An evolutionary approach to myelin proteins and ... - Semantic Scholar

Nervous System-Specific Proteins

neuronal cells. Why differentiating Schwann cells should synthesize a NF-like protein remains unclear. At any rate, the possibility that NF-like proteins exist in other non-neuronal cells should now be considered seriously. This work was supported by the Multiple Sclerosis Society of Great Britain in Scotland.

I . Chin, S. & 1,iem. K. (lOO0) J. Neurosci. 10, 37 14-3726 2. Steinert, 1’. M. & I,iem, K.(1000) Cell 60, 521-523 3. Georgatos. S. 11.. Weber, K., Geisler, N. & 13lobel, G. (1087) I’roc. Natl. Acad. Sci. I1.S.A.84. 6780-0784 4. (kiger, H. (1083) Hiochim. Hiophys. Acta 737, 3-4 5 . Kaff. M. C., Miller. K. 11. & Noble, M. I). (1083) Nature (London) 303, 390-3Oh 0. Jessen, K. K., Morgan, I + Stewart. H. & Mirsky, K. (1000) Development 109, 01-103

7. Hunge, M. H., Hunge, K. P., Carey, L). J.? Cornbrooks, C. J,?Eldridge, C. F.,Williams, A. K. & Wood, 1’. M. (1 983) in Developing And Regenerating Vertebrate Nervous Systems (Coates, P. W., Markwald, K. K. & Kenny, A. D.?eds.), pp. 71-105, I,iss, New York 8. Sobue, G., Yasuda, T., Mitsuma, T. & Pleasure, I). (1986) Neurosci. Lett. 72,253-257 9. Morgan, I,., Jessen, K. R. & Mirsky, K. (1991) J. Cell l h l . 112, 457-407 10. Wilson, K. & l h p h y , P. J. (1080) J. Neurosci. Kes. 22.439-448 11. Mirsky, K., l h b o i s , C., Morgan, I,. & Jessen, K. K. (1990) Development 109, 105-1 I 0 12. Chin, S. & Liem, K. (1989) Eur. J. Cell Hiol. 50, 475-490

Keccived 16 April 1002

An evolutionary approach to myelin proteins and myelin-forming cells in the vertebrate brain Gunnar ]eserich* and Thomas V. B. Waehneldt-Kreysingt *Universitat Osnabruck, D-4500 Osnabruck, Germany, and +Max Planck-lnstitut fur Experimentelle Medizin, D-3400 Gottingen, Germany I t is well documented through studies of mam-

malian species that myelin consists of spirally fused stacks of intimately packed glial membrane pairs, which help to increase the velocity of nerve impulses via insulation [ 1. 21. Although myelin is known to be extremely rich in lipids our emphasis is on the relatively simple proteinaceous make-up of myelin membranes in the central (CNS) and peripheral (PNS) nervous systems. CNS myelin is characterized by the presence of large proportions of the unglycosylated. strongly hydrophobic proteolipid protein (P1,P); it is synthesized by oligodendrocytes [ 1, 31. PNS myelin. by contrast, is characterized by massive amounts of the glycosylated PI, protein, also displaying strongly hydrophobic features; this is synthesized by Schwann cells [ 1, 31. Both PLP and PI, are transmembrane proteins of about 30000 Da; however, they display no sequence similarity and their genomic arrangements as well as their membrane topology are entirely different [ 4-81. Nevertheless, they presumAbbreviations used: CNP. 2’,3‘-cyclic nucleotide 3’phosphodiesterase; CNS, central nervous system; MAG, myelin-associated glycoprotein; MHI’, myelin basic protein; P I P 3 proteolipid protein; I’NS, peripheral nervous system.

ably function in an analogous way, maintaining tight appositions of the external glial faces. In spite of marked structural differences (see below) oligodendrocytes and Schwann cells share the capacity to synthesize substantial amounts of a hydrophilic protein component of about 18 000 I)a, myelin basic protein (MBP). This protein is presumed to be instrumental as a ‘molecular glue’ in the adhesion of glial cytoplasmic surfaces [91.

Evolution of myelin proteins The proteins mentioned above, i.e. PLP, P,, and MHI’, are highly conserved within the mammalian class and hence appear well adapted functionally. Hecause mammals constitute only about 10% of all living vertebrates, the question arises of whether the protein composition described above is also valid for the other vertebrate classes, including birds, reptiles, amphibians and bony (teleostei) and cartilaginous fishes (chrondrostei). When brain tissues of all myelin-forming vertebrate classes are examined, the following picture emerges (Fig. 1): CNS myelin proteins of birds, reptiles and amphibians (comprising the group of tetrapods, including mammals) show essentially the same pattern as those of the mammalian class [lo], the only exception being the absence of DM-20 in amphibians, a PLP-related

I992

617

Biochemical Society Transactions

Fig. I Occurrence of myelin proteins in the CNS of the major vertebrate classes

618

For abbreviations see text.

c

DM-20

pLp

-CNP

oMBP+MAG

Po

component carrying an internal deletion of 35 amino acids. The proteins of bony and cartilaginous fishes differ in several respects. Firstly, aside from the omnipresent MBP, at least two hydrophobic components ranging from 23-30000 Da are observed. Secondly, in contrast to mammalian PLP these components are glycosylated. Thirdly, the two trout proteins IP1 and IP2 [ 111 and the electric ray proteins T1 and T2 [ 121 react strongly with antibodies against anti-mammalian Po upon immunoblotting, whereas anti-mammalian PLP antibodies show no reaction [13]. The Po-relatedness of hydrophobic fish CNS myelin glycoproteins is substantiated further by cDNA sequencing of shark P,, ~141. Another feature that differs from tetrapods and even from cartilaginous fishes is the presence of an additional unglycosylated hydrophobic component of about 36 kDa in teleostean fishes [ 15, 161. Upon immunoblotting, the 36 kDa protein reacts with none of the known myelin protein components [ 111; moreover, recent cDNA sequence studies have revealed no sequence similarity with other protein sequences (T. V. Waehneldt et al., unpublished work). Thus, these observations suggest a surprising break or hiatus regarding the major hydrophobic CNS myelin proteins when progressing from rayfinned (actinopterygian) fishes (Po) to land-living tetrapod classes (PLP; Fig. 1). At this stage it was of great interest to examine the CNS myelin proteins of lobe-fined (sarcopterygian) fishes. Today only two representatives

Volume 20

are extant: lungfish and coelocanth. Both groups carry PLP in their brains but no P,).Lungfish PLP is glycosylated [ 17, 181 and reacts only weakly with anti-mammalian PLP antibodies, whereas coelocanth PLP reacts more strongly and is not glycosylated [13, 191. On this immunological basis the coelocanth Latimeria chalumnae rather than the lungfish appears to be more closely related to tetrapods [13, 191. CNS myelin from all vertebrate classes is characterized by the presence of myelin-associated glycoprotein (MAG) [20] . In contrast to MAG’S CNS myelin ubiquity, the myelin-related enzyme 2’,3’-cyclic nucleotide 3’-phosphodiesterase (CNP) is expressed in sizeable amounts only in tetrapods [21] and does not occur in CNS myelin of fishes, including PLP-carrying coelocanth and lungfish (Fig. 1). Taken together, and with caution, CNS myelin proteins may be used as phylogenetic markers. Apart from MHP and MAG, which are found in myelin of all vertebrate classes, P,, in cartilaginous and bony fishes is replaced by PLP not only in tetrapods but also in lungfkh and coelocanth, thus linking these sarcopterygian fishes to tetrapods. Only CNP appears to be a truly tetrapodal marker [19]. However, the role of CNP in tetrapods remains enigmatic, particularly in view of fully functional myelin in lungfish and coelocanth. In this context, the excessively high CNP values in amphibians [19], and also in larval stages, should be mentioned. The functional purpose of the teleostean 36 kDa protein is entirely unknown to date. Nevertheless its functional significance appears obvious, considering that more than half of all extant myelin-forming vertebrates (i.e. the huge class of bony fish) carry this protein. The glycosylation of lungfish PLP [17, 181, on the other hand, may be an ‘oddity’ or at best a dead-end specialization and of no importance for far-reaching cladogrammatical considerations. The following very crucial question now arises: could the Po-producing cells in CNS of bony and cartilaginous fishes possibly be called ‘Schwann cells’, quasi in analogy to the P,,-producing cells in the PNS of tetrapods?

Evolution of myelin-forming cells Oligodendrocytes and Schwann cells, the myelinforming cells of the mammalian CNS and PNS, respectively, are markedly different from each other in terms of morphology, biochemistry and glial cell lineage, although in principle they share the same cell function. Whereas the Schwann cell typically


lies in close apposition with its rnyelinated axon and elaborates only a single myelin segment [ 2 2 ] ,an oligodendrocyte can ensheath as much as 40 axonal internodes and is connected to nerve fibres via long slender processes [ 22, 23 I . This morphological diversity is complemented by specific molecular differences in the myelin membranes synthesized by either cell type: as mentioned above the glycoprotein P,, is the major gene product of SchLvann cells I I , 31, whilst biosynthesis of P I 2 is a property uniclut, to oligodendrocytes [ I , 31. Furthermore, both cell types basically differ in their dependence on an ongoing axonal contact for the induction and nxiintenance of their myelinogenic features during development [24-263 and regeneration [27]. Finally, Schwann cells and oligodendrocytes belong to cwmpletelr. different cell lineages, the former heing derivatives of the population of neural crest cells 128, 201 and the latter originating from proliferative precursor cells of the subventricular zones I 30 1, being phenotypically characterized by the occurrence of the ganglioside epitope A2HS on their cell surface 131 I. As the myelin-forming glia in the CNS of fishes, a s pointed out, seem t o share specific biochemical traits with Schwann cells in the mammalian I’M, the above-mentioned criteria were applied to define more closely the cellular nature of these cells. Electron microscopical analyses [ 1 1, .32] as well as immunocytochemical studies [ 331 using ;intibodies against myelin-specific proteins carried out mitt1 the CNS of a bony fish (trout) have reve;iled ;i close structural similarity with mammalian oligodendrocytes, in that the cells did not show ;I one by one relationship with myelinated axons, but extended rnultiple slender processes to nerve fibres in their vicinity. I Ience, on the basis of morphological criteria these cells have to be classified a s ‘oligodendrocytes’. ‘1’0 further characterize the physiological and biochemical properties of this type of oligodendrocytes, cell cultures of brain tissue from the teleostean trout were employed. Oligodendrocytes dissociated from mature trout brain and highly enriched by I’ercoll density centrifugation 1351 were taken in culture and were analysed irnmunocytochemically using a panel of monoclonal and polyclonal antibodies raised against each of the major myelin constituents of trout CNS 1.15. 361. After a few, days in vitro the cells adopted a multiple branched morphology (Fig. 2a), resembling their mammalian counterparts. When kept in cell culture for longer times in the absence of neuronal contacts, oligodendrocytes of

Fig. 2 Identifying fish oligodendrocytes in culture (a) lrnrnunocytochemical labelling of cultured trout oligodendrocytes after 4 days in vitro using the monoclonal antibody 6D2, which recognizes the fish myelin glycoproteins IPI and IP2 Bar indicates 40 ,urn (b) Phase-contrast micrograph of oligodendroglial progenitor cells isolated from larval trout brain and cultured for 3 days Bar indicates 50 ,urn

the mature trout CNS, in a similar manner to

Schwann cells, rapidly ceased to express galactocerebroside, a glycolipid characteristic to myelin [ 25, 361. The myelin 36 kI>a protein and IPI were also rapidly down-regulated, but gradually reappeared after prolonged culture periods [35]. In contrast, expression of the myelin glycoprotein IP2 was maintained constantly at high levels over weeks in culture in the entire absence of neuronal influences, perfectly mimicking the in vitro behaviour of mammalian oligodendrocytes 12.5, 371. During in vivo development the biochemical differentiation of trout oligodendrocytes proceeds in a sequential manner, the individual steps being phenotypically reflected by the consecutive appearance of myelin-specific antigens: the myelin glycoprotein IP2 together with galactocerebroside emerge first, followed by the 36 kDa protein and

I992

619

Biochemical Society Transactions

620

finally IP1 [ 381. Cell culture experiments using glial cells derived from early larval stages of trout brain revealed that immature oligodendrocytes were unable to differentiate in vitro beyond the level of IP2 expression - the first step of maturation in vivo - when cultured in the absence of neurons [39]. Hence the myelin-forming glia of fish CNS, in a similar manner to mammalian Schwann cells, seem to require appropriate inducing signals from other cells, e.g. neurons, for proper elaboration of their developmental programme [24, 261. This is in overt contrast to the in vitro behaviour of mammalian oligodendrocytes which, independently from the influence of other cell types in culture, go through a complete biochemical differentiation on the same time scale seen in the intact brain [40, 411. In order to identify oligodendroglial precursor cells of fish CNS, the monoclonal antibody A2H5 was employed [42]; this is known to label mammalian progenitor cells of oligodendrocytes [ 3 11. A large proportion of cells dissociated from larval trout brain were A2B5+, a few of them being labelled concurrently by the monoclonal antibody 6D2, an oligodendroglial marker of bony fish CNS [34, 351. In cell culture as well as during in vivo development the population of A2H5+/6D2+ cells strongly expanded initially, thereafter being replaced by A2H5 -/6D2+ oligodendrocytes [ 39, 431. On the basis of these results, an immunomagnetic cell separation technique was designed, which allowed the isolation of A2H5+ cells from larval trout brain in very high purity. When taken in culture (Fig. 2b) many of these cells in fact differentiated into 6D2 oligodendrocytes. In addition glial fibrillary acid protein positive cells with an astrocyte morphology developed, suggesting that A2R 5 cells can give rise to both cell types. It is not known, however, if they originate from a common precursor cell or belong to separate lineages accidentally sharing the A2B5 epitope on their cell surface. Taken together, both the morphology and cell lineage of the myelin-forming cells of bony fish CNS closely match oligodendrocytes of the mammalian CNS. In terms of biochemistry, however, especially regarding the regulation of antigenic expression related to myelinogenesis, the cells strikingly resemble Schwann cells of the mammalian PNS. In this context, it should be further stressed that bony fish oligodendrocytes have a remarkable capacity for remyelination [44, 451 and are able to support neuritic regenerative outgrowth in vitro [46], in a similar manner to mammalian Schwann cells. Therefore, studies on the specific molecular +

+

Volume 20

properties of fish oligodendrocytes and the mechanisms regulating myelin gene expression in these cells appear of particular future interest. In contrast with the situation in bony fish, much less is known about the biochemical and physiological features of myelin-forming cells in the remaining lower vertebrate groups, e.g. cartilaginous fish, lungfish or amphibia. Thus, a comparative characterization of glial cells in these species is necessary to allow a more comprehensive description of the phylogenetic relationships among the myelin-forming glia in the different vertebrate classes. The authors would like to thank Mrs. 13. 1;lenker for skilful technical help. The work presented here was supported by the Lkutsche Forschungsgemeinschaft (grants Je 114/4-1 and SFB 171/C13).

1. Raine, C. S. (1084) in Myelin (Morell, I’.. ed.), pp. 1-50, Plenum I’ress, New York 2. Ritchie, J. M. (1084) in Myelin (Morell, I)., ed.), pp. 117-145, I’lenum I’ress, New York 3. Braun, 1’. E. (1984) in Myelin (Morell. I’., ed.), pp. 97-1 16, Plenum I’ress, New York 4. Laursen, K. A,. Samiullah, M. & Lees, M. 13. (1984) I’roc. Natl. Acad. Sci. IJ.S.A. 81,20 12-20 1h 5. Stoffel, W., Hillen, II. & Giersiefen, 11. (1084) I’roc. Natl. Acad. Sci. lI.S.A.81. 5012-5010 6. I,emke, G. & Axel, K.(1985) Cell 40, 501-508 7. Diehl, H. J., Schaich, M., Hudzinski, K.-M. & Stoffel, W. (1986) I’roc. Natl. Acad. Sci. 1J.S.A. 83, 9807-98 1 1 8. I,emke, G., Lamar. E. & I’atterson, J. (1088) Neuron 1,73-83 9. Lees, M. H. & Hrostoff, S. W. (1984) in Myelin (Morell. P., ed.). pp. 197-224 10. Waehneldt, T. V., Malotka, J.. Karin, N. J. & Matthieu, J.-M. (1985) Neurosci. Lett. 57, 97- 102 11. Jeserich, G. & Waehneldt, T. V. (1980) J. Neurochem. 46, 525-533 12. Waehneldt, T. V.. Kiene, M.-I,., Malotka, J., Kiecke, C. & Neuhoff, V. (1084) Neurochem. Int. 6,223-235 13. Waehneldt, T. V.. Malotka, J., Gunn, C. A. & I,inington, C. (1990) in NATO AS1 Series, Vol. 1143, Cellular and Molecular Biology of Myelination Ueserich, G., Althaus, H. H. & Waehneldt, T. V., eds.), pp. 361-372. Springer Verlag, Herlin 1Ieidelberg 14. Saavedra, K.A., Fors, I,., Aebersold, K.I I., Arden, B., Horvath. S., Sanders, J. & Hood, I,. (1089) J. Mol. Evol. 29, 149- 156 15. Jeserich, G. (1983) Neurochem. Hes. 8,057-070 16. Waehneldt, T. V.. Stoklas, S.,Jeserich, G. & Matthieu, J.-M. (1986) Comp. Hiochem. I’hysiol. 84.273-278 17. Waehneldt, T. V., Matthieu, J.-M. & Jeserich. G. (1986)J. Neurochem. 46. 1387- 139 1 18. Waehneldt, T. V., Matthieu, J.-M., Malotka, J. & Joss, J. (1987) Comp. Hiochem. I’hysiol. 88, 1200- 1212


10. Waehneldt, T. V. (1000) in Ann. N. Y. Acad. Sci., Vol. 605, Myelination and Llysmyelination (Duncan, I., Skoff, K. & Colman, U. K., eds.), pp. 15-28. 20. Matthieu, J.-M., Waehneldt. T. V. & Eschmann, N. (1086) Neurochern. Int. 8, 521-526 21. Waehneldt, T. V.. Matthieu, J.-M. & Jeserich, G. (1086) Neurochem. Int. 9, 463-474 22. I’eters, A,, I’alay, S. & Webster, H. deF. (1976) The Fine Structure of the Nervous System, W. €3. Saunders, Philadelphia 23. Sternherger, N. lL, Itoyama, Y., Kies, N. W. & Webster, H. deF. ( 1978)J. Neurocytol. 7,25 1-203 24. Hrockes. J. I’.. Kaff, M. C., Nishiguchi, I). H. & Winter, J. (1070) J. Neurocytol. 9, 60-77 25. Mirsky, K., Winter. J., Abney, E. K.,I’russ, R. M., Gavrilovic, J. & Kaff, M. C. (1980) J. Cell Hiol. 84, 483-494 20. 1,ernke. G. & Chao, M. C. (1988) L>evelopment 102, -100-504 27. I’olitis, M. J., Sternberger, N., Ederle, K. & Spencer, 1’. S.(1082) J. Neurosci. 2, 1252-1260 28. I,e Douarin, N. M. (1982) The Neural Crest, Cambridge [Jniversity I’ress, Cambridge 20. Jessen. K. K. & Mirsky, K. (1991) Glia 4, 185-194 30. I,eVine, S. I,. & Goldman. J. E. (1988) J. Neurosci. 8, 3092-4006 31. Kaff, M. C., Miller. K. H. & Noble, M. (1083) Nature (Idondon)303, 300-300 32. Kruger, I,. & Maxwell, L). S. (1067) J. Comp. Neurol. 129, 115-142 33. Jeserich, G. & Waehneldt, T. V. (1086) J. Neurosci. Kes. 15, 147-1 58

34. Jeserich, G. & Kauen, T. (1990) Glia 3, 05-74 35. Jeserich, G., Kauen, T. & Stratmann, A. (1990) in NATO AS1 Series, Vol. H 43, Cellular and Molecular Biology of Myelination (Jeserich, G., Althaus, H. H. & Waehneldt, T. V.. eds.), pp. 343-359, Springer Verlag, Berlin Heidelberg 30. Kanscht, H., Clapshaw. 1’. A., Price, Noble, M. & Seifert, W. (1982) Proc. Natl. Acad. Sci. USA. 79, 2709-27 13 37. Szuchet, S. (1987) in NATO AS1 Series, Vol. H2, Glial-Neuronal Communication in 1)evelopment and Kegeneration (Althaus, H. H. & Seifert, W., eds.), pp. 756-777, Springer Verlag. Berlin Heidelberg 38. Jeserich, G., Muller, A. & Jacque, C. (1990) Dev. Hrain Kes. 51.27-34 30. Jeserich, G. & Stratmann, A. (1992) 1)ev. Hrain Kes., in the press 40. Duhois-Dalcq, M. E., Hehar, T. N., Hudson, I,. & 1,azzarini. K. A. ( I O X h ) J. Cell Hiol. 102, 384-392 41. Gard, A. I,. & I’feiffer, S. E. (1989) Development 106, 119-132 42. Eisenharth, G. S.,Walsh, F. S. & Nirenberg, M. (1979) Proc. Natl. Acad. Sci. IJSA. 76, 491 3-4917 43. Sivron, T., Jeserich, G., Nona, S. & Schwartz, M. (1992) Glia, in the press 44. Murray, M. (1976) J. Comp. Neurol. 168, 175-196 45. Wolburg, H. ( 198 1) Exp. Brain Kes. 43, 100-206 46. Hastmeyer, M., Heckmann, M., Schwab, M. E. & Stuermer, C. A. 0.(1001) J. Neurosci. 11, 026-640 J.?

Received 16 April 1 002

2’,3’-Cyclic nucleotide-3’-phosphohydrolase and signal transduction in central nervous system myelin R. J. Thompson University Clinical Biochemistry, Level D, South Laboratory and Pathology Block, Southampton General Hospital, Tremona Road, Southampton SO9 4XY, U.K. Although 2’,3’-cyclic nucleotide 3’-phosphohydrolase (EC 3.1.4.37-CNPase) activity was initially described in bovine spleen and pancreas, 30 years have now elapsed since it was realized that high levels of activity were present in the vertebrate nervous system [ 11. T h e enzyme hydrolyses 2’,3’cyclic nucleotides (but not 3’,5’-cyclic nucleotides) to produce exclusively the 2’derivatives. This property, together with the lack of requirement of a metal ion cofactor and the ability to hydrolyse all four 2’,3‘-cyclic nucleotides, distinguishes CNPase

Abbreviations used: CNI’ase, 2’,3’-cyclic nucleotide 3’~liosptiohydrolase;CNS, central nervous system; MAG, myelin-associated glycoprotein; MBI’, myelin basic protein; 1’1 J’, proteolipid protein; PI’, protein phosphatase.

from most other known phosphodiesterases [2, 31. CNPase also hydrolyses oligonucleotides with a 2’,3’-cycIic terminus [Z, 31. Purine-containing nucleotides are hydrolysed at a faster rate than pyrimidine-containing nucleotides [ 31. Fluorescent derivatives of 2’,3’-cyclic nucleotides o r 2’,3’-cyclic NADP are commonly used as in vztro substrates of the enzyme; K,,,values for such substrates (depending on the substrate and the source of the CNPase) range from 0.2 m v to as high as 10 m v [ 31. What appears to be true CNPase activity can be detected in a wide range of vertebrate organs; however, the specific activity in brain a n d spinal cord is usually at least one and often two orders of magnitude greater than that found in other tissues [ 3 ] . During evolution, significant CNPase activity in the central nervous system (CNS) first appears in amphibians

I992

62 I