Rapid ligand-regulated gating kinetics of single inositol 1,4,5-trisphosphate receptor Ca²⁺ release channels

Abstract

The ubiquitous inositol 1,4,5-trisphosphate receptor (InsP₃R) intracellular Ca²⁺ release channel is engaged by thousands of plasma membrane receptors to generate Ca²⁺ signals in all cells. Understanding how complex Ca²⁺ signals are generated has been hindered by a lack of information on the kinetic responses of the channel to its primary ligands, InsP₃ and Ca²⁺, which activate and inhibit channel gating. Here, we describe the kinetic responses of single InsP₃R channels in native endoplasmic reticulum membrane to rapid ligand concentration changes with millisecond resolution, using a new patch-clamp configuration. The kinetics of channel activation and deactivation showed novel Ca²⁺ regulation and unexpected ligand cooperativity. The kinetics of Ca²⁺-mediated channel inhibition showed the single-channel bases for fundamental Ca²⁺ release events and Ca²⁺ release refractory periods. These results provide new insights into the channel regulatory mechanisms that contribute to complex spatial and temporal features of intracellular Ca²⁺ signals.

Introduction

The ubiquitous endoplasmic reticulum (ER)-localized inositol 1,4,5-trisphosphate receptor (InsP₃R) Ca²⁺ channel has a central role in regulating numerous cell physiological processes by modulating cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_i) through release of Ca²⁺ from the ER lumen in response to production of cytoplasmic InsP₃ caused by extracellular stimuli (Taylor & Richardson, 1991; Berridge, 1993). InsP₃R channels respond rapidly (on the order of milliseconds; Marchant & Taylor, 1998) to dynamic changes in the concentrations of their primary ligands—cytoplasmic Ca²⁺ and InsP₃ (Foskett et al, 2007). However, kinetic information about single InsP₃R channel responses to ligand concentration changes is lacking. Here, we record, for the first time to our knowledge and with millisecond resolution, the responses of single InsP₃R channels in native ER membranes to controlled changes in [Ca²⁺]_i and [InsP₃] using rapid perfusion techniques (Hinkle et al, 2003) applied to cytoplasmic-side-out (cyto-out) nuclear membrane patches (Mak et al, 2005). The kinetics of activation and deactivation by [InsP₃] and [Ca²⁺]_i changes have shown novel high-affinity inhibitory Ca²⁺ binding and unexpected cooperativity between InsP₃ and Ca²⁺. Such ligand regulation has not previously been accounted for in most quantitative models of InsP₃-mediated Ca²⁺ signalling (Swillens et al, 1994, 1998; Tang et al, 1996; Falcke et al, 2000; Fraiman & Dawson, 2004; Sneyd & Falcke, 2005). The kinetics of Ca²⁺-mediated inhibition have now shown the single-channel bases for fundamental Ca²⁺ release events (Berridge, 1997) and Ca²⁺ release refractory periods (Fraiman et al, 2006). This study provides new insights into the channel regulatory mechanisms (Foskett et al, 2007) that contribute to complex spatial and temporal features of intracellular Ca²⁺ signals (Putney & Bird, 1993).

Results And Discussion

Channel activation and deactivation kinetics

Single-channel gating of homo-tetrameric InsP₃R channels in native ER membranes was studied by applying patch-clamp techniques to outer membranes of nuclei isolated from Sf9 cells (Ionescu et al, 2006). In the nucleus-attached configuration, with the cytoplasmic aspect of the channel containing the binding sites for InsP₃ and Ca²⁺ facing the pipette solution (Fig 1A), InsP₃R channel activities were observed in the current traces (Fig 1D) when the pipette solution contained activating ligand concentrations—2 μM Ca²⁺ and 10 μM InsP₃ (Ionescu et al, 2006). In excised nuclear membrane patches in the cyto-out configuration (Fig 1B), achieved here for the first time to our knowledge, InsP₃R channels were not observed when the local perfusion solution lacked InsP₃, but their presence was verified by observation of channel activities (Fig 1E) when the patches were perfused with a solution containing 10 μM InsP₃ and 2 μM Ca²⁺ (Fig 1C).

Figure 1.

Nuclear patch-clamp studies of single InsP₃R channel responses to ligand concentration changes using rapid perfusion techniques applied to cyto-out membrane patches. (A,B) Schematic diagrams illustrating orientation of InsP₃R channels in nucleus-attached and excised cyto-out configurations, respectively. In the nucleus-attached configuration, the cytoplasmic side of InsP₃R channels (black) faces the pipette solution. When the cyto-out configuration is obtained, the cytoplasmic side of different InsP₃R channels (purple) faces the bath solution. (C) Micrograph showing experimental set-up for rapid perfusion. B₁ and B₂ are square glass barrels used to deliver different bath solutions, to which the InsP₃R channels in the cyto-out nuclear membrane patch at the tip of the micropipette P were alternately exposed. (D) Single-channel current trace showing InsP₃R channel activity with 10 μM InsP₃ and 2 μM [Ca²⁺]_i in the pipette solution in the nuclear-attached configuration. Closed-channel baseline current levels are indicated by arrows in all current traces. (E–L) Single-channel current traces from cyto-out patches showing the responses of InsP₃R channels to abrupt changes in ligand concentrations, indicated by coloured bars above the traces. Ligand concentrations were changed by switching the perfusing bath solutions, which also caused a change in the level of the closed-channel baseline current because of the different [KCl] in the perfusing solutions (supplementary information online). This allowed the time course of the perfusing solution switch to be monitored. Response latencies are denoted by grey bars below the traces. Cyto-out, cytoplasmic-side-out; InsP₃R, inositol 1,4,5-trisphosphate receptor.

We first explored the kinetic responses of the channel to rapid changes in [InsP₃]. The mean activation latency in response to a step change of [InsP₃] from 0 to 10 μM in the constant presence of optimal (2 μM) [Ca²⁺]_i was 500±60 ms (Fig 1E; number of solution switches (n)=458; number of cyto-out patches used (N)=14). Because this and all other activation and deactivation latencies showed non-gaussian distributions (Fig 2), non-parametric statistics were used in latency analyses (supplementary information online). A shorter (P=0.002) but still large latency (350±40 ms, n=168, N=4) was observed in response to a step change of [InsP₃] from 0 to 100 μM, indicating that InsP₃ binding and subsequent conformation changes required for channel opening both contribute significantly to the total activation latency. The mean deactivation latency in response to rapidly dropping [InsP₃] from 10 to 0 μM in the constant presence of optimal [Ca²⁺]_i (690±50 ms; Fig 1F, n=260, N=14) was similar (P=0.06) to that in response to [InsP₃] dropping from 100 μM (700±70 ms, n=123, N=4), as expected. The latency of channel activation by InsP₃ in the presence of 2 μM cytoplasmic Ca²⁺ (Ca²⁺_i) is long compared with those observed for other ligand-activated channels in excitable cells (on the order of microseconds; Maconochie et al, 1994). It is also long compared with the latency of Ca²⁺ release resulting from the [InsP₃] jump owing to photolysis of caged InsP₃ in cells (Khodakhah & Ogden, 1995; Parker et al, 1996), or rapid perfusion of permeabilized cells or microsomes (Dufour et al, 1997). This is not surprising because in such experiments, the latency of Ca²⁺ release through a population of InsP₃R channels was observed. In a system with a large number of InsP₃R channels, the latency to first Ca²⁺ release will be significantly shorter than the activation latency for single InsP₃R channels in response to the same ligand concentration change (supplementary information online). Furthermore, the non-gaussian distribution of single-channel InsP₃ activation latency (Fig 2A) means that most of the InsP₃ activation latencies observed were shorter than 60 ms even though the mean latency was 500 ms (discussed further in the supplementary information online).

Figure 2.

Latency histograms of InsP₃R channels responding to changes in [InsP₃] and [Ca²⁺]_i. [InsP₃] and [Ca²⁺]_i are denoted by the coloured bars above each histogram. For each type of ligand concentration change investigated, both the linearly binned histogram (left) covering a limited latency range and the logarithmically binned histogram (right) showing the complete range of latencies over all times are shown. In each logarithmically binned histogram, each blue curve is the theoretical fit generated using one linear Markov chain scheme with the tabulated number of functionally indistinguishable conformations (supplementary information online). The black curve is the sum of the theoretical fits if several linear Markov chains were required. In each linearly binned histogram, the blue continuous curve is evaluated using the fastest linear Markov chain scheme and therefore is equivalent to the blue continuous curve in the corresponding logarithmically binned histogram. The range of the x-axis was selected so that the histogram includes at least 95% of the events described by the fastest linear Markov chain scheme. InsP₃R, inositol 1,4,5-trisphosphate receptor.

We next explored the kinetics of Ca²⁺ regulation by applying rapid [Ca²⁺]_i jumps in the continuous presence of a saturating concentration (10 μM) of InsP₃. Channel activation by a jump from low (<10 nM) to optimal (2 μM) [Ca²⁺]_i had a mean latency of 40±3 ms (Fig 1G, n=532, N=11), an order of magnitude faster than the response to InsP₃ activation (P≪0.001). The latency for [Ca²⁺]_i jumping to 300 μM was shorter (P≪0.001) but detectable (24±6 ms, n=94, N=3), indicating that Ca²⁺ binding and subsequent conformation changes contribute appreciably to the Ca²⁺ activation latency. Channel deactivation when Ca²⁺ was returned to <10 nM had a mean latency of 160±20 ms (Fig 1H, n=376, N=11), also significantly faster than channel deactivation owing to removal of InsP₃ (P≪0.001). These results show that InsP₃-bound channels respond rapidly to changes in [Ca²⁺]_i, a property ideally suited for Ca²⁺-induced Ca²⁺ release (CICR) to couple InsP₃R channels in Ca²⁺ signalling (Berridge, 1997).

Remarkably, the mean activation latency in response to simultaneous jumps of [InsP₃] (from 0 to 10 μM) and [Ca²⁺]_i (from <10 nM to 2 μM) was 59±4 ms (Fig 1I, n=638, N=7)—much shorter than that for InsP₃ activation in the continuous presence of optimal [Ca²⁺]_i (P≪0.001) and comparable with the latency for Ca²⁺ activation in the continuous presence of saturating [InsP₃]. If InsP₃ binding was required before Ca²⁺ can bind, then the activation latency for a simultaneous jump of the two ligands should be comparable with that for InsP₃ in the presence of optimal [Ca²⁺]_i. These results indicate that there is no requirement for strict sequential InsP₃ and Ca²⁺ binding to activate the channel, invalidating the assumption used in many models that Ca²⁺ can only bind to InsP₃-bound channels (Tang et al, 1996; Marchant & Taylor, 1997; Sneyd & Falcke, 2005). The activation latencies observed here are also incompatible with the independent binding of InsP₃ and Ca²⁺ to activate the channel that is assumed in many other models (Swillens et al, 1994, 1998; Tang et al, 1996; Hirose et al, 1998; Fraiman & Dawson, 2004; Sneyd & Falcke, 2005), because independence predicts that the activation latency for simultaneous [InsP₃] and [Ca²⁺]_i jumps should be longer than the activation latencies for individual [InsP₃] or [Ca²⁺]_i jumps. Instead, the kinetic data are consistent with an unexpected negative cooperative interference, in which high-affinity Ca²⁺ binding to a channel not bound to InsP₃ retards subsequent activation of the channel by InsP₃.

The mean deactivation latency of 69±5 ms for simultaneous removal of InsP₃ (from 10 μM to 0) and Ca²⁺ (from 2 μM to <10 nM) (Fig 1J, n=458, N=7) was significantly shorter than the deactivation latencies for removal of either InsP₃ (P≪0.001) or Ca²⁺ (P=0.002) alone. This result implies that unbinding of InsP₃ and Ca²⁺ from the InsP₃R is neither sequential nor independent.

Channel activation and deactivation latency distributions

The distributions of InsP₃R channel activation and deactivation latencies showed distinct deficits of short latencies (linearly binned histograms in Fig 2A–F) when compared with a single exponential distribution, similar to those observed during activation of voltage-gated channels (Patlak & Horn, 1982; Hoshi et al, 1994). Such activation latency deficits can be well accounted for by a linear Markov chain scheme connecting C₁—the closed inactive conformation adopted by the channel before the ligand concentration change—to O—the first open conformation adopted by the channel after the jump in ligand concentration—through one or more intervening closed conformations: C₁↔C₂↔…↔C_m↔O, where m (⩾2) is the total number of functionally indistinguishable closed conformations in the chain. The deficits of short deactivation latencies can be similarly accounted for by linear Markov chain schemes connecting the initial active open channel conformation to the first inactive closed conformation through several intervening, functionally indistinguishable active conformations (Fig 2; supplementary information online).

The logarithmically binned histogram (Sigworth & Sine, 1987) showing the latency distribution over all timescales of a channel activation or deactivation process described by a linear Markov chain scheme has a peak shape with a single maximum at the dominant time constant. This peak becomes narrower as the number of functionally indistinguishable conformations (m) present in the linear chain increases. However, logarithmically binned histograms of most InsP₃R channel activation and deactivation latencies show multiple maxima or broad shapes that cannot be described with a single linear Markov chain (Fig 2A,B,D–F). These can be satisfactorily described by branched Markov chain schemes such as the one shown below: in which the channel responding to the ligand concentration change passes through M conformation transition pathways, each corresponding to a linear Markov chain that generates a single peak in the logarithmically binned latency histogram (Fig 2; supplementary information online).

Kinetic schemes accounting for channel latencies

The observed response latencies for single InsP₃R channels point to complex and seemingly paradoxical interactions between InsP₃ and Ca²⁺ activation of the channel. The long activation latencies in response to jumps in [InsP₃] in the constant presence of optimal [Ca²⁺]_i compared with the activation latencies in response to simultaneous jumps in [InsP₃] and [Ca²⁺]_i indicate that Ca²⁺ can bind with high affinity to the channel in the absence of InsP₃. However, unlike the high-affinity Ca²⁺ binding that activates the channel along with InsP₃ binding that has been established from steady-state studies (Foskett et al, 2007), this Ca²⁺ binding interferes negatively with the activation of the channel by InsP₃. It does this by predisposing the channel to activate through slower conformation transition pathways or by reducing the rate of channel activation through available conformation transition pathways in the branched Markov chain scheme (cf. Fig 2A,E).

Conversely, the significantly more rapid deactivation of the channel in response to simultaneous drops in [InsP₃] and [Ca²⁺]_i compared with channel deactivation caused by drops in either [InsP₃] or [Ca²⁺]_i alone indicates that the InsP₃ and the high-affinity Ca²⁺ sites have positive cooperative coupling in an activated channel. Occupation of the InsP₃ or Ca²⁺ sites causes the channel to deactivate through slower conformation transition pathways in the branched Markov chain (cf. Fig 2B,D,F), thereby stabilizing the channel in active conformations.

Fig 3A depicts a kinetic scheme that can account for these observations qualitatively, assuming that the high-affinity inhibitory Ca²⁺ sites are distinct from the high-affinity activating Ca²⁺ sites. Ca²⁺ binding to these inhibitory Ca²⁺ sites precludes InsP₃ binding to the channel, and vice versa. Pre-exposure to optimal [Ca²⁺]_i in the absence of InsP₃ retards InsP₃ activation of the channel because InsP₃ cannot bind to the channel until Ca²⁺ dissociates from the inhibitory sites, even after the jump in [InsP₃]. In this scheme, the activating Ca²⁺ sites and the InsP₃ sites show a positive allosteric interaction in an active channel. This is a result of the different channel deactivation rates by various pathways in the branched Markov chain generated by multiple InsP₃ and Ca²⁺ binding to the tetrameric channels, shown partly in Fig 3A.

Figure 3.

Kinetic schemes accounting for the observed InsP₃R channel response latencies. (A) A qualitative model stipulating distinct high-affinity inhibitory and activating Ca²⁺ sites in each InsP₃R channel. C and O represent closed inactive and active channel conformations. Channel conformations with different ligand binding states are explicitly shown in the scheme by superscript and subscript numbers in different colours: number of occupied high-affinity Ca²⁺ inhibitory sites in red, occupied high-affinity activating Ca²⁺ sites in black and occupied InsP₃ sites in green (as indicated by the key at the top right). Ligand binding and dissociation are indicated by arrows: InsP₃ binding in green, Ca²⁺ binding to inhibitory sites in red and Ca²⁺ binding to activating sites in black. Yellow arrows represent conformation changes. Because the channel is a tetrameric unit, more than one Ca²⁺ can bind to either the activating or inhibitory sites. Those ligand-binding states are not shown in the scheme for clarity. Multiple InsP₃ binding to the channel (dashed lines) is shown to demonstrate that the scheme can accommodate branched Markov chains. (B) A qualitative allosteric model stipulating only one kind of high-affinity Ca²⁺ site in an InsP₃R channel. Occupation of the InsP₃ binding sites causes the channel to adopt the C and D conformations instead of the A and B conformations, so that Ca²⁺ binding to the high-affinity Ca²⁺ site becomes activating instead of inhibitory. Unlike (A), only conformation changes are shown in this kinetic scheme. InsP₃ and Ca²⁺ binding to the channel, and the ligand-binding state of the channel conformations are omitted for clarity.

Alternatively, a model stipulating that only one high-affinity Ca²⁺ site is present in each InsP₃R monomer is also qualitatively consistent with all the kinetic responses observed here. The model, a simplified version of a quantitative model proposed to account for steady-state InsP₃ and Ca²⁺_i regulation of channel activity (Mak et al, 2003), assumes only basic allosteric interactions between the two ligands that arise solely from preferential binding of InsP₃ and Ca²⁺ to various channel conformations. In its simplest form (Fig 3B), the model hypothesizes that the channel can adopt four different conformations. In the absence of InsP₃, it is energetically favourable for the channel to adopt two conformations, one active (A) and one inactive (B), with Ca²⁺ binding to the high-affinity sites stabilizing the inactive B conformation. Thus, in the absence of InsP₃, the high-affinity Ca²⁺ sites are actually inhibitory. InsP₃ binding to the channel energetically stabilizes the other two conformations, also one active (C) and one inactive (D). Ca²⁺ binding to the high-affinity sites stabilizes the active C conformation relative to the inactive D conformation. Consequently, the high-affinity Ca²⁺ sites are activating when the channel is InsP₃-bound. Thus, because of the different relative affinities of the high-affinity Ca²⁺ sites in the four channel conformations, InsP₃ binding to the channel changes the nature of the high-affinity Ca²⁺ sites from inhibitory to activating. This transformative effect of InsP₃ on the nature of the high-affinity Ca²⁺ binding sites can account for the seemingly paradoxical channel response latencies measured here. Because Ca²⁺ binding to the high-affinity sites in the absence of InsP₃ stabilizes the inactive B conformation, which has low affinity for InsP₃, pre-exposure of the channel to 2 μM Ca²⁺_i slows down binding of InsP₃ to the channel and consequent activation kinetics compared with activation by simultaneous [Ca²⁺]_i and [InsP₃] jumps. Conversely, occupation of the high-affinity Ca²⁺ and InsP₃ sites stabilizes an active channel in the C conformation with high affinity for InsP₃ and Ca²⁺. This accounts for the shorter mean latency for deactivation by simultaneous removal of InsP₃ and Ca²⁺_i compared with that owing to removal of either Ca²⁺_i or InsP₃ alone. The model hypothesizes that the channel can adopt the different conformations in any ligand-binding state, and InsP₃ binding and Ca²⁺ binding regulate channel activity by affecting the probability of the channel being in various conformations. Thus, active and inactive conformations are connected by many branched pathways through various ligand binding and conformation changes.

Channel activity level after activation

It has been speculated that spontaneous changes in channel activity after the initial activation might account for biphasic Ca²⁺ release kinetics and graded release of Ca²⁺ from intracellular stores in response to incremental levels of stimulation (Muallem et al, 1989; Meyer & Stryer, 1990; Marchant & Taylor, 1998). However, no change in single-channel P_o was observed after the InsP₃R was activated by jumps in [InsP₃], [Ca²⁺]_i or both simultaneously (Fig 4). Thus, behaviours akin to ‘adaptation' observed in the ryanodine receptor (Fill et al, 2000) are unlikely to account for InsP₃R-mediated graded release of Ca²⁺.

Figure 4.

Steady level of InsP₃R channel activity observed after channel activation by abrupt jumps in ligand concentration. (A) Current traces of InsP₃R channels after activation by step change in [InsP₃] (top), [Ca²⁺]_i (middle) and both simultaneously (bottom). (B) Averaged time course of P_o (in 0.1 s bins) of one InsP₃R channel after step change of [Ca²⁺]_i at t=0 in one membrane patch. A total of 72 [Ca²⁺]_i jumps were averaged. InsP₃R, inositol 1,4,5-trisphosphate receptor.

Channel inhibition and recovery kinetics

The steady-state gating activity of the InsP₃R is biphasically regulated by [Ca²⁺]_i, indicating that the channel also has low-affinity inhibitory Ca²⁺ binding sites (Foskett et al, 2007). Ca²⁺ inhibition can provide a negative feedback mechanism to either terminate or prevent Ca²⁺ release as the local [Ca²⁺]_i is raised by InsP₃R-mediated Ca²⁺ release. This can have significant roles in generating Ca²⁺ spikes and oscillations, and in creating highly localized Ca²⁺ signals in subcellular compartments by preventing runaway Ca²⁺ release due to CICR. Switching [Ca²⁺]_i from 2 μM (optimal) to 300 μM (inhibitory; Ionescu et al, 2006) in the presence of 10 μM InsP₃ inhibited channel gating (Fig 1K). This shows that InsP₃ binding does not render the low-affinity inhibitory Ca²⁺ sites inaccessible for Ca²⁺ binding, disproving previous claims (Adkins & Taylor, 1999). The mean latency for Ca²⁺ inhibition (290±40 ms, n=186, N=7) was substantially longer than that for Ca²⁺ activation (P≪0.001).

The mean latency for channel recovery from Ca²⁺ inhibition was 2.4±0.2 s (Fig 1L, n=197, N=7). This remarkably long latency defines a refractory period between successive Ca²⁺ release events mediated by the same InsP₃R channel or channel cluster. A refractory period has been invoked to account for mutual annihilation of colliding intracellular Ca²⁺ waves (Lechleiter et al, 1991) and the periodicity of Ca²⁺ oscillations (Parker & Ivorra, 1990), but its molecular basis has remained unknown. The observed latency is in quantitative agreement with the mean refractory periods (∼2.5 s) between two successive spontaneous InsP₃R-mediated elementary Ca²⁺ release events (puffs) observed in the presence of saturating [InsP₃] (Fraiman et al, 2006). This result indicates that the latency for channel recovery from Ca²⁺ inhibition, rather than other mechanisms such as lingering high [Ca²⁺]_i in the vicinity of the Ca²⁺ release sites after a Ca²⁺ release event, is responsible for the refractory behaviour of Ca²⁺ release sites observed in vivo (Ilyin & Parker, 1994).

In contrast to the lack of short latencies in channel activation and deactivation, latency distributions of channel inhibition by Ca²⁺ and recovery (Fig 2G,H) have no appreciable deficits of short events. This indicates that the active conformation adopted by the channel in optimal (2 μM) [Ca²⁺]_i is in effect connected directly to an inactive conformation through the binding of a single Ca²⁺ to one inhibitory low-affinity site. More remarkably, even in the presence of 300 μM Ca²⁺_i, the channel is in an inactive conformation that is connected directly to an active conformation through the dissociation of a single Ca²⁺, despite the tetrameric channel structure with possibly four equivalent inhibitory Ca²⁺ sites. Thus, Ca²⁺ inhibition of the channel can be described by a simple two-state kinetic scheme: O↔I. This is consistent with the observation that Ca²⁺ inhibition of steady-state gating of Sf9 InsP₃R in the presence of 10 μM InsP₃ showed a Hill coefficient of approximately 1 (Ionescu et al, 2006). It is possible that the multiple Ca²⁺ inhibitory sites have strong negative cooperativity so that binding of Ca²⁺ to one of them prevents Ca²⁺ from binding to the remaining vacant sites.

Kinetics of single-channel Ca²⁺ release burst

In cells, Ca²⁺ released by an individual channel raises [Ca²⁺]_i to inhibitory (∼100 μM) levels in the microdomain surrounding it. Our kinetic results predict that this Ca²⁺ will feed back to inhibit it and terminate release after a burst of activity. [Ca²⁺]_i jumps from low sub-activating (<10 nM) levels directly to inhibitory (300 μM) levels in the presence of 10 μM InsP₃ produced just such a burst of channel activity. This is due to Ca²⁺ first binding to the high-affinity sites to activate the channel before the low-affinity inhibitory sites become occupied (Fig 5A). Mean activation and inhibition latencies of channel activity bursts were 24±6 ms (Fig 5B, n=94, N=3) and 470±90 ms (Fig 5C, n=70, N=3), respectively. The mean single-channel burst duration was 470±100 ms (Fig 5D, n=50, N=3). The time course of these channel activity bursts represents the kinetics of the fundamental Ca²⁺ release event—the basic building block from which local concerted Ca²⁺ release by clusters of InsP₃R channels (puffs) and global propagating InsP₃R-mediated Ca²⁺ signals (waves) are generated through CICR (Berridge, 1997).

Figure 5.

Kinetics of InsP₃R channel activity bursts in response to [Ca²⁺]_i changes between sub-activating (<10 nM) and inhibitory (300 μM) levels in the presence of 10 μM InsP₃. (A) A typical current trace of InsP₃R channel activity burst in response to [Ca²⁺]_i increasing from <10 nM to 300 μM. The burst activation and termination latencies are denoted by the dark and light grey bars, respectively. (B–D) Logarithmically binned histograms of burst activation and termination latencies, and burst durations, respectively. The latency distributions were fitted (black curves) using more than one linear Markov chain schemes (blue curves). The number of functionally indistinguishable conformations in each scheme is tabulated next to the corresponding blue curve. Short latencies of burst termination were not observed in (C) because when Ca²⁺ binds rapidly to the inhibitory sites before it binds to the activating sites, no channel activity follows. (E,F) Typical current traces showing different InsP₃R channel responses to rapid decrease in [Ca²⁺]_i from 300 μM to <10 nM.

A considerable fraction (9 out of 103) of [Ca²⁺]_i jumps from <10 nM to 300 μM failed to generate a burst (although [Ca²⁺]_i jumps before and after those jumps did generate bursts), substantially greater than that (7 out of 607) for [Ca²⁺]_i jumps from <10 nM to 2 μM. Thus, Ca²⁺ binding to the high-affinity activating sites and low-affinity inhibitory sites is not sequential, and the failure of [Ca²⁺]_i jumps (<10 nM to 300 μM) to generate an activity burst reflects the non-zero probability of Ca²⁺ binding to an inhibitory site before the activating sites becomes occupied. The mean latency for recovery from Ca²⁺ inhibition is over an order of magnitude longer than the mean latency of deactivation by Ca²⁺_i removal (P≪0.001). Therefore during most [Ca²⁺]_i drops (from 300 μM to <10 nM), no channel activity was observed as the channels were deactivated by Ca²⁺ dissociation from the high-affinity activating sites before they recovered from Ca²⁺ inhibition (Fig 5E). However, in 6 out of 94 [Ca²⁺]_i drops, the channel recovered from Ca²⁺ inhibition first, generating a brief burst of activity before it was deactivated by Ca²⁺_i unbinding (Fig 5F). This indicates that Ca²⁺ dissociation from the inhibitory and activating sites is also not sequential.

Conclusions

The rapid kinetic responses of single InsP₃R channels in native ER membranes to [InsP₃] and [Ca²⁺]_i changes observed here have provided new insights into the molecular mechanisms of channel activation and inhibition by these ligands. These measurements have shown that there is no requirement for sequential InsP₃ and Ca²⁺ binding during channel activation or deactivation, but binding of the two ligands is nevertheless not independent. A complex interaction involving both positive and negative cooperativity exists between InsP₃ and high-affinity Ca²⁺ binding sites, which was not anticipated in most theoretical models of the InsP₃R. The kinetics of high [Ca²⁺]_i inhibition of channel activity have provided insights into the molecular basis of fundamental Ca²⁺-release events and the refractory period that accounts for [Ca²⁺]_i oscillation frequency and wave annihilation. These single-channel kinetic data thus provide novel quantitative details for future construction of realistic models to account for the role of InsP₃R channel kinetics in generating cytoplasmic Ca²⁺ signals (Shuai et al, 2007). The approaches described here can also be used in future studies to define the kinetics of other types of InsP₃R channel regulation, and facilitate better understanding of [Ca²⁺]_i signalling by the ubiquitous InsP₃ pathway.

Methods

Single-channel patch-clamp studies of InsP₃R channels in a native membrane environment using isolated nuclei from cultured insect Ispodoptera frugiperda (Sf9) cells were performed in cyto-out excised patch configuration. The cytoplasmic side of the nuclear membrane patch and of any InsP₃R channels therein was exposed to different perfusing solutions containing various ligand concentrations during a continuous patch-clamp recording using a rapid perfusion set-up. Activation latency is the duration between the solution switch time and the first observed channel opening event in response to the solution switch. Similarly, deactivation latency is the duration between solution switch time and the last observed channel closing. The analysis allows latencies >5 ms to be clearly and consistently resolved. Current traces from cyto-out membrane patches containing one to several (⩽10) active InsP₃R channels were selected for latency analysis (supplementary information online).

Supplementary information is available at EMBO reports online (http://www.emboreports.org).

Supplementary Material

supplementary Information