key: cord-008293-5cwb5g3h authors: Shaw, Shyh-Yu; Laursen, Richard A.; Lees, Marjorie B. title: Analogous amino acid sequences in myelin proteolipid and viral proteins date: 1986-10-27 journal: FEBS Lett DOI: 10.1016/0014-5793(86)81502-2 sha: doc_id: 8293 cord_uid: 5cwb5g3h Computer analysis of the intrinsic membrane protein, myelin proteolipid, shows strong sequence similarities between the putative extramembrane segments of the proteolipid protein and a number of viral proteins, several of which infect humans. These similarities are even more striking than those reported previously between viral proteins and the encephalitogenic myelin basic protein (MBP). These findings, along with other reports of molecular mimicry by viruses, suggest that immunological cross‐reactions between virus‐induced antibodies or T‐cells and analogous antigenic determinants (epitopes) in myelin proteolipid could be involved in the pathophysiology of multiple sclerosis or post‐infectious demyelinating syndromes. Current evidence suggests that the etiology of multiple sclerosis and other demyelinating diseases involves a combination of viral and autoimmune factors [1] . A particularly well studied model for multiple sclerosis is EAE which can be induced by immunization of laboratory animals with CNS tissue or with MBP, one of the major myelin proteins [2] . Recently it has been reported that MBP shows homology with certain viral proteins [3] . Furthermore, immunization with peptides having regions common to MBP and hepatitis B virus DNA polymerase showed histological EAE in rabbits [4] . However, it has long been recognized that components other than MBP contribute to the encephalitogenic response to CNS tissue [5] . * To whom correspondence should be addressed Abbreviations: MBP, myelin basic protein; EAE, experimental allergic encephalomyelitis; CNS, central nervous system; MLV, murine leukemia virus; AIDS, acquired immune deficiency syndrome; HTLV-III/LAV, human T-lymphotropic retrovirus The most abundant protein of CNS myelin is myelin proteolipid protein, which is embedded in the myelin membrane. Its amino acid sequence has been reported recently [6, 7] , and two different models have been proposed for its conformation in the myelin membrane [8, 9] . Immunization of rabbits with the proteolipid apoprotein leads to the development of a chronic, progressive or relapsing form of EAE [10] . However, the epitopes in myelin proteolipid which canse the encephalitogenic response are not known. In the present study we compared decapeptide sequences from proposed [8, 9] extramembrane hydrophilic regions of myelin proteolipid with various viral proteins by computer analysis, with the aim of searching for similarities, such as found between MBP and viral proteins, that might shed light on the origin of demyelinating diseases. Decapeptides were chosen for comparison with the extramembrane segments of the proteolipid protein, the rationale being that decapeptides are more than sufficient to induce an immune response [11] and that epitopes of membrane proteins are generally located at extramembrane hydrophilic regions. Comparisons and statistical scorings were made using the database (release 6.8 of 28 August 1985, containing 3309 sequences) and programs of the Protein Identification Resource of the National Biomedical Research Foundation. The search program compared every decapeptide in bovine brain myelin proteolipid with every decapeptide in the database. The calculation of similarity scores was accomplished using the Mutation Data Matrix [12] program. For each proteolipid decapeptide, 709253 comparisons were made and the highest scoring segments and their standard deviations were printed out. We selected those viral decapeptides, similar to extramembrane hydrophilic segments of myelin proteolipid, with scores higher than 30, and which also exceed the average by 5 standard deviations; i.e., the probability that the match is due to random chance is <3 x 10 -7 . A VAX 11/750 computer was used for searching and calculations. Sequence alignments with scores of 30 or greater were observed between proteolipid and 249 viral decapeptides, excluding identities seen with proteins from similar strains of virus. These include decapeptides from a variety of types of viruses, such as adenoviruses and Epstein-Barr, influenza and measles viruses. Only those with a score of 38 or greater, or which show at least 60% identity are shown in table 1. The degree of similarity between proteolipid and viral proteins is significantly greater than that observed [3] between MBP and viral proteins. Furthermore, although only 10% of the proteins in the protein sequence database are viral proteins, they account for over 30% of the similarities. It is interesting and perhaps significant that many of the strong similarities lie between residues 135 and 150 in the proteolipid. This region is predicted (Shaw, S.-Y., unpublished) to have a strong tendency to form an amphipathic helix (hydrophobic on one side and polar on the other), and thus might be regarded as a prime candidate for an antigenic site. Likewise, many similarities are found with the N-terminus of the proteolipid, a region which has also been predicted to be an amphipathic helix [8l. The two proposed models [8, 9] for the organization of the proteolipid in the myelin membrane are Table 1 Comparison of myelin proteolipid decapeptide sequences with those of certain viral proteins All of the peptides shown have alignment scores of 38 or greater and/or have at least 6 identities. Amino acids which are identical with the corresponding residues in the proteolipid are underlined similar in that they depict the protein as threading through the lipid bilayer several times with the polar segraents external to the bilayer; they differ, however, in the sidedness of some of the extramembrane segments. While both models put residues 1-10 and 220-235 on the extracellular face of the membrane, the model of Laursen et al. [8] puts residues 35-60 on the extracellular face and 90-150 on the cytoplasmic face, whereas the model of Stoffel et al. [9] predicts the opposite orientation for the latter two segments. Regardless of which (if either) model is correct, similarities between viral proteins and proposed extracellular and cytoplasmic segments of the proteolipid are seen. It has been proposed [3, 4] that EAE-like diseases can arise when virus-evoked antibodies and/or sensitized T-cells cross-react with homologous amino acid sequences (epitopes) in MBP. However, since MBP is located entirely on the cytoplasmic face of the membrane and is therefore within the oligodendroglial cell, it would be expected to be relatively inaccessible to immune surveillance. On the other hand, the proteolipid has potential antigenic sites not only on the cytoplasmic face but also on the outer surface of the membrane. Recognition of these sites by antivira[ antibodies or sensitized cells could induce a cascade of immunological events leading to cell destruction. In the process, MBP would be released, which might result in further inflammatory consequences. The similarities between MBP and virus proteins have been discussed [3, 4] previously in terms of 'homology', which implies a common ancestor for the proteins. When comparing myelin proteins with viral proteins, it is difficult to make a compelling case for homology, since the similarities fall off after a few residues. While it is not uncommon to encounter five or six identical residues in a stretch of ten (see table 1), extended regions of similar sequence are encountered only rarely. For example, although a 55% identity is seen between the first 20 residues of the proteolipid and the late 100 kDa protein of adenovirus 2 or 5, i.e., GIA.E('CA R('I V(;AP.I ASI.VA illll II ! ii~ Adenovirus (561-580) GI.I.I;('HCR('NI_(.'TPI, tRSI.VC, more likely that the virus and myelin proteins are 'analogous', i.e., that similar portions arose by convergent evolution to a common sequence or secondary structure. This idea is consistent with the well-known propensity of viruses to undergo rapid mutation, particularly of their surface antigenic sites [13] . The tryptophan containing region of MBP has long been recognized as being highly encephalitogenic [2, 5] . In particular, the nonapeptide, Phe-Ser-Trp-Gly-Ala-Glu-Gly-GIn-Lys, has been shown to be highly effective in inducing EAE in guinea pigs, and studies with peptide analogs indicate that the essential features of an encephalitogenic determinant are X-X-Trp-X-X-X-X-Gln/Asn-Lys/Arg. The proteolipid contains two regions that have a Trp and a Lys residue separated by 5 residues, but otherwise they are not remarkably similar to the encephalitogenic MBP peptide. It is interesting, however, that one of the tryptophans (Trp-144) occurs in a region that shows several similarities with viral proteins (table 1). In the course of our analysis, we also found a strong similarity, not previously reported [3] , between the encephalitogenic MBP protein and the gag-pol polyprotein of murine leukemia virus It would not be surprising if antibodies against the MLV peptide were found to cross-react with MBP. There are now at least two reports of mimicry of normal human proteins by viruses: MBP by hepatitis B virus polymerase [4] and c~-thymosin by the AIDS virus HTLV-III/LAV [14] . In addition Pruijn et al. [15] have reported that sera from patients suffering from autoimmune diseases can inhibit adenovirus DNA replication, suggesting a similar phenomenon. The similarities between viral proteins and myelin proteolipid may represent yet another case. In this instance, it seems likely that portions of the protein which are potential antigenic determinants are located on the outer surface of the cell membrane, making it more understandable how the cell membrane could come under immunological attack. the similarity does not extend further. It seems Viral Infections of the Nervous System Experimental Allergic Encephalomyelitis: A Useful Model for Multiple Sclerosis Hoppc Seyler's Proc. Natl. Acad. Sci. USA Proc. Natl. Acad. Sci. USA This work was supported by grants from the National Institutes of Health (NS-13649 and NS 16945) and National Science Foundation (DMB 85-03940).