Imaging Ca2+ Nanosparks in Heart With a New Targeted BiosensorNovelty and Significance
Rationale: In cardiac dyads, junctional Ca2+ directly controls the gating of the ryanodine receptors (RyRs), and is itself dominated by RyR-mediated Ca2+ release from the sarcoplasmic reticulum. Existing probes do not report such local Ca2+ signals because of probe diffusion, so a junction-targeted Ca2+ sensor should reveal new information on cardiac excitation–contraction coupling and its modification in disease states.
Objective: To investigate Ca2+ signaling in the nanoscopic space of cardiac dyads by targeting a new sensitive Ca2+ biosensor (GCaMP6f) to the junctional space.
Methods and Results: By fusing GCaMP6f to the N terminus of triadin 1 or junctin, GCaMP6f-triadin 1/junctin was targeted to dyadic junctions, where it colocalized with t-tubules and RyRs after adenovirus-mediated gene transfer. This membrane protein-tagged biosensor displayed ≈4× faster kinetics than native GCaMP6f. Confocal imaging revealed junctional Ca2+ transients (Ca2+ nanosparks) that were ≈50× smaller in volume than conventional Ca2+ sparks (measured with diffusible indicators). The presence of the biosensor did not disrupt normal Ca2+ signaling. Because no indicator diffusion occurred, the amplitude and timing of release measurements were improved, despite the small recording volume. We could also visualize coactivation of subclusters of RyRs within a single junctional region, as well as quarky Ca2+ release events.
Conclusions: This new, targeted biosensor allows selective visualization and measurement of nanodomain Ca2+ dynamics in intact cells and can be used to give mechanistic insights into dyad RyR operation in health and in disease states such as when RyRs become orphaned.
- biosensing techniques
- calcium signaling
- excitation-contraction coupling
- ryanodine receptor calcium release channel
Cardiac contraction is caused by intracellular Ca2+ release as a result of signaling in a nanoscopic domain, the dyad, formed from the close apposition of the sarcoplasmic reticulum (SR) and surface membranes.1 Two types of Ca2+ channels, the voltage-gated L-type Ca2+ channels, which control Ca2+ influx elicited by action potentials, and ryanodine receptors (RyRs), which mediate Ca2+ release from intracellular stores, reside on the surface and SR membranes of the dyad, respectively. Orthograde L-type Ca2+ channel-to-RyR coupling is mediated by the Ca2+-induced Ca2+ release mechanism, whereas retrograde coupling controls the L-type Ca2+ channel current within the same beat.2
In This Issue, see p 395
Editorial, see p 396
Whole-cell Ca2+ transients during cardiac excitation–contraction (EC) coupling were first seen as aequorin bioluminescence in frog cardiac muscle3 and in canine Purkinje fibers.4 The creation of small-molecule chemical fluorescence indicators, such as fura-2,5 catalyzed an explosion of Ca2+ imaging activity in single cardiac myocytes under normal conditions6 and during Ca2+ overload when propagating waves of elevated Ca2+ were observed.7 With the advent of confocal microscopy and the fast, high- contrast Ca2+ indicator fluo-3, the discovery of Ca2+ sparks,8 which is the result of single-dyad Ca2+ signaling, has revolutionized our understanding of cardiac EC coupling and the spatiotemporal summation of ≈104 elemental Ca2+ sparks results in a global Ca2+ transient that controls cell contraction. From the viewpoint of control theory, such digital behavior is quintessential for enabling high-speed, high-gain amplification with stability in EC coupling.9,10 The process of cardiac EC coupling is therefore a tale of signaling in cellular nanodomains whose regulation remains a focus of intense research worldwide.
However, fluorescent Ca2+ imaging techniques with diffusible indicators (eg, see ref. 11) do not reveal either the underlying Ca2+ signaling process or the behavior of the RyRs at the level of the dyad without considerable computational difficulty and uncertainties (eg, see ref. 11). Furthermore, Ca2+ sparks are not easily detected once whole-cell release starts and ≈104 dyads are activated near synchronously. The measurement of Ca2+ spikes12 provided a way to overcome this problem, but it requires the inclusion of high concentrations of an exogenous, slow Ca2+ buffer (eg, EGTA), which will disturb the Ca2+-induced Ca2+ release process as well as other signaling processes that depend on Ca2+. A new method to probe dyadic space dynamics directly could therefore give insightful information into how RyRs are regulated, as well as check the veracity of new models for RyR array operation and release termination.13,14
The dyadic space is ≈10–3 fl (1 al),15 a volume that would preclude obtaining useful dyadic Ca2+ signals using conventional fluorescent probes. Here, we describe a new approach to this problem using a targeted high-sensitivity biosensor construct for Ca2+-GCaMP6f.16 By attaching GCaMP6f to the dyad junctional proteins junctin or triadin,17 we show that high signal-to-noise records of Ca2+ changes in the dyad, which we call Ca2+ nanosparks, can be obtained without complications from indicator diffusion. Ca2+ nanosparks are ≈50× smaller in volume than Ca2+ sparks and can display substructure and gradation at the single dyad level. We also present exemplar data from adult rat cardiac myocytes showing evoked Ca2+ signals at single dyads, which should afford a new way of investigating spatiotemporal synchrony (or dyssynchrony) in EC coupling in heath and disease.
Adult Rat Cardiac Myocytes Isolation, Culture, and Adenovirus Infection
Animals were treated in compliance with the Guide for the Care and Use of Laboratory animals published by the US National Institutes of Health (Publication No.85-23, revised 1996) and approved by the Institutional Animal Care and Use Committee of Peking University (accredited by Association for Assessment and Accreditation of Laboratory Animal Care international). Single ventricular myocytes were enzymatically isolated from the hearts of adult male Sprague–Dawley rats (200–250 g), as described previously.18 Freshly isolated cardiac myocytes were plated on culture dishes coated with laminin (Sigma) for 1 hour and then the attached cells were cultured in M199 medium (Sigma) added with (in mmol/L) 5 creatine, 2 L-carnitine, 5 taurine, 25 HEPES (all from Sigma), and insulin-transferrin- selenium-X (Gibco). Cardiac myocytes were then infected with adenovirus carrying the target gene at an multiplicity of infection of 20 and experiments were performed after 48 hours in culture.
Recombinant Adenovirus Production
The GCaMP6f gene was amplified from pGP-CMV-GCaMP6f (Addgene, plasmid 40755) and inserted into pEGFP-C1 (Clontech). The rat triadin 1 and junctin genes were cloned and integrated into pEGFP-C1-GCaMP6f with GCaMP6f at the N terminus. The fusion genes GCaMP6f-triadin 1 (GCaMP6f-T) and GCaMP6f-junctin (GCaMP6f-J) were then amplified and inserted into pENTR/TEV/D-TOPO vector (Invitrogen). The adenovirus was produced using the Invitrogen adenoviral expression system (Invitrogen).
Cardiac myocytes were permeabilized with 0.5% Triton X-100 and blocked with 10% normal goat serum. The sections were incubated with anti–RyR monoclonal antibody (Sigma, R128) for 2 hours at room temperature, washed with PBS, and then incubated with tetrarhodamine isothiocyanate–conjugated goat anti-rabbit IgG (Santa Cruz) for 1 hour at room temperature. The immunofluorescence staining was visualized using Zeiss LSM710 confocal microscope at 488 nm (GCaMP6f-T/J) and 543 nm (tetrarhodamine isothiocyanate) excitation and 490 to 550 nm and >560 nm emission, respectively.
Unmixing of Di-4 AN (F) PPTEA and GCaMP6f-J Signals
Di-4 AN (F) PPTEA (Di-4, 0.5 μg/mL, 5 minutes), a lipophilic dye,19 was used to stain the surface membrane and t-tubules in cultured cardiac myocytes expressing GCaMP6f-J. With excitation at 488 nm, spectral images covering 486 to 719 nm were collected with an inverted confocal microscope (Zeiss NLO710) equipped with a 34-Channel QUASAR Detection Unit. For unmixing the Di-4 and GCaMP6f-J components, the reference spectra were obtained in Di-4–stained, uninfected cells and GCaMP6f-J–expressing, Di-4–unstained cells, respectively.
In Situ Calibration of the Biosensor
Adult rat cardiac myocytes expressing GCaMP6f-T/J were permeabilized with 50 μg/mL saponin and then treated with internal solution containing (in mmol/L) 10 KOH, 100 aspartic acid, 20 KCl, 0.81 MgCl2, 3 Mg-ATP, 0.5 EGTA, 5 phosphocreatine ditris, 10 phosphocreatine-Na, 5 creatine phosphokinase, 10 glutathione, 8% dextran, and 20 HEPES and different concentrations of Ca2+ (from 3×10–9 to 2×10–5 mol/L at pH 7.2). The fluorescence of GCaMP6f-T/J was recorded with Zeiss LSM710 confocal microscope with 488 nm excitation and 490 to 550 nm emission as was also used for fluo-4 measurement. The relationship of Ca2+ concentration (C) and normalized fluorescence (R) or (F–Fmin)/(Fmax–Fmin), where Fmin (minimum biosensor fluorescence) and Fmax (maximal fluorescence of the biosensor) were obtained at 3×10–9 and 2×10–5 mol/L, respectively, was fitted with the equation R=CnH/(KdnH+CnH), where nH is the Hill coefficient and Kd is the dissociation constant.
Confocal imaging was performed with a Zeiss LSM710 microscope with a 63×, 1.4 NA oil immersion objective, and linescan speed of 1.53 ms/line, the pinhole was set for a nominal 1 μm optical section. For simultaneous measurement of GCaMP6f-T/J and Rhod-2, excitation was at 488 and 543 nm and their fluorescence emission was collected at 490 to 550 and >560 nm, respectively. For chemical Ca2+ indicator loading, cultured cardiac myocytes were incubated with 5 µmol/L rhod-2 AM (Invitrogen) or fluo-4 AM (Invitrogen) for 10 minutes at room temperature.
Image Processing and Data Analysis
Digital images were processed using customer-devised routines written in Interactive Data Language (ITT, New York, NY). To obtain the time course of fluorescence corrected for the biosensor’s turn-off kinetics, deconvolution was performed with Wiener filter in the frequency domain. The degenerate kernel used in the Wiener filter was a single-exponential decay function with time constant of 60 ms reflecting the turn-off rate of the biosensor.
Data are reported as the mean±SE. Student t test and the Mann–Whitney U test for nonparametric distributions were applied, when appropriate, to determine the statistical significance of differences. P<0.05 was considered statistically significant.
Targeting the Ca2+ Biosensor GCaMP6f to Junctional SR
The rationale for developing a sensitive (instead of low affinity)-targeted Ca2+ biosensor is to probe dyad junctional Ca2+ dynamics, which has not been possible with cytoplasmic probes. We reasoned that a genetically encoded Ca2+ probe linked to known junctional proteins (triadin1 and junctin) should concentrate the biosensor into the high Ca2+ nanodomain associated with RyR activity (Figure 1A). This approach should avoid the potential complication arising from tagging RyRs themselves which might alter their gating and/or assembly in the dyad. As to the biosensor of choice, the latest generation of fluorescent protein–based Ca2+ probe GCaMP6f seems to be ideal because of its fast kinetics (the fastest among all currently known GCaMPs), high affinity (relative to junctional calcium), and superb contrast factor.16 In a recent study, it was successfully implemented in measuring single synapse events,16 leading to an important breakthrough in understanding dendritic integration.
In cultured rat ventricular myocytes 48 hours after adenoviral- mediated gene transfer, confocal imaging revealed that GCaMP6f-T and GCaMP6f-J (triadin 1 and junctin constructs, respectively) were enriched at punctate sites forming striated sarcomeric pattern (Figure 1B and 1C). This result suggests that both triadin 1 and junctin can target correctly their N-terminal–fused GCaMP6f to the dyadic clefts. This notion was confirmed by colocalization of GCaMP6f-J and GCaMP6f-T (latter data not shown) with type 2 RyR immunofluorescence (Figure 1C) and the pattern of labeling was similar to that reported in high- resolution studies of RyR distribution.20 The t-tubular structure of the myocytes was well maintained with GCaMP6f-J and Di-4 fluorescent signals (a surface membrane marker) strongly colocalizing (Figure 1D) after spectral unmixing (Online Figure I) again supporting the idea that the construct correctly trafficked to the dyadic junctions.
GCaMPs are circularly permuted variants of enhanced green fluorescent protein coupled to the Ca2+-binding protein calmodulin at the C terminus as the Ca2+ sensor, and a calmodulin-binding M13 peptide at the N terminus.21 Crystal structures of GCaMP2 have shown that, on Ca2+ binding, the calmodulin moiety interacts and interlocks with the M13 peptide, causing a significant structural reorganization in proximity to the chromophore of cpEGFP and increasing its fluorescence, probably by deprotonating the chromophore.22 To assess the performance of the dyad-targeted GCaMP6f-T/J, we measured fluorescence changes in saponin-permeabilized cells when bathing Ca2+ was increased from 3 nmol/L to 20 μmol/L. GCaMP6f-T/J in cardiac dyads displayed a contrast factor of 42.5±11 (n=6 cells) defined as the ratio of Fmax to Fmin, similar to that observed in vitro.16 Nonlinear fitting yielded a Kd of 632 nmol/L and a Hill coefficient (n) of 1.77 (Figure 1E). Interestingly, GCaMP6f-T/J displayed an off rate ≈4-fold faster than that of native GCaMP6f (see below). Thus, we have succeeded in creating a novel Ca2+ biosensor that colocalizes with t-tubules and RyRs at dyadic clefts, offering a new approach to detecting nanodomain Ca2+ changes.
Imaging Dyadic Ca2+ Nanosparks With GCaMP6f-T/J
Spatially discrete, sudden, and transient GCaMP6f-T/J fluorescence increases (Ca2+ nanosparks; see Discussion) arose spontaneously from biosensor-labeled dyads in resting cardiac myocytes (Figure 2A and 2B). Nanosparks rose abruptly, attained a peak of 3.0 (F/F0) in ≈22 ms, and then returned to the baseline with single-exponential kinetics (time constant of the decay of biosensor fluorescence [τdecay]≈63 ms; Figure 2; Table). Individual nanosparks were confined to focal regions of the cell (width≈540 nm; Table), with no evidence of a diffusive fluorescence signal affecting ambient space (or neighboring sites). Given the limited confocal resolution, the size of these regions seems consistent with recent tomographic data on dyadic junctions.23 When compared with Ca2+ sparks reported by rhod-2 (Figure 2C), Ca2+ nanosparks occupied a 50× smaller optical volume but were significantly brighter (Table). Ca2+ nanosparks were completely abolished by SR-Ca2+ depletion with 20 mmol/L caffeine but persisted in the presence of the L-type Ca2+ channel inhibitor nifedipine (10 μmol/L; data not shown). Hence, these GCaMP6f-T/J fluorescent nanosparks reflect the first direct visualization of Ca2+ signals in the dyad junctional space.
The immobility of GCaMP6f-T/J implies that Ca2+ nanospark formation is simply driven by the local Ca2+ turning ON the biosensor, convolved with the biosensor’s OFF kinetics. Specifically, the ON process reflects the rise of local Ca2+, the Ca2+ binding to the calmodulin moiety of the biosensor, and the deprotonating rearrangement around the chromophore before fluorescence emission, whereas the OFF process (dissociation rate constant [KOFF]) reflects the deactivation of the chromophore. After RyR closure, local Ca2+ gradients collapse rapidly24,25 so that the onset of nanospark decline should give a measure of the time taken for complete RyR closure, and the decline rate of nanosparks should be mainly determined by the OFF kinetics of the biosensor. Furthermore, the time course and amplitude (F/F0) of nanosparks should not depend on relative position of the dyad to the confocal plane, providing that contaminating background fluorescence from other sources is negligible (see below).
As shown in the Table, the time to peak of nanosparks was longer than the typical time to peak of a Ca2+ spark.8 The histogram of τdecay exhibited a major Gaussian component that centered around 60 ms, with a small tail reflecting a subpopulation of prolonged release events (Online Figure II). These data suggest that the turn-off rate of the targeted biosensor in situ was ≈17/s, a ≈4-fold improvement compared with its tag-free counterpart.16 Importantly, by simultaneously measuring rhod-2 Ca2+ signals, we found that the presence of the biosensors did not significantly alter the time course or amplitude of spontaneous Ca2+ sparks (Table). This result indicates that biosensor expression exerted negligible buffering effects on the evolution of dyadic Ca2+ that controls EC coupling.
Based on the nanospark formation mechanism described above, the time course measurement can be improved by temporal deconvolution using a kernel reflecting the turn-off kinetics of the biosensor. Using KOFF=17/s to approximate the OFF kinetics of the biosensor, we deconvolved the nanospark and used the resulting trace, (F/F0)d, as a more direct indicator of junctional Ca2+ dynamics (Figure 2B). A representative linescan image of kinetically deconvolved Ca2+ nanospark (F/F0)d and its corresponding spatially averaged line plot are shown in Figure 2B. On average, (F/F0)d reached its peak at ≈12 ms from onset, in reasonable agreement with detailed modeling and release flux calculations for rat with the release flux reaching a peak at ≈5 ms and lasting for ≈20 ms.24,26
RyR Array Operation at Single Dyads
With the precise colocalization of the biosensor and the Ca2+ source, the superior dyad-to-background contrast, and the high sensitivity provided by targeted GCaMP6f-T/J, we investigated RyR array operation at the single-dyad level. Although Ca2+ nanosparks generally appeared as a single homogeneous region, in some cases distinct substructure could be resolved. Figure 2D illustrates such an event with the fluorescence increase occurring in 2 linked regions nearly simultaneously (as far as can be determined from the limited time resolution of the microscope and signal-to-noise ratio). We suggest that such events reflect the activation of separate RyR clusters within the same junctional space, which coactivate in <4 ms. Furthermore, we found that the Ca2+ nanospark amplitude, measured at the origin of Ca2+ release and relatively immune to out-of-focus blurring, displayed a broad distribution (Online Figure IIIA), ranging from 1. 8 to 4.8 (F/F0 at 5 and 95 percentiles) with a mean value of 3.04. Similarly, after temporal deblurring, peak (F/F0)d varied from 4.1 (at 5%) to 14.9 (at 95%; Online Figure IIIB). Taken together, these data provide novel evidence for the possibility of stochastic recruitment of different numbers of RyRs within Ca2+ nanosparks.
We also examined dyads displaying >1 Ca2+ nanospark. Figure 3A and 3B illustrates variability in peak F/F0 and (F/F0)d in consecutive Ca2+ nanosparks. The smallest Ca2+ nanosparks are similar to single-quanta13 and quarky Ca2+ release events as reported previously.27 For pairs of Ca2+ nanosparks at the same dyads, their peak F/F0 ranged from 0.75 to 1.33 (at 5 and 95 percentiles, respectively) with peak (F/F0)d from 0.48 to 2.2 (Figure 3C). Thus, the gradation of amplitude reflects an event-to-event stochastic recruitment of individual or subclustered RyRs.
Imaging Dyadic Activation During EC Coupling
Next, we examined junctional Ca2+ dynamics during normal (ie, electrically evoked) EC coupling. Representative results in Figure 4 illustrate several novel features. First, in cardiac myocytes undergoing electric pacing, the Ca2+ transients occurred in discrete domains that did not merge (unlike other Ca2+ reporter signals). Because of this, timing of activation at individual dyads was clearly resolved in vigorously contracting cells. Second, dyads displaying 2-fold difference in F0, perhaps reflecting in-focus and slightly out-of-focus sites had similar F/F0 (Figure 4B and 4C). This is because of the probe being fixed so that the relative change in fluorescence is unaffected by the sensitivity of the optical detector (in this case being determined by the limited extent of the microscope point spread function). Third, at dyads that were activated late, we observed a slow transient preceding a sudden, rapid upstroke of the dyadic signal. We interpret the first phase to result from cytosolic Ca2+ transient that invades the junctional space, and the second phase to reflect evoked dyadic release in the form of a Ca2+ nanospark. Remarkably, the properties of these late Ca2+ nanosparks were similar to those of early ones, directly showing (for the first time) that inferred SR depletion in adjacent sites during EC coupling28 has minor effects on the proximal SR store on the time scale of normal Ca2+ release. Furthermore, the fully activated dyadic Ca2+ signal was much higher than the global Ca2+ transient alone. The peak (F/F0)d in an evoked nanotransient was 7.8-fold higher than that because of global Ca2+ elevation, providing direct evidence for a much higher cleft than global Ca2+ concentration in EC coupling (as expected from modeling studies).
We have developed a novel Ca2+ imaging method that has enabled the first detection of Ca2+ signals arising in the submicroscopic (nanoscopic) space of the dyadic junction between the surface membrane and junctional SR. By analogy with Ca2+ sparks and to reflect the nanoscopic origin of the signal from our nondiffusible targeted probes, we introduce the neologism Ca2+ nanosparks to describe them. We show that targeted GCaMP6f-T/J provides a new and sensitive way of detecting dyadic activity, that is, when, where, and for how long a dyad is activated, even in cells undergoing vigorous contraction. Individual Ca2+ nanosparks are confined to a volume 50× smaller than conventional Ca2+ sparks as measured with diffusible indicators. Our approach is somewhat analogous to the use of GCaMP6f to count single neuronal synapse events.16 However, we are dealing with a much smaller structure than a synaptic spine and, in this regard, it is remarkable that our nanoscopic signals had such good fidelity.
Imaging Ca2+ nanosparks has been made possible by the availability of a new generation of genetically encoded, ultrasensitive Ca2+ biosensor (GCaMP6f) and by the choice of proper targeting strategy. The fast kinetics, the high affinity, the high contrast factor of GCaMP6f collectively confers high sensitivity, whereas the fusion of biosensor to known dyadic proteins assures precise colocalization to the dyadic Ca2+ nanodomain. That the Ca2+ biosensor is immobile carries major advantages: not only does it permit a simple deconvolution to correct for the biosensor’s turn-off kinetics, but also it obviates spatial blurring because of out-of-focus sampling that has confounded Ca2+ spark measurements. Serendipitously, we also found that the fusion of triadin/junctin to the C terminus of the biosensor also accelerated the turn-off rate of the biosensor, perhaps by facilitating the undocking of the M13–calmodulin complex. This suggests that further C-terminal modification might present a new strategy to further improve the kinetic properties of GCaMPs.
Although being highly sensitive, this new nanodomain Ca2+ measurement method also has a wide dynamic range. For example, it allows resolution of quarky Ca2+ release, spontaneous Ca2+ nanosparks, and even evoked Ca2+ nanosparks during the cell-wide Ca2+ transient. Analysis of the variability of Ca2+ nanospark amplitude provided new evidence that Ca2+ release is not necessarily all or none at the level of single dyads. Even at the same dyads, amplitudes of consecutive Ca2+ nanospark can vary ≈ 2-fold. These results imply stochastic recruitment of different numbers or subclusters of RyRs in a single functional dyadic junction. Our observation of some substructure within occasional Ca2+ nanosparks further supports the idea of possible subcluster activation within groups of RyRs29 and previous reports of graded Ca2+ spark amplitude.13,14 However, we have not yet observed temporally separable Ca2+ release events behavior within detectable subregions of single junctions, suggesting that RyR subclusters are spatiotemporally (and functionally) coupled within the ≈2 ms/0.3 μm resolution of the microscope. From this, we suggest that intrajunctional activation propagation delays should be a minor component of the time to peak of the Ca2+ spark, consistent with Monte-Carlo simulations of the effect of RyR (re)organization on release time course within a single junctional region.30
By tracking individual dyad activation within the whole-cell contraction, the current work represents a substantial improvement over the previous Ca2+ spike method using heavy Ca2+ buffering.12 It is especially important to note that the local biosensor did not disrupt normal EC coupling (as evidenced by a lack of effect on Ca2+ sparks). Therefore, although there may be some limited junctional buffering effect because of the sensor, cell-wide Ca2+ signals and Ca2+-dependent regulation should be minimally perturbed.
Despite the biosensor’s moderate Kd and the high Ca2+ concentration expected at the active dyads, we detected no sign of GCaMP6f-T/J saturation during a nanospark: the amplitude of nanosparks is smaller than that of nanotransients during action potential-elicited EC coupling and both were smaller than maximal GCaMP6f-T/J that could be obtained in steady-state calibrations. A possible explanation is that the relatively slow kinetics of the biosensor do not allow the biosensor to equilibrate during the saturating but short-lived dyadic cleft Ca2+ transient. Thus, a high-affinity, slowly responding probe may present relatively low-affinity behavior during highly dynamic Ca2+ signals. Furthermore, nonlinear interplay among multiple sensor Ca2+-binding sites and temporal disparity between Ca2+-binding and biosensor fluorescence may also contribute to this peculiar nonequilibrium property of the biosensor. These possibilities warrant future investigation and tuning of biosensor properties but are outside the scope of this article.
As a general problem that applies to all conventional fluorescent Ca2+ indicators, Ca2+ spark measurement with diffusible indicators grossly underestimates the local peak Ca2+ levels and reports a highly distorted spatial profile of local Ca2+ gradients.31 Although the targeted sensors GCaMP6f-T/J can reveal Ca2+ release timing in the smallest domains (eg, a single dyad), we have not yet achieved the ambitious goal of directly reporting local Ca2+ levels within the junctional space, for reasons discussed above. Nevertheless, with the highly sensitive biosensor expressed at levels that did not significantly affect junction function, we obtained a Ca2+ signal (after temporal deblurring) that should provide an accurate measure of the local Ca2+ release duration and semiquantitative measurement of local Ca2+ fluxes.
Because heart failure and cardiac arrhythmias are often associated with dyssynchronous Ca2+ release, the ability of the probe to report release timing with good fidelity should facilitate investigation of the synchrony (or dyssynchrony) among release sites in health and diseases.32,33 Our targeted Ca2+ biosensor approach may be extended to whole tissues and living organisms by transfection or transgenic techniques, and the lack of apparent effect of the biosensor on Ca2+ signaling suggests that it should not prevent normal Ca2+-dependent signaling processes (eg, nuclear and cytoplasmic calmodulin signaling).
In summary, we have shown that high-quality intrajunctional Ca2+ release signals can be recorded as Ca2+ nanosparks with a novel targeted ultrasensitive biosensor. These Ca2+ nanosparks reveal junctional Ca2+ signaling in a way that is not possible with conventional diffusible Ca2+ indicators. Furthermore, this approach should be generally applicable to probing Ca2+ nanodomains (eg, clusters of inositol trisphosphate receptor Ca2+ release channels) in many cell types and even in living animals by transfection or transgenic approaches.
We thank Dr Looger from the Howard Hughes Medical Institute for the generous gift of GCaMP6 vectors and Dr Leslie M. Loew from University of Connecticut Health Center for Di-4 AN(F)PPTEA.
Sources of Funding
This work was supported by the National Key Basic Research Program of China (2011CB809100 and 2013CB531200), the National Science Foundation of China (31221002, 31130067, 31123004, 81370203, and 30900264), and the Royal Society UK.
In October 2013, the average time from submission to first decision for all original research papers submitted to Circulation Research was 12.81 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.114.302938/-/DC1.
- Nonstandard Abbreviations and Acronyms
- EC coupling
- excitation–contraction coupling
- normalized fluorescence where F0 refers to the resting fluorescence
- F/F0 deconvolved with a single-exponential kernel reflecting turn-off rate of the biosensor
- GCaMP6f fused to triadin 1
- GCaMP6f fused to junctin
- ryanodine receptor
- sarcoplasmic reticulum
- Received October 29, 2013.
- Revision received November 19, 2013.
- Accepted November 20, 2013.
- © 2013 American Heart Association, Inc.
- Grantham CJ,
- Cannell MB
- Wier WG
- Grynkiewicz G,
- Poenie M,
- Tsien RY
- Wier WG,
- Cannell MB,
- Berlin JR,
- Marban E,
- Lederer WJ
- Cheng H,
- Lederer WJ,
- Cannell MB
- Cheng H,
- Lederer WJ
- Wang SQ,
- Wei C,
- Zhao G,
- Brochet DX,
- Shen J,
- Song LS,
- Wang W,
- Yang D,
- Cheng H
- Wang SQ,
- Stern MD,
- Ríos E,
- Cheng H
- Zhang L,
- Kelley J,
- Schmeisser G,
- Kobayashi YM,
- Jones LR
- Huang X,
- Sun L,
- Ji S,
- Zhao T,
- Zhang W,
- Xu J,
- Zhang J,
- Wang Y,
- Wang X,
- Franzini-Armstrong C,
- Zheng M,
- Cheng H
- Yan P,
- Acker CD,
- Zhou WL,
- Lee P,
- Bollensdorff C,
- Negrean A,
- Lotti J,
- Sacconi L,
- Antic SD,
- Kohl P,
- Mansvelder HD,
- Pavone FS,
- Loew LM
- Soeller C,
- Crossman D,
- Gilbert R,
- Cannell MB
- Hayashi T,
- Martone ME,
- Yu Z,
- Thor A,
- Doi M,
- Holst MJ,
- Ellisman MH,
- Hoshijima M
- Brochet DX,
- Xie W,
- Yang D,
- Cheng H,
- Lederer WJ
- Picht E,
- Zima AV,
- Shannon TR,
- Duncan AM,
- Blatter LA,
- Bers DM
- Baddeley D,
- Jayasinghe ID,
- Lam L,
- Rossberger S,
- Cannell MB,
- Soeller C
- Litwin SE,
- Zhang D,
- Bridge JH
Novelty and Significance
What Is Known?
In cardiac dyads, junctional Ca2+ directly controls the gating of the ryanodine receptors (RyRs) and is itself dominated by RyR-mediated Ca2+ release from the sarcoplasmic reticulum.
The stochastic timing of individual dyad Ca2+ release events that produce Ca2+ sparks governs the rising phase of the whole-cell Ca2+ transient.
Conventional diffusible fluorescent Ca2+ indicators are unable to report Ca2+ in the dyadic junction, making it impossible to resolve individual dyad events during a fully activated Ca2+ transient.
What New Information Does This Article Contribute?
A targeted, nondiffusible Ca2+ probe was developed by fusing the Ca2+ sensor GCaMP6f to the N terminus of triadin 1 and junctin, GCaMP6f-T/J, which resulted in its colocalization with RyRs.
GCaMP6f-T/J reported Ca2+ nanosparks, which are Ca2+ signals associated with dyadic activation. Ca2+ nanosparks were spatially 50× smaller (by volume) than conventional Ca2+ sparks.
Ca2+ nanosparks report when, where, and for how long an individual dyad is activated, even during the whole-cell Ca2+ transient and can reveal coactivation of subclusters of RyRs in the junction.
Ca2+ signaling is the result of nanoscopic Ca2+ kinetics in, for example, synapses and cardiac dyads. Imaging with diffusible Ca2+ indicators has, to date, visualized microscopic events such as Ca2+ sparks and Ca2+ puffs, which reflect the diffusion of Ca2+ from the sources but does not resolve the underlying nanoscopic Ca2+ kinetics. By fusing GCaMP6f (a new fast calmodulin-based fluorescent protein) to the N terminus of triadin 1 or junctin, which are known to traffic to dyads, we developed a nondiffusible Ca2+ probe, GCaMP6f-T/J. This probe colocalizes with RyRs and displays 4× faster off-kinetics than native GCaMP6f. This approach allowed the first detection of Ca2+ nanosparks at individual dyads, which are 50× smaller (by volume) than conventional Ca2+ sparks. Ca2+ nanosparks report when, where, and for how long a dyad is activated, even in cells undergoing vigorous contraction. In addition, we showed coactivation of subclusters of RyRs as well as all-or-none behavior in a single junctional region. This imaging strategy should be generally applicable to probing Ca2+ nanodomains in many cell types and even in living animals by transfection or transgenic approaches.