Physiological consequences of the P2328S mutation in the ryanodine receptor (RyR2) gene in genetically modified murine hearts

Aim To explore the physiological consequences of the ryanodine receptor (RyR2)-P2328S mutation associated with catecholaminergic polymorphic ventricular tachycardia (CPVT). Methods We generated heterozygotic (RyR2p/s) and homozygotic (RyR2s/s) transgenic mice and studied Ca2+ signals from regularly stimulated, Fluo-3-loaded, cardiac myocytes. Results were compared with monophasic action potentials (MAPs) in Langendorff-perfused hearts under both regular and programmed electrical stimulation (PES). Results Evoked Ca2+ transients from wild-type (WT), heterozygote (RyR2p/s) and homozygote (RyR2s/s) myocytes had indistinguishable peak amplitudes with RyR2s/s showing subsidiary events. Adding 100 nm isoproterenol produced both ectopic peaks and subsidiary events in WT but not RyR2p/s and ectopic peaks and reduced amplitudes of evoked peaks in RyR2s/s. Regularly stimulated WT, RyR2p/s and RyR2s/s hearts showed indistinguishable MAP durations and refractory periods. RyR2p/s hearts showed non-sustained ventricular tachycardias (nsVTs) only with PES. Both nsVTs and sustained VTs (sVTs) occurred with regular stimuli and PES with isoproterenol treatment. RyR2s/s hearts showed higher incidences of nsVTs before but mainly sVTs after introduction of isoproterenol with both regular stimuli and PES, particularly at higher pacing frequencies. Additionally, intrinsically beating RyR2s/s showed extrasystolic events often followed by spontaneous sVT. Conclusion The RyR2-P2328S mutation results in marked alterations in cellular Ca2+ homeostasis and arrhythmogenic properties resembling CPVT with greater effects in the homozygote than the heterozygote demonstrating an important gene dosage effect.

. Each ($565 kDa) RyR2 monomer is associated with a range of accessory, potentially regulatory, proteins. These include calmodulin (CaM), the 12.6 kDa protein FKBP12.6 (also known as calstabin 2), protein kinase A (PKA), protein phosphatase 1 and 2A (PP1 and PP2A), and calmodulin-dependent protein kinase II (CaMKII) bound to its cytoplasmic scaffold domain at the N-terminal region (Bers 2002). It has been suggested that FKBP12.6 binding to RyR2 stabilizes the closed state of the channel during diastole and promotes the coupling of RyR2 subunits enabling channel opening during excitation contraction coupling (Yano et al. 2000, 2005a, Bers 2002. The gene encoding RyR2s has been implicated in the autosomal dominant inherited condition, catecholaminergic polymorphic ventricular tachycardia (CPVT), which is associated with presentations of adrenergically (exercise)-induced bi-directional and polymorphic ventricular tachycardia (VT) potentially leading to sudden cardiac death (SCD) in the absence of structural cardiac disease (Laitinen et al. 2001, Priori et al. 2001, 2002, Wehrens & Marks 2003. It has been suggested that an increased release of SR Ca 2+ into the cytosol causes delayed diastolic afterdepolarizations (DADs) in the action potential leading to triggered activity in common with findings in digitalis-induced VT (Rosen & Danilo 1980, Leenhardt et al. 1995, Tunwell et al. 1996, Priori et al. 2001, Kummer et al. 2006. Thus recent reports have similarly implicated alterations in SR calsequestrin in such pathology (Chopra et al. 2007).
Two groups first identified RyR2 mutations in CPVT patients; neither reported structural cardiac abnormalities or echocardiographic evidence of cardiac failure (Laitinen et al. 2001, Priori et al. 2001). The first reported four missense mutations of which S2246L, R2427S and N4104K were sporadic and R4497C was found in five clinically affected mutant carriers. The second group demonstrated three unrelated Finnish families carrying mutations in P2328S, Q4201R and V4653F. Seventeen members of the first of these families carried the P2328S mutation. Of these, 10 of the 12 studied showed exercise-induced VT. Two of these P2328S carriers were asymptomatic but showed abnormalities on clinical testing. The carriers within the family showed normal QTc intervals, and mortality rates as reflected in Kaplan-Meier analyses of 30-33% by the age of 35 years in common with the Q4201R and V4653F carriers.
Genetically modified murine hearts are increasingly used as models of human cardiac disease (London, 2001). However, translation of results from isolated murine hearts to the in vivo clinical setting must consider differences between both (1) species as well as (2) the electrophysiological assessment protocols employed. Thus (1) despite similar anatomy and fibre organization, murine and human hearts are markedly different in basal heart rate, heart size and ion channel expression (Vaidya et al., 1999). Although murine and human ventricular APs share a steep upstroke reflecting rapid Na + channel gating (Guo et al. 1999) and a strong influence of transient outward repolarizing K + current (I to ) on AP recovery, differences in repolarizing currents cause murine ventricular APs to overshoot and fall sharply in contrast to the pronounced plateau phase of the human ventricular AP (Danik et al. 2002).
(2) A significant proportion of the physiological assay techniques examined isolated perfused as opposed to in situ hearts for limited (h) rather than prolonged (years) intervals. Nonetheless, at the very least, murine hearts provide useful resemblances with human hearts the extent of which merits basic investigation. Thus, they have successfully been used to replicate arrhythmogenic properties associated with Brugada (BrS) and LQT3 syndromes (Stokoe 2007a,b;Thomas et al. 2007a,b). Similarly, a CPVT-like arrrhythmogenesis has been observed under conditions of adrenaline and/or caffeine perfusion in a recent murine model with RyR2 modifications targeted instead at the C-terminal luminal sites implicated in store-overload-induced Ca 2+ release (SOICR) (Jiang et al. 2004, Cerrone et al. 2005. This article reports the generation of murine hearts containing P2328S mutations in the central region of the RyR2 gene close to the FKBP12.6 binding site, and represents the first time both heterozygotic (p/s) and homozygotic (s/s) transgenic mice are compared and studied. Previous studies in HEK293 cells had associated the P2328S variant with altered FKBP12.6 binding following PKA phosphorylation ). This variant therefore directly complements genetically modified murine hearts with alterations targeted at the C-terminal luminal sites thought to be involved in SOICR (Jiang et al. 2004, Cerrone et al. 2005. The present experiments examined the physiological consequences of both p/s and s/s mutations not only through cytosolic Ca 2+ measurements in isolated cardiac myocytes but also electrophysiological recordings in Langendorffperfused hearts, and the extent to which the intact murine systems show arrhythmogenic properties. Our experiments thus clarified the physiological consequence of the P2328S (p/s) and (s/s) mutation, and defined the extent to which results from physiological studies demonstrate that this murine system resembled the human phenotype.

Generation of RyR2 P2328S knock-in mice
The targeting vector for homologous recombination consisted of a 5.89 kb genomic DNA fragment spanning a region from intron 44 to intron 47 of the RyR2 genomic sequence. The 5¢ and 3¢ flanks were amplified from 129Sv/Ev genomic DNA by PCR with PfuTurbo (Stratagene, Amsterdam, the Netherlands) using primers designed to the mouse genome reference sequence ENSMUSG00000021313. The codon change leading to the P2328S variant occurs in exon 45. The 5¢ flank was amplified using primers on either side on exon 45 to produce a 965-bp fragment. The forward primer contained an AvrII site which could be used to linearize the final targeting construct (5¢flankror: GTAAAGAAT-GAGACCTAGGATATAGTT; 5¢ flankrev: GCTTAG-CTAGCCTGAAAATACCTACA). The 3¢ flank was a 4925-bp PCR product amplified to continue the genomic homology from a NheI site at the 3¢ end of the 5¢ flank. Both forward and reverse primers contained NotI sites for cloning into the targeting construct. (3¢ flankfor: GCTAAGCGGCCGCAAAGAACACAGT TCAGAAACCTAT; 3¢flankrev: ATGGTAGCGGCC-GCTACCTAATATCACAAACACA).
Both flanks were cloned into pCR-BluntII-Topo (Invitrogen, Paisley, UK) and sequenced. Clones whose exon sequence was 100% identical to that predicted by the database were chosen to form the targeting construct. Single base changes to intron sequence (compared to the mouse genome sequence, C57Bl/6) were accepted if they occurred in more than one clone. To introduce the required mutation into the 5¢ flank the Invitrogen GeneTailor site-directed mutagenesis system was used. The forward mutagenesis primer (GGTGAGGCTGCT CATCCGGAGAtCtGAGTGCTT) changes the coding sequence from proline (CCC) to serine (tCt) and introduces a novel BglII site for screening for introduction of the mutation. (reverse primer: TCTCCGGATGAGC AGCCTCACCACAACATTT). Plasmids which had been correctly mutagenized were identified by sequencing and digestion with BglII (pCR-RyR25¢P2328S).
To create the targeting construct, the 3¢ flank was cloned into the NotI site of pCR-RyR25¢P2328SDTA. To reduce non-homologous recombination in the ES cells the PGK promoter-driven Diptheria toxin-A chain gene (from pKO3¢/DT-Anew, a kind gift of Dr Michael Snaith) was introduced 3¢ of the 3¢ flank. The loxP flanked thymidine kinase/neomycin selection cassette (loxPtkneo) (from plasmid pNT4) was added between the two flanks to give the final targeting construct, pCR-RyR25¢P2328SloxPtkneo3¢DTA (Fig. 1a). The targeting construct was checked by sequencing and restriction digestion at each stage of its construction.
Two electroporations of the linearized targeting construct into 129Sv/Ev ES cells were performed from which 132 ES clones were picked. After PCR screening for correct homologous recombination of the 5¢ (Hotstar Taq: Qiagen, West Sussex, UK) and 3¢ flank (Long Range Taq: Fermentas, York, UK) only one clone was found to be positive. The primer positions for this screening are indicated in Figure 1a. The PCR products of each screen were cloned and sequenced to show that there had been no sequence rearrangements and that the only detectable sequence change was the desired introduction of the codon change. This was confirmed by BglII digestion. The ES clone was karyotyped to confirm it had the correct number of chromosomes (40).
To remove the loxPtkneo selection cassette this clone was electroporated with plasmid expressing Cre Recombinase (pPGKnlsCreSV40polyA, a kind gift from Dr Sam Aparicio) and selected in gancyclovir to promote the loss of the loxPtkneo selection cassette. Twelve colonies were picked and screened by PCR to check for loss of the selection cassette (Primers RyR2f: GGGGAGTTGGCTGCTGAGAA; RyR2r: CGGGCA-CACACTCACCACCAA). The PCR product was cloned and sequenced to confirm that it contained the expected RyR2 genomic sequence with 204 bp of vector, including a single loxP site, introduced into intron 45 at the NheI site at the end of the 5¢ flank. The karyotype of the new clone was reconfirmed and then this clone was injected into C57Bl/6 blastocysts to produce chimeras. Confirmed germline transmitting chimeras were mated to 129Sv/Ev females. The line was then maintained as a mixture of RyR2 p/s heterozygous matings or RyR2 s/s homozygous matings.
Reverse transcription PCR. Total RNA was isolated from mouse ventricle or skeletal muscle using the Qiagen RNeasy fibrous tissue mini kit. The RNA was quantified using a Nanodrop ND-1000 (NanoDrop Technologies, Inc., Wilmington, DE, USA, supplied by LabTech International, East Sussex, UK) and 0.5 lg of RNA was reverse transcribed using random primers and AMV Reverse Transcriptase (both from Promega, Southampton, UK) in the following reaction: RNA was incubated at 70°C for 10 min, then placed on ice and the following reagents added: 2 lL 0.1 m DTT, 4 lL 5· RT buffer, 1 lL 10 mm dNTPs, 1 lL lg )1 Random primers, 0.5 lL RNAguard (Pharmacia, Sandwich, UK), 10 U AMV RT in a total volume of 10 lL. The reaction mix was then incubated at 25°C for 10 min, 42°C for 60 min and 70°C for 15 min. Control reactions in which the RT enzyme was replaced with water were included to exclude amplification from genomic DNA. The resulting cDNA was then diluted to 50 mL and stored at )80°C. To set up the PCR reaction 2 lL of cDNA was added to 0.5 lL 10 mm RTRev, CCAGCATGAATCAAGTGCATTTC; mm b actin for, GTTACCAACTGGGACGACATGG, mm b actin rev, CCATACCCAAGAAGGAAGGCTG. The PCR products were digested by addition of BglII to the PCR mix and incubation at 37°C for 2 h. The digest was analysed by electrophoresis on a 2% agarose gel.
All mice used were originally supplied by Harlan (Oxon, UK) and were housed at room temperature and subjected to 12-h light/dark cycles. They were fed sterile rodent chow and had free access to water at all times. Mice used were wild-type (WT), heterozygote (RyR2 p/s ) and homozygote (RyR2 s/s ) all aged between 3 and 6 months. Experimental mice were bred from a 129 genetic background, which, along with C57 mice, are less susceptible to PES-induced arrhythmias than FBV or Black Swiss animals (Maguire et al. 2003). All mice used for subsequent experimental work were killed by cervical dislocation [Schedule 1: UK Animal (Scientific Procedures) Act 1986].

Measurements of cell [Ca 2+ ]
Single-cell experiments used mouse ventricular myocytes isolated adapting an established enzymatic digestion as described previously (Harding et al. 1992). Following killing, hearts were excised rapidly and cannulated before being connected to a Langendorff perfusion system. The heart was then perfused for 1.5 min at 3 mL min )1 with perfusion buffer containing (mm), 120 NaCl, 5 MgSO 4 , 5.4 KCl, 5 sodium pyruvate, 10 Hepes, 5 glucose, 20 taurine, 5 nitrilotriacetic acid (NTA; Sigma-Aldrich, Poole, UK), and 1-2 lm Ca 2+ , pH 7.0 at 37°C. Subsequently the heart was perfused for 25-28 min at 3 mL min )1 with digestion buffer containing (mm), 120 NaCl, 5 MgSO 4 , 5.4 KCl, 5 sodium pyruvate, 10 Hepes, 5 glucose, 20 taurine, to which was added collagenase type II (final concentration 1 mg mL )1 ; Worthington Biochemical, Lakewood, NJ, USA), hyaluronidase (final concentration 1 mg mL )1 ; Sigma-Aldrich) and 13 lm Ca 2+ and set at pH 6.8 at a temperature of 37°C. The pale and swollen heart was then removed and cut below the atria. The digested tissue was teased gently with fine forceps and the dissociated cells were transferred to a conical tube containing 12.5 mL digestion buffer with 50 mg bovine serum albumin (BSA) to inactivate digestion. The resulting myocytes were subject to further trituration using sterile plastic transfer pipettes, and allowed to sediment by gravity for 10 min.
The cell suspension was then centrifuged for 1.5 min at 200 rpm and the pellets re-suspended twice in wash buffer containing (mm): 119 NaCl, 4.2 KCl, 0.94 MgSO 4 , 1.2 KH 2 PO 4 , 20 Hepes, 11.5 glucose, 20 taurine and 1 mg mL )1 BSA, and bubbled with 95% O 2 + 5% CO 2 , pH 7.4, to remove all traces of NTA, collagenase and hyaluronidase. CaCl 2 was then cautiously re-introduced by stepwise additions of CaCl 2 to reach an approximate concentration of 1.25 mm by centrifuging the preparation at each Ca 2+ reloading stage to remove dead or swollen cells. During this process, myocytes were examined periodically under the microscope to confirm a good yield of rod-shaped myocytes (better than 60%) before continuing the experiments. Cells were then transferred into a conical tube (BD Biosciences, Oxford, UK) and incubated at room temperature. The isolated myocytes were then placed in a control Hepes-buffered Krebs-Henseleit (KH) solution in a Perspex chamber 10 · 4 · 6.25 mm (length · width · depth): the myocytes spontaneously attached to the glass cover slip (4 · 4 cm, grade 1 cover slip; (Merck, Herts, UK) forming the floor of the chamber. The myocytes were stimulated to contract using two in-built platinum field electrodes running the length of the chamber through which the periodic field stimulation was applied using a home-built squarewave stimulator (R. Montgomery, Imperial College, London, UK). This applied successive 40-60 V steps of 2.2 ms duration, immediately followed by a step of similar duration and amplitude but opposite polarity at a pacing frequency of 0.5 Hz.
During this stimulation cells were loaded with the acetoxymethyl (AM) ester of the Ca 2+ -sensitive dye Fluo-3 (F-14242; Invitrogen, Paisley, UK; 50 lg vial diluted in 30 lL pluronic to give a stock concentration of 1.476 mm) by incubating the cells in a bath of volume 500 lL containing KH buffer (with 1.25 mm Ca 2+ ) with 2 lL of Fluo-3 AM solution for 15 min at room temperature in the dark. The bath with the myocytes was then transferred onto the stage of a Zeiss LSM-510 laser scanning confocal microscope system (Carl Zeiss, Welwyn Garden City, UK) with a ·20 air objective (NA 0.5; confocal aperture 1000 lm, slice thickness <42.4 lm) on a Zeiss Axiovert 100 M inverted microscope. The fluorescence (F) signals resulting from Ca 2+ binding to fluo-3 were excited by a 488nm argon laser and detected between wavelengths of 505-550 nm. Images were analysed using an in-house custom-written program. Series of 500 frames sampled at 98 ms per frame (128 · 128 pixels per frame) were used initially to monitor fluorescence changes over time. The appropriateness of the chosen frame capture rate was corroborated in some experiments by confirming similar results following use of the faster line-scan mode with a sampling rate of 960 ls per frame. Fluorescence signals were quantified within three successive regions of interest (ROI) placed at 7 lm intervals along the long axis of each myocyte. Each ROI was circular, with an area of $50 lm 2 . Ca 2+ transients obtained in response to the regular pacing stimuli were quantified in terms of their peak values that were normalized to the baseline fluorescence level F 0 , to give peak F/F 0 values following each response to each stimulus. All values were calculated from three ROIs for each myocyte and the mean peak F/F 0 values as well as the baseline diastolic values were calculated. All traces had stable baselines confirming consistent laser intensities and detector gains. The results were expressed as mean AE SEM and compared using anova (spss software, SPSS, Chicago, IL, USA).

Whole heart electrophysiological experiments
The electrophysiological experiments on arrhythmogenesis at the whole heart level used a Langendorffperfused preparation adapted for the murine model (Papadatos et al. 2002, Balasubramaniam et al. 2003. The contracting heart was stimulated with platinumstimulating electrodes applied to the epicardial surface, usually 15 min at 10 Hz, and was allowed to reach its physiological steady state. The heart was paced from its right ventricle at three times its threshold level (between 3 and 5 V) using square-wave stimuli of 2 ms duration (Grass S48 stimulator; Grass Telefactor (UK), Slough, UK). Monophasic action potentials (MAPs) were recorded from the epicardial basal surface of the left ventricle using a MAP electrode (Linton Instruments, Harvard Apparatus, Edenbridge, UK).
Hearts were subject to two stimulation protocols. First, they were regularly paced at 6, 8, 10 and 12 Hz for periods of up to 30 min to assess for tendencies to spontaneous arrhythmogenesis. Secondly, programmed electrical stimulation (PES) of the heart was then carried out using a variation of an existing clinical technique (Saumarez & Grace 2000. This applied stimulation sequences that each consisted of a drive train of pacing stimuli (S1) applied with a 200 ms cycle length with an extra stimulus (S2) inserted every eighth beat. The coupling interval (S1S2) was successively reduced by 1 ms with each cycle until the ventricular effective refractory period (VERP) was reached. The MAP signals were amplified, filtered (band-pass filter 30 Hz to 1 kHz; Gould-Nicolet Technologies, Ilford, Essex, UK) and digitized using an analog-to-digital converter (Model 1401plus; Cambridge Electronic Design, Cambridge, UK).
The MAP signals were analysed with Spike II software (Cambridge Electronic Design) and MAP traces were selected according to previously documented criteria requiring stable baselines with rapid upstrokes and consistent amplitudes. The action potential durations (APD) at 30, 50, 70 and 90% repolarization (APD 30, APD 50 , APD 70 and APD 90 ) were fully identified and tabulated in Excel (Microsoft, Redmond, WA, USA). Results were all expressed as mean AE SEM comparing different groups of experimental animals using anova (spss software).
Isoproterenol was dissolved initially in distilled water to make a stock of 1 mm and the final drug concentration was achieved by dilution with the control KH buffer. The stocks were stored at )20°C and made fresh on the day of each experiment. These agents were obtained from Sigma-Aldrich.

RyR2 P2328S mice
The three injections of embryonic stem cells into C57B/ 6 blastocysts resulted in a total of eight male chimeras. Each of these males was mated with Ca 57Bl/6 and a 129Sv/Ev female. Three of the chimeras showed germ line transmission of the knock-in allele, two produced no litters and two produced litters but did not show germ line transmission. Genotyping of pups was performed using primers RyR2f and RyR2r (Fig. 1a). These span the vector sequence left in intron 46. A second PCR can be used to confirm the presence of the RyR2 P2328S allele. The reverse primer binds to vector sequence left in intron 45 and the primers span the introduced mutation to produce a 740-bp product which can be cut by BglII to produce two, 524 and 216 bp, products (Fig. 1b). Pups were genotyped at 3-4 weeks of age. Results of genotyping 404 mice gave 98 WT, 189 RyR2 p/s and 117 RyR2 s/s . This is not statistically different from the expected numbers of 101, 202 and 101. The oldest mice in the colony were 18 months of age. There has been no significant additional mortality of either RyR2 p/s or RyR2 s/s compared with WT.
To check that both alleles produced mRNA, RT-PCR was performed on total RNA isolated from mouse ventricle tissue. Figure 1c shows the results of RT-PCR after the product had been digested by BglII and the digest run on a 2% agarose gel. For each genotype the RT-PCR reaction was run in triplicate. RNA from WT mice produced a single product at 694 bp. RNA from the heterozygous RyR2 p/s mice produced three products, uncut WT allele and digested RyR2 P2328S allele. RNA from the homozygous RyR2 s/s mice produced only two products (438 and 256 bp) showing that the RT-PCR product is derived from the RyR2 P2328S allele only. The b-actin gene was amplified in all samples as a positive control. PCR reactions using the )RT cDNA were also included as negative controls. No amplification was seen in these reactions from either the RyR2 or b-actin primers (data not shown). To show that the RyR2 primers were not amplifying from RyR1-derived cDNA, skeletal muscle-derived cDNA was used as a template. No amplification from the RyR2 primers was seen in these reactions, but the b-actin primers were able to amplify the expected product from the same samples (data not shown).
The presence of the P2328S allele alters Ca 2+ homeostasis Differences in cellular Ca 2+ homeostasis that might relate to the presence of the P2328S allele were first investigated through changes in the amplitude and pattern of Ca 2+ transients in response to regular stimulation (0.5 Hz) in isolated Fluo-3-AM loaded WT (Fig. 2a,b), RyR2 p/s (Fig. 2c,d) and RyR2 s/s (Fig.  2e,f) myocytes. These were studied in KH buffer in both the absence and presence of isoproterenol at a concentration (100 mm) at which it would be possible to assess the extent to which the RyR2 p/s or RyR2 s/s mutation might either increase or decrease the incidence of ectopic or subsidiary Ca 2+ events. In all three groups, the traces showed stable baseline F/F 0 and reproducible responses to regular stimulation with minimal evidence of fluorophor bleaching over the sampling periods. Thus the average values of the diastolic baseline F/F 0 were statistically indistinguishable (P >> 5%) between all three groups whether in the presence and absence of isoproterenol. Similar stability conditions applied to the corresponding peak F/F 0 amplitudes. We further quantified the characteristics of such peak F/F 0 (Table 1) obtained from Ca 2+ transients of the kind illustrated in Figure 2. In the absence of isoproterenol, Ca 2+ signals observed immediately after the commencement of stimulation in WT (Fig. 2a) and RyR2 p/s myocytes (Fig. 2c) consisted of a series of responses each directly following the individual stimuli. WT (Fig. 2a) and RyR2 p/s myocytes (Fig. 2c) similarly did not show either subsidiary events during the recovery time courses of the evoked transients or spontaneous ectopic peaks in the intervals between these regular responses to the imposed stimuli. In contrast, the RyR2 s/s myocytes showed an irregular pattern of responses that included subsidiary events (4.65 AE 1.46 events counted over a standard sampling period of 50 s; n = 36 cells, three mice) (Fig. 2e). After the addition of isoproterenol, WT myocytes showed both ectopic peaks (8.00 AE 2.99; n = 6 cells, two mice) and subsidiary events (9.00 AE 1.34, n = 6 cells, two mice) (Fig. 2b), and RyR2 s/s myocytes showed an increased incidence of ectopic peaks (15.83 AE 6.76) compared to the correspondingly isoproterenol treated WT, but no subsidiary events. RyR2 p/s myocytes showed neither ectopic peaks nor subsidiary events ( Fig. 2 and Table 1). In contrast, Table 1 indicates that the peak F/F 0 of Ca 2+ transients directly evoked by each of the individual stimuli were statistically indistinguishable on independently weighted anova in WT, RyR2 p/s , RyR2 s/s regardless of whether or not isoproterenol was present with the exception of a reduced peak F/F 0 in isoproterenoltreated RyR2 s/s , as expected for a depletion of SR Ca similarly produced by isoproterenol and caffeine (Venetucci et al. 2007).
Further observations were made in cells following continual stimulation extending over $10 min. Resumption of scanning then confirmed a similar appearance of subsidiary events in RyR s/s but not WT or RyR2 p/s myocytes (Fig. 3a,c,e). Addition of isoproterenol then similarly resulted in the appearance of both ectopic peaks and subsidiary events in WT (Fig. 3b) and their continued absence in RyR2 p/s myocytes (Fig. 3d). Such results agreed with the findings made immediately following the onset of stimulation (Fig. 2) despite the now reduced peak F/F 0 values likely reflecting progressive SR Ca 2+ depletion by such prolonged stimulation. RyR2 s/s myocytes showed irregular patterns of release events (Fig. 3f). These were further studied (Fig. 4) under conditions of exposure to 0 (n = 7), 20 (n = 4), 50 (n = 4) and 100 nm (n = 5) isoproterenol respectively: this gave progressive (and significant) reductions in peak F/F 0 from 4.045 AE 1.021 to 3.846 AE 0.068 (n = 5), 1.63 AE 0.0048 (n = 5) and 1.470 AE 0.058 (n = 5; P = 0.03 on independent weighted analysis), in which a reduction in the frequency of subsidiary peaks from 16.33 AE 2.08 to 6.57 AE 1.72 and 5.8 AE 4.82, following a series of 22 stimulations in 20 and 50 nm isoproterenol, respectively, was replaced by the irregular pattern of release events at 100 nm isoproterenol (Fig. 4d).
RyR2 s/s cells demonstrated Ca 2+ waves hitherto associated with arrhythmogenic properties in caffeinetreated murine myocytes , following the addition of isoproterenol. As described above, Ca 2+ signals, synchronized to the applied stimuli, were observed in the presence of KH buffer with 1.25 mm Ca 2+ for the WT and RyR2 p/s myocytes before and following addition of 100 nm isoproterenol. However, addition of 100 nm isoproterenol resulted in an appearance of propagated Ca 2+ waves along the longitudinal axis of the cell in the RyR2 s/s (Fig. 5). The arrows highlight the path taken by the Ca 2+ wave, starting from the upper end of the cell and moving along its length. Similar results were observed in n = 5 cells.

Action potentials from hearts containing the P2328S allele during regular pacing
Experiments on isolated Langendorff-perfused whole hearts compared electrophysiological properties of WT, RyR2 p/s and RyR2 s/s hearts, exploring for arrhythmogenic tendency both before and following addition of 100 nm isoproterenol in each heart. Mice were aged between 3 and 6 months (WT: one female, two males; RyR2 p/s : four females, four males; and RyR2 s/s : one female, six males). The experiments used both regular pacing and PES procedures, which imposed an S2 extra stimulus following eight-beat (S1) conditioning Symbols indicate values subject to one-way anova of findings obtained between designated groups, or in each group before and after introduction of isoproterenol with the following results: P = ns, *P < 0.05, **P < 0.01, ***P < 0.001.
The experiments first examined MAP waveforms in hearts regularly paced at 10 Hz for at least 30 min in KH buffer before and after addition of isoproterenol. They assessed for the presence or absence of changes in APD of the kind reported for LQTS mice models on recent occasions (Head et al., 2005;Stokoe et al., 2007a,b;Thomas et al., 2007a,b). APD values measured at 30, 50, 70 and 90% recovery (APD 30,50,70,90 ) from regularly paced WT (n = 3 hearts), RyR2 p/s (n = 7 hearts) and RyR2 s/s (n = 7 hearts) before and 20 min after addition of isoproterenol indicated uniform MAPs (Fig. 6ai,ii, bi,ii and ci,ii respectively), with no significant differences in all these APD durations between WT (Fig. 6aiii), RyR2 p/s (Fig. 6biii) and RyR2 s/s hearts (Fig. 6ciii). Such findings parallel reports on a previously described R4496C mutant where QT intervals, of WT and the mutant mice did not present significant differences whether before or after adrenergic stimulation (Cerrone et al. 2005). These data therefore parallel findings that CPVT carriers of the RyR2-P2328S mutation (Paavola et al. 2007) showed increases in endocardial MAP durations considerably smaller than the clinical QTc prolongations thought to account for the LQTS phenotype (Keating & Sanguinetti, 2001;Scoote & Williams, 2002;Laurita & Katra, 2005;.

Arrhythmic properties of hearts containing the P2328S allele during regular pacing and programmed electrical stimulation at varying frequencies
The presence or absence of an arrhythmogenic phenotype was then investigated in the WT, RyR2 p/s and RyR2 s/s hearts subject to both regular pacing and PES. VT was identified as a series of regular electrical deflections recurring at high frequencies independent of any pacing spikes that were present. Arrhythmias lasting >30 s were defined as being sustained (sVT) and those lasting <30 s as being non-sustained (nsVT). On these criteria: (1) regularly paced WT hearts showed no arrhythmic incidents at all frequencies (6, 8, 10 and 12 Hz) explored whether isoproterenol was present (n = 3) or absent (n = 4 hearts) (data not shown).
(2) The RyR2 p/s hearts showed relatively few arrhythmogenic phenomena, when regularly paced, in the absence of isoproterenol. Thus, of seven hearts, only one then showed an episode of sVT at a pacing rate of 12 Hz. However, addition of isoproterenol (100 nm) enhanced arrhythmic tendency with 1, 1 and 0 episodes of nsVT at pacing frequencies of 6, 8 and 10 Hz and 0, 1, 0 and 1 episodes of sVT at a pacing frequency of 6, 8, 10 and 12 Hz respectively (n = 5 hearts).
(3) RyR2 s/s hearts showed noticeably greater arrhythmic tendencies even prior to addition of isoproterenol (Fig. 7). Two hearts then showed episodes of nsVT at 8 and 10 Hz (Fig. 7b), and two further hearts showed episodes of sVT at 10 Hz (n = 11 hearts). (4) Addition of isoproterenol increased this arrhythmic tendency. Thus, its introduction resulted in (n = 5 hearts) 2, 1, 0 and 0 episodes of sVT at pacing frequencies of 6, 8, 10 and 12 Hz respectively. Finally, spontaneous after-depolarizations, previously implicated in the triggering of arrhythmias, were observed in hearts paced at both 8 and 10 Hz (n = 2) (Fig. 7c). The subsequent electrophysiological experiments explored for arrhythmogenic substrate as reflected in initiation of VT by extrasystolic (S2) stimulations applied during PES procedures (Fig. 8). This gave results that correlated well with the pacing studies. They demonstrated that (5) WT hearts were not arrhythmogenic at all pacing frequencies, whether before (n = 4, Fig. 8a) or after (n = 3, Fig. 8b) introduction of isoproterenol. In contrast, (6) both RyR2 p/s and RyR2 s/s hearts showed an arrhythmogenicity exacerbated by isoproterenol. Thus, five of seven RyR2 p/s hearts (Fig. 8c) showed nsVT at all applied pacing frequencies. Isoproterenol (100 nm) increased this incidence in four of five hearts. The RyR2 s/s hearts showed not only higher incidence of nsVT, but also an occurrence of sVT (n = 11 hearts, Fig. 8e). Addition of 100 nm isoproterenol resulted in full sVT (n = 5) (Fig. 8f). Taken together, these results suggest that the RyR2-P2328S mutation results in increased tendencies to nsVT as well as sVT in mice. This is  (Swan et al. 1999, Laitinen et al. 2001, Priori et al. 2001, 2002.

RyR2 s/s hearts additionally show arrhythmic behaviour during intrinsic activity
Spontaneous arrhythmogenic phenomena were only observed in intrinsically paced RyR2 s/s and not in WT (n = 3) or RyR2 p/s (n = 7). Six of 11 RyR2 s/s hearts showed early after depolarizations (EADs) that intercepted the recovery phase of their preceding action potentials (Fig. 9a), and 11 of 11 showed coupled beats consisting of pairs of full action potentials (Fig. 9b). There was one example of nsVT following the coupled beats. When stimulation was ceased and re-introduced at a pacing rate of 10 Hz, it resulted in an episode of ventricular fibrillation (Fig. 9c). These findings are consistent with reports in which CPVT patients exhibited episodes of polymorphic VT leading to SCD while asleep or at rest (Allouis et al. 2005, d'Amati et al. 2005, Postma et al. 2005).
In contrast to murine models directed to the C and N-terminal regions, there have been no such models involving the important central region (Cerrone et al. 2005, Kannankeril et al. 2006. This study reports the generation of mice with a knock-in mutation in the Monophasic action potential waveforms of whole isolated perfused hearts in Krebs-Henseleit buffer from RyR2 s/s (a). Of the three groups studied, RyR2 s/s hearts showed the highest incidence of ventricular tachycardias (b) which could be either sustained (>30 s) or non-sustained (<30 s) and early after depolarizations (EADs) (asterisks) (c). Results shown for hearts subject to regular pacing at 8 Hz.
central domain of the RYR2 gene (P2328S). The variant was first identified, in common with carriers for the Q4201R and V4653F in three unrelated Finnish families. The P2328S mutation is located in the large central cytoplasmic domain of RyR2, near the binding site of the regulatory protein FKBP12.6 (George et al. 2007); the Q4201R and V4653F mutations occur in the carboxy-terminal region of the receptor associated with the pore region (Tunwell et al. 1996). All the human carriers showed mortality rates of %33% between the ages of 10 and 35 years and threshold heart rates of $130 beats min )1 for the triggering of ventricular arrhythmogenesis during exercise despite an absence of structural abnormalities or heart failure (Laitinen et al. 2001). Recent clinical studies of carriers for the P2328S, as well as the Q4201R and V4653F, mutations reported runs of polymorphic tachycardia during exercise with a mean age of onset of symptoms of 34 AE 16 years. Furthermore, endocardial MAP recordings at the right ventricular septum from patients with one of the three RyR2 mutations showed delayed afterdepolarizations (DADs) in three of 15 CPVT patients. These occasionally coincided with premature ventricular complexes both before and during adrenaline infusion that were never observed in control subjects (Paavola et al. 2007).
There has been extensive discussion about the association between mutations in the RyR2, including that of P2328S, and the CPVT phenotypes, particularly concerning possible roles of FKBP12.6-RyR2 binding (Brillantes et al., 1994;Masumiya et al. 2003, Zissimopoulos & Lai 2005 Hz respectively) showed nsVT. In four of five RyR2 p/s hearts studied in the presence of 100 nm isoproterenol, 0, 3, 3, and 1 hearts showed nsVT at pacing frequencies of 6, 8, 10 and 12 Hz, respectively, and one heart showed sVT both at 10 and 12 Hz. The RyR2 s/s hearts showed 3, 3, 3, and 5 episodes of nsVT and 0, 1, 3 and 1 episodes of sVT at pacing rates of 6, 8, 10 and 12 Hz respectively (n = 11 hearts). Addition of 100 nm isoproterenol (n = 5) resulted in a high incidence (2, 3, 3 and 4 episodes) of sVT. In the example shown, the S2 extra-stimuli initiated an episode of non-sustained VT lasting for less that 30 s in KB buffer alone (e). Sustained VT lasting for more than 30 s was often observed following addition of 100 nm isoproterenol in the same heart during PES at 10 Hz (f). Scale bar 200 ms.
V4653F) mutations suggested a gain-of-function defect with leaky Ca 2+ release channels showing significant rightward shift in half-maximal inhibitory Mg 2+ concentration and decreased FKBP12.6 binding , Paavola et al. 2007. However, other studies have reported that mutations in the RyR2 central region variously showed increased (N2386I and Y2392C) and decreased (R2427S, S2246L and P2328S) FKBP12.6 binding affinity (Tiso et al. 2002 and respectively), although a decreased affinity with C-terminal mutations (Q4201R, R4496C and V4653F) . S2808D-myotubes did not show spontaneous Ca 2+ oscillations at any concentration of FKBP12.6 (Goonasekera et al., 2005). Fluorescence resonance energy transfer investigations of the functional impact of C-terminal and central domain RyR2 mutations suggested effects on RyR2 channel activity independent of FKBP12.6 binding and PKA phosphorylation and reviewed in George et al. 2003, Jiang et al. 2004, Liu et al. 2006. Such discrepancies led to alternative suggestions implicating elevations of SR store Ca 2+ content to a threshold level producing spontaneous SOICR with unaltered FKBP12.6-RyR2 interaction in CPVT. SOICR occurred in HEK293 cells expressing WT RyR2 exposed to elevated extracellular [Ca 2+ ] 0 (Diaz et al. 1997, Jiang et al. 2004. It also occurred in cells expressing RyR2 containing the N4104K, R4496C and N4895D mutations, as well as further RyR2 mutations from central (S2246L and R2474S), C-(Q4201R and I4867M) and N-terminal (R176Q (T2504M) and L433P) regions, in the form of increased frequencies of Ca 2+ oscillations and decreased SR Ca content. Single mutant RyR2 channels showed increased activation sensitivities with a luminal Ca 2+ of 300 lm and increased basal [ 3 H] ryanodine binding (see also Jiang et al. 2005). Such findings were corroborated using the first RyR2-R4496C (+/)) knock-in mouse model equivalent to the R4497C mutation identified in CPVT patients. Five of 14 such mice showed either nsVT or sVT only with exercise stress testing followed by adrenaline administration. Four of eight of a second group developed arrhythmias following adrenaline and caffeine administration, as did four of five pre-treated with K201 (Cerrone et al. 2005). Recent optical mapping studies further characterized these arrhythmogenic characteristics (Cerrone et al. 2007). Subsequent analyses demonstrated that pacing protocols of increasing frequencies induced DADs in RyR2-R4496C (+/)) cardiomyocytes but not in the WT. Superfusion with isoproterenol induced both DADs and triggered activity in the RyR2-R4496C (+/)) but not in the WT. Finally, the heterozygote showed slightly increased RyR2-FKBP12.6 interaction compared with WT that was unchanged in the presence or absence of caffeine and adrenaline (Liu et al. 2006).
Our data provide complementary mouse models using the P2328S mutation, at a central site remote from the luminal C-terminal, R4496C site. We have for the first time compared all three genotypes (WT, RyR2 p/s  Figure 9 Monophasic action potentials (APs) recorded from the epicardium of intrinsically paced initially non-arrhythmogenic RyR2 s/s hearts. Experiments were performed in the absence (a, b) and presence of electrical stimulation prior to (c). Of the spontaneous events observed both (a): early-after depolarizations (EADs) (*) that produced episodes of sustained monomorphic VT (sVT) when stimulation resumed, and (b): coupled beats (**) were observed in six of 11 hearts. (c) Persistent ventricular fibrillation (VF) occurring following the cessation of regular S1 pacing after two intrinsic MAPs. and RyR2 s/s ) at both the single-cell and whole heart level. The presence of P2328S allele appears to impart an arrhythmogenic tendency, greater in both RyR2 p/s and RyR2 s/s compared with WT. We report successively greater effects on Ca 2+ transients, arrhythmogenicity, the effect of b-adrenergic stimulation, and the presence and absence of afterpotentials in the order WT < RyR2 p/s < RyR2 s/s , adding important observations concerning the effects of gene dosage. Thus, there was an increased amplitude of Ca 2+ transients evoked by regular stimulation in isolated ventricular myocytes with ectopic Ca 2+ peaks occurring specifically in the RyR2 s/s myocytes as was observed following administration of isoproterenol to the WT. Isoproterenol reduced such transients in RyR2 s/s , suggesting Ca 2+ store depletion and an appearance of propagated Ca 2+ waves associated with arrhythmogenic tendencies on earlier occasions (Venetucci et al. 2007). In contrast, MAPs from isolated Langendorff-perfused WT, RyR2 p/s and RyR2 s/s hearts had similar waveforms and refractory periods, whether in the presence or absence of isoproterenol. Nevertheless, both regular pacing through a range of frequencies and extrasystolic stimuli demonstrated a progressively increasing tendency to the development of non-sustained, then sustained, arrhythmogenesis, in isolated, Langendorff-perfused RyR2 p/s and RyR2 s/s but not WT hearts. These effects were more marked at higher pacing frequencies and in the RyR2 s/s hearts. The latter showed the greatest arrhythmogenic tendency as well as extrasystolic events often followed by spontaneous sustained VT even during intrinsic beating. In addition to establishing an arrhythmogenic phenotype, and associating this with altered Ca 2+ homeostasis, these data therefore associated a physiological CPVT phenotype with the clinically described P2328S mutation, in the cytoplasmic central region of the RyR2.

Conflicts of interest
We report no conflicts of interest.
Ca 2+ transients as obtained from F/F 0 values obtained over fixed regions of interest, that were obtained from WT myocytes (Fig. 2b). These were made using frame captures of fluorescence (F) signals resulting from Ca 2+ binding to fluo-3 excited by a 488-nm argon laser and detected between wavelengths of 505-550 nm sampled at 98 ms/frame (128 · 128 pixels/frame). The appropriateness of the chosen frame capture rate was corroborated in some experiments by confirming similar results following use of the faster line-scan mode with a sampling rate of 960 ls/frame. The first series of images are from unstimulated RyR2 s/s myocyte studied in the presence of isoproterenol. It illustrates a propagating increase of local Ca 2+ fluorescence running along the cells concerned. This is in contrast to the second series of images from a regularly stimulated WT myocyte in which the evoked Ca 2+ is synchronized over the whole cell surface.
Video Clip S1. (1) Video Clip S2. (2) Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.