Sie sind auf Seite 1von 12

Molecular Cell, Vol.

6, 11–22, July, 2000, Copyright 2000 by Cell Press

Junctophilins: A Novel Family of Junctional


Membrane Complex Proteins

Hiroshi Takeshima,*†k Shinji Komazaki,‡ channel converts the depolarization signal into Ca2⫹ re-
Miyuki Nishi,*† Masamitsu Iino,† and Kenji Kangawa§ lease from the SR (Schneider, 1994). Previous studies
* Division of Cell Biology using mutant mice have demonstrated that the dihydro-
Institute of Life Science pyridine receptor (DHPR) and the ryanodine receptor
Kurume University (RyR) function as the voltage sensor and the SR Ca2⫹
Aikawa-machi, Kurume release channel, respectively, during excitation–con-
Fukuoka 839-0861 traction (E–C) coupling (Tanabe et al., 1988; Takeshima
† Department of Pharmacology et al., 1994). However, Chinese hamster ovary cells
Faculty of Medicine transfected with both DHPR and RyR cDNAs showed
University of Tokyo and neither Ca2⫹ release in response to depolarization nor
CREST, Japan Science and Technology Corporation close association between the PM and ER (Takekura et
Hongo, Bunkyo-ku al., 1995; Suda et al., 1997). Moreover, the triad junctions
Tokyo 113-8654 are still present in mutant skeletal muscle cells lacking
‡ Department of Anatomy either DHPR or RyR (Franzini-Armstrong et al., 1991;
Saitama Medical School Ikemoto et al., 1997). These results suggest that the
Moroyama-machi physiological coupling between DHPR and RyR requires
Saitama 350-0495 a junctional membrane complex formed by an as-yet-
§ Department of Biochemistry unidentified molecule. Although several transmembrane
National Cardiovascular Center Research Institute proteins with no established physiological roles have
Suita, Osaka 565-8565 been identified as components of triad junctions (Fan
Japan et al., 1995; Jones et al., 1995; Nishi et al., 1999), none
is a candidate for the molecule mediating physiological
coupling of the junctional membranes at present.
Summary In cardiac muscle, T-tubular and SR membranes form
junctional complexes, called “diad junctions,” in which
Junctional complexes between the plasma membrane DHPR and RyR function as the voltage-gated Ca2⫹ chan-
(PM) and endoplasmic/sarcoplasmic reticulum (ER/ nel and the Ca2⫹ release channel, respectively (Rayns
SR) are a common feature of all excitable cell types et al., 1968; Fabiato and Fabiato, 1978). In cardiac E–C
and mediate cross-talk between cell surface and intra- coupling, therefore, Ca2⫹ flows through DHPR, binds
cellular ion channels. We have identified the junctophi- to open RyR, and triggers the subsequent larger Ca2⫹
lins (JPs), a novel conserved family of proteins that release from the SR. This Ca2⫹ signal amplification is
are components of the junctional complexes. JPs are called Ca2⫹-induced Ca2⫹ release (Endo, 1985), and a
composed of a carboxy-terminal hydrophobic seg- similar mechanism has been found in smooth muscle
ment spanning the ER/SR membrane and a remaining cells and neurons (Verkhratsky and Shmigol, 1996).
Functional coupling between Ca2⫹-dependent chan-
cytoplasmic domain that shows specific affinity for
nels/transporters on the PM and Ca2⫹ release channels
the PM. In mouse, there are at least three JP subtypes:
on the ER/SR has also been proposed in several excit-
JP-1, -2, and -3. JP-2 is abundantly expressed in the
able cell types. Intracellular Ca2⫹ increases occur tran-
heart, and mutant mice lacking JP-2 exhibited embry-
siently in restricted regions, and the mobility of Ca2⫹ is
onic lethality. Cardiac myocytes from the mutant mice
highly restricted in the cytoplasm (Finch et al., 1991;
showed deficiency of the junctional membrane com-
Berridge, 1998). Thus, the Ca2⫹ signal amplification us-
plexes and abnormal Ca2ⴙ transients. Our results sug-
ing the intracellular Ca2⫹ store is thought to require the
gest that JPs are important components of junctional
junctional complexes between the PM and ER/SR, often
membrane complexes.
called the “subsurface cisternae” or “peripheral cou-
pling.” However, the molecular mechanisms underlying
Introduction the formation of the junctional membrane structures are
totally unknown. We describe here the identification and
Junctional membrane complexes between the plasma characterization of junctophilins (JPs), which are likely
membrane (PM) and the endoplasmic reticulum (ER) are to contribute to the stabilization of the junctional mem-
found in all excitable cells and are thought to provide a brane complexes by anchoring the ER/SR to the PM,
structural foundation for cross-talk between ion chan- thus providing a structural basis for physiological cou-
nels (Pozzan et al., 1994; Berridge, 1998). In skeletal pling between cell surface and intracellular channels.
muscle, invaginated cell surface (transverse [T] tubule)
and sarcoplasmic reticulum (SR) membranes form junc- Results and Discussion
tional complexes, designated “triad junctions” (Flucher,
1992), where it is proposed that direct coupling between Identification of JP-1 from Rabbit Skeletal Muscle
the cell surface voltage sensor and the Ca2⫹ release Triad Junction
To search for proteins supporting the structure of the
k To whom correspondence should be addressed (e-mail: takeshim@ triad junction, monoclonal antibodies (mAbs) were pre-
lsi.kurume-u.ac.jp). pared from mice immunized with membrane vesicles
Molecular Cell
12

Figure 1. Analysis of JP-1 in Rabbit Skeletal Muscle


A fluorescence-microscopic image of the longitudinal cryosection of skeletal muscle labeled by mAb2510 (A). Each sarcomere unit contains
a pair of immunolabeled rows assigned on the A–I junction. EM analysis of the longitudinal section ([B], arrow indicates the T tubule) and
triad junction ([C], arrow indicates the SR terminal cisternae) labeled with mAb2510 by immunocolloidal gold procedure. Z, Z line; T, transverse
tubule. Scale bars: (A), 10 ␮m; (B), 1 ␮m; (C), 0.5 ␮m. Of a total of 414 gold particles detected in the sections including 205 triad junctions,
409 particles were localized within 22 nm of the junctional-faced SR membrane cytoplasmically. In control experiments without mAb2510, 25
particles were detected in sections containing 648 triad junctions, which were distributed in cytoplasmic regions randomly. Immunoblotting
of membrane preparations from skeletal muscle with mAb2510 (D). Preparations of total microsome, T tubule, light SR, and heavy SR containing
enriched junctional SR (15 ␮g protein in each lane) were separated on an SDS-polyacryamide (4%–20%) gel and analyzed using mAb2510.
Size markers are indicated in kDa. Amino acid sequence of rabbit JP-1 deduced from cloned cDNA (E). The residues, numbered from the
initiating methionine, are shown in one-letter code, and numbers of the residues at the right-hand end of the individual lines are given. The
MORN motif (I–VIII) and transmembrane (TM) sequences are indicated by underlines and an open box, respectively. Hydropacithity profile of
rabbit JP-1 (F). The average hydropathicity index of the nonadecapeptide composed of amino acid residues i ⫺ 9 to i ⫹ 9 is plotted against
i, the amino acid number. The positions for MORN motif sequences and a putative transmembrane segment are indicated by boxes. Schematic
representation of proteins carrying MORN motif sequences (G). JP-1 and hypothetical proteins predicted from the C. elegans, A. thaliana,
and Cyanobacterum genomes share MORN motif sequences found eight times in the molecules. Putative transmembrane segments (TM) are
Identification of Novel Junctional Membrane Proteins
13

from rabbit skeletal muscle. Of the antibodies screened Generation of Junctional Membrane Complex
using the longitudinal skeletal muscle cryosections, in Cells Expressing JP-1
mAb2510 labeled intracellular rows oriented trans- To examine the biological functions of JP-1, amphibian
versely between the A and I bands, where triad junctions embryos, composed of large cells advantageous for an-
are localized (Figure 1A). In sections labeled with immu- alyzing morphological changes, were utilized as an ex-
nocolloidal gold for electron microscopy (EM), specific pression system. In the cells of the embryos injected
labeling was detected near the triad junction and fre- with cRNA for full-length JP-1, both the cytoplasmic foci
quently in the junctional gap between the T tubule and and PMs were specifically labeled with mAb2510 (Figure
SR (Figures 1B and 1C). Immunoblot analysis of micro- 2A). Cells localized in the peripheral region of the injec-
somal proteins (Figure 1D) showed that mAb2510 recog- tion site, where the cRNA level is presumably low, lacked
nized a protein of ⵑ72 kDa, which was highly enriched the cytoplasmic foci but retained the immunoreactive
in heavy SR preparations with a high content of junc- PMs. EM analysis revealed stacks of ER cisternae (Fig-
tional SR vesicles derived from triad junctions. These ure 2B) and junctional complexes between the ER and
results indicate that the antigen protein for mAb2510 is PM (Figures 2C and 2D) in the cells expressing JP-1; no
localized on the junctional SR in the triad junction. We such ultrastructures of ER elements could be detected in
call this antigen protein mitsugumin72 or junctophilin control cells (Komazaki et al., 1999). The abnormal ER
type 1 (JP-1). No obvious effects of N-ethylmaleimide stacks correlated well in size and structure with the
and N-glycosidase F on the migration of JP-1 were de- immunoreactive foci in the optical sections. As irregular
tected (data not shown), suggesting the absence of in- accumulation of lamellar ER membranes, sharing some
tra/intermolecular disulfide bridges and N-glycosylation morphological features with the ER stacks, has been
in the molecule. reported from cDNA expression of several ER mem-
After solubilization from microsomal preparations, brane proteins (for example, Jingami et al., 1987; Zimmer
JP-1 was immunoaffinity purified and subjected to se- et al., 1997), the ER stacks could be due to artificial
quence analysis. An oligonucleotide probe, synthesized effects common to overexpression of certain ER mem-
on the basis of the amino-terminal sequence, was used brane proteins. However, the junctional complexes be-
to screen a rabbit skeletal muscle cDNA library. The tween the ER and PM most likely correspond to the
cDNA clones thus isolated carry an open reading frame immunoreactive PMs observed histochemically, be-
that encodes 662 amino acid residues (Figure 1E). Hy- cause EM observation of ultrathin sections demon-
dropathicity analysis predicted a single transmembrane strated that more than 20% of the PMs were associated
segment in the carboxy-terminal end and no amino- with the junctional ER elements in the JP-1-expressing
terminal signal sequence in JP-1 (Figure 1F), suggesting cells (see figure legend). The junctional complexes
that the bulk of the protein is located in the cytoplasmic showed structural characteristics similar to those of the
region of the junctional gap between the SR and T-tubu- peripheral coupling detected commonly in muscle cell
lar membranes. Similarity matrix analysis revealed motif types (Flucher, 1992). The averaged gap size between
sequences of 14 residues, namely “MORN motif” se- the PM and ER in the junctional structures was ⵑ7.6
quences (membrane occupation and recognition nexus), nm (Figure 2D), which was narrower than that between
repeated eight times in the amino-terminal region. ER membranes in ER stacks (Figure 2B). Furthermore,
BLAST and FASTA searches found additional hypotheti- when the soluble form of JP-1 lacking the carboxy-termi-
cal proteins predicted from genomic sequences, which nal transmembrane segment was expressed, immuno-
carry MORN motif sequences (Figure 1G). We assigned labeling was detected specifically on the PM (Figure
the protein deduced from the C. elegans genome as the 2E), and neither the ER stacks nor ER-PM junctions were
nematode counterpart of JP based on overall homology detected in the cells. The subcellular distribution of the
and shared structural features. The protein deduced truncated JP-1 demonstrates the specific binding affin-
from the A. thaliana genome is composed of two do- ity of the cytoplasmic domain for the PM and that the
mains; an amino-terminal MORN motif region and a pu- membrane-spanning segment is essential for the gener-
tative carboxy-terminal catalytic region. The protein pre- ation of the junctional membrane structures.
dicted from the Cyanobacterium genome might be an To locate the sequence responsible for the interaction
adapter-like protein bearing MORN motif sequences. with the PM, JP-1-green fluorescent protein (GFP) fusion
Sequence comparison reveals a putative consensus se- proteins were constructed (Figure 2E). The fusion pro-
quence of MORN motif, “Tyr-Gln/Glu-Gly-Glu/Gln-Trp- teins carrying all the MORN motifs (FP-1 and -2) were
x-Asn-Gly-Lys-x-His-Gly-Tyr-Gly” (Figure 1H). These localized at the PM. FP-6 and -7 lacking the amino-
observations suggest that MORN motif is a novel pro- terminal two and four copies of the motif were also
tein-folding module shared by functionally different pro- detected on the PM. Surprisingly, FP-3 bearing the
teins and might have a specific physiological role. amino-terminal six copies of the motif was detected

indicated by filled boxes, and the region exhibiting high sequence similarities with mammalian phosphatidylinositol-4-phosphate (PIP)-5 kinase
subtypes is indicated by a shadowed box. Alignment of MORN motif sequences shared by rabbit JP-1, the nematode JP counterpart, plant
putative PIP kinase, and Cyanobacterium putative protein (H). In each protein, eight copies of MORN motif (I–VIII) are numbered from
the amino-terminal side. Sets of identical residues (ⱖ8/32 identity) are indicated by shadowing. The sequence data for the JP-1 counterpart
from C. elegans, putative PIP-5 kinase from A. thaliana, and hypothetical protein from Cyanobacteria have been taken from the databases
under accession numbers Z75550, AC004261, and D90899, respectively. Several variants of the putative PIP 5-kinase were deduced from the
A. thaliana genome in the database. The sequence data for rabbit JP-1 has been deposited in the databases under accession number
AB023447.
Molecular Cell
14

Figure 2. Analysis of Amphibian Embryonic Cells Expressing Various Forms of Rabbit JP-1
An immunofluorescence image of the cells expressing full-length JP-1 (A). Unique membrane features in the JP-1 expressing cells (B–D).
Immunofluorescence-positive cytoplasmic foci (arrows) and PMs in (A) are likely to correspond to ER stacks in (B) and junctional complexes
between the ER and PM in (C) and (D), respectively. Arrows indicate the ER-PM junctions in (C), and electron-dense circles are pigment
granules in (B) and (C). The gap sizes between the ER membranes in the ER stacks in (B) and between the PM and ER in the junctional regions
in (C) and (D) were 16.1 ⫾ 1.0 nm (mean ⫾ SD, n ⫽ 16) and 7.6 ⫾ 0.6 nm (n ⫽ 22), respectively. An immunofluorescence image of the cells
expressing the truncated form of JP-1 lacking the transmembrane segment (E). In the truncated JP-1, the carboxy-terminal 32 amino acid
residues were deleted. Subcellular distribution of expressed JP-1-GFP fusion proteins (F). The position for connecting with GFP is schematically
represented in the JP-1 structure (upper left). For all the fusion proteins, translation started from the initiation methionine in JP-1 and the 3⬘
terminus of the coding sequence was joined with GFP cDNA in frame. Scale bars: (A) and (E), 100 ␮m; (B) and (C), 1 ␮m; (D), 0.1 ␮m. In animal
polar cells from the embryos expressing full-length JP-1, out of a total length of 823 ␮m plasma membrane observed, 180 ␮m portions were
closely associated with the ER as shown in (C) and (D), and 226 junctional membrane structures were found. However, no junctional complexes
were observed in plasma membranes of cells from embryos injected with antisense JP-1 RNA (0 junction/1186 ␮m plasma membrane),
embryos expressing the truncated JP-1 in (E) (0/1339 ␮m), and embryos expressing FP-1 in (F) (0/1166 ␮m).

predominantly in nucleus, FP-4 carrying the amino-ter- fusion proteins was observed in cultured mammalian
minal four copies was expressed mainly in the nucleus cells (data not shown). These results indicate that the
and partly in the cytoplasm, and FP-8 carrying the car- MORN motifs contribute to the PM binding capacity of
boxy-terminal two copies was expressed on the PM and JP-1 and that the proposed interaction of JP-1 with
in the nucleus at similar density. FP-5 bearing a single the target component(s) on the PM depends upon the
copy of the motif was expressed mainly in the cyto- number and/or sequence of MORN motifs.
plasm. Essentially similar subcellular distribution of the Based on these results, we propose that JP-1 is in-
Identification of Novel Junctional Membrane Proteins
15

Figure 3. Mouse JP Subtypes in Excitable Tissues


The phylogenetic tree representing the molecular evolution of the JP family (A). On the basis of aligned amino acid sequences, the tree was
inferred by the neighbor-joining method using the computer program (Clustalw, the DDBJ database), and the branch lengths are proportional
to the numbers of amino acid substitutions. Confocal images of Madin-Darby canine kidney cells expressing JP–GFP fusion proteins (B).
Amino acid residues of JP subtypes constituting the fusion proteins are indicated in parentheses. Fluorescence signals were mainly detected
on the PMs in the cells. In about 30% of the cells expressing the JP-2-GFP fusion protein, fluorescence-positive nuclei were observed.
Northern blot analysis of JP subtypes in mouse tissues (C). Total RNA preparations (25 ␮g) were analyzed using subtype-specific hybridization
probes. Size markers are indicated in kb. Hybridization probes derived from the protein-coding and 3⬘-noncoding sequences produced
essentially the same results in each JP subtype. The accession numbers in the databases are AB023447 for rabbit JP-1, AB024445 for mouse
JP-1, AB024447 for mouse JP-2, AB024449 for mouse JP-3, and Z75550 for the nematode JP homolog.

volved in the construction of skeletal muscle triad junc- JP Subtypes in Excitable Tissues
tions by spanning the junctional SR membrane and inter- In an attempt to search for the JP family members by
acting with the T tubule. The gaps between T-tubular cross-hybridization, we isolated cDNAs encoding JP-1,
and SR membranes in the triad junctions from normal JP-2, and JP-3 from libraries derived from mouse skele-
muscle and mutant muscle lacking RyR are ⵑ12 nm and tal muscle, heart, and brain, respectively (Figure 3A).
7 nm, respectively, so the RyR probably restricts the The defined mouse JP subtypes, composed of 660–744
gap size (Ikemoto et al., 1997). The “foot” structure of amino acid residues, show homology in sequence (over-
RyR could not be detected in the junctional structures all ⵑ40% identity) and share structural features with
from the embryonic cells expressing JP-1 (Figure 2D), rabbit JP-1. The regions containing the MORN motif
and the gap size of 7.6 nm between PM and ER could sequences (ⵑ80% identity) and carboxy-terminal hy-
correspond to that of the mutant triad junction lacking drophobic segment spanning the ER/SR membrane
RyR. These observations might imply some elasticity in (ⵑ50% identity) are conserved well among the JP sub-
JP-1 depending on the presence or absence of RyR in types, whereas the regions of 211–286 residues immedi-
the junctional membrane complexes. ately preceding the transmembrane segment are highly
divergent (ⵑ6% identity). The high identity in eight
Our results from the JP-1 expression experiments,
MORN motif sequences suggests that all of the JP sub-
together with the fact that the proteins sharing MORN
types can interact with the PM like rabbit JP-1. Indeed,
motif sequences are found in organism of various spe-
fusion proteins, composed of the cytoplasmic regions
cies, further imply that the binding partner(s) of JP-1 on
of mouse JP subtypes and GFP, were all localized pre-
the PM is structurally conserved among eukaryotic cell
dominantly on the PM when expressed in cultured mam-
types. Protein overlay and surface plasmon assays sug- malian cells (Figure 3B).
gested that JP-1 can interact with phospholipids, espe- RNA blot experiments in mouse tissues (Figure 3C)
cially sphingomyelin and phosphatidylcholine (H. T., un- indicated that JP-1 is predominantly expressed in skele-
published data), and no protein associated with JP-1 tal muscle, JP-2 is distributed in both skeletal muscle
was detected in the immunoaffinity purification. Al- and heart, and JP-3 is abundantly expressed in the brain.
though the molecular mechanism underlying the spe- The lung and stomach, tissues containing smooth mus-
cific interaction of JP-1 with the PM remains to be further cle, showed weak hybridization signals with the JP-2-
investigated, it is likely that an affinity for phospholipids specific probe, suggesting expression of JP-2 in smooth
rather than protein–protein interaction is primarily re- muscle cells as well. Therefore, the JP subtypes seem
sponsible for the membrane association property of to be distributed throughout excitable tissues. The RNA
JP-1. species hybridizing with the JP-3-specific probe in the
Molecular Cell
16

Figure 4. Immunochemical Analysis of JP-2 and Generation of JP-2 Knockout Mice


Immunoblot analysis of JP-2 in adult mouse tissues (A). Total microsomal preparations (15 ␮g) were analyzed, and size makers are indicated
in kDa. In skeletal muscle, the immunoreactive band is rather broad because of its comigration with Ca2⫹-Mg2⫹ ATPase, the major protein
component in the SR. Immunohistochemical analysis of the hearts using JP-2-specific antibody (B and C). The cytoplasmic rows assigned
on the Z lines were immunofluorescence positive in mature cardiac myocytes from adult mice (B). Immunolabeling was observed at cell
periphery of cardiac myocytes from E9.5 embryonic mice (C). In cryosections of the whole embryos, the cardiac tubes were highly fluorescence
positive, but no obvious signals were detected in other tissues. Scale bar, 20 ␮m. Targeted mutation introduced into the mouse JP-2 gene
by homologous recombination (D). Restriction enzyme maps of the wild-type allele, targeting vector, and expected targeted allele are illustrated.
The first exon in the gene (E1), neomycin resistance gene (neo), and diphtheria toxin gene (DTA) are indicated. The genomic DNA probe used
for Southern blot screening with EcoRI (Probe 1) and PCR primers utilized for mouse genotyping (PJP2-5 and PJP2-4) are shown; the predicted
sizes of the detected DNA fragments are also given. The nucleotide sequence encompassing the exon 1 has been deposited to the nucleotide
databases with the accession number AB024448. Detection of the mutant JP-2 gene in E9.5 embryos (E). PCR with the primer set of PJP2-4
and PJP2-5 was carried out using genomic DNAs from embryos, and the resulting amplified products were analyzed on an agarose gel. The
size markers are indicated in base pairs. Detection of the JP-2 protein in the hearts of E9.5 embryos (F). Total proteins from the embryonic
hearts were examined in the immunoblot analysis using JP-2-specific antibody. Size markers are indicated in kDa.

testis was slightly smaller than JP-3 mRNA in the brain dancy might occur in a JP-1 knockout. The junctional
and seemed to be an aberrant product based on cDNA membrane structures have not been well characterized
cloning results (H. T., unpublished data). in the brain, which predominantly expresses JP-3. Thus,
to determine the basic biological function of JP sub-
Expression of JP-2 in Hearts and Preparation types, we focused on targeted disruption of the JP-2
of JP-2 Knockout Mice gene. In adult mouse heart, antibody to JP-2 mainly
Because both JP-1 and JP-2 are abundantly expressed recognized cytoplasmic rows along the Z lines within
in skeletal muscle (Figures 3C and 4A), functional redun- myocytes (Figure 4B). In mice on embryonic day (E)
9.5, the looped cardiac tube of the immature heart was
immunofluorescence positive, and the labeled region
Table 1. Genotype of Pups Obtained by Crossing Mice
was the periphery of cardiac myocytes (Figure 4C); such
Heterozygous in the Mutation Introduced to the JP-2 Gene
immunostaining signals were not detected in mutant
⫹/⫹ ⫹/⫺ ⫺/⫺ E9.5 embryos lacking JP-2 (see below). The SR forms
E9.5 44 72 40a the junctional complex, called the diad, with the T tubule
E10.5 10 27 11b running toward the Z line in mature cardiac myocytes
E11.5 5 18 5c (Rayns et al., 1968). Peripheral coupling between the
P0 (neonates) 18 35 0
PM and SR occurs in immature striated muscle cells
a
Embryos exhibited infirm heartbeats. (Flucher, 1992). The immunohistochemical results there-
b
More than half of embryos (7/11) exhibited cardiac arrest and con- fore suggest that JP-2 is localized on the junctional
gestive tissues. SR in the diad and peripheral coupling. Immuno-EM
c
Embryos exhibited autolysis after death.
analysis further indicated the localization of JP-2 to the
Identification of Novel Junctional Membrane Proteins
17

Figure 5. Junctional Membrane Structures in


Embryonic Cardiac Myocytes
Cardiac myocytes from wild-type E9.5 em-
bryos contained two types of junctional com-
plexes between the cell surface membrane
and the SR with gap sizes of ⵑ12 (A) and ⵑ30
nm (B). In the diad observed in mature cardiac
myocytes from adult mice, the gap size be-
tween the T-tubular and SR membranes was
ⵑ12 nm (C). Therefore, the junctional mem-
brane complex bearing the 12 nm gap is likely
to correspond to the structural foundation for
the physiological cross-talk between the
DHPR and RyR in the embryonic myocytes.
The connection between the SR and Z line is
common to striated muscles, and the struc-
tures are also found in immature cardiac myo-
cytes from E9.5 embryos (D). PM, plasma
membrane; SR, sarcoplasmic reticulum; T,
transverse tubule; Z, Z line. Scale bars:
(A)–(C), 0.1 ␮m; (D), 0.5 ␮m.

diad in adult cardiac myocytes (S. K. and H. T., unpub- Deficiency of Peripheral Coupling in JP-2 Knockout
lished data). Cardiac Myocytes
In the plasmid used for introducing a targeted muta- Junctional complexes containing the PM and SR were
tion, the genomic DNA segment containing the protein- examined in cardiac myocytes from E9.5 embryos. Car-
coding sequence of the first exon and partial sequence diac myocytes from wild-type embryos contained two
of the first intron in the JP-2 gene was replaced by the types of junctional membrane complexes with gap sizes
neomycin resistance gene (Figure 4D). In mutant mice of ⵑ12 and ⵑ30 nm (Figures 5A and 5B). Consistent
established using embryonic stem cells positive in with previous reports (for example, Fawcett and McNutt,
Southern blot screening, the targeted allele was de- 1969), the gap size in the diad from mature cardiac
tected by polymerase chain reaction (PCR) using prim- myocytes is always ⵑ12 nm (Figure 5C), suggesting
ers derived from the sequence of the JP-2 gene (Figure that the 12 nm junctions correspond to the functional
4E). The introduced mutation is thought to be a null peripheral couplings in embryonic myocytes. Statistical
mutation, because there was no detectable JP-2 protein analysis demonstrated that in the JP-2 knockout myo-
in hearts from homozygous E9.5 mutants (Figure 4F). cytes the appearance of the 12 nm junction was reduced
Targeted disruption of the JP-2 gene induced embryonic to only ⵑ10% of the control value (Table 2). Furthermore,
lethality in the homozygous state (Table 1). In the homo- the average length of the 12 nm junctional membrane
zygous E9.5 mutants, the hearts showed spontaneous complex in the mutant myocytes (0.17 ⫾ 0.06 ␮m, n ⫽
contractions, but the heartbeats were often weak and 18; mean ⫾ SD) was significantly shorter than that in
irregular. More than half of the E10.5 mutants exhibited controls (0.37 ⫾ 0.16 ␮m, n ⫽ 23). No differences be-
cardiac arrest and congested peripheral tissues, and tween the genotypes (Figure 5D) were detected in either
the E11.5 mutants showed autolysis after death. Al- the 30 nm junction or the close association between the
though JP-2 expression is observed throughout muscle SR and Z line, namely “Z tubule” (for example, Simpson
cell types in adult mice, functional skeletal and smooth and Rayns, 1968).
muscle cells are not developed in the early embryonic The deficiency of the peripheral coupling, demon-
stages, so the lethality seen in the mutant embryos is strated in the mutant myocytes prior to cardiac arrest,
most likely caused by the damage to the cardiac myo- supports our hypothesis that the JP subtypes contribute
cytes. to the formation of junctional membrane complexes in

Table 2. Characteristic SR Features in Cardiac Myocytes from Wild-Type and JP-2 Knockout E9.5 Mice
12 nm Junctiona 30 nm Junctiona SR–Z Connectionb
Wild-type myocytes 12.4 ⫾ 0.2 2.2 ⫾ 0.3 91 ⫾ 2.2
JP-2-deficient myocytes 1.5 ⫾ 0.7** 2.2 ⫾ 0.9 91 ⫾ 2.0
a
Number of junctions per 100 ␮m plasma membrane.
b
Percentage of SR-bearing Z line.
The data (mean ⫾ SD) were obtained from EM observations of cardiac myocytes from three embryos; the data from each embryos were
combined to examine the statistical differences. In wild-type myocytes (82 cells), total 4220 ␮m of the plasma membrane and 156 of Z lines
were analyzed. In JP-2-deficient myocytes (90 cells), 4266 ␮m of the plasma membrane and 135 of Z lines were observed. The junctional
membrane structures bearing 11–15 nm and 22–37 nm gaps were assigned to the 12 and 30 nm junctions, respectively. Statistical differences
are indicated by asterisks (**p ⬍ 0.01 in Student’s t test).
Molecular Cell
18

Figure 6. Abnormal Ca2⫹ Transients in Cardiac Myocytes from the E9.5 JP-2 Knockout Embryos
Infirm Ca2⫹ transients during spontaneous oscillations in the JP-2 knockout hearts (A). Averaged intracellular Ca2⫹ concentrations of myocytes
including a portion of embryonic hearts (ⵑ0.1 mm2) were measured in the normal bathing solution containing 2 mM Ca2⫹. Representative data
are shown from experiments in wild-type (n ⫽ 3 out of 8 embryos) and mutant hearts (n ⫽ 6 out of 10). The random Ca2⫹ transients in the
mutant hearts and the resistance to Ca2⫹-free conditions (B). Intracellular Ca2⫹ changes during spontaneous oscillations in wild-type (upper
Identification of Novel Junctional Membrane Proteins
19

various cell types. Generation of the junctional mem- transients in the mutant hearts appear to be independent
brane complexes may require at least two processes: on Ca2⫹ influx through DHPR.
(1) approach of the SR toward the PM, and (2) formation RyR-mediated Ca2⫹ release from the SR amplifies the
of a stable complex between the membranes. The ultra- Ca2⫹ transient more than 10-fold during E–C coupling
structural data in the mutant myocytes, together with in mature myocytes from the adult hearts (Nabauer et al.,
our biochemical data on JP-1, suggest that JPs are 1989). However, in embryonic cardiac myocytes, Ca2⫹
responsible for the latter process. It is a reasonable influx is the predominant component in the spontaneous
assumption that the former process requires complex oscillations (Fisher, 1995). Therefore, the amplitude of
cellular functions including vesicular transport and is the spontaneous Ca2⫹ transients was slightly reduced
not affected by the loss of JP-2 in cardiac myocytes. under store-depleted conditions by combined caffeine
The few peripheral couplings retained in the mutant and ryanodine application (caffeine opens RyR channels
myocytes might correspond to unstable forms of the and ryanodine locks RyR channels in their open states)
junctional membrane complexes. Alternatively, other JP in wild-type hearts from E9.5 embryos (Figure 6C). How-
subtypes might contribute to the formation of the periph- ever, the random Ca2⫹ transients in the mutant hearts
eral couplings in the mutant myocytes, because low were abolished by the application of caffeine and rya-
levels of JP-1 expression in the adult hearts were de- nodine. Because only the RyR type 2 (RyR-2) is ex-
tected by Northern blot analysis (Figure 3C). pressed in the embryonic cardiac myocytes (Takeshima
et al., 1998a), the random transients observed in the
mutant myocytes must have been evoked by Ca2⫹ re-
Random Ca2ⴙ Transients in JP-2 Knockout
lease from the SR through RyR-2. To further characterize
Cardiac Myocytes
the random Ca2⫹ transients, mutant cardiac myocytes
To survey the functional abnormalities in the JP-2
were analyzed at the subcellular level (Figure 6D). The
knockout mice, the hearts from E9.5 embryos were sub-
random Ca2⫹ transient was always generated within a
jected to Ca2⫹-imaging analysis using an intracellular
single cell and did not propagate to neighboring myo-
Ca2⫹ indicator. In wild-type hearts, the Ca2⫹ transients
cytes. Furthermore, intracellular Ca2⫹ waves were ob-
in spontaneous oscillations corresponding to rhythmic
served during the random Ca2⫹ transients with the slow
contractions were almost constant. By contrast, in the
time courses in the mutant myocytes.
JP-2-deficient hearts, the Ca2⫹ transients were varied
in amplitude among the embryos, and some of them
were almost eliminated (Figure 6A). The impaired Ca2⫹ Relationship of Ultrastructural and Physiological
transients probably cause the infirm beating observed Defects in JP-2 Knockout Cardiac Myocytes
in the mutant heart, and the variations of the Ca2⫹ tran- RyR-2 expressed in embryonic cardiac myocytes is
sients among the mutant embryos likely reflect the prone to regenerative Ca2⫹ release responses even un-
stages of a progressive lethal defect. der resting Ca2⫹ levels because of its high Ca2⫹ sensitiv-
Because cardiac E–C coupling requires Ca2⫹ influx ity. RyR-2 produces Ca2⫹ waves and oscillations inde-
via DHPR, heartbeats are abolished under Ca2⫹-free pendent of Ca2⫹ influx when expressed in cultured
conditions. In wild-type hearts, all of the myocytes skeletal myocytes, in which the SR has high Ca2⫹ con-
showed synchronized Ca2⫹ transients, and the tran- tents (Yamazawa et al., 1996). Moreover, abnormal intra-
sients disappeared in a Ca2⫹-free bathing solution (Fig- cellular Ca2⫹ waves have been observed in cardiac myo-
ure 6B). However, in mutant hearts from the JP-2 knock- cytes when the Ca2⫹ loading of the SR was increased
out embryos, a large number of myocytes showed and have been implicated in arrhythmia (Wier et al.,
irregular Ca2⫹ transients that were not synchronized with 1987; Berlin et al., 1989). The random Ca2⫹ transients
the heartbeats and occurred randomly. Normal Ca2⫹ accompanied by Ca2⫹ waves observed in the JP-2
transients in wild-type hearts were completed within 0.9 knockout hearts share characteristic features with those
s, whereas the random transients in the mutant hearts reported previously, suggesting that an overloaded SR
lasted for 1.7 ⫾ 0.45 s (mean ⫾ SD). Although the random was generated in the mutant myocytes (Figure 7A).
transients were observed in all mutant hearts examined, In mutant cardiac myocytes from RyR-2 knockout
the frequency of myocytes showing the abnormal tran- mice, which exhibit embryonic lethality at ⵑE10, the SR
sients was higher in the mutant hearts with the most was vacuolated and Ca2⫹ overloaded (Takeshima et al.,
infirm heartbeats. The random transients were observed 1998a). In other words, SR Ca2⫹ levels can be increased
even in Ca2⫹-free bathing solution, but the frequency by the reduction of RyR-2-mediated Ca2⫹ release in em-
was significantly reduced. Therefore, the abnormal Ca2⫹ bryonic cardiac myocytes. It is possible that the loss of

traces) and JP-2 knockout (lower traces) cardiac myocytes were measured in the normal bathing solution and the Ca2⫹-free solution containing
1 mM EGTA. The myocytes, examined for traces (1–3), are indicated by white circles in the fluorescence image of the embryonic heart.
Representative data are shown from the experiments on wild-type (n ⫽ 7) and mutant hearts (n ⫽ 8). Abolition of the random Ca2⫹ transients
in the mutant hearts under Ca2⫹-store-depleted conditions (C). The effects of the application of 20 mM caffeine (Caf) and 100 ␮M ryanodine
(Ry) on the Ca2⫹ transients were examined in the normal bathing solution. Representative data are shown from the experiments on wild-type
(n ⫽ 7) and mutant hearts (n ⫽ 7). In wild-type hearts, rhythm disturbances of heart beating were often observed after the caffeine applications
with or without ryanodine. Spatial and temporal patterns of the random Ca2⫹ transient in a single JP-2 knockout cardiac myocyte (D). The
upper traces show the time courses of changes in fluorescence intensity during a Ca2⫹ transient; the intracellular regions examined are
indicated in the fluorescence image of the myocyte (a–c). Intracellular Ca2⫹ concentrations during the Ca2⫹ transient are shown in pseudocolor
images at the frames indicated in the upper panel (1–8). The vertical scales are indicated in ⌬F/F0.
Molecular Cell
20

Figure 7. Proposed Functional Abnormality


and Observed Ultrastructural Defects in JP-2
Knockout Cardiac Myocytes
Proposed roles of JP-2 and RyR-2 on the SR
in embryonic cardiac myocytes (A). Based on
the results of gene targeting studies, major
intracellular Ca2⫹ flows are illustrated. Even
though it contributes only slightly to Ca2⫹
transients during E–C coupling, RyR-2-medi-
ated Ca2⫹ release coupled with Ca2⫹ influx
via DHPR is essential for regulating Ca2⫹ con-
tents of the SR to maintain the cellular func-
tions in embryonic cardiac myocytes. JP-2-
mediated formation of peripheral coupling
appears to be essential for functional cross-
talk between DHPR and RyR (see text). Ultra-
structural defects of intracellular organelles
in the JP-2 and RyR-2 knockout cardiac myo-
cytes (B). Structures of the rough ER and mi-
tochondria were analyzed in cardiac myo-
cytes from E10.5 wild-type embryos, E10.5
JP-2 knockout embryos retaining infirm
heartbeats and E9.5 RyR-2 knockout em-
bryos. In the JP-2 knockout myocytes, the
rough ER (rER) was fragmented, membrane-
bound ribosomes were significantly reduced,
and mitochondria showed abnormal struc-
tures. In the RyR-2 knockout myocytes, the
vacuolated ER/SR carried a few membrane-
bound ribosomes and the inner structures of
swollen mitochondria were degraded. Arrows
indicate ER membranes carrying ribosomes.
Scale bar, 1 ␮m.

JP-2 disconnects physiological coupling between cell What is the cellular mechanism underlying the diminu-
surface DHPR and RyR-2 on the SR, because a defi- tion of spontaneous Ca2⫹ oscillations in the mutant myo-
ciency in functional peripheral coupling could prevent cytes? Ultrastructural defects of intracellular organelles
close association of the channel molecules and interfere (Figure 7B) suggest that the disruption of Ca2⫹ homeo-
with effective activation of RyR-2 by DHPR-mediated stasis produces toxic effects in embryonic cardiac myo-
Ca2⫹ influx. The resulting reduction of SR Ca2⫹ release cytes. In cardiac myocytes from the E9.5 JP-2 knockout
could produce a Ca2⫹-overloaded SR in the mutant myo- embryos, some mitochondria carried irregular struc-
cytes. The overloaded SR could release Ca2⫹ randomly tures, but no other obvious abnormalities were detected.
through RyR-2, and the generated Ca2⫹ rise could be In the myocytes from the E10.5 JP-2 knockout embryos
expanded to neighboring SR regions by Ca2⫹-induced exhibiting infirm heartbeats, severe ultrastructural ab-
Ca2⫹ release (CICR) mechanism to produce the intracel- normalities were developed (e.g., swollen mitochondria
lular Ca2⫹ wave. The random Ca2⫹ transients in the mu- with degraded cristae, fragmented rough ER elements,
tant heart probably produce local extrasystoles and in- and the reduction of membrane-bound ribosomes). Sim-
firm heartbeats. Therefore, our results support the ilar ultrastructural abnormalities were detected in car-
recent proposal that the functional uncoupling between diac myocytes from the RyR-2 knockout embryos. The
DHPR and RyR may be involved in cardiac hypertrophy defects were specific to cardiac myocytes, and they
and heart failure (Cannell and Soeller, 1998). Further- could not be detected in cells from neural tubes, primi-
more, it is possible that mutations in JP-2 are a risk tive connective tissues, or epithelia (data not shown).
factor for the heart disease. Deficiency of the ribosomes likely reduces protein syn-
Why could the JP-2 knockout myocytes not recover thesis, and mitochondrial dysfunction may cause energy
Ca2⫹ transients under the store-depleted conditions? depletion. Moreover, irregular intracellular Ca2⫹ signal-
Identification of Novel Junctional Membrane Proteins
21

ing could interfere with Ca2⫹-dependent enzyme system hamster ovary and Madin-Darby canine kidney cells were trans-
and gene expression in the mutant myocytes. Such ab- fected with the expression plasmids using a commercial reagent
(Lipofectamine, GIBCO-BRL). After 2–3 days, the cells were fixed
normalities could cause unrecoverable damage to the
with acetone for immunostaining or were observed directly by a
mutant cardiac myocytes. In the JP-2 knockout mice, confocal microscope (FV300, Olympus).
therefore, both dysfunction of cardiac myocytes and
impaired pump function of the heart may contribute to Immunochemical Analysis of Mouse JP-2
the embryonic lethality. The mouse cDNA fragment encoding the divergent regions of JP-2
(amino acid residues 436–664) was cloned into the polylinker site in
pGEX-2T (Amersham Pharmacia Biotech) to produce the glutathione
Experimental Procedures S–transferase fusion proteins. Antibody was affinity purified using
Protein G– and fusion protein–conjugated resins from the antiserum
Immunochemical Analysis and Purification of Rabbit JP-1 derived from rats immunized with the fusion protein. To detect JP-2
Membranes of total microsome, T tubule, light SR, and heavy SR in E9.5 embryos in immunoblotting, the cardiac tubes were removed
were prepared from adult rabbit skeletal muscle as reported pre- and homogenized in an SDS-PAGE sampling buffer. After electro-
viously (Takeshima et al., 1998b). In our membrane preparations, phoresis, blotting, and reactions of primary and peroxidase-con-
DHPR as a T-tubular maker protein is abundant in the T tubule jugated secondary antibodies, immunological signals on resulting
fraction and is also distributed in the heavy SR fraction, while RyR filters were visualized after high enhancement using a blotting ampli-
as a junctional SR marker protein is specifically found in the heavy fication system (BLAST, NEN Life Science Products). For histochem-
SR fraction. Preparation of a mAb library to the triad junction, screen- ical analysis, the antibody reacted with cryosections of acetone-
ing methods of mAbs, immunoelectronmicroscopic, and immu- fixed tissues was detected with the secondary antibody conjugated
noblot analyses in skeletal muscle were described previously with TRITC, and the sections were then observed by the fluores-
(Takeshima et al., 1998b). cence microscopy.
JP-1 was highly resistant to solubilization from membrane vesicles
by ordinarily used detergents (NP-40, Triton X-100, or CHAPS) under Generation and Morphological Analysis of JP-2 Knockout Mice
various NaCl concentrations tested. Total microsomal preparations The generation of JP-2 knockout mice was carried out as described
from rabbit skeletal muscle were solubilized in the buffer containing previously (Takeshima et al., 1994). The targeting vector was con-
1% SDS and 50 mM Tris-HCl (pH 7.4). After ten times dilution with structed with the genomic DNA fragments from the JP-2 gene, the
the buffer containing 1% Triton X-100 and 50 mM Tris-HCl (pH 7.4), neomycin resistance gene from pMC1 Neo polyA, the diphtheria
the solubilized proteins were reacted with mAb2510-affinity resin. toxin gene from pMC1-DT-A, and pBluescript SK(⫺) as shown in
After extensive washing, the absorbed proteins were recovered with Figure 4D. In Southern blot screening, the J1 ES cells transfected
an elution buffer containing 1.5% SDS. Eluted JP-1 was transferred with the vector, several clones carrying the expected homologous
to a Problot membrane (Applied Biosystems) and was subjected to mutation were isolated. Chimeric mice produced with the ES clone
amino acid sequence analysis using a protein sequencer (Model numbered 121 were crossed with C57BL/6J mice and could transmit
494, Applied Biosystems). The determined amino-terminal sequence the mutant gene to the pups. The JP-2 knockout and wild-type mice
of 19 residues corresponds to that of the residues 2–20 in the primary obtained by crossing the heterozygous mutants were used for the
structure deduced by cDNA cloning, demonstrating that the mature analyses in this report. To determine the mouse genotypes, PCR
form of JP-1 lacks the initiating methionine. was carried out using the synthetic primers; the sequences are
PJP2-5 (26-mer, GTGAGTCTAGGGGCTCAACCTGAAGG) and PJP2-4
Cloning of JP Subtypes (26-mer, TAAAATCTAGCTGCTTACATACCTCG). Tissues from mouse
A rabbit skeletal muscle cDNA library (Takeshima et al., 1989) was embryos were analyzed at the histological and ultrastructural levels
screened with the 5⬘-32P-labeled probe (32 species-mixed 17-mer: as described previously (Takeshima et al., 1998a).
5⬘-TT[T/C]GA[T/C]TT[T/C]GA[T/C]GA[T/C]GG-3⬘) to yield several hy-
bridization-positive clones. Although there are no base changes Ca2ⴙ Measurements in Embryonic Cardiac Myocytes
in the protein-coding sequence among the clones, three types of Mouse embryonic heart tubes were fixed on silicone rubber, sub-
splicing variants in the 3⬘-noncoding sequence were found (see the merged in Fluo-3AM-containing medium, and then placed on a
databases). cDNA libraries, prepared using poly(A)⫹ RNAs derived glass-bottomed perfusion chamber containing a physiological salt
from mouse tissues, were screened with rabbit JP-1 cDNA frag- solution (PSS: 150 mM NaCl, 4 mM KCl, 2 mM CaCl2, 1 mM MgCl2,
ments as probes under low-stringency hybridization conditions. The 5 mM HEPES, and 5.6 mM glucose [pH 7.4]), as described previously
protein-coding sequences in clones hybridized with the probes were (Takeshima et al., 1998a). A cooled CCD camera mounted on the
deduced by comparing with the rabbit JP-1 cDNA sequence. Re- microscope equipped with a polychromator (Merlin system, Olym-
striction enzyme fragments from isolated cDNAs were used as hy- pus) was used to capture the fluorescence images with excitation
bridization probes for RNA blot analysis to examine the tissue- at 470 nm and emission at ⬎510 nm at approximately four or six
specific distribution of the JP subtypes as described previously frames per second at room temperature.
(Takeshima et al., 1998b). A mouse genomic library (Stratagene)
was screened with isolated JP-2 cDNA as a probe, and genomic Acknowledgments
segments derived from the JP-2 gene were obtained.
We thank Ms. I. Dobashi and Misses A. Yoshida, A. Sakamoto, and
Expression from JP Subtype cDNAs M. Matsunaga for their expert assistance. This work was supported
For expression of the full-length rabbit JP-1, the truncated form in part by grants from the Ministry of Education, Science, Sports,
lacking the putative transmembrane segment and the JP-1 deletion and Culture of Japan, the Ministry of Health and Welfare of Japan,
mutants fused to GFP derived from pEGFP-1 (Clontech), the corre- the TMFC, the Ichiro Kanehara Memorial Foundation, and the Ina-
sponding cDNAs were cloned into pBluescript SK(⫺) (Stratagene). mori Foundation.
Embryos of the Japanese newt, Cynops pyrrhogaster, had injections
of cRNAs prepared by in vitro transcription of the plasmids and Received December 6, 1999; revised May 8, 2000.
were incubated for 1–2 days as described previously (Komazaki et
al., 1999). The embryos were fixed in acetone for immunostaining References
or direct observation under the fluorescence microscope (IX-FLA,
Olympus). For the ultrastructural analysis under EM (JEM-200CX, Berlin, J.R., Cannell, M.B., and Lederer, W.J. (1989). Cellular origins
JEOL), the embryos were fixed and treated as described previously of the transient inward current in cardiac myocytes. Role of fluctua-
(Takeshima et al., 1998a). For cDNA expression in cultured cells, tions and waves of elevated intracellular calcium. Circ. Res. 65,
the protein-coding sequences for various forms of JP subtypes were 115–126.
cloned into the expression vector pcDNA3.1(⫹) (Invitrogen). Chinese Berridge, M.J. (1998). Neuronal calcium signaling. Neuron 21, 13–26.
Molecular Cell
22

Cannell, M.B., and Soeller, C. (1998). Sparks of interest in cardiac K., Minamino, N., Matsuo, H., Ueda, M., Hanaoka, M., Hirose, T.,
excitation–contraction coupling. Trends Pharmacol. Sci. 19, 16–20. and Numa, S. (1989). Primary structure and expression from comple-
Endo, M. (1985). Ca2⫹ release from sarcoplasmic reticulum. Curr. mentary DNA of skeletal muscle ryanodine receptor. Nature 339,
Topic. Membr. Transp. 25, 181–230. 439–445.
Fabiato, A., and Fabiato, F. (1978). Calcium triggered release of Takeshima, H., Iino, M., Takekura, H., Nishi, M., Kuno, J., Minowa,
calcium from the sarcoplasmic reticulum of skinned cells of adult O., Takano, H., and Noda, T. (1994). Excitation-contraction uncou-
human, dog, cat, rabbit, rat and frog heart and fetal and newborn pling and muscular degeneration in mice lacking functional skeletal
rat ventricles. Ann. NY Acad. Sci. 307, 491–522. muscle ryanodine-receptor gene. Nature 369, 556–559.
Fan, H., Brandt, N.R., and Caswell, A.H. (1995). Disulfide bonds, Takeshima, H., Komazaki, S., Hirose, K., Nishi, M., Noda, T., and
N-glycosylation and transmembrane topology of skeletal muscle Iino, M. (1998a). Embryonic lethality and abnormal cardiac myocytes
triadin. Biochemistry 34, 14902–14908. in mice lacking ryanodine receptor type 2. EMBO J. 17, 3309–3316.
Fawcett, D.W., and McNutt, N.S. (1969). The ultrastructure of the Takeshima, H., Shimuta, M., Komazaki, S., Ohmi, K., Nishi, M., Iino,
cat myocardium. I. Ventricular papillary muscle. J. Cell Biol. 42, 1–45. M., Miyata, A., and Kangawa, K. (1998b). Mitsugumin29, a novel
synaptophysin family member from the triad junction in skeletal
Finch, E.A., Turner, T.J., and Goldin, S.M. (1991). Calcium as a coag-
muscle. Biochem. J. 331, 317–322.
onist of inositol 1,4,5-trisphosphate-induced calcium release. Sci-
ence 252, 443–446. Tanabe, T., Beam, K.G., Powell, J.A., and Numa, S. (1988). Restora-
tion of excitation-contraction coupling and slow calcium current in
Fisher, D.J. (1995). Recent insights into the regulation of cardiac
dysgenic muscle by dihydropyridine receptor complementary DNA.
Ca2⫹ flux during perinatal development and in cardiac failure. Curr.
Nature 336, 134–139.
Opin. Cardiol. 10, 44–51.
Verkhratsky, A., and Shmigol, A. (1996) Calcium-induced calcium
Flucher, B.E. (1992). Structural analysis of muscle development:
release in neurons. Cell Calcium 19, 1–14.
transverse tubules, sarcoplasmic reticulum, and the triad. Dev. Biol.
154, 245–260. Wier, W.G., Cannell, M.B., Berlin, J.R., Marban, E., and Lederer, W.J.
(1987). Cellular and subcellular herterogeneity of [Ca2⫹]I in single
Franzini-Armstrong, C., Pincon-Raymond, M., and Rieger, F. (1991).
heart cells revealed by fura-2. Science 235, 325–328.
Muscle fibers from dysgenic mouse in vivo lack a surface compo-
nent of peripheral couplings. Dev. Biol. 146, 364–376. Yamazawa, T., Takeshima, H., Sakurai, T., Endo, M., and Iino, M.
(1996). Subtype specificity of the ryanodine receptor for Ca2⫹ signal
Ikemoto, T., Komazaki, S., Takeshima, H., Nishi, M., Noda, T., Iino,
amplification in excitation–contraction coupling. EMBO J. 15, 6172–
M., and Endo, M. (1997). Functional and morphological features of
6177.
myocytes from mutant mice lacking both ryanodine receptors type
1 and type 3. J. Physiol. 501, 305–312. Zimmer, T., Vogel, F., Ohta, A., Takagi, M., and Schunck, W.H. (1997).
Protein quality. A determinant of the intracellular fate of membrane-
Jingami, H., Brown, M.S., Goldstein, J.L., Anderson, R.G., and Luskey,
bound cytochromes P450 in yeast. DNA Cell Biol. 16, 501–514.
K.L. (1987). Partial deletion of membrane-bound of 3-hydroxy-3-
methylglutaryl coenzyme A reductase eliminates sterol-enhanced
DDBJ/GenBank/EMBL Accession Numbers
degradation and prevents formation of crystalloid endoplasmic re-
ticulum. J. Cell Biol. 104, 1693–1704.
Sequence data reported in this article will appear in the DDBJ/
Jones, L.R., Zhang, L., Sanborn, K., Jorgensen, A.O., and Kelley, J. GenBank/EMBL nucleotide databases under accession numbers
(1995). Purification, primary structure, and immunological character- AB023447 for rabbit JP-1 cDNA, AB024445 for mouse JP-1 cDNA,
ization of the 26-kDa calsequestrin binding protein (junctin) from AB024447 for mouse JP-2 cDNA, AB024449 for mouse JP-3 cDNA,
cardiac junctional sarcoplasmic reticulum. J. Biol. Chem. 270, and AB024448 for the mouse JP-2 genomic segment.
30787–30796.
Komazaki, S., Nishi, M., Kangawa, K., and Takeshima, H. (1999).
Immunolocalization of mitsugumin29 in developing skeletal muscle
and effects of the protein expressed in amphibian embryonic cells.
Dev. Dynam. 215, 87–95.
Nabauer, M., Callewaert, G., Cleenman, L., and Morad, M. (1989).
Regulation of calcium release is gated by calcium current, not gating
charge, in cardiac myocytes. Science 244, 800–803.
Nishi, M., Komazaki, S., Kurebayashi, N., Ogawa, Y., Noda, T., Iino,
M., and Takeshima, H. (1999). Abnormal features in skeletal muscle
from mice lacking mitsugumin29. J. Cell Biol. 147, 1473–1480.
Pozzan, T., Rizzuto, R., Volpe, P., and Meldolesi, J. (1994). Molecular
and cellular physiology of intracellular calcium stores. Physiol. Rev.
74, 595–636.
Rayns, D.G., Simpson, F.O., and Bertaud, W.S. (1968). Surface fea-
tures of striated muscle I. Guinea-pig cardiac muscle. J. Cell Sci.
3, 467–474.
Schneider, M.F. (1994). Control of calcium release in functioning
skeletal muscle fibers. Annu. Rev. Physiol. 56, 463–484.
Simpson, F.O., and Rayns, D.G. (1968). The relationship between
the transverse tubular system and other tubules at the Z disc levels
of myocardial cells in the ferret. Am. J. Anat. 122, 193–208.
Suda, N., Franzius, D., Fleig, A., Nishimura, S., Bodding, M., Hoth,
M., Takeshima, H., and Penner, R. (1997). Caffeine-induced release
of intracellular Ca2⫹ from Chinese hamster ovary cells expressing
skeletal muscle ryanodine receptor. J. Gen. Physiol. 109, 619–631.
Takekura, H., Takeshima, H., Nishimura, S., Takahashi, M., Tanabe,
T., Flockerzi, V., Hofmann, F., and Franzini-Armstrong, C. (1995).
Co-expression in CHO cells of two muscle proteins involved in exci-
tation–contraction coupling. J. Muscle Res. Cell Motil. 16, 465–480.
Takeshima, H., Nishimura, S., Matsumoto, T., Ishida, H., Kangawa,

Das könnte Ihnen auch gefallen