Identification of Novel Interactions Between Domains of
Myosin Binding Protein-C That Are Modulated by
Hypertrophic Cardiomyopathy Missense Mutations
Johanna Moolman-Smook, Emily Flashman, Willem de Lange, Zhili Li, Valerie Corfield,
Charles Redwood, Hugh Watkins
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Abstract—Cardiac myosin binding protein-C (cMyBPC) is a modular protein consisting of 11 domains whose precise
function and sarcomeric arrangement are incompletely understood. Identification of hypertrophic cardiomyopathy
(HCM)– causing missense mutations in cMyBPC has highlighted the significance of certain domains. Of particular
interest is domain C5, an immunoglobulin-like domain with a cardiac-specific insert, which is of unknown function yet
is the site of two HCM-causing missense mutations. To identify interactors with this region, a human cardiac cDNA
library was screened in a yeast two-hybrid (Y2H) assay using the C5 sequence as bait. Screening ⬎7⫻106 clones
surprisingly revealed that domain C5 preferentially bound to clones encoding C-terminal fragments of cMyBPC; the
interacting region was narrowed to domain C8 by deletion mapping. A surface plasmon resonance assay using purified
recombinant cMyBPC domains was used to measure the affinity of C5 and C8 in vitro (Ka⫽1⫻105 mol/L⫺1). This
affinity was decreased about 2-fold by the HCM mutation R654H, and by at least 10-fold by the mutation N755K.
Further Y2H assays also demonstrated specific binding between domains C7 and C10 of cMyBPC. Based on these novel
interactions, and previous biochemical and structural data, we propose that cMyBPC molecules trimerize into a collar
around the thick filament, with overlaps of domains C5-C7 of one cMyBPC with C8-C10 of another. We speculate that
this interaction may be dynamically formed and released, thereby restricting or favoring cross-bridge formation,
respectively. We suggest that the HCM mutations act by altering the cMyBPC collar, indicating its importance in thick
filament structure and regulation. (Circ Res. 2002;91:704-711.)
Key Words: cardiac myosin binding protein-C 䡲 hypertrophic cardiomyopathy
䡲 missense mutations 䡲 ligand binding
M
yosin binding protein-C (MyBPC) was first identified
as a contaminant of muscle myosin preparations in the
1970s1 and has since been shown to be an essential component of vertebrate striated muscle. Immunolocalization studies have shown that MyBPC is bound to thick filaments in
distinct stripes (between 7 and 11) in the C-zone of the
A-bands in each half-sarcomere.2 Its precise quaternary
structure on the myosin filament, however, remains controversial, as does its function, with both structural and regulatory roles proposed.3,4
Fast and slow skeletal muscle isoforms of MyBPC, and a
single cardiac isoform, are each encoded by a separate
gene.5,6 The sequences of these genes reveal that MyBPCs
consist of repeated domains with homology to immunoglobulin C2 (Ig-like) or fibronectin type 3 (FN3) domains (Figure
1). The first and second Ig-like domains (C1 and C2) are
separated by a unique 103-residue stretch, termed the MyBPC
motif,6 while linkers of varying sizes separate other domains
(Figure 1).7 Cardiac MyBPC (cMyBPC) differs from the
skeletal isoforms in that it contains an additional aminoterminal Ig-like domain (C0), as well as an insertion of
LAGGGRRIS residues and three cardiac isoform-specific
phosphorylation sites in the MyBPC motif, and a proline/
charged residue-rich insertion into domain C5 (Figure 1).
Knowledge of the gene and protein structure facilitated
investigation of the roles of MyBPC in the sarcomere and led
to elucidation of the functions of specific domains. A structural role was first confirmed when the primary myosin and
titin binding regions were localized to within C10 and
C8-C10, respectively.7,8 It was also shown that, for MyBPC
to be correctly incorporated into the thick filament, domain
C7 is required in addition.9 A regulatory role for the cardiac
isoform was indicated by a correlation between the rate of
twitch relaxation and the catecholamine-sensitive phosphor-
Original received April 29, 2002; revision received August 28, 2002; accepted August 29, 2002.
From the US/MRC Centre for Molecular and Cellular Biology (J.M.-S., W.d.L., V.C.), University of Stellenbosch, Tygerberg, South Africa; and the
Department of Cardiovascular Medicine (E.F., Z.L., C.R., H.W.), University of Oxford, Oxford, UK. Present address for Z.L. is the Department of
Cardiology, Fourth Military Medical University, Xi’an City, People’s Republic of China.
Correspondence to Hugh Watkins, MD, PhD, Dept of Cardiovascular Medicine, John Radcliffe Hospital, Oxford, UK OX28AF. E-mail
[email protected]
© 2002 American Heart Association, Inc.
Circulation Research is available at http://www.circresaha.org
DOI: 10.1161/01.RES.0000036750.81083.83
704
Moolman-Smook et al
HCM Mutations Modulate MyBPC Interactions
705
Figure 1. Schematic representation of
cardiac myosin binding protein-C, illustrating its modular structure. Cardiacspecific regions, domains with known
function, and HCM-causing mutations are
indicated.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
ylation of cMyBPC.6,10 It has since been shown that the
MyBPC motif of cMyBPC can bind to myosin-S211 and that
this association is abolished by cAMP-dependent protein
kinase phosphorylation of the MyBPC motif.12 This may
provide a mechanism for regulating the attachment of myosin
heads to actin and altering the contractility of muscle fibers.
The degree of phosphorylation of cMyBPC also correlates
with changes in thick filament backbone diameter and the
degree of order of the myosin cross bridges, and so likely
affects the probability of cross-bridge formation.13,14 It has
also been proposed that domain C0 sterically hinders the
myosin head; thus, mechanically constraining cross-bridge
movement and providing cardiac muscle with yet another
mechanism to regulate the development of force.15
Interest in the function, structure, and interplay of sarcomeric proteins has increased with the recognition that HCM is
a disease of the sarcomere.16 This clinically and genetically
heterogeneous disease is caused by more than 150 different
mutations in 9 different sarcomeric protein genes.17 Mutational analysis has helped elucidate the pathophysiology of
the disease, by directing in vitro and in vivo investigations of
the effects of these mutations (see review18). Mutations in
cMyBPC are responsible for a clinically recognizable subclass of HCM19,20 characterized by late onset hypertrophy that
progresses throughout life.21 Most of the mutations in
MYBPC3 are splice-site or deletion/insertion mutations that
lead to truncation of the protein17 (see review3) with loss of
the C-terminal titin and/or myosin binding sites. It has been
suggested that these may affect the incorporation of cMyBPC
into the sarcomere, a notion that has been largely confirmed
by a number of studies.22,23 However, at least 14 missense
mutations in MYBPC3 have also been reported to cause
HCM.21,24 –27 Although some of these cause amino acid
substitutions within the known myosin and titin binding
domains, a number occur in domains of unknown function.
These “experiments of nature,” the individually rare disease-
causing mutations, direct attention to areas of the protein not
previously known to be functionally important.
Two such missense mutations, R654H24 and N755K,26
have been described in domain C5 of cMyBPC, occurring on
either side of the 28 amino acid cardiac-specific insertion.
This suggests that this region has an important role in
cMyBPC structure/function in the heart. We hypothesized
that domain C5 of cMyBPC, being an Ig-like domain, might
interact with one or more specific binding partners and so
used the yeast two-hybrid (Y2H) system to search systematically for cardiac-expressed ligands. We identified, unexpectedly, an interaction between domains C5 and C8 of cMyBPC
and subsequently used surface plasmon resonance to confirm
the interaction and to demonstrate the impact of the known
HCM-causing missense mutations in C5 on ligand binding.
We have used our data along with previous biochemical and
structural findings to propose a revised model for the arrangement of cMyBPC in the sarcomere based on cMyBPC
multimerization.
Materials and Methods
Bacterial and Yeast Strains
Saccharomyces cerevisiae strains PJ69-2A and Y187 were used in
the Y2H library screen, while S. cerevisiae strain AH109 was used
in some Y2H direct protein-protein interaction assays. Escherichia
coli KC8 was used to isolate prey plasmids from yeast (all strains
from Clontech).
Constructs
Baits
The region of MYBPC3 encoding domain C5 was PCR-amplified
from a MYBPC3 cDNA clone from an adult cardiac ␭ZapII cDNA
library (Stratagene No. 936208) and cloned in-frame with the
GAL4-DNA binding domain (GAL4BD) in pAS2.1 (pAS2.1-C5).
The region of MYBPC3 encoding domain C7 was PCR-amplified
from the MYBPC3 cDNA clone and the product cloned in-frame with
the GAL4BD in pGBKT7 (pGBK-C7).
C5 bait constructs containing either the R654H24 (pAS2.1C5H654) or the N755K26 (pAS2.1-C5K755) point mutations were
706
Circulation Research
October 18, 2002
generated by PCR-based site-directed mutagenesis, as previously
described,28 with fragments cloned in-frame with GAL4-BD in
pAS2.1.
The integrity of the cloning sites, reading frame, and MYBPC3
sequence of all constructs were verified by bidirectional sequencing.
Preys
Cardiac cDNA Library
A pretransformed MATCHMAKER library (Clontech), consisting of
S. cerevisiae strain Y187 transformed with a cardiac cDNA library
constructed in pACT2, was used for Y2H library screening.
cMyBPC
Prey constructs encoding fusions of specific cMyBPC domains and
GAL4AD were generated by PCR-amplification of the MYBPC3
region(s), followed by in-frame cloning of these products in pACT2.
Yeast Two-Hybrid Assay
Library Screening
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
The MATCHMAKER two-hybrid system 2 (Clontech) was used
according to the manufacturer’s protocols. Briefly, pAS2.1-C5 was
transformed into S. cerevisiae strain PJ69-2A, and this bait line was
mated with the S. cerevisiae Y187 pretransformed cardiac cDNA
library line. Diploid clones were grown on medium lacking leucine
(Leu), tryptophan (Trp), and histidine (His) (TDO plates) for 10
days, and then transferred to medium lacking adenine (Ade) as well
as Leu/Trp/His (QDO plates) for a further 8 days. Colonies positive
for the reporter genes ADE2 and HIS3 were tested for expression of
the reporter gene LacZ by ␤-galactosidase filter assay.
Colonies expressing all reporter genes were further analyzed:
duplicate interactor clones were identified by HaeIII digestion
profiles of nested PCR products from the prey plasmid inserts. Prey
plasmids from representative clones were rescued from diploid yeast
cells via E. coli KC8 cells, retransformed into S. cerevisiae Y187,
and tested for activation of reporter genes autonomously or in the
presence of heterologous baits. Putative true interactors were sequenced and the sequences analyzed with BLASTN (GenBank). The
deduced amino acid sequence of open reading frames (ORFs)
in-frame with the GAL4AD were analyzed with BLASTP.29
Deletion Mapping of Interacting Domains
Prey constructs of fusion genes encoding individual domains, or
combinations of domains, of cMyBPC and GAL4AD in pACT2 were
individually transformed into S. cerevisiae Y187. These lines were
mated with the bait line (pAS2.1-C5) transformed into S. cerevisiae
AH109 and diploid cells selected on medium lacking Leu and Trp.
These colonies were replicated to TDO plates (2 to 4 days);
thereafter, interacting clones were selected on QDO plates (6 days).
Expression and Purification of cMyBPC Domains
in E. coli
Expression constructs (either pMW17230 or pET28a; Novagen) were
assembled encoding the wild-type domains C5, C5-C7, C4, and
C8-C10 and mutant C5 and C5-7 domains containing the R654H and
N755K amino acid substitutions. Recombinant proteins were overexpressed and isolated by standard chromatographic techniques at a
final purity of ⬎95%.
BIAcore Experiments
Proteins were dialyzed into myosin binding buffer, and binding
studies were performed on a BIAcore X biosensor (Biacore AB). The
C8-C10 fragment was covalently coupled to one flow cell of a B1
sensor chip using the BIAcore Amine Coupling Kit (BIAcore AB) at
a buffer flow rate of 20 ␮L/min, 25°C, until the response units had
increased by between 1000 and 2000. The control flow cell underwent the same treatment without addition of C8-C10. Proteins were
injected over both flow cells for 150 seconds at the same flow rate.
Equilibrium binding constants were calculated for C5 wild-type,
R654H, and N755K (His-tagged) and C5-C7 wild type, R654H, and
N755K (all His-tagged) using the BIAevaluation software (BIAcore
AB). Repeat experiments with C5 domains using all His-tagged
proteins from the pET-28a constructs demonstrated comparable
binding patterns.
An expanded Materials and Methods section can be found in the
online data supplement available at http://www.circresaha.org.
Results
Library Screening With C5 Bait
Screening more than 7⫻106 cDNA clones with domain C5 of
cMyBPC as bait, using a yeast mating-based Y2H analysis
method, yielded 13 independent interactor clones protrophic
for Ade, as evidenced by robust light pink to white colonies
on QDO medium and a positive ␤-galactosidase filter assay
within 3 hours (Table). Eleven of these clones, 7 of which
were represented by additional, apparently identical, cDNAs
(Table), showed no activation of the three reporter genes
autonomously or in the presence of heterologous baits.
BLASTN comparisons of the nucleotide sequences derived
from representative prey cDNAs of these 11 interactors
revealed significant matches to genomic sequences in GenBank (Table). However, 10 were discarded as being implausible, because the predicted in-frame ORFs encoded either
very short peptides with no significant database matches, or
proteins that are physically separated from cMyBPC in vivo
(Table).
The remaining clone, designated clone 77, was the most
abundantly represented among the positive clones (1 of 4
apparently identical clones), and gave the strongest and
fastest lacZ expression on ␤-galactosidase filter assay. Sequencing of clone 77 revealed that it encoded the C-terminal
608 amino acids of cMyBPC itself, from residue 667, which
is located within the C5 domain, through to the stop codon.
Deletion Mapping of C5 Ligand
To define the domains involved in this interaction and to
confirm specificity, deletion mapping of this region of
cMyBPC was undertaken. Diploid yeast cells harboring both
pAS2.1-C5 and pACT2-C5 were unable to activate any of the
reporter genes, but were able to activate all three reporter
genes when pAS2.1-C5 and pACT2-C6-C10 were present
(Figure 2A). This ability was retained in the absence of C10,
as seen with pACT2-C6-C9 (Figure 2A). Further deletion
mapping of the C6-C10 region demonstrated that all deletion
constructs that included domain C8 were able to activate the
reporter genes in the presence of pAS2.1-C5, but could not do
so autonomously (Figure 2A). In contrast, constructs lacking
domain C8 failed to activate the three reporter genes, whether
in the presence or absence of pAS2.1-C5 (Figure 2A). These
results are summarized in Figure 2B.
Direct Protein-Protein Interaction Assay of C7
Using the results obtained from the C5-library screen, we
proposed an arrangement of cMyBPC in the sarcomere based
on an overlapping, parallel, multimerization bringing C5 of
one molecule into contact with C8 of the next (see Discussion). In such an arrangement domain C7 of one cMyBPC
molecule would be juxtaposed to domain C10 of another. To
establish whether specific interaction also occurred between
these two domains, an Y2H direct protein-protein interaction
Moolman-Smook et al
HCM Mutations Modulate MyBPC Interactions
707
Putative True Interactor Clones Obtained From the Yeast Two-Hybrid Library Screen With cMyBPC Domain C5
Colony Phenotype
(with pAS2.1-C5 bait)
Clones and Their
Replicates
77, 32, 28, 34
25, 45, 46
20, 33, 56
QDO
LacZ
Genomic Hit/Accession No.
In-Frame ORF Protein
Hit/Accession No.
Cellular Localization
Light pink
2 hours/⫹⫹⫹
cMyBPC/NM_000256
cMyBPC/AAC04620
C-zone sarcomere
White
2 hours/⫹⫹
M-protein/XM_005198
M-protein/XP005198
M-line sarcomere
None
White
2 hours/⫹⫹
Genomic clone/AC046186
4, 5, 48
Light pink
2 hours/⫹⫹
Genomic clone/AL122015
None
60
Light pink
2 hours/⫹⫹
Vitellogenic carboxypeptidase-like
protein/AF282617
Vitellogenic carboxypeptidase-like
protein/AF282617_1
68, 58, 79
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Light pink
2 hours/⫹⫹
Genomic clone/Hs18_25232
None
66, 75
White
3 hours/⫹⫹
Tenascin-X/HSA17867
Tenascin-X/A40701
64, 21, 26
White
3 hours/⫹
Genomic clone/AK024570
None
70
White
3 hours/⫹
Similar to bHLH Zip transcription
factor/BC000015
Similar to bHLH-Zip transcription
factor/AAH00015
55
Light pink
3 hours/⫹
Genomic clone/Hs1_4843
None
8
Light pink
3 hours/⫹
Collagen 1A2/XM_004658
Collagen 1A2/XP_004658
Lysosome
Extracellular
Nuclear
Extracellular
ORF indicates open reading frame; QDO, colony color on solid medium lacking Leu, Trp, Ade, and His; lacZ, relative amount of, and time taken
to produce, blue color by colony during ␤-galactosidase filter assay.
⫹Light blue; ⫹⫹Blue; ⫹⫹⫹Deep blue.
assay was performed. Diploid yeasts containing both
pGBK-C7 and pACT2-C10 were protrophic for Ade and His
as demonstrated by their ability to grow on QDO medium
(Figure 3). In contrast, yeasts containing pGBK-C7 and either
pACT2-C5, pACT2-C6, or pACT2-C8 were not able to grow
on TDO or QDO medium (Figure 3).
Surface Plasmon Resonance Confirmation of
C5:C8 Interaction
Surface plasmon resonance-based measurement of the C5-C8
interaction in vitro revealed binding of wild type recombinant
C5 to C8-C10 over a range of concentrations in myosin
binding buffer (Figure 4A). Analysis of the association and
dissociation phases of the binding curves allowed estimation
of the association (ka) and dissociation (kd) rate constants and
calculation of the association equilibrium constant (Ka) for
this interaction. For wild-type C5, the Ka was 9.83⫾1.06⫻104
mol/L⫺1 (n⫽3). The interaction was shown to be specific to
the C5 domain, as C4 (the flanking Ig-like domain in
cMyBPC) does not bind significantly to C8-C10 (Figure 4B).
Both the C5 R654H and N755K mutants showed a significantly weaker binding to C8-C10 compared with wild type; a
comparison at 13 ␮mol/L is shown in Figure 4C. For R654H
the Ka was 6.00⫾2.55⫻104 mol/L⫺1 (n⫽2). In 4 independent
experiments, C5 N755K was shown to have an affinity for
C8-C10 at least 10 times lower than wild type; the weakness
of this interaction was such that the Ka can only be approximated at 8.68⫻103 mol/L⫺1. Interestingly, when C5-C7
(rather than C5 alone) was passed over C8-C10, the Ka
increased to 1.41⫾0.09⫻106 mol/L⫺1 (n⫽3). Binding of
C5-C7 R654H was not significantly weaker than wild type,
with a Ka of 1.36⫾0.07⫻106 mol/L⫺1 (n⫽4); however, C5-C7
N755K bound with a Ka of just 6.15⫾1.41⫻105 mol/L⫺1
(n⫽3). There seems therefore to be a stronger interaction
between this larger fragment and C8-C10, which is also
affected by the HCM-causing mutations, though to a lesser
extent than for C5 binding alone. This is consistent with the
additional interaction between C7 and C10 demonstrated by
the Y2H analysis.
Discussion
In the present study, we used Y2H analysis to identify a
ligand for domain C5 of cMyBPC and found, unexpectedly,
that the binding partner was cMyBPC itself; the binding
sequence lay within the C-terminal domains (C5-C10). The
region of interaction was subsequently narrowed down to
domain C8 by deletion mapping of the original interactor
clone. Evidence for the involvement of C5 in cMyBPC:cMyBPC interaction was compelling for a number of reasons.
First, when domain C5 was used as bait to screen more than
7⫻106 cardiac cDNA clones, it preferentially bound to
C-terminal cMyBPC clones containing the C8 domain. Moreover, of the 11 representative interactor clones identified in
the library screen, only clone 77 contained an insert with an
extended ORF in-frame with GAL4AD and encoded a protein
whose in vivo localization was compatible with that of the
bait. Reporter gene activation by clone 77 was specific to the
pAS2.1-C5 bait and could not be elicited by heterologous bait
constructs.
A surface plasmon resonance-based assay was then developed to demonstrate C5-C8 binding in vitro. This methodology has been frequently used to study the binding interactions
of antibodies and cell surface receptors containing Ig domains. With the recombinant fragment C8-C10 immobilized
on the chip surface, wild-type recombinant C5 bound with an
estimated Ka of ⬇1.0⫻105 mol/L⫺1. This is a comparatively
weak interaction but is of a similar affinity to the association
of other Ig-like domains measured by the same methodology,
for example NCAM IgI and IgII.31 A binding constant of this
magnitude is fully compatible with interaction of these
domains in vivo, as in the thick filament the effective
708
Circulation Research
October 18, 2002
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Figure 3. Yeast two-hybrid reporter gene activation by prey
cMyBPC domains paired with bait domain C7. Growth of diploid
yeast on QDO medium indicates interaction of C7 with a prey
domain. Representative direct protein-protein interaction assay
demonstrates that domain C10 activates the reporter genes in
the presence of domain C7.
Figure 2. A, Yeast two-hybrid reporter gene activation by different prey cMyBPC domains paired with bait domain C5. Growth
of diploid yeast on QDO medium indicates interaction of C5
with a prey protein, leading to activation of the ADE2 and HIS3
reporter genes. Only prey constructs that included domain C8
of cMyBPC activated these reporter genes in the presence of
domain C5. B, Summary of interactions of cMyBPC domain C5
with other carboxy-terminal domains of cMyBPC. Domains of
cMyBPC encoded by the prey plasmids are schematically illustrated, whereas His⫹, Ade⫹, lacZ⫹ indicate the ability of these
domains, in conjunction with domain C5, to activate all three
reporter genes.
concentration of cMyBPC will be high due to its attachment
via its myosin and titin binding sites.
Previous investigations have suggested that both skeletal
and cardiac isoforms of MyBPC multimerize.1,32 At physiological ionic strength, the sedimentation coefficient of skeletal MyBPC increased with rising concentration,1 consistent
with dimerization of the protein and also with formation of
higher multimers. Furthermore, cMyBPC eluted faster on gel
filtration than would be expected on the basis of its chain
molecular weight, indicating that the protein was asymmetric
and/or rapidly and reversibly dimerised.32
The dimensions of the v-shaped particles of purified
cMyBPC (arm lengths 22.1⫾4.5 nm), as determined by
electron microscopy of rotary shadowed protein,32 suggest
that the hinge of the v-shape most likely coincides with the
region between domains C4 and C5; the presence of additional residues between these two domains also suggests that
this is a flexible “linker.”32 Domains C5 and C8 are, therefore, expected to occur on the same arm, such that any
cMyBPC:cMyBPC interaction appears more likely to be
inter- than intramolecular. If the interaction is, indeed, intermolecular this could involve parallel or antiparallel orientation of the cMyBPC proteins. The orientation would determine which additional domains are in register, and which
might, therefore, interact to stabilize the dimer structure. In an
antiparallel arrangement, it is plausible that domains C5 and
C8 of each cMyBPC would interact with domains C8 and C5,
respectively, of the other. Such an orientation would also
place domain C7 of each cMyBPC in register with domain C6
of the other. However, in a parallel arrangement, domains C7
of one cMyBPC would be in register with domain C10 of the
other. To test these alternatives, domain C7 was used as bait
in Y2H-based direct protein-protein interaction assay against
other C-terminal cMyBPC domains. Only domain C10 demonstrated binding to C7, supporting a staggered parallel
arrangement of cMyBPC during multimerization.
Early data suggested that the amount of MyBPC in the
thick filament corresponds to 2 to 4 molecules per every third
crown of myosin heads in the C-zone of the sarcomere.1,2 It
would seem most likely that the number of cMyBPC molecules would be three, considering that myosin rods are most
probably arranged in a three-stranded helix,33 that six, prob-
Moolman-Smook et al
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Figure 4. The interaction of C5, C5-C7, and C5 R654H and
N755K mutants with immobilized C8-C10. C8-C10 was
covalently immobilized onto a BIAcore B1 chip and association
with other cMyBPC fragments measured in buffer containing
100 mmol/L KCl, 1 mmol/L DTT, and 20 mmol/L imidazole pH
7.0 at 25°C at a flow rate of 20 ␮L/min. 50 ␮L (150 sec) injections were performed, and dissociation of bound protein was
monitored for at least a further 100 sec. Typical real-time binding curves are shown for the following: A, C5 wild type over a
range of concentrations; B, C5 (10 ␮mol/L), C5-C7 (10 ␮mol/L),
and C4 (13 ␮mol/L); C, C5 wild type, R654H, and N755K, each
at 100 ␮mol/L.
ably paired, titin strands lie along the surface of the myosinbackbone,33,34 and that the position of domain C10, which
binds to both titin and myosin, would be dictated by the
register of the first Ig-like domain of the titin superrepeats and
the MyBPC binding site on the LMM component of myosin.7,8 Indeed, many different models of the arrangement of
MyBPC in the thick filament have been based on three
MyBPC molecules. A subset of models have also incorporated evidence indicating that MyBPC probably lies on the
outer surface of the myosin filaments35 and the immunocytochemical measurement of the axial extent of MyBPC (⬇7
nm), implying that part of MyBPC is perpendicular to the
thick filament axis.36,37 Within these parameters, a number of
different quaternary structures have been proposed, ranging
from a tight38,39 to a loose4,40 MyBPC trimer collar wrapped
around the outside of the thick filament. Others have also
proposed lateral projection of part of the molecule into the
interfilamental space,41 which accommodates data from x-ray
diffraction studies thought to indicate that, in resting muscle,
HCM Mutations Modulate MyBPC Interactions
709
a part of MyBPC is attached to the thick filament while
another part is probably floating.33
Incorporating these previous observations with our own
data, we propose a model (Figure 5) in which three cMyBPC
molecules trimerize to form a collar around the thick filament
backbone, with overlaps of domains C5-C7 of one cMyBPC
molecule with domains C8-C10 of the ensuing molecule, with
staggered parallel orientation of the proteins. In this model,
domains C5-C10 of each cMyBPC molecule form part of the
two molecule–thick collar that is arranged perpendicularly to
the thick filament axis, while the N-terminal domains C0-C4
are extended into the interfilamental space. This model of
cMyBPC quaternary structure is compatible with the finding
that, in an analysis of C-terminal fragments, C8-C10 could
not target MyBPC to the correct location in the sarcomere,
whereas a C7-C10 fragment incorporated normally into the A
band of the sarcomere.9 This model is also compatible with
previous structural data. The point at which the proposed
N-terminal arms veer away from the C-terminal collar coincides with the hinge of the v-shaped purified cMyBPC
particles.32 The length predicted for these N-terminal arms
(⬇26 nm) would be sufficient for the MyBPC motif to reach
the S2 domain of myosin from the preceding crown. The
length of each Ig-like domain and FN3 domain has been
estimated at approximately 4 nm,37 and thus, the expected
circumference of the proposed 9 domain collar is broadly
compatible with the measured diameter of 13 to 15 nm for the
intercross-bridge spaces of myosin filaments.33 In addition to
the dimensions of the MyBPC domains, factors such as the
structure of the linker amino acids between domains7 and the
possibility that the alignment of the interacting domains is
staggered may also influence the actual circumference of the
proposed collar. Recent studies have suggested that when
cMyBPC is phosphorylated the thick filament diameter may
increase, altering the position and increasing the degree of
order of the myosin cross bridges, thereby potentiating their
binding to actin.13,14 These data imply that the cMyBPC
collar is probably not static, but may have to widen on
phosphorylation, and that the C5:C8 and the C7:C10 interactions might have to be abrogated and reformed in a manner
compatible with the rate of phosphorylation of the MyBPC
motif.
The surface plasmon resonance data suggest that the two
HCM-causing missense mutations in domain C5 significantly
weaken the C5:C8 interaction and hence in vivo may perturb
the structure of the proposed MyBPC collar. If a function of
the collar is to provide the correct positioning of the myosin
S2-binding N-terminal region of MyBPC with respect to the
myosin neck, then its disruption may reduce this interaction
in the absence of MyBPC phosphorylation. Mutations that
disrupt the collar may therefore have an activating effect by
favoring cross bridge formation. In this way, these mutations
may increase the rate of ATP consumption as has been seen
for other classes of HCM mutations that appear to act by
increasing the cost of force production.18,42 The degree of
disruption of the C5:C8 interaction by the R654H and N755K
missense mutations, as measured by surface plasmon resonance, appears to mirror the severity of clinical features
associated with each. Although only described in a small
710
Circulation Research
October 18, 2002
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Figure 5. Proposed arrangement of cMyBPC
around thick filament backbone. Based on
previous data and our own results, we propose that three cMyBPC molecules trimerize
to form a collar around the thick filament,
with overlaps of domains C5-C7 of one
cMyBPC molecule with domains C8-C10 of
the ensuing molecule, with staggered parallel
orientation of the proteins. Domains C5-C10
of each cMyBPC molecule form part of the
two molecule–thick collar, while the
N-terminal domains C0-C4 are extended into
the interfilamental space where the cMyBPC
motif binds to myosin S2.
family, the R654H mutation was associated with low penetrance and only moderate hypertrophy even in adults,24
whereas the N755K mutation was associated with near
complete penetrance, quite marked hypertrophy even in
children, and sudden cardiac death in 1 of 8 affected
patients.43
The site of interaction of C8 on the C5 domain remains
speculative. The R654H and N755K mutations may result in
a reduced affinity for C8 either because they are directly
involved in the interaction or because they induce a conformational change that is transmitted to the binding interface.
Molecular modeling of this Ig-like domain suggests these two
residues lie on the same surface of C5 and in close proximity
to the 28 residue cardiac-specific insertion. It is tempting to
speculate that this sequence may be important for modulating
the C5:C8 interaction. Interestingly, a recent preliminary
NMR study of C5 has suggested that this insertion forms a
mobile region on the outside of the domain.44
In conclusion, the observation of disease-causing mutations in a region of cMyBPC not previously associated with
a particular function has highlighted novel structure/function
aspects for this protein. The approach of integrating mutational data with functional and structural investigations of the
relevant individual domains is likely to lead to important
insight into both the quaternary structure of cMyBPC within
the sarcomere and the manner in which it regulates muscle
contractility in both health and disease.
Acknowledgments
We acknowledge the Wellcome Trust, British Heart Foundation, and
South African Medical Research Council for financial support, and
thank Dr Shoumo Bhattacharya for his advice and gift of pAS2.1p35 for use as a heterologous bait.
References
1. Offer G, Moos C, Starr R. A new protein of the thick filaments of
vertebrate skeletal myofibrils. Extractions, purification and characterization. J Mol Biol. 1973;74:653– 676.
2. Bennett P, Craig R, Starr R, Offer G. The ultrastructural location of
C-protein, X-protein and H-protein in rabbit muscle. J Muscle Res Cell
Motil. 1986;7:550 –567.
3. Bennett PM, Furst DO, Gautel M, The C-protein (myosin binding protein
C) family: regulators of contraction and sarcomere formation? Rev
Physiol Biochem Pharmacol. 1999;138:203–234.
4. Winegrad S. Cardiac myosin binding protein C. Circ Res. 1999;84:
1117–1126.
5. Weber FE, Vaughan KT, Reinach FC, Fischman DA. Complete sequence
of human fast-type and slow-type muscle myosin- binding-protein C
(MyBP-C): differential expression, conserved domain structure and chromosome assignment. Eur J Biochem. 1993;216:661– 669.
6. Gautel M, Zuffardi O, Freiburg A, Labeit S. Phosphorylation switches
specific for the cardiac isoform of myosin binding protein-C: a modulator
of cardiac contraction? EMBO J. 1995;14:1952–1960.
7. Okagaki T, Weber FE, Fischman DA, Vaughan KT, Mikawa T, Reinach
FC. The major myosin-binding domain of skeletal muscle MyBP-C (C
protein) resides in the COOH-terminal, immunoglobulin C2 motif. J Cell
Biol. 1993;123:619 – 626.
8. Freiburg A, Gautel M. A molecular map of the interactions between titin
and myosin-binding protein C. Implications for sarcomeric assembly in
familial hypertrophic cardiomyopathy. Eur J Biochem. 1996;235:
317–323.
Moolman-Smook et al
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
9. Gilbert R, Cohen JA, Pardo S, Basu A, Fischman DA. Identification of
the A-band localization domain of myosin binding proteins C and H
(MyBP-C, MyBP-H) in skeletal muscle. J Cell Sci. 1999;112(Pt
1):69 –79.
10. Hartzell HC, Titus L. Effects of cholinergic and adrenergic agonists on
phosphorylation of a 165,000-dalton myofibrillar protein in intact cardiac
muscle. J Biol Chem. 1982;257:2111–2120.
11. Gruen M, Gautel M. Mutations in ␤-myosin S2 that cause familial
hypertrophic cardiomyopathy (FHC) abolish the interaction with the
regulatory domain of myosin-binding protein-C. J Mol Biol. 1999;286:
933–949.
12. Gruen M, Prinz H, Gautel M. cAPK-phosphorylation controls the interaction of the regulatory domain of cardiac myosin binding protein C with
myosin-S2 in an on-off fashion. FEBS Lett. 1999;453:254 –259.
13. Weisberg A, Winegrad S. Relation between crossbridge structure and
actomyosin ATPase activity in rat heart. Circ Res. 1998;83:60 –72.
14. Levine R, Weisberg A, Kulikovskaya I, McClellan G, Winegrad S.
Multiple structures of thick filaments in resting cardiac muscle and their
influence on cross-bridge interactions. Biophys J. 2001;81:1070 –1082.
15. Witt CC, Gerull B, Davies MJ, Centner T, Linke WA, Thierfelder L.
Hypercontractile properties of cardiac muscle fibers in a knock-in mouse
model of cardiac myosin-binding protein-C. J Biol Chem. 2001;276:
5353–5359.
16. Thierfelder L, Watkins H, MacRae C, Lamas R, McKenna W, Vosberg
HP, Seidman JG, Seidman CE. Alpha-tropomyosin and cardiac troponin
T mutations cause familial hypertrophic cardiomyopathy: a disease of the
sarcomere. Cell. 1994;77:701–712.
17. DNA Mutation Database: Familial Hypertrophic Cardiomyopathy. Australian National Genomic Information Service. Available at: http://
morgan.angis.su.oz.au/Databases/Heart/heartbreak.html. Accessed
August 28, 2002.
18. Redwood CS, Moolman-Smook JC, Watkins H. Properties of mutant
contractile proteins that cause hypertrophic cardiomyopathy. Cardiovasc
Res. 1999;44:20 –36.
19. Watkins H, Conner D, Thierfelder L, Jarcho JA, MacRae C, McKenna
WJ, Maron BJ, Seidman JG, Seidman CE. Mutations in the cardiac
myosin binding protein-C gene on chromosome 11 cause familial hypertrophic cardiomyopathy. Nat Genet. 1995;11:434 – 437.
20. Bonne G, Carrier L, Bercovici J, Cruaud C, Richard P, Hainque B, Gautel
M, Labeit S, James M, Beckmann J, Schwartz K. Cardiac myosin binding
protein-C gene splice acceptor site mutation is associated with familial
hypertrophic cardiomyopathy. Nat Genet. 1995;11:438 – 440.
21. Niimura H, Bachinski LL, Sangwatanaroj S, Watkins H, Chudley AE,
McKenna W, Kristinsson A, Roberts R, Sole M, Maron BJ, Seidman JG,
Seidman CE. Mutations in the gene for cardiac myosin-binding protein C
and late- onset familial hypertrophic cardiomyopathy. N Engl J Med.
1998;338:1248 –1257.
22. Flavigny J, Souchet M, Sebillon P, Berrebi-Bertrand I, Hainque B, Mallet
A, Bril A, Schwartz K, Carrier L, COOH-terminal truncated cardiac
myosin-binding protein C mutants resulting from familial hypertrophic
cardiomyopathy mutations exhibit altered expression and/or incorporation in fetal rat cardiomyocytes. J Mol Biol. 1999;294:443– 456.
23. Yang Q, Sanbe A, Osinska H, Hewett TE, Klevitsky R, Robbins J. In vivo
modeling of myosin binding protein C familial hypertrophic cardiomyopathy. Circ Res. 1999;85:841– 847.
24. Moolman-Smook JC, Mayosi B, Brink P, Corfield VA. Identification of
a new missense mutation in MyBP-C associated with hypertrophic cardiomyopathy. J Med Genet. 1998;35:253–254.
HCM Mutations Modulate MyBPC Interactions
711
25. Moolman-Smook JC, De Lange WJ, Bruwer EC, Brink PA, Corfield VA.
The origins of hypertrophic cardiomyopathy-causing mutations in two
South African subpopulations: a unique profile of both independent and
founder events. Am J Hum Genet. 1999;65:1308 –1320.
26. Yu B, French JA, Carrier L, Jeremy RW, McTaggart DR, Nicholson MR,
Hambly B, Semsarian C, Richmond DR, Schwartz K, Trent RJ.
Molecular pathology of familial hypertrophic cardiomyopathy caused by
mutations in the cardiac myosin binding protein C gene. J Med Genet.
1998;35:205–210.
27. Niimura H, Patton KK, McKenna WJ, Soults J, Maron BJ, Seidman JG,
Seidman CE. Sarcomere protein gene mutations in hypertrophic cardiomyopathy of the elderly. Circulation. 2002;105:446 – 451.
28. Elliott K, Watkins H, Redwood CS. Altered regulatory properties of
human cardiac troponin I mutants that cause hypertrophic cardiomyopathy. J Biol Chem. 2000;275:22069 –22074.
29. BLAST Database. National Center for Biotechnology Information Web
Site. Available at: http://ncbi.nlm.nih.gov/blast. Accessed August 28,
2002.
30. Way M, Gooch J, Pope B, Weeds AG. Expression of human plasma
gelsolin in Escherichia coli and dissection of actin binding sites by
segmental deletion mutagenesis. J Cell Biol. 1989;109:593– 605.
31. Kiselyov VV, Berezin V, Maar TE, Soroka V, Edvardsen K, Schousboe
A, Bock E. The first immunoglobulin-like neural cell adhesion molecule
(NCAM) domain is involved in double-reciprocal interaction with the
second immunoglobulin-like NCAM domain and in heparin binding.
J Biol Chem. 1997;272:10125–10134.
32. Hartzell HC, Sale WS. Structure of C protein purified from cardiac
muscle. J Cell Biol. 1985;100:208 –215.
33. Squire J, Cantino M, Chew M, Denny R, Harford J, Hudson L, Luther P.
Myosin rod-packing schemes in vertebrate muscle thick filaments. J
Struct Biol. 1998;122:128 –138.
34. Liversage AD, Holmes D, Knight PJ, Tskhovrebova L, Trinick J. Titin
and the sarcomere symmetry paradox. J Mol Biol. 2001;305:401– 409.
35. Moos C, Offer G, Starr R, Bennett P. Interaction of C-protein with
myosin, myosin rod and light meromyosin. J Mol Biol. 1975;97:1–9.
36. Starr R, Offer G. The interaction of C-protein with heavy meromyosin
and subfragment-2. Biochem J. 1978;171:813– 816.
37. Pfuhl M, Pastore A. Tertiary structure of an immunoglobulin-like domain
from the giant muscle protein titin: a new member of the I set. Structure.
1995;3:391– 401.
38. Offer G. C-protein and the periodicity in the thick filaments of vertebrate
skeletal muscle. CSH Symp Quant Biol. 1972;37:87–95.
39. Swan RC, Fischman DA. Electron microscopy of C-protein molecules
from chicken skeletal muscle. J Muscle Res Cell Motil. 1986;7:160 –166.
40. Bennett P, Starr R, Elliott A, Offer G. The structure of C-protein and
X-protein molecules and a polymer of X- protein. J Mol Biol. 1985;184:
297–309.
41. Luther P. Three dimensional structure of the A-band and localization of
MyBPC-C. Biophys J. 1998;74:A352. Abstract.
42. Montgomery DE, Tardiff JC, Chandra M. Cardiac troponin T mutations:
correlation between the type of mutation and the nature of myofilament
dysfunction in transgenic mice. J Physiol. 2001;536:583–592.
43. Yu B, French JA, Jeremy RW, French P, McTaggart DR, Nicholson MR,
Semsarian C, Richmond DR, Trent RJ. Counselling issues in familial
hypertrophic cardiomyopathy. J Med Genet. 1998;35:183–188.
44. Idowu SM, Gautel M, Pfuhl M. Sequence specific resonance assignment
of the central domain of cardiac myosin binding protein C (MyBP-C).
J Biomol NMR. 2002;22:199 –200.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Identification of Novel Interactions Between Domains of Myosin Binding Protein-C That
Are Modulated by Hypertrophic Cardiomyopathy Missense Mutations
Johanna Moolman-Smook, Emily Flashman, Willem de Lange, Zhili Li, Valerie Corfield,
Charles Redwood and Hugh Watkins
Circ Res. 2002;91:704-711; originally published online September 12, 2002;
doi: 10.1161/01.RES.0000036750.81083.83
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2002 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/91/8/704
Data Supplement (unedited) at:
http://circres.ahajournals.org/content/suppl/2002/10/13/91.8.704.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/