Calcium in Biological System

3
Calcium in Biological Systems

STURE FORSEN AND JOHAN KORDEL
Physical Chemistry 2,
Chemical Centre,
University of Lund
I. INTRODUCTION
Calcium, like many other "inorganic elements" in biological systems, has dur-
ing the last decade become the subject of much attention both by scientists and
by the general public.
1
The presence and central role of calcium in mammalian
bones and other mineralized tissues were recognized soon after its discovery as
an element by Davy in 1808. Much later, the insight arrived that Ca
2
+ ions
could play an important role in other tissues as well. Experiments of great his-
torical influence were performed by the British physiologist Sidney Ringer a
little over a century ago.
2
He was interested in the effects of various cations on
frog-heart muscle and somewhat serendipitously discovered that Ca 2+ ions, ev-
erpresent in the tap water distributed in central London, in millimolar concen-
trations were necessary for muscle contraction and tissue survival.
Today it is widely recognized that Ca2 + ions are central to a complex intra-
cellular messenger system that is mediating a wide range of biological pro-
cesses: muscle contraction, secretion, glycolysis and gluconeogenesis, ion
transport, cell division and growth (for definitions of terms in boldface, see
Appendix A in Section IX). The detailed organization of this messenger system
is presently the subject of considerable scientific activity, and some details are
already known. One of the links in the system is a class of highly homologous
Ca 2+ -binding proteins, to be discussed later on in this chapter, that undergo
Ca 2+ -dependent conformational changes and respond to transitory increases in
intracellular Ca 2+-ion concentrations. A prerequisite for the proper function of
the calcium messenger system in higher organisms is that the cytosolic Ca 2+
concentration in a "resting" cell be kept very low, on the order of 100 to
200 nM. Transitory increases in the Ca 2 + concentration that may result from
hormonal action on a membrane receptor must rapidly be reduced. Several
transport proteins, driven either by ATP hydrolysis or by gradients of some
other ion like Na +, are involved in this activity. 107
108 3 / CALCIUM IN BIOLOGICAL SYSTEMS
Ca 2+ ions are also known to play various roles outside cells. In the plant
kingdom Ca2+ ions often form links between individual cells and are required
for maintaining the rigidity of whole plants; some seaweeds are typical ex-
amples. In the blood plasma of mammals, in which the Ca 2+ concentration
exceeds the intracellular by a factor of about 10
4
, Ca2+ ions are instrumental
in joining certain proteins in the blood-clotting system with membrane surfaces
of circulating cells. Many extracellular enzymes also contain Ca 2+ ions, some-
times at the active site but most often at other locations. It is generally believed
that Ca 2+ ions confer on proteins an increased thermal stability, and indeed
proteins in heat-tolerant microorganisms often hold many such ions.
Vertebrates require much calcium in their food; in the USA the recom-
mended daily allowance (RDA) for adult humans is 800 mg, and most other
countries have comparable recommendations. During gestation in mammals,
calcium must be transported across the placenta into the fetus, in particular
during those phases of pregnancy when bone formation is most rapid. Interest-
ingly, there appear to be some parallels between intestinal and placental trans-
port that will be discussed further below. The role of calcium in biominerals is
a vast subject that we can treat only superficially in this chapter.
To provide a background to the more biologically oriented sections that
follow, we begin with a brief recapitulation of some basic facts about calcium.
Then we continue with an outline of calcium distribution in biological tissues
and organelles, and of the methods that can be used to obtain this information.
After this follows a brief section on Ca2+ transport, and an account of the
mechanism of intracellular Ca 2+ release as it is presently understood. A discus-
sion of some selected Ca
2
+-binding proteins of general interest, both intracel-
lular and extracellular, then follows. Before we conclude the chapter, we will
summarize some recent observations on Ca
2
+-binding proteins in prokaryotes.
II. BASIC FACTS ABOUT CALCIUM: ITS COMPOUNDS
AND REACTIONS
A. Basic Facts
Calcium was first recognized as an element in 1808 by Humphry Davy, and the
name was given after the Latin for lime: calx. Several isotopes of calcium are
known. The stable isotopes are, in order of decreasing natural abundance, 40Ca
(96.94%), 44Ca (2.1%), 42Ca (0.64%), and 43Ca (0.145%). 43Ca is the only
isotope with a nuclear spin (I = ~ ) different from zero, which makes it amen-
able to NMR studies. 45Ca is a radioactive isotope of some importance ({3
decay; 8.8 min half life).3 It has been used in studies of calcium localization
and transport in biological systems.
Calcium constitutes about 3 percent by weight of the Earth's crust, mostly
in the form of sedimentary rocks of biological origin dating back some three
II. BASIC FACTS ABOUT CALCIUM: ITS COMPOUNDS AND REACTIONS 109
Table 3.1
Ca
2
+ concentrations in fluids and tissues.6-9
Specimen
Sea water
Fresh water
Rain water
"Hard" tap water
"Good" beer
Adult human serum
Serum of other vertebrates
Nematote body fluids
Molluscan serum-marine
-fresh water
-land
Milk
Bone
Mitochondria from rat liver
Endoplasmatic reticulum
Cytoplasm of a resting mammalian cell
Cytoplasm of E. coli
Units are mM if not otherwise stated
10
0.02-2
0.002-0.02
1.5
4
2.45 0.05
1.5-5
6
9-15
1.5-7.8
3.3-12.3
70
0.8-1.0
0.8 0.1 mmol/kg
8-10 mmollkg
0.0001
0.0001
billion years. In sea water the total concentration of calcium ranges from 5 to
50 times higher than in fresh water, which, in tum, has a calcium concentration
ten times that of rain water (see Table 3.1). This explains the pleasant feeling
when ordinary soaps are used in rain water. The calcium concentration in ordi-
nary tap water varies with location; calcium is usually added to water in distrib-
uting networks in order to prevent corrosion of iron pipes. Tap water with a
calcium concentration above 1.5 mM is usually classified as "hard." Interest-
ingly, the taste of beer seems related to the calcium concentration, and it is
claimed that "good" beer should have a concentration higher than that of "hard"
tap water.
In the body fluids of higher organisms the total calcium concentration is
usually on the order of a few millimolar (see Table 3.1). In adult human serum,
the concentration is observed to be, within narrow limits, 2.45 mM.
B. Essentials of Ca
2
+ Chemistry
Since the Ca2 + ion accomplishes its biological tasks in an environment with 1
to 3 mM Mg 2 +, it is of particular interest to compare the properties of these
two ions in order to understand how a discrimination is made in biological
systems. In addition, the coordination chemistry of Ca 2+ is closely related to
that of Mg 2 + (as well as Cd2 +), though there are several obvious differences.
First of all, the ionic radius of a Ca
2
+ ion with a given coordination number
(CN) is always higher than that of an Mg
2
+ or Cd
2
+ ion with the same CN.
11 a 3 / CALCIUM IN BIOLOGICAL SYSTEMS
At CN = 6, the ionic radii of Ca
2
+, Cd
2
+, and Mg
2
+ are 1.00, 0.95, and
0.72 A, respectively, whereas at eN = 8 they are 1.12, 1.10, and 0.89 A,
respectively.
4
Ligand preferences of Ca 2+ depend on the fact that it is a hard metal ion.
Thus Ca 2+ strongly prefers oxygen ligands over nitrogen or sulfur ligands;
Ca 2+ .... N bonds are about 0.25-0.3 A longer than Ca 2+ 0 bonds.
5
,6,10
Large differences in coordination number and geometry have been observed for
Ca
2
+ complexes. In a study of 170 x-ray structures of Ca
2
+ complexes involv-
ing carboxylate groups, 11 binding was found to be either (i) unidentate, in which
the Ca
2
+ ion interacts with only one of the two carboxylate oxygens, (ii) biden-
tate, in which the Ca2+ ion is chelated by both carboxylate oxygens, or (iii)
mixed ("a-mode") in which the Ca 2+ ion is chelated by one of the carboxylate
oxygens and another ligand attached to the a-carbon (see Figure 3.1). The Ca2+-
oxygen distances span a range from 2.30 to 2.50 A, with the average distance
being 2.38 A in the unidentate and 2.53 A in the bidentate mode, respec-
tively.II Observed coordination numbers follow the order 8 > 7 > 6 > 9. By
contrast, Mg
2
+ nearly always occupies the center of an octahedron of oxygen
atoms (CN = 6) at a fixed Mg2+ -oxygen distance of 2.05 0.05 A.
In Table 3.2, stability constants for the binding of Ca
2
+and Mg
2
+ to var-
ious ligands are collected. We may note that selectivity of Ca2+ over Mg2+ is
not very great for simple carboxylate ligands, but that it tends to increase for
large multidentate ligands, such as EDTA and in particular EGTA. The Ca
2
+
sites in many intracellular proteins with "EF-hand" binding sites (see Section
V. C) bind Ca
2
+ about 10
4
times more strongly than Mg
2
+.
Another difference in ligand-binding properties of Mg 2+ and Ca 2+ can be
seen by comparing the rates of substitution of water molecules in the inner
/ ,
/ ,
/ ,
/ ,
/
.-'
/
(A)
(B) (C)
Figure 3.1
The three commonly observed modes of calcium carboxylate ligation. (A) The unidentate mode,
in which the calcium ion interacts with only one of the two carboxylate oxygens. (B) The biden-
tate mode, in which the calcium ion is chelated by both oxygen atoms. (C) The a-mode, in
which the calcium ion is chelated by one carboxylate oxygen, and another ligand is attached to
the a-carbon. Adapted from Reference II.
II. BASIC FACTS ABOUT CALCIUM: ITS COMPOUNDS AND REACTIONS 111
Table 3.2
Ca
2
+ and M
g
2+ (where available) stability
constants (log K) for different organic and
biochemical ligands. Most values are at ionic
strength 0.1 and 25C.5,6,12-15
Ligand
Acetate
Lactate
Malonate
Aspartate
Citrate
Nitrilotriacetate
EGTAa
EDTA
b
Glycine (Gly)
y-Carboxyglutamic acid (Gla)
Gly-Gly dipeptide
Gla-Gla dipeptide
Macrobicyclic amino cryptate [2.2.2]
Fluo-3
Fura-2
BAPTA
Quin-2
Phospholipase A2
Thrombin fragment I
Trypsinogen
Chymotrypsinogen
Chymotrypsin
Calmodulin, N-terminal
Trypsin
Calmodulin, C-terminal
Protein kinase C
a-Lactalbumin
Rabbit skeletal muscle
Troponin C, Ca 2 + /Mg 2 + sites
Carp parvalbumin
Bovine calbindin D
9k
0.5
l.l
1.5
1.6
3.5
6.4
10.9
10.6
1.4
1.3
1.2
3.2
4.5
6.2
6.9
7.0
7.1
3.6
3.7
3.8
3.9
4.1
4.5
4.6
5.3
~ 7
~ 7
7.3
~ 8 . 5
8.8
0.5
0.9
2.1
2.4
3.4
5.5
5.3
8.8
3.4
2.0
2.0
1.8
2.7
3.0
3.3
3.6
4.2
~ 4 . 3
" EGTA: ethylenebis(oxyethylenenitrilo)tetraacetate
b EDTA: ethylenedinitrilotetraacetate
hydration sphere by simple ligands, according to
This rate (log k, with k in s- 1) has been determined to be 8.4 for Ca2 + and 5.2
for Mg
2
+ ,16
112 3 I CALCIUM IN BIOLOGICAL SYSTEMS
The formation of biominerals is a complex phenomenon. In order to obtain
a feeling for the conditions under which inorganic solid phases in biological
systems are stable, it is of some interest to look at solubility products. Solubility
products, K ~ p , have a meaning only if the composition of the solid phase is
specified. For a solid compound with the general composition (A) k(B) ,(C) m the
solubility product is defined as
(3.1)
where [A], [B], etc., denote activities of the respective species, usually ionic,
in equilibrium with the solid. Activities are concentrations multiplied by an ac-
tivity coefficient, 'Y, nearly always less than unity. Activity coefficients for ions
in real solutions can be estimated from Debye-Hiickel theory 17 if the ionic strength
of the solution is known. In human blood plasma, the ionic strength, I, is about
0.16, and the activity coefficient for Ca
2
+ at 37C is 0.34. In many discussions
it may be sufficient to equate concentrations with activities.
The solid phase involved is essentially assumed to be an infinitely large,
defect- and impurity-free crystal with a well-defined structure. Microscopic crystals
have higher solubilities than large crystals, a well-known phenomenon that leads
to "aging" of precipitates, in which larger crystals grow at the expense of
smaller ones.
Many anionic species appearing in the solubility products may also be in-
volved in protonation equilibria in solution, such as those of phosphoric acid:
H
2
P0
4
- ~ H+ + HP0
4
2
-; HP0
4
2
- ~ P0
4
3
+ H+; etc. When the pros-
pects for the formation of a solid phase under certain solution conditions are
investigated, the activity, or concentration, of the particular anionic species
specified in the solubility product must be known, not only "total phosphate"
or "total calcium," etc. The data in Table 3.3 show that, at pH > 5, the most
stable (i.e., insoluble) solid calcium phosphate is hydroxyapatite.
Table 3.3
Solubility products, at pH 5 and 25C, for solid calcium phosphates.
Solid phase
CaS0
4
' 2H
2
0 (sulfate, "gypsum")
Ca(OH)2 (hydroxide)
CaHP0
4
' 2H
2
0 (hydrogen phosphate)
CaC0
3
(carbonate, "calcite," "aragonite")
CaC
2
0
4
' H
2
0 (oxalate, "whewellite")
{3-Ca3(P04h ({3-phosphate)
Cas(P04hOH (hydroxyapatite)
log /C,p
5.1
5.3
6.6
8.5
10.5
29
58
-log /C,p of
corresponding
Mg 2 + compound
where applicable
< 1.0
10.7
7.5
5.0
III. CALCIUM IN LIVING CELLS: METHODS FOR DETERMINING
CONCENTRATIONS AND SPATIAL DISTRIBUTIONS
Much of our present knowledge of the biological role of Ca 2+ ions in the reg-
ulation and modulation of cellular activities rests on the development of analyt-
ical techniques in three different areas: our ability to measure the low concen-
tration levels in the cytoplasm of resting cells, follow the concentration changes,
both temporally and spatially, that may occur as a result of an external stimulus,
and measure the distribution of Ca 2+ in various compartments of a cell. The
last decade has seen the emergence of many such new techniques, and the im-
provement of old ones, which has had a major impact on our understanding of
the detailed molecular mechanisms and dynamics of the Ca 2+ messenger sys-
tem. In this section, we will survey some of the most important techniques and
results obtained using these. Broadly speaking there are two main groups of
experimental techniques: those that aim at measuring the concentration of "free"
(or uncomplexed) Ca
2
+ -ion concentrations (or activities), and those that mea-
sure total calcium.
A. Measurements of "Free" Calcium Concentrations
1. Ca2 + -selective microelectrodes
Ion-selective electrodes can be made from a micropipette (external diameter
O.l-l/Lm) with an ion-selective membrane at the tip.18.19 For Ca
2
+ the mem-
brane can be made of a polyvinyl chloride gel containing a suitable Ca2+ -selec-
tive complexing agent soluble in the polymer gel. A commonly used complex-
ing agent is "ETH 1001" (see Figure 3.2A). An additional "indifferent" reference
electrode is needed. For measurements inside cells, the reference electrode can
also be made from a micropipette filled with an electrolyte gel. Often the ion-
selective and reference electrodes are connected in a double-barrelled combina-
tion microelectrode.
21
The whole assembly can then be inserted, using a micro-
manipulator, into a single cell typically 30-50 /Lm across. The arrangement is
depicted in Figure 3. 2B. With proper care, Ca 2+ microelectrodes can be used
to measure Ca2+-ion concentrations down to 10 ~ 8 M.
19
, 21 One limitation of
the technique is that the response time is usually in seconds or even minutes,
making rapid concentration transients difficult to follow.
2. Bioluminescence
Several living oganisms are able to emit light. The light-emitting system in
the jellyfish (Aequorea) is a protein called aequorin (M
r
:::=: 20 kDa). The light
is emitted when a high-energy state involving a prosthetic group (coelenterazine)
returns to the ground state in a chemical reaction that is promoted by Ca
2
+
113
114
(A)
(B)
" N ~ O "
~ O 0
o
X
o
ly0 0
/ N ~ /
o
"ETH 1001"
reference
electrode
(KCI- gel)
plasma membrane
of cell
II
II
II
II
II
II
II
II
II
-10 mM CaCI
2
porous plug containing
Ca
2
+- specific ligand
Figure 3.2
(A) Structure of a commonly used neutral Ca
2
+ chelator in Ca
2
+ -selective electrodes, "ETH
1001" (N,N'-di[ll-(ethoxycarbonyl) undecul]-N,N'-4,S-tetramethyl-3,6-dioxaoctan-l ,8-diam-
ide).20 (B) Schematic arrangement for the measurement of the activity (or concentration) of
Ca2+ ions in cells using a Ca
2
+-selective double-barreled microelectrode. Frequently the mi-
croelectrode is supplemented by a third, indifferent, electrode inserted into the bathing medium
surrounding the cell.
21
ions. At Ca
2
+ concentrations below ~ O . 3 ILM the emission is weak, but in the
range 0.5-10 ILM the emission is a very steep function of the concentration
(roughly as [Ca
2
+V
5
).18,19,22 The response to a Ca
2
+-concentration transient is
rapid (TI/2 = 10 ms at room temperature), and the light emitted can be accu-
rately measured even at very low light levels by means of image intensifiers
and/or photon counting. For measurements of Ca
2
+ concentrations inside cells,
aequorin has usually been introduced either through microinjection or through
some other means. A novel idea, however, is to utilize recombinant aequorin
reconstituted within the cells of interest, thus circumventing the often difficult
injection step.174
III. CALCIUM IN LIVING CELLS: DETERMINING CONCENTRATIONS AND SPATIAL DISTRIBUTIONS 115
3. Complexing agents with Ca
2
+ -dependent light absorption or
fluorescence
An important advance in the field of Ca
2
+-ion detennination was made by
R. Y. Tsien, who in 1980 described23 the synthesis and spectroscopic properties
of several new tetracarboxylate indicator dyes that had high affinity and reason-
able selectivity for Ca2+. All these dye molecules have a high UV absorbance
that is dependent on whether Ca 2 + is bound or not; a few also show a Ca 2 +-
dependent fluorescence. Tsien has also demonstrated that these anionic chelators
can be taken up by cells as tetraesters, which, once inside the cells, are rapidly
enzymatically hydrolyzed to give back the Ca2+ -binding anionic forms. Flu-
orescent tetracarboxylate chelators with somewhat improved Ca2+ selectivity
such as "BAPTA," "Quin-2," and "Fura-2" (Figure 3.3) were later de-
scribed.
24
These chelators are very suitable for measurement of Ca2+ -ion con-
centrations in the range 1 J.LM to 10 nM in the presence of 1 mM Mg2 + and
100 mM Na + and/or K +-i.e., conditions typically prevailing in animal cells.
Recently a new set of chelators that are more suitable for measurements of
calcium concentrations above 1 J.LM was presented.
25
The most interesting of
these is "Fluo-3," with a calcium-binding constant of 1.7 x 10
6
.
Whereas the emission spectrum for Fura-2 (Figure 3.3B), which peaks at
505-510 nm, hardly shifts wavelength when Ca
2
+ is bound, the absorption
spectrum shifts toward shorter wavelengths. In studies of free Ca2+ concentra-
tions where internal referencing is necessary, for example, in studies of single
cells, it is therefore advantageous to excite alternately at ~ 350 and 385 nm,
and to measure the ratio of fluorescence intensity at ~ 510 nm.
The use of fluorescent chelators has recently permitted studies in single cells
of rapid fluctuations or oscillations of free Ca 2 + and the formation of Ca 2 +
concentration gradients. Using a fluorescence microscope coupled to a low-light-
level television camera feeding a digital image processor, Tsien et al.
26
have
been able to reach a time resolution of about 1 s in single-cell studies. The
results of some highly informative studies made using this instrument are shown
in Figure 3.4. (See color plate section, page C-7.) The concentration of free
Ca2 + is presented in pseudocolor, and the Fura-2 concentration inside cells is
50-200 J.LM, as indicated in the figures. We see a Ca
2
+ gradient diffusing through
an entire sea-urchin egg ( ~ 1 2 0 J.L across) in 30 s. The free Ca 2+ concentration
of the resting egg ( ~ 1 00 nM) is increased to about 2 J.LM as Ca2 + diffuses
through the egg. The mechanism of propagation is believed to be a positive
feedback loop with inositol trisphosphate releasing Ca 2+ and vice versa (see
Section V).
A pertinent question concerning the uses of intracellular Ca 2 + chelators is
whether or not the chelator significantly perturbs the cell. The chelator will
obviously act as a Ca 2 + buffer in addition to all other Ca2 +-binding biomole-
cules in the cell. The buffer effect is probably not of any major consequence,
since the cell may adjust to the new situation by an increase in total Ca2 + ,
especially if the chelator concentration is in the J.LM range. The chelators could,
however, interact with and inhibit intracellular enzymes or other molecules, an
116
(A)
(B) 10
C0
2
H C0
2
H
l )
C0
2
H C0
2
H
l )
"BAPTA"
0.1 mM
C0
2
H C0
2
H
l )
"Quin-2"
8
6
C
"iii
c
OJ
. ~
4
2
300
wavelength / nm
400
Figure 3.3
Molecular structure of three chelators frequently used in measurements of "free" Ca2+ -ion con-
centrations. They may all be regarded as aromatic analogues of the classical chelator "EDTA";
their optical spectroscopic properties change upon binding Ca
2
+ ions. (A) For "BAPTA" the
spectral changes are confined to the absorption spectrum, whereas "Quin-2" and the "Fura-2"
in (B) show Ca2+ -dependent changes in their fluorescence spectra. 23, 24 (B) The relative fluores-
cence intensity of "Fura-2" at 505 nm as a function of the wavelength of the excitation light at
different Ca2+ concentrations, The data
26
refer to a solution containing 115 mM KCI, 20 mM
NaCl, and I mM Mg
2
+ at 37C and pH 7,05, At increasing Ca2+ concentration, the excitation
efficiency at ~ 350 nm is increased, but that at ~ 385 nm is decreased, In order to eliminate (as
much as possible) variations in fluorescence intensity in biological samples due to slight varia-
tions in dye concentrations and/or cell thickness, it is often advantageous to measure the inten-
sity ratio at 345 and 385 nm excitation wavelengths.
III. CALCIUM IN LIVING CELLS: DETERMINING CONCENTRATIONS AND SPATIAL DISTRIBUTIONS 117
effect that could result in aberrant cellular behavior. It is not unlikely that BAPTA
will bind to certain proteins. 27
4. Complexing agents with Ca
2
+-dependent NMR spectra
A series of symmetrically substituted fluorine derivatives of BAPTA (see
Figure 3.3A) has been synthesized.
28
,29 One of these chelators is 5F-BAPTA
(Figure 3.5A), which has a binding constant for Ca
2
+, of 1.4 X 10
6
M-
1
and a 19F NMR chemical shift, 0, that in the free ligand is different from that
in the complex with Ca 2 + = 6 ppm). The rate of Ca2 + dissociation,
k
off
' is 5.7 x 10 2 S - I, which gives the rate of association, k
on
' as 8 x 108
M 1 S - 1 according to
(3.2)
This exchange rate means that we are approaching the slow exchange limit in
19F NMR, and in subsaturating concentrations of Ca 2 + two 19F signals are seen
(see Figure 3.5B).
Since the areas of the NMR signals from the bound (B) and free (F) forms
of the ligand are proportional to their concentration, the free Ca 2+ concentration
(A) COOH eOOH eOOH eOOH

N N
P-LJ-Q
F F
(B)
9
I
B
6
I
ppm
3
I
F
o
I
Figure 3.5
(A) Molecular structure of the calcium chelator 5F-BAPTA whose 19F
chemical shift changes upon calcium binding. (B) 19F NMR spectrum
of a solution containing 5F-BAPTA and Ca 2+ in a molar ratio of 3: I.
Signal B originates from the Ca2+-5F-BAPTA complex, and F from
free 5F-BAPTA. Adapted from Reference 29.
is obtained simply as
2+ _ ~ . 1
[Ca ]free - F K
B
'
(3.3)
An additional beneficial property of SF-BAPTA and other fluorinated analogues
of BAPTA is that they will also bind other metal ions with a 19F chemical shift
of the complex that is characteristic of the metal ion.
29
Under favorable circum-
stances, it is thus possible to measure simultaneously the concentrations of sev-
eral cations.
For SF-BAPTA the selectivity for Ca
2
+ over Mg
2
+ is very good
(K
B
M
g
2
+ = 1 M ~ l ) . In applications of SF-BAPTA to intracellular studies, the
same protocol is used as with the parent compound and its fluorescent deriva-
tives: some esterified derivative, e.g., the acetoxymethyl ester, is taken up by
the cells and allowed to hydrolyze in the cytoplasm. The intracellular concentra-
tions of SF-BAPTA needed to get good 19F NMR signals depend on the density
of cells in the sample tube and the number of spectra accumulated. With accu-
mulation times on the order of ten minutes (thus precluding the observation of
concentration transients shorter than this time), Ca
2
+ concentrations of the order
of llLM have been studied in perfused rat hearts using SF-BAPTA concentra-
tions of about 201LM.34
B. Measurements of Total Calcium Concentrations
The measurement of total calcium in a biological sample can be made by any
method sensitive only to the element and not to its particular chemical form.
Atomic absorption spectroscopy is excellent as such a method. Obviously, the
spatial resolution that can be obtained with this method is limited, and it is hard
to imagine its application to elemental mapping of single cells. The techniques
discussed in this subsection have been limited to those that permit a spatial
resolution of at least 1 ILm on samples usually prepared by sectioning the frozen
biological specimens.
1. Electron probe and electron energy-loss techniques
When the electron beam in an electron microscope hits a thin sample, some
atoms in the sample will be excited or ionized, and returning to their ground
state will emit characteristic x-rays. The x-ray emission at different wavelengths
may then be measured by a photon-energy-sensitive detector. This is the basis
of electron probe x-ray microanalysis (EPMA). The electrons that pass through
the sample, and that give the transmission image in electron microscopy, will
suffer energy losses that depend on the nature (to some extent also, the chemical
state) and distribution of different elements. The outcome of these phenomena
forms the basis of electron energy-loss spectroscopy (EELS; see Figure 3.6).
scanned electron beam energy dispersive
______ detector (gives "EPMA"
....--- spectra)
Ill,
f i\ \ -+-- transmitted electron beam;
tttt gives energy loss spectra ("EELS")
Figure 3.6
Principles of electron probe microanalysis ("EPMA") and electron energy loss spectroscopy
("EELS" ).
A thin slice of a freeze-dried sample is exposed to a collimated beam of electrons that may be
scanned across the sample. Atoms ionized by electron collisions will emit x-radiation at wave-
lengths characteristic of their nuclear charge (i.e., characteristic of each element). With the use
of an energy dispersive x-ray detector, signals from different elements can be differentiated.
Apart from the characteristic x-rays, a broad-spectrum background x-ray emission is also present
because of inelastic scattering of the incident electrons. Some of the electrons that pass through
the sample will have lost energy because of ionization of atoms in the sample. The energy loss
is again characteristic for each element, and if the energy distribution of the transmitted elec-
trons is analyzed, this will have "peaks" at certain characteristic energies. If the energy resolu-
tion is pushed far enough I eV), the loss peaks even become sensitive to the chemical state
of the element.
The EPMA technique as applied to calcium has been improved by Somlyo
in particular. 30 Typically samples are rapidly frozen and sectioned at low tem-
peratures (- 130C) to preserve the in vivo localization of diffusible ions and
molecules. Spatial resolutions of 10 nm or better have been attained on 2:100 nm
thick freeze-dried cryosections. The minimal detectable concentration, which
requires some signal averaging, is approximately 0.3 mmol Ca per kg dry spec-
imen (i.e., 10 ppm). The calcium content of mitochondria and endoplasmic
reticulum in rat liver cells has been studied by EPMA (see Table 3.1).8
The high calcium content of endoplasmic reticulum (ER) is consistent with
the view that this organelle is the major source of intracellular Ca2 + released
119
through the messenger inositol trisphosphate (see Section IV.C). Other EPMA
studies have shown mitochondria to have a large capacity for massive calcium
accumulation in cells where cytoplasmic Ca 2 + concentrations have been abnor-
mally high, for example, as a result of damage of the cell membrane. 30
EELS is presently less well-developed than EPMA. Two of the major dif-
ficulties in the use of EELS for quantitative analysis of calcium and other ele-
ments are (i) large background, since it is a difference technique, and (ii) sen-
sitivity to specimen thickness. The major advantage of EELS is that the spatial
resolution is potentially much better than in EPMA, and can be I to 2.5 nm in
favorable specimens.
2. Proton-induced x-ray emission (PIXE)
A specimen exposed to a beam of high-energy (1 to 4 MeV) protons will
also emit characteristic x-rays just as in EPMA. The advantage of using protons
instead of electrons is that protons are more likely to collide with an atom, thus
producing excited atoms emitting x-rays. The sensitivity in detecting a particular
element is therefore much higher in PIXE than in EPMA or EELS. The PIXE
technique, which was developed at the University of Lund, Sweden, in the late
1960s, was originally used mainly for studies of fairly large objects.
9
In 1980 a group at Oxford University succeeded in focusing the proton
beam to a diameter of 1 f.Lm with sufficient energy (4 MeV) and beam intensity
(lOa pAJf.Lm 2) to allow elemental mapping at ppm concentrations. 31 Similar beam
performances ( ~ 0 . 5 f . L m diameter) are now also available at the University of
Lund and other laboratories. Beam diameters of 0.1 f.Lm are likely to be achieved
in the near future. Like EPMA, the PIXE method allows the simultaneous ob-
servation of several elements in the same sample. The biological applications of
the microbeam PIXE technique are limited, but it is clear that its potential is
great. Some representative results obtained with the Oxford microbeam are shown
in Figure 3.7. (See color plate section, page C-8.)
3. Ion microscopy
Ion microscopy is another technique capable of detecting all elements at the
ppm level. The basic idea is to expose a freeze-fixed, cryofractured, and freeze-
dried sample, which has been put onto a conducting substrate in a vacuum
chamber, to a beam of ions (e. g., D
2
+ or Ar + ). These ions will remove the top
two or three atomic layers of the sample surface by sputtering. A certain fraction
of the removed atoms will leave as ions. This secondary ion beam is accelerated
into a double-focusing mass spectrometer, where the ions are separated accord-
ing to their mass-to-charge ratio. The ion optics are designed to preserve the
spatial distribution of the emitted secondary ions, and an element image of the
sample can thus be produced with a spatial resolution of ~ 0 . 5 f.Lm.
32
The ion-
microscope technique can form images of a particular isotope of an element. In
principle, then, one could perform isotope labeling or "isotope chase" studies
IV. THE TRANSPORT AND REGULATION OF Ca
2
+ IONS IN HIGHER ORGANISMS 121
and follow, say, the fate of isotope-enriched 43Ca externally applied to a cell.
The ion-microscope technique has not yet come into widespread use, but the
quality of element (or ion) images obtained on single cells is impressive. 33
C. Summary
Much of our present knowledge about the biological role of Ca 2+ rests on de-
tailed measurements of the concentration, distribution, and chemical nature of
Ca
2
+ and its complexes. Concentrations of uncomplexed, or "free," Ca
2
+ can
be measured by Ca
2
+ -selective microelectrodes, bioluminescence and com-
plexing agents with Ca
2
+ -dependent light absorption, fluorescence, or NMR
spectra. An outcome of such studies is that the "free" Ca2+ concentration in
resting eukaryotic cells generally is very low, on the order of 100 to 200 nM.
Total Ca 2+ concentrations, uncomplexed and complexed, can be measured by
a variety of physical techniques. Some techniques, like atomic absorption, are
sensitive but give poor spatial resolution. Others involve the bombardment of
the sample with electrons or charged atoms, and can yield spatial resolutions of
the order of a few nm; however, there is a trade-off between detectability and
resolution.
2
+ IONS
IN HIGHER ORGANISMS
All living organisms need calcium, which must be taken up from the environ-
ment. Thus, Ca 2 + ions have to be distributed throughout the organism and
made available where needed. In higher organisms, such as humans, the blood-
plasma level of total calcium is kept constant (=2.45 mM) within narrow limits,
and there must be a mechanism for regulating this concentration. On a cellular
level we have already seen in the preceding section that the basal cytoplasmic
Ca 2+ concentration, at least in eucaryotic cells, is very low, on the order of
100 nM. At the same time the concentrations of Ca 2+ in certain organelles,
such as endoplasmic (or sarcoplasmic) reticulum or mitochondria, may be con-
siderably higher. If Ca 2+ ions are to be useful as intracellular "messengers,"
as all present evidence has it, Ca 2+ levels in the cytoplasm would have to be
raised transitorily as a result of some stimulus. Ca2+ ions may enter the cyto-
plasm either from the extracellular pool or from the Ca 2+ -rich organelles inside
the cell (or both). We could imagine Ca
2
+ channels being regulated by chemi-
cal signaling, perhaps by a hormone acting directly on the channel, or by a
small molecule released intracellularly when a hormone is attached to a mem-
brane-bound receptor. Some channels may be switched on by voltage gradients,
and both these mechanisms may operate concurrently.
Increased intracellular Ca 2 + levels must eventually be brought back to the
basal levels, in some cells very quickly. The ions could be transported out of
the cell or back into the Ca2+ -rich organelles. This transport will be against an
electrochemical potential gradient, and thus requires energy. There are many
possibilities for different forms of Ca 2+ transport and. regulation in living sys-
tems, and we still know fairly little about the whole picture. Detailed studies
are also complicated by the fact that, in higher organisms, cells are differen-
tiated. Nature is multifarious, and what is valid for one type of cell may not be
relevant for another. With these words of caution we will start out on a macro-
scopic level and continue on toward molecular levels.
A. Ca
2
+ Uptake and Secretion
The uptake of Ca 2+ from food has mostly been studied in typical laboratory
animals, such as rats, hamsters, chickens, and humans. In humans, uptake oc-
curs in the small intestine, and transport is regulated by a metabolite of vitamin
D, calcitriol (l,25-dihydroxy vitamin D
3
).34 The uptake process is not without
loss; roughly 50 percent of the calcium content in an average diet is not ab-
sorbed. To maintain homeostasis and keep the calcium level in blood plasma
constant, excess Ca 2+ is excreted through the kidney. The main factor control-
ling this phenomenon in vertebrates is the level of the parathyroid hormone that
acts on kidney (increases Ca2 + resorption), on bone, and, indirectly, via stim-
ulated production of calcitriol, on the intestinal tract (increases Ca
2
+ uptake).
Calcium enters the cells from the outside world, i.e., the intestinal lumen, by
traveling through the brush-border membrane of the intestinal epithelial cells,
through the cytosolic interior of these cells, and into the body fluids through the
basal lateral membranes of the same cells. The molecular events involved
need to be studied further. Figure 3.8 outlines the Ca2 + transport processes
known or thought to occur.
Transfer through the brush-border membrane is assumed to be "passive"
although indirectly facilitated by calcitriol. The calcitriol effect may be due to
synthesis of a carrier protein,35 but could also be an effect of altered membrane
lipid composition.
36
The fate of Ca
2
+ ions, once inside the epithelial cell, is a
much-debated subject. What appears clear is that the Ca2 + ions entering through
the brush-border membrane do not cause an increase of the low cytosolic Ca
2
+
concentration. It is thus quite likely that the Ca 2 + ions are carried through the
cell but the means of transportation is unknown. One plausible carrier is the
intracellular low-molecular-weight Ca2+ -binding protein calbindin D
9K
(M
r
= 9
kDa) formerly known as ICaBP (see Section V.C).35 Its synthesis is induced by
vitamin D, and it is mainly found in mammalian intestines. The porcine and
bovine calbindin D
9K
has a Ca2 + binding constant of KB = 3 X 10
8
M - 1 in
low ionic strength media 37 and K
B
= 2 X 10
6
M -1 in the presence of 1 mM
Mg2 + and 150 mM K + .38 The concentration of calbindin D
9K
in epithelial cells
can reach millimolar levels, 35 which could allow it to facilitate Ca2 + diffusion
across the cytosol. This was first suggested by Williams, subsequently elabo-
rated by Kretsinger et aI. in 1982,39 and later demonstrated in a model cell by
Feher. 40 The basic idea is that, although the diffusion rate of Ca 2 + ions
( ~ 1 O ~ 5 cm2 s -1) is higher than for the (Ca 2+)2 calbindin complex
123
S
~
~
Ca
2
+
c
Ca
2
+
OJ
.2
0
vesicles?
0
7
Na+
I

CaBP?
~
mitochondria
brush
cytosolic compartment
basal-
border
lateral
membranes
Figure 3.8
A scheme representing some of the known and hypothetical molecular participants in the trans-
port of Ca 2 + across intestinal epithelial cells. Transport across the brush-border membrane is
generally assumed to be passive or to be facilitated by a carrier (1M Cal), and is also influenced
by vitamin D. Transport through the cell may be in vesicles and/or in association with Ca2+_
binding proteins (CaBP), notably calbindins D
9k
(mammals) or D
28k
(avians). Temporary storage
or buffering of Ca2 + may be through cytosolic CaBPs, mitochondria, endoplasmic reticula
(ER), or other organelles. Transport of Ca2 + out of the cell through the basal-lateral membranes
is energetically uphill, and appears primarily accomplished by a Ca
2
+-ATPase and possibly to
some extent by a Na
2
+ _Ca
2
+ antiport. Adapted from Reference 35.
( ~ 0.2 X 10 - 5 cm2 s - I), the fact that the concentration of the latter complex
may be about 10
3
times higher than that of free Ca 2+ will result in an increased
net calcium transport rate. Calbindin would, in fact, act very much like myo-
globin in facilitating oxygen transport through muscle tissue.
Plausible as the above mechanism may seem, it may, however, not be the
whole truth. An alternative mechanism is vesicular transport. In chicken intes-
tine it has been shown that the only epithelial organelles that increased in Ca 2+
content as a result of calcitriol treatment were the lysosomes. 41 The result lends
support to a transport mechanism involving Ca2+ uptake across the brush-bor-
der membrane by endocytic vesicles, fusion of these vesicles with lysosomes,
and possibly also delivery of Ca 2+ to the basal lateral membrane of the epithe-
lial cell by exocytosis. This process would also explain the vitamin-D-induced
alterations in brush-border-membrane lipid compositions as a consequences of
preferential incorporation of certain types of lipids into the vesicles. Interest-
ingly, the lysosomes in the chicken studies also contained high levels of calbin-
din D
28k
-a type of vitamin-D-induced Ca2+ -binding protein found in avian
intestines-making it conceivable that this protein acts as a "receptor" for Ca2 +
at the brush-border membrane and upon Ca
2
+ binding could become internal-
ized in endocytic vesicles.
41
The basal lateral plasma membrane contains at least two types of Ca 2+
pumps that also may playa role in Ca
2
+ uptake, one ATP-driven, one driven
by a concurrent flow of Na + ions into the cytoplasm (i.e., a Na + -Ca 2+ anti-
port; see Figure 3.8). We discuss these types of transporting proteins in the
next subsection.
There are some apparent analogies between intestinal Ca 2 + transport and
that occurring in the placenta. Transplacental movements of Ca 2+ increase dra-
matically during the last trimester of gestation.
42
In mammalian placental troph-
oblasts, high concentrations of calbindin D
9K
are found.
43
,44 The protein synthe-
sis also in this tissue appears to be under calcitriol regulation. Ca 2 + ions have
to be supplied by mammalian females, not only to the fetus during pregnancy,
but also to the newborn child through the mother's milk. The molecular details
of Ca
2
+ transport in the mammalian glands have not been extensively studied.
In milk, Ca
2
+ is bound mainly to micelles of casein, and the average Ca
2
+
content is reported to be 2.5 glliter (see Table 3.1).
B. Intracellular Ca2+ Transport
In order to provide a better understanding of the role of Ca2 + as an almost
universal regulator of cellular function, we need to take a brief look at the many
ways by which Ca2 + ions can be transported in or out of eukaryotic cells.
Although various transport pathways have been elucidated, the present picture
is probably not complete, since the molecular structures and properties of the
transport proteins are only partially known. The major pathways for Ca2 + trans-
port across cellular membranes involve three membrane systems: the plasma
membrane, the inner mitochondrial membrane, and the membrane of the endo-
plasmic reticulum (ER) (or, in striated muscle cells, a specialized form of ER
called the sarcoplasmic reticulum (SR): (Figure 3.9). Two of the membrane-
bound transport systems are Ca
2
+-ATPases, since they derive their main energy
from the hydrolysis of ATP (l and 2 in Figure 3.9). Their properties do, how-
ever, differ in many other respects, as we will see.
1. The Ca2 + -ATPases
The plasma membrane Ca
2
+-ATPase (PM Ca
2
+-ATPase) of erythro-
cytes-first recognized by Schatzmann in 1966
45
-was isolated in pure form
by Niggli et at. in 1979, using an affinity column with an ATP-ase binding
protein, calmodulin (see Section V.A), coupled to the ge1.
46
Ca
2
+-ATPases
purified from other types of plasma membranes appear to be very similar. The
schematic structure of the erythrocyte membrane Ca 2 + -ATPase is presented in
PM
125
ER(SR)
ATP
Figure 3.9
Schematic representation of the major pathways for the transport of Ca2 + across cellular mem-
branes. PM, plasma membrane; ER(SR), endoplasmic reticulum (sarcoplasmic reticulum); M,
mitochondria; ~ 'IT, difference in membrane potential. The transport proteins shown are: I and 2,
PM and ER(SR) Ca
2
+-ATPases; 3 and 4, PM and ER(SR) receptor-mediated Ca
2
+ channels; 5
and 6, PM and M (inner-membrane) Na + ICa2+ exchangers; 7 and 8, PM and M voltage-sensi-
tive Ca2+ channels. In addition, some not-wellCdefined "passive" transport pathways are indi-
cated by dashed arrows.
Figure 3.10.
47
The sarcroplasmic reticulum in muscle cells is abundant in Ca2 +-
ATPase. It is estimated that this protein constitutes more than 80 percent of the
integral membrane proteins, and covers a third of the surface area.
The sarcoplasmic reticulum Ca
2
+ -ATPase (SR Ca
2
+-ATPase) was first pu-
rified by MacLennan in 1970.
48
Presently it is the best characterized Ca2 + -
ATPase. A schematic model and a summary of some properties are given in
Figure 3.11.
49
Ten hydrophobic segments of about 20 amino-acid residues each
are revealed by hydropathy plots, and these segments are assumed to span the
membrane as a-helices. (For the one-letter codes for amino acids, see Appendix
B in Section IX.) The phosphorylation site has been identified as Asp-35I, and
the nucleotide binding domain is following the phosphorylation domain. The
Ca2 +-binding sites are located within the predicted trans-membrane domains
126
N
1 phospholipid sensitive domain
2 calmodulin binding domain
3 cAMP phosphorylation domain
4 hinge
Figure 3.10
Schematic structure of the calmodulin (CaM)-activated plasma membrane Ca2 +-ATPase of
erythrocytes. Some molecular characteristics are: My = 138,000: transport rate (30C), 20-70
Ca
2
+ ions per protein molecule per second; K
M
(Ca
2
+) = 0.5 J.LM (cytoplasmic side in high-
affinity form); Ca
2
+/ATP ratio, I(?); activated not only by CaM but also by acidic phospho-
lipids and unsaturated fatty acids. Figure kindly provided by R. Moser and E. Carafoli.
(see Figure 3.11). This was shown through a series of site-directed mutations in
which likely Ca 2+ -liganding residues like Asp, GIu, and Thr were mutated into
residues lacking possible side-chain ligands (e.g., Asn, GIn, and Ala).50
The presently accepted reaction cycle involves two main alternative confor-
mations, E] and E
2
, the former with two high-affinity sites (K
m
;5 1 J.LM)4 on
the cytoplasmic side, which in E
2
are open to the luminal side with
K
m
~ 1 mM.
49
,5] The mechanism suggested for Ca
2
+ transport (Figure 3.12)
has many features similar to that suggested by Williams for H + translocation in
the mitochondrial ATPase.
52
It is instructive to consider briefly the thermodynamic limits of the transport.
(The discussions about the thermodynamics behind Ca2+INa + transport pertain
to Na +IK + gradients in excitable tissues as well). Let us define an "inside"
and an "outside" separated by a membrane, as shown in Figure 3.13, where
[Ca
2
+] and l/J denote activities and membrane potentials, respectively. The dif-
ference in electrochemical potential, /).p" , across the membrane for a Ca 2+ ion
is given by
(3.4)
127
ATP
\ I

/.
C
E982
0981
K972
E90
R63
stalk
domain
N
trans-
membrane
domain
Figure 3.11
Schematic structure of the Ca 2 +- ATPase of sarcoplasmic reticulum. Some molecular characteris-
tics are: My = 110,000; K
m
< li-LM (two Ca 2+ sites on cytoplasmic side in high-affinity form);
Ca
2
+/ATP ratio, 2; Mg
2
+ required for activity. The amino-acid residues labeled were mutated
to a residue lacking side chains capable of binding Ca2 + . Mutations at the circled positions
resulted in complete loss of Ca2+ transport activity, suggesting that the circled residues partici-
pate in Ca
2
+ binding. Adapted from Reference 50.
2Ca
2
+ ATP(1
\ ATP AOPP j P
E
1
=====> E
1
(Ca
2
+)2 =====> E:(Ca
2
+)2=====> E: (Ca
2
+)2
a b
c
e
d
P
1
E2 <===, l2
Pi (10 mM)
Figure 3.12
Simplified schematic reaction cycle of the Ca
2
+ -ATPase of sarcoplasmic reticulum (SR).49,51
The transport protein is assumed to be in either of two states, E] on the cytoplasmic side, or E
2
on the side of the SR lumen. Starting from E] in the upper left comer, the reactions steps shown
are: (a) binding of Ca
2
+ and ATP (approximate dissociation constants within parentheses); (b)
rapid phosphorylation of Asp-351 of the protein (E] -P) and release of ADP; (c) transformation
from an energy-rich, high-Ca2+ -affinity conformation (E] -P) (Ca2+)z to a low-energy, low-af-
finity conformation (E
2
-P)(Ca
2
+)z; (d) hydrolysis of the phosphorylated protein and release of
the phosphate into the lumen; (e) return to the original state.
128
outside (0)
(cytoplasm)
Ca
2
+===t
inside (i)
(SR Lumenj
Figure 3.13
Schematic representation of Ca
2
+ transport through a membrane
by a Ca
2
+ -ATPase molecule. ' denotes membrane potentials.
where F is Faraday's constant, T the temperature, and R the gas constant. If we
assume = 0, which appears reasonable for the SR membrane according to
experimental evidence, we may calculate the free-energy change, at 25C
for transferring moles of Ca
2
+ across the membrane. This becomes =
X = X 4.1 kcallmol if [Ca
2
+]j[Ca
2
+]j = 10-
3
and =
X 5.4 kcallmol if [Ca
2
+lo/[Ca
2
+]j = 10 -4. Under the pertinent cellular
conditions, the free-energy change associated with ATP hydrolysis to ADP and
Pi has been calculated by Tanford to be = - 13 to - 14 kcallmop3 In the
absence of a membrane potential, it is thus possible to transport two Ca2 + ions
for every ATP molecule hydrolyzed against a concentration (or activity) gradient
of 10
4
or more. This treatment says nothing, of course, about the molecular
details of this transport. A more detailed model for the transport cycle has been
proposed by Tanford.
53
In the specialized cells of muscle tissue, the sarcoplasmic reticulum may
contain much calcium, and if all were "free" Ca
2
+, the concentration could be
as high as 30 mM.
54
This value would cause an osmotic pressure difference
across the membrane, as well as put a high demand on the SR Ca
2
+-ATPase.
A lowering of the free Ca
2
+ concentration inside the SR would clearly be ben-
eficial. In the presence of oxalate or phosphate ions in the external medium,
calcium oxalate or phosphate may precipitate inside the sarcoplasmic reticulum,
but under normal circumstances it appears that Ca2 + ions inside the SR are
bound to a very acidic protein, calsequestrin.
54
Each molecule (M
r
= 40 kDa)
is able to bind 40 to 50 Ca2 + ions with an effective dissociation constant of
about 1 mM (at I = O. 1). The protein has a low cation specificity and behaves
in many respects like a negatively charged polyelectrolyte. It has been crystallized55
and we may soon have access to its x-ray structure.
outside (0) inside (i)
(cytoplasm)
129
Figure 3.14
Schematic representation of the Ca
2
+INa + exchanger of the
plasma membrane. 'I' denotes membrane potentials.
2. The Na +ICa
2
+ exchanger of the plasma membrane
Presently available information on the Na +ICa2+ exchanger has mainly been
obtained from studies of the large cells of the giant squid axon and of plasma-
membrane vesicles from various other tissues.
56
,57 In heart plasma-membrane
vesicles, the exchanger has the following characteristics: K
m
= 1.5-5,uM for
Ca
2
+ and ~ 2 0 nM for Na+; V
max
= 20 nmol Ca
2
+/mg protein. 58 The stoichi-
ometry is at least 3: 1 Na +ICa2+. Very few molecular details of the exchanger
are available at present. We may again briefly consider the thermodynamic
framework for an Na +ICa 2+ exchanger (Figure 3. 14). The difference in elec-
trochemical potential for Na + and Ca 2+ across the membrane is:
(3.5)
(3.6)
The free-energy change, LlGfa
2
+, associated with a transfer of LlnCa2+ moles of
Ca 2+ from the inside to the outside is LlGf
a2
+ = Llnca2+ x Ll,uca2+, and the
corresponding change associated with the movement of LlnNa+ moles of Na +
from the outside in is L l G ~ a + = - LlnNa+ X Ll,uNa+' If these free-energy changes
are coupled via the exchanger, there will be a net flux of Ca2+ as long as the
free-energy difference,
is less than zero. We can write for the transport of 1 mol Ca 2+ as
[
[Ca2+]0 [Na+]o]
= 2.303 RT log 2+] - X log [ +]
[Ca i Na i (3.8)
+ (2 - X
Equating ion activities with concentrations, we note that in a typical mammalian
cell [Na +]0 = 110-145 mM, and [Na +]i = 7-15 mM, or [Na +]J[Na +]i = 10.
In the absence of a membrane potential difference = 0), Equation (3.8)
can thus be simplified to
(3.9)
To pump one Ca2+ ion out of a cell against a concentration gradient of about
10
3
(l fLM 1 mM) requires that at least 3 Na + ions pass in the opposite
direction, thus maintaining < O. What then will be the effect of a mem-
brane potential difference? Most animal cells, particularly excitable cells such
as nerve and muscle cells, have resting potential differences, '\(f, over the plasma
membrane of 30 to 90 mV (cytoplasm negative). For this value we find the
change in free energy, for the transport of one mol Ca 2+ to be
[
[Ca2+]0 ]
= 2.3 RT log 2+ - + (2 - 0.1 F.
[Ca ]i
(3.10)
Thus for > 2, we have < 0, and the transport of Ca 2+ against a
concentration gradient of about 103 will be promoted. This is another good
reason for having a Na +ICa2+ exchange stoichiometry of 3: 1.
3. Mitochondrial Ca
2
+ transport: influx
Mitochondria isolated from various types of animal cells-but, interest-
ingly, not those from plant cells--can rapidly accumulate exogenous Ca
2
+.59
The transporter is located in the inner membrane and the driving force behind
the Ca
2
+ transport appears to be merely the high potential difference across this
membrane = 150 to 180 mY, negative in the inner matrix). This potential
difference is fairly closely maintained by the pumping out of H + from the
matrix by cell respiration. For the transport of 1 mol Ca2+ from the "outside"
(= cytoplasm) to the "inside" (= inner mitchondrial matrix), we may deduce
from Equation (3.4) that the free-energy change may be written
= - 1)
[Ca
2
+]0
= - RT In C 2+ - 2F
[a ]i
(3.1l)
2
+ IONS IN HIGHER ORGANISMS 131
From this analysis it may be inferred that the limiting Ca 2+ concentration (or
activity) ratio that can be achieved by this electrogenic pump (i.e., /)"G = 0)
IS
e -2F!!.P/RT
(3.12)
With /)"'l' = 150 mV, this ratio is calculated to be 8.4 X 10 - 6 at 25C. It is
evident that, as long as the Ca2+ influx would not lower the membrane potential
difference, the Ca 2+ uniporter has a very high pumping potential. Measured
values of the pumping rate, Vmax, are indeed high (>10 nmol/mg protein59) and
probably limited only by the rate of electron transport and H + extrusion in the
mitochondria.
Mitochondria may accumulate large quantities of Ca
2
+, probably to main-
tain electroneutrality. To prevent the buildup of high concentrations of free Ca2+
(and of osmotic pressure), phosphate ions are also transported into the inner
matrix, where an amorphous calcium phosphate--or possibly a phosphocitrate
60
-
is formed. The equilibrium concentration of free Ca2+ in the mitochondrial
matrix may as a result be comparatively low, on the order of 1 p,M.
The molecular nature of the mitochondrial Ca2+ uniporter continues to be
elusive, and needs to be studied further.
4. Mitochondrial Ca2+ transport: efflux
Mitochondria, as well as SR, release Ca2+ ions by mechanisms other than
"back leakage" through the pumps. In mitochondria from excitable cells, the
efflux occurs mainly through an antiport, where 2 Na + ions are transported
inward for every Ca2+ ion departing for the cytosolic compartment.6l In other
cells there is evidence for the dominance of a 2H + - Ca2+ antiport .59 In all
likelihood the Ca2+ efflux is regulated, possibly by the redox state of pyridine
nucleotides in the mitochondria. As with the Ca2+ uniporter, few details on the
molecular nature of the antiporters are presently available.
5. Ca2+ efflux from non-mitochondrial stores
Release of Ca
2
+ from ER and SR presently appears to be the prime effect
of the new intracellular messenger 1,4,5-triphosphoinositol (l,4,5-IP3) released
into the cytoplasm as a result of an external hormonal stimulus (see Section
IV.C). It seems that receptors for 1,4,5-IP
3
have been established on ER, and
that the binding of 1,4,5-IP
3
causes a release of Ca2+ stored in this orga-
nelle.
62
,63,170,J71 In addition to the receptor-controlled Ca
2
+ efflux, there may be
other pathways for Ca
2
+ release, and Ca
2
+ mobilization may be regulated by
other intracellular entities, the Ca
2
+ ions themselves included.
6. Other voltage-gated or receptor-activated Ca
2
+ channels
In addition to the transport pathways already discussed, some cells seem to
have Ca 2 + channels in the plasma membrane that can be opened by the action
of an agonist on a receptor or that are gated in response to changes in membrane
potential. 64 For example, Ca
2
+ channels can be opened by nicotinic cholinergic
agonists
65
or by the excitatory amino acid N-methyl-D-aspartate (NMDA).66
Endochrine cells and also some muscle and neuronal cells have voltage-sensitive
Ca 2 + channels. 67,68 We will not discuss these further, but merely point to their
existence. We finally note that during the last few years knowledge about the
mechanisms of Ca
2
+ entry and release to and from extracellular and intracellular
pools has increased dramatically, and we refer the reader to recent reviews of
the field. 175,176
C. Inositol Trisphosphate and the Ca
2
+ Messenger System
A "second" messenger is an entity that inside a cell mediates the action of
some hormone at the plasma membrane, the hormone being considered the "first"
messenger. The first such second messenger to be discovered-in fact, the very
molecule that led to the formulation of the whole concept-was cyclic AMP. 69
During the decade following the discovery of cAMP, it was gradually realized
that intracellular release of Ca2 + ions also accompanied hormonal stimuli, and
the Ca2+ ion slowly became regarded as a second messenger. This idea was
first clearly enunciated by Rasmussen70 as early as 1970, and gained general
acceptance when the ubiquitous intracellular Ca 2+ -binding protein calmodulin
(see Section V.A) was discovered. In the mid-1970s this protein was shown to
be a Ca 2 +-dependent regulator of a large number of Ca2 +-dependent enzymes,
transport proteins, etc., establishing a molecular basis for Ca 2+ action in cells.
There were some puzzling facts, however. Although a transitory increase in
intracellular Ca2 + concentration in response to the binding of a hormone or
transmitter substance to a surface receptor could result from extracellular Ca2+
being released into the cytoplasm, there was compelling evidence for muscle
cells that the main Ca2 + source was the sarcoplasmic reticulum (SR). This
result led to the hypothesis of ''Ca2+ -induced Ca2+ release," i.e., that upon
stimulation of the cell, a small amount of Ca2 + entered into the cytoplasm and
triggered the release of greater amounts of Ca
2
+ from the SR. For some cell
types it could, however, be shown that transient increases in intracellular Ca 2 +
could occur even when extracellular Ca
2
+ was removed, although prolonged
responses required the presence of extracellular Ca
2
+. Although some special-
ized cells have gated plasma-membrane Ca
2
+ channels, release of Ca
2
+ into
the cytoplasm from intracellular stores appears to be of at least equal impor-
tance. Furthermore, there is now overwhelming evidence
63
,70-72 that intracellu-
lar Ca2 + is released in response to the formation of a new type of intracellular
messenger: 1,4,5-IP
3
. Receptors for this messenger have recently been found in
the membranes of intracellular organelles, and binding of 1,4,5-IP
3
to these
receptors results in the release of Ca 2 + ions. 73
hormonal
receptor plasma
membrane
II
II
II
II
II
II
II ?
II
II
II
II
II
II
II
II
II
I
I
V
133
Figure 3.15
Outline of the presumed role of inositol phosphates in the intracellular mobilization of Ca
z
+.
Upon binding of an agonist to a plasma-membrane receptor, phosphatidylinositol 4,5-bisphos-
phate (PIP
z
) membrane lipids are hydrolyzed to give diacylglycerol (DG) and inositol 1,4,5-
trisphosphate (1,4,5-IP
3
). The latter interacts with specific receptors on the endoplasmic reticu-
lum (ER) membrane that trigger the release of Ca
z
+ into the cytosol. Ca2+ may be returned to
the ER through the Ca2+-ATPase of the ER membrane (see Figure 3.9) and also by a direct
influx of Ca
2
+ from the extracellular medium.
70
,8l
1,4,5-IP
3
is formed as a product in the hydrolysis of a special phospholipid
present in the cell membrane: phosphatidyl-inositol-4,5-bisphosphate. This re-
action, then, is the initial receptor-stimulated event. The newly formed 1,4,5-
IP
3
is assumed to diffuse into the cytoplasm, and eventually reach intracellular
1,4,5-IP
3
receptors on the ER, thereby triggering the release of Ca2+. A sim-
plified reaction scheme is shown in Figure 3.15. A diacylglycerol (DG) is also
formed in the hydrolysis step. DG can also act as an intracellular messenger,
and stimulates the activity of a membrane-bound protein kinase, known as pro-
tein kinase C (PKC). As a result, PKC may phosphorylate certain key proteins
and influence their activity. Protein kinase C is also activated by Ca2 + ions, a
fact that illustrates Nature's knack in designing regulatory networks! 1,4,5-IP
3
is either directly degraded in a series of enzymatic steps back to inositol, which
is then used to resynthesize the phospholipid, or it may be further phosphor-
ylated to inositol-l,3,4,5,-tetraphosphate (1,3,4,5-IP4)' which may undergo de-
phosphorylation to form inositol-I,3,4-trisphosphate (l,3,4-IP
3
). The biological
functions of the latter compounds are now being investigated.
The intracellular levels of Ca2 + are restored back to the normal low resting
values (loa to 200 nM) via transport back into the SR, and/or into mitochon-
dria, or out through the plasma membrane by the pumping mechanisms dis-
cussed in Section IV.B. As was briefly mentioned above, depriving a cell of
extracellular Ca 2+ will eventually make the cell incapable of prolonged re-
sponses to external stimuli. It appears that the intracellular Ca2+ stores may
become depleted if not replenished. It has been suggested that the intracellular
ER Ca 2+ pool has a direct route of access to the extracellular pool, a route that
is closed when the ER pool is full. 74
In a sense, then, Ca 2+ seems to have been downgraded by the inositolphos-
phates from a "second" to a "third" messenger; however, the pivotal role of
Ca2 + as a regulator of cellular activities remains undisputed.
D. Summary
The fluxes of Ca2 + ions and their regulation in higher organisms, as well as in
microorganisms, depend on several transport proteins in addition to vesicular
and gated processes. An important class of transport proteins are the Ca2 +-
ATPases, which are particularly abundant in muscle cells. These proteins trans-
locate Ca 2 + ions against large activity (or concentration) gradients through the
expenditure of ATP. Transport of Ca2 + ions against activity gradients across
membranes may also be accomplished by coupled transport of other ions, like
Na +, with a gradient in the opposite direction.
As a result of some external stimulus-the action of a hormone, for ex-
ample-the "free" Ca
2
+-ion concentrations in the cytoplasm of many cell types
may transiently increase several orders of magnitude. This increase largely re-
sults from the release of Ca2+ from intracellular stores (ER, SR) in response to
the initial formation of a new type of messenger, 1,4,5-IP
3
. The activity of
Ca2 +-transport proteins eventually restores the Ca2 + concentration levels to resting
levels. This sequence of events forms the basis for Ca2 + , s role in the regulation
of a wide variety of cellular activities (see Section V).
V. MOLECULAR ASPECTS OF Ca
2
+-REGULATED
INTRACELLULAR PROCESSES
So far we have mainly discussed the routes and means by which the concentra-
tion of Ca 2+ ions in the cytoplasm can be transiently increased and brought
back to resting levels. But changing the cytoplasmic Ca2 + concentration is not
enough. In order to influence the cellular machinery, the Ca2 + ions must inter-
act with different proteins, intracellular Ca
2
+ receptors if you like. These in-
tracellular Ca2 + -receptor proteins must have certain properties in order to func-
tion.
2
'-REGULATED INTRACELLULAR PROCESSES 135
(i) Their Ca 2 + -affinity must be such that their Ca2 + -binding sites are essen-
tially unoccupied at resting levels of free Ca2 + ( ~ 1 O -7 M) and occupied at
levels reached upon stimulus (generally assumed to be 10 -5 to 10 -6 M). This
means that the binding constants Kf?2+ should be ~ 10
6
M -1.
(ii) We should also remember that Ca
2
+ must exert its function in the pres-
ence of a number of other ions; in mammalian cells the intracellular concentra-
tion of "free" Mg
2
+ ions is around 1 mM, and that of K + ions around 100 to
150 mM. The receptors must therefore have an adequate selectivity for Ca 2+ .
(iii) In response to Ca
2
+ binding, a Ca
2
+ receptor must undergo some kind
of conformation change that either alters its interaction with other molecules or
changes its activity if it is an enzyme.
(iv) Finally, there are kinetic considerations. In many cells a rapid response
is essential, and therefore the receptors must be able to interact swiftly-within
milliseconds-with incoming Ca
2
+ ions, and the ions must also be able to de-
part almost as rapidly.
A few proteins have been discovered that qualify as intracellular Ca 2+ re-
ceptors. The best known of these is calmodulin (CaM), which appears to be
present in all eukaryotic cells. Most of the cellular responses elicited by Ca2 +
appear to result from interactions between the Ca 2+ -calmodulin complex and
various other target enzymes and proteins. 75 Another important Ca2 +-receptor
protein is troponin C (TnC), which occurs in muscle cells and is instrumental
in mediating muscle contraction.
76
These two types of proteins are highly ho-
mologous, as we shall see, and may be considered members of a superfamily
of closely related intracellular Ca2 +-binding proteins. This superfamily has been
given the name' 'the calmodulin superfamily," and close to 200 distinct family
members are presently known.
77
Not all members of the superfamily may qual-
ify as Ca
2
+ receptors; some like parvalbumins and calbindins (see Section IV.A)
appear to have a role in intracellular transport and/or Ca 2+-buffering. For oth-
ers, such as the 5-100 proteins 78 found predominantly in brain tissue, and cal-
cimedins,79 isolated from smooth muscle, the biological function is still unclear.
One Ca
2
+ receptor with enzymatic activity is protein kinase C. Its activity
is markedly increased in the presence of Ca2+ , and it has a high calcium-bind-
ing constant (see Table 3.2) in the presence of diacylglycerol or phorbol es-
ters.
80
During recent years, groups interested in the role of Ca2 + in secretion and
in the control of membrane cytoskeleton have identified some intracellular Ca2+/
phospholipid-binding proteins that appear to be distinct from the calmodulin
superfamily; these include lipocortin, endonexin, calelectrin, p36, and calpac-
tin. 81-83 These membrane-binding proteins are collectively called annexins, 84 and
contain repeated domains distinct from EF-hands. The Ca
2
+ sites are very sim-
ilar to that observed in phospholipase A
2
, as shown by the recently determined
x-ray structure of annexin V.
172
A condensed overview of the interaction of
Ca
2
+ with intracellular proteins is shown in Figure 3.16. We will now go on to
discuss the molecular properties of some of the proteins mentioned above, start-
ing with calmodulin.
136
,/ ....... / // I
-' / J.
;:;- r
phosphorylated
target proteins
protein kinase C plasma membrane
mitochondrion
dephosphorylated
target proteins
muscle ...... l---- Troponin C -- parvalbumins
C,','m,""
. . 1
secretion-regulating proteases
proteins (calpain... )
Ca
2
+/calmodulin- Ca
2
+-calmodulin calmodulin-activated or
00"/1'\" /1\ 00'
calmodulin-dependent
protein kinases

phosphorylated target
protein
Figure 3.16
Condensed overview of the interaction of Ca2+ with intracellular proteins.
A. Calmodulin
Calmodulin is a small acidic protein (My = 16,700), the amino-acid sequence
of which has been remarkably preserved during evolution. Early on, an analysis
of its amino-acid sequence indicated that it should have four Ca 2+ -binding sites,
a deduction that proved to be correct. The three-dimensional x-ray structure of
bovine brain calmodulin85 has been solved to a resolution of 2.2 A. A space-
filling model is shown in Figure 3. 17. (See color plate section, page C-9.) The
molecule has a dumbbell-like shape, with two globular domains connected by
an eight-tum a-helix-an unusual structural feature. In the crystal structure,
there are no direct contacts between the two globular domains, each of which
contains two Ca 2+ -binding sites. The Ca2+ sites are all constructed in the same
way: two a-helices separated by a calcium-binding loop, 12 amino acids long,
and wrapped around the Ca
2
+ ion. This structural arrangement is nearly iden-
2
+-REGULATED INTRACELLULAR PROCESSES 137
tical with that first observed in the x-ray structure of carp parvalbumin, and is
colloquially termed "the EF-hand." 86 This structural unit is also observed in
all available x-ray structures of proteins of the calmodulin superfamily (see Sec-
tions V.B and V.C). The Ca
2
+ ligands are all oxygen atoms, located approxi-
mately at the vertices of a pentagonal bipyramid.
The binding of Ca
2
+ and other cations to CaM has been extensively inves-
tigated.
87
The first two Ca
2
+ ions are bound in a cooperative manner, with an
average binding constant of about 2 X 10
5
M -I in 150 mM KCI and
1 mM Mg
2
+. The third and fourth Ca
2
+ ions are bound with binding constants
of about 3 X 10
4
M-I under the same conditions. Spectroscopic evidence has
shown that the first two Ca
2
+ -ions are bound in the C-terminal domain. Mg
2
+
has been shown to bind primarily to the N-terminal domain (see Table 3.2).88
The rates of dissociation of Ca
2
+ from the (Ca
2
+)4 CaM complex have
been studied by both stopped-flow and NMR techniques. 89,90 Fast and slow pro-
cesses are observed, both corresponding to the release of two Ca2 + ions. At an
ionic strength I = 0.1 and 25C, the rates for the two processes differ by a
factor of 30 (see Table 3.4).
A body of biophysical measurements, mostly made before the advent of x-
ray structures, indicated that CaM is constructed from two largely independent
domains.
87
This conclusion emanated from studies of the two tryptic fragments,
TR1C and TR
2
C. The major site of cleavage is between Lys-77 and Asp-78 of
the central helix, and results in N-terminal and C-terminal fragments of nearly
equal size. To a good approximation, the biophysical properties of the intact
CaM molecule-NMR, UV and CD spectra, kinetic properties, thermochemical
data, etc.-are the sum of the same properties of the fragments TR1C and TR
2
C.
This means that we may assign the slow dissociation process, k ~ f f , to the C-
terminal domain, and the fast, k;ff' to the N-terminal domain of CaM. Combin-
ing binding constants and off-rates, we may calculate that the rates of Ca 2+
binding to CaM are on the order of 107 M -I S - 1 at high ionic strength, and
Table 3.4
Rates of Ca2+ -dissociation and -association of some enzymes and
proenzymes.
Macrobicyclic amino cryptate [2.2.2]
Phospholipase A2
sTroponin C: Ca
2
+ sites
Ca
2
+ _Mg
2
+ sites
Trypsin
Trypsinogen
Chymotrypsin
Chymotrypsinogen
Calmodulin: N-terminal
Calmodulin: C-terminal
0.3
1.1 x 10
3
300
5
3
,,; 10
70
350
300-500
10-20
1.1 X 10
5
6 X 10
4
~ 10
6
2.8 X 10
5
10
7
10
8
M -I S -lor higher at low ionic strength. Recently the x-ray structure of the
C-terminal fragment TR
2
C was solved, and indeed showed a structure nearly
identical with C-terminal domain of intact CaM.
91
The structural changes occurring in CaM as Ca
2
+ ions are bound are asso-
ciated with pronounced changes in IH NMR, UV, fluorescence, and CD spec-
tra.
87
The observed changes in CD and fluorescence spectra in the presence of
Mg
2
+ are only about 20 to 25 percent of those induced by Ca 2+ . A comparison
of the CD spectra of CaM and its tryptic fragments indicates that the structural
changes induced by Ca2+ are substantially greater in the C-terminal than in the
N-terminal half.
92
By and large, few structural details of the conformation changes
have as yet been obtained. However, one aspect of the Ca
2
+ -induced confor-
mation change is that hydrophobic sites, probably one on each domain of the
molecule, become exposed. In the presence of excess Ca
2
+, CaM will bind to
other hydrophobic molecules, e.g., phenyl-Sepharose, a variety of drugs, many
small peptides, and-last but not least-its target proteins. This brings us to the
question of how CaM recognizes and interacts with the latter. We may suspect
that the hydrophobic sites on each domain are somehow involved, but the role
played by the central helix is still not clear. To explain small-angle x-ray scat-
tering data, the interconnecting helix needs to be kinked, bringing the intact
globular domains closer. 93
A putative CaM-binding segment (27 amino acids long) of myosin light-
chain kinase (MLCK), an enzyme activated by CaM, has been identified.
94
The
interaction between the segment peptide ("M13") and CaM has been studied95
by CD spectroscopy and IH NMR. From these studies it appears that a unique
1: 1 complex is formed, and that secondary and tertiary structural changes occur
not only in the peptide M13 but also in both halves of the CaM molecule.
Further NMR studies 96,97 of the interaction between CaM and naturally occur-
ring peptides (mellitin and mastoparan) that share some structural features of
M13---clusters of basic residues, hydrophobic residues adjacent to the basic res-
idues, and a predicted high a-helical content-show very much the same re-
sults. Based on these results, a model, shown in Figure 3.18, for the interaction
between CaM and M13 has been proposed. In this model the central helix is
kinked at position 81, allowing the two domains to wrap around the assumed
a-helical M13. Preliminary structure calculations of calcium-loaded CaM, based
on NMR data, indicate that the central helix in solution indeed is kinked and
very flexible,99 and comparisons 100 of chemical shifts in calmodulin with and
without M13 complexed supports the model in Figure 3.18. Recent structural
studies using NMR spectroscopy and x-ray diffraction have essentially con-
firmed the general features of this model, although the orientation of the peptide
is found to be reversed. 173
In conclusion, two important features of the protein should be recognized.
(i) The binding of Ca
2
+ to CaM (and to its complex with the target protein)
is quite likely cooperative, meaning that the switch from inactive to active con-
formation may occur over a much more narrow Ca 2 + -concentration interval
than otherwise.
K-148
139
K-1
Figure 3.18
A model for the interaction between calmodulin (CaM) and the assumed a-helical ( = - 57,
l/J = - 47) peptide M13. To produce this model, the backbone dihedral angles of Ser-81 in the
central a-helix of CaM have been changed (to = - 54, l/J = + 98), allowing the hydropho-
bic patches of both globular domains (green in Figure 3.17) of CaM to interact with the peptide
simultaneously. Figure kindly provided by R. Kretsinger; see also Reference 98.
140
G
P + CaM + 4Ca
2
+
--: ~ J ~ ___p. CaM + 4Ca
2
+
t.G,
P + CaM(Ca
2
+)4 - - - -1-
t.G
II
,
- ------......-
P CaM(Ca
2
+)4
Figure 3.19
Scheme depicting the standard free energies of different states in a system consisting of Ca 2 + ,
calmodulin (CaM), and a target protein (P). p. CaM denotes a complex between calcium-free
CaM and P, PCaM(Ca
2
+)4 denotes a complex with Ca
2
+ -loaded CaM. If the affinity of the
Ca2+-loaded CaM with the target protein P is higher than that of the Ca
2
+ -free form-
i.e., I ~ G I I I I > I ~ G I I I - i t follows that the Ca2+ affinity of the complex p. CaM is higher than
that of CaM itself.
(ii) The effective Ca2+ affinity will be different in the presence of the target
proteins. To illustrate this second point, consider the standard free energies in
the minimum scheme depicted in Figure 3.19. If the affinity of the Ca
2
+-cal-
modulin complex (CaM(Ca)4) for the target protein (P) is greater than that of
Ca 2+ -free calmodulin (CaM)-i.e., ILlGml > ILlGnl-it follows that the Ca 2+
affinity of the complex between P and CaM (p. CaM) must be higher than in
CaM itself. This effect is also found experimentally in model systems. 101
B. Troponin C
The contraction of striated muscle is triggered by Ca2+ ions. Muscle cells are
highly specialized, and contain two types of filaments that may slide past each
other in an energy-consuming process. One of the filaments, the thin filament,
is built up by actin molecules (M
r
= 42 kDa) polymerized end-to-end in a dou-
ble helix. In the grooves of this helix runs a long rod-like molecule, tropomyo-
sin; and located on this molecule at every seventh actin, is a complex of three
proteins, troponin. The three proteins in the troponin complex are troponin I
(TnI) , troponin T (TnT), and troponin C (TnC). A schematic picture of the
organization of the thin filament is shown in Figure 3.20.
Troponin C is the Ca 2+ -binding subunit of troponin, and it is structurally
highly homologous to calmodulin. Skeletal-muscle troponin C (sTnC; M
r
= 18
kDa) can bind four Ca 2+ ions, but cardiac-muscle troponin C (cTnC) has one
of the four calcium sites modified, so that it binds only three Ca
2
+ ions. The
troponin complex
TnT
actin
I
tropomyosin
141
+Ca
2
+
TnT
Figure 3.20
Schematic diagram of the organization of skeletal muscle thin filament, showing the position of
tropo-myosin and the troponin complex on the actin filament. The binding of Ca 2+ to TnC, the
calcium-binding subunit of the troponin complex, removes TnI, the inhibitory subunit, from
actin and thus permits an interaction with a specialized protein, myosin, on neighboring thick
muscle filaments (not shown). An ATP-driven conformation change in the myosin head group
makes the thick and thin filaments move relative to one another, so that muscle contraction
occurs.
x-ray structures of sTnC from turkey and chicken skeletal muscle have been
determined to resolutions of 2.8 and 3.0 A, respectively. 102,103 The structure of
turkey sTnC is shown in Figure 3.21. The similarity between the structures of
CaM (Figure 3.17) and sTnC is obvious. In sTnC we again find two domains,
each with two potential Ca2 + sites, separated by a 9-turn a-helix. The crystals
were grown in the presence of Ca2 + at a low pH (pH = 5), and only two Ca2 +
ions are found in the C-terminal domain. The two Ca 2+ -binding sites in this
domain have the same helix-loop-helix motif that is found in CaM, and they
both conform to the archetypal EF-hand structure. The interhelix angles between
helices E and F and between G and H are close to 110
0
By contrast, the helices
in the N-terminal domain, where no Ca
2
+ ions are bound, are closer to being
142
c
III
Figure 3.21
A ribbon backbone representation of the three-dimensional structure of turkey-
skeletal-muscle troponin C according to Herzberg and James. 102 The crystals
were grown at pH 5 in the presence of excess Ca2+, and at this low pH only
Ca2+ ions bound to the high-affinity domain (the C-terminal domain) are
observed. Note the high structural homology with calmodulin (Figure 3.17).
antiparallel, with interhelix angles of 133 (helices A and B) and 151 (helices
C and D).
Both sTnC and cTnC have two high-affinity Ca2+ -binding sites (see Table
3.2) that also bind Mg
2
+ ions competitively, although with a much lower affin-
ity. These two sites are usually called "the Ca 2 + -Mg2 + sites." 76,104 In sTnC
there are also two (in cTnC, only one) Ca2+ -binding sites of lower affinity
(Kjia
2
+ = 105 M - I) that bind Mg2 + weakly and therefore have been called "the
Ca 2 + -specific sites." Since Ca2 + binding to the latter sites is assumed to be
2
< -REGULATED INTRACELLULAR PROCESSES 143
the crucial step in the contractile event, they are often referred to as "the reg-
ulatory sites" (see below). The existence of additional weak Mg
2
+ sites
(K
B
= 300 M -I) on sTnC, not in direct competition with Ca 2+, has also been
inferred. 76,104, 105 Spectroscopic studies have shown that the two strong Ca 2+ -
Mg
2
+ sites are located in the C-terminal domain, and the weaker Ca
2
+-specific
sites in the N-terminal domain of sTnC.
106
This pattern is similar to that ob-
served with CaM. NMR spectroscopic studies strongly suggest that binding of
Ca
2
+ to both sTnC and cTnC is cooperative. 107 In sTnC the C-terminal domain
binds Mg
2
+ much more strongly than the N-terminal domain, by contrast to
CaM, where the reverse is true.
The rates of dissociation of Ca2+ and Mg 2+ from sTnC have been mea-
sured by both stopped-flow and 43Ca NMR techniques.
76
,108 As with CaM, the
actual numbers depend on the solution conditions, ionic strength, presence of
Mg
2
+, etc. (see Table 3.4). On the rate of Mg
2
+ dissociation from the Ca
2
+-
Mg
2
+ sites, quite different results have been obtained by stopped-flow studies76
of fluorescence-labeled sTnC (k!;lt+ = 8 s - I) and by 25Mg NMR
(k!;lt+ = 800-1000 S -1).109 This apparent discrepancy seems to have been re-
solved by the observation that both binding and release of Mg
2
+ ions to the
Ca 2+ _Mg
2
+ sites occur stepwise, with k!;lt+ < 20 s -I for one of the ions, and
k!;lt+ ?:: 800 s - I for the other. 110 The rates of dissociation of the Mg 2+ ions
are important, since under physiological conditions the Ca 2+-Mg 2+ sites of
sTnC are likely to be predominantly occupied by Mg
2
+ ions, release of which
determines the rate at which Ca2+ can enter into these sites.
Spectroscopic and biochemical data III indicate that upon binding Ca2+, sTnC
and cTnC undergo significant conformation changes. Comparisons of NMR
spectroscopic changes on Ca2+ binding to intact sTnC, as well as to the two
fragments produced by tryptic cleavage (essentially the N-terminal and C-ter-
minal halves of the molecule, just as was the case with CaM), have shown that
the conformation changes induced are mainly localized within the domain that
is binding added ions.
llO
,112 Thus the central a-helix connecting the domains
seems unable to propagate structural changes from one domain to the other. It
has been suggested that the structural differences found in the x-ray structure of
turkey sTnC between the C-terminal domain, which in the crystal contains two
bound Ca
2
+ ions, and the N-terminal domain, in which no Ca
2
+ ions were
found, may represent these conformational changes.
113
This rather substantial
conformational change is schematically depicted in Figure 3.22. However, pre-
liminary structure calculations 114 of the calcium-saturated and calcium-free forms
of calbindin D
9k
indicate that much more subtle conformational changes take
place upon binding Ca
2
+ in calbindin D
9k
. Interestingly, IH NMR spectroscopy
has provided evidence for the concept that the structural change induced by
Mg
2
+ binding to the C-terminal domain of sTnC must be very similar to that
induced by Ca2+ ions. Another result obtained by 113Cd NMR studies 108 is that
the cadmium-loaded N-terminal domain of sTnC in solution undergoes a rapid
interchange between two or more conformations, with an exchange rate on the
order of 103_10
4
S - I.
144
(A) (8)
Figure 3.22
Diagrammatic representation of the proposed conformational changes in the N-terminal domain
of troponin C upon Ca
2
+ binding. 113 Helices are depicted as cylinders, and I, II denote the
Ca 2+ -binding sites. Helices N, A, and D retain their relative positions, and the relative disposi-
tion of helices Band C are also kept constant. (A) Ca 2+ -free conformation as determined by x-
ray crystallography. 102 (B) Proposed Ca 2 +-saturated conformation based on the structure of the
highly sequence-homologous Ca 2+-loaded C-terminal. Figure kindly provided by N. C. J. Stry-
nadka and M. N. G. James.
Just as CaM exerts its biological function in complexes with other proteins,
TnC participates in the three-protein troponin complex. It presently appears that
TnC and Tnl form a primary complex that is anchored by TnT to a binding site
on tropomyosin. 115 In the troponin complex the Ca2+ affinity is increased by a
factor of about ten over that in isolated sTnC, both at the Ca
2
+ _Mg
2
+ sites and
at the Ca
2
+-specific sites. A similar increase in affinity is found for Mg
2
+.
Given the amounts of "free" Mg
2
+ inside muscle cells (1 to 3 mM), it seems
likely that the Ca
2
+_Mg
2
+ sites in the resting state of troponin are filled with
Mg2 +, so that a transitory release of Ca 2 + leads primarily to rapid Ca2 + bind-
ing to the Ca2 +-specific sites, and subsequently to conformation change and
contraction.
C. Parvalbumin and Calbindins D
9K
and D
28K
A few intracellular Ca2 +-binding proteins have been discovered that by se-
quence homology clearly belong to the CaM-TnC family with Ca
2
+ sites of the
"EF-hand"-type, but that do not appear to exert a direct regulatory function.
Parvalbumins (M
r
= 12 kDa), calbindin D
9K
(M
r
= 8.7 kDa) and calbindin D
28K
/
D
/
,\
E
\
145
Figure 3.23
Structure of the Ca
2
+ -binding sites ofcarp parvalbumin. The Ca
2
+ ions are depicted as regular
octahedra making six ligand contacts with oxygen atoms at each vertex, labeled x, y, z, - x,
- y, - z. The helix-loop-helix structure that forms a Ca2+ -binding site can be regarded as a
hand with the forefinger representing one helix (e.g., the E-helix) in the plane of the figure, the
thumb oriented perpendicular to the plane representing the second helix (the F-helix), and the
remaining fingers make up the Ca2+ -binding loop. After Kretsinger and Barry. 118
(M
r
= 28 kDa) belong to this group. Parvalbumin(s) exist in two main types,
a and {3, found in large quantities in the white muscle of fish, amphibia, and
reptiles, but also in different mammalian tissues, 116,117 including neurons of the
central and peripheral nervous system. The molecule has two fairly strong Ca2 + -
binding sites (see Table 3.2). The x-ray structure of carp parvalbumin was solved
in 1973 by Kretsinger et al., 118 and for a decade provided the basis for all
discussions on intracellular Ca2 +-binding proteins. The concept of the canonical
"EF-hand" Ca2+ -binding site originated from the parvalbumin work, and the
name "EF" derives from the labeling of the two helices that flank the second
of the two Ca2 + sites in parvalbumin, as shown in Figure 3.23. If the first Ca2 +
x y z -y -x -z
En n n n G I En.",. n n n
Figure 3.24
One consensus EF-hand sequence including residues in the flanking a-helices; x, y, z, -x, -y,
- z denote positions in the octahedral Ca2 + coordination sphere. E-glutamic acid residue, G-
glycine residues, I-isoleucine residue, n-nonpolar residue, -a residue with a nonaromatic
oxygen-containing side chain (i.e., Glu, Gin, Asp, Asn, Ser, or Thr), and -nonspecific resi-
due.
ligand in the approximately octahedral coordination sphere is given number 1
(or "x") the others come in the order 3( "y"), 5( "z"), 7(" - y"), 9(" - x"),
and 12(" - z"). In the second site of parvalbumin, "-x" is actually a H
2
0
molecule, but in the first site it is the carboxylate of a Glu. Studies 118 of puta-
tive Ca 2+ -binding sites in other proteins with known primary sequences led to
the generalized EF-hand structure-including residues in the flanking a-hel-
ices-shown in Figure 3.24. This sequence, with minor modifications, has been
widely used in searching for "EF-hands" in libraries of amino-acid (or DNA)
sequences of new proteins with unknown properties. In this way, calbindin D
28b
a protein with unknown function, initially discovered in chicken intestine, but
later found also in brain, testes, and other tissue, has been shown to have four
EF-hand sites.
ll9
Recently two structures of carp parvalbumin, both with a resolution of 1.6 A,
were published. 120 One of these structures is the native calcium-loaded form of
the protein; the second is the structure of parvalbumin in which Ca2+ has been
replaced by Cd2+. No significant differences are observed upon replacement of
calcium by cadmium. 113Cd has a nuclear spin of I = i, making it much more
amenable to NMR studies than the quadrupolar 43Ca (I = i), This study sup-
ports the use of 113Cd NMR as a tool for the study of calcium-binding pro-
teins.
121
The function of parvalbumin has long been assumed to be that of buffering
Ca
2
+ in muscle cells, i.e., taking up Ca
2
+ ions released from Ca
2
+-troponin
complexes, thereby ensuring that the cytoplasmic levels of free Ca2+ are always
kept very low, even during short bursts of muscle activity,122 The widespread
occurrence of parvalbumin in non-muscle tissue indicates that it probably has
other roles as well.
Calbindin D
9k
(M
r
= 8.7 kDa) is another intracellular Ca
2
+-binding protein
with unknown function. It was briefly mentioned in connection with Ca 2+ up-
take and transport in the intestine and placenta (Section IV.A). Like the avian
calbindin D
28b
the D
9k
calbindin has been observed in many types of tissue.
The homology between the D
9k
and D
28k
calbindins is much less than the name
suggests; both their syntheses are, however, regulated by vitamin D. The x-ray
structure of bovine calbindin D
9k
has been determined 123 and refined to a reso-
lution of 2.3 A, and a three-dimensional solution structure of porcine calbindin
2
+-REGULATED INTRACELLULAR PROCESSES 147
D
9k
is also available.
124
The average solution structure calculated from NMR
data is shown in Figure 3.25 (See color plate section, page C-lO.)
The protein has four main a-helices and two Ca
2
+-binding loops (I and II).
The interior of the molecule shows a loose clustering of several hydrophobic
side chains; in particular, three phenylalanine rings come ,very close in space.
The Ca
2
+ -binding loops constitute the least-mobile parts of the molecule. The
crystallographic temperature factors have pronounced minima in these regions,
with the lowest overall B-factor observed in loop II. Both Ca2+ ions are roughly
octahedrally coordinated with protein oxygen atoms. There are some striking
differences between the two sites, however. Whereas the C-terminal site (II) has
a general structure very similar to the archetypal "EF-hand," as observed in
CaM, sTnC, and parvalbumin, the N-terminal site (I) has an extra amino-acid
residue inserted between vertices x and y, and z and - y (see Figure 3.24). As
a consequence, the peptide fold in site I is different from that in site II. Three
carboxylate groups are ligands in site II, but in site I there is only one.
Despite this marked difference in charge and peptide fold, the Ca
2
+ affinity
of both Ca 2 + sites is remarkably similar, as has been shown in a study in which
site-directed mutagenesis was combined with different biophysical measure-
mentsY Cooperative Ca
2
+ binding in the native calbindin D
9k
(the "wild type")
was first demonstrated at low ionic strength by means of the values of the two
stoichiometric Ca
2
+-binding constants, K
1
and K
2
, which could be measured
with good accuracy (K
1
= 4.4 X 10
8
M -1 and K
2
= 7.4 X 10
8
M -I). The
effects of amino-acid substitutions in Ca 2+ site I were primarily localized to
this site, with virtually no effects on the structure or other biophysical properties
pertinent to site II. The appearance of sequential Ca2 + binding in some of the
calbindin mutants did allow the identification of 1H NMR resonances that re-
spond primarily to binding of Ca 2+ to either one of the sites. This result in tum
permitted an estimate of the ratio between the site-binding constants (K
A
and
K
B
) in the wild-type protein and in one of the mutant proteins (Tyr-13 ~ Phe).
In this way the reseachers 125 could assess, to within narrow limits, the free
energy of interaction, LlLlG, between the two Ca2 + sites as 7.7 kJ/mol at low
ionic strength and 4.6 kJ/mol in the presence of 0.15 M KCl. How this site-site
interaction is transmitted on a molecular level is still unknown.
Through a combination of site-specific mutations and biophysical measure-
ments, it has recently been demonstrated that carboxylate groups at the surface
of the protein, but not directly ligated to the bound Ca 2+ ions, have a profound
effect on the Ca2 + affinity. 126 Neutralization of the surface charges reduces
affinity and increases the stability of the protein toward unfolding by urea. 127
A surprising discovery about the structure of bovine calbindin D
9k
in solu-
tion has also been made recently. 128 Detailed analysis of the 2D IH NMR spec-
trum of wild-type calbindin has revealed that it exists as a 3: 1 equilibrium
mixture of two forms, corresponding to a trans and cis conformation around the
Gly-42-Pro-43 peptide bond. The global fold appears essentially the same in the
two forms, and structural differences are primarily located in the inter-domain
loop in which Pro-43 is located.
D. Sarcoplasmic Calcium-Binding Protein from Nereis diversicolor
The calmodulin superfamily of proteins also includes' sarcoplasmic Ca
2
+ -bind-
ing proteins (SCPs) that can be found in both vertebrate and invertebrate mus-
cle.
129
The function of SCPs is not yet known, but their sequence h011;lology
with Ca2+ -binding proteins of known tertiary structure suggests that they orig-
inally contained four helix-loop-helix Ca 2 +-binding domains. Ca2 + binding has
been preserved in the first and third domains of all known SCPs, but only one,
if any, of domains II and IV is functional. The three-dimensional crystal struc-
ture of an SCP from the sandworm Nereis diversicolor analyzed at 3.0 A
resolution 130 can be seen in Figure 3.26. (See color plate section, page C-ll.)
The C-terminal half (domains III and IV) of the molecule contains two Ca 2 + -
binding EF-hands (green and red in Figure 3.26) similar to calbindin D
9k
and
the globular domains of troponin C and calmodulin. The N-terminal half is, on
the contrary, markedly different from the normal helix-loop-helix geometry. Do-
main I binds Ca
2
+ with a novel helix-loop-helix conformation, whereas domain
H lacks calcium-binding capacity. The two halves are packed closely together,
and are not, as in troponin C or calmodulin, connected by a solvent-exposed
a-helix.
E. Membrane Cytoskeleton and Phospholipid Binding Proteins
It has long been suspected that Ca 2+ ions are somehow involved in exocytosis.
Recently several groups 131 have isolated intracellular proteins that associate with
membranes, and/or membrane cytoskeleton proteins, in a Ca
2
+-dependent man-
ner, and that seem able to mediate vesicle fusion or aggregation at Ca2 + con-
centrations above 200 ILM. These proteins--endonexin, calelectrin, p36, and
pH-have stretches of consensus amino-acid sequences that are also found in a
phospholipase A
2
inhibitor protein, lipocortin.
132
It appears that further studies
of this new class of proteins, known as annexins, will lead to new insights into
cell-signaling pathways. Multiple functions have been proposed for the annex-
ins, but no cellular role has yet been defined. 133. The first crystal structure of
an annexin, human annexin V-which in vitro will form voltage-gated Ca
2
+
channels-has been determined recently. 172 In annexin, the three Ca
2
+-binding
sites are located on the side of the molecule that is involved in membrane bind-
mg.
F. Ca
2
+-Dependent Proteases
An interesting Ca 2+ -activated intracellular protease, sometimes called calpain,
was discovered during the last decade. 134 The ending -pain refers to its relation
with other proteolytic enzymes like papain. It may seem dangerous to have a
proteolytic enzyme loose inside a cell, and it must have rather specialized func-
tions and be under strict control. The complete primary structure of the calcium
protease (My = 80,000) in chicken tissues has recently been deduced from the
nucleotide sequence of cloned DNA. 135,136 The findings are quite unexpected.
2
+ -REGULATED INTRACELLULAR PROCESSES 149
The protein contains four distinct domains. The first and third domains have no
clear sequence homologies with known protein sequences, but the second do-
main has a high homology with the proteolytic enzyme papain, and the fourth
domain is highly homologous to calmodulin. This fourth domain thus has four
EF-hand-type Ca2 +-binding sites, although the third site has a somewhat un-
usual loqp sequence. Here we apparently are faced with an unusual invention
by Nature: by fusing the gene for a protease with that of the canonical Ca
2
+
receptor, she has created a molecule in which a regulatory protein is covalently
linked to its target enzyme!
G. Protein Kinase C
Before we leave our brief survey of intracellular Ca 2+ -binding proteins, we
must write a few lines about an important Ca2+ -regulated kinase (a phospho-
rylating enzyme), i.e., protein kinase C (PKC). The activity of this enzyme, or
rather family of enzymes, 137 appears to be regulated by three factors: phospho-
lipids, in particular phosphatidylserine; diacyl-glycerols, one of the products of
inositol lipid breakdown; and Ca 2+ ions. The high-activity form of PKC, which
appears responsible for much of the phosphorylation activity of many cells, is
presumably membrane-bound, whereas the low-activity form may be partly cy-
tosolic (Figure 3.27). The schematic structure of rabbit PKC (M
r
= 77 kDa)
inactive
forms of
protein kinase C
+ Ca
2
+ ll-
ca2
+
active form of
protein kinase C
phosphorylated
protein
Figure 3.27
Outline of the cellular events that result in the activation of protein kinase C (PKC). The en-
zyme apparently exists in at least two states. Recent sequence work indicates that it has a Ca2+_
binding site of the EF-hand type. When no Ca2+ ion is bound, and when the "concentration"
of diacylglycerol (DG) in the inner layer of the plasma membrane is low, the kinase exists in a
low-activity form, possibly dissociated from the membrane. When a hormone binds to a plasma-
membrane receptor (R), cleavage of phosphoinositol into 1,4,5-IP
3
and DG is induced. The lat-
ter lipid may bind to and activate the calcium-loaded form of PKC. The active form of protein
kinase C will now phosphorylate other cytoplasmic proteins, and in this way modify their bio-
chemical properties. R = receptor; PL-C = phospholipase C; G = a GTP-binding protein that
is assumed to act as an intermediary between the receptor and the membrane bound PL-C.
150
protein kinase
1 ~ .. r_eg=-U_la_to_rY=--do_m_a_in - + - ~ 1 1 ~ ..__d_o_m_a_in - - - - - j ~ ~ 1
ATP binding
site
catalytic site &
target recognition(?)
Figure 3.28
Schematic representation of the structure of rabbit protein kinase C. 138 Three highly homologous
protein kinases C were actually identified with M
r
= 76,800. The kinase region shows clear
similarity with other kinases. The regulatory domain should contain binding sites for Ca
2
+,
phosphatidyl serine (PS), and diacylglycerol (DG).
according to Ohno et al.
138
is shown in Figure 3.28. The Ca
2
+ site(s) are pre-
sumably in the regulatory domain. No typical "EF-hand" pattern has been found
in the amino-acid sequence. A protein kinase that requires Ca2+ but not phos-
pholipids nor calmodulin for activity has been purified from soybean. From the
amino-acid sequence the protein appears to have a calmodulin-like Ca2+ -bind-
ing domain, very much as in calpain.
139
H. Summary
Many different biological processes in eukaryotic cells are regulated by intra-
cellular Ca2 + concentration levels. Examples of such processes are muscle con-
traction, transport processes, cell division and growth, enzyme activities, and
metabolic processes. A link in this regulatory chain is a number of intracellular
Ca2 + receptors with Ca 2 +affinities such that their binding sites are largely un-
occupied at resting Ca 2 + concentration levels, but are occupied at Ca2 + levels
reached as a result of some external stimulus. This class of Ca 2+ receptors is
often called the "calmodulin superfamily" and includes the well-known mem-
bers troponin C (regulating muscle contraction in striated muscle) and calmo-
dulin (playing an important role in the regulation of many cellular processes).
Amino-acid sequence determinations as well as x-ray and 2D IH NMR studies
have revealed a strong homology between the regulatory Ca2+ -binding proteins.
The Ca
2
+-binding sites are located in a loop flanked by two helices, and the
Ca2 + ions are ligated with approximately octahedral or pentagonal bipyramidal
symmetry. The ligands are six or seven oxygen atoms that are furnished by side-
chain carboxylate or hydroxyl groups, backbone carbonyls, and water mole-
cules. Pairs of these Ca 2+ sites, rather than individual sites, appear to be the
functional unit, and a common consequence of their arrangement is cooperative
Ca
2
+ binding. Ca
2
+ binding to the intracellular receptor proteins is accompa-
VI. EXTRACELLULAR Ca
2
+ -BINDING PROTEINS 151
nied by structural changes that expose hydrophobic patches on their surfaces,
thereby enabling them to bind to their target proteins.
VI. EXTRACELLULAR Ca
2
+-BINDING PROTEINS
The Ca2 + concentration in extracellular fluids is usually orders of magnitude
higher than intracellular concentrations. In mammalian body fluids, the "free"
Ca
2
+ concentration is estimated to be 1.25 mM (total Ca
2
+ is ~ 2 . 4 5 mM) with
only minor variations. 140 We would thus expect that Ca2 + ions in extracellular
fluids playa very different role from that inside cells. To ensure Ca 2+ binding
the macromolecular binding sites need have only a modest Ca2 + affinity
(Kfi
a2
+ = 103 to 10
4
M - 1), and since extracellular Ca2 + does not seem to have
a signaling function, the rates of Ca 2+ association or dissociation in protein-
binding sites need not be very high.
One particularly important aspect of Ca 2 + in mammals is its role in the
blood coagulation system. Here we will meet a new type of amino acid, y-
carboxyglutamic acid ("Gla" )-see Figure 3.29, that seems to have been de-
COOH COOH
" / CH
I
CH
2
(i)NH - 6H - COo
e
3
(Gla)
COOH
I
HC-OH
(i)NH - 6H - COo
e
3
(Hya)
Figure 3.29
Chemical structures of two novel amino acids believed to bind
calcium in, e.g., blood-clotting proteins.
signed by Nature as a Ca
2
+ ligand with rather special functions. Gla-containing
proteins are also encountered in some mineralized tissues. The formation of
bone, teeth, and other calcified hard structures is an intriguingly complicated
phenomenon that will be dealt with in Section VII. We start, however, with a
brief discussion of the role of Ca2 + in some extracellular enzymes.
A. Ca
2
+-Binding in Some Extracellular Enzymes
Several extracellular enzymes have one or more Ca2+ ions as integral parts of
their structure. In a very few of them the Ca 2+ ion is bound at or near the
active cleft, and appears necessary for maintaining the catalytic activity (phos-
pholipase A
2
, a-amylase, nucleases), whereas other enzymes show catalytic ac-
tivity even in the absence of Ca2+ (trypsin and other serine proteases). In the
latter proteins, the Ca
2
+ ion is usually ascribed a "structural" role, although
its function may be rather more related to "dynamics" and so be more subtle
and complex.
Trypsin has one Ca2+-binding site with four ligands (two side-chain and
two backbone oxygens) donated by the protein (Glu-70, Asn-72, Val-75, and
Glu-80) and two ligating water molecules, making the site roughly octahed-
ral. 141 The binding constant of Ca2+ to trypsin and its inactive precursor
"proenzyme," trypsinogen, has been measured (see Table 3.2). The binding
constant is slightly smaller for the precursor, as is also true for chymotrypsin
and chymotrypsinogen. 142 The Ca 2+ affinities of the serine proteases and their
proenzymes are such that their Ca2+ sites will be largely occupied in extracel-
lular fluids, but would be unoccupied inside a cell. It has been suggested that
this phenomenon constitutes a safeguard against unwanted conversion of the
proenzymes into the active enzymes as long as they still are inside the cells
where they are synthesized.
The rates of Ca2+ dissociation of the above enzymes and proenzymes have
been measured by 43Ca NMR and stopped-flow techniques,142 and are collected
in Table 3.4. We note that the values of k
on
and k
off
are generally much smaller
than in the intracellular regulatory EF-hand proteins discussed in Section VI.
Whereas the latter have dynamic and equilibrium properties similar to those of
flexible low-molecular- weight chelators such as EDTA and EGTA, the serine
proteases are more similar to the more-rigid cryptates, such as the macrobicyclic
amino cryptate [2.2.2] (see Tables 3.2 and 3.4).
As mentioned above, there are a few enzymes in which a Ca 2+ ion is pre-
sent in the active cleft and essential for activity. Pancreatic phospholipase A
2
(M
r
= 14 kDa) is an enzyme of this type. The x-ray structure is known to high
resolution, and a single Ca2+ ion is found to be surrounded by six ligands, four
presented by the protein (Tyr-28, Glu-30, Glu-32, and Asp-49) and two water
molecules.
143
A mechanism for the action of phospholipase A
2
has been
proposed 144 and is shown in Figure 3.30. This mechanism is based on three
high-resolution x-ray crystal structures of phospholipase A
2
with and without
transition-state analogues bound. The binding constant for Ca 2+ together with
the rate of dissociation found from variable-temperature 43Ca NMR studies 145
can be used to calculate k
on
= 4 X 10
6
M-I S -I, again lower than in EF-hand
proteins. Recent IH NMR studies indicate that the global structure of the lipase
is very much the same in the Ca
2
+-free and the Ca
2
+-bound forms. Structural
changes upon Ca2+ binding appear primarily located in the region of the bind-
ing site. 112,146
The mammary glands produce, among other substances, a Ca 2+ -binding
enzyme activator, a-lactalbumin, that has about 40 percent sequence identity
with lysozyme. This protein , which is involved in the conversion of glucose
into lactose, is secreted in large quantities, and in human milk constitutes some
15 percent of total protein. The Ca2+-binding constant of bovine or human a-
lactalbumin is on the order of 10
7
M-I under physiological conditions. In ad-
dition to Ca2+, the enzyme also binds Zn 2+. It appears that Ca2+-ion binding
affects enzymatic activity, and somehow controls the secretion process, but the
biological role of metal-ion binding to a-lactalbumin needs to be studied further.
The x-ray structure of a-lactalbumin from baboon milk (M
r
= 15 kDa) has been
(A)
(B)
153
Figure 3.30
Catalytic mechanism144 of phospholipase A
z
. (A) Catalytic attack on substrate bound in a pro-
ductive mode. (B) The tetrahedral intermediate as it collapses into products. (C) Products
formed by "productive collapse" in which three water molecules move into the active site to
replace the products. Two of these water molecules will coordinate the calcium ion. Figure
kindly provided by P. B. Sigler.
determined 147 to a high resolution ( ~ 1 . 7 A). The Ca
2
+ -binding site has an in-
teresting structure. The ion is surrounded by seven oxygen ligands, three from
the carboxylate groups of aspartyl residues (82, 87, and 88), two carbonyl oxy-
gens (79 and 84), and two water molecules. The spatial arrangement is that of
a slightly distorted pentagonal bipyramid with the carbonyl oxygens at the ap-
ices, and the five ligands donated by the proteins are part of a tight "elbow"-
like tum. The a-lactalbumin site has a superficial structural similarity to an
"EF-hand," although the enzyme presumably has no evolutionary relationship
with the intracellular Ca2 +-binding regulatory proteins.
Blood clotting proceeds in a complicated cascade of linked events involving
many enzymes and proenzymes. About a decade ago it was shown that several
of these proteins contained a previously unknown amino acid, y-carboxyglu-
tamic acid (Gla), and more recently yet another new amino acid, f3-hydroxyas-
partie acid (Hya), has been discovered (see Figure 3.29). The former is formed
postribosomally by a vitamin-K-dependent process in the liver. 148 Presently the
most-studied Gla protein in the blood-clotting system is prothrombin (M
r
= 66
kDa). Ten Gla residues are clustered pairwise in the N-terminal region, essen-
tially lining one edge of the molecule, forming a highly negatively charged
region.
149
A small (48 residues) proteolytic fragment (Fl) that contains all ten
Gla amino acids can be prepared. Prothrombin can bind about 10 Ca 2+ ions,
but Fl binds only 7. Binding studies to Fl show that the Ca 2+ ions bind at
three high-affinity cooperative sites and four noninteracting sites, 150 and that this
binding takes places in conjunction with a spectroscopically detectable confor-
mational change (see Table 3.1).
In the presence of Ca 2 + ions, prothrombin and other vitamin-K-dependent
proteins in the blood-coagulation system will bind to cell membranes containing
acidic phospholipids, in particular, the platelet membrane, which is rich in phos-
phatidylserine. A proposed model for the prothrombin-membrane interaction is
shown in Figure 3.31.
It has long been known that calcium ions are involved in cell-to-cell and
cell-to-extracellular matrix interactions, but the molecular details largely remain
to be unraveled. In the late 1980s a large, adhesive, calcium-binding matrix
glycoprotein (M
r
~ 420 kDa) named thrombospondin was characterized. This
multifunctional adhesion molecule is composed of three polypeptide chains, each
with 38 amino-acid-Iong repeats that are homologous with the calcium-binding
helix-loop-helix sites of the calmodulin superfamily. 152 Each thrombospondin
molecule is reported to bind 12 calcium ions with an affinity of about 10
4
M -I,
and the removal of calcium is accompanied by a conformational change. 153,154
B. Summary
In higher organisms, the Ca 2+ concentration in extracellular fluids generally is
considerably higher than the intracellular concentrations. In mammalian body
fluids, the Ca 2+ concentration is typically on the order of a few mM. The
Figure 3.31
Proposed features of the interaction between the prothrombinase complex and a membrane lipid
bilayer. lSI K] and K
2
are the kringle domains of prothrombin, and EGFI and EGF2 are the two
epidermal growth factor units of factor Xa' Prothrombin and factor X
a
form a heterodimer com-
plex harbored within the membrane protein factor Va- The proposed interaction between pro-
thrombin and factor Xu involves hydrophobic interactions between two helices and bridging by a
Ca
2
+ ion between two Gla residues. The N-terminal Gla residues attach the heterodimer com-
plex to the phospholipid surface. Figure kindly provided by C. C. F. Blake.
extracellular concentration levels are highly regulated and undergo only minor
variations. A consequence of these high levels of Ca2 + in extracellular fluids is
that the binding constant need be only 10
3
to 10
4
M -1 in order for a protein
site to be highly occupied by Ca 2+. Several extracellular enzymes and enzyme
activators have one or more Ca2+ ions as integral parts of their structures. Some
Ca2 + ions are bound at, or near, the active cleft and may take part in the
enzymatic reactions (e.g., phospholipase A
2
, a-amylase). In other molecules,
for example, serine proteases like trypsin and chymotrypsin, the Ca 2+ ion is not
essential for enzymatic activity, and may play more of a structural role. Ca2+
ions are involved in the cascade of enzymatic events that results in blood clot-
ting in mammals. Several of the proteins in this system contain two new amino
acids, y-carboxyglutamic acid (Gla) and {3-hydroxyaspartic acid (Hya) , which
155
are strongly suspected to be involved as ligands in Ca
2
+ binding. In the pres-
ence of Ca2 + ions, prothrombin and other Gla-containing proteins will bind to
cell membranes containing acidic phospholipids, in particular, the platelet mem-
brane. It appears likely that Ca 2+ ions form a link between the protein and the
membrane surface.
VII. CALCIUM IN MINERALIZED TISSUES
The formation of calcified tissue-shells, bone, and teeth-is a very complex
process that is under strict regulatory control. Despite the obvious importance
of this field, relatively little research has been directed toward elucidation of the
underlying mechanisms, perhaps because the field spans a broad range of sub-
jects, from inorganic solution and solid-state chemistry to cellular physiology. 155
Historically, it was long held that formation of biological minerals such as
bone was simply the nucleation and growth of calcium hydroxyapatite within an
extracellular matrix of collagen. Many proteins other than collagen have now
been discovered in appreciable quantities in bone and other biological minerals.
It is also apparent that the pattern of calcification differs in shells, bone, teeth,
and other mineralized tissues; so it is not likely that there is only one underlying
mechanism. Considering the immensity of the subject, we will here only make
a few brief comments, mainly about bone and teeth.
As was briefly mentioned earlier in this chapter, the inorganic matter of
bone and teeth in many ways resembles apatite minerals (Cas(OH)(P0
4
h). Ta-
ble 3.5 summarizes inorganic solid components of other biominerals. A detailed
Table 3.5
A summary of the main inorganic solid component
8
of the most-common biominerals in
living systems.
148
Anion Formula Crystal form Occurrence Main function
Carbonate CaC0
3
Calcite Sea corals, molluscs, Exoskeleton; Ca-store; eye
Aragonite and many animals lens
Valerite and plants
Oxalate Ca(COOh . H
2
O Whewellite Insect eggs; verte- Deterrent; cytoskeleton; Ca
Ca(COO)2 2H20 Weddellite brate stones store
Phosphate (Ca)IO(P0
4
MOHh Hydroxyapatite Bones; teeth; shells; Skeletal; Ca storage; pres-
(unit cell comp.) intracellular in some sure-transducer (piezo-elec-
bacteria tric)
Sulfate CaS0
4
'H
2
O Gypsum Jelly fish; plants Gravity device; Sand Ca
store
a Most real biominerals are actually nonstoichiometric, and contain a number of additional cations (e. g., Mg 2 ') or anions
(e.g., F -). In addition, the inorganic phase may be interpenetrated by a biopolymer.
VII. CALCIUM IN MINERALIZED TISSUES 157
analysis 156 shows that, apart from Ca
2
+ and P0
4
3
-, many other cations and
. . b M 2 + N + K + S 2 + CO 2 - F - Cl - d
amons occur III one, e.g., g , a, ,r, 3, , , an
citrate. X-ray diffraction patterns and electron-microscope pictures of bone show
that the inorganic phase is made up of many very small and imperfect crystals.
By contrast, dental enamel is made up of much larger and uniform thin crystals.
Although the solubility product of calcium hydroxyapatite (see Section II) is
such that the equilibrium Ca2 + concentration should be in the low micromolar
range, bone mineral appears to be in equilibrium with much higher Ca2+ con-
centrations (0.8-1 .0 mM). 157 This discussion brings us to the question of how
the inorganic crystallites are formed. Obviously both Ca 2 + and P0
4
3
- ions
must be concentrated in cells or organelles bordering on the regions where mi-
neralization is to take place. Fresh layers of bone matrix are formed by a con-
tinuously replenished layer of cells called osteoblasts (Figure 3.32A), which, in
addition to apatite crystallites, also secrete collagen, and large specific proteins
called osteonectin, osteocalcin (a Gla protein), proteoglycans, and phosphopro-
teins. In tissues undergoing rapid mineral deposition, the crystallites appear to
be formed in vesicles that may have peeled off from the adjacent cell layers.
These vesicles seem able to concentrate calcium and phosphate in a manner not
well understood.
Bone, unlike diamond, is not forever. It can be remodeled and dissolved.
A serious medical problem, which affects some women after menopause, is
osteoporosis, i.e., the decalcification of bone. This loss of bone mass, which
occurs with increasing age, makes bones more susceptible to breaking under
stress. About 50 percent of American women, and 25 percent of American men,
over 45 years of age are affected by osteoporosis. 158 Whereas osteoblast cells
handle bone formation, another type of cells, osteoclasts, can erode it (Figure
3.32B). These macrophage-like cells can form deep tunnels in a bone matrix,
and the cavities left behind are rapidly invaded by other cells forming blood
vessels and new layers of osteoblasts. The modus operandi of osteoclast cells is
not well understood at present. They may secrete calcium-chelating organic an-
ions, such as citrate, to assist in the solubilization of the bones, as well as
extracellular proteases that degrade the organic part of the matrix.
Summary
Calcium is, along with iron, silicon, and the alkaline earth metals, an important
constituent of mineralized biological tissues. Some Ca2 + -based biominerals,
like bone or mother-of-pearl, can be regarded as complex composites with mi-
croscopic crystallites embedded in a protein matrix. The formation of calcified
biominerals is a highly regulated process, and human bone, for instance, is
constantly being dissolved and rebuilt. When the rates of these two counteract-
ing processes are not in balance, the result may be decalcification, or osteopo-
rosis, which seriously reduces the strength of the bone.
158
(A)
osteogenic cell
(osteoblast -------,-
precursor)
osteoblast
osteoid
(uncalcified
bone matrix)
calcified bone _
matrix
cell process
in canaliculus
osteocyte
(8)
quiescent
osteoblast
small
blood -----.!!llli!ii
vessel
endothelial
cell
old bone
osteocyte
osteoblast about to
lay down new bone
to fill in the excavated
tunnel
osteoclast excavating
tunnel through
old bone
newly deposited
bone matrix
loose
connective
tissue
inward
growing
capillary
sprout
VIII. Ca
2
+-BINDING PROTEINS IN MICROORGANISMS 159
Figure 3.32 (facing page)
Schematic diagram depicting the roles of the most important cell types in bone formation. (A)
The osteoblast cells line the bone surface and secrete the inorganic and organic components
(collagen, etc.) that will form new bone. Some osteoblast cells gradually become embedded in
their own secretion. A particular secreted bone-specific protein, osteonectin, forms strong links
between calcium hydroxyapatite and collagen. The bone-forming cells that become trapped in
the bone matrix are now called osteocytes. (B) The osteoclast cells function to remodel compact
bone. A group of cells acting together excavate a tunnel through old bone at a rate of about 50
fLm per day. Behind the advancing osteoclasts follow a contingent of osteoblasts that line the
wall of the tunnel and start to form new bone. Concurrently a capillary vessel is formed along
the center of the tunnel and provides the cells with nutrients. Eventually the tunnel will become
filled with concentric layers of new bone with only a narrow canal remaining. It is apparent that
bone is far from a dull inorganic deposit, and very much a site of continuous activity. It is
estimated that 5 to 10 percent of the bone in an adult mammal is replaced per year. Adapted
from Reference 159.
VIII. Ca
2
+-BINDING PROTEINS IN MICROORGANISMS:
THE SEARCH FOR A PROKARYOTIC CALMODULIN
Since Ca2 + ions evidently play an important role in regulating a variety of
cellular responses in animals and higher organisms, one may ask whether this
use of Ca2+ is a recent discovery of Nature, or if it was invented early in
evolution. It now appears well-established that the key intracellular "Ca2 +-
receptor" protein calmodulin (CaM; see Section V.A) is present in all eukar-
yotic cells. Even in a unicellular eukaryote like common yeast (Saccharomyces
cerevisiae), Ca 2+ has an important regulatory role, and recently yeast CaM, as
well as the single-gene encoding for it, was isolated. 160
The amino-acid sequence of the yeast CaM (147 a.a.; M
r
= 16.1 kDa) is
60 percent identical with the sequences of all other CaMs known. In fact, if
generally accepted conservative amino-acid replacements are allowed, the ho-
mology increases to 80 percent or more, the most highly conserved portions
being the four putative Ca 2 + -binding sites. Sites I and III match the EF-hand
test sequence (see Figure 3.24) very well; in site a His occurs after the "z"-
ligand instead of the archetypal Gly; and in site IV there is no amino acid
between the residues that usually make up ligands "x" and "y." The effect of
these alterations on the Ca 2+ affinity of yeast CaM is not yet known.
That CaM is essential for the growth of yeast cells was shown by deletion
or disruption of the gene. This constitutes, in fact, the first demonstration in
any organism that CaM is an essential protein. (Deletions of genes in mammals
are ethically questionable research procedures!)
In the biochemically less sophiscated (than eukaryotes) prokaryotic cells, a
regulatory role of Ca 2+ is not well-established. What is known is that calcium
is massively accumulated during sporulation in many bacteria, for example, in
strains of Bacillus, Streptomyces, and Myxococcus. In Myxococcus xanthus a
development-specific protein called protein S assembles at the surface of myxos-
pores in the presence of Ca
2
+. The DNA sequence of the gene that encodes this
protein has been deciphered.
161
The primary sequence of protein S (175 a.a.,
M
r
= 19.2 kDa) turns out to closely resemble mammalian CaM. It has four
internally homologous regions with putative Ca2+ sites. At least two of these
are partly similar to the typical EF-hand, but uncharacteristically there are many
more prolines in the M. xanthus protein than in bovine CaM (12 versus 2); so
it is questionable if the bacterial protein really has the repeated helix-loop-helix
structure found in mammalian CaM.
162
One candidate for a prokaryotic CaM was reported by Leadlay et al.
163
in
Streptomyces erythreaus, the bacterium that produces the well-known antibiotic
"erythromycin." The amino-acid sequence of a low-molecular-weight Ca
2
+-
binding protein, as determined from the gene encoding it, revealed a high ho-
mology with mammalian CaM. The protein is made up of 177 amino acids
(M
r
= 20.1 kDa), and has four regions that are predicted to have the helix-
loop-helix secondary structure typical of EF-hand proteins. The aligned se-
quences of the 12 residues in each of the four potential calcium-binding loops
in the S. erythreaus protein are compared with those of human calmodulin in
Table 3.6. The pattern of residues in the S. erythraeus protein is typical of an
EF-hand at least in sites I, III, and IV. Site II is unusual in having Gly at both
positions I and 3. I13Cd NMR studies show that the bacterial protein binds three
metal ions strongly (K :2: 105 M -1) with chemical shifts close to those expected
for EF-hands, and lH NMR studies show that it undergoes a Ca
2
+- dependent
conformational change. 1M
Although the S. erythraeus protein has a homology with eukaryotic CaM, it
has been pointed out that the protein has an even higher homology with a group
of eukaryotic sarcoplasmic Ca2+ -binding proteins 165 (see Section V.D). The
search for a prokaryotic CaM analogue continues, and the prospect of success
has been improved after recent reports of a 21-amino-acid-Iong polypeptide from
an E. coli heat-shock protein 166 that shows the typical structural features of
CaM-binding domains in other eukaryotic proteins. 167
Table 3.6
Aligned EF-hand sequences for the prokaryotic and human calmodulins.
Ligands 3 5 7 9 12
S. erythraeus protein I D F D G N G A L E R A D
II G V G S D G S L T E E Q
III D K N A D G Q I N A D E
IV D T N G N G E L S L D E
Human calmodulin I D K D G D G T I T T K E
II D A D G N G T I D F P E
III D K D G N G
y
I S A A E
IV D I D G D G Q V N
y
E E
IX. APPENDIXES 161
Summary
The role of Ca2 + ions in the regulation of biological activities of procaryotic
organisms is still largely unsettled. Over the last decade, however, evidence has
gradually accumulated that calcium ions are involved in diverse bacterial activ-
ities, such as chemotaxis and substrate transport, sporulation, initiation of DNA
replication, phospholipid synthesis, and protein phosphorylation.
168
An impor-
tant landmark is the recent demonstration that the intracellular Ca2 + concentra-
tion in E. coli is tightly regulated to about 100 nM, a level similar to that typical
of resting eukaryotic cells. 169 Furthermore, increasing numbers of calcium-bind-
ing proteins, some of which also have putative EF-hand Ca2 + sites character-
istic of the calmodulin superfamily of intracellular regulatory proteins, have been
isolated in bacteria. 168
IX. APPENDIXES
A. Definition of Biochemical Terms
Antiport
Basal lateral membrane
Cytosol
Electrogenic
Endocytosis
Endoplasmic reticulum (ER)
Epithelial cells
Erythrocytes
Eukaryotic cells
A transport protein that carries two ions or
molecules in opposite directions across a
membrane.
The membrane in intestinal epithelial cells that
is located on the base of the cells, opposite the
microvilli that face the intestinal lumen.
The unstructured portion of the interior of a
cell-the cell nucleus excluded-in which the
organelles are bathed.
A biological process driven by electric field
gradients.
The process by which eukaryotic cells take up
solutes and/or particles by enclosure in a por-
tion of the plasma membrane to (temporarily)
form cytoplasmic vesicles.
Sheets of folded membranes, within the cyto-
plasm of eukaryotic cells, that are the sites for
protein synthesis and transport.
Cells that form the surface layer of most, if
not all, body cavities (blood vessels, intestine,
urinary bladder, mouth, etc.).
Red-blood corpuscles.
Cells with a well-defined nucleus.
Exocytosis
Gluconeogenesis
Glycolysis
Hydropathy
Lamina propria mucosae
Mitochondrion
Organelle
Osteoporosis
Phorbol esters
Prokaryotic cells
Sarcoplasmic reticulum
Trophoblasts
Tryptic digest
Uniporter
The process by which eukaryotic cells release
packets of molecules ..(e.g., neurotransmitters)
to the environment by fusing vesicles formed
in the cytoplasm with the plasma membrane.
Metabolic synthesis of glucose.
Metabolic degradation of glucose.
A measure of the relative hydrophobic or hy-
drophilic character of an amino acid or amino-
acid side chain.
The layer of connective tissue underlying the
epithelium of a mucous membrane.
A double-membrane organelle in eukaryotic
cells that is the center for aerobic oxidation
processes leading to the formation of energy-
rich ATP.
A structurally distinct region of the cell that
contains specific enzymes or other proteins that
perform particular biological functions.
Brittle-bone disease.
Polycyclic organic molecules that act as ana-
logues to diacylglycerol and therefore are strong
activators of protein kinase C.
Cells lacking a well-defined nucleus.
The ER of muscle cells.
The cells between the maternal and fetal cir-
culation systems.
Fragmentation of proteins as a result of treat-
ment with the proteolytic enzyme trypsin.
A transport protein that carries a particular ion
or molecule in one direction across a mem-
brane.
B. One-Letter Code for Amino-Acid Residues
A-alanine, C-cysteine, D-aspartate, E-glutamate, F-phenylalanine, G-
glycine, H-histidine, I-isoleucine, K-Iysine, L-Ieucine, M-methionine,
N-asparagine, P-proline, Q-glutamine, R-arginine, S-serine, T-thre-
onine, V-valine, W-tryptophan, Y-tyrosine.
X.REFERENCES
C. The Activity of a Transport Protein
This is usually described in terms of the classical Michaelis-Menten scheme:
163
v (= transport rate)
where [S] is the concentration of the solute to be transported and K
m
= (L] +k
2
)1
k] is the Michaelis constant (dimension "concentration") for the reaction
~ k
2
E + S +--'- ES ----? P.
k_
1
Approximated as the reciprocal ratio between on- and off-rate constants rel-
evant to the solute-protein complex, lIK
m
= k]/k_] may be taken as a lower
limit of the affinity of the protein for the solute.
X. REFERENCES
1. See, for example, "Going Crazy over Calcium" in the Health and Fitness section of Time, February
23, 1987, p. 49, or C. Garland et al., The Calcium Diet, Penguin Books, 1990.
2. S. Ringer, J. Physiol. 3 (1883), 195.
3. Handbook of Chemistry and Physics, 64th ed., CRC Press, 1984.
4. R. D. Shannon, Acta Cryst. A, 32 (1976), 751.
5. R. J. P. Williams, in Calcium in Biological Systems, Cambridge Univ. Press, 1976, p. 1.
6. B. A. Levine and R. J. P. Williams, in L. J. Anghileri and A. M. T. Anghileri, eds., Role of Calcium
in Biological Systems, CRC Press, 1982, pp. 3-26.
7. A. K. Campbell, Intracellular Calcium: Its Universal Role as Regulator, Wiley, 1983.
8. A. P. Somlyo, M. Bond, and A. V. Somlyo, Nature 314 (1985), 622.
9. T. B. Johansson, R. Akselsson, and S. A. E. Johansson, Nucl. Inst. Methods 84 (1970), 141.
10. R. B. Martin, in H. Sigel, ed., Metal Ions in Biological Systems, Dekker, 17 (1984), 1.
11. H. Einspahr and C. E. Bugg, in ibid., pp. 52-97.
12. L. G. Sillen and A. E. Martell, eds., Stability Constants of Metal Ion Complexes, Chemical Soc.,
London, 1964.
13. A. E. Martell and R. M. Smith, eds., Critical Stability Constants, Plenum Press, 1, 1975.
14. J. D. Potter and 1. Gergely, 1. Bioi. Chem. 250 (1975), 4628.
15. W. Marki, M. Oppliger, and R. Schwyzer, Helvetica Chemica Acta 60 (1977), 807.
16. C. M. Frey and J. Stuehr, in H. Sigel, ed., Metal Ions in Biological Systems, Dekker, 1 (1974), 51.
17 W. J. Moore, Physical Chemistry, Longman, 5th ed., 1972, Chapter 10.
18. M. V. Thomas, Techniques in Calcium Research, Academic Press, London, 1982.
19. T. J. Rink, Pure Appl. Chem. 55 (1988), 1977.
20. W. Simon et al., Ann. N.Y. Acad. Sci. 307 (1987),5269.
21. R. C. Thomas, Ion-Sensitive Intracellular Microelectrodes, Academic Press, London, 1987.
22. J. R. Blinks et al., Prog. Biophys. Mol. Bioi. 4 (1983), 1.
23. R. Y. Tsien, Biochemistry 19 (1980), 2396.
24. G. Grynkiewicz, M. Poenie, and R. Y. Tsien, J. Bioi. Chem. 260 (1985), 3440.
25. A. Minta, J. P. Y. Kao, and R. Y. Tsien, J. Bioi. Chem 264 (1989), 8171.
26. R. Y. Tsien and M. Poenie, Trends Biochem. Sci. 11 (1986),450.
27. E. Chiancone et al., 1. Bioi. Chem. 26 (1986), 16306.
28. G. A. Smith et al., Proc. Natl. Acad. Sci. USA 80 (1983),7178.
29. J. C. Metcalfe, T. R. Hesketh, and G. A. Smith, Cell Calcium 6 (1985), 183.
164
3 I CALCIUM IN BIOLOGICAL SYSTEMS
30. A. P. Somlyo, Cell Calcium 6 (1985), 197.
31. G. W. Grime et al., Trends Biochem. Sci. 10 (1985), 6.
32. G. H. Morrison and G. Slodzian, Anal. Chem. 47 (1975), 932A.
33. S. Chandra and G. H. Morrison, Science 228 (1985), 1543.
34. E. Murphy et al., Circulation Research 68 (1991), 1250.
35. R. H. Wasserman and C.S. Fullmer, in W. Y. Cheung, ed., Calcium and Cell Function, Academic
Press, 2 (1982), 176.
36. H. Rasmussen, O. Fontaine, and T. Matsumoto, Ann. N. Y. Acad. Sci. 77 (1981),518.
37. S. Linse et al., Biochemistry 26 (1987),6723.
38. D. T. W. Bryant and P. Andrews, Biochem. J. 219 (1984), 287.
39. R. H. Kretsinger, J. E. Mann, and J. G. Simmonds, in A. W. Norman et al., eds., Vitamin D, Chem-
ical, Biochemical, and Clinical Endocrinology of Calcium Metabolism, W. de Gruyter, pp. 232-248.
40. J. J. Feher, Am. 1. Physiol. 244 (1983), 303.
41. 1. Nemere, V. Leathers, and A. W. Norman, 1. Bioi. Chem. 261 (1986), 16106.
42. G. E. Lester. Feder. Proc. 45 (1986),2524.
43. P. Marche, C. LeGuern, and P. Cassier, Cell Tissue Res. 197 (1979), 69.
44. M. Warembourg, C. Perret, and M. Thomasset, Endocrinol. 119 (1986), 176.
45. H. J. Schatzman, Experientia 22 (1966), 364.
46. V. Niggli, J. T. Penniston, and E. Carafoli, 1. Bioi. Chem. 254 (1979),9955.
47. E. Carafoli, M. Zurini, and G. Benaim, in Calcium and the Cell, CIBA Foundation Symposium no.
122, Wiley, 1986, pp. 5 8 ~ 6 5 .
48. D. MacLennan, 1. Bioi. Chem. 245 (1970), 4508.
49. N. M. Green et al., in Reference 47, pp. 93.
50. D. M. Clarke et aI., Nature 339 (1989), 476.
51. G. Inesi, Annu. Rev. Physiol. 47 (1985), 573.
52. R. J. P. Williams, Eur. J. Biochem. 150 (1985), 231.
53. C. Tanford, Proc. Nat!. Acad. Sci. USA 79 (1982), 6527.
54. D. H. MacLennan, K. P. Campbell, and R. A. Reithmeier, in Reference 35, pp. 151-173.
55. A. Maurer et al., Proc. Natl. Acad. Sci. USA 82 (1985), 4036.
56. M. P. Blaustein and M. T. Nelson, in E. Carafoli, ed., Membrane Transport of Calcium, Academic
Press, 1982, pp. 217-236.
57. P. F. Baker, in Reference 47, pp. 73-84.
58. E. Carafoli, G. Inesi, and B. P. Rosen, in Reference 10, pp. 140-143.
59. G. Fiskum, in Reference 10, pp. 187-214.
60. A. L. Lehninger et al., in F. Bronner and M. Peterlik, eds., Calcium and Phosphate Transport Across
Membranes, Academic Press, 1981, pp. 73-78.
61. E. Carafoli, in Reference 56, pp. 109-139.
62. M. J. Berridge, in Reference 47, pp. 39-49.
63. R. F. Irvine, Brit. Med. Bull. 42 (1986), 369.
64. E. W. McCleskey et al., J. Exp. BioI. 124 (1986), 177.
65. D. J. Adams, T. Dwyer, and B. Hille, J. Gen. Physiol. 75 (1980), 493.
66. A. B. MacDermott et al., Nature 321 (1986), 519.
67. R. J. Miller, Science 235 (1987), 46.
68. R. W. Tsien, Annu. Rev. Physiol. 45 (1983), 341.
69. E. W. Sutherland, Science 177 (1972), 401.
70. M. J. Berridge and R. F. Irvine, Nature 312 (1984),315.
71. M. J. Berridge, Biochem. J. 212 (1983), 849.
72. M. J. Berridge, J. Exp. Bioi. 124 (1986), 323.
73. C. C. Chadwick, A. Saito, and S. Fleischer, Proc. Natl. Acad. Sci. USA 87 (1990), 2132.
74. A. R. Hughes and J. W. Putney, Envir. Health Perspec. 84 (1990), 141.
75. A. S. Manalan and C. B. Klee, in Adv. Cycl. Nucl. Prot. Phospho Res. (1984),227.
76. J. D. Potter and J. D. Johnson, in Reference 35, pp. 145.
77. (a) R. H. Kretsinger, Cold Spring Harbor Symp. Quant. Bioi. 52 (1987), 499; (b) C. W. Heizmann and
K. Braun, Trends in Neurosciences 15 (1992), 259.
78. A. Marks et al., J. Neurochem. 41 (1983), 107.
79. P. B. Moore and J. R. Dedman, in H. Hidaka and P. J. Hartshorne, eds., Calmodulin Antagonists and
Cellular Physiology, Academic Press, 1985, pp. 483-494.
80. Y. Nishizuka, Phil. Trans. Roy. Soc. Lond. (B) 302 (1983), 101.
X. REFERENCES
81. M. J. Geisow, FEBS Lett. 203 (1986) 99.
84. M. J. Crumpton and J. R. Dedman, Nature 345 (1990), 212.
85. Y. S. Babu, C. E. Bugg, and W. J. Cook, J. Mol. Bioi. 204 (1988), 191.
86. R. H. Kretsinger and C. E. Nockolds, J. Bioi. Chem. 248 (1973), 3313.
87. S. Forsen, H. J. Vogel, and T. Drakenberg, in Reference 35,6 (1986) 113.
88. M. D. Tsai et aI., Biochemistry 26 (1987), 3635.
89. S. R. Martin et al., Eur. 1. Biochem. 151 (1985), 543.
90. A. Teleman, T. Drakenberg, and S. Forsen, Biochim. Biophys. Acta 873, (1986), 204.
91. L. Sjiilin, Acta Cryst. 846 (1990), 209.
92. S. R. Martin and P. M. Bayley, Biochem. J. 238 (1986), 485.
93. D. B. Heidorn and J. Trewhella, Biochemistry 27 (1988), 909.
94. D. K. Blumenthal et al., Proc. Natl. Acad. Sci. USA 82 (1985),3187.
95. R. E. KIevit et al., Biochemistry 24 (1985),8152.
96. S. Linse, T. Drakenberg, and S. Forsen, FEBS Lett. 199 (1986), 28.
97. S. H. Seeholzer et al., Proc. Nat!. Acad. Sci. USA 83 (1986),3634.
98. A. Persechini and R. H. Kretsinger, J. Cardiovasc. Pharm. 12 (supp!. 5, 1988), S1.
99. G. Barbato et al., Biochemistry, in press.
100. M. Ikura et al., Biochemistry 30 (1991),5498.
101. C. H. Keller et al., Biochemistry 21 (1982), 156.
102. O. Herzberg and M. N. G. James, Nature 313 (1985), 653.
1m. K. A. Satyshur et al., J. Bioi. Chem. 263 (1988), 1628.
104. E. D. McCubbin and C. M. Kay, Ace. Chem. Res. 13 (1980), 185.
lOS. M. T. Hincke, W. D. McCubbin, and C. M. Kay, Can. J. Biochem. 56 (1978),384.
106. P. C. Leavis et al., J. BioI. Chem. 253 (1978), 5452.
107. O. Teleman et al., Eur. J. Biochem. 134 (1983), 453.
108. H. J. Vogel and S. Forsen, in L. J. Berliner and J. Reuben, eds., Biological Magnetic Resonance,
Plenum Press, 7 (1987), 247.
109. S. Forsen et al., in B. de Bernard et al., eds., Calcium Binding Proteins 1983, Elsevier, 1984, pp.
121-131.
110. T. Drakenberg et al., J. Bioi. Chem. 262 (1987), 672.
111. K. B. Seamon and R. H. Kretsinger, in T. G. Spiro, ed., Calcium in Biology, Wiley, 6 (1983), 1;
review of the literature.
112. B. A. Levine and D. C. Dalgamo, Biochim. Biophys. Acta 726 (1983), 187.
113. O. Herzberg, J. Moult, and M. N. G. James, 1. Bioi. Chem. 261 (1986), 2638.
114. W. Chazin et al., personal communication.
115. M. L. Greaser and J. Gergely, J. Bioi. Chem. 246 (1971), 4226.
116. W. Wnuk, J. A. Cox, and E. A. Stein, in Reference 35, pp. 243-278.
117. C. W. Heizmann, in Reference 109, pp. 61-63.
118. R. H. Kretsinger, CRC Crit. Rev. Biochem. 8 (1980), 119; R. H. Kretsinger and C. D. Barry, Biochim.
Biophys. Acta 405 (1975), 4051.
119. W. Hunziker, Proc. Natl. Acad. Sci. USA 83 (1986), 7578.
120. A. L. Swain, R. H. Kretsinger, and E. L. Arnma, J. Bioi. Chem. 264 (1988), 16620.
121. M. F. Summers, Coord. Chem. Rev. 86 (1988), 43; review of the literature.
122. J. M. Gillis et al., J. Musc!. Res. Cell. Mot. 3 (1982), 377.
123. D. M. E. Szebenyi and K. Moffat, J. Bioi. Chem. 261 (1986), 8761.
124. M. Akke, T. Drakenberg, and W. J. Chazin, Biochemistry 31 (1992), 1011.
125. S. Linse et al., Biochemistry 30 (1991), 154.
126. S. Linse et al., Nature 335 (1988),651.
127. M. Akke and S. Forsen, Proteins 8 (1990), 23.
128. W. Chazin et al., Proc. Natl. Acad. Sci. USA 86 (1989), 2195.
129. W. Wnuk, J. A. Cox, and E. A. Stein, in Reference 35, pp. 243-278.
130. W. J. Cook et al., 1. Bioi. Chem. 266 (1991), 652.
131. J. J. Geisow and H. H. Walker, Trends Biochem. Sci. 11 (1986),420.
132. R. H. Kretsinger and C. E. Creutz, Nature 320 (1986), 573.
133. M. R. Crompton, S. E. Moss, and M. J. Crumpton, Cell S5 (1988), 1.
134. J. Kay in A. Heidland and W. H. Horl, eds., Protease Role in Health and Disease, 1984, pp. 519-
570.
135. S. Ohno et al., Nature 312 (1984),566.
165
136. D. E. Croall and G. N. DeMartino, Physiol. Rev. 71 (1991), 813.
137. D. Carpenter, T. Jackson, and M. J. Hanley, Nature 325 (1987), 107.
138. S. Ohno et al., Nature 325 (1987), 161. '.
139. J. F. Harper et al., Science 252 (1991), 951.
140. H. Rasmussen, Am. J. Med. 50 (1971),567.
141. W. Bode and P. Schwager, 1. Mol. Bioi. 98 (1975), 693.
142. E. Chiancone et al., 1. Mol. Bioi. 185 (1985), 201.
143. B. W. Dijkstra et al., J. Mol. Bioi. 147 (1983), 97.
144. D. L. Scott et al., Science 250 (1990), 1541.
145. T. Drakenberg et al., Biochemistry 23 (1984),2387.
146. R. J. P. Williams, in Reference 47, pp. 144-159.
147. D. I. Stuart et al., Nature 324 (1986),84.
148. G. L. Nelsestuen, in Reference 10, pp. 354-380.
149. M. Soriano-Garcia et al., Biochemistry 28 (1989), 6805.
150. D. W. Deerfield et al., J. Bioi. Chern 262 (1987),4017.
151. K. Harlos et al., Nature 330 (1987), 82.
152. J. Lawler and R. O. Hynes, J. Cell. Bioi. 103 (1986), 1635.
153. V. M. Dixit et aI., J. Bioi. Chern. 261 (1986), 1962.
154. J. M. K. Slane, D. F. Mosher, and C.-S. Lai, FEBS Lett. 229 (1988), 363.
155. R. E. Wuthier, in Reference 10, pp. 411-472.
156. F. G. E. Pautard and R. J. P. Williams, in Chemistry in Britain, pp. 188-193 (1982).
157. W. E. Brown and L. C. Chow, Annu. Rev. Material Sci. 6 (1976),213.
158. O. R. Bowen, Publ. Health Reports Suppl. Sept.-Oct. 1989, pp. 11-13.
159. B. Alberts et al., Molecular Biology of the Cell, Garland, 1983, pp. 933-938.
160. T. N. Davis et al., Cell 47 (1986), 423.
161. S. Inouye, T. T. Franceschini, and M. Inouye, Proc. Natl. Acad. Sci. USA 80 (1983),6829.
162. G. Wistow, L. Summers, and T. L. Blundell, Nature 315 (1985), 771.
163. P. F. Leadlay, G. Roberts, and J. E. Walker, FEBS Lett. 178 (1984), 157; D. G. Swan et al., Nature
329 (1987), 84.
164. N. Bylsma et al., FEBS f"ii. 299 (1992), 44.
165. J. A. Cox and A. Bairoch, Nature 331 (1988), 491.
166. M. A. Stevenson and S. K. Calderwood, Mol. Cell Bioi. 10 (1990), 1234.
167. K. T. O'Neil and W. F. DeGrado, Trends Biochem. Sci. 170 (1990),59.
168. V. Norris et al., Mol. Microbial. 5 (1991), 775.
169. M. R. Knight et al., FEBS Lett. 282 (1991), 405.
170. S. Supattapone et al., J. Bioi. Chern. 263 (1988), 1530.
171. C. D. Ferris et al., Nature 342 (1989), 87.
172. R. Huber et al., J. Mol. Bioi. 223 (1992), 683.
173. (a) M. Ikura et al., Science 256 (1992), 632; (b) W. E. Meador, A. R. Means, and F. A. Quiocho,
Science 257 (1992), 1251.
174. M. R. Knight et al., FEBS Lett. 282 (1991), 405.
175. R. W. Tsieu and R. Y. Tsien, Annu. Rev. Cell Bioi. 6 (1990), 715.
176. T. E. Gunter and D. R. Pfeiffer, Am. 1. Physiol. 258 (1990), c755.
177. The authors would like to express their warm gratitude to the many students, colleagues, and coworkers
who, during the preparation of this chapter, have supplied helpful comments, preprints of unpublished
work, background material for figures, etc. Their encouragement is much appreciated. Special thanks
are due to Drs. R. J. P. Williams and G. B. Jameson, who critically read and commented on an early
version of the chapter.

Calcium in Biological System

Uploaded by

Copyright:

Available Formats

Calcium in Biological System

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calcium in Biological System

Uploaded by

Copyright:

Available Formats

3

Calcium in Biological Systems

You might also like