Diseases caused by mutations in Na_v1.5 interacting proteins

Synopsis

Sodium current in the heart flows principally through the pore protein Na_V1.5, which is part of a complex of interacting proteins that serve both to target and localize the complex in the membrane, and to modulate function by such post-translational modifications as phosphorylation and nitrosylation. Multiple mutations in seven different Na_V1.5 interacting proteins have been associated with dysfunctional sodium current and inherited cardiac diseases, including long QT syndrome, Brugada syndrome, atrial fibrillation, and cardiomyopathy, as well as sudden infant death syndrome (SIDS). Mutations in as yet unidentified interacting proteins may account for cardiac disease for which a genetic basis has not yet been established. Characterizing the mechanisms by which these mutations cause disease may give insight into etiologies and treatments of more common acquired cardiac disease, such as ischemia and heart failure.

Clinical Importance

In this review we summarize current understanding of cardiac diseases caused by mutations in proteins that interact with the cardiac sodium channel Na_V1.5. Na_V1.5 is encoded by the gene SCN5A and forms the pore through which flows the majority of sodium current (I_Na). More than 30 sodium channel interacting proteins (SCIPs) have been identified^{1, 2} (Table 1). Excluding the four β-subunits, which are covered in Chapter 13, mutations in seven additional SCIPs have been associated with cardiac diseases (Table 2). These diseases include inherited arrhythmia syndromes such as long QT syndrome (LQTS), Brugada syndrome (BrS) and atrial fibrillation (AF), as well as inherited cardiomyopathy. We also include sudden infant death syndrome (SIDS) as a disease where deaths may be presumed to have a cardiac cause and where ~10% have been associated with channelopathies³. Mutations in SCIPs are a rare cause of inherited arrhythmia (<1% of LQTS and BrS), but they account for ~5% of SIDS overall³ or half of the SIDS cases associated with channelopathies. Mutations in SNTA1 alone represent ~25% of SIDS associated channelopathies (~2.5% of SIDS overall). Despite the overall rarity, the study of these mutations has provided insight into the regulation of I_Na and the pathophysiology of arrhythmia in common acquired cardiac diseases, such as ischemia and heart failure. In addition, an important percentage of BrS and SIDS remain without a linked genotype and SCIPs, both known (Table 1) or as yet unidentified, could account for the etiology of an important percentage of these syndromes.

Table 1.

Na_v1.5 interacting proteins (SCIPs)

Protein	Gene	Size (aa)	Evidence for NaV1.5 Association	Effect on INa
14-3-3η	Ywhah	246	Co-IP64	(↓)RecR, (−) SSI64
α-Actinin 2	Actn2	894	Co-IP65	(↑) INa Peak65
Ankyrin G	Ank3	4377	Co-IP66, 67	(↑)Surface Density, (↑)INa Peak66, 67
α1-Syntrophin	Snta1	505	GST-PD, Co-IP18, 19, 23	Regulate INa Late 18, 19, 23
β1, β1b	Scn1b	β1 218 β1b 268	Co-IP68	Kinetics conflicting (↓) INa Late69, 70 (↑)RecR71 (↑)INa Peak72, 73
β2	Scn2b	215	Co-IP74	Sialylation Status75, (↑)Late Current76
β3	Scn3b	215	Co-IP77	(↑)INa, (↑)RecR78(+) SSI, (↑)RefractPer.,79
β4	Scn4b	228	Co-IP80	(↑) AP Upstroke Velocity, (+) SSI81
β4 Spectrin	Sptbn4	2,564	Co-IP82	(↑)INa Peak, (+) SSI via Phos.@ S57182
Calmodulin	Calm1	152	GST-PD83 NMR84	(+)SSI85, 86
CaMKinase IIδ isoform 3	Camk2d	510	Co-IP82	(−) SSI, Phos. @ S571S, 516, T594 (↑) INa Late87
Caveolin-3	Cav3	151	Co-IP32	(↓) INa Late29, 30, 32
Connexin 40	Gja5	358	Co-loc88	?
Connexin 43	Gja1	382	Co-IP89	(↑) INa Peak90
Desmoglein 2	Dsg2	1118	Co-IP91	(↑) INa Peak91
Dystrophin	Dmd	3,685	Co-IP92	(↑)INa Peak92
Fibroblast growth factor FGF12B, 13	Fgf12 Fgf13	181	Co-IP93	(+) SSI, (↑) RecR93
λ2 syntrophin	Sntg2	539	GST-PD94	(+) Activ Ou, 2003 175/id}
GPD1L	Gpd1l	351	GST-PD37	(↑)INa by phosphorylation37 80 or ROS39
MOG1	Rangrf	218	GST-PD, co-IP44	(↑)Surface Density, (↑)INa Peak44
Nedd4-2	Nedd4l	975	GST-PD8	(↑) INa Peak8
nNOS	Nos1	1434	GST-PD, Co-IP19	(↑)INa Late19 (↑) INa Peak22, 23
Plakophilin-2	Pkp2	881	Co-IP54	(↑) INa Peak, (+) SSI56
PMCA4b	Atp2b4	1241	GST-PD, Co-IP19, 20	(↓) INa Late19
SAP97	Dlg1	904	Co-IP10	(↑) INa Peak10
SLMAP	Slmap	828	*59	(↑) INa Peak59
Telethonin	Tcap	167	Co-IP13	(−) Activ13
PTPH1	Ptpn3	913	GST-PD95	(−) SSI95
Utrophin	Utrn	3433	Co-IP92	(↓) INa Peak92
Zasp	Ldb3	727	GST-PD63	(↑) INa Peak 63

Adapted from Adsit GS, Vaidyanathan R, Galler CM, Kyle JW, Makielski JC. Channelopathies from mutations in the cardiac sodium channel protein complex. J Mol Cell Cardiol 2013.

^*SLMAP has not yet been shown to associate with NaV1.5.

Abbreviations: Co-IP = co-immunoprecipitation; GST-PD = pull-down with a GST fusion protein; (↓) = decrease in; (↑) = increase in; (+) = depolarizing shift; (−) = hyperpolarizing shift; Activ.= activation; RecR = recovery rate; RefractPer.=refractory period; SSI=steady state inactivation; INa Peak= peak sodium current; INa Late =late sodium current or persistent current

Table 2.

Sodium channel interacting protein (SCIP) with disease-causing mutations that have been shown to affect I_Na

SCIP	Gene name	Common Name	Chromosome	MW kDa	Splice Variants	Disease	Disease mutations
SNTA1	SNTA1	α1-Syntrophin	20q11.2	54	1	LQT1219, 25 SIDS23	A257G, A261V, 390V T262P, T372M, G460S
CAV3	CAV3	Caveolin-3	3p25	17	1	LQT929 SIDS30	A85T, F97C, S141R V14L, T78M, L79R
GPD11	GPD1L	Glycerol -3- Phoshphate deydrogenase 1-Like	3p22.3	38	1	BrS12, 38 SIDS37, 96	A280V E83K, I124V, R273C
MOG1	RANGRF	MOG1	17p13.1	20	4	BrS46 AF47	E83D E61X
PKP2	PKP2	Plakophillin-2	12p11	97	2	BrS56	S183N, M365V, R635Q T526A
ZASP	LDB3	Z-band alternatively spliced PDZ motif protein	10q23.2	77	7	LVNC61, 63	D117N (isoform2)
SLMAP	SLMAP	Sarcolemma Membrane- Associated Protein	3p21.2	95	8	BrS59	V269I, E710A
NaV1.5	SCN5A	Cardiac Sodium Channel	3p21	227	6	LQT3, BrS, SIDS, CM, and others	>400

LQT, long QT; SIDS, sudden infant death syndrome; BrS, Brugada syndrome; AF, atrial fibrillation; LVNC, left ventricular non-compaction; CM, cardiomyopathy. Underlined mutations reported to be associated with more than one syndrome. Information on splice variants can be found at the Uniprot web site, specifically for SNTA1 http://www.uniprot.org/uniprot/Q13424, CAV3 http://www.uniprot.org/uniprot/P56539, GPD1L http://www.uniprot.org/uniprot/Q8N335 MOG1 http://www.uniprot.org/uniprot/Q9HD47 PKP2 http://www.uniprot.org/uniprot/Q13835 SLMAP http://www.uniprot.org/uniprot/Q14BN4 ZASP http://www.uniprot.org/uniprot/O75112

Definition of SCIPs and the Sodium Channel Complex (SSC)

SCIPs are defined as proteins that co-localize with Na_V1.5 as part of a Sodium Channel Complex (SCC) and where a physical association has been shown by co-immunoprecipitation, GST-fusion protein pull-down, or other binding assay, such as yeast-two hybrid assay. SCIPs have been variously called and classified as accessory proteins, auxiliary proteins, associated proteins, anchoring proteins, β-subunits, scaffolding proteins, adaptor proteins, and regulatory proteins^{2, 4–6}. Each term has intuitive implications for structure or function, but the terms are not usually rigorously defined and classifications may overlap. Co-localization may occur through direct interaction with Na_V1.5 or indirectly through other SCIPs. For example both α1-syntrophin and ankyrin-G serve to connect SCIPs to Na_V1.5 and are called adapter or scaffolding proteins. In many cases the sites of physical interaction are known, and most SCIPs have also been shown to affect I_Na function. On the other hand some proteins affect I_Na or “interact functionally” but have not been shown to be physically associated with the SCC and do not meet this formal definition of a SCIP. For example, PKC affects Na_V1.5 density⁷ but has not been shown to be associated with the SCC. A broader definition of SCIPs might include proteins such as chaperones that associate during trafficking to the membrane or tags for recycling/degradation, such as ubiquitin⁸, but for this review we use the narrower definition. Another issue with the definition is that through the cytoskeleton and scaffolding proteins the SCC may be attached more distantly to many other complexes in the cell. As the field evolves, more precise definitions of “interacting” based on proximity, time, location, and function will emerge.

Although more than 30 SCIPs have been identified (Table 1), they are not present in every complex at all times. Indeed two distinct “pools” of SSCs have been identified in cardiac myocytes, one at the lateral membrane associated with SNTA1 and a second at the intercalated disc with SAP97/plakophilin-2 (PKP2)^{9, 10}. Also, the dynamic nature of the SCC is not known; are the components “permanent” for the ~35 hour half-life¹¹ of the channel at the cell surface, or are they transient over some time period?

Mutations, disease, and causality

When a mutation is identified in a patient with a particular disease, perhaps the strongest evidence of causality is the strength of genetic linkage analysis. As an example, GPD1L was discovered through a strong genetic linkage to the BrS phenotype¹², however, most mutations in SCIPs do not have this level of genetic linkage evidence for causality. Absence of a mutation in “control” populations and high conservation of the residue across species provides additional genetic evidence for causality. Plausible pathogenicity of a candidate disease mutation is often built on the functional importance of the mutation by studies at the molecular and cellular level. The mechanisms by which I_Na dysfunction causes clinical syndromes is covered in detail in other articles in this issue, but can be briefly conceptualized as a gain of function where I_Na, particularly late I_Na corresponding to phase 2 and 3 of the action potential is increased, or as “loss of function” where I_Na, particularly peak or early I_Na corresponding to phases 0 and 1 of the action potential, is decreased.

Scope of this review

We feature five SCIPs that meet the definition above and also include SLMAP and ZASP as two more putative SCIPs (Table 2 and Fig. 1) that have evidence of causation of a cardiac disease through changes in I_Na. SCIPs that are not considered include mutations in the SCIP telethonin associated with irritable bowel syndrome¹³, but not associated with cardiac disease, which is the scope of this review. A mutation within Na_V1.5 disrupted association with the SCIP ankyrin-G in BrS patients¹⁴, but this is not included because no disease causing mutations in ankyrin-G itself have been yet identified. Mutations (D130G and F142L) in the SCIP calmodulin (CALM1) (Table 1) were discovered in infants with long QT arrhythmia¹⁵; both showed decreased sensitivity to calcium, but effects on I_Na are not yet established.

Figure 1.

Diagrams of SCIPs with locations of mutations causing Na_V1.5-related cardiac disease and select functional domains. The N- and C-termini are indicated as N and C with the amino acid number beneath. The amino acids corresponding to each domain are shown below (amino acid range for domains are assigned based on http://www.uniprot.org/). Disease-causing mutations are indicated at the top of each panel with the box color coded for diseases long QT syndrome (LQTS), Brugada syndrome (BrS), sudden infant death syndrome (SIDS), and atrial fibrillation (AF) as noted at the figure bottom, with references for these mutations given in Table 2. A. SNTA1: Two parts of the split plextrin homology domain are shown as PH1_N and PH1_C with the PDZ domain located between them. PH2 is the second plextrin homology domain and SU is the syntrophin unique domain. http://www.uniprot.org/uniprot/Q13424 B. CAV3: SD is the CAV3 scaffold domain and IMD is the intramembrane domain. http://www.uniprot.org/uniprot/P56539 C. GPD1L: NAD binding sites are indicated as diamonds, substrate binding regions are red boxes and the active site is indicated as *. http://www.uniprot.org/uniprot/Q8N335 D. MOG1: RanBD/NR is the Ran binding/GTP release domain⁴² http://www.uniprot.org/uniprot/Q9HD47 E. PKP2: The N-terminal domain is marked NT and the eight armadillo repeats are indicated as red boxes. http://www.uniprot.org/uniprot/Q13835 F. SLMAP: The long cytoplasmic N-terminus with a forkhead-associated domain is marked FHA, coiled-coiled leucine zipper domains labeled CC are shown as coils, and a transmembrane domain is labeled TMD. http://www.uniprot.org/uniprot/Q14BN4 G. ZASP: The PDZ domain at the N-terminal end is marked PDZ and the 3 Lim zinc binding domains labeled Lim1, Lim2 and Lim3. http://www.uniprot.org/uniprot/O75112

SNTA1 and LQTS/SIDS

SNTA1 or α1-syntrophin is a 54 kDa cytoplasmic membrane-associated adaptor protein that is a member of the multigene syntrophin family and is coded for by the SNTA1 gene¹⁶. Syntrophins contain three protein interacting domains: 1) a PDZ domain (postsynaptic density protein-95/disc large/zona occludens-1), 2) a plextrin homology (PH) domain, one of which is split by the PDZ domain in SNTA1, and 3) a syntrophin unique domain (SU) (Fig. 1A). SNTA1 also contains a second intact PH domain following the split PH domain. SNTA1 serves as an adaptor to link proteins with signaling molecules, such as nNOS or calmodulin, or even other adaptor proteins, such as dystrophin.¹⁶ An adaptor protein itself does not possess intrinsic activity, but localization of bound signaling molecules to the microenvironment provides specificity¹⁷. The PDZ domain on SNTA1 interacts with a PDZ binding domain on the last 3 amino acids of the C-terminus of Na_V1.5 as shown by pull-down of SNTA1 by a GST-fusion protein containing 66 amino acids of the Na_V1.5 C-terminus (Fig. 2).¹⁸ When the last 3 amino acids (SIV) of the Na_V1.5 C-terminus were deleted syntrophins and dystrophins were no longer pulled down.

Figure 2.

Cartoon of the Na_V1.5/SNTA1/nNOS/PMCA4b complex. The PDZ binding domain at the C-terminus of Na_V1.5 binds to the PDZ domain of SNTA1. One of the PDZ domains from the nNOS dimer associates with the PDZ domain of SNTA1. PMCA4b interacts with SNTA1 at two locations, the 4–5 cytoplasmic loop binds to the PH2/SU region of SNTA1 and the PDZ binding domain in the C-terminus of PMCA4b can also associate with PDZ in SNTA1. The A390V-SNTA1 BrS mutation that disrupts binding of PMCA4b to SNTA1 is shown. The release of PMCA4b from the complex results in upregulation of nNOS and increased nitrosylation of Na_V1.5 and increased late I_Na.

A screen for mutations in SNTA1 in a cohort of 50 unrelated LQTS patients negative for 11 established LQTS genes yielded a missense mutation (A390V-SNTA) in an 18 year old male with syncope¹⁹. Previous work on neuronal nitric oxide synthase (nNOS) in brain established a physical association between SNTA1, nNOS, and the plasma membrane calcium ATPase subtype 4b calcium pump (PMCA4b)²⁰. Other studies in heart established that PMCA4b inhibited nNOS²¹. SNTA1 also associated with Na_V1.5 in heart¹⁸, and together these studies raised the possibility that SNTA1 was part of an SCC in heart with SNTA1 connecting nNOS and PMCA4b to Na_V1.5 (Fig. 2). Confirmation of Na_V1.5/SNTA1/nNOS/PMCA4b complex was demonstrated in transfected HEK cells where GST- Na_V1.5 C-terminal fusion constructs pulled-down WT-SNTA1, nNOS and PMCA4b. When A390V-SNTA1 was co-expressed instead of WT-SNTA1, the association with PMCA4b was lost¹⁹. The A390V mutation is within the second PH2 domain in SNTA1 (Fig. 1A & Fig. 2) and is thought to be the stronger of two association sites between PMCA4b and SNTA1, the weaker association being between a PDZ domain on the C-terminus tail of PMCA4b and the PDZ domain within the split PH1 domain on SNTA1¹⁶. The loss of association of PMCA4b with A390V-SNTA1 resulted in increased S-nitrosylation of Na_V1.5 and increased late I_Na and both increases were prevented by the addition of L-NMMA, an arginine analogue and specific inhibitor of nNOS¹⁹. These results support the idea that late I_Na was increased secondary to S-nitrosylation of Na_V1.5 caused by A390V-SNTA1 disruption of association of the nNOS inhibitor PMCA4b (Fig. 2).

A screen of 39 LQTS patients negative for mutations in known LQTS susceptibility genes identified a single missense mutation in SNTA1 (A257G-SNTA1) in 3 unrelated patients²². A257G is in a highly conserved region of SNTA1 and it was not found in controls. Expression studies with A257G-SNTA1 showed increased peak I_Na but no change in late I_Na. Kinetic effects with co-expression of A257G-SNTA1 included a negative shift of 7–10 mV in steady-state activation. Computer modelling supported the role of A257G-SNTA1 in triggering arrhythmia²². It should be noted that the expression studies did not include nNOS or PMCA4b, which were shown¹⁹ to be required for increased late I_Na with A390V-SNTA1, and this could explain why no disproportionate increase in late I_Na was observed. Together these studies^{19, 22} established SNTA1 as a cause of LQTS and was designated LQT12.

SNTA1 was screened for mutations in 282 SIDS cases and 6 rare missense mutations (G54R, P56S, T262P, S287R, T372M and G460S) were found in 8 cases that were absent in 800 controls²³. When these were co-expressed with Na_V1.5, PMCA4b, and nNOS the results for S287R, T372M and G460S were similar to A390V-SNTA showing increased late I_Na that was blocked by nNOS inhibitors, thereby establishing mutations in SNTA1 as a plausible cause of SIDS. Missense mutations in both SNTA1 and Na_V1.5 (A261V-SNTA and R800L-Na_V1.5) were found in a three generation family with LQTS²⁴. Co-expression of both mutations showed increased late I_Na that was the sum of each mutation expressed alone, suggesting additive effects to produce the phenotype. In another example of interaction, the P74L-SNTA1 polymorphism mitigated the deleterious effect of the LQT12 A257G-SNTA1 mutation²⁵.

CAV3 and LQTS/SIDS

Caveolin-3 (CAV3) is a small ~17 kDa integral membrane protein highly expressed in skeletal muscle and heart^26–28. CAV3 is a member of a family of caveolins which are the major proteins of caveolae, cholesterol-enriched invaginations of the sarcolemma (Fig 3). CAV3 mutations (Fig. 1B) were identified in LQTS (classified as LQT9)²⁹ and SIDS³⁰ and showed increased late I_Na when co-expressed with Na_V1.5. CAV3 has also been shown to associate with and inhibit nNOS³¹ (Fig. 3), raising the possibility that LQT9, analogous to LQT12, is caused by a loss of function by CAV3 to inhibit nNOS leading to increased S-nitrosylation of Na_V1.5 and increased late I_Na.. Expression of the LQT9 mutation F97C-CAV3 in both HEK cells and rat myocytes showed increased late I_Na and rat myocytes showed increased action potential duration that were abrogated by the NOS inhibitor L-NMMA³². Evidence for increased S-nitrosylation of Na_V1.5 was provided by experiments in HEK cells where S-nitrosylation of Na_V1.5 determined by biotin switch assay was increased by F97C-CAV3 compared to WT-CAV3 ³². Interestingly F97C-CAV3 remained associated with the SCC but lost the nNOS inhibition activity as determined by enzymatic assay³². This suggested that CAV3 and SNTA1 mutations share a common mechanism to increase late I_Na by releasing inhibition of nNOS leading to an increase in S-nitrosylation of Na_V1.5 (or other SCC) and an increased late I_Na. It is important to note that CAV3 interacts with multiple ion channels and transporters, and ion currents other than I_Na may also be affected by these mutations. Indeed these mutations decrease the inward rectifier current by decreasing KIR2.1 channel protein expression at the surface³³, and this likely contributes to the pathogenesis of LQT9.

Figure 3.

The CAV3/nNOS/Na_V1.5 complex in caveolae. The triangles on CAV3 represents amino acids 62–72 and are homologous to the CAV1 region that has been shown to bind to and inhibit eNOS activity⁹⁷. The binding site on nNOS shown is located at W681 and W683 and is based on the homologous site on eNOS⁹⁷.

GPD1L and BrS

GPD1L (Glycerol Phosphate Dehydrogenase 1 Like) is a 38 kDa protein (Fig. 1C) with very high homology to human cytoplasmic GPD1(cGPD1)³⁴, an enzyme that catalyzes the reversible reaction converting dihydroxyacetone phosphate to glycerol 3-phospate with the subsequent oxidation of NADH to NAD⁺. cGDP1 is an NAD-dependent cytosolic enzyme that is an important link between the glycolytic pathway and triglyceride synthesis. GPD1L was unknown as a human protein until the mutation A280V-GPD1L was implicated in BrS by linkage analysis and gene walking in a large family³⁵. This BrS mutation¹² and three novel SIDS-associated GPDL1 mutations (E83K, I124V and R273C)³⁶ decreased I_Na by >50% when the mutant GPD1L cDNAs were transfected into HEK cells expressing Na_V1.5^{12, 36}. The SIDS mutations transfected in mouse myocytes³⁶ also increased late I_Na. In heterologous expression systems GPD1L co-localized with Na_V1.5 and was shown to be associated with Na_V1.5 by co-immunoprecipitation, and Na_V1.5 was pulled down by GST-GPD1L (both WT and GPD1L mutants)³⁷, establishing GPD1L as a SCIP. GPD1L was localized at the cell surface¹² and the A280V mutant decreased cell surface expression ~30%¹², but the mechanism for the decrease was unknown. How does a defect in an enzyme that is presumably involved in metabolic pathways specifically target Na_V1.5? Two different but not mutually exclusive mechanisms have been proposed. One hypothesis³⁷ has the A280V-GPD1L mutation leading to an increased concentration of glycerol-3-phospate which would increase other substrates in the pathway to diacyl-glycerol (DAG). DAG concentration would increase primarily in the area near the NaV1.5/GPD1L SCC and upregulate PKC, which is known to decrease I_Na by direct channel phosphorylation at S1503⁷. In support of this hypothesis E83K-SNTA1 and A280V-SNTA1 decreased both I_Na and cell surface expression of Na_V1.5 and these decreases were abrogated both by pharmacological blockers of the PKC-related pathway and also by co-expression with the PKC phosphorylation deficient Na_V1.5 mutant S1503A³⁷. This mechanism could account for specificity by co-localization of GPD1L and intermediates with Na_V1.5 substrate and direct phosphorylation of Na_V1.5. Another proposed mechanism has the mutant increase NADH, and through PKC effects on mitochondria, to increase ROS, which then decreases I_Na by unspecified mechanisms^{38, 39}. It is not clear however, how this mechanism would provide specificity for Na_V1.5 as a general increase in ROS would have wide effects. Either or both mechanisms could be operant depending upon the direction of the reversible reaction that GPD1L regulates. GPD1L causing BrS is very rare⁴⁰ and is only slightly more common in SIDS³, but of wider interest is linkage of a SNP upstream of GPD1L that has been associated with sudden cardiac death in patients with coronary artery disease⁴¹. Whether or not this association occurs through Na_V1.5 is unknown, but the localization to Na_V1.5, the functional interaction, and the genetic associations suggest an important role for GPD1L in regulating cardiac excitability.

MOG1 and BrS/AF

MOG1 is a 20 kDa protein (Fig. 1D) coded for by the RANGRF gene (Ran GTP release factor). The yeast homolog, scMOG1, is primarily located in the nucleus and acts as a Ran GTP release factor regulating nuclear import and export of proteins. Human MOG1 binds to both yeast and human Ran, has a GTP release activity that has been mapped to the first 45 amino acids⁴² (Fig 1D), and can partially rescue the growth defect in yeast cells lacking scMOG1, showing that it retains some of the activity of the yeast homologue⁴³. MOG1 was identified as a SCIP in human heart by a yeast two-hybrid screen using a human heart cDNA library as prey⁴⁴ and the second intracellular loop of Nav1.5 between repeats II-III as bait⁴⁴. Although MOG1 is mostly found in the nucleus it also co-localizes with Na_V1.5 at the intercalated disc⁴⁴. Na_V1.5 and MOG1 co-immunoprecipitation and GST pull-down assays confirmed the association between the Na_V1.5 cytoplasmic loop II and in vitro translated MOG1. The GST pull-down assay using in vitro translated proteins is notable because it demonstrates a direct interaction without the need for intermediates to link MOG1 to Na_V1.5, although the precise interaction sites are not known. Co-expression of Na_V1.5 and MOG1 in HEK cells increased I_Na 2-fold without affecting steady-state activation, steady-state inactivation or recovery from inactivation. Similarly, overexpression of MOG1 in mouse neonatal cardiomyocytes increased I_Na with no change in single channel conductance, suggesting that the increased I_Na was caused by an increase in channel number at the cell surface. Knock-down of MOG1 by siRNA in neonatal mouse cardiomyocytes decreased I_Na^{45, 45}.

A screen of 246 BrS patients⁴⁶ yielded a missense mutation (E83D-MOG1) in a female patient with BrS that was absent in controls. Co-expression of E83D-MOG1 with Na_V1.5 in HEK cells caused ~50% reduced I_Na without a change in kinetics, and it exerted a dominant-negative effect on WT-MOG1 in co-expression experiments. The E61X nonsense mutation causing a premature stop (E61X) was reported in 4 patients in a screen of 197 patients with lone AF and in one of 23 patients with BrS⁴⁷, but this mutation was also detected in two control subjects. Expression of E61X-MOG1 with Na_V1.5 in CHO-K1 cells showed ~50% decreased I_Na but when E61X-MOG1 was co-expressed with WT-MOG1 there was no dominant-negative effect and the levels of I_Na were comparable to control. The pathogenicity of E61X was further questioned when it was detected in an asymptomatic patient with a BrS ECG and in five other asymptomatic family members⁴⁸.

PKP2 and BrS and Cardiomyopathy

Plakophilin (PKP2) is a 98 kDa protein and a member of a family of desmosomal proteins localized primarily at the intercalated disc in cardiomyocytes. PKP2 has a 335 amino acid N-terminus that is the site of interaction with binding partners (Fig. 1E)^{49, 50}. There are 8–9 armadillo repeats, a signature feature of the family, and a characteristic bend occurs between armadillo repeats 5 and 6 followed by a short C-terminus^{49, 50}. PKP2 is a scaffolding protein that with other desmosomal proteins forms a bridge between cadherens and intermediate filaments. Mutations in PKP2 have been linked to familial arrhythmogenic cardiomyopathy (AC), also called arrhythmogenic right ventricular cardiomyopathy⁵¹. As with heart failure⁵² and other cardiac diseases, Na_V1.5 is often reduced in AC⁵³. In autopsy samples from 5 AC patients and 5 normal controls, immunohistochemistry at the intercalated discs showed that levels of Na_V1.5, Cx43 and plakoglobin were reduced in most patients (65–74%) while PKP2 was not affected unless there was a PKP2 mutation⁵³. These results show that levels of expression of Na_V1.5 at the intercalated disc can be reduced independently of PKP2. Evidence that PKP2 is a SCIP was first provided in siRNA knock down experiments of PKP2 in rat cardiomyocytes⁵⁴ where PKP2 protein was reduced and I_Na reduced ~50% with no apparent loss of total Na_V1.5 protein. This suggested a redistribution of NaV1.5 from the cell surface to intracellular locations. Evidence for association between PKP2 and Na_V1.5 was provided by pull-down of Na_V1.5 by a GST-fusion construct containing the N-terminus of PKP2⁵⁴. The kinetics of Na_V1.5 were altered after siRNA knock down of PKP2 with a negative shift in steady-state inactivation and slower recovery from inactivation⁵⁴. Optical mapping showed a slowing of conduction⁵⁵ suggesting PKP2 as a potential causative agent in BrS. Subsequently a screen of 200 BrS patients without evidence of AC and negative for mutations in other genes linked to BrS yielded 5 PKP2 missense mutations (Q62K, S183N, M365V, T526A and R635Q)⁵⁶ (Fig. 1E). I_Na and the number of Na_V1.5 channels found at the intercalated disc were decreased when PKP2 mutants were expressed in a rat atrial cell line (HL1 cells), human embryonic stem cell derived cardiomyocytes, or in induced pluripotent stem cell derived cardiomyocytes from an AC patient. A dominant-negative effect was absent when WT and mutant PKP2 were co-expressed. In contrast to PKP2 siRNA knock down studies in cardiomyocytes, sodium channel kinetics were not affected⁵⁶. Single channel properties were also unaffected and taken together with immunohistochemical data are consistent with a reduction of Na_V1.5 protein at the intercalated discs⁵⁶. These data support a role for PKP2 in targeting and transport of Na_V1.5 to the intercalated disc⁵⁶ as recently reviewed⁵⁷.

SLMAP and BrS

Sarcolemma membrane-associated protein (SLMAP or SLAP) is a 95 kDa protein that is localized to the sarcolemma and t-tubules near the sarcoplasmic reticulum. SLAMP3, the longest of three alternatively splice forms (37, 46, 74 kDa respectively) is predominant in heart (Fig. 1F)⁵⁸. SLMAP and ZASP (see below) are recently identified disease causing SCIPs included in this review despite falling outside the strict definition of a disease causing SCIP put forth earlier. A screen of 190 unrelated BrS patients identified two missense mutations (V269I and E710A) in SLMAP⁵⁹. Both mutations decreased I_Na and also cell surface expression of both SLMAP and Na_V1.5 in HEK cells, with no change in I_Na gating kinetics. SLMAP and Na_V1.5, failed to co-immunoprecipitate suggesting that SLMAP may not be a tightly associated part of the SCC ⁵⁹. SLMAP may play a role in targeting of Na_V1.5 and is a candidate for further study as a BrS linked gene, but it has not yet been firmly established as a SCIP.

ZASP and Dilated Cardiomyopathy/Left Ventricular Noncompaction Syndrome

ZASP (Z-band alternatively spliced PDZ motif protein) is encoded by the gene LDB3Z4 (Lim Domain-Binding Protein 3) and is a member of a group of 10 genes that code for proteins that contain both PDZ and multiple LIM domains (Fig. 1G)⁶⁰. Multiple alternative spliced forms of ZASP exist and the protein coded by longest transcript is 77 kDa (Table 2). All alternatively spliced forms of ZASP have an N-terminal PDZ domain but some forms lack the C-terminal Lim domains (Table 2). LIM domains contain a cysteine-rich consensus Zinc-finger sequence and are protein interaction domains⁶⁰. Mutations in ZASP have been identified in patients with dilated cardiomyopathy and left ventricular non-compaction⁶¹ and cardiac specific loss of the murine ZASP homologue results in a severe dilated cardiomyopathy and premature death⁶². In addition to a role in stabilizing the sarcomere structure, ZASP can act as an adaptor protein as, for example, to bridge alpha-actinin-2 through its N-terminal PDZ domain with PKC or PKA through its C-terminal LIM domains⁵ or to bridge the L-type calcium channel through its C-terminal PDZ binding motif with PKA through the LIM binding domain⁵. Mutations in ZASP have been associated with myofibrillar myopathy, and dilated cardiomyopathy⁶¹. D117N-ZASP, a mutation reported previously⁶¹, was found in a patient with Left Ventricular Noncompaction Syndrome⁶³ and when co-expressed with Na_V1.5 in HEK293 cells and in rat neonatal myocytes it caused a ~30% reduction in I_Na. Steady-state activation was shifted +9 mV in HEK cells and +15 mV in myoctes and inactivation was shifted by + 4 mV in HEK cells and +10 mV in myocytes, and recovery from inactivation was slowed⁶³. Computer modeling supported the hypothesis that the reduction in I_Na mediated by ZASP-D117N would generate arrhythmias. Association with Nav1.5 was demonstrated by co-immunoprecipitation using purified ZASP (or D117N- ZASP) produced in E. coli. mixed with homogenates from Na_V1.5 transfected HEK cells or myocytes⁶³. The reported association of ZASP with Na_V1.5 ⁶³ and subsequently with L-type calcium channels ⁵ acts as a reminder that SCIPs can be promiscuous (see also Cav3) and the pathogenesis may be complex.

Summary

SCIPs serve multiple functions including targeting of the SCC to the sarcolemma and regulating function through such mechanisms as post-translational modification (phosphorylation and nitrosylation). Specificity to Na_V1.5 and I_Na can be achieved by directly interacting with the Na_V1.5 channel protein but also by localizing signaling pathway components to the local milieu ¹⁷. When this regulation is disturbed by mutations in SCIPs the resulting dysregulation of I_Na can be a mechanism for diseases such as inherited arrhythmia syndromes and SIDS. Although at present mutations in SCIPs are a relatively rare cause of cardiac disease, they are prime candidates to account for BrS syndrome and other inherited arrhythmia syndromes, as well as SIDS and cardiomyopathies, where genetic causes are suspected but not yet demonstrated. At a more basic level, understanding the mechanisms of how mutations in SCIPs cause disease may give insight into the etiology and treatment options of the more common acquired cardiac diseases, including the contribution of subtle genetic variations as susceptibility variants to cardiac disease.

Key points.

• Sodium current, which underlies cardiac excitability, flows through a pore protein Na_V1.5 which is part of a larger complex of interacting proteins

• Mutations in 7 Na_V1.5 interacting proteins have been associated with dysfunctional sodium current and inherited cardiac diseases.

• The mechanisms by which mutations in interacting proteins cause specific dysfunction involve targeting/trafficking and phosphorylation/nitrosylation of the Na_V1.5 complex.

• Mutations in as yet unidentified interacting proteins may account for cardiac disease for which a genetic basis is not yet known