bb10b31.qxd
02/08/2000
86
10:43
Page 86
Research Paper
Differential codes for free Ca2+–calmodulin signals in nucleus
and cytosol
Mary N. Teruel*, Wen Chen*, Anthony Persechini† and Tobias Meyer*‡
Background: Many targets of calcium signaling pathways are activated or
inhibited by binding the Ca2+-liganded form of calmodulin (Ca2+–CaM). Here,
we test the hypothesis that local Ca 2+–CaM-regulated signaling processes
can be selectively activated by local intracellular differences in free
Ca2+–CaM concentration.
Results: Energy-transfer confocal microscopy of a fluorescent biosensor was
used to measure the difference in the concentration of free Ca2+–CaM between
nucleus and cytoplasm. Strikingly, short receptor-induced calcium spikes
produced transient increases in free Ca2+–CaM concentration that were of
markedly higher amplitude in the cytosol than in the nucleus. In contrast,
prolonged increases in calcium led to equalization of the nuclear and cytosolic
free Ca2+–CaM concentrations over a period of minutes. Photobleaching
recovery and translocation measurements with fluorescently labeled CaM
showed that equalization is likely to be the result of a diffusion-mediated net
translocation of CaM into the nucleus. The driving force for equalization is a
higher Ca2+–CaM-buffering capacity in the nucleus compared with the cytosol,
as the direction of the free Ca2+–CaM concentration gradient and of CaM
translocation could be reversed by expressing a Ca2+–CaM-binding protein at
high concentration in the cytosol.
Addresses: *Department of Cell Biology and
Department of Pharmacology and Cancer Biology,
Box 3709, Duke University Medical Center,
Durham, North Carolina 27710, USA. †Department
of Pharmacology and Physiology, Box 711,
University of Rochester School of Medicine and
Dentistry, Rochester, New York 14642, USA.
Present address: ‡Department of Molecular
Pharmacology, Stanford University School of
Medicine, 300 Pasteur Drive, SUMC L333
Stanford, California 94305-5332, USA.
Correspondence: Tobias Meyer
E-mail: [email protected]
Received: 17 September 1999
Revised: 16 November 1999
Accepted: 2 December 1999
Published: 11 January 2000
Current Biology 2000, 10:86–94
0960-9822/00/$ – see front matter
© 2000 Elsevier Science Ltd. All rights reserved.
Conclusions: Subcellular differences in the distribution of Ca 2+–CaM-binding
proteins can produce gradients of free Ca 2+–CaM concentration that result in
a net translocation of CaM. This provides a mechanism for dynamically
regulating local free Ca2+–CaM concentrations, and thus the local activity of
Ca2+–CaM targets. Free Ca2+–CaM signals in the nucleus remain low during
brief or low-frequency calcium spikes, whereas high-frequency spikes or
persistent increases in calcium cause translocation of CaM from the
cytoplasm to the nucleus, resulting in similar concentrations of nuclear and
cytosolic free Ca2+–CaM.
Background
The protein calmodulin (CaM), a ubiquitous mediator of
intracellular signaling, can bind cytosolic and nuclear targets
in the absence or presence of bound calcium ions [1–4].
Examples of proteins that are regulated by the Ca2+-liganded form of CaM (Ca2+–CaM) include ion channels,
protein kinases and phosphatases, calcium pumps, adenylate cyclases, and phosphodiesterases [5–7]. Nuclear and
cytosolic proteins that bind Ca2+–CaM mediate diverse signaling pathways that regulate processes as diverse as
cytoskeletal structure changes, gene expression, and
synaptic plasticity. The large number of putative functions
for Ca2+–CaM raises the important question of how it
selectively activates its targets in the cell. As the relevant
parameter for activation of a particular CaM target is the
free concentration of Ca2+–CaM, important insights into
local CaM function would be obtained if this parameter
could be measured locally in nucleus and cytosol.
Given that the total concentration of CaM in the cell is
limiting [5–7], and that complexes of Ca2+–CaM and its
targets have dissociation constants that range from less
than 1 nM to greater than 100 nM [8], the free
Ca2+–CaM concentration is likely to be regulated by the
relative concentrations of CaM and its various targets.
Three mechanisms are likely to be in operation at the
same time. First, the total concentration of CaM can be
controlled by its regulated expression and degradation
[9]. Second, the free Ca2+–CaM concentration can be
regulated by increasing or decreasing the affinity of
abundant CaM-binding proteins. This hypothesis is
derived from the finding that phosphorylation can
produce changes in the binding affinities of several
Ca2+–CaM-binding proteins [8,10,11]. Third, the local
concentration of Ca2+–CaM might be regulated by
regional differences in the Ca2+–CaM-binding capacity.
For example, as a higher CaM-binding capacity has been
bb10b31.qxd
02/08/2000
10:43
Page 87
Research Paper Differential Ca2+–calmodulin signals in nucleus and cytosol Teruel et al.
measured in the nucleus [12,13], receptor-triggered Ca2+
signals should produce lower levels of free Ca 2+–CaM
concentration in the nucleus than in the cytosol.
Can local differences in free Ca2+–CaM concentration
persist in cells? Although, according to diffusion theory,
a uniform increase in the free calcium concentration will,
over time, lead to a uniform increase in the free
Ca2+–CaM concentration, local gradients may exist for
time periods that are limited by the diffusion of CaM.
Such gradients in free Ca2+–CaM concentration have not
yet been directly measured. Earlier studies showed that
an increase in calcium concentration can, nevertheless,
trigger the translocation of CaM to the nucleus and other
cell regions [12–15], a process which may require free
diffusion [12], facilitated diffusion [16], or, possibly,
active transport of CaM [14]. Here we ask whether the
observed nuclear translocation of CaM may in fact
reflect the formation of a dynamic cellular gradient in
the free Ca2+–CaM concentration.
A potential alternative mechanism for generating different concentrations of nuclear and cytosolic free
Ca2+–CaM is the existence of persistent gradients in
calcium concen-tration between nucleus and cytosol, as
suggested by previous studies (reviewed in [17,18]).
Such differences are, however, unlikely to be relevant
to the generation of persistent free Ca 2+–CaM gradients
in the rat basophilic leukemia (RBL) cell line we are
studying. Studies with fluorescent calcium indicators in
RBL cells, as well as in other cell systems, have shown
rapid nuclear responses after calcium was released from
cytosolic stores or entered cells across the plasma membrane [19–24]. This suggests that, at least in these
cell types, differences in the calcium concentration
between nucleus and cytosol are equilibrated in less
than a second.
The recent development of a fluorescent energy-transfer
probe that reports the concentration of free Ca2+–CaM
now allows the direct investigation of subcellular differences in free Ca2+–CaM concentration [25,26]. This
probe works by monitoring alterations in fluorescence
resonance energy transfer (FRET) between cyan and
yellow fluorescent proteins (CFP and YFP, respectively)
linked by M13, a short Ca2+–CaM-binding peptide
(CFP–M13–YFP or FIP–CBSM). Using confocal energytransfer imaging of FIP–CBSM, we found that cytosolic
CaM-dependent processes may indeed be regulated differently from nuclear ones as a result of subcellular differences that can be produced in the free Ca2+–CaM
concentrations. In particular, short calcium transients
trigger lower nuclear versus cytosolic Ca2+–CaM signals,
whereas persistent calcium increases lead to a slow
equalization of the Ca2+–CaM concentrations in nucleus
and cytosol.
87
Results
Measurement of changes in nuclear and cytosolic free
Ca2+–CaM concentrations
Receptor-induced calcium spikes or persistent calcium
increases were generated in RBL cells, a tumor mast-cell
line. As shown in measurements with the calcium indicator
Fluo3-AM, stimulation of the P2-type purinergic receptor
with ATP triggered brief individual calcium spikes
(Figure 1a, left panel) [27], whereas crosslinking of the IgE
Fc receptor (FcεRI) triggered repetitive calcium spikes or
persistent increases in calcium concentration, depending on
the concentration of antigen added and the particular cell
studied (Figure 1a, center and right panels) [28,29]. The
local concentration of free Ca2+–CaM corresponding to
these different patterns of calcium signals was measured by
expressing FIP–CBSM, an energy-transfer-based fluorescent
indicator that binds Ca2+–CaM with 0.4 nM affinity. The
basic properties of this fluorescent probe are described in
[25], and are shown schematically in Figure 1b. The change
in nuclear versus cytosolic calcium concentration was monitored by confocal imaging of FIP–CBSM using a 442 nm
helium–cadmium laser for excitation. We found this laser
line to be suitable for exciting the probe, while producing a
relatively small amount of cellular autofluorescence. At
each data point, the CFP emission (between 450–500 nm)
and the YFP emission (> 522 nm) were measured simultaneously and then ratioed (Figure 1c).
Strikingly, the FcεRI-triggered calcium signals led to an
initial rapid increase in cytosolic Ca2+–CaM concentration
and to a near-simultaneous, but markedly smaller increase in
the amplitude of the nuclear free Ca2+–CaM concentration
(Figure 1d). An approximate calibration of the free concentration of Ca2+–CaM [26] is shown in color code on Figure 1d.
Time course of nuclear and cytosolic free Ca2+–CaM
signals after receptor stimulation
The time course of nuclear versus cytosolic free Ca2+–CaM
signals was analyzed by taking a series of simultaneous
dual images and plotting the ratio of the CFP and YFP
emissions for the nucleus and the cytosol as a function of
time. For ATP-receptor stimuli that trigger an individual
calcium spike, the increase in the free Ca2+–CaM concentration in the nucleus was markedly lower than in the cytoplasm (Figure 2a; n = 6). For repetitive calcium spikes
triggered by low concentrations of antigen (Figure 2b;
n = 6), the peak nuclear amplitudes of free Ca2+–CaM
remained lower than those in the cytosol, albeit with a
baseline increase in the free Ca2+–CaM concentration in
many of the experiments. In contrast, persistent increases
in the free Ca2+ concentration produced lower nuclear
versus cytosolic free Ca2+–CaM signals that slowly equalized over a time scale of several minutes (Figure 2c; n = 6).
The same kinetic differences in nuclear and cytosolic free
Ca2+–CaM concentrations were observed when stimulation
bb10b31.qxd
02/08/2000
88
10:43
Page 88
Current Biology Vol 10 No 2
Figure 1
Measurement of the free Ca2+–CaM
concentration in RBL cells after
antigen-induced crosslinking of FcεRI.
(a) Calcium responses in RBL cells measured
by confocal microscopy using Fluo-3AM as a
fluorescent indicator after: (left panel) addition
of 100 µM ATP; (center panel) minimal antigen
stimulation (2 ng/ml DNP–BSA, final
concentration); and (right panel) maximal
antigen stimulation (2 µg/ml DNP–BSA, final
concentration). (b) Schematic view of the
FIB–CBSM probe used to measure free
Ca2+–CaM concentration (modified from [25]).
(c) Fluorescence images in (left panel) the
cyan channel and (middle panel) the yellow
channel simultaneously measured by confocal
microscopy. The right panel is a gray-scale
representation of the ratio image of the left
and middle channels. (d) False-color confocal
ratio images of CFP/YFP fluorescence in RBL
cells stimulated with 2 µg/ml DNP–BSA.
Images of the free Ca2+–CaM concentration
are shown: (left panel) before stimulation;
(center panel) at t = 50 sec; and (right panel)
at t = 120 sec after stimulation. Initially, the
free Ca2+–CaM concentration is lower in the
nucleus (arrows) than in the cytosol.
protocols using calcium ionophores were used instead of
receptor stimulation. Figure 3a shows ratio images
acquired before the addition of ionomycin (left panel),
60 seconds after addition of ionomycin in the presence of
extracellular calcium (center panel), and then immediately
after the addition of EGTA (right panel). When sustained
increases in the free Ca2+ concentration were generated by
the addition of ionomycin in the presence of extracellular
calcium (Figure 3b; n = 20), the measured nuclear-tocytosolic signal ratio again showed an equalization
between nuclear and cytosolic free Ca2+–CaM concentration that occurred over several minutes. Figure 3c
shows that brief calcium spikes also triggered markedly
lower peak free Ca2+–CaM concentrations in the nucleus
than in the cytosol (n = 6). These brief calcium spikes
were generated by first placing the cells in a low-calcium
Figure 2
Time course of the changes in nuclear versus cytosolic free
Ca2+–CaM signals after physiological stimulation. The concentration
of free Ca2+–CaM in each compartment was measured by the ratio
of CYP to YFP fluorescence from the reporter construct FIB–CBSM.
Blue lines represent nuclear concentrations and green lines,
cytosolic concentrations. (a) Typical example of the time course of
changes in nuclear and cytosolic free Ca2+–CaM concentrations
triggered by an ATP-induced single calcium spike. (b) Typical
example of the time course of nuclear versus cytosolic free
Ca2+–CaM concentrations after a minimal antigen stimulus
(2 ng/ml DNP–BSA, final concentration). (c) Typical example of the
time course of nuclear versus cytosolic free Ca2+–CaM
concentrations after a maximal antigen stimulus (2 µg/ml DNP–BSA,
final concentration).
bb10b31.qxd
02/08/2000
10:43
Page 89
Research Paper Differential Ca2+–calmodulin signals in nucleus and cytosol Teruel et al.
89
Figure 3
Comparison of nuclear and cytosolic free
Ca2+–CaM responses after short and
persistent calcium increases induced by
calcium ionophore. (a) Confocal ratio images
of the free Ca2+–CaM concentration in RBL
cells (left panel) before the addition of 1 µM
ionomycin, (center panel) 60 sec after addition
of 1 µM ionomycin in the presence of
extracellular calcium, and (right panel)
immediately after the addition of 10 mM
MgEGTA. (b) Time course of nuclear (blue)
and cytosolic (green) free Ca2+–CaM
concentrations after the addition of 1 µM
ionomycin (final concentration) at t = 0 sec.
(c) Time course of nuclear (blue) and cytosolic
(green) free Ca2+–CaM concentrations after a
short calcium spike (1 µM ionomycin added at
t = 0 sec, followed 3 sec later by the addition
of 10 mM MgEGTA). (d) Time course of
nuclear (blue) and cytosolic (green) free
Ca2+–CaM concentrations after the addition
of 1 µM ionomycin (final concentration) at
t = 0 sec, measured using the FIP–CBSM-39
probe (Kd = 400 nM).
extracellular buffer for 5 minutes to deplete internal
calcium stores, and then adding back 1.5 mM calcium and
1 µM ionomycin, followed 3 seconds later by the addition
of 10 mM MgEGTA buffer.
Because the expression of high-affinity Ca2+–CaM indicators is known to reduce the amplitude of the measured
free Ca2+–CaM responses [26], we also measured the free
Ca2+–CaM concentration using a lower-affinity probe,
FIP–CBSM-39, which has a Kd of 400 nM for Ca2+–CaM.
Although the fluorescence signal from this indicator did
not saturate in response to calcium ionophore, its relative
fluorescence responses are small and difficult to resolve,
particularly for physiological stimuli. Figure 3d shows that
equalization of nuclear and cytosolic free Ca2+–CaM
signals can, nevertheless, also be observed with this indicator after ionomycin addition (n = 7). The response
observed in this experiment is probably due to a biphasic
change in the free Ca2+ concentration caused by the transient release of Ca2+ from intracellular stores followed by a
persistent influx via the plasma membrane.
Calmodulin rapidly exchanges between nucleus and
cytosol at low and high calcium concentrations
The difference between cytosolic and nuclear free
Ca2+–CaM concentrations that persisted for several minutes
may reflect differences in free calcium concentrations or
the presence of a nuclear permeability barrier for CaM.
Since nuclear and cytosolic calcium signals in RBL cells
equilibrate in less than a second [24], persistent differences in their amplitudes are an unlikely explanation for
the observed minute-long gradient in free Ca2+–CaM concentration. This leaves the alternative possibility, that the
nuclear envelope significantly slows the exchange of CaM
which, in turn, would delay the equalization of cytosolic
and nuclear free Ca2+–CaM concentrations.
A photobleaching protocol [12,30,31] using either GFP (as
a control) or CaM fluorescently labeled with fluorescein
iso-thiocyanate (FITC–CaM) was used to measure the
exchange rate of proteins between the nucleus and
cytosol at low and high calcium concentrations. A previous
photobleaching measurement at low calcium concentrations has shown an exchange time of CaM in the order of
minutes [12]. We directed a brief, 2 µm diameter, 488 nm
laser pulse at the nucleus to rapidly reduce nuclear
FITC–CaM or the control GFP fluorescence (Figure 4a;
the pulse was applied between the first and second
panels). The time for equalization between cytosol and
nucleus was nearly indistinguishable for FITC–CaM and
GFP at both low and high calcium concentrations. This
comparison of the relative photobleaching recovery of the
bb10b31.qxd
02/08/2000
90
10:43
Page 90
Current Biology Vol 10 No 2
Figure 4
Nuclear photobleaching recovery studies of
FITC-labeled CaM (FITC–CaM) compared
with GFP. (a) Sequential images from a typical
photobleaching recovery experiment. The
series shown is for RBL cells transfected with
GFP. Between the first and second panels, a
2 µm diameter, 488 nm laser spot was
focused onto the cell’s nucleus, and nearly all
nuclear fluorescence was reduced by a
10 msec photobleaching pulse. The third and
fourth panels show the same cell 1 and 4 min
after the photobleaching pulse. (b) Time
course of the relative photobleaching recovery
for nuclear FITC–CaM (blue line) and nuclear
GFP (green line), measured at baseline
calcium concentration. (c) Time course of
photobleaching recovery in the presence of
ionomycin. FITC–CaM (blue line); GFP (green
line). Whereas the nuclear permeability rate for
CaM as well as GFP is increased in the
presence of ionomycin, the relative
permeability rates of GFP and CaM remain the
nucleus is shown at basal calcium concentration in
Figure 4b (n = 15) and in the presence of 1 µM ionomycin
in Figure 4c (n = 15). The faster equilibration time of the
two proteins at high calcium concentration may reflect
calcium-induced differences in nuclear pore permeability.
Independently of the free calcium concentration, the
measured exchange times are more rapid than those
reported for active nuclear transport of CaM-binding proteins [31,32]. This suggests that the relatively small sizes
of GFP and CaM allow these proteins to equilibrate
between the nucleus and cytosol by diffusion.
Because a gradient in free Ca2+–CaM concentration exists
between the nucleus and cytosol, and because CaM diffuses in and out of the nucleus, a gradient in free
Ca2+–CaM concentration should induce a net translocation
of CaM into the nucleus. Indeed, Figure 5a shows that the
concentration of fluorescently labeled CaM was markedly
same at low and high calcium concentration.
The y-axis values (nuclear/cytosol fluorescence
intensities) were normalized.
enhanced in the nucleus after ionomycin addition.
Figure 5a shows the distribution before (left panel), and
5 minutes after (right panel), ionomycin addition. This
observed translocation of CaM is consistent with previously reported nuclear enrichment of CaM in other cell
types [12,27,31,32].
Figure 5b shows that the CaM translocation to the
nucleus is reversible when a transient calcium signal is
applied. Calcium levels were lowered at the end of the
experiment by addition of extracellular EGTA. The
delay times for CaM translocation into and out of the
nucleus were both several minutes (n = 15). Figure 5c
shows that the same time course of nuclear CaM translocation is also observed in response to receptor stimulation
(n = 6). As the time required for CaM translocation into
the nucleus is similar to the time required for photobleaching recovery of fluorescently labeled CaM in the
Figure 5
Calcium-induced reversible translocation of fluorescently labeled CaM
into the nucleus. (a) Addition of ionomycin (final concentration 1 µM)
leads to a marked increase in the total nuclear FITC–CaM concentration.
RBL cells loaded with FITC–CaM were imaged (left panel) before, and
(right panel) 3 min after, the addition of ionomycin. The relative increase
in nuclear fluorescence (arrows) was variable between cells and ranged
from approximately 30 to 300%. (b) Comparison of calcium responses
with the induced CaM translocation. The top panel shows the time
course of the whole-cell change in Ca2+ concentration, measured by
Fluo-3AM, after the addition of 1 µM ionomycin followed by 10 mM
MgEGTA after 6 min. The bottom panel shows the corresponding time
course of the nuclear translocation of FITC–CaM. Translocation into the
nucleus was reversed by lowering the Ca2+ concentration. (c) Time
courses of (top panel) the increase in Ca2+ concentration and
(bottom panel) the corresponding translocation of CaM to the nucleus,
measured in RBL cells stimulated with platelet-activating factor, a
G-protein-coupled receptor agonist (PAF; 100 nM, final concentration).
All y-axis values were normalized.
bb10b31.qxd
02/08/2000
10:43
Page 91
Research Paper Differential Ca2+–calmodulin signals in nucleus and cytosol Teruel et al.
91
Figure 6
Overexpression of a Ca2+–CaM-binding
peptide targeted to the plasma membrane
causes an inversion of the nuclear and
cytosolic Ca2+–CaM gradient and a reversal
of the direction of CaM translocation. The
uniformly distributed Ca2+–CaM-binding
peptide FIP–CBSM was targeted to the
plasma membrane by addition of an aminoterminal myristoylation/palmitoylation
sequence from Lyn. The resulting construct,
PM–FIB–CBSM, was expressed in cells at
high concentrations. (a) Confocal
fluorescence image of an RBL cell transfected
with PM–FIB–CBSM cDNA, showing plasma
membrane localization of PM–FIB–CBSM.
(b) Expression of the uniformly distributed
FIB–CBSM with a fivefold excess of the
plasma-membrane-targeted PM–FIB–CBSM
suppresses the peak Ca2+–CaM
concentration in the cytosol (green) to below
the nuclear concentration (blue). Nuclear and
cytosolic Ca2+–CaM responses are shown
after addition of 1 µM ionomycin. The resulting
signals are effectively the inverse of the
normal situation. (c) Microporation of
FITC–CaM with an excess of PM–FIB cDNA
leads to translocation of FITC–CaM out of the
nucleus and to the time taken for equalization of nuclear
and cytosolic free Ca2+–CaM, it can be argued that
nuclear and cytosolic concentrations of free Ca2+–CaM
are equilibrated by a diffusion-mediated translocation of
CaM into the nucleus.
High levels of a CaM-binding protein in the cytosol reverse
the Ca2+–CaM gradient and direction of CaM translocation
Our results suggest that a higher Ca2+–CaM-binding
capacity in the nucleus compared with the cytosol [12] is
responsible for the transiently lower free Ca2+–CaM concentration in the nucleus. If this interpretation is correct,
artificial overexpression of a Ca2+–CaM-binding protein
in the cytosol should reverse the gradient of free
Ca2+–CaM concentration between nucleus and cytosol
and should also reverse the direction of CaM translocation between nucleus and cytoplasm. To obtain a cytosolic buffer for Ca2+–CaM, we made a plasma membrane
sink for CaM by fusing a palmitoylation/myristoylation
sequence [33] to the high-affinity Ca2+–CaM-binding
M13 construct. The expression of this plasma-membrane-targeted construct (PM–FIP–CBSM) is shown in
Figure 6a. When a fivefold excess of this construct was
coexpressed with the normal, uniformly distributed,
FIP–CBSM, the nuclear–cytosolic gradient was inverted,
and the initial concentration of free Ca2+–CaM in the
nucleus was markedly higher than that in the cytosol
(Figure 6b; n = 4). Even more strikingly, when this
plasma membrane CaM-binding protein was present in
the cell together with fluorescent CaM, calcium signals
triggered CaM translocation out of the nucleus instead of
into it (Figure 6c; n = 8). This shows that the concentration of free Ca2+–CaM can indeed be regulated in
different cell regions by local expression of Ca 2+–CaMbinding proteins.
nucleus after an increase in calcium instead of
into it. FITC–CaM (2 mg/ml) and
PM–FIB–CBSM cDNA (5 mg/ml) were mixed
in a 1:1 ratio and microporated into the cells.
The values obtained (crosses) were
normalized and fitted to an exponential curve
(solid line).
Discussion
Differential control of local free Ca2+–CaM concentration in
nucleus and cytosol
Our measurements suggest that differential regulation of
the local free Ca2+–CaM concentration can provide a way
of preferentially activating CaM targets in a particular cell
region. In this generalized model, transient calcium
increases will generate a lower concentration of free
Ca2+–CaM in cell regions with a higher CaM-binding
capacity. In contrast, persistent calcium signals will lead to
a diffusion-mediated equilibration of the free Ca2+–CaM
concentration. The resulting delayed increase in free
Ca2+–CaM concentration in regions with high
CaM-binding capacities provides a mechanism for the
local decoding of calcium signals.
This model is particularly relevant to nuclear versus cytosolic Ca2+–CaM signals. We found that the nuclear membrane
markedly prolongs the equilibration time for free
Ca2+–CaM. The resulting differential codes for nuclear and
cytosolic free Ca2+–CaM signals are shown in Figure 7.
Whereas persistent calcium signals rapidly increase the
cytosolic free Ca2+–CaM concentration, free Ca2+–CaM concentrations in the nucleus only reach their maximal level
after an equalization time of several minutes. For short
calcium spikes or low-frequency repetitive calcium spikes,
no equalization occurs, and nuclear amplitudes of free
Ca2+–CaM concentration remain small compared with those
in the cytosol. It is likely that such differences in nuclear
versus cytosolic free Ca2+–CaM signals exist in many cell
types, as calcium-triggered nuclear accumulation of CaM
has been observed in several previous studies [12–15].
Although we observed markedly suppressed amplitudes of
nuclear Ca2+–CaM signals in most RBL cells, the study
bb10b31.qxd
02/08/2000
92
10:43
Page 92
Current Biology Vol 10 No 2
Figure 7
Model view of how persistent calcium increases and repetitive calcium
spikes may differentially control nuclear Ca2+–CaM function. If
individual calcium spikes (red) are shorter than the equalization time for
CaM, the nuclear free Ca2+–CaM signal (blue) remains lower than the
cytosolic one (green).
using a plasma membrane Ca2+–CaM sink (overexpressed
PM–FIP–CBSM) showed that an increase in the relative
Ca2+–CaM-buffering capacity of the cytosol can lead to a
higher nuclear Ca2+–CaM signal compared with that in the
cytosol. Thus, by altering the balance of nuclear and
cytosolic Ca2+–CaM targets, cells may be able to enhance
nuclear Ca2+–CaM signals over cytosolic ones. It will be
interesting to learn whether cells indeed change their
nuclear and cytosolic CaM-binding capacities to either
enhance or inhibit relative nuclear Ca2+–CaM signals.
Where could such a selective requirement for nuclear and
cytosolic free Ca2+–CaM signals be important? It is suggestive to consider that the previously shown selective
induction of NFAT-mediated gene expression by high
frequency repetitive calcium spikes [34,35] may in fact
result from a requirement to sufficiently raise the nuclear
free Ca2+–CaM concentration close to the cytosolic one.
Although not tested with the same direct methods, selectivity mechanisms that depend on oscillations may also
exist for CREB activation [36,37]. Our studies now
provide direct evidence for a ‘nuclear frequency code’
hypothesis by showing that short calcium spikes are
insufficient to trigger equilibration of the nuclear and
cytosolic free Ca2+–CaM concentration, whereas high-frequency calcium spikes and prolonged calcium signals
enable equilibration.
Molecular mechanism for the control of the nuclear free
Ca2+–CaM concentration
Although earlier studies suggested that CaM can be
actively transported into the nucleus along with calcineurin [32], Ca2+–CaM kinase II isoforms [38], or possibly by other processes [14], the small size of CaM
(17 kDa) makes it likely that the flux of CaM between
nucleus and cytosol is mainly by diffusion. A diffusion or
facilitated diffusion mechanism has been suggested previously from nuclear CaM influx and efflux studies which
used permeabilized cells combined with blockers of active
transport and externally added CaM sinks [12,16]. Further
evidence for a diffusion mechanism came from photobleaching recovery measurements of nuclear-localized fluorescent CaM in smooth muscle cells which were
performed at low calcium concentration [12]. In RBL
cells, the measurements of nuclear photobleaching recovery at low and high calcium concentration (Figure 5) show
exchange times sufficiently fast to support a passive diffusion model. Active transport mechanisms could, nevertheless, be relevant in neurons and other cells in which a
higher Ca2+–CaM buffering capacity may slow the diffusion times to near or below the rates of active import.
Our results can best be explained by the hypothesis that
the initially lower nuclear free Ca2+–CaM concentration is
a result of a higher buffering capacity in the nucleus compared with the cytosol. High concentrations of nuclear
CaM-binding proteins have been measured biochemically
(reviewed in [13]), and a recent study has shown a relatively higher CaM-buffering capacity in nucleus compared
with cytosol in permeabilized smooth muscle cells [12].
The latter study also proposed that a higher nuclear
binding capacity for CaM, combined with passive diffusion, could be responsible for the observed translocation of
CaM into the nucleus. We investigated whether differential binding capacities could explain the gradients in free
Ca2+–CaM concentration between cytosol and nucleus by
coexpressing the uniformly distributed Ca2+–CaMbinding M13 peptide (FIB–CBSM) with an excess of a
plasma-membrane-targeted version (PM–FIB–CBSM). In
these cells, the initial gradient of free Ca2+–CaM after
ionophore addition was effectively inverted, with a cytosolic free Ca2+–CaM concentration markedly lower than
that in the nucleus (Figure 6b). Even more strikingly, the
direction of the CaM flux was reversed, and CaM now
translocated out of the nucleus into the cytosol
(Figure 6c). Together with the previous measurements,
this finding strongly supports the hypothesis that the
initial gradient in the free Ca2+–CaM concentration results
from a high Ca2+–CaM-buffering capacity in the nucleus.
From a regulatory perspective, this suggests that cells may
control the amplitude of nuclear Ca2+–CaM signals by
increasing or decreasing the concentration and composition of CaM-binding proteins in different cellular regions.
Conclusions
Our study shows that free Ca2+–CaM signals in the nucleus
are suppressed in response to brief or low-frequency
calcium spikes. In contrast, high-frequency calcium spikes
or persistent calcium increases induce maximal nuclear
activation by a delayed equalization of nuclear and cytosolic free Ca2+–CaM signals. Our results are consistent with
the hypothesis that the nucleus has a higher binding
bb10b31.qxd
02/08/2000
10:43
Page 93
Research Paper Differential Ca2+–calmodulin signals in nucleus and cytosol Teruel et al.
capacity for Ca2+–CaM than the cytosol and that CaM can
readily diffuse in and out of the nucleus either alone or
bound to small targets. Thus, the gradient in the free
Ca2+–CaM concentration between cytosol and nucleus
serves as a driving force for a calcium-induced net translocation of CaM to the nucleus, and for an equalization of
nuclear and cytosolic free Ca2+–CaM signals.
Materials and methods
Calmodulin labeling and cDNA constructs
CaM (gift of A.R. Means, Duke University; and Sigma) was fluorescently labeled with fluorescein isothiocyanate (FITC; Sigma) at a ratio
of less than 1 mole FITC per mole CaM. As we found that labeling CaM
with more than 1 mole FITC per mole CaM resulted in a protein that diffused much more slowly (presumably due to artifactual binding of the
fluorophores to cell components), we ensured that all labeling
remained at or below one fluorophore per CaM molecule. FITC–CaM
labeled at a ratio of less than 1 mole FITC per mole CaM diffuses
approximately twice as slowly as the highly mobile GFP. Cells were
loaded with FITC–CaM using a microporator which uses 1 µl sample
and typically loads between 0.1 and 10% of extracellular protein into
adherent cells [39]. As estimated from the relative fluorescence intensity measured in the microscope and the fraction loaded, the total CaM
concentration in the cell used for fluorescence recovery after photobleaching (FRAP) measurements was in the range 1–20 µM.
FIP–CBSM and FIP–CBSM-39 are described in [25] and [26], respectively. For the present study, the fluorescent proteins EBFP and
EGFP in the orginal FIP–CBSM were switched to ECFP and EYFP
(see also [26]). FIP–CBSM was also used as a cytosolic Ca2+–CaM
sink by being targeted to the plasma membrane by the addition of a
conserved myristoylation and palmitoylation sequence from the protein
kinase Lyn (MGCIKSKRKD, single-letter amino-acid notation) at the
amino-terminal end ([33]; D. Holowka, personal communication). This
construct (PM–FIP–CBSM) was expressed at a cDNA ratio of 5:1 with
the cytosolic probe in the studies shown in Figure 6. A concentration of
plasma-membrane-targeted CaM-binding peptide higher than 10 µM
(relative to total cell volume) was estimated to be present in all studies
shown in Figure 6.
For all studies, rat basophilic leukemia (RBL) cells were transfected
by microporation of mRNA, cDNA, or protein [39]. Cells were incubated overnight with 0.2 µg/ml anti-dinitrophenol (DNP) IgE before
antigen stimulation. Dinitrophenylated bovine serum albumin
(DNP–BSA) and anti-DNP monoclonal mouse IgE were obtained from
Sigma. Platelet-activating factor (PAF) was obtained from Sigma, and
RBL cells transfected with the PAF receptor from R. Snyderman’s
group, Duke University.
Confocal microscopy and confocal energy-transfer imaging
Confocal imaging of fluorescence energy transfer between CFP and
YFP was performed with a self-built microscope that used hardware
and software modified from a design by Stephen Smith (Stanford University) and Noam Ziv (Technion, Haifa, Israel). Imaging of CFP and
YFP was done using 442 nm (HeCd laser, Liconix) and 514 nm (argonion laser, Coherent) excitation respectively, and appropriate emission
filters (Chroma). FITC–CaM translocation and Fluo3-AM (Molecular
Probes) responses were measured on an inverted Nikon Diaphot
microscope coupled to an Odyssey confocal imaging system (Noran
Instruments) controlled by Metamorph imaging software (Universal
Imaging Corporation). The 488 nm line from a Coherent Enterprise
laser was coupled directly into the Nikon Diaphot and was used to
carry out the photobleaching measurements of FITC–CaM and GFP.
During imaging, the culture medium was replaced with an extracellular
buffer consisting of 135 mM NaCl, 5 mM KCl, 1.5 mM CaCl2, 1.5 mM
MgCl2, 20 mM glucose, and 20 mM HEPES pH 7.4. All experiments
were performed at room temperature (~25°C).
93
Acknowledgements
We thank S. Smith (Stanford University) and N. Ziv (Technion, Haifa, Israel)
for providing support in the construction of the confocal microscope used
for energy transfer imaging, D. Holowka (Cornell University) for suggesting
the use of the Lyn domain for membrane targeting, and K. Shen and A.
Means (Duke University) for valuable discussion. This work was supported
by National Institutes of Health grants GM-48113 and GM-51457 (T.M.),
1F32NS10767-01 (M.N.T.), and DK-53863 (A.P.).
References
1. Ikura M, Clore GM, Gronenborn AM, Zhu G, Klee CB, Bax A:
Solution structure of a CaM-target peptide complex by
multidimensional NMR. Science 1992, 256:632-638.
2. Nakayama S, Kretsinger RH: Evolution of the EF-hand family of
proteins. Annu Rev Biophys Biomol Struct 1994, 23:473-507.
3. Ikura M: Calcium binding and conformational response in EF-hand
proteins. Trends Biochem Sci 1996, 21:14-17.
4. Rhoads AR, Friedberg F: Sequence motifs for CaM recognition.
FASEB J 1997, 11:331-340.
5. Manalan AS, Klee CB: Calmodulin. Adv Cyclic Nucleotide Protein
Phosphorylation Res 1984, 18:227-278.
6. Means AR, VanBerkum MF, Bagchi I, Lu KP, Rasmussen CD:
Regulatory functions of CaM. Pharmacol Ther 1991, 50:255-270.
7. Weinstein H, Mehler EL: Ca2+-binding and structural dynamics in
the functions of CaM. Annu Rev Physiol 1994, 56:213-236.
8. Meyer T, Hansen P, Stryer L, Schulman H: Calmodulin trapping by
calcium-CaM-dependent protein kinase. Science 1992,
256:1199-1202.
9. Chafouleas JG, Lagace L, Bolton WE, Boyd AE 3rd, Means AR:
Changes in CaM and its mRNA accompany reentry of quiescent
(G0) cells into the cell cycle. Cell 1984, 36:73-81.
10. Liu YC, Storm DR: Regulation of free CaM levels by neuromodulin:
neuron growth and regeneration. Trends Pharmacol Sci 1990,
11:107-111.
11. Skene JH: GAP-43 as a ‘CaM sponge’ and some implications for
calcium signalling in axon terminals. Neurosci Res Suppl 1990,
13:S112-S125
12. Liao B, Paschal BM, Luby-Phelps K: Mechanism of Ca2+-dependent
nuclear accumulation of CaM. Proc Natl Acad Sci USA 1999,
96:6217-6222.
13. Bachs O, Agell N, Carafoli E: Calmodulin and CaM-binding proteins
in the nucleus. Cell Calcium 1994, 16:289-296.
14. Deisseroth K, Heist EK, Tsien RW: Translocation of CaM to the
nucleus supports CREB phosphorylation in hippocampal neurons.
Nature 1998, 392:198-202.
15. Craske M, Takeo T, Gerasimenko O, Vaillant C, Torok K, Petersen O,
Tepiken AV: Hormone-induced secretory and nuclear translocation
of CaM: oscillations of CaM concentration with the nucleus as an
integrator. Proc Natl Acad Sci USA 1999, 96:4426-4431.
16. Pruschy M, Ju Y, Spitz L, Carafoli E, Goldfarb DS: Facilitated nuclear
transport of CaM in tissue culture cells. J Cell Biol 1994,
127:1527-1536.
17. Lee MA, Dunn RC, Clapham DE, Stehno-Bittel L: Calcium regulation
of nuclear pore permeability. Cell Calcium 1998, 23:91-101.
18. Hardingham GE, Cruzalegui FH, Chawla S, Bading H: Mechanisms
controlling gene expression by nuclear calcium signals. Cell
Calcium 1998, 23:131-134.
19. Meyer T, Oancea E, Allbritton NL: Nuclear calcium signals. Ciba
Found Symp 1995, 188:252-266.
20. Lipp P, Thomas D, Berridge MJ, Bootman MD: Nuclear calcium
signalling by individual cytoplasmic calcium puffs. EMBO J 1997,
16:7166-7173.
21. Nakazawa H, Murphy TH: Activation of nuclear calcium dynamics
by synaptic stimulation in cultured cortical neurons. J Neurochem
1999, 73:1075-1083.
22. Parkinson N, Bolsover S, Mason W: Nuclear and cytosolic calcium
changes in osteoclasts stimulated with ATP and integrin-binding
peptide. Cell Calcium 1998, 24:213-221.
23. Zimprich F, Torok K, Bolsover SR: Nuclear CaM responds rapidly to
calcium influx at the plasmalemma. Cell Calcium 1995, 17:233-238.
24. Albritton NL, Oancea E, Meyer T: Source of nuclear calcium signals.
Proc Natl Acad Sci USA 1994, 91:12458-12462.
25. Romoser VA, Hinkle PM, Persechini A: Detection in living cells of
Ca2+-dependent changes in the fluorescence emission of an
indicator composed of two green fluorescent protein variants
linked by a CaM-binding sequence. A new class of fluorescent
indicators. J Biol Chem 1997, 272:13270-13274.
bb10b31.qxd
02/08/2000
94
10:43
Page 94
Current Biology Vol 10 No 2
26. Persechini A, Cronk B: The relationship between the free
concentrations of Ca2+ and Ca2+-CaM in intact cells. J Biol Chem
1999, 274:6827-6830.
27. Osipchuk Y, Cahalan M: Cell-to-cell spread of calcium signals
mediated by ATP receptors in mast cells. Nature 1992,
359:241-244.
28. Millard PJ, Ryan TA, Webb WW, Fewtrell CJ: Immunoglobulin E
receptor cross-linking induces oscillations in intracellular free
ionized calcium in individual tumor mast cells. J Biol Chem 1989,
264:19730-19739.
29. Oancea E, Meyer T: Reversible desensitization of inositol
trisphosphate induced calcium release provides a mechanism for
repetitive calcium spikes. J Biol Chem 1996, 271:17253-17260.
30. Luby-Phelps K, Lanni F, Taylor DL: Behavior of a fluorescent analogue
of CaM in living 3T3 cells. J Cell Biol 1985, 101:1245-1256.
31. Subramanian K, Meyer T: Calcium-induced changes in the
structure of endoplasmic reticulum and nuclear envelope calcium
stores. Cell 1997, 89:963-971.
32. Timmerman LA, Clipstone NA, Ho SN, Northrop JP, Crabtree GR:
Rapid shuttling of NF-AT in discrimination of Ca2+ signals and
immunosuppression. Nature 1996, 383:837-840.
33. Zhang FL, Casey PJ: Protein prenylation: molecular mechanisms and
functional consequences. Annu Rev Biochem 1996, 65:241-269.
34. Dolmetsch RE, Xu K, Lewis RS: Calcium oscillations increase the
efficiency and specificity of gene expression. Nature 1998,
392:933-936.
35. Li W, Llopis J, Whitney M, Zlokarnik G, Tsien RY: Cell-permeant
caged InsP3 ester shows that Ca2+ spike frequency can optimize
gene expression. Nature 1998, 392:936-941.
36. Bito H, Deisseroth K, Tsien RW: CREB phosphorylation and
dephosphorylation: a Ca2+- and stimulus duration-dependent
switch for hippocampal gene expression. Cell 1996 87:1203-1214.
37. Westphal RS, Anderson KA, Means AR, Wadzinski BE: A signaling
complex of Ca2+ CaM-dependent protein kinase IV and protein
phosphatase 2A. Science 1998, 280:1258-1261.
38. Heist EK, Srinivasan M, Schulman H: Phosphorylation at the nuclear
localization signal of Ca2+/CaM-dependent protein kinase II
blocks its nuclear targeting. J Biol Chem 1998, 273:19763-19771.
39. Teruel M, Meyer T: Electroporation induced formation of calcium
entry sites in cell body and processes of adherent cells.
Biophys J 1997, 73:1785-1796.
Because Current Biology operates a ‘Continuous Publication
System’ for Research Papers, this paper has been published
on the internet before being printed. The paper can be
accessed from http://biomednet.com/cbiology/cub — for
further information, see the explanation on the contents page.