Structure and Stability of Internodal Myelin in Mouse Models of Hereditary Neuropathy

Abstract

Peripheral neuropathies often result in abnormalities in the structure of internodal myelin, including changes in period and membrane packing, as observed by electron microscopy (EM). Mutations in the gene that encodes the major adhesive structural protein of internodal myelin in the peripheral nervous system of humans and mice-P0 glycoprotein-correlate with these defects. The mechanisms by which P0 mutations interfere with myelin packing and stability are not well understood and cannot be provided by EM studies that give static and qualitative information on fixed material. To gain insights into the pathogenesis of mutant P0, we used x-ray diffraction, which can detect more subtle and dynamic changes in native myelin, to investigate myelin structure in sciatic nerves from murine models of hereditary neuropathies. We used mice with disruption of one or both copies of the P0 gene (models of Charcot-Marie-Tooth-like neuropathy [CMT1B] or Dejerine-Sottas-like neuropathy) and mice with a CMT1B resulting from a transgene encoding P0 with an amino terminal myc-tag. To directly test the structural role of P0, we also examined a mouse that expresses P0 instead of proteolipid protein in central nervous system myelin. To link our findings on unfixed nerves with EM results, we analyzed x-ray patterns from unembedded, aldehyde-fixed nerves and from plastic-embedded nerves. From the x-ray patterns recorded from whole nerves, we assessed the amount of myelin and its quality (i.e. relative thickness and regularity). Among sciatic nerves having different levels of P0, we found that unfixed nerves and, to a lesser extent, fixed but unembedded nerves gave diffraction patterns of sufficient quality to distinguish periods, sometimes differing by a few Å. Certain packing abnormalities were preserved qualitatively by aldehyde fixation, and the relative amount and structural integrity of myelin among nerves could be distinguished. Measurements from the same nerve over time showed that the amount of P0 affected myelin's stability against swelling, thus directly supporting the hypothesis that packing defects underlie instability in “live” or intact myelin. Our findings demonstrate that diffraction can provide a quantitative basis for understanding, at a molecular level, the membrane packing defects that occur in internodal myelin in demyelinating peripheral neuropathies.

Electron microscopy, Membrane interactions, myc epitope tag, Myelin, P0 glycoprotein, P0 null, Proteolipid protein, PLP null, X-ray diffraction

Introduction

Transgenic (tg) mice are now available in which the presence, absence, or amounts of myelin proteins can be regulated. Such tg mice have been developed, for example, to elucidate the roles in myelin structure of the major adhesion proteins P0 glycoprotein and proteolipid protein (PLP) (1-6) (Yin et al, unpublished observations), myelin membrane interactions and paranodal axon-glia interactions (7), sensory deficits (8), gene expression and morphogenesis (9), and axonal degeneration (10) (Yin et al, unpublished observations). These tg mice include PLP^null (or PLP⁻), P0 heterozygous null (P0^+/−), P0^null (or P0⁻), and myc-tag P0 (P0^myc/P0^+/+), which contains 13 extra residues at the N-terminus of the normal P0 protein (4). The morphologic similarities between certain of these tg mice and biopsies from patients having Charcot-Marie-Tooth peripheral neuropathy (4) and Pelizaeus-Merzbacher disease (5) demonstrate that such mice can be used as model systems for elucidating human demyelinating disorders. Electron microscopy (EM) of thin sections of sciatic and optic nerves from P0 null or PLP null mice show, surprisingly, the presence of multilamellar membrane arrays, but the stacking of myelin lamellae is disordered (i.e. less regular) and the extracellular apposition of the intraperiod line is often widened (4, 5). Packing defects of myelin are also apparent in EMs from peripheral nerve containing myc-tag P0 and resemble subtypes of Charcot-Marie-Tooth with hypomyelination and altered intraperiod lines (4, 5).

Electron microscopy provides an essential description of myelin morphology, that is, the relation among the sheath's differentiated regions (internode, paranode, node); however, owing to chemical fixation and dehydration, the processing of tissue for EM introduces considerable structural artifacts (11). For example, as demonstrated by x-ray diffraction, the “major dense line” is not “major” or “dense”-i.e. this moniker is a misnomer because it implies the obliteration of a hydrated space. Moreover, EM involves limited and possibly biased sampling because the microscopist examines only those regions of sections that are recognizable as multilamellar myelin. Such “preserved” membrane arrays are often thought to represent native structure. Despite these shortcomings, previous EM studies on transgenic mice containing altered myelin-related genes as well as on biopsy/autopsy material from humans with peripheral neuropathies describe either no changes or significant changes in myelin period, often with packing defects (as exemplified previously) that are attributed to the genetic defect (1, 2, 5, 12, 13). The packing defects have been construed as indicating myelin instability, but there has been no direct testing of this hypothesis.

To understand structure/function relationships involving myelin proteins that are thought to underlie membrane-membrane adhesion and membrane packing stability of internodal myelin, we have used x-ray diffraction to analyze internodal myelin-structure, membrane packing, and interactions-in unfixed nerves from tg mice. In the current article, we studied myelin in tg mice having genotypes P0^+/−, P0^null, P0^myc/P0^+/+, cnsP0/PLP^null, cnsP0/PLP⁺, and PLP^null. Our study was undertaken to address questions about genetically altered myelins that are in the native state, i.e. unfixed, hydrated, unstained, and unembedded. To assess intermembrane interactions in these myelins, we treated some of the nerves at lowered ionic strength and alkaline pH. Such treatments, which are examples of “electrostatic stressing,” have been shown by EM to cause changes in myelin period owing to altered membrane packing (14, 15). To begin our study, we linked the diffraction findings with previous EM results by analyzing x-ray patterns from some of the nerves after aldehyde fixation (but not further processing) and from some nerves embedded in plastic blocks (i.e. after complete processing). From the x-ray patterns, we calculated the electron density projection along the membrane stacking direction, the extent of disorder in the stacking of the membranes, and the amount of compact myelin. Our study demonstrates that we can rigorously characterize myelin period and packing in transgenic mice by analyzing x-ray diffraction from unfixed whole nerves and from fixed nerves. For example, we discerned periods (sometimes differing by a few Å) among sciatic nerves having different levels of P0 and showed that aldehyde fixation qualitatively preserves some packing abnormalities. We have demonstrated that the relative amount and structural integrity of myelin in a whole nerve could be quantitated and that the amount of P0 in the myelin affected the stability of its packing. Our findings directly support the hypothesis, intuited from EM on embedded nerves, that packing defects underlie instability in “live” or intact myelin. Thus, membrane diffraction provides a unique tool to quantitate membrane packing defects in molecularly defined demyelinating neuropathies.

Materials and Methods

Transgenic Mice

For characterizing peripheral nervous system (PNS) myelin, wild-type mice (P0^+/+) as well as transgenics, including P0 heterozygous null (P0^+/−), P0^null (P0⁻, or P0^null), and myc-tagged P0 (P0^myc/P0^±/±, or mycP0^±/±), were obtained from the colonies at the San Raffaele Scientific Institute (HSR) (Milano, Italy [4]). The myc-tagged P0 protein contains an extra 13 residues (IEQKLISEEDLNA) at the N-terminus of the normal P0 protein; and the mice used in the current experiments corresponded to line Tg88.4P0myc. These myc-tagged mice show a total P0 expression approximately 1.45 normal levels of mRNA; however, the total amount of P0 protein is reduced because the amount of bulk myelin in nerves is diminished as a result of neuropathy (4). For characterizing central nervous system (CNS) myelin, wild-type (PLP^+/+) as well as transgenics (including PLP^null [PLP⁻], cnsP0/PLP⁺, and cnsP0/PLP^null) were obtained from the mouse breeding colonies at the Cleveland Clinic Foundation (CCF) (5) (Yin et al, unpublished observations). The P0 transgene in the CNS, which we refer to as cnsP0, is under the control of myelin basic protein promoter (Yin et al, unpublished observations). We have previously described some of our diffraction results on the transgenic optic nerves from CCF (16) (Yin et al, unpublished observations). At CCF, HSR, and Boston College, we have adhered to the standards set forth by the National Research Council's Guide for the care and use of laboratory animals.

Sample Preparation

Sciatic and optic nerves were dissected from wild-type and transgenic mice that had been killed by cervical dislocation. During dissection, the tissue was continually rinsed with physiological saline (154 mM NaCl, 5 mM Tris buffer, pH 7.4). Nerves were tied off at both ends with fine silk suture. For immediate diffraction, a freshly dissected nerve was slightly stretched and inserted into a 0.7-mm quartz capillary tube (Charles Supper Co., Natick, MA) with medium and the capillary was sealed at both ends with wax. For electrostatic stressing before diffraction, a nerve was incubated overnight at room temperature and with gentle agitation in 20 mL medium of known pH and ionic strength, and subsequently mounted in the capillary. Unembedded, fixed nerves from HSR were prepared by immersion in 2% glutaraldehyde in 0.12 M phosphate buffer at pH 7.4, whereas nerves from CCF were prepared by vascular perfusion of mice with 2.5% glutaraldehyde, 4% paraformaldehyde, 0.08 M Sorenson's buffer at pH 7.4 (Yin et al, unpublished observations). Furthermore, some of the embedded nerves that had been examined in the light microscope or by EM (4) were trimmed from the plastic blocks and examined by diffraction.

X-Ray Diffraction and Myelin Structure Analysis

Diffraction experiments (Fig. 1) were carried out using nickel-filtered, single-mirror focused CuKα radiation from a fine-line source on a 3.0 kW Rigaku x-ray generator (Rigaku/MSC Inc., The Woodlands, TX) operated at 40 kV by 14 to 22 mA. In a typical experiment, the x-ray diffraction pattern was recorded for 1 hour (except P0^null for 4 hours) using a linear, position-sensitive detector (17, 18) (Molecular Metrology, Inc., Northampton, MA). Detector characteristics and specimen-to-film distance were established by recording diffraction patterns from a standard consisting of silver behenate powder, which has a fundamental spacing of d₀₀₁ = 58.38 Å (19, 20). Diffraction data here is shown as either the raw spectra or as the total intensity (after background subtraction; see subsequently) as a function of reciprocal coordinate. For purposes of calculation, the specimen-to-film distance (approximately 200 mm) was expressed as channel number. The integral width of the direct beam in Gaussian form was 7.4 channels (or 8.2 × 10⁻⁴ Å⁻¹).

FIGURE 1.

Overview of diffraction method and the information it provides. The schematic summarizes, from top to bottom, the steps in undertaking and analyzing x-ray diffraction from myelinated nerves. Note that whole nerves can be examined and that the x-ray scatter from the nerve can be recorded on either film (which is digitized) or using an electronic detector. From the x-ray pattern, the myelin period can be directly measured from the positions of the intensity maxima (Bragg orders 1-5 and 11 are most apparent here). After background subtraction, the areas of the peaks (or integrated intensities) are used to calculate structure amplitudes, which are then used in a Fourier synthesis to calculate the membrane profile. From the profile are measured the distances between membranes at the cytoplasmic and extracellular appositions (cyt, ext) and the thickness of the lipid bilayer (lpg). The molecular models above the profile indicate the positions of polar lipids and cholesterol in the membrane array and show that there is considerable water space between the lipid bilayers. Protein has been omitted for clarity. Finally, the strip along the bottom of the figure shows how an electron micrographic image of myelin, with its major dense line (MDL) and intraperiod line (IPL), relates to the membrane profile based on diffraction data. (The schematic of a myelinated nerve, the lipid molecules, and the EM were redrawn from [22.])

Measurement of the positions of the reflections in the myelin diffraction pattern gives direct information about the periodicity of myelin and quantitation of the intensities above background allows calculating membrane profiles (see subsequently; Fig. 1; reviewed [21-23]) as well as the relative amount of multilamellar myelin in the sample. The diffracted intensity from whole myelinated nerve was analyzed using PeakFit (Jandel Scientific, Inc.) or in-house software (24). After background subtraction, the intensity in the resulting peaks was integrated to yield integral areas I(h) and integral widths w(h) for each reflection of order h. As previously detailed (25-27), the intensities I(h) were used in the calculation of the membrane profiles, from which we measured the widths of the intermembrane spaces at the extracellular and cytoplasmic appositions (ext and cyt, respectively) and the thickness of the membrane bilayer (lpg). These structural parameters are defined as the distances between the middles of the head group layers across the intermembrane (extracellular and cytoplasmic) spaces and within a single bilayer (22). The disorder in the stacking of membranes in multilamellar myelin was determined by plotting the integral widths w² (see previously) as a function of h⁴, in which the intercept on the ordinate axis is inversely related to the number of repeating units N (the coherent domain size), and the slope is proportional to the fluctuation in period Δ (lattice or stacking disorder) (24). An alternative algorithm for analyzing disorder and domain size in myelin (28, 29) does not yield any better predictions that can be validated by traditional morphologic methods.

To estimate the relative amount of multilamellar myelin among the samples, we calculated (for samples having equivalent exposure times) the total integrated intensity (M) above “background” (B, after excluding the small-angle region of the pattern around the beam stop as well as the wide-angle region of the pattern). The stability of the x-ray generator and its x-ray output, which declines very slowly over a period of years, obviated our need to continually monitor the intensity of the direct beam. Moreover, this normalization allowed us to compare the diffraction from nerves having different thicknesses and therefore different absorptions. As we discovered and illustrate subsequently, a scatterplot of the fraction of total, scattered, integrated intensity that is the result of myelin (M/(M+B)) versus myelin period (d) facilitated distinguishing among the different genotypes.

Results

Diffraction From Plastic-Embedded Nerves: Noteworthy for Their Paucity of Information

Because EM examination of thin sections is routinely used to assess the ultrastructural integrity of myelinated tissue in cases in which demyelinating neuropathy is suspected, we asked whether x-ray diffraction patterns from whole nerves embedded in plastic might confirm the ultrastructural observations or provide additional details. X-ray diffraction has the sampling capability of probing in a single exposure a relatively large volume of tissue (approximately 0.8 mm³ based on the size of the beam and the thickness of the nerve). EM of thin-sectioned material, by contrast, besides being limited to what is judged to be “preserved” after processing, can sample at one time only approximately 2 × 10⁻⁸ mm³ (based on the approximate dimensions of a thin section). Furthermore, the resolving power of x-rays in membrane diffraction is several Å or more, whereas that for EM is approximately 15 to 20 Å, which is the approximate difference in period between CNS and PNS myelins. Based on EM of “preserved” multilamellar myelin, altered myelin-related genes are thought to result in either no changes or significant changes in myelin period, in which the membrane packing is irregular or the staining of the major dense and/or intraperiod lines are altered.

Using x-ray diffraction from embedded nerve (wild-type controls, mycP0/P0^+/+, mycP0/P0⁻, and P0^null), we observed very weak, broad, poorly resolved intensity maxima (Fig. 2A) by contrast with the strong, well-defined intensity maxima from unembedded (fixed or unfixed) nerves. The fraction of total diffracted integrated intensity accounted for by multilamellar myelin (i.e. M/[M+B]) was <1% for embedded nerve compared with 25% to 30% for fixed or unfixed WT (see subsequently). From the positions of the intensity maxima for embedded nerve, we calculated myelin periods of ∼160 Å, ∼160 Å, ∼180 Å, and ∼190 Å, respectively, which were substantially different from that in fixed or unfixed nerves (see subsequently). The fact that the peaks were broad also indicated extensive heterogeneity or disorder (irregular packing) and presumably reflects the lack of structural integrity of the internodal “compact” myelin that is examined by EM. Detailed analysis of the successive changes in myelin structure with processing has been described previously (11); here we are focusing on assessing packing defects in tg myelin.

FIGURE 2.

Diffraction from sciatic nerves chemically treated for electron microscopy. (A) Plastic-embedded whole nerves. Very weak-intensity maxima are detected compared with that for unembedded aldehyde-fixed nerve (dotted), and no discrete scatter at all is apparent for P0^null. The arrows indicate the positions of the broad second orders from the WT, mycP0^+/+, and mycP0⁻, which also have fourth orders detectable. The P0^null nerve, at the same y-axis scale as the other, shows only the barest hint of a broad intensity maximum at a position that corresponds to the fourth order of an expanded lattice (at channels 700-800 and 1200-1300; brackets). The patterns from embedded nerve were each recorded for 2 hours, whereas that from the fixed WT was for 1 hour; and the spectra have been shifted along the y-axis for clarity. Periods are summarized in Figure 4A. (B) Unembedded, aldehyde-fixed nerves. Scatterplot of M/(M+B) versus myelin period. The error bars are sample standard deviations. The Student t-test indicates that, except for p < 0.04, comparing the mean period of mycP0^+/+ with WT, all significance levels comparing the transgenics in either period or relative amount of myelin to their respective control (i.e. mycP0^+/− to P0^+/− and mycP0⁻ to P0⁻) or to the WT (P0^+/+) are < 0.0001. The tremendous data spread in period for the P0⁻ is the result of the very poorly resolved Bragg peaks in the diffraction patterns; however, the relative amount of myelin in these samples is always substantially less than in all the other nerves. Inset: The x-ray patterns, from 1-hour exposures, have been shifted vertically for clarity. As reference marks to indicate the shift in positions of the reflections, the second and fifth order intensities for the WT and mycP0⁻ are denoted with numerals and with short vertical bars above and below the spectra.

Diffraction from Aldehyde-Fixed Nerves Revealed Differences Among tg Nerves

EM is well-suited to examining morphology of myelinated nerve and can detect relatively large changes in internodal period, but, as noted previously, may not be appropriate for measuring subtle changes in the widths of intermembrane spaces or in probing the stability of the myelin arrays. To assess the usefulness of nerves that had been initially fixed but not further processed (where most of the structural changes and distortion are introduced [11]), we recorded diffraction from aldehyde-treated nerves. This part of the study was prompted also by the cost of transporting live animals from the tg mouse facilities to our lab for dissection and analysis of fresh, myelinated tissues. For these experiments, the nerves were fixed at the transgenic facility (HSR and CCF; see “Materials and Methods”) and maintained in the fixative during shipment and during the x-ray exposures. The diffraction patterns from such nerves (Fig. 2B, inset) showed x-ray scatter that was nearly as strong as from unfixed samples and intensity maxima that were considerably sharper than from embedded nerves but not as sharp as from unfixed nerves (see subsequently). The reflection broadening indicated increased disorder or distortion caused by the fixative. In addition, changes in the positions and intensities of the reflections were indicative of alterations in myelin structure (11), specifically, narrowing of the cytoplasmic space, which became partially filled with density, and widening of the extracellular space, which became more fluid-filled. Overall, however, the aldehyde treatment did not introduce modifications that were as extreme as those shown by embedded nerves.

Despite changes caused by chemical fixation, we detected reproducible differences in the relative extent of myelination and myelin periods for the various genotypes both for tg sciatics (Fig. 2B) and for tg optics (Figs. 3, 4). The relative strength of diffraction from multilamellar myelin in whole nerve (calculated as M/[M+B]; see “Materials and Methods”) was ranked from highest to lowest for the sciatic nerves as WT (at ∼25%) > mycP0/P0^+/+ (∼20%) ≥ mycP0/P0^+/− (∼15%) > mycP0/P0^null (∼5%) > P0^null (∼1%), and for the optic nerves as WT (at ∼25%) > cnsP0/PLP^null, cnsP0/PLP⁺, and PLP^null (with each ∼15%).

FIGURE 3.

Diffraction from aldehyde-fixed optic nerves. The position of the second order reflection from WT is indicated (solid vertical lines), and the spectra from 1-hour exposures have been shifted vertically for clarity. The spectra are shown from top to bottom as indicated in the key. For comparison, the diffraction from fixed WT sciatic nerve (dotted) is also shown, with the position of its second order indicated (dashed vertical lines). Note that the cnsP0/PLP⁺ is similar to but weaker than the WT optic, and that the cnsP0/PLP^null and PLP^null resemble the WT sciatic nerve pattern. Discrete fourth and fifth order reflections in cnsP0/PLP^null are seen only as a smear in the PLP^null owing to irregular membrane packing in the latter.

FIGURE 4.

Myelin period and intermembrane spaces. Results for optic nerves are shown in the upper part and for sciatic nerves in the lower part of each panel. (Left) Comparison of periods among embedded, fixed, and unfixed nerves. Not every type of sample preparation for each kind of nerve was available. Generally, nerves that had undergone complete processing for electron microscopy (embedded) had periods substantially less than that for fixed and for unfixed nerves, owing to the large shrinkage when the water is replaced with plastic (11). Also, fixed nerves had periods that were larger than for unfixed myelin; and the exception was the myelin of PLP^null optic nerve. (Right) Widths of intermembrane spaces in myelin of unfixed nerves whose periods are indicated in the left panel. Note how the extracellular space in cnsP0/PLP^null optic nerve closely resembles that of WT sciatic and that the extracellular spaces in the P0 and PLP null myelins are substantially swollen.

The period for fixed WT sciatic nerve myelin was ∼185 Å (Fig. 2B), and that for nerves with mycP0 on WT or heterozygous backgrounds was greater than WT by up to 5 Å. Fixed sciatics from the 2 other genotypes (P0^null and mycP0/P0^null) showed myelin with swollen periods of 195 Å and >200 Å, respectively. By comparison, as evident from electron micrographs, complete processing increased their difference in periods by 12% (from 156 Å for WT to 175 Å for mycP0/P0^null), with the increase resulting from a widened intraperiod line (4). Electron density profiles of myelin calculated from the diffraction data of unembedded, fixed nerves showed that the extracellular apposition was wider by 7 Å (Fig. 4B).

For the fixed optic nerves, the periods were ∼160 Å for WT, ∼5 Å greater for tg cnsP0/PLP⁺, and ∼180 Å for the cnsP0/PLP^null and PLP^null (Figs. 3, 4). Myelin in the latter 2 optic nerves-one containing only P0 and the other without P0 or PLP-had similar periods as fixed sciatic nerve (Fig. 3, dotted curve). Furthermore, cnsP0/PLP^null had similar although weaker second to fifth order intensities as sciatic, indicating similar membrane packing between the tg optic and WT sciatic myelins. This set of experiments demonstrated that quantitation of diffracted intensity and period of myelin from aldehyde-fixed nerves can inform about the relative extent of myelin among different whole nerves and can detect subtle but reproducible differences in period.

To What Extent Does Aldehyde Fixation Preserve Abnormal Myelin Packing?

To determine how well aldehyde fixation preserves altered myelin packing, and to help link further the results from EM with those of diffraction, we incubated freshly dissected sciatic nerves from wild-type mice (inbred DDY strain) in normal or hypotonic saline (controls), or phosphate buffer isosmolar to the saline, and then fixed them in 2% glutaraldehyde in 0.12 M phosphate buffer at pH 7.4. We also used saline or phosphate buffer at 60 mM, which leads to myelin swelling, and then fixed the nerves. Analysis of the diffraction patterns showed that myelin in 154 mM NaCl or in 120 mM phosphate buffer gave virtually identical patterns having 179 Å periods, and when the latter nerve was fixed, the pattern changed-the period swelled to 191 Å (data not shown) and the third and fifth order intensities became prominent, indicating that within the swollen lattice, the membrane packing became more asymmetric-that is, the difference in widths of the cytoplasmic and extracellular appositions became greater (11). Thus, although these ionic conditions had little or no effect on myelin, the aldehyde treatment did, increasing the period by ∼7%. The diffraction patterns also showed that the packing abnormality (myelin swelling to 240 Å) caused by incubation in 60 mM NaCl or phosphate buffer was somewhat preserved by glutaraldehyde fixation (∼230 Å); however, the sharp reflections become considerably broader, indicating a major disruption of the regular packing of membranes. Because fixation introduces its own packing artifacts, we investigated the structure and stability of myelin in unfixed whole nerves from some of the transgenic mice strains used here.

Myelin Structure in Unfixed Control Nerves from Wild-Type and Transgenic Mice

To establish a baseline for analyzing the freshly dissected nerves from the transgenic mice, we recorded and analyzed the myelin diffraction patterns from control nerves. These included sciatics and optics from wild-type mice, as well as optic nerves from HSR mutant mice having an altered PNS phenotype and sciatic nerves from CCF mutant mice having a CNS phenotype. All of these mice were of the FVB/N strain. Sciatic and optic nerves in physiological medium (154 mM NaCl, pH 7.4) gave x-ray patterns like those previously recorded (reviewed in [22]), with periods 173.0 ± 1.2 Å (n = 5) and 153.7 ± 1.0 Å (n = 7) (Fig. 5A; Table 1). As previously established, measurement of the membrane profiles (Fig. 5B) showed that the substantial difference in period between sciatic and optic myelins was mostly attributed to the larger separation at the extracellular apposition (ext) in the sciatic (Table 1). By contrast, the thickness of their lipid bilayers (lpg) was about the same and the width of the cytoplasmic appositions (cyt) was marginally greater in sciatic. Analysis of the total integrated intensity indicated that the relative amount of ordered myelin in sciatic and optic nerves (i.e. M/[M+B]) were similar (Table 1). Analysis of the widths of the reflections (Fig. 5C) showed that, compared with optic nerve myelin, sciatic myelin consistently was more crystalline or regularly packed (smaller integral widths of reflections), had a greater number of repeating units (smaller y-intercept value), and less stacking disorder (shallower slope).

TABLE 1.

Summary of X-ray Structural Data for Unfixed Myelin of Wild-Type and Control Nerves*

FIGURE 5.

X-ray diffraction analysis of unfixed, wild-type, and control nerves for the transgenic mice studied here. (Diffraction patterns for the CCF mice have appeared elsewhere [16].) (A) Total x-ray intensity as a function of reciprocal coordinate (R, in Å^-1) after background subtraction for wild-type sciatic (solid) and optic (dotted) nerves. The Bragg orders (h) for the 2 patterns are indicated. Inset: The raw diffraction data before background subtraction. (B) Electron density profiles (on a relative scale) as a function of distance from the middle of the cytoplasmic apposition for wild-type sciatic (solid) and optic (dotted) nerve myelins, calculated from the data shown in (A), show the periods (d) and the widths of the intermembrane spaces at the cytoplasmic (C) and extracellular (E) appositions, and the thickness of the lipid bilayer (L). The large difference in the widths of their extracellular appositions is the major structural basis for their difference in period. (C) Quantitation of the widths of the reflections (peak widths), expressed as the square of the integral half-width w²(h) versus the fourth power of the Bragg order h⁴, provides information about the relative number of myelin layers in a diffracting region of the sheath (∝ [(intercept)^-1]) and the regularity of the membrane packing (∝ slope) (24). Linear least square lines have been fit to the data points, and the equations and R² are indicated. Note that unfixed versus fixed myelin, and sciatic versus optic nerve myelin are readily distinguished by the slopes of their w² versus h⁴ lines. Line widths from fixed optic nerve myelin were too variable for line fitting. (D) Scatterplot for intercepts versus slopes from the analyses of peak widths for the myelin diffraction from the different transgenic mice reported here. Note that smaller intercepts and smaller slopes correspond, respectively, to increased thickness and increased regularity of the membrane packing. Inset: The product (slope^-1 × intercept^-1), which we suggest as a “quality index” for the structural integrity of the myelin in a whole nerve, is indicated for the sciatic and optic nerves studied here. The values have been normalized to that for wild-type sciatic nerve myelin (lowest bar). Note that the “quality” index for wild-type optic is approximately 10% that of the sciatic, owing to less regularity in its membrane packing and thinner diffracting domains. Furthermore, the structural integrity of fixed myelin in sciatic nerve is very low (open bar), although the amount of myelin in the nerve gives a strong x-ray pattern (Fig. 2A, B).

Myelin Structure in Unfixed Sciatic Myelin of P0 Transgenic Mice

The myelin from heterozygous (P0^+/−) or null (P0⁻) mice that were approximately 100 days old diffracted more weakly than the controls (Fig. 6A). The P0^+/− myelin had half or less the intensity of the WT, a period of ∼180 Å, and seemed to be unstable, because it swelled by ∼30 Å within 24 hours after nerve dissection. Because a nerve will typically show a native or native-like period for several days after dissection, the presence of swollen membrane arrays is indicative of inadequate membrane-membrane adhesion-i.e. instability. The instability of P0^+/− myelin at ∼3 months was also evident at ∼17 months of age, where ∼200 Å-period, swollen arrays were detected (Fig. 7). Widening of the extracellular space accounted for the increased period of P0^+/− myelin. Sciatic nerve from P0^null mice diffracted at ∼10% the wild-type level and the strongest of the reflections corresponded to the second through the fifth orders of a 191 Å-period structure. Compared with wild-type, the number of repeating units was approximately one half and the stacking disorder was 2- to 4-fold greater (Table 2).

TABLE 2.

Transgenic Myelin Structure: Physiological Conditions and Hypotonic Saline at pH 8*

FIGURE 6.

Diffraction data from unfixed, transgenic sciatic nerves. (A) Physiological conditions (154 mM NaCl, pH 7.4); (B) swelling conditions (30 mM NaCl, pH 8). The myelin periods are indicated in the legend boxes, and the Bragg orders are shown for the peaks. X-ray data was collected for 1 hour from each sample. Note the difference in scale for the y-axis and that the swollen myelin gave much weaker x-ray diffraction than the native. Measurements from the patterns are summarized in Table 2.

FIGURE 7.

Diffraction evidence that myelin heterozygous for P0 is less stable than wild-type or mycP0^+/+ myelins. Sciatic nerves (unfixed) were dissected from mice older than 15 months and inserted in contact with physiological saline at pH 7.4 in x-ray capillaries, which were then sealed. X-ray patterns were recorded from the nerves immediately (A) and again after 4 hours (B) and after overnight (C). Bragg orders for the reflections are indicated above or below the spectra. The patterns show that compared with the WT and mycP0^+/+ nerves, which have native periods, the P0^+/− is initially swollen. At 4 hours, the WT and mycP0^+/+ are showing a weak, new reflection (*) at the position of the prominent third order reflection of the P0^+/−. After overnight, the predominant structure in all 3 nerves is the swollen pattern and remnants of the native reflections are evident (Bragg orders 2, 3, 5) for WT and mycP0^+/+.

Addition of myc-tag P0 onto a P0^+/+ background resulted in sciatic nerve that diffracted less than half as strongly but had a similar myelin period as the control (Fig. 6A). Analysis of the integral widths of the reflections (see “Materials and Methods”; Fig. 5D) showed that the mycP0^+/+ myelin was less regular than the wild-type, although its period and intermembrane distances were the same (Table 2).

Swelling of Myelin in Transgenic Sciatic Nerves in Hypotonic Saline at pH 8 Was Normal

Because the equilibrium spacing between myelin membranes is determined by a balance of attractive and repulsive forces and by specific molecular contacts (30), the changes in the pH and ionic strength of the incubation medium (“electrostatic stressing”) may affect this balance and reveal differences between the membranes in wild-type versus genetically altered myelins. EM reveals the swelling of myelin at its extracellular appositions in hypotonic saline (15, 31) and changes in membrane packing after treatment at nonphysiological pH (14). As a prelude to a more complete study of the intermembrane interactions in genetically modified sciatic nerves, we incubated dissected nerves from tg mice in hypotonic saline (30 mM and 60 mM) at pH 8, conditions that have been shown to cause myelin membrane swelling (Fig. 6B; Table 2).

The swelling of peripheral nerve myelin in hypotonic media is the result of electrostatic repulsion at the extracellular apposition, and the amount of swelling increases as ionic strength is lowered (as counterion concentration is reduced) and as pH is increased (as acidic groups are deprotonated, increasing the net negative charge) (30). For example, at pH 8 in 60 mM and 30 mM NaCl, the period for mouse sciatic myelin increases by more than 60 Å. For sciatic nerves from P0 heterozygotes (P0^+/−) incubated at pH 8 in the dilute saline, the myelin periods were 231 and 248 Å at 60 mM and 30 mM, respectively. MycP0^+/+ myelin after treatment at pH 8 in the dilute saline swelled to similar extents (237 Å and 247 Å).

Typically, for treatments that resulted in myelin swelling, the diffracting power and number of repeating units decreased and the amount of packing disorder increased (Table 2; Fig. 5C, D). The membrane profiles showed that the major change in structure underlying the considerable differences in period was mostly the result of a widening of the extracellular space (Table 2). By contrast, the packing of membranes at their cytoplasmic apposition and the thickness of the membrane bilayer were constant, even in the transgenic myelin.

Residual Multilamellar Structures in P0^null Sciatic and PLP^null Optic Nerves Are Similar

As recently reported (16), x-ray results have substantiated both quantitatively and structurally that the introduction of P0 into PLP⁻ optic nerve myelin generates a “rescued” CNS myelin that resembles PNS myelin in having the same period, intermembrane spacing (cyt, ext) and low stacking disorder, but like CNS myelin, has a smaller number of units in the coherent domain. By contrast, PLP^null (PLP⁻) optic myelin (in 154 mM NaCl at pH 7.4) shows a diffraction pattern that is substantially swollen compared with wild-type optic (194 Å vs 155 Å) (16) (Yin et al, unpublished observations). Here, we found that P0^null (P0⁻) sciatic myelin under the same conditions also was notably swollen compared with wild-type sciatic, and moreover, its diffraction pattern was remarkably similar to that of PLP^null optic nerve (Fig. 8). Moreover, for both P0^null sciatic and PLP^null optic myelins, the diffracting power was very small, and the sheaths were thin and disordered (Fig. 5D), like that for other swollen myelins (Table 2). Profiles for the membrane pairs from P0^null sciatic and PLP^null optic show that the bilayer thickness was essentially the same as in the wild-type and that the main increase in period from the wild-type values came from widening of the space at the extracellular apposition (to 68 Å from 31 Å in the optic and to 71 Å from 48 Å in the sciatic).

FIGURE 8.

Comparison of unfixed P0^null sciatic and PLP^null optic nerve myelins. (Left) X-ray diffraction patterns. Bragg orders 2 to 5 are indicated. Broadening of the reflections is likely the result of considerable disorder arising from abnormal membrane-membrane adhesion. (Right) Membrane profiles (on a relative scale) as a function of distance from the middle of the cytoplasmic apposition. Comparison of these profiles with those from native sciatic and optic myelins shows how similar are the dimensions of the widths of the spaces between the membranes.

Discussion

Rationale for Diffraction Approach and Comparison With Electron Microscopy

X-ray diffraction analysis of membranes can provide a measure of the bilayer dimensions, establish the localization and distribution of particular proteins and lipids, and monitor dynamic changes in structure owing to the effects of the milieu (21, 23). For multilamellar systems such as myelin, the diffraction technique can provide highly accurate measurements on the packing distances between the membranes and thus may provide insight on the function of intermembrane adhesion proteins in myelin (22). For example, the correlation of biochemical findings with x-ray diffraction analysis of myelin in naturally occurring mouse mutants (e.g. jimpy, quaking, msd, trembler, shiverer, and rumpshaker) (32-36) has contributed to our understanding of the role and localization of specific proteins and lipids in myelin compaction. The shiverer mouse, for example, lacks myelin basic protein in both the CNS and PNS; however, although dysmyelination is observed in the CNS, typical multilamellar compact myelin populates the PNS (34, 37), suggesting that myelin basic protein is less important for myelin formation and compaction in the PNS (12). Moreover, x-ray observations that shiverer PNS myelin swells to a greater extent than wild-type myelin as a function of pH at low ionic strength correlate with an elevated content of sulfatide (30, 32, 38). Studies on the rumpshaker mouse (jp^rsh; in which there is an Ile186Thr substitution in PLP (39)) also exemplify why diffraction is essential in characterizing internodal myelin. EM of fixed jp^rsh CNS myelin shows a decrease in periodicity and a lack of clear distinction between the MDL and the IPL (40); however, diffraction from unfixed nerves of jp^rsh reveals myelin periods 5 to 15 Å wider than normal and less stable packing of the membranes (33).

In the current study, we report x-ray diffraction measurements of myelin structure in nerves from transgenic mice. Peripheral myelin was studied in P0 heterozygous- or homozygous-null mice with or without the P0myc transgene (4); CNS myelin was studied in PLP-null mice with or without a CNS-expressed P0 transgene (16) (Yin et al, unpublished observations). Previous analysis of myelin structure in nerves from these mice relied on EM to evaluate period and packing. In using x-ray diffraction to complement the EM studies, we were able to provide quantitative data not only on myelin membrane structure and the widths of intermembrane spaces, but also on the thickness of internodal compact myelin and its extent of ordered packing and on the stability of the myelin. Based on the breadths of the reflections (24), we developed here a “quality index” for comparing myelins-i.e. the product of the relative thickness of the sheaths and the regularity of the multilamellar stacking was taken to indicate the overall structural integrity or state of the myelin. Furthermore, by determining what fraction of total diffracted intensity was from ordered myelin, we could estimate the relative amount of myelination among the different nerves. To link our approach with the more familiar one of EM, we also analyzed diffraction patterns from glutaraldehyde-fixed nerves and from plastic-embedded nerves. Because the x-rays sampled a large volume of tissue-that subtended by the beam (∼200 μm × 3 mm) times the thickness of the nerve-we obtained a comprehensive picture of internodal myelin at the level of the multilamellar assembly of the sheaths as well as at the level of molecular organization of the membranes in the nerve bundles.

In the x-ray diffraction experiments described here, the x-rays that are scattered by the scores or hundreds of fibers in a nerve segment come from multilamellar internodal myelin (which includes regularly ordered as well as disordered membranes), noninternodal myelin (e.g. paranode), and nonmyelinic tissue constituents (e.g. basement membrane, collagen, microtubules, neurofilaments; and macromolecules, metabolites, salts, water, and so on). These constituents may or may not be arrayed, may be present in amounts dwarfed by the amount of myelin, or may not be oriented appropriate to the x-ray camera geometry to give coherent scatter. For example, the one-dimensional (line-focus) camera used here is suited to recording x-ray patterns from the multilamellar myelin when the nerve axis is parallel to the beam length but not from endoneurial collagen fibers whose 695 Å-repeat is perpendicular to the myelin repeat. Heterogeneity of the myelin, owing to diverse pathologic stages in a neuropathy, could result in multilamellar arrays having different repeats. Thus, a classic CMT1 demyelinating neuropathy may have several types of fibers: (i) fibers with packing abnormalities, (ii) fibers with active myelin degeneration (e.g. ovoids that contain myelin debris), (iii) fibers with thin remyelination and packing abnormalities, (iv) fibers with failed remyelination and onion bulb formation, and (v) lost axon/myelin units substituted by endoneurial collagen accumulation. The manner in which such a nerve contributes to the diffraction patterns will depend on whether the particular fiber type contains periodic arrays sufficient to produce reflections and on what proportion of total nerve volume that type of fiber occupies. Of the 5 types of fibers cited, one might expect (i), (iii), and (v) to produce reflections, and (i) and (iii)-(v) to occupy significant proportions of volume. Type (ii) fibers would not produce reflections or occupy significant proportion of volume, and (iv) would not produce reflections, that is, fiber types (ii) and (iv) would effectively be invisible. Types (i) and (iii) would figure in measures of periodicity and lattice stability, and (iii) and (v) would figure in the relative quantity of myelin as measured by M/(M+B). When 2 different packings are present in a significant proportion of the fibers, then 2 overlapping patterns of reflections are produced (Fig. 7).

For analyzing the diffraction data, the “quality index” for comparing myelins is based on direct measurements of the integral widths of the x-ray reflections, irrespective of the strength of the pattern. By contrast, the relative amount of myelin in a nerve is based on the relative strength of the myelin diffraction compared with the total diffracted intensity. In calculating the latter, only a portion of the total pattern was used because we omitted the intense spillover on either side of the central beam-stop and the flat intensity outside the major portion of the diffraction pattern. Thus, because the background (B) is underestimated, the calculated fraction of total diffracted intensity that is from ordered myelin is an overestimate of the actual value and should be regarded only as a relative measurement for comparing different nerves.

Myelin Structure and Amount of P0

Consistent with the reported hypomyelination (1, 4), peripheral myelin from the P0^+/−, P0⁻, and mycP0^+/+ mice all diffracted more weakly than the wild-type, with the P0 null myelin diffracting at the lowest level (WT > P0^+/− ≈ mycP0^+/+ > P0^null; Table 2). Analysis of the breadths of the x-ray reflections, the calculated number of repeating units in a diffracting domain and the amount of disorder within the multilamellar array, showed that the P0^+/− and P0⁻ sciatic nerves had approximately half or fewer the number of repeating units and considerably more disorder than the wild-type (Table 2; Fig. 5B). MycP0^+/+ had a similar number of repeating units as the wild-type but more packing irregularity (greater disorder). Based on the available biochemical analysis of isolated myelin from these transgenics, we can conclude that adequate amounts of P0 protein (but not too much, like in the mycP0^+/+ mouse, which expresses approximately 1.45 the normal level) are required for generating normal numbers of myelinated sheaths in a nerve. Based on its thickness (size of diffracting domains), the myelin, however, seemed able to incorporate the excess P0 (3). Moreover, more regular or orderly stacking of membranes seemed also to be dependent on having P0 present. This is particularly evident from diffraction studies demonstrating that the introduction of P0 into PLP^null optic nerve generates CNS myelin that has low stacking disorder like PNS myelin (16) (Yin et al, unpublished observations) (Fig. 5D). The greater packing disorder in mycP0^+/+ myelin is likely the result of its additional N-terminal residues. Added molecular mass toward the apical end of the extracellular domain could affect homophilic packing of the putative, 2-fold adhesion interface (41), causing packing disorder without affecting the number of layers or myelin period. In tg optic nerve, the addition of P0 to PLP-containing myelin (i.e. in cnsP0/PLP⁺) did not affect the period but did reduce the “quality index” from that of WT, owing mostly to fewer layers. It is likely that PLP in this myelin precludes the formation of the homotypic P0-P0 interactions that are responsible for the widened, PNS-like period seen in cnsP0/PLP^null CNS myelin.

The amount of P0 also affected the myelin period, mostly from changes in the width of the extracellular space between membranes. For the null genotype, which is the most extreme case (i.e. no P0), unfixed nerve myelin showed an increase in period from 174 Å in wild-type to 191 Å. By contrast, the heterozygous (P0^±) myelin was at most only a few Å wider than wild-type. This increase in period was reflected by an increase in ext by only a few Å in P0^+/−, but by >20 Å in P0⁻. EM measurements on plastic-embedded myelin from the null genotype report different values for its period, owing most likely to inconsistent fixation and processing among laboratories and/or to instability of the myelin; in one report, the regular compact membranes in the PNS of P0⁻ mice had the same period as wild-type (∼120 Å), but there were also 3 other regular arrays with various periods (1), whereas in another report, EMs suggest a doubling in period in the null compared with wild-type (4).

Myelin Period Heterogeneity in Embedded Nerves

The considerable variability in measurements from EM is the result of processing effects, not only from the use of different initial fixatives, but also from the subsequent steps. Our x-ray measurements on plastic-embedded sciatic nerves from the tg mice demonstrate the profound deterioration in structural integrity of the multilamellar myelin (Fig. 1); the amount of ordered myelin was ∼3% or less that in unfixed or glutaraldehyde-fixed nerves, and the broadened reflections indicated considerable heterogeneity caused by disordering of the regular packing of membranes. Thus, measurements from EMs are based on a very small sample of the myelin, specifically whatever has retained a regular multilamellar structure despite the effects of crosslinking and oxidizing reagents, dehydration, and plastics. For example, as previously shown from observations on semithin and thin sections cut from plastic-embedded tissue (4), sciatic nerves from mycP0⁻ mice show altered periodicity compared with WT, whereas mycP0^+/+ shows hypomyelination and tomacula. These changes are virtually impossible to detect with diffraction on embedded nerves owing to processing-induced structural distortion added to any intrinsic structural heterogeneity resulting from pathologic processes. By contrast, in freshly dissected nerve and, to a lesser extent, in fixed nerve, the change in myelin period would likely be detected by diffraction as altered positions of the reflections and the hypomyelination as a decrease in the M/(M+B) ratio; however, the tomacula would be undetected unless they comprised a significant portion of the nerve volume.

Utility of X-Ray Diffraction Analysis From Fixed Nerve

Our analysis establishes the utility of diffraction measurements from glutaraldehyde-“stabilized” myelin, although myelin spacing and membrane packing is altered (11). Certain other treatments cause phase separation in myelin (e.g. glycerol [42], DMSO [43-45], calcium and tetracaine [44, 45]) such that the resulting, coexisting protein-rich and lipid-rich membrane arrays have periods that differ by as much as ∼60 Å; however, simple glutaraldehyde fixation does not induce such a phase change (11), and does not result in a significant decrease in diffracting power (compare M/(M+B) in Tables 1, 2; Fig. 2B). Thus, as shown here, even in aldehyde-fixed nerves, we could discern small differences in period and diffracting power between nerves having different levels of P0 or PLP (Figs. 2B, 3). For example, EM does not detect a difference between WT and mycP0^+/− in thin sections from embedded nerve (4); however, the diffraction from aldehyde-fixed nerves clearly shows an ∼5-Å increase in the tg (Fig. 2B, green triangles). When EM does detect a difference in period (e.g. a 12% increase in mycP0/P0⁻ compared with WT), x-ray diffraction is much more effective at quantitating the specific alteration in period and also the localization of that difference. Moreover, we showed that although the fixation promotes disorder in membrane packing, particularly in swollen myelin, the fact of swelling was still preserved and measurable by diffraction. One explanation for the multiple myelin periodicities for P0^null nerve reported in EM studies (1) would be that the aldehyde enhances the differences among myelin sheaths having different membrane packing (periods) owing to deficiency in the amount of P0 adhesion protein.

The greater lability of the extracellular apposition in the transgenic mice having less P0 is indicated by the formation, after overnight treatment in physiological saline, of a swollen, second phase (212 Å) in P0^+/− (Table 2), which subsequently became the only structure. In wild-type myelin, a slightly expanded structure of 180 Å period was observed (data not shown) after overnight treatment. This effect was particularly manifest in sciatic nerves dissected from animals greater than 15 months of age (Table 2; Fig. 7). In this case, the P0^+/− myelin was least stable compared with the WT and mycP0^+/+ myelins, presumably owing to weakened intermembrane interactions arising from the loss or absence of appropriate P0-P0 contacts in the heterozygote.

In hypotonic saline at alkaline pH, both P0^+/− and mycP0^+/+ swelled like the wild-type, likely because the amount of P0 in the membranes is not absolutely crucial for defining a precise equilibrium separation within the broad net energy minimum for swollen membranes (30).

Genetic Background Affects Period

The 173 Å-period that we measured for control sciatic nerve myelin from mice with the FVB/N background differs somewhat from those of mice of other strains, for example, 176 Å for C57BL/6J or BALB/CWt (11) and 179 Å for DDY (R. L. Avila, unpublished observations). Are these differences significant? Assuming accurate calibration of the specimen-to-detector distance, then from myelin diffraction studies published over the past approximately 70 years, we know that measurements on individual sharp diffraction spectra are accurate to ±0.5 Å (35, 46-51). Moreover, under apparently identical conditions (animal species and strain, pH, ionic strength, temperature, medium composition), sample-to-sample differences in period are ±1 to 2 Å for myelins having the greatest packing regularity (25, 35, 52), whereas for less well-ordered myelin, the period is within ±3 of the mean value (11). Previous x-ray diffraction results also suggest that genetic background affects the period (36), and genetic background has also been proposed to account for the observed differences in behavior, biochemistry, and white matter morphology between 2 different breeding stocks of shi^mld mice (53). Because the period of native myelin is determined by molecular contacts between apposed surfaces as well as by a balance between attractive and repulsion forces, then strain-dependent differences in period could arise from small but significant differences in biochemical composition.

Summary and Future Perspectives

Static images from EM analysis of myelin structure in fixed and embedded material do not reveal the range and complexity of packing defects and, in fact, can be misleading or wrong (as cited previously). As shown here, x-ray diffraction analysis can reveal this complexity and distinguish types of packing defects, allowing us to better correlate the severity of phenotype with particular types of mispacking; and monitoring myelin period in whole, unfixed nerve as a function of time postdissection can be used to gauge myelin stability. Moreover, the analysis of diffraction from aldehyde-fixed nerves can be of great utility as an initial “screen” for abnormalities in internodal myelin.

Two novel parameters for assessing myelin structural integrity in whole nerve were developed in this article. The first one measures the relative amount of ordered myelin in whole nerve by calculating the fraction of total diffracted intensity that is the result of the myelin or M/(M+B). By normalizing to the total diffracted intensity of myelin (M) plus background (B), this calculation should be independent of differences in sample size, exposure times, incident beam shape and intensity, and detector. Typically, we observed a decrease in the parameter with increasing period (see Fig. 2B for fixed myelin and Table 2 for unfixed myelin), which can be accounted for by an increase in disorder of the membrane packing arising from a shallower or broader potential energy minimum between interacting membranes when the width of the aqueous space between their surfaces increases (30, 38). Comparison of this measure among myelins having the same period indicates whether a nerve is hypomyelinated (or hypermyelinated). For example, comparing unfixed nerves, we note that periods for mycP0^+/+ and P0^+/− myelins are the same or nearly the same as for WT, but the M/(M+B) values are almost half that for WT, suggesting significant hypomyelination in the tg nerves. Similarly, comparison of this measure among nerves having the same “relative” amount of myelin can indicate whether modification of some myelin constituent affects the period. Variation of M/(M+B) values among WT optic nerve was considerable. Although the HSR optic nerves appeared to be about half the value for HSF sciatics, the CCF optic was comparable in value to that for the CCF sciatics (Table 1). An explanation for this difference is not apparent at this time. Nonetheless, application of this measurement to nerves during development could provide a useful and novel quantitative approach to tracking the degree of myelination.

The second parameter for assessing myelin structural integrity is what we termed the “quality index,” which is calculated as the product of parameters related to the relative number of myelin layers and the regularity of the packing. These parameters, which are based on the integral widths of the reflections, are measured from a plot of w² versus h⁴ (see “Materials and Methods”). Thus, myelin with the “highest” quality would consist of highly ordered and thick sheaths. Based solely on the breadths of the reflections and the period, this measure should also be independent of differences in sample size and exposure times. In the current study, this proved to be a useful way to rank comparatively myelin structure among the unfixed and fixed tg nerves.

In future studies, we anticipate exploring to what extent x-ray diffraction measurements of myelin instability correlate with pathologic findings in nerves of acute demyelinating fibers and onion bulbs. Furthermore, comparing the packing defect induced by mutant P0 in PNS myelin to that of the same mutant P0 in PLP^null-CNS myelin might indicate how other PNS- or CNS-specific myelin proteins could modulate the packing defects. A broader understanding of packing defects, their effects on myelin stability, and their amelioration by other myelin constituents could help to provide a firmer basis for strategies aimed to improve myelin stability in neuropathies. Moreover, validation of results from x-ray crystallographic analysis depends crucially on the dimensions of PNS myelin in its chemically unmodified, intact state, as demonstrated in rationalizing the 3-dimensional structure for the extracellular domain of P0 (41). Thus, future crystallographic analyses on mutant forms of P0 that attempt to understand demyelination in terms of abnormal P0-P0 adhesion will also require the appropriate membrane diffraction data from intact tg systems that model the human disease. Our analysis of internodal myelin in transgenic mice demonstrates the utility of x-ray diffraction in quantitating the amount and structural integrity of internodal myelin, and the packing, interactions, and stability of the constituent membranes. Advancing this approach by probing the stability of myelin to changes in pH and ionic strength is expected to yield further insights on the relation between mutations in the genes for myelin membrane adhesion proteins and alterations in intermembrane interactions involving these proteins.

Acknowledgments

The authors thank Ms. Cinzia Ferri, Dr. Deepak Sharma, and Mr. Xiao (Tony) Luo for helpful discussions and experimental assistance; Mr. Zaid Haddadin for his assistance in calibrating the position-sensitive detector; Prof. Rudolf Martini for the gift of the P0^null mice; Prof. Tom Seyfried for the gift of the control DDY mice; and the anonymous reviewers for their constructive comments.

Author notes

Supported by NIH NINDS NS39650, institutional support from Boston College, and a Fulbright Senior Research Scholar Award from the Binational US-Italian Fulbright Scholar Program (DAK); NIH NS41319 and NS45630, Telethon, Italy, Great Britain Multiple Sclerosis Society, and the European Community (LW, MLF); and NIH NS38186 (BDT).