|Home | About | Journals | Submit | Contact Us | Français|
Caspr and Caspr2 regulate the formation of distinct axonal domains around the nodes of Ranvier. Caspr is required for the generation of a membrane barrier at the paranodal junction (PNJ), whereas Caspr2 serves as a membrane scaffold that clusters Kv1 channels at the juxtaparanodal region (JXP). Both Caspr and Caspr2 interact with protein 4.1B, which may link the paranodal and juxtaparanodal adhesion complexes to the axonal cytoskeleton. To determine the role of protein 4.1B in the function of Caspr proteins, we examined the ability of transgenic Caspr and Caspr2 mutants lacking their 4.1-binding sequence (d4.1) to restore Kv1 channel clustering in Caspr and Caspr2 null mice, respectively. We found that Caspr-d4.1 was localized to the PNJ and is able recruit the paranodal adhesion complex components contactin and NF155 to this site. Nevertheless, in axons expressing Caspr-d4.1, Kv1 channels were often detected at paranodes, suggesting that the interaction of Caspr with protein 4.1B is necessary for the generation of an efficient membrane barrier at the PNJ. We also found that the Caspr2-d4.1 transgene did not accumulate at the JXP, even though it was targeted to the axon, demonstrating that the interaction with protein 4.1B is required for the accumulation of Caspr2 and Kv1 channels at the juxtaparanodal axonal membrane. In accordance, we show that Caspr2 and Kv1 channels are not clustered at the JXP in 4.1B null mice. Our results thus underscore the functional importance of protein 4.1B in the organization of peripheral myelinated axons.
The segregation of ion channels to distinct membrane domains at and around the nodes of Ranvier is essential for fast and efficient nerve conduction in myelinated axons. In each of these domains, ion channels are found in protein complexes that include cell adhesion molecules (CAMs), adapter proteins and cytoskeletal components (Poliak and Peles, 2003; Salzer, 2003; Hedstrom and Rasband, 2006). The nodes of Ranvier are bordered by the paranodal junctions (PNJ) that form between the axon and the paranodal loops of oligodendrocytes or myelinating Schwann cells. The formation and maintenance of the PNJ depends on the presence of an axonal CAM complex containing Caspr and contactin, as well as on the presence of the 155 kDa isoform of neurofascin (NF155) at the paranodal glial membrane (Bhat et al., 2001; Boyle et al., 2001; Gollan et al., 2003; Sherman et al., 2005; Zonta et al., 2008; Pillai et al., 2009). The PNJ separates the nodal Na+ channels from Shaker-like Kv1.1 and Kv1.2 K+ channels at the juxtaparanodal region (JXP) found at the ends of the compact myelin internode. At this site, Kv1 channels are localized and associated with an axonal cell adhesion complex consisting of Caspr2 and TAG-1, which bind to TAG-1 present on the glial membrane (Poliak et al., 1999; Traka et al., 2002; Poliak et al., 2003; Traka et al., 2003). Caspr and Caspr2 are both important for the accumulation of K+ channels at the JXP. Caspr is required for the formation of a membrane barrier at the PNJ, in its absence, Kv1 channels aberrantly accumulate adjacent to the nodes (Bhat et al., 2001; Boyle et al., 2001; Gollan et al., 2003). In contrast, Caspr2 serves as a membrane scaffold that clusters Kv1 channels at the JXP. Accordingly, in Caspr2 null mice, these channels are distributed along the internodes (Poliak et al., 2003; Traka et al., 2003).
The cytoplasmic region of Caspr and Caspr2 both contains a short sequence that mediates their binding to 4.1 proteins (Menegoz et al., 1997; Peles et al., 1997; Poliak et al., 1999; Gollan et al., 2002; Denisenko-Nehrbass et al., 2003). 4.1 proteins are cytoskeletal adapters that link adhesion molecules (Nunomura et al., 1997; Ward et al., 1998; Biederer and Sudhof, 2001; Zhou et al., 2005; Yang et al., 2009), ion channels (Coleman et al., 2003; Li et al., 2007; Stagg et al., 2008; Lin et al., 2009) and receptors (Binda et al., 2002; Lu et al., 2004; Beekman et al., 2008; Rose et al., 2008) to the actin/spectrin cytoskeleton, thereby playing an important role in membrane organization (Bennett and Baines, 2001). Of the four members of the protein 4.1 family (i.e., 4.1R, 4.1N, 4.1G and 4.1B), protein 4.1B is expressed in myelinated axons, where it is located along the internodes, PNJ and the JXP region (Ohara et al., 2000; Poliak et al., 2001; Gollan et al., 2002; Arroyo et al., 2004; Ogawa et al., 2006). We have previously suggested that the association between Caspr and protein 4.1B stabilizes Caspr at the plasma membrane by linking the PNJ adhesion complex to the actin-rich cytoskeleton found at this site (Gollan et al., 2002). Interestingly, Neurexin IV, the fly homologue of Caspr and Caspr2 (Baumgartner et al., 1996), as well as Coracle, the fly orthologue of protein 4.1, are both necessary for the generation of septate junctions (Baumgartner et al., 1996; Ward et al., 1998), suggesting that the function of Caspr and Caspr2 in the organization of myelinated axons may require protein 4.1B. In the present work, we use an in vivo transgenic rescue and gene knockout approaches to elucidate the role 4.1 proteins play in the organization of myelinated axons in the peripheral nervous system (PNS) by Caspr and Caspr2.
HA-tagged constructs were generated by PCR using human Caspr and Caspr2 cDNAs: Caspr full-length (C1FL), Casprd4.1 (C1d4.1, deletion of amino acids Q1307 to Y1315), Caspr2 full-length (C2FL) and Caspr2d4.1 (deletion of amino acids R1298 to H1304). In Caspr constructs, the HA tag (amino acids YPYDVPDYAS) was inserted at the carboxy terminus at position E1384 for C1FL and after E1375 for C1d4.1. In Caspr2 constructs, the same sequence was inserted at the extracellular region adjacent to the transmembrane domain between positions N1255 and G1256. For expression and analysis in mammalian cultured cells, these constructs were cloned into the expression vector pCi-neo (Promega, Inc). Generation of a myc-tagged protein 4.1B has been previously described (Gollan et al., 2002). For pull-down experiments, the cytoplasmic tails of Caspr and Caspr2 (full lengths and the respective 4.1 deletions) were cloned into pGEX-6P1 (GE Healthcare). For generation of transgenic mice, Caspr and Caspr2 constructed genes were cloned into a Thy1.2 expression cassette (Gollan et al., 2002; Horresh et al., 2008), linearized, and introduced by pronuclear injection into fertilized eggs derived from CB6F1 mice. Pseudo pregnant CD-1 albino females were used as foster mothers for embryo transfer. Founder mice were genotyped for the presence of the transgenes by PCR using primers 5’-CTGCTGAGCTAGCTGAGGC-3’ and 5’-TCACGATGCGTAGTCAGGC-3’, and for Caspr2-transgenes using primers 5’-GCGTAGTCAGGCACATCGTATGGG-3 ’ and 5 ’-ATACCACGGTTACTCCTGCGATTGC-3’. Founders were further crossed with CB6F1 mice and interbred to generate lines. Transgenic mice were routinely identified by PCR of tail genomic DNA, using the appropriate primers derived from human Caspr or Caspr2 and hemagglutinin (HA) -tag. The same primers were also used for RT-PCR analyses of dorsal root ganglia (DRG) RNA isolated from the transgenic mice. Based on immunoprecipitation performed from brains of transgenic mice, two lines from each genotype were chosen to be further crossed with Caspr-null or Caspr2-null mice (Gollan et al., 2003; Poliak et al., 2003) until homozygote-null mice for the endogenous Caspr or Caspr2 and positive for the relevant transgene were obtained. Generation and genotyping of 4.1B-null mice have been described before (Yi et al., 2005). Wild type control animals were derived from the same litters. All experiments were performed in compliance with the relevant laws and institutional guidelines and were approved by the Weizmann Institute’s Institutional Animal Care and Use Committee.
Dorsal root ganglia (DRG) were removed from WT, transgenic mice and 4.1B-null mice and were immediately frozen in liquid nitrogen. Total RNA was isolated using TRI reagent (Sigma-Aldrich) and cDNA was synthesized with Super-Script II reverse transcriptase (Invitrogen). PCR reactions were done with the following primers: 4.1B: 5’-AGACCAGGAGGCTGAGCC-3 ’ and 5 ’-GCCGTACTCTGACCCATCCA-3 ’, 4. 1 G : 5 ’-AAGAAGGGCTCTGACCAGGC-3 ’ and 5 ’-CCACCTCTGCCTTCAGCTCA-3’, 4. 1 N : 5 ’-AAGGCTCAGGAGGAGACTCC-3’ and 5’-CGGCCGTGCTTCTCCACTT-3 ’, 4. 1 R : 5 ’-AGAGTTTAGCGGCTGAAGC-3’ and 5’-CGGGCTTCTGGGAGGCTT-3’, Actin: 5’-TCATACCGGGCACACTTCAC-3’ and 5’-CATTTGGGAGGGTCTGTGGC-3’, Caspr 5’-ATTGCGGAGCGCTGGGGAGAGG-3’ and 5’ GAGAGGGAAGGGTGGATAAGGAC-3’, and for the endogenous Caspr2 gene 5’-TGCTGCTGCCAGCCCAGGAACTGG-3’ and 5’ TCAGAGTTGATACCCAGCGCC-3’.
For pull-down experiments HEK293 were transfected with cDNA encoding myc-tagged protein 4.1B by calcium phosphate. After 48h, the cells were lysed using a buffer containing 50mM Tris, pH 7.5, 150mM NaCl, 1mM EGTA, 10% glycerol and 1% T×100, supplemented with a protease inhibitors cocktail (Sigma-Aldrich). Cleared lysates were allowed to bind GST-fusion proteins pre-bound to Agarose-Glutathione beads (Sigma-Aldrich) overnight at 4°C. Complexes were then washed three times in a buffer containing 0.1% T×100, separated on 7.5% SDS-PAGE, transferred to a nitrocellulose membrane and reacted with anti-myc antibody. For comparison of protein levels, three pairs of sciatic nerves of each genotype were dissected, desheathed in cold PBS and frozen in liquid nitrogen. The nerves were then ground with a mortar and pestle, solubilized in 150µL of buffer containing 25mM Tris, pH=8.0, 1mM EDTA and 2% SDS, boiled for 15 min and then spun for 20 min at 15°c. Protein concentration was determined using a BCA protein assay reagent (Pierce Chemical Co.). Sample (50 µg) were loaded onto a 10% SDS-gel followed by Western blot analysis (Poliak et al., 1999). Cell surface biotinylation of HEK-293 transfected cells was performed as described before (Gollan et al., 2002).
A detailed list of the antibodies used in this study is provides in the (supplementary material Table S1). A rat polyclonal antibody against NrCAM was generated by immunizing rats with an Fc-fusion protein containing the 6 Immunoglobulin and 2 of the fibronectin-like domains of mouse NrCAM. For the generation of anti-4.1N antibody, rabbits were immunized with a GST fusion peptide containing amino acids 1–321 of mouse protein 4.1N. A mouse polyclonal antibody to the extracellular domain (ECD) of Caspr2 was made by immunizing Balb/C mice with Caspr2-Fc fusion protein (Poliak et al., 2003). A rabbit polyclonal antibody against TAG1 was generated by immunizing rabbits with an Fc fusion protein containing the fibronectin-like domains of rat TAG-1 (residues 610–1040). Immunofluorescence analysis of transfected cells and teased sciatic nerves was done as previously described (Horresh et al., 2008). Sciatic nerves were dissected and immersed in freshly made 4% paraformaldehyde for 30 minutes at 4°C, or for 10 minutes in Zamboni’s fixative at ambient temperature. After desheathing, nerves were teased using fine forceps on Super-Frost Plus slides, air-dried for 2–3 hours and then kept frozen at −20°C until used. Slides were postfixed for 2–10 minutes using cold methanol, washed with PBS, incubated for 45 minutes in blocking solution (5% fish skin gelatin and 0.1% triton X-100), and then incubated overnight at 4°C with the different primary antibodies diluted in blocking solution. For staining with anti-NF155 antibody, nerves were postfixed for 1 minute with Bouin’s fixative (Sigma) instead of methanol. For protein 4.1B staining, slides were left in blocking solution containing 0.5% Triton X-100 for 2 hours at ambient temperature before adding the antibody. Slides were washed with PBS and incubated for 1 hr in blocking solution containing the secondary antibodies: anti mouse-Cy3, anti rat Cy5 (Jackson Immuno Research), or anti rabbit-Alexa-488 (Invitrogen). Triple immunolabeling was done using antibodies from three different species. Fluorescence images were obtained using an Axioskop 2 microscope equipped with Apotom imaging system (Carl Zeiss), or a Nikon eclipse E1000 microscope fitted with a Hamamatsu ORCA-ER CCD camera. Quantification of the paranodal size was performed using Volocity 4.2.1 image analysis software (Improvision). Mean paranodal area was calculated for each genotype (n=600) and statistical significance was validated using a two-tail student T-Test. Quantification of Kv1.2 distribution was done by counting the number of sites in which it was present at the JXP, PNJ, or both sites, and presented as a percentage of the total number of nodal sites examined (n=150).
In myelinated axons, protein 4.1B is located at the PNJ and JXP regions together with Caspr and Caspr2, respectively (Ohara et al., 2000; Denisenko-Nehrbass et al., 2003; Ogawa et al., 2006). Furthermore, it has been previously suggested that the localization of protein 4.1B in these axons might be regulated by Caspr proteins (Poliak et al., 2001; Gollan et al., 2002; Arroyo et al., 2004). To directly determine the role Caspr and Caspr2 play in the localization of protein 4.1B in myelinated axons, we compared the distribution of protein 4.1B in sciatic nerves of wild type and Caspr or Caspr2 null mice. In wild type animals, protein 4.1B immunoreactivity was detected along the internodes, and was often stronger at the PNJ and JXP (Fig. 1A–B). Notably, protein 4.1B immunoreactivity was highly specific as determined by labeling of teased sciatic nerves isolated from protein 4.1B null mice (Yi et al., 2005) using the same antibody (Fig. 1G–H). Similar to wild type nerves, in sciatic nerve fibers isolated from Caspr−/− or Caspr2−/− mice, protein 4.1B was absent from the nodes but detected at the PNJ, JXP and the internodes (Fig. 1C–F). These results demonstrate that the localization of protein 4.1B in PNS myelinated axons occurs independently from Caspr or Caspr2.
Caspr and Caspr2 interact with 4.1 proteins through a short sequence found in their cytoplasmic domain (Menegoz et al., 1997; Peles et al., 1997; Gollan et al., 2002; Denisenko-Nehrbass et al., 2003). To determine whether this interaction is required for the function of Caspr proteins in the molecular organization of the axolemma, we have constructed deletion mutants of Caspr and Caspr2 lacking their 4.1-binding motif (C1d4.1 and C2d4.1; Fig. 2A), and tested whether they could restore the clustering of Kv1 channels at the JXP when placed on the genetic background of Caspr or Caspr2 null mice. As controls, we used full length Caspr (C1FL) and Caspr2 (C2FL) cDNAs. All constructs also contained a hemagglutinin (HA) tag (Fig. 2A). We next performed a pulldown experiments using GST-fusion proteins containing the cytoplasmic domains of the four constructs from HEK-293 cells expressing protein 4.1B. As shown in Fig. 2B, protein 4.1B was associated with the cytoplasmic domains of the full length, but not with the mutant Caspr and Caspr2. Transfection of Caspr2 constructs, or Caspr constructs along with contactin cDNA into HEK-293 cells revealed that all proteins were efficiently expressed on the cell surface, as was evident by cell surface biotinylation, followed by immunoprecipitation with an antibody to HA-tag and blotting with streptavidin-HRP (Fig. 2C). Similarly, antibodies directed to the extracellular region of Caspr and Caspr2 recognized all the respective proteins in intact, non-permeabilized cells (Fig. 2D), further demonstrating that they were expressed on the cell surface.
In order to express Caspr and Caspr2 transgenes in neurons, they were cloned downstream from a Thy1.2 promoter (Gollan et al., 2002; Horresh et al., 2008). Transgenic mice were generated by pronuclear injection, and crossed with mice lacking either Caspr or Caspr2, in order to obtain Caspr−/−/C1FL, Caspr−/−/C1d4.1, Caspr2−/−/C2FL, and Caspr2−/−/C2d4.1 transgenic animals. RT-PCR analysis of dorsal root ganglia mRNA isolated from the four genotypes using specific primer sets that recognize each of the transgenes confirmed that they are expressed in sensory peripheral neurons (Fig. 2E). In contrast to wild type mice, no expression of the endogenous Caspr and Caspr2 transcripts was detected in either Caspr2−/−/tg or Caspr2−/−/tg lines, respectively, further demonstrating that the transgenes were expressed on a null background. To examine the expression of the transgenes in peripheral nerves, we performed immunoblot analysis of sciatic nerves isolated from wild type, Caspr or Caspr2 nulls, or the four transgenic mice lines using an HA tag antibody, as well as antibodies directed against the intracellular domain of the two Caspr proteins. As depicted in Fig. 2F, all the respective transgenic proteins were detected in sciatic nerves isolated from the transgenes but not from the wild type or the null mice. Thus, similarly to the full-length transgenes (i.e., C1FL and C2FL), Caspr and Caspr2 mutants lacking their protein 4.1-binding sequence (i.e., C1d4.1 and C2d4.1) are targeted to the axon in mice lacking their respective endogenous protein.
To determine whether the interaction between Caspr and 4.1 proteins is required for its targeting to the PNJ, we immunolabeled teased sciatic nerves isolated from wild type mice, Caspr−/−, or Caspr nulls expressing the C1d4.1 (Caspr−/−/C1d4.1) or the FL (Caspr−/−/C1FL) transgenes, using antibodies to Caspr (Fig. 3A–D). Similarly to wild type nerve, but in contrast to Caspr null, both C1FL and C1d4.1 were located at the PNJ in axons expressing these transgenes (about 70% and 45% of the axons expressed the C1FL and C1d4.1 transgenes, respectively). Nevertheless, while the average area of Caspr-positive paranodes in Caspr−/−/C1FL nerves was 23.4±3 µm2 (n=600), it only reached the size of 15.9±2 µm2 (n=600) in Caspr−/−/C1d4.1 mice, suggesting that stable expression of Caspr at the PNJ requires its binding to protein 4.1B (Gollan et al., 2002). Consistent with previous studies (Bhat et al., 2001; Gollan et al., 2003), both contactin and NF155, which are part of the paranodal adhesion complex (Charles et al., 2002; Gollan et al., 2003; Bonnon et al., 2007), were not concentrated at the PNJ in Caspr null mice (Fig. 3F,J). In contrast, both were present at the PNJ in sciatic nerves isolated from Caspr−/−/C1d4.1 and Caspr−/−/C1FL mice. These results demonstrate that the expression of either C1FL or the C1d4.1 transgene on the background of Caspr null nerves resulted in the recovery of all paranodal adhesion components at the PNJ.
The presence of Caspr is required for the generation of a paranodal membrane barrier that separates Kv1 channels and other juxtaparanodal proteins from the nodal axolemma (Bhat et al., 2001; Pedraza et al., 2001; Gollan et al., 2003; Rosenbluth, 2009). Accordingly, we proceeded to examine whether expression of the C1FL and C1d4.1 transgenes could rescue the membrane barrier function of the PNJ. In peripheral nerves of wild type mice Kv1.2 was clustered at the JXP, while in Caspr−/− nerves this K+ channel subunit accumulated at the paranodal region adjacent to the nodes of Ranvier (Fig. 4A–B). Immunofluorescence analysis of sciatic nerves isolated from the two transgenic mice showed that while Kv1.2 was sequestered at the JXP in Caspr−/−/C1FL as in wild type mice, in Caspr−/−/C1d4.1 nerves it was detected at the JXP and more weakly at the PNJ, where it partially overlapped with the C1d4.1 transgene (Fig. 4C–D). In general, the distribution of Kv1.2 in sciatic nerves of Caspr−/−/C1d4.1 was variable; it was either detected at both JXP and the PNJ (Fig. 4E–F), at the JXP and in a line bordering the nodes (Fig. 4G), or only at the JXP (Fig. 4H). Quantitation of the different cases is presented in Fig. 4I. In addition to Kv1.2, similar results were obtained using antibodies to other juxtaparanodal proteins, including Caspr2 and TAG-1 (data not shown). These results indicate that in contrast to the full-length protein, axonal expression of a Caspr mutant lacking its protein-4.1 binding sequence is not sufficient to completely restore the clustering of K+ channels at the JXP, and suggest that the interaction of Caspr with protein 4.1B is necessary for its proper accumulation at the PNJ and generation of an efficient membrane barrier at this site.
We next examined the significance of the interaction between Caspr2 and 4.1 proteins for its function as a membrane scaffold that retains K+ channels at the JXP (Poliak et al., 2003; Traka et al., 2003). To this end, we labeled teased sciatic nerves isolated from wild type, Caspr2 nulls, or Caspr2 nulls expressing the C2FL or C2d4.1 transgenes, using antibodies to Caspr2 and either Caspr (to mark the PNJ) or to Na+ channels (to mark the location of the nodes) (Fig. 5A–H). In agreement with a recent report (Horresh et al., 2008), the C2FL transgene was localized at the JXP similar to the endogenous protein in the wild type mice, while the Caspr2 mutant lacking its protein 4.1-binding sequence (C2d4.1) was absent from the JXP. The latter transgene was either undetectable, or only weakly detected along the axon. The expression of C2FL at the JXP resulted in a full rescue of Kv1.2 channels at this site (Fig. 5I–P). C2FL-positive JXP also contained the two other juxtaparanodal proteins, TAG-1 (Fig. 5T) and PSD93 (Horresh et al., 2008). In contrast, Kv1.2 and TAG-1 immunoreactivity was reduced to absent in the JXP of myelinated axons from C2d4.1 transgenic mice (Fig. 5K,O, S). Given that the C2d4.1 transcript was present in DRG neurons (Fig. 2E), and that the transgenic C2d4.1 protein was detected in sciatic nerves isolated from Caspr2−/−/C2d4.1 mice (Fig. 2F), these results indicate that the protein 4.1-binding sequence of Caspr2 is essential for its targeting to the JXP.
The above results indicate that the interaction between Caspr proteins and protein 4.1 is required for their function. To further determine the role of these cytoskeletal linkers in myelinated axons, we have examined the organization of the nodal environs of mice lacking protein 4.1B (Yi et al., 2005), a member of the 4.1 protein family that overlaps and associates with both Caspr and Caspr2 (Ohara et al., 2000; Gollan et al., 2002; Denisenko-Nehrbass et al., 2003; Arroyo et al., 2004). These mice developed normally and showed no signs of neurological abnormalities (Yi et al., 2005). Immunofluorescence analysis of teased sciatic nerves isolated form adult 4.1B−/− mice, showed that Na+ channels and gliomedin were present at the nodes of Ranvier (Fig. 6A–J). Caspr and contactin were present at the PNJ, which had comparable size to wild type nerves. Similar results were obtained using an antibody to NF155 (data not shown). However, in marked contrast to wild type animals, Caspr2, TAG-1, PSD93 and Kv1.2 were not concentrated at the JXP in 4.1B−/− nerves. In accordance with previous observations (Arroyo et al., 1999), Kv1.2 was located throughout the internodal region in a double strand that apposes the inner mesaxon of the myelin sheath and forms a circumferential ring just below the inner aspect of the Schmidt-Lanterman incisures (Fig. 6K–M). Remarkably, 4.1B−/− mice exhibit an identical phenotype to mice lacking Caspr2 or TAG1 (Poliak et al., 2003; Traka et al., 2003), further supporting the conclusion that the interaction of Caspr2 with protein 4.1B is required for its localization at the JXP and the assembly of the JXP complex (Fig. 6N).
The finding that 4.1B−/− mice exhibit no paranodal abnormalities is unexpected given the inability of C1d4.1 transgene to fully rescue the Caspr null phenotype (Fig. 4), and thus may suggest that other 4.1 proteins compensate for the absence of protein 4.1B. To test this possibility, we first analyzed the expression of all four protein 4.1 genes in dorsal root ganglia neurons (Fig. 7A). Of the four genes examined (i.e., 4.1R, 4.1N, 4.1G and 4.1B), protein 4.1R, 4.1N and 4.1B transcripts were present in sensory peripheral neurons. We next performed RT-PCR analysis on DRG mRNA isolated from 4.1B−/− and wild type mice using specific primers that recognize the three neuronal 4.1 genes. This analysis revealed a slight increase in the amount of 4.1R transcript in 4.1B−/− compare to wild type mice. Lastly, we immunolabeled teased sciatic nerves of wild type or 4.1B−/− mice using antibodies to protein 4.1R or protein 4.1N (Fig. 7C). Surprisingly, although protein 4.1R was almost undetectable in wild type axons, it was concentrated at the PNJ in the absence of protein 4.1B. These results indicate that protein 4.1R compensates for the loss of protein 4.1B at the PNJ in myelinated 4.1B−/− axons.
A functional nervous system requires the precise localization of ion channels and receptors at distinct membrane domains, where they are optimally positioned for their function. Formation and maintenance of these specialized membrane domains in neurons are achieved by a combination of different sorting and targeting mechanisms, selective endocytosis, and the restriction of lateral diffusion of membrane proteins, as well as by specific retention mechanisms (Lasiecka et al., 2009). The latter two processes are often controlled through the association of transmembrane proteins with the underlying cytoskeleton (Susuki and Rasband, 2008). Three specialized membrane domains are found at and around the nodes of Ranvier, and include the nodes, where Na+ channels are clustered, the PNJ, and the JXP, where Kv1 channels are accumulated (Poliak and Peles, 2003). The formation of the PNJ and the assembly of the JXP complex depend on axon-glia interaction mediated by Caspr and Caspr2, respectively (Bhat et al., 2001; Poliak et al., 2003; Traka et al., 2003). In this study we used a complimentary in vivo transgenic/rescue and gene knockout approaches to show that the cytoskeletal linker protein 4.1B, which interact with both of these CAMs, plays an essential role in their membrane expression and consequently, in the organization of myelinated axons.
Protein 4.1B belongs to a group of cytoskeletal proteins that serves as linkers between the plasma membrane and the actin cytoskeleton (Parra et al., 2000). In myelinated axons in both the PNS and the CNS, protein 4.1B colocalized with Caspr and Caspr2 at the PNJ and the JXP, respectively (Ohara et al., 2000; Poliak et al., 2001; Denisenko-Nehrbass et al., 2003), and thus may provide a link between the axolemma and the axonal cytoskeleton at these sites. During development of myelinated axons, Caspr and Caspr2 appear prior to protein 4.1B, suggesting that they may recruit it to the axolemma (Denisenko-Nehrbass et al., 2003). In agreement, clustering of the Caspr–contactin complex on the cell surface results in the immobilization of protein 4.1B into these clusters (Gollan et al., 2002). Furthermore, the distribution of protein 4.1B is altered in myelin mutants (Poliak et al., 2001; Gollan et al., 2002), as well as after experimental demyelination (Arroyo et al., 2004). Similarly, null mutations in neurexin IV, the Drosophila homologue of Caspr and Caspr2, leads to the complete absence of protein 4.1/coracle from septate junctions (Baumgartner et al., 1996). In contrast, we show that protein 4.1B immunoreactivity in sciatic nerves isolated from either Caspr or Caspr2 null mice was indistinguishable from wild type nerves, arguing against a role for these CAMs in the membrane recruitment of Protein 4.1B. However, it should be noted that the presence of protein 4.1B in Caspr−/− paranodes may be due to the accumulation of Caspr2 at this site. Similarly, the presence of protein 4.1B at the JXP in Caspr2−/− axons could be attributed to the expression of Necl1 adhesion molecule that is located at this site (Maurel et al., 2007; Spiegel et al., 2007). In line with this latter possibility, Necls were shown to interact with 4.1 proteins (Yageta et al., 2002) and to recruit protein 4.1B to synapses (Hoy et al., 2009).
The PNJ serves as a membrane barrier that excludes juxtaparanodal components from the nodes of Ranvier (Rosenbluth, 2009). The formation of such membrane barriers and the assembly of specialized membrane at nodes of Ranvier and the axonal initial segment depend on the underlying actin/spectrin cytoskeleton mesh (Bennett and Healy, 2008; Susuki and Rasband, 2008). At the PNJ, protein 4.1B colocalized and associates with αII and βII spectrin, as well as with ankyrin B, suggesting that these cytoskeletal proteins may be important constituents of this junction (Ogawa et al., 2006). To begin analyzing the contribution of the axonal cytoskeleton to the formation of the PNJ, and particularly to its membrane barrier function, we examined the function of a mutant Caspr lacking its protein 4.1-binding site. We show that this mutant was targeted to the PNJ and was able to assemble the paranodal adhesion complex (i.e., to recruit contactin and NF155 to the PNJ). In accordance with a previous study (Gollan et al., 2002), we found that the paranodal area occupied by Caspr-d4.1 was in average 28% smaller than the area occupied by the Caspr-FL transgene, suggesting that the interaction of Caspr with protein 4.1B is required for its stabilization at the cell surface. In marked contrast to nerves expressing the full length Caspr molecule, in which Kv1 channels were completely eliminated from the PNJ in 80% of the sites, in Caspr-d.4.1 expressing nerves these channels were mainly detected (78% of the sites) at both the PNJ and the JXP. During development of peripheral nerves, Kv1 channels are first detected at the paranodes and are translocated to the JXP with the gradual formation of the PNJ (Vabnick et al., 1999), suggesting that in Caspr-d4.1 expressing axons only a fraction of the paranodal Kv1 channels were excluded towards the JXP. These results indicate that the interaction between Caspr and protein 4.1B is required for the stable expression of Caspr on the paranodal axolemma and the generation of an efficient membrane barrier at this site. Given that Kv1.2 are still well segregated in heterozygous Caspr mice (unpublished results), it is likely that the impaired PNJ barrier in Caspr-d4.1 transgene is due to its inability to interact with protein 4.1B rather than its reduced expression at the PNJ. In support of this suggestion, Kv1 channels are located at paranodes in sciatic nerves of the shambling (shm) mutant mouse, which expresses a truncated Caspr protein lacking its entire cytoplasmic domain (Sun et al., 2009).
Analysis of sciatic nerves isolated form 4.1B−/− mice revealed that Caspr, contactin and NF155 were present at the PNJ. Similarly, Kv1.2, as well as Caspr2 and TAG-1 were well separated from the nodes, indicating that protein 4.1B is not required for the formation of a functional, barrier-forming PNJ. In addition, we found that the localization of ankyrin B, another component of the paranodal cytoskeleton that associate with protein 4.1B (Ogawa et al., 2006), was indistinguishable from wild type nerves (data not shown), further suggesting the formation of a normal PNJ in these mice. We provide a possible explanation for the discrepancy between the 4.1B null and the Caspr-d4.1 transgene by showing that protein 4.1R accumulates at the PNJ in 4.1B−/− mice (Fig. 7C). 4.1R was demonstrated to bind Caspr (Menegoz et al., 1997), and thus could potentially compensate for the absence of 4.1B. Interestingly, an increase in the expression of 4.1G in the heart, as well as 4.1N in T cells, was reported in mice lacking 4.1R (Stagg et al., 2008). Redundancy between 4.1 proteins was also suggested to explain the difference between in vitro (Lin et al., 2009) and in vivo (Wozny et al., 2009) role of 4.1N and 4.1G in the formation of glutamatergic synapses. Given that protein 4.1R accumulates at the PNJ in mice lacking 4.1B, it will require the generation of double 4.1B/4.1R null mice to further understand the role of these proteins in myelinated axons.
Confinement of K+ channels at the JXP is regulated by axon-glia interaction mediated by an axonal adhesion complex of Caspr2 and TAG-1, which bind to TAG-1 present at the adaxonal membrane of myelinating Schwann cells and oligodendrocytes (Poliak et al., 2003; Traka et al., 2003). An important finding of this work is the observation that protein 4.1B is required for the localization of Caspr2 and hence for that of Kv1 channels at the JXP. This conclusion is based on the inability of a Caspr2 mutant lacking its protein 4.1B-binding sequence to accumulate at the JXP, as well as on the absence of Caspr2 at this site in sciatic nerves isolated from protein 4.1B−/− mice. In both Caspr2−/−/C2d4.1 transgenic and 4.1B−/− mice, Kv1 channels, TAG-1 and PSD-93 were not accumulated at the JXP. Notably, Western blot analysis of sciatic nerve lysates using an antibody to Kv1.2 showed that similar amounts of K+ channels are present in wild type and 4.1B−/− mice (data not shown), indicating that the reduction of these channels at the JXP resulted from their redistribution along the internodes. Caspr2 and Kv1.2 were still associated in 4.1B−/− brains (data not shown), indicating that protein 4.1B does not mediate the interaction between these two proteins. Remarkably, in terms of the molecular organization of myelinated axons, 4.1B−/− mice exhibit an identical juxtaparanodal phenotype to Caspr2−/− and TAG-1−/− mutants (Poliak et al., 2003; Traka et al., 2003). In all of these mutants, the nodes or Ranvier and the PNJ formed regularly, while the assembly of the juxtaparanodal domain was impaired.
Protein 4.1B may control the accumulation of Caspr2 at the JXP by regulating its transport, membrane insertion and/or by stabilizing it at the axolemma. In this regard, epithelial cells express a Golgi-specific variant of protein 4.1B that is required for targeting of Na+/K+-ATPase to the plasma membrane (Kang et al., 2009). However, the presence of a Caspr2 mutant lacking its 4.1-binding sequence in axons argues against a role for 4.1B in axonal transport of Caspr2. In addition, protein 4.1B associates with Caspr and Caspr2 only after they have already reached the PNJ and JXP, respectively (Denisenko-Nehrbass et al., 2003), further supporting a role for protein 4.1B in membrane insertion and stabilization of Caspr2 at the juxtaparanodal axolemma.
This work was supported by grants from the National Multiple Sclerosis Society, the NIH (NINDS NS50220), the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation, the Israel Science Foundation, the Shapell Family Biomedical Research Foundation at the Weizmann Institute, and the Wolgin Prize for Scientific Excellence. E.P. is the Incumbent of the Hanna Hertz Professorial Chair for Multiple Sclerosis and Neuroscience.