Changes in protein structure of myelin sheaths throughout vertebrate evolution

The vertebrate brain is composed of many tiny cells that communicate with each other in order to aggregate sensory information, oversee homeostatic processes, and make decisions. These cells are known as neurons, and while this organization is often taken for granted, it relies on an intricate infrastructure in order to function properly. Additional cells, known as glial cells, aid this neuronal communication. Glial cells are present in all vertebrates, and while their roles are probably not completely understood, they are believed to be involved with protection and nutrition of neurons. One of the uses of glial cells is to produce myelin sheaths, but there are other ways that glial cells are used in the nervous systems of vertebrates.

Evolutionarily, the first appearance of a glial cell used as a myelin sheath appears to have been an ancestor of modern-day gnathostomes, probably a placoderm (Hartline and Colman, 2007). Neurons communicate with one another through a process called saltatory propogation. Electrical current is sent “down” the axon to the synapse, where a neurotransmitter is released, which can interact with the recipient neuron in myriad different ways. Myelin expedites this process by ensheathing much of the axon and clustering ion channels at specific places called nodes of Ranvier, where chemical exchanges take place that modulate the electrical current inside the axon in just a few milliseconds. The electrical current will then “hop” down the axon (due to diffusion and electromagnetism) to the next node of Ranvier, where the same chemical exchange will take place once again. Clustering the ion channels at specific locations along the axon reduces the time necessary for the current to propogate down the cell and saves energy, since the ATP-intensive ion channels only have to be present at the nodes of Ranvier of the axon, in lieu of working throughout the entire length of the axon. Oligodendrocytes are the glial cells that ensheath the myelin in the CNS, while Schwann cells are the glial cells that do the ensheathing in the PNS. Myelin sheaths do not cover every neuron in the vertebrate brain, and neurons with longer axons are generally the ones covered with myelin. The portion of the brain dominated by axons covered by myelin is known as white matter, as opposed to grey matter, which is composed of neuron cell bodies.

Myelin provides a tremendous adaptive advantage for the vertebrates that possess it by speeding conduction speed up to a 100-fold increase (Hartline and Colman, 2007). The alternative way to increase conduction speed would be to increase the diameter of the axon itself, the most famous example of which occurs in the giant squid. However, this structure would require substantially more energy consumption, and given that in humans the CNS already takes up 20% of our energy use, we don’t have much to spare (Hartline and Colman, 2007). A complete understanding of the structure of these myelin sheaths would be a great advancement for science and the study of the brain. Demyelination and other deficiencies of myelin function is a common cause of human disease, particularly Multiple Sclerosis and the group of disorders known as leukodystrophy. Additionally, some vertebrates such as goldfish and zebrafish have myelin sheaths that can regenerate (Oertle et al., 2003), which intrigues researchers for reasons that will be intuitive to any human that has ever considered the possibility of brain degeneration.

This paper will focus on the changes to the protein structure and function of myelin sheathing throughout vertebrate evolution. Although lipids make up the majority of the dry weight of myelin, this fatty layer is essentially homogenous throughout vertebrate taxa. The remaining 20% of myelin, made up of various proteins, is where the intrataxa variance lies (Schweigreiter et al., 2006). This paper will define a character as one of these proteins that make up myelin, such as proteolipid protein (PLP), among others. A character state will be defined as one homologous manifestation of these proteins in a given taxon. For example, M6B is a PLP-related protein present in tetrapod myelin, and DMg is the homologous protein present in fish and amphibians.  Researchers generally use immunohistology, nucleotide and amino acid sequencing, and “knock-out” species with one or more genes removed in order to determine the whether these proteins are homologous from species to species, and whether they are identical or have differences that lead to character state polarity. Character state polarity is particularly difficult to decipher in the study of myelin since there is no fossil record, but whenever possible it will be included.

The main proteins that this paper will focus on are myelin basic protein (MBP), proteolipid proteins (PLP), protein zero (P0), myelin-associated glycoprotein (MAG), 2’, 3’-cyclic nucleotide 3’ phosphodiesterase (CNP), and myelin/oligodendrocyte glycoprotein (MOG). Other myelin proteins that have received some study are peripheral myelin protein 22 (PMP-22), the 3K protein of zebrafish (36-kDa), and oligodendrocyte myelin glycoprotein (OMgp). PMP-22 is abundant in mammalian PNS myelin, whose mutations are a cause of many human myelin pathologies (Gould et al., 2005). 36-kDa is a recently characterized protein of myelin in teleosts which has recently been found to have a homologous protein in humans (Morris et al., 2004). OMgp is expressed in the CNS and is believed to have a role in myelin formation and maintenance, based on recent in vitro studies (Vourch and Andres, 2004). This paper will not focus extensively on these last three proteins because there has not been much study on the changes in their character states throughout vertebrates. First, I will review the evidence for homology from each character, then I will explain what researchers believe the evidence for polarity of each character state is, and finally I will present some of the hypotheses for the functions for each of these proteins and how or why these characters may have changed over the course of evolution.

Myelin Basic Protein

There is substantial evidence that MBP is one of the major proteins necessary for the myelination process. Immunological studies have been carried out which show that MBP is present in all gnathostomes, including cartilaginous fish, whose ancestor is believed to be the first animal to express myelination (Schliess and Stoffel, 1991). It is even easier to see its importance to vertebrate myelination by looking at where it is not present: tunicates. Gould et al. (2005) analyzed protein sequences from sea squirts (Ciona intestinalis) by comparing its genome database with the largest MBP protein sequences available from 8 different vertebrate species. They found no homologous proteins between the sea squirt and any of these vertebrates. Tunicates are close to vertebrates evolutionarily, which signifies that MBP development in the gnathostome ancestor was a key component to the development of myelin sheathing, although the other possibility is that the gene that codes for the protein was lost during ascidian evolution. Within vertebrates, MBP is highly conserved, without much change between taxa. The one exception are isoforms of MBP that contain exon 2, which only appear in mammals (Schweigreiter, 2006). Therefore, the only two distinct character states for MBP in vertebrates are those that do not contain exon 2 (non-mammals) and those that do contain exon 2 (mammals).

Allinquant et al. (2001) attempted to determine the function of the exon 2-containing MBP in mammals. In order to do so, they took MBP deficient mice (known as “shiverers”), and transfected into them either exon 2-containing isoforms 21.5- and 17-kD MBP, or non exon 2-containing isoforms 18.5- and 14-kD MBP. The mice were immunostained 48 hours later and examined under electron microscopy in order to determine where the MBP isoforms had traveled. What they found was that the mice with the exon 2-containing isoforms were expressed primarily in the cytoplasm and the nucleus of the transfectant cells, which is what you’d expect to see in the oligodendrocytes of 9-day old normal mice, whose myelin is just developing. The mice whose MBP lacked the exon 2-containing isoform were expressed primarily in the plasma membrane, and were similar to what you’d expect to see in normal adult oligodendrocytes (see figure 1). Their results suggest that the exon 2-containing isoform of MBP is involved first and primarily in maturating and differentiating oligodendrocytes. The study does not suggest any evolutionary scenarios for why this deletion may have occurred, and possibly further research transfecting the exon 2-containing isoform of MBP into non-mammalian vertebrates could yield some results about what advantage this additional exon provided for mammals when it was derived. This study also provided further evidence for the belief that “classic” MBP—that is, the MBP containing the 6 other exons—is necessary for holding surfaces of adjacent myelin lamellae together (Schweigreiter et al., 2006).

Protein Zero

Protein Zero, a glycoprotein, is the major protein of fish CNS myelin, and of all vertebrate PNS myelin (Schliess and Stoffel, 1991). Jeserich et. al. (1987) used bovine P0 antibodies and tested their reactivity in trout. They found two glycoproteins related to P0, termed IP1 and IP2, and they both reacted strongly to the bovine P0 antibody when examined by electro-immunoblotting. This design determined that P0 shows homology throughout vertebrates, and it also helped differentiate its distinct character states. Cartilaginous and bony fishes have two or more isoforms of P0 (like IP1 and IP2), while mammals only express one isoform of P0 (studying myelin evolution). In trout, the larger isoform, IP2, is present during the early stages of oligodendrocyte differentiation, while the smaller isoform, IP1, is present only in mature stages of myelin development (Lanwert and Jeserich, 2001). This is eerily similar to the function of myelin basic protein isoforms in mammals (see section on MBP), although the connection has apparently not been made in a peer-reviewed journal article.

The other change in the expression of P0 and P0-like proteins across vertebrate evolution is that it is expressed less in the CNS of amphibians as compared to fish, and it is not at all expressed in the CNS of mammals, although it is expressed in the PNS in all vertebrate classes (Yoshida and Colman, 1996). Its absence in the CNS of mammals raises the question of whether or not it is essential at all. Lanwert and Jeserich (2001) set out to discover the function of the IP1 isoform in trout and this same isoform when combined with the P0 tail of rat to make a “chimeric” protein. They hypothesized based on past research that the proteins would have adhesive properties, and designed an experiment where they would transfect the proteins into Chinese Hamster Ovary (CHO) cells in vitro and test to see whether the proteins would cause more of the cells to clump together after a given amount of time. The researchers also included a control group of CHO cells that did not include either protein . The researchers were attempting to find the “adhesiveness” of each of the proteins, and operationalized adhesiveness as a function of the number of particles before and after transfection and a 90 minute waiting period. The IP1 protein isoform on its own was only slightly more adhesive than the control, while the chimeric protein was substantially more adhesive (see Figure 2). These results suggest that P0 and its isoforms in fish have adhesive properties that presumably are involved with bringing the lamellae of the myelin together in its formation. It is unreasonable to conclude that mammalian P0 is “more” adhesive solely on the basis of this study, as the longer isoform IP2 of trout was not included. Regardless, P0 and its isoforms undoubtedly express some adhesion-like properties, which makes sense in its hypothesized role of myelin compaction. If mammals were to not express it entirely, they would have to have something to replace it with evolutionarily (Yoshida and Colman, 1996).

Proteolipid Proteins

Proteolipid proteins (PLPs) are hydrophobic proteins that integral to the protein make-up of myelin throughout teleosts and tetrapods. It was originally thought that PLP was the protein that “replaced” P0 in the CNS of myelin, and while that is true to an extent, a closer analysis reveals that there are PLP homologues in fish. It is more likely that a common ancestor of teleosts and tetrapods developed earlier versions of these homologues, and that they have split off evolutionarily since then (Yoshida and Colman, 1996). There are three states of PLP evolution in myelin: DMa, which is present in bony fish and is homologous to PLP, PLP itself, first present in amphibians, and PLP/DM20, which is the character state for amniotes. Amniotes express the same PLP isoform as amphibians, but they also express an isoform of the protein known as DM20, which contains a deletion of 35-36 amino acid residues (Schweitzer et al., 2006).

The evidence for homology of PLPs between these vertebrate taxa is substantial. The first evidence brought to the table was PLP immunoreactivity in bony fish, lobe-finned fish, and lungfish. Researchers transfected mammalian PLP antibodies into these fish, and when they saw that there was a reaction to the antibodies, they knew that there had to be some sort of homology (Yoshida and Colman, 1996). Additional evidence for homology has come by way of analyzing amino-acid sequences of the proteins in various vertebrate classes and comparing them. Schliess and Stoffel (1991) conducted such a study, comparing the amino acid sequences of humans, chickens, and frogs. Doing so, they were able to locate the exact splice spot, 5’, where they hypothesize that amniote PLP/DM20 split off from amphibian PLP. DM20 originates from “alternative splicing” when a cryptic 5’ donor site within exon 3B is activated and deletes 35-36 amino acid residues downstream (Schliess and Stoffel, 1991). Figure 3 illustrates the difference between DM20 and PLP. The character state polarity here is a function of which isotopes are present in these animals. It is clear that the PLP/DM20 of amniotes is derived from the PLP of amphibians, because DM20 is an additional isoform stemming from a gene deletion, while the PLP isoform is necessarily ancestral. (Scweitzer, et al., 20056). Given the complexity of this relationship, it is certainly possible that an ancestor of both modern amphibians and amniotes developed PLP from DM20, and it is modern amphibians that have lost the DM20 isoform. See Figure 4 for an attempt to reconstruct this complicated evolutionary process, based on levels of gene divergence.

Another myelin-related homologous protein present in many vertebrates is M6B. Although it is a general a proteolipid protein used in various processes, its use in oligodendrocytes and in part of myelin-enriched membranes (Schweitzer et al., 2006) qualifies it for discussion as a myelin protein. The homologous protein in fish and amphibians is called DMg, and sequence analysis has determined that it is homologous to the M6B of amniotes, while distinct enough to qualify as its own character state. The character state polarity is established because fish and amphibian DMg, as a group, differ substantially less from one another than they do from the M6B of amniotes, as a group. DMg is assumed to be ancestral to M6B presumably based on the pattern of gene changes. Once again, see Figure 4 for a reconstruction of the evolutionary pattern based on gene sequencing and the percentage divergence of these genes.

The next step is to attempt to decipher the function of these proteins, and look for reasons that these evolutionary switches may have occured. One interesting study by Yin et al. (2006) genetically engineered mice to express P0 instead of PLP in CNS myelin. As you recall from the P0 section, P0 is expressed less in the CNS of amphibians than fish, less still in the CNS of aves, and not at all in the CNS of mammals (Schweigreiter et al., 2006). The researchers in this study took PLP “knockout mice”, referred to as PLP-null, and engineered into them different groups. One group they left as PLP-null and did not add additional proteins to the CNS, one group they bred to express a combination of PLP and P0 in the CNS, one group they bred to express only P0 in the CNS, and one group was the wildtype, meaning that they expressed normal amounts of PLP in the CNS. Their results were striking: the P0-CNS mice died at a much faster rate than each of the other groups, and had less motor control while they were alive. Their hypothesis was their defficiences stemmed from axonal pathology in P0-CNS mice, which they measured based on the immunohistory reactivity to amyloid precursor protein (APP). As you can see in Figure 5, the P0-CNS mice along with the PLP-null mice show the highest levels of APP reactivity. Their results suggest that in the CNS of amphibians, amniotes, and mammals, PLP plays a role in protecting the axons of myelinated neurons.

The other character state polarity of functional interest to researchers is DM20, which is an isoform of PLP that has a homologous protein in fish, is not present in amphibians, and is present as an additional isoform of PLP in modern amniotes. Stecca et al. (2000) created PLP-knocked out mice by eliminating the gene that codes for the protein in the CNS. They then derived a second group of mice that replaced PLP with DM20, at similar levels to how much PLP would be expressed in wildtype mice. Finally, they had a control group of wildtype mice that expressed normal amounts of PLP in the CNS. One of the tests that they conducted on these mice was the rotarod motor performance test, which is used to measure the motor abilities of mice. Although the motor abilities of each group were initially similar, as the mice aged the wildtype became more proficient than the other groups. The wildtype was significantly more proficient than the PLP-null mice, which in turn was significantly more proficient than the DM20 mice (Stecca et al., 2000). Their results suggest that there are functional differences between DM20 and PLP, and that DM20 cannot functionally replace PLP. Previous X-ray diffraction data shows that intermembrane space is 0.5 nanometers wider in amphibians than mammalian myelin, which indicates that the DM20 isoform might lead to a more compact myelin structure (Schliess and Stoffel, 1991). From this data some have suggested that while the DM20 isoform does not confer any hugely adaptive function to amniotes, it may have been differentially expressed through evolution as an attempt to save space in the skull, as complexity increased faster than the skull could expand (Schweitzer et al., 2006).


Myelin-associated glycoprotein (MAG), also known as siglec-4 (for sialic acid binding protein) is one of 11 siglecs in the human genome. It has been known to be associated with myelination in humans for quite some time (Lehmann et al., 2004). Initially, immunology evidence suggested that there were a number of homologous proteins to MAG in nonmamallian vertebrates, but that evidence was later disputed (Schweigreiter, 2006). In 2004, Lehmann et al. resolved the debate by screening the genomes of pufferfish and zebrafish and looking for homologous siglecs in those species. They found a high degree of similarity between the amino acid sequences of siglec-4 in fish and mammals. Interestingly, this was the only siglec of the human genome that was highly conserved between these fish and humans.

In order to determine that these amino acid sequences were indeed indicative of a human siglec-4 ortholog in fish, Lehmann et al. performed binding studies. They transfected pufferfish cDNA into human red blood cells (erthrocytes), which in humans will lead to cell binding when sialidase enzyme is not present. These transfected erthrocytes were placed in the presence of COS-7 cells (often used in molecular biology to test recombinant proteins) in vitro. As you can see in Figure 6, the binding occurred exactly as would have been expected in mammalian MAG. The pretreatment of erthrocytes with sialidase led to inhibition of binding. From this study, it is clear that there are nonmammalian homologues of MAG, although they are still technically polar character states since their amino acid and genome sequences show substantial divergence.

One of the major reasons that the finding of this homology was so important is that MAG is thought to be one of the major inhibitors of myelin regeneration in mammals (Spencer et al., 2003). Based on the developmental stage of a neuron (as in, “developing” or “already developed”), MAG will either promote growth of the neuron or inhibit it (Schweigreiter et al., 2006). However, if there is a nonmammalian ortholog of MAG, and moreover if the two act practically the same (Lehmann et al., 2004), then the presence of MAG in mammalian myelin cannot be the sole reason for the lack of regenerative abilities of adult mammals. Granted, there is still the question of whether or not MAG binds to fish Nogo receptors, where it is believed that neurite growth is inhibited in humans (Schweigreiter et al., 2006). It is assumed that mammalian siglec-4/MAG is ancestral to the siglec-4 of teleosts and ray-finned fish because after the switch to tetrapods animals lost the ability to regenerate, which is one of the functions believed to be somehow mediated by MAG. The function and evolutionary patterns of MAG has yet to be fully elucidated, and when it is it may lead to some novel implications.


Another protein of myelin, 2’, 3’-cyclic nucleotide 3’phosphodiesterase (CNP), whose last three letters gives it away as an enzyme, has recently been the subject of research on myelin regeneration. In 1995, Ballastero et al. identified a protein in goldfish that they believed was involved with optic nerve regeneration. They analyzed the peptide sequences of the related protein doublet, p68/70, and predicted the amino acid structure of this new protein. They found that it showed high amounts of amino acid similarity with mammalian CNPase, so they termed this new protein gRICH for goldfish regeneration-induced CNPase homolog. These amino acid sequences show enough similarity, 53.4% (Ballastero et al., 1995), to be homologous, but not enough to be considered the same character state. CNPase is assumed to be derived from the ancestral gRICH because regenerative abilities in the CNS were lost after the evolutionary switch from fish to tetrapods.

In order to determine the function of CNP in mammals, Lee et al. (2003) used CNP-knockout rats, isolated their oligodendrocytes three months after birth, and stimulated their growth. They found that those rats whose oligodendrocytes were CNP-null showed less process outgrowth and that the processes that did grow were on average shorter. From these results it is evident that CNP is used in mammals to stimulate growth of oligodendrocyte processes. The question is whether this growth is turned off depending on the age of the mammal. If it is, and this inhibition does not occur in the homologous gRICH of teleosts, then that would implicate these two proteins highly in the phylogenetic loss of regeneration in tetrapods. Regardless, gRICH is expected to play an important role in CNS regeneration in fish (Schweigreiter et al., 2006).


Myelin/oligodendrocyte glycoprotein (MOG) is encoded by a gene in the major histocompatibility complex and is present only in mammals. It is a relatively minor protein, but its selective appearance phylogenetically only in mammals and the fact that its expression is restricted to myelin makes it of particular interest to researchers. Delarasse et al. (2006) performed an analysis of the cDNA of 5 species of mammals: mouse, bovine, human, macaca, and marmoset. They found that primates have a number of isoforms for MOG that contain deletions at certain exons that other mammals do not possess, and that humans specifically express still more unique isoforms of MOG. Importantly, it should be noted that macaques also have two exons containing isoforms that are not present in any other mammals. Figure 8 has a chart summarizing these results. Since each of these species contains the major a1 isoform, it is safe to say that the MOG is homologous for each of these species, and since primates, humans, and marmosets contain high variance in the exons they possess, it is fair to say that these are distinct character states, probably ancestral to the MOG seen in mice which express very few exon-deletion based isoforms. Interestingly, the major a1 isoform is the only one expressed in humans after 40 days, while all of the other isoforms are present after 2 years. The researchers use this fact to draw some inferences of its role evolutionarily. They hypothesize that the MOG isoform variance in primartes and specifically humans may be relevant in inflammatory and demyelinating diseases, since the immune system may or may not ignore these isoforms during development. This could explain the high variance of myelin-related diseases in humans like multiple sclerosis. The evolution of MOG is likely to continue to be studied in order to flesh out this hypothesis.


There are so many individual proteins working together to make up myelin, and our understanding of each of them is at such different places (see the nuanced research into proteolipid versus the relatively less detailed research into CNP/gRICH) that it is difficult to aggregate each of the characters and their states into one large phylogenetic tree. Nevertheless, I have attempted to do so, the results of which you can see in Figure 9.  An additional copy of the tree should be available so that you can read the about the character changes without flipping pages back and forth. The number here corresponds to a number on the phylogenetic tree, and since there are many character changes at each node, they are aggregated here.

12: From agnathans to a common ancestor of gnathostomes, vertebrates developed myelin itself, myelin basic protein, multiple isoforms of protein zero (which is expressed in the CNS and PNS), the PLP-like myelin-related DMg isoform, and the ability to regenerate their myelin sheaths throughout their lifespan, which is present in all fish.

13: Between cartilaginous fish and a common ancestor of bony fish, vertebrates developed the PLP-like isoform DMa, and are likely to have developed gRICH (the CNPase homolog) and siglec-4 (the MAG homolog), although the presence of siglec-4  has not been confirmed in lobe-finned fish and the presence of gRICH has not been confirmed in either lobe-finned or ray-finned fish.

15: Between lobe-finned fish and a common ancestor of ray-finned fish and teleosts, siglec-4 may have developed, depending on whether or not it is present as well in lobe-finned fish (see 13).

4: Between ray finned fish and a common ancestor of teleosts, gRICH may have developed, depending on whether or not it is present as well in lobe-finned and ray-finned fish (see 13).

16: Between fish and a common ancestor of tetrapods, vertebrates began to express just one isoform of protein zero, and began to express it less in the CNS, although it was expressed the same in the PNS. In this same common ancestor, PLP was developed in CNS myelin, and regeneration of myelin became restricted within the lifespan to pre-metamorphosis.

17: Between amphibians and a common ancestor of amniotes, vertebrates began to express the DM20 isoform of PLP in the CNS, they began to express M6B protein in myelin, and they completely lost their abilities to regenerate myelin during adulthood.

18: Between amphibians and a common ancestor of mammals, vertebrates began to express an additional isoform of MBP, completely stopped expressing P0 in the CNS, began to express MOG in the myelin of the CNS, and definitely began to express CNP and MAG, whereas in other taxa it is ambiguous—see the explanation for #13.

19: Between mammals and a common ancestor of primates, vertebrates began to express the second major isoform of MOG, b1 (see Figure 8).

10: Between primates and an ancestor of humans, vertebrates began to express additional isoforms of MOG.

A cursory glance at the date that many of the referenced studies were published is indicative of the way that the field is heading. New techniques for gene sequencing have already led to new discoveries. For example, the sequencing of the zebrafish genome was helpful to study of PLPs (Schweitzer et al., 2006), and to the study of siglec-4/MAG (Lehmann et al., 2004). Although the research that has already been done has laid a crucial groundwork, there is the potential for much more to be learned about the evolution of myelin sheaths in the coming years. Epigenetics, for instance, has not been discussed extensively in the literature of myelin evolution until the past few years, but it has huge implications for the field as molecular genetics continues its paradigm shift.  And while we can squabble over whether or not it is an overall plus for science, the reality remains that the neuroscience studies most likely to be highly funded are those researching topics easily applicable to the treatment of human mental disorders. The study of the evolution of myelin is therefore lucky in the sense that there are many demylinating diseases in humans, including multiple sclerosis and leukodystrophy, whose treatments could benefit from an understanding of how myelin works and how it develops. Moreover, the fact that some vertebrates such as teleosts are able to regenerate their adult CNS myelin means that there is even more reason for neuroscientists to study the evolution of myelin, in order to determine which morphological changes undermined the ability of tetrapods to regenerate adult CNS myelin. There’s good reason to suspect that in 10 years this paper will be almost completely obsolete. If you are not yet excited about the possibilities for research in the evolution myelin, it may be high time to get excited.

Figure 1A:

Figure 1B:

These are confocal microscopy scans of “shiverer” oligodendrocytes in vitro, after the cell has grown. “A” shows a group of slides (with 4 different pixel sizes) of a shiverer mouse transfected with 21.5-kD MBP. “B” shows a group of slides (with 3 pixel sizes) of a shiverer mouse transfected with 18.5-kD MBP. A shows high levels of fluorescence at the ends of the processes, while B shows high levels of fluorescence in the plasma membrane, with no fluorescence in the nucleus (see the star). The slides for A resemble 9-day old mice whose myelin is maturing, while the slides for B resemble normal adult expression of MBP. Citation: Allinquant B, Staugaitis SM, D’Urso D, Colman DR. 1991. The ectopic expression of myelin basic protein isoforms in Shiverer oligodendrocytes: implications for myelinogenesis. Journal of Cell Biology 113 (2). Pages 395 (A) and 397 (B).

Figure 2:

Microscopic images of immunohistory surface staining with IP1 antibodies of CHO cells. The researchers transfected the proteins and waited 90 minutes, these are the images after waiting. Left (A) shows CHO cells with solely the IP1 protein insertion, it had a small difference in particle percentage (36% +/- 9) from baseline (pre-waiting) as compared to the control (not shown, 45% +/- 8). The lower the particle percentage, the greater adhesive properties the protein is assumed to have. Right (B) shows CHO cells with the IP1 protein insertion along with a P0 “tail” from rat myelin, creating a chimeric protein. It was substantially more adhesive than the control, with a particle percentage of 24% +/- 5 after 90 minutes. These results suggest that P0 may be slightly more adhesive than the IP1 isoform. Citation: Lanwert C, Jeserich G. 2001. Structure, heterologous expression, and adhesive properties of the P0-like myelin glycoprotein IP1 of trout CNS. Microscopy Research and Technique 52 (6). Page 640.

Figure 3:

The amino acid sequence of exon 3 (top, white) and 4 (bottom, grey) of hydrophilic proteolipid proteins is shown. On the left is human PLP, in the center is frog (amphibian) PLP, and on the right is DM20, which is only present in amniotes. As you can see, PLP in frogs and humans is highly similar, while DM20 clearly lacks an amino acid sequence in exon 3. This additional isoform in amniotes is thought to be a major change in the myelin protein make-up of vertebrates. (The image may turn out a little bit blurry, which is the result of a photocopied PDF. Please e-mail the author at for a cleaner image if necessary). Citation: Schliess F, Stoffel W. 1991. Evolution of the myelin integral membrane proteins of the central nervous system. Biological Chemistry Hoppe-Seyler 372 (9). Page 870.

Figure 4:

Evolutionary relationships of proteolipid proteins among vertebrates are reconstructed in a phylogenetic tree. On top the researchers illuminate the relationship of PLP/DM20/DMa, next they reconstruct the relationship of M6B/DMg, and finally they reconstruct the relationship of M6A and its ortholog. Note that a discussion of M6A was not included in this paper since the protein is not intimately involved in myelination itself. The researchers compared the gene sequences that code for these proteins in order to construct the chart. Citation: Schweitzer J, Becker T, Schachner M, Nave K-A, Werner H. 2006. Evolution of myelin proteolipid proteins: gene duplication in teleosts and expression pattern divergence. Molecular Cellular Neuroscience 31. Page 167.

Figure 5:

Yin et al. (2006) used knockout mice and bred them to either express only P0 in the CNS (D) or a combination of PLP and P0 (C). They also had groups of mice with no protein added (“PLP-null”, B) and wildtype (A). Slides A-D show immunoreactivity to amyloid precursor protein (APP), which is highest in D, and shows up in the slides as black dots. Slide E is a close-up of one of these black dots, identifying it as indicative of swollen axons. F shows a chart with the rat’s age on the x-axis and the number of APP-positive ovoids on the y-axis; P0-CNS is shaded in black and has clearly more APP reactivity than any other group. The 4 time periods are 1 month, 3 months, 6 months, and 12 months. Nearly all P0-CNS mice were dead by 16 months, a much faster rate than another of the other groups (data not shown here). These results suggest that PLP plays a neuroprotective role in the axons of myelinated neurons. Citation: Yin X, Baek RC, Kirschner DA, Peterson A, Fujii Y, Nave KA, Macklin WB, Trapp BD. 2006. Evolution of a neuroprotective function of central nervous system myelin. Journal of Cell Biology 172(3). Page 473.

Figure 6:

Lehmann et al. performed binding studies of pufferfish siglec-4 in order to determine if it shares the same binding properties as its human ortholog. Here they show COS-7 cells transfected with cDNA encoding for siglec-4. In A, the COS-7 cells were treated with sialidase before exposure to nontreated red blood cells. In B, the cells were not treated with sialidase and exposed to red blood cells that were also untreated. Finally in C, transfected COS-7 cells were untreated with salidilase and exposed to red blood cells that had been treated with sialidase. A and B act as controls for C. In C, it is clear that binding of the cells has been inhibited. This is the same pattern that we would expect to see with human siglec-4/MGA. Citation: Lehmann F, Gäthje H, Kelm S, Dietz F. 2004. Evolution of sialic acid–binding proteins: molecular cloning and expression of fish siglec-4. Glycobiology 14. Page 964.

Figure 7:

Lee et al. (2003) isolated 3 month old oligodendrocytes in CNP-null and wildtype rats and grew them in vitro. “C” is a graphical representation of the number of branches in each type of rat. The y-axis counts the number of branches, and ranges from 0 to 80. Wildtype, on the left, has 70.3 on average per oligodendrocyte, while CNP-null has on average 26.8. In “D”, the researchers selected representative cells for morphological analysis. It is clear that the wildtype oligodendrocyte (left) has many more branches than the CNP-null oligodendrocytes. These results suggest that CNP is related to oligodendrocyte outgrowth in the CNS of mammals. Citation: Lee J, Gravel M, Zhang R, Thibault P, Braun PE. 2005.  Process outgrowth in oligodendrocytes is mediated by CNP, a novel microtubule assembly myelin protein. Journal of Cell Biology 170 (4). Page 669.

Figure 8:

Delarasse et al. (2006) analyzed cDNA sequences isolated from 5 mammalian species. They found a number of MOG isoforms present in humans and primates (represented by macaques) that are not presented in other mammalian species. a1 is the first major isoform, while b1 is the second major isoform, expressed only in primates. Interestingly, a1 is expressed at all time points, while after 40 days none of the others are present in human (and presumably primate) CNS. In a 2-year old brain sample, all isoforms are expressed. This late development of MOG isoforms correlates with the high variability of diseases present in human CNS. Citation: Delarasse C, Gaspera BD, Lu CW, Lachapelle F, Gelot A, Rodriguez D, Dautigny A, Genain C, Pham-Dinh D. (2006) Complex alternative splicing of the myelin oligodendrocyte glycoprotein gene is unique to human and non-human primates. Journal of Neurochemistry 98 (6). Page 1712.

Figure  9:

This is a phylogenetic reconstruction of protein character state evolution in vertebrates using MacClade. The tree was made to emphasize the shortest possible tree given all of the character traits. Explanations of the changes at each of these numbers are presented in the “overview” section of this paper.


Alka A, Vyas A, Schnaar RL. 2001. Brain gangliosides: functional ligands for myelin stability and the control of nerve regeneration. Biochimie 83(7): 677-682.

Allinquant B, Staugaitis SM, D’Urso D, Colman DR. 1991. The ectopic expression of myelin basic protein isoforms in Shiverer oligodendrocytes: implications for myelinogenesis. Journal of Cell Biology 113 (2):393-404.

Ballestero RP, Wilmot GR, Leski ML, Uhler MD, Agranoff BW. 1995. Isolation of cDNA clones encoding RICH: A protein induced during goldfish optic nerve regeneration with homology to mammalian 2′,3′-cyclic-nucleotide 3′-phosphodiesterases.

Proceedings of the National Academy of Sciences of the United States of America 92 (19): 8621-8625.

Bizzozero, OA. Lees, MB. 1999. Fatty acid composition of myelin proteolipid protein during vertebrate evolution. Neurochemical Research 24(2): 269-274.

Delarasse C, Gaspera BD, Lu CW, Lachapelle F, Gelot A, Rodriguez D, Dautigny A, Genain C, Pham-Dinh D. 2006. Complex alternative splicing of the myelin oligodendrocyte glycoprotein gene is unique to human and non-human primates. Journal of Neurochemistry 98 (6): 1707–1717.

Denarier E, Forghani R, Farhadi HF, Dib S, Dionne N, Friedman HC, Lepage P, Hudson TJ, Drouin R, Peterson A. 2005. Functional organization of a Schwann cell enchancer. Journal of Neuroscience 25(48): 11210-11217.

Diekmann H, Klinger M, Oertle T, Heinz D, Pogoda HM, Schwab ME, Stuermer CAO. 2005. Analysis of the reticulon gene family demonstrates the absence of the neurite growth inhibitor Nogo-A in fish. Molecular Biology and Evolution 22(8):1635-1648.

Greenfield EA, Reddy J, Lees A, Dyer CA, Koul O, Nguyen K, Bell S, Kassam N, Hinojoz J, Eaton MJ, Lees MB, Kuchroo VK, Sobel RA. 2006. Monoclonal antibodies to distinct regions of human myelin proteolipid protein simultaneously recognize central nervous system myelin and neurons of many vertebrate species. Journal of Neuroscience Research 83(3): 415-431.

Gould RM, Morrison HG, Gilland E, Campbell RK. 2005. Myelin tetraspan family proteins but no non-tetraspan family proteins are present in the ascidian (Ciona intestinalis) genome. Biological Bulletin 209(1):49-66.

Hartline DK, Colman DR. 2007. Rapid conduction and the evolution of giant axons and myelinated fibers. Current Biology 17(1): 29-35.

Jeserich G, Waehneldt TV. 1987. Antigenic sites common to major fish myelin glycoproteins (IP) and to major tetrapod PNS myelin glycoprotein (P(o)) reside in the amino acid chains. Neurochemical Research 12(9): 825-829.

Kirschner DA, Blaurock AE. 1992. Organization, phylogenetic variations and dynamic transitions of myelin. In: Martenson RE. Myelin: biology and chemistry. Boca Raton: CRC Press. p. 3-78.

Klinger M, Taylor JS, Oertle T, Schwab ME, Stuemer CAO, Diekmann H. 2004. Identification of Nogo-66 receptor (NgR) and homologous genes in fish. Molecular Biology and Evolution 21(1):76-85.

Lanwert C, Jeserich G. 2001. Structure, heterologous expression, and adhesive properties of the P0-like myelin glycoprotein IP1 of trout CNS. Microscopy Research and Technique 52(6): 637-644.

Lee J, Gravel M, Zhang R, Thibault P, Braun PE. 2005.  Process outgrowth in oligodendrocytes is mediated by CNP, a novel microtubule assembly myelin protein. Journal of Cell Biology 170(4): 661-673.

Lehmann F, Gäthje H, Kelm S, Dietz F. 2004. Evolution of sialic acid–binding proteins: molecular cloning and expression of fish siglec-4. Glycobiology 14: 959 – 968.

Lieserson WM, Harkins EW, Keshishian H. 2001. Fray, a Drosophila serine/threonine kinase homologous to mammalian PASK, is required for axonal ensheathment. Neuron 28(3):793-806.

Morris JK, Willard BB, Yin X, Jeserich G, Kinter M, Trapp BD. 2004. The 36K protein of zebrafish CNS myelin is a short-chain dehydrogenase. Glia 45(4): 378-391.

Niemann S, Sereda MW, Suter U, Griffiths IR, Nave K-R. 2000. Uncoupling of Myelin Assembly and Schwann Cell Differentiation by Transgenic Overexpression of Peripheral Myelin Protein 22. Journal of Neuroscience 20(11): 4120-4128.

Oertle T, Klinger M, Stuermer CAO, Schwab ME. 2003. A reticular rhapsody: phylogenic evolution and nomenclature of the RTN/Nogo gene family. FASEB Journal 17(10):1238-1247.

Schweitzer J, Becker T, Schachner M, Nave K-A, Werner H. 2006. Evolution of myelin proteolipid proteins: gene duplication in teleosts and expression pattern divergence. Molecular Cellular Neuroscience 31: 161-177.

Spivack WD, Zhong N, Salerno S, Saavedra RA, Gould RM. 1993. Molecular cloning of the myelin basic proteins in the shark, Squalus acanthias, and the ray, Raja erinacia.  Journal of Neuroscience Research 35(6): 577-584.

Saavedra RA, Fors L, Aebersold RH, Arden B, Horvath S, Sanders J, Hood L. 1989. The myelin proteins of the shark brain are similar to the myelin proteins of the mammalian peripheral nervous system. Journal of Molecular Evolution 29(2): 149-156.

Schliess F, Stoffel W. 1991. Evolution of the myelin integral membrane proteins of the central nervous system. Biological Chemistry Hoppe-Seyler 372(9): 865-874.

Stecca B, Southwood CM, Gragerov A, Kelley KA, Friedrich VL, Gow A. 2000. The evolution of lipophilin genes from invertebrates to tetrapods: DM-20 cannot replace proteolipid protein in CNS myelin.Journal of Neuroscience 20(11): 4002-4010.

Spencer T, Domeniconi M, Cao Z, Filbin MT. 2003. New roles for old proteins in adult CNS axonal regeneration. Current Opinion in Neurobiology 13(1): 133-139.

Schweigreiter R, Roots BI, Bandtlow CE, Gould RM. 2006. Understanding myelination through studying its evolution. International Review of Neurobiology 73: 219-273.

Vourch P, Andres C. 2004. Oligodendrocyte myelin glycoprotein (OMgp): evolution, structure and function. Brain Research Reviews 45(2): 115-124.

Wulf P, Bernhardt RR, Suter U. 1999. Characterization of peripheral myelin protein 22 in zebrafish (zPMP22) suggests an early role in the development of the peripheral nervous system. Journal of Neuroscience Research 57(4): 467-478.

Yin X, Baek RC, Kirschner DA, Peterson A, Fujii Y, Nave KA, Macklin WB, Trapp BD. 2006. Evolution of a neuroprotective function of central nervous system myelin. Journal of Cell Biology 172(3):469-479.

Yoshida M, Colman DR. 1996. Parallel evolution and coexpression of the proteolipid proteins and protein zero in vertebrate myelin. Neuron 16(6): 1115-1126.

2 thoughts on “Changes in protein structure of myelin sheaths throughout vertebrate evolution

  1. Hi, I just started reading your blog – thanks for writing. I wanted to inform you that it’s not showing up properly on the BlackBerry Browser (I have a Bold). Either way, I’m now subscribed to your RSS feed on my PC, so thank you!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s