Differential regulation of ion channels function by proteolysis

Abstract

Ion channels are pore-forming protein complexes in membranes that play essential roles in a diverse array of biological activities. Ion channel activity is strictly regulated at multiple levels and by numerous cellular events to selectively activate downstream effectors involved in specific biological activities. For example, ions, binding proteins, nucleotides, phosphorylation, the redox state, channel subunit composition have all been shown to regulate channel activity and subsequently allow channels to participate in distinct cellular events. While these forms of modulation are well documented and have been extensively reviewed, in this article, we will first review and summarize channel proteolysis as a novel and quite widespread mechanism for altering channel activity. We will then highlight the recent findings demonstrating that proteolysis profoundly alters Inositol 1,4,5 trisphosphate receptor activity, and then discuss its potential functional ramifications in various developmental and pathological conditions.

Introduction

Cells and intracellular organelles are bounded by biological membranes that function as electrical insulators. Ion channels are pore-forming protein complexes embedded in biological membranes that allow the rapid flux of ions in a direction dependent on their electrochemical gradient [1, 2]. By virtue of selectivity and exquisite control over activity, ion channels establish the resting cellular membrane potential and subsequently their behavior in response to cellular stimuli controls a multitude of physiological and pathological events [3–5].

In order to selectively control a diverse array of biological processes with fidelity and selectivity, ion channel activity is strictly regulated [1, 3]. Fundamentally, channel gating can be initiated following the binding of ligands, in response to changes in membrane potential or pH, together with sensing of mechanical stretch [1–3, 6] (Figure 1A). Channel activity is also frequently further fine-tuned by a myriad of regulatory molecules and inputs. A common theme is that the activity of specific channels is selectively regulated by the binding of ions, nucleotides and accessory proteins, together with phosphorylation and redox status and additional post-translational modifications [3, 7, 8]. A further consideration is that many ion channels are protein complexes assembled from multiple individual channel subunits [9, 10]. Thus, the particular subunit composition and their specific regulation can be a major determinant of the overall biophysical properties of an individual channel [3, 9].

Figure 1.

Schematic diagram summarizing different potential forms of proteolytic regulation of ion channels. In normal condition, ion channels are activated by the corresponding stimulus (A). Proteolytic fragmentation functions as a regulatory event, which can disable (B), activate (C), or regulate (C) channel activity in a channel specific manner.

While these “conventional” forms of regulation are well documented, recently a less orthodox, novel form of regulation has emerged as an event, widely employed to alter ion channel activity. Intracellular proteases, including caspase and calpain introduce peptide cleavage at specific sites and result in channel and isoform specific regulation of channel activity. Although, intuitively it might be expected that proteolysis might simply always function as a binary switch to deactivate ion channels, proteolysis can also alter the biophysical properties of channels to enhance or attenuate activity or influence the pharmacological properties of specific ion channels [11–15]. With major reference to proteins involved in Ca²⁺ signaling, or where proteolysis is regulated by Ca²⁺, we first highlight the diversity of ion channels subject to proteolysis as regulatory events and discuss the corresponding functional consequences of receptor cleavage. We then focus on recent findings which suggest that proteolytic cleavage of the ubiquitous intracellular Ca²⁺ release channel, the inositol 1,4,5 trisphosphate receptor (IP₃R), profoundly alters activity in a subtype-specific fashion.

Disabling channel activities by proteolysis

Somewhat instinctively, a major consequence of ion channel proteolysis is the abrogation of activity as the proteins are disabled following proteolytic cleavage (Figure 1B). In this scenario, the abundance of functional ion channels and thus the macroscopic current density decrease with no influence per se on their individual biophysical properties. Examples are found for various types of ion channels including voltage-gated potassium channels (hERG) [16–18], acid-sensing ion channels (ASIC-1) [19], NMDA receptors (NR2A subunits) [20], the Store-Operated Calcium Entry (SOCE) machinery (STIM1) [21, 22], and chloride channels (CFTR) [23]. (See Table 1). Exemplars illustrative of the disabling of channel function from the calcium signaling field are described below.

Table 1:

Summary of the proteolytic regulation of ion channels

Protein	Protease(s)	Cleavage site(s)	Functional outcome of proteolysis
hERG [16]	calpain, proteinase K, proteinase XIV and XXIV	In the S5-pore linerGly- 603 for calpain	Disable channel activity
NMDA receptor: NR2A subunit [20]	calpain	After the amino acid 1051 in the C-terminal region	Disable channel activity
STIM1 [21, 22, 26]	calpain, γ-secretase and casepase-3		Disable channel activity
CFTR [23]	calpain	Between the first nucleotide-binding site and the regulatory domain	Disable channel activity
Voltage gated sodium channel: b2 subunit [89]	BACE1 and γ secretase		Disable channel activity
ENaC [90]	furin	Arginine 205 and arginine 231 on the a subunit; arginine 143 on the γ subunit for furin	Activate the channel
TRPC5 [31]	calpain	Threonine 857	Activate the channel
Nav. 1.6 [11]	calpain		Decrease the activation threshold of INaP and increase its amplitude
ASIC-1a [12, 91]	trypsin, chymotrypsin and proteinase K	Arginine 145 for trypsin	Shift both the pH dependence of channel activation and the steady- state channel inactivation to lower pH values; be resistant to the inhibition of venom of P. cambridgei and display faster recovery from inactivation than wild type channels
EAG2 [14]	calpain		Decrease current density; a positive shift in voltage dependence of channel activation; a lack of EAG2 signature fast activation
Cav 1.2 [32, 34]	calpain, proteasome		Alter channel voltage-current relationship and voltage-dependent channel inactivation
TRPM7 [13]	caspase	Aspartic acid 1510	Potentiate TRPM7 channel activity
NMDA receptors: NR1 and NR2B subunits [39, 40]	tissue plasminogen activator	Arginine 260 in NR1 Arginine 67 in NR2B	Enhance NMDA-mediated Ca2+ influx for NR1; reduce ifenprodil inhibition and increase glycine EC50 for NR2B
CNG channel [36]	metalloproteinase 2 and 9		Increase the apparent affinity for cGMP and the efficacy of cAMP to regulate CNG channel activity
R1 [15]	caspase and calpain	Aspartic acid 1891 for caspase; Glutamic acid 1917 for calpain	Increase the frequency of Ca2+ oscillations and augment single channel open probability; abolish the PKA regulation of R1
R2 and R3 [65]	Digestive enzymes	Third and fourth solvent exposed regions	Decrease the frequency of Ca2+ oscillations; decrease the single channel open probability.

NMDA receptors (NR2A subunits).

Prolonged stimulation of NMDA receptors results in calpain activation and the NR2A subunit of the NMDA receptor is a substrate for this protease [20]. Calpain activity ultimately leads to a reduction in NR2A protein levels. Inhibition of calpain by calpastatin expression prevents NMDA receptor degradation, promotes Ca²⁺ uptake, and subsequently potentiates agonist-induced cell death. These observations strongly suggest that NR2A proteolysis by calpain is protective and that this event prevents Ca²⁺ overload and thus abrogates excitotoxicity [20].

Store operated Ca²⁺ entry machinery: STIM1.

STIM1 is an ER transmembrane protein that closely monitors the Ca²⁺ concentration in the endoplasmic reticulum (ER) [24]. Upon ER Ca²⁺ depletion, STIM1 aggregates to form oligomers which interact with Orai1, the pore-forming subunit on the plasma membrane to result in Ca²⁺ influx. This process is termed store operated Ca²⁺ entry (SOCE) [25]. Calpain, γ-secretase and casepase-3 can cleave STIM1 thus regulating SOCE [21, 22, 26]. The Michalak group reported that calpain dynamically regulated STIM1 turnover and consequently SOCE through proteolysis of STIM1 proteins in both basal and apoptotic conditions. Silencing the regulatory subunit of calpain, CAPN4, inhibited calpain activity and blocked STIM1 fragmentation. This was associated with a substantial increase in the SOCE [22]. The transmembrane region of STIM1 has a similar sequence motif to that of amyloid precursor protein (GGVXIX) which is a substrate of γ-secretase. Indeed, this protease can cleave STIM1 to markedly reduce STIM1 oligomerization, and the degree of SOCE. Altered proteolysis of STIM1 is associated with detrimental effects. For example, mutations in PS1, a component of γ-secretase, were discovered in Familial Alzheimer and these patients exhibited a higher γ-secretase activity, enhanced STIM1 cleavage and lower SOCE [21]. Further, cultured hippocampal neurons expressing mutant γ-secretase had destabilized dendritic spines that were rescued by inhibition of γ-secretase activity or overexpression of STIM1 [21]. In addition, it has been shown that in submandibular gland acinar cells, exposure to ionizing radiation led to caspase-3 activation and consequently STIM1 cleavage through a TRPM2-dependent pathway. Proteolysis of STIM1 further resulted in a marked loss of SOCE, and consequently decrease in fluid secretion [26]. In summary, proteolytic fragmentation of STIM1 by calpain, γ-secretase and caspase-3 regulates the amount of functional STIM1 on the ER membrane and hence the extent of Ca²⁺ influx via the SOCE mechanism.

Cystic fibrosis transmembrane conductance regulator (CFTR).

CFTR is a chloride channel that plays a crucial role in fluid transport across epithelial cells [27, 28]. Calpain has been reported to initiate the turnover of CFTR by cleaving the full-length channel [23]. Although the fragmentation site was not identified, based on the size of the proteolytic products, the cleavage was predicted to occur between the first nucleotide-binding site and the regulatory domain. Intriguingly, calpain proteolysis does not lead to the immediate disassembly of the protein fragments but appears to only “mark” the protein for internalization. This concept was supported by the observation that fragmented CFTR peptides remained associated and were mainly evident in cytosolic fractions [23]. Subsequently, internalized CFTR was completely degraded via the lysosomal pathway. The turnover of CFTR is a dynamic process and is regulated by [Ca²⁺]_i and the chaperone HSP90. A role for [Ca²⁺]_i in calpain mediated CFTR turnover is not unexpected because Ca²⁺ is a major regulator of calpain activity. This also implies a correlation between CFTR function and intracellular Ca²⁺ homeostasis. In addition, in the presence of the chaperone, HSP90, degradation of fragmented CFTR was almost completely inhibited while the removal of full length CFTR was markedly decreased. These regulatory events define the ratio of full-length channel to the fragmented channel. In the future, it will be important to investigate whether the fragmented channel preserved by HSP90 has altered biophysical properties. Further, given that dysregulation of CFTR localization, expression and function are implicated in several serious diseases [27], it is conceivable that calpain mediated CFTR turnover may have fundamental biological significance.

Activating channels by proteolysis

The structures of the majority of ion channels, can be generally divided into two domains: 1) a channel domain, containing the ion conducting pore and 2) a regulatory domain that receives modulatory input and in so doing adjusts channel activity. In many cases, the regulatory domain inhibits channel activity by maintaining the channel in a closed state in the absence of the stimulatory input. Interestingly, there are examples showing that region specific proteolysis cleaves in the regulatory domain to eliminate any inhibitory effect, and thereby facilitates channel function (Figure 1C). Examples where cleavage results in an increase in activity include the electroneutral (epithelial) sodium channel (ENaC) and the Ca²⁺ conducting channel TRPC5 (see Table 1).

Transient receptor potential canonical 5 (TRPC5).

TRPC5 is a non-selective cation channel that is involved in regulating the number of neuronal processes and growth cones [29, 30]. Calpain 1 and 2 cleave TRPC5 at the C-terminus to produce a fragmented receptor approximately 15 kD smaller than the full-length channel [31]. Threonine 857 appears to be an essential sites for calpain 2 as mutation of this residue to alanine greatly reduced proteolysis [31]. Calpain mediated receptor fragmentation has been convincingly demonstrated to increase TRPC5 channel activity in multiple experimental settings. First, whole cell patch clamp recording showed that co-expression of constitutive active calpain with TRPC5 significantly increased the basal channel activity. This result was then confirmed in excised patches where purified calpains activated TRPC5 in a manner dependent on the catalytic activity of calpain. Furthermore, truncated TRPC5 channel was constitutively active, and the activity was further potentiated by an increase in [Ca²⁺] following stimulation of muscarinic receptors. Proteolytic fragmentation of TRPC5 has also been suggested to participate in semaphoring 3A-induced neuronal growth cone collapse as either inhibition of calpain activity or genetic knockout of TRPC5 attenuated this process [31].

Regulation of ion channels activities by proteolysis

In contrast to the previously described roles of proteolytic cleavage, emerging evidence suggests that region specific fragmentation can regulate channel activity by altering the biophysical properties or susceptibility of channels to pharmacological reagents (Figure 1D). It is likely that particular regulatory events require peptide continuity for the appropriate communication between the regulatory domains and the ion conducting pores. In this scenario, disruption of peptide continuity may eliminate, or alter, propagation of regulatory signals to the channel domain and thereby impact channel activity. To date, Na⁺ channels (Nav 1.6, ASIC-1) [11, 12], K⁺ channels (EAG2) [14], Ca²⁺ channels (Cav 1.2, TRPM7) [13, 32–34], and non-selective cation channels (CNG channels, NMDA receptors) [35, 36] have been reported as substrates whose biophysical properties are regulated by proteolysis (see Table 1). We describe several examples of Ca²⁺ conducting channels where proteolysis alters their properties and finally highlight how this form of regulation alters IP₃R activity in a subtype dependent manner.

Voltage gated Ca²⁺ channel: Cav1.2.

Cav1.2 is a voltage-gated Ca²⁺ channel that contributes to Ca²⁺ influx in excitable cells [37]. Multiple sites in the Cav1.2 pore-forming subunit are subject to calpain and proteasome mediated proteolytic cleavage in an activity dependent manner [32–34]. Despite diverse fragmentation sites, all cleaved Cav1.2 channels share the common feature that they remain functional and membrane associated but nevertheless display a remarkable reduction in macroscopic current density [32]. Further, by expression of channels assembled from various complementary peptides representing cleavage products, it was shown that proteolysis at particular sites also dictates the current-voltage relationship and the extent of voltage-dependent channel inactivation [32]. Moreover, by simultaneously visualizing the N- and C- termini of these channels using confocal microscopy, it was shown that cleaved Cav1.2 could either remain associated or become separated into its peptide constituents. Remarkably, the dissociated channel fragments could also associate with intact channels and alter their biophysical properties [32]. Altogether, both the fragmentation site and the mobility of the fragmented channel contribute to regulation of Cav1.2 activity by proteolysis.

Transient receptor potential melastatin 7 (TRPM7).

Transient receptor potential melastatin 7 (TRPM7) is a unique Ca²⁺ channel which harbors serine/threonine kinase activity at its C-terminus. Knockout of TRPM7 inhibited activation and Fas-induced T cell apoptosis, thus indicating a prominent role of TRPM7 in this process [13]. Fas stimulation resulted in caspase mediated channel fragmentation at aspartic acid 1510, with the N-terminal fragment containing the complete sequence of the channel domain and the C-terminal fragment comprising the kinase domain. Cleaved TRPM7 channels displayed a significant higher current density but showed no alteration in their current-voltage relationship [13]. This suggests that caspase cleaves and potentiates TRPM7 channel activity. Further, while TRPM7 knockout T cells ectopically expressing a caspase cleavage resistant TRPM7 were impervious to Fas application, expression of cleaved TRPM7 effectively induced apoptosis even in the absence of Fas stimulation [13]. These data suggest that in response to Fas stimulation, caspase mediated channel fragmentation promotes TRPM7 channel activity and this event subsequently contributes to T cell apoptosis.

NMDA receptors (NR1 and NR2B subunit).

Tissue plasminogen activator (tPA) forms a complex with the NR1 subunit of the NMDA receptor and regulates the receptor activity through proteolytic cleavage [38, 39]. The Buisson group reported that tPA activity was dynamically increased in cortical neurons [38]. In response to depolarization, biologically active tPA was released into the extracellular space and fragmented NR1 subunit at Arginine 260 in the N-terminus (39). tPA pretreatment enhanced NMDA-mediated Ca²⁺ influx, which was blocked by mutation of Arg 260, which renders the channel resistant to tPA proteolysis (39). More importantly, tPA potentiated the NMDA mediated excitotoxicity in a dose-dependent manner presumably as a result of NMDA receptor fragmentation.

The NR2B subunit of the NMDA receptor is also subject to proteolytic regulation by tissue plasminogen activator (tPA) [40]. A truncated NR2B protein, designed based on the predicted site of tPA cleavage was unaffected in terms of glutamate activation but exhibited reduced propensity to be inhibited by the antagonist ifenprodil and displayed an increased affinity for glycine. In total, this information demonstrates that tPA can cleave both NR1 and NR2B, but exerts distinct regulation of individual subunits; fragmentation of NR1 markedly enhances channel activity while cleavage of NR2B dramatically alters this subunits pharmacological properties.

Cyclic nucleotide-gated channel (CNG channel).

Cyclic nucleotide gated ion channels (CNG channels) couple the light induced alteration in cGMP concentration with electrical signals and underlie mammalian phototransduction [41, 42]. It has been reported that metalloproteinase 2 and 9 (MMP2 and MMP9) cleave, and thereby alter, both the biophysical and the pharmacological properties of CNG channels [35, 36]. Specifically, extracellular exposure to MMP2 or MMP9, significantly increased the apparent affinity for cGMP and the efficacy of cAMP to regulate CNG channel activity [36]. This is consistent with the observation that exposure of CNG channels to MMP9 substantially augmented the single channel open probability following cGMP exposure [29]. Interestingly, the degree of MMP-dependent CNG channels fragmentation is dictated by the extent of channel glycosylation, as this post-translational modification prevents CNG channel cleavage [35]. Of note, in mouse retina, aged mice exhibited an elevated level of MMP9 activity with a concomitant higher level of fragmented CNGA1 channels [36] and thus, the impact of cleaved CNG channels in the retina during aging remains an exciting field for further investigation.

Proteolytic regulation of IP₃R

IP₃R are ubiquitous, intracellular Ca²⁺ release channels predominantly expressed on the endoplasmic reticulum membrane [3, 7, 43] IP₃R are capable of regulating a diverse array of cellular activities including, gene expression, cell proliferation, differentiation, migration, muscle contraction, digestive enzyme secretion, and cell death [3, 7, 44–46]. To regulate the diverse activities mentioned above with high specificity and fidelity, IP₃R activate downstream signaling cascades by generating Ca²⁺ signals with distinct spatial and temporal characteristics. The versatility of Ca²⁺ signals mediated by IP₃R is conferred by a series of complex regulatory events. For example, in addition to IΡ₃ [47], Ca²⁺ acts as a co-agonist and regulates IP₃R activity in a biphasic manner [48]. In particular, low concentrations of Ca²⁺ activate, whereas high concentrations of Ca²⁺, inhibit IP₃R activity. This underlies the ability of IP₃R to contribute to the generation of oscillatory Ca²⁺ signals even when the IP₃ concentration is constant [49]. Further, nucleotides can bind to and modulate IP₃R channel activity in the presence of elevated IP₃ and Ca²⁺ [8, 48]. For example, ATP regulates IP₃R in an isoform-specific manner. Specifically, IP₃R type 1 (R1) and IP₃R type 3 (R3) are regulated by ATP at all concentrations of IP₃, while the activity of IP₃R type 2 (R2) is only regulated at sub-maximal concentration of IP₃ [8, 48]. Numerous proteins function as binding partners and modulate IP₃R properties including the IP₃ binding affinity, Ca²⁺ dependency, cellular localization, receptor turnover, and channel activity [7, 44, 50]. In addition, post-translational modification such as phosphorylation, ubiquitination, and redox state also have profound effects to alter and fine-tune IP₃R activity [8, 51, 52].

In the last decade, our lab and others have investigated how proteolytic cleavage of IP₃R influences activity [15, 53–56]. All three isoforms of IP₃R are substrates of proteases. Early in vitro studies investigating IP₃R domain structure demonstrated that exposure of R1 to low concentrations of trypsin resulted in five predictable and reproducible receptor fragments (Figure 2A/B) [57–59]. These data were interpreted to indicate that R1 comprises five compact globular domains, interconnected by four solvent exposed linker regions. Notably, the first four tryptic fragments comprise the IΡ₃ binding core and much of the cytosolic domain, while the fifth fragment constitutes the transmembrane domain consisting of six helices encompassing the Ca²⁺ permeation pore and the cytosolic C-terminal tail [57]. Notably, IP₃R can also be cleaved in vivo and in vitro by intracellular proteases. Both caspase and calpain have been shown to cleave R1 at the fourth solvent exposed linker region and produce at least two fragments (Figure 2C/D) [56, 60, 61].

Figure 2.

Proteolytic fragments mapped to the Cryo-EM structure of R1. A and C are based on pdb 3JAV. A depicts a single subunit of R1 in which the five tryptic fragments generated by limited trypsin digestion are color coded. B shows these fragments in the linear structure. C Depicts a dimer of R1 in which the soluble (red) and membrane fragments (blue) formed by calpain and caspase activity in vivo are shown. D shows the calpain and caspase fragments in the linear structure of R1.

Given the high sequence homology among all three isoforms, R2 and R3 are predicted to have the same overall structure as R1 and be subject to protease cleavage [57, 62, 63]. For example, although the proteases were not definitively identified, in an early study it was shown that R2 and R3 were proteolytically processed into lower mass products in pancreatic acinar cells in models of the inflammatory disease, acute pancreatitis [64]. Remarkably, inhibition of proteasome activity failed to eliminate receptor fragmentation, indicating that other proteases, likely inappropriately and prematurely activated digestive enzymes, account for the proteolytic cleavage in acinar cells. Consistent with this idea, using an in vitro acute pancreatitis model where isolated pancreatic acinar cells were incubated with bile salts, R2/3 fragmentation was also readily evident [65]. Notably, in this case, fragmentation was ameliorated by pre-incubation with cell permeable trypsin inhibitors. These data indicate that IP₃R fragmentation may be a common event in acute pancreatitis and that trypsin activity is either directly or indirectly involved in the process of R2 and R3 fragmentation. Furthermore, these observations provide evidence that proteases can cleave R2 and R3 at specific sites and produce receptor fragments in “primary” cells. R2 and R3 have also been reported to be substrates of caspase and calpain [66]. However, whether there are reproducible fragmentation patterns resulting from the activity of these enzymes requires further investigation.

The biochemical characteristics of fragmented IP₃R were first thoroughly investigated in cells expressing R1 in isolation. Staurosporine, a general kinase inhibitor, which promotes caspase and calpain activation and apoptosis resulted in IP₃R fragmentation in DT40–3KO cells (null for IP₃R) stably re-expressing R1. Notably, native gel analysis demonstrated that all fragmented R1 migrated at an identical molecular weight to the intact R1 tetramer, indicating that receptors retain tetrameric architecture even after proteolysis [56]. These data are consistent with a previous report showing that R1 fragments generated following exposure to trypsin remained associated, presumably through non-covalent interactions [58]. Moreover, receptor fragments were not present in the cytosol, but were retained in ER membranes [56]. Taken together, these data demonstrated that R1 are retained in the ER as a tetramer after proteolytic fragmentation. The native structure of the proteolytically cleaved R2 and R3 was also investigated. Given the same overall topology is shared among all three isoforms of IP₃R, it was not unexpected that R2 and R3, like R1 were shown to retain tetrameric architecture and remain membrane associated after proteolysis [65].

Studying the function of fragmented IP₃R is a major challenge for a number of reasons. Principally, following protease activation, many proteins and molecules associated with IP₃R regulation may also be protease substrates and thus any functional readout is likely influenced by cleavage of a multitude of factors which potentially influence IP₃R activity. In addition, induction of protease activation using conventional methods, such as by staurosporine treatment and in acute pancreatitis models [56, 64], only cleaves an undefined portion of the IP₃R present. Thus, the functional readout is the net output of the activities of full length and fragmented IP₃R. These issues can be largely negated by expression of complementary receptor fragments from individual cDNAs, designed based on predicted protease sites and studying their function following stable expression in IP₃R null cells. By exploiting this strategy, we aimed to answer a number of major unanswered questions. For example, are tetrameric IP₃R assembled from complementary peptide chains and are fragmented IP₃R functionally gated by IP₃? Further, do fragmented IP₃R have distinct biophysical properties compared to full length IP₃R? To address the first issue, co-expression of pairs of complementary peptide chains was demonstrated to result in the assembly of the tetrameric IP₃R as determined by the native gel analysis [56]. This finding also suggested that the function of the fragmented IP₃R could be investigated using this general approach. Indeed, activation of cell surface G_q protein-coupled receptors were shown to evoke robust Ca²⁺ signals in cells expressing any pair of complementary IP₃R fragments representing various proteolytic cleavage events [15, 56, 67]. More strikingly, it was demonstrated that functional IP₃R can be assembled from 20 individual polypeptides, encoded by individual cDNA’s designed based on the five fragments that result from in vitro trypsin activity [15]. An important general implication of these data is that peptide continuity per se is obviously not necessary for IP₃R gating.

While R1 gating does not appear to require peptide continuity, it is possible that particular forms of allosteric regulation require communication of a conformational change throughout the peptide chain. It is well established that when IP₃R isoforms are expressed in isolation, that individual subtypes support characteristic temporal patterns of agonist-evoked Ca²⁺ release. For example, stimulation of cells solely expressing R1 always results in a few, irregular Ca²⁺ transients, (Figure 3A) whereas stimulation of R2 and R3 initiates robust Ca²⁺oscillations [68–70]. We interpret this observation as reflecting the net integrated output of the specific regulation of the activity of each isoform. One intriguing question is whether disruption of peptide continuity, while not markedly impacting IP₃-induced gating of the channel, influences communication between regulatory inputs and the channel domain, and consequently regulates overall IP₃R activity. To address this question, we expressed complementary peptide R1 fragments designed based on proteolytic products observed under conditions where R1 is fragmented in vivo. Intriguingly, cells expressing complementary fragments representing caspase, calpain or tryptic (I-III+IV-V) (Figure 2) fragmented R1, supported, in stark contrast to the intact R1, robust oscillatory Ca²⁺ signals following stimulation (Figure 3C and D) [15, 71]. This effect appears to be site specific, as receptor fragments mimicking fragmentation introduced at the second solvent exposed linker region showed no alteration in the temporal profile of Ca²⁺ signals (Figure 2B) [67]. In addition, caspase mediated receptor fragmentation significantly increased single channel open probability [15]. In total, these data suggest strongly that proteolytic fragmentation increases R1 channel activity and subsequent alterations in the effect of regulatory input results in an increase in oscillatory activity.

Figure 3.

Region-specific proteolysis regulates R1 activity. Single cell Ca²⁺ imaging was performed using cells expressing DT$) 3KO cells stably expressing R1 WT (A), R1 I-II+III-V (tryptic fragments) (B), R1 I-III+IV-V (tryptic fragments) (C) and R1 I-IV+V (caspase fragments) (D). Cells were loaded with fura-2AM followed by anti-IgM stimulation, which cross-linked the cell surface B cell receptors and induced continuous production of intracellular IP₃. In response to anti-IgM stimulation, cells expressing R1 WT (A) and R1 I-II+III-V (tryptic fragments) (B) evoked few Ca²⁺ transients while cells expressing R1 I-III+IV-V (tryptic fragments) (C) and R1 I-V+V (caspase fragments) (D) exhibited robust Ca²⁺ oscillations. Two representative single cell calcium traces (black and red) were shown for each condition. In (E), pre-incubation of cells with 20 μM forskolin for 2 min activated PKA and phosphorylated R1. This resulted in significant increases in the amplitude of Ca²⁺ signals mediated by the full length R1 and fragmented R1 I-III+IV-V, but not R1 I-IV+V. Each point represents one experiment. * indicates p< 0.05. Adapted from [15] with permission.

Based on these observations, we asked whether specific forms of regulation were altered by fragmentation. Phosphorylation by protein kinase A (PKA) is an established and important regulator of R1 activity that potentiates R1 mediated Ca²⁺ signals [15, 72]. PKA-phosphorylation in tryptic fragment IV at S1755 and S1589 is responsible for this effect. Interestingly, receptor fragmentation introduced to separate the phosphorylation sites and the channel domain into discrete fragments (e.g. R1 I-IV+V) abolished PKA regulation. Notably, when fragmentation was introduced more toward to the N-terminus (e.g. R1 I-III+IV-V), which left the phosphorylation sites and the channel domain in the same peptide fragment, PKA regulation was maintained [15]. One explanation for this region-specific receptor regulation is that some forms of regulation rely on direct peptide continuity for appropriate communication between the regulatory sites and the channel domain. In this scenario, proteolysis disrupts this communication and subsequently either abolishes or alters the regulatory outcome.

Notably proteolytic fragmentation also regulates R2 and R3 activity in a region-specific manner [66]. Using a similar experimental strategy, it was shown that while intact R2 and R3 support robust Ca²⁺ oscillations in response to sustained production of IP₃ (Figure 4A and D), receptor fragmentation introduced at the third or fourth, but not the second, solvent exposed regions suppress the Ca²⁺ oscillatory activity (Figure 4B, C, E, F). Thus, remarkably IP₃R fragmentation results in isoform specific effects. Based on the cryo-EM structure of R1 [73], these data suggest that proteolytic fragmentation must occur more toward to the C-terminal domain, between the ARM (armadillo solenoid folds) domains and the channel domain, to alter the temporal profile of Ca²⁺ signals mediated by IP₃R. The ARM domains are proposed to generate interfaces which bind regulatory molecules. Similar to the effect on PKA regulation of R1, proteolytic fragmentation in the solvent exposed region three and four may disrupt the appropriate transmission of the regulatory inputs from the ARM domains to the channel domain, which impact the signature temporal profile of Ca²⁺ signals of all three isoforms of IP₃R, but in an isoform specific manner.

Figure 4.

Region-specific proteolysis regulates R2 and R3 activity. Single cell Ca²⁺ imaging was performed using cells DT40–3KO cells stably expressing R2 WT (A), R2 I-III+IV-V (tryptic fragments) (B), R2 I-V+V (calpain fragments) (C), R3 WT (D), R3 I-III+IV-V (tryptic fragments) (E), R3 I-IV+V (tryptic fragments) (F). In response to anti-IgM stimulation, cells expressing R2 and R3 WT (A and D) evoked robust Ca²⁺ oscillations, while cells expressing R2 I-III+IV-V (tryptic fragments) (B), R2 I-V+V (calpain fragments) (C), R3 I-III+IV-V (tryptic fragments) (E), or R3 I-IV+V (tryptic fragments) (F) exhibited only a few irregular transients. Two representative single cell calcium traces (black and red) were shown for each condition. Adapted from [15] with permission.

Why does receptor fragmentation differentially regulate different isoforms of IP₃R? A current mathematical model suggests that the frequency of oscillatory Ca²⁺ signals mediated by IP₃R can be dictated by the rate at which Ca²⁺ activates and inactivates the receptor [71]. Full-length IP₃R may be regulated by Ca²⁺ in a subtype-specific manner and therefore show distinct temporal characteristics of Ca²⁺ signals [69, 74]. We speculate that proteolysis differentially affects the Ca²⁺ regulation of each isoform and exerts subtype-specific regulation. Support for this hypothesis is suggested by the alteration in the subtype-specific effect of single channel open probability (Po) in response to receptor fragmentation [15]. Specifically, upon receptor fragmentation, an increase in the Po for R1 or a decrease for R2 may result from the alteration of the rate of IP₃R to be activated or inactivated by Ca²⁺.

Physiological and pathophysiological roles for fragmented IP₃R.

The differential regulation of each IP₃R by proteolysis may have both physiological and pathological ramifications. For example, emerging evidence shows that both caspase and calpain are activated and play essential roles in developmental processes such as neural differentiation and skeletal muscle myoblast differentiation [75–77]. Of note, in these tissues, the activity of R1 is also pivotal. For example, mice with global knockout of R1 largely died in utero, and the surviving animals developed severe ataxia and epileptic seizures, and eventually died soon after weaning [78]. This strongly suggests that R1 is necessary for brain development and function. Consistent with this idea, Ca²⁺ release through R1 is critical for neuronal sprouting and arborization [79, 80]. It is tempting to speculate that low level caspase and/or calpain leads to R1 fragmentation, altered Ca²⁺ signals and initiation of a developmental gene program. Similarly, there also appears to be a role for R1 in skeletal muscle myoblast differentiation which occurs on a background of increased cellular protease activity. Interestingly, both the mRNA and protein levels of R1 are greatly increased during myoblast differentiation and knockdown of R1 results in a marked reduction in the expression of muscle-specific transcription factors [81]. These experimental systems provide potential platforms to investigate crosstalk between R1 and protease activity and future work is required to investigate the potential functional role of fragmented R1 during these developmental processes.

Fragmentation of both R2 and R3 were first described in pancreatic acinar cells in a model of acute pancreatitis [64], however, the functional significance has remained elusive. Ca²⁺signals, specifically those that result in the prolonged cytosolic Ca²⁺ overload, have been suggested to play important roles in the etiology of acute pancreatitis [82–84]. In support of this contention, various strategies aimed at inhibiting the Ca²⁺ overload, including inhibition of IP₃R, show protective effects [85–88]. Of note, region-specific receptor fragmentation occurs at an early stage during acute pancreatitis and substantially decreases the total Ca²⁺ release mediated by IP₃R, which consequently reduces the possibility of ER Ca²⁺ depletion and the subsequent Ca²⁺ influx [65]. These data suggest that proteolytic fragmentation of R2 and R3 may have a protective effect by preventing pathological Ca²⁺ signals and consequently delaying the progress of acute pancreatitis.

Summary

In conclusion, we summarize evidence showing that proteolysis acts as a regulatory event for a diverse array of ion channels. Proteolytic fragmentation can induce, inhibit, or mediate channel activity both in physiological and pathological conditions. This modulation is accomplished by cleaving either the pore-forming subunits or auxiliary proteins. Further, proteolytic regulation has a high specificity and selectivity; many proteases only cleave channels harboring unique motifs and thus leave others intact. This ensures precise regulation of downstream signaling cascades without uncontrolled cellular damage. We also summarize the evidence showing that all three isoforms of IP₃R are substrates of proteases and fragmented IP₃R are still properly gated by IP₃ binding. Further, region-specific receptor fragmentation differentially regulates all three isoforms of IP₃R. While proteolytic fragmentation significantly increases the frequency of Ca²⁺oscillations mediated through R1 activation, disruption of peptide continuity can decrease the oscillatory activity mediated through R2 and R3. In total, burgeoning evidence indicates that proteolysis may be a common regulatory event for a large number of ion channels. This irreversible form of modification adds a unique regulatory mechanism by expanding the repertoire of modes of channel regulation that a particular channel is subject to.

Highlights.

• Ion channels are exquisitely regulated to fine tune function.

• A novel form of channel regulation is by proteolysis

• Proteolysis alters channel function to disable, activate of change channel activity

• Calcium regulated and permeable channels are subject to this form of regulation

• As examples, Inositol 1,4,5 trisphosphate receptors are regulated in a subtype specific manner following proteolysis by calpain, caspase and trypsin.