Post-translational Modifications of Chicken Myelin Basic Protein Charge Components

Purified myelin basic protein (MBP) from various species contains several post-translationally modified forms termed charge components or charge isomers. Chicken MBP contains four charge components denoted as C1, C2, C3 and C8. (The C8 isomer is a complex mixture and was not investigated in this study.) These findings are in contrast to those found for human, bovine and other mammalian MBP’s. Mammalian MBP’s, each of which contain seven or eight charge components depending on the analysis of the CM-52 chromatographic curves and the PAGE gels obtained under basic pH conditions. Chicken MBP components C1, C2 and C3 were treated with trypsin and endoproteinase Glu-C. The resulting digests were analyzed by capillary liquid chromatography combined with either an ion trap tandem mass spectrometer or with a Fourier transform ion cyclotron resonance mass spectrometer. This instrumentation permitted establishing the amino acid composition and the determination of the post-translational modifications for each of the three charge components C1-C3. With the exception of N-terminal acetylation, the post-translational modifications were partial. The C1 component lacks any phosphorylated sites, a finding in agreement with the analysis of other MBP species. It also had a single methylation at R105 as did the components C2 and C3. The C2 component contains ten phosphorylated sites (S7, S18, S33, S64, S73, T96, S113, S141, S164, and S168), and modified arginine to citrulline residues at R24, and R165. Component C3 contains eight phosphorylated sites (S7, S33, S64, T96, S113, S141, S164, and S168), and citrulline residues at Arginine 41, R24 and R165. Partial deamidation of glutamine residues Q71, Q101 and Q146 were present in addition to asparagine N90 that was found in all three charge components. The glutamine at residue 3 is partially deamidated in isomers C1 and C2, whereas glutamine 74 and asparagine 83 were found not to be deamidated. Comparison of the PTM’s of MBP’s isolated from several vertebrate species reveals marked differences in their phosphate content. Chicken MBP does not share any phosphorylated sites with dogfish MBP; However, it does contain phosphorylated serine and threonine residues in common with mammalian MBP.     

Studies of posttranslational modifications in spiny dogfish myelin basic protein

The objective of this investigation was to determine whether nonmammalian myelin basic protein contained charge isomers resulting from extensive posttranslational modifications as seen in mammalian MBP. Four charge isomer components from dogfish MBP have been isolated. These forms arise by phosphorylation and deamidation modifications. Components C1, C2 and C3 have been characterized. We are currently characterizing component C8. Dogfish MBP is less cationic than mammalian MBP and has about 50% lower mobility on a basic pH gel electrophoresis relative to human and to bovine MBP. The mammalian component C1, which is unmodified, is modified in the dogfish by phosphorylation. The reduced electrophoretic mobility is largely attributable to the charge reduction resulting from phosphorylation in serine 72, 83, and 120 or 121 in C1, and C3. In component C2, two or three phosphate groups were distributed among residues 134, 138 and 139. It was found that dogfish amino acid residue 30 was a lysine residue and not a glutamate residue as reported in the literature.     

Myelin basic protein undergoes a broader range of modifications in mammals than in lower vertebrates

Myelin basic protein (MBP) is an important component of the myelin sheath surrounding neurons, and it is directly affected in demyelinating diseases. MBP contains a relatively large number of post-translational modifications (PTMs), which have been reported to play a role in multiple sclerosis, while MBPs from lower vertebrates have been reported to be incapable of inducing multiple sclerosis or allergic encephalitis. This study reveals the extent of differences in PTM patterns for mammalian and nonmammalian MBPs. This included intact mass and de novo sequence analysis of approximately 85% of rattlesnake MBP, the first reptile MBP to be characterized, and of bovine MBP. We identified 12 PTMs at 11 sites in the five bovine MBP charge components, which include both previously reported and novel modifications. The most notable modification is an acetylation of lysine 121. Other modifications found in bovine MBP include N-terminal acetylation in components C1, C2, and C3; oxidation of methionine 19 in all five components; all charge isomers having both a mono- and dimethylated (symmetric) arginine at position 106; deimination in arginines 23 and 47 found only in component C8b; deimination of arginine 96 and deamidation in glutamine 102 found in components C2, C3, C8a, and C8b; phosphorylation in threonine 97 restricted to charge components C2 and C3; deimination in arginine 161 only found in component C3; deamidation of glutamine 120 was only observed in C3. All four deiminated arginines and one acetylated lysine were first experimentally revealed in this study for bovine MBP. Mascot database searching combined with de novo sequence analysis of rattlesnake MBP provided more than 85% sequence coverage. A few PTMs were also revealed in rattlesnake MBP: mono- and dimethylated Arg, protein N-terminal acetylation, and deiminated Arg. Overall, snake MBP was found to undergo less modification than bovine MBP on the basis of the mass heterogeneity of the intact protein, the bottom-up structure analysis, and the limited complexity of rattlesnake MBP chromatography. The combined data from this study and information from previous studies extend the known MBP PTMs, and PTMs unique to higher vertebrates are proposed.     

The contribution of phosphorylation and loss of COOH-terminal arginine to the microheterogeneity of myelin basic protein.

Two guinea pig myelin basic protein preparations which differed markedly in their contents of high pH electrophoretic or chromatographic forms were studied in an attempt to elucidate the causes of their microheterogeneity. Both total preparations and components isolated therefrom were examined for their amino acid compositions, NH2-terminal and COOH-terminal residues, total phosphorus contents, amd contents of phosphamino acids. The results showed that the five components differed sequentially by a single charge and that the microgeterogeneity arose as a result of secondary modifications of a single secies (Component 1) Of basic protein. Two modifications were demonstrated; viz. phosphorylation of serine and threonine and loss of COOH-terminal arginine. These two modifications were insufficient to account completely for the observed microheterogeneity; an additional cause, deamidation, was postulated. From the relationship between the number of components present in the total basic protein, the phosphorus and phosphoamino acid contents of the components, and the changes in relative electrophoretic mobility of the components which accompanied their phosphorylation and dephosphorylation we conclude that in the native basic protein no more than two sites in any polypeptide chain are phosphorylated.     

Primary structure of equine myelin basic protein by mass spectrometry

Equine myelin basic protein (MBP) has been isolated from spinal cord and shown to consist of a number of components (charge isomers) by alkaline-urea gel electrophoresis. Mass analyses of several of these components showed that each was posttranslationally modified and some have been identified. Component 1, the most cationic charge isomer, was sequenced by a combination of liquid chromatography and mass spectrometry of peptides obtained by proteolytic digestion. At 172 residues it is slightly larger than the bovine (169) and the human (170). A major difference between bovine and equine sequences was the replacement of AQGH (bovine residues 76-79) by SRDG (equine). A number of other replacements involving single amino acids were also found. Methylated arginine (residue 108 equine) was found as both the mono- and the dimethylated derivative and represents the first MS/MS evidence for this modification in any MBP.     

Effect of bovine basic protein charge microheterogeneity on protein-induced aggregation of unilamellar vesicles containing a mixture of acidic and neutral phospholipids

Two of the charge isomers (components 1 and 2) normally found as microheteromers of myelin basic protein were isolated, and their abilities to aggregate vesicles consisting of mixed phospholipids were studied. Component 1 (the most cationic of the microheteromers) aggregated phosphatidylcholine (PC) vesicles containing 7.8 mol% phosphatidylserine (PS) more rapidly and at lower protein concentrations than component 2, which differs from component 1 by 1 net positive charge. Modification of components 1 and 2 in vitro by phosphorylation with rabbit muscle protein kinase decreased the ability of both components to aggregate vesicles. The greater the extent of phosphorylation, the less effective were the isomers at inducing aggregation. Decreasing the charge of either component 1 or component 2 by removal of the two C-terminal arginyl residues also decreased the ability of the isomers to induce aggregation. Therefore, charge microheterogeneity, whether arising in vivo or generated in vitro, markedly affected the ability of these microheteromers to aggregate PC vesicles containing 7.8 mol% PS. Because a small difference in the charge of the protein had a marked effect on vesicle aggregation, we propose that charge microheterogeneity may play an important and dynamic role in the structure and function of normal myelin.     

The binding of calmodulin to myelin basic protein and histone H2B.

1. A calmodulin-binding protein of apparent mol.wt. 19 000 has been purified from chicken gizzard. Similar proteins have been isolated from bovine uterus, rabbit skeletal muscle and rabbit liver. 2. These proteins migrated as an equimolar complex with bovine brain calmodulin on electroporesis on polyacrylamide gels in the presence of Ca2+ and 6M-urea. The complex was dissociated in the presence of EGTA. 2. The chicken gizzard calmodulin-binding protein has been shown to be identical with chicken erythrocyte histone H2B on the basis of partial amino acid sequence determination. 4. The calmodulin-binding proteins of apparent mol.wt. 22 000 isolated previously from bovine brain [Grand & Perry (1979) Biochem. J. 183, 285-295] has been shown, on the basis of partial amino-acid-sequence determination, to be identical with myelin basic protein. 5. The activation of bovine brain phosphodiesterase by calmodulin is inhibited by excess bovine uterus calmodulin-binding protein (histone H2B). 6. The phosphorylation of myelin basic protein by phosphorylase kinase is partially inhibited, whereas the phosphorylation of uterus calmodulin-binding protein (histone H2B) is unaffected by calmodulin or troponin C. 7. The subcellular distribution of myelin basic protein and calmodulin suggests that the two proteins do not exist as a complex in vivo.     

On basic proteins in bovine peripheral nerve myelin.

1. Myeline proteins in bovine peripheral nerve migrated as two main band-(BF and BR protein) and one faint middle band (BM protein) on sodium dodecyls sulfate-polyacrylamide gel electrophoresis. The relative mobility of these two main bands differed from those of myelin proteins in the central nervous system. 2. The acid extract of the myelin fraction from bovine peripheral nerve was separated into one main peak and two minor peaks on a Sephadex G-75 column. The major component of the second minor peak was the BM protein; the major component of the main peak was the BF protein. The BR protein was not extractable by acid solution. 3. Molecular weights of the BF, the BM and the BR protein were determined as around 13 000, 20 000 and 28 000, respectively, by sodium dodecylsulfate-polyacrylamide gel electrophoresis. 4. The amino acid composition of the BF protein was quite different from the encephalitogenic protein and the Folch-Lees type proteolipid protein in the central nervous system. However the BM protein showed similar amino acid composition to the encephalitogenic protein. 5. The tryptic peptide maps of the BF protein and of the encephalitogenic protein were quite different. The results suggested that the amino acid sequences of these two proteins are different and that they contain no common tryptophan-containing peptide.     

Studies on the Wolfgram High Molecular Weight CNS Myelin Proteins: Relationship to 2'',3''-Cyclic Nucleotide 3''-Phosphodiesterase

Evidence is presented that the major protein components of the high molecular weight CNS myelin proteins designated as the Wolfgram protein doublet (W1 and W2) contain the enzyme 2',3'-cyclic nucleotide 3'-phosphodiesterase (EC 3.1.4.37, CNP). CNP is a basic hydrophobic protein containing about 830 to 840 amino acid residues. When electrophoresed on SDS polyacrylamide gels, CNP appears as a protein doublet, separated by a molecular weight difference of about 2500-3000 in bovine, human, rat, guinea pig, and rabbit. A similar protein doublet has been identified as the Wolfgram proteins W2 and W1 in myelin and in the chloroform-methanol-insoluble pellet obtained from myelin. Moreover, the relative Coomassie blue staining intensity of the CNP2 plus CNP1 protein doublet among the species examined was remarkably similar to that observed for electrophoresed myelin and chloroform-methanol-insoluble pellet derived from myelin. Antisera raised against purified bovine CNP recognized the W1 and W2 proteins isolated from bovine and human brain. The amino acid composition of pure bovine CNP is presented and compared with the compositions of several rat and bovine Wolfgram proteins obtained by other investigators. Our electrophoretic, compositional, and immunological data support the contention that the enzyme CNP is a major component of the Wolfgram protein doublet.     

Soluble bovine galactose-binding lectin. cDNA cloning reveals the complete amino acid sequence and an antigenic relationship with the major encephalitogenic domain of myelin basic protein.

A full-length cDNA clone for the 13-14 kDa soluble beta-galactoside-binding lectin was isolated from a bovine fibroblast cDNA library. The derived amino acid sequence shows eight differences from a preliminary partial amino acid sequence given previously for the bovine heart lectin. This observation led to a re-examination of the data and correction of the heart lectin protein sequence. Except for a possible polymorphism of the heart lectin at position 57, the fibroblast and heart lectin sequences are considered identical. The epitope recognized by two monoclonal anti-(bovine lectin) antibodies, 36/8 and 9/5, was identified as the tetrapeptide sequence W-G-A/S-E/D by the isolation of several different cDNA clones from a human intestine cDNA library. A similar tetrapeptide is present in all of the soluble beta-galactoside-binding animal lectins sequenced thus far. It is also found in myelin basic protein, which we show is antigenically cross-reactive with the lectin. In myelin basic protein the tetrapeptide is a part of the major domain previously shown to be responsible for the induction of experimental allergic encephalomyelitis.     

The role of charge microheterogeneity of human myelin basic protein in the formation of phosphatidylglycerol multilayers

Human myelin basic protein (HBP) was fractionated into its various charge isomers by chromatography on CM-52 columns at pH 10.6. Components 1,2,3 and "8" (C-1, C-2, C-3, and C-"8") were cleanly separated. Each component was combined with phosphatidylglycerol (PG) vesicles, at neutral pH at a concentration of 30% (w/w), protein/lipid. C-1, the most cationic of the components was the most effective at inducing the formation of multilayers when studied by liquid X-ray diffraction. C-3, which differs from C-1 by 2 positive charges was less effective than C-2. C-"8" was totally ineffective since the scattering pattern with this component was no different from that of the pure lipid. Thus a seemingly small change in net charge of the protein had a dramatic effect on the ability of the protein to organize the lipid into a crystalline, multilayer arrangement characteristic of compact myelin.     

A New Form of Myelin Basic Protein Found in Human Brain

Human myelin basic protein was subjected to ion-exchange chromatography at high pH to separate the differently charged components. Polyacrylamide gel electrophoretic patterns of the fractions showed that the less basic fractions 3, 4, and 5 contained significant amounts of a protein somewhat smaller than the more common 18.5-kDa form. Fraction 3 consisted of approximately equal amounts of this smaller polypeptide and component 3, the 18.5-kDa form found in other mammalian myelin basic protein preparations. The two proteins in fraction 3 were separated by fast protein liquid chromatography. Both have blocked N termini and identical C termini (-Met-Ala-Arg-Arg). When the tryptic digests of the two proteins were fractionated by HPLC, the elution profiles were similar, except that four peaks found in the chromatogram of the larger protein were missing from the chromatogram of the smaller one. In addition, an extra peak was found in the elution pattern of the latter chromatogram. Amino acid analysis of the individual tryptic peptides indicated that the smaller protein lacked residues 106-116 (-Gly-Arg-Gly-Leu-Ser-Leu-Ser-Arg-Phe-Ser-Trp-). The deleted portion corresponds exactly to the amino acid sequence encoded by exon 5 of the mouse basic protein gene. This new form of myelin basic protein has a molecular weight of 17,200, calculated from its amino acid composition.     

Membrane interactions in nerve myelin: II. Determination of surface charge from biochemical data.

In our accompanying paper (Inouye and Kirschner, 1988) we calculated the surface charge density at the extracellular surfaces in peripheral and central nervous system (PNS; CNS) myelins from observations on the dependency of the width of the extracellular space on pH and ionic strength. Here, we have determined the surface charge density of the membrane surfaces in myelin from its chemical composition and the localization of some of its molecular components. We then analyzed the attractive and repulsive forces between the apposed surfaces and calculated equilibrium periods for comparison with the measured values. The biochemical model accounts for the observed isoelectric range of the myelin period and, with the surface charge reduced (possibly by divalent cation binding or a space charge approximation), the model also accounts for the dependency of period on pH above the isoelectric range. At the extracellular (and cytoplasmic) surfaces the contribution of lipid (with pI approximately 2) to the net surface charge is about the same in both PNS and CNS myelin, whereas the contribution of protein depends on which ones are exposed at the two surfaces. The protein conformation and localization modulate the surface charge of the lipid, resulting in positively-charged cytoplasmic surfaces (pI approximately 9) and negatively-charged extracellular surfaces (pI approximately 2-4). The net negative charge at the extracellular surface is due in CNS myelin to lipid, and in PNS myelin to both lipid and (PO) glycoprotein. The net positive charge at the cytoplasmic surface is due in CNS myelin mostly to basic protein, and in PNS myelin to PO glycoprotein and basic protein. The invariance of the cytoplasmic packing may be due to specific short-range interactions. Our models demonstrate how the particular myelin proteins and their localization and conformation can account for the differences in inter-membrane interactions in CNS and PNS myelins.     

Secondary structure of charge isomers of myelin basic protein before and after phosphorylation

Human myelin basic protein (MBP) was fractionated into several of its charge isomers (components). Of these, the secondary structures of four isomers before and after phosphorylation have been studied by circular dichroism (CD). None of the four showed any alpha-helical structure. All of the components showed varying amounts of beta-structure, random structure, and turns. Component 1 (C-1), the most cationic of the components, showed 13%; component 2 (C-2) had 19%; C-3, 17%; and C-4, 24% of beta-structure. Each of the four components was phosphorylated with protein kinase C, from human brain. The extent of phosphorylation varied considerably from 2.8 +/- 0.6 mol of PO4/mol of protein in C-1 to 5.2 +/- 0.8 mol of PO4/mol of protein in C-4. The effect of phosphorylation on the secondary structure was to induce beta-structure in all the components. The largest change in beta-structure was in C-1 and the least in C-4. The surprising result is that although the components were phosphorylated to different extents, the amount of beta-structure in all four components increased to a final proportion of 35-40%. Treatment of phosphorylated C-1 with acid phosphatase removed 50% of the total radioactivity. Although the remainder represented approximately 1 mol of PO4/mol of protein, the proportion of beta-structure was unaltered. We concluded that a single phosphorylation site identified as residues 5-13 represented a critical size for stabilization of beta-structure of MBP in solution and that phosphorylation at the other sites had little influence on secondary structure.     

MICROHETEROGENEITY AND PHOSPHOAMINO ACIDS IN THE CARBOXY-TERMINAL HALF OF MYELIN BASIC PROTEIN

—Three chromatographically distinct peptic peptides (F80-1, F80-2 and F80-3) derived from the C-terminal half of the bovine and guinea-pig myelin basic proteins were characterized. The three peptides of each animal species had the same N-terminal residue (phenylalanine) and essentially the same amino acid composition, but they differed in electrophoretic mobility at alkaline pH. The least basic peptide (F80-3) differed from the others in showing a deficit of C-terminal arginine residues and in containing phosphorus, 0·37 and 0·46 g-atom/mol of bovine and guinea-pig peptides, respectively. Other peptic peptides. derived from the N-terminal half of the basic protein, were essentially phosphorus-frcc. Analyses of partial acid hydrolyzates of peptide F80-3 by high voltage electrophoresis showed the presence of both phosphoserine and phosphothreonine. After incubation with E. coli alkaline phosphatase (EC 3.1.3.1). 34 and 40% of the bovine and guinea-pig F80-3 peptides. respectively, were converted to peptide F80-1. This reaction involved the loss of 2 net negative charges, and its extent corresponded to loss of essentially all of the phosphate originally present in the peptide. This result indicated that the phosphorylated species of peptide F80-3 contained one phosphate group per molecule.     

Charge effects modulate actin assembly by classic myelin basic protein isoforms

Myelin basic protein (MBP), a highly cationic structural protein of the myelin sheath, is believed to be associated with the cytoskeleton in vivo and interacts with actin in vitro, but little is known about the regulation of this interaction. The rate and extent of actin polymerization induced by 18.5 kDa MBP charge isomers were correlated to charge reduction by post-translational modifications. Increased ionic strength attenuated the initial rate but not the final extent of polymerization achieved. Reduced pH enhanced the rate and extent of polymerization, presumably via partial protonation of intrinsic histidyl residues. The polymerizing activities of the 21.5, 17, and 14 kDa MBP splice variants were not proportionate to their net charges or charge densities. The presence of at least one region derived from exon II or VI of the "classic" MBP gene was required for effective bundling as assessed by light scattering and transmission electron microscopy.     

Role of phosphorylation in conformational adaptability of bovine myelin basic protein.

Controlled thrombic digestion of a preparation of components 2 + 3 isolated from the 18.5 kDa bovine myelin basic protein (MBP) yielded a polypeptide that was monophosphorylated on threonine 97 (component 3pT97). This is the first posttranslationally phosphorylated MBP isolated in pure form. We studied the effect of this single phosphate on the conformational adaptability of 18.5 kDa bovine MBP by comparing the circular dichroism (CD) spectrum of component 3pT97 with the spectra of highly purified nonphosphorylated components 1 and 2. The CD spectra of nonphosphorylated component 1 and component 2 [monodeamidated form(s) of component 1] were indistinguishable, while component 3pt97 exhibited a different spectrum. The singly phosphorylated MBP component exhibited 13% more ordered conformations than that adopted by nonphosphorylated MBP in dilute aqueous solutions. This was estimated from the CD spectra, and apparently involved about 17 additional amino acid residues in beta-structure(s).     

Amino acid sequence of the smaller basic protein from rat brain myelin

1. The complete amino acid sequence of the smaller basic protein from rat brain myelin was determined. This protein differs from myelin basic proteins of other species in having a deletion of a polypeptide of 40 amino acid residues from the centre of the molecule. 2. A detailed comparison is made of the constant and variable regions in a group of myelin basic proteins from six species. 3. An arginine residue in the rat protein was found to be partially methylated. The ratio of methylated to unmethylated arginine at this position differed from that found for the human basic protein. 4. Three tryptic peptides were isolated in more than one form. The differences between the two forms of each peptide are discussed in relation to the electrophoretic heterogeneity of myelin basic proteins, which is known to occur at alkaline pH values. 5. Detailed evidence for the amino acid sequence of the protein has been deposited as Supplementary Publication SUP 50029 at the British Library (Lending Division) (formerly the National Lending Library for Science and Technology), Boston Spa, Yorks. LS23 7BQ, U.K., from whom copies may be obtained on the terms given in Biochem. J. (1973) 131, 5.     

Deimination of myelin basic protein. 2. Effect of methylation of MBP on its deimination by peptidylarginine deiminase

Deimination of myelin basic protein (MBP) has been implicated in the chemical pathogenesis of multiple sclerosis (MS). Degradation of bovine MBP by cathepsin D, a myelin-associated protease, was increased when 6 arginyl residues were deiminated and became very rapid when all 18 arginyl residues were deiminated. Since MBP contains a number of modifications, including methylation, phosphorylation, etc., we studied the effect of methylation, an irreversible modification, to determine how this modification affected deimination. Methylation of Arg 106 in bovine MBP (Arg 107 in human), a naturally occurring modification of MBP, has been shown to affect the deimination of arginyl residues in the present study. Since fractionation of MBP into unmethylated, monomethylated, and dimethylated species cannot be done readily on a preparative scale, mass spectrometry with the Q-TOF instrument resolved these species readily since each differed from the other by 14 atomic mass units (amu). Examination of five different hMBP samples, two from normal brain and three from MS brain, revealed that increased deimination of arginyl residues correlated with a decreased methylation of Arg 107 (human sequence). To study this process in vitro, bovine MBP (bMBP) was used. Component 1 (C-1) is the most cationic of the MBP "charge isomers" and the most unmodified, in which all arginyl residues are intact. It was deiminated to various extents with purified bovine brain peptidylarginine deiminase, generating a number of species containing 0-13.7 mol of citrulline/mol of bMBP. Mass spectrometry of each of these species permitted us to determine the influence of methylation of Arg 106 (bovine sequence) on deimination by this enzyme. We found that bMBP with unmethylated arginine was deiminated at a rate of 0.081 mol of citrulline/min, with monomethylarginine, 0.068 mol of citrulline/min, and with dimethylarginine, 0.036 mol of citrulline/min. We suggest that the methylated arginyl residue becomes sequestered in the hydrophobic beta-sheet structure and disrupts the three-dimensional structure of the protein so that other arginyl residues are less accessible to peptidylarginine deiminase.     

Characterization of Basic Proteins from Goldfish Myelin

Abstract: Myelin basic protein (MBP) from common goldfish ( Carassius auratus ) myelin was extracted with dilute mineral acid. Immunological cross-reactivity of the goldfish MBP, with polyclonal antisera raised against bovine MBP, suggested that the goldfish protein has epitopes for these antibodies. It also reacted with a monoclonal antibody specific for a seven amino acid epitope (130–137) conserved in the MBP of most mammalian species. To characterize the charge heterogeneity of this protein, we iodinated the protein with ¹²⁵I and chromatographed it on a carboxymethyl cellulose-52 column together with a nonlabeled acid soluble fraction prepared from human white matter as a carrier protein. All of the goldfish protein was recovered in the unbound fraction, demonstrating that it was less cationic than the carrier protein (human MBP). We have also examined the urea alkaline gel profile of the goldfish MBP together with the human C-1, C-2, C-3, C-4, and C-8 components. The results from these experiments indicated that this MBP extracted from goldfish brain myelin lacked the microhet-erogeneity that is associated with MBPs from higher vertebrates. The MBPs from goldfish myelin were separated into their isoforms by reversed-phase HPLC. Amino acid compositions were determined for both the 17- and 14-kDa goldfish proteins. Amino acid analysis revealed similarities with the compositions of other MBPs; however, the serine content in both the 17- and 14-kDa proteins was higher than that of the human C-1, the mouse C-1 protein, and the shark proteins. The HPLC-purified 14-kDa goldfish protein was chemically cleaved with CNBr for partial sequence analysis. Even from the limited sequence obtained, the sequence ATAST was found in goldfish, which is also present in human, rabbit, and guinea pig MBPs.