We compared biochemical properties of control and CIM sodium channels to find candidates that might account for the hyperpolarized shift in inactivation gating seen in the acute phase of CIM. We identified several biochemical changes in sodium channel in CIM, but the most promising candidates appeared to be alterations in sodium channel-associated proteins in CIM. In particular, nNOS is a promising candidate that was more associated with sodium channels from CIM muscle. In mice lacking nNOS, the normal reduction in excitability following denervation was greatly reduced. These data are consistent with the possibility that increased association of nNOS with sodium channels is involved in triggering loss of muscle excitability in CIM.
We identified an increase in NaV
1.5 in CIM muscle such that it is approximately 28% of the entire channel population. This is similar to our earlier estimate of 21% obtained by measuring current densities [9
]. In our previous study, both the TTX-insensitive (NaV
1.5) and TTX-sensitive (NaV
1.4) channels demonstrated similar hyperpolarizing shifts in inactivation gating in CIM [9
], so increased expression of NaV
1.5 per se
cannot be responsible for the shift.
A second change identified in membranes from CIM muscle was reduced glycosylation of the NaV
1.4 sodium channel. Removal of the entire carbohydrate ‘tree’ at the asparagine-linkage eliminated the molecular weight difference between the control and CIM channel. However, selective removal of sialic acid moieties with neuraminidase did not eliminate the size difference. Given that the NaV
1.4 channel is known to have multiple carbohydrate trees [16
], the simplest explanation for these observations is that some but not most of the carbohydrate trees are removed in CIM, removing some but not most of the sialic acids. Previous work shows that removal of sialic acid from sodium channels shifts inactivation gating in a depolarizing direction [16
]. This is opposite of the hyperpolarizing shift we observed in CIM [9
]. Thus, removal of some carbohydrate trees may be a compensatory mechanism that moves the voltage dependence of inactivation towards more depolarized potentials.
One change we found in CIM muscle that could underlie the hyperpolarized shift in the voltage dependence of inactivation was an alteration in the composition of the dystrophin protein associated complex (DAPC) as summarized in Figure D. This figure is based not only on work in this paper, but also on work carried out by other investigators that identified the components of the DAPC (reviewed in [32
]). In our hands, the DAPC appeared to dissociate more easily in control samples such that all components (except sodium channels) were present at higher levels in the CIM samples. This was especially true for β-dystroglycan and nNOS, which are present at much higher levels in CIM CoIPs. These observations suggested to us that the sodium channel-DAPC complex is bound more tightly in CIM, perhaps indicating that the sodium channel and cytoskeleton are in a different and more strongly ‘locked’ conformation in the disease. Alternatively, the constituent members of the DAPC may be dynamically regulated. In either case, the presence of the important signaling protein nNOS in the DAPC of CIM muscle suggests that NO signaling through this protein could contribute to the altered inactivation gating in CIM.
The protein components that we identified in the DAPC are consistent with those identified by other investigators [26
]. In skeletal muscle, the consensus C-termini (S/TXV-COOH) of NaV
1.4 and 1.5 sodium channels bind the PDZ domain of syntrophin at a site overlapping and/or closely adjacent to the binding site for nNOS [26
]. Through syntrophin, both sodium channels and nNOS bind the C-terminus of dystrophin [26
]. In cardiac muscle, the dynamic nature of this complex was shown by comparative analysis of control vs.
syntrophin point mutation that causes Long QT syndrome. The syntrophin point mutation altered the complex constituents, such that the plasma membrane Ca2+
ATPase no longer bound syntrophin. This released inhibition of nNOS, allowed S-nitrosylation of the NaV
1.5 sodium channel, and altered gating [27
Dystrophin is part of the muscle cytoskeletal system. In mdx mice, which lack dystrophin, sodium channel inactivation gating is shifted 10
mV more positively than that of control mice [33
]. This observation suggests that loss of cytoskeletal components shifts inactivation gating in a depolarizing direction, a finding consistent with our hypothesis that sodium channel inactivation gating is hyperpolarized in CIM because it is more tightly associated with cytoskeletal components. However, acute disruption of cytoskeleton by pressure during formation of seals during patch clamp measurements has been found to trigger hyperpolarized shifts in the voltage dependent of NaV
1.4 and NaV
1.5 activation and fast inactivation [34
]. Thus, while it is clear that changes in cytoskeleton can have profound effects on the voltage dependence of sodium channel gating, we currently do not know which changes in cytoskeleton will translate into changes in sodium channel gating.
Our finding that nNOS is present at higher levels in the sodium channel-DAPC complex in CIM raises the possibility that increased signaling through NO-dependent pathways contributes to loss of muscle excitability in CIM (see however, [38
]). There are several cell signaling pathways that are regulated by NO, including protein phosphorylation through cGMP-protein kinase [39
] and direct nitrosylation of cysteine or other amino acid side chains, as discussed above for the cardiac NaV
]. We measured phosphorylation changes in CIM and found no overall difference (Figure ).
To determine whether increased association of nNOS with sodium channels could be involved in inducing inexcitability of muscle, we measured excitability following denervation in control and nNOS-null mice. In rats, denervation alone induces inexcitability in only a minority of fibers, so addition of corticosteroids is necessary to induce inexcitability [8
]. In control mice, denervation alone was sufficient to induce inexcitability so it was not necessary to co-administer corticosteroids. In nNOS-null mice, a greater percentage of muscle fibers remained excitable following denervation. There are multiple mechanisms that could account for maintenance of excitability following denervation in the absence of nNOS. Further study will be necessary to determine if the contribution of nNOS to inexcitability is mediated by its association with sodium channels as part of the DAPC.