TALKING POINT
TIBS 25 – MAY 2000
Regulation of mitochondrial
metabolism by ER Ca21
release: an intimate connection
Guy A. Rutter and Rosario Rizzuto
New live-cell imaging techniques indicate that mitochondria exist in the
living cell as a continuous interconnected mitochondrial reticulum, or
‘MR’, closely associated with the endoplasmic reticulum (ER). Ca21 ions
released from the ER in response to hormonal stimulation might thus be
preferentially transferred into the mitochondrial matrix causing the local
activation of ATP synthesis. Ca21 uptake into the MR might also subtly
modify the activity of ER Ca21 release channels and thus the dynamics of
cytosolic Ca21 oscillations and waves.
EMERGING CELL IMAGING techniques,
combined with the targeting of recombinant probes, are revealing complex and
dynamic interactions between organelles
that were previously considered functionally distinct. In particular, it now appears that mitochondria might exist as a
continuous reticulum (MR). Close contacts between this and the endoplasmic
reticulum (ER) appear likely to allow
high local concentrations of Ca21 to be
generated between the two reticulae
without deleterious effects. Further, the
MR might then influence the cytosolic
Ca21 concentration ([Ca21]c), through
the local regulation of ER inositol
trisphosphate (IP3) receptors. Uptake of
Ca21 by the ‘target’ mitochondrion probably activates oxidative metabolism, in
turn allowing the generation of high concentrations of ATP (‘microdomains’),
just where (and when) they are required. An intriguing possibility is that
this activation of ATP synthesis, which
outlives the mitochondrial Ca21 increase, might result from changes in the
MR–ER structure.
The ‘mitochondrial reticulum’ and its
association with the endoplasmic reticulum
The classical ‘text book’ view of mitochondria is of discrete spheroidal or
sausage-like entities, as observed in electron micrographs of fixed cells and mitoG.A. Rutter is at the Dept of Biochemistry,
School of Medical Sciences, University of Bristol,
Bristol, UK BS8 1TD; and R. Rizzuto is at the
Dept of Experimental and Diagnostic Medicine,
Section of General Pathology, University of
Ferrara, Via Borsari 46, 44100 Ferrara, Italy.
Email: [email protected]
chondrial suspensions. However, early
three-dimensional (3D) reconstitution
studies of electron micrographs1, transmission electron tomography (TEM)2
(see Table 1), as well as the use of mitochondria-seeking fluorescent probes3,
and now targeted green-fluorescent protein (GFP)4 in living cells, suggest that
this is an inaccurate picture. GFP, from
the jellyfish Aequoria victoria, was targeted to mitochondria by exploiting the
cell’s own high-fidelity systems for sorting proteins. Namely, a cleavable mitochondrial presequence, derived from
subunit VIII of cytochrome c oxidase,
was added to the N terminus of the GFP,
thus allowing, after recombinant expression of the chimeric polypeptide, the import of the latter into the mitochondrial
matrix. By combining the expression of
mitochondrially targeted GFP and rapid
deconvolution microscopy (Table 1), a
picture now emerges in which for many
(if not most) cell types, ‘mitochondria’
exist as a continuous, interlinked ‘mitochondrial reticulum’ (MR) that weaves
its way through the cell (Fig. 1). This
technique avoids artefacts due to the
considerable movement of the mitochondria (and other organelles), by
allowing extremely rapid acquisition of
optical sections (60/s)4 – faster than can
be achieved using other imaging techniques (e.g. confocal microscopy, see
Table 1 and Fig. 1).
It should be stressed that the existence of mitochondria as an interlinked
reticulum does not exclude the possibility of functional compartmentalization,
as suggested by the presence of
‘hotspots’ of high membrane potential
0968 – 0004/00/$ – See front matter © 2000, Elsevier Science Ltd. All rights reserved.
(DC), and the obvious ‘flickering’ of DC
within apparently single mitochondria,
imaged
with
membrane-potentialsensitive probes (Table 1)5. Such observations have interesting implications for
the chemi-osmotic theory, as they suggest the existence of localized proton
circuits. Neither do the new structural
data exclude the possibility that the
reticulum can undergo transient collapse (e.g. the ‘pinching off’ of a single
‘mitochondrion’) or a more radical
collapse during cell division.
How does this mitochondrial reticulum interact with other organelles and
reticulae within the living cell? This
question has required simultaneous
analysis of the structures of both
mitochondria and the ER, and has revealed, unexpectedly, that the MR makes
tight, synaptic-like contacts with the ER.
After expressing GFP targeted to the ER,
and a spectrally shifted, blue-fluorescent
protein (BFP) in the mitochondrial matrix, Rizzuto et al.4 used digital imaging
microscopy to investigate the spatial relationship between these two organelles in
three dimensions. This approach demonstrated that the ER and MR were in close
contact (at the limit of resolution, estimated to be ,60 nm) over ~5% of the surface of each reticulum (Fig. 1, shown in
white).
Mitochondrial [Ca21] changes
The view of mitochondrial calcium,
and its role in cell biology, has evolved
dramatically over the last quarter of the
20th century. The early view, dating
from work on isolated mitochondria in
the 1960s6, suggested that mitochondria
probably acted in the cell as large,
possibly mobilizable, reservoirs of Ca21,
capable of buffering dramatic increases
in [Ca21]c7. However, four key observations were soon to alter this view.
First, in cells or tissues that had not
been exposed to stimuli known to elevate [Ca21]c, mitochondria contained
minute total concentrations of elemental calcium, detected either after isolation of the organelles8, or in situ, using
electron probe X-ray microanalysis (see
Table 1)9. Second, hormones and neurotransmitters, such as adrenaline and
histamine, that can signal into the cell
by raising [Ca21]c from ~100 nM to .1
mM, appeared to do so by mobilizing Ca21
ions from a store associated with the
ER10. Thus, located on this organelle, as
well as the Golgi apparatus11, receptors
for D-myo-inositol 1,4,5-trisphosphate
(IP3) were found. This inositol phospholipid-derived ‘second messenger’ is
PII: S0968-0004(00)01585-1
215
TALKING POINT
TIBS 25 – MAY 2000
Table 1. Specialized microscopy techniques to investigate the role of mitochondria in cellular calcium homeostasis
Microscopy technique
Sample
Signal (probe)
Feature studied
Essential hard- and soft-ware
Digital deconvolution. Uses digital analysis Single living cells Fluorescence
to remove out-of-focus information from
(mitochondria or ERstacks of fluorescence images42
targeted GFPs, etc.)
3D structure of
organelles in vivo4
Inverted fluorescence microscope with z
(vertical drive) control; charge-coupled device
(CCD) imaging camera. Advantage: rapid data
acquisition. Disadvantage: requires knowledge
of spatial blurring and computationally
intensive, non-linear iterative algorithms
(‘deconvolution’)
Transmission electron tomography (TEM); Fixed and stained Electron beam
enhanced (3D) electron microscopy
sections
absorption
(endogenous lipids
and other cellular
components)
3D structure of
organelles in fixed
samples2,44
Electron microscope; reconstitution algorithms
Confocal scanning laser. Obtains
fluorescence images from a single
focal plane
Single living cells Fluorescence
Laser source and moveable scan head.
[Rhod-2;
[Ca21]m18,24;
Advantage: off-line deconvolution not
tetramethylrhodamine mitochondrial
generally required. Disadvantage: slow scan
(TMRE);
membrane potential25,29; speeds preclude generation of 3D images of
autofluorescence]
NAD(P)H24
mitochondria using GFP
Photon-counting imaging. Detects ultralow-level luminescence from expressed
photoproteins
Single living cells Bioluminescence
(aequorin;
firefly luciferase)
Electron probe X-ray microanalysis43
Ultra-thin
(,200 nm)
cryo-sections
generated when receptors for Ca21mobilizing hormones are occupied at
the cell surface and cause the activation
of phospholipase C (Ref. 10). Binding of
IP3 to its receptor, a ligand-gated ion
channel, leads to a massive flux of Ca21
from the ER and Golgi down a large concentration gradient into the cytosol.
Third, three intramitochondrial citrate
cycle dehydrogenases (pyruvate, 2-oxoglutarate and the NAD1-dependent
1 µm
1 µm
Ti BS
216
X-ray emission
(endogenous metals)
[Ca21]m18;
[ATP]m29
Total elemental
metal content43
isocitrate dehydrogenases) were shown
to be allosterically stimulated by Ca21
ions, and regulated by agonist-induced
increases in cytosolic, and thence
mitochondrial
Ca21
concentration
21
12,13
([Ca ]m) . Activation of these
enzymes by Ca21 could thus stimulate
mitochondrial ATP synthesis at the
same time as there is an increased need
to supply ATP-requiring processes in the
cytosol (secretion, motility, protein synthesis, etc.). Fourth, isolated mitochondria displayed very limited efficiency as
Ca21 buffers at physiological Ca21 concentrations12 and occupy a relatively
small proportion of the intracellular volume (~10% of total). As a result of these
observations, the opinion that mitochondria had any significant impact on
Figure 1
Close
association
between
the
mitochondrial reticulum [MR, red; visualized by excitation of mitochondrial bluefluorescent protein (BFP) with 360 nm
light] and endoplasmic reticulum [ER,
green; excitation of ER-targeted greenfluorescent protein (GFP), excitation at
488 nm] in HeLa cells. Optical sections
were rapidly captured and each image
was enhanced using digital deconvolution
techniques (see Table 1), as described
in Refs 4 and 42, to remove out-of-focus
information. This allowed restoration of
the 3D image with software developed at
the University of Massachusetts Medical
Center, Worcester, Massachusetts, USA.
Note the close association of the MR and
ER (in the areas shaded white) over part
of each reticulum. Modified from Ref. 4.
Inverted microscope; CCD camera. Intensifier
uses multiple microchannel plate arrays to
amplify light output .106 fold. Limitation:
relatively low temporal and spatial resolution
Electron microscope. Analyses the distribution
of elements between organelles with high
spatial resolution
cytosolic Ca21 concentration under
normal physiological circumstances
was essentially silenced by the late
1980s.
However, the view that mitochondria
have no role to play in the regulation of
[Ca21]c in healthy cells has gradually
been modified over the last ten years.
As a first hint, NAD1-dependent isocitrate
dehydrogenase was found to respond
under physiological conditions to intramitochondrial [Ca21] values between 10 and
100 mM (Ref. 14), one to two orders of
magnitude higher than peak [Ca21]c values. However, the major breakthrough
came in 1992 with the ability to measure
free [Ca21]m dynamically in living cells
using aequorin, a Ca21-sensitive photoprotein, as a Ca21 probe targeted specifically to the mitochondrial matrix15.
Aequorin (like GFP from the jellyfish A.
victoria) is a tetrameric photoprotein
with three Ca21-binding sites, which oxidizes a cell-permeant cofactor, called
coelenterazine, but only at permissive
Ca21 concentrations. As with GFP, the
sorting of aequorin into mitochondria is
highly efficient, and nearly 100% of the
photoprotein is localized to this compartment. Thus, targeted aequorin is far
more selective than all other mitochondrial Ca21 probes, including positively
charged fluorescent dyes, such as Rhod-2
(Table 1)18,24. Photon production by
aequorin proceeds via a stable
enzyme–coelenteramide intermediate
that decays rapidly, producing a photon
of light, when [Ca21] rises16. Using this
TALKING POINT
TIBS 25 – MAY 2000
observed for an equivalent
(a)
global increase in extramitoHistamine
21
chondrial [Ca ] in the cu10
vette, or after cell permeabi8
lization15,20. Physiologically,
FCCP
this could be important to
6
Control
ensure the rapid decoding of
21
4
[Ca ]c increases, and also to
permit the initiation of oscil2
lations in [Ca21]m. What
mechanism might explain
0
1 min
this rapid mitochondrial
(b)
Ca21 uptake? One possibility
ATP
is the existence of a diffusible
cytosolic factor that stimu4
0s
lates the Ca21 uniporter in
living cells. However, an in5 µm
triguing alternative explanation, with implications be8
12
16
yond the control of organelle
signalling, is that the data in
fact report the generation of
high local Ca21 concenTi BS
trations (Fig. 3) at the points
of close contact between the
Figure 2
ER and MR (Fig. 1). Thus, loTime courses of [Ca21] changes in the mitochondrial
calized release of Ca21 ions
matrix ([Ca21]m), detected with targeted aequorin. (a)
(a ‘Ca21 puff’) from one or a
[Ca21]m was followed in populations of HeLa cells,
transfected with cDNA encoding mitochondrial aefew Ca21 release channels
quorin, using a photomultiplier and perifusion sys(IP3 receptors on the ER or
tem16,45. Histamine, a hormone involved in inflamma21
plasma membrane Ca
tory responses, induces a large increase in cytosolic
channels) might be ‘targeted’
[Ca21] in this cell type, and thus an increase in
onto a nearby mitochon[Ca21]m. Note the near-complete elimination of the
drion or mitochondrial uni[Ca21]m increase after incubation with the mitochondrial
uncoupler p-trifluoromethoxyphenyl hydrazone
porter (Fig. 3). Indeed,
(FCCP), which collapses the proton gradient across
Hajnoczky and colleagues21
the inner mitochondrial membrane and thus the drivsuggest that IP3 receptor
ing force for mitochondrial Ca21 uptake. (b) A similar
channel opening exposes the
experiment
performed in a single transfected Chinese
mitochondrial Ca21 unihamster
ovary
(CHO) cell, stimulated with ATP (which
porter to 20-fold higher Ca21
increases in [Ca21]c by binding to purinergic receptors
concentrations than those in
at the surface of this cell type), followed using an
intensified charge-coupled device camera. The first
the bulk cytosol, and have
image (top left) shows the stimulated cell under white
compared this situation with
light, and subsequent images show the rate of photon
localized release of neuroproduction, measured in darkness before and after
transmitters at the synaptic
addition of ATP, as indicated. Pseudocolour indicates
cleft. Complementing these
luminescence intensity, which is proportional to
data, targeting of aequorin to
[Ca21]m: black (,1 mM) → dark blue → yellow (.10
the outer surface of the inner
mM) → red → white (.15 mM) over different regions
of the cell surface. Luminescence images were
mitochondrial
membrane
recorded during 3 s integration periods, each starting
(by fusion with mitochonat the time (s) after the addition of stimulus, as
drial glycerol phosphate deshown. Note that the technique does not permit indihydrogenase) allowed detecvidual tubules of the mitochondrial reticulum to be
21
tion of [Ca ] changes at this
spatially resolved, but does indicate subcellular
site that were, as predicted,
heterogeneities in the rate of [Ca21]m increase and
decrease. These could reflect local differences in
approximately 5–20% higher
MR–ER overlap. Modified from Ref. 18.
than those in the bulk cy4
tosol . One interpretation of
the latter result is that it reflects a [Ca21]m several-fold higher than the mean
several-fold higher local [Ca21]c in the [Ca21]c increase. However, it should be
small portion of the mitochondrial reticu- mentioned that, even in suspensions of
lum surface (<5%), which is in close isolated mitochondria, Ca21 uptake
contact with the ER (Figs 1,3). Diffusion displays markedly sigmoidal kinetics with
of Ca21 within the mitochondrial matrix respect to extramitochondrial [Ca21],
could then generate an increase in average possibly due, in part, to the activity of a
[Ca2+]m (µM)
novel probe, [Ca21]m measurements
were carried out in numerous cell types
differing in embryological origin and calcium signalling patterns, demonstrating
that large increases in [Ca21]m always
parallel the Ca21 signals evoked in the
cytoplasm by extracellular agonists. In
particular, this technique demonstrated
that Ca21-mobilizing agonists can increase [Ca21]m from values close to
those in the cytosol of resting cells
(~100 nM) to values well in excess of the
cytosolic levels (.10 mM, compared
with 1–2 mM in the cytosol) (Fig. 2a).
Furthermore, this approach revealed a
striking discrepancy between the slow
kinetic properties of mitochondrial Ca21
uptake systems studied in vitro17 and
the fast dynamics observed in intact
cells (see below).
Because less than one photon is produced with the breakdown of each preformed aequorin–colenteramide complex, this limits the total number of
photons that can be produced to the
total number of aequorin molecules present in the cell. Thus, although luminescence signals can readily be obtained
from large populations (.10 000) of cells
using a relatively simple photomultiplier tube15,16, imaging aequorin bioluminescence at the single-cell level (as can
readily be done with fluorescent
probes) provides a more formidable
technical challenge. Nevertheless, photon-counting imaging systems (see
Table 1) have now been used to quantitate the feeble light production by aequorin (and other photoproteins; see
below), and to demonstrate subcellular
heterogeneities in agonist-induced
[Ca21]m changes18 (Fig. 2b).
But how do Ca21 ions cross the mitochondrial inner membrane? Because of
the lack of molecular information, most
of what is known still stems from classical biochemistry. Isolated mitochondria
possess systems for both Ca21 uptake
(a proton motive force (DC)-dependent
uniporter) and egress (principally
Na1/Ca21 exchange)17, which are operative at ‘normal’ [Ca21]c levels (0.1–2
mM). At higher Ca21 concentrations, and
in the absence of adenine nucleotides,
loading of mitochondria with large
amounts of Ca21 also appears to lead to
the opening of a non-specific pore (the
‘mitochondrial permeability transition’,
MPT), which is blocked by the immunosuppressant cyclosporin19.
A surprising finding of the studies
with mitochondrial aequorin was that increases in [Ca21]m appear to occur much
more rapidly in living cells than is
217
TALKING POINT
TIBS 25 – MAY 2000
ER
IP3
MR
2
1
High [Ca2+]
microdomain
Lower [Ca2+]
3
4
Ti BS
Figure 3
Generation of microdomains of high local [Ca21] (blue), close to mitochondria, during mobilization
from the ER. (1) Shown schematically is a region of the ER (vertical tube) in close contact with
the MR. (2) Binding of a Ca21-mobilizing agonist, such as histamine, to a cell surface receptor results in IP3 production by phospholipase C and the subsequent binding of IP3 to intracellular receptors on the ER membrane. (3) Release of Ca21 through the IP3-gated channels will lead to a high
local [Ca21] in their immediate vicinity. Where these regions of the ER are located in the vicinity of
the MR, as shown here, this will encourage rapid uptake of Ca21 into the mitochondrial matrix. (4)
Diffusion and binding of Ca21 in the cytosol will gradually lead to the formation of a larger volume
of lower [Ca21] and thus a decrease in the rate of mitochondrial Ca21 uptake (see Fig. 2a).
‘rapid’ mitochondrial (RaM) Ca21 uptake
pathway22. This appears to be a second,
separate pathway from that catalysed by
the Ca21 uniporter (see above), which is
transiently activated by high [Ca21].
Ca21 uptake via both pathways is likely
to cause progressive saturation of the
mitochondrial Ca21 efflux pathway, and
thus a dramatic further increase in
[Ca21]m as the Ca21 influx rate exceeds
efflux17. Thus, a local microdomain of
only slightly elevated [Ca21]c might produce a dramatically greater rate of
[Ca21]m increase than predicted from the
bulk [Ca21]c. The relative contributions
of each of these factors (i.e. ER–MR overlap and RaM) might well be cell-typedependent, with the former more prevalent in certain cell types (e.g. RBL-2H3
mast cells)21. An important future goal is
therefore to investigate the relative importance for rapid in vivo mitochondrial
Ca21 uptake of ER–MR contacts, and the
intrinsic properties of the mitochondrial
Ca21 transport systems. This might be
achievable by disrupting the exquisite
interlacing of the mitochondrial and
endoplasmic reticulae, for example by
interfering with critical regulatory
proteins involved in their
structural integrity23.
Figure 4
Relationship between [Ca21]m and [ATP]m.
(a) Buffering of [Ca21]m (red trace)
[Ca2+]m
increases during different periods of incubation with the Ca21 chelator BAPTAAM (10 mM; 0 s, 90 s and 300 s, from
left to right), reduces mitochondrial free
[ATP] increase (blue trace) when HeLa
cells are treated with histamine (arrows).
[ATP]m
Vertical bar represents a 1 mM increase
in [Ca21]m or a 20% increase in ATPdependent luciferase luminescence.
Modified from Ref. 30. (b) An increase in
(b)
(ii)
extracellular glucose concentration (from
(i)
(ii)
3 mM (i) to 30 mM (ii), as indicated)
(i)
leads to distinct patterns of [ATP] change
Cyt
as monitored by targeted luciferase bioluminescence (pseudocolour: blue, low intensity; red, high intensity) in the cytosol
(Cyt), sub-plasma membrane domain
Pm
(Pm) and within the mitochondrial matrix
(Mit) of MIN6 b-cells. These [ATP]
increases involve activation of mitochondrial metabolism, both due to the
Mit
increased supply of a metabolizable fuel
and an increase in [Ca21]m (see Ref. 29).
Luminescence from single b-cells was
Ti BS
recorded over the cell surface at video
rate (40 Hz) and luminescence time
courses (left-hand panel) were generated by integrating the light output from the single cell shown over
1 s (broken lines) or 10 s intervals (solid lines) using time-resolved imaging software provided by the
camera manufacturer (Photek, St Leonard’s on Sea, UK). [ATP] in each compartment was estimated
by cell permeabilization and perifusion with medium of known [ATP]. Resting [ATP] in each compartment was ~1 mM. The vertical bar represents a 10% increase in luminescence and the horizontal bar,
120 s. Images (right-hand panel) were generated off-line from the time-course data by integration of
light production over 30 s immediately before the increase in glucose concentration (i), and at the peak
of the [ATP] increase in each compartment (ii). Scale bar, 5 mm. Modified from Ref. 29.
(a)
218
Mitochondrial ATP and cytosolic
ATP microdomains
Are there other consequences of the existence of
tight contacts between the
ER and MR, or between the
MR and the plasma membrane, for mitochondrial function? Mitochondrial oxidative
metabolism can be monitored
qualitatively in living cells
through changes in cellular
NAD(P)H
autofluorescence,
which largely reflects the activity of intramitochondrial Ca21sensitive dehydrogenases24.
Monitoring
of
NAD(P)H
changes suggests that increases in [Ca21]m do indeed
regulate mitochondrial dehydrogenases in vivo, albeit in a
complex fashion, and also
regulate
the
respiratory
chain25 (G.A.R. and E. Ainscow,
unpublished). Importantly, it
would appear that small and
transient depolarizations of
the mitochondrial membrane
potential (DC), which can
occur upon mitochondrial
Ca21 uptake26, are quickly
compensated by a dramatic
TALKING POINT
TIBS 25 – MAY 2000
stimulation of mitochondrial oxidative
metabolism.
Ultimately, the stimulation of mitochondrial metabolism is expected to increase intramitochondrial and, subsequently, cytosolic ATP concentration
([ATP]c). Unfortunately, measurements of
total ATP levels in cell homogenates give
only a very poor guide to free concentrations of the nucleotide, as a result of its
extensive compartmentalization (for example, in secretory vesicles), and binding
to proteins. A solution to this problem
was provided by the studies of Cobbold
and co-workers28, who demonstrated that
the photoprotein luciferase from the firefly
Photinus pyralis can act as an intracellular
ATP probe. Indeed, luciferase oxidizes a
cell-permeant cofactor, luciferin, following the formation of luciferyl–AMP intermediate. This step requires ATP (Ref. 27)
and can therefore be used to sense local
ATP concentration, given that the low
affinity for ATP of luciferase in a ‘cytosolic
milieu’ matches the physiological range
of intracellular ATP concentrations well28.
Upon addition of appropriate targeting
sequences, the photoprotein can be
specifically addressed to defined intracellular locations. Thus, a mitochondrial
ATP probe was obtained by fusing it to
the cytochrome c oxidase subunit VIII mitochondrial presequence29,30, as described above for GFP and aequorin.
With this probe, it was possible to show
that the agonist-dependent [Ca21]m
increases are responsible for the large
increases in intramitochondrial [ATP]
([ATP]m); when the former are buffered,
the latter are virtually abolished (Fig. 4a).
Changes in [ATP]c are believed to be
particularly important in triggering insulin secretion from islet b-cells (see
Ref. 29 and Refs therein). In derived MIN6
b-cells, recombinant targeted luciferases
and photon-counting microscopy29
(Table 1) revealed no difference in [ATP]c
and [ATP]m (~1 mM in each compartment) when cells were incubated at the
low (,5 mM) glucose concentrations
normally expected for the blood of a
fasted human. Conversely, when the extracellular glucose concentration was increased over the range experienced after
a carbohydrate-containing meal (3–16
mM), which activates insulin secretion,
[ATP]m increases were longer-lasting
than the global [ATP]c changes. However,
[ATP] changes immediately beneath the
plasma membrane (detected with luciferase linked to a protein of the exocytotic machinery called SNAP25)31
matched closely those of [ATP]m
(Fig. 4b). This might suggest that in these
iii
(a)
ii
i
[ATP]m
(b)
[Ca2+]m
Ti BS
Figure 5
Long-lived changes in [ATP]m can be provoked by Ca21-mobilizing hormones. (a)
Populations of HeLa cells expressing
cDNA for mitochondrially targeted luciferase were perifused in the presence
of luciferin during continuous monitoring
of light output with a simple photomultiplier tube: (i) in medium containing 5.5
mM glucose only, (ii) in the same medium
plus histamine (100 mM), (iii) in medium
without glucose or histamine but with 10
mM pyruvate and 1 mM lactate (red
trace). Cells were not exposed to histamine, but were maintained in glucoseonly medium for the same period before
transfer into pyruvate and lactate (iii)
(blue trace). The change from glucose- to
pyruvate- and lactate-containing medium
(period iii) results in a large increase in
mitochondrial oxidative metabolism, and
thus [ATP]m, because in these glycolytic
cells, glucose alone furnishes very little
intracellular pyruvate for oxidation by mitochondria. (b) Changes in [Ca21]m were
monitored in similar cell populations
transfected with cDNA encoding mitochondrial aequorin and reconstituted with
coelenterazine. The vertical bar represents a 10% increase in luciferase luminescence (a) or 1 mM Ca21 (b).
Horizontal bar, 120 s. Modified from
Ref. 30.
cells, Ca21 influx, and the generation of a
gradient of Ca21 beneath the plasma
membrane, preferentially activates peripherally located mitochondria, which
might also form ‘synaptic-like’ contacts
with the plasma membrane, as observed
for the ER. Similarly, evidence for the
existence of confined domains of [ATP],
close to ER Ca21 ATP pumps, has come
from studies of Ca21 homeostasis in
permeabilized cells32.
A mitochondrial ‘memory’?
Could changes in the structure of the
ER–MR interface, perhaps following
cell stimulation, occur and play a role
in cell physiology? At present, direct
measurements of these changes (by digital deconvolution imaging of targeted
GFPs) are not available. However, indirect evidence supports the idea that
they might be important. Thus, an
intriguing result that emerged from
studies using both DC-sensitive
fluorochromes, and now recombinant
targeted luciferase, is that increases in
intramitochondrial [Ca21] might produce
rather long-lived increases in mitochondrial metabolism. In particular, when
cells are rendered strongly glycolytic,
and then incubated in medium lacking
glucose, pre-challenge with [Ca21]c-raising
agonists leads to larger increases in
[ATP]m when a mitochondrial fuel (pyruvate plus lactate) is re-supplied. This
phenomenon is still apparent many minutes (up to 1 h) after the removal of the
agonist and the return of [Ca21]m to
prestimulatory level30 (Fig. 5). This feature, which appears to represent a novel
form of cellular plasticity, might also be
especially important in cells designed to
sense changes in nutrient concentration
(e.g. pancreatic islet b- and a-cells, and
certain glucose-sensitive hypothalamic
neurones). Indeed, activation of Ca21
influx sensitizes islet b-cells to subsequent increases in glucose concentration,
with little perceptible effect on glucoseinduced Ca21 increases (G.A.R. and E.
Ainscow, unpublished). What could be
the molecular bases of these phenomena? Whilst stable increases in the activity of pyruvate dehydrogenase (which is
regulated by reversible phosphorylation) seemed likely, it now appears that
a more probable candidate is the respiratory chain. Indeed, recent studies by
Robb-Gaspers et al.25, as well as much
work on isolated mitochondria from
Halestrap and colleagues33, have
demonstrated that increases in [Ca21]m
can produce rather long-lived increases
in respiratory-chain activity. Quite how
this is achieved remains a mystery, although changes in mitochondrial structure, and its relationship to the ER,
might well be involved.
Mitochondria and the control of cytosolic
Ca21 concentration
Although the above data demonstrate
a direct role of agonist-mobilizable Ca21
in modulating mitochondrial function,
the other side of the coin – that is,
whether mitochondrial Ca21 uptake has
an effect on cytosolic Ca21 homeostasis
– is still debated. As discussed above,
recent data have now changed the view
that the net contribution of mitochondrial to intracellular calcium handling is
219
TALKING POINT
marginal. Indeed, data from several
sources21,34,35 have demonstrated that
mitochondrial Ca21 uptake actively participates in shaping the spatio-temporal
complexity of calcium signalling in various ways. Firstly, mitochondrial Ca21
uptake in the microenvironment at the
mouth of calcium channels influences
[Ca21]c changes36, by subtly modifying
the open-state probability of IP3-gated
release channels37, or store-operated
channels in the plasma membrane38.
Secondly, spatially restricted mitochondrial clusters might, in polarized cells
such as pancreatic acinar cells, act as
local buffers. Thus, in these cells, there
appears to exist a high density of mitochondria or mitochondrial reticulum at
the interface between the apical (secretory) and the basolateral poles of the
cell. Uptake of Ca21 ions by this mitochondrial density, at the front of an advancing Ca21 wave, might thus form a
cordon sanitaire, preventing the further
spread of the wave into the nonsecretory pole39. Finally, it has also been
proposed that mitochondria could represent an ‘excitable’ medium, capable of
conducting Ca21 waves via transient activation and inactivation of the mitochondrial permeability transition pore
(MPT)40. Such a mechanism can be
imagined as analogous to the propagation of nerve impulses along an axon.
The above data strongly suggest that
the mitochondrial reticulum, through
mechanisms that might differ depending
on the physiology of the various cell
types, can control the pattern and velocity of intracellular Ca21 waves34. An intriguing consequence of this observation is that, in sensitive cells,
nutrients might be able to alter the mitochondrial Ca21 uptake capacity, by
hyperpolarizing the mitochondrial inner
membrane, thereby altering the local
behavior of [Ca21]c. For example, Ca21dependent processes such as secretion
or gene expression could be altered if
the shape of the Ca21 increases close
to secretory vesicles or the nucleus,
respectively, were to be affected.
Conclusions and future prospects
The use of selective probes has allowed a drastic expansion of our knowledge of how mitochondria participate in
local signalling routes controlling, on
the one hand, mitochondrial function in
response to cell needs and, on the other
hand, the propagation and the net effect
of cytosolic Ca21 changes. We can envision, however, that the development of
similar targeted probes for other
220
TIBS 25 – MAY 2000
organelles and cell districts, perhaps expressed in primary cell types via new
viral vectors, will further refine this
scheme. A question that leads from the
above results is whether, in addition to
the ER, there are other ‘downstream’
organelles that might detect microdomains of released Ca21, or ATP, close
to the MR? It is now evident that secretory vesicles that are close to the
plasma membrane of the cell experience
higher Ca21 concentrations than the surrounding cytosol41. It appears likely that
this is due to the privileged association
between these vesicles and plasma
membrane Ca21 channels (or ER located
close to the plasma membrane).
Whether these peripheral vesicles also
experience higher local [ATP]c, as a result of hyperstimulation of ATP synthesis in the region of the MR beneath the
plasma membrane, will now need to be
examined carefully. Whether other organelles (lysosomes, peroxisomes, etc.)
could similarly be the target of localized
changes in mitochondrial ATP production warrants further investigation.
Acknowledgements
Work in the authors’ laboratories is
supported by grants from support by
the Wellcome Trust, the Medical
Research Council (UK), the Royal
Society, the British Diabetic Association,
Italian ‘Telethon (project no. 850), the
‘Biomed’ program of the European Union
and the Italian University Ministry.
References
1 Bakeeva, L.E. et al. (1978) Mitochondrial framework
(reticulum mitochondriale) in rat diaphragm muscle.
Biochim. Biophys. Acta 501, 349–369
2 Mannella, C.A. et al. (1998) Electron microscopic
tomography of rat-liver mitochondria and their
interaction with the endoplasmic reticulum. Biofactors
8, 225–228
3 Reers, M. et al. (1991) J-aggregate formation of a
carbocyanine as a quantitative fluorescent indicator of
membrane potential. Biochemistry 30 4480–4486
4 Rizzuto, R. et al. (1998) Close contacts with the
endoplasmic reticulum as determinants of
mitochondrial Ca21 responses. Science 280,
1763–1766
5 Duchen, M.R. et al. (1998) Transient mitochondrial
depolarizations reflect focal sarcoplasmic reticular
calcium release in single rat cardiomyocytes. J. Cell
Biol. 142, 975–988
6 Lehninger, A.L. (1967) Energy-linked ion movements in
mitochondrial systems. Adv. Enzymol. 29, 259–320
7 Nicholls, D.G. (1978) The regulation of
extramitochondrial free calcium ion concentration by
rat liver mitochondria. Biochem. J. 176, 463–474
8 McCormack, J.G. and Denton, R.M. (1984) Role of
Ca21 ions in the regulation of intramitochondrial
metabolism in rat heart. Evidence from studies with
isolated mitochondria that adrenaline activates the
pyruvate dehydrogenase and 2-oxoglutarate
dehydrogenase complexes by increasing the
intramitochondrial concentration of Ca21. Biochem. J.
218, 235–247
9 Somlyo, A.P. and Somlyo, A.V. (1986) Electron probe
analysis of calcium content and movements in
sarcoplasmic reticulum, endoplasmic reticulum,
mitochondria, and cytoplasm. J. Cardiovasc.
Pharmacol. 8, S42–S47
10 Berridge, M.J. (1993) Inositol trisphosphate and
calcium signalling. Nature 361, 315–325
11 Pinton, P. et al. (1998) The Golgi apparatus is an
inositol 1,4,5-trisphosphate-sensitive Ca21 store, with
functional properties distinct from those of the
endoplasmic reticulum. EMBO J. 17, 5298–5308
12 Denton, R.M. and McCormack, J.G. (1998) On the role
of the calcium transport cycle in the heart and other
mammalian mitochondria. FEBS Lett. 119, 1–8
13 Hansford, R.G. (1985) Relation between mitochondrial
calcium transport and control of energy metabolism.
Rev. Physiol. Biochem. Pharmacol. 102, 1–72
14 Rutter, G.A. and Denton, R.M. (1988) Regulation of
NAD1-linked isocitrate dehydrogenase and 2oxoglutarate dehydrogenase by Ca21-ions within
toluene-permeabilized rat heart mitochondria.
Biochem. J. 252, 181–189
15 Rizzuto, R. et al. (1992) Rapid changes of
mitochondrial Ca21 revealed by specifically targeted
recombinant aequorin. Nature 358, 325–327
16 Cobbold, P.H. and Rink, T.J. (1987) Fluorescence and
bioluminescence measurement of cytoplasmic free
calcium. Biochem. J. 248, 313–328
17 Crompton, M. (1985) The regulation of mitochondrial
calcium transport in heart. Curr. Top. Membr.
Transport 25, 231–286
18 Rutter, G.A. et al. (1996) Subcellular imaging of
intramitochondrial Ca21 with recombinant targeted
aequorin: significance for the regulation of pyruvate
dehydrogenase activity. Proc. Natl. Acad. Sci. U. S. A.
93, 5489–5494
19 Halestrap, A.P. (1991) Calcium-dependent opening of a
non-specific pore in the mitochondrial inner membrane
is inhibited at pH values below 7. Implications for the
protective effect of low pH against chemical and
hypoxic cell damage. Biochem. J. 278, 715–719
20 Rizzuto, R. et al. (1993) Microdomains with high Ca21
close to IP3-sensitive channels that are sensed by
neighbouring mitochondria. Science 262, 744–747
21 Csordas, G. et al. (1999) Quasi-synaptic calcium signal
transmission between endoplasmic reticulum and
mitochondria. EMBO J. 18, 96–108
22 Gunter, T.E. et al. (1998) The Ca21 transport
mechanisms of mitochondria and Ca21 uptake from
physiological-type Ca21 transients. Biochim. Biophys.
Acta 1366, 5–15
23 Yaffe, M.P. (1997) Mitochondrial morphogenesis:
fusion factor for fly fertility. Curr. Biol. 7, R782–R783
24 Hajnoczky, G. et al. (1995) Decoding of cytosolic
calcium oscillations in the mitochondria. Cell 82,
415–424
25 Robb-Gaspers, L.D. et al. (1998) Integrating cytosolic
calcium signals into mitochondrial metabolic
responses. EMBO J. 17, 4987–5000
26 Loew, L.M. et al. (1994) Physiological cytosolic Ca21
transients evoke concurrent mitochondrial
depolarizations. Proc. Natl. Acad. Sci. U. S. A. 91,
12579–12583
27 DeLuca, M. (1976) Firefly luciferase. Adv. Enzymol.
Relat. Areas Mol. Biol. 44, 37–68
28 Allue, I. et al. (1996) Evidence for rapid consumption of
millimolar concentrations of cytoplasmic ATP during
rigor-contracture of metabolically compromised single
cardiomyocytes. Biochem. J. 319, 463–469
29 Kennedy, H.J. et al. (1999) Glucose-induced ATP
microdomains in single islet beta-cells. J. Biol. Chem.
274, 13281–13291
30 Jouaville, L.S. et al. (1999) Regulation of mitochondrial
ATP synthesis by calcium: evidence for a long term,
Ca21 triggered, mitochondrial activation. Proc. Natl.
Acad. Sci. U. S. A. 96, 13807–13812
31 Sollner, T.H. and Rothman, J.E. (1996) Molecular
machinery mediating vesicle budding, docking and
fusion. Experientia 52, 1021–1025
32 Landolfi, B. et al. (1998) Ca21 homeostasis in the
agonist-sensitive internal store: Functional interactions
between mitochondria and the ER measured in situ in
intact cells. J. Cell Biol. 142, 1235–1243
33 Halestrap, A.P. (1982) The nature of the stimulation of
the respiratory chain of rat liver mitochondria by
glucagon pretreatment of animals. Biochem. J. 204,
37–47
34 Jouaville, L.S. et al. (1995) Synchronisation of calcium
waves by mitochondrial substrates in Xenopus laevis
oocytes. Nature 377, 438–441
35 Simpson, P.B. and Russell, J.T. (1996) Mitochondria
support inositol 1,4,5-trisphosphate-mediated Ca21
waves in cultured oligodendrocytes. J. Biol. Chem.
271, 33493–33501
36 Hajnoczky, G. et al. (1999) Mitochondria suppress local
feedback activation of inositol 1,4,5-trisphosphate
receptors by Ca21. J. Biol. Chem. 274, 14157–14162
37 Adkins, C.E. and Taylor, C.W. (1999) Lateral inhibition
OBITUARY
TIBS 25 – MAY 2000
of inositol 1,4,5-trisphosphate receptors by cytosolic
Ca21. Curr. Biol. 9, 1115–1118
38 Hoth, M. et al. (1997) Mitochondrial regulation of
store-operated calcium signaling in T lymphocytes. J.
Cell Biol. 137, 633–648
39 Tinel, H. et al. (1999) Active mitochondria surrounding
the pancreatic acinar granule region prevent spreading
of inositol trisphosphate-evoked local cytosolic Ca(21)
signals. EMBO J. 18, 4999–5008
40 Ichas, F. et al. (1997) Mitochondria are excitable
organelles capable of generating and conveying
electrical and calcium signals. Cell 89, 1145–1153
41 Emmanouilidou, E. et al. (1999) Imaging [Ca21]
changes at the secretory vesicle surface with a
recombinant targeted cameleon. Curr. Biol. 9, 915–918
42 Rizzuto, R. et al. (1998) Digital imaging microscopy of
living cells. Trends Cell Biol. 8, 288–292
43 Somlyo, A.V. et al. (1989) Electron probe X-ray
microanalysis of Ca21, Mg21, and other ions in rapidly
frozen cells. Methods Enzymol. 172, 203–229
44 Mannella, C.A. et al. (1997) Reconsidering
mitochondrial structure: new views of an old organelle.
Trends Biochem. Sci. 22, 37–38
45 Brini, M. et al. (1995) Transfected aequorin in the
measurement of cytosolic Ca21 concentration
([Ca21]). A critical evalulation. J. Biol. Chem. 270,
9896–9903
of the bacterial trp repressor. The Sigler
laboratory reported the structure of the
trp repressor complexed to DNA in
1988, a structure that was instrumental
in revising our view of the role of water
molecules at protein–nucleic acid and
Paul Sigler (Fig. 1), one of the most influ- Department of Molecular Biophysics protein–protein interfaces4. This strucential of contemporary structural biolo- and Biochemistry, Yale University, ture was one of the first to be detergists, died suddenly of a heart attack on where he was Henry Ford II Professor mined of a transcription factor recognizJanuary 11, 2000. Paul was one of those and an Investigator of the Howard ing a specific DNA target. Much to the
rare individuals who shaped the devel- Hughes Medical Institute.
surprise of all concerned, the structure
opment of structural biology not just by
Paul Sigler’s trajectory through struc- revealed that a sheet of water molecules
the example set by his own research ca- tural biology matches that of the field as a was interposed between the protein and
reer but also by the force of his person- whole, moving from the early and the DNA. Paul Sigler suggested that the
ality. His unexpected death leaves a void painstaking investigations of individual protein achieved specificity by ‘reading
that extends far beyond the boundaries protein structures to the present-day out’ the pattern of water molecules that
of his own research interests. None of unveiling of the architecture and mechan- are bound specifically to the DNA. This
us worked closely with Paul, who was ics of large macromolecular assemblies. hypothesis met with considerable resismore senior to us, but we were influ- Paul’s early work (with David Davies and tance because of the prevailing view
enced by him in different ways. By writ- David Blow) was mainly focused on crys- that water molecules were too diffuse in
ing this obituary we hope to communi- tallographic studies of chymotrypsin. structure to promote specificity. Paul
cate to some degree the manner in Along with Brian Matthews, David Blow was engaged, over several years, in a
which Paul Sigler engaged the hearts and Richard Henderson he published the concerted attempt to resolve this conand minds of the scientists who were structure of chymotrypsin and an analy- troversy. He ultimately accumulated a
fortunate enough to encounter him.
sis of its structural mechanism1,2. A paper considerable weight of evidence pointPaul Sigler was born in Richmond, on the iodination of tyrosine residues for ing towards the importance of water
Virginia, on February 19, 1934. After his the generation of crystallographic deriva- molecules in mediating specific interacundergraduate studies in chemistry at tives is still rewarding to read3.
tions between proteins and nucleic
Princeton University in 1955, Paul
Paul subsequently worked on the acids and this is now widely recogmoved to Columbia University in 1959, structures of tRNA and phospholipases. nized5. These water molecules are usuwhere he obtained a medical degree. He also developed an interest in tran- ally missing in structures of macroOne thing that Paul appears to have scription, working first on the structure molecular assemblies determined at low
been particularly proud of in his
resolution, and we think of Paul
medical training was his champiSigler when we envision the waonship status in delivering
ters that would be there but for
babies; he was pictured in a medical
the lack of resolution.
textbook demonstrating the proper
After moving to Yale University,
way to hold a newborn infant after
Paul became interested in cellular
delivery. Despite this innate skill in
signal transduction, and his labomidwifery, Paul felt the pull of
ratory made contributions that
structural biology, and he moved
are central to our current underto the National Institutes of Health
standing of how heterotrimeric G
(NIH) in 1961, where he worked on
proteins function6. Other work
chymotrypsin with David Davies.
from Yale in this area includes the
From 1964 to 1967 he was at the
structural analysis of hormoneMedical Research Council (MRC)
binding
nuclear
receptors.
Laboratory of Molecular Biology in
Perhaps most exciting among the
Cambridge, UK, and he obtained a
recent results from the Sigler labPhD in biochemistry from the
oratory has been the elucidation
University of Cambridge in 1967.
of structures of the molecular
He was appointed an Associate
chaperone GroEL–GroES7. These
Professor of Biophysics at the
marvelous engines that drive proUniversity of Chicago in 1967, and
tein folding are enormously intriFigure 1
was promoted to Professor in
cate in their inner workings, and
Paul Sigler, 1999. Photograph used, with permission,
1973. In 1989 he moved to the
the crystal structures of various
from Yale University.
Paul Sigler (1934–2000)
0968 – 0004/00/$ – See front matter © 2000, Elsevier Science Ltd. All rights reserved.
PII: S0968-004(00)01587-5
221