|Home | About | Journals | Submit | Contact Us | Français|
Despite early demonstrations of myosin binding protein C’s (MyBP-C) interaction with actin, different investigators have reached different conclusions regarding the relevant and necessary domains mediating this binding. Establishing the detailed structure-function relationships is needed to fully understand cMyBP-C’s ability to impact on myofilament contraction as mutations in different domains are causative for familial hypertrophic cardiomyopathy. We defined cMyBP-C’s N-terminal structural domains that are necessary or sufficient to mediate interactions with actin and/or the head region of the myosin heavy chain (S2-MyHC). Using a combination of genetics and functional assays, we defined the actin binding site(s) present in cMyBP-C. We confirmed that cMyBP-C’s C1 and m domains productively interact with actin, while S2-MyHC interactions are restricted to the m domain. Using residue-specific mutagenesis, we identified the critical actin binding residues and distinguished them from the residues that were critical for S2-MyHC binding. To validate the structural and functional significance of these residues, we silenced the endogenous cMyBP-C in neonatal rat cardiomyocytes (NRC) using cMyBP-C siRNA, and replaced the endogenous cMyBP-C with normal or actin binding-ablated cMyBP-C. Replacement with actin binding-ablated cMyBP-C showed that the mutated protein did not incorporate into the sarcomere normally. Residues responsible for actin and S2-MyHC binding are partially present in overlapping domains but are unique. Expression of an actin binding-deficient cMyBP-C resulted in abnormal cytosolic distribution of the protein, indicating that interaction with actin is essential for formation and/or maintenance of normal cMyBP-C sarcomeric distribution.
Cardiac myosin binding protein C (cMyBP-C) is a large, 140 kDa thick-filament protein that shows a unique pattern of localization to the A-band in the sarcomere and has both structural and regulatory functions . cMyBP-C’s role in normal cardiac function has attracted extensive attention as it became apparent that mutations in the protein were responsible for a substantial number of the total mutations found in familial hypertrophic cardiomyopathy (FHC) . Of the 196 known FHC-causing mutations distributed throughout MYBPC3, 54 are located in the region that encodes the four, N-terminal domains (C0-C1-m-C2 or C0C2; Fig. 1) of the protein .
cMyBP-C binds to all three filament systems in the contractile apparatus: the thin filament (predominantly made up of actin), the thick filament (containing myosin) and the giant protein titin, which makes up the structural, third filament system . MyBP-C‘s interaction(s) with the actin thin filament was first defined via biochemical analyses over 30 years ago . However, the data are contradictory with respect to the actual domains mediating the proteins’ interaction(s). While biochemically-based assays point to the C0 domain (Fig. 1) [4, 5], the Pro-Ala-rich linker between the C0 and C1 domains was also identified as being critical , with other investigators defining a combination of N-terminal domains located within the C0-C2 domains as playing critical roles in mediating cMyBP-C:actin interactions [7–9]. A recent study concluded these putative, N-terminal actin binding sites were due to weak, nonspecific electrostatic interactions, implicating cMyBP-C’s C-terminal half (C5-C10; Fig. 1) as containing the specific actin binding site . The experimental data as well as modeling studies have failed to provide a consensus for either the domains or residues that are necessary and sufficient to mediate cMyBP-C:actin interactions.
We decided to use the power of genetics to approach these issues and have carried out a comprehensive series of yeast 2-hybrid (Y2H) screens in which colony growth on highly selective media is dependent upon a functional interaction between the protein domains. To systemically define the binding site(s) and specific residues responsible for actin binding, we used the individual domains and combinations of domains derived from the both the N-terminal and C-terminal portions of the protein to distinguish cMyBP-C’s actin and S2-MyHC binding regions. The genetic data were then confirmed by complementary biochemical assays and the specific residues in cMyBP-C responsible for actin/S2-MyHC binding defined using site-directed mutagenesis and genetic assays. We then validated the structural importance of the identified residues in the actin/cMyBP-C interaction by replacing the endogenous cMyBP-C with a protein in which the actin binding domain had been ablated.
Mouse cMyBP-C domains were cloned by standard PCR methods from a mouse cDNA library. Constructs were then subcloned into the GAL4 DNA binding domain of pGBKT7 (Clontech) so as to serve as “bait.” The “prey” was either the mouse α-cardiac actin, or mouse S2-MyHC fragment cloned next to the GAL4 activation domain of pGADT7 (Clontech). The bait and prey were transformed into AH109 individually and in combination. The yeast were initially plated on medium stringency media containing SD/-His/-Leu/-Trp and then plated onto SD/-Ade/-His/-Leu/-Trp/X-α-gal plates.
We used a yeast-3-hybrid system to determine whether binding of cMyBP-C’s m domain with actin and/or S2-MyHC are mutually exclusive . The m domain was fused to the GAL4 DNA-binding domain under the control of the ADH promoter in the pBridge vector (Clontech), whereas the S2-MyHC was introduced under control of the Met25 promoter, resulting in the plasmid “m-S2-pBridge” (Fig. S6A), which served as “bait.” The “prey” was the mouse α-cardiac actin, cloned next to the GAL4 activation domain behind the ADH promoter the of pGADT7 vector (Clontech). Alternatively, we also designed complementary “bait” vectors where the α-cardiac actin was introduced under the control of the Met25 promoter, resulting in the plasmid “m-Act-pBridge” (Fig. S6B). The “prey” was the S2-MyHC, cloned next to the GAL4 activation domain behind the ADH promoter the of pGADT7 vector (Clontech). The bait and prey were transformed into AH109 sequentially, with the bait transformed in first. Functional interactions were tested by growth on SD/-Leu/-Trp plates, while expression of the putative inhibitor domain, such as S2-MyHC in m-S2-pBridge (Fig. S6A) are repressed. The yeast were plated on media containing SD/-His/-Leu/-Trp and then plated on SD/-His/-Leu/-Met/-Trp plates, which activates expression of the repression protein.
Primary NRCs were isolated from the ventricles of 1–2-day old Sprague-Dawley rat pups and grown in DMEM (Cellgro) containing 10% FBS and 1% penicillin/streptomycin. Twenty-four hours after plating, they were infected with purified, wild type cMyBP-C (control) or actin binding-ablated cMyBP-C adenoviruses for 2 hours in DMEM media. Adenovirus titers were determined using the plaque forming unit assay method. PFU’s: Wt-cMyBP-C = 4x1010; mut-cMyBP-C = 2x109. Infection of parallel plates with adenovirus expressing β-galactosidase served as a control for all experiments. Each 10 cm plate was infected with 3.75x106 PFU. Chamber slides were infected with 2.5X105 PFU/chamber.
We prepared adenoviral constructs containing wild type cMyBP-C as well as cMyBP-C with the postulated critical actin-binding residues in the C1 domain mutated. Both wild type and actin binding-ablated cMyBP-C mutants were cloned into a pShuttle-CMV vector and replication-deficient recombinant adenoviruses made using the AdEasy system (Stratagene). In order to distinguish the expression of exogenous cMyBP-C from the endogenous cMyBP-C protein, the cMyBP-C constructs were myc-tagged at the N-termini.
A pool of siRNAs (Invitrogen) was tested for their capacity to reduce cMyBP-C protein levels in NRCs. We used 3 siRNAs for cMyBP-C to silence endogenous cMyBP-C expression. The most potent silencing siRNA, siRNA-1, (GCA UGU UCU GCA AGC AGG GAG UAU U) was used. A nonspecific siRNA was used for a negative control in all silencing experiments. Twenty-four hours after plating, cells were transfected with siRNA with Lipofectamine 2000 (Invitrogen) in OptiMem (Invitrogen) media overnight. In experiments using both siRNAs and adenovirus, cells were transfected first, returned to growth media for 6 hours, and then infected for 2 hours.
Data are expressed as mean±SEM. All statistical tests were done with SigmaPlot 9.0 software. Comparisons between 2 groups were analyzed with Student’s t test (P<0.05).
Identification of the actin binding region by different investigators using biochemically-based assays have resulted in conflicting data [8–10]. The lack of consensus speaks to the necessity for testing those conclusions using independent methods. We chose to study binding structure function relationships in vivo using Y2H analyses, in which protein domain-domain interactions are needed to achieve growth under nutritional restriction. In the present study, we used Y2H to explore cMyBP-C’s interactions with both the thin filament (actin) and the myosin head region (S2-MyHC). We prepared nucleotide fragments containing the individual domains C0, L, C1, m and C2 (Fig. 1). In order to study potential domain interactions that might enhance or diminish the ability of an adjacent domain to interact, we also prepared various combinations encoding: C0L, C0C1, C0C2, LC1, LC2, C1C2, C1m and mC2, using standard PCR methods (Fig. 1A). To fully explore the entire protein and to rule out the possibility that we were observing non-specific interactions in our assays, we also prepared fragments encoding C3,C4, C5, C6, C7, C9, C10 and the N-terminal domains deleted from the full length cMyBP-C, encoding C2-C10, C3-C10, C5-C10 and full length cMyBP-C (Fig. 1A, Fig. S1A). The cMyBP-C fragments were subcloned into pGBKT7 and mouse cardiac actin cloned into the GAL4 activation domain of pGADT7. All cMyBP-C fragments were stably expressed within yeast cells via Western blot analysis using GAL4 DNA-BD antibody (Clontech) and appeared to phosphorylated normally as well, as determined by positive reactivity with residue-specific phosphorylation-dependent antibodies (Fig. S2). An intrinsic property of the system appeared to be that the rate of growth is somewhat dependent upon the size of the fragment. The longer fragments of cMyBP- and the full-length protein result in slow growth of the yeast colonies on restrictive media- although they do grow. Indeed, a limitation of the assay is that subdomains often interact better than full-length clones, probably reflecting some loss of domain accessibility due to protein folding of the protein. Thus slow growth, particularly of longer fragments containing multiple domains, still indicates positive interaction.
Domain interactions were confirmed by growth on SD/-His/-Leu/-Trp plates, with growth on SD/-Leu and SD/-Trp plates used to test for false positives. Growth on more stringent medium containing SD/-Ade/-His/-Leu/-Trp/X-α-gal indicates a specific interaction, eliminating false positives. The data show that the C1 and m domains productively bind to actin (Fig. 1B). Notably, in light of recent data supporting the C-terminal of cMyBP-C interacting with actin , we did not find any C-terminal domains that were able to support growth on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates (Fig. 1B, Fig. S1B). It is noteworthy that there appear to be two domains that are binding actin independently of one another. We also determined that using a fragment containing the binding domain in m, but linked to C2, abolishes any productive cMyBP-C-actin interaction mediated by that domain (Fig. 1B).
Distinguishing cMyBP-C’s S2-MyHC binding site from that of actin’s is critical for understanding the basic interactions responsible for cMyBP-C’s ability to impact on myofilament contraction. To resolve current ambiguities with respect to the sites in the myosin head region responsible for interacting with cMyBP-C, we extended our genetic approach to studying cMyBP-C:S2-MyHC binding by cloning the cMyBP-C fragments into pGBKT7 and mouse cardiac S2-MyHC into the GAL4 activation domain of pGADT7. Growth on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates indicated strong interaction with the m domain and a subset of fragments containing it, as observed for C1m and C1C2 for S2-MyHC binding (Fig. 1C). Again, when the binding domain located in m was linked to C2, the adjacent fragment on the carboxyl-terminus side (Fig. 1), all interaction was abolished. The data indicate that both S2 and actin binding are mediated, at least in part, through interactions with the cardiac-specific m domain but that multiple domains for actin interaction exist in the N-terminal portion of cMyBP-C.
Using homology algorithms, we identified putative actin binding residues in C1 (Fig. 2A) [12, 13]. Sequence alignment of the C1 domain of mouse cMyBP-C revealed 5 amino acid similarities in the actin binding domain-1 of these proteins [12, 13]. Studies of complementary interfaces for actin binding in depactin, myosin, troponin and cofilin suggested that basic amino acid-rich domains in these proteins were responsible for actin binding. We therefore mutated three lysines (K190, K193, K195) in the C1 domain of cMyBP-C to alanines (A190, A193, A195) in an effort to abolish actin binding. We subcloned the mutated-C1, mutated-C1m and mutated-C1C2 into pGBKT7, and mouse cardiac actin was cloned into the GAL4 activation domain of pGADT7 (Fig. 2B). We confirmed expression of the mutant cMyBP-C fragments and their stability within yeast cells by Western blot analysis using GAL4-BD antibody (Clontech) (Fig. S3). Growth on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates indicated strong interaction with cMyBP-C’s C1, C1m and C1C2 domains. Strikingly, the mutated-C1, mutated-C1m and mutated-C1C2 constructs did not grow on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates, confirming the importance of the 3 lysines for actin binding (Fig. 2C). We also determined that ablation of actin binding did not interfere with S2-MyHC binding as C1m and C1C2 as well as mutated-C1m, and mutated-C1C2 grew on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates (Fig. 2D).
We wished to confirm the genetic functional data using an independent, biochemical assay. We prepared recombinant His-tagged cMyBP-C protein fragments for the C1C2, and mutated-C1C2 fragments (three lysines K190, K193, K195 to alanine A190, A193, A195) as above. We then tested the actin binding properties of mutated-C1C2 recombinant proteins by high-speed co-sedimentation analysis using 1μM F-actin and increasing concentrations (0–30μM) of the mutated-C1C2 recombinant proteins [14, 15]. The data (Fig. S4) confirmed the genetically-detected interactions. That is, while wild type C1C2 co-sedimented with actin and both proteins were found in the pelleted material, mutated-C1C2 was not detected in the pellet.
As the genetic study indicated that both S2-MyHC and actin binding are mediated, at least in part, through interactions with the m domain, we mutated sets of basic amino acids to alanine in this domain in order to test their necessity for actin or S2-MyHC binding. We subcloned eight separate, mutated-m domain constructs into pGBKT7, and mouse cardiac actin and S2-MyHC were cloned into the GAL4 activation domain of pGADT7 (Fig. 3A). Stable, steady state levels of the mutated fragments within the yeast were confirmed by Western blot analysis using GAL4-BD antibody (Clontech) (Fig. S5). Growth on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates indicated ablation of the cMyBP-C:actin interaction with mut-4 (R279A and R280A), mut-5 (R279A, R280A and T281A) and mut-8 (R304A and R305A) (Fig. 3B). Similarly, the lack of growth, or greatly reduced growth on SD/-Ade/-His/-Leu/-Trp/X-α-gal plates for mut-1 (R266A, R270A and R271A), mut-5 (R279A, R280A and T281A), mut-6 (K298A, K299A and R300A) and mut-7 (K298A, K299A) with the S2-MyHC construct pointed to these residues as being critical for MyBP-C specific binding to S2 (Fig. 3C). These data point to the importance of specific residues in the m domain for both actin and S2-MyHC binding. However, with the possible exception of residues 279–281, as defined by mut-5, the residues involved in cMyBP-C’s binding to actin versus S2-MyHC, appear to be unique.
The genetic data confirmed that cMyBP-C’s m domain interacts with both actin and S2-MyHC. Lacking crystallographic data, the structural definition of MyBP-C’s binding to the thick and thin filaments remains opaque, but we questioned whether the close proximity of these binding sites in cMyBP-C rendered actin and S2-MyHC binding through this domain mutually exclusive or whether m could interact simultaneously with the two proteins. We again utilized a genetic approach to study these potential interactions: the yeast-3-hybrid system can define proteins involved in promoting or inhibiting a complex formation, as well as identify interacting domains of a protein complex . The system is designed to test if expression of a third protein results in abrogation of a complex that is normally formed when the other two proteins are expressed. The third protein is expressed under the control of the Met25 promoter, which is only active in the absence of methionine in the media. Yeast cells co-transformed with the m-S2-pBridge and Act-pGADT7 as well as m-Act-pBridge and S2-pGADT7 (Fig. S6) were plated on SD/-His/-Leu/-Trp/X-α-gal plates. As expected, growth occurred in both transformants, indicating the interaction of the m domain with either actin or S2-MyHC (Fig. 3D). The transformants were subsequently plated on SD/-His/-Leu/-Met/-Trp/X-α-gal plates, where methionine deficiency activates bridge protein synthesis. The data show that, under these conditions, growth of the yeast is severely inhibited (Fig. 3E). Thus, when actin is expressed, productive interactions between the m domain and S2-MyHC is negatively affected. Similarly, expression of S2-MyHC precludes a productive interaction between the m domain and actin, resulting in inhibition of cell growth (Fig. 3E). Together, these genetic data show that the third polypeptide, either S2 (Fig. S6A), or actin (Fig. S6B) can prevent or dramatically inhibit transcriptional activation via domain interaction of the other two and subsequent colony growth.
Data from mice homozygous for the null allele of cMyBP-C confirm that the protein is not essential for sarcomere formation [16, 17]. We wished to determine if normal actin interaction was necessary for cMyBP-C insertion into the sarcomere and whether insertion of an actin binding-defective mutant would act as a “poison peptide,” disrupting normal sarcomeric structure. To obtain complete incorporation of the mutant protein, we first defined an siRNA that could effectively knock-down expression of endogenous cMyBP-C. We defined a sequence, siRNA-1, effectively reduced endogenous cMyBP-C levels at both 3 days and 5 days after transfection (Fig. S7). Western blot analysis indicated 90% knockdown of endogenous cMyBP-C in NRCs 5 days post-transfection with 10nM cMyBP-C-siRNA-1 (Fig. 4A). Immunofluorescent staining also indicated efficient silencing of endogenous cMyBP-C expression (Fig. 4B). Notably, as was observed in cMyBP-C nulls [16, 17], sarcomeric structure was not significantly disrupted, as indicated by TnI staining (Fig. 4B) or ultrastructural examination using transmission electron microscopy (Fig. S8).
We then infected cMyBP-C-siRNA treated NRCs with wild type (wt-cMyBP-C) or actin binding deficient, C1-mutated (K190A, K193A, and K195A) cMyBP-C (mut-cMyBP-C) using adenovirus constructs. Western blot analysis confirmed endogenous cMyBP-C knock down using cMyBP-C-siRNA and replacement with roughly equivalent amounts of either wt-cMyBP-C or mut-cMyBP-C (Fig. 5A). Immunocytochemistry showed that replacement of the endogenous cMyBP-C with virus-driven expression of wt-cMyBP-C conserved normal sarcomeric structure and cytoplasmic organization (Fig. 5B). Interestingly, replacement of the endogenous cMyBP-C with the same level of mut-cMyBP-C resulted in punctate accumulations of extra-sarcomeric protein but also somewhat disrupted gross sarcomeric architecture as evidenced by less-defined striations (Figure 6B). Electron microscopy confirmed disruption of normal sarcomeric structure (Fig. S9). Sarcomere disruption was not due to general cytotoxicity induced by recombinant adenovirus transfection and/or siRNA expression, as both adenylate kinase and lactate dehydrogenase assays showed normal, control levels (Fig. S10). Thus, cMyBP-C lacking a functional actin-binding domain is stably expressed, but is unable to incorporate normally and appears to interfere with maintenance of normal sarcomere structure.
Although the ability of cMyBP-C’s N-terminal regulatory domains to influence actomyosin interaction is supported by numerous studies, the protein’s functional domains underlying its most important functions are more controversial. In light of this protein’s long history, spanning almost 40 years, it is surprising that the field still lacks clarity with respect to the site or sites on cMyBP-C responsible for thick and thin filament binding, and direct in vivo evidence for the physiological relevance of these interactions remains obscure. In the present study, we defined restricted domains by which cMyBP-C interacts with both the myosin heavy chain’s head region and the major thin filament protein, actin, using functional genetic assays, and were able to distinguish regions containing the individual binding site(s) as well as critical residues responsible for binding.
Data from a number of laboratories [4, 18–20], as well as modeling studies, have indicated that cMyBP-C’s actin binding region resides within the N-terminal regulatory domain. Similarly, genetic, biochemical and NMR spectral evidence are all consistent with cMyBP-C’s N-terminal terminal regulatory domains, C1, m, C2 and combinations thereof  as interacting with S2-MyHC to modulate myocardial contraction. Several recent studies using cosedimentation binding assays , small angle solution X-ray scattering , electron microscopy [8, 23] and NMR spectroscopy  implicated C0 as an important binding partner with actin. In contrast, biochemical experiments using cMyBP-C’s N-terminal fragments showed that C1C2 and C0C2 can bind with F-actin, while C0C1 did not have any effect. Moreover, high speed cosedimentation assays showed almost similar affinities of C0C2 and C1C2 for actin [20, 25]. Our Y2H experiments, which assay for functional binding, using the C0 domain and C0 domain-containing C0C1 fragment, did not show any interaction with actin, which is consistent with the observations that C0C1 did not inhibit actin filament velocity in the laser trap assay  and had no effect on actin filament velocity in the in vitro motility assay. A recent study using frog skeletal muscle lacking the C0 domain provides evidence for MyBP-C’s N-terminal actin-binding by electron tomography suggesting that the C0 domain might not function in F-actin-binding . Using Y2H assays, we confirmed that cMyBP-C’s N-terminal C1- and m domains are responsible for cMyBP-C/actin binding and that the m domain is responsible for cMyBP-C/S2-MyHC binding. A recent report in conflict with these data maintained that the actin binding region resided in the C-terminal portion of cMyBP-C and proposed that the previously observed N-terminal binding is non-specific in nature . However, our genetic experiments were unable to detect any functional interactions in the carboxyl half of the molecule. Thus, our data, as well recent modeling studies from other laboratories [5, 23], are not consistent with COOH-terminal binding to the thin filament.
The genetic data demonstrating the presence of multiple binding sites (C1 and m) for actin within the cMyBP-C’s N-terminal domains (Fig. 1B) are also supported by numerous studies. Using electron tomography , the single molecule laser trap assay  and electron microscopy , multiple groups have determined that the N-terminal domains can interact with F-actin. Biochemical analyses  and in vitro actin filament motility assays , implicated the C1-m region, while electron microscopy and 3D reconstruction experiments with a C0C3 fragment implicated the C0, C1 and m domains bind on or near subdomain 1 on F-actin . Interestingly, our Y2H experiments using a fragment containing the C0C1 domains did not show any interaction with actin, which is consistent with the observations that C0C1 did not inhibit actin filament velocity in the laser trap assay  and had no effect on actin filament velocity in the in vitro motility assay .
Our genetic experiments confirmed that both the C1 and m domains can independently bind to actin. While this approach was able to define unique binding sites, it also highlighted the effects of spatially separate domains on these protein interactions. For example, while the C0-L-C1 fragment was unable to interact with actin, C1 alone interacted strongly. Similarly, while the m domain resulted in robust growth under stringent selection, when coupled to C2, the m-C2 fragment was unable to bind. This implies that intra-molecular interactions as well as intermolecular proximity and binding are probably important in mediating cMyBP-C-actin binding. These data illustrate both the strengths and limitations of reductionist approaches in defining protein-protein interactions in terms of individual domains. At this point our data indicate the importance of both intra-molecular interactions as well as intermolecular proximity and binding in mediating cMyBP-C-actin binding. The data do point to the critical domains and residues that can be targeted within the context of the whole molecule and that mutated cMyBP-C used to replace endogenous protein via transgenic approaches [27–30]. Using this approach, the in vivo functionality of the domain and residues can be unambiguously confirmed.
Using an in silico approach, we identified potentially critical actin binding residues in the cMyBP-C’s C1 domain based on similarities to a consensus actin binding site in α-actinin, human-β-spectrin, human-dystrophin and human-utrophin. A recent, biophysically-based study is consistent with these residues forming a bona fide actin binding site. Using15N-heteronuclear single quantum coherence NMR spectroscopy, four residues within the C1 domain were identified as being critical for actin binding . We identified 3 lysines in this region (K190, K193, K195), and mutation of the 3 residues to alanine (A190, A193, A195) abolished actin binding, confirming their essential role in thin filament interaction (Fig. 2). Their functional significance was confirmed by replacing endogenous cMyBP-C with either normal or actin binding-ablated (K190A, K193A and K195A) protein. Expression of the K190/193/195A mutation resulted in aberrant cMyBP-C trafficking and partial sarcomere dysgenesis that was evident both by immunohistochemical observation and at the ultrastructural level.
NMR spectroscopy and small-angle neutron scattering data using human cMyBP-C fragments suggested that C0 and C1 interact with myosin and actin using a common set of binding determinants . The Y2H assay data do not support a strong interaction of the mouse C0 domain with actin, or the C0-C1 domains with S2-MyHC as proposed by other laboratories using biochemically-based experiments [24, 31]. Rather, our data, which depend upon productive, domain-domain specific interactions to drive transcription, point to cMyBP-C’s regulatory m domain as being essential for binding to S2-MyHC but containing a second, distinct site that binds actin. This domain, which is phosphorylatable, is unique to the cardiac isoform. Phosphorylation of the regulatory m-domain modulates the interaction of cMyBP-C with S2-MyHC: phosphorylation-induced ablation of S2-MyHC binding accelerates cross-bridge cycling and stretch activation by (presumably) modulating the proximity and subsequent interaction of myosin and actin .
The importance of an intact cMyBP-C S2-MyHC interaction is emphasized by the identification of mutations linked to FHC in the S2-MyHC, which leaves the structure of the coiled-coil intact but interferes with binding to cMyBP-C . In the present study we were able to define cMyBP-C’s m domain as being both necessary and sufficient for S2-MyHC binding using a combination of site directed mutagenesis and genetic assays. The basic residues preceding the three major phosphorylation sites (Ser273, Ser282 and Ser302) mediate cMyBP-C’s interaction with the myosin head. The spatial proximity of the binding sites to three sites that undergo rapid and reversible phosphorylation are consistent with previous data from our laboratory indicating the importance of cMyBP-C phosphorylation in modulating the protein’s overall phosphorylation and myocardial function .
Three arginine residues (R266, 270, 271) within the m domain are critical for S2 binding but appear to be dispensable for actin interaction. A recent study  suggested that mutation of residues R266, R270A, and R271 to alanines should decrease actin binding (shown as mutant 1 in Fig. 3). However, our data indicate that those residues do not, by themselves affect actin binding although their mutation completely abolished binding to the S2-myosin fragment. We hypothesize that the primary basis of actin filament motility inhibition measured in their assays could be due to cMyBP-C’s actin and/or S2-MyHC binding. At a higher level of resolution, R266 appears to be a critical residue for S2 binding: if we mutate R270 and 271, S2 binding can occur. Two proximal arginine residues (R279, R280) are critical for actin binding but have no effect on S2 binding, confirming that the sites being detected are unique. However, mutation of the adjoining T281 to alanine ablates S2 binding as well, implying that at least part of what may constitute a minimum actin or S2 binding site may functionally overlap or influence one another. R304 and R305 within the m domain also help to mediate actin binding but mutating them to alanine did not affect S2 myosin binding. This close proximity of portions of the actin and S2-MyHC binding sites implied the potential for interference and the yeast-3-hybrid assays (Fig. 4) were consistent with this hypothesis, as expression of either actin or the S2-MyHC fragment interfered with the ability of the domain to interact with the other protein.
Definition of the residues in cMyBP-C’s N-terminal regulatory domains responsible for cMyBP-C/actin and cMyBP-C/S2-MyHC binding will inform our understanding of the pathophysiology of cMyBP-C associated cardiomyopathy. For example, the missense mutation, R282W causes moderate ventricular hypertrophy, with sudden death reported in two elder family members . This region lies within the m domain, which participates in both myosin and actin binding and clearly, ablation of actin binding can cause significant sarcomeric disarray.
While interaction data must be treated cautiously, we are encouraged by these first data sets. First, the data are consistent with the majority of previously published studies using a variety of biophysical and biochemical approaches; these comparisons are shown and referenced in Table 1. Second, if non-specificity were an issue, we would expect to see some interactions with fragments containing the carboxyl terminus. Finally, confirmation that the actin site-ablated cMyBP-C protein shows aberrant cytoplasmic trafficking and appears to actively impact on sarcomere integrity, lends credence to the hypothesis that these assays are, in fact, detecting functionally important residues. However, the data’s relevance to human disease must be treated cautiously as species-specific differences in cMyBP-C’s N-terminal regulatory domains exist . Species-specific sequence analysis indicated that cMyBP-C’s N-terminal 4 domains share more than 80% identity between human and mouse isoforms (with the C1, M, and C2 domains each >90% identical) while the Pro-Ala rich linker regions are less well conserved showing only 46% homology . Sequence analysis of the Pro-Ala rich regions indicated that the mouse Pro-Ala rich region was composed of 28.2% proline and alanine residues, while the human isoform contains 51.0%. These differences in species-specific sequence homology in the human and mouse Pro-Ala rich regions might explain some of current, conflicting data that were obtained using either the human  or mouse protein  in terms of defining the functional domains.
Our data illustrate a complementary approach and a first iteration in identifying cMyBP-C’s critical domains and residues. The mutations in the relevant domains can now be subsequently be tested using in vivo transgenic systems in order to precisely define the structure function relationships that underlie normal and abnormal cMyBP-C function within the functional context of the whole organ and animal.
This work was supported by National Institutes of Health grants P01HL69779, P01HL059408, R01HL05924, R011062927 and a Trans-Atlantic Network of Excellence grant from Le Fondation Leducq (to J.R.) as well as American Heart Association Postdoctoral Fellowship grants (to S.B., M.G.).
The authors declare that there are no conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.