E-Conference: Defining Calcium Entry Signals


E-Conference: Defining Calcium Entry Signals

Overview and Objectives of the E-Conference

May 24 2004 9:07AM

STKE Editors

The E-Conference is designed to focus upon current understanding onthe function, control, and role of calcium entry signals in mammaliancells. One central question to be addressed is how calcium entry channels are activated in response to depletion of calcium from stores. Likewise,it is critical to know which channels contribute to such calcium entry.The focus is primarily upon calcium entry signals in nonexcitable cells in which voltage changes are not the predominant means by which calciumsignals are induced. However, nonvoltage-dependent mechanisms also playan important role in calcium entry in excitable cells, hence the focus isto include all cell types.

The conference will focus on three central questions, each of whichwill have its own thread or topic in the forum and which can be thought of as separate "sessions" within the conference. Several scientists havebeen invited to start the discussion on each topic, much like invitedspeakers at a meeting. The community is also encouraged to share theirquestions and comments on each topic by responding either to the maintopic in each session or responding to a specific response in the session.

We hope this approach stimulates discussion and leads to new ideasfor unraveling some of the remaining questions related to control ofcalcium entry.

E-Conference Chairpersons,
Donald L. Gill
Randen L. Patterson

STKE Editors,
L. Bryan Ray
Nancy R. Gough
Elizabeth M. Adler

Invited Participants

May 24 2004 1:28PM

E-Conference Chairpersons

Several researchers were asked to serve as invited participants for each of the three sessions, much like the invited speakers at a meeting. The chairpeople, the invited participants, and the signaling community are welcome to comment in any of the sessions.

Session 1: Physiological Processes that Mediate Activation of Calcium Entry Channels

  • James W. Putney, Jr.
  • Thomas Gudermann
  • Victoria M. Bolotina

Session 2: The Cellular Domains that Contribute to the Calcium Entry Events

  • Michael J. Berridge
  • Indu S. Ambudkar

Session 3: Defining the Plasma Membrane Calcium Channels

  • Reinhold Penner
  • Bernd Nilius
  • Trevor J. Shuttleworth

Session 1: Physiological Processes that Mediate Activation of Calcium Entry Channels

May 24 2004 1:22PM

E-Conference Chairpersons

This session is focused on the main question of what thephysiological transduction processes that mediate activation of calciumentry channels are.

Specific issues include

(a) The role of calcium stores in activating entry:

Are changes in intracellular stores necessary or sufficient input, or both, to activate calcium entry? Are intracellular calcium storesphysiologically depleted under normal cellular conditions?

(b) Distinctions between receptor-operated and store-operated calcium entry:

Do receptors activate distinct calcium entry events from thosestimulated by store-emptying? Are the calcium entry events activated byreceptors and emptied stores independent or integrated?

(c) The role of intracellular calcium release channels:

Do InsP3 receptors or ryanodine receptors, or both, mediate or modify calcium entry events?

(d) The function of mediators and adaptors in calciumentry:

What role is played by structural or adaptor proteins (includingHomer, NHERF, InsP3Rs, and PLC-gamma), or by soluble cellular messengers(such as InsP3, cyclic nucleotides) and lipid messengers (includingdiacylglycerol, arachidonic acid, and PIP2)?

TRPC channels and the second messengers that regulate them

Jun 4 2004 8:57AM

Thomas Gudermann

More than 20 years have elapsed since the suggestion was made that receptor activation could lead to Ca2+ entry into smooth muscle cells by mechanisms independent of membrane depolarization, and the concept of receptor-operated Ca2+ entry was developed. Receptor-stimulated cation channels are gated in response to agonist binding to a cell membrane receptor that is molecularly distinct from the channel protein itself.

Over the past couple of years a large family of mammalian homologues of the Drosophila transient receptor potential (TRP) visual transduction channel have been identified. These channels, in particular those of the canonical TRPC subfamily, are likely molecular correlates of receptor-operated cation channels as defined before. In order to add some spice to the discussion, I shall refer to TRPC proteins primarily as receptor-operated cation channels, because in my personal view none of them has been rigorously proven to be a store-operated Ca2+ channel in a direct sense. Thus, although TRPCs like any Ca2+-permeable ion channel may have an impact on store-operated Ca2+ influx, there appears to be insufficienthard evidence to classify them as genuinely store-operated. Therefore, in my opinion a clear distinction can be made between receptor-operated and store-operated Ca2+ entry with respect to TRPC proteins as potentialmolecular correlates.

Receptor-operated cation entry plays an eminent physiological role in vascular smooth muscle cells. I shall, therefore, use this cell type as a kind of physiological springboard for further arguments and views. In vascular smooth muscle cells, cation influx through non-selective cation channels is thought to be required for cell membrane depolarization in response to vasoconstrictors resulting in the activation of voltage-gated Ca2+ channels, Ca2+ influx, and constriction of blood vessels.

As TRPC6 is highly expressed in vascular smooth muscle cells, it represents a likely molecular candidate for the vasoconstrictor-activated Ca2+-permeable cation channel. As characterized in native vascular smooth muscle cells, the latter channels are activated via phospholipase C (PLC)-coupled receptors and by diacylglycerol (DAG) independent of protein kinase C. Upon heterologous expression in CHO-K1 cells, TRPC6 behaves as a receptor-activated non-selective cation channel insensitive to the depletion of internal stores being activated by DAGs independent of protein kinase C.

To summarize, TRPC3, TRPC6, and TRPC7 form a structural and functional subfamily of DAG-sensitive cation channels coupling receptor/PLC signalling pathways to cation entry. Although these TRPC proteins are generally classified as the DAG-responsive subfamily, it is still a highly contentious issue as to whether DAG can be regarded as the physiological activator of native channel complexes, while there is no doubt that all members of the TRPC3/6/7 subfamily can principally be activated by DAG. As deduced from pharmacological inhibition of DAG lipase and DAG kinase, endogenously generated DAG is sufficient for channel activation. Most notably, receptor agonists and DAGs do not display additive effects on TRPC3 and TRPC6 current amplitudes, suggesting that the same TRPC channelsare activated by DAG and by PLC-linked receptors and that DAG may be the decisive second messenger generated by PLC.

However, so far a direct interaction of DAG with TRPC3/6/7 proteins has not been demonstrated. Postulating such a direct contact between the lipid messenger and the channel protein as demonstrated for the interaction of capsaicin with TRPV1, possible interaction sites in thechannel protein might be located within the first intracellular loop and neighbouring portions of transmembrane helices 2 and 3. In the absence ofa mapped DAG-contact site in the channel protein, TRPC3/6/7 activation by C1 domain-containing proteins, such as chimaerins, MUNC13s, RasGRPs, and even DAG kinases, cannot be excluded and deserves experimental clarification.

Also, the role of PKC is more complex than initially assumed. Pharmacological inhibition of PKC, as well as enzyme down-regulation by long-term phorbol ester treatment, showed that PKC activity is not required for DAG-dependent TRPC3/6/7 activation. However, short-term PKC activation prior to DAG addition completely blocks channel gating, and PKC inhibition results in decreased TRPC deactivation. Thus, while not necessary forchannel activation, PKC is intrinsically involved in TRPC channel regulation.

As yet, we cannot satisfactorily answer the question as to whether DAG alone is sufficient for TRPC3/6/7 activation. Phorbol ester and DAG treatment of many cells is frequently accompanied by the engagement of tyrosine kinase-dependent signaling pathways. In fact, the Src-family tyrosine kinase Fyn physically associates with TRPC6 and phosphorylates the protein thereby substantially enhancing channel activity. However, it has not been reported whether ablation of Fyn-dependent TRPC6 tyrosinephosphorylation negatively impacts on DAG-induced channel activation.

Finally, one may even pose the heretical question whether the subsummation of TRPC3/6/7 as the DAG-sensitive subfamily is correct at all. Recently, patch-clamp recordings on vomeronasal neuron dendrites provided evidence that TRPC2 may also be a DAG-gated cation channel. Along these lines, DAG-activated currents were observed in TRPC5-expressing HEK 293 cells. An important methodological difference between these recent findings and other studies on DAG-sensitive TRPCs is the use of electrophysiology, which is a highly sensitive technique, on the one hand, and fluorescence imaging on the other hand. Thus, it appears to be worth examining whether in fact all TRPCs respond to DAG, albeit with different potency and efficacy.

Another lipid messenger that has been implicated as a modulator of TRP channel function is PIP2. The latter phospholipid inhibits Drosophila TRP and TRPL, as well as mammalian TRPV1. On the contrary, constitutive TRPM7 activity is enhanced upon PIP2 addition and rapidly inactivated by PIP2 hydrolysis. Clearly, further experimentation is required to come up with a unifying hypothesis to reconcile the discrepant effects of PIP2 onvarious TRP channels.

Some TRP channels (TRPC6, TRPV2, C. elegans TRP-3) have been described to be rapidly translocated to the plasma membrane in response to various stimuli like G-protein-coupled receptor agonists (TRPC6), growthfactors (TRPV2), and sperm activation (TRP-3). These observations give rise to the concept that some TRP ion channels may be held in reserve in intracellular vesicles. So far, however, a mechanistic understanding of the stimulus-induced translocation to the plasma membrane still remains elusive.

For the sake of brevity, I shall not go into detail to discuss the barrage of published data on the activation of TRPC proteins by store depletion. Suffice it to say, store-dependent and -independent gating mechanisms have been postulated for nearly each member of the TRPC family. However, there appears to be a kind of consensus allowing us to surmise that TRPC proteins are receptor-operated cation channels sharing a common gating mechanism that is contingent on PLC activation. Under particularcircumstances -- defined by the endowment of a special cell line with signaling proteins, TRP protein expression level, species and methodology chosen -- certain TRPCs might be part of store-operated calcium entry channel complexes.

One important reason for the discrepant results is related to different methods used. Fluorescence imaging is quite an indirect measure of channel activity, because it mirrors the accumulated free Ca2+ concentration irrespective of the source. It may even reflect reversed- mode Na+/Ca2+ exchanger contribution rather than TRPC channel activity. As overexpression of TRPC proteins frequently results in constitutive Ca2+ entry, store depletion and recalcification protocols that do not accountfor this fact will inevitably lead to false positive results. Patch-clamp recording as a direct approach to monitor channel activity has the potential to overcome most of these difficulties, but is prone to depletethe cell of diffusible messengers that might regulate ion channel function. Ideally, several independent methodological approaches should be applied and the results be critically compared.

Because the composition of functional TRP channel complexes is largely unknown, it has proven difficult to ascribe receptor-activated cation currents to molecularly defined TRP proteins. Several studies have recently shown that various TRP channels can assemble as heteromericcomplexes which differ in their biophysical properties. Systematic analyses on the principles of TRP channel formation offer the conceptual framework to assess the gating mechanism, regulation and physiological role of distinct TRP proteins in their native environment. So far, two principally different themes have evolved from TRP channel heteromultimerization: On the one hand, heteromultimerization can alter the biophysical properties of channel complexes whose individual memberscan also be functionally expressed alone (e. g. TRPV5/6), on the other hand, heteromultimerization is necessary to measurably transport certain TRP channels subunits to the cell membrane (e. g. TRPC1/5, TRPM6/7). In most instances, the daunting task to define the composition of channelcomplexes under physiological native conditions still remains to be addressed.

A last question that I think needs to be answered is: Why are there so many genetically and functionally related TRP proteins. For instance, why do we need three classical DAG-sensitive TRPC channels? Future experiments, preferably in vivo, will have to address the issue as to whether TRP proteins belonging to a certain subfamily are functionallyredundant or not and whether unique and indispensable physiological roles can be ascribed to individual members. I am absolutely sure that answering these questions will keep the scientific community busy for quite sometime.

TRPC channels and the second messengers that regulate them

Jun 4 2004 10:01AM

Randen L Patterson

I have a question about this statement "As deduced frompharmacological inhibition of DAG lipase and DAG kinase, endogenouslygenerated DAG is sufficient for channel activation. Most notably, receptor agonists and DAGs do not display additive effects on TRPC3 and TRPC6current amplitudes, suggesting that the same TRPC channels are activatedby DAG and by PLC-linked receptors and that DAG may be the decisive second messenger generated by PLC. "

1)Ma et al. (Science 2000, Figure 2), clearly demonstrated that the DAG could increase the flow of Strontium (Sr2+) through TRPC3 after agonist stimulation.......I would argue that additive effects of DAG can be seen, which would argue for independent mechanisms of DAG and receptor activation of TRPCchannels.

2)Pharmacological blockade of DAG breakdown is, in my opinion, no different than the external addition of DAG analogues. The levels that accumulate in response to this inhibition could be micromolar. Similar effects can be seen when using pervanadate to block tyrosine phosphatases. In only a few minutes, ALL tyrosine kinase substrates are highly phosphorylated.

Until direct binding studies of DAG are performed with TRPC channels, it can not be concluded that DAG plays a significant role in their physiological activation, although it is clear that DAG analogues can beused as a tool to manipulate TRPC channel activity.

TRPC channels and the second messengers that regulate them

Jun 8 2004 6:29AM

Thomas Gudermann

I completely agree that as yet there is insufficient experimental evidence around to state that DAG is the physiological second messenger of TRPC3/6/7 gating.

Further milestones in our understanding of TRPC activation would be the identification of DAG binding sites in TRPC channels (if direct binding actually occurs), genetic manipulation of these sites and thentesting for in vivo consequences. There is still a long way to go on this path.

However, there is nothing wrong with the speculation that "DAG may be the decisive second messenger generated by PLC". In the scientific literature, it is discussed controversially how DAG-sensitive TRPC channels are activated. I am very puzzled by the finding that some investigators claim that TRPC3 stably expressed in HEK293 cells (T3-clones) behaves as a store-operated Ca2+ permeable channel gated by conformational couplingwith the IP3 receptor, while others working with the same cell clone provide evidence that the same ion channel is gated store-independently, subsequent to PLC activation by a mechanism involving DAG. At present, I have no explanation for these discrepant results.

Honestly, I tend to be a little bit reserved with regard to the conclusions drawn from Figures 2H and 2I of Ma et al., Science 287:1647-1651, 2000. Simply looking at the morphology of cells treated with calyculin A illustrates that many processes are going on in these cells, one of which may be the disruption of TRPC/IP3 receptor contacts. Considering these drastic morphological changes, I personally find distinct functional differences between untreated and treated cells very difficult to interpret. Figure 2H would indeed indicate that receptor and OAG stimulation of TRPC3 are independent phenomena. However, if a different receptor were chosen, for instance the endogenous muscarinic receptor in HEK cells or a coexpressed H1 histamine receptor, OAG challenge subsequent to receptor stimulation would not elicit anymeasurable Sr2+ or Ca2+ influx. Thus, most probably the quantitative aspect of receptor-induced PLC activation determines whether a secondary OAG challenge has an effect or not.

This general view is further substantiated if one examines TRPC3 or 6activity electrophysiologically. The pro's and con's of fluorescence imaging versus electrophysiology have already been discussed in detail inother contributions to this E-conference, so I will not elaborate on this issue again. Suffice it to say, conclusions drawn gain credibility if both methods yield compatible results. Along these lines, at least two labshave shown independently by means of electrophysiology that receptor and OAG stimulation of cells have no additive effect on TRPC3 or TRPC6 current amplitude. Therefore, I am not yet convinced that receptor and DAGstimulation of TRPCs rely on independent mechanisms.

It does not come as a surprise that the role of lipid messengers for the activation of TRP and TRPL in the Drosophila eye is also a moot issue. Flies impaired in DAG kinase activity (leading to an increase in the local DAG concentration upon receptor stimulation) show enhanced spontaneous currents and light responses consistent with the concept that DAGs or metabolites, such as polyunsaturated fatty acids (PUFAs), gate the cationchannels. An alternative explanation is based on the role of PUFAs as metabolic uncouplers and the observation that metabolic inhibition activates light-sensitive TRP and TRPL channels. However, there is recent evidence that metabolic inhibition primarily impairs DAG kinase activityconsistent with the notion of TRP channel gating by DAG.

Last not least, can we really be sure that exogenously added membrane-permeable DAG analogs like OAG and DOG are not at all different from endogenously produced DAGs like SAG (steaoryl-arachidonoyl-glycerol) orSLG (stearolyl-linoleoyl-glycerol) with regard to their biological effects? How do we know about the local concentration of these DAGs following PLC stimulation? Chemically and biophysically SAG and SLG are clearly distinct from DAG and DOG. Thus, as stated before I would only feel confident with experimental results if external addition of membrane-permeable DAGs yield results compatible with those obtained after pharmacological blockade of DAG lipase and kinase and after addition ofmembrane-impermeable endogenous DAGs. Needless to say, the best way to study these "natural" DAGs is electrophysiology...

TRPC channels and the second messengers that regulate them

Jun 12 2004 2:16PM

Randen L. Patterson

I only have a few comments:

"I completely agree that as yet there is insufficient experimentalevidence around to state that DAG is the physiological second messenger of TRPC3/6/7 gating.

Further milestones in our understanding of TRPC activation would bethe identification of DAG binding sites in TRPC channels (if directbinding actually occurs), genetic manipulation of these sites and thentesting for in vivo consequences. There is still a long way to go on thispath."

1)This is true. Until the molecular biology and protein biochemistryis performed to ascertain whether or not DAG or its analogues can have adirect effect on TRPC channels, all other discussion is merely hypothesisdriven, without strong experimental evidence to support it.

"However, if a different receptor were chosen, for instance theendogenous muscarinic receptor in HEK cells or a coexpressed H1 histaminereceptor, OAG challenge subsequent to receptor stimulation would notelicit any measurable Sr2+ or Ca2+ influx. Thus, most probably thequantitative aspect of receptor-induced PLC activation determines whethera secondary OAG challenge has an effect or not."

2) We used the endogenous muscarinic receptor in that figure. Thesetypes of explanations are merely handwaving, see 1).

"However, there is nothing wrong with the speculation that "DAG maybe the decisive second messenger generated by PLC". In the scientificliterature, it is discussed controversially how DAG-sensitive TRPCchannels are activated. I am very puzzled by the finding that someinvestigators claim that TRPC3 stably expressed in HEK293 cells (T3- clones) behaves as a store-operated, Ca2+-permeable channel gated byconformational coupling with the IP3 receptor, while others working withthe same cell clone provide evidence that the same ion channel is gatedstore-independently, subsequent to PLC activation by a mechanism involving DAG. At present, I have no explanation for these discrepant results."

3)As I have stated in previous comments, one of the biggest problemswith this field is that very few researchers are looking at endogenousTRPC channels. With the advent of siRNA, there is no excuse that thosewho use fluorescence OR ELECTROPHYSIOLOGY to specifically measure calciumentry via TRPC channels, do not to delete endogenous TRPC channels fromtheir system. This would allow one to look for specific endogeous effects of TRPC channels, which could then be rescued with overexpression ofexogenous TRPC channels, and more importantly TRPC channel mutants.

"It does not come as a surprise that the role of lipid messengers for the activation of TRP and TRPL in the Drosophila eye is also a moot issue. Flies impaired in DAG kinase activity (leading to an increase in the local DAG concentration upon receptor stimulation) show enhanced spontaneouscurrents and light responses consistent with the concept that DAGs ormetabolites, such as polyunsaturated fatty acids (PUFAs), gate the cationchannels. An alternative explanation is based on the role of PUFAs asmetabolic uncouplers and the observation that metabolic inhibitionactivates light-sensitive TRP and TRPL channels. However, there is recentevidence that metabolic inhibition primarily impairs DAG kinase activityconsistent with the notion of TRP channel gating by DAG."

4)Since TRPC channels have evolved quite significantly from fly, itis unreasonable to think that they would have to retain the sameactivation mechanisms that are used in fly. As humans have already doneaway with TRPC2, which was an evolution of TRP from fly, it is my opinionthat even when the activation of TRP in fly eye is determined, it willprovide little insight into the activation and regulation of mammalianTRPC channels.

TRPC channels and the second messengers that regulate them

Jun 21 2004 3:38PM

Raghu Padinjat

As Randen Patterson has insightfully pointed out (12 June 2004,comment 3), “very few researchers are looking at endogenous TRPCchannels.” As one of those “few researchers” I would like tocontribute my perspective to this E-conference:

The Drosophila phototransduction cascade is perhaps the oldest andbest studied signalling pathway in which light stimulated, G-proteincoupled, PLC activity ends with calcium influx through two well-definedclasses of endogenous TRPC channels, TRP and TRPL. The total light-inducedcurrent in the fly eye is completely eliminated by protein null mutants in these two genes.

It has been known for over 15 years that phospholipaseC activity isessential for TRP and TRPL activation(analysis of norpA mutants); indeedit is fair to suggest that this finding was the conceptual basis for thewidely accepted and largely undisputed idea that mammalian TRPC channelscan be activated via phospholipaseC activity.

However, in the eye, despite a number of rigorous studies over the years, it has been difficult tounderstand the biochemical basis of this requirement for PLC activity(reviewed in detail elsewhere). In particular the requirement of InsP3receptor (InsP3R) activity (with or without calcium release) in TRP andTRPL activation have been difficult to confirm.

Most recently and ofimmediate relevance to this debate we have analysed flies in which theonly InsP3R in the completed version of the Drosophila genome has beendeleted (complete gene missing). When analysed by patch-clampelectrophysiology the activation of TRP and TRPL is completely normal.This suggests that InsP3R activity is not required to activate TRPCchannels in Drosophila photoreceptors.

In the meantime, although an enormous variety of thoughtfulexperiments have been done in mammalian cell lines to address theequivalent question, no one seems to have bothered to do the followingexperiment; analyse the activation of an endogenous TRPC channel in loss- of-function mutants in the three mammalian InsP3Rs within the context of a well-defined endogenous signalling cascade. Therefore with respect to therole of InsP3R activity in activating TRPC channels, until such anexperiment is done, it is perhaps less meaningful to conclude that“Since TRPC channels have evolved quite significantly from fly, it is it is unreasonable to think that they would have to retain the sameactivation mechanisms that are used in fly.” (Randen Patterson, 12 Junecomment 4).

With regard to the potential role of lipid second messengers inactivating TRPC channels I would once again like to place in perspectivethe analysis of this question in Drosophila phototransduction.

On thebasis of the numerous studies that Roger Hardie and I have now publishedon this issue, it seems most reasonable to conclude that in the fly eyediacylglycerol kinase plays a crucial role in inactivating the lightresponse (TRPC channel activity) during light induced PLC activity. Whatis less clear is how it does this,i.e the identity of the second messenger whose levels it regulates; candidates include diacylglycerol, itsmetabolites PUFA’s (polyunsaturated fatty acids), PIP4,5 or perhaps indeed more distant lipidmetabolites that have so far not caught the attention of the calciumsignalling community.

Until mammalian TRPC channel investigators can showthat within the context of an endogenous signalling cascade (see paragraph above) diacylglycerol kinase does not play a role in inactivating TRPCchannels it will be difficult to make “evolutionarly statements” about the equivalence or otherwise of the mechanisms by which lipid secondmessengers activate TRPC channels.

In summary it is worth reflecting that (1) Two key elements thatunderpin the mammalian TRPC signalling cascade namely the channelsthemselves and the requirement for phospholipaseC activity were discovered in the fly eye. (2) Until equivalent experiments have been done inmammalian systems, the evolution of different signalling mechanismsdownstream of PLC activity in fly v mammals must remain an untestedhypothesis.

TRPC channels and the second messengers that regulate them

Jun 23 2004 1:58PM

Randen L. Patterson

A few comments in response to Raghu,

Perhaps I was not clear when I stated very few researchers arelooking at endogenous TRPC channels, I should have included mammalian TRPC channels.........

As for this comment"In the meantime, although an enormous variety of thoughtful experimentshave been done in mammalian cell lines to address the equivalent question, no one seems to have bothered to do the following experiment; analyse theactivation of an endogenous TRPC channel in loss- of-function mutants inthe three mammalian InsP3Rs within the context of a well-definedendogenous signalling cascade. Therefore with respect to the role ofInsP3R activity in activating TRPC channels, until such an experiment isdone, it is perhaps less meaningful to conclude that “Since TRPCchannels have evolved quite significantly from fly, it is it isunreasonable to think that they would have to retain the same activationmechanisms that are used in fly.” (Randen Patterson, 12 June comment4)."

These experiemnts have been done in DT40 IP3R k/o cells(Vazquez et al JBC 2003). They demonstrated that overexpressed TRPC3 channels cannot beactivated by receptor alone in these cells. Overexpression of type3 IP3Rwith TRPC3 rescued TRPC3 activity in response to agonist. We also showedthat all calcium entry in response to IgM in these cells was dependentupon IP3 binding to the IP3R, not calcium release from the IP3R (vanRossum et al PNAS 2004) so I would say that this has in fact been tested,and demonstrates a requirement for the IP3R for the activation ofmammalian TRPC channels.

In light of this data, coupled to the rigorous work of Kiselyovdemonstrating by electrophysiology that the N-terminus of IP3R gates TRPC3 through its binding to IP3, there seems to be a large divergance from thefly TRPC channels to the mammalian channels.

TRP channels as non-selective cation channels

Jun 18 2004 11:11AM

Kenneth L. Byron

As the focus of this forum is defining Ca2+ entry signals, we haveperhaps limited our view of the roles of Ca2+-permeable ion channels.

Forexample, consider the statement "TRPC6 is highly expressed invascular smooth muscle cells, it represents a likely molecular candidatefor the vasoconstrictor-activated Ca2+-permeable cation channel." Although I agree with that statement, I wonder about the emphasis on Ca2+permeability of TRPC6.

The electrophysiological evidence from eitherexogenously expressed TRPC6 channels or endogenously expressed channelsthat have been attributed to TRPC6 in vascular smooth muscle cellsindicates that these are non-selective cation channels. I am not aware ofany evidence that, under physiological agonist concentrations and ionicconditions, the entry of Ca2+ via these channels directly elicits aphysiological response in vascular smooth muscle cells.

I think we ought to keep in mind that in excitable cells such asvascular smooth muscle, the primary role of activation of non-selectivecation channels may be to depolarize the membrane and consequentlyactivate voltage-sensitive Ca2+ channels. In fact Na+ entry may be moreimportant than Ca2+ entry under physiological conditions.

Some members ofthe TRPC family, e.g. TRPC4 and TRPC5, at least as homotetramers, mayfunction as monovalent cation channels, but may nonetheless contribute toCa2+ entry via membrane potential effects in excitable cells.

Even in non- excitable cells Na+ entry via non-selective Ca2+-permeable channels may be important for physiological responses. Focusing specifically on Ca2+entry, a consequence of Na+ entry may be reverse mode Na+/Ca2+ exchange,which may have important localized effects at regions of close appositionbetween plasma membrane and ER and may even account for some of thesustained Ca2+ signals measured using fluorescent Ca2+ indicators.

The role of Ca2+ stores in activating Ca2+ entry:

Jun 4 2004 9:04AM

Victoria M. Bolotina

Indeed, are the changes in intracellular stores necessary or sufficient (or both) to activate Ca2+ entry?

This is a very good question, and the answers are... "Yes and No", which is one of many reasons why this pathway remains such a complicated and controversial issue for such a long time.

Physiologically depletion of the stores definitely provides a major trigger for activation of store-operated channel (SOC), and in this respect changes in Ca2+ stores are indeed necessary for activating Ca2+ entry. Refilling of the stores is also needed for termination of Ca2+influx, and acceleration of CaATPase (SERCA)-dependent refilling indeed shuts down Ca2+ influx, as we (and others) have demonstrated as a mechanism for physiological effect of nitric oxide on Ca2+ influx in smooth muscle cells (SMC), platelets and other systems.

On the other hand, under experimental conditions, you donot necessarily need to deplete the stores-- you may activate the channels by affecting the pathway downstream from the stores. For example, in our new model in which iPLA2 is a crucial determinant of the pathway (Smani et al., Nature Cell Biology 2004), activation of Ca2+-independent phospholipase A2 (iPLA2) (bycell dialysis with 10 mM BAPTA, by calmodulin (CaM) displacement with CMZ or CaM inhibitory peptide, or by something else that is yet to be determined) results in activation of SOC (and Ca2+ influx) that is identical to that activated by depletion of the stores with thapsigargin (TG). In these cases, depletion of the stores was not necessary for activation of SOC and Ca2+ influx, because a short cut in this pathway was used. Can short cuts be present and used under some physiological /pathological conditions? We don't know, and this is something very interesting to look for in different cellular systems.

Is depletion of the stores sufficient for activation of SOC? In most cases yes, but in some cases it may not be enough. For example, inhibition of iPLA2 (disrupting the pathway), or simple down regulation of iPLA2 excitability (if it would be technically possible), makes depletion of the stores insufficient for activation of SOC. This disruption can be reversed by application of iPLA2 product, lysophospholipids, which we have shown to activate SOC channels even when iPLA2 is inhibited. Thus, regulation ofiPLA2 activity can serve as a major on-off switch (as well as a tuning device) for the whole pathway that connects depletion of the stores with activation of SOC and Ca2+ influx. The prediction may be that in experimental conditions, direct activation of iPLA2 (by any means) willproduce activation of SOC without depletion of the stores, and inhibition of iPLA2 will make store depletion insufficient for SOC activation.

Also, one should consider production and degradation of Lysophospholipids as another store-independent way for regulation of SOC activity. It is important to emphasize, that we presently know nothing about how lysophospholipid products of iPLA2 activate SOC. Activation may occur as a result of their direct interaction, as well as through some other still unidentified membrane delimited machinery, which awaits its discovery.

Distinctions between receptor-operated and store-operated Ca2+ entry:

Jun 4 2004 10:28AM

Victoria M. Bolotina

Do receptors activate distinct Ca2+ entry events from those stimulated by store-emptying? Are the Ca2+ entry events activated by receptors and emptied stores independent or integrated?

Here is another point of major confusion. Many do not discriminate receptor- and store-operated pathways simply because both are triggered by agonist-induced activation of the same receptors, and result in entry of the same Ca2+ ions into the cell.

I strongly believe receptor-operated and store-operated pathways are independent, and may be viewed as separate and very specific branches of a global receptor-triggered Ca2+ influx machinery. These pathways may haveone starting point (such as receptor activation), but definitely present separate signaling cascades that involve different reactions leading to activation of different ion channels. They can function independently of each other, but they certainly may be (and most probably are) integrated into more complex interactions. The nature of these interactions (their cross-talk) may depend on the physiological requirements of specific cell types. It may strongly depend on whether two signaling cascades arespatially co-localized, or separated into different structural/functional compartments (domains).

For example, G proteins, IP3, PIP2, DAG (and any other immediate products) may determine a short signaling pathway (co-localized with the receptor), which is designed to activate specific receptor-operated channels (many belonging to TRP family). This "short" pathway co-localized with the receptors could be called a receptor-operated pathway. But the same receptor may trigger another cascade of reactions: IP3 can travel to ER, and release Ca2+, producing store depletion (partial or global, localized of diffuse is still to be determined). This may be a crucial stimulus for production of another major signal (like CIF), which may be released in totally different cellular compartments that are restricted to another plasma membrane-delimited signaling domain, which we believe includes iPLA2 and store-operated channels.

There certainly may be other major components that eventually will be determined and placed in the same signaling cascade/domain. SOC may never see IP3 (or any other immediate products released in the vicinity of the receptors), but through a "long" and complex pathway (going through the stores), SOC activation will still physiologically depend on agonist-induced IP3 production.

For me, the term "receptor-operated" Ca2+ influx means that specific Ca2+-conducting channels (not SOC) are closely associated with the receptor, and are regulated by its immediate products.

The term "store-operated" points to the fact that the signal goes much further, and downstream there is yet another major triggering event (store-depletion) that is needed to activate a specific type of Ca2+ influx channels (store-operated channels, or SOC), which may be located far away from the receptors.

Why did nature create several different and independent cascades that may be triggered by the same receptor, but then diverge into different pathways, which eventually lead to the same event, Ca2+ influx? The answer is simple -- nature needs diversity and precision. We all know that Ca2+ signaling is not the same when calcium enters through different and highly specialized channels localized to different structural/functional domains. So, it is important to discriminate between store-operated and receptor-operated Ca2+ influx, (as well as many other type of global Ca2+ entry mechanism).

First, we need to get a clear picture of how each of them may work independently of all others. And only then we can ask the next question, how those pathways may talk to each other? This is the next level of system complexity, which is hard to address correctly without first obtaining a total clarity on how each individual pathway works. Right now,we are still at the level of arguing "whether store-operated and receptor-operated Ca2+ influx pathways are the same or different", and each of us has a different opinion. My opinion is only an invitation for the discussion in this Forum...

Distinctions between receptor-operated and store-operated Ca2+ entry:

Jun 11 2004 1:29PM

Kenneth L. Byron

I agree with Victoria that there are distinct channels that mediatereceptor-operated Ca2+ entry and store-operated (a.k.a. capacitative) Ca2+ entry. Several years ago, Colin Taylor and I provided evidence, usingfura-2 fluorescence techniques, that vasopressin receptor activation leads to divalent cation entry via a noncapacitative Ca2+ entry (NCCE) pathwayin A7r5 vascular smooth muscle cells (1). The NCCE pathway was apparentlyactivated in addition to capacitative entry (CCE) and had differentdivalent cation permeabilities than CCE. More recently, my lab has beenusing electrophysiological methods to examine cation currents in A7r5cells and our results (soon to be published, I hope) are consistent withthe fura-2 studies. We find that vasopressin activates at least threedistinct non-selective cation currents. One of these is a store-operatedcurrent, which can also be activated by passive store depletion (e.g. with thapsigargin or internal BAPTA). The other two currents are not activatedby store depletion, but robustly activated by vasopressin.

Victoria has already speculated on reasons for having such distinctmechanisms. One aspect that she didn’t elaborate on is the concentration-dependence for activation of these different pathways by agonists. Itseems likely that we will find that different Ca2+ entry pathways areactivated over different ranges of agonist concentration and may allow for a broader range of biological responses with varying agonistconcentrations. I have argued previously (2) that vasopressin elicits very different Ca2+ signals at low (picomolar) versus high (nanomolar)concentrations and that these signals may in turn trigger differentphysiological responses (Ca2+ spiking at low [vasopressin] may producerhythmic vasoconstriction, Ca2+ release and store-operated Ca2+ entry athigh [vasopressin] may be involved in mitogenic stimulation). It will bevery interesting to learn whether there are distinct structures within the cells that allow not only activation of these channels by distinctsignaling cascades, but may also direct the entering Ca2+ to particularsubcellular targets, thereby promoting a selective functional responsedepending on which complement of channels are activated at any givenconcentration of agonist.

1. Byron, K.L. and Taylor, C.W. (1995) Vasopressin stimulation ofCa2+ mobilization, two bivalent cation entry pathways, and Ca2+ efflux inA7r5 rat smooth muscle cells. J. Physiol. (London) 485: 455-468.
2. Byron, K.L. (1996) Vasopressin stimulates Ca2+ spiking activity in A7r5 vascular smooth muscle cells via activation of phospholipase A2. Circ.Res. 78: 813-820.

The role of intracellular Ca2+ release channels:

Jun 4 2004 10:32AM

Victoria M. Bolotina

Do InsP3 receptors or RyR, or both, mediate or modify Ca2+ entryevents?

Either of these two receptors may affect Ca2+ entry, but they do not seem to be required and/or sufficient for activation of store-operated calcium entry (SOC). From all the data accumulated so far, it looks like whatever you do with InsP3 receptors or Ryanodine receptors (RyR), you still need a stimulus (depletion of the stores) to activate SOC and Ca2+ influx. So, intracellular Ca2+ release channels may modulate (shape) capicitative calcium entry (CCE) response (still to be determined how), but they do not seem to trigger it.

If one will look at different schemes in the papers devoted to the direct coupling model -- it is amazing how this model has evolved from a simple "physical link of InsP3R with SOC" into increasingly complicated and rather confusing arrangements that are now needed to tighten the growing number of loose ends.

So, does such complexity meanthat physical coupling does not exist? Not at all! It can certainly exist, but most probably not in a way how it is presently viewed. One of theoretical possibilities is that conformational coupling may play a totally different role in the store-operated pathway -- for example, itmay be not a trigger, but a "positioning", or some other kind of tuning or accessory device. This is, of course, pure speculation at this point, which should be viewed as a simple invitation for re-interpretation of many valuable pieces of experimental data from some totally differentpoint of view, which may result in a new break through in this intriguing direction.

What are the mediators of the store-operated calcium influx pathway?

Jun 4 2004 10:52AM

Victoria M. Bolotina

I guess no one will be surprised if I will talk about the role of CIF, iPLA2 and lysophospholipids...but I am not going to repeat what we have recently described in our papers (Smani et al. 2003, 2004). What I want is to spark a discussion on what we presently know and what wedon't know (but definitely want to know) about this pathway.

So, how much do we really know? When you are able to put several simple steps into a cascade of reactions that may connect depletion of the stores with activation of plasma membrane channels, you think it is a lot. But nextyou start to ask specific questions about each individual step, and you realize how little we presently know, and how much more work (not by one, but by many labs) will be required to generate a complete detailed picture of this "simple" pathway. I will start with mentioning only few questions that we have about major mediators of the pathway:

- CIF identity remains a major issue; not existence, but identity. I am sure there always will be some one who will continue doubting CIF existence, simply because we do not know yet what it is. The good news is that now we at least know what it does to activate SOC, which is a major relief to all of us working with CIF model, and a great stimulus to keep us going in this direction.

- Ca2+-independent phospholipase A2 (iPLA2) appeared to be a perfect link that accommodates major requirements (and experimental observations) in store-operated pathway, but do we know everything about how it works? No, many important features of iPLA2 remain poorly understood. Forexample, how CaM binds to this protein, and what molecular interactions are crucial for its displacement? We have accumulating evidence that its binding and unbinding does not follow the same pattern, and is much more complex than in other CaM-binding systems. The good news are that thiscomplexity seems to meet the physiological requirements for activation of store-operated channels when CIF is present, and inactivation when CIF is gone, independently of whether Ca2+ is high or low (physiologically low,not 10mM BAPTA conditions). Further studies are needed to fully understand how this ON-OFF switch in store-operated pathway (and certainly its fine-tuning device) works in different cellular systems.

- Lysophospholipids emerged as a next important class of lipid second messengers. Which lysophospholipids are more effective and how they activate SOC? We don't know, and it will be great to find out. So far, we did not find any differences in LysoPC and LysoPI in activating store- operated channels (Ca2+ influx) in RBL cells (Ca2+ -SOC or CRAC-type) and in SMC (cat-SOC type). The most important fact is that it looks like both types of store-operated channels, SOC (Ca2+ selective SOC /CRAC andnonselective cation SOC, whether Reinhold likes this idea, or not) react identically on LysoPC and LysoPI. This further supports the idea that store-operated channels exist in different flavors with Ca2+ selectivitybeing the major (if not the only) variable, which is very important for adjustment of the pathway to the specific needs of different type of the cells.

- The most important characteristic that is common for all SOCs tested so far is that they are all operated via the same CIF -iPLA2-LysoPL pathway. The crucial role of iPLA2 in SOC activation has been already confirmed not only in RBL, Jurkat, SMC and platelets, but also in mostrecent studies by Prevarskaya et al (JBC, May 2004) in epithelial cells, and in several other pending studies that we will soon hear about. Isn't it exciting that there is a new simple pathway identified, which openstremendous possibilities for moving forward in many different directions? But is this the only pathway that can connect SOC with the stores? We don't know... and I am sure there will be many different opinions on this question, which certainly will be interesting to discuss in thisForum. The truth is waiting to be discovered.

It would be great to hear different opinions and certainly questions about what else we all want to know.

Processes that mediate activation of calcium entry channels

Jun 5 2004 6:44AM

James W Putney

What are the physiological transduction processes that mediate activation of Ca2+ entry channels?

In my opinion, regulated Ca2+ channels consist, as far as we know, of two basic types: Those which are gated by the binding of ligands, and those which are gated by physical forces.

The latter would include voltage-gated channels, stretch- or contact-activated channels (perhaps), and channels gated by changes in temperature. The ligand-gated channelsinclude those gated by actions of extracellular ligands or intracellular ligands. The intracellular ligands may be ions, small molecules or second messengers, or proteins or, more likely, specific sequences within proteins.

Although we do not know the mechanism of gating store-operated or capacitative calcium entry channels, it is presumed that they are ligand-gated in some manner, either by a diffusible mediator or by another cellular protein. The physiological mechanisms and contexts for activation of channels regulated by physical forces is straightforward, as is the case for those gated by extracellular ligands. The chairs of the conference have made it clear that the current discussion is to be focused on the mechanisms of activation of channels gated by intracellular ligands, whether store-operated or otherwise, the modes of interactions between these pathways, and the physiological relevance of these different mechanisms of channel regulation.

(a) The role of Ca2+ stores in activating entry:

Are changes in intracellular stores necessary or sufficient (or both) to activate Ca2+ entry? Are intracellular Ca2+ stores physiologically depleted under normal cellular conditions?

The easiest part of this question is: Are changes in intracellular stores sufficient to activate Ca2+ entry? The clear answer to this is yes, as demonstrated by any number of strategies, including discharge of stores with messengers such as IP3, passive depletion of stores with SERCA inhibitors like thapsigargin, or passive depletion of stores with Ca2+ chelators. This ability to activate entry channels simply by depleting Ca2+ from intracellular stores defines the store-operated or capacitative pathway. This mode of Ca2+ entry activation is the most pervasive mechanism known, and as far as we know is present in all eukaryotes (although not necessarily in every cell type in multicellular organisms).

Speculatively, this mechanism may have evolved to support the requirement for Ca2+ within the endoplasmic reticulum for normal protein synthesis, folding and trafficking, and may have been exploited later as a means of generating or supporting Ca2+ signals. Since we have not as yet identified these channels at the molecular level, it is more difficult to say whether or not depletion of stores is necessary. That is, we cannot definitively rule out activation by other signaling mechanisms of the channels which are apparently store-operated. In one recent report on smooth muscle cells, store-operated channels were identified which could be activated by protein kinase C in the absence of store depletion. A major difficulty isthat we do not know with certainty the downstream events following Ca2+ store depletion, and so we must entertain the possibility that these undisclosed steps could be activated or regulated by other means.

As to whether intracellular Ca2+ stores are physiologically depleted under normal cellular conditions depends upon one's definition of "normal" cellular conditions. Clearly, neurotransmitters that activate receptors coupled to phospholipase C in many instances activate store-operated Ca2+ entry mechanisms. In many experimental situations, supramaximal concentrations of agonists are used to generate maximal release of Ca2+, which in turn generates maximal activation of store-operated entry.Physiologically, it is unlikely that these saturating concentrations of neurotransmitters ever occur, and receptor signaling mechanisms probably operate in the linear range, below the binding Kd. Although this is thesubject of some debate, it does appear that in many cases a close almost stoichiometric relationship exists between the extent of Ca2+ release and the extent of activation of store-operated entry. Thus, it is likely that whenever sufficient IP3 is generated to cause significant release, no matter how small, a corresponding activation of entry will occur. Therefore, the question of the physiological relevance of the role of store-operated channels is tied in a way to the question of the physiological role of IP3-induced Ca2+ release, an issue which would engender far less debate among scientists in the Ca2+ signaling field.

An important case in point is the situation in which the [Ca2+]i in cells oscillates through a mechanism involving cyclical IP3-induced Ca2+ discharge. The case for the physiological relevance of Ca2+ oscillations to the fundamental processes of Ca2+ signaling in a variety of cell types is very strong. Because the duration of these release events is brief, and the extent of Ca2+ loss is small, it is difficult to demonstrate directly that store-operated entry accompanies Ca2+ oscillations. Yet the oscillations quickly run down in the absence of extracellular Ca2+, indicating a need for Ca2+ from theoutside for maintenance of the releasable stores. In instances in which this run down process was compared to maneuvers that inhibited store-operated entry, it appeared that indeed it was the store-operated entry of Ca2+, which served to support the pools of Ca2+ required for oscillations.

(b) Distinctions between receptor-operated and store-operated Ca2+ entry:

Do receptors activate distinct Ca2+ entry events from those stimulated by store-emptying? Are the Ca2+ entry events activated by receptors and emptied stores independent or integrated?

The arguments in the preceding section are not intended to imply that every phospholipase C-linked receptor activates Ca2+ entry solely by the store-operated pathway. A case in point is the TRPC channels which, atleast when ectopically expressed in cell lines, can be activated in a phospholipase C-dependent manner often independently of store depletion. And in such instances, the channels can be activated by very low concentrations of agonists that do not appear to cause significant Ca2+release.

There are clear cases in the literature of specific cell lines that utilize endogenous non-store-operated pathways for Ca2+ entry, and, in many instances, these pathways have properties similar to those which we expect from TRPC channels. A strong case has be made for such a signaling mechanism in smooth muscle cells. On the other hand, a number of studies have indicated that in T-lymphocytes the store-operated pathway constitutes the major, if not exclusive, route for activated Ca2+ entrythat is essential for T-cell activation. The extent to which different mechanisms account for the physiological regulation for Ca2+ entry in other cell types has not been extensively investigated, but such studies are clearly needed.

Are the two mechanisms independent or integrated? Little work has focused on this issue, but in a few studies a somewhat reciprocal relationship between the two pathways was demonstrated. That is, there appeared to be mechanisms, not always clearly defined, through which the activation of one pathway inhibited concomitant activation of the alternative pathway.

(c) The role of intracellular Ca2+ release channels:

Do InsP3 receptors or ryanodine receptors, or both, mediate or modify Ca2+ entry events?

I can only give a rather short answer to this question. There is fragmentary evidence in the literature for regulation of store-operated channels by IP3 receptors or ryanodine receptors, and at least one report of the regulation of non-store-operated channels by IP3 receptors. However, some of these findings have not be readily reproduced in other laboratories. I think it is fair to say that a consistent pattern of behavior or mechanism has not developed as yet. Additional studies areneeded to assess the roles of intracellular release channels in various Ca2+ entry mechanisms.

(d) The function of mediators and adaptors in Ca2+ entry:

What role is played by structural or adaptor proteins (including Homer, NHERF, InsP3Rs, and PLC-gamma), or by soluble cellular messengers (such as InsP3, cyclic nucleotides) and lipid messengers (including diacylglycerol, arachidonic acid, and PIP2)?

This question seems to be an expanded version of the one above. Again, some very interesting studies have been published implicating each and every one of the listed regulatory substances, but in most cases a well-established and general paradigm of mechanism has not beenestablished. One might also include the elusive calcium influx factor and lysolipids in the list of putative mediators for which there is encouraging, but as yet limited, published support. Exceptions would be thecyclic nucleotides and diacylglycerol, and possibly arachidonic acid, for which a growing body of evidence supports their roles as regulators of specific classes of Ca2+-permeable channels.

Gating of calcium entry channels

Jun 16 2004 9:34AM

Donna L Cioffi

I have a couple of questions for Jim Putney regarding the gating ofSOCE channels.

Do you envisage a single gating mechanism for eachdifferent type of SOCE channel, perhaps in a cell-type specific manner?

And are there certain criteria you feel should be met in order todemonstrate that something (protein, peptide, ion, etc) is indeed part ofthe gating mechanism.....as opposed to simply modulating channel function?

Gating of calcium entry channels

Jun 16 2004 9:45AM

Jim Putney

These are important questions, without easy answers. Since we do notas yet have any one gating mechanism that is proven and wellcharacterized, we do not have criteria for evaluating one or anothermechanism in any given cell type.

There is certainly considerable circumstantial evidence for each of the two general mechanisms: diffusible signal, and conformational coupling. In fact, I have often stated that byselective choice of literature references, a rather solid case can be made for either one.

It is possible that this reflects differentmechanisms indifferent cell types, or for different channel types. A case in point isthe mechanism of interaction between plasma membrane Ca2+ channels andintracellular ryanodine release channels in two cell types, skeletalmuscle and cardiac muscle. In one case the interaction is clearly a direct one, in the other, clearly the interaction involves a diffusible signal,in this case Ca2+ itself.

As to what the criteria are for demonstrating thatsomething is partof the gating mechanism, I think these are not different from thosealready discussed for identifying the channels. One would like to identify or speculate on candidates by whatever basis, and investigate theirinvolvement by a combination of positive (add it, see if it works) andnegative (knock it out, or inhibit it) approaches. There has been someencouraging work along these lines for the case of the CIF - PLA2 (calcium inlux factor- phospholipase A2) hypothesis, already discussed in this forum.

A final criterion Iwouldinvoke for both the gating mechanism question and the channel question isthat we need to eventually achieve a consensus of experimentalobservations and interpretation among a number of different laboratories.

Processes that mediate activation of calcium entry channels

Jun 5 2004 7:00AM

Donald Gill

The terms "store-operated channel" and "capacitative Ca2+ entry" are becoming increasingly familiar in the literature and are finding their way into the vocabulary of an ever greater diversity of papers. Thus, investigators in quite different fields may feel confidentthat Ca2+ signals in their particular cell types are mediated by the activation of store-operated channels. However, those of us studying this process are becoming increasingly unsure about the basis of store-operated Ca2+ entry, that is, both the signals that generate such entry and the channels that mediate it. It was this disparity between the increased acceptance of the concept of store-operated channels and the decreased understanding of the function of store-operated channels that promptedRanden and I to instigate the E-Conference. This was to allow experts and other interested investigators to air their views in an uninhibited yet highly visible forum.

A major focus of this E-Conference is to highlight the distinction between physiological receptor-activated Ca2+ signals and what we frequently study as "store-operated" Ca2+ signals. The latter are generated mostly by a nonphysiological emptying of stores using ionophores, SERCA blockers, or infusion of InsP3 into the cells. Increasingly, we learn that the substantial depletion of stores may be a nonphysiological event. An exception to this may be small blood cells (such as lymphocytes) with large nuclei and limited cytoplasm, in which activation of PLC-coupled receptors may cause a significant decrease inluminal Ca2+. In most other cell types in which the endoplasmic reticular network is extensive and pervasive, the capacity of Ca2+ within the lumen is great and luminal depletion in response to physiological agonists maybe very small.

There are a number of known channel proteins that have been suggested to function in a "store-operated" mode. Notable among these candidates are the TRPC channels denoted as being the mammalian TRP channels most similar to the original TRP channels that mediate Drosophila phototransduction. There seems little question that the TRPC channels are receptor-operated. Indeed, the rhodopsin/G protein/PLC photosignaling system of fly retinas is a prototype for such signaling. At least threeclosely resembling members of the mammalian TRPC family (TRPC3, TRPC6, and TRPC7) can be activated by the PLC product, diacylglycerol (DG) and it would appear that this PLC product could explain activation of these channels.

However, in fly retinas, the TRP channels present do not appear to be activated by DG. Moreover, other mammalian TRPC channels (including TRPC1, TRPC4, and TRPC5) are not activated by DG. Yet, somewhat enigmatically, when expressed in cells, all of these TRPC channels can beactivated by PLC-coupled receptors in what appears an almost identical fashion to the TRCP3/6/7 family.

Strangely, we find that in A7r5 smooth muscle cells endogenously expressing high levels of TRPC6 channels, there is little activation of these channels upon addition of DG. In contrast, in the same cells, exogenously overexpressed TRPC6 is activated by DG. Are we therefore to believe that physiologically the TRPC family ofchannels respond to a common alternative PLC-derived signal? One such signal could be a decrease in PIP2, but, while this is an attractive hypothesis, there is no direct evidence for this action at present.

We are starting to believe that the transduction mechanism for store-operated channels is perhaps a more "general" phenomenon rather than a highly specific and directed signal. Whereas the conformational couplingmodel has been an attractive scheme and has derived some support from the possible role of InsP3Rs, we consider that the analogy of this model with the functioning of the triad junction in skeletal muscle is somewhat flawed. Thus, many believe the simple direct interaction between a Ca2+ entry channel and a Ca2+ release channel may be a universal paradigm for Ca2+ signals in all cells.

In reality, however, the triad junction in muscle serves as an exceedingly rapid means for sensing and transducingthe electrical signal in the plasma membrane into a Ca2+ release signal from the SR. In contrast, the store-operated signal takes many seconds or even minutes to develop and is, hence, unlikely to require a similar rapidand direct coupling process.

The role of the InsP3R in mediating the store-depletion signal to activate store-operated channels has been a popular area of investigation and has received considerable experimental support. Nevertheless, the complete elimination of InsP3Rs throughknockout procedures results in little influence on the activation of store-operated channels. Although apparently not necessary for the coupling process, it does appear that InsP3Rs may influence coupling and hence may have a regulatory rather than obligatory role on the activation of SOCs.

Whether direct conformational coupling is involved or not, our own experiments and those of others indicate that it is likely that close interactions between the plasma membrane and ER membrane are required for activation of the "store-operated" coupling process. Thus, we consider it likely that there is a spatially directed "area ofinfluence" between of the plasma membrane and the ER. This may be through areas of close association between the ER and plasma membrane (PM), which have been observed to exist in many cell types. However, these "junctional" structures have proven extremely elusive. In most cells they are sparse at best and there is no direct evidence that they are the sites at which Ca2+ entry signals are activated.

Given that TRPC channels are interesting candidates for the process of receptor-induced signals and given the numerous reports suggesting that these channels are influenced by store-emptying, it appears that they mayat least be components of store-operated Ca2+ signals in some cells. However, we are developing a somewhat different perspective on the whole process.

We are realizing that TRPC channels (as well as other TRPchannel members) are channels receiving multiple inputs. We think of them as "integrators" of a number of signals, their opening reflecting the summation of multiple distinct signals. Thus, the TRPC channels can be modified by diacylglycerol (and perhaps unsaturated fatty acids), by InsP3 receptors (and perhaps ryanodine receptors), by adaptor proteins such as homer, NHERF and PLC-gamma, by protein kinase C, and by Ca2+ itself. By analogy with the TRP channels of fly retina, it would appear that themammalian TRPC channels function within domains containing numerous structural, adaptor and modifier proteins, as well as the receptor/G protein/PLC proteins transducing the agonist-induced signals.

We would like to think that this structural domain is also a site "under theinfluence" of the filling status of Ca2+ stores. This would be one further input. Under conditions in which stores are totally emptied, albeit an unphysiological circumstance, then the store-derived signal may become a dominant one. Under conditions in which physiologicallyactivated receptors are inducing store-release, small changes in store-Ca2+ may be sufficient to subtly modify the TRPC channels in a way in which larger sustained entry of Ca2+ may occur, sufficient, for example, to increase the frequency of Ca2+ oscillations.

This being said, we really do not know what this store-derived signal is. If areas of close association are the communicating sites betweens stores and PM, then it is possible that locally released factors may be mediating the communication process. In this case, we may be hard-pressed to identify and measure such mediators. However, by carefully dissecting the "tool-box" of TRPC modifiers and by changing the function of these modifiers by a combination of pharmacology and knock-down strategies, wemay be able to begin to pinpoint the nature of this elusive coupling mechanism.

Comment on processes that mediate activation of calcium entry channels

Jun 8 2004 2:41PM

Kenneth L. Byron

I’d like to respond briefly to comments made by Don and by JimPutney.

Don writes: “Whereas the conformational coupling model has been an attractive scheme and has derived some support from the possible role of InsP3Rs, weconsider that the analogy of this model with the functioning of the triadjunction in skeletal muscle is somewhat flawed… In reality, however, the triad junction in muscle serves as an exceedingly rapid means for sensingand transducing the electrical signal in the plasma membrane into a Ca2+release signal from the SR. In contrast, the store-operated signal takesmany seconds or even minutes to develop and is, hence, unlikely to require a similar rapid and direct coupling process.”

I’m not sure what “store-operated signal” Don refers to or in factwhat we ought to monitor to determine the kinetics of store-operatedsignaling. The Ca2+ overshoot measured in fura-2 fluorescence studies mayreflect a non-physiological spillover of Ca2+ that either cannot be takenup into the stores because we’ve inhibited the SERCA pumps or thatexceeds normal physiological concentrations because we’ve usedsupraphysiological concentrations of agonists. Under physiologicalconditions (low agonist concentrations) perhaps, as postulated by Jim,even a small release of Ca2+ is sufficient to rapidly trigger a small(undetectable by conventional methods) Ca2+ entry to refill the ER Ca2+stores.

I think it’s possible that in oscillatory Ca2+ responses thecoupling between stores and plasma membrane may be very rapid, perhapsapproaching the muscle triadic junction model. There may be some built-indifference in kinetics that is unrelated to conformational coupling perse, which relates to the fact that the signal goes from inside out ratherthan outside in. There may be a temporal lag for example if theconcentration of Ca2+ inside the stores must be reduced to some thresholdlevel before the store-operated signal is activated. Electrophysiological measurements of store-operated currents are generally conducted under even more non-physiological conditions, so the kinetics of their activation may not reflect the normal coupling between storedepletion and Ca2+ entry.

With regard to the statement “We are realizing that TRPC channels (aswell as other TRP channel members) are channels receiving multiple inputs. We think of them as "integrators" of a number of signals, their openingreflecting the summation of multiple distinct signals.”, I don’tparticularly like the notion that individual TRPC channels integratesignals from so many different areas. I don’t think there is sufficientexperimental evidence to support either case, but I would favor the notion that distinct channel assemblies can respond to localized receptoractivation, while others are positioned (and probably comprised ofdifferent auxiliary proteins or have different TRP channel homologuestoichiometry) to respond to store depletion or other signals.

Session 2: Cellular Domains that Contribute to the Calcium Entry Events

May 24 2004 1:22PM

E-Conference Chairpersons

This session is focused on the main question of what is known aboutthe cellular domains contributing to the transduction of signals intocalcium entry events.

Specific issues include

(a) The role of specific organelles:

Do endoplasmic reticulum, mitochondria, or other cellular organelleshave direct or indirect input into the activation of plasma membranecalcium entry channels?

(b) Local signaling domains:

Do localized domains (lipid rafts or organized signaling complexes)define operational coupling domains required for the activation of calcium entry?

(c) The role of trafficking events:

Do dynamic changes in the translocation or organization of membranesor the proteins within them contribute to control of calcium entrysignals?

(d) The influence of local calcium signals:

Are diffusionally restricted calcium release or entry eventsimportant in the activation or deactivation of calcium entry channels?

Agonist-mediated calcium entry - contiguous ER or ER resident proteins?

Jun 1 2004 12:05PM

Randen L Patterson

A role for the ER in the activation and modulation of receptoractivated (G-protein coupled or tyrosine kinase coupled) calcium entry atthe plasma membrane is questionable; however a role for resident proteinsof the ER seems indisputable. In particular, the number of papers demonstrating a physical and functional interaction of the IP3-receptorwith plasma membrane ion channels, in particular TRPC channels, is growing rapidly. A quick summary of this work details a direct physicalinteraction of the IP3-recpetor with TRPC channels, and that thisinteraction is capable of gating these channels in the presence of IP3 inresponse to plasma-membrane receptor stimulation. The undefined variablein this equation is whether these IP3-receptors are part of the contiguous ER, or are they in other membrane pools, perhaps even the plasma membrane.

A reasonable hypothesis, building from the model proposed in the 1995review by Michael Berridge, is that IP3-recpetors coupling to calciumentry would be in a “release-dead” conformation, or alternatively on membranes/vesicles which do not contain high levels of calcium. Thismodel is supported by the work of many, including us, showing that the N-terminusof the IP3-receptor is capable of gating calcium entry independent of itscalcium release functions. Additionally, as calcium entry channels tendto be inhibited by calcium, release of calcium from IP3-receptors coupledto this process would likely be counter-productive.

These same statements would hold true for a role of the ryanodine receptor coupling to calciumentry channels, although the data supporting this idea is tenuous at best.

A role for SERCA, another major calcium handling component of the ER, isalso under-defined in receptor-activated calcium entry. A few manuscripts have demonstrated changes increases in SERCA activity following agoniststimulation, but a molecular mechanism has yet to be elucidated. The onlyhypothesis that can be drawn currently is that SERCA activity is used inconjunction with the plasma-membrane calcium ATPase, sodium/calciumexchanger, mitochondria, and other calcium storage units to tightlyregulate intracellular calcium levels during stimulation, with theorchestration of these many components allowing for multiple cellularoutputs (growth, gene transcription, hormone secretion etc.).

Store-operated Calcium Entry- Contiguous ER or ER resident proteins?

Jun 1 2004 12:15PM

Randen L. Patterson

A role for the contiguous ER in the activation of store-operatedcalcium (SOC) entry at a glance would seem undeniable, although molecularevidence defending this hypothesis is minimal. In general, SOC isactivated by the application of:

1) pathophysiological plant defense toxins (thapsigargin) that inhibit SERCA, thus passively depleting calcium storagepools, or

2) application of ionophores to release calcium from internalmembrane compartments.

Either of these strategies would deplete anycalcium pool that was actively maintained; therefore these studies do not provide conclusive evidence for the role of the contiguous ER in SOC.

Both of these methodologies have been examined using knock-down or knock- out strategies to define molecular constituents of SOC, in particular theIP3-receptor and SERCA. A “minimal at best” alteration of SOC has been observed under these conditions, ruling out a requirement for theseproteins in the activation or maintenance mechanisms of store-operatedcalcium entry.

The strongest data supporting a role of the contiguous ERin SOC comes from two recent studies looking at the sarcoplasmicreticulum, not the ER, in skeletal muscle. These studies demonstrate afunctional link between the calcium storage protein calsequestrin and the synaptophysin-family-related protein Mitsugumin 29 (located in thejunction between the PM and SR of skeletal muscle), with both influencing store-operatedcalcium entry. Unfortunately, other studies looking at reduction ofcalreticulin in hepatocytes did not support this hypothesis. Takentogether, although pathophysiological depletion of internal calcium stores absolutely activates SOC, the calcium pool(s) responsible has not beendefined.

Twenty years of studying the effects of thapsigargin andiononphores to elicit SOC has not provided any concrete evidence as to the molecular entities in the ER, or anywhere for that matter, responsible for this phenomenon. More rigorous molecular approaches (gene-chip of ERdepleted cells vs. control, sequential siRNA deletion of calcium relatedER proteins, real-time chromophore inactivation of ER resident proteinsetc.) using physiological methods for depleting calcium pools (chronicagonist stimulation) are likely to provide molecular entities that can be identified, modified, and tested for a definitive role is SOC.

Lipid Rafts- Hotspots for agonist mediated calcium entry?

Jun 2 2004 10:02AM

Randen L. Patterson

There are more than 3000 G-protein and receptor tyrosine kinasecoupled receptors encoded by the human genome. Although not all of thesereceptors are encoded by all cells, each cell does contain numerousreceptors, even specialized cells such as B and T cells. This observation alone predicts that micro-domains of calcium entry must exist. To obtainspecificity of receptor responses for cellular functions, it is logicalthat the cell must segregate the receptor response into discreet areas, to grant the receptor a specific response.

Evidence to support this hypothesis is rapidly increasing. Some ofthe most convincing work comes from Delmas et al (Neuron 2002), using TRPC channels as biosensors to test muscarinic receptors and bradykininreceptors. Not surprisingly, each receptor demonstrated a functionallydistinct output. We also observed functionally distinct receptor specific calcium entry responses when studying RACK1 interactions with the IP3- receptor. The work from other laboratories, in particular Ambudkar’sgroup at NIH, provide evidence for spatially restricted expression of TRPC channels into lipid raft/caveolar/actin-rich domains which contain many of the proteins implicated in agonist-mediated calcium entry (PLC, SERCA, IP3 -receptors, calmodulin, heterotrimeric G-proteins, scaffolding proteinsetc.).

I think the question is not whether calcium entry occurs within micro -domains, rather how do we investigate these domains? How do we determine the constituents of these regions within the cell? Although lipid raftsare still loosely defined, biochemical methods have been developed toisolate and enrich these areas for study. One of the most important areas of study which has been largely overlooked in the calcium entry field IStrying to determine the molecular environment around TRP channels. Whatare the other scaffolding proteins tethering these channels? Which othereffecter molecules are binding to this complex? Biochemical isolation ofTRP channel complexes from discreet regions of the cell, coupled to massspectrometry sequencing would provide invaluable information as to themolecular environment in which these channels exist.

Why would TRP channels localize to these regions? Lipids, inparticular PIP2 whose levels can be rapidly regulated via receptorstimulation, regulate the activity of a wide variety of channels andexchangers in the plasma membrane, from Na/H exchangers to potassiumchannels (which are found in lipid rafts). Lipids have been demonstratedto modify the function of TRPC 3, 6, 7, in addition to the TRPV and TRPMfamilies. It seems reasonable these channels would localize to areas where lipid organization could be tightly controlled. TRPV and TRPM familyproteins can be regulated by PIP2, but the TRPC channels regulation bylipid is still a mystery. TRPC 3, 6, 7 can be activated by theapplication of exogenous DAG analogues in overexpression systems, although endogenous calcium channel activity observed by the application ofexogenous DAG analogues has been observed in only a few reports. Asdirect lipid binding to any of these channels has yet to be demonstrated,this seems a reasonable place to begin addressing this question. Determination of the lipids for which these channels have affinity willprovide information not only about their activity, but likely theirtrafficking as well.

All of this is not to say that functional TRP channels do not existoutside of lipid rafts, but lipid rafts provide a functional micro-domainwhich can be focused on, and then expanded to other regions of the plasmamembrane.

What are lipid rafts?

Jun 7 2004 8:35AM

Peter J Lockyer

Define this first Randen?

Are we talking the artifacts of cold detergent preps and cholesteroldepletion, or the ms timescale clustering of lipid microdomains?

The rafting landscape is being challenged and is changing rapidly.The idea of stable lipid domains on the time-scale of SOCE is not veryconsistent with results from clever fluorescence-based assays of lipidmicrodomains.

They exist, and at least for GPI-anchored reporters, return to thesame 'site' in the membrane. But these are remarkably transient events and the biochemical methods have been proved flawed.

What are lipid rafts?

Jun 7 2004 11:00AM

Randen L Patterson

A good comment. Yes, I was thinking of cold-detergent preps etc, which can be used to isolate what WERE traditionally thought of as lipid rafts. Perhaps a better way of describing these regions to which I am referring is as "cytoskeletal rich, detergent insoluble fractions", atleast from a biochemical aspect.

From a physiological aspect, it seems that lipids-rich regions would have to be in the vicinity of calcium entry channels, not only for receptor signaling in local domains, but likely for channel regulation as well. Although speculative, the lipids demonstrated to be involved inthese processes(PIP2, PIP3, PA, etc.) are also found in high levels in "lipid rafts"... Might they be related?

I would argue one point:
Although lipid domains change on a ms timescale, what is to say that the lipid domain need be present for theentire length of a signal for a long-term signal to be generated. Couldn't the transient "passing-by" of a lipid-rich domain provide enough substrate for the process, until the "passing-by" of the next lipid rich region?

What are lipid rafts?

Jun 17 2004 9:35AM

Indu S. Ambudkar

Lipid rafts:

The idea that lipid rafts are involved in Ca2+ signaling is veryappealing and in view of the biochemical data that have been reported itappears that signaling and Ca2+ entry can in fact occur in suchmicrodomains.

However, the methods that have been used to assess protein- association with lipid-raft domains are by no means rigorous. Thetechniques and reagents used for these studies are fairly non-specific and unless carefully controlled can lead to artifacts.

One such example isthe detergent solubilization protocol. Insolubility in Triton-X-100 is not always an indicator of LRD association. Just as detection at the bottomof the gradient does not mean that it is not associated with LRD. Themost simplistic case is a protein that is somehow anchored to thecytoskeleton. This interaction will render it insoluble in detergent andcause it to sink to the more dense fractions of the gradient. That itself does not exclude LRD association unless one disrupts the interaction withthe cytoskeleton and demonstrates that it either floats to the top orremains in the lower fractions.

Very few people try to use otherdetergents that are known to disrupt LRDs to show differentialsolubilization. These experiments are tedious and require meticulousdetail. Further, one has to discriminate between lipid rafts in theplasma membrane versus those in the trafficking membranes or golgi.

Regarding the very dynamic nature of the lipids in the rafts, yesthis is true. But, there are the relatively “stable” LRDs, such ascaveolar lipid raft domains. Another way could be a scaffold that isclosely associated with the LRD which can hold the proteins in place.

Thus, although the lipids themselves might be dynamic, the proteins(specifically polytopic plasma membrane proteins) might actually be morestably anchored. The question of course is does the dynamic lipidenvironment impact on the physiological function of the protein.

Anotherquestion I have asked the LRD aficionados is whether such proteins moveinto LRDs or whether the lipid around the protein changes and in doing sobrings other regulatory proteins (e.g. GPI-anchored) or lipids that areassociated with the LRD to the protein.

GTP-gamma(S), small G-proteins, and SOC- Any connection?

Jun 3 2004 7:47AM

Randen L. Patterson

One of most fascinating observations during my career as a graduatestudent learning about calcium entry was that GTP-gamma(S) was capable ofinhibiting SOC. I was amazed by this, because at the time (1997) therewere very few things that were capable of inhibiting SOC (which is stilltrue today). Obviously, GTP-gamma(S) inhibits essentially all G-proteinsat some level, so this is equivalent to hitting the cell with asledgehammer. BUT, G-proteins are also the major component of mosttrafficking systems. It is surprising that this line of research has notprogressed significantly since 1997. Small G-proteins such as Rac, Rho,and Cdc42 are all directly linked to intracellular trafficking, as well as cytoskeletal remodeling. Cytoskeletal remodeling has also been effectivein disrupting SOC, as reported by numerous laboratories.

This begs the question, are small G-proteins an avenue for mappingthe activation of SOC?

Yao et al (Cell 1999) demonstrated that SOC couldbe inhibited by constitutively active Rho, and potentiated when Rho wasinhibited. Consistent with this idea, Ito et al (Am J Respir Cell MolBiol. 2002) and Ghisdal et al (J. Physiol 2003) performed studies whichsupported these findings. They demonstrated that noradreniline activation of agonist-induced calcium entry required rho-dependent kinase activation, but SOC did not. Furthermore, TRPC1 association with the IP3-receptor has been demonstrated to be regulated by rhoA in endothelial cells (Mehta etal, JBC 2003).

From this data, it seems reasonable that Rho signalingpathways are a key regulator of calcium entry, and perhaps are adiscriminating factor between agonist-induced calcium entry and calciumentry induced by pathophysiological store depletion.

Is Rho activity down -regulated in response to calcium store depletion? If so, could this beused as a way of increasing the protein(s) responsible for SOC in theplasma membrane, which might be identified using S-35 methionine pulse- chase experiments, or some other genomic approach? For example, the TG2DDT cell line which was developed to be resistant to thapsigargin, and has constitutive SOC. Does this cell line have depressed rho activity versuscontrol?

It is also surprising that the TG2 DDT cell line has never been analyzed by microarrayversus the control cell line. This might provide candidate genes whichcould be tested for SOC activity.

Finally, could a trafficking eventindeed be why SOC is slow to activate after calcium store depletion? Further investigation of the Rho pathway will likely shed new light on the activation mechanism of not only SOC, but agonist-mediated calcium entryas well.

GTP-gamma(S), small G-proteins, and SOC- Any connection?

Jun 7 2004 6:50AM

Peter J Lockyer

Considering Rho is fine, but I would suggest there are a couple of other more interesting groups of small GTPases.

I would say, having recently spoken to Alan Dawson who first discovered the GTP effect (reported in FEBS 1985), that much is probably due to the regulation of ER integrity and continuity, which would include cytoskeletal remodelling via Rho, but may be more pertinent to Rabs andArfs. Alan would be a much better source of information than me, and so would Jim and Gary Bird.

Using RNAi would seem a simple method to start having a look, but imagine the number of indirect effects this would produce on agonist-induced calcium entry, e.g. receptor numbers, receptor trafficking, phosphoinositide levels, you name it really.

Ras is very often upstream of Rho family member activation and there are very specific influences of oncogenic Ras on bradykinin-induced calcium entry - all in transformed NIH3T3 cells first reported by Hans H Grunicke(and later others). From my perspective, Ras is one of the most interesting small GTPases to look at because there are 2 specific calcium sensors (CAPRI and RASAL) that regulate Ras activity. There is nothing quite so specific and dynamic for any other small GTPase.

I would say that yes, small GTPases must be involved in the regulated of SOCE. At multiple levels, probably mainly indirect. So the question is how to delineate all the indirect effects, and whether this is possibleconsidering the number of small GTPases that could be involved; upstream and downstream of each other.

BTW, I haven't noticed obvious effects of CAPRI on calcium signalling, so this is a neat tool but with negative results (for Ras).

Calcium entry regulation of calcium release

Jun 4 2004 3:48PM

Randen L. Patterson

The importance of calcium entry in regulation and control of calciumoscillations has been one of considerable debate. Many groups havedemonstrated that modulation of calcium entry can change the frequency and amplitude of agonist-mediated calcium oscillations, which are key forprocesses such as gene expression.

These experiments support the existence of functional IP3 receptors in extremely close opposition to the plasmamembrane, generating CICR, which would allow these local signals togenerate and control global signals, a reasonable argument.

The opposition to these observations is from experiments demonstratingincreased agonist concentration can ablate the need for external calciumentry. Although for many years I myself have used the removal of external calcium during experiments to separate calcium release from calcium entry, this experimental paradigm does not model physiological conditions.

Cells are never exposed to low, or zero calcium. Therefore, althoughoscillations can occur in the absence of calcium entry, this does nothappen in physiology. Furthermore, the removal of external calcium, although commented on little in the literature, can actually obscure the interplay between calciumrelease and entry.

During the last 3 years, I have been examiningproteins from a yeast-2-hybrid screen with the IP3-receptor for theirfunctional effects on calcium release and entry. I used the experimentalparadigm of overexpressing these proteins in PC12 cells, and then testingfor a functional output by first adding agonist in the absence of external calcium to observe calcium release, and then adding back calcium to lookfor changes in calcium entry. What I failed to realize until recently, is that any protein that was modulating calcium entry "coupled" tocalcium release or visa versa would not be represented.

A thirdexperiment must be added to this paradigm for it to work, which is merelyadding agonist in the presence of calcium to observe the integratedcalcium signal.

I think it is safe to say, we have all seen phenomena that occur onlywhen calcium is removed from the external media, or performed experimentsthat were sensitive to an order of operation. These observations shouldnot be removed from manuscripts, rather they should be highlighted. Weknow so little about the mechanisms coordinating calcium release and entry events, any reproducible data is important to the community fordeciphering these complex processes, and should be published.

Comments on cellular domains that contribute to calcium entry events

Jun 4 2004 11:00AM

Michael Berridge

Comments on Question 2: What is known about the cellular domains contributing to the transduction of signals into Ca2+ entry events?

I shall comment on this question by re-interpreting my earlier conformational coupling hypothesis (Berridge 1995 Biochem. J. 312:1) in the light of new developments. The basic idea, as originally proposed by Robin Irvine, is that stimulation of the IP3 receptor (IP3R) induces a conformation change that is transferred to the channel in the plasma membrane through a direct protein-protein interaction (Irvine 1990 FEBS Letters263:5). The entry channels have yet to be defined and current contenders are ICRAC and some of the Trp channels. Just how this coupling complex is activated is a critical issue.

In the hypothesis I developed earlier, I proposed that the IP3R could be activated either by IP3 or by store depletion and both might be operating. Strong evidence for the former was provided recently in a study on DT40 cells that lack IP3Rs and fail to induce entry or release following agonist stimulation. However, Ca2+ entry was restored upon expression of an IP3R even when it had a C-terminal truncation preventing it from releasing Ca2+ (van Rossum et al 2004 PNAS 101: 2323). The IP3R was able to respond to IP3 and to induce the conformational change necessary to promote an entry of Ca2+ even though it was unable to release Ca2+ from the store. This observation is in line with another prediction of the original model that suggested that when the IP3R is coupled to the entry channel it is non-conducting and thus not able to release Ca2+.

It is the operation of this conformational coupling mechanism at physiological agonist concentrations that will be a primary focus of the comments outlined below that will be based on a series of questions:

  1. What are the Ca2+ entry signals in primary cells and does conformational coupling play any role as an entry mechanism?

Studies on cloned cell lines have identified a number of putative entry signals such as DAG, IP3, arachidonic acid, store emptying etc, but what is the evidence that any of these actually play a role in primary cells? Studies on various primary cells certainly indicate that agonist can induce Ca2+ entry and in some cases this has been linked to Trp channels, but the precise nature of the coupling mechanism is largely obscure. There are an increasing number of examples of agonist-induced entry:

  1. It Mast cells, receptor activation stimulates a Ca2+ release activated current (ICRAC) (Hoth & Penner 1992 Nature 355: 353).
  2. In Purkinje neurons, metabotropic glutamatergic receptors activate Ca2+ entry through TRPC1 channels apparently independently of IP3 (Kim et al 2003 Nature 426:285).
  3. In thalamic interneurons, 5-HT2 receptors activate TRPC4 receptors to induce Ca2+ entry that triggers the dendritic release of GABA (Munsch et al 2003 PNAS 100:16065) .
  4. Fertilization of mammalian oocytes results in the activation of Ca2+ entry responsible for driving Ca2+ oscillations (McGuiness et al 1996 Development 122:2199).
  5. In pontine neurons, BDNF stimulates Ca2+ entry through TRPC3 channels (Li et al 1999 Neuron 24:261).
  6. Information on additional examples of Ca2+ entry mechanisms in primary cells would be welcome.

It is clear from the above that primary cells may employ a number of different mechanisms for coupling receptor activation to Ca2+ entry that can occur independently of store depletion. However, as indicated in the conference perspective, there is a prevalent assumption that "Receptor-induced Ca2+ signals involve two closely coupled events: Ca2+ release from ER stores, and Ca2+ entry across the plasma membrane." The way in which release might be coupled to entry is through activation of store-operated channels (SOCs) as first proposed by Putney (1986 Cell Calcium 7:1). There is no doubt that such a store-operated mechanism is unveiled when stores are depleted using thapsigargin or high doses of agonist, but is there any evidence that such a mechanism exists following agonist stimulation at physiological doses?

Is it correct to assume that these two processes are always coupled or is it possible for agonists to activate entry independently of release? The answer to this question is probably yes and the model outlined earlier suggests one possible mechanism i.e. the conformational change in the IP3R responsible for stimulating the entry channel can be activated either by IP3 or by store depletion. While both mechanism can activate entry, it is possible that their primary functions are different: the direct IP3 activation pathway may be dedicated to coupling receptor activation to entry whereas the store-operated pathway may be a homeostatic mechanism that ensures the internal stores remain filled. Having such a dual activation mechanism may also explain how entry can be activated by receptors that generate IP3 or by mechanism that deplete stores independently of IP3 as occurs following activation of ryanodine receptors.

To fully understand how entry is regulated it is necessary to consider the critical question of the sensitivity of these entry mechanisms and how they are activated at physiological agonist concentrations.

  1. What is the dose response relationship for the activation of calcium entry signals relative to the downstream physiological responses?

Dose-response relationships have been largely ignored even though they are crucial in order to understand the physiological mechanisms of Ca2+ entry. I am not aware of any dose-response curves for the sequential events of agonist-induced IP3 formation, Ca2+ entry and subsequent downstream cellular responses. I attempted to do this with the insect salivary gland by measuring the EC50 values for 5-HT induced fluid secretion (3x10-9M), the transepithelial flux of Ca2+ as an indirect measure of Ca2+ entry (1.5 x 10-8 M) and inositol efflux to monitor IP3 formation (5x10-7 M) (Fain and Berridge 1979 Biochem. J. 178:45). The curves for Ca2+ entry and IP3 formation were displaced to the right of the physiological response such that a 50% activation of secretion was seen at a 15% activation of entry and a 2.5% activation of IP3 formation. The point of this example is to stress that a very small activation of IP3 formation is sufficient to induce the small activation of Ca2+ entry necessary to fully activate a downstream physiological response. A similar conclusion emerges if one considers the agonist concentrations responsible for setting up Ca2+ oscillations.

C. What is the role of Ca2+ entry signals in driving Ca2+ oscillations?

For some cell types, the low agonist concentrations that activate physiological responses also induce Ca2+ oscillations. It is often assumed that Ca2+ entry signals and, in particular the SOCs, function to maintain such oscillations by recharging the internal store following each spike. It is relevant to ask, therefore, how much of the store is released during the course of a spike and what is the level of Ca2+ entry? When the ER lumenal level of Ca2+ was monitored in a pancreatic cell spiking regularly in response to a low dose of acetylcholine, it declined by about 5% during each spike (Park et al (2000) EMBO J. 19:5729). This small loss was then gradually replenished during the next interspike interval. In many cells, this refilling of the partially depleted store occurs during the course of the interspike interval even though the cytosolic level of Ca2+ is close to its resting level. This implies that the rate of Ca2+ entry is very low and is rapidly taken up by the ER to set the stage for the next spike. Indeed, I consider that it is the rate of Ca2+ entry that is the timing mechanism for the frequency-modulated (FM) signaling mechanism seen in many cell types.

As the entry of Ca2+ necessary to maintain oscillations occurs with a minimal depletion of the internal store this would seem to rule out a role for a store-operated signalling mechanism. However, this may not be the case as will be discussed in the following section.

  1. Is there a role for conformational coupling mechanisms in regulating entry at physiological agonist concentrations?

As Ca2+ entry in many cells (Section A) is driven by receptors that stimulate phospholipase C, the most likely entry signals are either IP3 or DAG or a related lipid intermediate. Since DAG has been shown to activate certain Trp channels, it cannot be ruled out. However, it would seem that IP3-induced entry is a more likely mechanism because the introduction of IP3 into the cell can activate Ca2+ signaling including the more complex patterns of oscillations that depend on entry. Since there is no evidence for IP3 activating entry channels directly (except in some sensory cells), it is argued that it acts through the conformational coupling mechanism outlined earlier.

E. Conformational coupling- a revised hypothesis

In order to understand how the conformational coupling mechanism might operate to regulate entry, especially at low agonist concentrations, it is necessary to consider the structural organization of the coupling unit within the context of the tubular ER network. As proposed previously (Berridge,1995), small fingers of this network form flattened sacks that make close contact with the plasma membrane to form the specialized junctional zones where the conformational coupling units are located. The10 nM gap separating the ER sacks from the plasma membrane has periodic densities that are thought to be the large cytoplasmic heads of the IP3R, which communicate information to the entry channels. These junctional zones appear to be few in numbers, which may explain why the rate of entry at physiological concentrations is so low.

An important issue arises as to whether these junctional zones are coupled to the bulk of the ER as is assumed in this hypothesis or whether there is a separate store associated with the plasma membrane that is dedicated to controlling entry as some have suggested. My argument against the latter mechanism is that it is inconsistent with the homeostatic function of store-operated entry, which is to ensure that the store remains filled. I would argue that a small store may fill up quickly and entry would cease even when the bulk of the ER is empty and requires further Ca2+.

As described in Section B the formation of IP3 at physiological agonist concentrations is very low raising the possibility that it occurs it is punctate i.e. IP3 may appear as brief highly localized pulses in small domains beneath the plasma membrane. If these localized domains of IP3 coincide with a region where a junctional zone is located, conformational coupling will be activated. This localized activation could be mediated by IP3 in two ways. Firstly, IP3 could act directly to stimulate the IP3Rs that are coupled to the entry channels. Secondly, it could act indirectly by stimulating uncoupled IP3Rs in the vicinity of the junctional zone to induce a localized depletion of the store within the ER to switch on a store-operated mechanism. This latter case could explain how a store-operated mechanism could operate to regulate entry during the low agonist conditions that induce Ca2+ oscillations. The highly localized depletion in the junctional zones control entry while the bulk of the Ca2+ is kept within the main body of the ER where it is available for the generation of Ca2+ spikes.

Whatever the activation mechanism turns out to be, the critical point is that entry is driven by a conformational coupling mechanism restricted to the junctional zones that are responsible for bringing Ca2+ into the cell prior to the onset of the Ca2+ spike. In the case of the insect salivary gland, I obtained some indirect evidence that 5-HT might be capable of stimulating entry during the latent period before the first spike occurred (Berridge 1994 Biochem J. 302:545). At low agonist concentrations, the earliest Ca2+ response to receptor activation may be the stimulation of entry, which is then responsible for charging up the internal store to prime the Ins3Rs for the large-scale regenerative release of Ca2+ that occurs during each spike.

Is it possible that the long held assumption the entry follows release might in fact be quite the opposite? Perhaps the earliest Ca2+ response at physiological agonist concentrations is an increase in entry that then sets the stage for Ca2+ release.

Calcium entry at physiological agonist concentrations

Jun 4 2004 3:46PM

Trevor Shuttleworth

Nobody who knows our work will be surprised when I say that Icompletely agree with Mike's point emphasizing the importance ofdetermining which mechanism(s) for Ca2+ entry operate at physiologicalagonist concentrations. However, I would just make a a couple of isolatedcomments.

RE: In response to the question raised that "studies on cloned celllines have identified a number of putative entry signals such as DAG, IP3, arachidonic acid, store emptying etc, but what is the evidence that any of these actually play a role in primary cells?".We have shown the presence of the ARC channels in parotid acinar cells(from both mouse and human), and demonstrated their crucial role in Ca2+entry at low agonist (carbachol) concentrations.

RE: The discussion regarding the mechanism for the entry of Ca2+necessary to maintain oscillations occurs with a minimal depletion of theinternal store seeming to rule out a role for a store-operated signalingmechanism. Also, the point made that "perhaps the earliest Ca2+ responseat physiological agonist concentrations is an increase in entry that thensets the stage for Ca2+ release". We, of course, agree and have made this same point repeatedly, ourselves.However, our data on Ca2+ entry during oscillations in HEK cells are notconsistent with the "revised conformational coupling" model proposed. Inthese cells, although the muscarinic receptors are coupled, as usual, toPLC, the entry of Ca2+ is independent of PLC activity. It is, instead,entirely dependent on the generation of arachidonic acid which, in the HEK cells at least, is exclusively via a cPLA2. This enzyme is also coupled to the activation of the muscarinic receptors, but a way that is separate,but parallel, to PLC. Of course, different cells may operate differently,and we don't have to consider that there is just one option. The realpoint I am trying to make here is that although we usually talk about PLC- coupled receptors, it is important to remember that such receptors may not couple exclusively to this enzyme, and other independent signalingpathways may be activated.

Calcium entry mechanisms

Jun 11 2004 2:09PM

Michael Berridge

Like Ken Byron, Trevor has raised an important question concerning the operation of other entery signals. The hypothesis I proposed certainly does not preclude other entry mechanisms especially with regard to their integration into the larger issue of how they contribute to the generation of Ca2+ spikes. The key point is that any entry mechanism must operate in a dose dependent manner over the physiological range that induced normal Ca2+ spiking. In my hypothesis, I tried to explain how the conformational coupling mechanism functions to feed Ca2+ into the ER to sensitize the uncoupled IP3Rs for a subsequent Ca2+ spike. This hypothesis certainly does not preclude other mechanisms as long as they can operate at these low agonist concentrations where entry plays a priming role. This is a somewhat different role for Ca2+ entry than Trevor has proposed recently where he placed most emphasis on entry contributing directly to the spike.

Calcium entry mechanisms

Jun 15 2004 8:47AM

Trevor Shuttleworth

Mike –Re. your comment that I had recently proposed that Ca2+ maybe "contributing directly to the spike".

Just to clarify, if I impliedthat the Ca2+ entering actually formed part of the increase in Ca2+associated with the spike, then I apologize. That is certainly not what Imeant.

Rather, we have always believed that the Ca2+ spike is entirely aresult of Ca2+ released from the stores (an exception may be T lymphocytes where, I understand, Rich Lewis has argued for a direct contribution ofCa2+ to the Ca2+ spike).

The evidence we had, indicated that Ca2+ entry did not seem to bemodulating oscillation frequency by any detectible influence on the rateof store refilling. This implied an effect on the release process.

Wethought that this probably resulted from the entering Ca2+ acting as a co- agonist with IP3 at the IP3 receptors to trigger release, and often stated as much.

However, as you pointed out to me (I think) at the Liverpoolmeeting last year, there are problems with this interpretation, and weagree with you. Consequently, we have begun thinking about possiblealternative mechanisms – although, based on our evidence, I stillbelieve that we should be focusing on an effect on release from thestores.

By the way, in a recent collaboration with James Sneyd (PNAS 101:1392-1396) he came up with a model that suggested that the "criticalparameter" for the effect of entry on oscillation frequency was the TOTALCa2+ load in the cell (not specifically in the stores), and how this wasinfluenced by the balance between influx and efflux across the plasmamembrane. We are still struggling to work out what this may mean in abiological/physiological sense, and there are still some aspects that themodel does not fully describe, but it has certainly given us something tothink about!

How many stores are there?

Jun 7 2004 7:35AM

Reinhold Penner

I would like to raise the issue about "how many stores are there?". I believethat this is a confusing issue and directly relates to some of the pointsraisedby Mike Berridge and Trevor Shuttleworth when they talk aboutphysiologicalconcentrations of agonists.

I think it is safe to say different agonists are capable ofgenerating differentCa2+ signals even in the same cell. Moreover, these signals change as afunction of agonist concentration even for the same agonist. Furthermore,wemust consider different "receptor concentrations", since receptors may not necessarily be expressed in equivalent numbers.

To complicate matters, itisalso clear that we cannot regard the ER as a single homogenous store,sincestore heterogeneity has been observed in numerous cell types. If we thenfurther consider that receptors and stores may have specific spatialarrangements, which would certainly have important consequences fordownstream second messenger production, we can easily envisage a complexsituation that defies simplistic interpretations.

Nevertheless, I'd like to point out a few results we have obtained in assessingsome of the intricacies of SOC activation in RBL (rat basophilic leukemia) cells. These cells, justlikeJurkat T cells (the other well-established model system for studyingCRAC),offer some advantages to at least appreciate these problems. In thesecells,we know that agonist stimulation in patch-clamp experiments activates asingle Ca2+ entry pathway, namely Icrac (yes, there could always beothers,but they simply do not show up even under the most physiologicalexperimental conditions that patch-clamp recordings can provide).

Withthisin mind, we can try to assess the dose-response relationship for Icracactivation by intracellular IP3 concentrations. We have shown that thisrelationship is rather steep and proceeds essentially in an all-or-nonemanner, with IP3 concentrations of 3 micromolar or more required to triggerIcrac.When we assess the store release by IP3 under similar experimentalconditions, IP3 concentrations of 1 micromolar or less already empty thebulkof IP3-sensitive stores, without triggering significant CRAC. This hasprompted us to propose that CRAC channels are under the control offunctionally (and possibly physically) distinct "CRAC stores" that do notcontribute significantly to cytosolic Ca2+ release transients.

Obviously, perfusing IP3 into a cell via the patch pipette is not how IP3 isgenerated physiologically by receptor agonists. This process occurs at the plasma membrane and it is therefore obvious that IP3 concentrations arehighest underneath the plasma membrane, which is also likely to be thelocation of the "CRAC stores". So how can we link store release and CRACactivation at different agonist concentrations? I would like to proposethefollowing scenario and am happy to receive flames ;-)

1. Very low, subtreshold agonist concentrations: Very low carbacholconcentrations (30 nM) in RBL cells will not trigger a visible Ca2+releasetransient and will not activate CRAC. This does not mean that signaling or IP3production do not occur, since giving a second identical subthresholdstimulus within a specific time window will produce a Ca2+ influx withoutgenerating a Ca2+ transient.

We believe this reflects the activation ofCRACby localized IP3 increases triggering the CRAC stores. The firstsubthresholdstimulus acts as a priming step, which then changes the apparentsensitivityof the CRAC stores by inhibiting the IP3-metabolizing enzymes thatnormallykeep IP3 levels low around the CRAC stores (presumably via IP4-mediatedinhibition of IP3 5-phosphatase). The second stimulus can then reach theIP3threshold required to activate CRAC. But the process is very slow andrequiresperfect timing of agonist stimulation.

So, Ca2+ entry through CRAC appears to occur at very, very low agonist concentrations. Such low concentrations of agonis do not trigger abulk ERrelease transient, because IP3 remains localized below the plasma membrane. The IP3 never reaches the deep cytosolic ER to generate a significant Ca2+releasetransient.

2. Low agonist concentrations: These are presumably the ones thatMike andTrevor allude to as "physiological". Here the cell responds to low "above- threshold" agonist concentrations by generating single or repetitive Ca2+transients, but without necessarily producing a significant plateauresponse.Surely, it seems surprising that the plateau phase does not show up wheneven subtrheshold agonist concentrations can lead to a detectable calcium transient (see point 1 above).Well,one could interpret this as a cross talk between release, uptake, andinflux. Iwould hypothesize that the low agonist concentrations produce enough IP3toreach both the CRAC stores and the deep ER. Since the ER is more sensitive toIP3 (possibly due to less IP3 metabolism and/or different IP3 receptorswithhigher sensitivity towards IP3), it can release Ca2+ very quickly atsubmicromolar IP3 concentrations, thereby generating the Ca2+ transient.

Although the CRAC stores may also be initially depleted by the above- threshold stimulus, they could immediately refill (possibly due to higherpump rates into a smaller-volume storage compartment). This may rapidlyterminate the ongoing slow activation of CRAC channels, curtailing anydevelopment of a plateau phase.

3. High agonist concentrations: Here both stores are maximally andrapidlydepleted and neither IP3 metabolism nor uptake mechanisms can counteractthis, giving rise to the typical biphasic signals (Ca2+ transient followed byinflux plateau).

So I think that store-operated influx could potentially occur at alllevels ofagonist stimulation and is not necessarily a feature of high (or even non- physiological) agonist concentration. The specific circumstances thendetermine how much of it we actually see reflected in the overall Ca2+signal.If we accept that a specialized CRAC store exists, which may not evengeneratemuch of a release transient, then we face even more challenges inevaluatingfura-2 signals...

How many stores are there?

Jun 8 2004 2:28PM

Trevor Shuttleworth

Certainly, the mechanism proposed by Reinhold could operate in theway described – and this may be the real situation in some cells, but itclearly is not the case in all cells.

We have demonstrated, for exampleboth in HEK293 cells and in mouse parotid cells, that low concentrationsof agonists (those that typically produce oscillatory Ca2+ signals)activate a conductance (the ARC channels) that displays fundamentalbiophysical characteristics that are entirely distinct from theconductance activated by high agonist concentrations (where sustainedelevated Ca2+ signals are produced).

Of course, as yet, we can only saythis for the cell types we have studied in such detail, and maybe the RBLcells operate differently. But, as we have published, the ARC channels are certainly present in RBL cells.

How many stores are there?

Jun 11 2004 2:09PM

Michael Berridge

Reinholt Penner

Reinholt, there is no doubt that the system is complex but one has to start with a "simplistic interpretation" to create a robust working hypothesis that attempts to bring together most of the salient observations. The main object of my working hypothesis was to describe a mechanism of Ca2+ entry and then to explain how it might function in the development of a typical Ca2+ transient. Many of the points that you made seemed to me to be exactly in line with this hypothesis except for two points. Firstly, you propose the existence of two stores, the CRAC store and the much larger release store and I shall discuss this later. Secondly, you do not mention the role for lumenal Ca2+, which I consider as an important component of the hypothesis in that it sets the sensitivity of IP3Rs. With these two points in mind, let me comment on your three points:

  1. This point about the action of very low subthreshold doses is well taken. I very much like your observation that CRAC can be activated at 30nM carbachol without inducing release, has this been published? An essential component of my hypothesis is that entry occurs before release and I suggested a possible mechanism for this in my hypothesis. I suggested that the formation of IP3 was highly localized in the immediate vicinity of the receptor, exactly as you do, where it could have two possible actions. Either it could activate conformation coupling directly or it could induce a local release to switch on entry through local store emptying.
  2. You make a distinction between subthreshold and threshold on the basis that the latter is the concentration that begins to give a meaningful Ca2+ signal in the form of a Ca2+ spike, which relates to a point made by Ken Byron. I am comfortable with this distinction because it fits in very nicely with my hypothesis. The two are really part of a continuum. At the subthreshold dose the entry is not large enough to set up the conditions for release to occur for two reasons. Firstly the level of IP3 reaching the uncoupled receptors is not high enough to trigger them. Secondly, and more importantly, the level of entry is not large enough to increase the lumenal level of Ca2+ sufficiently to sensitize the uncoupled IP3Rs to the low level of IP3.
  3. You seem to be concerned about why there is no plateau phase during this threshold condition. As I argued before, the reason why there is no plateau phase is because the SERCA pumps are very efficient at taking up the Ca2+ and it this uptake that sensitizes the IP3Rs.

  4. With regard to high doses, I agree with you completely.

Perhaps the main difference in our interpretation is whether there is a single or two separate stores (a CRAC store and a larger release store). I dealt with this question in my comments. It seems that we both agree that there has to be two functional stores and the question therefore is whether or not they are physically separate. I argue that the ER is continuous but that it can be separated physiologically into two regions. The junctional zone that responds to the high dose of IP3 near the membrane and the remainder of the store where the uncoupled IP3Rs are located. At physiological agonist concentrations, these uncoupled receptors do not respond directly to IP3 because the concentration is too low. Therefore, before they can respond to the low ambient IP3 level they have to be sensitized by the Ca2+ coming in from the outside through the entry pathway.

As I mentioned in my comments, the problem with having two separate stores is what happens when the small CRAC store fills up and the main release pool is still empty?

How many stores are there?

Jun 12 2004 5:59AM

Reinhold Penner


Thanks for the thoughtful reply. I will try and answer some of the questions you raised and comment on some of the points you made.

"1. This point about the action of very low subthreshold doses is well taken. I very much like your observation that CRAC can be activated at 30nMcarbachol without inducing release, has this been published?"

Yes, but it's in a rather obscure journal [Hermosura et al. (2000) Nature 408,735-740] ;-)

"An essential component of my hypothesis is that entry occurs before release and I suggested a possible mechanism for this in my hypothesis. I suggestedthat the formation of IP3 was highly localized in the immediate vicinity of the receptor, exactly as you do, where it could have two possible actions. Either itcould activate conformation coupling directly or it could induce a local release to switch on entry through local store emptying."

Yes, I completely agree that local signaling under the plasma membrane is crucial at low agonist concentrations. I'm not a big fan of the directcoupling hypothesis via IP3R, since SOC works just fine without IP3R (in DT40 cells with triple IP3R KO) and huge doses of heparin completely block IP3-inducedactivation of CRAC without affecting CRAC that is activated by e.g. ionomycin (I know, you can always argue that the coupling is independent of IP3, but I would expect heparin to at least hinder the interaction of IP3R with other proteins).

"2. (...) You seem to be concerned about why there is no plateau phase during this threshold condition. As I argued before, the reason why there is no plateau phase is because the SERCA pumps are very efficient at taking upthe Ca2+ and it this uptake that sensitizes the IP3Rs."

I'm not really concerned about the lack of the plateau at threshold agonist concentrations, because I do have an explanation for it, although it's differentfrom the one you propose. If I understand you correctly (please correct me if I'm wrong), you seem to assume that influx always precedes release and that the influx supercharges the bulk ER so it becomes more sensitive to IP3. Our data would argue against that. As mentioned in my previous post, the bulk ER responds very sensitively to IP3. 1 micromolar IP3 administered via patch pipette (i.e. globally) will essentially empty the entire IP3-sensitivestore, but that is not enough to trigger CRAC [see Parekh et al. (1997) Cell 89, 973-980]. So the bulk of the ER is clearly more sensitive to IP3 than the CRAC store (as to whether this is functionally or physically separate is notimportant in this context, although I would argue that this experiment is more consistent with physical separation). In any case, a reasonable conclusion from this is that for whatever reasons, the CRAC store has a higher threshold for IP3 than the bulk of the ER. Yet, when we go to subthreshold agonist concentrations it can be emptied first and cause influx. Based on a number observations, we have proposed that this is due to local metabolism of IP3 in the vicinity of the CRAC stores [see Hermosura et al. (2000) Nature 408, 735-740]. So, let me briefly compare the subtreshold and the threshold scenario:

Subthreshold: CRAC stores are selectively emptied due to local inhibition of IP3 metabolizing enzymes. This only works when subthreshold agonist doses are delivered repeatedly. The first application does not trigger anything (neither influx nor release; subthreshold). However, it primes the system by generating IP4, which is a very potent inhibitor of IP3 5-phosphatase and more long-lived than IP3 itself. Deeper ER is not affected, because IP3 remains local and subthreshold. Now, when the second subthreshold agonist application occurs, the IP3 metabolism underneath the plasma membrane iscompromised and this causes an apparent sensitization towards IP3. This now is above the threshold to trigger the CRAC stores, resulting in only Ca2+ influx, but no release transient, because the IP3 increase remains local.

Threshold: Here, the higher IP3 concentrations escape the subplasma membrane domain to trigger the more sensitive bulk ER release, and we get a Ca2+ transient. It is possible that the higher IP3 concentrations alsooverwhelm the metabolizing enzymes around the CRAC store and trigger the CRAC stores. However, the significant Ca2+ release from the bulk ER may cause a rapid shutdown of the slowly activating CRAC due to the uptake of Ca2+ into the CRAC store (but see my comment at the very end of thispost). That is my favorite explanation for the lack of aplateau.

"Perhaps the main difference in our interpretation is whether there is a single or two separate stores (a CRAC store and a larger release store). I dealt withthis question in my comments. It seems that we both agree that there has to be two functional stores and the question therefore is whether or not they are physically separate. I argue that the ER is continuous but that it can be separated physiologically into two regions. The junctional zone that responds to the high dose of IP3 near the membrane and the remainder of the store where the uncoupled IP3Rs are located. At physiological agonistconcentrations, these uncoupled receptors do not respond directly to IP3 because the concentration is too low. Therefore, before they can respond to the low ambient IP3 level they have to be sensitized by the Ca2+ coming in from the outside through the entry pathway."

I guess we have different ways of explaining the same phenomenon. You say that the bulk ER is initially less sensitive to IP3 than the CRAC store and requires Ca2+ entry to increase its sensitivity, whereas I argue theopposite, namely that the bulk ER is more sensitive than the CRAC store, but the CRAC store can gain apparent sensitivity when IP3 metabolism is altered. I wouldargue that the IP3 dose response curve obtained with patch pipette-loaded IP3 appears to support a higher sensitivity of the bulk ER. Adding IP4 changes the CRAC store sensitivity in these very same conditions.

"As I mentioned in my comments, the problem with having two separate stores is what happens when the small CRAC store fills up and the main release pool is still empty?"

Another question is whether the bulk ER ever gets emptied while the CRAC store is full. If so, then we need to ask the question whether filling the CRAC actually turns off CRAC. Really, we have no clue what activates CRAC and that represents a wealth of knowledge compared to what we know about the mechansims that would turn it off :-)

CRAC activity at low agonist concentrations?

Jun 15 2004 8:33AM

Trevor Shuttleworth

Reinhold – I have a question.

I looked again at the paper youmentioned by Hermosura et al. and saw that you recorded a Ca2+ influx at very low CCh concentrations in RBL cells using fura-2 fluorescence(even when in whole-cell mode), but I could not see anything about current recordings.

Did I miss something?

If I read your posting of June 7correctly, I understand that you believe that this entry is likely to bevia CRAC (as Mike Berridge later restated). However, without specificidentification of the responsible conductance as CRAC, I would questionthis assumption.

This may seem trivial, but I don't think it is. As statements get repeated in the literature, there seems to be an inevitable tendency to increasetheir impact or specificity such that a general measurement of Ca2+ entry, or even a poorly characterized current, becomes "morphed" into a specificassignment to a particular channel.

To be specific, I am not aware of any published evidence showing that CRAC channels are activated at very low agonist concentrations.

ARC channels,on the other hand, are activated by agonist concentrations at which we are just able to see the very weakest of Ca2+ signals, and continue to provide the predominant source of Ca2+ entry until sustained, elevated Ca2+signals are seen (JBC 276: 35676-35683).

CRAC activity at low agonist concentrations?

Jun 16 2004 9:19AM

Reinhold Penner


The experiments with subthreshold agonist concentrations were donewithfura-2 in intact cells and in patch-clamp experiments. Of course, the fura -2data only imply that CRAC is responsible for the observed increase in[Ca2+]i.In the patch-clamp experiments, there is no detectable increase in current (that's the nature of CRAC under unbuffered conditions). So yes, we havenoabsolute proof that the signal we observe is indeed due to CRAC.

However, I would like to point out that we have never seen an agonist - induced current under buffered conditions other than CRAC. We tried hardtofind ARC, but so far have never seen it.

I realize that you state you have seenARC currents in RBL cells (as unpublished observations), but we have notbeen able to get a 0.5-pA/pF increase, even with high CCh (carbachol) concentrations ontop of IP3-induced CRAC. (We have not tried this with low CChconcentrations,but I can't see why high CCh would fail to activate ARC under buffered[Ca2+]i.)

So, if you have an RBL cell line that produces a robust ARC, Iwouldcertainly be interested in getting it.

CRAC activity at low agonist concentrations?

Jun 17 2004 2:25PM

Trevor Shuttleworth

Reinhold A minor point: Re. ARC in RBL cells - in fact, our report of ARC currents in RBL cells is published in JBC 278: 10174-10181 (Fig. 1). The macroscopic currents areapproximately 0.5-0.6 pA/pF. As for the cell source, we simply used thestandard cell-line available from ATCC.

CRAC activity at low agonist concentrations?

Jun 18 2004 8:53AM

Reinhold Penner


thanks for the clarification about ARC in RBL-1 cells.

Do you haveany data onagonist activation in RBL, as that is the experiment I was referring to in mypost? We simply cannot get more current with agonists on top of CRAC.

CRAC activity at low agonist concentrations?

Jun 18 2004 1:12PM

Trevor Shuttleworth

No, I'm afraid we have never tried agonist-induced responses in theRBL cells.

We are currently (no pun intended!) looking at responses in parotid cellsin an attempt to obtain data on a "real" cell. We know the ARC channelsare there, and Ca2+ entry at low concentrations of agonists is AA-dependent, but we have yet to try and record current responses to agonists - staytuned!

Comments on cellular domains that contribute to calcium entry events

Jun 8 2004 7:04AM

Kenneth L. Byron

Let me first endorse Mike's arguments that we must consider physiological concentrations of agonists to understand the mechanisms governing Ca2+ entry. Despite the practical difficulties that are involved in resolving submaximal signaling events in the laboratory, work from anumber of labs, including my own, have revealed some remarkable differences in Ca2+ signaling patterns comparing low and high agonist concentrations.

I really wanted to ask Mike a couple of questions about the conformational coupling model, which I have always found appealing. I have two questions related to your vision of the cellular architecture. First, there is more and more experimental evidence supporting mechanisms for activation of Ca2+ entry via signaling pathways (e.g. CIF, iPLA2) that can apparently operate independently of PLC, IP3 receptors, or Ca2+ release. Do you think these mechanisms converge on the same cellular structures that are involved in conformational coupling?

Second, the arrangement of Ca2+ transporters (Ca2+-ATPases, Na+/Ca2+ exchangers) and Ca2+ release sites is not clearly described, so I wonder if you would expand on your ideas about which subsets of IP3 receptors are functioning as release channels at different agonist concentrations and the question of whether there is a particular orientation that would direct the released Ca2+ toward its presumed effector targets, which may be away from the plasma membrane. Are SERCA pumps arranged to buffer Ca2+ that diffuses toward the membrane? It seems as though your model would necessarily have the Ca2+ entrydirected toward the IP3 receptors of the junctional ER and also perhaps toward a concentration of SERCA pumps that would act to refill the stores.

If Ca2+ entry precedes Ca2+ release, does the local elevation of [Ca2+]i sensitize the junctional IP3 receptors or act as a co-agonist to activate the release channels? Or is there a different subset of release channels away from the junction that allow the bulk of released Ca2+ to more effectively reach its targets?

Comments on cellular domains that contribute to calcium entry events

Jun 11 2004 2:10PM

Michael Berridge

I was pleased to see that Ken agrees with the need to consider physiological doses of agonist. This is a very important issue that cannot be over stressed.

In answer to the question of whether some of the alternative entry mechanisms (e.g. CIF, AA) impinge upon the conformational coupling (CC) mechanism, I doubt it. This does not mean that the do not play a role as I indicate in my answer to Trevor’s qiestion outlined above. It seems that these other entry signals act directly on channels so there is no need to suppose that they act through the CC mechanism.

The second question concerns the organisation of other components that function together with the CC mechanism. It is difficult to be precise about this. With regard to the IP3Rs, it is important to make a distinction between the coupled IP3Rs that occur in the junctional zone and the uncoupled IP3Rs that lie elsewhere on the ER. It is the positioning of the latter that are critical and, as Ken suggests, some of these may be situated close to downstream effectors well away from the cell surface. Indeed it is clear from various cell types that the initiation site where release begins is often located away from the cell surface. As pointed out by Indu Ambudkar, this spatial separation of entry and release is particularly evident in pancreatic acinar cells where entry occurs at the basal side while release occurs in the apical region. This relates to the question raised by Ken at the end of his comments, what is the relationship between entry at the junctional zone and the subsequent release? In the hypothesis I outlined in my comments, I proposed that the ER takes up the calcium that enters at the junctional zone. This Ca2+ can then tunnel throughout the ER network to increases the lumenal level of Ca2+ which in turn sensitizes all the uncoupled receptors so that they can release Ca2+ during each spike. A critical part of the hypothesis is the role that lumenal Ca2+ plays in sensitizing IP3Rs to set the stage for release to occur (see also my comments to Reinholt Penner).

Comments on cellular domains that contribute to calcium entry events

Jun 15 2004 8:54AM

Jim Putney

I am commenting in response to Mike’s contribution, but really Ihave something general to say that is relevant to several of thecontributions. I wish to speak to two issues that have been mentioned onseveral occasions. These are (1) it is best to address signalingmechanisms for physiological concentrations of hormones andneurotransmitters, and (2) it is always better to measure current directly than to deal with fluxes based on fluorescent indicators.

Of course, no one can be opposed to studying physiologicalconcentrations of agonists; this would be akin to being opposed tomotherhood. Nonetheless, at lower agonists concentrations somedifficulties arise.

For the case of the store-operated channels, I reallydo not know of a single instance in which anyone has measured the currentunderlying these channels under truly physiological conditions. In manycell types, even the extreme manipulations of high intracellular Ca2+buffering do not bring up enough current to be reliably measured. Yet thefluorescence experiments indicate it is clearly there.

The reason for this is that with all of the potentials problems and artifacts from the use offluorescent indicators (all of which can be dealt with by use ofappropriate controls), this is still by far a much more sensitivemeasurement, for Ca2+ anyway, than measurement of current. Thus as theagonist concentrations get lower and lower, the experiments and evidencefor and against various proposed mechanisms will of necessity get lessdirect.

Under these conditions, it is useful to employ specificpharmacological tools to identify specific pathways. I am not sure that Iagree with the general statement that we do not have a specific blocker of SOCs.

It is true we do not have a single, highly specific and potentinhibitor, such as the dihydropyridines or TTX. However, I would suggestthat a complete block by both very low (1 um and below) concentrations oflanthanides together with block by 2APB (2-aminoethoxydiphenyl borate) in the 20-50 uM range isdiagnostic for store-operated channels. I do not know of any other channel with these pharmacological properties (but perhaps one of the otherparticipants does).

Calcium entry at low agonist concentrations

Jun 16 2004 9:25AM

Trevor Shuttleworth

A couple of points:

1) I too am not aware of any published report showing SOC channel activity under "truly physiological" conditions. However, in contrast, we ARE ableto show ARC channel activity under these conditions (i.e. at the sameagonist concentrations that we are just able to see the very weakest Ca2+signal, and in perforated patch experiments etc.).

These findings wouldcontradict your comment that fluorescence methods are (necessarily) a much more sensitive measurement- for Ca2+ than a measurement of current - atleast under these conditions. It is precisely because we CAN measure theactivation of a Ca2+-selective conductance at the identical agonistconcentrations at which Ca2+ signals can begin to be seen that led us to theconclusion we have been proposing for the past few years, namely that SOCchannels are not active at such concentrations, but ARC channels are!

2) A minor comment re. the pharmacology. Lanthanides at 1 uM willblock ARC channels (by approximately 50%), so submicromolar would berequired if ARC and SOC channels are to be clearly distinguished.

I agreewith the "selectivity" of 2-APB for the channels themselves, but whenstudying agonist-stimulated responses, our experience indicates that almost anything can happen! I think this is because such responses involve somany steps at which 2-APB can be having any number of nonspecific effects.

Calcium entry at low agonist concentrations

Jun 17 2004 7:47AM

Reinhold Penner


I would be interested in hearing your opinion on the followingquestions:

1. How general (ubiquitous) do you think ARC is in comparison to SOC? Doyou envisage these two mechanisms to always act as a tandem (ying-yangingat different agonist concentrations)? Is ARC regulated (i.e. do you seelarger/ smaller ARC in say preactivated cells)?

2. How do you see the signaling cascades for typical PLC-coupledagonists(considering that in RBL and Jurkats both PLCbeta and PLCgamma convergeon the same SOC)? Does every PLC-coupled receptor produce both AA andIP3? And how would that be transduced?

3. What about the Ca2+ dependence of ARC? You say it's inactivated by Ca2+(via calcineurin). But CRAC is also very strongly inhibited by Ca2+ entry, somuch so that we can't detect it when we leave the cytosol unbuffered.Wouldn't ARC actually limit CRAC if it preceded it rather than CRACinhibitingARC?

Calcium entry at low agonist concentrations

Jun 18 2004 9:02AM

Trevor Shuttleworth

1. This is obviously difficult to answer categorically, and ofcourse my opinion is probably not entirely objective! So let me try tojust give you the information.

So far, we have identified ARC currents inHEK293, HeLa, COS, CHO, RBL, and DT40 cells. It is also in parotid acinarcells (mouse and human).

In the cells we use most (the HEK 293 cellsstably transfected with the m3 muscarinic receptor), we begin to see Ca2+signals in intact cells at around 0.2 uM CCh. These are typically small,rather irregular, oscillations that are often not sustained.

At this sameconcentration we can measure a detectible ARC current in whole-cell patchrecordings equivalent to around 40-50% of maximum ARC current. An exampleof these data was published in Biochem Soc Trans. 31: 916-919. Of course,AA-activated Ca2+ entry has been demonstrated in a wide variety of cells,but this may not always involve ARC channels.

Re. the "ying-yang"..... our proposal for the reciprocal regulation of ARC and SOC channels is clearlysupported by a wealth of evidence from the m3-HEK cell system. But it isalso consistent with demonstrations that Ca2+ entry at low agonistconcentrations is entirely (or, at least, predominantly) dependent on thegeneration of AA in several cell types.

It also explains why the originalwork from Jim Putney's lab (Takemura et al) failed to detect anyadditional route of agonist-activated Ca2+ entry in intact cells afterthapsigargin treatment. This was critical in leading Jim to propose thatthere was only one type of entry, i.e. capacitative entry.

As to the"preactivation" question, if you are referring to the type of protocol you used in the Hermosura paper, then we haven't tried that. However,something we have been looking at recently is that, if we activate a large Ca2+ signal in an intact cell with a supramaximal agonist concentrationand then look at ARC currents, they are much smaller (consistent with thecalcineurin-dependent inhibition story).

2. I see no problem for both PLCbeta and PLCgamma converging on the same SOC. They will both result in the generation of IP3 which, ifproduced in sufficient quantity, will deplete the stores and activate SOC.

As for the second part of your question, this obviously relates, at leastin part, to the first question re. how ubiquitous we think the ARC channel pathway is.

In the HEK cell case, the agonist-induced generation of AAseems to be exclusively via a cPLA2 that is activated simultaneously, andin parallel, to the PLC pathway. However, it is clear that different cells have several different ways of making AA in response to agonist action – some are inndependent of PLC, some are downstream of PLC. Dissecting allthese possibilities out pharmacologically would likely be difficult.

Ithink the key point to remember, as I stated in a comment I made inresponse to a posting by Mike Berridge earlier in this discussion, is that although we all routinely talk about PLC-coupled receptors, we shouldalways be aware that this does not necessarily represent the solesignaling pathway activated by such receptors.

3. My understanding of the inhibition of CRAC by Ca2+ entry suchthat it cannot be detected in an unbuffered situation is that this islargely (although possibly not entirely) due to refilling of the stores by the avid SERCA pumps. Isn't this essentially what the data from AnantParekh's lab says? If we hold the IP3 receptors open (e.g. withadenophostin) we can measure robust SOC (CRAC-like) currents in HEK cellswith the cytosol buffered at 400 nM (see JBC 276: 35676), and I thinkAnant had showed the same effect at even higher concentrations (J.Physiol. 522: 247).

As for ARC limiting CRAC I think the key here is thatduring agonist activation, the Ca2+ entry through ARC does not, itself,raise global cytosolic Ca2+ - it only helps to induce the repetitivetransient (and probably partial) IP3-induced release of Ca2+ from stores.Only when sufficient IP3 is generated to produce a sustained depletion ofthe stores is SOC activated. The resulting sustained elevation incytosolic Ca2+ then turns off the ARC channels and Ca2+ entry is entirelyvia the SOC/CRAC channels – as is demonstrated by its independence fromAA generation.

I am sorry these are all rather long answers to short questions, butI hope they help clarify our view of what is happening.

Pharmacology of SOCs and ARC

Jun 17 2004 11:44AM

Donald Gill

This is a response to Trevor and Jim's discussion on thepharmacological distinctions between SOC and ARC.

2-APB may be useful but Trevor is not so sure that the lanthanides are useful in discriminating.

Another group of compounds, the 3,5-bistrifluoromethyl pyrazoles (papersby Ishikawa et al in J.Immunol and by Zitt et al, JBC) are interestingmodifiers of ICRAC. We find they also modify SOCs in most cells and andalso TRP channels. We are questioning whether their effects are direct or indirect, but we certainly find their potency is quite high (IC50 around100 nM for both SOCs and TRPC channels). It might be useful for Trevor to look at these on ARC in case he hasn't already done so!

Pharmacology of SOCs and ARC

Jun 18 2004 8:54AM

Reinhold Penner


the 3,5-bistrifluoromethyl pyrazole (BTP2) compound you mentioned isveryproblematic. I must issue a big warning to anyone using this compound as a potent and selective inhibitor of CRAC. It isn't!!!

Unfortunately, our paper on this compound just got rejected byScience, so Ican't be more specific about the mechanisms involved in its action. It istruethat BTP2 will completely shut off Ca2+ influx in lymphocytes at lownanomolar concentrations, but this effect is due to an as yet unpublishedmechanism... This novel mechanism is very important to consider whenrelying only on fura-2 measurements to assess SOC.

Sorry to be so secretive about this, but please be advised that BTP2shouldnot be considered a SOC inhibitor at low concentrations.

Pharmacology of SOCs and ARC

Jun 18 2004 8:55AM

Trevor Shuttleworth

Don: No, we haven't tried them, but we will put them on the "to do" list....... so many drugs, and so little time!!

Physiological agonist-concentrations

Jun 17 2004 9:03AM

Indu S. Ambudkar

I think the discussion we have been having regarding concentrationsof agonists used for Ca2+ influx measurements is a critical issue.

Iagree with Reinhold and Jim (although, I would like to discuss the analogy to motherhood in greater detail !!!) that there are problems associatedwith our ability to accurately measure Ca2+ influx at low agonistconcentrations. With the exception of one paper and different [IP3] inthe pipette (Parekh) no-one has really reported the characteristics ofchannel activity at the so-called “physiological concentrations” ofagonists.

What we do know from fura-2 measurements and readouts using K-or Cl channel activity is that typically these low concentrations induceoscillations, while higher concentrations give sustained [Ca2+]ielevations.

In polarized epithelial cells there is another level ofcomplexity since the oscillations are spatially restricted. This might be due to the large abundance of IP3R in the luminal region or, as has beensuggested, a more “sensitive” population of IP3Rs.

More importantly, and pertinent to this discussion, is the fact that therequirement for external Ca2+ in these oscillations appears to bedifferent in different cells. In some cells, Ca2+ entry is absolutelyrequired for the oscillations whereas in some cases the oscillations arerelatively independent of external Ca2+. In the former, it is not entrythat is oscillating per se, but rather each oscillation is due to thecycle of Ca2+ release via IP3R - activation of influx (meaning there mustbe inhibition of release, likely due to feed-back regulation of IP3R orPLC) - refill of stores - next release.

This in my opinion is the mostinteresting system. Ca2+ influx occurs without “global” Ca2+ changeseven in the continued presence of the agonist. So, here is a store thatcan be depleted at low stimulus levels and is coupled to influx (seeReinhold’s point about IP3 sensitivity of the “Ca-store”).

Thecritical question of course is whether this Ca2+ entry is “store- operated” or activated by local changes in DAG/other messenger. Thiscan be answered only by measurements of currents at these lowconcentrations to see if they are the same as those produced at higheragonist-concentrations. This is an important experiment that needs to bedone to move forward in this field.

I also wanted to throw out one more point for discussion. What isreally known about “physiological” concentrations of agonists ?

Infact, the junctional architecture of the region where neurotransmittersare released near the plasma membrane of cells can in fact enable fairlyhigh local concentrations of agonists. If receptors are strategicallylocalized in this region, depending on the affinity and number of suchreceptors, one could elicit a “High” but extremely localized response. Any thoughts……?

Physiological agonist-concentrations

Jun 18 2004 10:13AM

Kenneth L. Byron

In regard to Indu’s query: “What is really known about“physiological” concentrations of agonists?

In fact, the junctionalarchitecture of the region where neurotransmitters are released near theplasma membrane of cells can in fact enable fairly high localconcentrations of agonists. If receptors are strategically localized inthis region, depending on the affinity and number of such receptors, onecould elicit a “High” but extremely localized response. Anythoughts……?”

Neurotransmitter release, as well as other paracrine or autocrinesignaling paradigms may very well produce quite high local concentrationsof agonists and it’s important to consider whether these are thephysiological mechanisms under which Ca2+ signals are being generated inthe cells under investigation.

On the other hand, many of us investigatetruly endocrine signaling phenomena where concentrations of agonists inthe systemic circulation probably never approach the concentrationstypically used to study store-operated Ca2+ entry pathways.

I think weought to test a range of agonist concentrations that may reasonably coverthe full range of concentrations expected physiologically in the systemunder investigation. Perhaps we should also focus more attention on trying to measure what concentrations are actually achieved in paracrine orautocrine signaling systems.

My guess is that store-operated channels are activated even at very lowagonist concentrations, but the difficulty in detecting very low-levelactivity of these channels especially under physiological ionic conditions has prevented us from examining this in any detail.

Physiological agonist-concentrations

Jun 18 2004 1:20PM

Indu S. Ambudkar

Ken; I agree with your response. Everything gets more complicated, doesn't it?

I think each agonist has to be examined within the context of

(i) how/where the cell "senses" it and

(ii) cell type.

Thus, there could be quite a difference between the effective "concentrations" of truly endocrine signals vs paracrine or neurocrine signals. Definitely some tissues such as salivary glands recieve their primary stimuli through neurotransmitter release near the cell. In fact, earlier EM has shown nerve endings in close proximity to the basolateral plasma membrane in salivary gland cells.

In these glands, there is a resting level of activity (stimulii unknown)and then a "regulated" level where fluid secretion increases 10-100 fold (in response to oral and sensory stimulii). We should go back and re-assess some of the old morphology as well as pharmacokinetic data.

My feeling is that the cells "custom design" each signaling pathway not only with regard to how (concentration, time, etc.) they recieve the signal but also which downstream function is regulated. In fact the latter might determine the former.

My suggestion is that we should stop generalizing our findings !!

Physiological agonist-concentrations

Jun 18 2004 1:17PM

Trevor Shuttleworth

Indu: some comments...

1. "With the exception of one paper and different [IP3] in thepipette (Parekh) no-one has really reported the characteristics of channel activity at the so-called “physiological concentrations” of agonists."

I would disagree. As I have already noted on this forum, we havemeasured the Ca2+-selective currents responsible for Ca2+ entry at agonist concentrations that are just able to produce the very weakest Ca2+ signals (see JBC 276: 35676; and Biochem Soc Trans. 31: 916). These currents showall the unique specific characteristics of the ARC channels, and areentirely distinct from the currents activated by store-depletion (SOCchannel currents).

2. "In polarized epithelial cells there is another level ofcomplexity since the oscillations are spatially restricted. This might bedue to the large abundance of IP3R in the luminal region or, as has beensuggested, a more “sensitive” population of IP3Rs."

Again, I would disagree. This is certainly true for pancreatic acinar cells. However, in parotid acinar cells, Dave Yule's group have shown that although the Ca2+ signals originate apically, they quickly spreadthroughout the cell in less than 200 ms (J. Physiol. 540:469).

3. "More importantly, and pertinent to this discussion, is the fact that the requirement for external Ca2+ in these oscillations appears to be different in different cells. In some cells, Ca2+ entry is absolutelyrequired for the oscillations whereas in some cases the oscillations arerelatively independent of external Ca2+."

I think this is more of an apparent difference than a real one. Asyou note later in your posting, in almost all cases, the oscillationsthemselves result from the cyclical release and re-uptake of Ca2+ into the stores. Based on the evidence we, and others, have accumulated over recent years, the role of Ca2+ entry seems to be simply to facilitate thelikelihood of this happening.

In other words, the main effect of Ca2+entry under these conditions is to modulate the frequency of theoscillations. In some cases this influence is "absolute" in thatoscillations cease immediately external Ca2+ is removed. In other casesoscillations continue, often for a considerable period, in the absence ofexternal Ca2+. But even in these latter cases, the rate of Ca2+ entrymodulates oscillation frequency.

Simply put, although the fundamentalmechanism underlying the oscillations themselves may be independent ofentry, their frequency is almost always very dependent on the rate ofentry.

4. "This can be answered only by measurements of currents at theselow concentrations to see if they are the same as those produced at higher agonist-concentrations. This is an important experiment that needs to bedone to move forward in this field."

We have done precisely this, at least in HEK293 cells, and theresults are published (see my response #1 above). As noted, the currentsactivated at low agonist concentrations are NOT the same as thoseactivated at high concentrations. Moreover, their mode of activation isalso entirely different.

Physiological agonist-concentrations

Jun 18 2004 2:06PM

Indu S. Ambudkar

Trevor; Point noted. I am glad you corrected me. I was actually writing within the narrow context of agonist-stimulated "store-operated" calcium entry channels.

I think your studies reveal a lot. First of all they show that different signals are generated by the same stimulus and that these signals can activate different calcium channels.

Secondly, low concentrations of agonists produce signals that activate calcium channels distinct from SOC.

The question I am trying to deal with, and what has come up in this discussion, is whether store-operated calcium entry works at these low levels of stimulii or not.

If it is seen only at higher agonist concentrations, then is it physiologically relevant (I know your answer to that one already!!!). Which is why I was trying to figure out what might be the actual agonist concentration that is sensed by the cell. After chatting with a few people, I came to the conclusion that in some cases it can be fairly high. But, of course, "in vivo" experiments will be needed to prove this. Good luck to us!!

Calcium signaling and microdomains

Jun 4 2004 4:06PM

Indu S. Ambudkar

Is store-operated calcium entry organized in functionally and structurally distinct microdomains?

Since other sections of this forum address more didactic issues relating to the actual definition and measurements of these channels, modes of regulation and their molecular composition, let's start this discussion by agreeing on the fact that calcium influx is activated in cells upon stimulation by agonists coupled to the stimulation of PIP2 hydrolysis. For now, I will set aside the question of whether this calcium influx is comprised of one type of channel or several, and examine the evidence that agonist-stimulated calcium influx occurs within specific spatially restricted microdomains.

Early studies using fluorescence measurements showed that Ca2+ entry continues to occur after removal of an agonist and this leads to store refill without any significant change in cytosolic calcium levels, other than the small overshoot observed upon re-addition of calcium to cells treated with agonists. This lead to the proposal that Ca2+ that enters the cell is rapidly taken up into the store.

Electrophysiological studies using Ca2+-activated Cl- and K+ channels as readouts also showed that there is minimal diffusion of Ca2+ in the plasma membrane region. Ca2+ entry during refill did not activate these channels, unless SERCA was inhibited. Thus, SERCA and Ca2+ entry channels can be suggested to be present in close proximity to each other, which also then suggests that the ER membrane has to be juxtaposed with the PM.

If the Ca2+ status within the ER regulates activation of this channel, then apposition of the two membranes not only determines the route for Ca2+ once it enters the cell, but could also regulate channel activation. Subsequently, it was shown that not only SERCA, but also PMCA and mitochondria affect calcium entry by regulating the calcium concentration in the vicinity of the channel. These organelles together act as a buffer to keep calcium concentration near the channel low and prevent Ca2+-dependent inactivation of the channel inactivation, a regulation that has been well documented. SERCA of course also enables refill of the internal calcium stores, i.e. its activity leads to channel inactivation.

Questions that can be raised for further discussion during this forum are:

(i) Does [Ca2+] in this region change during the refill process and whether this directly affects gating of calcium entry (distinct from the Ca2+-dependent inactivation, which is reversible)?

(ii) Whether this microdomain organization is universally found in all cell types or whether it is dictated by the architecture of the cell and the localization of the channel?

Compartmentalization of calcium signaling and calcium entry:

Biochemical and morphological data support the suggestion that Ca2+ signaling proteins are assembled in multiprotein complexes that are localized in distinct regions of the cells and that Ca2+ signaling events occur is spatially segregated domains of the cell. The best data available in this regard come from studies with polarized epithelial cells (see reports from the Petersen and Muallem labs). Exocrine gland acinar and ductal cells are polarized epithelial cells that also demonstrate a functional polarity. Thus, the cellular location of the Ca2+ signal has important physiological implications in the regulation of function.

A number of calcium signaling proteins have been shown to be localized in the apical region of these cells. Further, initial Ca2+ signals are generated in this region and then spread to the basal region. Importantly, at low concentrations of agonist, calcium spikes are observed in the apical region and do not spread to the basolateral region. These Ca2+ spikes are sufficient for functional regulation of ion channels that are present in the apical membrane. It was also recently reported by Muallem and co-workers that distinct RGS proteins regulate Ca2+ signaling events at different cellular locale. Thus, each Ca2+ signaling complex has a specific function within a cell. Its activity and location are strategically designed in order to facilitate and optimally enable the cellular function it regulates.

Our knowledge about localization of calcium influx is rather sparse.In nonpolarized cells, which most of us use for our studies, we do nottake into consideration the locale of the channel. When working with amore physiological system, such as primary cultures or freshly dispersedcells, this is crucial for understanding function and regulation ofcalcium entry. I think such understanding ultimately depends on knowing

(i)the site of calcium influx,

(ii) the function carried out by the calcium entering the cell, and

(iii) the calcium signaling cascade it is associated with.

Studies reported with acinar cells present an interesting problem. Low concentrations of agonists that produce apical calcium spikes, have not thus far been associated with calcium entry. Higher agonist concentrations or thapsigargin that induce substantial depletion of Ca2+ stores appear to stimulate Ca2+ entry via the basolateral membrane. An interesting point to consider is whether in fact apical Ca2+ entry does occur in response to apical Ca2+ store depletion, and, if so, whether it is "store-operated." Alternatively, signals within the ER might transmit this apical Ca2+ depletion to the basal region for SOCE activation. Receptors that activate Ca2+ signaling have been detected in the apical region of cells. PIP2 hydrolysis resulting from activation of such receptors produces local signals, which, in addition to inducing internal release, could activate calcium channels in the same signaling domain. Such channels might not only provide calcium for refilling the local stores, but could also modify local Ca2+ signals and recruit other signaling proteins. Thus, agonist-activated Ca2+ channels might serve functions in cells other than refilling of internal Ca2+ stores. As we give some thought to these calcium channels in the coming weeks, it will be important to consider and discuss some of these other possibilities. Ultimate understanding of this will depend on the identification of the channels and defining their location in cells.

Physiological relevance of SOCE:

Since we are defining cellular signals for activation of calcium entry, I think it is important to consider the function and regulation of SOCE in the context of what we have discussed above. The first point that needs to be addressed is whether SOCE serves any other purpose other than to replenish internal Ca2+ stores. Although this will require further studies and our definition of this calcium influx pathway might change as we learn more, we should consider the conditions under which SOCE is activated. Typically, high levels of stimulus are required to deplete the stores. In acinar cells, local Ca2+ signals with low levels of agonists do not appear to activate it. This is also true for ICRAC, where relatively high agonist concentrations are needed for activation. Thus, it is quite relevant to ask whether SOCE in fact has any role under normal physiological conditions of cell stimulation or whether it is a "response" to more extreme conditions that induce substantial depletion of stores, which might put the cells in a "stress mode". This suggestion does not in any way detract from the concept of SOCE but might give us another perspective in understanding its mode of regulation.

Assembly and trafficking of Ca2+ signaling proteins:

It is becoming increasingly clear that Ca2+ signaling proteins are associated physically to form multiprotein complexes. Specific complexes exist that are coupled to different receptors. Functional distinctions will be dictated by the location of these proteins, the intensity of the signals they perceive, and functional cross-talk between the various signaling pathways. Both protein and lipid components are structurally involved in the assembly of these signaling complexes.

PDZ-domain containing and several other proteins, e.g. RACK1, have been shown to act as scaffolds for the assembly of receptor associated signaling complexes in the plasma membrane. These scaffolding proteins provide a framework that brings functionally related proteins in close proximity to each other.

In addition, lipid raft domains (LRD) in membranes might also have a role in assembly of these complexes. Caveolae are a specialized form of LRD that contain caveolin-1, a cholesterol binding protein. Key protein and non-protein molecules involved in the Ca2+ signaling cascade, such as phosphatidylinositol -(4,5)-bisphosphate (PIP2), Galphaq/11, muscarinic receptor, PMCA, and IP3R-like protein, and Ca2+ signaling events such as receptor-mediated turnover of PIP2 have been localized to caveolar microdomains in the plasma membrane. An interesting study showed that agonist-stimulated Ca2+ signal in endothelial cells originates in specific areas of the plasma membrane that are enriched in caveolin-1 (Issihiki et al., 1998). Furthermore, we and others have reported that intact lipid rafts are required for activation of SOCE (Lockwich et al 2000, Issihiki et al 2002, Kunzelmann-Marche et al 2002). Thus, caveolae might regulate the spatial organization of calcium signaling by contributing to the localization of Ca2+ signaling complex as well as the site of Ca2+ entry.

However, exactly how caveolae regulate SOCE is not yet known. Added to the problem is the observation that cells that lack caveolin appear to be functionally intact. However, possible compensatory mechanisms have not yet been assessed in this case.

Nevertheless, it is interesting to propose, that at least in some cases, caveolar LRD might facilitate and coordinate the signals that lead to activation of SOCE via two possible mechanims.

(1) Since Ca2+ signaling proteins that lead to the activation of SOCE are colocalized in the same microdomain, caveolae could mediate interactions between the SOC channels and proposed regulatory proteins such as IP3R, or PLC-gamma and PLC-beta. The invaginated morphology of caveolae would uniquely enable the plasma membrane in this region to have access to regulatory components located further inside the cells, such as the ER.

(2) Since caveolae are also found as subplasma membrane vesicles, it is interesting to speculate that they might act as holding platforms for pre-assembled Ca2+ signaling complexes, or key components of this complexes, which upon stimulation of the cell are recruited to the plasma membrane via vesicle fusion. It is important to note that proteins involved in docking and membrane fusion are enriched in caveolae (Issihiki and Anderson, 1999). Such regulation of SOCE would be consistent with the secretion-coupling model. Finally, caveolae could also function as regulators of SOCE inactivation. Caveolin-1 is known to act as a tonic inhibitor of a number of signaling proteins. Additionally, caveolae have been shown to undergo dynamic internalization and the internalized vesicles have been shown to fuse with the ER. Thus, during prolonged activation of calcium influx, channels could be down-regulated via internalization.

An interesting idea proposed by Andersen is that this process would allow the external Ca2+ to be trapped in the vesicles and delivered to the ER. Thus, recycling of plasma membrane calcium channels would both limit the number of "active" channels and provide a route for the refill of internal Ca2+ stores with external Ca2+.

Another possible mechanism for inactivation of SOCE would be exit from caveolae and internalization via clathrin-coated pits. Channels internalized this way would be routed to endosomes for degradation.

There is almost no information regarding trafficking and assembly of calcium entry channels. Since these mechanisms can determine their surface expression as well as regulation, it is important to examine this is in our future studies. One interesting observation in MDCK cells is that caveolar-lipid rafts are found basolaterally while non-caveolar LRDs are found apically. This can then form the basis for functional segregation of proteins in such cells.

What can we learn from TRPC channels?

In our quest for the identity of the store-operated calcium channels, many of us have focused our attention on TRPC channels. While it might be possible that none of the TRPCs might in fact be the SOC channel (further discussed in sections 1 and 2 of this forum), our studies have demonstrated that TRPCs are the only presently available candidate proteins for calcium influx channels activated by agonist-stimulation of calcium signaling.

1. Such studies have identified at least two possible types of channels that can be activated in response to agonist-stimulation of PIP2 hydrolysis. One set of TRPC channels appears to be independent of internal Ca2+ store depletion, but requires IP3 generation and likely involvement of IP3R. Another set of channels can be activated by conditions used to deplete intracellular Ca2+ stores.

2. Knock-down studies suggest that some TRPCs might be components for calcium entry stimulated in response to PIP2 hydrolysis or internal Ca2+ store depletion (TRPC1, TRPC4). Thus, TRPCs can be used to identify regulatory mechanisms that are initiated by Ca2+ signaling events.

3. TRPCs appear to form complexes with Ca2+ signaling proteins and be localized within the same microdomains as Ca2+ signaling and SOCE. Thus, determining the localization of TRPC channels, their assembly into protein complexes, and their trafficking in cells will provide us valuable information.

4. TRPC channels have been proposed to exist as heteromers, based on studies, most of them done with heterologous expression systems. This has been used to account for the distinct characteristics of SOCE in different cell types. However, this is yet to be supported by data. For example, are the TRPCs that form heteromers in fact co-localized in cells? More importantly, we need to examine this with endogenous proteins and see which TRPCs are spatially segregated, which Ca2+ signaling complexes they interact with, and whether Ca2+ entry mediated by them has any physiological role in the cell.

5. TRPC proteins can be used to fish out novel proteins; e.g. the/an as yet unknown calcium channel or key scaffolding and regulatory proteins.

I am sure we will continue to discuss these issues vociferously in the coming weeks and beyond.

Session 3: Defining the Plasma Membrane Calcium Channels

May 24 2004 1:22PM

E-Conference Chairpersons

This session is focused on the main question of can we define theplasma membrane channels implicated in mediating calcium entry signals.

Specific issues include

(a) The properties of calcium channels:

How are specific conductances defined with respect to ion-selectivity, feedback regulation, and resemblance to or divergence fromthe "CRAC channel"?

(b) The role of known channel proteins:

Are proteins within the TRP superfamily of ion channels (includingmembers of the TRPC, TRPV, and TRPM channel families) mediators of calcium entry signals?

(c) The assembly and organization of channels:

Are different channel subunits assembled into heteromeric functionalcalcium entry channels? If so, how are they assembled and how doesassociation with regulatory elements mediate control over channelactivity?

How many SOCs are there?

Jun 1 2004 8:18AM

Reinhold Penner

Forum discussions often benefit from an initial controversial"spark". So letme make a provocative initial assessment that relates to the question "How many store-operated currents/channels are we dealing with?"

My provocative hypothesis would be that, although there is anabundance ofpapers that use the term store-operated for various (mostly non-selective) currents and channels, IMHO so far only one current has been establishedbeyond reasonable doubt as store-operated and that would be Icrac (Calcium Release-Activated Calcium-selective current).

I do not wish to be misunderstood. I do not exclude the possibilitythat otherstore-operated channels exist. However, none of the non-CRACconductances have been subjected to the same rigorous experimental testsfrom a number of independent labs that establish the final activation mechanism as unambiguously store-dependent. So what are the criteria one might want tosee fulfilled for a genuine store-operated current?

1. Obviously, there has to be a current that can be measured directly withelectrophysiological techniques. Fura-2 experiments or pharmacologicalmanipulations are indirect and not sufficient to demonstrate a store- operatedcurrent (even if the cell under investigation has been demonstrated topossess a store-operated current).

2. Independent and unrelated store-depletion protocols should alllead tonon-additive recruitment of the same conductance (e.g., IP3, BAPTA,ionomycin, thapsigargin, receptor agonists). Either alone is notsufficient toestablish store-dependence.

3. While cell lines can be good models for store-operated mechanisms, it isimperative that a store-operated mechanism is also found in native cellsthecells are derived from.

4. Although it is conceivable that calcium-impermeable channels could betriggered by store-depletion, the ultimate outcome is likely calcium entry andthe current should have some calcium permeability if not high selectivityforthe ion.

5. The currents should activate in a calcium-independent manner. That isunbuffered intracellular solutions should not lead to activation of SOCandmechanisms that prevent store-depletion (e.g. heparin or IP3 antagonistsshould prevent SOC activation without blocking the channels directly,whereas non-IP3-dependent store-depletion protocols such as ionomycin should still function under those conditions).

So, from my point of view, here is an ideal experiment that would goa longway to establish a SOC: In a whole-cell experiment, perfuse a cell with[Ca]ibuffered slightly above resting levels (say 200-300 nM using 10 mM BAPTA/ EGTA, add physiological MgATP and ideally K+ as physiological cation). Intheabsence of a store-depletion stimulus in control cells, this should causenoSOC activation for 10 min of whole-cell recording. Then perform the sameexperiment and induce store-depletion after 3-5 min by at least threedifferent protocols (e.g. brief ionomycin application, receptor agonist,releaseof caged IP3). All of these stimuli should cause a rapid and supramaximalSOC current. Finally, apply all three stimuli together (or bettersequentially) todemonstrate that they all converge on the current in a non-additivemanner.

6. Heterologuous expression or knock-out systems are dangerous forseveralreasons. Compensation mechanisms, up- or downregulation of endogenousSOC, aberrant signaling, reorganization of signaling mechanisms are allpossible caveats to keep in mind when interpreting results from suchsystems.

So in summary, I believe that it is possible to activate a number ofchannelsfollowing store depletion, but the crucial question as to whether thechannelsare truly store-operated is difficult to answer unless many differenttypes ofexperimental conditions are tested. A case in point that illustrates thisproblem is the fact that TRPM7 channels (termed MagnuM for Magnesium- Nucleotide regulated Metal ion currents in native cells) have beenmistaken asCRAC channels, simply because those channels are activated under the verysame experimental conditions that were previously considered tospecificallyonly activate CRAC channels. It takes some effort to dissociate theseconductances even under controlled patch-clamp conditions. Another case in point is that of TRPV5 (ECaC) and TRPV6 (CaT1), which both exhibit a great degree of similarity to CRAC channels, yet on close scrutiny are clearlydistinct and not store-operated.

I hope that these thoughts have stirred the pot and I would like tohear someopinions about which channels other than CRAC the forum members consideras established store-operated channels or which are close to beingclassifiedas such.

Aloha, -Reinhold


Reinhold Penner MD, PhD Center for Biomedical Research The Queen's Medical Center & University of Hawaii 1301 Punchbowl St. - UHT 8 Honolulu, HI 96813

How many SOCs are there?

Jun 1 2004 12:35PM

Randen L. Patterson

The experimental paradigm presented here is extremely rigorous forthe identification of store-operated currents or Icrac. Additionally,pointing out how many other molecularly identified channels have beenmistaken as SOC channels seems reasonable.

My question is how do we knowthat Icrac is even a calcium channel?

What is the experiemntal evidencethat this is indeed a calcium channel, not a high-speed exchanger, or co- transporter?

Can this be determined from whole-cell currents?

Can theidentity of a channel for Icrac be deduced from experiments where sodiumwas used as a current carrier not calcium?

As I am not anelectrophysiologist, I do not know, but I am curious about this topic.

Defining plasma membrane calcium channels

Jun 1 2004 8:52AM

Bernd Nilius

(a) The properties of calcium channels:

1. Do we ask the right question? The main question in this topic,“can we define plasma membrane channels mediating Ca2+ entry signals”, seems to be somewhat redundant because all Ca2+-permeable channels willmediate Ca2+ entry signals. Obviously, we will focus on “store-operatedCa2+-permeable channels” (SOCs). However, we should be aware that even a monovalent channel could couple to Ca2+ entry, e.g. in case where a“store-operated” Na+-permeable channel is tightly coupled to NCX (1).Possible candidates could therefore be searched in the whole range ofcationic channels, Ca2+-permeable or even Ca2+-impermeable non-selectivecation channels.

We should also be aware that manipulating “global” or “domain” [Ca2+]i will probably affect a plethora of store–dependentprocesses as well as Ca2+-permeable channels and we have to carefullydefine a feedback-independent measuring system (2). Needless to say,identifying the molecular nature of store-operated Ca2+ entry pathwayswill help to answer the questions how are SOCs activated.

2. What is required to define a “store-operated channel”? The maindifficulty therefore is: Can properly define that activation ofa Ca2+-permeable channel is indeed mediated by an intracellular Ca2+store? Even this question has not been solved properly.

a) It must be shown that a “channel” is activated by store depletionin a native system at a physiological temperature. This should bepreferentially done in a patch-clamp approach and the best would be toshow a close correlation between current activation/deactivation andchanges in the Ca2+ content of an intracellular Ca2+ store.

b) The most straightforward protocol would be showing activation of awhole-cell current -preferentially measured under perforated patchconditions- caused by store depletion and to show a single channelequivalent (as has been done for all voltage-operated Ca2+ channels).Under these conditions, currents through a highly selective Ca2+ channelmight be difficult to dissect. In this case, non-stationary noise analysis should be used (3). In case of channels with a low permeability for Ca2+,as for most of the TRP candidates, the whole cell approach should be moststraightforward and it is still surprising that this approach is widelyneglected. Single channel data alone are often worthless.

c) A quantitative correlation must be shown between intracellular Ca2+signals and SOC entry of Ca2+ taking into account Ca2+ buffering,extrusion and sequestration (4,5). This also requires that we knowbiophysical properties of SOC, e.g permeation and gating features, andthat we know its regulation under the chosen conditions, especiallymodulation by [Ca2+]i.

d) For identification of a molecular candidate, the critically defined“SOC” phenomenon should disappear when the candidate protein or geneis eliminated in vivo or in a native cell. This all together has not been accomplished yet.

3. Let us assume that we have clearly identified a SOC: The question isnow what should be the biophysical properties of such a channel. From thebiophysical point of view, it should be Ca2+ permeable. Obviously, thebest characterized store operated channel is CRAC, which has been measured in several cell types, particularly in blood cells(6). CRAC ischaracterized by a PCa/PNa ~ 1000, but a very low single channelconductance for monovalent cations (~0.1 pS) and in the fS range in thepresence of Ca2+(7,8). So far there is no real molecular candidate forsuch a channel. All other channels described as SOCs only partially matchthe criteria defined above and the question remains, whether they arereally SOCs.

(b) The role of known channel proteins:1. Are TRPs SOCs? Most intriguingly, TRP channels have been overwhelmingly welcomed as the missing molecular candidates for SOCs(2,9). But are TRPsreally SOCs? The evidence that depletion of intracellular Ca2+ storesinitiates or modulates activation of various TRP channels is overwhelming. Such experiments have been shown for all TRPCs (2) and also for TRPV6(10). TRPA1 can be activated by a store-depletion protocol but is unlikely a SOC (11). Most TRP channels have a low Ca2+ selectivity with a PCa/PNabetween 0.1 and 10 (2,12) and contribute undoubtly to Ca2+ entry. TRPM4and TRPM5 are very likely Ca2+ impermeable (13,14), however, TRPM5 hasalso been considered as store operated (15). Most of the studies were done in heterologous expression systems. Needless to say, we cannot rely on cell systems in which the signalling cascade between store and plasmamembrane might be altered, the correct stoichiometry might be violated, orthe correct players (subunits) might be missing. In addition, in anyoverexpression system the significance of local changes in Ca2+concentration in a domain around the channel might be dramatic, considering the extraordinary high Ca2+ sensitivity of nearly all TRPs.

The questionremains, whether TRPs fulfil the above-described criteria for SOCs.Referring to the last critrion (1d), only a few studies have been performed. In a trpc4-deficient model a SOC phenomenon disappeared (16). However, the biophysical properties of heterogeneously expressed TRPC4 channels do notmatch the “lost” currents (17,18). TRPC1 has been described in a knock -down approach as essential component of SOC (19). Recently, TRPC1, TRPC4, TRPV6 have all been described in native cells and using a knock-downapproach they were all found to participate in SOCs, although withdifferent modes of activation (20). Surprisingly, none of these studiesuses cell models from the respective knock-out models. None of thestudies, which are only a selection from many reports, clearly defines a“store-operated channels” according to the above-mentioned criteria.With no doubt, we cannot sufficiently answer this simple question yet.

2. TRPs and CRAC: The same dilemma is true for answering the question: Are TRPs CRACs? The only highly Ca2+ selective channels in the TRP family areso far TRPV5 and TRPV6 with PCa/PNa > 100 (2,21). Several features ofTRPV6 (and TRPV5) are identical with those of CRAC. However, single channelconductance, open pore block by intracellular Mg2+, and permeation for Cs+ (which all reflect pore properties) and also several pharmacologicalproperties substantially differ between TRPV6 and CRAC (22). Therefore,TRPV6 is very likely not CRAC. However, endogenous CRAC was markedlydepressed by expression of N-terminal TRPV6 fragments, indicating apossible modulatory role of TRPV6 on CRAC. However, these findingsunderline again that TRPV6 is not CRAC (23). So far, no CRAC channel seems to be present in the TRP family, at least judged from our knowledge fromexpressed channels including heteromers. In any case, the superficialidentification of TRPs as SOCs must be avoided.

In an alternative approach, however, we should consider that SOCs could be attributed to transporters (remember that most of the ATP driven primarytransporters, many exchangers, and even ClC-ec1 (24,25), the prokaryoticpredecessor of ClC channels, are electrogenic and see the very smallconductance of CRAC for monovalent cations (7)). Two other excitingexamples for a possible involvement of transporters in Ca2+ entry havebeen recently publisehd. The divalent metal transporter-1(DMT1/DCT1/Nramp2) can easily be converted into a Ca2+ channel by a single mutation, G185R, inducing a constitutive open, highly Ca2+ permeablechannel (26), and the mitochondrial Ca2+ uniporter is a highly Ca2+permeable channel (27). We should be aware that permeation pathways arealso coupled to transporters. Should these examples not already draw ourattention on novel candidates?

(c) The assembly and organization of channels:Is it worthwhile discussing this issue if we cannot answer the previousquestion? Can the dilemma that no reliable candidate for SOC is availablebe solved by assuming that known channels, including TRPs, formheteromers, assemble with subunits or form signalplexes with regulatoryproteins? Nobody will deny this possibilty, and from the philosophicalview of Karl Popper, this question cannot be falsified. Focusing on CRAC,which is the only reliably SOC described so far, the crucial experimentwill be whether the correct pore properties of CRAC will emerge in aheteromer or a signalplex. This has not yet been shown.

There is no doubtthat multimerisation occurs for many TRPs, such as TRPC1/4/5, TRPC3/6/7,TRPV1/3, TRPV5/6, TRPM4/5, TRPM6/7 and that heteromer formation changespermeation and kinetic properties of those channels. Partnership withother proteins acting in a signalling network has been most impressivelyshown for Drosophila TRPs (28). So far, various modulators of mammalianTRPs have been identified, including calmodulin, the inositol (1,4,5)trisphosphate receptor, ankyrin, NHERF, PI3-kinase, annexin 2, S100A10,caveolin-1, 80KH, PLCg, TrkA, RGA, MAP7, which probably act by directbinding to TRPs (2,12). The most intriguing assembling function might becredited to the adapter protein homer, which couples TRPC1 to the IP3R.Disassembly of the TRPC1-homer-IP3R complex parallels TRPC1 activation(29). However, even this example provides more evidence for TRPC1modulation by protein-protein interaction than for TRPC1 being a SOC.

Again, the bottom line, none of these many interactions described so farin the TRP family has solved the SOC dilemma. Again, we have to be openlooking beyond the now somewhat absorbing TRP hypothesis. Then thequestion remains, what else, if not TRPs?


1. Rosker, C., Graziani, A., Lukas, M., Eder, P., Zhu, M. X., Romanin, C., and Groschner, K. (2004) J Biol Chem 279, 13696-13704

2. Clapham, D. E. (2003) Nature 426, 517-524

3. Alvarez, O., Gonzalez, C., and Latorre, R. (2002) Adv Physiol Educ 26,327-341

4. Bautista, D. M., and Lewis, R. S. (2004) J Physiol 556, 805-817

5. Bautista, D. M., Hoth, M., and Lewis, R. S. (2002) J Physiol 541, 877- 894

6. Parekh, A. B., and Penner, R. (1997) Physiol Rev 77, 901-930

7. Prakriya, M., and Lewis, R. S. (2003) Cell Calcium 33, 311-321

8. Clapham, D. E. (2002) J Gen Physiol 120, 217-220

9. Montell, C. (2001) Science's STKEhttp://stke.sciencemag.org/cgi/content/full/OC-sigtrans;2001/90/re1 [Abstract] [Full Text]

10. Yue, L., Peng, J. B., Hediger, M. A., and Clapham, D. E. (2001) Nature 410, 705-709.

11. Jordt, S. E., Bautista, D. M., Chuang, H. H., McKemy, D. D., Zygmunt,P. M., Hogestatt, E. D., Meng, I. D., and Julius, D. (2004) Nature 427,260-265

12. Nilius, B. (2003) Cell Calcium 33, 293-298

13. Launay, P., Fleig, A., Perraud, A. L., Scharenberg, A. M., Penner, R., and Kinet, J. P. (2002) Cell 109, 397-407

14. Prawitt, D., Monteilh-Zoller, M. K., Brixel, L., Spangenberg, C.,Zabel, B., Fleig, A., and Penner, R. (2003) Proc Natl Acad Sci U S A 100,15166-15171

15. Perez, C. A., Huang, L., Rong, M., Kozak, J. A., Preuss, A. K., Zhang, H., Max, M., and Margolskee, R. F. (2002) Nat Neurosci 5, 1169-1176

16. Freichel, M., Suh, S. H., Pfeifer, A., Schweig, U., Trost, C.,Weissgerber, P., Biel, M., Philipp, S., Freise, D., Droogmans, G.,Hofmann, F., Flockerzi, V., and Nilius, B. (2001) Nat Cell Biol 3, 121- 127.

17. Schaefer, M., Plant, T. D., Obukhov, A. G., Hofmann, T., Gudermann,T., and Schultz, G. (2000) J Biol Chem 275, 17517-17526.

18. Plant, T. D., and Schaefer, M. (2003) Cell Calcium 33, 441-450

19. Vaca, L., and Sampieri, A. (2002) J Biol Chem 277, 42178-42187

20. Vanden Abeele, F., Lemonnier, L., Thebault, S., Lepage, G., Parys, J., Shuba, Y., Skryma, R., and Prevarskaya, N. (2004) J Biol Chem

21. Vennekens, R., Voets, T., Bindels, R. J., Droogmans, G., and Nilius,B. (2002) Cell Calcium 31, 253-264

22. Voets, T., Prenen, J., Fleig, A., Vennekens, R., Watanabe, H.,Hoenderop, J. G. J., Bindels, R. J. M., Droogmans, G., Penner, R., andNilius, B. (2001) J Biol Chem 276, 47767-47770

23. Kahr, H., Schindl, R., Fritsch, R., Heinze, B., Hofbauer, M., Hack, M. E., Mortelmaier, M. A., Groschner, K., Peng, J. B., Takanaga, H., Hediger, M. A., and Romanin, C. (2004) J Physiol 557, 121-132

24. Gadsby, D. C. (2004) Nature 427, 795-797

25. Accardi, A., and Miller, C. (2004) Nature 427, 803-807

26. Xu, H., Jin, J., DeFelice, L. J., Andrews, N. C., and Clapham, D. E.(2004) PLoS Biol 2, E50

27. Kirichok, Y., Krapivinsky, G., and Clapham, D. E. (2004) Nature 427,360-364

28. Montell, C. (2003) Cell Calcium 33, 409-417

29. Yuan, J. P., Kiselyov, K., Shin, D. M., Chen, J., Shcheynikov, N.,Kang, S. H., Dehoff, M. H., Schwarz, M. K., Seeburg, P. H., Muallem, S.,and Worley, P. F. (2003) Cell 114, 777-789

The importance of terminology and electrophysiology

Jun 3 2004 7:58AM

Trevor Shuttleworth

I think it is clear that the field is suffering under a profoundconfusion in terminology – a fact that was probably a major drivingforce for initiating this forum.

In some areas, terminology can devolveinto arcane semantics, but this is certainly not the case for Ca2+ entrychannels where often the same term is used for obviously widely differentphenomena. Part of the problem is that we are often trying to define achannel, or at least a Ca2+ entry pathway, by techniques that are reallyhopelessly inadequate to make such definitions. Reinhold and Bernd havealready alluded to this in their postings.

The use of changes in cytosolic Ca2+ based on fluorescence measurements is simply not suitable to define a channel. Similarly, all the pharmacological tools currently available forsuch channels are so poor that they cannot be used with any confidence todiscriminate between different pathways for entry.

The only way we caneven begin to characterize the pathways operating under various conditions in any given cell type is by measuring the channel properties themselvesusing electrophysiological techniques. However, even this is not withoutsignificant problems. Because it appears that many of these Ca2+ entrychannels have such small conductances, almost all studies have to usewhole-cell techniques, with the associated problems of contamination fromother currents etc. that has already bedeviled the field.

So, based on what we know, what criteria can we use to definethese Ca2+ entry pathways?

First, we know something, but not nearlyenough, about their Ca2+ selectivity. Thus, there are entry channels that, under normal conditions, show a high selectivity for Ca2+ [e.g. CRAC (Ca(2+) release-activated Ca2+ channels) andARC (arachidonate-regulated Ca2+ channels)], in contrast to others that are essentially nonselectivecation channels. Although both can contribute to overall Ca2+ entry, oneimportant difference is that a major physiological effect of the latterwill be to depolarize the cell. This could have an important consequencefor Ca2+ signaling in that it would result in the opening of voltage-gated Ca2+ channels in those cell types that possess such channels (e.g. smoothmuscle cells).

The real problems arise when we try to use the mode of activation asa criterion for classification – because even our basic understanding of these processes is so poor. Most importantly, the direct means ofactivation is almost always unknown.

There do seem to be channels that are activated only as a consequence of store depletion, and others that areactivated only as a result of receptor activation. The key word here, ineach case, is "only". Obviously, any member of the first group could beactivated by receptor activation, providing the receptor couples to thegeneration of IP3 (or some other store-depleting messenger) and that thisis produced in sufficient quantity to effect an adequate store depletion.However, these should be distinguished from those channels whoseactivation is solely dependent on receptor activation and is entirelyindependent of store depletion. An appropriate test here might be whetherthe channel activity can be increased by appropriate receptor activationafter maximal depletion of the intracellular stores.

As to the role of known channel proteins – here we are obviouslytalking about the TRP family, as these are currently "the only game intown". In this, I will just focus on the TRPC family.

Reinhold and Berndhave already alluded to the many potential problems of expression andknock-down approaches to the study of these proteins. However, to me, thefundamental problem with the "TRPC hypothesis" is that it is clear that,in many cell types, the endogenous Ca2+ entry channels – both store- operated, and noncapacitative – are of the highly Ca2+-selective type(see above). As such, none of the known TRPC family members fit the billas candidates for these channels.

For some time now, the standard argument has been that the TRPCs form heteromultimers that may possess uniquecharacteristics that differ from either, or any, of their individualcomponents. However, Ca2+ selectivity is likely to be a fundamentalproperty of the channel pore itself, and none of the few published reports on such TRPC heteromeric channels show the required levels of Ca2+selectivity. It seems curious, given the widespread use to the"heteromeric channel" argument, that so few studies have appeared wheresuch channels have been investigated.

Fluorescent calcium imaging or electrophysiology, or both?

Jun 4 2004 8:34AM

Randen L. Patterson

As a researcher who uses fluorescence and fura to study calcium signaling, I feel it necessary to retort a few of Trevor's comments.

First, these statements,
"The use of changes in cytosolic Ca2+ based on fluorescence measurements is simply not suitable to define a channel. Similarly, all the pharmacological tools currently available for such channels are so poor that they cannot be used with any confidence to discriminate between different pathways for entry. The only way we can even begin to characterize the pathways operating under various conditions in any given cell type is by measuring the channel properties themselves using electrophysiological techniques."
is not correct.

Electrophysiology (particularly whole-cell patching) can only provide the biophysical properties of the channel, not the molecular identity, unless coupled with other techniques, just as fluorescent calcium entry research. In addition, even if single channel recordings could be taken of these channels, it is performed in the ABSENCE of every other channel. Calcium signals are integrated into the receptor response, and have inputs and outputs to and from other ion channels. Therefore, study of singlechannels, although important to understanding channel properties, is a poor system for determining the "behavior" of these channels, or their role in cellular physiology.

In my opinion, fluorescent calcium imagingand electrophysiology are extremely complementary, and perhaps more "fura-people" should be teaming up with more "patchers", each lab using their expertise to truly begin getting a grip on what these channels do in physiology.

Secondly, what is the definition of calcium selective channels anyways?
Many of the researchers that study calcium entry channels can see ions, such as strontium and barium, passing into cells after receptor stimulation AT MILLIMOLAR CONCENTRATIONS!!! Strontium is found in human blood below 1 micromolar, and barium at 41-96 micromolar, while calcium is millimolar. Therefore, even if these channels are "non-selective" for the purposes of defining biophysical properties, in the body they are fluxing calcium, and should not be considered "non-selective" when trying to determine their role incellular function.

I do agree with Trevor's comments on heteromultimer TRPC channels. As a field, this argument is bandied about all the time, although very few studies have addressed this exact question, not that it's an easyquestion to address. Only by trying to isolate homomultimer and heteromultimer TRPC channels in vivo, will we truly know the selectivity of these channels. Even with the advent of siRNA, it's still difficultto look at the loss of a current, as opposed to gaining a current.

Unfortunately, overexpression studies of TRPC channels are in part why we are in this quandary today. Each lab uses their own protocol for transient expression, all of which will differ in their expression as homomultimers and heteromultimers, and therefore provide differentresults. There is no easy answer to this problem, but that does not negate its importance.

Re: Randen's comments on defining channels

Jun 4 2004 3:45PM

Trevor Shuttleworth

I would have to take issue with a couple of the responses Randon hasmade to my original posting (but the this is what this all supposed to beabout, right?)

1. "Electrophysiology (particularly whole-cell patching) can onlyprovide the biophysical properties of the channel, not the molecularidentity".This is my point precisely. In the absence of any clear molecular clues,our best bet is to precisely characterize the biophysical fingerprint ofthe channel. At least that way, if any molecular candidate is proposed wehave the ability to compare the molecular properties of the candidate with those of the endogenous channels in a preciuse way. It exactly this kindof analysis that leads me to have a problem with the "TRPC argument".There is simply no way that "fluorescent calcium entry" recordings canprovide such a fingerprint.

2. "In addition, even if single channel recordings could be taken of these channels, it is performed in the ABSENCE of every other channel.Calcium signals are integrated into the receptor response, and have inputs and outputs to and from other ion channels. Therefore, study of singlechannels, although important to understanding channel properties, is apoor system for determining the "behavior" of these channels, or theirrole in cellular physiology." I agree, but you are talking about the behavior and role of the channel in signaling. Whilst I would be the first to agree that these are essentialcomponents of our understanding of the channel, I do not agree that theyoffer much to help in its identification.

3. "Secondly, what is the definition of calcium selective channelsanyways?" I think I may have misled you here. I was using the terms "Ca2+-selective" and "non-selective" in relation to the relative permeabilities of Ca2+versus monovalent cations (i.e. Na+ in normal physiological conditions). I would say that channels such as CRAC, ARC, and the ECaCs (TRPV5 and 6) are clearly highly Ca2+-selective (pCa2+:pNa+ >>100). In contrast, manyof the non-selective conductances seen, for example, in smooth musclecells, as well as most of the reports on TRPCs indicate pCa2+:pNa+ <10. Although these will conduct Ca2+ and may make a contribution to overallCa2+ entry, I would say a reasonable classification for these is "non-selective cation channels". Personally, I find the Ba2+/Sr2+ story formost entry pathways very confusing and not very informative.

Re: Randen's comment on defining channels

Jun 7 2004 8:05AM

Randen L. Patterson

A small rebuttle 1)"There is simply no way that "fluorescent calcium entry" recordings canprovide such a fingerprint"

Wouldn't siRNA deletion of a calcium channel, and looking at a change in bulk calcium flow using fura provide a molecular fingerprint?

2)"I agree, but you are talking about the behavior and role of thechannel in signaling. Whilst I would be the first to agree that these areessential components of our understanding of the channel, I do not agreethat they offer much to help in its identification. "

So should we as a field abandon the ideas of proteomic approaches tolink proteins which we know can influence calcium entry to the channelsthemselves?

3)"Personally, I find the Ba2+/Sr2+ story for most entry pathwaysvery confusing and not very informative."

I agree entirely on this point. There are papers that exist in theliterature that talk about TRPC channels being calcium channels, yet in notone place in the entire manuscript do they look at calcium flux throughthese TRPC channels. In fact, has anyone demonstrated that anoverexpressed TRPC channel carries calcium at all? How can they be surethat the currents they are observing are not from a heteromultimer withendogenous machinery? (back to Trevor's original arguement)


Jun 7 2004 11:01AM

Peter J Lockyer

Systematic RNAi seems the way to clearly define the channel responsible for Icrac.

It's also the way to sort out the messy heterologous TRP expression data produced over more than a decade in various cell lines.

Presumably all of this is being done (probably has been already in big pharma) and the identity of ARCs can also be determined this way.

I don't keep a close enough eye on the TRP field, but there doesn't seem to have been a great deal of knockdown data presented in support of any TRP channels mediating forms of Ca entry, much still relies on overexpression. Is this a reasonable statement?

Re: Randen's comment on defining channels

Jun 8 2004 2:34PM

Trevor Shuttleworth

Re: Tthe channel fingerprint -- I think maybe we are talking about two different things. The experiment you describe would certainly helpidentify the contribution of a particular channel in cellular responses(its "functional" fingerprint if you like).

However, specific siRNAtechniques (as opposed to systematic siRNA screening) require knowledge of the molecular identity of the channels involved. In the case of at least two widely-expressed, highly Ca2+-selective channels involved in Ca2+ entry (CRAC and ARC channels), this is not known.

I think of our currentsituation much like a criminal investigation. If you already have aclearly identified suspect in custody, then an appropriate DNA test mayput the suspect at the scene of the crime and clinch the case (the siRNAexperiment).

But, if you have no real idea who the perpetrator is, youhave to start by building a file of critical bits of information based onthe evidence available that may help narrow down the field and direct your search in an appropriate direction -- if entry to the scene of the crimewas achieved via a 12 inch hole-- it is no good looking at any NFLlinemen as potential suspects!