OUP user menu

Properties of mutant contractile proteins that cause hypertrophic cardiomyopathy

Charles S. Redwood , Johanna C. Moolman-Smook , Hugh Watkins
DOI: http://dx.doi.org/10.1016/S0008-6363(99)00213-8 20-36 First published online: 1 October 1999

Abstract

Hypertrophic cardiomyopathy (HCM) is one of the most frequently occurring inherited cardiac disorders, affecting up to 1 in 500 of the population. Molecular genetic analysis has shown that HCM is a disease of the sarcomere, caused by mutations in certain contractile protein genes. To date seven disease-associated genes have been identified, those encoding β-myosin heavy chain, both regulatory and essential myosin light chains, myosin binding protein-C, cardiac troponin T, cardiac troponin I and α-tropomyosin. Here we review the analyses of how these mutations affect the in vitro contractile protein function and the hypotheses derived to explain the development of the disease state.

Keywords
  • Contractile proteins
  • Heart muscle
  • Cardiomyopathy
  • Hypertrophy
  • Mutation

Time for primary review 22 days.

1 Introduction

1.1 Background

Hypertrophic cardiomyopathy (HCM) is one of the most frequently occurring inherited cardiac disorders [1], and the first to be elucidated at a molecular genetic level. HCM has been studied both because it is a clinically important condition, being the commonest identified cause of sudden death in young adults, and because of its potential relevance as a genetic model of cardiac hypertrophy, in itself an indicator of increased risk of death [2]. Therefore, it is hoped that an understanding of the molecular aetiology of HCM will be of broad relevance. Over the last 40 years, HCM has been described in detail as a primary cardiac muscle disorder, with certain characteristic features among affected families, but with marked clinical heterogeneity [3,4]. The hallmark abnormality, that of unexplained cardiac hypertrophy, is variable in degree and severity, as is the extent of the underlying myocardial disarray; accordingly, the time of onset and nature of symptoms are also very variable. In some families HCM is a malignant disease with a high incidence of sudden death, but in the main (and particularly in families ascertained outside of referral centres) the natural history is more benign [5].

1.2 Molecular genetics – a disease of the sarcomere

Over the last 6 years, seven different disease-genes have been identified for HCM. These genes have, in the main, been defined by linkage analyses in large families, followed by a positional-candidate gene approach. The genes encode proteins which are components of either the thick filament (β-myosin heavy chain (gene MYH7) [6,7], regulatory myosin light chain (MYL2) [8], essential myosin light chain (MYL3) [8] and myosin binding protein-C (MYBPC3)) [9,10] or the thin filament (cardiac troponins T (TNNT2) [11,12] and I (TNNI3) [13], α-tropomyosin (TPM1) [11]) of striated muscle (see Fig. 1). Moreover, within each of these disease-genes a variety of individual mutations have been found, and it appears that most families have ‘private’ mutations. Thus, this apparently single clinical syndrome, in fact, represents a group of genetically distinct disorders. Nevertheless, it remains reasonable to consider HCM as a single entity, as all of the known disease genes confirm that HCM is a disease of the cardiac muscle sarcomere.

Fig. 1

Diagrammatic representation of the molecular structure of the sarcomere showing the thick (myosin) and thin (actin) filaments and the location of the proteins–β myosin heavy chain, myosin light chains, myosin binding protein-C, α-tropomyosin, troponin T and troponin I – that are mutated in HCM. From Ref. [5] with permission.

In order to understand how mutations in these different contractile proteins all cause cardiomyopathy and hypertrophy, it will be necessary to understand the functional consequences of the mutations at a molecular level. It might be anticipated that the different mutant proteins cause similar functional abnormalities, which subsequently activate the same disease pathways. However, the unifying hypothesis – that HCM is a disease of the cardiac sarcomere – belies the fact that the different proteins involved, although members of the same functional apparatus, have very different properties and roles. Some have enzymatic and force-generating roles (e.g. myosin heavy chain), while others play structural roles (e.g. myosin binding protein-C) or have regulatory functions (e.g. troponins T and I and α-tropomyosin). This review will, therefore, survey the functional consequences of HCM mutations in these diverse proteins as a starting point for understanding the initiation of, and subsequent pathways involved in, the development of hypertrophy.

In addition, genotype:phenotype analyses of HCM have indicated that the molecular genetic heterogeneity is a major determinant of the clinical heterogeneity seen in this condition. Although this review will only briefly look at the insights gained from such correlations, it is instructive to focus on the chief characteristics of HCM arising from the different disease genes and mutations, as this informs interpretation of the functional studies at a molecular level. Conversely, clearer understanding of the different paradigms by which HCM is produced is helpful in understanding variants of the clinical disease and may lead to the design of more rational therapies.

1.3 Insights from genotype:phenotype analyses

Clinical studies that preceded knowledge of the molecular genetic basis of HCM indicated that HCM families could be categorised by certain quantitative differences in presentation – such as families with either a ‘malignant’ or ‘benign’ natural history. It has, however, become clear that intrafamilial variation is also marked, particularly with regard to the morphological features of the disease [14]. As relatives in a given family must share an identical causal mutation, this variability indicates a significant modifier role for genetic background and/or environmental influences in the disease phenotype. It was, therefore, not unexpected that heterogeneity at the level of the primary genetic abnormality could only partly explain what was seen clinically. Nevertheless, significant quantitative differences in various aspects of the disease, such as the risk of sudden death, length of survival, severity of hypertrophy (and, hence, disease penetrance, that is, the proportion of individuals with clinical identifiable disease), have been identified. These differences exist both between different disease genes, and between different mutations within a given disease gene.

Thus, distinct missense mutations in the gene for β-myosin heavy chain (βMHC) appear to result in major differences in survival, which are not obviously mirrored by matching differences in other clinical features in these families [7,15–17]. The degree of hypertrophy seen with MHC mutations is very variable and demonstrates at best a weak correlation with the risk of sudden death. However, in families with βMHC mutations that are associated with a poor prognosis, the hypertrophy tends to be quite marked and penetrance nearly complete [16]. In contrast, in individuals with mutations in cardiac troponin T (cTnT), commonly associated with a high incidence of sudden death comparable to that seen with severe βMHC mutations, the hypertrophy is most often mild and sometimes clinically undetectable [12,18]. This observation underscores the clinical utility of genetic testing in TnT HCM families and, importantly, also suggests that the pathways leading to hypertrophy differ from those that determine the risk of sudden death.

Recent data indicate that individuals bearing cardiac myosin binding protein-C (MyBPC) mutations tend to develop hypertrophy at a later age, and this hypertrophy continues to progress through adult life [19,20]. This is in marked contrast to all previous clinical descriptions of this disease which describe the onset of hypertrophy in adolescence, with little or no subsequent progression of hypertrophy in later life.

Although individual mutations seem to be associated with significant quantitative differences in a given parameter, it does not appear that specific mutations or disease-genes are associated with qualitatively different forms of the clinical disease. Early data regarding mutations in the myosin essential and regulatory light chains (MELC and MRLC) suggested that these may be associated with a specific sub-type of HCM with papillary muscle thickening and mid-cavity obliteration [8]; however, mutations in these genes have since been associated with classical patterns of HCM [21]. Likewise, apical HCM is sometimes seen with mutations in disease genes more typically associated with classical forms of the disease [13,22].

1.4 The mechanism of the dominant phenotype

Most HCM-causing mutations in contractile protein genes are relatively subtle mutations, typically missense mutations changing just one DNA nucleotide and resulting in the substitution of just one amino acid in the particular protein. This raises the question of the mechanism by which such seemingly minor changes result in disease. It appears that most, and perhaps all, mutations that cause HCM do so by a dominant-negative action [23]. In this model, the mutant peptide is stably expressed and is able to incorporate into the sarcomere, but then results in abnormalities either in the further assembly, or in the function, of the multimeric array of contractile proteins that constitute the sarcomere [24]. Direct experimental evidence has been necessary to document this model for each of the disease genes and includes direct observation of incorporation into the sarcomere [25,26], demonstration that null alleles which completely inactivate the gene do not produce the disease [27], and, finally, demonstration that introduction of the mutant peptide through gene targeting or transgenesis in the mouse can recapitulate aspects of the phenotype [28,29].

Proof of the dominant-negative mode of action of HCM-causing mutations has important implications. If, instead of following the dominant-negative model, HCM had resulted from an imbalance in stoichiometry of components needed for self-assembly of the sarcomere, there would be no merit in in vitro analyses of mutated proteins implicated in HCM. However, in the light of the dominant-negative pathogenesis, the starting point for an understanding of this disease (and its relevance to other forms of hypertrophy) must be a careful biochemical, biophysical and physiological analysis of the mutant proteins and the deficits they cause.

2 Analyses of mutant proteins in HCM

2.1 β-myosin heavy chain (βMHC) and myosin light chains (MLCs)

2.1.1 Peptide structure and sites of mutations

Muscle myosin is a hexamer consisting of two heavy chains (MHC) along with two regulatory light chains (RLC) and two essential light chains (ELC). Each MHC consists of a globular head joined to an α-helical rod by a hinge region. The tails of the two MHCs dimerise to form a coiled-coil rod of about 15 nm. The heads each contain a catalytic ATPase site and an actin-binding site; additionally, one RLC and one ELC are bound to each head at the head-rod junction region, referred to as the lever arm helix. The head and head-rod junction region (referred to as subfragment-1 [S1] because of its generation by proteolysis of whole myosin) of the heavy chain has been further subdivided, based on the results of limited tryptic digestion, into an N-terminal 25 kDa fragment, a central 50 kDa fragment and a C-terminal 20 kDa fragment. These were initially thought to be discrete tertiary domains and, although that is no longer thought to be the case, are still used to describe regions of the protein. Both light chains are members of the EF-hand protein superfamily, in common with troponin C and calmodulin [30], and have both structural and regulatory roles in myosin function [31].

The structure of S1 from chicken skeletal muscle myosin has been solved by X-ray crystallography [32]. Since there is high sequence conservation between different myosin molecules [33,34], the chicken skeletal muscle S1 model has been used to locate HCM mutations in the three-dimensional structure of the protein [35] (Fig. 2). More than 50 MHC mutations have been described in families or probands with HCM; the bulk of these are found clustered in four particular locations in myosin S1 (see Figs. 2 and 3). These are the actin-binding interface; the nucleotide-binding pocket, an area adjacent to the region which connects two reactive thiols in the hinge region at the base of the lever arm helix; on the lever arm helix, near the site of interaction with ELC (20 kDa subdomain). Additionally, a few mutations are clustered in the incipient rod region of the myosin peptide [35]. Whilst an HCM-associated deletion has been detected in the carboxy-terminal region of βMHC, it was of undetermined significance as it did not appear to segregate with the disease in the affected family [36].

Fig. 3

A diagram showing the ‘domain’ structure of β myosin heavy chain based on proteolytic digestion and the location of selected HCM-associated mutations [6,7,16,27,35,103–106], and a table summarising the reported changes in in vitro properties of the HCM-mutant proteins compared with those of the wild type.

Fig. 2

Ribbon diagram of the structure of chicken skeletal muscle myosin subfragment 1 showing the functional domains and the location of residues homologous to certain HCM-associated mutations in β myosin heavy chain and the regulatory and essential myosin light chains. The S1 tryptic 50, 25 and 20 kDa fragments are shown in red, green and blue, respectively; the RLC is shown in purple, the ELC in yellow. Adapted from Ref. [8].

Visualisation of the position of described mutations in the three-dimensional structure of myosin S1 has facilitated conjecture of how the mutations might lead to dysfunction of the myosin protein. Mutations located in the ATP-binding pocket (e.g. Thr124Leu, Phe244Leu) were hypothesised to alter either the water-structure of the active site, or the position of the critical phosphate-binding residues, and so compromise the catalytic function of the S1 peptide [35]. Mutations located at the actin interface of myosin (e.g. Arg403Gln) may influence actin–myosin interaction, either by direct association with actin or by modifying the closure of a cleft, which connects the actin- and the ATP-binding sites [35]. Mutations located close to the helix which connect the two reactive thiols (e.g. Phe513Cys, Gly584Arg) are likely to lead to altered conformation of this region and so modify the nature and timing of conformational changes during the contractile cycle [35]. A number of mutations are found on, or are closely associated with, structural elements in the lever-arm helix that form the interface with ELC (e.g. Arg719Trp, Gly741Arg). As some ELC mutations have been postulated to lie in a hinge region between ELC and βMHC in the intact sarcomere, and as a βMHC mutation hotspot lies spatially adjacent to these ELC mutations [8], it seems that the interface between MHC and ELC is important for mechano–chemical coupling [35]. Mutations such as Leu908Val, that occur in the incipient part of the MHC rod, which links S1 to the backbone of the thick filament, have been less easy to interpret as the role of this domain was not known; it had been proposed that these mutations may lead to defective transmission of force from the myosin heads to the thick filament by affecting either filament stability or rigidity of the rod [35]. Recent data from Gruen and Gautel [37] reveal that the N-terminal residues of the S2 segment of the MHC rod in fact bind to the regulatory domain of cardiac myosin binding protein-C (MyBPC). Moreover, HCM missense mutations in this region dramatically decrease the affinity of that association. This finding suggests both the way in which the regulatory function of MyBPC is mediated and the reason why the proximal S2 region, but not the remainder of the rod, is a site for HCM mutations. Of note, missense mutations are not found in the most critical residues such as the consensus ATP-binding sequence; it is presumed that these would produce too severe a defect to be viable.

2.1.2 Functional studies of mutant MHC peptides

Much work has been reported on the effects of the HCM mutations on the in vitro function of myosin. In addition to ATPase and actin-binding assays, many groups have used an in vitro motility assay in which fluorescently-labelled actin filaments are translocated by myosin bound to a coverslip [38]; in a range of experimental conditions the filament sliding-velocity in this assay is a good analogue of unloaded shortening velocity in muscle fibres [39]. The myosin used for these in vitro analyses has come from a variety of sources: slow skeletal muscle also expresses cardiac βMHC and hence soleus muscle biopsies from patients have been used as a source of mutant βMHC; baculovirus-based systems have been developed for the co-expression of truncated MHC and both MLCs to yield a soluble recombinant myosin similar to heavy meromyosin (a soluble proteolytic fragment of myosin S1 lacking the rod); Dictyostelium discoideum myosin with the equivalent of human HCM mutations has been expressed in a MHC null background; and mutant myosin has been extracted from the hearts of transgenic mice.

Most attention has been paid to the Arg403Gln mutation, which lies close to the actin-binding interface [35] (see Fig. 2); studies of this mutation illustrate the type of information gleaned about the pathogenesis of HCM by biochemical and biophysical analyses. This is a common, severe mutation (100% penetrance in adults and high frequency of sudden death) and, fortuitously, it is possible to distinguish the mutant myosin from the wild type since the mutation causes disruption of an Arg-C endoproteinase cleavage site. Using this distinction, the Arg403Gln mutant myosin has been shown to be stably expressed in slow skeletal muscle of HCM-affected individuals [40]. The influence of the mutation on sarcomere structure and myofibrillar assembly has been examined in vitro. Although in COS cells [41] and in feline myocytes [42] there was some evidence of decreased myofibrillar assembly and sarcomere disruption caused by expression of the human Arg403Gln mutant myosin, more recent work has shown that expression of two mutant human βMHCs (Arg249Gln and Arg403Gln) in rat ventricular cardiomyocytes resulted in normal thick filament assembly, producing myofibrils with well defined I bands, A bands and H zones [43].

In in vitro motility assays, the Arg403Gln mutation has been shown to result in reduced filament sliding velocity in experiments using βMHC purified from patients’ skeletal muscle [40], recombinant rat αMHC [44], recombinant human βMHC [45] and in D. discoideum myosin [46]. This mutation has also been shown to result in a diminished rate of actin-activated myosin–ATPase activity, accompanied by an increased Km for actin, while the intrinsic myosin–ATPase activity (K+EDTA or Ca2+) remained unaltered [44–46]. (The actin- and ATP-binding capacities of two recombinant βMHC fragments, one corresponding to the 50 kDa ‘domain’, the other the 25 and 50 kDa ‘domains’, were found to be unaffected by the introduction of the Arg403Gln mutation; however, since neither fragment possesses the full binding capacity of intact S1, these results are of limited value [47]).

The contractile properties of skinned fibres with the βMHC Arg403Gln mutation have been measured and found to have a lowered force/stiffness ratio (56% of normal), a decreased velocity of shortening (50% of normal) and a reduced maximum force (71% of normal) [48]. Similarly, the force generated by D. discoideum Arg403Gln-equivalent myosin on a single actin filament measured using a laser-trap apparatus was found to be 57% of that generated using wild type myosin [46].

Other important data have emerged recently from using a laser-trap to measure the step size (unitary displacement, Duni) of αMHC from normal mice and mice homozygous for the Arg403Gln mutation. Using mean variance analysis of the data it was found that the wild type myosin Duni was distributed into two populations, one of mean c.5 nm and the other c.10 nm. It is suggested that the 5 nm step size is due to the action of a single myosin head, while the 10 nm step size is generated by the action of two myosin heads acting co-operatively. Interestingly, for mutant myosin the proportion of Duni=c.10 nm was considerably diminished, thus suggesting that the Arg403Gln mutation disrupts head-to-head co-operativity [49]. If one assumes that the attachment time following the powerstroke remains the same, such a reduced Duni (stepsize) would explain the observed drop in filament velocity in the motility assay.

To date, it can be concluded that the net effect of the βMHC Arg403Gln mutation on myosin function is to decrease the force exerted on the actin filament, through an altered interaction with actin and a slowing of the crossbridge cycle. As patients are heterozygous for HCM mutations, their thick filaments will contain both normal and mutant myosin heads; these heads are effectively connected in series, and thus the mutant myosin heads will exert a drag on the whole filament – and thus a dominant negative effect. The important result of these analyses of the Arg403Gln mutation is the finding that this mutation leads to a primary hypocontractile state, which, in turn, provides a stimulus for the development of compensatory hypertrophy.

Many in vitro analyses have been performed on myosins with other HCM-associated mutations and these have largely yielded qualitatively similar results to those obtained with the Arg403Gln mutant; these are summarised in Fig. 3. In general, the HCM βMHC mutations result in myosin which has diminished actin-activated ATPase activity and which produces less force and a lower velocity of actin filament translocation, either by reducing the step size, by reducing the crossbridge cycling rate or by providing additional internal load. Since the HCM mutations lie in different functional domains of the protein it is likely that there are different molecular mechanisms of functional impairment. The most straightforward hypothesis for the pathogenesis of the disease is that the reduced force provides a stimulus for compensatory hypertrophy. Analysis of the hearts of αMHC403/+ mice which showed diastolic dysfunction has also suggested that the mutant myosin may affect cardiac energetics by generating less force per ATP and which contribute to the disease state.

2.1.3 Mutations in myosin light chains

Poetter assessed the functional effect of the Met149Val mutation in ELC and of Glu22Lys in RLC (Fig. 4), utilising myosin obtained from cardiac biopsies of HCM patients in an in vitro motility assay [8]. Whereas myosin with the RLC Glu22Lys substitution was indistinguishable from the wild type, myosin with the ELC Met149Val mutation translocated actin faster than the normal control myosin. This is clearly in contrast with the impact of HCM mutations in MHC, with the possible exception of the spatially adjacent βMHC Arg719QGln mutation for which data are conflicting [8,46]. If ELC Met149Val also results in decreased force production with concomitantly weak actin-interaction, analogous to that observed for the βMHC Arg719Gln mutation [46], a faster sliding velocity may not be unexpected.

Fig. 4

A diagram showing the location of the HCM-associated mutations in myosin essential and regulatory light chains [8,21] and a table showing the reported changes in in vitro properties of the HCM-mutant proteins.

While Glu22 and Pro94 on RLC are located at an interface between the carboxy- and amino-terminal domains of RLC, the mutations Glu22Lys, Arg13Thr and Pro94Arg are, in addition, all located close to a very highly conserved serine residue [8]. When the equivalent serine was mutated in Drosophila, it produced normal isometric tension, but led to a loss of the stretch-activation response in indirect flight muscle, and consequently a loss of flight [8,50]. It is thought that cardiac papillary muscle resembles insect indirect flight muscle in terms of longitudinal arrangement, generation of oscillatory motion, and a pronounced stretch-activation response [51]. Furthermore, if, as postulated, the elastic element in series with the force generator of muscle is located in the 20 kD fragment of myosin S1, these mutations in RLC may affect the stretch-activation response which is thought to be fundamental to an increase in oscillatory power output [8,52]. It was thus of some interest to find these mutations in families where some individuals manifested striking papillary muscle hypertrophy [8]; it is now known, however, that this is not a consistent feature of these mutations [21].

Levine et al. [53] measured the contractile properties of slow twitch deltoid muscle fibres from a patient bearing a RLC Glu22Lys mutation in comparison with those from a normal relative. Although no significant difference in mean force/cross-sectional area was found, the Ca2+-sensitivity of force production was shifted to the left, suggesting that the mutation results in increased force at submaximal concentrations of Ca2+. This result and the increased motility obtained with ELC Met149Val suggest that the MLC mutations may cause HCM by a different mechanism to that of the bulk of the βMHC mutations.

3.1 Myosin binding protein-C (MyBPC)

3.1.1 Peptide structure and function

MyBPC is probably the HCM-associated protein in which the functional domains are least well characterised. The protein consists mainly of modules with high homology to immunoglobulin C2 (Ig-like) and fibronectin type 3 domains (see Fig. 5). It has been shown that the binding of MyBPC to titin depends on the presence of the three carboxy-terminal modules (modules C8–C10) [54]. Furthermore, the Ig-like module C10 had previously been demonstrated to be chiefly responsible for the binding of MyBPC to myosin filaments [55].

Fig. 5

A diagram showing the domain structure of myosin binding protein-C and the location of selected HCM-associated missense and truncation mutations [9,10,19,64–66]. Int=intron, DS=donor site, AS=acceptor site.

The cardiac isoform of MyBPC has a unique insertion of 28 residues, abundant in proline and charged residues, into the Ig-like module C5. Such proline/charged residue-rich stretches often form the basis of specific ligand interactions of signal transduction molecules, such as the ligands of SH3 domains [56]. Because this is a unique feature of the cardiac isoform, it may act as a scaffold for binding of the cardiac-specific MyBPC-associated kinase which co-purifies with MyBPC [57–59]. It is known that MyBPC becomes phosphorylated during adrenergic stimulation of the myocardium, with a concomitant increase in systolic tension [58,60,61], and becomes dephosphorylated by cholinergic agonists, again paralleled by a decrease in systolic tension [62]. There are a total of four putative phosphorylation sites located in the region between modules C1 and C2, including the characteristic MyBPC motif and a feature, called the LAGGGRRIS loop, specific to cardiac MyBPC. Three of these sites can be phosphorylated by ATP-dependent protein kinase (PKA) and MyBPC-associated kinase in vitro [59]. Phosphorylation of one of these sites, situated in the LAGGGRRIS loop, is a prerequisite for phosphorylation at the other two sites, presumably by inducing conformational changes which make the other sites accessible [59]. Because phosphate turnover on cardiac MyBPC is dynamic and related to changes in contractility in the intact heart, one might postulate that the MyBPC motif is involved in mechanisms where phosphorylation of MyBPC in vivo alters muscle function, such as contraction/relaxation rates.

Moreover, in in vitro studies, high concentrations of MyBPC are capable of binding F-actin, and can compete for actin binding with myosin S1 fragments [63], although the domains involved in these interactions are unknown. In active muscle, MyBPC could modify the activity of myosin crossbridges located near MyBPC binding sites by reducing the frequency of actin–myosin contacts or by altering their kinetics [63].

Therefore, it would seem that MyBPC does not have a solely structural role, but that it is also able to participate in regulatory signals, potentially in several pathways, linked through protein kinase cascades [59].

3.1.2 Functional inferences from HCM mutations in MyBPC

The majority of mutations reported in MyBPC, to date, have been splice site mutations, or deletion and insertion mutations which cause a shift in the reading frame, that would result in variable degrees of truncation of the protein and loss of the important C8–C10 binding domains [64]. However, HCM-associated missense mutations have increasingly been reported in other parts of the gene [19,65,66]; a selection of the truncation and missense mutations are shown in Fig. 5. As some of these mutations do not affect splicing, the main myosin- and titin-binding domains of MyBPC are presumably retained, and hence they presumably adversely affect other functional domains of MyBPC, as yet unidentified. As with MHC, the total number of HCM-causing mutations in MyBPC is large; approximately 25 have been reported to date and most newly ascertained families continue to have novel, rather than known, mutations.

Truncation mutations affecting the two defined functional domains, namely those involved in binding titin and myosin, have been evaluated, with contradictory results. Gilbert and co-workers [67] performed transfection studies in skeletal muscle myoblasts with constructs encoding truncation mutants of MyBPC (ΔC9–10). Certain truncated peptides incorporated properly into the A-band, and the presence of the truncated protein, migrating with the expected mobility, could be demonstrated on Western blots with antibodies directed against the amino-terminus of MyBPC. Of note, one truncated mutant produced a clear dominant negative disruption of myofibril assembly. However, when the same technique was used by different investigators to study the occurrence of truncated proteins in myocardial biopsies obtained from two patients with different MyBPC splice site mutations [68,69], no truncated proteins could be detected on Western blots, despite the presence of the corresponding altered mRNAs. These results suggest enhanced proteolysis of the truncated proteins, which may alter the stoichiometry of sarcomeric proteins, which may, in turn, affect the assembly of the sarcomere [68]. In addition, the authors speculate about an additional potential mechanism whereby the mutant mRNA inhibits wild type mRNA processing and thus acts as ‘dominant negative mRNA’. In contrast, Yang et al. [70] were able to show the presence of the stable truncated protein by both Western blot and SDS–PAGE, using protein extracts from a transgenic mouse model which expresses mutant MyBPC. They were able to demonstrate that this truncated protein, lacking the myosin- and titin-binding domains, did not incorporate correctly into the sarcomere and was only weakly associated with the rest of the sarcomeric proteins [70]. Together these results have been interpreted as suggesting that, unlike the bulk of HCM-causing mutations, MyBPC mutants lacking the myosin and titin-binding domains may, in fact, cause HCM by haploinsufficiency, rather than by dominant-negative action. Further work is needed to test this hypothesis as low levels of truncated protein that may be sufficient to cause a dominant negative phenotype may escape detection. There are currently few published data relating to the functional impact of these mutations on contractility. The transgenic model did, however, reveal an effect of the truncated MyBPC allele; skinned fibres from left ventricular papillary muscle from MyBPC mutant transgenic mice showed increased calcium sensitivity of force development and decreased maximum power compared with control fibres [70].

3.2 Cardiac troponin T (cTnT)

3.2.1 Peptide structure and sites of mutations

The TNNT2 gene gives rise to a number of different cTnT isoforms by alternative splicing. The composition changes from foetal to adult and has been found to be altered in heart failure. The principal isoform in normal adult heart is composed of 288 amino acids and the residue numbers of human cTnT given below are with respect to this isoform. TnT consists of an extended amino-terminal portion (T1; residues 1–187 in the human cardiac protein) which lies alongside Tm on the thin filament and a globular carboxy-terminal domain (T2; residues 188–288) which binds to Tm near Cys190, as well as to the amino-terminal portion of TnI and an unidentified domain on TnC [71,72]. It would seem that TnT is involved in distributing the inhibitory effect of the Tn complex, via Tm, to the seven actin monomers with which it interacts, in the absence of Ca2+, and removing this inhibitory effect from all seven actin monomers, as well as activating the actomyosin ATPase, in the presence of Ca2+ [73]. To date, at least ten mutations that cause HCM have been reported in cTnT, situated in both the T1 and T2 portions of molecule (see Fig. 6). In T1 these include missense mutations Ile79Asn, Phe110Ile and Glu163Lys and a three base-pair deletion which causes the in-frame loss of a glutamic acid at codon 160 [11,12]; Arg92 is apparently a hotspot, with mutations to Gln, Leu and Trp reported [11,18,74]. The missense mutations Glu244Asp and Arg278Cys lie in T2. A splice site mutation, Int15 G1→A, inactivates a 5 ́ splice donor site, leading to either skipping of exon 15, or activation of a cryptic splice site, which results either in a truncated protein with the C-terminal 28 amino acids replaced by seven new residues, or in a truncated protein lacking the C-terminal 14 amino acids [11,12].

Fig. 6

A diagram showing the location of selected HCM-associated mutations in cardiac troponin T [11,12,18,74] and the location of putative functional domains, and a table summarising the reported changes in in vitro properties of the HCM-mutant proteins compared with those of the wild type. * The Int15 Gl→A mutation gives rise to two transcripts, one encoding a truncated protein with the C-terminal 28 amino acids replaced by seven new residues, the other a truncated protein lacking the C-terminal 14 amino acids (for details see text).

3.2.2 Functional studies of mutant TnT peptides

The in vitro properties of recombinant embryonic rat cardiac TnT with a mutation equivalent to the human Ile79Asn have been assessed [75]. The mutated TnT was indistinguishable from the wild type protein in assays measuring the affinity of the Tn complex for Tm, Tn-induced binding of Tm to actin, the binding of myosin S1 to the thin filament, and actin-activated myosin ATPase activity. However, in an in vitro motility assay, thin filaments containing the mutant TnT were translocated over myosin heads with 50% increased velocity compared to filaments containing wild type TnT. This increased sliding velocity could imply an increased distance per crossbridge cycle, or an increased rate of release of myosin S1 from the thin filament. The fact that alteration of the amino acid sequence by alternative splicing in the nearby hypervariable region of TnT can result in subtle changes in actin–myosin interaction and Ca2+-sensitivity would seem to support the latter suggestion. It was proposed that increased actomyosin cycling would directly cause hypertrophy via hypercontractility, rather than creating it through a compensatory mechanism [75]. In the light of this suggestion, although the Ile79Asn mutation has not been investigated in extensive families, it is noteworthy that TnT-associated HCM is seldom correlated with remarkable hypertrophy.

The Ile79Asn mutant has been further analysed in a motility assay using a glass microneedle attached to a single thin filament to measure the force exerted by the myosin heads [76]. Using this technique it was established that the maximum force at pCa5 exerted on thin filaments containing Ile79Asn TnT was 9.38±0.35 pN/μm, compared with 12.36±0.4 pN/μm exerted on thin filaments containing wild type TnT. This suggests that the presence of the Ile79Asn mutation in vivo may in fact lead to reduced contractility, which, in turn, may provide the hypertrophic stimulus.

The effects of the HCM-associated TnT truncations on in vitro TnT function have been analysed in both recombinant bovine cardiac TnT [77] and recombinant human cardiac TnT [78] backgrounds. In both cases it was shown that both the truncated TnTs had a profound effect on actin-activated myosin–ATPase activity. Mukherjea et al. [78] found that the Tn complex containing wild type TnT activated unregulated actomyosin ATPase activity by 51±7% at pCa5 whereas the two truncated TnTs that result from this mutation gave activation of only 13±3 and 20±5%. The truncated bovine cardiac TnT mutants were found not to activate at all [77]. Clearly these data suggest that incorporation of these mutations into the thin filament in vivo would give rise to a hypocontractile state.

The effects of certain HCM-associated TnT mutations on contractility in cells and in skinned muscle fibre preparations have also been studied. Watkins et al. [25] used a novel quail myotube system in which quail myoblasts were transfected with plasmids encoding either wild type or mutant human cardiac TnT. After selection and differentiation, the myotubes were permeabilised and mounted for measurement of mechanical properties. Using this system it was found that the endogenous TnT was completely replaced in the differentiated myotube by either wild type human cardiac TnT, or by truncated mutant TnT [25], or by Ile79Asn, Arg92Gln or ΔGlu160 mutant TnT [79]. This demonstrates the ability of all of these mutant TnTs to incorporate correctly into the sarcomere in vivo; even though the truncation mutation had been predicted to create a null allele, in analogy to a similar mutation in Drosophila. Moreover, the maximum isometric force generated by these myotubes in the presence of Ca2+ was found to be diminished in the case of three of the mutants studied. The force generated by the truncation mutant myotubes was 80% reduced compared to wild type, that generated by Ile79Asn myotubes 25% and that by ΔE160 myotubes 40%, whereas Arg92Gln myotubes gave the same maximum force as wild type. The stiffness:force ratio was unchanged in each case and thus the decreased force is due to a reduction in the number of attached crossbridges rather than a decrease in force per crossbridge. Additionally, the missense mutations Ile79Asn and Arg92Gln caused a nearly twofold increase in unloaded shortening velocity. The decrease in isometric force generation and the increase in unloaded shortening velocity caused by the Ile79Asn mutation in the myotube system parallel the reduction in maximum force and increased rate of motility measured in the motility assays [75,76]. Furthermore, all the mutants analysed in the quail myotube system caused a decrease in the Ca2+-sensitivity of force production.

Morimoto et al. [80] have analysed the effects of the Ile79Asn and Arg92Gln mutations on the mechanical properties of skinned rabbit trabeculae preparations. Purified recombinant wild type or mutant human cardiac TnT was used to displace endogenous Tn; exogenous TnC and TnI were subsequently added to restore Ca2+-dependent contractility [81,82]. In contrast to the myotube system results of Sweeney et al. [79], it was found that both the Ile79Asn and the Arg92Gln mutants had Ca2+-sensitising effects compared with the wild type protein. It is not clear whether the discrepancy in the results of these studies relates to species and/or muscle type differences, or whether it, in fact, reflects a lowered efficiency of exchange when using mutant protein (and hence less of the rightward shift associated with the exchange process).

Wild type and the Arg92Gln mutant human cardiac TnT have also been expressed in adult feline cardiac myocytes using adenovirus vectors [83]. Although no significant difference was observed between the cells transfected with wild type constructs and those transfected with mutant constructs within the first 24 h, fractional shortening and the peak velocity of shortening of cells transfected with the mutant construct was significantly reduced, compared to wild type, after 48 h. These values were even further reduced by 72 h after transfection. These results indicated that the expression of the Arg92Gln cardiac TnT mutant impaired intact adult cardiac myocyte contractility.

The functional data are summarised in Fig. 6; in skinned myotubes or fibres incorporation of the HCM TnT mutants gives rise to an increased unloaded shortening velocity (mirrored by the increased velocity of actin filament translocation in the in vitro motility assay) and diminished maximum force. The decrease in maximum force is brought about by a reduction in the number of attached crossbridges whereas the increased shortening velocity is caused by an accelerated crossbridge detachment rate. This is, therefore, akin to an uncoupling of the crossbridge cycle. The decrease in force observed for the missense mutations is less severe than that observed for myosin mutations with a similarly poor prognosis (and hence there would be less stimulus for compensatory hypertrophy – in keeping with the clinical phenotype). It has been suggested that the pathogenesis of HCM due to TnT mutations may instead be largely attributable to altered cardiac energetics resulting from an increase in the cost of force production [79].

3.3 α-Tropomyosin (αTm)

3.3.1 Peptide structure and sites of mutations

Tropomyosin (284 amino acids) is an elongated α-helical molecule, which forms a parallel coiled-coil dimer that lies along the length of the groove of the thin filament, binding to seven actin monomers and to TnT. The interaction of Tm–Tn with actin is proposed to act as a Ca2+-sensitive switch, allowing, or sterically blocking, the attachment of myosin S1 fragments to actin grooves [84,85]. The molecular flexibility needed for such movement is made possible by the inherent slight irregularities in αTm structure, especially in the C-terminal part, where there may be a region akin to a hinge (possibly involving the conserved cysteine residue at codon 190). The T2 portion of TnT, to which TnI and TnC are also bound, binds αTm in the vicinity of Cys190, while the nearby amino acids Val170 and Ile171 form a hydrophobic patch predicted to be part of the TnT-binding site [85]; Cys190 can be labelled with fluorophores which are sensitive to Tn-binding [86]. The elongated T1 portion of TnT binds towards the C-terminus of αTm, independently of Ca2+, and extends onto the N-terminus of the adjacent αTm [72]. Two of the four HCM mutations, Asp175Asn and Glu180Gly, introduce changes in surface charge in the region of αTm that has been implicated in Ca2+-sensitive TnT-binding; the other two, Ala63Val and Lys70Thr, lie outside the TnT-binding sites and have been suggested to alter actin binding (see Fig. 7).

Fig. 7

A diagram showing the location of the HCM-associated mutations in α-tropomyosin [11,107,108] and the location of putative functional domains, and a table summarising the reported changes in in vitro properties of the HCM-mutant proteins compared with those of the wild type.

3.3.2 Functional studies of mutant αTm peptides

Studies of the effect of the HCM mutations on in vitro Tm function have been carried out using wild type and mutant α-Tms overexpressed in E. coli, with an Ala-Ser N-terminal tag (AS-Tm) to mimic the N-terminal acetylation [87] that is essential for end-to-end interactions of adjacent Tms [67]. In an in vitro motility assay, it was found that thin filaments reconstituted with Asp175Asn and Glu180Gly mutant AS-Tms gave a greater increase in velocity upon the addition of Tn (18–21%) than did filaments containing wild type AS-Tm (5%), at pCa5 [88]. When NEM-S1 (which switches αTm to the fully on state) was added, there was no further increase in velocity of actin filaments containing either wild type or mutant AS-Tm [89]. This suggests that these thin filaments were already fully switched on and that the αTm mutations lead to an altered on state. In all other parameters measured in the motility assay, including the sliding velocity of thin filaments lacking Tn, the αTm mutations had no effect; hence, the observed difference between these mutant and wild type αTms is likely to be caused by altered interaction with TnT.

Similarly, Golitsina et al. [90] used myosin S1 to switch pyrene-labelled recombinant wild type or Asp175Asn and Glu180Gly mutant α-Tms, lacking the N-terminal leader, to the fully on state. The change in excimer fluorescence of the wild type was found to be different than that observed with either mutant αTm, thus also suggesting that the HCM mutations affect the structure of αTm in the on state.

The expression of α-Tm in fast twitch muscle has been exploited to examine the effects of the Asp175Asn mutation on the contractile properties of vastus lateralis muscle from two patients [26]. Asp175Asn mutant αTm migrates slightly faster upon SDS–PAGE and hence it was also possible to quantitate the relative levels of wild type and mutant proteins; Asp175Asn αTm was found to be stably expressed at the same level as wild type protein. The Ca2+-sensitivity of single skinned fibres containing Asp175Asn αTm was significantly increased compared with that seen in fibres from unaffected controls (ΔpCa50=0.09). However, no significant differences in maximum force, maximum shortening velocity or co-operativity were found. Similar results were obtained using cardiac myofilaments from transgenic mice expressing Asp175Asn αTm; both myofibrillar ATPase activity and fibre force generation showed increased Ca2+-sensitivity (ΔpCa50=0.11 and 0.08, respectively), with no difference in maximum force [29].

Thus, in contrast to the βMHC mutants, the α-Tm HCM-associated mutants do not appear to cause a depression of maximum force; instead, these proteins cause an increase in the Ca2+-sensitivity of force production and hence give an increase in force at submaximal Ca2+ concentrations. It has been postulated that mutations in this gene may cause hypertrophy by a more direct ‘hypercontractile’ mechanism. In addition, increased Ca2+-sensitivity might be expected to produce deficits in diastolic relaxation.

3.4 Troponin I

3.4.1 Peptide structure and position of mutations

Troponin I is the inhibitory component of the Tn complex and the human cardiac isoform (cTnI) is a protein of 210 amino acids encoded by the TNNI3 gene. Studies using peptide fragments of TnI have identified a highly basic inhibitory region (corresponding to residues 129 to 149 of human cardiac TnI) which is capable of inhibition of actomyosin ATPase [91–94]. At low Ca2+ concentrations this region binds to actin and inhibits myosin S1-binding; under activating conditions it binds to the carboxy-terminal domain of TnC–Ca2+, thus releasing the inhibition of actomyosin ATPase [93]. Further work has identified regions C-terminal to the inhibitory sequence which are also important in regulation; residues 140–148 of rabbit skeletal TnI (173–181 in the human cardiac protein) bind actin–Tm and enhance the inhibitory effect, while the amino acids immediately C-terminal to the inhibitory region (149–164 in human cardiac TnI) are important for the interaction with TnC [95]. In addition, deletion of 23 residues from the C-terminus of TnI has been shown to cause a marked reduction in inhibition, thus implying that an additional site for actin–Tm-binding exists in the C-terminus [96].

TnI forms an anti-parallel dimer with TnC [97,98] with multiple sites of interaction; the inhibitory region plus the C-terminus of TnI binds to the N-terminal half of TnC in a Ca2+-dependent manner, while the N-terminal domain of TnI binds the C-terminal half of TnC independent of Ca2+ [97]. Two of the reported HCM-associated TnI mutations are at Arg145 (to either Gly or Gln) [13] which falls within the inhibitory region. The other mutations are Arg162Trp (which lies within the ‘second TnC-binding site’ [95]), the deletion of Lys183, Gly203Ser and Lys206Gln [13] (see Fig. 8).

Fig. 8

A diagram showing the location of the HCM-associated mutations in cardiac troponin I [13] and the sites of putative functional domains.

3.4.2 Functional studies of mutations

The cTnI HCM mutations are the ones most recently reported and there is little in the literature at time of writing concerning their effect on TnI function. Van Eyk et al. [93] found that mutation of any of the basic amino acids to glycine in the inhibitory peptide, including the equivalent of the Arg145 residue, had a severe effect on the ability of the peptide to inhibit actomyosin ATPase. Work using recombinant full length TnI has found that the Arg145 mutations do significantly reduce the potency of TnI inhibition of actomyosin ATPase; the Arg162Trp and ΔLys183 were found to cause a small reduction in the potency of inhibition whereas the Gly203Ser and Lys206Gln mutations were found not to affect inhibitory function at all [99].

4 Discussion

4.1 The identification of a specific mutation as a predictor of severity of disease

Collectively, the results of in vitro functional assays suggest that the dysfunction caused by distinct point mutations is responsible for the pathogenesis of HCM and supports the notion that knowledge of different HCM mutations could be used to stratify HCM into different risk classes. Quantitative variation in the degree of functional impairments caused by the different HCM-associated mutations may well determine the severity and the penetrance of the disease, a theory born out by the correlation between the degree of dysfunction caused by various mutations and the prognosis of HCM in patients with these mutations [45]. Notably, the extent of the decrease in myosin motor activity for MHC Arg403Gln (70–80%), Arg453Cys (70–80%), Arg249Gln (40–50%) and Val606Met (10–15%) correlates well with the cumulative survival rates at 50 years for these mutations (36, 34, 79 and 94%, respectively) [7,45]. Also, the decrease in actin-activated myosin–ATPase rates of rat myosins with the Arg249Gln, Arg403Gln and Val606Met mutations was found to correlate with the severity of clinical phenotype (moderate, severe and mild, respectively) [100]. In studies in skinned fibres, Gly741Arg mutant myosin gave a decreased velocity of shortening (39% of normal) as well as reduced isometric force generation (42% of normal); in contrast, no abnormal contractile properties were found for the Gly256Glu mutation, which is paralleled by the benign prognosis of this mutation in HCM families [48]. It has even been suggested that such in vitro functional assays be used to predict the prognostic implication of mutations for which no genotype–phenotype correlations could be established due to a lack of subjects [45]. However, studies of 100% mutant protein may not be relevant to the heterozygous disease state, and the relative expression and efficiency of incorporation of mutant protein in the heart will vary between different mutations such that studies of 50:50 mixtures equally may not be appropriate (transgenic mouse studies suggest that profoundly dominant negative peptides may be incorporated at quite low levels). Moreover, disease pathogenesis may be influenced by other genetic or environmental factors, thus modulating the pathogenesis of the specific mutation and leading to variability in penetrance and severity of disease. Therefore, although in vitro studies are decidedly useful in understanding the functional defect caused by mutations at a protein level, care should be exercised in extrapolating their results to a physiological counterpart of myofibrillar disarray and selective thick filament lysis seen in HCM-patients [101].

4.2 Summary of the models of disease pathogenesis

Biochemical and physiological studies into the ways in which HCM-causing mutations alter the in vitro properties of contractile proteins have provided a number of clues to the mechanism by which these mutations may give rise to a disease state. The bulk of the work has concerned the mutations to βMHC, TnT and α-Tm as described above. The βMHC studies have shown that the HCM mutations in general result in myosin which generates less force and these data have led to the ‘hypocontractile’ hypothesis by which the decreased force provides the stimulus for compensatory hypertrophy. In contrast to this, the α-Tm HCM mutants do not appear to cause a depression of maximum force; instead, these proteins cause an increase in the Ca2+-sensitivity of force production and hence give an increase in force at submaximal Ca2+ concentrations. Mutations in this gene, and possibly those in RLC or ELC, may cause hypertrophy by a more direct ‘hypercontractile’ mechanism. In addition, increased Ca2+-sensitivity might produce abnormalities of relaxation. Although the TnT mutants give an increased velocity in the in vitro motility assay and an elevated unloaded shortening velocity in skinned myotubes, they have also been shown to result in reduced maximum force and hence they may act via a hypocontractile route. In addition to mechanisms in which altered contractility provides the primary hypertrophic stimulus, a further hypothesis has been proposed in which altered cardiac energetics gives rise to hypertrophy. Data from both αMHC403/+ mouse hearts and myotubes transfected with HCM TnT mutant constructs have suggested that the HCM mutations increase the cost of force production (less force per ATP) and this increased rate of energy consumption could result in energy demands that cannot be met [79,102]. Clearly, both the increased energy demands and altered contractility could contribute to the disease progression; potentially, the mechanical deficits underlie the compensatory hypertrophy, while the metabolic deficits underlie the propensity to ischaemia and arrhythmia, and hence sudden death.

5 Conclusion

Work to date on the in vitro properties of HCM-associated contractile protein mutants has identified discrete parameters by which the mutant proteins differ from their wild type counterparts and this has led to the construction of models to describe the development of the disease state. Evidence has also emerged to show that the degree of functional impairment caused by the particular HCM-associated mutation may well determine the severity and the penetrance of the disease. Further biochemical and physiological work, particularly looking at the more recently reported MLC and Tn I mutations, along with the studies on transgenic and gene targeted animals is now needed to refine the current tentative models of pathogenesis.

Acknowledgements

The authors wish to acknowledge the British Heart Foundation and the Wellcome Trust for financial support. Dr. Moolman-Smook is the recipient of a Wellcome Trust Travelling Fellowship.

Footnotes

  • 1 These authors contributed equally to this work.

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
  88. [88]
  89. [89]
  90. [90]
  91. [91]
  92. [92]
  93. [93]
  94. [94]
  95. [95]
  96. [96]
  97. [97]
  98. [98]
  99. [99]
  100. [100]
  101. [101]
  102. [102]
  103. [103]
  104. [104]
  105. [105]
  106. [106]
  107. [107]
  108. [108]
View Abstract