HYPERCONTRACTILE PROPERTIES OF CARDIAC MUSCLE FIBERS IN A KNOCK-IN MOUSE MODEL OF CARDIAC MYOSIN-BINDING PROTEIN-C

Myosin-binding protein-C is a component of all striated-muscle sarcomeres, with a well-established structural role and a possible function for force regulation. Multiple mutations within the gene for cardiac MyBP-C, one of three known isoforms, have been linked to familial hypertrophic cardiomyopathy. Here we generated a knock-in mouse model that carries N-terminally shortened cardiac MyBP-C. The mutant protein was designed to have a similar size as the skeletal MyBP-C isoforms, while known myosin and titin binding sites, as well as the phosphorylatable MyBP-C motif, are not altered. We show that mutant cardiac MyBP-C is readily incorporated into the sarcomeres of both heterozygous and homozygous animals and can still be phosphorylated by cAMP-dependent protein kinase. Although histological characterization of wildtype and mutant hearts did not reveal obvious differences in phenotype, left ventricular fibers from homozygous mutant mice exhibited an increased Ca 2+ sensitivity of force development, particularly at lower Ca 2+ concentrations, while maximum active force levels remained unchanged. The results allow us to propose a model of how cMyBP-C may affect myosin-head mobility and to rationalize why N-terminal mutations of the protein in some cases of familial hypertrophic cardiomyopathy could lead to a hypercontractile state.


Introduction
Myosin-binding protein-C (MyBP-C) (1, 2) is a myofibrillar protein that contributes to the structural integrity of the sarcomere and possibly is involved in the regulation of contraction (3). Three different isoforms of MyBP-C have been identified: skeletal (sMyBP-C) fast and slow, and a cardiac-specific variant (cMyBP-C; Fig. 1A), each of these being coded for by a distinct gene (4,5). All isoforms interact at the C-terminus with the rod portion of myosin (e.g., Ref. 6), as well as with titin (e.g., Ref. 7), and thus, help maintain an ordered thick-filament structure (reviewed in Ref. 3). MyBP-Cs are modular polypeptides that belong to the intracellular immunoglobulin (Ig) superfamily. Whereas the skeletal variants consist of 10 globular domains of the Ig-like or fibronectin-type-III-like fold (4), cMyBP-C contains an additional N-terminal Ig module termed C0 (5). Between the Ig domains C1 and C2, MyBP-Cs also contain a stretch of about 100 residues, the so-called MyBP-C motif (Fig. 1A), which in cardiac muscle can be phosphorylated at three sites by cAMP-dependent protein kinase (8).
The MyBP-C motif was shown to bind to the S-2 segment of myosin, close to the lever arm domain of the myosin head (9). Interestingly, this interaction is dynamically regulated by phosphorylation/dephosphorylation of the MyBP-C motif (10). Moreover, the controlled interaction with the myosin hinge region appears to affect the contractile behavior of muscle fibers (11) and thus, could represent a potential regulatory mechanism of contractility (12).
These hints notwithstanding, direct evidence for a role of cMyBP-C in force regulation has been difficult to obtain.
Uncovering the functions of cMyBP-C is interesting from a clinical point of view as the protein is involved in the pathophysiology of familial hypertrophic cardiomyopathy (FHC) (13,14). This inherited disease occurs in autosomal-dominant fashion and affects ~0.2% of the general population. FHC is known to be a disease of the sarcomere: mutations in at least eight different sarcomeric protein genes have been identified so far (14,15). Mutations in cMyBP-C account for approximately 15-20% of genetically defined FHC cases, but the cMyBP-C-linked types of FHC present as relatively benign phenotypes with mild hypertrophy at mid-life (16,17). Most cMyBP-C lesions show C-terminally truncated polypeptides lacking either the myosin or myosin and titin binding sites, but some lesions are also due to missense mutations occurring in more N-terminal regions of the protein (16). Genetical  of cMyBP-C (18)(19)(20). These model systems have demonstrated the importance of the Cterminus of cMyBP-C for a regular sarcomeric structure and normal contractility of the heart.
In the present study we used knock-out-knock-in technology to generate mice (hereafter termed "knock-in mice") with N-terminal deletion of a region of cMyBP-C comprising one Ig domain and a linker sequence next to the MyBP-C motif. The shortened cMyBP-C (Fig. 1A) thus has a domain structure similar to that of sMyBP-C. Notably, within the region affected by the knock-in, a missense mutation has been described for a family of FHC patients exhibiting a distinct phenotype (16). We show that the cMyBP-C deletion variant is expressed in both homozygous and heterozygous mice at the protein level and is readily incorporated into the sarcomere. Animals carrying the deletion are viable, show no significant ultrastructural changes of the heart, and appear to have a normal life span. Mutated cMyBP-C could still be phosphorylated by cAMP-dependent protein kinase, but skinned muscle fibers from homozygous mutant hearts revealed a leftward shift in the force-pCa curve and a decreased slope of that curve. The increased Ca 2+ sensitivity may result from decreased steric hindrance of myosin-head mobility due to the expression of the shorter cMyBP-C. We discuss the possibility that the additional N-terminal Ig domain present in cardiac versus skeletal MyBP-C could be included by nature to aid force regulation at the crossbridge level in the heart. Our findings also provide a starting point to explain the development of hypertrophied cardiac tissue in FHC cases with N-terminal mutations of cMyBP-C.

Gene targeting
A P1 clone containing the murine cardiac MYBP-C sequence was obtained from a mouse 129 P1 genomic library (Genome Systems, St. Louis, MO). A 9.1 kb EcoRI fragment from the P1 clone was isolated and subcloned, and found to contain the 5' prime end of the gene from exon 1-20 (Fig. 1B). A 1.7 kb StuI/EcoRV fragment (including exon 2) located upstream of exon 3 and a 5.3 kb NsiI/EcoRI fragment (including exon 7-20) located downstream of exon 6 were used as the 5' and 3' homology units.
The targeting vector was constructed by standard recombinant techniques. A genomic fragment of the MYBP-C gene (1.3 kb) including exons 3-6 was deleted and replaced by a neomycin resistance gene (Fig. 1B). The vector contained a herpes simplex thymidine kinase cassette for negative selection of single recombinant embryonic stem (ES) cell clones. Also, the vector included a unique ClaI restriction site for linearization of the plasmid. Homologous recombination between targeting vector and cognate cMyBP-C locus deleted exons 3-6.
Colony selection and target clone identification were done as described elsewhere (21).
Targeting vector (20 µg) was introduced into 1.2 * 10 7 ES cells by electroporation. Genomic DNA was prepared as described (22). Correct targeting of G418-resistant clones was analyzed by Southern blotting.
Clones were subsequently tested by long PCR assay (Combi Pol/InViTek, Berlin-Buch). To check for the occurrence of new recognition sites on the amplificates, Southern blotting was employed. Correctly targeted clones were microinjected into C57/BL 6 blastocysts, which were implanted into pseudo-pregnant CB6 mice bred to produce heterozygous or homozygous mutant animals.
MyBP-C mRNA was assessed by nucleotide sequence analysis of RT-PCR-amplified DNA fragments according to standard protocols. The following primer pairs were used: Northern blotting of cardiac/skeletal-muscle RNA was performed as described (23).

Morphology and microscopy
Excised hearts were rinsed in 4% paraformaldehyde and weighed, and cardiac tissue was examined for pathological alterations (24). Examined parameters included heart weight, left ventricular wall thickness and cavity size, and myocyte nuclear size (measured by outlining the nucleus in 150 cardiac cells cut in their long axis).

Indirect immunofluorescence microscopy
Bundles of myofibrils prepared from left ventricle essentially as described (25) were examined under a Zeiss Axiovert 135 microscope. MyBP-C was visualized by using antibodies against the MyBP-C motif (25).

Cardiac fiber mechanics
Freshly excised mouse hearts were retrogradely perfused through the aorta with 4°C total ionic strength, 130) containing 0.25-0.5% Triton X-100 (26). A relatively low buffer pH used in skinned fiber mechanical studies was reported to be beneficial for the functional preservation of the regulatory system (27). After washes with fresh relaxing buffer, fiber bundles 150-200 µm thin and 3-4 mm long were mounted isometrically between a positioncontrolled rigid post and a force transducer (AME AE 801, Horten Electronics, Norway) with nitroacetate glue (26). Sarcomere length was adjusted to 2.2 µm by laser diffractometry. After removal of BDM and addition of ATP regenerating system (creatine phosphate, 10 mM; creatine kinase, 150 U/ml) to the solution, fibers were activated by transfer from relaxing to activating buffer, in which EGTA was substituted by Ca-EGTA. The desired Ca 2+ concentration was calculated as described (26). Experiments were carried out at room temperature.
The normalized force-pCa relationships, in which force was expressed relative to the maximum force usually developed at pCa 4.34, were fitted to the Hill equation: where HC (the Hill coefficient, a measure of cooperativity) and K c are constants.

Gel electrophoresis and 32 P autoradiography
Fiber bundles prepared as described above were washed with relaxing solution (ATP, 4 mM). Specimens were incubated with the catalytic subunit of protein kinase A (SIGMA, 500 U/ml relaxing buffer) in the presence of [γ-32 P]ATP (specific activity, 250 µCi/µM) for 45 min at room temperature (26,28). Proteins were then denatured, dissolved, and electrophoresed on 8% SDS-polyacrylamide gels. Major myofibrillar proteins were identified by Coomassie staining. 32 P incorporation was visualized by autoradiography, using a 4-12 hour exposure time with standard Kodak x-ray film (28,29).

Generation of mutant cMyBP-C mice
To target the MYBPC gene, a 9.1 kb fragment of the murine cardiac MYBPC locus encompassing exons 1-20 was isolated and subcloned (Fig. 1B). The targeting construct was designed to selectively remove exons 3-6 (1.3 kb), thus producing a deletion of the Ig domain C1 and the linker region between domains C0 and C1 of cMyBP-C (Fig. 1A). Ninety six G418-resistant ES cell clones were analyzed, and genomic Southern blotting of DNA from ES cell clones was performed to detect the targeting event. Correct targeting was found in 6 clones: as shown in Fig. 2A, a band corresponding to a 9.1 kb fragment was detected as the wildtype allele and a band corresponding to a 2.7 kb fragment as the targeted allele. Correctly targeted clones were used for blastocyst-mediated transgenesis and production of chimeric animals. Appropriate breeding produced mice either homozygous or heterozygous for the cMyBP-C deletion. These mice were fertile, produced normal litter sizes, and survived for >1 year. We also tested the correctly targeted clones in a long PCR assay. Primers were designed such that a 2.2 kb fragment was produced specific for the wildtype allele and a 1.9 kb fragment specific for the targeted allele (data not shown).
Additionally, a PCR was done with a 4.8 kb product (Fig. 1B, top). The analysis showed that the restriction enzyme EcoRI cut the amplificate only of the targeted allele into a 2.8 kb and a 2.0 kb fragment (Fig. 2B), indicating the introduction of a new recognition site.

cMyBP-C expression in mutant mice
To determine the expression of cMyBP-C transcripts in mutant mice, we performed RT-PCR analysis with various primer pairs from different regions of heart cDNA (Fig. 2C).
With a primer pair encompassing domains C0 to C1, a signal was obtained only for cDNA from wildtype or heterozygous mice, whereas in homozygous mutant animals, RNA encoding the C1 domain plus N-terminal adjacent linker was not expressed (Fig. 2C, panel a). This observation is consistent with the results of Northern blot analyses (Fig. 2D): no signal was 9 detected when total RNA, isolated from homozygous hearts, was hybridized with a C1+linker probe. By contrast, a signal was detectable in heterozygous hearts. In comparison, when skeletal-muscle RNA was used, no signal was present (Fig. 2D). Hybridization with a probe of the MyBP-C motif revealed a normal signal for heart RNA in all types of animals, but none for skeletal-muscle RNA. Controls with a GAPDH probe showed a signal in all lanes.
By RT-PCR, using primer pairs encompassing the C0 domain and the MyBP-C motif, we detected the expected deletion of 474 base pairs (Fig. 2C, panel b). Cloning and sequencing of the products highlighted by the asterisks (Fig. 2C) revealed that the upper band corresponds to the wildtype DNA sequence, whereas the lower band product has the same flanking sequence but contains the predicted deletion. We note that in competitive PCR's, a shorter product tends to show a larger signal than a longer product, as seen in Fig. 2C, panel b.
This figure, as well as the RT-PCR at the 3' prime UTR region (Fig. 2C, panel c), demonstrate that the transgenic RNA is stable and well expressed; no degradation or lowered expression was detectable. Thus, regulation at the transcription level seems unlikely.
Both homozygous and heterozygous mice were found to express the deletion mutant also at the protein level. Western blot analyses with muscle protein obtained from all types of animals revealed a distinct band stained by a polyclonal antibody against the MyBP-C motif (Fig. 3A). Moreover, cardiac myofibrils labelled with fluorophore-marked α-MyBP-C antibodies exhibited the expected staining pattern in the sarcomeric A-band; no obvious difference in staining intensity or regularity of labelling was found between wildtype and homozygous mutant animals (Fig. 3B). Thus, mutant mice stably expressed the shortened cMyBP-C protein.

Histological characterization
Hearts from several months (up to ~1 year) old animals (n=7, for each animal type) were examined for histological and morphological abnormalities ( Table I). None of the parameters investigated differed between animal types in a statistically significant manner, although one homozygous mutant heart revealed an abnormal phenotype with strongly increased values for all four parameters. In general, however, no evidence was found for cardiac hypertrophy, myocyte loss or inflammation. Thus, the histological appearance of heterozygous or homozygous mutant hearts appeared to be normal.

Fibers from homozygous mutant hearts show increased Ca 2+ sensitivity of force generation
Left ventricular muscle strips obtained from wildtype and mutant, litter-matched, animals were probed for their contractile properties by measuring the active force of skinned fiber bundles as a function of the Ca 2+ concentration. Since the Ca 2+ sensitivity varies with sarcomere length, laser diffractometry was used to set this length to 2.2 µm in all experiments.
A typical example demonstrating the force rise with increasing [Ca 2+ ] (i.e., decreasing pCa) is shown in Fig. 4, inset. Fibers from 5 wildtype, 2 cMyBP-C (+/-) , and 6 cMyBP-C (-/-) mice were included in the analysis, and 5-7 fiber bundles per animal were investigated. A summary of results is presented in the main Fig. 4

The N-terminal deletion mutant is still phosphorylated by cAMP-dependent protein kinase
To find out whether the altered Ca 2+ sensitivity could be related to an altered response of the mutant cMyBP-C to activation by cAMP-dependent protein kinase (since the deletion is close to the MyBP-C motif), we tested the protein's ability to be phosphorylated by this kinase and C1 (Fig. 1). Thus, the mutant mice contain a shorter-than-normal cMyBP-C molecule whose size and domain architecture resemble those of the skeletal isoforms.
We bred the mice bearing mutant MyBP-C alleles to homozygosity, because we expected a relatively mild effect on cardiac structure and/or function: MyBP-C is not found in the entire A-band but forms 7-9 stripes in the A-band's C-zone on either side of the M-line (32,33). Indeed, hearts from both heterozygous and homozygous mutant mice showed no statistically significant changes at the ultrastructural level, and no differences were found between wildtype and heterozygous animals in terms of the force response of skinned cardiac fibers to Ca 2+ -dependent activation. In contrast, fibers from homozygous mutant mice showed 12 an increased Ca 2+ sensitivity of force production (Fig. 4) while maximum force levels remained unchanged. This finding is consistent with that of an earlier study reporting that active tension at submaximal Ca 2+ concentrations was increased, but maximum tension was not affected, upon partial extraction of MyBP-C from rat skinned cardiomyocytes (34). We note that extraction of MyBP-C was shown to slightly increase Ca 2+ sensitivity at low to intermediate Ca 2+ concentrations also in rabbit psoas muscle fibers, but the effect was much smaller than in cardiac cells (34). Our results extend the previous findings, suggesting that at least part of the change in Ca 2+ sensitivity of cardiac sarcomeres may be related to a functional role of the N-terminal cMyBP-C domains.
One family of FHC patients has been described bearing a missense mutation (Glu258Lys) in the region just N-terminal to the MyBP-C motif (16). It is not unlikely that an altered function of N-terminal MyBP-C domains is responsible for the hypertrophy phenotype found in some members of this family. However, also clinically healthy individuals can carry the mutant MYBPC3 allele (5,16). Moreover, mutations in MYBPC3 are frequently characterized by a mild phenotype particularly in young patients and a delayed age at the onset of symptoms (16,17). Then, since the physiological background of human and mouse is very different, it is possible that the life span of the knock-in mice of this study is not long enough for significant changes of cardiac ultrastructure (and contractile properties of heterozygous animals) to occur. On the other hand, we demonstrated that the mutant protein is expressed and incorporated into the sarcomeres, which was associated in homozygous mutant animals with enhanced contractile performance. Taken together, N-terminal cMyBP-C mutations, if occurring in human heterozygous FHC patients, might in some instances determine a "hypercontractile" state that could induce cardiac hypertrophy directly. In this context we point out that a hypercontractile hypothesis has been put forth for some FHC cases in which other sarcomeric proteins are mutated, such as α-tropomyosin (35,36). FHC might therefore be a disease induced by mutations causing either functional cardiac impairment followed by compensatory hypertrophy (apparently the majority of all cases) or functional enhancement followed by direct cardiac hypertrophy (14).
What mechanism(s) could be envisioned to explain the observed functional effect of the N-terminal deletion in cMyBP-C? Although the exact layout of MyBP-C in the thick filament is still subject to debate, structural details known to date (3,10,37)  a plausible model demonstrating the protein's regulatory input. As shown in Fig. 6, cMyBP-C binds to the myosin rod (and titin) at the C-terminus and also to the myosin neck region with the MyBP-C motif (in the unphosphorylated state). In the wildtype protein (Fig. 6A), the C0 domain at the N-terminus could well interfere with the myosin head region, either by proximity to the regulatory light chains as proposed (10) or through specific interaction with the head. Indeed, a recent study suggested that the C0 domain of human cMyBP-C contains a novel putative myosin-binding site (38). Thus, in cardiac sarcomeres, MyBP-C could mechanically constrain crossbridge movement in a manner not found in skeletal muscle.
Phosphorylation-induced unbinding of the MyBP-C motif from the myosin neck region (Fig.   6A, asterisk) would release some constraints from the myosin head, thereby providing cardiac cells with an additional mechanism to regulate force development. In the case of N-terminally shortened cMyBP-C (Fig. 6B), the molecule may not be able to reach the myosin head region, which would change the flexibility or mobility of the crossbridge permanently. Even though the number of myosin heads whose mobility can be affected by cMyBP-C is limited (because many heads lie outside of the C-zone), mechanical coupling of crossbridges within a thick filament (12)       In contrast, Ca 2+ sensitivity of homozygous mutant fibers (n=38) was significantly increased at modest to high pCa. The pCa 50 value was shifted leftward by 0.07 pCa units; the slope of the curve was decreased. Statistically significant differences to wildtype specimens were confirmed by unpaired Student's t-test (*, p<0.05; **, p<0.001). Values are mean ± S.E.M. Phosphorylation of TnI is known to cause a distinct decrease in Ca 2+ sensitivity of force.