., and Chan, D. S. (1978) J. Neurochem.
31,1095-1099. 18. Vogel, 1677.
U. S., andThompson,
19. Harris, H., and Hopkinson, Electrophoresis in Human
R. J. (1988)
J. Neurochem.
D. A. (1976) Handbook Genetics, North-Holland,
50,1667of Enzyme Amsterdam.
20. Needleman, S. B., and Koenig, H. (1974) Biochim. Biophys. Acta 379,43-56. 21. Kiihnl, P. (1981) in Electrophoresis ‘81 (Allen, R. C., and Arnaud, P., Eds.),
pp. 563-571, Koenig,
22. Patel, A., and 23. Lojda, Z., and 24. Latner, A. L., 40,1-7. J. 25. Benjamins, 26.
de Gruyter, Berlin. H. (1976) Neurochem.
Res.
1,275-298.
Kulich, J. (1981) Histochemistv 73,311-319. Parsons, M., and Skillen A. W. (1971) Enzymologia
A., and Smith, M. E. (1984) in Myelin (Morell, P., Ed.), pp. 225-258, Plenum, New York/London. Hogan, E. L., and Greenfield, S. (1984) in Myelin (Morell, P., Ed.), pp. 489-534, Plenum, New York/London.
27. Corless, J. K., and Middleton, H. M., III (1983) Arch. Intern. Med. 143,2291-2294. W. A., and Thorpe, J. W. (1980) Med. Sci. 28. Randall, T., Harland, zmI20,43-47. 29. Young, C. W., and Bittar, E. S. (1973) Cancer Res. 33,2692-2700. 30. Parenti, G., Willemsen, R., Hoogeveen, A. T., Verleun-Mooyman, 31.
M., Van Dongen, J. M., and Galjaard, H. (1987) 43,121-127. Braands, R., Slot, J. W., and Geuze, H. J. (1982)
Eur.
J. Cell Biol.
Eur.
J. Cell Biol.
27.213-220. 32. Van Dongen,
REFERENCES 1. Drummond, 1165.
2. Sprinkle, (1980)
R. J., and Dean,
G. (1980)
T. J., Wells, M. R., Garver, J. Neurochem. 35,1200-1208.
J. Neurochem. F. A., and
35,1155Smith,
D. B.
3. Miiller, H. W., Clapshaw, P. A., and Seifert, W. (1981) FEBS Lett. 131,37-40. 4. Miiller, H. W. (1982) FEBS L&t. 144,77-80. 5. Bradbury, J. M., and Thompson, R. J. (1984) Biochem. J. 221, 361-368. 6. Domanska-Janik, K., Gajkowska, B., de Nechaud, B., and Bourre, J. M. (1988)
J. Neurochem.
50,122-130.
7. Sinha, P., and Gossrau, R. (1984) Histochemistry 81,161-165. 8. Ogata, M., and Satoh, Y. (1988) Ekctrophoresis 9,128-131. 9. Norton, W. T., and Poduslo, S. E. (1973) J. Neurochem. 21,749757.
33. 34.
J. M., Barneveld, R. A., Geuze, H. J., and Galjaard, H. (1984) Histochem. J. 16,941-954. Suda, H., and Tsukada, Y. (1980) J. Neurochem. 34,941-949. Sprinkle, T. J., and Eller, A. G. (1979) Trans. Znt. Sot. Neurochem.
7,123. 35. Sprinkle, them.
T. J., Grimes, 34,880-887.
M. J., and Eller,
A. G. (1980)
J. Neuro-
36. Drummond, 37. 38. 39. 40.
R. J., Hamill, E. B., and Guha, A. (1978) J. Neurochem.3 1,871-878. Kurihara, T., Nishizawa, Y., Takahashi, Y., and Odani, S. (1981) Biochem.J. 195.153-157. Pleasure, D., Kim, S. U., and Silberberg, D. H. (1984) in Oligodendroglia (Norton, W. T., Ed.), pp. 175-190, Plenum, New York. Thompson, R. J., Bradbury, J. M., and Richardson, P. J. (1985) Biochem. Sot. Trans. 14,351-352. Bradbury, J. M., andThompson, R. J. (1984) Biochem. Sot. Trans.
12,1044-1045. J. M., and Thompson, 42. Bradbury, 351-359.
R. J. (1984)
Biochem.
J.
221,
BIOCHEMISTRY
186,%%94
(1990)
Separation and Identification of 2’,3’-Cyclic Nucleotide 3’-Phosphodiesterase on Isoelectric Focusing Gels H&kan
Persson’
and Olle Corneliuson
Section of Neuroanatomy, Department of Anatomy, Gothenburg University, Box 33031,S-400 33 G6teborg Sweden
Received
October
5,1989
A method is presented for the separation and detection of the myelin marker enzyme 2’,3’-cyclic nucleotide 3’-phosphodiesterase on isoelectric focusing gels and by immunoblotting. The gel staining procedure is a modification of a method used to demonstrate enzyme activity on blots after sodium dodecyl sulfate-polyacrylamide gel electrophoresis and two-dimensional polyacrylamide gel electrophoresis. The results show that immunologically active 2’,3’-cyclic nucleotide 3’phosphodiesterase can be separated under equilibrium conditions on isoelectric focusing gels with an expanded alkaline pH range after solubilization in a mixture of nonionic/zwitterionic detergents and urea. Enzymatically active 2’,3’-cyclic nucleotide 3’phosphodiesterase focused as two closely spaced bands at pIaPP 8.1 and 8.8, respectively, while 2’,3’-cyclic nucleotide 3’-phosphodiesterase immunoreactivity was detected as four distinct bands at PI,,, 4.2,7.4,8.8, and 9.3 and a diffuse band at pIaPP 7.9-8.2. By two-dimensional separation these five bands showed molecular weights of about 43-47 kDa, i.e., corresponding to reported values for immunologically active 2’,3’-cyclic nucleotide 3’-phosphodiesterase. Since enzyme activity is associated with only two of the bands, nonspecific and artifactual banding due to, e.g., detergent micelle formation, is unlikely. o 1990 Academic press, IN.
Extraction of purified myelin with chloroform-methanol 2:l (v/v) yields an insoluble protein residue. When this fraction is subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)2 under rei To whom correspondence should be addressed. ’ Abbreviations used: CHAPS, 3-(3-cholamidopropyl)-dimethylammonio)-l-propanesulfonate; CNP, 2’,3’-cyclic nucleotide 3’-phosphodiesterase (immunologically active); CNPase, 2’,3’-cyclic nucleotide 3’.phosphodiesterase (enzymatically active); HRP, horseradish peroxidase; IEF, isoelectric focusing; MES, 4-morpholineethanesulfonic acid; NEPHGE, nonequilibrium pH gel electrophoresis; PAGE, polya&amide gel electrophoresis; PBS, phosphate-buffered saline; pl.,,, apparent isoelectric point; SDS, sodium dodecyl sulfate; TEMED, N,N,N’JV’-tetramethylethylenediamine. 90
ducing conditions, two proteins with apparent molecular weights between 43 and 50 kDa, depending on species, are prominent. These have been identified as subunits of an enzyme, 2’,3’-cyclic nucleotide 3’-phosphodiesterase (CNP; EC 3.1.4.37) which constitutes 4% of total myelin proteins (1,2). CNP is located almost exclusively in mammalian central nervous system, although low levels have been detected in other mammalian tissues and in bacteria (3). Sedimentation rate and activity analyses under reducing and nonreducing conditions have established a monomer-dimer relationship in CNP with catalytically active subunits (4). The previously employed methods for electrophoretic separation of CNP, i.e., SDS-PAGE and two-dimensional PAGE with nonequilibrium pH gel electrophoresis (NEPHGE) in the first dimension (5,6), have inferior separation characteristics as compared to isoelectric focusing (IEF) in analysis of other enzymes [see, e.g., (7,8)]. Therefore, an IEF procedure is presented for the analysis of CNP and CNPase activity with high resolution and under less denaturing conditions than previously used methods. The CNPase activity staining used is a modification of a method for Western blots (5). In the present work the immunologically identifiable CNP is referred to as CNP and the term CNPase is used for enzymatically active CNP. MATERIALS
AND
METHODS
Animals. Adult rabbits (New Zealand white) were used. The animals showed no signs of disease and had free accessto food and water. Chemicals. Acrylamide, N,N-methylenebisacrylamide, N,N,N’,N’-tetramethylethylenediamine (TEMED), 3 - ((3 - cholamidopropyl)dimethylammonio) - 1 - pro panesulfonate (CHAPS), EDTA, 4-morpholineethanesulfonic acid (MES), and 2’,3’-cyclic NADP were from Sigma (St Louis, MO). Ampholine 3.5-9.5, Repel-Silane, and Silane A-174 were from LKB Pharmacia (Bromma, Sweden). Glucose 6-phosphate and glucose6-phosphate dehydrogenase (grade II from yeast) were 0003-2697/90
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
ISOELECTRIC
FOCUSING
OF 2’,3’-CYCLIC
NUCLEOTIDE
3’-PHOSPHODIESTERASE
91
obtained from Boehringer-Mannheim (Lewes, East Sus- pensions were loaded in the stack gel sample wells and sex, U.K.). Gelatin (microbiological quality, article No. electrophoresed under standard conditions (13) in a 1.51.4070) and all other chemicals were from Kebo Lab AB mm-thick 12% acrylamide gel. (Spanga, Sweden). Nitrocellulose (0.22 pm) was from Myelin to be separated directly by SDS-PAGE was delipidated in ether-ethanol 3:2 (v/v) and solubilized in Schleicher & Schuell (Dassel, F.R.G.). Antisera. Goat anti-CNP antiserum was provided by SDS sample buffer at a concentration of 1 mg/ml. One hundred micrograms of protein was loaded in each samDr. T. J. Sprinkle. Horseradish peroxidase (HRP)-conple well and electrophoresed as above. jugated rabbit anti-goat antibodies were from Dakopatts A/S, Glostrup, Denmark. Gel staining. Reference IEF gels were fixed for 30 Subcellular fractionation. Myelin was isolated from min in 20% trichloroacetic acid, cleared for 5 min in derabbit spinal cord as described by Norton and Poduslo staining solution (ethanol-acetic acid-water 25:7.5:67.5, (9). Protein content was determined using bovine serum v/v/v), stained for 10 min at +6O”C in 0.1% Coomassie brilliant blue R-250 in destaining solution, and dealbumin as standard (10). The myelin was lyophilized stained overnight. Coomassie blue staining was used and stored desiccated at -70°C. since the chemicals used as glass repellents and adherIsoelectric focusing. Myelin samples to be electrofoents interfered with silver staining (see below). cused were prepared by partial delipidation of 1 mg lyStaining for CNPase activity was done by a modificaophilized myelin in ether-ethanol 3:2 (v/v) (11) and subtion of a previously published method (5). After the IEF sequent incubation for 15 min at room temperature in a glass homogenizer with 200 ~1 of a medium consisting of run, the gel was fixed in methanol-water 40:60 (v/v) for 1 h. The gel was then placed in water for 10 min and 7% of a mixture of Triton X-100 and CHAPS in a ratio soaked for 1 h in 50 mM MES buffer, pH 6.1, with 30 of 4:3 (w/w), 9 M urea, 1% dithiothreitol, and 2% amphomM MgClz, 0.1% Triton X-100, 1 mM EDTA, and 1 mM lines 3.5-9.5 with intermittent homogenization. dithiothreitol. The gels were then stained in a reaction IEF gels were prepared by dissolving 10.6 g urea and mixture consisting of 100 mM MES buffer, pH 6.1, con0.6 g CHAPS in a mixture of 2.33 ml 30% stock bisacryltaining 30 mM MgCl,, 0.1% Triton X-100,0.05 mM 2’,3’amide (28.38 g acrylamide and 1.62 g bis dissolved in 100 cyclic NADP, glucose-6-phosphate (2 mg/ml), nitroblue ml H,O), 8 ml 10% Triton X-100 stock solution, 1.25 ml tetrazolium (0.4 mg/ml), phenazine methosulfate (0.02 Ampholine 3.5-9.5, and 35 ~1 TEMED. The extra addimg/ml), and glucose-6-phosphate dehydrogenase (0.05 tion of TEMED together with appropriate electrode sounit/ml). A brownish-red color usually developed within lutions expanded the alkaline end of the gradient to over pH 10 (12). The solution was adjusted to 20 ml with HzO, 30 min. The reaction was terminated by immersing the deaerated, and mixed with 0.25 ml 1% ammonium per- gels in 10% acetic acid. SDS-PAGE gels were silver stained according to sulfate. Slab gels (180 ym thick, 6 X 10 cm) were molded Wray et al. (14). between glass plates treated with Silane A-174 and Repel-Silane. Electroimmunoblotting. SDS-PAGE gels were elecElectrofocusing was performed on an LKB Multiphor troblotted as described by Towbin et al. (15) except for at +lO”C. After 30 min prefocusing at 10 mA, constant the use of 0.22~pm instead of 0.45-pm nitrocellulose. IEF current, 200 pg protein was loaded onto each lane using gels were blotted as described by Johnson et al. (16) with 3 X lo-mm pieces of Whatman GF/D glass filter pieces minor modifications. Briefly, after the IEF run the gel as sample applicators. Control lanes with gel pieces with was equilibrated for 30 min in 50 ml of a solution of 3% blank solubilization mixture without myelin protein SDS in 25 mM Tris buffer, pH 6.8. The gel was then sample were run in parallel. The applicators were placed transferred to blotting buffer (0.192 M glycine and 25 mM 3 cm from the anode and removed after 1 h. After prefoTris, pH 8.3, with 20% (v/v) methanol) for 10 min and cusing, the samples were focused for 3 h at 120 V/cm electroblotted for 18 h at 150 mA in an LKB Transblot constant voltage (=3800 V h), with 0.04 M aspartate and unit with a 3-mm layer of 1% agarose gel equilibrated in 2% ampholines as analyte and 0.5 M NaOH, 0.05 M argi- blotting buffer inserted between the polyacrylamide gel nine, and 2% ampholines as catholyte. After the run, the and the nitrocellulose. pH gradient was measured with a surface electrode at After transfer, the nitrocellulose blots were cut into 4+25”C. mm strips and blocked overnight in phosphate-buffered SDS-PAGE. After focusing, the lo-mm-wide sam- saline (PBS) with 1.5% gelatin, 0.5% Tween-20, and 1% ple lane on the gel was cut into 8-mm-long pieces along normal rabbit serum at 4°C. After a 30-min rinse in PBS the pH gradient, yielding ten pieces from each lane. The the strips were incubated with primary antibody diluted gel pieces were placed in methanol-water 50:50 (v/v) for 1:2000 in PBS-0.5% Tween-20 for 30 min, rinsed 4 X 15 1 h and subsequently homogenized in SDS sample buffer min in PBS, and incubated with HRP-conjugated sec(1% SDS, 0.01% EDTA, 1.5% dithiothreitol, 9 M urea, ond antibody diluted 1:500 in PBS-0.5% Tween-20. 8% sucrose, and 0.001% bromphenol blue in 0.01 M TrisControl strips with PBS instead of diluted primary antiHCl, pH 8.0). After equilibration for 30 min the gel sus- body were run in parallel. After rinsing 4 X 15 min in
92
PERSSON
AND
CORNELIUSON
PBS, the strips were covered with substrate solution (1 part of a stock solution of 3 mg/ml4-chloro-l-naphthol in methanol and 3 parts PBS with 0.01% Hz02 added just before use). The reaction was terminated by rinsing with distilled water. IEF control strips for CNPase activity were made by running unblotted nitrocellulose strips through the standard staining procedure. IEF control CNP strips were made by incubating myelin blots as described but without primary antibody.
94
67
RESULTS
Delipidation of the myelin samples proved essential since experiments with focusing of undelipidated myelin sometimes totally abolished CNPase activity staining and instead resulted in a nonspecific staining of the lipids. However, this nonspecific lipid “staining” took considerably longer time to develop. It did not appear until after 3-4 h. The protein content of the myelin fraction was 2531%. Figure 1 shows the normal SDS-PAGE pattern of myelin proteins (lane A) and the corresponding nitrocellulose blot incubated with anti-CNP antibody (lane B). The antiserum stains a protein doublet with estimated relative molecular weights of 45-47 kDa corresponding to the CNP doublet on the silver-stained reference lane. In Fig. 2, lane A is a reference myelin lane stained with Coomassie brilliant blue. Lane B is a gel run in parallel and stained for CNPase activity, showing that enzyme activity is associated with two bands focusing at PI,,, 8.1 and PI,,, 8.8, respectively. The more alkaline of the two bands cofocuses with a protein focusing as a broad band with the same pl as the myelin proteolipid protein [see Fig. 3 and (16)]. Lane C shows the binding of CNP antibody. Four distinct bands at PI,,, 4.2, 7.4, 8.8, and 9.3 are seen. Furthermore, a diffuse band at PI,,, 7.9-8.2 is recognized by the antibody. CNPase activity thus coincides with the diffuse PI,,, 7.9-8.2 band and the pIapp8.8 band. Lane D is a control gel strip for CNPase, which is blank. Lane E is a control strip for CNP immunoreactivity where primary anti-CNP antibody was omitted. Figure 3 shows the results of submitting the focused proteins in the excised IEF gel pieces to SDS-PAGE. This method does not result in the familiar “spot” appearance of the O’Farrell type two-dimensional gels, but instead a continuum of bands at different heights result; for example, two spots with very similar molecular weight but situated say 8 mm apart can, with the described method, be “caught” within the same gel piece and thus become pooled in the same SDS-PAGE sample well after being homogenized in SDS sample buffer. However, in our hands the resolution and reproducibility of the method are superior as compared with preliminary experiments with “true” two-dimensional separation. Lanes a-j show the protein patterns after silver staining, lane a being the more anodal. Lane f represents the region of the IEF gel where the more acidic of the
43
30
20
FIG. 1.
SDS-PAGE of myelin proteins. Lane A shows the separation pattern of myelin after silver staining. Lane B shows the pattern obtained after incubation of a nitrocellulose blot of the proteins in lane A with anti-CNP antibodies and detection with HRP-conjugated second antibodies.
CNPase activity-stained bands focused, and lane h, the region where the more alkaline CNPase activity band focused. As can be seen in Fig. 2, the CNPase activity coincides with the appearance, just above 43 kDa, of the CNP doublet on the SDS-PAGE gel in Fig. 3, with the more acidic subunit being the smallest. However, both subunits are represented in the IEF gel in a far broader pH region than indicated by the enzyme activity stain. This is demonstrated by the immunoblot pattern, showing one band at ~1~~~4.2 (which can also be seen in Fig. 3, lane b, and four bands between PI,,, 7.4 and 9.3 which are also represented in Fig. 3, lanes f-i.
ISOELECTRIC
4
5
fi
FOCUSING
‘I
OF
8
2’,3’-CYCLIC
$3
10
PH
t FIG. 2. IEF of myelin proteins. Lane A shows the pattern obtained after Coomassie blue staining. Lane B shows a sample run in parallel and stained for CNPase as described in the text. Lane C shows the pattern obtained after incubation of a nitrocellulose blot of the proteins in lane A with anti-CNP antibody and detection with HRP-conjugated second antibody. Lane D shows a control gel strip without myelin sample, but incubated as in lane B. Lane E shows a control gel incubated as in lane C except that the primary anti-CNP antibody was omitted. CorrespondingpH gradient shown on top was measured with a Beckman surface electrode.
DISCUSSION
The biological significance and function of CNP are unknown. So are its biosynthesis and degradation pathways [see (18) for a current review on these topics]. Previous electrophoretic studies of CNP in gels have been limited to SDS-PAGE or two-dimensional PAGE with NEPHGE in the first dimension (5,6), presumably due to difficulties in solubilization of myelin proteins and in obtaining a sufficiently alkaline, stable pH gradient for IEF in slab gels. However, IEF in slabs gives considerably higher resolution than SDS-PAGE and two-dimensional PAGE (7,8), making it the method more often used for studying the isoforms of various enzymes (19,20), genetic polymorphism, and enzyme microheterogeneity (21,22) and for evaluating enzyme substrate specificity (23). The inhibition of alkaline phosphatase activity exerted by the chelating effect of ampholines is well documented (24), but metal cofactors have never been demonstrated to be of importance for CNPase activity. The reason for staining the IEF gel itself instead of a blot was that since the gel is only 0.18 mm thick, diffusion time for reaction products and background is reduced so much that blotting with accompanying losses and risk of SDS denaturation would not mean any advantages. Post-translational modifications has been demonstrated in all myelin proteins investigated in this regard (25). Since this processing consists of phosphorylation, methylation, glycosylation, etc., with no studies describing cleavage of peptide bonds, this implies that the net charge rather than the molecular weight would be changed during these processes. Thus, for further studies of the metabolism of CNP as well as studies of its genetic variations, possible polymorphism and microheterogeneity in normal animals as well as in the array of frequently used dysmyelinating murine (26), canine, and rabbit (6) mutants, IEF seems
NUCLEOTIDE
93
3’-PHOSPHODIESTERASE
the method of choice. As can be seen in Fig. 3, protein bands with molecular weights corresponding to the CNP doublet (45 and 47 kDa) are present in a far broader region than the CNPase activity staining in Fig. 2B. Since CNPase activity staining of two-dimensional gels (5) gives a pattern corresponding to Fig. 2B rather than to Fig. 3, an explanation would be the presence of inactive CNP subunits with slightly different surface charges not resolvable by SDS-PAGE. Since immunoreactivity and enzyme activity are not equally distributed on the IEF gels/blots the possibility that the focusing pattern is an artifact, as a result of unspecific interactions of CNP and ampholines or due to detergent micelle formation, is highly unlikely. As a rule, enzymes resolved on IEF gels give rise to multiple bands, and this fact has been exploited in a number of ways by many different workers [e.g., (27-29)]. However, substrate specificity and enzyme activity have remained problems, which have led to efforts to circumvent them by using immunocytochemistry. This has indeed proven to be a powerful tool which, in addition, permits studies of inactive precursors and their intracellular pathways and compartmentalization (30-32). Also, the greater sensitivity of the immunological methods as compared with enzyme histochemistry will most probably sharpen the clinical value of enzyme fractionation. Thus it is not surprising that the pattern produced by the immunoblots in the present work is more complex than the enzyme activity staining pattern. The CNP separation pattern and reported PI,,, of active enzyme in the present work, 8.1 and 8.8, adds to and partly explains the previous confusion [i.e., 8.6 (33), ap-
94 67
43
30
20
FIG. 3. Silver-stained SDS-PAGE of excised gel pieces from a lane of myelin proteins as shown in Fig. 2, lane A. Each piece was approx 8 mm and lanes a-j cover the IEF gel pH gradient from left (= most acidic gel piece) to right (= most alkaline gel piece). The figures to the right indicate molecular weights as determined with protein standards.
94
PERSSON
AND
prox 9 (34,35), approx 9.7 (36), and 10.33 (37)]. Maybe the more alkaline CNP subspecies, inactive when stained for activity as described here and in (5), retain their activity when treated and detected otherwise. Particularly, this could be an additional factor to take into account besides contamination by myelin membranes when considering the problem of the high variability obtained by different laboratories in measuring CNPase activity in oligodendrocyte preparations [reviewed in (38)]. Calcium-dependent protein kinases, apparently endogenous to the myelin membrane, and cyclic AMP-dependent protein kinases have been shown to actively phosphorylate the larger subunit of CNP both in vitro (39) and in uiuo (5,40,41). Thus, one explanation for the discrepancy in the number of bands seen on SDS-PAGE and IEF gels could be charge heterogeneity due to different degrees of phosphorylation. A number of nervous system- and myelin-affecting pathological conditions are accompanied by changes in CNPase activity [reviewed in (IS)]. Studies are underway in our laboratory to investigate the separation pattern during experimental allergic encephalitis (a myelinaffecting pathological condition used as an “animal model disease” for multiple sclerosis) to correlate changes in separation pattern with the clinical course of the disease. The results are to be published in a forthcoming study. ACKNOWLEDGMENTS This work was supported by grants from the Swedish Medical Research Council (Project 03157) and Anna Ahrenbergs stiftelse. The authors thank Rita Grander and Marieanne Eriksson for excellent technical assistance. We are grateful to Dr. Terry Joe Sprinkle for generously providing the antiserum against CNP and Professor C-H. Berthold for always being an enthusiast and a support.
CORNELIUSON 10. Peterson, G. L. (1983) in Methods in Enzymology (Hirs, C. H. W., and Timasheff, S. N., Eds.), Vol. 91, pp. 95-119, Academic Press, New York. 11. Greenfield, S., Norton, W. T., and Morell, P. (1971) J. Neurochem. l&2119-2128. 12. Yao-Jun, G., andBishop, R. (1982) J. Chromatogr. 234,459-462. 13. Laemmli, U. K. (1970) Nature (London) 227,680-685. 14. Wray, W., Boulikas, T., Wray, V. P., and Hancock, R. (1981) Anal. Biochem. 118,197-203. 15. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. N&l. Acad. Sci. USA 76,4350-4354. 16. Johnson, T. K., Leonard Yuen, K. C., Denell, R. E., and Consigli, R. A. (1983) Anal. Biochem. 133,X26-131. 17. Draper, M., Lees, M. B., and Chan, D. S. (1978) J. Neurochem.
31,1095-1099. 18. Vogel, 1677.
U. S., andThompson,
19. Harris, H., and Hopkinson, Electrophoresis in Human
R. J. (1988)
J. Neurochem.
D. A. (1976) Handbook Genetics, North-Holland,
50,1667of Enzyme Amsterdam.
20. Needleman, S. B., and Koenig, H. (1974) Biochim. Biophys. Acta 379,43-56. 21. Kiihnl, P. (1981) in Electrophoresis ‘81 (Allen, R. C., and Arnaud, P., Eds.),
pp. 563-571, Koenig,
22. Patel, A., and 23. Lojda, Z., and 24. Latner, A. L., 40,1-7. J. 25. Benjamins, 26.
de Gruyter, Berlin. H. (1976) Neurochem.
Res.
1,275-298.
Kulich, J. (1981) Histochemistv 73,311-319. Parsons, M., and Skillen A. W. (1971) Enzymologia
A., and Smith, M. E. (1984) in Myelin (Morell, P., Ed.), pp. 225-258, Plenum, New York/London. Hogan, E. L., and Greenfield, S. (1984) in Myelin (Morell, P., Ed.), pp. 489-534, Plenum, New York/London.
27. Corless, J. K., and Middleton, H. M., III (1983) Arch. Intern. Med. 143,2291-2294. W. A., and Thorpe, J. W. (1980) Med. Sci. 28. Randall, T., Harland, zmI20,43-47. 29. Young, C. W., and Bittar, E. S. (1973) Cancer Res. 33,2692-2700. 30. Parenti, G., Willemsen, R., Hoogeveen, A. T., Verleun-Mooyman, 31.
M., Van Dongen, J. M., and Galjaard, H. (1987) 43,121-127. Braands, R., Slot, J. W., and Geuze, H. J. (1982)
Eur.
J. Cell Biol.
Eur.
J. Cell Biol.
27.213-220. 32. Van Dongen,
REFERENCES 1. Drummond, 1165.
2. Sprinkle, (1980)
R. J., and Dean,
G. (1980)
T. J., Wells, M. R., Garver, J. Neurochem. 35,1200-1208.
J. Neurochem. F. A., and
35,1155Smith,
D. B.
3. Miiller, H. W., Clapshaw, P. A., and Seifert, W. (1981) FEBS Lett. 131,37-40. 4. Miiller, H. W. (1982) FEBS L&t. 144,77-80. 5. Bradbury, J. M., and Thompson, R. J. (1984) Biochem. J. 221, 361-368. 6. Domanska-Janik, K., Gajkowska, B., de Nechaud, B., and Bourre, J. M. (1988)
J. Neurochem.
50,122-130.
7. Sinha, P., and Gossrau, R. (1984) Histochemistry 81,161-165. 8. Ogata, M., and Satoh, Y. (1988) Ekctrophoresis 9,128-131. 9. Norton, W. T., and Poduslo, S. E. (1973) J. Neurochem. 21,749757.
33. 34.
J. M., Barneveld, R. A., Geuze, H. J., and Galjaard, H. (1984) Histochem. J. 16,941-954. Suda, H., and Tsukada, Y. (1980) J. Neurochem. 34,941-949. Sprinkle, T. J., and Eller, A. G. (1979) Trans. Znt. Sot. Neurochem.
7,123. 35. Sprinkle, them.
T. J., Grimes, 34,880-887.
M. J., and Eller,
A. G. (1980)
J. Neuro-
36. Drummond, 37. 38. 39. 40.
R. J., Hamill, E. B., and Guha, A. (1978) J. Neurochem.3 1,871-878. Kurihara, T., Nishizawa, Y., Takahashi, Y., and Odani, S. (1981) Biochem.J. 195.153-157. Pleasure, D., Kim, S. U., and Silberberg, D. H. (1984) in Oligodendroglia (Norton, W. T., Ed.), pp. 175-190, Plenum, New York. Thompson, R. J., Bradbury, J. M., and Richardson, P. J. (1985) Biochem. Sot. Trans. 14,351-352. Bradbury, J. M., andThompson, R. J. (1984) Biochem. Sot. Trans.
12,1044-1045. J. M., and Thompson, 42. Bradbury, 351-359.
R. J. (1984)
Biochem.
J.
221,
