The p.P888L SAP97 polymorphism increases the transient outward current (Ito,f) and abbreviates the action potential duration and the QT interval

Synapse-Associated Protein 97 (SAP97) is an anchoring protein that in cardiomyocytes targets to the membrane and regulates Na+ and K+ channels. Here we compared the electrophysiological effects of native (WT) and p.P888L SAP97, a common polymorphism. Currents were recorded in cardiomyocytes from mice trans-expressing human WT or p.P888L SAP97 and in Chinese hamster ovary (CHO)-transfected cells. The duration of the action potentials and the QT interval were significantly shorter in p.P888L-SAP97 than in WT-SAP97 mice. Compared to WT, p.P888L SAP97 significantly increased the charge of the Ca-independent transient outward (Ito,f) current in cardiomyocytes and the charge crossing Kv4.3 channels in CHO cells by slowing Kv4.3 inactivation kinetics. Silencing or inhibiting Ca/calmodulin kinase II (CaMKII) abolished the p.P888L-induced Kv4.3 charge increase, which was also precluded in channels (p.S550A Kv4.3) in which the CaMKII-phosphorylation is prevented. Computational protein-protein docking predicted that p.P888L SAP97 is more likely to form a complex with CaMKII than WT. The Na+ current and the current generated by Kv1.5 channels increased similarly in WT-SAP97 and p.P888L-SAP97 cardiomyocytes, while the inward rectifier current increased in WT-SAP97 but not in p.P888L-SAP97 cardiomyocytes. The p.P888L SAP97 polymorphism increases the Ito,f, a CaMKII-dependent effect that may increase the risk of arrhythmias.

Three cardiac-specific transgenic-like mouse models on the basis of adeno-associated virus (AAV) gene transfer [5] were created trans-expressing or not, wild type (WT) and p.P888L SAP97, respectively. To this end AAV vectors driven from the cardiomyocyte-specific cardiac troponin T proximal promoter and encoding or not (empty vector [Sham]), WT human I3-I1A SAP97 (SAP97 WT) or p.P888L SAP97 were constructed. The use of the cardiac troponin T proximal promoter guarantees the expression of the delivered genes mostly in the heart [5].
AAV vector production and purification. AAV vectors were all produced by the triple transfection method, using HEK 293T cells as described previously [5]. AAV plasmids were cloned and propagated in the French; University of Virginia, USA) and packaged into AAV-9 capsids with the use of helper plasmids pAdDF6 (providing the three adenoviral helper genes) and pAAV2/9 (providing rep and cap viral genes), obtained from PennVector. The AAV shuttle and helper plasmids were transfected into HEK 293T cells by calcium-phosphate co-precipitation. A total of 840 µg of plasmid DNA (mixed in an equimolar ratio) was used per Hyperflask (Corning) seeded with 1.2x10 8 cells the day before. Seventy-two hours after transfection, the cells were collected by centrifugation and the cell pellet was resuspended in TMS (50 mM Tris HCl, 150 mM NaCl, 2 mM MgCl 2 ) on ice before digestion with DNase I and RNaseA (0.1 mg/mL each; Roche) at 37 °C for 60 minutes. Clarified supernatant containing the viral particles was obtained by iodixanol gradient centrifugation. Gradient fractions containing virus were concentrated using Amicon UltraCel columns (Millipore) and stored at -70ºC.
Determination of AAV vector titer. Titers for the AAV vectors (vg per mL) were determined by quantitative real-time PCR as described [5]. Known copy numbers (10 5 -10 8 ) of the respective plasmid (pAAV empty vector, pAAV-SAP97 WT and pAAV-SAP97 p.P888L) carrying the appropriate cDNA were used to construct standard curves.
Four-to 6-week-old wild-type C57BL6/J male mice were injected with 3.5 × 10 10 viral genomes encoding the empty vector, WT or p.P888L SAP97 through the intravenous femoral route. This method ensures the expression of the protein in the cardiac tissue for more than 6 months, a period of time in which mice did not show any apparent illnesses or increased mortality. Importantly, the use of this mouse model avoids the maintenance of large colonies of genetically modified animals and, thus, decreases the number of animals used, which fits with public concerns and the minimal-use concept expressed in the 3 Rs (3Rs) principle for the rational use of animals in research: Replacement of animals by alternatives wherever possible; Reduction in the number of animals used; and Refinement of experimental conditions and procedures to minimize the harm to animals. Ten to twenty weeks after infection with AAV particles, animals were used for the subsequent analyses (ECG or cellular electrophysiological recordings, and protein expression).
Surface ECG. Mice were anaesthetized using isoflurane inhalation (0.8-1.0% volume in oxygen), and efficacy of the anaesthesia was monitored by watching breathing speed. Four-lead surface ECGs were recorded, for a period of 5 minutes, from subcutaneous 23-gauge needle electrodes attached to each limb using the MP36R amplificator unit (BIOPAC Systems) [5]. ECG parameters such as the duration of the P wave, and of the PR, QRS, QT, and RR intervals were measured by using the Acknowledge 4.1 analysis (BIOPAC Systems) software.

Cell culture and transfection. Chinese Hamster Ovary (CHO) cells (ATCC, LGC Standards, UK)
were cultured in 60 mm dishes at 37ºC in an atmosphere of 5% of CO 2 , with humidity of ≈95%, as previously described [2,3,[6][7][8][9]. CHO cells were transfected with the cDNA encoding Kv4.3 (3 g) and cotransfected or not with the cDNA encoding WT, or p.P888L SAP97 (1.6 g) plus the cDNA encoding the CD8 antigen (0.5 µg) using FUGENE XtremeGENE (Roche Diagnostics, Switzerland) according to the manufacturer´s instructions [2,3,[6][7][8][9]. Forty eight h after transfection, cells were incubated with polystyrene microbeads precoated with anti-CD8 antibody (Dynabeads M450; Life Technologies). Most of the cells that were beaded also had channel expression. Mouse fibroblasts or Ltk − cells stably expressing hKv1.5 channels were grown in Dulbecco's modified Eagle medium (Invitrogen, USA) supplemented with 10% horse serum, 0.05 mg/mL gentamicin, and 0.25 mg/mL G418 (a neomycin analog; Gibco, USA) in a 5% CO 2 atmosphere [7,10]. Micropipette resistance was kept below 1.5 MΩ for I Na , below 3.5 MΩ for the rest of the currents or above 7 MΩ for action potentials when filled with the internal solution and immersed in the external solution. In all the experiments, series resistance was compensated manually by using the series resistance compensation unit of the Axopatch amplifier, and ≥80% compensation was achieved.
The remaining access resistance after compensation and cell capacitance were 1.3±0.8 MΩ and 10.7±0.5 pF (n=80) and 1.4±0.7 MΩ and 10.4±0.8 pF (n=40) in Ltkand CHO cells, and 2.5±1.3 MΩ and 125±2.9 pF (n=222) in ventricular myocytes from mice, respectively. Therefore, under our experimental conditions no significant voltage errors (<5 mV) due to series resistance were expected with the micropipettes used. To minimize the contribution of time-dependent shifts of channel availability during I Na recordings, all data were collected 5-10 min after establishing the whole-cell configuration.
Under these conditions current amplitudes and voltage dependence of activation and inactivation were stable during the time of recordings [2,3,8,9,12,15]. The current recordings were sampled at 4 kHz (except for I Na that was sampled at 50 kHz), filtered at half the sampling frequency and stored on the hard disk of a computer for subsequent analysis. and, thus, it was added to the internal solution that dialyzes the cells [9]. In these experiments, the tip of the pipette was filled with AIP-free internal solution, in order to obtain ''control'' current records.

Pulse protocols (mouse ventricular myocytes).
To construct the current-voltage relationships for I Na , 50ms pulses in 5 mV increments from -120 mV to potentials between -100 and +30 mV were applied. To construct the inactivation curves, I Na was recorded by applying 500-ms pulses from -120 mV to potentials between -140 and -20 mV in 10 mV increments followed by a test pulse to −40 mV. To analyze the recovery from inactivation of I Na , two 50-ms pulses (P1 and P2) from -120 to -40 mV were applied at increasing coupling intervals (0.05-500 ms). A monoexponential function was fitted to the data to measure the reactivation kinetics.
The protocol to record I K1 consisted of 250-ms steps that were imposed in 10 mV increments from -80 mV to potentials ranging -120 and -40 mV. I K1 was always measured at the end of the 250-ms pulses.
Current-voltage relationships for I to,f were constructed by applying 500-ms pulses in 10 mV increments from -80 mV to potentials ranging -120 and +50 mV. The I to,f was measured as the difference between peak current and the current at the end of the 500-ms pulse and the I to,f charge was estimated from the integral of the current traces comprising the area between the peak and the current at the end of the 500-ms pulse [13]. In another group of experiments, the I to,f and the rapidly activating slowly inactivating component (I K,slow ) were recorded by applying 4-s pulses from a holding potential of -80 mV to potentials ranging -90 and +50 mV in 10 mV increments. A biexponential function was fitted to the decay of currents generated by pulses to +50 mV. As previously described [17], under these conditions the amplitude of the fast component of the exponential can be identified as the amplitude of the I to,f , whereas the amplitude of the slow component can be identified as that of the I K,slow . Furthermore, the fast and slow time constants of current decay were assimilated as the inactivation time constants of the I to,f and the I K,slow , respectively. Superfusion of TEA (25 mM) inhibited I K,slow and I ss currents by ~60%, leaving the I to,f unaffected [17].
Inactivation curves for I to,f were constructed by applying 500-ms pulses in 10 mV increments from -80 mV to potentials ranging -120 and +50 mV followed by a test pulse to +50 mV. To analyze the recovery from inactivation of I to,f , two 500-ms pulses (P1 and P2) from -80 to +50 mV were applied at increasing coupling intervals (5-4000 ms). A monoexponential function was fitted to the data to measure I to,f , reactivation kinetics in myocytes isolated from Sham and WT SAP97 mice. I to,f , reactivation data from p.P888L SAP97 mice were fitted by a biexponential function.

Pulse protocols (CHO cells).
To obtain current-voltage relationships for I Kv4.3 , 500-ms pulses in 10 mV increments from -80 mV to potentials between -90 and +50 mV were applied. I Kv4.3 was measured as the maximum peak amplitude at each potential. I Kv4.3 was also measured as the total charge crossing the membrane estimated from the integral of the current traces elicited at each potential. To obtain the inactivation curves for I Kv4.3 , a two-step protocol was used consisting of a first 500-ms conditioning pulse from -80 mV to potentials between -90 and +50 mV, followed by a test pulse to +50 mV. To analyze the recovery from inactivation of Kv4.3 channels, two 500-ms pulses (P1 and P2) from -80 to +50 mV were applied at increasing coupling intervals (5-4000 ms). A monoexponential function was fitted to the data (P2/P1 plotted as a function of the time interval) to obtain the time constant that defines the process.
Activation/conductance voltage curves for I Kv4.3 recorded in CHO cells and for I Na and I to,f recorded in mouse ventricular myocytes were constructed by plotting the normalized conductance as a function of the membrane potential. The conductance was estimated for each experiment by the equation: where G is the conductance at the test potential V m , I represents the peak maximum current at V m , and E rev is the reversal potential. To determine the E rev , I Na density-voltage relationships obtained in each experiment were fitted to a function of the form: where I is the peak current elicited at the test potential V m , G max is the maximum conductance, and k is the slope factor.
For I Kv4. 3 and I to,f , the E rev introduced for the calculation of the conductance was the value described in previous studies [14].
This procedure yields conductance curves for I Kv4. 3 and I to,f with less steep slopes than those obtained when the voltage dependence of the activation is measured with the full envelopes of tails. However, it is a procedure widely used for the comparative analysis of the effects on the voltage-dependence of activation [1,14].
Activation/conductance-voltage and inactivation curves were fitted with a Boltzmann distribution according to the following equation: where A is the amplitude term, V h is the midpoint of activation/inactivation (in mV), V m is the test potential and k represents the slope factor for the curve.
In each experiment, current amplitudes were normalized to membrane capacitance to obtain current densities. Action potentials were recorded using the current clamp configuration and elicited by depolarizing-current pulses of 2 ms in duration at 1.5 times the current threshold at a frequency of 1 Hz.
Pulse protocols (Ltkcells). The protocol used to obtain I Kv1.5 current-voltage relationships consisted of 500-ms pulses in 10 mV increments from a holding potential of -80 mV to potentials ranging -80 and +60 mV. Between −80 and −40 mV, only passive linear leak was observed and least-squares fits to these data were used for passive leak correction. Deactivating 'tail' currents were recorded on return to −40 mV. The activation curves were constructed by plotting tail current amplitude as a function of the membrane potential and were fitted with a Boltzmann function (see above). A monoexponential function was fitted to the current traces at +50 mV to obtain the time constant of activation and inactivation and to the tail current at -40 mV after pulses to +50 mV to determine the time course of deactivation.

Western-blot analysis.
Detection of SAP97, Nav1.5, and Kir2.1 proteins was carried out in ventricular samples from Sham, WT or p.P888L SAP97 mice by Western-blot following previously described procedures [2,3,6,8,9,11,16].  Table III). For each cluster, Cluspro yields the size (number of conformations/members), the weighted energy score of the cluster center (i.e. the structure that has the highest number of neighbour structures in the cluster), and the energy score of the lowest energy structure in the cluster. The model is based on cluster size rather than on energy. Indeed, low energy regions tend to generate large clusters of docked structures, and the size of a cluster is approximately proportional to its probability, therefore, the higher the size, the more likely the conformation of the complex.

Statistical analysis.
Results are expressed as mean±SEM. Unpaired t-test or one-way ANOVA followed by Tukey's test were used where appropriate. In small-size samples (n<15), statistical significance was confirmed by using nonparametric tests. To take into account repeated sample assessments, data were analyzed with multilevel mixed-effects models. Comparisons between categorical variables were done using Fisher´s exact test. A value of P<0.05 was considered significant.       Total protein gel WB Nav1.   A negative SVM score indicates a deleterious substitution, considering that the larger the score, the more confident the classification. The higher the entropy score and PSSM scores are, the better tolerated is a mutation. PROVEAN scores classified the variants as deleterious if the cut-off value is ≤ −2.5 and as neutral if the cut-off value is ≥ −2.5. When the SIFT score is below the threshold (0.05) it is considered that the variant "affects protein function" and "tolerant" otherwise.
Up to now, 836 DLG1 variants have been annotated in the Exome Aggregation Consortium database (http://exac.broadinstitute.org/). More than 50% of these variants are located in non-coding regions (intronic variants). Among the variants located in coding regions, up to 250 are non-synonymous or missense substitutions. We selected the 5 most frequent missense DLG1 variants according to their total allele frequency value obtained from the Exome Aggregation Consortium database. We used 3 well recognized bioinformatics tools (SNP3d, PROVEAN, and SIFT) to predict the impact on the protein structure and function produced by these variants. The result of the analysis consistently shows that among the 5 most frequent missense DLG1 variants, the p.P888L substitution is the most deleterious one i.e., the variant with the highest probability of changing structure, stability and function of the SAP97 protein.
inactivation yielded by the fit of a biexponential function to the K + current traces recorded by applying 4s pulses to +50 mV. τ finact and τ sinact = fast and slow time constants of current inactivation yielded by the fit of a biexponential function to the I Kv4.3 traces recorded by applying 500-ms pulses to +50 mV; τ inact = time constant of inactivation yielded by the fit of a monoexponential function to the I Kv1.5 traces recorded by applying 500-ms pulses to +60 mV. τ react = time constant of recovery from inactivation for I to,f and  τ act = time constant of activation measured at peak maximum current; τ finact and τ sinact = fast and slow time constants of current inactivation measured at peak maximum current; τ react = time constant of recovery from inactivation; V hact and k act = midpoint and slope values of conductance-voltage curves; V hinact and k inact = midpoint and slope values of the inactivation curves. Each value represents mean±SEM of >6 experiments in each group.