1: %\documentclass[twocolumn,11pt]{article}
2: \documentclass[11pt]{article}
3: \usepackage{graphicx,epsfig}
4: \usepackage{amsmath,chemarrow}
5:
6: \setlength{\textwidth}{160mm} \setlength{\textheight}{200mm}
7: \setlength{\topmargin}{0mm} \setlength{\oddsidemargin}{0mm}
8:
9: \begin{document}
10:
11:
12: \begin{center}
13: {\LARGE Modeling stochastic Ca$^{2+}$ release from a \\
14: cluster of IP$_3$-sensitive receptors} \vspace{1cm}
15:
16: {\large L. Diambra \\ Instituto de F\'{\i}sica de S\~ao Carlos,
17: Universidade de S\~ao Paulo, \\
18: Caixa Postal: 369, cep: 13560-970, S\~ao Carlos SP, Brazil. \\
19: diambra@if.sc.usp.br}
20: \end{center}
21:
22: \begin{abstract}
23: We focused our attention on Ca$^{2+}$ release from the endoplasmic
24: reticulum through a cluster of inositol 1,4,5-trisphosphate (IP$_3$)
25: receptor channels. The random opening and closing of
26: these receptors introduce stochastic effects that have been observed
27: experimentally. Here, we present a stochastic version of Othmer-Tang
28: model for IP$_3$ receptor clusters. We address the average behavior
29: of the channels in response to IP$_3$ stimuli. We found, by
30: stochastic simulation, that the shape of the receptor response to
31: IP$_3$ (fraction of open channels versus [IP$_3$]), is affected by
32: the cytosolic Ca$^{2+}$ level. We also study several aspects of the
33: stochastic properties of Ca${2+}$ release and we compare with
34: experimental observations.
35: \end{abstract}
36: \parskip=0.5\baselineskip
37: \newpage
38:
39: \section{Introduction}
40:
41: Ionized calcium (Ca$^{2+}$) not only represents the most common
42: signal transduction element relaying information within cells to
43: control a wide array of activities including secretion,
44: contraction and cell proliferation, but also is invariably
45: involved in cell death (Berridge, 1997). To coordinate all of
46: these functions, cytosolic Ca$^{2+}$ needs to be precisely
47: regulated in space, time and amplitude. Normal concentrations of
48: cytosolic Ca$^{2+}$ ([Ca$^{2+}$] $\sim$ 100 nM) is 20000 fold
49: lower than the 2 mM concentration found extracellularly. Under
50: resting conditions this gradient is maintained by active extrusion
51: of cytosolic Ca$^{2+}$ by Ca$^{2+}$ATPases present in the plasma
52: membrane and in the endoplasmic reticulum (ER), or in the
53: sarcoplasmic reticulum in electrically excitable cells. These
54: pumps counterbalance the leak of Ca$^{2+}$ into the cytosol from
55: both the extracellular space and the ER Ca$^{2+}$ store.
56:
57: A wide variety of extracellular stimuli cause the increase of
58: [Ca$^{2+}$] to exert their effect. In non-excitable cells, this
59: increase is triggered by inositol(1,4,5)-trisphosphate (IP$_3$),
60: produced upon activation of phospholipase C (Berridge, 1997).
61: IP$_3$ rapidly diffuses into the cytosol, where it interacts with
62: the inositol(1,4,5)-triphosphate receptors (IP$_3$R), to promote
63: the release of Ca$^{2+}$ into the cytosol. Depending on the cell
64: type, the resulting cytosolic Ca$^{2+}$ signal can have a complex
65: spatio-temporal composition (Berridge, 1997). It is generally
66: recognized that, in the presence of a constant external stimulus,
67: the Ca$^{2+}$ displays spiking behavior. These Ca$^{2+}$
68: oscillations can be spatially localized or extended, spreading as
69: a wave throughout the entire cell.
70:
71: The process of Ca$^{2+}$ released from the ER through channels is
72: nonlinear since, increased Ca$^{2+}$ concentration in the cytosol
73: favors channel opening. This autocatalytic amplification is called
74: calcium-induced calcium release (CICR). There are a variety of
75: channels showing CICR. Ca$^{2+}$ release is terminated by closure
76: of the channels at high Ca$^{2+}$ levels, after which Ca$^{2+}$ is
77: removed from the cytosol by the action of the Ca$^{2+}$ ATPases.
78:
79: It has been observed that Ca$^{2+}$ release channels are spatially
80: organized in clusters. Single release channel, named Ca$^{2+}$
81: blips, has been observed in experiments (Bootman et al., 1997).
82: Collective opening and closing of several Ca$^{2+}$ channels in a
83: cluster, named puffs, have also been observed in experiments
84: (Callamaras et al., 1998). That suggests a hierarchy of calcium
85: signaling events from small blips to large puffs (Lipp and Niggli,
86: 1998; Bootman et al., 1997). Improved spatial and temporal
87: resolution in recording reveal that there is not a clear
88: distinction between fundamental blips and elementary puffs.
89: Channels open and close in a stochastic way. Ca$^{2+}$ release by
90: one channel increase the open probability for the neighboring
91: channels. Thus at high levels of IP$_3$, neighboring clusters
92: become functionally coupled by Ca$^{2+}$ diffusion and CICR
93: support the formation of spatiotemporal patterns of intracellular
94: Ca$^{2+}$ release.
95:
96: Watras et al. (1991) found that Ca$^{2+}$ channels in the
97: cerebellum exhibits four conductance levels that are multiples of
98: a unit conductance step. This observation suggests that the number
99: of interacting receptors in one complex can vary from one to four
100: and supports the hypothesis that the channel is a tetramer.
101: Examination of the IP$_3$ dependence of channel, yielded Hill
102: coefficients of 1-1.3 (Watras et al., 1991). However, in
103: hepatocytes the response to IP$_3$ are positively cooperative with
104: a Hill coefficient of 3.0-3.4 (Marchant and Taylor, 1997; Dufour
105: et al., 1997) or. In our numerical simulation we found that the Hill
106: coefficient increase from 1.15 to 2.2 when total Ca$^{2+}$
107: (cytosolic + stored) level increase. However, the model
108: is not able to explain higher values of $n_h$ as that reported by
109: Marchant and Taylor (1997). This finding suggest that feedback
110: from cytosolic Ca$^{2+}$ plays a key role in the channel response
111: to IP$_3$, but is not able to explain high Hill coefficients found in
112: some tissues.
113:
114: Experimental finding of Bezprozvanny et al. (1991) clear up
115: several aspects of the kinetic of the receptors. Thus several
116: models for Ca$^{2+}$ dynamics in systems involving IP$_3$R have
117: been proposed (De Young and Keizer, 1992; Othmer and Tang, 1993;
118: Bezprozvanny and Ehrlich, 1994). These models of IP$_3$R assume a
119: regulatory site for IP$_3$ on the channel complex, one activating
120: site and one inhibiting site for Ca$^{2+}$. Experimental findings
121: suggest that the channel is open when IP$_3$ ion is bound to its
122: corresponding domain and Ca$^{2+}$ is bound to its activating
123: domain and not bound to its inhibiting site. In the aforementioned
124: models, Ca$^{2+}$ binding to the activating regulatory site is a
125: fast process compared with that of inhibiting binding and all
126: these models reproduce a bell-shaped curve when one plots the
127: fraction of open channels as a function of $\log $[Ca$^{2+}$].
128:
129: De Young and Keizer model (DKM) for IP$_3$R assumes that the
130: ligands can bind to any unoccupied site on the IP$_3$R
131: independently of the binding status of the other sites, which
132: leads to eight possible states of the receptor (De Young and
133: Keizer, 1992). A more general scheme is to assume state-dependent
134: binding to each site. In this framework, the simplest model was
135: developed by Othmer and Tang (1993) which assumes that the binding
136: order of ions and molecules to the receptor is not free. In the
137: Othmer-Tang model (OTM) it is assumed that the binding process is
138: sequential (Othmer and Tang, 1993). Ca$^{2+}$ binds to the
139: activating site on the receptor only after IP$_3$ has bound, and
140: that the binding of calcium to inhibitory site occurs only after
141: calcium is bound to the activating site. This sequential binding
142: leads to four possible states for the receptor. Bezprozvanny and
143: Ehrlich (1994) proposed a variation to the OTM which incorporates
144: an additional closed state corresponding to activating site
145: occupied which decay rapidly to the open state.
146:
147: The bindings of IP$_3$ and Ca$^{2+}$ to the regulatory sites are
148: stochastic events rendering the opening and closing of the channel
149: a stochastic process. The small number of calcium channels
150: in a cluster indicates that deterministic models might be insufficient.
151: In fact, abortive waves cannot be understood in terms of deterministic
152: models, since in these models an excitation travels steadily if it
153: travels at all. The stochastic effects are relevant for modeling Ca$^{2+}$
154: wave propagation. Furthermore, the observation of localized
155: stochastic Ca$^{2+}$ puffs and the rather small number of channels
156: creating the localized event, suggest that it is mandatory take into
157: account the binding processes as stochastic events. The stochastic
158: dynamics of clustered IP$_3$R has been studied by Swillens et al.
159: (1999) which model has 14 states. There are also several studies
160: considering stochastic dynamics focusing on the onset of the
161: saltatory propagation of calcium waves due to intercluster
162: diffusion of Ca$^{2+}$ (Keizer et al., 1998; Falcke et al., 2000).
163: Shuai and Jung (2002b) have been study the statistical properties
164: of Ca$^{2+}$ release in a stochastic and simplified version of the DKM.
165:
166: In this paper we regard a Markov-stochastic version of the
167: OTM to study the stochastic properties of the calcium release of
168: IP$_3$R clusters. We decided to stick to the Othmer and Tang
169: approach because the OTM has the least number of channel states
170: and kinetic parameters, yet it adequately explains the
171: experimentally observed channel kinetics. After a brief
172: description of the model in Section 2, we discuss the role of
173: cytosolic calcium and IP$_3$ in regulating the fraction of open
174: channels. We show that the fraction of open channels versus IP$_3$
175: concentration fit a Hill curve which coefficient increase with the
176: average intracellular calcium concentration.
177: The amplitude, inter-puff interval (IPI) and size distribution of
178: calcium puffs are also discussed. Finally, we show that there are
179: no correlation between two consecutive IPIs, but there are some
180: correlation between the puff duration and the puff amplitude.
181:
182: \section{Materials and Methods}
183: \subsection{Stochastic version of OTM}
184:
185: In the Othmer and Tang model there is only one subunit/receptor in
186: a each channel. The three binding sites on the receptor and the
187: sequential binding scheme allows four different states for the
188: receptor X$_{i,j,k}$. The index $i$ stands for the IP$_3$ site,
189: $j$ for the activating Ca$^{2+}$ site, and $k$ for the inhibiting
190: Ca$^{2+}$. An index is 1 if an ion is bound and 0 if not. The
191: state transition scheme is given by the pathway
192: \begin{displaymath}
193: {\ensuremath{\text{X}_{\text{0,0,0}}}} \ \
194: \autorightleftharpoons{$k_1$[IP$_3$]}{$k_{-1}$} \ \
195: {\ensuremath{\text{X}_{\text{1,0,0}}}} \ \
196: \autorightleftharpoons{$k_2$[Ca$^{2+}$]}{$k_{-2}$} \ \
197: {\ensuremath{\text{X}_{\text{1,1,0}}}} \ \
198: \autorightleftharpoons{$k_3$[Ca$^{2+}$]}{$k_{-3}$} \ \
199: {\ensuremath{\text{X}_{\text{1,1,1}}}},
200: \end{displaymath}
201: in which the $k_i$, $i=\pm 1,\pm 2,\pm 3$, are the rate constant
202: of the state transitions. The channel is only conductive when the
203: receptor is in the state {\ensuremath{\text{X}_{\text{1,1,0}}}}.
204: The channels are assumed to be close enough so that Ca$^{2+}$
205: concentrations can be considered homogeneous throughout the
206: cluster. We neglect Ca$^{2+}$ diffusion between cluster and
207: environment without accounting for spatial aspects of formation
208: and collapse of Ca$^{2+}$ waves. We will not consider calcium
209: transport across the plasma membrane. Thus the Ca$^{2+}$ simply
210: flow or is pumped back and forth between the ER and the cytoplasm,
211: and the total intracellular calcium concentration is constant.
212: Thus, the dynamic of Ca$^{2+}$ in the cytoplasm is governed by the
213: equation
214: \begin{equation}
215: \frac{d\left[{\rm Ca}^{2+}\right]}{dt}=J_{channel}+J_{leak}-J_{pump},
216: \end{equation}
217: where $J_{channel}$ is the calcium flux from the ER to the cytoplasm
218: through the IP$_3$R channel, $J_{pump}$ is the calcium flux pumped
219: from the cytoplasm into ER, and $J_{leak}$ is the calcium flux being
220: leakage from the ER to the cytoplasm. The expression for the
221: fluxes are given by
222: \begin{eqnarray}
223: J_{channel}&=& v_r \gamma _1 \frac{N_{open}}{N} \left( \left[{\rm Ca}^{2+}\right]_{\rm ER} -\left[{\rm Ca}^{2+}\right]\right) \nonumber \\
224: J_{leak}&=&v_r \gamma _0 \left( \left[{\rm Ca}^{2+}\right]_{\rm ER} -\left[{\rm Ca}^{2+}\right]\right) \nonumber \\
225: J_{pump}&=&\hat{p}_1 \frac{\left[{\rm
226: Ca}^{2+}\right]^4}{\left[{\rm Ca}^{2+}\right]^4+\hat{p}_2^4},
227: \end{eqnarray}
228: where $v_r$ is the ratio of the ER volume to the cytoplasm volume,
229: $\gamma _1 $ is the maximal Ca$^{2+}$ fluxes. $N_{open}$ is a
230: random variable that represent the number of channels open
231: (receptors in the state {\ensuremath{\text{X}_{\text{1,1,0}}}})
232: and $N$ is the total number of channels in the receptor. $\gamma
233: _0$ is the basal permeability of the ER membrane in the absence of
234: IP$_3$. For the study of Ca$^{2+}$ dynamics in many cell types,
235: Ca$^{2+}$ exchange with the extracellular medium is much smaller
236: than the Ca$^{2+}$ flux across the ER membrane (Koster et al.,
237: 1993; Thastrup, 1990). For this reason, Eqs. (1-2) can be
238: simplified by using the volume average intracellular calcium
239: concentration [Ca$^{2+}$]$_{ave}=$([Ca$^{2+}$]$+ v_r$
240: [Ca$^{2+}$]$_{\rm ER}$)$/(1+v_r)$. The [Ca$^{2+}$]$_{ave}$ is a
241: control parameter whose value we can choose, but that it is not a
242: dynamical variable. We then define
243: $C=$[Ca$^{2+}$]/[Ca$^{2+}$]$_{ave}$ and rewrite the Eqs. (1-2) in
244: the form
245: \begin{equation}
246: \dot{C}=\left(1+v_r \right) \left(\gamma_0 +\gamma_1 N
247: _{open}\right) \left(1-C\right)-p_1 \frac{C^4}{C^4+p_2^4},
248: \end{equation}
249: where $p_1 = \hat{p}_1/$[Ca$^{2+}$]$_{ave}$ and
250: $p_2=\hat{p}_2/$[Ca$^{2+}$]$_{ave}$.
251:
252: This equation was numerically integrated using fourth order
253: Runge-Kutta algorithm, with the actual configuration of open
254: channels $N_{open}$, which is obtained from stochastic simulations
255: of the kinetics of the receptors. The number of open channels
256: changes due to six stochastic processes: the IP$_3$ binding
257: (unbinding) to (from) receptors; the activation (deactivation) of
258: receptors by binding (unbinding) Ca$^{2+}$ to (from) their
259: activation domains; the inhibition (activation) of receptors by
260: binding (unbinding) Ca$^{2+}$ to (from) their inhibition domains.
261: Transitions involving binding (but not those involving unbinding)
262: depend on the concentration of IP$_3$ or Ca$^{2+}$. Let us define
263: the probability distribution vector $\mathcal{P}\left(t
264: \right)=\left(P_{0,0,0} \left(t \right),
265: P_{1,0,0} \left(t \right), P_{1,1,0} \left(t \right),
266: P_{1,1,1} \left(t \right)\right)^T$,
267: %\begin{equation}
268: %\mathcal{P}\left(t \right)=\left(%
269: %\begin{array}{c}
270: % P_{0,0,0} \left(t \right)\\
271: % P_{1,0,0} \left(t \right)\\
272: % P_{1,1,0} \left(t \right)\\
273: % P_{1,1,1} \left(t \right)\\
274: %\end{array}%
275: %\right),
276: %\end{equation}
277: where $P_{i,j,k}$ denotes the probability that the receptor is in
278: the state X$_{i,j,k}$. The dynamic of the vector $
279: \mathcal{P}\left(t \right)$ is described by the master equation,
280: \begin{equation}
281: \dot{\mathcal{P}}= Q\mathcal{P}, \label{me}
282: \end{equation}
283: where $Q$ represents a ($4 \times 4$) matrix of transition
284: probabilities,
285:
286: \begin{equation}
287: Q=\left(
288: \begin{array}{cccc}
289: 1-k_1\left[{\rm IP}_3 \right] & k_{-1} & 0 & 0 \\
290: k_1\left[{\rm IP}_3 \right] & 1- k_{-1} - k_{2} [ {\rm Ca}^{2+}] & k_{-2} & 0 \\
291: 0 & k_{2} [ {\rm Ca}^{2+}]& 1- k_{-2} - k_{3} [ {\rm Ca}^{2+}] & k_{-3} \\
292: 0 & 0 & k_{3} [ {\rm Ca}^{2+}] & 1-k_{-3} \\
293: \end{array}
294: \right),
295: \end{equation}
296: For the stochastic part of the simulation, the state of each receptors
297: at time $t+\Delta t$ were independently updated from the previous state
298: and $Q$. If the time step $\Delta t$ is sufficiently small to allow at
299: most one transition, the
300: probabilities for the receptor to transform into the various
301: possible states within this short time interval are given by
302: \begin{eqnarray}
303: P_{{0,0,0} \to {1,0,0}}=k_1\left[{\rm IP}_3 \right]\Delta t \ \ \quad &
304: P_{{1,0,0} \to {0,0,0}}=k_{-1}\Delta t \nonumber \\
305: P_{{1,0,0} \to {1,1,0}}=k_2 [ {\rm Ca}^{2+}]\Delta t \quad &
306: P_{{1,1,0} \to {1,0,0}}=k_{-2}\Delta t \nonumber\\
307: P_{{1,1,0} \to {1,1,1}}=k_3 [ {\rm Ca}^{2+}]\Delta t \quad &
308: P_{{1,1,1} \to {1,1,0}}=k_{-3}\Delta t.
309: \end{eqnarray}
310: For this model, there is only one possible transition out of the
311: states X$_{0,0,0}$ and X$_{1,1,1}$, while there are two possible
312: transitions out of the states X$_{1,0,0}$ and X$_{1,1,0}$. In the
313: last case a random number $\rho$ between 0 and 1 was drawn from a
314: uniform distribution for each update step. Then, a channel
315: initially in state, for example, X$_{1,0,0}$, will be updated to
316: state X$_{1,1,0}$, if $\rho<P_{{1,0,0} \to {1,1,0}}$, and to state
317: X$_{0,0,0}$ if $P_{{1,0,0} \to {1,1,0}}< \rho \leq P_{{1,0,0} \to
318: {0,0,0}}+ P_{{1,0,0} \to {1,1,0}}$, and
319: remained in its current state otherwise.
320:
321: All over the paper we have used $\Delta=0.001$ s, and the number
322: of receptors in a cluster, $N$, was setted to 20. This number is
323: based on the theoretical predictions of the cluster size by
324: Swillens et al. (1999), from numerical studies regarding the
325: requirements of interchannel communication. Furthermore, found
326: that the optimal cluster size for coherence resonance is around
327: $N=20$ (Shuai and Jung, 2002a). The values of the parameters for
328: the kinetic model used in this paper are given in Table 1.
329:
330: To observe puffs in experiments, calcium diffusion is
331: suppressed by intracellular loading with the Ca$^{2+}$ buffer EGTA
332: (Thomas et al., 1998; Marchant et al., 1999; Cheng et al., 1999;
333: Callamaras and Parker, 2000). With a large loading of EGTA, the
334: clusters become functionally isolated (Callamaras and Parker,
335: 2000). Under these condition, our model is valid.
336:
337: \section{Results}
338: \subsection{The channel response to IP$_3$ and Ca$^{2+}$}
339:
340: Before to study the statistical properties of the intracellular
341: Ca$^{2+}$ release, we firstly address the average behavior of
342: channels in response to IP$_3$ stimuli. In this sense, we simulate
343: the stochastic dynamics (Eq. 4-5) with the parameters in Table 1
344: and using [Ca$^{2+}$]$_{ave}$ as control parameter. We compute the
345: mean value of cytosolic Ca$^{2+}$ concentration in a long time
346: period (10,000 seconds) as a function of [IP$_3$] for several
347: values of [Ca$^{2+}$]$_{ave}$ (0.4, 0.8, 1.2, 1.6, 1.9 and 2.4
348: $\mu$M). The results are sigmoidal curves as is shown in Fig. 1.
349: It should be noted that the cytosolic mean Ca$^{2+}$ invariably
350: increases with the level of [Ca$^{2+}$]$_{ave}$. The vertical grey
351: lines both in Fig. 1 and Fig. 2 are indicating the IP$_3$ levels
352: for which we have also studied several statistical properties of
353: the puffs. Further insight can be obtained computing the effect of
354: the IP$_3$ over the mean fraction of open channels. The mean
355: fraction of channel that is activated increase monotonically as
356: the IP$_3$ level increases following Hill curves ($r^2\geq 0.995$
357: in all cases) as shown in the Fig. 2. The Hill curve is defined by
358: \begin{equation}
359: \frac{N_{open}}{N}=V_{max}\frac{[{\rm IP}_3]^{n_h}} {k^{n_h}+[{\rm
360: IP}_3]^{n_h}}, \label{hill}
361: \end{equation}
362: where $k$ is the half maximal value, $V_{max}$ is the maximum
363: fraction of channels open, and $n_h$ is Hill coefficient. In the
364: top panel of Fig. 3 we depict the Hill coefficient $n_h$ and in
365: the bottom panel we depict $k$ and $V_{max}$ for different values
366: of [Ca$^{2+}$]$_{ave}$. The Hill coefficient increases as
367: [Ca$^{2+}$]$_{ave}$ increases. At physiological values of
368: [Ca$^{2+}$]$_{ave}$, the $n_h$ found in our simulation range from
369: 1.15 to 2.2 in very well agreement with the experimental values
370: which range from 1.0 (Volpe et al., 1990) to greater than 3.7
371: (Delisle, 1991). In the bottom panel of Fig. 3 we depict the
372: others two fitted parameters $V_{max}$ and $k$. The maximum
373: fraction of open channels $V_{max}$ are between [0.15,0.30] with a
374: maximum at 0.80 of [Ca$^{2+}$]$_{ave}$, this range is slightly
375: higher than the values found by Watras et al. (1991), around 0.14
376: in cerebellar cells (with $n_h=1.3$). It should be noted that the
377: half-maximal $k$ decrease with the level of [Ca$^{2+}$]$_{ave}$
378: and range from 0.10 to 0.7 $\mu$M IP$_3$. The half-maximal found
379: in the Watras experiment was 0.15 $\mu$M IP$_3$.
380:
381: Bezprozvanny et al. (1991) shown, in experiment where Ca$^{2+}$
382: channels reconstituted in lipid bilayers, that the fraction of
383: open cannel as a function of $\log$[Ca$^{2+}$] has an asymmetry
384: bell-shaped curve, in experiment where calcium was clamped. The
385: mathematical analysis of the master equation (\ref{me}), allows to
386: determine the open channel probability (to find the receptor in
387: the state X$_{1,1,0}$) for a given [Ca$^{2+}$] and [IP$_3$]. The
388: steady-state distribution vector $\mathcal{P}^{st}$ correspond to
389: the normalized eigenvector associated to the eigenvalue zero of
390: the equation of (\ref{me}). Some algebraic manipulations lead to
391: \begin{equation}
392: P_{1,1,0}^{st} \left( \left[{\rm IP}_3 \right], [ {\rm Ca}^{2+}]\right)=
393: \frac{K_1 K_2 \left[{\rm IP}_3 \right] [ {\rm Ca}^{2+}]}
394: {1+K_1 \left[{\rm IP}_3 \right]+K_1 K_2 \left[{\rm IP}_3 \right]
395: [ {\rm Ca}^{2+}] +K_1 K_2 K_3\left[{\rm IP}_3 \right] [ {\rm Ca}^{2+}]^2 },
396: \label{result}
397: \end{equation}
398: where $K_1=k_1/k_{-1}$, $K_2=k_2/k_{-2}$ and $K_3=k_3/k_{-3}$.
399: Fig. 4 depicts the open channel probability for three different
400: values of IP$_3$ concentration as a function of [Ca$^{2+}$]. The
401: result (\ref{result}) predicts the bell-shaped curve obtained
402: experimentally by Bezprozvanny et al. (1991), with the symmetry
403: axe correctly shifted to left. It should be noted the difference
404: between Eq. (\ref{hill}) and (\ref{result}). The first one can be
405: understood as a open channel probability in physiological
406: conditions described by Eq. (1-2). Thus the fraction of open
407: channels is computed as the average of fraction of open channels
408: over a period of time (10,000 seconds in this paper), where
409: calcium is released (removed) to (from) cytosol many times. On the
410: other hand, Eq. (\ref{result}) is a steady state probability for a
411: fix level of calcium and IP$_3$. This situation emulate the
412: Bezprozvanny et al. experimental conditions, where the calcium was
413: clamped.
414:
415: \subsection{Statistical properties of puffs}
416:
417: We now study the statistical properties of puffs obtained with the
418: stochastic version of the OTM, the deterministic version of OTM is
419: not able to address these properties. It is known that IP$_3$
420: stimuli evokes repetitive spikes in the intracellular [Ca$^{2+}$]
421: rather than simply increase [Ca$^{2+}$]. In the Fig. 5, we can see
422: the typical spiking behavior of calcium release by a cluster with
423: 20 channels ($N=20$) and some statistical properties of the
424: observed puffs 10,000 seconds of simulations. In this case, we
425: performed the simulation setting [Ca$^{2+}$]$_{ave}$=1.6 $\mu$M
426: and IP$_3$=0.35 $\mu$M. At the top panel of Fig. 5 we depict the
427: first 100 seconds of the temporal course of the intracellular
428: calcium concentration. At the bottom-left panel, the scatter plot
429: of IPI$_i$ versus IPI$_{i+1}$ shown that there is not temporal
430: correlation between two consecutive IPIs. At the bottom-center
431: panel: scatter plot of puff amplitude versus puff duration shown
432: that there is correlation between puff amplitudes and durations
433: ($r^2$=0.84). Small puff amplitudes are correlated with brief
434: puffs, the correlation decreases as amplitude puff increases.
435: Experimental observations of Thomas et al. (1998) indicate similar
436: correlations, but with a smaller linear regression coefficient
437: ($r^2$=0.69). At the bottom-right panel we shown the puff size
438: distribution, it should be noted that there are Ca$^{2+}$ releases
439: of all sizes. All plots in the bottom panel of Fig. 5, and
440: hereafter, were computed with puffs which amplitudes exceed the
441: threshold 0.20 $\mu$M, in order to avoid the background
442: fluctuation. In this paper, the puff size is defined as the area
443: between the [Ca$^{2+}$] time course and the threshold.
444:
445: It was experimentally established that when the IP$_3$
446: concentration increases, the mean frequency of spiking raises, but
447: the amplitude remain essentially constant. Fig. 6 displays the
448: mean frequency of puffs as a function of the IP$_3$ concentration
449: for the some values of [Ca$^{2+}$]$_{ave}$ showed in Figs. 1-2
450: (0.8, 1.2, 1.6, 1.9 and 2.4 $\mu$M) obtained by stochastic
451: simulations. We can observe that the onset of spiking behavior in
452: response to the IP$_3$ stimuli depends on the [Ca$^{2+}$]$_{ave}$.
453: For high levels of calcium, even low levels of IP$_3$ are able to
454: produce repetitive puffs. These responses are well characterized
455: by Hill curves. The Hill coefficients associated to these curves
456: were essentially independent of [Ca$^{2+}$]$_{ave}$ level. For
457: [Ca$^{2+}$]$_{ave}$=2.4 $\mu$M the Hill coefficient was 2.1 and
458: decrease to 1.8 for [Ca$^{2+}$]$_{ave}$=0.8 $\mu$M.
459:
460: Furthermore, the number of channel recruited in the puffs varies
461: stochastically from puff to puff. Important, and experimentally
462: recorded, characteristics of these variabilities are amplitude,
463: inter-puff interval, duration and size distributions (Bootman et
464: al., 1997; Sun et al., 1998; Thomas et al., 1998; Callamaras and
465: Parker 2000; Marchant and Parker 2001). The shape of the puff
466: amplitude distribution depends on the concentration of IP$_3$ and
467: the [Ca$^{2+}$]$_{ave}$. In Fig. 7 we shown a diagram of the puff
468: amplitude distributions in the [Ca$^{2+}$]$_{ave}$-IP$_3$ plane.
469: In order to comparison the scale of the axis are the same for all
470: plots in the diagram, between [0.20,1.80] $\mu$M for the
471: amplitudes of puffs (horizontal axe), and between [0,6000] for the
472: number of events (vertical axe). No events (NE) were recorded for
473: [IP$_3$]= 0.07 $\mu$M and low level of [Ca$^{2+}$]$_{ave}$=0.8
474: $\mu$M). For lower [Ca$^{2+}$]$_{ave}$, the amplitudes of the
475: spontaneous puffs are typically smaller than 0.20 $\mu$M and are
476: regarded as fluctuations. For [Ca$^{2+}$]$_{ave}$=0.8 $\mu$M, or
477: [IP$_3$]= 0.07 $\mu$M, monotonically decreasing amplitude
478: distribution are found. In the other cases a two-peak amplitude
479: distributions are mainly found. The characteristic amplitudes
480: depend on [Ca$^{2+}$]$_{ave}$, but not on [IP$_3$]. For
481: [Ca$^{2+}$]$_{ave}$=1.6 $\mu$M, the typical amplitude was around
482: 0.80 $\mu$M. The shape of the puff amplitude distributions shown
483: in the Fig. 7 differ from those observed in experiments, which
484: have one-peak shape (Sun et al., 1998; Thomas et al., 1998;
485: Marchant and Parker,2001).
486:
487: Another important characteristic of the calcium dynamics is the
488: distribution of time interval between two consecutive puffs.
489: Recent experimental investigation with {\it Xenopus} oocytes has
490: revealed that IPI distribution exhibits a single peak. In Fig. 8
491: we display another diagram in the [Ca$^{2+}$]$_{ave}$-IP$_3$
492: plane, but now with IPI distributions. In this diagram the range
493: of the vertical axis are the same for all distributions, between
494: [0,6000] events. While the ranges of IPI (horizontal axis) are
495: between [0,80] seconds for the first column ([Ca$^{2+}$]$_{ave}$
496: =0.8 $\mu$M), and between [0,20] seconds for all other cases. For
497: [Ca$^{2+}$]$_{ave}$ =0.8 $\mu$M, the IPI distribution seems to
498: decay exponentially. However, when the level of
499: [Ca$^{2+}$]$_{ave}$ is increased, skewed single-peak distribution
500: characterize the IPI distributions. The position of the single
501: peak range from 1 to 2 seconds increasing with the
502: [Ca$^{2+}$]$_{ave}$. The shape of the distributions shown in the
503: Fig. 8 are consistent with observed by Marchant et al. (1999), but
504: with a typical IPI smaller than that observed in the experiment
505: (around 2.5 seconds). It is apparent that for high values of
506: [IP$_3$] and [Ca$^{2+}$]$_{ave}$, the IPI are quite regular.
507:
508: We also want to illustrate the typical temporal behavior and some
509: statistical properties, as we shown in the Fig. 5, for two
510: different situations: i) high [IP$_3$], low [Ca$^{2+}$]$_{ave}$
511: (Fig. 9); and ii) low [IP$_3$], high [Ca$^{2+}$]$_{ave}$ (Fig.
512: 10). It should be noted that different scale axis were used. In
513: Fig. 9, where [IP$_3$]=21 $\mu$M, and [Ca$^{2+}$]$_{ave}$=0.8
514: $\mu$M, we found the distribution of puffs size decay
515: exponentially, and there are very few great calcium release events
516: (bottom-right panel). More broad distributions of IPI are yielded
517: for this case than those shown in bottom-left panel of Figs. 5 and
518: 10. There are also less IPI around 1-2 seconds (bottom-left panel)
519: than those shown in Fig. 5 or in Fig. 10. The correlation between
520: puff amplitude and duration seems to be similar than that shown in
521: bottom-right panel of Fig. 5. When increase the
522: [Ca$^{2+}$]$_{ave}$ until 2.4 $\mu$M and decrease the [IP$_3$] to
523: 0.07 $\mu$M (Fig. 10), we found a slower exponential decay in the
524: distribution of puffs size than in Fig. 9 (bottom-right panel). In
525: this case the distributions of IPI (bottom-center panel of Fig.
526: 10) looks like those shown in bottom-left panel of Figs. 5. Puff
527: with small amplitude seems to be correlated with short duration.
528:
529: \section{Discussion and conclusion}
530:
531: We have developed a stochastic version of the Othmer-Tang model to
532: study dynamical and statistical properties of Ca$^{2+}$ release of
533: clusters of IP$_3$-sensitive receptors. In comparison to others
534: stochastic models (Falcke et al., 2000; B\"ar et al., 2000;
535: Swillens et al., 1999), our receptor model is simpler and is
536: represented by only four states. In this stochastic clustered
537: IP$_3$R model, the Ca$^{2+}$ diffusion between the cluster and the
538: environment is ignored so that an isolated cluster can be
539: discussed. However, the channels are assumed to be close enough
540: and the instantaneous Ca$^{2+}$ diffusion within a cluster is so
541: fast that the calcium concentration within a cluster is always
542: homogeneous.
543:
544: We found that the model reproduce experimental result by Watras et
545: al. (1991) where it is shown that the channel response to the
546: IP$_3$ follow a Hill curve. In intact cells, the long latency and
547: subsequent sigmoidal increase in cytosolic [Ca$^{2+}$] after
548: addition of agonists (Berridge, 1997), are consistent with either
549: positively cooperative activation of IP$_3$Rs by IP$_3$ alone or
550: by feedback from cytosolic [Ca$^{2+}$]. Distinguishing between
551: these mechanisms, however, is difficult in intact cells, in which
552: IP$_3$ receptor activation inevitably increases cytosolic
553: [Ca$^{2+}$]. In broken cells, the use of fluorescent indicators to
554: measure rapid rates of Ca$^{2+}$ efflux is prone to similar
555: difficulties. Results from a variety of tissues using superfusion
556: (Finch et al., 1991), bilayer recording (Watras et al., 1991;
557: Besprozvanny et al., 1991), fluorescent indicators (Champeil et
558: al., 1989) or flash photolysis of caged IP$_3$ in intact cells
559: (Parker et al., 1996) suggest that responses to IP$_3$ are either
560: not positively cooperative (Finch et al., 1991; Parker et al.,
561: 1996; Watras et al., 1991) or only slightly so (Besprozvanny et
562: al., 1991; Champeil et al., 1989). Each approach has its
563: limitations and together the results provide no clear indication
564: of whether responses to IP$_3$ are positively cooperative in the
565: absence of positive feedback from cytosolic Ca$^{2+}$. In our
566: numerical simulation we found that, even in absence of
567: cooperativity in the model, the Hill coefficient increase from
568: 1.15 to 2.2 when [Ca$^{2+}$]$_{ave}$ increase. However, the model
569: is not able to explain higher values of $n_h$ as that reported by
570: Marchant and Taylor (1997). This results suggest that feedback
571: from cytosolic Ca$^{2+}$ plays a key role in the channel response
572: to IP$_3$, but to explain high values of $n_h$ found in some
573: tissues, a cooperative mechanism could be mandatory.
574:
575: Other important aspect is the frequency encoding. It is well known
576: that IP$_3$ stimuli evoke repetitive spikes in the intracellular
577: Ca$^{2+}$ concentration rather than simply elevating the level of
578: Ca$^{2+}$. Moreover, when the [IP$_3$] is increased, the mean
579: frequency of spiking raises, but the amplitude remains essentially
580: constant. Thus a continuous level of stimuli signal is converted
581: into a frequency-encoded signal: the number of spikes per time
582: unit (Tang and Othmer, 1995). A common problem found in this kind
583: of converter devices is the competition between two desirable
584: goals: i) high sensitivity, the system ideally should be able to
585: detect even weak signals; and ii) large dynamic range, the system
586: should not saturate over various orders of magnitude of input
587: intensity. For this reason the Hill coefficient (when the response
588: is well-fitted by a Hill curve) its desirable to be less than 1.
589: In the results derived from our numerical simulation, we can
590: observe that the mean frequency of the spiking behavior in
591: response to the IP$_3$ stimuli is well characterized by Hill
592: curves. The Hill coefficient, between 1.8-2.1, was essentially
593: independent of [Ca$^{2+}$]$_{ave}$. This value is too high given
594: narrow dynamical range in order to frequency-encode the IP$_3$
595: signal. However, this narrow dynamical range observed in isolated
596: cluster, could be translates into a wider dynamical range as
597: result of the collective phenomenon by considering an intercluster
598: communication mechanism. This interesting discussion is the
599: subject of our current research and will be discussed in a
600: forthcoming paper.
601:
602: The small number of the IP$_3$Rs in a cluster introduce stochastic
603: oscillation into the calcium release dynamics. Experimental data
604: suggest that the localized calcium release varies in a continuous
605: fashion due to stochastic variation in both numbers of channels
606: recruited and durations of channel openings. It is not obvious
607: which aspects of Ca$^{2+}$ puffs are originally determined by the
608: dynamics of the Ca$^{2+}$ channels, which properties are
609: determined by the diffusion and Ca$^{2+}$ binding kinetics, and
610: which properties are induced from the measurement procedures
611: (Cheng et al., 1999; Izu et al., 1998). Although in several
612: experiments one-peak amplitude distributions were observed, Cheng
613: et al., (1999) suggested that the original calcium puffs should
614: have an exponential decaying decreasing amplitude distribution.
615: Our stochastic simulations shown that the amplitude distributions
616: varies in a continuous fashion from exponential decay to two-peak
617: distributions. Even for a small threshold (as 0.20 $\mu$M) a peak
618: is observed in the amplitude distributions for high values both of
619: [IP$_3$] and [Ca$^{2+}$]$_{ave}$. However, the puffs size (i.e.
620: amplitude $\times$ duration) distributions have essentially an
621: exponential decay distribution over the range studied here. These
622: results indicate that a fixed puff morphology which is sometimes
623: assumed in literature (Pratusevich and Balke, 1996; Izu et al.,
624: 1998; Cheng et al., 1999) is not a good assumption for Ca$^{2+}$
625: puff analysis. Furthermore, in our markovian simulation, we found
626: that the amplitude of puff is correlated with its duration in
627: agreement with experimental observation reported by Thomas et al.,
628: (1998).
629:
630: Other useful way to characterize the calcium puffs dynamics is
631: using the IPI distributions. The stochastic simulation of the OTM
632: shown that the IPI distributions varies from exponential decay to
633: a skewed bell-shape distributions. For physiological values of
634: IP$_3$ and [Ca$^{2+}$]$_{ave}$ the distributions have a maximum
635: around 1.5 seconds. The shape is consistent with distribution
636: observed by Marchant et al. (1999), however, the typical IPI
637: observed in the experiment, around 2.5 seconds, was slightly
638: higher. In contrast to the deterministic models, we shown that two
639: consecutive IPI are not correlated. For this reason, the random
640: spiking behavior or stochastic oscillations have an essentially
641: different nature that the periodic oscillations generated by
642: deterministic models.
643:
644: We want to remark that the stochastic oscillations are different
645: from the stochastic excitability discussed by Keizer and Smith
646: (1998). For the stochastic excitability, once [Ca$^{2+}$] randomly
647: becomes larger than a threshold, a fast release
648: (action-potential-like) of [Ca$^{2+}$] followed by a refractory
649: period can be observed. For the stochastic oscillation studied
650: here there is not such a threshold. More broad distributions of
651: puff amplitudes and IPI are yielded for stochastic oscillation.
652:
653: \section*{Acknowledges}
654: This research was partially supported by the Deutsche
655: Forschungsgemeinschaft. The author thanks to Gerhard Thiel and
656: Nara Guisoni for valuable suggestions. L.D. is grateful for the
657: hospitality of Professor Friedemann Kaiser and Nonlinear Dynamics
658: Group at the Darmstadt University of Technology where these ideas
659: were developed.
660:
661: %\begin{thebibliography} {99}
662: \section*{References}
663:
664: \noindent B\"ar, M., M. Falcke, H. Levine, and L. S. Tsimring. 2000.
665: Discrete stochastic modeling of calcium channel dynamics. {\it
666: Phys. Rev. Lett.} 84:5664-5667.
667:
668: \noindent Berridge, M. J., 1997. Elementary and global aspect of
669: calcium signaling. {\it J. Physiol.} 499:291-306.
670:
671: \noindent Bezprozvanny, I., J. Watras, and B. Ehrlich. 1991.
672: Bell-shaped calcium response curves of Ins(1,4,5)P$_3$- and
673: calcium-gated channels from endoplasmic reticulum of cerebellum.
674: {\it Nature.} 351:751-754.
675:
676: \noindent Bezprozvanny, I., and B. E. Ehrlich. 1994. Inositol
677: 1,4,5-trisphopate (IP$_3$)-gated Ca channels from cerebellum:
678: conduction properties for divalent cations and regulation by
679: intraluminal calcium. {\it J. Gen. Physiol.} 104:821-856.
680:
681: \noindent Bootman, M., E. Niggli, M. Berridge, and P. Lipp. 1997.
682: Imaging the hierarchical Ca$^{2+}$ signaling system in HeLa cells.
683: {\it J. Physiol.} 499:307-314.
684:
685: \noindent Callamaras, N., J. S. Marchant, X. Sun, and I. Parker.
686: 1998. Activation and co-ordination of InsP$_3$-mediated elementary
687: Ca$^{2+}$ events during global Ca$^{2+}$ signals in {\it Xenopus}
688: oocytes. {\it J. Physiol.} 509:81-91.
689:
690: \noindent Callamaras, N., and I. Parker. 2000. Phasic
691: characteristic of elementary Ca$^{2+}$ release sites underlies
692: quantal responses to IP$_3$. {\it EMBO J.} 19:3608-3617.
693:
694: \noindent Champeil, P., L. Combettes, B. Berthon, E. Doucet, S.
695: Orlowski, M. Claret. 1989. Fast kinetics of calcium release
696: induced by myo-inositol trisphosphate in permeabilized rat
697: hepatocytes. {\it J. Biol. Chem.} 264:17665-17673.
698:
699: \noindent Cheng, H., L. Song, N. Shirokova, A. Gonzalez, E. G.
700: Lakatta, E. Rios, and M. D. Stern. 1999. Amplitude distribution of
701: calcium sparks in confocal images: theory and studies with an
702: automatic detection method. {\it Biophys. J.} 76:606-617.
703:
704: \noindent Delisle, S. 1991. The four dimensions of calcium
705: signaling in {\it Xenopus} oocytes. {\it Cell Calcium.}
706: 12:217-227.
707:
708: \noindent De Young, G. W., and J. Keizer. 1992. A single-pool
709: inositol 1,4,5- trisphosphate-receptor-based model for
710: agonist-stimulated oscillations in Ca$^{2+}$ concentration. {\it
711: Proc. Natl. Acad. Sci. USA.} 89:9895-9899.
712:
713: \noindent Dufour, J. F., I. M. Arias, and T. J. Turner. 1997.
714: Inositol 1,4,5-trisphopate and calcium regulate the calcium
715: channel function of the hepatic inositol 1,4,5-trisphopate
716: receptor. {\it J. Biol. Chem.} 272:2675-2681.
717:
718: \noindent Falcke, M., L. Tsimring, and H. Levine. 2000. Stochastic
719: spreading of intracellular Ca$^{2+}$ release. {\it Phys. Rev. E.}
720: 62:2636-2643.
721:
722: \noindent Finch, E.A., T.J. Turner, S. M. Goldin. 1991. Calcium as
723: a coagonist of inositol 1,4,5-trisphosphate-induced calcium
724: release. {\it Science.} 252:443-446.
725:
726: \noindent Izu, L. T., W. G. Wier, and C. W. Balke. 1998.
727: Theoretical analysis of the Ca$^{2+}$ spark amplitude
728: distribution. {\it Biophys. J.} 75:1144-1162.
729:
730: \noindent Keizer, J., and G. D. Smith. 1998. Spark-to-wave
731: transition: saltatory transmission of calcium waves in cardiac
732: myocytes. {\it Biophys. Chem.} 72:87-100.
733:
734: \noindent Keizer, J., G. D. Smith, S. Ponce-Dawson, and J. E.
735: Pearson. 1998. Saltatory propagation of Ca$^{2+}$ waves by
736: Ca$^{2+}$ sparks. {\it Biophys. J.} 75:595-600.
737:
738: \noindent Koster, H., C. van Os, and R. Bindels. 1993. Ca$^{2+}$
739: oscillations in the rabbit renal cortical collecting system
740: induced by Na$^+$ free solutions. {\it Kidney Int.} 43:828-836.
741:
742: %\noindent Li, Y., and J. Rinzel. 1994. Equations for InsP3
743: %receptor-mediated [Ca$^{2+}$]i oscillations derived from a
744: %detailed kinetic model: a Hodgkin-Huxley like formalism. J. Theor.
745: %Biol. 166:461-473.
746:
747: \noindent Lipp, P., and E. Niggli. 1998. Fundamental calcium
748: release events revealed by two-photon excitation photolysis of
749: caged calcium in guinea-pig cardiac myocytes. {\it J. Physiol.}
750: 508:801-809.
751:
752: \noindent Marchant, J. S., and C. W. Taylor. 1997. Cooperative
753: activation of IP$_3$ and Ca$^{2+}$ safeguards against spontaneous
754: activity. {\it Current Biology} 7:510-518.
755:
756: \noindent Marchant, J. S., N. Callamaras, and I. Parker. 1999.
757: Initiation of IP$_3$-mediated Ca$^{2+}$ waves in {\it Xenopus}
758: oocytes. {\it EMBO J.} 18:5285-5299.
759:
760: \noindent Marchant, J. S., and I. Parker. 2001. Role of elementary
761: Ca$^{2+}$ puffs in generating repetitive Ca$^{2+}$ oscillations.
762: {\it EMBO J.} 20:65-76.
763:
764: %\noindent Moraru, I. I., E. J. Kaftan, B. E. Ehrlich, and J.
765: %Watras. 1999. Regulation of type 1 inositol
766: %1,4,5-trisphosphate-gated calcium channels by InsP3 and calcium:
767: %simulation of single channel kinetics based on ligand binding and
768: %electrophysiological analysis. J. Gen. Physiol. 113: 837-849.
769:
770: \noindent Othmer, H. G., and H. Tang. 1993. In Oscillations and
771: waves in a model of calcium dynamics. {\it In} Experimental and
772: Theoretical Advances in Biological Pattern Formation. H. Othmer,
773: J. Murray, and P. Maini, editors. Plenunm Press, London. 295-319,
774:
775: \noindent Parker, I., Y. Yao, V. Ilyin. 1996. Fast kinetics of
776: calcium liberation induced in {\it Xenopus} oocytes by
777: photoreleased inositol trisphosphate. {\it Biophys J.} 70:222-237.
778:
779: \noindent Pratusevich, V. R., and C. W. Balke. 1996. Factors
780: shaping the confocal image of the calcium spark in cardiac muscle
781: cells. {\it Biophys. J.} 71:2942-2957.
782:
783: %\noindent Roberts, W. M. 1993. Spatial calcium buffering in
784: %saccular hair-cells. Nature. 363:74-76.
785:
786: \noindent Shuai, J. W., and P. Jung. 2002a. Optimal intracellular
787: calcium signaling. {\it Phys. Rev. Lett.} 88:68102-1-68102-4.
788:
789: \noindent Shuai, J. W., and P. Jung. 2002b. Stochastic properties
790: of Ca$^{2+}$ release of inositol 1,4,5-trisphosphate receptor
791: cluster. {\it Biophys. J.} 83:87-97.
792:
793: \noindent Sun, X., N. Callamaras, J. S. Marchant, and I. Parker.
794: 1998. A continuum of InsP$_3$-mediated elementary Ca$^{2+}$
795: signaling events in {\it Xenopus} oocytes. {\it J. Physiol.}
796: 509:67-80.
797:
798: \noindent Swillens, S., G. Dupont, L. Combettes, and P. Champeil.
799: 1999. From calcium blips to calcium puffs: theoretical analysis of
800: the requirements for interchannel communication. {\it Proc. Natl.
801: Acad. Sci. USA.} 96:13750-13755.
802:
803: \noindent Tang, Y., and H. Othmer. 1995. Frequency encoding in
804: excitable systems and its application to calcium oscillations.
805: {\it Proc. Natl. Acad. Sci. USA.} 92:7869-7873.
806:
807: \noindent Tang, Y., J.L. Stephenson, and H. Othmer. 1996.
808: Simplification and analysis of the models calcium dynamics based
809: on IP$_3$-sensitive calcium channel kinetics. {\it Biophis. J.}
810: 70:246-263.
811:
812: \noindent Thastrup, O., P. Cullen, B. Drobak, M. Hanley, and A.
813: Dawson. 1990. Thapsigargin, A tumor promoter, discharges
814: intracellular Ca$^{2+}$ stores by specific inhibition of the
815: endoplasmic reticulum Ca$^{2+}$ ATPase. {\it Proc. Natl. Acad.
816: Sci. USA.} 87:2466-2470.
817:
818: \noindent Thomas, D., P. Lipp, M. J. Berridge, and M. D. Bootman.
819: 1998. Hormone-evoked elementary Ca$^{2+}$ signals are not
820: stereotypic, but reflect activation of different size channel
821: clusters and variable recruitment of channels within a cluster.
822: {\it J. Biol. Chem.} 273:27130-27136.
823:
824: \noindent Volpe, P., B. H. Alderson-Lang, and G. A. Nickols. 1990.
825: Regulation of inositol 1,4,5-trisphospate-induced Ca$^{2+}$
826: release. I. Effect of Mg$^{2+}$. {\it American J. of Physiology.}
827: 258:C1077-C1085.
828:
829: \noindent Watras, J., I. Bezprozvanny, and B. E. Ehrlich. 1991.
830: Inositol 1,4,5-trisphopate-gated channels in cerebellum: presence
831: of multiple conductance states. {\it J. Neurosci.} 11:3239-3245.
832:
833:
834: %\end{thebibliography}
835:
836: \newpage
837:
838: \section*{Tables and Figures}
839:
840: \begin{table} [h]
841: \begin{tabular}{|ll||ll|} \hline
842: \multicolumn{2}{|c||}{\underline{Kinetic parameters}} &
843: \multicolumn{2}{|c|}{\underline{Other parameters}}\\
844: name&value& name& value \\ \hline
845: $k_{1}$ & 12.0 $(\mu$M$\times$ s$)^{-1}$ & $v_r$ & 1.85 \\
846: $k_{-1}$ & 8.0 s$^{-1}$ & $\gamma _{0}$ & 0.1 s$^{-1}$\\
847: $k_{2}$ & 23.4 $(\mu$M$\times \ s)^{-1}$ & $\gamma _{1}$ & 20.5 s$^{-1}$\\
848: $k_{-2}$ & 1.65 s$^{-1}$ & $\hat{p}_1$ & 8.5 $\mu$M$\times $ s$^{-1}$ \\
849: $k_{3} $ & 2.81 $(\mu$M$\times \ s)^{-1}$ & $\hat{p}_2$ & 0.065 $\mu$M\\
850: $k_{-3}$ & 0.21 s$^{-1}$ & $N$ & 20 \\
851: \hline
852: \end{tabular}
853: \caption{Parameters values for the OTM.}
854: \end{table}
855:
856: %\end{document}
857:
858: \vspace{3cm}
859: \begin{figure}[ht]
860: \includegraphics[width=15cm]{Fig_1.eps} \caption{Mean value
861: of [Ca$^{2+}$] versus IP$_3$ concentration, for six different
862: concentrations of the average intracellular calcium
863: [Ca$^{2+}$]$_{ave}$, computed over 10,000 seconds.}
864: \end{figure}
865:
866: \begin{figure}[ht]
867: \includegraphics[width=15cm]{Fig_2.eps}
868: \caption{Fraction of open channels versus IP$_3$ concentration,
869: for six different concentrations of the average intracellular
870: calcium [Ca$^{2+}$]$_{ave}$, computed over 10,000 seconds.}
871: \end{figure}
872:
873: \begin{figure}[ht]
874: \includegraphics[width=10cm]{Fig_3.eps}
875: \caption{Top panel: Hill coefficient $n_h$ for IP$_3$-induced
876: calcium release versus [Ca$^{2+}$]$_{ave}$, obtained fitting data
877: displayed in figure 2. Bottom panel: $k$ and $V_{max}$ values of
878: the Hill curves for IP$_3$-induced calcium release versus
879: [Ca$^{2+}$]$_{ave}$, obtained fitting data displayed in Fig. 2.}
880: \end{figure}
881:
882: \begin{figure}[ht]
883: \includegraphics[width=10cm]{Fig_4.eps}
884: \caption{The theoretically predicted probability to find the open channel
885: as a function of cytosolic [Ca$^{2+}$] for three different values of IP$_3$
886: concentration (0.01 $\mu$M solid line, 0.1 $\mu$M dotted line,
887: and 1$\mu$M dot-dashed line).}
888: \end{figure}
889:
890: \begin{figure}[ht]
891: \includegraphics[width=15cm]{Fig_5.eps}
892: \caption{Top panel: 100 seconds of the stochastic temporal
893: behavior of the calcium release from a cluster with 20 IP$_3$Rs
894: ($N=20$, [Ca$^{2+}$]$_{ave}$=1.6 and IP$_3$=0.35). Bottom-left
895: panel: scatter plot of IPI$(i)$ versus IPI$(i+1)$ shown that there
896: is not temporal correlation between two consecutives puffs.
897: Bottom-center panel: scatter plot of puff amplitude versus puff
898: duration shown that there is some correlation between puff
899: durations and amplitudes of puffs. Bottom-right panel: puff area
900: distribution shown that there is Ca$2+$ release of all sizes. All
901: plots in the bottom panel were computed with puffs which
902: amplitudes exceeds the threshold 0.20 $\mu$M, recorded when
903: simulate 10,000 seconds.}
904: \end{figure}
905:
906: \begin{figure}[ht]
907: \includegraphics[width=15cm]{Fig_6.eps}
908: \caption{Mean frequency versus IP$_3$ concentration, for five
909: different values of [Ca$^{2+}$]$_{ave}$. All plots were computed
910: with puffs which amplitudes exceeds the threshold 0.20 $\mu$M,
911: recorded when simulate 10,000 seconds.}
912: \end{figure}
913:
914: \begin{figure}[ht]
915: \includegraphics[width=15cm]{Fig_7.eps}
916: \caption{Diagram of the puff amplitude distributions is shown in
917: the [Ca$^{2+}$]$_{ave}$-IP$_3$ plane. The distributions was
918: computed with all puffs, which amplitudes exceeds the threshold
919: 0.20 $\mu$M, recorded in 10,000 seconds. In this condition no
920: events (NE) were recorded for [IP$_3$]= 0.07 $\mu$M and
921: [Ca$^{2+}$]$_{ave}$ = 0.8 $\mu$M. The range of the axis are the
922: same for all distributions, between [0.20,1.8] $\mu$M for the
923: amplitudes of puffs (horizontal axe), and between [0,6000] for the
924: number of events (vertical axe).}
925: \end{figure}
926:
927: \begin{figure}[ht]
928: \includegraphics[width=15cm]{Fig_8.eps}
929: \caption{Diagram of the inter puff interval (IPI) distributions is
930: shown in the [Ca$^{2+}$]$_{ave}$-IP$_3$ plane. The distributions
931: was computed with all puffs, which amplitudes exceeds the
932: threshold 0.20 $\mu$M, recorded in 10000 seconds. In this
933: condition no events (NE) were recorded for [IP$_3$]= 0.07 $\mu$M
934: and [Ca$^{2+}$]$_{ave}$ = 0.8 $\mu$M. The range of the vertical
935: axis are the same for all distributions, between [0,6000] events.
936: While the ranges of IPI (horizontal axis) are between [0,80]
937: seconds for the first column ([Ca$^{2+}$]$_{ave}$=0.8 $\mu$M), and
938: between [0,20] seconds in all other cases.}
939: \end{figure}
940:
941: \begin{figure}[ht]
942: \includegraphics[width=15cm]{Fig_9.eps}
943: \caption{Top panel: 100 seconds of the stochastic temporal
944: behavior of the calcium release from a cluster with 20 IP$_3$Rs
945: ($N=20$, [Ca$^{2+}$]$_{ave}$=0.8 $\mu$M and IP$_3$=21.0 $\mu$M).
946: Bottom-left panel: scatter plot of IPI$(i)$ versus IPI$(i+1)$
947: shown that there is not temporal correlation between two
948: consecutives puffs. Bottom-center panel: scatter plot of puff
949: amplitude versus puff duration shown that there is some
950: correlation between puff durations and amplitudes of puffs.
951: Bottom-right panel: puff area distribution shown that there is
952: Ca$^{2+}$ release of all sizes. All plots in the bottom panel were
953: computed with puffs which amplitudes exceeds the threshold 0.20
954: $\mu$M, recorded in 10,000 seconds.}
955: \end{figure}
956:
957: \begin{figure}[ht]
958: \includegraphics[width=15cm]{Fig_10.eps}
959: \caption{Top panel: 100 seconds of the stochastic temporal
960: behavior of the calcium release from a cluster with 20 IP$_3$Rs
961: ($N=20$, [Ca$^{2+}$]$_{ave}$=2.4 $\mu$M and IP$_3$=0.07 $\mu$M).
962: Bottom-left panel: scatter plot of IPI$(i)$ versus IPI$(i+1)$
963: shown that there is not temporal correlation between two
964: consecutives puffs. Bottom-center panel: scatter plot of puff
965: amplitude versus puff duration shown that there is some
966: correlation between puff durations and amplitudes of puffs.
967: Bottom-right panel: puff area distribution shown that there is
968: Ca$^{2+}$ release of all sizes. All plots in the bottom panel were
969: computed with puffs which amplitudes exceeds the threshold 0.20
970: $\mu$M, recorded in 10,000 seconds.}
971: \end{figure}
972:
973:
974: \end{document}
975: \begin{figure}[ht]
976: \includegraphics[width=15cm]{Fig_2124.eps}
977: \caption{Top panel: 100 seconds of the stochastic temporal
978: behavior of the calcium release from a cluster with 20 IP$_3$Rs
979: ($N=20$, [Ca$^{2+}$]$_{ave}$=2.4 and IP$_3$=21). Bottom-left
980: panel: scatter plot of IPI$(i)$ versus IPI$i+1$ shown that there
981: is not temporal correlation between two consecutives puffs.
982: Bottom-center panel: scatter plot of puff amplitude versus puff
983: duration shown that there is some correlation between puff
984: durations and amplitudes of puffs. Bottom-right panel: puff area
985: distribution shown that there is Ca$2+$ release of all sizes. All
986: plots in the bottom panel were computed with puffs which
987: amplitudes exceeds the threshold 0.15 $\mu M$, recorded in 10000
988: seconds. The others parameters are the same that top panel
989: ($N=20$, [Ca$^{2+}$]$_{ave}$=2.4 and IP$_3$=21). }
990: \end{figure}
991: