JPhysiology (1992) 86, 135-138 © Elsevier, Paris
135
Presynaptic calcium concentration microdomains and transmitter release R L l i n f i s a, M S u g i m o r i a, R B S i l v e r b a Department of Physiology and Biophysics, New York University Medical Center, New York, New York; b Section and Department of Physiology, Cornell University, Ithaca, New York, USA
Summary - n-Aequorin J, a luminescent protein which responds to calcium concentration changes in the order of several hundred micromoles, was injected into the preterminal fiber in the squid giant synapse. The activation of the presynaptic terminal leading to release of transmitter was accompanied by light emission at well-defined sites at the active zone in
the presynaptic terminal. Location of these light emission sites was very much the same from one stimulus to the next, indicating that light emission was triggered by the inward calcium current occurring at specific and invariant locations. The distribution, size and number of these QEDs (quantum emission domains) coincides well with the location and number of active zones in the presynaptic terminal. The results imply that transmitter release is triggered by very well-localized calcium concentration changes that may be as high as several hundred micromoles. aequorin / transmitter release / video imaging
It is known that local elevation of intracellular calcium concentration is the trigger for neurotransmitter release at most chemical synapses, however, the precise mechanisms by which calcium activates such release are presently unknown (Katz, 1969). The initial working hypothesis couching the relationship between intracellular calcium concentration ([Ca2+]i) and transmitter release was based on results from the neuromuscular junction (Dodge and Rachamimoff, 1967), and was proposed as a single-compartment release model. According to this model, vesicles are exocytosed in a fourth-order relationship with respect to [Ca 2+ ]i. While offering a useful paradigm that was state-of-the art for the time, the single-compartment release model did not incorporate the many variables that regulate factors such as [Ca2+]i or vesicular availability. Within this paradigm, calcium was thought to flow into the preterminal uniformly across the preterminal membrane through synchronously activated channels. Furthermore, this model of transmitter release did not adequately consider the ultrastructure of the presynaptic terminal. Present information regarding such parameters demands a more complete, multicompartment model.
Indeed, the possibility that calcium channels are localized at discrete sites in the preterminal was suggested in the late '70s (Llin~is, 1977). Resuits of voltage-clamp experiments revealed the latency between the tail-current calcium entry and transmitter release to be as short as 200 gs (Llin~is, 1977; Llin~s et al, 1981). Recent histological evidence confirmed the morphological prerequisite of synaptic vesicles being directly apposed to the calcium channels (Robitaille et al, 1990). From voltage-clamp data the m a x i m u m [Ca2+]i against the membrane was calculated to the about 10-ZM (Llinfts et al, 1981). As a result of the information elicited via these studies the experimental focus shifted from cytosolic residual calcium to the problem of what is now known as calcium microdomains (Chad and Eckert, 1964; Simon et al, 1984; Simon and Llin~is, 1985; Fogelson and Zucker, 1985). The term 'calcium microdomains' refers to a very precise distribution of spatially limited sites for [Ca2+]i change. These microdomains, which may be described as small domes of increased [Ca 2+]i, were expected to occur at the cytoplasmic surface of the presynaptic terminal membrane within an active zone. In fact, each active calcium channel is thought to produce a
136 rapid (microseconds) increase in [Ca2+]i, lasting for the duration of the average open time of the channel (Simon and Llin~is, 1985) which then rapidly returns to the pre-opening value when the Ca 2÷ channel closes (Simon and Llin~s, 1985). The calcium influx is thought to generate a [Ca2+]i profile as high as 200-300 gM in the proximity of the calcium channels. In this case, transmitter release would be triggered by extraordinarily high transient changes in localized calcium concentration within the immediate vicinity of the presynaptic release sites where the vesicles are lodged (Ltin~is et al, 1981; Simon and Llin~s, 1985). To test this hypothesis and to study directly the existence of such [Ca2+]i microdomains, a special type of signaling methodology was introduced. A hybrid synthetic n-aequorin J, having a sensitivity to [Ca2+]i in the order of 10-~M, was developed by Drs Shimomura, Inouye and Kishi, (Shimomura et al, 1990) and generously provided through Dr Shimomura. This aequorin - a photoprotein that emits light upon binding free Ca 2÷ was injected presynaptically in the squid giant synapse. Of the total aequorin injected, 5% was a fluorescein-conjugated recombinant apoaequorin, whose fluorescence emission facilitated intracellular localization (Llin~is et al, 1992). Use of the 'low sensitivity' aequorin preparation permitted selective detection of high calcium concentration microdomains (Llin~s et al, 1992). Distribution of the injected aequorin following impalement and injection of the presynaptic terminal of the giant synapse of the squid Loligo pealii was visualized with a fluorescence microscope using a 40 X water-immersion lens. Aequorin luminescence was then detected with a VIM camera operated in the photon-counting mode; the images were stored on videotape and characterized by digital image processing and analysis methods. Results were obtained from 27 different synapses bathed in artificial seawater (10 mM Ca2+); one synapse was injected with normal aequorin and the other 26 were injected with the n-aequorin J. Upon tetanic stimulation, small points of light localized in time and space were detected over the preterminal region in the area of the 'active zone' (Llin~is et al, 1992). These points had an average diameter of approximately 0.5 Ixm and were distributed over roughly 5-10% of the total area of the presynaptic membrane (with an average of 8.4 gm 2 per 100 gin2). Resuits from all synapses were quite similar. They demonstrate that the portion of the presynaptic terminal forming the active zone emits light
during presynaptic activation, indicating that the calcium concentration is elevated in the range of 10-4M during this active period. The location of these small light sources was determined by two methodologies. First, images of light-emission points were digitized and integrated into frame buffer over several seconds of stimulation and the spatial and temporal distribution of the blips was studied. Second, to ensure that the location of the patches of blips was similar within repeating stimuli, sets of successive temporal image integrations of the recording were compared. Both methods illustrated a similar distribution of microdomains and roughly the same size and intensity of light points. Detailed analysis of the temporal and spatial distribution of quantum emission domains (QEDs), which appeared as white spots, suggests that once a microdomain is activated the probability of a second near-term activation is decreased. Under most conditions QEDs occurred as singlets or doublets and did not occur at hight frequency at any particular microdomain. Analysis of the activation dynamics of individual microdomains suggested a quasisequential activation, that is, a low probability of immediate re-activation (fig 1). One possible mechanism for this refractory period may be related to the high [Ca2+]i. The results also suggest that microdomains may belong to one of two varieties: 1) those frequently activated; and 2) those that, while repeating, activate less often. Analysis of digitized subregions of the image field demonstrated that the temporal and spatial distribution of microdomain action during stimulation over time reflect the temporal cycling of active sites from one point of membrane to the next. This finding suggests that an even more complex modulation of activity may be present in these terminals. Indeed, in addition to calcium microdomains, compartmentalization of other parameters such as the degree of phosphorylation of synapsin I (Llin~s et al, 1985) may also show dynamic interaction, which would ensure a fine control of transmitter release from one impulse to the next. Of interest is the possibility that a certain number of calcium channels may have to be active simultaneously in order to activate the aequorin signals observed here. In such a case, the high level of [Ca2+]i obtained at the microdomain may itself reduce, by a 'lateral inhibition' type effect, the probability of further calcium channel activation in a given active site.
137
Also, it is well documented that normally an average of one vesicle is released per active zone (Pumplin et al, 1981; Korn et al, 1982; Triller and Korn, 1982). Invoking 'lateral inhibition' by
A
B
C
D
activation of a certain number of channels in an active zone patch (average measured area, 0.313 btm2) would temporarily depress local channel activity. If this were so a special kinetics
10
QED l0 7.5 um Fig 1. Graphical representation of the probability of calcium entry and elevation of intracellular calcium concentration ([Ca2+]i) among consecutive 30-s periods during continuous stimulation of the nerve. Pseudo color rendering was used to aid in visualization; black and purple and blue are lowest probabilities, while red and white represent the highest probabilities for Ca 2+ entry to a particular region of the preterminal. A. A pair of 30-s integrations of QEDs from the same portion of a preterminal which have been processed to show the probability of Ca 2+ entry to a given region during that period of stimulation. Experimental details are presented in Llin~is et al (1992). Left. The pattern of QEDs recorded during the first 30 s of stimulation. Right. ,+ The pattern of QEDs recorded during the following 30-s period. Note the similarity in the overall patterns of microdomains of [Ca- ]i between the paired panels left and right. B. The intensity profile pattern of the images displayed in A. In this display, increasing probability of a QED occurring at any site in the image field was projected vertically, yielding an intensity relief map of the probability of QEDs occurring during the sampling period. Left. The intensity profile pattern for QEDs recorded during the first 30-s of stimulation. Right. The intensity profile pattern of QEDs recorded during the following 30-s period. Note the selective elevation of . . . . . . . . . . . 2+ [Ca2+]i within particular mlcrodomams, as well as the stmdanty m the overall patterns of mmrodomams of [Ca ]i between the paired panels left and right. C. An intensity profile display depicting the probability of QEDs occurring at a particular location within the preterminal during the same 30-s period represented in A, left and B, left. Here, as in D, a grid of white vertical marker 'poles' was embedded into the image as an aid in establishing spatial registration for individual domes of QEDs between C and D. These displays have also been projected with a 50 ° offset from the perspective of B to aid visualization of the QED domes and the marker poles. D. A similar intensity profile display depicting the probability of QEDs occurring at a particular location within the preterminal during the 30-s period represented in A, right and B, right. As in C, a grid of white vertical marker poles was embedded into the image, and the image was projected with a 50 ° offset as aids to establishing spatial registration for individual domes of QEDs between C and D.
138 would operate in which the probability of release is related tO the previous activitj n any given active zone. This prospect adds an ;~:teresting new variable to consider and incorporate into testable models of transmitter release. The distribution of Ca 2÷ microdomains suggests that these sites are active zones where increased [Ca2+]i triggers neurotransmitter release by binding to a low-affinity Ca 2÷ binding site at the presynaptic vesicles and activating the release process (Llin~s et aL 1992). Such a mechanism would safeguard the synapse from large amounts of spontaneous transmitter release, since probably more than one Ca 2÷ channel must be activated per active zone to trigger exocytosis (Simon and Llin~is, 1985). Such a scheme would also permit sufficient time to replace the expended vesicles (Simon and Llin~s, 1985). The short refractory delay between Ca 2+ entry and transmitter release suggests that only vesicles near the QEDs would be released by a given action potential (Llings, 1977). Lastly, the fact that [Ca2+]i attains such high concentrations at release sites must be taken into account in the study of the mechanisms of membrane fusion.
References Katz B (1969) The Release of Neurotransmitter Substances: The Sherrington Lectures X. Charles C Thomas, Springfield, IL Dodge FA, Rachamimoff R (1967) Co-operative action of calcium ions in transmitter release sites at the neuromuscular junction. J Physiol (Lond) 193, 419-432 LliMs R (1977) Calcium and transmitter release in squid synapse. In: Approaches to the Cell Biology of Neurons. (Cowan WM, Ferendelli JA eds) Society for Neurosciences, Bethesda, MD, 139-160 LliMs R, Steinberg I, Walton K (1981) Presynaptic calcium currents in squid giant synapse. Biophys J 33, 28%322
Robitaille R, Adler EM, Charlton M (1990) Strategic location of calcium channels at transmitter release sites of frog neuromuscular synapses. Neuron 5, 773-779 Chad JE, Eckert R (1984) Calcium domains associated with individual channels may account for anomalous voltage relations of Ca-dependent responses. Biophys J 45, 993-999 Simon SM, Sugimori M, Llin~s R (1984) Modelling of submembranous calcium-concentration changes and their relation to rate of presynaptic transmitter release in the squid giant synapse. Biophys J 45, 264a Simon SM, Llinfis R (1985) Compartmentalization of the submembrane calcium activity during calcium influx and its significance in transmitter release. Biophys J 48, 485-498 Fogelson AL, Zucker RS (1985) Presynaptic calcium diffusion from various arrays of single channels. Implication for transmitter release and synaptic facilitation. Biophys J 48, 1003-1007 Shimomura O, Inouye S, Musicki B, Kishi Y (1990) Recombinant aequorin and recombinant semi-synthetic aequorins: cellular Ca2+ ion indicators. Biochem J 270, 309-312 Llintis R, Sugimori M, Silver B (1992) Microdomains of high calcium concentration in a presynaptic terminal. Science 256, 677-679 LliMs R, McGuinness TL, Leonard CS, Sugimori M, Greengard P (1985) Intraterminal injection of synapsin I or calcium/calmodulin-dependent protein kinase II alters neurotransmitter release at the squid giant synapse. Proc Natl Acad Sci USA 82, 30353039 Pumplin DD, Reese TW, Llin~s R (1981) Are the presynaptic membrane particles the calcium channels? Proc Natl Acad Sci USA 78, 7210-7213 Korn J, Mallet A, Triller A, Faber DS (1982) Transmission at a central synapse. II. Quantal description of release with a physical correlation for binomial n. J Neurophysiol 48, 679 Triller A, Korn H (1982) Transmission at a central inhibitory synapse. III. Ultrastructure of physiologically identified terminals. J Neurophysiot 48 708-736
135
Presynaptic calcium concentration microdomains and transmitter release R L l i n f i s a, M S u g i m o r i a, R B S i l v e r b a Department of Physiology and Biophysics, New York University Medical Center, New York, New York; b Section and Department of Physiology, Cornell University, Ithaca, New York, USA
Summary - n-Aequorin J, a luminescent protein which responds to calcium concentration changes in the order of several hundred micromoles, was injected into the preterminal fiber in the squid giant synapse. The activation of the presynaptic terminal leading to release of transmitter was accompanied by light emission at well-defined sites at the active zone in
the presynaptic terminal. Location of these light emission sites was very much the same from one stimulus to the next, indicating that light emission was triggered by the inward calcium current occurring at specific and invariant locations. The distribution, size and number of these QEDs (quantum emission domains) coincides well with the location and number of active zones in the presynaptic terminal. The results imply that transmitter release is triggered by very well-localized calcium concentration changes that may be as high as several hundred micromoles. aequorin / transmitter release / video imaging
It is known that local elevation of intracellular calcium concentration is the trigger for neurotransmitter release at most chemical synapses, however, the precise mechanisms by which calcium activates such release are presently unknown (Katz, 1969). The initial working hypothesis couching the relationship between intracellular calcium concentration ([Ca2+]i) and transmitter release was based on results from the neuromuscular junction (Dodge and Rachamimoff, 1967), and was proposed as a single-compartment release model. According to this model, vesicles are exocytosed in a fourth-order relationship with respect to [Ca 2+ ]i. While offering a useful paradigm that was state-of-the art for the time, the single-compartment release model did not incorporate the many variables that regulate factors such as [Ca2+]i or vesicular availability. Within this paradigm, calcium was thought to flow into the preterminal uniformly across the preterminal membrane through synchronously activated channels. Furthermore, this model of transmitter release did not adequately consider the ultrastructure of the presynaptic terminal. Present information regarding such parameters demands a more complete, multicompartment model.
Indeed, the possibility that calcium channels are localized at discrete sites in the preterminal was suggested in the late '70s (Llin~is, 1977). Resuits of voltage-clamp experiments revealed the latency between the tail-current calcium entry and transmitter release to be as short as 200 gs (Llin~is, 1977; Llin~s et al, 1981). Recent histological evidence confirmed the morphological prerequisite of synaptic vesicles being directly apposed to the calcium channels (Robitaille et al, 1990). From voltage-clamp data the m a x i m u m [Ca2+]i against the membrane was calculated to the about 10-ZM (Llinfts et al, 1981). As a result of the information elicited via these studies the experimental focus shifted from cytosolic residual calcium to the problem of what is now known as calcium microdomains (Chad and Eckert, 1964; Simon et al, 1984; Simon and Llin~is, 1985; Fogelson and Zucker, 1985). The term 'calcium microdomains' refers to a very precise distribution of spatially limited sites for [Ca2+]i change. These microdomains, which may be described as small domes of increased [Ca 2+]i, were expected to occur at the cytoplasmic surface of the presynaptic terminal membrane within an active zone. In fact, each active calcium channel is thought to produce a
136 rapid (microseconds) increase in [Ca2+]i, lasting for the duration of the average open time of the channel (Simon and Llin~is, 1985) which then rapidly returns to the pre-opening value when the Ca 2÷ channel closes (Simon and Llin~s, 1985). The calcium influx is thought to generate a [Ca2+]i profile as high as 200-300 gM in the proximity of the calcium channels. In this case, transmitter release would be triggered by extraordinarily high transient changes in localized calcium concentration within the immediate vicinity of the presynaptic release sites where the vesicles are lodged (Ltin~is et al, 1981; Simon and Llin~s, 1985). To test this hypothesis and to study directly the existence of such [Ca2+]i microdomains, a special type of signaling methodology was introduced. A hybrid synthetic n-aequorin J, having a sensitivity to [Ca2+]i in the order of 10-~M, was developed by Drs Shimomura, Inouye and Kishi, (Shimomura et al, 1990) and generously provided through Dr Shimomura. This aequorin - a photoprotein that emits light upon binding free Ca 2÷ was injected presynaptically in the squid giant synapse. Of the total aequorin injected, 5% was a fluorescein-conjugated recombinant apoaequorin, whose fluorescence emission facilitated intracellular localization (Llin~is et al, 1992). Use of the 'low sensitivity' aequorin preparation permitted selective detection of high calcium concentration microdomains (Llin~s et al, 1992). Distribution of the injected aequorin following impalement and injection of the presynaptic terminal of the giant synapse of the squid Loligo pealii was visualized with a fluorescence microscope using a 40 X water-immersion lens. Aequorin luminescence was then detected with a VIM camera operated in the photon-counting mode; the images were stored on videotape and characterized by digital image processing and analysis methods. Results were obtained from 27 different synapses bathed in artificial seawater (10 mM Ca2+); one synapse was injected with normal aequorin and the other 26 were injected with the n-aequorin J. Upon tetanic stimulation, small points of light localized in time and space were detected over the preterminal region in the area of the 'active zone' (Llin~is et al, 1992). These points had an average diameter of approximately 0.5 Ixm and were distributed over roughly 5-10% of the total area of the presynaptic membrane (with an average of 8.4 gm 2 per 100 gin2). Resuits from all synapses were quite similar. They demonstrate that the portion of the presynaptic terminal forming the active zone emits light
during presynaptic activation, indicating that the calcium concentration is elevated in the range of 10-4M during this active period. The location of these small light sources was determined by two methodologies. First, images of light-emission points were digitized and integrated into frame buffer over several seconds of stimulation and the spatial and temporal distribution of the blips was studied. Second, to ensure that the location of the patches of blips was similar within repeating stimuli, sets of successive temporal image integrations of the recording were compared. Both methods illustrated a similar distribution of microdomains and roughly the same size and intensity of light points. Detailed analysis of the temporal and spatial distribution of quantum emission domains (QEDs), which appeared as white spots, suggests that once a microdomain is activated the probability of a second near-term activation is decreased. Under most conditions QEDs occurred as singlets or doublets and did not occur at hight frequency at any particular microdomain. Analysis of the activation dynamics of individual microdomains suggested a quasisequential activation, that is, a low probability of immediate re-activation (fig 1). One possible mechanism for this refractory period may be related to the high [Ca2+]i. The results also suggest that microdomains may belong to one of two varieties: 1) those frequently activated; and 2) those that, while repeating, activate less often. Analysis of digitized subregions of the image field demonstrated that the temporal and spatial distribution of microdomain action during stimulation over time reflect the temporal cycling of active sites from one point of membrane to the next. This finding suggests that an even more complex modulation of activity may be present in these terminals. Indeed, in addition to calcium microdomains, compartmentalization of other parameters such as the degree of phosphorylation of synapsin I (Llin~s et al, 1985) may also show dynamic interaction, which would ensure a fine control of transmitter release from one impulse to the next. Of interest is the possibility that a certain number of calcium channels may have to be active simultaneously in order to activate the aequorin signals observed here. In such a case, the high level of [Ca2+]i obtained at the microdomain may itself reduce, by a 'lateral inhibition' type effect, the probability of further calcium channel activation in a given active site.
137
Also, it is well documented that normally an average of one vesicle is released per active zone (Pumplin et al, 1981; Korn et al, 1982; Triller and Korn, 1982). Invoking 'lateral inhibition' by
A
B
C
D
activation of a certain number of channels in an active zone patch (average measured area, 0.313 btm2) would temporarily depress local channel activity. If this were so a special kinetics
10
QED l0 7.5 um Fig 1. Graphical representation of the probability of calcium entry and elevation of intracellular calcium concentration ([Ca2+]i) among consecutive 30-s periods during continuous stimulation of the nerve. Pseudo color rendering was used to aid in visualization; black and purple and blue are lowest probabilities, while red and white represent the highest probabilities for Ca 2+ entry to a particular region of the preterminal. A. A pair of 30-s integrations of QEDs from the same portion of a preterminal which have been processed to show the probability of Ca 2+ entry to a given region during that period of stimulation. Experimental details are presented in Llin~is et al (1992). Left. The pattern of QEDs recorded during the first 30 s of stimulation. Right. ,+ The pattern of QEDs recorded during the following 30-s period. Note the similarity in the overall patterns of microdomains of [Ca- ]i between the paired panels left and right. B. The intensity profile pattern of the images displayed in A. In this display, increasing probability of a QED occurring at any site in the image field was projected vertically, yielding an intensity relief map of the probability of QEDs occurring during the sampling period. Left. The intensity profile pattern for QEDs recorded during the first 30-s of stimulation. Right. The intensity profile pattern of QEDs recorded during the following 30-s period. Note the selective elevation of . . . . . . . . . . . 2+ [Ca2+]i within particular mlcrodomams, as well as the stmdanty m the overall patterns of mmrodomams of [Ca ]i between the paired panels left and right. C. An intensity profile display depicting the probability of QEDs occurring at a particular location within the preterminal during the same 30-s period represented in A, left and B, left. Here, as in D, a grid of white vertical marker 'poles' was embedded into the image as an aid in establishing spatial registration for individual domes of QEDs between C and D. These displays have also been projected with a 50 ° offset from the perspective of B to aid visualization of the QED domes and the marker poles. D. A similar intensity profile display depicting the probability of QEDs occurring at a particular location within the preterminal during the 30-s period represented in A, right and B, right. As in C, a grid of white vertical marker poles was embedded into the image, and the image was projected with a 50 ° offset as aids to establishing spatial registration for individual domes of QEDs between C and D.
138 would operate in which the probability of release is related tO the previous activitj n any given active zone. This prospect adds an ;~:teresting new variable to consider and incorporate into testable models of transmitter release. The distribution of Ca 2÷ microdomains suggests that these sites are active zones where increased [Ca2+]i triggers neurotransmitter release by binding to a low-affinity Ca 2÷ binding site at the presynaptic vesicles and activating the release process (Llin~s et aL 1992). Such a mechanism would safeguard the synapse from large amounts of spontaneous transmitter release, since probably more than one Ca 2÷ channel must be activated per active zone to trigger exocytosis (Simon and Llin~is, 1985). Such a scheme would also permit sufficient time to replace the expended vesicles (Simon and Llin~s, 1985). The short refractory delay between Ca 2+ entry and transmitter release suggests that only vesicles near the QEDs would be released by a given action potential (Llings, 1977). Lastly, the fact that [Ca2+]i attains such high concentrations at release sites must be taken into account in the study of the mechanisms of membrane fusion.
References Katz B (1969) The Release of Neurotransmitter Substances: The Sherrington Lectures X. Charles C Thomas, Springfield, IL Dodge FA, Rachamimoff R (1967) Co-operative action of calcium ions in transmitter release sites at the neuromuscular junction. J Physiol (Lond) 193, 419-432 LliMs R (1977) Calcium and transmitter release in squid synapse. In: Approaches to the Cell Biology of Neurons. (Cowan WM, Ferendelli JA eds) Society for Neurosciences, Bethesda, MD, 139-160 LliMs R, Steinberg I, Walton K (1981) Presynaptic calcium currents in squid giant synapse. Biophys J 33, 28%322
Robitaille R, Adler EM, Charlton M (1990) Strategic location of calcium channels at transmitter release sites of frog neuromuscular synapses. Neuron 5, 773-779 Chad JE, Eckert R (1984) Calcium domains associated with individual channels may account for anomalous voltage relations of Ca-dependent responses. Biophys J 45, 993-999 Simon SM, Sugimori M, Llin~s R (1984) Modelling of submembranous calcium-concentration changes and their relation to rate of presynaptic transmitter release in the squid giant synapse. Biophys J 45, 264a Simon SM, Llinfis R (1985) Compartmentalization of the submembrane calcium activity during calcium influx and its significance in transmitter release. Biophys J 48, 485-498 Fogelson AL, Zucker RS (1985) Presynaptic calcium diffusion from various arrays of single channels. Implication for transmitter release and synaptic facilitation. Biophys J 48, 1003-1007 Shimomura O, Inouye S, Musicki B, Kishi Y (1990) Recombinant aequorin and recombinant semi-synthetic aequorins: cellular Ca2+ ion indicators. Biochem J 270, 309-312 Llintis R, Sugimori M, Silver B (1992) Microdomains of high calcium concentration in a presynaptic terminal. Science 256, 677-679 LliMs R, McGuinness TL, Leonard CS, Sugimori M, Greengard P (1985) Intraterminal injection of synapsin I or calcium/calmodulin-dependent protein kinase II alters neurotransmitter release at the squid giant synapse. Proc Natl Acad Sci USA 82, 30353039 Pumplin DD, Reese TW, Llin~s R (1981) Are the presynaptic membrane particles the calcium channels? Proc Natl Acad Sci USA 78, 7210-7213 Korn J, Mallet A, Triller A, Faber DS (1982) Transmission at a central synapse. II. Quantal description of release with a physical correlation for binomial n. J Neurophysiol 48, 679 Triller A, Korn H (1982) Transmission at a central inhibitory synapse. III. Ultrastructure of physiologically identified terminals. J Neurophysiot 48 708-736
