doi:10.1016/S0022-2836(02)00890-2 available online at http://www.idealibrary.

com on

w
B

J. Mol. Biol. (2002) 323, 8598

3D Structure of the Skeletal Muscle

Dihydropyridine Receptor
Ming-Chuan Wang1, Giles Velarde1, Robert C. Ford1
Nicholas S. Berrow2, Annette C. Dolphin2 and Ashraf Kitmitto1*
1

Department of Biomolecular
Sciences, University of
Manchester Institute of Science
and Technology (UMIST)
P.O. Box 88, Manchester
M60 1QD, UK
2

Department of Pharmacology
University College London
Gower Street, London, WC1E
6BT, UK

The dihydropyridine receptors (DHPR) are L-type voltage-gated calcium

channels that regulate the flux of calcium ions across the cell membrane.
resolution
Here we present the three-dimensional (3D) structure at , 27 A
of purified skeletal muscle DHPR, as determined by electron microscopy
and single particle analysis. Here both biochemical and 3D structural
data indicate that DHPR is dimeric. DHPR dimers are composed of two
across and , 75 A
thick, that interact
arch-shaped monomers , 210 A
very tightly at each end of the arch. The roughly toroidal structure of the
diameter, which is
two monomers encloses a cylindrical space of , 80 A
then closed on each side by two dome-shaped protein densities reaching
over from each monomer arch. The dome-shaped domains have a length
and a maximum height of , 45 A
. Small orifices punctuate
of , 80 90 A
their exterior surface. The 3D structure disclosed here may have important
implications for the understanding of DHPR Ca2 channel function. We
also propose a model for its in vivo interactions with the calcium release
channel at the junctional sarcoplasmic recticulum.
q 2002 Elsevier Science Ltd. All rights reserved

*Corresponding author

Keywords: dihydropyridine receptor; electron microscopy; L-type voltagegated calcium; 3D reconstruction; negative staining

Introduction
L-type voltage-gated calcium channels (VGCC)
mediate the flux of calcium ions across cell membranes, in response to changes in membrane potential. In excitable cells such as neurons, heart and
skeletal muscle membranes, VGCC are responsible
for coupling membrane potential changes to biological functions such as secretion, neurotransmitter release, and muscle contraction.1 One group
of drugs binding to these calcium channels with
nanomolar binding constants are the dihydropyridines,2 and thus L-type VGCC are also commonly referred to as dihydropyridine receptors
(DHPR); the nomenclature we will use here. In
Present address: G. Velarde, School of Biological
Sciences, University of Manchester, Stopford Building,
Oxford Road, Manchester, M13 9PT, UK.
Abbreviations used: DHPR, dihydropyridine receptor;
EM, electron microscopy; Chaps, (3-[(cholamidopropyl)dimethyl-ammonio]-1-propanesulphonate); WGA, wheat
germ agglutinin; 2D, two-dimensional; 3D, threedimensional; FSC, Fourier shell correlation; DPR,
differential phase residual; RyR, ryanodine receptor.
E-mail address of the corresponding author:
a.kitmitto@umist.ac.uk

addition to functioning as calcium channels the

DHPRs serve, in skeletal muscle, as voltage sensors
for initiation of muscle excitation contraction
(E C) coupling.3 However, as yet the exact mechanism of muscle E C coupling at the molecular
level is not fully understood.4,5
Purification of the skeletal muscle DHPR by a
number of groups6 8 determined that these ion
channels are hetero-oligomeric membrane protein
complexes, composed of at least four non-covalently associated polypeptides, encoded by four
separate genes:9 a1(, 170 kDa), a2d (, 170 kDa), b
(, 55 kDa), and g (, 30 kDa). The subunit forming
the ion channel pore, and responsible for the influx
of Ca2 across the lipid bilayer is the a1 polypeptide. Topological analysis of the a1 subunit
amino acid sequence, reveals a phylogenetically
conserved motif found in a variety of different
cation channels such as the voltage-gated sodium
and potassium channels.10 Sequence analysis of
the a1 polypeptide predicts 24 a-helical transmembrane segments arranged as four repeats.11
The a2d subunit is composed of two proteins a2
(, 143 kDa) and d (, 25 17 kDa) linked via disulphide bonds.12,13 The d peptide, containing a single
transmembrane segment,14 is thought to anchor

0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved

86

Quaternary Structure of the DHPR

Figure 1. Characterisation of an oligomeric DHPR complex. (a) Lane 1, Native gel of purified DHPR showing a
silver-stained protein band at ,750 kDa; lane 2, immunoblot (anti-beta antibody) of the native gel showing labelling
of the , 750 kDa band. (b) Densitometric scan of DHPR polypeptides after the denaturation of the ,750 kDa protein
complex and SDS-PAGE.

the highly glycosylated extracellular a2 protein to

the lipid bilayer.15,16 Hydropathy plots indicate
that the intracellular b subunit is a hydrophilic
polypeptide,17 whereas the primary sequence of
the g subunit suggests it is composed of four
putative transmembrane segments.18 The functions
of these auxiliary polypeptides (a2d, b and g) are
less clear, though roles in membrane targeting,
stabilisation and channel gating have been
reported by a large number of groups.15,19 23

The DHPRs have been extensively studied in

terms of electrophysiological and pharmacological
properties, however there is a paucity of data pertaining to structure. This has been probably due
in part to the low natural abundance of the ion
channel in most tissue types, and difficulties in isolation, purification, and expression. Freeze-fracture
images of t-tubular membranes24,25 showed a tetrad
arrangement of the DHPR, i.e. groups of four particles, organised in two rows, that are spatially

Quaternary Structure of the DHPR

orientated to allow superimposition and interdigitation with the calcium release channel, also
termed the ryanodine receptor (RyR), which is
localised to the junctional sarcoplasmic recticulum
(jSR) and bridges the junctional gap. These data
provided strong evidence for a muscle E C
coupling model involving a direct interaction of
the DHPR and the RyR (see review26). Rotary
shadowed images of freeze-dried DHPR, purified
from solubilised triad junctions,27 showed particles
ovoid in shape, with a central indentation, and
with short and long dimensions of approximately
16 nm and 22 nm. Four of these oval complexes
were proposed to be required to form the tetrad
structures observed in freeze-fracture images.
Single particle image analysis (SPA) is a powerful technique for the structural study of large
protein complexes that are not readily amenable
to crystallisation.28 31 The development of image
processing techniques designed to analyse noncrystalline large macromolecular complexes such
as membrane proteins has allowed the determination of a number of 3D structures32,33 at
) with SPA of
medium/low resolutions (, 30 A
some viruses sufficient to reveal details of secondary structure.34 Recently Murata and co-workers35
presented two dimensional (2D) projection maps
of negatively stained purified rabbit skeletal
muscle DHPR determined by SPA techniques.
resolution, showed a monomeric
Structures at 30 A
in height and
DHPR complex approximately 200 A
in width. The use of negative staining is wide90 A
spread and it has been successfully employed for
studying the size, shape, and oligomeric form of
macromolecules36 as well as for the investigation
of subunit subunit interactions,37 and large scale
conformational changes.38 Here we have used
negative staining coupled with single particle
analysis to study large dimeric DHPR complexes
with very similar dimensions and gross structural
features to those first observed by Leung and
co-workers.27 These dimeric complexes are equivalent to two of the monomeric particles observed
by Murata et al.35 We present here a three dimensional (3D) structure for this VGCC.

Results
Purification
DHPR was purified using an established
protocol.6 8 Typical yields of purified DHPR were
found to range between 90 and 140 mg/400 g wet
skeletal muscle. DHPR samples were characterised
by SDS-PAGE (Figure 1) and by immunoblotting
with affinity-purified antibody Ab2491 (reactive
with skeletal muscle b subunit from rat and rabbit
t-tubule preparations and also with all rat b isoforms 1 4)39 as well as antisera raised against the
a2/d protein (Ab2079) (an anti-VGCC a2/d antiserum reactive with skeletal muscle from rat and
rabbit t-tubule preparations).15 Purified DHPR

87

complexes were also examined by native gel electrophoresis displaying a single high molecular
mass band at , 750 kDa (Figure 1(a), lane 1).
Immunoblotting with anti-beta peptide antibody
Ab2491 gave a single band corresponding to the
same mass (Figure 1(a), lane 2). The native gel
band (750 kDa) was extracted and the recovered
protein examined by SDS-PAGE (Figure 1(b) (see
Materials and Methods). Due to the varying background intensity from the high to low mass end of
the silver-stained broad-range gel the gel was
digitised and the background subtracted. The
polypeptide composition corresponded to the published literature6 8 for skeletal muscle DHPR. Two
high molecular mass bands at , 170 kDa (peak a)
and , 140 kDa (peak b) correspond to the a1 and
a2 polypeptides. The lower molecular mass band
(e) at , 55 kDa is the expected position of the b
subunit. A further four bands (denoted f i)
between , 32 kDa and 17 kDa were also observed,
corresponding to g, d1, d2 and d3 polypeptides,
respectively. The presence of the d1, d2 and d3
polypetides indicates a heterogeneous d subunit
population due to variable glycosylation as
reported.14 Peaks (c) and (d) at , 66 kDa and
, 60 kDa, respectively, corresponded to silver
stain artefacts and were observed in the absence
of protein. The appearance of these two bands has
been previously documented.40 An aliquot of the
electroeluted protein from the , 750 kDa native
gel band was examined by EM, revealing a homogeneous particle population, almost spherical in
(data not
shape with a diameter of , 200 A
shown). These dimensions and overall shape were
found to be virtually identical with electron micrographs of the purified DHPR shown in Figure 2.
Structural analysis of purified DHPR complexes
Presented in Figure 2(a), top panel, is an electron
micrograph of the DHPR preparation in negative
stain. The particles are homogeneous and uniform
in shape, suggesting a favoured orientation of
DHPR to the EM grid. Preferred orientations of
protein associating with EM grids is a well established phenomenon.41 Factors influencing this
orientation are complex and involve properties of
the negative stain, interactions of the protein with
the stain and the EM grid support film, charge
effects and protein surface topology. Due to the
presentation of a preferred view of the DHPR
complex the random-conical tilt method31 was
used for the 3D reconstruction. This technique
is widely employed for 3D reconstructions of
macromolecules,42 and allows reconstructions to
be determined from one orientation of a
complex.43,32
The micrograph presented in Figure 2(b) represents the same area of the grid after tilting to 508.
2D alignment and classification of the untilted
data set revealed six major classes with estimated
and 35 A

spatial resolutions ranging between 25 A

(see Figure 2, bottom panel, row A, and Table 1).

88

Quaternary Structure of the DHPR

Figure 2. Top panel, electron micrographs of purified, negatively stained, DHPR complexes. (a) Untilted micrograph
showing particles approximately 20 nm in diameter (arrows). (b) The same area as shown in (a) tilted by 508. Examples
of corresponding particles in each micrograph (tilt pairs) are indicated by arrows. Scale bar represents 50 nm. Bottom
panel, row A: 2D projections of the six class averages (untilted dataset). Row B: corresponding 3D volumes (tilted dataset) reconstructed for each class (see Table 1). Scale bar represents 10 nm.

The projections calculated from these six classes

appeared to display roughly the same view with
slightly different orientations with respect to the
EM grid support film and each other (see Table 1).
Each 2D average displays an almost circular shape
to 195 A
and
with diameters ranging from 220 A
distinguished by a large central stain-filled cavity
across, though some protein density can be
, 50 A
seen in each image to be partially obscuring the
cavity. Surface rendered 3D volumes reconstructed
(at 1.2 times the standard deviation (s) above the
mean density) for each of these group averages
from the corresponding tilted data, prior to merging, are displayed in row B, bottom panel, Figure
2. Each volume displays a very similar toroidal
structure with a smaller protruding density in the
centre. Back projections of each of the 3D volumes
gave good correlation to the averaged 2D projec-

tion maps in row A (data not shown) as would be

expected, since the methodology employed uses
the simultaneous iterative reconstruction technique
(SIRT) that works to minimise the discrepancy
between the two maps. This approach is reported
to reduce elongation effects during the reconstruction as a result of the missing cone.44

A final 3D reconstruction of DHPR at , 27 A

resolution (see Table 1) was generated from 1887
images using standard methodology31 as described
in Materials and Methods. The 3D structure of the
purified DHPR is displayed in Figure 3 using
mesh of different colours to delineate different
density levels. In (a) (c) of the top panel in Figure
3 yellow and red mesh correspond to density levels
of 1.6 and 1.25 s above the mean density, respectively. The pink netting in Figure 3(a) is at 2.6 s
above the mean. No symmetry has been applied

89

Quaternary Structure of the DHPR

Table 1. Summary of image processing data

Euler angles (8)a
Class average

No. of particles

2D projections (Figure 2)
1
2
3
4
5
6

557
426
304
268
192
140

3D volume (Figure 3)
No. of particles
) (FSC)
Resolution (A

1887
25

Netting colour
Pink
Yellow
Red

x s above
mean density
2.6
1.6
1.25

)
Resolution (A
26
27
27
28
29
34

Phi

Theta

Psi

0
0.0292
0.1523
275.8631
90.3224
20.2394

0
20.0018
20.0044
5.2138
24.7481
20.0043

0
20.0298
20.1526
94.3392
290.3353
0.2396

Protein massb enclosed by

different thresholding (MDa)
1.40
0.98
0.12

Upper part of the Table gives data on class averages and the Euler rotation angles applied for merging into the 3D reconstruction.
The lower portion relates to the 3D DHPR volume giving the approximate protein mass that could be incorporated in the structure
at different thresholds.
a
Relative to class average 1.
b
Using 0.73 cm3 g21.61

at this stage. The montage begins in Figure 3(a)

with a view perpendicular to the grid support
film where the structure resembles the projection
classes (Figure 2). However, it is clear that the cen 85 A
85 A
in size)
tral stained region (, 85 A
observed in the particles shown in Figure 2 is not
devoid of protein density (arrow). After rotation
around the vertical axis by 458 a much clearer
impression of the 3D structure becomes apparent
(Figure 3(b)). A finger-like protrusion of density
(arrow) (yellow and red mesh combination)
extends upwards and over the upper surface of
the central cavity formed by the main body of protein. Further rotation by 1808 around the vertical
axis (Figure 3(c)) shows a very similar view of a
second finger-like protrusion (yellow mesh only)
that is clearly related by 2-fold symmetry to the
first. In Figure 3(d) with red mesh corresponding
to 1.25 s, and pink to 2.4 s above the mean density,
the finger-like protrusions are seen to partially seal
the central stain cavity from the exterior, forming
a dome-shaped density with a base length of
and a maximum height of , 45 A
. With
, 80 90 A
thresholds used here for display small holes can
be seen to punctuate the surface of the finger/
dome domains (broken arrow). In Figure 3(d) it
can also be seen that the main body of density is
split into two arch-shaped densities that are joined
at each end of the complex (arrows). The archshaped densities are themselves segregated into
about four or five lesser domains, which also map
onto corresponding domains in the other half of
the complex. Thus a C2 axis running through the
DHPR complex is readily identified. Figure 3(e)
shows a view rotated by 908 with respect to Figure
3(a), i.e. a view parallel with the grid support film

and perpendicular to the C2 axis. Some effects due

to differential staining can be noted in this view,
with the side labelled B slightly less contrasted
than side A. In the 3D reconstruction this side of
the particle is the one furthest from the grid
support film, agreeing with the usual expectations
for stain distribution.45,46 Figure 3(f) shows the
same view of the DHPR dimer after the imposition
of 2-fold symmetry on the map. Apart from some
smoothing, there is little change to the overall
shape of the complex, as would be expected for a
homodimeric structure. These results also suggest
that the effects of differential staining are relatively
minor, and do not have a detrimental effect on the
3D reconstruction.
The resolution of the 3D volume (unsymmetrised) was estimated by the Fourier shell correlation (FSC)47 and differential phase residual
and 30 A
were
(DPR)48 techniques. Values of 25 A
calculated, respectively, which are in line with
expected resolutions from structures determined
from negatively stained samples. The true resolution of the volume probably lies midway
between the FSC and DPR values, though this will
be direction-dependent with the resolution along
the z-axis (i.e. perpendicular to the specimen
plane) likely to be lower due to the well established
missing-cone phenomenon.
The volume of the DHPR complex has been
measured at various cut-off density levels for the
3D map (see Table 1). Estimations of the protein
mass for the complex will depend on the density
level chosen. At intermediate thresholds, as indicated by the yellow mesh, the mass corresponding
to this volume could accommodate a DHPR dimer
structure.

90

Quaternary Structure of the DHPR

Figure 3. 3D reconstruction of DHPR viewed at three different density thresholds. (a) Pink 2.6s, yellow 1.6s, red
1.25s above the mean density displaying a view perpendicular to the EM grid support film and resembling the projection images in Figure 2. Density overlaying the central cavity is indicated by the arrow. (b) Rotation of the view in (a)
by 458 ccw around the vertical axis; yellow and red mesh illustrate a finger-like protrusion. (c) Similar view of a second
finger protrusion after rotation of (b) by 1808 around the vertical axis. (d) At lower thresholding (red) the finger-like
protrusion can be seen to form a dome-shaped domain, the broken arrow points to a hole punctuating the dome. The
main body of density enclosed by the pink mesh (2.4s) is formed by two arches of density, the arrows indicate where
the two arches contact each other. (e) View rotated by 908 round the vertical axis with respect to (a). (f) The same
view as (e) after imposition of 2-fold symmetry. Scale bar represents 5 nm.

Quaternary Structure of the DHPR

91

Figure 4. Lectin-gold labelled

purified DHPR complexes.
Row A, Negatively stained
(ammonium molybdate) DHPR
complexes labelled with a
single WGA-gold conjugate
(5 nm colloidal gold). Row B,
DHPR complexes with two
WGA-gold conjugates bound.
Scale bar represents 20 nm.

Lectin-gold labelling of DHPR complexes

Wheat germ agglutinin(WGA) (Triticum vulgaris )
binds to glycosylated residues and therefore will
label the extracellular regions of DHPR.8 Goldlabelled WGA was therefore incubated with
DHPR and then the samples were examined by
electron microscopy (EM). For these labelling
experiments, ammonium molybdate was found to
be the best stain. The majority of DHPR complexes
were labelled with a single WGA-gold conjugate
bound, (Figure 4, row A). Dual labelled DHPR
complexes were also observed (Figure 4, row B).
The two WGA-gold particles appeared to be
approximately the same distance apart and apparently equidistant from the central axis of the
ellipsoidal DHPR complexes, which would be
expected if the WGA binding sites were symmetry-related. Intriguingly, the central cavity was
not stained by the ammonium molybdate (see
Discussion). The dual labelled particles suggest
that the 2-fold symmetry axis runs parallel with
the plane of the support film, which may be
important for differentiating between alternative
explanations for the formation of the DHPR
dimer, as discussed later.

Discussion
Monomer or dimer?
VGCC have been the subject of significant
biochemical, physiological and pharmacological
study but until recently very little data has been
available pertaining to the 3D structure. More
success has been obtained with the potassium49
and sodium voltage-gated channels.50 Sato and
co-workers50 presented a 3D structure of the
voltage-sensitive sodium channel isolated from
the eel Electrophorus electricus. The polypeptide
containing the ion channel (, 260 kDa) has the
same predicted topology as a1 DHPR, i.e. four
homologous domains that each contain six transmembrane helices,11 although the skeletal muscle
DHPR a1 subunit is , 70 kDa lower in mass
than the sodium channel. The sodium channel is
monomeric and has a bell-shape of 13.5 nm in

height, and a square-shaped base of side length

10 nm. Grigorieff and co-workers49 have recently
shown by EM/negative stain that the tetrameric
potassium channel is mushroom-shaped, with a
thickness of approximately 7 nm as viewed from
the cytoplasmic side of the membrane and 11 nm
as viewed from the other, with an , 11 nm span
perpendicular to the membrane. Each monomer in
this channel contains six transmembrane helices
giving a similar topology to the sodium and
calcium channels. The DHPR 3D volume is large
enough (Table 1) to accommodate two sodium
channels or two potassium channels. Murata and
colleagues35 have presented projection structures
corresponding to particles present in skeletal
muscle DHPR preparations. These projections did
not display any obvious 2-fold symmetry, and
their estimated dimensions of 9 nm 10 nm 20
nm, approximate to half of the 3D DHPR structure
presented here. Thus we suggest that the particles
studied by Murata et al., probably correspond to
monomeric DHPR.
The 3D structure presented here displays a
noticeable 2-fold symmetry, it has a volume that
can accommodate at least twice the expected mass
of the DHPR polypeptides, and with native PAGE
it runs at a mass equivalent to a dimer. Thus we
appear to have isolated a dimer of DHPR. It
remains to be established whether the DHPR
dimer is of functional and physiological significance. Dimerisation occurring after membrane
solubilisation by detergent is not unknown, and
can be manifested in both head-to-head and a
head-to-tail structures. The latter conformation,
readily identified as non-physiological, is ruled
out by the lectin-gold labelling studies (Figure 4),
which appear to show a head-to-head conformation. Thus we conclude that either the DHPR
dimer is the true in vivo form of DHPR or that
detergent solubilisation and purification has
resulted in a head-to-head dimer. Cross-linking
studies prior to solubilisation might help to clarify
this issue in the future. A finding of possible
relevance to this issue is that point mutations in
the human skeletal a1 subunit that result in
hypokalaemic periodic paralysis give rise to a
dominant disease,51 whereas truncation mutations

92

Quaternary Structure of the DHPR

Figure 5. 3D structure of the DHPR after 2-fold symmetry has been applied, displayed at decreasing density
thresholds ((a) (c)). (a) The DHPR complex at the highest threshold with two possible locations for the lipid bilayer
indicated by broken and dotted lines (I and II, respectively, see Discussion). (d) DHPR complex at the same thresholds
as that in (b) but tilted so as to reveal the interior and depth of the central cavity. The arrows in the panel indicate
the position of the C2 symmetry axis. Meshing: pink 2.5s, yellow 1.5s, and red 1.2s above the mean density. Scale
bar represents 5 nm.

that produce a complete loss of expression of one

allele result in a recessive disease.
Location of the transmembrane region?
The volumes in Figure 5 show the DHPR dimer
after 2-fold symmetry has been applied. Figure
5(a) (c) presents the same view of the volume
with different threshold combinations. Two possible locations of the lipid bilayer with respect to
the 3D volume are indicated in Figure 5(a), the
broken lines representing one position for the
transmembrane region in DHPR (I), and the dotted
lines another (II). Another possible model (III)
could be where one of the finger domains is
located on the intracellular side of the membrane
and the other on the extracellular side. A DHPR
monomer is predicted to contain , 29 transmembrane helices and by taking an average of 25
residues for a transmembrane helix, it is clear that
only a small (17 25%) proportion of the total
mass of each DHPR monomer will actually be
located in the lipid bilayer. Moreover, a footprint
in the membrane of about 50 100 nm2 would be

expected for , 24 transmembrane helices (a1 subunit), dependent on their packing. These factors
would be compatible with models I and II, but not
model III, where most of the mass of the protein is
in the membrane. In addition model III would not
fit with DHPR polypeptide topology predictions,
which indicate an asymmetric distribution of
protein mass across the t-tubular membrane.
Figure 5(d) is displayed at threshold levels comparable to those in (b). In this Figure the symmetrised structure has been rotated so as to reveal
the interior and depth of the cavity extending into
the finger/dome domains. The arrows indicate the
symmetry axis through the volume.
Sections through the 3D volume orientated
as shown in Figure 5(a) (c) (surface rendered in
SPIDER, 1s above the mean density) taken perpendicular to the 2-fold symmetry axis, displayed at
intervals, are shown in Figure 6 (direction of
11 A
slicing running top to bottom). Model I would
place the boundaries of the lipid bilayer between
slices 4 and 8, corresponding to a thickness of
, whereas in model II the membrane
, 40 50 A
would be placed between slices 4 and 8. It is not

Quaternary Structure of the DHPR

93

intervals through the DHPR 3D structure taken perpendicular to the C2 axis, i.e. as viewed
Figure 6. Slices at 11 A
looking down onto the membrane. Slices start from the top through to the bottom through to the top of the structure
(as defined in the legend to Figure 5). Box size, 26 nm 26 nm.

currently possible to distinguish between models I

and II. However, since the majority of the mass of
DHPR is on the extracellular side (which is glycosylated), then single particle analysis of the lectinlabelled complexes should in principle offer a
route to resolve this problem. The slices also clearly
show the finger domains, each one emerging from
one side of the complex and interacting with the
opposing monomer (indicated by arrows). The
contacts between the two DHPR monomers are
seen to be most intimate at the top and bottom
of the dimeric structure, as also noted in the
orientations displayed in Figure 3.
Interactions with the ryanodine
receptor complex
The DHPR complex is a key component of
a complex signal-transduction mechanism. In
skeletal muscle a number of studies have suggested that the DHPR interacts on its intracellular
side with a number of protein components of
the triad junction including the RyR26 and
calmodulin.52 As discussed earlier, in skeletal
muscle the direct interaction of the DHPR and the

RyR is believed to form the basis for E C coupling.

The RyR complex (a tetramer) has four large cytoplasmic domains which extend into the gap that
separates the t-tubular and jSR membranes as concluded from cryo-EM studies of purified RyR.53,34
The four globular domains form a square shape
with dimensions of 28 nm 28 nm and 12 nm
thick, attached to a central transmembrane region.
On each of the four cytoplasmic domains are
regions termed the clamps and handles that
form the corners of the square, surrounding a
region of low density, and are approximately
19.5 nm apart side to side. A recent study by
Samso et al.54 mapped the four binding sites of
imperatoxin A (IpTxa) (believed to mimic the
cytoplasmic II III loop of the DHPR a1 subunit)
on the RyR to a crevice between the clamp and
handle regions. The distance between the four
binding sites was determined to be 15 nm. Figure
7 shows how dimeric DHPR complexes could be
positioned onto the tetrameric RyR (adapted from
Radermacher et al.32). It is intriguing to note that
the dimeric DHPR structure could superimpose
on the RyR clamp regions roughly overlaying the
IpTxa binding domains, and whilst speculative,

94

Quaternary Structure of the DHPR

Figure 7. Cartoon depicting possible interaction sites of the DHPR

complex with the tetrameric RyR
complex (adapted from32). (a) The
complex as viewed parallel with
the membrane planes (side view).
(b) Superposition of two DHPR
dimers on the opposing RyR extracellular face (with respect to the
jSR).

this model could now be tested by further structural analyses and cross-linking studies.
Structure/function relationships
We have presented data from both structural and
biochemical studies demonstrating that skeletal
muscle DHPR can be isolated as a dimer. An
intriguing feature of the dimeric DHPR structure
is the large central cavity, which is formed by the
two halves of the dimer. If this chamber were to
have a function in gating or selectivity, then a
physiological role for dimerisation is implicit.
Of interest in this respect is the observation of
different behaviour of negative stain with respect
to the cavity. Whilst the positively charged uranyl
ions readily penetrate the cavity, the behaviour
of negatively charged molybdate ions is quite
different, with little or no staining of the cavity. A
larger survey of heavy atom stains is underway.
Future work is now aimed towards the determination of a higher resolution 3D structure of the
DHPR, as well as reconstitution of the complexes
into lipid bilayers, to further develop and expand
upon the data and proposals for the quaternary
organisation of the skeletal muscle L-type VGCC
presented here.

Materials and Methods

Purification
Triads/t-tubules were prepared as described55 from
New Zealand White rabbit skeletal muscle (,2.5 3 kg)
and then photoaffinity labelled.7 DHPR was isolated
using a four-step purification based upon the methods
described by a number of groups6 8 i.e. two steps of
WGA affinity chromatography, sucrose gradient centri-

fugation, and DEAE-chromatography. All steps were

carried out at 4 8C and buffers at each step supplemented
with the following protease inhibitors: 0.1 mM phenylmethylsulphonyl fluoride (PMSF), 1 mM pepstatin A,
1 mM L -trans-epoxysuccinyl-leucylamide-(4-guanidino)butane (E64), and 1 mM benzamidine. Triad membranes
(2 mg ml21) were solubilised with 1% (w/v) 3-[(cholamidopropyl)dimethyl-ammonio]-1-propanesulphonate)
(Chaps) (Sigma), 0.5% (w/v) asolecithin (L -a-phosphatidylcholine from soyabean, Type II-S, Sigma-Aldrich) at
a protein: detergent ratio of 1:5 (w/w). Membranes
solubilised with Chaps included the addition of asolecithin in buffers at all stages of the purification.27
Image processing was carried out on purified DHPR in
a final buffer of 20 mM tris(hydroxymethyl)aminomethane (Tris) (pH 7.5), 0.5% (w/v) Chaps, 0.25% (w/v)
asolecithin and protease inhibitors.
Gel electrophoresis
Samples were prepared for denaturing electrophoresis
(SDS-PAGE) by incubation of protein with sample buffer
(60 mM Tris (pH 6.8), 25% (v/v) glycerol, 2% (w/v) SDS,
14.4 mM 2-mercaptoethanol, 40 mM dithiothreitol (DTT),
1% (w/v) bromophenol blue) at 90 8C for five minutes.
4% 20% gradient gels (Bio-Rad Ltd) were run at 200 V
for 45 minutes.
Gels were stained with either 0.1% (w/v) Coomassie
blue (in 10% (v/v) methanol, 10% (v/v) glacial acetic
acid) or silver stained using a Bio-Rad silver staining kit.
Covalent incorporation of [3H]azidopine into triad
proteins was detected by liquid scintillation counting of
gel slices. Protein bands were visualised by Coomassie
staining and then excised from the gels, cut into small
pieces and then digested overnight27 in 0.5 ml of 30%
H2O2, at 50 8C. Samples were centrifuged at 13,000g
(Eppendorf centrifuge 5415C) and supernatants collected
for liquid scintillation counting (Packard 2000CA liquid
scintillator).
Native gels with a 4% stacking and 5% separating gel
were run at 60 V for five hours (Bio-Rad Mini-Protean II
electrophoresis system) at 4 8C. Protein samples in 0.5%

95

Quaternary Structure of the DHPR

Chaps, 0.25% asolecithin, supplemented with a protease

cocktail, as described earlier were combined with an
equal volume of sample buffer (60 mM Tris (pH 6.8),
25% glycerol, 1% bromophenol blue). The running buffer
was composed of 16.7 mM tricine, 192 mM glycine (pH
8.3) and 0.1% SDS. Molecular mass markers for running
on native gels were purchased from Pharmacia.
Electroelution
Bands were excised from native gels and the protein
was electroeluted (five hours) at 4 8C in a BioRad
Electro-eluter. Protein was eluted into a 20 mM Tris,
160 mM boric acid (pH 8.4), 0.5 mM EDTA buffer
containing 0.5% Chaps, 0.25% asolecithin, and supplemented with protease inhibitors as described above.
Eluted protein was examined by EM as described below.
Immunoblotting
Protein bands from gels were transferred to nitrocellulose using a Biometra Electrophoretic transfer cell.
Ponceau S staining (0.2% (w/v) Ponceau S, 3% (v/v) trichloroacetic acid (TCA), 3% (v/v) sulphosalicylic acid)
of the nitrocellulose indicated that 90 minutes at 20 V
was required for protein transfer. The nitrocellulose was
incubated for one hour with the primary antibody
(dilution 1:200). An anti-rabbit IgG secondary antibody
(Sigma-Aldrich) with an alkaline-phosphatase conjugate
was employed for immunodetection. Antisera and affinity purified antibodies raised against anti-beta peptide
(Ab2491) and anti-alpha2 peptide (Ab2079) were prepared as described.15,39 Protein was extracted from nitrocellulose paper by incubating the nitrocellulose in 20%
acetonitrile for three hours at 37 8C, after which time the
nitrocellulose paper was discarded and the acetonitrile/
protein mix lyophilised. The lyophilised protein extract
was resuspended in a small volume of SDS sample
buffer and the polypeptide composition determined by
SDS-PAGE electrophoresis as described above. Silverstained gels were digitised in transmissive mode
(EPSON GT-9600 flatbed scanner) and protein peaks
quantified using ImageQuant (v 5.1) after background
subtraction.
Electron microscopy
Samples at 100 mg/ml of protein, were loaded onto
carbon-coated grids and negatively stained with 4%
(w/v) uranyl acetate as described56 or 5% (w/v)
ammonium molybdate and 1% (w/v) trehalose.57
Samples were examined in a Philips Tecnai 10 transmission electron microscope at an accelerating voltage
of 100 kV under low dose conditions. Electron micrographs were typically taken at 1.7 mm defocus, and
were recorded at a calibrated magnification of 43,400
on Kodak Electron Image film SO-163. Micrographs
were digitised on a UMAX PowerLook 3000 scanner
with a scan step of 16 mm, corresponding to a pixel size
at the specimen level.
of 3.7 A
Lectin-gold labelling of purified DHPR
Lectin (T. vulgaris ) conjugated to 5 nm colloidal gold
(British Biocell Ltd) was added in excess to the purified
DHPR. Samples were incubated for 24 hours (in the
presence of protease inhibitors, described above) at 4 8C.
Aliquots were taken from the incubation mix, negatively

stained (ammonium molybdate/trehalose) and examined by EM as described above.

3D Image processing
The random conical reconstruction method using
SPIDER and WEB image processing packages31 was
employed for the calculation of a 3D structure of the
DHPR. A total of 35 micrograph pairs (taken at 08 and
508) of negatively stained (uranyl acetate) DHPR complexes were selected and scanned as described above.
Particles were selected from the untilted and tilted
micrographs (2405 tilt pairs) and an individual contrast
transfer function (CTF) correction was applied to each
particle in both data sets based upon its positional
co-ordinates. The focus gradient across the tilted
micrographs was determined using the software
ctffind2.com.58 The focus gradient parameters for each
micrograph were then applied using the SPIDER CTF
correction batch files.
This file also compensates for the amplitude contrast
for negative stain.
The untilted data set was rotationally and translationally aligned using a reference-free alignment algorithm.
However, due to the ellipsoidal shape of the DHPR
complexes, with axes very similar in length, alignment
of the dataset involved initially filtering the particles to
followed by reference-free alignment. This gave
45 A
good alignment of the dataset as judged by correspondence analysis and hierarchical ascendant classification.
The shifts and rotations used for aligning these heavily
filtered particles were then applied to the original data . Hierarchical clusterset, after low pass filtering to 20 A
ing, with complete linkage as a merging criterion, was
used to separate out the major groups. Only those
averages (six classes: 557, 426, 340, 248, 204, 192, and
140 particles, Figure 2, lower panel, row A) with reso were selected for reconstruction of
lutions 25 35 A
volumes by back projection from the respective tilted
particles. Resolution assessment of each major class was
determined by plotting the Fourier ring correlation
(FRC) between two independent averages. As a measure
of resolution for FRC 0.5 was taken as a cut-off.59
Each of the six volumes was individually refined six
times by optimising the original volume by an angular
refinement step (Figure 2, lower panel, row B). In brief,
back projections calculated from the initial 3D volume
were used to translationally realign the original tilted
particles. The realigned particles were then used to
create a new volume. The refined volumes from each
separate class were then merged after determining the
correct Euler angles for combination (Table 1). The
groups were merged in order of size, the largest first.
The final reconstruction included 1887 images representing 79% of the original tilted dataset. The resolution of
this final volume was assessed by the standard method
of calculating the FSC60 between two subvolumes,
formed by two half-sets of the selected images, with the
resolution limit taken to be at the point where the FSC
fell below 0.5.47 In addition, resolution was assessed by
plotting the DPR48 with 458 used as a cut-off measure.
The 2-fold rotational symmetry of the DHPR complexes isolated here was confirmed by cross-correlation
of the unsymmetrised volume with itself after rotation
http://www.wadsworth.org/spider_doc/spider/
docs/techs/ctf/ctf.html

96

by 1808 around the C2 symmetry axis. Despite the effects

of partial staining of one side of the complex, good
agreement was found with a value of 0.922. These data
are consistent with those from biochemical studies,
presented here; indicating that this complex is dimeric.
The approximate protein mass encapsulated at each
threshold level used to display the 3D volumes in Figure
3 were determined using the SPIDER commands TH M
and FS to calculate the volume fraction above the
threshold. This value was then converted to cubic
Angstroms. Using a partial specific volume of protein61
of 0.73 cm3 g21 the molecular mass of the complex at
each threshold was calculated.
SPOCK (developed by JA Christopher at Texas A&M)
molecular graphics software was used to display 3D
volumes using a chicken wire effect to illustrate the
structure at different thresholding.

Quaternary Structure of the DHPR

10.
11.

12.

13.

14.

Acknowledgements
This work was funded by the British Heart
Foundation via a Basic Science Lectureship
Award to Dr A. Kitmitto (BS/97002). The authors
acknowledge Dr Richard Collins and Professor
Alan Eddy (UMIST) and Dr Anthony Davies
(UCL) for useful discussions.

References
1. Hagiwara, S. & Byerly, L. (1981). Calcium channel.
Annu. Rev. Neurosci. 4, 69 125.
2. Janis, R. A. & Triggle, D. J. (1984). 1,4-Dihydropyridine Ca-2 channel antagonists and activatorsa
comparison of binding characteristics with pharmacology. Drug Dev. Res. 4, 257 274.
3. Rios, E. & Brum, G. (1987). Involvement of dihydropyridine receptors in excitation contraction coupling
in skeletal-muscle. Nature, 325, 717 720.
4. Grabner, M., Dirksen, R. T., Suda, N. & Beam, K. G.
(1999). The II III loop of the skeletal muscle dihydropyridine receptor is responsible for the bi-directional
coupling with the ryanodine receptor. J. Biol. Chem.
274, 21913 21919.
5. Wilkens, C. M., Kasielke, N., Flucher, B. E., Beam,
K. G. & Grabner, M. (2001). Excitation contraction
coupling is unaffected by drastic alteration of
the sequence surrounding residues L720-L764 of the
alpha(1S) II III loop. Proc. Natl Acad. Sci. USA, 98,
5892 5897.
6. Curtis, B. M. & Catterall, W. A. (1984). Purification of
the calcium antagonist receptor of the voltage sensitive calcium channel from skeletal muscle transverse
tubules. Biochemistry, 23, 2113 2118.
7. Sharp, A. H., Imagawa, T., Leung, A. T. & Campbell,
K. P. (1987). Identification and characterization of
the dihydropyridine-binding subunit of the skeletalmuscle dihydropyridine receptor. J. Biol. Chem. 262,
12309 12315.
8. Florio, V., Striessnig, J. & Catterall, W. A. (1992).
Purification and reconstitution of skeletal-muscle
calcium channels. Methods Enzymol. 207, 529 546.
9. Leung, A. T., Imagawa, T. & Campbell, K. P. (1987).
Structural characterization of the 1,4-dihydropyridine receptor of the voltage-dependent Ca-2

15.

16.

17.
18.

19.
20.

21.

22.

23.

24.
25.

channel from rabbit skeletal-muscleevidence for 2

distinct high-molecular-weight subunits. J. Biol.
Chem. 262, 7943 7946.
Catterall, W. A. (1988). Structure and function of
voltage-sensitive ion channels. Science, 242, 50 61.
Tanabe, T., Takeshima, H., Mikami, A., Flockerzi, V.,
Takahashi, H., Kangawa, K. et al. (1987). Primary
structure of the receptor for calcium-channel blockers
from skeletal-muscle. Nature, 328, 313 318.
Ellis, S. B., Williams, M., Ways, N. R., Brenner, R.,
Sharp, A. H., Leung, A. T. et al. (1988). Sequence
and expression of messenger-RNAs encoding the
alpha-1 subunit and alpha-2 subunit of DHP-sensitive calcium channel. Science, 241, 1661 1664.
De Jongh, K. S., Warner, C. & Catterall, W. A. (1990).
Subunits of purified calcium channelsalpha-2 and
delta are encoded by the same gene. J. Biol. Chem.
265, 14738 14741.
Jay, S. D., Sharp, A. H., Kahl, S. D., Vedvick, T. S.,
Harpold, M. M. & Campbell, K. P. (1991). Structural
characterization of the dihydropyridine-sensitive
calcium-channel alpha-2-subunit and the associated
delta-peptides. J. Biol. Chem. 266, 3287 3293.
Brickley, K., Campbell, V., Berrow, N., Leach, R.,
Norman, R. I., Wray, D. et al. (1995). Use Of sitedirected antibodies to probe the topography of the
alpha(2) subunit of voltage-gated Ca2 channels.
FEBS Letters, 364, 129 133.
Gurnett, C. A., De Waard, M. & Campbell, K. P.
(1996). Dual function of the voltage-dependent
Ca2 channel alpha(2)delta subunit in current stimulation and subunit interaction. Neuron, 16, 431 440.
Walker, D. & De Waard, M. (1998). Subunit interaction sites in voltage-dependent Ca2 channels:
role in channel function. Trends Neurosci. 21, 148 154.
Jay, S. D., Ellis, S. B., McCue, A. F., Willimas, M. E.,
Vedvick, T. S., Harpold, M. M. & Campbell, K. P.
(1990). Primary structure of the gamma-subunit of
the DHP-sensitive calcium-channel from skeletal
muscle. Science, 248, 490 492.
Hofmann, F., Biel, M. & Flockerzi, V. (1994).
Molecular-basis for Ca2 channel diversity. Annu.
Rev. Neurosci. 17, 399 418.
DeWaard, M., Scott, V. E. S., Pragnell, M. & Campbell, K. P. (1996). Identification of critical amino
acids involved in alpha(1) beta interaction in
voltage-dependent Ca2 channels. FEBS Letters,
380, 272 276.
Brice, N. L., Berrow, N. S., Campbell, V., Page, K. M.,
Brickley, K., Tedder, I. & Dolphin, A. C. (1997).
Importance of the different beta subunits in the
membrane expression of the alpha 1A and alpha 2
calcium channel subunits: studies using a depolarization-sensitive alpha 1A antibody. Eur. J. Neurosci.
9, 749 759.
Gurnett, C. A., Felix, R. & Campbell, K. P. (1997).
Extracellular interaction of the voltage-dependent
Ca2 channel alpha(2)delta and alpha(1) subunits.
J. Biol. Chem. 272, 18508 18512.
Gao, T., Chien, A. J. & Hosey, M. M. (1999). Complexes of the alpha(1C) and beta subunits generate
the necessary signal for membrane targeting of class
C L-type calcium channels. J. Biol. Chem. 274,
2137 2144.
Franzini-Armstrong, C. (1984). Freeze-fracture of
frog slow tonic fibres:structure of surface and
internal membranes. Tissue Cell, 16, 146 166.
Block, B. A., Imagawa, T., Leung, A., Campbell, P. &
Franzini-Armstrong, C. (1988). Structural evidence

97

Quaternary Structure of the DHPR

26.
27.

28.

29.
30.

31.

32.

33.
34.
35.

36.
37.

38.

39.

40.

41.

for direct interaction between the molecularcomponents of the transverse tubule sarcoplasmic
reticulum junction in skeletal-muscle. J. Cell Biol.
107, 2587 2600.
Franzini-Armstrong, C. & Jorgenson, A. O. (1994).
Structure and development of E C coupling units
in skeletal muscle. Annu. Rev. Physiol. 56, 509 534.
Leung, A. T., Imagawa, T., Block, B., FranziniArmstrong, C. & Campbell, K. P. (1988). Biochemical
and ultrastructural characterization of the 1,4-dihydropyridine receptor from rabbit skeletal-muscle
evidence for a 52,000 Da subunit. J. Biol. Chem. 263,
9941001.
van Heel, M., Gowen, B., Matadeen, R., Orlova, E. V.,
Finn, R., Pape, T. et al. (2000). Single-particle electron
cryo-microscopy: towards atomic resolution. Quart.
Rev. Biophys. 33, 307 369.
Radermacher, M. (1988). 3-Dimensional reconstruction of single particles from random and nonrandom
tilt series. J. Electron Microsc. Tech. 9, 359 394.
Schatz, M., Orlova, E. V., Dube, P., Jager, J. & van
Heel, M. (1995). Structure of lumbricus-terrestris
resolution determined using
hemoglobin at 30 A
angular reconstitution. J. Struct. Biol. 114, 28 40.
Frank, J., Radermacher, M., Penczek, P., Zhu, J., Li,
Y. H., Ladjadj, M. & Leith, J. (1996). SPIDER and
WEB: processing and visualization of images in 3D
electron microscopy and related fields. J. Struct. Biol.
116, 190 199.
Radermacher, M., Rao, V., Grassucci, R., Frank, J.,
Timerman, A. P., Fleischer, S. & Wagenknecht, T.
(1994). Cryoelectron microscopy and 3-dimensional
reconstruction of the calcium-release channel ryanodine receptor from skeletal-muscle. J. Cell Biol. 127,
411 423.
Saibil, H. (2000). Molecular chaperones: containers
and surfaces for folding, stabilising or unfolding.
Curr. Opin. Struct. Biol. 10, 251 258.
Chiu, W. & Schmid, M. F. (1997). Pushing back the
limits of electron cryomicroscopy. Nature Struct. Biol.
4, 331333.
Murata, K., Odahara, N., Kuniyasu, A., Sato, Y.,
Nakayama, H. & Nagayama, K. (2001). Asymmetric
arrangement of auxiliary subunits of skeletal muscle
voltage-gated L-type Ca2 channel. Biochem. Biophys. Res. Commun. 282, 284 291.
Harris, J. R. & Horne, R. W. (1994). Negative staining:
a brief assessment of current technical benefits,
limitations and future possibilities. Micron, 25, 5 13.
Holzenburg, A., Flint, T. D., Shepherd, F. H. & Ford,
R. C. (1996). Photosystem II: mapping the locations
of the oxygen evolution-enhancing subunits by
electron microscopy. Micron, 27, 121 127.
Rosenberg, M. F., Velarde, G., Ford, R. C., Martin, C.,
Berridge, G., Kerr, I. D. et al. (2001). Repacking of the
transmembrane domains of P-glycoprotein during
the transport ATPase cycle. EMBO J. 20, 5615 5625.
Berrow, N., Tedder, I., Page, K., Brice, N., Amos, B. &
Dolphin, A. C. (1995). Properties of alpha(1a)
calcium-channel currents is expressed in the Cos-7
cell-line. J. Physiol. (Lond.), 489P, P52 P53.
Flockerzi, V., Oeken, H.-J. & Hofmann, F. (1986).
Purification of a functional receptor for calciumchannel blockers from rabbit skeletal-muscle microsomes. Eur. J. Biochem. 161, 217 224.
Tischendorf, G. W., Zeichhardt, H. & Stoffler, G.
(1974). Determination of the location of proteins
L14, L17, L18, L19, L22, and L23 on the surface of
the 50 S ribosomal subunit of Escherichia coli by

42.

43.

44.
45.
46.
47.

48.
49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

immunoelectron microscopy. Mol. Gen. Genet. 134,

187 204.
Venien-Bryan, C., Lowe, E. M., Boisset, N., Traxler,
K. W., Johnson, L. N. & Carlson, G. M. (2002).
Three-dimensional structure of phosphorylase
resolution and its complex with glycokinase at 22 A
gen phosphorylase b. Structure, 10, 33 41.
Mayanagi, K., Miyata, T., Oyama, T., Ishino, T. &
Morikawa, K. (2001). Three-dimensional electron
microscopy of the clamp loader small subunit from
Pyrococcus furiosus. J. Struct. Biol. 134, 35 45.
Penczek, P., Radermacher, M. & Frank, J. (1992).
Three-dimensional reconstruction of single particles
embedded in ice. Ultramicroscopy, 40, 33 53.
Frank, J. (1982). New methods for averaging nonperiodic objects and distorted crystals in biologic
electron microscopy. Optik, 63, 67 89.
Boekema, E. J. (1991). Negative staining of integral
membrane proteins. Micron Microsc. Acta, 22,
361 369.
Bottcher, B., Wynne, S. A. & Crowther, R. A. (1997).
Determination of the fold of the core protein of
hepatitis B virus by electron microscopy. Nature,
386, 88 91.
Frank, J., Verschoor, A. & Boublik, M. (1981). Computer averaging of electron micrographs of 40 S ribosomal subunits. Science, 214, 1353 1355.
Sokolova, O., Kolmakova-Partensky, L. & Grigorieff,
N. (2001). Three-dimensional structure of a voltagegated potassium channel at 2.5 nm resolution. Structure, 9, 215 220.
Sato, C., Ueno, Y., Asai, K., Takahashi, K., Sato, M.,
Engel, A. & Fujiyoshi, Y. (2001). The voltage-sensitive
sodium channel is a bell-shaped molecule with
several cavities. Nature, 409, 1047 1051.
Lapie, P., Lory, P. & Fontaine, B. (1997). Hypokalemic
periodic paralysis: an autosomal dominant
muscle disorder caused by mutations in a voltagegated calcium channel. Neuromusc. Disord. 7,
234 240.
Zuhlke, R. D., Pitt, G. S., Deisseroth, K., Tsien, R. W.
& Reuter, H. (1999). Calmodulin supports both
inactivation and facilitation of L-type calcium
channels. Nature, 399, 159 162.
Serysheva, I. I., Orlova, E. V., Chiu, W., Sherman,
M. B., Hamilton, S. L. & van Heel, M. (1995). Electron
cryomicroscopy and angular reconstitution used
to visualize the skeletal-muscle calcium-release
channel. Struct. Biol. 2, 18 24.
Samso, M., Trujillo, R., Gurrola, G. B., Valdivia, H. H.
& Wagenknecht, T. (1999). Three-dimensional
location of the imperatoxin A binding site on the
ryanodine receptor. J. Cell Biol. 146, 493 499.
Glossman, H. & Ferry, D. R. (1983). Solubilization
and partial-purification of putative calcium channels
labeled with nimodipine-H-3. Naunyn-Schmiedebergs
Arch. Pharmacol. 323, 279 291.
Holzenberg, A., Shepherd, F. H. & Ford, R. C. (1994).
Localization of the oxygen-evolving complex of
photosystem-II by Fourier difference analysis.
Micron, 25, 447451.
Harris, J. R., Hoeger, U. & Adrian, M. (2001). Transmission electron microscopical studies on some
haemolymph proteins from the marine polychaete
Nereis virens. Micron, 32, 599 613.
Crowther, R. A., Henderson, R. & Smith, J. M. (1996).
MRC image processing programs. J. Struct. Biol. 116,
9 16.

98

Quaternary Structure of the DHPR

59. vanHeel, M. (1987). Similarity measures between

images. Ultramicroscopy, 21, 95 99.
60. Harauz, G. & van Heel, M. (1986). Exact filters for
general geometry three-dimensional reconstruction.
Optik, 73, 146 156.

61. Holzenburg, A., Bewley, M. C., Wilson, F. H.,

Nicholson, W. V. & Ford, R. C. (1993). Three dimensional structure of photosystem II. Nature, 363,
470 472.

Edited by I. B. Holland
(Received 7 May 2002; received in revised form 9 August 2002; accepted 9 August 2002)