Cell, Vol. 40, 501-508,
March
1985, Copyright
0092-8674/85/030501-08
0 1985 by MIT
$02.00/O
Isolation and Sequence of a cDNA Encoding
the Major Structural Protein of Peripheral Myelin
Greg Lemke and Richard Axe1
Howard Hughes Medical Institute and
Institute of Cancer Research
College of Physicians and Surgeons
Columbia University
New York, New York 10032
Summary
The myelin sheath is a multilayered
membrane, unique
to the nervous system, which functionsas
an insulator
to increase greatly the velocity of axonal impulse conduction. We have used the techniques
of differential
screening and hybrid selection to identify a cDNA clone
encoding the Schwann cell glycoprotein
P,, the major
structural protein of the peripheral myelin sheath. The
sequence of this protein, deduced from the nucleotide
sequence of the cloned cDNA, indicates that PO is an
integral membrane protein containing
a single membrane-spanning
region, a large hydrophobic
extracellular domain, and a smaller basic intracellular
domain.
The structure of the protein suggests that each of
these domains plays an essential role in generating
the highly ordered structure of the myelin sheath. Furthermore, we find that the induction of P, mRNA coincides with the initiation
of myelin formation,
and we
propose a model in which the glycoprotein
serves as
a molecular guidepost
for this process.
Introduction
Vertebrate peripheral nerves contain two types of axons:
rapidly conducting myelinated axons and slowly conducting unmyelinated axons (Morell, 1984). Myelinated axons
are sheathed in a compact, multilamellar extension of the
plasma membranes of Schwann cells (Geren, 1954). This
structure, the myelin sheath, serves principally as an electrical insulator, greatly increasing the velocity of impulse
propagation (Ranvier, 1878; Ritchie, 1983). Formation of
the myelin sheath results in a striking reorganization
of
Schwann cell morphology, including a several thousand
fold increase in the plasma membrane surface area of the
myelinating cell (Webster, 1971). The characteristic structure of the sheath is formed by repeated spiral wrapping
of newly synthesized Schwann cell membrane about the
nerve axon and subsequent
compaction of adjacent
lamellae at both apposed cytoplasmic and extracellular
surfaces (Robertson, 1962). These morphologic changes
are accompanied by a dramatic induction of a set of proteins unique to myelin.
The glycoprotein P, is the most abundant of these proteins, accounting for over 50% of the protein present in purified myelin sheaths (Greenfield et al., 1973). Although P,
is expressed at these high levels in the peripheral nervous
system, it is absent from the myelin-insulating
axons of the
central nervous system (Greenfield et al., 1973). It is a
Schwann cell specific, integral membrane glycoprotein of
apparent molecular weight 28-30 kd, the expression of
which is restricted to myelinating cells (Lees and Brostoff,
1984; Politis et al., 1982; Poduslo, 1984). It has been suggested that P,, as an amphipathic protein, serves as a
structural element to link adjacent lamellae and thereby
stabilize the myelin assembly (Braun, 1984).
We have isolated and sequenced a cDNA clone encoding the rat P, glycoprotein and have deduced its complete
amino acid sequence. We have used this clone to quantitate the level of P, RNA during development. The structure of the glycoprotein suggests a model in which P,
plays an essential role in both the elaboration and the subsequent organization of the myelin sheath.
Results
Isolation of a cDNA Clone Encoding
P,
P, is abundant in peripheral myelin but is absent from the
central nervous system. If this tissue-specific pattern of
expression of P, protein also reflected the levels of P,
mRNA, it should be possible to use procedures of differential screening with cDNA prepared from the mRNA of
brain and Schwann cells to isolate P, clones from a peripheral nerve cDNA library. We therefore performed a
series of in vitro translation experiments to determine
whether PO mRNA is restricted to Schwann cells. Poly(A)
RNA was isolated from both the sciatic nerves and brains
of 7 day rats. Several differences are observed when the
RNAs from the two tissues are translated in vitro (Figure
1, lanes a and b). A prominent protein of molecular weight
31.5 kd (Figure 1, arrow) is expressed in sciatic nerve but
not in the brain. lmmunoprecipitation
of the in vitro translation products with a monospecific rabbit antiserum prepared against purified P, (Brockes et al., 1980) indicates
that this protein is indeed P, and that its synthesis is restricted to the peripheral nerve (Figure 1, lanes c-e).
Given these observations, we constructed a peripheral
nerve cDNA library using the A phage cloning vector gtl0
(Huynh et al., 1985) and poly(A)’ RNA from the sciatic
nerve of rats sacrificed at 8-10 days after birth (a time of
rapid peripheral myelination [Webster, 19711). A portion of
this library was then screened for clones encoding
mRNAs expressed in sciatic nerve, but not in brain, by
replica filter hybridizations with 32P-labeled cDNAs from
the two tissues. Approximately
2% of the cDNA clones
hybridized intensely with sciatic nerve cDNA but not with
brain cDNA. Upon further characterization,
70% of these
clones were found to cross-hybridize. The 1.85 kb insert
from the largest member of this homologous set of cDNAs
was subcloned into the plasmid vector pUC8 (Vieira and
Messing, 1982) for further analysis.
This plasmid (pSN63) was immobilized on nitrocellulose filters and tested for its ability to hybrid-select poly(A)
RNA (Parnes et al., 1981) which when translated in vitro,
Cell
502
a
b
c
d
a
e
b
c
-68
-45
-68
-30
-
45
-
21
by Hybridization
to pSN63
-21
Figure
1. Localization
of P, mRNA
to Peripheral
Nerve
The 3Wabeled
proteins obtained from in vitro translation
(IVT) of 30
ng of poly(A)’ RNA from sciatic nerve (lane a) and brain (lane b) were
analyzed by SDS-polyacrylamide
gel electrophoresis.
The arrow indicates the most prominant IVT product, which is expressed
in peripheral nerve, but not in brain. Sciatic nerve IVT products (from translation
of 200 ng of RNA) were immunoprecipitated
with both anti-P, rabbit
antiserum (lane c) and normal rabbit serum (lane d). Brain IVT products were also immunoprecipitated
with anti-P, antiserum
(lane e).
Indicated size markers are in kilodaltons.
would direct the synthesis of PO. A single in vitro translation product is enriched by hybridization with pSN63 DNA
(Figure 2, lanes a and b). This 31.5 kd protein is specifically immunoprecipitated
by anti-P, antiserum (Figure 2,
lane c). No in vitro translation products are observed to be
enriched when hybridization is carried out in the presence
of pUC8 DNA alone (data not shown). These data, together with the nucleotide sequence data presented below, demonstrate that we have isolated a cDNA encoding
the PO glycoprotein.
P, mRNA Expression
In the rat sciatic nerve, the synthesis of peripheral
myelin
is first
migra-
observed
shortly
before
birth,
following
the
tion of Schwann cells from the neural crest. These cells
proliferate in situ and eventually envelop all peripheral
axons (Asbury, 1975; Raine, 1984). Myelination is thought
then to be triggered by a signal imparted by a subset
of axons (Aguayo and Bray, 1984; Bunge and Bunge,
1984). Myelin formation reaches peak synthetic levels
around
14 days and subsequently
falls to lower
maintenance levels in the adult. We have characterized the expression of P, mRNA as a function of postnatal development by Northern blot analysis using the P, cDNA as
probe (Figure 3b). P, is encoded in the same single mRNA
species at all developmental
stages examined. This
mRNA is readily detected in sciatic nerve at birth, increases approximately lo-fold to a peak level around day
14, and then falls to a steady state level in the adult (Table
1). P, mRNA is not detectable in control tissues such as
liver. Thus the level of PO mRNA parallels the level of myelin synthesis.
Figure
2. Selection
of P, mRNA
DNA
The 35S-labeled IVT products translated from unselected sciatic nerve
poly(A)’ RNA (lane a) and from poly(A)’ RNA selected by hybridization
to pSN63 DNA (lane b) were analyzed by SDS-gel electrophoresis.
The
enriched 31.5 kd protein is identified as P, by immunoprecipitation
of
the selected IVT products with anti-P, antiserum
(lane c). (1.6 pg of
sciatic nerve poly(A)’ RNA was hybridized
to 20 pg of immobilized
pSN63 DNA. Lanes b [before immunoprecipitation]
and c [after immunoprecipitation]
correspond
to proteins translated from 27% and 67%,
respectively,
of the RNA selected by this hybridization.)
Note that the
bands at 45 kd and 56 kd in lane b are not enriched by hybrid selection.
Indicated size markers are in kilodaltons.
P,, however, is only one of several enzymes and structural proteins required for the production of the myelin
membrane. For example, the massive increase in surface
area of the Schwann cell membrane requires the biosynthesis of large quantities of cholesterol. Since the
rate-limiting step in de novo cholesterol biosynthesis is
catalyzed by the enzyme HMG-CoA reductase (Brown and
Goldstein, 1980) we asked whether the expression of
mRNA encoding this enzyme was coordinately induced
with P, message. The time course of expression of HMGCoA reductase mRNA during postnatal development
parallels that of P, (Figure 3c, Table l), although RNA levels for the enzyme are less than 1% of those for the structural protein. It therefore appears that these two proteins
are part of a larger set of coordinately regulated gene
products required for the synthesis and maintenance of
the myelin membrane.
Northern blot analyses indicate that the P, gene is transcribed as a single mRNA species approximately
1.9 kb
in length (Figure 3a). Furthermore, this RNA encodes a
single in vitro translation product. In vivo, however, lower
molecular weight forms of PO (fragments at 23 kd and 19
kd) are often present in purified myelin preparations
(Roomi and Eylar, 1978; lshaque et al., 1980). Our data
suggest that these lower molecular weight species arise
Isolation
503
and Sequence
A
of a cDNA
Encoding
B
P,
of the sort characteristic
of the myelin basic proteins
(Roach et al., 1983; N. Takahashi, A. Roach, and L. Hood,
personal communication).
Furthermore, the intensity of
hybridizing bands suggests that this gene is present only
once in the haploid genome.
C
L 0
7
14 21 26 A
L 0
7
14 21 26 A
,,$
.j*
9444-
-265
23-
06-
Figure
3. Expression
of PO mRNA
(A) Northern
blot of the P, message. 0.5 rg of poly(A)’ RNA from the
sciatic nerves of E-10 day rats was electrophoresed
through a 1%
agarose-formaldehyde
gel, blotted on to Gene Screen (NEN), and
probed with a2P-labeled P, DNA (pSN63 insert). Indicated size markers
are in kilobases.
(6) Northern
blot of P, RNA levels during development. 3 rg of total RNA, isolated from the sciatic nerves of rats at various times after birth, was probed with “‘P-labeled
P, cDNA. L, adult
rat liver total RNA (3 rg); A, adult sciatic nerve RNA; numbers above
each lane represent the time, in postnatal days, at which sciatic nerves
were dissected for RNA isolation. Cellular proliferation
in the rat sciatic
nerve has largely ceased by postnatal day 3 (Diner, 1965; Asbury,
1967). Exposure
time: 5 hr with an intensifying
screen. (C) Northern
blot of HMG-CoA
reductase
levels during development.
The RNA
samples in Figure 38 were electrophoresed
in parallel lanes of the
same gel and independently
probed with the 3zP-labeled cDNA insert
from the plasmid pRed-10, which corresponds
to the 3’ region of the
HMG-CoA reductase mRNA (Liscum et al., 1983). Exposure
time: 92
hr with an intensifying
screen.
either through posttranslational
modifications or are generated by proteolysis during purification. They are not
likely to be generated by alternative RNA splicing events.
Genomic Representation
We have performed genomic blotting experiments to characterize the P, gene in rat chromosomal DNA. Figure 4 is
a Southern blot of total rat genomic DNA digested by four
different restriction endonucleases. Two of these enzymes
(Eco RI and Hind Ill) do not cut within the P, cDNA, one
(Nco I) exhibits a single site, and the remaining enzyme
(Pst I) cuts twice within the cDNA. Eco RI and Hind Ill each
give a single band upon digestion of genomic DNA (Figure 4, lanes a and b), while Nco I and Pst I yield two and
three bands, respectively. Since the cDNA probe we have
used is almost full length, these data suggest that the P,
gene is relatively small and does not contain large introns
Table
1. Relative
RNA Levels
during
Postnatal
The Sequence of P,
The 1.85 kb cDNA we have cloned hybridizes with a single
species of RNA 1.9-2.0 kb in length, suggesting that we
have isolated a nearly full-length cDNA. The restriction
map of this cDNA as well as the strategy employed in its
sequencing (using the chemical cleavage reactions described by Maxam and Gilbert, 1980) is shown in Figure
5. The primary structure of P, was deduced from the
nucleotide sequence of the only open reading frame long
enough to generate the complete protein. This deduced
sequence was confirmed by comparison with previously
published sequences of fragments of the isolated protein.
lshaque and his colleagues (1980) have determined the
17 N-terminal amino acids of rabbit P,. This sequence,
with amino acid substitutions at four positions, is present in the deduced sequence beginning with the Ile encoded at nucleotide 119. The variation between these
sequences probably reflects evolutionary divergence between rabbit and rat. lshaque et al. have also presented
amino acid sequence data on two additional proteolytic
fragments of the P, protein. The first of these, the so-called
“PO glycopeptide:’ begins at Asp 80 (numbering from Ile-1
at nucleotide 119) of the deduced sequence. The first 21
amino acids of this glycopeptide (determined by sequenator analysis) agree completely with the deduced sequence, except for the interposition of His-86 and Asp-87.
The sequence of the seven remaining amino acids of this
fragment was determined by carboxypeptidase
analysis
and is not present anywhere in the deduced sequence.
Since the order of these seven residues was ambiguous
in the protein sequence, it seems likely that they are artifactual. The third sequence published by lshaque and
his colleagues is a ten residue stretch, purportedly located at the amino terminus of a 19 kd tryptic fragment,
which the authors assert is identical with a P, fragment
(the “19K glycoprotein”) commonly observed in isolated
myelin. This decapeptide is not present at any position in
the deduced sequence, nor is it present in the amino acid
sequences produced by translating the pSN63 cDNA insert sequence in the two nontranslating
reading frames.
Development
Day
RNA Transcript
0
7
14
21
PO
1.0
6.9 + 0.5
9.8 k 0.4
9.3 f
1 .o
4.2
7.0
4.4
HMG-CoA
reductase
0.1
26
Adult
7.8 + 0.2
1.7 * 0.1
3.1
1.4
To quantitate levels of P, RNA during development,
the autoradiogram
of Figure 38 was scanned (full width of each lane) with a Joyce-Lobe1
densitometer, and the signal across the entire peak of P, hybridization
was integrated. The level of P, RNA at day 0 was arbitrarily set to 1.O and RNA
levels at later times were normalized
to this value. Values given are averages
of scans of three different exposures
(3, 5, and 6 hr) + standard
deviation. A similar densitometer
analysis was performed for the HMG-CoA reductase blot (Figure 3C). These data were normalized to the HMG-CoA
reductase
RNA level at day 0. Note that this level is less than 1% of the PO RNA present at this time. No signal above background
was observed
for liver RNA in the P, blot, X-ray film was preflashed
prior to autoradiography.
Cell
504
a
b
c
d
23.
9.4
6.5
4 .4
2.3
from Ile-1 through Arg-124, and an intracellular domain
from Arg-151 through Lys-219. The extracellular domain
contains several hydrophobic stretches. It also contains
(at Asn-93) the canonical acceptor sequence for N-linked
(Asn-Gly-Thr)
glycosylation:
Asn-X-A$:. This sequence
occurs only once in the deduced P, structure. Its assignment is consistent with several experiments which indicate that POis glycosylated at asingle position via N-linked
attachment, and that the attached oligosaccharide
is extracellularly
positioned (Wood and McLaughlin,
1975;
Peterson and Gruener, 1978; lshaque et al., 1980). Of the
89 residues comprising the cytoplasmic (C-terminal) domain, 21 are basic and only 6 are acidic. Thus, the cytoplasmic domain carries a strong (+15) positive charge.
The stop codon at nucleotide 776 was confirmed by sequencing through the corresponding
region of a shorter
P, cDNA clone, independently
isolated from the AgtlO
library.
Discussion
Figure
4. Southern
Blot of Rat DNA with P, cDNA
Twenty micrograms
of Sprague-Dawley
rat genomic DNA was digested
with restriction
endonuclease.
electrophoresed
through a 0.8% agarose gel, blotted on to Gene Screen, and probed with 32P-labeled P,
cDNA. Lane a, Hind III; Lane b, Eco RI; Lane c. Nco I; and Lane d, Pst
I. The arrowhead
marks the position of a faint band apparent in the
autoradiogram,
but barely visible in this photograph.
Indicated size
markers are in kilobases.
Given that this tryptic fragment was purified by repeated
chromatography
on SDS-agarose columns, it is likely that
the sequence published represents the amino terminus of
a contaminating
peptide.
Several features of the primary structure of P, are relevant to its localization in the myelin membrane. The initiator methionine at nucleotide 32 is followed by 28 uncharged and/or nonpolar amino acids which precede the
N-terminal isoleucine of the mature protein. This domain
almost certainly comprises the signal sequence necessary for the translocation and appropriate insertion of nascent P, into the Schwann cell membrane. (This rather long
signal sequence may account for the slightly reduced
SDS-gel mobility of the P, in vitro translation product relative to the mature protein.) Hydrophobicity
profiles, obtained using the program of Kyte and Doolittle (1982) define a single highly hydrophobic
membrane-spanning
region from Tyr-125 through lie-150, bounded by charged
anchors on either side. Assuming a conventional polarity
of membrane insertion, this single traverse of the bilayer
spatially divides the protein into an extracellular domain
During the course of myelination,
newly synthesized
Schwann cell plasma membrane is continually wrapped
about the peripheral axon, The final structure is highly ordered. When fixed by conventional methods and viewed
in cross section in the electron microscope, the mature
sheath appears as a compact spiral of membrane bilayers
separated by regularly alternating bands which arise from
the tight association of apposed cytoplasmic and extracellular membrane faces (Napolitano and Scallen, 1969;
Peterson and Pease, 1972). These alternating bands have
been termed the “major dense line” (formed by the apposition of cytoplasmic faces) and the “intraperiod
line”
(formed by the apposition of extracellular
membrane
faces). Several features of the deduced sequence of P,,
together with available anatomical and genetic data, suggest that this protein may be responsible for the formation
of both of these tight associations.
The appearance of myelin in the mouse mutant shiverer (shi) suggests a mechanism for the association of
cytoplasmic membrane surfaces. This mutant is characterized phenotypically by a generalized action tremor presumably due to markedly diminished myelination of axons
within the central nervous system (Bird et al., 1978). The
shiverer phenotype results from the deletion of a large
portion of the gene coding for the myelin basic proteins
(Roach et al., 1983). This family of cytoplasmic membrane
proteins (21.5, 18.5, 17, and 14 kd) is normally expressed
in both central and peripheral nervous system myelin, but
is undetectable in shi/shi homozygotes (Kirschner and
Ganser, 1980). Although little myelin is apparent in the
brains and spinal cords of these animals, loose wrappings
of myelin-like membrane, adherent at apposed extracellular surfaces but full of cytoplasm between these appositions, are occasionally observed (Ganser and Kirschner,
1980). In contrast, peripheral nervous system myelin from
the same animals is fully elaborated, has a compact structure with normal periodicity, and is functionally normal
(Kirschner and Ganser, 1980). These findings have led to
Isolation
505
and Sequence
of a cDNA
Encoding
P,
Figure 5. DNA
Sequence
and Deduced
P, Amino
Acid
(A) Sequencing
strategy. Box represents
translated region. Arrows
indicate length of sequence, read from a 3’ or 5’ labeled end, as
noted. The DNA sequence presented
extends
from the 5’end of the pSN63 insert to the Xba
I site positioned
at nucleotide
1024. (B) DNA
and amino acid sequence.
The P, precursor
is
encoded
in nucleotides
32-775. Uncharged
amino acids are underlined.
+++
denotes
basic amino acids. The putative membranespanning domain from Tyr-125 through He-150
(nut. 491-566)
is boxed. Numbers
in parentheses at the end of each line are amino acids
of the mature protein.
C&C ?a~ WI' mw,G
GK 'ICC F.X AAGCCC UX CGG CAGXCCCAQTG
~1s LYS ser ser L~S asp Ser
Lys Arq &
Arq Gln 'I% Pro Vd
-+++
+t+
+++
720
'XG TATCCCATG CTG
Iru Wr Ala Net Leul194)
the suggestion that the myelin basic proteins are involved
in compaction at apposed cytoplasmic faces and that
some functionally related component of peripheral myelm, not present in the central nervous system, serves this
function in the periphery. This component is likely to be
the cytoplasmic domain of P,, given its abundance, its
localization at the cytoplasmic surface of the bilayer, and
its strong positive charge.
A computer-assisted
comparison of the sequence of the
cytoplasmic domain of P, to the myelin basic proteins re-
Cell
506
Figure 6. Diagrammatic
Representation
Orientation
in the Myelin Membrane
major
dense
(inside)
of P,
Semicircles
(0)
represent the P. extracellular
domain, and’tgangles
(A)
the Pi cytoplasmic
domain. Cytoplasmic
domains are illustrated
as forming the major dense line through electrostatic
interaction
with the head groups of
acidic lipids ( fl ) present in the cytoplasmic
leaflet of the apposed
bilayer. Extracellular
domains are illustrated as interacting with each
other to form the intraperiod line.
line
Cytoplasmic
domain
intraperiod
line
(outside)
veals short patches of incomplete homology. The mouse
14 kd basic protein sequence from residue 2 through 8, for
example, shares homology with a P, sequence from residue 160 through 166:
MBP
PO
Ala-Ser-Gin-LysArg-Pro-Ser
Ala-Leu-Gin-ArgArg-Leu-Ser
The extent of these homologies, however, is inadequate to
suggest a common origin for these two genes. The myelin
basic proteins and the P, cytoplasmic domain do share a
highly basic character: the ratio of basic to acidic amino
acids is 3.3 for the 14 kd basic protein of the mouse, and
3.5 for the rat P, cytoplasmic domain.
How might the cytoplasmic domain of P, serve as a
membrane adhesion molecule? One possibility is that the
domain interacts electrostatically with acidic lipids present
in the cytoplasmic face of the apposed bilayer. This is
schematically shown in Figure 6. Acidic lipids are present
in myelin (Norton and Cammer, 1984) and there is good
evidence in other systems (e.g., the erythrocyte membrane) that these negatively charged lipids may be preferentially localized to the cytoplasmic leaflet of the bilayer
(Verkleij et al., 1973). Furthermore, when PO is purified
from peripheral myelin, it is complexed with a set of very
tightly bound acidic lipids (Ishaque et al., 1980). The most
abundant of these associated acidic lipids is phosphotidylserine, which in the red cell, is confined almost exclusively to the cytoplasmic leaflet of the bilayer. Although
this sort of electrostatic protein-lipid
interaction seems to
us most straightforward, it remains possible that apposed
PO cytoplasmic domains connect intracellular membrane
surfaces with each other via protein-protein
interactions.
The structure of the extracellular domain of POsuggests
that this protein may mediate the association of extracellular membrane faces through a different set of molecular
interactions. The extracellular domain of P, is both hydrophobic and glycosylated. Structures of this sort are
often “self-adhesive”(Edelman,
1983). It is therefore possible that the association of apposed extracellular mem-
Extracellular
domain
brane surfaces occurs via homophilic interactions between facing P, molecules. This is also illustrated in
Figure 6. Support for this hypothesis comes from the observation that the purified protein readily aggregates and
becomes insoluble in aqueous solutions and neutral organic solvents (Brostoff et al., 1975). Examination of the
sequence of the extracellular domain reveals several series of uncharged/hydrophobic
residues (e.g., 9-26, 4754, 110-119) through which hydrophobic bonding could
occur. The P, oligosaccharide
is also present within this
extracellular domain, and it is possible that interactions
between facing sugar moieties also mediate the association of P, molecules. Such interactions appear to play an
important role in the function of the neural cell adhesion
molecule (N-CAM) (Edelman, 1983). Thus, P, may serve
to organize the myelin membrane by mediating different
associations at apposed extracellular
and cytoplasmic
membrane faces. The possible role of P, in the structural
organization of the mature myelin sheath is summarized
diagrammatically
in Figure 6.
The Elaboration
of Myelin
During development,
a given myelin-forming
Schwann
cell envelops a single axon. Newly formed membrane
then follows a relatively uniform course around the axon
many times, generating a thick sheath with the Schwann
cell body outermost (Webster, 1971). Electron micrographs
of the growing sheath suggest a model in which the
Schwann cell body remains “stationary” during this process, while the growing myelin membrane
migrates
around the axon as a two-dimensional
fluid, continually
tucking under the adjacent, previously deposited layer
(Raine, 1984; Webster and Favilla, 1984). If this “tuck under” model is correct, what defines the initial circumferential path and what maintains it? It has been suggested, for
example, that specific recognition and adhesion between
the axonal surface and the growing myelin membrane,
coupled with the force of continuous new membrane biosynthesis, would allow for appropriate growth of the myelin sheath (Raine, 1984). This mechanism by itself, how-
Isolation
507
and Sequence
of a cDNA
Encoding
P,
ever, is probably not sufficient to account for the ordered
biosynthesis
of new membrane.
We suggest that
homophilic
interactions
between the extracellular
domains of facing P, molecules would serve to guide the
growing myelin membrane
appropriately.
During the
wrapping process, the growing membrane would therefore interact on one extracellular face with the axon, and
on the other, with the adjacent (previously deposited) layer
of myelin via P,-P, bonding. Although lamellae adjacent
to the growing membrane are displaced, they will maintain their circumferential disposition by virtue of this association. Data presented above indicate that PO mRNA
levels in rat sciatic nerve are already markedly elevated at
birth, a point at which myelin formation has just begun.
Recent fine structure immunocytochemical
studies of
Trapp et al. (1981) further indicate that P, protein is expressed at the earliest stages of myelination, and that it
may be preferentially directed to the myelin membrane
and excluded from the Schwann cell plasma membrane.
The molecule would therefore appear to be expressed appropriately, both temporally and spatially, to act as a
guidepost for myelin wrapping.
The structure of the P, glycoprotein, along with the time
of expression of its mRNA, suggests that P, may be essential both for the elaboration of the myelin membrane
as well as for the maintenance of the compact structure
of mature myelin. We suggest that the hydrophobic extracellular domain creates a “self-adhesive” extracellular
membrane face which guides the wrapping process and
ultimately compacts adjacent lamellae. The highly basic
intracellular domain appears to be functionally related to
the myelin basic proteins. It may serve to compact apposed cytoplasmic membrane faces by association with
acidic lipids in the cytoplasmic leaflet of the apposed
bilayer. The availability of cDNA and genomic clones encoding P, and the ability to insert this gene into new cellular environments
should now permit us to test these
models of PO function.
Experimental
Hybrid Selection
and lmmunoprecipitation
Hybrid selections of SN poly(A)* RNA were carried out by the method
of Parnes et al. (1981) as modified by Maniatis & al. (1982), except that
the formamide concentration
was lowered to 45% and the salt raised
to 0.5 M. SN poly(A)’ RNA was hybridized
to both pSN63 and pUC8
DNA. Selected RNAs were translated
in a rabbit reticulocyte
lysate.
purchased
from New England Nuclear, according
to procedures
provided by the manufacturer.
lmmunoprecipitations
of in vitro translation
products were performed
according to standard procedures
(Kessler,
1981). using a well-characterized
monospecific
rabbit antiserum to rat
P, (Brockes et al., 1980). SDS-polyacrylamide
gel electrophoresis
was
performed
according
to the procedure
of Laemmli (1970). Fluorography (Figures 1 and 2) was performed
by equilibrating
SDS gels in Enlighten (New England Nuclear), according
to procedures
supplied by
the manufacturer.
Autoradiography
was conducted
with Kodak XAR-5
film at -70%.
Southern
and RNA Blot Analysis
Rat genomic DNA was a gift from Dr. Ashok Kulkarni. Southern and
Northern
blots were carried out as previously
described
(Southern,
1975; Goldberg,
1980), except that Northern blots were performed via
transfer to Gene Screen (New England Nuclear), following procedures
supplied by the manufacturer.
Probes were a2P-labeled nick-translated
DNA fragments
(Rigby et al., 1977). P, blots were probed with the full
length pSN63 insert.
DNA Sequence
Analysis
DNA sequences
were determined
by the Maxam-Gilbert
previously
described
(Maxam and Gilbert, 1980).
method,
as
Acknowledgments
The authors are grateful to Dr. Jeremy Brockes for the generous gift
of anti-P, antiserum, to Dr. Joe Goldstein for the HMG-CoA reductase
cDNA clone, to Dr. Henry Webster for commenting
on the manuscript,
and to Phyllis Kisloff for secretarial
assistance.
This work was sup
ported by a Muscular Dystrophy Association
postdoctoral
fellowship to
G. L., by NIH grants to R. A., and by the Howard Hughes Medical Institute.
The costs of publication of this article were defrayed in part by the
payment
of page charges.
This article must therefore
be hereby
marked “advertisement”
in accordance
with 18 U.S.C. Section 1734
solely to indicate this fact.
Received
November
21, 1984; revised
December
14, 1984
Procedures
RNA Isolation
RNA was isolated from the brains and sciatic nerves of 7-10 day
Sprague-Dawley
rats by homogenization
in 5 M guanidinium
thiocyanate, followed by ultracentrifugation
through a 5.7 M CsCl cushion
(Chirgwin et al., 1979). Poly(A)’ RNAs were enriched for by oligo(dT)
cellulose chromatography
(Aviv and Leder, 1972).
cDNA Library
Construction
Double-stranded
cDNA was synthesized
from rat sciatic nerve (SN)
poly(A)* RNA. First strand synthesis was carried out with reverse transcriptase in the presence of actinomycin
D (40 pglml), using oligo(dT)
as primer. Double-stranded
cDNA was generated by a modification
of
the procedure
described
by Okayama and Berg (1982), using RNAase
H, DNA polymerase
I, and DNA ligase. After treatment with T4 polymerase and Eco RI methylase, the double-stranded
cDNA was cloned
into the imm434 Eco RI insertion vector lgtl0 (Huynh et al., 1985) using
Eco RI linkers. Forty nanograms
of cDNA produced a library of 2.5 x
lo5 plaques. Replica filters of a portion of this library (5,000 plaques)
were screened in duplicate with 32P-labeled cDNAs synthesized
from
brain and SN poly(A)* RNAs (Maniatis et al., 1982). Sciatic nervespecific clones were sorted into non-cross-hybridizing
+I- sets. The
1.85 kb insert from the largest clone (ASN63) of the major set was subcloned into the Eco RI site of pUC8 (Vieira and Messing, 1982). This
plasmid, designated
pSN63, was used for further analysis.
Aguayo, A. J. and Bray, G. M. (1984). Cell interactions
studied in peripheral nerves of experimental
animals. In Peripheral Neuropathy,
2nd
ed., P J. Dyck, I? K. Thomas,
E. H. Lombert,
and R. Bunge, eds.
(Philadelphia:
W. B. Saunders),
pp. 360-377.
Asbury, A. K. (1967). Schwann cell proliferation
sciatic nerve. J. Cell Biol. 34, 735-743.
in developing
mouse
Asbury, A. K. (1975). The biology of Schwann cells. In Peripheral
ropathy, Vol. 1, P J. Dyck, P. K. Thomas, and E. H. Lambert,
(Philadelphia:
W. B. Saunders),
pp. 201-212.
Neueds.
Aviv, H., and Leder, I? (1972). Purification of biologically active globin
messenger
RNA by chromatography
on oligothymidylic
acid-cellulose.
Proc. Natl. Acad. Sci. USA 69, 1408-1412.
Bird, T. O., Farrell, D. F., and Sumi. S. M. (1978). Brain lipid composition
of the shiverer mouse: genetic defect in myelin development.
J. Neurochem. 37, 387-394.
Braun, P E. (1984). Molecular organization
of myelin. In Myelin,
ed., P Morell, ed. (New York: Plenum Press), pp. 97-113.
2nd
Brockes, J. P., Raff, M. C.. Nishiguchi,
D. J., and Winter, J. (1980).
Studies on cultured rat Schwann cells. Ill. Assays for peripheral myelin
proteins. J. Neurocytol.
9. 67-77.
Brostoff, S. W.. Karkhanis, Y. D., Carlo, D. J., Reuter, W.. and Eylar, E.
H. (1975). Isolation and partial characterization
of tha major proteins
of rabbit sciatic nerve myelin. Brain Res. 86, 449-458.
Cell
506
Brown, M. S., and Goldstein, J. L. (1980). Multivalent feedback
tion of HMG CoA reductase, a control mechanism
coordinating
noid synthesis
and cell growth. J. Lipid Res. 27, 505-517.
regulaisopre-
Bunge, Ft. P, and Bunge, M. B. (1984). Tissue culture observations
relative to peripheral
nerve development,
regeneration
and disease. In
Peripheral Neuropathy,
2nd ed., P J. Dyck, P K. Thomas, E. H. Lombert, and R. Bunge, eds., (Philadelphia:
W. 6. Saunders), pp. 378-399.
Chirgwin, J. M., Przybyla, A. E., MacDonald,
R. J., and Rutter, W. J.
(1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease.
Biochemistry
18, 5294-5299.
Diner, 0. (1965). Les cellules de Schwann en mitose et leur rapports
avec les axones au tours du developpement
du nerf sciatique chez le
rat. Compt. Acad. Sci., Paris 261, 1731-1737.
Edelman,
450-457.
G. M. (1983).
Cell
adhesion
molecules.
Science
219,
Ganser, A. L., and Kirschner, P A. (1980). Myelin structure
in the absence of basic protein in the Shiverer mouse. In Neurological
Mutations Affecting Myelination,
N. Baumann,
ed. (Amsterdam:
Elsevied
North Holland), pp. 171-176.
Geren, B. B. (1954). The formation of the Schwann cell surface of myelin in peripheral nerves of chick embryos. Exp. Cell Res. 7, 558-562.
Goldberg, D. A. (1980). Isolation
sophila alcohol dehydrogenase
5794-5799.
Greenfield,
S., Brostoff,
composition
of myelin
rochem. 20, 1207-1216.
and partial characterization
gene. Proc. Natl. Acad.
S., Eylar, E. H., and Morell,
of the peripheral
nervous
of the DroSci. USA 77,
P (1973). Protein
system.
J. Neu-
Huynh, T., Young, Ft. A., and Davis, R. W. (1985). Constructing
and
screening
cDNA libraries in AgtlO and Agtll. In DNA Cloning Techniques, A Practical Approach,
D. M. Glover, ed. (Oxford:
IRL Press),
in press.
Ishaque, A., Roomi, M. W., Szymanska,
I., Kowalski, S., and Eylar, E.
H. (1980). The PO glycoprotein
of peripheral nerve myelin. Can. J. Biothem. 58, 913-921.
Kessler, S. W. (1981). Use of protein-A-bearing
mune precipitation and isolation of antigens
73, 442.
staphylococci
for the imfrom cells. Meth. Enzymol.
Kirschner, D. A., and Ganser, A. L. (1980). Compact myelin exists in the
absence of basic protein in the Shiverer mutant mouse. Nature 283,
207-209.
Kyte, J., and Doolittle,
hydropathic
character
R. F. (1982). A simple method for displaying
of a protein. J. Mol. Biol. 757; 105-132.
Laemmli, U. K. (1970). Cleavage of structural
proteins
sembly of bacteriophage
T,. Nature 227, 680-685.
during
the
the as-
Lees, M. B., and Brostoff, S. W. (1964). Proteins of Myelin. In Myelin,
2nd ed., P Morell, ed., (New York: Plenum Press), pp. 197-224.
Liscum, L., Luskey, K. L., Chin,
Brown, M. S. (1983). Regulation
zyme A reductase and its mRNA
antibody and a cDNA probe. J.
D. J., Ho, Y. K., Goldstein, J. L., and
of 3-hydroxy-3-methyl-glutaryl
Coenin rat liver studied with a monoclonal
Biol. Chem. 258, 8450-8455.
Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). Molecular Cloning.
(Cold Spring Harbor, New York: Cold Spring Harbor Laboratory),
pp.
326-328.
Maxam, A. M., and Gilbert,
with base specific chemical
Morell,
P (1984).
Myelin,
W. (1980).
cleavages.
2nd ed. (New
Sequencing
end-labeled
DNA
Meth. Enzymol. 65, 499-560.
York: Plenum
Press).
Napolitano, L. M., and Scallen, T. J. (1969). Observations
on the fine
structure
of peripheral
nerve myelin. Anat. Rec. 763, l-23.
Norton, W. T., and Cammer, W. (1984). Isolation and Characterization
of Myelin. In Myelin, 2nd ed. P Morell. ed., (New York: Plenum Press),
pp. 147-195.
Okayama, H., and Berg, P (1982). High-efficiency cloning
cDNA. Mol. Cell. Biol. 2, 161-170.
Parries.
Apella,
clones.
of full-length
J. R., Velan, B., Felsenfeld,
A., Ramanathan,
L., Ferrini, V.,
E., and Sidman, J. G. (1981). Mouse /I2 microglobulin
cDNA
Proc. Natl. Acad. Sci. USA 78, 2253-2258.
Peterson, R. G., and Grevner, R. W. (1978). Morphological
of PNS myelin proteins. Brain Res. 152, 17-29.
localization
Peterson, R. G., and Pease, D. C. (1972). Myelin embedded in polymerized glutaraldehyde-urea.
J. Ultrastruct.
Res. 47, 115-132.
Poduslo, J. F. (1984). Regulation of myelination: biosynthesis
of the major myelin glycoprotein
by Schwann cells in the presence and absence
of myelin assembly. J. Neurochem.
42, 493-503.
Politis, M. J., Sternberger,
N., Ederle,
Studies on the control of myelogenesis.
K., and Spencer, P. S. (1962).
J. Neuroscience
2. 1252-1266.
Raine, C. S. (1984). Morphology
of Myelin and Myelination.
In Myelin,
2nd ed., P. Morell, ed. (New York: Plenum Press), pp. l-50.
Ranvier, M. L. (1878). Lecons
(Paris: Librarie F Savy).
sur I’Histologie
du Systeme
Nerveux.
Rigby, P. W. J., Dieckman,
M.. Rhodes, C., and Berg, P (1977). Labeling
deoxyribonucleic
acid to high specific activity in vitro by nick translation with DNA polymerase
I. J. Mol. Biol. 773, 237-251.
Ritchie. J. M. (1983). On the relation
duction velocity in myelinated nerve
B 277, 29-39.
between fibre diameter and confibres. Proc. R. Sot. London Ser.
Roach, A., Boylan, K., Horvath, S.. Prusiner, S. B., and Hood, L. E.
(1983). Characterization
of cloned cDNA representing
rat myelin basic
protein. Cell 34, 799-806.
Robertson,
J. D. (1962). The unit membrane of cells and mechanisms
of myelin formation. In Ultrastructure
and Metabolism of the Nervous
System, S. R. Korey, A. Pope, and E. Robins, eds. (Baltimore: William
and Wilkins), pp. 94-158.
Roomi, M. W., and Eylar, E. H. (1978). Isolation of a product from the
trypsin-digested
glycoprotein
of sciatic nerve myelin. Biochim. Biophys. Acta 536, 122-129.
Southern, E. (1975). Detection of specific
ments separated
by gel electrophoresis.
sequences
among DNA fragJ. Mol. Biol. 98, 503-511.
Trapp, B. D., Itoyama, Y., Sternberger,
N. H.. Quarles, R. H., and Webster, H. de F. (1981). lmmunocytochemical
localization of P, protein in
Golgi complex
membranes
and myelin of developing
rat Schwann
cells. J. Cell Biol. 90, l-6.
Verkleij, A. J., Zwaal, R. F. A., Roelofsen, B., Cornfurius,
P., Kastelijn,
D., and Van Deenen, L. L. M. (1973). The asymmetric
distribution
of
phospholipids
in the human red cell membrane.
Biochim. Biophys.
Acta 323, 178-193.
Vieira, J., and Messing, J. (1962). The pUC plasmids, an M13mp7 derived system for insertion mutagenesis
and sequencing
with synthetic
universal primers. Gene 79, 259-268.
Webster, H. de F (1971). The geometry
of peripheral
myelin sheaths
during their formation and growth in rat sciatic nerves. J. Cell Biol. 48,
348-370.
Webster, H. de F, and Favilla, J. T. (1984). Development
of peripheral
nerve fibers. In Peripheral
Neuropathy,
2nd ed., P. J. Dyck, I? K.
Thomas, E. H. Lombert. and R. Bunge, eds. (Philadelphia:
W. B. Saunders), pp. 329-359.
Wood, J. G., and McLaughlin,
J. (1975). The visualization
of concanavalin A binding sites in the intraperiod line of rat sciatic nerve myelin. J. Neurochem.
24, 233-235.