Shiverer and Normal Peripheral Myelin Compared: Basic Protein

1 downloads 0 Views 1MB Size Report
Journal of Neurochemistry ... Department of Neuroscience, Children's Hospital, and Department of ... teins and indicates that in normal myelin, the basic pro-.
Journal of Neurochemistry Raven Press, New York 0 1985 International Society for Neurochemistry

Shiverer and Normal Peripheral Myelin Compared: Basic Protein Localization, Membrane Interactions, and Lipid Composition Hideyo Inouye, Allen L. Ganser, and Daniel A. Kirschner Department of Neuroscience, Children’s Hospital, and Department of Neuropathology, Harvard Medical School, Boston, Massachusetts, U.S.A.

Abstract: We have correlated membrane structure and interactions in shiverer sciatic nerve myelin with its biochemical composition. Analysis of x-ray diffraction data from shiverer myelin swollen in water substantiates our previous localization of an electron density deficit in the cytoplasmic half of the membrane. The density loss correlates with the absence of the major myelin basic proteins and indicates that in normal myelin, the basic protein is localized to the cytoplasmic apposition. As in normal peripheral myelin, hypotonic swelling in the shiverer membrane arrays occurs in the extracellular space between membranes; the cytoplasmic surfaces remain closely apposed notwithstanding the absence of basic protein from this region. Surprisingly, we found that the interaction at the extracellular apposition of shiverer membranes is altered. The extracellular space swells to a greater extent than normal when nerves are incubated in distilled water, treated at a reduced ionic strength of 0.06 in the range of pH 4-9, or treated at constant pH (4 or 7) in the range of ionic strengths 0.02-0.20. To examine the biochemical basis of this difference in swelling, we compared the lipid composition of shiverer and normal

myelin. We find that sulfatides, hydroxycerebroside, and phosphatidylcholine are 20-30% higher than normal; nonhydroxycerebroside and sphingomyelin are 15-20% lower than normal; and ethanolamine phosphatides, phosphatidylserine, and cholesterol show little or no change. A higher concentration of negatively charged sulfatides at the extracellular surface likely contributes to an increased electrostatic repulsion and greater swelling in shiverer. The cytoplasmic surfaces of the apposed membranes of normal and shiverer myelins did not swell apart appreciably in the pH and ionic strength ranges expected to produce electrostatic repulsion. This stability, then, clearly does not depend on basic protein. We propose that PO glycoprotein molecules form the stable link between apposed cytoplasmic membrane surfaces in peripheral myelin. Key Words: Shiverer mutant mouseMyelin structure-Myelin basic protein-Myelin lipids - Sulfatide - PO glycoprotein - Membrane - membrane interactions-Mouse. Inouye H. et al. Shiverer and normal peripheral myelin compared: Basic protein localization, membrane interactions, and lipid composition. J . Neurochem. 45, 1911-1922 (1985).

The shiverer mutant mouse lacks the major myelin basic proteins (Dupouey et al., 1979; Kirschner and Ganser, 1980; Mikoshiba et al., 1981; Winter, 1982). In contrast to the severe myelin deficiency in the CNS, the myelin in the peripheral nervous system (PNS) is present in nearly normal amounts (Peterson and Bray, 1984) and appears to be grossly normal in ultrastructure as visualized by electron microscopy (Privat et al., 1979; Kirschner and Ganser, 1980; Rosenbluth, 1980~).Closer examination of the lamellar structure of shiverer myelin by x-ray diffraction, a technique uniquely suited for analyzing molecular organization in native myelin (Kirschner et al., 1984), reveals a deficit

of material in the cytoplasmic half of the membrane bilayer and in the cytoplasmic space. In addition, tFe cytoplasmic space between membranes is 1-2 A wider than normal (Kirschner and Ganser, 1980; Ganser and Kirschner, 1980). These changes in membrane structure and packing have been attributed to the absence of basic proteins from the cytoplasmic side of the membrane. Since basic protein comprises 15-20% of the total protein in normal mouse PNS myelin (Greenfield et al., 1980), one might expect that the absence of this protein would greatly alter the interactions between the cytoplasmic surfaces of shiverer myelin membranes. In a preliminary study we found that shiverer ~~

Received April 4, 1985; accepted June 21, 1985. Address correspondence and reprint requests to Dr. D. A. Kirschner at Department of Neuroscience, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115, U.S.A.

1911

Abbreviation used: PNS, peripheral nervous system.

H . INOUYE ET A L .

1912

myelin appeared to respond like normal myelin to hypotonic media, i.e., the cytoplasmic surfaces remained adherent, whereas the extracellular surfaces swelled apart (Ganser and Kirschner, 1980). Subsequently, we did find a difference in shiverer membrane interactions. Under certain conditions of pH or metal cation composition of the incubation medium, swelling occurred in both normal and shiverer peripheral myelin, but the extent of swelling was greater in shiverer. Moreover, contrary to our expectations, this greater swelling was localized to the extracellular and not to the cytoplasmic apposition (Inouye and Kirschner, 1984). Motivated by these unexpected findings, we decided to test our previous interpretation of the diffraction pattern from shiverer PNS myelin, to examine more systematically the interactions between the myelin membranes in shiverer as a function of pH and ionic strength, and to analyze the lipid composition of shiverer myelin. Our diffraction analysis substantiates localization of basic protein to the cytoplasmic half of the membrane and suggests that there is a greater electrostatic repulsion on the extracellular surfaces of shiverer myelin membranes. Analysis of the lipids indicates that shiverer peripheral myelin contains 20-25% more sulfatides than normal. We suggest that this increase in negatively charged glycolipid contributes in part to the altered membrane interactions at the extracellular apposition in shiverer. A preliminary report of some of this work has been presented (Kirschner and Inouye, 1984). MATERIALS AND METHODS

Specimens Sciatic nerves were dissected from mature mice, either normal ones ( + / + ) of the strain C57BL/6J, or mutant shiverers (shilshi)or proven heterozygotes ( + ishi) on hybrid backgrounds. For x-ray diffraction the nerves were tied off at both ends and incubated at room temperature in distilled water for variable times or in solutions of different pH and ionic strength for -24 h. The effects of pH were examined at a constant ionic strength of 0.06, which was obtained by adding buffer to solutions of 40-50 mM NaCI. The pH was controlled by: HCI at pH 2; glycine and HCI at pH 2.1-3.6; formic acid and NaOH at pH 3.1-4.3; sodium acetate and acetic acid at pH 4.2-5.1 ; sodium mono- and dibasic phosphates at pH 6.1-8.1; Tris base and HCI at pH 7.6-8.9; and glycine and NaOH at pH 9.0-10.6. The effects of ionic strength were studied at pH 4 and 7, in which case NaCl was added to 15 mM formic acid at pH 4 or to 10 mM phosphate buffer at pH 7. For lipid analysis and isolation of myelin, sciatic nerves were dissected out, immersed in normal saline, freed of adhering fat, blotted dry, frozen in liquid nitrogen, and stored at - 80°C. Whole nerve lipids were analyzed in the pooled, lyophilized sciatic nerves from five normal and from five shiverer mice. Myelin was isolated from sciatic nerves from 10 normal and 14 shiverer animals by the procedure of Wiggins et al. (1975). The isolated myelin

J . Neurochem.. Vol. 45, N o . 6 , 1985

was lyophilized and stored in vacuo over P,O, at room temperature until use. Dry weights of whole nerves and isolated myelin were determined on a Cahn electrobalance .

X-ray diffraction Diffraction experiments were carried out with nickelfiltered and single-mirror focused CuKa radiation from a fine-line source on a Rigaku generator (RigakulUSA, Danvers, MA, U.S.A.) operated at 40kV, 20mA. The diffraction patterns were recorded with Kodak NoScreen or Direct Exposure, or with CEA Reflex 25 (CEA America, Greenwich, CT, U.S.A.) x-ray films. Exposure times were for one hour except where indicated differently. The specimen-to-film distance was 208 mm. The spacings of the diffraction spectra were measured directly off the x-ray films viewed at 6x magnification. The optical densities of the films were determined on an Optronics Photoscan P-1000 microdensitometer (with a 25 p m raster) interfaced with a VAX 11/780 computer. Calculations of the electron density profiles of the membranes and measurement of their center-to-center separations across cytoplasmic and extracellular appositions were carried out as previously described (Inouye and Kirschner, 1984).

Lipid extraction and analysis Extraction of nerve lipids was carried out by homogenizing the tissue in ground-glass homogenizers with chloroform-methanol (2: 1, vol/vol) to which 0.9% KC1 in water had been added to make up 5% of the final volume (Folch et al., 1957). The homogenates were incubated 30 min at room temperature, methanol was added in the amount of 20% of the original volume, and the samples were centrifuged to yield a pellet and a supernatant containing the lipid extract. The pellet was treated again with the same quantities of solutions as for the original whole nerve sample. The combined supernatants gave the total lipid extract. The volume of each solution was adjusted so that the total lipid extract was 0.5 ml/mg dry weight of tissue. For isolated myelin a similar procedure was used to extract the lipids. Partitioning the extract into two phases (Folch et al., 1957) was unnecessary either for the separation or the quantitation of the lipids; no lipid-like material was present in upper phases of trial partitionings. Butylated hydroxytoluene was added to the solvents as an antioxidant. To avoid loss of lipids on the glass surfaces, all glassware in contact with the lipids was silylated with hexamethyldisilazane (Sigma) by a 180°C vapor phase procedure (Fenimore et al., 1976). Lipids were separated into classes by one-dimensional high performance thin layer chromatography. From the total lipid extracts, polar lipids were separated with a mixture of methyl acetate, 1-propanol, choloroform, methanol, 0.25% KCI, and acetic acid (25:25:25: 10:9:0.3) (modified from Vitiello and Zanetta, 1978). Neutral lipids were separated first with chloroform, methanol and acetic acid (93.1 :I .9:0.1) and second with hexane, ethyl ether, and acetic acid (89.35.7:O.l) (modified from Bitman et al., 1981). Separation of the lipids was improved and made more reproducible by varying the amount of acetic acid relative to that used in the original published chromatographic solvents. Whatman LHP-K and Merck Silica Gel 60 plates (10 x 20 cm) with preabsorbent zones were used. The extracts from normal and mutant animals

MEMBRANE INTERACTIONS AND LIPIDS IN SHIVERER MYELIN were applied to the same plate in duplicate in I-, 2-, 3-, and 4- or 5-pI aliquots for polar lipids and in 0.5,I-, 2-, and 3-1.1.1aliquots for neutral lipids. Lipids were visualized by spraying the plates to wetness with CuS0,-H,PO, and charring in an oven on an aluminum plate at 170°C for 4-5 min (Touchstone et al., 1981). For lipid quantitation the plates were photographed with Tech Pan 4 x 5 film using transmitted light, and the negatives were scanned with an Optronics P-1000 microdensitometer. Multiple scans covering the entire width of each lane of lipids were averaged. The areas under the lipid peaks were numerically integrated after computer-aided background subtraction. The degree of separation of lipids is evident from the densitometer tracings (Fig. 7). Linear regression lines for plots of peak area versus extract volume were computed (Bevington, 1969). The amount of a lipid in shiverer relative to normal was obtained by dividing the slope of the shiverer regression line by that of the normal. The uncertainty in the quotient for each lipid was calculated knowing the uncertainty in each regression line slope. When assays were repeated two or three times, the mean quotient and the mean uncertainty in the quotient were weighted according to the uncertainty of the individual quotients and individual un-

+/+

I913

certainties (Bevington, 1969). The individual or mean uncertainties are plotted as error bars (see Fig. 8). These methods have been used to separate and analyze the following lipids: the polar lipids galactosyldiglyceride, cerebrosides with nonhydroxy and hydroxy fatty acids, sulfatides with nonhydroxy and hydroxy fatty acids, ethanolamine phosphatides, phosphatidic acid and cardiolipin as a single band, phosphatidylinositol, phosphatidylserine, phosphatidylcholine, sphingomyelin, and lysophospholipids; and the nonpolar lipids cholesterol ester, triglyceride, diglycerides (the 1,3 and 1,2 isomers migrate as separate bands), cholesterol, monoglyceride, and free fatty acids. The identity of the extracted lipids was verified by comigration with commercial standard lipids. Phosphatidylinositol was such a minor lipid in the extracts from nerve and isolated myelin that we did not quantitate it. Sulfatide assay. The assay of Kean (1968), which detects sulfatide colorimetrically by its binding to and partitioning with azure A, was used to determine the level of sulfatide in total lipid extracts from pooled sciatic nerves and isolated myelin of normal and shiverer mice (see above). Kean determined that the presence of a 10fold excess of lipids that are present in myelin does not significantly affect the assay (Kean, 1968). The small amounts of our extracts required that the original assay be scaled down to one-half. The determinations were made on duplicate aliquots of 5 , 10, 20, 30, and 40 PI of each extract. The optical densities of the samples were measured at 645 nm on a Perkin-Elmer spectrophotometer and found to be linear with sample volume. Linear regression lines for plots of optical density versus extract volume were computed, and the relative amount of sulfatide in shiverer compared to normal was obtained by dividing the slope of the shiverer regression line by that of the normal. The uncertainty in the quotient for each lipid was calculated knowing the uncertainty in each regression line slope.

RESULTS

center Distance along film

of Patiern

.

-

Diffraction order

FIG. 1. Densitometer scans of x-ray diffraction patterns from normal (upper) and shiverer (lower) sciatic nerves. The optical density of the x-ray films is plotted as a function of the distance along the film. The orders of diffraction are indicated along the bottom of the lower scan. The second orders were overexposed, so the tops of their peaks were left open. Comparing the scans, note that the fourth orders have approximately the same intensity; however, whereas the third and fifth orders are weaker in the shiverer, the first order is increased. The magnitude of the changes in intensity in the third and fifth orders (which are normally strong) indicates that the phases of these reflections are the same in shiverer as in normal. In contrast, the large change in the intensity of the first order which is normally weak suggests that its phase may be opposite to that in normal myelin. This possibility prompted our reexamination of the phase assignment for the first-order x-ray reflection of shiverer myelin.

Structural difference between shiverer and normal PNS myelin The ?-ray patterns that we previously recorded to 15 A spacing from unfixed sciatic nerves from shiverer mice were interpreted on the basis of their overall similarity to the patterns from normal nerve (Kirschner and Ganser, 1980). Our analysis indicated that basic protein was localized in the cytoplasmic region of the membrane. However, in view of the altered extracellular interactions in shiverer (see below) and the fact that the electron density levels in the extracellular and cytoplasmic regions depend crucially on the magnitude and sign (phase) of the first-order harmonic IF(l)I,which is threefold more intense in shiverer than in normal (Fig. l ) , we now reexamined the phase assignment of IF(1)I in shiverer. Analysis of the diffraction data from swollen shiverer myelin provides a method of establishing the correct phase for the first-order harmonic. Each of the two possible structures for unswollen shiverer myelin [calculated by choosing - F(1) or

J . Neurochem., Vol. 45, N o . 6,1985

H . INOUYE ET AL.

I914

+ F ( I ) ; Fig. 2, top] corresponds to a different continuous transform (Fig. 2, bottom). The greatest difference between these transforms is that the - F( 1) tdansform passes through zero twice at R < 0.01 A - l , whereas the + F ( 1 ) transform never passes through zero in this region. The diffraction data measured from shiverer myelin swollen to different extents in distilled water is expected to map on the continuous transform which more closely corresponds to the correct structure. We found that the third ordcr from the swollen shiverer myelin with the 336 A period (arrows in Fig. 2, bottom) was

zero, very close to one of the expected zeroes in the calculated transform. This firmly establishes that the sign of IF( I)/ for shiverer is negative, which substantiates our previous interpretation of the x-ray pattern (Fig. 2, top left). Swelling of shiverer myelin in water We measured the period of myelin in nerves after different treatment times in distilled water (Fig. 3). Within the time resolution of our protocol, the tran180 A) to sition from the native period q a y s (d the swollen arrays (d > 220 A) was discrete, i.e.,

-

shi/shi cyloplasmr

0

extracellular boundary

- F ( 1)

boundary

20

40

A

a0

Bo

0

40

20

i

Bo

80

o o .02A-’ R

FIG. 2. Confirmation of choice of membrane profile for shiverer peripheral myelin. The upper profiles indicate the two possible membrane structures for shiverer using -F(l) (left) or +F(l)(right) in the Fourier synthesis. The difference profiles, calculated

by subtracting the shiverer profile (heavy curve) from the normal one (light curve), show the possible localization of differences in structure between shiverer and normal myelin membranes. Analysis of the diffraction data from swollen shiverer mye\in enabled us to choose between the two possibilities. The lower left plate shows a diffraction pattern from swollen shiverer sciatic nerve myelin (in distilled water) with a 336 A period; note the absence of the third-order reflections (arrows). The lower right plot shows that the continuous transform calculated from a swollen membrane pair with -F(1) (solid curve) passes through zero twice for R < 0.01 k’. In particular, the third order from the 336 A experimental data is in accord with this transform (arrow). Therefore, the correct profile for shiverer is the one shown in the upper left, and previously chosen (Kirschner and Ganser, 1980). [The transforms were calculated by first estimating a value for F(0) from the data of Blaurock (1971) for myelin swollen in water and then repacking the membranes at their cytoplasmic apposition (Kirschner and Ganser, 1982) to minimize the difference between the structure factors calculated for the swollen lattice and the structure amplitudes measured for swollen shiverer myelin. F(0) is related to the difference between the average electron density of the membrane pair and that of the fluid between them (Worthington and Blaurock, 1969a).]

J . Neurochem., Vol. 45, No. 6 , 1985

MEMBRANE INTERACTIONS A N D LIPIDS IN SHIVERER MYELIN

1801,

0

,

,

,

,

,

5

10

,

,

1915

,

20 Time (h)

FIG. 3. Swelling response of shiverer (shilshi) and normal (+/+) sciatic nerve myelin to treatment with distilled water. Ordinate: myelin period measured from x-ray diffraction patterns. Abscissa: treatment time for whole nerve in bulk distilled water. The diffraction patterns were recorded for 1-4 h. The curves were the visual best fit of the data.

native and swollen arrays did not coexist and no intermediate phases were observed. For both normal and shiverer swollen myelins, the altered intensities of the reflections indicated that the extracellular apposition between membranes was the site of swelling (Finean and Burge, 1963; Blaurock, 1971; Kirschner and Hollingshead, 1980; see note added in proof). With prolonged exposure to distilled water over the course of 24 h, !he myelinsontinued to swell gradu?lly, from 2Z!O A to -270 A for normal and from 240 A to -330 A for shiverer. Most striking, the period of shiverer myelin was consistently greater than that of normal at all times. Effect of pH In preliminary x-ray experiments normal and shiverer sciatic nerve myelin showed different structural responses to changes in pH between 4 and 10 in normal saline (Kirschner and Inouye, 1984). The electron density profiles showed that the separation of membranes across the cytoplasmis apposition in normal myelin increased by -5 A from pH 4 tp 10 and that this space for shiverer was always -2 A wider than normal. Another difference between normal and shiverer myelin was particularly evident at pH 5 , whqe the extracellular space in shiverer expanded 6 A more than normal. To study this pH-dependent difference between normal and shiverer myelins in more detail, we enhanced the interactions between the ionizable groups on the apposed membrane surfaces by reducing the concentration of counterions from the physiological ionic strength of 0.16 to 0.06 and then recorded x-ray diffraction patterns (Figs. 4 and 5). At pH >5 the myelin arrays in both normal and shiverer were swollen (Figs. 4c and 5). The x-ray reflections were broad, indicating considerable disorder in the membrane packing, and the intensity variation of the reflections (enhanced third and fifth orders) indicated that the membrane pairs swelled apart at their extracellular appositions. The density profiles showed that the center-to-center separation

FIG. 4. X-ray diffraction patterns from mouse sciatic nerves in low-ionic-strength (0.06) solutions at different pH values. a: pH 2 for 2 h, myelin period is 328 A; shiverer nerve. b: pH 3 for 24 h, period is 165 A; normal nerve. c: pH 8 for 24 h, period is 234 A; normal nerve. The diffraction patterns were obtained after -24 h incubation in bulk solution.

of membranes across theoextracellular apposition increased from 90 to 153 A in the range of pH 4 to 8, whereas the center-to-center separation across tbe cytoplasmic apposition increased from 74 to 81 A. At pH 2.5-4.5 (Figs. 4b and 5), where the myelin period was minimum and constant, the reflections were relatively sharp, indicating a more ordered packing of the closely apposed membranes; however, reflections higher than the fifth order were either weak or unobserved, suggesting that the close packing of membranes reduced the contrast within the structure. At pH 2 (Figs. 4a and 5 ) the prominent reflections were 1(1), I(4), and I(7). The relative intensities were sufficiently different than those we have obtained from peripheral myelin swollen to the same extent in distilled water to suggest that at pH 2 some swelling is also occurring at the cytoplasmic apposition. Assuming that the swelling involves the separation of units of two membranes [rather than the separation of units of four membranes (Worthington, 1979)], our density profiles show that the cytoplasmic space increased only about 7 A, whereas the center-to-center sepJ . Neurockem.. Vol. 45, N o . 6, 1985

H . INOUYE ET AL.

1916

9. 320

t

250 h

J

-W

L

D 0 E

2, 0

PH 7

I 200

.4

3

.06 I50

7

4

10

VH

FIG. 5. Myelin period in shiverer (0.shilshi), shiverer het-

erozygote (a), and normal (0, + I + )sciatic nerves as a function of pH at a constant ionic strength of 0.06. Treatment times were 24-30 h.

aration of membranes a u o s s the cytoplasmic apposition increased -13 A. The periods for both shiverer and norm51 rnyelins increased fr9m similar minima of -165 A at pH 4 to 250-260 A at about pH 9 (Fig. 5). The rate of increase in period with pH was greater for shiverer, and the shiverer period remained greater than normal from pH 4-9. Sciatic nerves from mice heterozygous for the shiverer mutation showed myelin periods that were intermediate between those for shiverer and normal mice. In the pH range 2.5-4SQthe periods were at a constant minimum of -165 A. Below pH 2.5 the periods increased discontinuously to around 300 A. The altered myelin at pH 2 was not reversible to the native structure by increasing the pH. The isoelectric range for the extracellular surface of peripheral myelin appears to be pH 2.5-4.5. Above and below this range the membranes swell apart at their extracellular appositions. We do not observe any separation at the cytoplasmic apposition until below pH 2.5; therefore, an isoelectric range cannot be determined for the cytoplasmic surface. Effect of ionic strength The period of myelin in shiverer and normal sciatic nerves was followed as a function of ionic strength at pH 7 and at pH 4 (Fig. 6). These measurements allowed us to probe the counterion-dependent interaction between membranes at physiological pH and at a pH at which the membranes J . Neurochem., Vol. 45, No. 6 , 1985

.I2 Ionic strength

.I0

FIG. 6 . Myelin period in shiverer (filled symbols, shiishi), shiverer heterozygote and normal (open symbols, +/ +)

(a),

sciatic nerves as a function of ionic strength at pH 7 (circular symbols) and pH 4 (square symbols). At pH 7 there were two phases in the ionic strength range of 0.10-0.15. Treatment times were 24-28 h.

exhibit their closest approach. At pH 7 and at an ionic strength of 0.12-0.14, which is slightly below the physiological value, a portion of both types of myelin swelled discontinouously from the native value of -175 A to -220 A. The rest of the myelin remained at the native period. As the ionic strength decreased from 0.14 to 0.10, a greater proportion of the myelins swelled; and while the normal myelin remained at -220 A, shiverer swelled further to -240 A. Below an ionic strength of 0.10 for normal and 0.12 for shiverer, only the swollen phases were observed. As the ionic strength decreased further to 0.02, both types of myelig continued to swell with shiverer always 20-25 A more swollen than normal. Greater swelling in shiverer at pH 7 in 0-80 mM NaCl has also been found by R. P. Rand (personal communication). Shiverer heterozygotes showed periods that were generally intermediate between the shiverer and normal values. At pH 4 over the ionic strength range studied, the periods for both shiverer and normal rnyelins were compacted to 162-168 A and were virtually indistinguishable. In addition, only single phases were observed. As the ionic strength was decreased from 0.20 to 0.01, there was a very slight decrease in periods for both types of myelin, which could result from increased dipole-dipole attraction at low ionic strength. Lipid composition of shiverer sciatic nerve and isolated myelin The increased swelling of shiverer myelin above pH 5 when the counterion concentration is reduced

MEMBRANE INTERACTIONS AND LIPIDS IN SHIVERER MYELIN might be accounted for by a greater level of charged lipids. To examine this possibility we first analyzed the lipid composition of whole sciatic nerves, which mostly reflects that of the myelin itself due to the large number of myelinated fibers and to the large proportion of lipid in myelin. Densitometer scans of the polar lipids (Fig. 7) showed that the same classes of lipids were present in normal and shiverer nerves, but that there were small differences in the amounts of some of the lipids. For example, among the negatively charged lipids, nonhydroxysulfatide was increased whereas phosphatidylserine was unchanged. Chromatography of neutral lipids indicated that only cholesterol and traces of triglyceride (most likely from fat adhering to the nerves) were present in both shiverer and normal extracts; no evidence of cholesterol ester was found. The differences between shiverer and normal lipids from whole sciatic nerve and isolated myelin were quantitated from the thin layer chromatograms (Fig. 8). Sulfatide was increased by 2625% in shiverer; both the nonhydroxy and the hydroxy forms appeared to be increased to similar extents (data not shown). Cerebroside, a glycosphingolipid related metabolically to sulfatide, was also changed in shiverer, but in this case the nonhydroxy form was reduced and the hydroxy form was increased. Among the phospholipids appreciable changes were seen in phosphatidylcholine, which was increased by 10-20% and sphingomyelin which was decreased by 20-25%. Changes in ethanolamine phosphatides, phosphatidylserine, and cholesterol were at about the 5% level. The alterations in lipid composition of shiverer whole nerve and isolated myelin were generally the same. The elevated level of sulfatide in shiverer was confirmed by use of the Kean (1968) assay. An in-

n

PE

CH

.*

1917

*

crease of 15 2% compared to normal was detected in extracts of whole sciatic nerve, and an increase of 20 2 2% was found in extracts of sciatic nerve myelin. DISCUSSION Absence of basic protein underlies alterations at cytoplasmic apposition in shiverer Our analysis of diffraction patterns from swollen shiverer myelin confirms the deficit in electron density from the cytoplasmic side of the membrane and the correlation of this deficit with the absence of basic protein. There is also a slight alteration in the interactions of shiverer membranes at their cytoplasmic apposition as revealed by measurement of membrane separations under different conditions. Previously, we obsgrved that the cytoplasmic space in shiverer is 1-2 A greater than normal in physiological saline (Kirschner and Ganser, 1980) as well as after treatment with acidic or alkaline saline (Inouye and Kirschner, 1984). Also, in alkaline saline containing 1 mM added divalent catiocs, the cytoplasmic space in shiverer is up to -4 A wider than normal (Inouye and Kirschner, 1984). What can account for the altered structure and interactions at the cytoplasmic apposition in shiverer? In addition to the absence of basic protein, we have now found changes in lipid composition of shiverer myelin. The lipids expected to be present in the cytoplasmic leaflets [i.e., cholesterol (Caspar and Kirschner, 1971), ethanolamine phosphatides (Kirschner and Ganser, 1982), and phosphatidylserine (Op den Kamp, 1979; Gwarsha et al., 1984)1, however, show only minor changes. Further, lipid is not likely to account for the deficit in material outside the headgroup region of the bilayer in the cytoplasmic space. Therefore, the absence of basic

- +/+ ...........

shi/shi

FIG. 7. Densitometer scans of chromatographed polar lipids from normal and shiverer whole sciatic nerves. The

J . Neurochem., Vol. 45, No. 6,1985

H . INOUYE ET A L .

1918 CN

CH

ST

PE

PS

PC

SM

co

I

FIG. 8. Comparison of the relative amounts of lipids in shiverer and normal PNS. Light stipple, whole sciatic nerves; dark

stipple, isolated myelin from sciatic nerves. Vertical bars represent errors of the measurements (see Materials and Methods for detailed explanation). Analyses were performed once on whole sciatic nerve and three times on isolated myelin extracts. For abbreviations see legend to Fig. 7; ST, total of nonhydroxy- and hydroxysulfatides; CO, cholesterol.

proteins rather than the changes in lipid composition is most likely responsible for the altered structure and interactions. Sulfatide and connective tissue contribute to increased extracellular swelling in shiverer When counterions are reduced or removed from peripheral myelin, the membranes swell apart at their extracellular appositions (Robertson, 1958) and attain an equilibrium separation determined by repulsive forces (hydration and electrostatic), attractive forces [van der Waals forces (Rand, 1981)], physical links between the membranes, and connective tissue constraints. We found that the swelling between extracellular surfaces is greater in shiverer than normal (Figs. 3-6; R. P. Rand, personal communication). Several explanations may account for this: a greater charge density on extracellular surfaces in shiverer resulting in greater electrostatic repulsion; weaker basement membrane or perineurium or less dense packing of endoneurial collagen in shiverer resulting in less mechanical resistance to swelling; and a change in PO conformation or self-association in the extracellular half of the membrane caused by the absence of basic proteins from the cytoplasmic side in shiverer. A greater electrostatic repulsion in shiverer may derive from its altered Lipid composition. Of the lipids that are negatively charged at neutral pH, only sulfatide is elevated. The 20-25% increase in sulfatide in shiverer myelin would increase the surface charge density (see below). Sulfatides are likely present in the external bilayer leaflet of PNS myelin membranes, since other glycolipids [cerebrosides in CNS myelin (Linington and Rumsby, 1980) and membrane glycolipids in general (Op den J . Neurorhem., Vol. 45. N o . 6 , 1985

Kamp, 1979)l are found in this leaflet. Therefore, the higher sulfatide level in shiverer myelin could contribute to its greater swelling. Connective tissue has been shown to restrict the swelling of peripheral myelin beyond a certain extent (Rand et al., 1979). For exaomple, frog myelin swells only to a period of -240 A unless the nerve is preincubated in collagenase, in which case it swelis indefinitely (Rand et al., 1979). Below this 240 A limit the connective tissue does not affect the amount of swelling. The myelin in normal mouse peripheral nerve with its connectiye tissue intact swells to a limiting &xtent of -270 A but in shiverer it swells to -330 A (Fig. 3). This appreciable difference in the maximum amount of swelling between shiverer and normal is likely due to a difference in the constraint of the connective tissue. Connective tissue does not appear to affect swelling below the swelling limit (Rand et al., 1979). The intermediate equilibrium states between the norm91 period and the m%ximum swollen period (270 A in normal, and 330 A in shiverer) must then be determined by van der Waals attractive forces (Rand, 1981) or physical links between apposed surfaces which counter the hydration and electrostatic repulsive forces. The greater swelling in shiverer in this intermediate range may be accounted for by greater electrostatic repulsion, weaker attractive forces or fewer physical links (e.g., between PO molecules). Increase in net charge at the extracellular surfaces of shiverer myelin We estimated the ratio of the surface charge density between shiverer and normal myelins in two

MEMBRANE INTERACTIONS AND LIPIDS IN SHIVERER MYELIN ways: (1) by comparing the difference in lipid composition between the two myelins (Fig. 8) and (2) by applying electric double-layer theory (Rand, 1981) to the period versus ionic strength data at pH 7 for the two myelins (Fig. 6). Our first estimate of the increase in charge density in shiverer depends on assuming that all the sulfatide is located in the extracellular half of the membrane and that the 20-25% increase in sulfatide is compensated for by a decrease of the zwitterionic or neutral lipids. Considering the lipids alone, the surface charge of the shiverer myelin lipids is therefore -1.2 times larger than that of the normal myelin lipid at pH 7 (where the zwitterionic lipids are neutral). However, sulfatide normally contributes only -60% of the net negative surface charge (with PO contributing the rest; see below). Therefore, the 20-25% increase in sulfatide in shiverer amounts only to -10% increase in surface charge density. Our second estimate of the increase in surface charge density in shiverer comes from a comparison of the swelling of normal and shiverer myelin in response to changes in ionic strength. Below the swelling limit, where the swollen period depends on the ionic strength at pH 7 (Fig. 6), the relation is like that for model charged lipid bilayer and water systems (reviewed by Rand, 1981). As the ionic strength increases, the repulsive electrostatic force decreases exponentially due to screening of the fixed surface charge by mobile counterions, and the membrane (or lipid bilayer) separation decreases. If the factors countering the swelling (i.e., van der Waals attractive forces and physical links between the membranes) are t h e same and constant for normal and shiverer nerves, then the ratio of surface charge densities for shiverer to normal can be expressed as exp[a(DShilshi - D+/+)],where D is the slope of the linear relationship between myelin period and (ionic strength)-''2 and a is a constant. The quantity (ionic strength)-'12 is proportional to the Debye length or width of the diffuse double layer. From our data at pH 7 (Fig. 6) we calculate the surface charge density of the shiverer to be 70% (?40%) greater than normal. This value is appreciably higher than the 10% increase estimated from the altered lipid composition. One explanation for the difference in estimations of the surface charge density ratios is that myelin components other than sulfatide contribute to a greater surface charge in shiverer myelin. For example, the extracellular projection of PO glycoprotein may have a greater net negative charge in shiverer, perhaps due to an altered posttranslational modification. Also, the absence of basic proteins from the cytoplasmic side of the shiverer membrane might affect the composition of lipids and the lipidto-protein ratio, or even the conformation of PO in the extracellular side. Another explanation is that

1919

the factors that counter the electrostatic repulsion (i.e., van der Waals attractive forces and physical links between the membranes) are less effective in shiverer myelin.

Surface charge and swelling: roles of PO The two types of membrane appositions in peripheral myelin respond very differently to hypotonic treatments-the extracellular surfaces readily swell apart, whereas the cytoplasmic surfaces remain together (Robertson, 1958). Can this difference be explained simply on the basis of the presence or absence of electrostatic repulsion between the two types of apposed membrane surfaces? To address this question, we have estimated the amount of charge on the extracellular and cytoplasmic surfaces from available compositional and structural data. Extracellular surfaces. From published compositions of rodent PNS myelin (Smith and Curtis, 1979; Greenfield et al., 1980; Carson et al., 1983; Norton and Cammer, 1984) we estimate that there are about one PO glycoprotein and one basic protein per 200 lipid molecules. [P2 protein is present at about one-tenth the level of basic protein (Greenfield et al., 1980), and is not considered further in our model.] Since PO is believed to be a transmembrane protein (Kirschner et al., 1984; Lemke and Axel, 1985), there would be about 100 lipid molecules associated with each of the extracellular and the cytoplasmic domains of PO. Our previous study on the interaction of ZnC1, with normal and shiverer peripheral myelins suggests that the extracellular domain of PO is acidic while the cytoplasmic domain is basic (Inouye and Kirschner, 1984). This is now confirmed in detail by the amino acid sequence deduced from the nucleotide sequence of cDNA for PO (Lemke and Axel, 1985). From the amino acid composition of the proposed extracellular domain of PO (Lemke and Axel, 1985), and the sialic acid and sulfate groups in this domain (Lees and Brostoff, 1984), the extracellular portion of PO has a net charge of about - 5 at neutral pH. The localization of sulfatides to the extracellular leaflet, as suggested by our study, would add an additional - 8 charges per PO molecule. Therefore, the observed swelling between extracellular surfaces of myelin at neutral pH when counterions are reduced is consistent with a net negative surface charge provided by both sulfatide and PO glycoprotein. Roles for PO at the extracellular apposition derive from the facts that PO is the principal protein present at these apposed membrane surfaces and is the only compoGent that could extend appreciably into the 40-50 A space between the membranes. One role of PO at this apposition is to maintain a certain amount of membrane separation most likely by steric hindrance and hydration forces (Kirschner et al., 1979; Melchior et al., 1979; Rand, 1979; J . Neurochem., Vol. 45, No. 6 , 1985

1920

H . INOUYE ET A L .

Ganser and Kirschner, 1980; Blaurock, 1981; Hollingshead et al., 1981). Another role for PO at the extracellular apposition is to maintain the close apposition of membranes. Evidence that the membranes are adherent derives from our data on the dependence of myelin period on ionic strength (Fig. 6). These data show that a portion of the myelin maintains the native period notwithstanding the development of electrostatic repulsion as the ionic strength is reduced to about 0.10 at pH 7. Evidence that the membrane adhesion may be dependent on divalent cations comes from swelling experiments on frog sciatic myelin (Worthington and Blaurock, 19696; Padron et al., 1979). Our present study shows that there is no apparent adhesion at lywer ionic strengths and beyon+ periods of -220 A tp the limit of swelling (270 A for normal and 330 A for shiverer). In this swelling range (i.e. >220 the equilibrium periods must be determined by a balance between van der Waals attraction and electrostatic repulsion. This balance is altered in shiverer, since shiverer always swells more than normal. The PO mediated adhesion at the extracellular apposition is relatively weak compared to that at the cytoplasmic apposition, since the adhesion is lost by simply reducing the ionic strength. Cytoplasmic surfaces. We have estimated the net charge at the cytoplasmic apposition. PO is likely to contribute a net charge of + 16 (Lemke and Axel, 1985) while phosphatidylserine, which is thought to be localized mainly to the cytoplasmic leaflet (Gwarsha et al., 1984), would contribute at most a charge of - 18 per PO molecule. Basic protein would contribute a net charge of about +2 0 (Carnegie and Moore, 1980; Carson et al., 1983). Thus, at neutral pH and low ionic strength, there should be a very large electrostatic repulsion. The expected swelling is not observed under these conditions, and this suggests that the cytoplasmic surfaces are tightly linked together. Is myelin basic protein or PO glycoprotein, or both, providing the adhesion? Observations on shiverer peripheral myelin indicate that PO glycoprotein is the link between the cytoplasmic surfaces. In the absence of basic protein the cytoplasmic surfaces would lack a significant net charge at neutral pH (assuming the same lipid asymmetries in shiverer as in normal). Up to pH 9, the basic groups on phosphatidylserine (Cevc et al., 1981) and on PO continue to be positively charged due to their high pK values, so that the cytoplasmic surfaces still have neither a net charge nor an electrostatic repulsion which could produce swelling. Around pH 4 the carboxyl groups on phosphatidylserine (Cevc et al., 1981) and on PO become uncharged resulting in a net positive surface charge. Moreover, from pH 4 down to pH 2 the phospholipids on the cytoplasmic surface become positively charged (Hauser and Phillips,

A)

J . Neurochem., Vol. 45, N o . 6 , 1985

1979). Therefore, below pH 4 the large net positive charge that develops on the cytoplasmic surfaces should lead to swelling. However, cytoplasmic swelling is not observed until below pH 2.5, aad this swelling is relatively small (only about 5-10 A), unlike the large swelling thFt occurs at the extracellular apposition (90-150 A). This stability of the cytoplasmic apposition in the absence of basic protein in shiverer must be due to linkages that involve the PO molecules on the apposed membranes. The short range forces that underlie this linkage must be so strong that even when basic protein is present in normal PNS myelin and is contributing a very large net positive charge at neutral pH (see above), there is still little or no cytoplasmic swelling at low ionic strength and very acidic pH.

Expression of the shiverer mutation The shiverer mutation is a large deletion in the myelin basic protein region of the genome (Roach et al., 1983; Kimura et al., 1985), and it results in a major deficit of this protein in both the CNS and PNS (Dupouey et al., 1979; Kirschner and Ganser, 1980; Mikoshiba et al., 1981; Winter, 1982). What is the consequence of this deficit? In the CNS the most obvious result is the lack of myelin (Bird et al., 1978). In addition, morphological abnormalities in the small amount of myelin that does form (Privat et al., 1979; Ganser and Kirschner, 1980; Rosenbluth, 19806) and in the oligodendrocytes (Inoue et al., 1983; Billings-Gagliardi and Wolf, 1984) have been observed. In the PNS the amount of myelin and its gross structure are virtually normal; however, subtle structural abnormalities have been observed (Ganser and Kirschner, 1980; Kirschner and Ganser, 1980; Rosenbluth, 1 9 8 0 ~Mikoshiba ; et al., 1981; Inouye and Kirschner, 1984; Peterson and Bray, 1984). Our present study demonstrates a number of additional abnormalities in shiverer PNS myelin. These include alterations in the lipid composition and in the interactions at the extracellular apposition between membranes. The altered membrane interactions are likely due in part to the increased sulfatide level in shiverer. In addition, changes in the PO glycoprotein (i.e., its charge, conformation, or disposition in the membrane), changes in lipid asymmetries, and/or alterations in the connective tissue of the nerve must also underlie the altered interactions at the extracellular apposition in shiverer myelin. The relationship of these numerous phenotypic abnormalities in shiverer to the lesion in its genome remains to be established. Note added in proof: Using electron microscopy, we determined more directly the site of swelling in shiverer myelin. Normal and shiverer sciatic nerves were incubated in 10 mM sodium cacodylate (pH 7.4) for 2-3 h at room temperature, and subsequently fixed in a mixture of 1.5% glutaraldehyde-1.0%OsO, in 10 mM cacodylate

MEMBRANE INTERACTIONS AND LIPIDS IN SHIVERER MYELIN at WC, and processed for thin sectioning. In both normal and shiverer myelin, membrane separation was observed only at the extracellular apposition. This confirms our interpretation of the x-ray diffraction patterns. Acknowledgment: We thank Dr. M. K. Wolf (Department of Anatomy, University of Massachusetts Medical School, Worchester, MA, U.S.A.) for providing some of the mutant and control mice used in this study, Betty Jane Brown for assistance with the lipid analysis, and William P. McIntosh for expert photographic services. We also thank Drs. G. Lemke and R. Axel for sending a preprint of their work. The research was supported by NIH Grants NS 20824 (to D.A.K.) and NS 11237 (to Dr. R. L. Sidman) from the National Institute of Neurological and Communicative Disorders and Stroke, a research grant from the Scottish Rite Schizophrenia Research Program, N.M. J., U.S.A. (D.A.K.), and a bridging grant from the Burroughs Wellcome Fund (D.A.K.). The work was carried out in facilities of the Mental Retardation Research Center at Children’s Hospital, and was supported by NIH Core Grant HD 06276.

1921

logical Mutations Affecting Myelination (Baumann N., ed), pp. 171-176. Elsevier/North Holland, Amsterdam. Greenfield S., Brostoff S . W., and Hogan E. L. (1980) Characterization of the basic proteins from rodent peripheral nervous system myelin. J . Neurochem. 34, 453455. Gwarsha K., Rumsby M. G., and Little C. (1984) On the disposition of phospholipids in freshly isolated myelin sheath preparations from bovine brain. Neurochem. I n t . 6, 599-606.

Hauser H. and Phillips M. C. (1979) Interactions of the polar groups of phospholipid bilayer membranes. Prog. Surf. Membr. Sci. 13, 297-413. Hollingshead C. J., Caspar D. L. D., Melchior V., and Kirschner D. A. (1981) Compaction and particle segregation in myelin membrane arrays. J . Cell B i d . 89, 631-644. Inoue Y., Inoue K., Terashima T., Mikoshiba K., and Tsukada Y. (1983) Developmental changes of oligodendroglia in the posterior funiculus of “shiverer” mutant mouse spinal cord, with special reference to myelin formation. Anat. Embryo/. 168, 159-171. Inouye H. and Kirschner D. A. (1984) Effects of ZnC1, on membrane interactions in myelin of normal and shiverer mice. Biochim. Biophys. Acta 776, 197-208. Kean E. L. (1968) Rapid, sensitive spectrophotornetric method for quantitative determination of sulfatides. J . Lipid Res. 9, 319-327.

REFERENCES Bevington P. R. (1969) Data Reduction and Error Analysis for the Physical Sciences. McGraw-Hill, New York. Billings-Gagliardi S . and Wolf M. K. (1984) Shiverer and its allele, myelin-deficient: alterations of morphology with advancing age. Soc. Neurosci. Abstr. 10 (I), 84. Bird T. D., Farrell D. F., and Sumi S. M. (1978) Brain lipid composition of the shiverer mouse: (genetic defect In myelin development). J . Neurochem. 31, 387-391. Bitman J., Wood D. L., and Ruth J. M. (1981) Two-stage, onedimensional thin layer chromatographic method for separation of lipid classes. J . Liquid Chromatogr. 4, 1007-1021. Blaurock A. E. (1971) Structure of the nerve myelin membrane: proof of the low resolution profile. J . Mol. Biol. 56, 35-52. Blaurock A. E. (1981) The spaces between membrane bilayers within PNS myelin as characterized by X-ray diffraction. Brain Res. 210, 383-387. Carnegie P. R. and Moore W. J. (1980) Myelin basic protein, in Proteins of the Nervous System, 2nd edit (Bradshaw R. A. and Schneider D. M., eds), pp. 119-143; Raven Press, New York. Carson J. H., Nielson M. L., and Barbarese E. (1983) Developmental regulation of myelin basic protein expression in mouse brain. Dev. B i d . 96, 485492. Caspar D. L. D. and Kirschner D. A. (1971) Myelin membrane structure at 10 8, resolution. Nature 231, 46-52. Cevc G., Watts A,, and Marsh D. (1981) Titration of the phase transition of phosphatidylserine bilayer membranes: effects of pH, surface electrostatics, ion binding, and head-group hydration. Biochemistry 20, 49554965. Dupouey P., Jacque C., Bourre J. M., Cesselin F., Privat A., and Baumann N. (1979) Immunochemical studies of myelin basic protein in shiverer mouse devoid of major dense line of myelin. Neurosci. Lett. 12, 113-1 18. Fenimore D. C., Davis C. M., Whitford J. H., and Hamngton C. A. (1976) Vapor phase silylation of laboratory glassware. Anal. Chem. 48, 2289-2290. Finean J. B. and Burge R. E. (1963) The determination of the Fourier transform of the myelin layer from a study of swelling phenomena. J . Mol. Biol. 7, 672-682. Folch J . , Lees M., and Sloane-Stanley G. H. (1957) A simple method for the isolation and purification of total lipids from animal tissues. J . Biol. Chem. 226, 497-509. Ganser A. L. and Kirschner D. A. (1980) Myelin structure in the absence of basic protein in the shiverer mouse, in Neuro-

Kimura M., Inoko H., Katsuki M., Ando A., Sato T., Hirose T., Takashima H., Inayama S., Okano H., Takamatsu K., Mikoshiba K., Tsukada Y., and Watanabe 1. (1985) Molecular genetic analysis of myelin-deficient mice: shiverer mutant mice show deletion of gene(s) coding for myelin basic protein. J. Neurochem. 44, 692-696. Kirschner D. A. and Ganser A. L. (1980) Compact myelin exists in the absence of basic protein in the shiverer mutant mouse. Nature 283, 207-210. Kirschner D. A. and Ganser A. L. (1982) Myelin labeled with mercuric chloride: asymmetric localization of phosphatidylethanolamine plasmalogen. J . Mol. Biol. 157, 635658. Kirschner D. A. and Hollingshead C. J. (1980) Processing for electron microscopy alters membrane structure and packing in myelin. J . Ultrastruct. Res. 73, 211-232. Kirschner D. A. and Inouye H. (1984) Altered membrane interactions in shiverer peripheral myelin. Trans. Am. Soc. Neurochem. 15, 285. Kirschner D. A., Hollingshead C. J., Thaxton C., Caspar D. L. D., and Goodenough D. A. (1979) Structural states of myelin observed by X-ray diffraction and freeze-fracture electron microscopy. J . Cell Biol. 82, 140-149. Kirschner D. A., Ganser A. L., and Caspar D. L. D. (1984) Diffraction studies of molecular organization and membrane interactions in myelin, in Myelin (Morell P., ed), pp. 51-95, Plenum Press, New York. Lees M. B. and Brostoff S. W. (1984) Proteins of myelin, in Myelin (Morell P., ed), pp. 197-224, Plenum Press, New York. Lemke G. and Axel R. (1985) Isolation and sequence of a cDNA encoding the major structural protein of peripheral myelin. Cell 40, 501-508. Linington C. and Rumsby M. G. (1980) Accessibility of galactosyl ceramides to probe reagents in central nervous system myelin. J . Neurochem. 35, 983-992. Melchior V., Hollingshead C. J., and Caspar D. L. D. (1979) Divalent cations cooperatively stabilize close membrane contacts in myelin. Biochim. Biophys. Acta 554, 20rC226. Mikoshiba K., Kohsaka S., Takamatsu K., and Tsukada Y. (1981) Neurochemical and morphological studies on the myelin of peripheral nervous system from shiverer mutant mice: absence of basic proteins common to central nervous system. Brain Res. 204, 455460. Norton W. T. and Cammer W. (1984) Isolation and characterization of myelin, in Myelin (Morell P., ed), pp. 147-195, Plenum Press, New York.

J . Neurochem., Val. 45, No.6, 198s

I922

H . INOUYE ET A L .

Op den Kamp J. A. F. (1979) Lipid asymmetry in membranes. Annu. Rev. Biochem. 48, 47-71. Padron R., Mateu L . , and Kirschner D. A. (1979) X-ray diffraction study of the kinetics of myelin lattice swelling: effect of divalent cations. Biophys. J . 28, 231-240. Peterson A. C. and Bray G. M. (1984) Hypomyelination in the peripheral nervous system of shiverer mice and shiverer normal chimaera. J . Comp. Neurol. 227, 348-356. Privat A., Jacque C., Bourre J. M., Dupouey P., and Baumann N. (1979) Absence of the major dense line in myelin of the mutant mouse shiverer. Neurosci. L e f t . 12, 107-1 12. Rand R. P. (1981) Interacting phospholipid bilayers: measured forces and induced structural changes. Annu. Rey. Biophys. Bioeng. 10, 277-314. Rand R. P., Fuller N. L., and Lis L . J. (1979) Myelin swelling and measurement of forces between myelin membranes. Narure 279, 258-260. Roach A., Boylan K., Horvath S . , Prusiner S . B., and Hood L. E. (1983) Characterization of cloned cDNA representing rat myelin basic protein: absence of expression in brain of shiverer mutant mice. Cell 34, 799-806. Robertson J. D. (1958) Structural alterations in nerve fibers produced by hypotonic and hypertonic solutions. 1.Biophys. Biochem. Cytol. 4, 349-364.

J . Neurochem., Vol. 45, N o . 6 , 1985

Rosenbluth J. (1980~)Peripheral myelin in the mouse mutant shiverer. J . Comp. Neurol. 193, 729-739. Rosenbluth J. (1980b) Central myelin in the mouse mutant shiverer. J . Comp. Neurol. 194, 639-648. Smith M. E. and Curtis B. M. (1979) Frog sciatic nerve myelin: a chemical characterization. J . Neurochem. 33, 447452. Touchstone J. C., Levin S. S . , Dobbins M. F., and Carter P. J. (1981) Differentiation of saturated and unsaturated phospholipids on thin layer chromatograms. J . High Res. Chromatogr. Col. Chromatogr. 4, 423424. Vitiello E. and Zanetta J.-P. (1978) Thin-layer chromatography of phospholipids. J . Chromatogr. 166, 637-640. Wiggins R. C., Benjamins J. A , , and Morel1 P. (1975) Appearance of myelin proteins in rat sciatic nerve during development. Brain Res. 89, 99-106. Winter J. (1982) Shiverer peripheral myelin contains P2. Nature 298, 471472. Worthington C. R. (1979) An X-ray study of the pH property of frog sciatic nerve. In?. J . Biol. Macromol 1, 157-164. Worthington C. R. and Blaurock A. E. (1969~)A structural analysis of nerve myelin. Biophys. J . 9, 970-990. Worthington C. R . and Blaurock A. E. (1969h) A low-angle X-ray diffraction study of the swelling behavior of peripheral nerve myelin. Biochim. Biophys. Acfa 173, 427435.