|
|
|
|
|
|
|
||
Article |
-dystrobrevin
:
roles in skeletal muscle and its neuromuscular and myotendinous junctions
2 Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, MO 63110
3 Department of Cell and Developmental Biology, State University of New York Upstate Medical University, Syracuse, NY 13210
Address correspondence to R. Mark Grady, Dept. of Pediatrics, Washington University School of Medicine, Pediatric Research Bldg., St. Louis, MO 63110. Tel.: (314) 286-2796. Fax: (314) 286-2892. E-mail: grady{at}kids.wustl.edu
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsReferences |
|---|
-Dystrobrevin (DB), a cytoplasmic component of the dystrophin–glycoprotein complex, is found throughout the sarcolemma of muscle cells. Mice lacking DB exhibit muscular dystrophy, defects in maturation of neuromuscular junctions (NMJs) and, as shown here, abnormal myotendinous junctions (MTJs). In normal muscle, alternative splicing produces two main DB isoforms, DB1 and DB2, with common NH2-terminal but distinct COOH-terminal domains. DB1, whose COOH-terminal extension can be tyrosine phosphorylated, is concentrated at the NMJs and MTJs. DB2, which is not tyrosine phosphorylated, is the predominant isoform in extrajunctional regions, and is also present at NMJs and MTJs. Transgenic expression of either isoform in DB-/- mice prevented muscle fiber degeneration; however, only DB1 completely corrected defects at the NMJs (abnormal acetylcholine receptor patterning, rapid turnover, and low density) and MTJs (shortened junctional folds). Site-directed mutagenesis revealed that the effectiveness of DB1 in stabilizing the NMJ depends in part on its ability to serve as a tyrosine kinase substrate. Thus, DB1 phosphorylation may be a key regulatory point for synaptic remodeling. More generally, DB may play multiple roles in muscle by means of differential distribution of isoforms with distinct signaling or structural properties.
Key Words: dystrobrevin; dystrophin; muscular dystrophy; myotendinous junction; neuromuscular junction
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
and ß), sarcoglycans (, ß, 
, 
, and 
), syntrophins (1, ß1, and ß2), dystrobrevins (DBs; and ß), and sarcospan (Ervasti et al., 1990; Yoshida and Ozawa, 1990; for review see Blake et al., 2002). Interest in the DGC stemmed originally from the finding that mutations in genes encoding numerous DGC components cause muscular dystrophy: loss of dystrophin results in Duchenne muscular dystrophy; deficiencies in four of the sarcoglycans have been linked to limb-girdle muscular dystrophies; and mutations in laminin 2, the main extracellular ligand for the DGC, lead to congenital muscular dystrophy (Cohn and Campbell, 2000; Blake et al., 2002).
In addition to its role in myofiber integrity, the DGC is also important in formation or maintenance of two specialized domains on the muscle fiber surface: neuromuscular junctions (NMJs), at which motor axons innervate muscle fibers, and myotendinous junctions (MTJs), at which muscle fibers form load-bearing attachments to tendons. The DGC is dispensable for initial steps in postsynaptic differentiation, but contributes importantly to the maturation and stabilization of the postsynaptic membrane. For example, myotubes lacking dystroglycan, utrophin, 1-syntrophin, or DB have alterations in the density and patterning of acetylcholine receptors (AChRs) embedded within the postsynaptic membrane (Deconinck et al., 1997a; Grady et al., 1997a, 2000; Adams et al., 2000; Jacobson et al., 2001; Akaaboune et al., 2002). Roles of the DGC have been less extensively studied at the MTJ, but several DGC components are concentrated in this region (e.g., dystrophin, utrophin, syntrophin, sarcospan, and nNOS [Chen et al., 1990; Byers et al., 1991; Khurana et al., 1991; Chang et al., 1996; Crosbie et al., 1999]), and MTJs are structurally abnormal in mice lacking dystrophin or both dystrophin and utrophin (Ridge et al., 1994; Deconinck et al., 1997b).
Little is known about how the DGC plays these disparate roles, but one important factor is that its composition varies among cell types (Straub et al., 1999; Loh et al., 2000; Moukhles and Carbonetto, 2001) and, for muscle cells at least, from site to site within a single cell. In muscle fibers, dystrophin is present throughout the sarcolemma, whereas its autosomal homologue, utrophin, is confined to the NMJ and MTJ (Khurana et al., 1991; Ohlendieck et al., 1991). Likewise, each of the three syntrophins has a distinct sarcolemmal distribution (Kramarcy and Sealock, 2000). Here, we focus on another DGC component, DB, to address the issue of how isoform diversity contributes to functional diversity. DB is found throughout the sarcolemma in vertebrate skeletal muscle, where it binds dystrophin, utrophin, and syntrophin (Carr et al., 1989; Wagner et al., 1993; Peters et al., 1997b, 1998; Sadoulet-Puccio et al., 1997). A homologue, ßDB, has been described but is expressed at low levels if at all in skeletal muscle (Peters et al., 1997a; Blake et al., 1998). Previously, we showed that DB-/- knockout mice exhibit both muscular defects (muscular dystrophy) and abnormal NMJs (Grady et al., 1999, 2000; Akaaboune et al., 2002). In addition, we show here that DB-/- mice have malformed MTJs. Thus, DB influences DGC function at three distinct locations within the muscle cell.
Several isoforms of DB are generated by alternative splicing, of which three, DB1–3, have been detected in skeletal muscle (Blake et al., 1996; Sadoulet-Puccio et al., 1996; Enigk and Maimone, 1999; Newey et al., 2001a) They are identical over most of their length (551 aa) but have distinct COOH termini (Fig. 1
A). The 188 aa COOH terminus of DB1 is a substrate for tyrosine kinases in vivo (Wagner et al., 1993; Balasubramanian et al., 1998), whereas common sequences and the short (16 aa) COOH terminus of DB2 do not appear to undergo phosphorylation. DB3 lacks the syntrophin- and dystrophin-binding sites present in the other isoforms and has not been studied in detail. Although both DB1 and DB2 are concentrated at the postsynaptic membrane, their detailed localization differs within the junction; furthermore, only DB2 is present at high levels in extrasynaptic regions (Peters et al., 1998; Newey et al., 2001a). Based on these differences, we hypothesized that DB1 and DB2 might have distinct functions. In addition, based on evidence that tyrosine phosphorylation regulates the plasticity of neuron–neuron synapses (Ali and Salter, 2001), we wanted to test the theory that tyrosine phosphorylation of DB1 affects neuromuscular maturation or structure.
|
DB1 or DB2 in DB-/- mice and analyzed the ability of each isoform to "rescue" the dystrophic, synaptic, and myotendinous phenotypes. We show that either DB1 or DB2 is able to maintain muscle stability but that DB1 is significantly better than DB2 in preventing synaptic and myotendinous defects. We also used sited-directed mutagenesis to show that the enhanced efficacy of DB1 depends in part on its tyrosine phosphorylation. |
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
DB1 or DB2. Transgenic mice were bred to DB-/- mice (Grady et al., 1999, 2000). We studied two transgenic lines that expressed DB1 (12 and 14) and three that expressed DB2 (11A, 11B, and 28). In the text and figures, the designations tgDB1 and tgDB2 refer to lines 12 and 28, respectively. However, the main results were confirmed with the other lines (see Materials and methods).
Immunoblotting with antibodies that recognize all DB isoforms showed that levels of transgene expression were similar in the tgDB1 and tgDB2 lines (Fig. 1 B). Levels of recombinant DB2 in DB-/-,tgDB2 were similar to levels of endogenous DB2 in control mice, whereas levels of recombinant DB1 in DB-/-,tgDB1 muscle were significantly higher then levels of endogenous DB1 but similar to levels of total DB in controls (Fig. 1 B). Immunohistochemical analysis showed that DB1 and DB2 were present in >95% of all muscle fibers in DB-/-,tgDB1 and DB-/-,tgDB2 mice, respectively (Fig. 1 C). In both lines, the transgene was expressed in all skeletal muscles tested, including tibialis anterior, sternomastoid, and diaphragm, and there were no detectable differences among fibers that correlated with fiber type (unpublished data).
Muscular dystrophy
DB-/- mice exhibit a mild muscular dystrophy characterized by degenerating myofibers, infiltrating monocytes, and centrally nucleated regenerating myotubes (Fig. 2)
(Grady et al., 1999). To test whether DB1 and DB2 differ in their ability to maintain muscular integrity, we examined the diaphragm, quadriceps, soleus, sternomastoid, and tibialis anterior muscles of DB-/-,tgDB1 and DB-/-,tgDB2 mice. The diaphragm provided a particularly stringent test because it is the most severely affected muscle in several models of muscular dystrophy, including DB-/- mice (Stedman et al., 1991; Grady et al., 1997b, 1999; Duclos et al., 1998). Rescue by both transgenes was dramatic. No degenerating fibers, regenerating (centrally nucleated) fibers, or infiltration by monocytes were detected in any muscles of either DB-/-,tgDB1 or DB-/-,tgDB2 mice (Fig. 2 and unpublished data). This was true in mice ranging in age from 1–7 mo, in muscles with predominantly fast (type IIB and II) fibers (tibialis anterior, quadriceps, diaphragm, and sternomastoid), and in muscles with predominantly slow (type I and IIA) fibers (soleus). Thus, either DB1 or DB2 alone is capable of maintaining muscle fiber integrity.
|
DB isoformsDB1 and DB2 (Fig. 1 A). Consistent with previous reports, both isoforms are present at higher levels at synaptic sites (marked with rhodamine -bungarotoxin [rBTX], which binds specifically and quasiirreversibly to AChRs) than in extrasynaptic regions of the sarcolemma (Peters et al., 1998; Grady et al., 2000). Likewise, recombinant DB1 and DB2 were concentrated at synaptic sites in DB-/-,tgDB1 and DB-/-,tgDB2 mice, respectively (Fig. 3
A). For DB2, the endogenous and recombinant proteins were distributed similarly. However, whereas endogenous DB1 was barely detectable extrasynaptically, recombinant DB1 was abundant extrasynaptically (Fig. 3 A). Thus, selective localization of DB1 to the NMJ in wild-type muscle cannot be explained solely by its primary sequence.
|
DB1 in DB-/-,tgDB1 muscle might result from its higher than normal levels (Fig. 1 B). An alternative, however, is that restriction of DB1 to synaptic sites requires the presence of DB2. For example, DB2 might have a higher affinity than DB1 for extrasynaptic-binding sites and thereby displace DB1 from such sites. To test this hypothesis, we generated doubly transgenic DB-/- mice that expressed DB1 and DB2 at similar levels. In DB-/-,tgDB1/DB2 mice, DB2 was clearly present extrasynaptically, but its presence failed to reduce extrasynaptic levels of DB1 (Fig. 3 A). Indeed, levels of membrane-associated DB2 immunoreactivity were decreased when DB1 was present, suggesting that DB2 did not bind selectively to extrasynaptic sites.
We also exploited the presence of extrasynaptic DB1 to test whether DB1 depends on the DGC for its localization to the sarcolemma. Levels of endogenous DB, primarily DB2, are greatly reduced in extrasynaptic regions of dystrophin-deficient (mdx) mice (Grady et al., 1997b). However, there is evidence for dystrophin-independent associations of DB with the sarcolemma (Crawford et al., 2000), and these associations might be sufficient to tether DB1 to the membrane. We found that levels of extrasynaptic DB1 remained low in mdx,DB+/-,tgDB1 mice despite the overexpression of recombinant DB1 (Fig. 3 B). Consistent with this result, there was no attenuation of dystrophic symptoms in mdx,DB+/-,tgDB1 compared with mdx,DB+/- mice (unpublished data). Thus unique sequences in DB1 neither mediate a DGC-independent association with the membrane nor attenuate dystrophy in the absence of dystrophin.
AChR distribution and density
Normal adult mouse NMJs are pretzel shaped with an AChR-rich postsynaptic membrane underlying each branch of the nerve terminal. DB deficiency does not greatly affect the overall topology of the NMJ but does affect the arrangement of its AChRs in three ways (Fig. 4)
(Grady et al., 2000; Akaaboune et al., 2002). First, AChRs smoothly outline each branch in controls, whereas branch borders in mutants are fragmented with long finger-like spicules that radiate beyond the terminal's edge. Second, AChRs in control NMJs are enriched along the crests of the junctional folds that invaginate the postsynaptic membrane, leading to a striated appearance within each branch. In DB-/- synapses, in contrast, the distinction between the crests and folds is blurred (shown ultrastructurally in Grady et al., 2000) with receptors forming small, irregularly spaced aggregates or clumps. Finally, the density of AChRs is reduced by 
70% at DB-/- synapses compared with controls.
|
DB1 in mice prevented all three of these defects: AChR distribution and density were indistinguishable from normal at >95% of synaptic sites examined (Fig. 4). Thus, DB1 alone can support postsynaptic maturation. In contrast, in DB-/-,tgDB2 mice examined at 4–6 wk of age the postsynaptic membrane remained abnormal at >90% of synaptic sites: AChRs formed small aggregates comparable to those seen at DB-/- synapses rather than the linear striations characteristic of normal muscle. On the other hand, most DB-/-,tgDB2 postsynaptic sites were distinguishable from those in mice lacking all DB, in that spicules radiating from the edge of the AChR-rich region were either absent or short and AChR density was slightly but significantly greater than in mutants (Fig. 4, B and C). Thus, DB2 supports normal postsynaptic maturation to a lesser degree than DB1.
The difference between DB-/-,tgDB1 and DB-/-, tgDB2 synapses was also evident in 3–6-mo-old animals. However, the percentage of normal-appearing synapses in DB-/-,tgDB2 increased to 50% at 6 mo of age. This age-related increase suggests that the continued presence of even the relatively ineffective DB2 can eventually normalize the structure of initially abnormal synapses. Moreover, it raises the possibility that differences between DB1 and DB2 are quantitative rather than qualitative. In this regard, it is important to note that levels of transgene expression were similar in DB-/-,tgDB1 and DB-/-, DB2 mice (Fig. 1 B). Moreover, studies of transfected muscle fibers, described below, provide independent evidence that DB1-specific sequences play a distinct role.
AChR turnover
AChRs migrate and turn over faster at DB-/- synapses than in controls, supporting the idea that DB acts in part by tethering AChRs to the postsynaptic cytoskeleton (Akaaboune et al., 2002). Because little is known about the molecular mechanisms that regulate AChR turnover, we quantitated rBTX fluorescence in vivo (see Materials and methods) to ask whether DB1 and DB2 had different effects on this process. Adult sternomastoid muscles were labeled with a single nonsaturating dose of rBTX, and individual synapses were imaged. 3 d later, the same synapses were located and imaged again to assess changes in fluorescence intensity. In wild-type mice, 20% of the labeled receptors were lost from the cell surface after 3 d. This degree of loss indicates a t1/2 of 10.5 d, similar to previous reports (Akaaboune et al., 1999, 2002). In DB-/- mice, nearly 60% of AChRs were lost over a similar time, indicating a t1/2 of 2.5 d (Fig. 5)
. AChR stability in DB-/-,tgDB1 mice did not differ significantly from that of controls (t1/2 = 9 d), whereas the t1/2 of AChRs in DB-/-,tgDB2 mice was intermediate between controls and mutants (t1/2 = 5.5 d). Thus, consistent with the results on AChR distribution and density, DB1 alone can support normal AChR turnover, whereas DB2 is only partially effective.
|
1-syntrophin and nitric oxide synthaseDB1 than DB2 to support synaptic maturation, we studied the distribution of two proteins whose concentration at synaptic sites requires DB: 1-syntrophin and neuronal nitric oxide synthase (nNOS). Levels of both are reduced synaptically and extrasynaptically in DB-/- mice (Grady et al., 2000). Moreover, 1-syntrophin mutants have synaptic defects similar to those seen in DB-/- mice (Adams et al., 2000), and studies in vitro have implicated nNOS as a modulator of AChR clustering (Jones and Werle, 2000).
In DB-/-,tgDB1 muscle, synaptic levels of 1-syntrophin and nNOS were similar to those in controls (Fig. 6)
. In contrast, levels of 1-syntrophin and nNOS remained low in many synapses of DB-/-,tgDB2 mice. Interestingly, when viewed en face, DB-/-,tgDB2 synapses with the lowest levels of 1-syntrophin also had the most abnormal AChR distribution (unpublished data). Thus, although either DB isoform can recruit 1-syntrophin and nNOS to the NMJ, the increased efficacy of DB1 over DB2 in maintaining synaptic architecture may reflect in part its increased recruiting ability.
|
DB1 functionDB1 is tyrosine phosphorylated in vivo (Balasubramanian et al., 1998), raising the possibility that tyrosine phosphorylation is important for its synaptic function. To test this idea, we fused GFP to the NH2 terminus of DB1 (GFP-DB1) or to a mutant form of DB1 in which the three major phosphorylated tyrosine residues (aa 698, 706, and 723 [Balasubramanian et al., 1998]) had been changed to phenylalanine by site-directed mutagenesis (GFP-DB1-P3-). Expression of these constructs in heterologous cells showed that only the GFP-DB1 fusion protein underwent tyrosine phosphorylation (Fig. 7
A). Thus, the addition of GFP to DB1 did not preclude its phosphorylation, whereas the mutation of its three tyrosine residues did. The two constructs were introduced into tibialis anterior muscles of 2-wk-old DB-/- mice by electroporation. Muscles were removed 10–14 d later, transfected fibers were identified by GFP fluorescence, and their postsynaptic membranes were analyzed (Fig. 7, B and C).
|
DB-/- myofibers resulted in significant restoration of postsynaptic structure at most synaptic sites, and AChR distribution was qualitatively normal at a subset of NMJs (Fig. 7, D–F). Expression of GFP-DB1-P3- also restored AChR distribution to some extent, but the recovery was significantly less than that seen with GFP-DB1 (p < 0.0001 by Chi-square test), and no synaptic sites appeared normal (Fig. 7, D–F). The variability of "rescue" seen among transfected fibers was greater than that seen in DB-/- transgenic mice, possibly because GFP interfered with DB function or because electroporation led to variable protein levels. Nonetheless, these results indicate a role for tyrosine phosphorylation in the function of DB1.
Interestingly, the qualitative differences between DB-/- fibers transfected with GFP-DB1 and GFP-DB1-P3- paralleled the differences described above between DB-/-, tgDB1 and DB-/-,tgDB2 fibers. Expression of DB1 either transgenically or by transfection resulted in postsynaptic sites with sharp, spicule-free borders and striated instead of granular interiors. In contrast, although expression of either DB2 or GFP-DB1-P3- in DB-/- muscle usually led to restoration of sharp (spicule-poor) borders, interiors remained granular (Fig. 4 B compared with Fig. 7 D). These parallels suggest that the enhanced ability of DB1 over DB2 to support synaptic structure depends on its tyrosine phosphorylation.
The MTJ
In initial studies, we found that both DB1 and DB2 were enriched at the MTJ in wild-type mice (Fig. 8)
. We therefore asked whether DB is required for the integrity of the MTJ. We used EM to address this issue. The muscle membrane at the MTJ is invaginated to form folds that run parallel to the myofiber's long axis (Fig. 9
A). These folds, which are deeper than those of the NMJ, create an interdigitation with the collagen fibrils of the tendon. Folds were also present at mutant MTJs but were significantly shorter than those in controls (Fig. 9, difference between control and DB-/- p < 0.0001 by ANOVA). Thus, DB is important for maintaining normal MTJ architecture.
|
|
DB1 and DB2 differed in their ability to maintain MTJ structure. In DB-/-,tgDB1 mice, the depth of the folds was not significantly different from wild-type (Fig. 9 B; p = 0.25), whereas folds in DB-/-,tgDB2 mice were significantly shorter than those of wild-type or DB-/-,tgDB1 animals (p < 0.0001). Thus, DB1 was significantly more effective than DB2 in maintaining the integrity of both MTJs and NMJs. Because we could not identify transfected fibers in the electron microscope, we were unable to test whether DB1 phosphorylation plays a role in MTJ maintenance. |
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
-sarcoglycan results in muscular dystrophy with little impact upon the NMJ (Hack et al., 1998; Grady et al., 2000; Akaaboune et al., 2002). Conversely, mice lacking utrophin or 1-syntrophin have abnormal synapses with no detectable dystrophy (Deconinck et al., 1997a; Grady et al., 1997a; Kameya et al., 1999; Adams et al., 2000). In contrast, DB is critical for all three DGC functions. In this paper, we show that DB's disparate effects are explained in part through alternative splicing of the DB transcript and selective localization of the two main muscle isoforms, DB1 and DB2.
Muscular dystrophy and DB
In normal muscle, DB2 is the predominant extrasynaptic isoform, suggesting that it plays the primary role in helping the DGC maintain muscle viability. Although we found that expression of DB2 alone in DB-/- mice prevented muscle fiber degeneration, DB1 was equally capable. These findings show that the unique COOH terminus of DB2 is not required for its effect upon the membrane and suggest that shared sequences suffice. Shared domains include sites that mediate binding of DB to dystrophin and syntrophin. Also included are two EF hand domains and a zinc finger region that can potentially mediate other protein–protein interactions (for review see Enigk and Maimone, 2001). Ligands for these sites are unknown but may include novel DB-binding proteins identified recently in yeast two-hybrid screens (Benson et al., 2001; Mizuno et al., 2001; Newey et al., 2001b). Thus, it is likely that both DB1 and 2 can maintain myofiber integrity by attracting similar binding partners to the DGC.
On the other hand, loss of DB1-specific functions at the MTJ may contribute to muscle pathology. DB1 is better able than DB2 to maintain the integrity of the MTJ. The MTJ is the major site of force transmission from muscle fibers to the skeleton, and disruption of this structure may be involved in the pathogenesis of some muscle disorders (Law et al., 1995; Miosge et al., 1999). Previous studies have implicated the DGC in maintenance of the MTJ (Ridge et al., 1994; Deconinck et al., 1997b), and our results suggest that a main role of the DGC at this site may be to recruit DB1.
Localization of DB isoforms
In normal muscle, DB1 is selectively associated with the NMJ and MTJ, but when overexpressed transgenically it was capable of association with extrasynaptic membrane. We therefore wondered what factors account for the normally distinct distributions of the two isoforms. Analysis of mdx mice has shown that the extrasynaptic localization of DB2 requires an intact DGC, so we considered the possibility that the extrasynaptic localization of DB2 seen in normal mice results from the preferential binding of DB2 over that of DB1 to the DGC. This competition might be accentuated by the higher levels of DB2 than DB1 seen in normal muscle (Fig. 1 B). However, analysis of the DB-/-,tgDB1/DB2 double transgenics, in which DB1 and DB2 were expressed at similar levels, showed that DB1 was not dislodged from its extrasynaptic position by DB2. Thus the extrasynaptic predominance of DB2 in normal muscle is not likely to be a result of competition between the isoforms.
Other reasons for the selective distribution of the two isoforms include the possibility that DB1 might be selectively transcribed from synaptic nuclei, analogous to the synaptic expression of AChR subunits (for review see Sanes and Lichtman, 2001). This is unlikely, however, because both isoforms appear to be transcribed from the same promoter (Newey et al., 2001a). Two remaining potential explanations are as follows: First, there may be posttranscriptional localization of DB1 mRNA either by synapse-specific stabilization of the mRNA or by synapse-specific transport of the mRNA using targeting information encoded in the 3' UTR (Newey et al., 2001a). Second, DB3 might play a role. DB3 lacks the syntrophin- and dystrophin-binding sites present in DB1 and 2 (Nawrotzki et al., 1998) but is nonetheless associated with the DGC (Yoshida et al., 2000). Thus, in normal mice DB3 may confine DB1 to the NMJ by blocking its access to extrasynaptic-binding sites. This interaction would not occur in DB-/-,tgDB1 muscle, which lacks DB1–3.
DB and the NMJ
DB is a component of the molecular machinery that stabilizes AChRs within the postsynaptic membrane. In DB mutants, the mobility of synaptic receptors is increased; as a result, there is enhanced flux from the synapse to perijunctional regions, which are sites of AChR internalization (Sanes and Lichtman, 2001; Akaaboune et al., 2002). This mechanism could explain, at least in part, the appearance of the DB-/- postsynaptic membrane in which AChR turnover is abnormally high, density is low, and the distinction between crests and troughs of junctional folds is blurred (Grady et al., 2000; Akaaboune et al., 2002). Here, we show that DB1 is better able than DB2 to rescue these synaptic defects in DB-/- mice, indicating that its unique COOH terminus is important for DB's synaptic function.
How might DB act? One possibility is that DB exerts two distinct effects on the postsynaptic membrane, one mediated by common sequences and one by DB1-specific sequences. Alternatively, common sequences might mediate all synaptic effects, with DB1-specific sequences serving to enhance their efficacy. For example, even though regions common to DB1 and 2 bind both utrophin and -syntrophin, there is some evidence that DB1 associates more tightly with both proteins than does DB2 (Balasubramanian et al., 1998; Peters et al., 1998) (Fig. 6). These differences are likely to be relevant because AChR density is decreased in the absence of utrophin (Deconinck et al., 1997a, Grady et al., 1997a) and mice lacking 1-syntrophin have postsynaptic defects similar to those in DB-/- mice (Adams et al., 2000). Thus, sequences in the unique COOH terminus of DB1 could enhance the ability of common sequences to bind utrophin and syntrophin, thereby enhancing DB1's ability to stabilize the postsynaptic membrane.
DB phosphorylation and synaptic plasticity
Replacement of three tyrosine residues in DB1's unique COOH-terminal domain by phenylalanine decreased its synaptic efficacy. These residues are the major, if not the only sites of tyrosine phosphorylation in DB1 (Balasubramanian et al., 1998) (Fig. 7 A). Our results, therefore, provide strong evidence that DB1 function is modulated by phosphorylation. Phosphorylation might alter the conformation of neighboring sequences to affect their affinity for other proteins (as suggested by modeling studies of Balasubramanian et al., 1998) or may serve to recruit adaptor or signaling proteins as occurs in numerous other phosphoproteins.
Our interest in DB1 phosphorylation stems from the growing realization that even though mature synapses are remarkably stable, they are not inert. Instead, several of their features, most notably the distribution and density of their postsynaptic receptors, can change rapidly and dramatically in response to altered input (Sheng and Lee, 2001). At the NMJ, the t1/2 of AChRs increases, and their density begins to decrease within 1 h after imposition of complete paralysis (Akaaboune et al., 1999). AChR turnover and density are similarly affected in DB-/- mice, suggesting that DB is part of the regulatory mechanism. However, actual loss of DB is unlikely to be a physiological mechanism for such rapid activity-dependent alterations. On the other hand, altered efficacy of DB1 by changes in its phosphorylation state could affect AChR mobility quickly, reversibly, and in an activity-dependent fashion.
Several tyrosine kinases have been implicated in postsynaptic structure: erbB and ephA kinases are concentrated in the postsynaptic membrane; MuSK plays a critical role in postsynaptic differentiation; src and fyn modulate AChR stability in vitro; and trkB affects AChR distribution in vivo (DeChiara et al., 1996; Gonzalez et al., 1999; Buonanno and Fischbach, 2001; Lai et al., 2001; for review see Sanes and Lichtman, 2001; Smith et al., 2001). It will be interesting to learn whether DB1 is a substrate for any of these kinases and whether activators of the kinases (neuregulin for erbBs, ephrinA for ephA, agrin for MuSK, and BDNF for trkB) affect DB1 phosphorylation. In addition, in view of numerous reports implicating tyrosine kinases in central synaptic plasticity and the presence of DGC components, including DB and ßDB, at central synapses (Blake and Kroger, 2000; Levi et al., 2002) it is intriguing to consider the possibility that similar mechanisms act there.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
DB1 expression vector, two cDNA clones (16.1A and 14.1.2 [Enigk and Maimone, 1999]) were isolated from a BC3H1 mouse muscle cDNA library and ligated. The 3' end of the cDNA, including the stop codon and 125 bp of the 3' UTR, was generated by RT-PCR from mouse muscle RNA. The entire coding region was then subcloned first into the SalI/NotI sites of pBK-RSV (Stratagene) and then into a pEtCAT vector, which contains 3.3 kb of regulatory sequences from the mouse muscle creatine kinase gene, extending from -3,300 to +7 with respect to the transcriptional start site (Jaynes et al., 1988) (a gift from S. Hauschka, University of Washington, Seattle, WA). In addition, the vector contains an SV40/polyadenylation sequence. To generate the DB2 expression vector, a full-length clone was isolated from the BC3H1 library, subcloned into the EcoRI site of pBluescript II SK(+) (Stratagene) and then into pEtCAT.
Linearized constructs were injected into C57BL6 oocytes at the Washington University Mouse Genetics Core. Four independent lines of mice carrying the DB1 construct (tgDB1) and six independent lines carrying the DB2 construct (tgDB2) were identified by PCR. Each line was bred onto an DB-/- background (Grady et al., 1999). DB-/-,tgDB1 line was also bred to mdx mice and to DB-/-,tgDB2 mice.
Histology
For bright field microscopy, muscles were frozen in liquid nitrogen–cooled isopentane and cross sectioned in a cryostat at 8 µm; sections were stained with hematoxylin and eosin. For immunohistochemistry, sections from the same muscles were stained with primary antibody diluted in PBS/1% BSA/2% normal goat serum for 2 h and then rinsed with PBS. Sections were then incubated 1 h with a mixture of fluorescein-conjugated goat anti–rabbit IgG (Alexa 488; Molecular Probes) and rBTX (Molecular Probes), rinsed in PBS, and mounted using 0.1% p-phenylenediamine in glycerol/PBS. For en face views, sternomastoid muscles were fixed in 1% PFA in PBS for 20 min, cryoprotected in sucrose, frozen, and sectioned en face at 40 µm. Rabbit polyclonal antibodies to DB1 (DB638), DB2, and 1-syntrophin (SYN17) were gifts from Stanley Froehner, University of Washington (Peters et al., 1997b). Rabbit polyclonal antibodies that recognize all forms of DB (called pan-DB here) were generated using a recombinant fragment of DB and affinity purified using the immunogen (Grady et al., 1997a). Rabbit polyclonal antibody to nNOS was purchased from Immunostar Inc. (no. 24287). Illustrations were prepared in Adobe Photoshop®.
For ultrastructural studies, tibialis anterior muscles were fixed in 4% glutaraldehyde/4% PFA in PBS, washed, refixed in 1% OsO4, dehydrated, and embedded in resin. Thin sections were stained with lead citrate and uranyl acetate. Sections were systematically scanned in the electron microscope, and all MTJs encountered were measured from the micrographs. Muscles from two to four animals were analyzed per genotype.
Immunoblotting
For immunoblots, sternomastoid muscle was homogenized and sonicated in extraction buffer (PBS, 5 mM EDTA, 1% SDS, and protease inhibitors [CompleteMini; Roche]). Protein concentration of whole muscle extract was determined by a BCA protein assay (Pierce Chemical Co.). Equal amounts of protein (50 µg) were resolved on 7.5% SDS–polyacrylamide gels and incubated with monoclonal antibody to DB (D62320; Transduction Laboratories). This antibody was detected with goat anti–mouse IgG1 peroxidase-conjugated secondary antibody (Roche) using ECL (NEN).
Quantitative fluorescence microscopy and in vivo imaging
AChR density at NMJs was calculated using the quantitative fluorescence imaging technique described by Akaaboune et al. (1999)(2002). Briefly, mice were anesthetized, and the exposed sternomastoid muscle was saturated with rBTX (5 µg/ml) for 60 min, and superficial NMJs were viewed with a fluorescence microscope. The fluorescence intensity at synapses was compared with that of a nonbleaching fluorescent standard viewed concurrently. To study loss rate of AChRs, a nonblocking dose of rBTX (0.1 µg/ml) was applied to the sternomastoid for 2–5 min and 1 d later, after unbound toxin had cleared, individual synapses were imaged. Total fluorescence intensity at the first view was expressed as 100%. Mice were then reexamined 3 d later, and total fluorescence intensity of the previously identified junctions was measured (Akaaboune et al., 1999). The use of a nonblocking dose ensured that the turnover rate was not affected by paralysis.
Comparison of transgenic lines
Of the four DB1 transgenic lines examined on a DB-/- background, two (lines 12 and 14) expressed detectable DB1. More than 95% of muscle fibers expressed DB in line 12 but only 50–70% in line 14. However, both lines were similar in that expression of the transgene prevented dystrophy in transgene-positive fibers and led to a more normal synaptic structure than observed in DB-/-,tgDB2 transgenic lines. Of the six DB2 lines examined, expression was detected in three (lines 11A, 11B, and 28). Results from DB-/-,tgDB2 lines 11B and 28 were similar in all respects tested (muscle and synaptic structure and AChR turnover). Levels of transgene expression were lower in line 11A, and only 70% of muscle fibers in DB-/-,tgDB2 line 11A mice were transgene positive. No central nuclei occurred in these transgene-positive fibers, indicating rescue of the dystrophic phenotype, but rescue of the synaptic phenotype was significantly less in this line than in any of the other lines tested. In summary, all five transgenic lines tested exhibited rescue of the dystrophic phenotype, and both tgDB1 lines rescued synaptic defects more effectively than any of the tgDB2 lines. Based on these results, we studied DB-/-,tgDB1 line 12 and DB-/-,tgDB2 line 28 in greatest detail.
In vitro and in vivo transfection
Expression vector GFP-DB1 was constructed by cloning the DB1 cDNA described above into a pEGFP-C1 plasmid (CLONTECH Laboratories, Inc.), generating a fusion protein with GFP attached to the NH2 terminus of DB1. To create GFP-DB1-P3-, PCR-directed mutagenesis was used to change three terminal tyrosine residues (aa 698, 706, and 723) to phenylalanine. These residues correspond to aa 685, 693 and 710, which were shown to be major if not sole sites of tyrosine phosphorylation in Torpedo DB (Balasubramanian et al., 1998).
The DB1 constructs were transfected into HEK293 cells using Lipofectamine Plus transfection reagent (Invitrogen) and 4 µg of DNA per 10-cm dish. Cells were harvested 48 h after transfection. Immediately before collection, cells were subjected to pervanadate stimulation to inhibit tyrosine phosphatases as described by Balasubramanian et al. (1998). Soluble fractions were collected and immunopurified by incubating with GFP specific antibodies (A-11120; Molecular Probes) for 1 h at 4°C, and then with protein G sepharose beads (Amersham Biosciences) for 3 h more. Beads were precipitated, washed, resuspended in 1x sample buffer, boiled for 4 min, and subjected to immunoblotting (see above). DB was detected using an mAb (D62320, Transduction Labs), and phosphotyrosine was detected using PY20 antibody (610000, Transduction Labs).
For in vivo electroporation, DNA was dissolved into normal saline (0.9% NaCl) at a concentration of 2 µg/µl. Mice were anesthetized, and their tibialis anterior muscles were injected transcutaneously with 25 µl of a 4 U/µL bovine hyaluronidase/saline solution (Sigma-Aldrich) as recommended by McMahon et al. (2001). 2 h later, the mice were reanesthetized, the tibialis was exposed, injected with 25 µl of DNA (50 µg), and electroporated (Aihara and Miyazaki, 1998). Electroporation was performed with a pair of 0.2-mm diameter stainless steel needle electrodes (Genetronics) held 4 mm apart and inserted on either side of the injection site parallel to the muscle fibers. Ten 80 V pulses, each 20 ms in duration, were delivered at a frequency of 1 Hz, giving a field strength of 200 V/cm (BTX electroporator). Muscles were dissected 10–14 d after injection and fixed for 20 min in 1% PFA. Fiber bundles exhibiting GFP fluorescence were isolated under a fluorescence dissecting microscope, stained with rBTX and viewed by both light (Carl Zeiss MicroImaging, Inc.) and confocal (Olympus) microscopy.
|
Footnotes |
|---|
-bungarotoxin. 
|
Acknowledgments |
|---|
This work was supported by grants from the National Institutes of Health (to R.M. Grady and J.R. Sanes) and a Howard Hughes Medical Institute undergraduate fellowship to A.L. Cohen.
Submitted: 10 September 2002
Revised: 17 January 2003
Accepted: 17 January 2003
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Adams, M.E., N. Kramarcy, S.P. Krall, S.G. Rossi, R.L. Rotundo, R. Sealock, and S.C. Froehner. 2000. Absence of alpha-syntrophin leads to structurally aberrant neuromuscular synapses deficient in utrophin. J. Cell Biol. 150:1385–1398.
Aihara, H., and J. Miyazaki. 1998. Gene transfer into muscle by electroporation in vivo. Nat. Biotechnol. 16:867–870.[CrossRef][Medline]
Akaaboune, M., S.M. Culican, S.G. Turney, and J.W. Lichtman. 1999. Rapid and reversible effects of activity on acetylcholine receptor density at the neuromuscular junction in vivo. Science. 286:503–507.
Akaaboune, M., R.M. Grady, S.G. Turney, J.R. Sanes, and J.W. Lichtman. 2002. Neurotransmitter receptor dynamics studied in vivo by reversible photo-unbinding of fluorescent ligands. Neuron. 34:865–876.[CrossRef][Medline]
Ali, D.W., and M.W. Salter. 2001. NMDA receptor regulation by Src kinase signalling in excitatory synaptic transmission and plasticity. Curr. Opin. Neurobiol. 11:336–342.[CrossRef][Medline]
Balasubramanian, S., E.T. Fung, and R.L. Huganir. 1998. Characterization of the tyrosine phosphorylation and distribution of dystrobrevin isoforms. FEBS Lett. 432:133–140.[CrossRef][Medline]
Benson, M.A., S.E. Newey, E. Martin-Rendon, R. Hawkes, and D.J. Blake. 2001. Dysbindin, a novel coiled-coil-containing protein that interacts with the dystrobrevins in muscle and brain. J. Biol. Chem. 276:24232–24241.
Blake, D.J., and S. Kroger. 2000. The neurobiology of duchenne muscular dystrophy: learning lessons from muscle? Trends Neurosci. 23:92–99.[CrossRef][Medline]
Blake, D.J., R. Nawrotzki, M.F. Peters, S.C. Froehner, and K.E. Davies. 1996. Isoform diversity of dystrobrevin, the murine 87-kDa postsynaptic protein. J. Biol. Chem. 271:7802–7810.
Blake, D.J., R. Nawrotzki, N.Y. Loh, D.C. Gorecki, and K.E. Davies. 1998. beta-dystrobrevin, a member of the dystrophin-related protein family. Proc. Natl. Acad. Sci. USA. 95:241–246.
Blake, D.J., A. Weir, S.E. Newey, and K.E. Davies. 2002. Function and genetics of dystrophin and dystrophin-related proteins in muscle. Physiol. Rev. 82:291–329.
Buonanno, A., and G.D. Fischbach. 2001. Neuregulin and Erbß receptor signaling pathways in the nervous system. Curr. Opin. Neurobiol. 11:287–296.[CrossRef][Medline]
Byers, T.J., L.M. Kunkel, and S.C. Watkins. 1991. The subcellular distribution of dystrophin in mouse skeletal, cardiac, and smooth muscle. J. Cell Biol. 115:411–421.
Carr, C., G.D. Fischbach, and J.B. Cohen. 1989. A novel 87,000-Mr protein associated with acetylcholine receptors in Torpedo electric organ and vertebrate skeletal muscle. J. Cell Biol. 109:1753–1764.
Chang, W.J., S.T. Iannaccone, K.S. Lau, B.S. Masters, T.J. McCabe, K. McMillan, R.C. Padre, M.J. Spencer, J.G. Tidball, and J.T. Stull. 1996. Neuronal nitric oxide synthase and dystrophin-deficient muscular dystrophy. Proc. Natl. Acad. Sci. USA. 93:9142–9147.
1% in category 4. In contrast, over 70% of synapses in fibers expressing GFP-DB1-P3- fell into category 3 or 4 with none in category 1.
