Molecular and Cellular Neuroscience 21, 341–351 (2002)
E-cadherin Is Required for the Correct Formationof Autotypic Adherens Junctions of the OuterMesaxon but Not for the Integrity of MyelinatedFibers of Peripheral Nerves
Peter Young,*,† Oreda Boussadia,‡ Philipp Berger,* Dino P. Leone,*Patrick Charnay,§ Rolf Kemler,‡ and Ueli Suter*,1
*Institute of Cell Biology, Department of Biology, Swiss Federal Institute of Technology,ETH Honggerberg, CH-8093 Zurich, Switzerland; †Department of Neurology, University ofMunster, Munster, Germany; ‡Max-Planck Institute of Immunobiology, Freiburg, Germany;and §INSERM U368, Ecole Normale Superieure, Paris, France
The calcium-dependent adhesion protein E-cadherin ispresent in noncompacted regions of myelin sheaths in theperipheral nervous system. There, it is localized to elec-tron-dense structures between membranes of the sameSchwann cell referred to as autotypic adherens junctions.It has been suggested that the failure of E-cadherin-me-diated adhesion might cause demyelination that proceedsin certain pathological states. To test the requirement ofE-cadherin in peripheral nerves, we used tissue-specificgene ablation techniques based on the Cre/LoxP system.We show that E-cadherin deficiency does not cause sig-nificant demyelination up to the age of 15 months. Immu-nostainings for nodal sodium channels, the paranodalprotein Caspr1, and the juxtaparanodal potassium chan-nels Kv1.1 and Kv1.2 revealed that E-cadherin is not nec-essary to maintain the general functional architecture ofthe nodal region. On the ultrastructural level, we detecteda widening of the outer mesaxon accompanied by a lossof electron-dense cytoplasmic areas. We conclude thatE-cadherin is required for the proper establishmentand/or the maintenance of the outer mesaxon in myelin-ated PNS fibers but is dispensable for proper nerve func-tion.
INTRODUCTION
Myelinated nerve fibers of the peripheral nervoussystem (PNS) consist of large-caliber axons that are
tightly enwrapped by continuous membrane layers ofsingle Schwann cells. Compaction of these layers leadsto the formation of the myelin sheath, which contains aunique set of proteins (Arroyo et al., 2001; Snipes andSuter, 1995). Compaction does not occur, however, insome parts of myelinated fibers. These noncompactedmyelin domains include the outer and the inner mes-axon, the Schmidt–Lanterman incisures, and the para-nodes flanking the nodes of Ranvier. These regionscontain a second specific set of Schwann cell membraneproteins (reviewed by Scherer and Arroyo, 2002). Sev-eral proteins located in noncompacted myelin domainsare involved in Schwann cell–axon interactions thatcontribute to the regulation of the accumulation of po-tassium channels in the juxtaparanodal region (Ras-band et al., 1998, 1999) as well as the assembly of clus-ters of sodium channels at the node of Ranvier (Vabnicket al., 1996; reviewed by Peles and Salzer, 2000; Schererand Arroyo, 2002). In addition, functional gap junctionsare present in Schmidt–Lanterman incisures and para-nodes, allowing small molecules to diffuse between theadaxonal and the perinuclear Schwann cell cytoplasm(Balice-Gordon et al., 1998).
How the Schwann cell membranes are maintained in aconnected but noncompacted shape is not clear. Myelin-associated glycoprotein (MAG) may contribute to thestructural integrity but a good candidate protein is alsoE-cadherin, a glycoprotein of the classical cadherin familythat is thought to mediate cell adhesion in epithelia (Schuh
doi:10.1006/mcne.2002.1177
MCN
et al., 1986; Takeichi, 1991; reviewed by Steinberg andTo whom correspondence should be addressed at the Institute of
Cell Biology, ETH Honggerberg, CH-8093 Zurich, Switzerland. Fax:
McNutt, 1999). E-cadherin is found in association with
341
subplasmalemmal electron-dense plaques resembling ad-herens junctions in the paranodes, in the Schmidt–Lanter-man incisures, and at the outer mesaxon (Fannon et al.,1995). In classical adherens junctions, E-cadherin plays amajor role due to its ability to form homophilic complexes,thereby connecting two neighboring membranes. In theunique situation of the myelinated peripheral nerve, how-ever, E-cadherin appears to form adherens structures be-tween the same contiguous Schwann cell membrane andthus, the term “autotypic adherens-type junction” hasbeen coined (Fannon et al., 1995). The hypothesis has beenput forward that adhesion mediated by E-cadherin is anessential component in the architecture of myelinated pe-ripheral nerves and E-cadherin-dependent junctionalcomplexes might be important stabilization sites in myeli-nation (Fannon et al., 1995). There is evidence from in vitrostudies that E-cadherin is necessary for cell adhesion (re-viewed by Kemler and Ozawa, 1989; Ozawa and Kemler,1998; Thoreson et al., 2000) but data from in vivo experi-ments are scarce. Mice lacking E-cadherin are not viabledue to their inability to form a trophoectoderm epitheliumand adhesion failure at the late morula stage (Larue et al.,1994). In human carcinogenesis, an association betweenE-cadherin deficiency and breast cancer, gastric cancer,and prostate cancer has been described (reviewed byArias, 2001).
To analyze the functional consequences of E-cadherindeficiency for myelinated PNS fibers in vivo, we gener-ated mice specifically lacking E-cadherin in myelinatingSchwann cells. We used the Cre/LoxP system and com-bined a floxed E-cadherin allele, an E-cadherin nullallele, and the Cre recombinase under the control of theKrox20 regulatory region (Voiculescu et al., 2001). In thePNS, Krox20 is expressed in myelinated nerve fibersand the Cre recombinase has been “knocked” into theendogenous Krox20 locus (Topilko et al., 1994; Voi-culescu et al., 2000; Garratt et al., 2000). The resultingE-cadherin ablation in peripheral nerves leads to theloss of electron-dense structures in the outer mesaxonof myelinated fibers and a widened gap in the outermesaxon between the two opposing membranes of thesame Schwann cell. However, no effect on myelin in-tegrity was observed.
RESULTS
Assessment of E-cadherin Ablation Efficiency
Mutant Krox20�/Cre; E-cad�/lox mice were behaviorallyindistinguishable from control animals with the geno-type Krox20�/Cre; E-cad�/lox. To obtain a measure for the
efficiency of E-cadherin ablation, we analyzed the sci-atic nerves of 20-week-old mice by immunohistochem-istry using teased fiber preparations (Figs. 1a–1d). Thistechnique has the distinct advantage that it allows res-olution at the cellular level due to the strictly main-tained 1:1 relationship between the myelinatingSchwann cell and the axon. Approximately 80% of themyelinated fibers were devoid of E-cadherin immuno-reactivity (Figs. 1b and 1e) while E-cadherin was de-tected in each myelinated fiber of control animals (Fig.1a). The recombination frequency was further evalu-ated by crossing the Krox20Cre/� animals into theR26R reporter mouse line (Soriano, 1999). We detected�-galactosidase activity, as visualized by X-gal staining,in almost all fibers of the sciatic nerve of the resultingdouble-transgenic animals (data not shown). Interest-ingly, we also observed reporter gene expression inlamina II of the spinal cord (Fig. 2a), consistent with Crerecombinase expression in a subpopulation of dorsalroot ganglia (DRG) neurons (G. Maro, O. Voiculescu, J.Cohen, P. Charnay, and P. Topilko, manuscript in prep-aration; Fig. 2a). Since E-cadherin is found in presyn-aptic terminals of DRG neurons on lamina II (Shi-mamura et al., 1992), we have also analyzed mutantKrox20Cre/�; E-cad�/lox mice for loss of E-cadherin in thisstructure. As expected, E-cadherin was detected in con-trol animals (Fig. 2b), but not in mutants (Fig. 2c). Inaddition, E-cadherin was absent from DRG (Fig. 2c anddata not shown).
Analysis of the Organization of the Nodeof Ranvier
Based on the prominent expression of E-cadherin inthe paranodes, E-cadherin deficiency might disturb theintegrity of the nodes of Ranvier. Thus, we next deter-mined the localization of the juxtaparanodal potassiumchannel Kv1.1 (Figs. 3a–3c) and of sodium channels atthe node (Figs. 3g–3i). Both proteins were correctlylocalized in fibers lacking E-cadherin. Similarly, stain-ing with antibodies against MAG, which colocalizeswith E-cadherin at the paranodes and in Schmidt–Lan-terman inscisures (data not shown) (Trapp et al., 1984),revealed no significant alteration of MAG distributionin E-cadherin-negative myelinated fibers (Figs. 3d–3f).To analyze the integrity of the nodal region in moredetail, we performed triple immunostainings for theparanodal protein Caspr1, the juxtaparanodal potas-sium channel Kv1.2, and E-cadherin. No detectable dif-ferences in the localization of these marker proteinswere found between control and E-cadherin-deficientmutant nerves (Fig. 4). �-Catenin, an intracellular bind-
342 Young et al.
FIG. 1. Immunohistochemical localization of E-cadherin in control (a) and E-cadherin-deficient (mutant) teased sciatic nerve fibers (b) with thecorresponding phase pictures (c, d). Arrows indicate nodes of Ranvier. Quantification is given in (e). “n” indicates number of animals analyzed.Size bar in d for a–d represents 20 �m.
343E-cadherin Deficiency in Peripheral Nerves
ing partner of E-cadherin (Fannon et al., 1995), showedalso a normal distribution in both mutant and controlanimals (data not shown).
Morphological Alterations in Peripheral Nerves
Genetic alterations of proteins located in the plasmamembrane of myelinating Schwann cells are often as-sociated with demyelination and axonal damage (re-viewed by Young and Suter, 2001; Martini, 2001; Bergeret al., 2002; Maier et al., 2002). Intriguingly, MAG andconnexin32, which, like E-cadherin, are both located innoncompacted regions of myelin, appear to be crucialfor myelin and axonal maintenance in older animals(Anzini et al., 1997; Fruttiger et al., 1995; Yin et al., 1998).Thus, we have analyzed the morphology of E-cadherin-deficient and control sciatic nerves using semithin sec-tions at three progressive ages, postnatal day 15 (P15),20 weeks, and 15 months. At P15, myelination wascomparable in control and mutant mice (Figs. 5a and5b). Also, in young adult animals (20 weeks; data notshown) and in 15-month-old mice, myelin thickness,axon caliber, and number of myelinated fibers werenormal in mutants and controls (Figs. 5c and 5d). Nosigns of axonal damage were observed.
To further evaluate the consequences of E-cadheringene ablation on peripheral nerves, we carried out ul-trastructural analysis of the sciatic nerves of mutant andcontrol animals. Close inspection of areas that normallyexpress E-cadherin revealed an alteration of the outermesaxon in mutant nerves suggesting an increased gapbetween the two apposing membrane layers (Figs. 6aand 6b). We quantified this difference in comparison tocontrol nerves by measuring the gap in three indepen-dent preparations from three different mutant and con-trol animals. This analysis revealed that the gap wassignificantly widened in mutant nerves compared tothose of control animals (Fig. 6c). Furthermore, none ofthe outer mesaxons in mutant mice with a widening ofthe intermembranous gap showed the electron-densestructures that are common in control nerves (Figs. 6a
FIG. 3. Immunohistochemical localization of E-cadherin, Kv1.1, MAG, and sodium channel in teased fibers of E-cadherin-deficient sciaticnerves of adult (20 weeks old) mice. While E-cadherin is lacking in all three preparations (a, d, g), antibodies against Kv1.1 (b), MAG (e), andNa channel (h) showed a normal distribution. Phase images (c, f, i) reveal the localization of nodal regions of the myelinated fibers as indicatedby arrows. Arrowheads indicate Schmidt–Lanterman incisures. Size bars in c, f, i, for a–i represent 20 �m.FIG. 4. Triple immunohistochemical localization of E-cadherin, Caspr1, and Kv1.2 on teased nerve fibers of adult (20 weeks old) control andE-cadherin-deficient (mutant) mice. (a, c, e, and g) Normal localization of all three proteins in control mice. In mutant mice, E-cadherin is missing(b), while unchanged localization of Kv1.2 (d) and Caspr1 (f) is observed at the juxtaparanode and paranode, respectively (merged in h). Arrowindicates node of Ranvier, arrowheads mark Schmidt–Lanterman incisures. Size bar in h represents 20 �m for a–h.
FIG. 2. Recombination visualized by X-gal staining in the spinalcord and DRG of an adult (20 weeks old) Krox20Cre/�; R26R reporteranimal (a). Immunohistochemical localization of E-cadherin in laminaII of control spinal cord (b) and ablation of E-cadherin expression inthe mutant (c). Blue staining in (b) and (c) indicates nuclei stained byDAPI. Arrowheads indicate lamina II, bidirectional arrow showsorientation. D, dorsal; V, ventral. Size bar in c for a–c represents 100�m.
344 Young et al.
345E-cadherin Deficiency in Peripheral Nerves
and 6b). No difference was found in the width of themajor dense line periodicity that was assessed as inter-nal control.
To detect potential alterations in paranodal loops ofE-cadherin-deficient mice, we compared these struc-tures from mutant mice to control mice. The generalappearance of the paranodal loops was normal (Fig.7b). In 44% of well-preserved loops (n � 9) of mutantnerves, we could not detect stacks of electron-denseadherens structures. However, this was also observedin 60% of the loops (n � 10) of control nerves, probablydue to technical reasons. Similarly, the structure ofSchmidt–Lanterman incisures was not obviously al-tered in mutant nerves (Fig. 7a). Although these datasuggest that the adherens junctions in peripheral nervesare not altered by the lack of E-cadherin (without reach-
ing statistical significance and considering the incom-plete elimination of E-cadherin; Fig. 1), minor defects inthe paranodal loops of large caliber axons and inSchmidt–Lanterman incisures cannot be formally ex-cluded due to potential preparation and fixation arti-facts of the examined specimens.
DISCUSSION
The maintenance and the integrity of the myelinsheath in the PNS are dependent on various proteinsthat are located in different domains of the myelinatednerve. The crucial functions of some of these proteinsare particularly evident for the myelin-associated pro-teins whose genes are mutated in demyelinating forms
FIG. 5. Semithin cross sections through the sciatic nerves from 15-day-old (P15) (a, b) and 15-month-old (c, d) control and E-cadherin-deficient(mutant) animals, stained with alkaline toluidine blue. No difference were observed. Size bar in d for a–d represents 10 �m.
346 Young et al.
of Charcot–Marie–Tooth disease (reviewed by Youngand Suter, 2001; Berger et al., 2002). Other importantcomponents have emerged recently from transgenicmouse studies aimed at the elucidation of Schwanncell–axon interactions and the molecular organizationof the node of Ranvier (reviewed by Martini, 2001; Pelesand Salzer, 2000; Bhat et al., 2001; Boyle et al., 2001).However, the role of some proteins that may be cru-cially involved in the structure and function of periph-eral nerves might not be assessed easily by generatingnull mutations in mice due to the requirement for anactive gene during embryogenesis. E-cadherin falls intothis latter category (Larue et al., 1994). Conditional geneablation techniques have been developed to overcome
these problems (reviewed by Lewandoski, 2001). In thisstudy, we have used the Cre/LoxP system to ablateE-cadherin in a tissue-specific manner from myelinatednerve fibers. Recombination and thus inactivation of theE-cadherin gene was mediated by Krox20-driven Cre-recombinase, and the subsequent loss of E-cadherinprotein could be assessed readily on the cellular leveldue to the advantages of teased nerve fiber prepara-tions. The high efficiency of our system, indicating ap-proximately 80% recombination in the adult sciaticnerve, is in line with an earlier report in which the sameKrox20/Cre line was used for tissue-specific ablation ofthe ErbB2 receptor causing a severe phenotype (Garrattet al., 2000).
FIG. 6. Electron microscopy of the outer mesaxon formation in myelinated nerves in cross sections of control (a) and E-cadherin-deficient(mutant) sciatic nerves (b) from 20-week-old animals. Quantification of the width of the outer mesaxon gap in three animals (c). Arrowheadsindicate the two apposed membranes of the outer mesaxon. Note the electron-dense deposits aligned with the membranes in (a) which arelacking in (b). M, compacted myelin. Asterisks indicate statistical significance (P � 0.05). OM, outer mesaxon; MDLP, major dense lineperiodicity. “n” indicates number of outer mesaxons analyzed. Size bar in b for a and b represents 20 nm.
347E-cadherin Deficiency in Peripheral Nerves
We have found that E-cadherin is necessary for aproperly formed outer mesaxon and its accompanyingelectron-dense structures. This is the region where theouter tongue of the Schwann cell cytoplasm is sealed tothe underlying abaxonal Schwann cell membrane. Whyis the phenotype restricted to this unique structure ofthe myelinated nerve while other normally E-cadherin-expressing structures appear unaffected? The lattercould be attributed to technical difficulties since theSchmidt–Lanterman incisures and paranodal loops arefragile structures to be analyzed in detail in mousenerves. Alternatively, other cadherins may be able tocompensate for the loss of E-cadherin as suggested, inanalogy, by the observation that N-cadherin deficiencycan be compensated for by E-cadherin in mice (Luo etal., 2001). We found this possibility difficult to analyzedue to the lack of reliable antibodies in the mousesystem. However, other cadherins that could poten-tially compensate are expressed by Schwann cells, in-cluding N-cadherin, which plays an important role inthe formation of Schwann cell–Schwann cell junctions
and Schwann cell process growth in alignment withaxons (Wanner and Wood, 2002). In addition, othernoncadherin adhesion proteins may substitute for theloss of E-cadherin in the mutant nerves. MAG emergesas a particularly good candidate since it is normallyexpressed in Schmidt–Lanterman incisures and theparanodes but is not present at the outer mesaxon(Trapp et al., 1984). Finally, the finding that the struc-tural integrity of the myelinated nerves was not affectedby the observed defects indicates that myelinatednerves are not crucially dependent on proper adherensjunctions in the outer mesaxon.
An important issue for the interpretation of the de-tected phenotype concerns whether the observationsare due to effects on development, maintenance, orboth. Our analysis revealed that Krox20/Cre-mediatedloss of E-cadherin is already complete before E-cad-herin expression can be observed in distinct structuresof noncompacted myelin in wild-type nerves (data notshown; Menichella et al., 2001). Therefore myelinationper se is likely not to be dependent on E-cadherin
FIG. 7. Electron microscopy analysis of Schmidt–Lanterman incisures (a) and the paranodal region (b) in E-cadherin-deficient 20-week oldmice. The inset in b shows paranodal loops of normal shape. Size bar a and b represents 20 nm. A, axon; N, node of Ranvier. Size bar in insetrepresents 1 �m.
348 Young et al.
expression. Further indirect support for such an inter-pretation is provided by the finding that E-cadherin-deficient mutant animals regenerate normally after sci-atic nerve crush, with remyelination and full behavioralrecovery (data not shown).
In summary, the specific lack of E-cadherin in periph-eral nerves is associated with a widening of the outermesaxon without apparent consequences for the molec-ular architecture and functional long-term integrity ofthe myelin sheath. Development and regeneration ofmyelinated nerve fibers are likely not to depend cru-cially on the presence of E-cadherin.
EXPERIMENTAL METHODS
Generation of Conditional E-cadherin-Deficient Mice
Animals carrying the Krox20/Cre allele (Voiculescu etal., 2000) and the E-cadherin null allele (Larue et al.,1994) were intercrossed to obtain Krox20Cre/�; E-cad�/wt
mice. These mice were subsequently crossed with ani-mals homozygous for LoxP sites flanking the E-cad-herin gene (E-cadlox/lox) (Boussadia et al., 2002). Mutantanimals from this breeding, KroxCre/�; E-cad�/lox (mu-tant), were compared to littermates with the genotypeKrox20Cre/�; E-cad�/lox (control). Genotyping was carriedout on DNA extracted from tail biopsies. The analysis ofconditional E-cadherin-deficient animals was per-formed on a mixed C57BL/6/129 genetic background.The ROSA26R reporter transgenic line was kindly pro-vided by Dr. P. Soriano (Soriano, 1999)
Immunohistochemistry
Animals were anesthetized by intraperitoneal injec-tion of sodium pentobarbital (100–150 �g/g bodyweight) and subsequently transcardially perfused with0.25 mg/ml heparin in phosphate-buffered saline (PBS)followed by 4% paraformaldehyde (PFA) in 0.1 M ca-codylate buffer (pH 7.4) for 20 min. Sciatic nerves weredissected and teased in PBS including 0.1% TritonX-100. Teased nerve fibers were mounted on glassslides and air dried overnight. For quantification, 50single fibers were analyzed in which at least one inter-nodal segment was detectable. Teased fibers were re-hydrated with PBS for 10 min followed by blockingnonspecific binding with 10% goat serum, Fab frag-ments of goat antibodies against mouse IgG (H�L, 0.13mg/ml), and IgM � chain-specific (0.14 mg/ml) (bothJackson Laboratories, West Grove, PA) in 0.1% Triton
X-100. Incubation with different combinations of pri-mary antibodies was carried out for 12 h at room tem-perature. Mouse monoclonal antibodies against MAG,myelin basic protein (MBP) (Boehringer Mannheim, In-dianapolis, IN), and Kv1.2 (Upstate Biotechnology,NY); rat monoclonal antibodies against E-cadherin(Vestweber and Kemler, 1984); and rabbit polyclonalantibodies against Kv1.1, Na channel (III–IV) (gifts fromDr. B. Tempel) (Wang et al., 1995), and Caspr1 (kindlyprovided by Dr. Elior Peles) (Arroyo et al., 1999; Poliaket al., 2001) were used. For MBP staining, fibers werepermeabilized for 1 h in 100% Triton X-100 prior toincubation with the primary antibody. After the incu-bation, nerve fibers were washed three times for 5 minin PBS containing 0.1% Tween 20 (Sigma). Donkey anti-rabbit FITC, goat anti-rat Cy3, and goat anti-mouse Cy3(Jackson Laboratories) were used as secondary antibod-ies. Incubation was carried out for 1 h at room temper-ature. Primary antibodies and secondary antibodieswere diluted in blocking buffer without Fab fragments.After the incubation with secondary antibodies, nervefibers were washed in PBS and subsequently mountedin AF1 (Citifluor, Canterbury, UK) containing DAPI tovisualize nuclei. Triple stainings for E-cadherin, Caspr1,and Kv1.2 were performed on nerve fibers which werefixed in Zambonies fixative for 10 min prior to teasing.Secondary anti-rabbit A594 antibodies and anti-mouseA488 antibodies were used (Molecular Probes, Leiden).Secondary anti-rat horseradish peroxidase-conjugatedantibodies were applied followed by tyramine amplifi-cation and final signal detection with Pacific blue-con-jugated streptavidin (Molecular Probes, Leiden) in anal-ogy to Mueller et al., (2000, 2001). Photographs weretaken with conventional fluorescence microscopy usinga Hamamatsu color chilled 3CCD camera or a SPOT IIcamera system (Visitron, Munich). Figures were assem-bled with Adobe PhotoShop 4.0.
Electron Microscopy and Quantification of theWidth of the Outer Mesaxon
For electron microscopy, three 20-week-old mutantanimals and as controls the corresponding littermateswere used. After anesthesia and initial perfusion with0.25 mg/ml heparin in 0.1 M PBS, a fixative solutioncontaining 4% PFA, 2% glutaraldehyde in 0.1 M caco-dylate buffer (pH 7.4) was applied. Sciatic nerves weredissected, postfixed for 24 h at 4°C, osmicated for 2 h in2% osmium tetroxide at room temperature, washed indistilled H2O several times, dehydrated in ascendingacetone, and finally embedded in Spurr’s medium(Martini et al., 1995). Half-micrometer sections were cut
349E-cadherin Deficiency in Peripheral Nerves
on an Ultracut 200 microtome (Leica, Wetzlar). Semi-thin sections were stained with alkaline toluidine blue.For electron microscopy, ultrathin sections were cut at70 nm, mounted on copper grids, and contrasted withuranyl citrate and lead citrate. Photographs were takenon a JEOL 200C electron microscope and negativeswere analyzed for quantification. The nearest distancebetween the two membranes of the outer mesaxon for-mation was measured. As an internal control and tominimize potential artifacts due to different prepara-tions, the average width between 10 subsequent majordense lines was determined in each sample and takenas reference for the major dense-line periodicity.
Statistical Analysis
Statistical analysis was performed with the Mann–Whitney U test (two-sided) using Statview 4.0 software.P values of �0.05 were considered statistically signifi-cant.
ACKNOWLEDGMENTS
We thank Drs. Ned Mantei, Lukas Sommer, and Verdon Taylor forhelpful discussions and for critically reading the manuscript. We arealso grateful to Elior Peles, Phil Soriano, and Bruce Temple for pro-viding reagents and Joke Nowitzki for excellent technical assistance.This work was supported by grants of the Swiss National ScienceFoundation, the Swiss Muscle Disease Foundation, the Wolfermann-Nageli Stiftung, and the Swiss Bundesamt for Science related to theCommission of the European Communities specific RTD programme“Quality of Life and Management of Living Ressources,” QLK6-CT-2000-00179. P.Y. was supported by a fellowship from the DeutscheForschungsgemeinschaft (DFG YO48/1-1).
REFERENCES
Anzini, P., Neuberg, D. H., Schachner, M., Nelles, E., Willecke, K.,Zielasek, J., Toyka, K. V., Suter, U., and Martini, R. (1997). Struc-tural abnormalities and deficient maintenance of peripheral nervemyelin in mice lacking the gap junction protein connexin 32. J.Neurosci. 17: 4545–4551.
Arias, A. M. (2001). Epithelial mesenchymal interactions in cancer anddevelopment. Cell 105: 425–431.
Arroyo, E. J., Xu, T., Poliak, S., Watson, M., Peles, E., and Scherer, S. S.(2001). Internodal specializations of myelinated axons in the centralnervous system. Cell Tissue Res. 305: 53–66.
Arroyo, E. J., Xu, Y. T., Zhou, L., Messing, A., Peles, E., Chiu, S. Y., andScherer, S. S. (1999). Myelinating Schwann cells determine theinternodal localization of Kv1.1, Kv1.2, Kvbeta2, and Caspr. J. Neu-rocytol. 28: 333–347.
Balice-Gordon, R. J., Bone, L. J., and Scherer, S. S. (1998). Functionalgap junctions in the Schwann cell myelin sheath. J. Cell Biol. 142:1095–1104.
Berger, P., Young, P., and Suter, U. (2002). Molecular cell biology ofCharcot–Marie–Tooth disease. Neurogenetics 4: 1–15.
Bhat, M. A., Rios, J. C., Lu, Y., Garcia-Fresco, G. P., Ching, W., StMartin, M., Li, J., Einheber, S., Chesler, M., Rosenbluth, J., et al.(2001). Axon–glia interactions and the domain organization of my-elinated axons requires neurexin IV/Caspr/Paranodin. Neuron 30:369–383.
Boussadia, O., Kutsch, S., Hierholzer, A., Delmas, V., and Kemler, R.(2002). E-cadherin is a survival factor for the lactating mouse mam-mary gland. Mech. Dev. 115: 53–62.
Boyle, M. E., Berglund, E. O., Murai, K. K., Weber, L., Peles, E., andRanscht, B. (2001). Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve.Neuron 30: 385–397.
Fannon, A. M., Sherman, D. L., Ilyina-Gragerova, G., Brophy, P. J.,Friedrich, V. L., Jr., and Colman, D. R. (1995). Novel E-cadherin-mediated adhesion in peripheral nerve: Schwann cell architecture isstabilized by autotypic adherens junctions. J. Cell Biol. 129: 189–202.
Fruttiger, M., Montag, D., Schachner, M., and Martini, R. (1995).Crucial role for the myelin-associated glycoprotein in the mainte-nance of axon–myelin integrity. Eur. J. Neurosci. 7: 511–515.
Garratt, A. N., Voiculescu, O., Topilko, P., Charnay, P., and Birch-meier, C. (2000). A dual role of erbB2 in myelination and in expan-sion of the Schwann cell precursor pool. J. Cell Biol. 148: 1035–1046.
Kemler, R., and Ozawa, M. (1989). Uvomorulin–catenin complex:Cytoplasmic anchorage of a Ca2�-dependent cell adhesion mole-cule. BioEssays 11: 88–91.
Larue, L., Ohsugi, M., Hirchenhain, J., and Kemler, R. (1994). E-cadherin null mutant embryos fail to form a trophectoderm epithe-lium. Proc. Natl. Acad. Sci. USA 91: 8263–8267.
Lewandoski, M. (2001). Mouse genomic technologies: Conditionalcontrol of gene expression in the mouse. Nat. Rev. Genet. 2: 743–755.
Luo, Y., Ferreira-Cornwell, M., Baldwin, H., Kostetskii, I., Lenox, J.,Lieberman, M., and Radice, G. (2001). Rescuing the N-cadherinknockout by cardiac-specific expression of N- or E-cadherin. Devel-opment 128: 459–469.
Maier, M., Berger, P., and Suter, U. (2002). Understanding Schwanncell–neuron interactions: The key to Charcot–Marie–Tooth disease?J. Anat. 200: 357–366.
Martini, R. (2001). The effect of myelinating Schwann cells on axons.Muscle Nerve 24: 456–466.
Martini, R., Zielasek, J., Toyka, K. V., Giese, K. P., and Schachner, M.(1995). Protein zero (P0)-deficient mice show myelin degenerationin peripheral nerves characteristic of inherited human neuropa-thies. Nat. Genet. 11: 281–286.
Menichella, D. M., Arroyo, E. J., Awatramani, R., Xu, T., Baron, P.,Vallat, J. M., Balsamo, J., Lilien, J., Scarlato, G., Kamholz, J., et al.(2001). Protein zero is necessary for E-cadherin-mediated adherensjunction formation in Schwann cells. Mol. Cell. Neurosci. 18: 606–618.
Mueller, M., Wacker, K., Hickey, W. F., Ringelstein, E. B., and Kiefer,R. (2000). Colocalization of multiple antigens and specific DNA. Anovel method using methyl methacrylate-embedded semithin se-rial sections and catalyzed reporter deposition. Am. J. Pathol. 157:1829–1838.
Mueller, M., Wacker, K., Ringelstein, E. B., Hickey, W. F., Imai, Y., andKiefer, R. (2001). Rapid response of identified resident endoneurialmacrophages to nerve injury. Am. J. Pathol. 159: 2187–2197.
Ozawa, M., and Kemler, R. (1998). The membrane-proximal region ofthe E-cadherin cytoplasmic domain prevents dimerization and neg-atively regulates adhesion activity. J. Cell Biol. 142: 1605–1613.
350 Young et al.
Peles, E., and Salzer, J. L. (2000). Molecular domains of myelinatedaxons. Curr. Opin. Neurobiol. 10: 558–565.
Poliak, S., Gollan, L., Salomon, D., Berglund, E. O., Ohara, R., Ranscht,B., and Peles, E. (2001). Localization of Caspr2 in myelinated nervesdepends on axon–glia interactions and the generation of barriersalong the axon. J. Neurosci. 21: 7568–7575.
Rasband, M. N., Trimmer, J. S., Peles, E., Levinson, S. R., and Shrager,P. (1999). K� channel distribution and clustering in developing andhypomyelinated axons of the optic nerve. J. Neurocytol. 28: 319–331.
Rasband, M. N., Trimmer, J. S., Schwarz, T. L., Levinson, S. R.,Ellisman, M. H., Schachner, M., and Shrager, P. (1998). Potassiumchannel distribution, clustering, and function in remyelinating rataxons. J. Neurosci. 18: 36–47.
Scherer, S. S., and Arroyo, E. J. (2002). Recent progress on the molec-ular organization of myelinated axons. J. Peripheral Nerv. Syst. 7:1–12.
Schuh, R., Vestweber, D., Riede, I., Ringwald, M., Rosenberg, U. B.,Jackle, H., and Kemler, R. (1986). Molecular cloning of the mousecell adhesion molecule uvomorulin: cDNA contains a B1-relatedsequence. Proc. Natl. Acad. Sci. USA 83: 1364–1368.
Shimamura, K., Takahashi, T., and Takeichi, M. (1992). E-cadherinexpression in a particular subset of sensory neurons. Dev. Biol. 152:242–254.
Snipes, G., and Suter, U. (1995). Molecular anatomy and genetics ofmyelin proteins in the peripheral nervous system. J. Anat. 186:483–494.
Soriano, P. (1999). Generalized lacZ expression with the ROSA26 Crereporter strain. Nat. Genet. 21: 70–71.
Steinberg, M. S., and McNutt, P. M. (1999). Cadherins and theirconnections: Adhesion junctions have broader functions. Curr.Opin. Cell Biol. 11: 554–560.
Takeichi, M. (1991). Cadherin cell adhesion receptors as a morphoge-netic regulator. Science 251: 1451–1455.
Thoreson, M. A., Anastasiadis, P. Z., Daniel, J. M., Ireton, R. C.,Wheelock, M. J., Johnson, K. R., Hummingbird, D. K., and Reyn-olds, A. B. (2000). Selective uncoupling of p120(ctn) from E-cad-herin disrupts strong adhesion. J. Cell Biol. 148: 189–202.
A., Chennoufi, A. B., Seitanidou, T., Babinet, C., and Charnay, P.(1994). Krox-20 controls myelination in the peripheral nervous sys-tem. Nature 371: 796–799.
Trapp, B. D., Quarles, R. H., and Suzuki, K. (1984). Immunocytochem-ical studies of quaking mice support a role for the myelin-associ-ated glycoprotein in forming and maintaining the periaxonal spaceand periaxonal cytoplasmic collar of myelinating Schwann cells.J. Cell Biol. 99: 594–606.
Vabnick, I., Novakovic, S. D., Levinson, S. R., Schachner, M., andShrager, P. (1996). The clustering of axonal sodium channels duringdevelopment of the peripheral nervous system. I. Neurosci. 16:4914–4922.
Vestweber, D., and Kemler, R. (1984). Rabbit antiserum against apurified surface glycoprotein decompacts mouse preimplantationembryos and reacts with specific adult tissues. Exp. Cell Res. 152:169–178.
Voiculescu, O., Charnay, P., and Schneider-Maunoury, S. (2000). Ex-pression pattern of a Krox-20/Cre knock-in allele in the developinghindbrain, bones, and peripheral nervous system. Genesis 26: 123–126.
Voiculescu, O., Taillebourg, E., Pujades, C., Kress, C., Buart, S., Char-nay, P., and Schneider-Maunoury, S. (2001). Hindbrain patterning:Krox20 couples segmentation and specification of regional identity.Development 128: 4967–4978.
Wang, H., Allen, M. L., Grigg, J. J., Noebels, J. L., and Tempel, B. L.(1995). Hypomyelination alters K� channel expression in mousemutants shiverer and trembler. Neuron 15: 1337–1347.
Wanner, I. B., and Wood, P. M. (2002). N-cadherin mediates axon-aligned process growth and cell–cell interactions in rat Schwanncells. J. Neurosci. 22: 4066–4079.
Yin, X., Crawford, T. O., Griffin, J. W., Tu, P., Lee, V. M., Li, C., Roder,J., and Trapp, B. D. (1998). Myelin-associated glycoprotein is amyelin signal that modulates the caliber of myelinated axons.J. Neurosci. 18: 1953–1962.
Young, P., and Suter, U. (2001). Disease mechanisms and potentialtherapeutic strategies in Charcot–Marie–Tooth disease. Brain Res.Brain Res. Rev. 36: 213–221.
